1 15 14 https://www.nal.usda.gov/exhibits/speccoll/files/original/83b69d30063b533807899b556076f45a.pdf 9c2fde559fa4c367f1167180d4bafbe0 PDF Text Text Item ID Number °1581 Wolfe, William H. Corporate Author Report/Article Title Air Force Health stud V (Ranch Hand II) Update Briefing Journal/Book Title Year 1986 Month/Day September 29 Color Number of Images n 27 Desoripton Notes Wednesday, May 23,2001 Page 1582 of 1608 �AfRjl=QRCE HEALTH STUDY HAND II) UPDATE BRIEFING 29 SEPTEMBER 1986 BRIEFER;: COLONEL WILLIAM H. WOLFE I �STUDY DEVELOPMENT 19-78-1979 PROPOSAL DEVELOPED 1979 PEER REVIEW AND REFINEMENT 19*79-1982 STUDY POPULATIONS IDENTIFIED QUESTIONNAIRES DEVELOPED 1982 PHYSICAL EXAMINATIONS ACCOMPLISHED 1983-PRESENT MORTALITY REPORTS RELEASED 1984 BASELINE MORBIDITY REPORT RELEASED 1985 FIRST FOLLOWUP PHASE BEGUN �STUDY DESIGN IDENTIFY EXPOSED POPULATION IDENTIFY AND SELECT COMPARISON POPULATION DETERMINE BASELINE HEALTH STATUS MORTALITY MORBIDITY COMPARE FINDINGS STATISTICALLY TO DELINEATE POSSIBLE HERBICIDE EFFECTS ACCOMPLISH FOLLOWUP STUDIES OF POPULATIONS �MATCHING PROCESS 8 COMPARISONS SELECTED FOR EFlCH EXPOSED INDIVIDUAL MATCHED FOR SEX. AGE. DUTY CODE. RACE RANDOM COMPARISONS FROM EACH MATCHED SET USED 5:1 IN FIRST FOUR MORTALITY REPORTS ALL COMPARISONS AFTER 1:1 1386 IN BASELINE AND SUBSEQUENT MORBIDITY OTHERS AVAILABLE FOR REPLACEMENT REPORTS �SUMMARY OF MORTALITY RESULTS CA/O 31 DEC 84) EVALUATION OF 55 EXPOSED AND 285 COMPARISON DEATHS RELATIVELY SMALL NUMBERS EMPHASIZE PRELIMINARY NATURE OF THESE RESULTS MORTALITY EXPERIENCE NEARLY IDENTICAL IN BOTH GROUPS CAUSE-SPECIFIC ANALYSES NOT STATISTICALLY DIFFERENT �ADDITIONAL MORTALITY CONTRASTS DOD RETIRED POPULATION ALL RANK-EXPOSURE GROUPS DOING BETTER EXPOSED-ENLISTED GROUP NOT STATISTICALLY BETTER ACTIVE C I V I L SERVICE PERSONNEL ALL SUBGROUPS EQUIVALENT TO C I V I L SERVICE 1978 U S WHITE MALES ALL SUBGROUPS DOING SIGNIFICANTLY BETTER �I -"-^ SUMMARY COUNTS OF DEATHS RANCH HAND AT RISK DEAD PERCENT OFFICERS 46 6 16 3 4 . ENLISTED 791 3 9 FLYING 66 4 GROUND COMPAR I SON AT RISK DEAD PERCENT 2 7 2 8 9 8 4 3 . 4 9 . 3 9 S 3 187 4 8 . 24 3 7 . 3163 161 5. 1 611 31 5. 1 3OO8 124 4. 1 1 5 2 7 55 4 4 . 6171 2 5 8 4 6 . RANK OCCUPATION TOTAL �DEATHS BY CRUSE r EXPOSED • COMPARISONS 19 9 6 SUICIDE/HOMICIDE 5 23 MALIGNANCY 6 51 CIRCULATORY 18 80 RESPIRATORY O DIGESTIVE 5 13 OTHER 2 15 TOTAL 5 5 2 5 8 ACCIDENTAL 7 1 �BASELINE MORBIDITY RESULTS r NO DEFINITIVE CLINICAL ENDPOINTS CONCLUSIVELY ATTRIBUTABLE TO HERBICIDE EXPOSURE NO STS, PCT, OR CHLORACNE IN THE EXPOSED GROUP SOME CLINICAL AND SUBCLINICAL DIFFERENCES OBSERVED RELEVANACE OF SOME DEPENDENT ON ADDITIONAL COVARIATE DATA MOST VALUES WITHIN NORMAL RANGES SCHEDULED FOLLOWUP EXAMINATIONS WILL PROVIDE DATA NECESSARY TO DEFINE RELEVANCE �SIGNIFICANT GROUP DIFFERENCES SELF-PERCEPTION OF HEALTH SKIN CANCER REPORTED BIRTH DEFECTS AND NEONATAL DEATHS BABINSKI REFLEXES SUBJECTIVE PSYCHOLOGICAL MEASURES HEPATIC FUNCTION TESTS (GGTP, LDH. CHOLESTEROL) PERIPHERAL PULSES THYROID AND TESTOSTERONE �PHASE II EXAMINATIONS PILOT EXAMINATIONS IN APRIL 85 1O EXAMINEES INITIAL EXAMINATIONS BEGAN MAY 85 2 GROUPS PER WEEK 9 GROUPS LOST DUE TO HOLIDAYS 2 0 PARTICIPANTS 3 9 COMPLETED MARCH 86 �PROGRAM MODIFICATIONS FROM BASELINE INTERVAL HISTORY (SUBJECT AND SPOUSE) PHONE INTERVIEW WITH ENTIRE COMPARISON GROUP BASELINE QUESTIONNAIRES TO NEW SUBJECTS AND SPOUSES DELETIONS: PULMONARY FUNCTION STUDIES NERVE CONDUCTION STUDIES SEMEN STUDIES IQ TESTING �PROGRAM ENHANCEMENTS SKIN CANCER EVALUATION IMPROVED ALCOHOL AND SMOKING ASSESSMENT COMBAT STRESS ASSESSMENT BIRTH DEFECT SEVERITY ASSESSMENT DOPPLER EXAMINATION OF PULSES IMPROVED IMMUNOLOGIC ASSESSMENT PORPHYRIN PROFILE BY HPLC �EXAMINATION QUALITY CONTROL. OBSERVATION BY MONITOR WITH CHECKLIST MARK-SENSE EXAMINATION FORMS MANUAL AND ADP QC REVIEWS STRICT LABORATORY QC TIGHT CV% FIRCUSUM NO PRIOR STANDARDS FOR IMMUNOLOGY QC PATIENT CRITIQUE FORMS BLINDNESS OF EXAMINERS �DAILY SCHEDULE O53O START 12-HOUR URINE COLLECTION O63O DEPART FOR CLINIC 07OOO8OO INITIAL BLOOD DRAW O800O9OO SKIN O8OO150O EXAMINATION TESTING PSYCHOLOGICAL TESTING INTERVIEW 1730 TURN IN URINE COLLECTION �RESTRICTIONS 3 DAY CARBOHYDRATE LOADING DIET 3 DAY ABSTINENCE FROM ALCOHOL NO ALCOHOL FOR 2 DAYS ONSITE FASTING AFTER MIDNIGHT NO CAFFIENE OR NICOTINE AFTER MIDNIGHT RESTRICTED UNTIL AFTER ECG AND DOPPLER �PARTICIPATION RANCH HANDERS ORIGINALS REPLACEMENTS PARTIAL 39 62 32 REFUSAL 6 9 9 NEW - 3 39 GAINS 45 74 80 LOSSES 74 64 21 971 872 267 PHASE II TOTAL 1016 946 3-47 BASELINE TOTAL 1045 936 288 RETURNEES �IMMUNOLOGY STUDIES RANDOM SELECTION OF 1082 PARTICIPANTS B AND T CELL STUDIES PERFORMED SKIN TESTS TO 1768 ( * % 77) 541 DELETED TO AVOID CONFOUNDING �SPOUSE QUESTIONNAIRES INTERVAL HISTORY 1915 COMPLETED 979 ONSITE 936 MAIL/PHONE BASELINE GIVEN TO 7-4 FERTILITY REVIEW CONTACTED 2108 SPOUSES 1613 RESPONDED (76.5%) �TELEPHONE r QUESTIONNAIRES TOTAL SUBJECTS 7.963 INTERVIEWS COMPLETE 7. -411 ( 3 1 ) 9 . % REFUSALS 335 < 4.2%) UNLOCATABLES 190 ( 2.4%) 26 ( O.3%) DECEASED �THIRD PHYSICAL EXAMINATION SCRIPPS CLINIC LA JOLLA, CA MAY 1 ^Q'? �ANTICIPATED EXAMINATION ADDITIONS COMPLIANCE ENHANCEMENT TONOMETRY VISUAL ACUITY AUDIOGRAMS STOOL FOR OCCULT BLOOD < PROCTOSCOPIC FOLLOWUP) MEDICAL CONTACT CARDS �ANTICIPATED EXAMINATION ADDITIONS SCIENTIFIC ISSUES PULMONARY FUNCTION SCLR-90 STUDIES PSYCHOLOGICAL. BATTERY MILLON CLINICAL SCALE AUTOMATED BLOOD PRESSURE MONITOR PROTEIN PROFILE PHENOTYPE MARKER FOR ACTIVATED T CELLS OPTIMIZE NATURAL KILLER WITH INTERLUKIN-2 CELL ASSAY �ANTICIPATED EXAMINATION DELETIONS MMPI CORNELL INDEX POKEWEED ANTIGEN HALSTEAD-REITAN FROZEN MLC POOL TIMED URINE COLLECTION PERIPHERAL DOPPLER PORPHYRIN PROFILES CORTISOL DETERMINATIONS �ADDITIONAL CONSIDERATIONS CARDIAC STRESS TESTING PERSONALLY FUNDED IF DESIRED D-GLUCARIC ACID ASSAY OF BANKED URINE SPECIMENS SERUM DIOXIN ASSAY OF SUBSET OF EXPOSED GROUP �MILESTONES JUNE 1386 FALL 1936 MAY 1987 FALL 1 8 9 7 FALL 1 8 9 7 FALL. 1 8 9 9 ANALYSIS BEGUN 1986 MORTALITY UPDATE THIRd EXAMINATION TO BEGIN RELEASE OF SECOND MORBIDITY REPORT MORTAL I TY REPORT RELEASE OF THIRD MORBIDITY REPORT � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 059 Folder The folder containing the original item. 1581 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Wolfe, William H. Date A point or period of time associated with an event in the lifecycle of the resource September 29 1986 Title A name given to the resource Air Force Health Study (Ranch Hand II) Update Briefing Subject The topic of the resource Air Force Health Study study protocol morbidity trends mortality trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/1f9b2eb8741a985801b0699f1d446087.pdf f8d1a09cc62f75b69d8f463ffd33d77a PDF Text Text Item ID Number °1584 AlltllOr Lathrop, George D. Corporate Author RBpOrt/ArtldO TltlB Air F°rce Health Study: An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides, Summary, First Followup Examination Results, January 1985-September 1987 Journal/BookTItlQ Year 1987 Month/Day October n Color Number of Images 50 Dnsnrlnton NotBS Contract no. F41689-85-D-0010 and SAIC Project no. 2816-XX-195/254-XX. Item also includes 4 page report that was found inserted into the summary. Report is titled "Air Force Completes Second Ranch Hand Health Study" and is a short summary of the summary. V Wednesday, May 23, 2001 Page 1585 of 1608 �AIR FORCE OCMPLETES SBOOND RANCH HAND HEALTH STODY The Air Force today released the second health, study of its Herbicide Orange study. The study is an investigation of the possible health, effects from herbicide exposure of Air Force members who conducted aerial spraying missions in Southeast Asia (Operation Ranch Hand). Like the initial 1982 study, this follow-up concludes that there is insufficient evidence to support a cause and effect relationship between herbicide exposure and health affects in the Ranch Hand group at this time. A number of minor medical findings have been reported that require continued surveillance. The Air Force is currently engaged in a collaborative study with the Centers for Disease Control to determine whether blood dioxin levels vary significantly in the Ranch Hand population. Initial serum analysis indicates clear and meaningful dioxin exposure did occur in the Ranch Hand group. The follow-up study examined 2,309 participants (1,016 Ranch Hands and 1,293 comparisons). It was conducted under contract to the Air Force by Science Applications International Corporation, in conjunction with the Soripps Clinic and Research Foundation and the National Opinion Research Center. This follow up study marks nearly 9 years of intensive Air Force research into the herbicide health question. Nearly 100 government, academic and industry scientists have guided and contributed to the study since its inception. The current advisory committee, (The Advisory Committee on Special Studies Related to the Possible Long-term Health Effects on Ehenoxy Herbicides and Contaminants) comprised of renowned epidemiologists and chaired by Dr. Robert W. Miller of the National Cancer Institute, has reviewed and approved this report. A detailed report on the study follows. �No significant differences between the Ranch Hand and comparison groups were seen in the.1982-1985 interval for skin or systemic cancers. However, when overall lifetime basal cell carcinoma rates were adjusted for risk factors involved in the cause of such cancers (e.g. sun exposure, skin color, skin reaction to sun), Ranch Hands had a significantly higher proportion of basal cell carcinoma (the most common neoplasm in the white population in the United States) than comparisons. Yet the degree of difference between the two groups is decreasing with the passage of time. No group differences were observed for systemic cancer, although one soft tissue sarcoma and one suspected lymphoma were noted in the Ranch Hands. Two similar cases were previously found in the comparison group bringing the lifetime total to two of these cancers in each group. Overall, the cancer findings were not viewed as disturbing but as a reason for continued medical surveillance. The neurological assessment of cranial nerve function, peripheral nerve function, and central nervous system coordination did not reveal any consistently significant group differences. There were fewer neurological abnormalities noted than in the baseline study. Age, alcohol, and diabetes had the medically expected effects on many neurological measures. There were no group differences for current or past psychological or psychiatric illness. Age, educational level, and alcohol history showed strong and expected effects on the psychological measures. In evaluating the gastrointestinal and hepatic functions, both groups had an equivalent history of liver disease and ulcers. The follow-up examination disclosed more statistically significant findings for tests of liver function than the Baseline examination, but they were equally divided between the two groups. There were no demonstrated clinical, statistical, or exposure patterns consistent with an herbicide-related health effect. No evidence was found to suggest an increased likelihood of porphyria cutanea tarda in the Ranch Hand group. In the dermatological assessment, no cases of chloracne were diagnosed on examination, nor was there a history of acne that suggested past chloracne in the Ranch Hand group. Current and past non-malignant skin disease was equivalent in both groups. Overall, there was similarity in the cardiovascular health between the Ranch Hand and comparison groups. The cardiovascular evaluation showed no significant group differences for reported or verified hypertension, reported heart disease, or reported or verified heart attacks. However, verified heart d1sea.se was significantly greater in the Ranch Hands than the Comparisons. This category included a wide variety of cardiac conditions. Comparisons of the renal, and endocrine systems were very similar. Extensive study of the immune system revealed group equivalence in this important body function. �The pulmonary assessment, consisting of past history, physical examination, and x-ray results did not indicate any consistently different disease patterns in the two groups. Age and lifetime smoking history were important risk factors for most pulmonary measures. The exposure index analyses, which were specific to each occupation, revealed sporadic differences between exposure levels; however, there were no consistent dose-response relationships that supported an herbicide effect for any clinical area. These analyses found a subtle but consistent narrowing of medical differences between the Ranch Hands and comparisons since the baseline study in 1982. Continued close medical surveillance of these military populations is strongly indicated. At this time, there is not sufficient plausible or consistent scientific evidence to implicate a causal relationship between herbicide exposure and health effects in the Ranch Band group. Exposure Index Refinements Since the development of the study protocol and the analysis of the baseline data, there has been concern over the accuracy and validity of herbicide exposure estimates. The current herbicide exposure index is based on an estimate of the gallons of herbicide sprayed per month ( 9 2 1 7 ) 16-91; the levels of dioxin contamination and the number of men assigned each month. A different index was established for each occupation group. It was recognized that a better method of determining exposure was needed. In mid 1986, strong correlations between dioxin levels in fat tissue and blood were demonstrated by the Centers for Disease Control (CDC) and other institutions. The Air Force is currently engaged in a collaborative study with the CDC to determine whether serum dioxin levels vary significantly in the Ranch Hand population. Approximately 200 volunteers have supplied a pint of blood for analysis at the CDC laboratories. Initial serum dioxin analysis indicates clear and meaningful exposure did occur in the Ranch Hand group. The range of dioxin exposure in the Ranch Hand group (3-314 parts per trillion) is significantly higher than the control group ( - parts per trillion.) A part per trillion is equivalent to one second 39 in 32 thousand years. It is anticipated that serum dioxin testing will be expanded to all study participants and a more meaningful exposure index based on total current dioxin contamination may be developed for each participant. It may also be possible to calculate a precise half-life of dioxin in humans. These expanded studies will allow the estimation of dioxin in the participants since their departure from Southeast Asia. Study participants are providing blood for this dioxin test during the fifth year follow-up study currently underway. �Within the next two years, another morbidity report, mortality reports, and an expanded birth, defects study will be completed. The Air Force Health Study is a planned twenty year research, effort, concluding in 2002. This follow-up morbidity report has been reviewed and approved by the Advisory Committee on Special Studies Related to the Possible Long-Terra Health Effects on Fhenoxy Herbicides and Contaminants. �Air Force Health Study An Epidenriologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides Air Force Team SAJC Team George D. Lathrop, M.D., M.P.H., Ph.D. Stella G. Machado, Ph.D. Theodore G. Karrison, Ph.D. William D. Grubbs, Ph.D. Wanda F. Thomas, M.S. COL Joel LTC LTC William H. Wolfe, M.D., M.P.H. E. Michalek, Ph.D. Judson C. Miner, D.V.M., M.P.H. Michael R. Peterson, D.V.M., M.P.H., Dr.P.H. Project Manager: W.F. Thomas Program Manager: R.W. Ogershok SCIENCE APPLICATIONS INTERNATIONAL CORPORATION 8400 Westpark Drive McLean, Virginia 22102 EPIDEMIOLOGY DIVISION USAF School of Aerospace Medicine Human Systems Division (AFSC) Brooks Air Force Base, Texas 78235 October 1987 SUMMARY First Followup Examination Results January 1985 to September 1987 Contract Number F41689-85-D-0010 SAIC Project Number 2-8I6-XX-195/254-XX (Distribution Unlimited) �INTRODUCTION FIRST FOLLOWUP MORBIDITY STUDY The Air Force Health Study is an epidemiological study conducted to determine whether adverse health effects exist and can be attributed to occupational exposure to Herbicide Orange. The study consists of mortality and morbidity components, with followup studies. The Baseline study was conducted in 1982, and the first followup physical examinations were performed in 1985. The purpose of this report is to present the results of the first followup study. In the Baseline study, each living Ranch Hand was matched to the first living and compliant member of a randomly selected Comparison group based on age, race, and military occupation. The Comparisons had served in numerous flying organizations that transported cargo to, from, and within Vietnam but were not involved in the aerial spray operations of Herbicide Orange. Recruitment for the first followup was in accordance with the Study Protocol: All previous participants and refusals, newly located study members, and replacements (matched to noncompliant Comparisons on self-perception of health) were invited. Of the living Baseline study participants, 99.2 percent were contacted to enroll in the followup on a strictly voluntary basis. Participation was very high, with 93 percent of both the Ranch Hands and the Comparisons who were fully compliant at Baseline also participating in the followup. Overall, the 2,309 followup participants (1,016 Ranch Hands and 1,293 Comparisons) represented a loss to the study of 159 individuals but a gain of 199 new participants since Baseline. Statistical analyses of selection and participation bias supported the use of the total Comparison group for the main analyses presented in this report. 1 The followup study was conducted under contract to the Air Force by Science Applications International Corporation, in conjunction with the Scripps Clinic and Research Foundation and the National Opinion Research Center. Most of the data were collected through face-to-face interviews and physical examinations conducted at the Scripps Clinic in La Jolla, California. Other data sources included medical and military records and the 1982 Baseline data base. As a contract requirement, all data collection personnel were blind to exposure status, and all phases of the study were monitored by stringent quality control. The statistical analyses were based on analysis of variance and covariance, chi-square tests, Fisher's exact tests, general linear models, Kolmogorov-Smirnov tests, logistic regression, proportional odds models, t-tests, and log-linear models. The release of this 1987 followup Morbidity Report marks more than 8 1/2 years of intensive Air Force research in the herbicide question. Since the commitment to Congress in October 1978 to conduct an epidemiologic investigation of Air Force.personnel who aerially disseminated herbicides in the Vietnam War (code-named Operation Ranch Hand), the United States Air Force Surgeon General has issued the following publications: a Study Protocol, four annual mortality reports, the Baseline Morbidity Report, and this first followup morbidity report. Within the next 2 years, the second followup morbidity report, other annual mortality reports, and an expanded birth defects study are expected for publication. This level of commitment has used approximately $40 million of contract research funds, excluding significant Air Force in-house expenditures. �Nearly 100 Government, academic, and industry scientists have guided and contributed to the Air Force Health Study (AFHS) since its inception. The Air Force's current advisory committee, chaired by Dr Robert W. Miller of the National Cancer Institute, is responsible for providing assistance on all scientific and medical matters pertaining to the AFHS. The chief strength of the AFHS is its design. The interwoven study elements of multiple mortality assessments, a Baseline morbidity study, and five followup morbidity studies over 20 years provide a comprehensive approach to the detection of adverse health effects. The weakest feature of the design is the mortality assessment which cannot detect group differences for very rare conditions (e.g., soft tissue sarcoma) because of the inherent constraints of the limited size of the population. The strength of the mortality studies should increase with the aging of the study populations and the increase in death with the passage of time. All four mortality assessments have shown that the Ranch Hand population is faring about the same as the Comparison group, with no unusual causes o-f death, increased frequency of death, or evidence suggesting death at younger ages. Because of the healthy veteran .effect, both groups are surviving significantly longer than similarly aged civilians. The morbidity assessment, released in 1984, disclosed only minor differences between the Ranch Hands and the Comparisons, and these differences were not traditional indicators of dioxin>-related disease. Both the content and the progress of the AFHS has been presented on many occasions to Congress, to the media, and to scientific meetings around the world. On the whole, the AFHS has been very well received in these circles, giving additional strength and credence to this work. The questionnaire and physical examination data were analyzed by major organ system. The primary focus was on the assessment of differences between the Ranch Hand and Comparison groups based on data from the first followup. Additionally, dose-response relationships within the Ranch Hand group were examined, and longitudinal assessments of differences in the changes of the two groups between the examinations were conducted for selected variables. Reported significant findings in subsequent major organ systems are understood to be significant at the 0.05 level or less. GENERAL HEALTH General physical health was evaluated by five measures, three of which were subjective (self^-perception of health, appearance of distress, and appearance of relative age), and two of which were objective (percent body fat and sedimentation rate). Table 1 presents a summary of all the unadjusted and adjusted analyses of these five variables. The mean age of the Ranch Hands was 46.9 years and the comparisons mean age was 46.8 years. The Ranch Hands rated their health as fair or poor more often than the Comparisons (9.1$ versus 7.3%, respectively), but this difference was not significant by categorical testing. However, further analysis revealed a significant group-by-occupation interaction; differences were largely confined to the enlisted groundcrew category. Both the Ranch Hand and Comparison groups noticeably improved their perceptions of health since the 1982 Baseline examination. �Only 10 individuals were reported as appearing acutely ill or distressed at the followup examination, 4 were Ranch Hands and 6 were Comparisons. This difference was not statistically significant and the data were insufficient for adjusted analyses. TABLE 1. Overall Suuury Results of Unadjusted and Adjusted Analyses of General Health Variables Unadjusted Variable Categorical Adjusted Mean Categorical Mean **** Self-Perception of Health NS Appearance of Illness/Disstress NS Appearance of Relative Age NS Sedimentation Rate 0.013 NS 0.011 NS Percent Body Fat NS 0.037 NS 0.035 - • **** —Analysis not performed. ****Group-by-covariate interaction. 'Analysis not possible due to sparse data. Appearance of relative age, as determined by the examining physician, showed 1.6 percent of the Ranch Hands appearing younger than their stated age, 94.3 percent appearing the same, and 4.1 percent appearing older (as contrasted to 0.7$, 95.4$, and 3.9$, respectively, in the Comparison group). There was a significant group-by-occupational interaction, but none of the estimated relative risks for the occupational categories was significant. This observation at the followup examination contrasted with the significant tendency at the Baseline for a higher percentage of Ranch Hands than Comparisons to appear younger than their stated ages. The geometric mean sedimentation rates (5.05 mm/hr Ranch Hand versus 4.93 mm/hr Comparison) did not differ significantly by group, either unadjusted or after adjustment for age, race, occupation, personality score, and an ageby-personality score interaction. However, in the dichotomous form, 5.8 percent of the Ranch Hands had sedimentation rate abnormalities as contrasted to 3.6 percent in the Comparison group. This difference was significant by �both unadjusted and adjusted tests. Also, this finding was opposite to that of the Baseline examination, where it was noted that younger Comparisons had significantly elevated sedimentation rates. The mean percent body fat of the Ranch Hands was significantly lower than the Comparisons (21.10J+0.15, 21.5^+0.14, respectively, p-0.037), and was of nearly the same magnitude after adjustment for age, race, and occupation. However, both unadjusted and adjusted categorical tests did not reveal significant group differences, although the percent obese was lower in the Ranch Hands than in the Comparisons. No group differences in percent body fat were noted at the Baseline examination. Detailed exposure analyses were done on four general health variables (appearance of acute distress was too sparse for testing). Only one analysis demonstrated statistical significance, i.e., a positive association of sedimentation rate abnormalities with increasing exposure in the enlisted flyer cohort. Overall, no consistent pattern of exposure effects was discernible, and the exposure findings at the thirds-year followup were similar to the findings at Baseline. Longitudinal differences between the 1982 Baseline and the 1985 followup examination were assessed by analyses of two discrete variables, selfperception of health and sedimentation rate. Perceived health showed no significant group differences over time, but both the Ranch Hand and Comparison groups paradoxically reported symmetrical improvements in their perceptions over the 3~year period. The sedimentation rate analysis revealed a highly significant group difference (p-0.002), due to a reversal of findings between examinations, i.e., a significant detriment in the younger Comparisons at the Baseline versus a significant detriment in the Ranch Hands at the followup. The cause(s) and biological relevance of this observation are unclear. In conclusion, a nonspecific assessment of general physical health has shown relatively close similarity between the Ranch Hand and Comparison groups, with the Ranch Hands continuing to perceive their health more negatively than the Comparisons, having a slightly more favorable percent body-<fat proportion, but a higher proportion of abnormal sedimentation rates that reflects a marked change since the Baseline examination. These findings must be placed in context with the organ and system-specific evaluations found in the succeeding chapters. MALIGNANCY The cancer analysis focused on cancer occurrences in the Baselinefollowup interval, and also included analyses of the Baseline plus interval cancer history. A listing of systemic malignancies occurring in the study particpants is shown in table 2 and a summary of the cancer findings is given in Table 3No significant unadjusted differences were found between nonblack Ranch Hands and Comparisons in the Interval (Baseline^Followup) incidence rates of basal cell carcinoma, melanoma, squamous cell carcinoma, all malignant skin cancers, sun^exposure related malignant neoplasms (comprising basal cell carcinoma, melanoma, and epithelial neoplasms NOS) or all malignant skin cancers as a group. The unadjusted group contrast of all skin neoplasms �TABLE 2. Summary of Pollowup Participants With Lifetime Incidence of Verified Malignant Systemic Neoplasms by Group Group Ranch Hand Comparison Total 0 1 Oral Cavity and Pharynx 1 3..b 0 3 Larynx 0 1 1 Thyroid Gland 0 2 2 Esophagus 0 lc 1 Bronchus and Lung 2 1 Colon 0 0 5d.. 5 Kidney and Bladder 4 3 7 Prostate 2 2 4 Testicles 3 0 3 Connective and Other Soft Tissue 1 1 2 Hodgkin's Disease 0 1 1 Ill-Defined Sites lf I9 2 Site Eye Total 17 17 34 'includes one Ranch Hand with separate malignancies of tongue and epiglottis and also malignant neoplasm of bone. b Includes one Ranch Hand with separate malignant neoplasms of tongue and oropharynx and secondary malignant neoplasm of other site. c Also has malignant neoplasm of bone. d lncudes one Comparison with secondary malignant neoplasms of liver and bone and bone marrow. 'includes one Comparison with secondary malignant neoplasm of liver. 'Malignant neoplasm of thorax. g Malignant neoplasm of face, head, or neck. �TABLE 3. Overall Siunary Table: Unadjusted and Adjusted Analysis of Interval and Lifetime Skin and Systeaic Cancer Incidence Cancer Type Baseline-Followup Interval UnadjustedAdjusted Lifetime (Baseline & Follovup) UnadjustedAdjusted Malignant Skin Cancer (Nonblacks only) Verified Basal Cell Carcinoma NS **** NS Verified plus Suspected Basal Cell Carcinoma NS **** NS Verified Melanoma NS —* NS Verified plus Suspected Melanoma NS —* NS Verified Squamous Cell Carcinoma NS —* NS Verified plus Suspected Squamous Cell Carcinoma NS —* NS Verified Sun Exposure Skin Cancers NS NS NS* Verified plus Suspected Sun Exposure Skin Cancers NS NS NS —" NS All Verified Malignant Skin Cancers NS Verified plus Suspected Malignant Skin Cancers NS NS Verified Skin Cancers of Any Type NS* S Verified plus Suspected Skin Cancers of Any Type NS NS* **** NS �TABLE 3. Overall Summary Table: Unadjusted and Adjusted Analysis of Interval and Lifetime Skin and Systemic Cancer Incidence (continued) Baseline-Followup Interval UnadjustedAdjusted Cancer Type Lifetime (Baseline & Follovup) Unadjusted Adjusted Malignant Systemic Cancer (Blacks and Nonblacks) Verified Systemic Cancer NS NS NS **** Verified plus Suspected Systemic Cancer NS **** NS **** All Neoplasms (Blacks and Nonblacks) Any Type, Any Location13 Verified NS* NS: Not significant (p>0.10). ****Group-by-covariate Interaction. —"Analysis not done. NS*: Borderline significant (0.05<p<0.10). Comprises malignant, benign, uncertain behavior. S: Significant (p<0.05). �(comprising malignant and benign neoplasms, and neoplasms of uncertain behavior or unspecified nature) was marginally significant, with a higher rate among Ranch Hands. When suspected malignant skin cancers (noted at Followup but not verified at the time of writing) were included in the analyses with the verified conditions, all the unadjusted group contrasts were nonsignificant. The covariates used for the adjusted analyses of basal cell carcinoma and the sun exposure related skin malignancies were age, occupation, skin color, reaction of skin to sun, and average latitude, all of which were highly associated with skin cancer incidence. Other host factors were related to skin cancer incidence, but not as strongly as those included in the analysis. A borderline association with smoking history was noted, and was determined to be partly an age effect. Analysis of the incidence of interval basal cell carcinoma revealed a significant group«-by-occupation interaction, due to a significant group difference for enlisted flyers, but not for officers or enlisted groundcrew. .Inclusion of suspected basal cell carcinoma resulted in a group-bysun reaction index interaction. This was due to Ranch Hands with an intermediate reaction to sun having a higher relative risk than the corresponding Comparisons. The adjusted group contrast of the incidence rates of verified sunexposure related skin cancers was not significant; inclusion of suspected conditions did not alter this lack of significance. There was no significant group difference for Blacks and nonblacks in the unadjusted incidence rates of all interval verified malignant systemic neoplasms combined, nor was there a significant difference in the adjusted group rates. Analysis of the verified plus suspected interval systemic cancers showed a nonsignificant unadjusted group difference, but a group by occupation interaction was found in the adjusted analysis. This was due to a significant group difference of verified plus suspected systemic malignancies among the enlisted flyers with five occurrences among the Ranch Hands, but none among the Comparisons. Age and a race^by-packyear interaction were important adjusting factors. The Baseline and Followup data were combined for the assessment of lifetime incidence of cancer; occurrences of cancer prior to Vietnam were excluded. There were no significant unadjusted group differences in lifetime incidence rates among nonblacks for basal cell carcinoma, melanoma, squamous cell carcinoma, the sun-exposure related skin cancers, or all malignant skin cancers combined. The unadjusted group contrast of all lifetime skin malignan*cies was significant, with a higher rate among Ranch Hands. Inclusion of suspected cancers with the verified cancers reduced the difference between the groups for all these malignant skin contrasts, except for the sun-exposure related skin cancers, for which a marginally significant group difference was found. However, the contrast of all skin malignancies remained close to significance. Adjusted analysis of the incidence rates of lifetime basal cell carcinoma revealed a significantly higher incidence rate among Ranch Hands (Adj. RR: 1.56, p-0.035). Significant effects of an occupation-by-age interaction, a �skin color-by-sun reaction index interaction, and a sun reaction index-byaverage residential latitude interaction were seen. The adjustment resulted in a significant relative risk that, moreover, was higher than the unadjusted relative risk. Average residential latitude, associated with both group and skin cancer, and skin color, which was associated with the disease and marginally associated with group, played a major part in the change from the unadjusted analysis due to confounding. Inclusion of suspected basal cell carcinoma in the adjusted analysis resulted in a group by sun reaction index interaction, as was noted for the interval analysis. The adjusted group contrast in incidence rates of the sun-exposure related skin cancers was also significant (Adj. RR: 1.54, p=0.030), which is not surprising since the majority are basal cell carcinoma. Inclusion of the suspected conditions resulted in a nonsignificant group contrast. The unadjusted group contrasts of the incidence rates of all systemic cancers combined were not significant, both for verified and verified plus suspected conditions. There was one new occurrence of a soft tissue sarcoma (Ranch Haind) and one suspected cancer of the lymphatic system (Ranch Hand), in addition to the one previously reported soft tissue sarcoma and one Hodgkin's disease in the Comparison group. Adjusted analysis of all lifetime malignant systemic neoplasms as a group, however, revealed a group by occupation interaction, due to a significantly higher rate for Ranch Hand enlisted flyers as contrasted to Comparisons. The same result was found for verified plus suspected systemic cancers. In conclusion, there were no adjusted or unadjusted differences between groups in basal cell carcinoma incidence in the Baseline-followup interval. At Baseline, a significantly higher rate of basal cell carcinoma was found for Ranch Hands when contrasted with Original Comparisons. When the Baseline data were combined with the interval data, adjusted analysis, but not the unadjusted analysis, revealed a significantly higher rate of basal cell carcinoma among the Ranch Hands than among all Comparisons. The relative risk of basal cell carcinoma appears to be declining over time. Relative risks of basal cell carcinoma and systemic cancer were found to be consistently larger than 1. Most of the skin cancers were basal cell carcinomas, upon which most of the skin cancer analysis focused, thus relative risks for swvexposure related skin neoplasms and all malignant skin cancers as a group were very similar to those for basal cell carcinoma. The number of occurrences of systemic cancer was small, in part because the cohort is relatively young, and although the relative risks (lifetime and interval) are greater than 1, the difference between groups is not significant. Sufficient time may not have elapsed since Vietnam to enable a group difference in systemic neoplasms, if one exists, to be apparent. NEUROLOGY Interval questionnaire data (1982 through 1985) on neurological illnesses, verified by medical records, revealed no significant group differr ences. These data were added to verified Baseline historical information to �assess possible differences in the lifetime experience of neurological disease. Again, there was no significant difference between the Ranch Hand and Comparison groups. A detailed neurological examination evaluated neurological integrity in three broad areas: cranial nerve function, peripheral nerve function, and central nervous system (CNS) coordination. The summary analytic results for all measurement variables comprising these three functional areas are presented in Table 4. Assessment of the 12 cranial nerves was based on the measurement of 1M variables. Two summary indices were constructed. Both the unadjusted and adjusted analyses did not disclose any statistically significant group differences, although two variables, speech and tongue position, were of borderline significance, with Ranch Hands faring worse than Comparisons. One of the two cranial nerve summary indices was marginally significant, again with the Ranch Hands at a slight detriment. The unadjusted and adjusted analyses of peripheral nerve function, as measured by eight variables (four reflexes, three sensory determinations, and muscle mass), did not reveal significant group differences. CNS coordination was evaluated by four measurements and a constructed summary variable. Hand tremor was found to be of borderline significance, with the Ranch hands faring slightly worse than the Comparisons. The CNS summary index showed a significant detriment to the Ranch Hands. The exposure analyses for neurological variables with reasonable counts of abnormalities showed only occasional statistically significant results. No consistent pattern with increasing exposure was evident for any occupational category of the Ranch Hand group. In a longitudinal analysis of the Romberg sign and the Babinski reflex, only the Babinski reflex revealed a significant difference between the Baseline and followup examination, with the Ranch Hands converting from significant adverse findings at Baseline to favorable nonsignificant findings at the followup examination. Overall, the followup examination findings are quite similar to the Baseline findings. However, several distinct patterns were evident from the analyses: (1) The followup examination detected substantially fewer abnormalities for almost all measurement variables, (2) the decrease in abnormalities was equivalent in both groups, (3) most of the covariate effects were classical, although exceptions were evident, CO the adjusted analyses were uniformly similar to the unadjusted analyses, (5) the constructed summary variables were generally statistically significant, or of borderline significance (however, some indices were created after the data were examined), and (6) although statistical significance at the pre^-assigned '-level of 0.05 was not achieved for any of the measurement variables, abnormalities tended to cluster in the Ranch Hand group. Of the three group-by-covariate interactions in the adjusted analyses, only one, a borderline group-by-insecticide exposure interaction for hand tremor, where Ranch Hands exposed to insecticides had a marginally significant adverse effect, was of probable biologic (and operational) significance. 10 �TABLE 4. Overall Suwury Results of Unadjusted and Adjusted Analyses of Neurological Variables Variable Unadjusted Adjusted Direction of Results** Questionnaire* Physical Examination Neurological Disease (Interval) Neurological Disease (History) NSb NS Cranial Nerve Function Smell Visual Fields Light Reaction Ocular Movements Facial Sensation Corneal Reflex Jaw Clench Smile Palpebral Fissures Balance Gag Reflex Speech Tongue Position Relative to Midline Palate and Uvula Movement Neck Range of Motion Cranial Nerve Function Index* Cranial Nerve Function Index (excluding Neck Range of Motion) NS NS NS NS NS —c NS NS NS NS NS NS* RH>C NS* NS NS NS NS NS NS* NS* RH>C Peripheral Nerve Function Pin Prick Light Touch Muscle Status Vibratory Sensation Patellar Reflex Achilles Reflex Biceps Reflex Babinski Reflex NS NS NS NS NS NS NS NS 11 **** NS NS NS RH>C �TABLE 4. (Continued) Overall Suuury Results of Unadjusted and Adjusted Analyses of Neurological Variables Variable Unadjusted Adjusted Direction of Results** Central Nervous System Coordination Tremor Coordination Romberg Sign Gait CNS Summary Index NS* NS NS NS 0.036 NS* Rfl>C — — NS O.OA2 RH>C **RH>C: More abnormalities in Ranch Hand group than in Comparison group. 'Disease categories include: inflammatory diseases, heriditary and degenerative diseases, peripheral disorders, disorders of the eye, disorders of the ear, and other disorders. NS:Not significant (p>0.10). b No inflammatory diseases noted; borderline significant (p-0.069, RH>C) for other disorders; not significant for remaining categories. —Analysis not performed because of sparse number of abnormalities. c No abnormalities present. NS*Borderline significant (0.05<p<0.10). Constructed variable. ****Group-by-covariate interaction. 12 �In conclusion, none of the 27 neurological variables demonstrated a significant group difference, although several showed an aggregation of abnor*malities in the Ranch Hand group, which merits continued surveillance. Historical reporting of neurologic disease was equal in both groups. The clinical sensitivity in detecting neurological deficits varied substantially between the Baseline and the followup examinations, but the number of statistically significant variables remained about the same. None of the exposure analyses revealed dose-response patterns in the Ranch Hand occupational categories. The longitudinal analyses disclosed a favorable reversal of significant Babinski reflex abnormalities at Baseline to nonsignificant findings at . the followup examination for the Ranch Hands. The similarity in results between unadjusted and adjusted statistical tests is evidence of group equality for the traditionally important neurological covariates of age, alcohol, and diabetes. Of three group-by-covariate interactions in the adjusted analyses, only the Ranch Hand insecticide interaction with hand tremor was biologically plausible. PSYCHOLOGY Questionnaire data (verified by medical record reviews) for the lifetime events of psychotic illness, alcohol dependence, anxiety, or other neuroses disclosed no significant differences between groups for these conditions. (Table 5). TABLE 5 Distribution of Reported Psychological Illness By Group - Baseline and Followup Studies Combined Type of Illness Psychoses Alcohol Dependence Anxiety Other Neuroses Abnormalities Ranch H a n d s C o m p a r i s o n s NumberPercent NumberPercent 14 9 7 72 9 8 13 74 1.4 0.9 0.7 7.1 *Fisher's Exact Test 13 0.7 0.6 1.0 5.7 Total 23 17 20 146 p-Value* 0.138 0.473 0.501 0.197 �Analyses of the followup psychological examination emphasized Hi scales from the Minnesota Multiphasic Personality Inventory (MMPI), 3 parameters of the Cornell Medical Index (CMI), and the Halstead-Reitan Battery (HRB) impairment index. The similarity of the group distribution for the 14 MMPI variables, each stratified by the 3 occupational categories, was examined, and only 2 of the 42 tests approached statistical significance. The group distributions of the total CMI score were similarly contrasted, with separate analyses performed with stratification by the five covariates of age, race, occupation, education, and current drinking status. For one stratum of each of these covariates, a significant difference in the distribution of the Ranch Hand and Comparison scores was found. In all cases for the CMI, the Ranch Hand mean was greater than the Comparison mean. Distributional analyses using Original Comparisons generally reflected the same results as those involving the total Comparison group. Results of unadjusted and adjusted analyses on all of the 18 psychological variables are given in Table 6. The unadjusted analyses showed a significant difference for the MMPI scales of denial (p<0.001) and masculinity/femininity (p-0.017), the total CMI (p<0.001), and the Section A^H area subscore (p-0.003). A borderline significant difference was observed for the MMPI scales of hysteria (p-0.067) and social introversion (p-0.069). Comparisons had a greater percentage of abnopmal scores for the denial and masculinity/femininity scales, whereas Ranch Hands showed adverse findings for the other four variables. The overall MMPI results have been interpreted in light of the significant increased denial in the Comparison group. The covariates age, education, drink-years, current alcohol use, and occupation had pronounced effects on the psychological variables, with a significant association or a borderline significant association with at least twof-thirds of the 18 psychological variables. Many dependent variables in this chapter were affected by age in an expected pattern. Very few variables exhibited this pattern of consistency with drink--years. The intermediate category of greater than 0 to 50 drink-years often had'the smallest proportion of abnormalities. The post?-traumatic stress disorder (PTSD) variable, derived from a subset of'the MMPI, was strongly associated with the CMI measures, but not with the HRB Impairment Index. Race and the Vietnam combat index (used for the MMPI subscales) had significant associations with a lesser amount of the psychological variables (6 of 18 variables and 3 of 14 variables, for race and combat index, respectively). The adjusted analyses were generally quite similar to the unadjusted analyses with respect to group differences, although a direct comparison of these analyses was often clouded by the presence of a substantial number of interactions (six group4bycovariate interactions were significant, and three interactions approached significance [0.05<p£0.10]). The MMPI scales of denial and masculinity/femininity were statistically significant in both the adjusted and unadjusted analyses, where Comparisons showed an adverse effect over Ranch Hands. The A--H area subscore of the CMI (suggesting diffuse medical problems) was also significant, where the Ranch Hands had higher mean scores than the Comparisons, suggesting the Ranch Hands had more illness. Education was often involved in significant group interactions with high 14 �TABLE 6. Overall Summary Results of Adjusted and Unadjusted Analyses of Psychological Variables Variable Unadjusted Questionnaire! Psychological Illness Adjusted Direction of Results* NS Psychological Examination; MMPI Anxiety Consistency Oefensiveness Denial Depression Hypochondria Hysteria Nania/Hypomania Masculini ty/Feminini ty Paranoia Psychopathic/Deviate Schizophrenia Social Introversion Validity CMI Total CMI M-R Subscore A-H Area Subscore HRfi Impairment Index • NS NS NS <0.001 NS NS b NS*b NS 0.017 NS NS NS **** NS <0.001 NS MS b NS* NS 0.020 **** ORH RH>C ORH NS NS **** h NS*b NS **** **** RH>C <0.001 NS 0.003 **** NS 0.040 RH>C RH>C NS NS *RH>C - more abnormalities in Ranch Hands; ORH - more abnormalities in Comparisons. b lllnesses include psychosis, alcohol dependence, anxiety, and other neuroses. —Analysis not performed. NS: Not significant. NS*: Borderline significant (0.05<p<0.10). ****Interaction involving group. 15 �school-educated Ranch Hands demonstrating a higher percentage of abnormal scores than high school-educated Comparisons. No group differences were observed in the college-educated stratum. The M-R subscore of the CMI, a broad indicator of emotional health, was not statistically different between the two groups. The HRB impairment index, a measure of central nervous system (CNS) functional integrity, did not d i f f e r significantly between the Ranch Hand and Comparison groups. Strong covariates in the adjusted analysis were age, race, and education. Because of alternate statistical models and slightly different psychological testing parameters, a direct contrast between the psychological results of the Baseline and followup examinations was not always possible,, However, several broad patterns were observed: ( 1 ) the discordance between distributional tests and results from traditional statistical models of the MMPI variables was noted with data from both examinations; (2) there was a narrowing of group differences at the followup examination for most subjective' variables, either by a decrease in Ranch Hand reporting, or by an increase in Comparison reporting; and (3) as at the Baseline, functional CNS testing, as measured by the HRB impairment index, showed no group differences, and did not support an organic basis for differences in self-reported symptomatology. The longitudinal analysis of two MMPI scales, depression and denial, showed a significant reversal of depression seen at Baseline in the high schooleducated Ranch Hands. The determination of PTSD in both Air Force cohorts by a relatively new MMPI scale showed a prevalence rate of less than 1 percent. This low rate is strongly influenced by characteristics of the study population (e.g., age, education, and officer ratio). Unadjusted exposure index analyses did not reveal any patterns consistent with a dose-response relationship. For the adjusted exposure analyses, approximately one-third presented exposure interactions with the covariates of race, education, and age, but no consistent pattern could be identified. In conclusion, some test measures of psychological health (MMPI and CMI) did not show substantial adverse effects for either group. Significant test results were present in both groups or were noted in specific subgroups of a covariate. Educational level, age, and alcohol use showed strong effects on the psychological scales and scores in this psychological assessment. There was a subtle but consistent trend for more favorable subjective test results at the followup examination for the Ranch Hands relative to the Comparisons. Testing of the CNS by the HRB demonstrated an almost identical prevalence of pathology in both groups. DIGESTIVE SYSTEM The interval questionnaire revealed sparse reporting of liver disorders from 1982 to 1985 that was not significantly different between groups. Historical liver disease was verified by medical records, and these data were added to the verified Baseline history to assess possible lifetime differences. No significant differences were found. The medical record verifica- 16 �tion process showed that the historical data were generally correctly reported and classified between groups, except for the category of enlarged liver which showed a higher verification rate in the Comparison group. Digestive system mortality showed an overall nonsignificant excess in the Ranch Hands, but a relative nonsignificant excess of malignant neoplasms in the Comparisons. No differences were found for past or current peptic ulcer disease for the Ranch Hand and Comparison groups, adjusted for standard covariates as well as blood type. The physical examination disclosed a borderline significant increase of hepatomegaly in the Ranch Hand group. Emphasis was placed on nine laboratory test variables measuring liver function, i.e., serum glutamic-oxaloacetic transaminase (SCOT), serum glutamic^pyruvic transaminase (SGPT), gammaglutamyl transpeptidase (GGTP), alkaline phosphatase, total and direct bilirubin, lactic dehydrogenase (LDH), cholesterol, and triglycerides. In addition, • uroporphyrin and coproporphyrin measurements were obtained to assess liver function and the likelihood of porphyria cutanea tarda (PCT). The nine hepatic variables were subjected to continuous and categorical statistical tests, and were adjusted for the covariates age, race, occupation, current alcohol consumption, and unprotected exposure to both industrial chemicals and degreasing chemicals. Final statistical models used only the significant covariates and two-way interactions for adjustment. The two porphyrin measurements were analyzed only in the continuous form. The overall summary re-' suits of the analyses of these 11 variables are given in Table 7. The results showed a significantly lower mean SGPT level, a greater mean alkaline phosphatase level, a lower mean uroporphyrin level for Ranch Hands as contrasted with Comparisons, and a marginally significant greater mean coproporphyrin level. Only in the instance of alkaline phosphatase did the discrete analysis approach statistical significance. No group differences were noted for SCOT, GGTP, total and direct bilirubin, LDH, cholesterol, or triglycerides. However, an analysis using only the Original Comparisons revealed a significantly greater mean cholesterol level in the Comparison group. A review of the covariate effects in the adjusted statistical models revealed that all covariates behaved as expected with the exception of alcohol consumption for the alkaline phosphatase analysis, which showed an inverse relationship with wine consumption. Exploration of group"by-group covariate interactions for alkaline phosphatase, direct bilirubin, triglycerides, SCOT, and uroporphyrins revealed significant group differences within specific covariate strata. In particular, Ranch Hands exposed to industrial chemicals had a significantly higher adjusted mean level of alkaline phosphatase and a significantly higher abnormal prevalence rate of direct bilirubin than similarly exposed Comparisons. For triglycerides, Ranch Hands born in or before 1922 had a significantly higher adjusted mean level than similar aged Comparisons, while Ranch Hand officers exhibited a significantly higher abnormal prevalence rate than Comparison officers. For SCOT, Ranch Hand moderate current drinkers (more than one to four drinks per day) had a significantly higher mean level than corresponding Comparisons. In the opposite direction, Comparisons with a mean BUN level less than or equal to 1*1 (median participants) were found to have a 17 �TABLE 7. Overall Stunary Results of Unadjusted and Adjusted Analyses of Nine Hepatic Function Variables and Two Porphyrin Metabolite Tests Unadjusted Adjusted* Mean Categorical Mean CC Variable Questionnaire Liver Disease (Lifetime History) Hepatitis Jaundice Cirrhosis Enlarged Liver Miscellaneous Liver Disorders Peptic Ulcer Disease CD NS Physical Examination Hepatomegaly Categorical Direction of Results** NS* NS NS NS NS NS NS' RH>C Laboratory Testing SCOT SGPT GGTP Alkaline Phosphate Total Bilirubin Direct Bilirubin LDH Cholesterol Triglycerides Uroporphyrin Coproporphyrin NS NS* NS 0.009 NS NS NS NS NS 008 .4 NS* NS NS NS NS NS NS NS NS NS NS **** NS 0.048 NS 0.008 NS NS **** NS **** **** 0.029 NS **** NS NS NS NS NS NS NS NS* NS **** NS NS **** NS* Questionnaire-Laboratory Correlation Skin Bruises, Patches,-0.001 and Sensitivity ORH RH>C ORH Rfl>C RH>C *C: Continuous D: Discrete **RH>C: more abnormalities, or higher mean value, in Ranch Hands, ORH: more abnormalities, or higher mean value, in Comparisons. 'Adjusted for blood type. NS: Not significant (p>0.10). NS*: Borderline significant (0.05<p<0.10). --Analysis not performed. ****Group-by-covariate interaction. 18 �significantly higher adjusted mean uroporphyrin level than similar Ranch Hands. These results did not disclose any common pattern detrimental to the Ranch Hand group. These findings were generally consistent with the 1982 Baseline data, which disclosed a significantly increased mean cholesterol level in the Comparisons and nonsignificant Ranch Hand mean elevations for GGTP and LDH. Slight differences in analytic results are probably due to the use of more fully adjusted models used for the followup examination data. Overall, the followup examination laboratory data showed no adverse clinical or exposure patterns in either group. Further, the detection of significant mean shifts (still within normal range) by the continuous statistical tests, not mirrored by the categorical tests, suggests a circumstance of statistical power rather than findings of biological relevance. Of the five significant or marginally significant results that were found in the adjusted exposure index analyses, four exhibited a trend suggest!ve of an increasing dose-response relationship. In the enlisted flyer cohort, the percentages of SGPT abnormalities were significantly different and increased from the low to the high exposure categories. The corresponding mean values were marginally significantly different among exposure levels. Also, the mean levels of total bilirubin were marginally significantly different among exposure levels, increasing with exposure level. For enlisted groundcrew, the percentage of SCOT abnormalities significantly differed among exposure levels. Within the enlisted flyer cohort, all nine laboratory tests of hepatic function had the lowest percentage of abnormalities in the low exposure category; correspondingly, six of the nine mean levels were lowest for the low exposure category. Of the ten group-by-covariate interactions that were found, three (SGOT, SGPT, and GGTP) supported a dose-response relationship in the enlisted groundcrew cohort. Exploration of these interactions revealed a trend that showed an increasing association between current alcohol consumption and the dependent variables for increasing exposure levels. Longitudinal analyses for SGOT, SGPT, and GGTP disclosed no statistically significant group differences in the mean shifts from the Baseline to the followup examination. Interval reporting of PCT-like symptoms of skin patches, bruises, and sensitivity was significantly increased in the Ranch Hands (p-0.001). However, when these historic data were contrasted to both uroporphyrin and coproporphyrin abnormalities, no correlation was apparent, nor were there any significant group differences. Since an elevation in the uroporphyrin level is required for a diagnosis of PCT, the historic data were retabulated with only uroporphyrin abnormalities; again, no group differences were apparent, and, in fact, uroporphyrin abnormalities in both groups were higher in those participants without a history of skin disorders than in those participants with such a history. The likelihood of bona fide PCT among study participants, and particularly among the Ranch Hands, appears to be remote. In conclusion, the followup examination disclosed more statistically significant findings for tests of liver function than the Baseline examination, but they were equally divided between the two groups and did not demon- 19 �strate clinical, statistical, or exposure patterns consistent with an herbicide-related effect on health. No evidence was found to suggest an increased likelihood of PCT among the Ranch Hand group. DERMATOLOGY Interval questionnaire data on the occurrence, time, and location of acne were analyzed to assess the possible historical diagnosis of chloracne. No significant difference was observed between groups for reported occurrence of acne, although the Ranch Hand cohort reported slightly more acne. The occur-' rence of acne relative to 1961 was comparable between groups. A marginally significant difference in the occurrence of post-1961 acne was found, with more Ranch Hands than Comparisons reporting acne strictly post-SEA. The duration of post-31961 acne was not significantly different between the two groups. For participants with post-SEA acne, the spatial eyeglass distribution of acne (suggesting chloracne) was observed to be similar for the Ranch Hand and Comparison groups, both for individual sites and the combination of acne on the eyelids, ears, and temples. This analysis suggested that the occurrence of skin disease compatible with chloracne was not different in the two groups. Analyses of the followup physical examination data, as with the Baseline examination, placed primary emphasis on six dermatologic disorders: comedones, acneiform lesions, acneiform scars, inclusion cysts, depigmentation, and hyperpigmentation. Secondary emphasis was given to 16 other minor conditions (generally not associated with chloracne) recorded at the physical examination. No significant findings occurred in any variable, as reflected in Table 8.• No significant difference was found for any of these variables in the unadjusted analyses. The variable consisting of the 16 secondary conditions, labeled "other abnormalities," had the largest difference in the prevalence of abnormalities for the Ranch Hand cohort over the Comparison group (Est. RR: 1.08, 95$ C.I.: [0.92,1.28], p-0.3^9), but the difference was clearly nonsignificant. The covariate effects of age, race, occupation, and the presence of pre-SEA acne were often profound with respect to the recorded dermatologic conditions. The adjusted analyses closely mirrored the unadjusted analyses, with no significance noted between groups for any variable. Only one group-bycovariate interaction was observed in the adjusted'analysis of the dermatology index, with a group~by-presence of pre-SEA acne interaction noted. However, further analysis of this interaction did not show an adverse effect in the Ranch Hand group. Exposure index analyses did support dose-response relationships for some of the variables in certain occupational strata, but did not reveal a strong pattern of results suggesting a relationship between skin disease and herbicide exposure. Overall the followup examination results paralleled the Baseline findings. Although the followup examination detected more dermatologic abnormalities than those present at Baseline, slightly more abnormalities were found in 20 �TABLE 8. Overall Summary Results of Unadjusted and Adjusted Analyses of Questionnaire and Physical Examination Dermatologies! Variables Variable Unadjusted Adjusted Questionnaire Incidence of Acne Occurrence NS Relative to 1961 NS Relative to SEA (Post -1961 Cases) NS* Duration of Acne NS Location of Acne NS NS Physical Examination Comedones NS NS Acneifora Lesions NS NS Acne i form Scars NS NS Depigmentation NS NS Inclusion Cysts NS NS Hyperpigmentation NS NS Other Abnormalities NS NS Dermatology Index NS **** NS: Not significant (p>0.10). — Analyses not performed. NS*: Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. �the Comparisons, and moat relative risks approached unity. The longitudinal analysis for the dermatology index showed no statistically significant differences between groups in the change in results from the Baseline to the followup examination. In conclusion, none of the questionnaire results disclosed an increased likelihood of past chloracne in the Ranch Hands. The physical examination did not diagnose a current case of chloracne. The dermatological data were similar between the Ranch Hand and Comparison groups, and the longitudinal analy.sis of the dermatology index suggested equivalence between the Baseline and followup examination results. CARDIOVASCULAR SYSTEM The cardiovascular health of both cohorts was assessed by collection of reported and record-verified heart disease events; measurement of central cardiac function by systolic blood pressure, abnormal heart sounds, and electrocardiograph (ECG) findings; and evaluation of peripheral vascular function by diastolic blood pressure, funduscopic examination, presence of carotid bruits, and detailed manual and Doppler measurements of five peripheral pulses. Table 9 presents the overall summary of the unadjusted and adjusted results. Where possible, the analyses used the covariates of age, race, occupation, percent body fat, cholesterol, high density lipoprotein (HDL) cholesterol, cholesterol-HDL ratio, smoking history (packryears and current smoking level), alcohol history (drink-years and current drinking level), personality score, and differential cortisol. The cardiovascular variables did not reveal significant group differences, with the exception of verified heart disease, for which the proportions of recorded cardiac events were 24 and 20 percent in the Ranch Hand and Comparison groups, respectively, (p-0.054 unadjusted, p-0.036 adjusted). This finding was not reinforced by results of individual questionnaire or examination variables showing impairment in the Ranch Hands. There was a remarkable balance in relative risks above and below unity between the groups. Other related analyses showed an absence of significant group differences in reported or verified hypertension, reported or verified heart attacks, and reported heart disease. There was good correlation between the verified cardiovascular history'and the central and peripheral cardiovascular abnormalities detected at the physical examination, supporting accuracy and valid-' ity of .the cardiovascular measurements. The adjusted analyses of central cardiac function disclosed a significant group^-byrage interaction involving systolic blood pressure in the Black cohort, with a mean systolic blood pressure greater in the Ranch Hands than the Comparisons at younger age levels, but a lower mean pressure at the older ages; the group-by-age interaction was not significant in the nonblack cohort. Additionally, there was a significant group-*by-pack-years of smoking interaction for the overall ECG findings, and significant group-by-pack-years of smoking and group^-by?-percent body fat interactions for arrhythmia, but they all generally pointed to lower adjusted relative risks in the Ranch Hands. 22 �TABLE 9. Overall Summary Results of Unadjusted and Adjusted Analyses Cardiovascular'Variables Variable Statistical/ Clinical Analysis Unadjusted Adjusted NS NS MS' Historical and Verified Heart Disease Reported Verified Reported Verified Reported Verified Hypertension Hypertension Heart Disease* Heart Disease* Heart Attack Heart Attack NS NS NS* NS NS NS NS Central Cardiac Function Systolic Blood Pressure Discrete Continuous Heart Sounds Electrocardiogram (Overall) ECG: RBBB ECG: LBBB ECG: Nonspecific T-Vave Changes ECG: Bradycardia ECG: Tachycardia ECG: Arrhythmia ECG: Other Diagnoses NS NS NS NS NS NS NS NS NS 23 NS **** NS **** NS N/A NS NS N/A **** NS �TABLE 9. (Continued) Overall Sunary Results of Unadjusted and Adjusted Analyses Cardiovascular Variables Statistical/ Clinical Analysis Variable Unadjusted Adjusted Funduscopic Examination NS NS NS NS NS NS Carotid Bruits NS NS ' Peripheral Vascular Function Oiastolic Blood Pressure Radial Pulses Femoral Pulses Popliteal Pulses Oorsalis Pedis Pulses Posterior Tibial Pulses Leg Pulses Peripheral Pulses All Pulses Discrete Continuous Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS:Not significant (p>0.10). NS*:Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. 'Excluding hypertension. b RH>C (Est. RR: 1.25; 95* C.I.:. (1.02, 1.54], p-0.054). NS NS NS NS **** NS **** NS **** NS **** NS **** NS **** NS �In the analysis of peripheral vascular function, no significant group differences were observed for abnormalities involving radial, femoral, popliteal, posterior tibial, dorsalis pedis, or three anatomic aggregates of these pulses, either by manual palpation or Doppler techniques. This overall f i n d i n g was in distinct contrast to the 1982 Baseline examination, which by the manual palpation method, showed significant peripheral pulse deficits in the Ranch Hands. This favorable pulse reversal over the two examinations is primarily attributed to the rigid 4-hour tobacco abstinence applied prior to Doppler testing, although other factors may be related. The lack of group differences for pulse abnormalities was noted even though the manual and Doppler techniques differed significantly (p<0,05, p<0.001 for most) in the detection of abnormalities for all but one of the pulses or pulse combinations. For manually-determined pulse abnormalities, there was a significant group-by-race interaction for the popliteal pulses, a significant groupby-percent body fat interaction for the leg pulses, and significant groupby-occupation interactions for the posterior tibial, dorsalis pedis, and the three pulse aggregates (leg, peripheral, and all pulses). No interactions were encountered in the adjusted analyses of the Doppler results, and none showed significant group differences. Statistical analyses involving the Original Comparisons also showed no significant differences in the cardiovascular measurements between groups, although slightly different interactions were detected in some of the adjusted analyses. For the exposure analyses, the only statistically significant effects were those pointing to less bradycardia and less reported and verified heart disease in the medium exposure level category, as contrasted to the low exposure category, among the enlisted groundcrew. In many cases there were too few abnormalities within the occupational categories to permit formal statistical tests. Overall, the exposure analyses were deemed as unsupportive of any meaningful dose-response relationships. The longitudinal analysis of the pulse index confirmed the significant difference in the change in the pattern of results from the Baseline examination to the followup examination, largely due to a relatively greater increase of pulse abnormalities in the Comparison group than in the Ranch Hand group. There was no significant change in pattern between the two groups in overall ECG findings between examinations. There was a similar distribution of the covariates between groups, except for a slightly higher level of current Ranch Hand smoking (also observed at Baseline), and a corresponding slightly lower mean percent body fat. The general covariate effects were strong and showed expected, classical associations with the cardiovascular measurements. However, unexpected effects were consistently noted for personality score, with higher proportions of various cardiovascular abnormalities associated with scores in the Type B direction, a. finding possibly attributable to the method of personality determination. Although smoking was positively associated with many of the cardiovascular measurements, negative associations were seen between current smoking and reported and verified essential hypertension and between pack-years of smoking and verified hypertension. 25 �In conclusion, of 27 cardiovascular variables, only one, verified heart disease, showed a significant excess in the Ranch Hands, but this finding was largely unsupported by other cardiac measurements. Both manual palpation and Doppler recordings of five peripheral pulses were similar in both groups, in marked contrast to the 1982 Baseline examination which found significant pulse deficits in the Ranch Hand group. This change at the followup examination was most likely due to required tobacco abstinence prior to the pulse measurements. Exposure index analyses did not support a consistent dose-response relationship for any variable. Overall, there was remarkable similarity in the cardiovascular health between the Ranch Hand and Comparison groups. HEMATOLOGY The functional integrity of the hematopoietic system was assessed by the measurement of eight peripheral blood variables: red blood cell count (RBC), white blood cell count (WBC), hemoglobin (HGB), hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and platelet count (PLT). These variables were analyzed in the discrete form to detect differences in the percentages of values outside the designated laboratory range, as well as in the continuous form to detect shifts in mean values between the two groups. A summary of all of these analyses, unadjusted and adjusted for the covariates of age, race, occupation, and smoking, is presented in Table 10. The unadjusted discrete analysis of the percent abnormal values, both low and high, showed no statistically significant differences between the Ranch Hand and Comparison groups for any of the hematological variables. Similarly, the adjusted categorical analysis disclosed that none of the adjusted relative risks was significant for either group, and that no significant group-by covariate interactions were present. The unadjusted continuous analysis did not detect any significant differences in group means for any of the eight variables. The adjusted continuous analysis found no significant group differences for HGB, HCT, MCV, MCH, and MCHC, but encountered significant three^factor interactions for WBC (group^ by-race-by-age, group-by-age-by-smoking history, and group~by-race-byoccupation), for PLT (group-byrace-by-smoking history and group^by^raceby-current level of smoking), and a borderline interaction for RBC (group?by-occupation-by-smoking history). Ranch Hand versus Original Comparison analyses revealed further significant interactions for HGB, HCT, MCV, and MCH. As no group strata demonstrated consistent patterns of hematologic impairment, biologic relevance was not assigned to the interactions. The covariate effects of age, race, occupation, and smoking history were highly significant for many of the hematological variables. The effect of race was particularly profound for all variables except PLT. There was fair consistency in the covariate effects upon the RBC-related variables. Generally, decreasing hematologic values were associated with increasing age and the Black race, and increasing hematologic values were associated with increasing smoking. The detection of these classical covariate effects lends credence to the overall finding of nonsignificant group differences for all of the hematological variables. Significant group differences found for MCV and MCH at the Baseline examination were not significant at the first followup. Other differences (e.g., covariate effects, interactions) 26 �TABLE 10. Overall SUBMIT? Results of Unadjusted and Adjusted Analyses of Heaatologlcal Variables Unadjusted Adjusted Mean Categorical Mean Categorical RBC NS NS NS* NS VBC NS NS **** NS HGB NS NS NS NS HCT NS NS NS NS HCV NS NS NS NS HCH NS NS NS NS HCHC NS PLT NS NS — NS *_-* • J WwWK — NS NS: Not sgnifleant (p>0.10). NS*: Borderline significant group-by-covariate interaction (0.05<p<0.10). —Analysis not performed due to sparse data. ****Group-by-covariate interaction. Note: Significant group-by-covariate interaction, Ranch Hands versus Original Comparisons only, for HGB, HCT, MCV, and MCH. 27 �between the Baseline and followup examinations may be due to small numeric shifts in the cohorts under study (see Chapter 2) and the selection of alternate statistical models, or due to chance. Unadjusted continuous exposure analyses in the Ranch Hand group revealed only one significant effect (RBC in enlisted groundcrew) and one borderline effect (HCT in enlisted groundcrew) but neither was consistent with a plausible dose-response relationship. The adjusted continuous exposure analyses found only one significant contrast (HCT, medium exposure versus low exposure, enlisted groundcrew). However, seven exposure level-by-covariate interactions were noted for four of the hematological variables. Discrete outcome analyses of the exposure level index revealed a significant result only for WBC in the enlisted flyers. The longitudinal analyses of MCV, MCH, and PLT found significant differences only for PLT values between the Baseline and followup examinations, with the Baseline group difference-in mean values closing to near equivalence at the followup examination. In conclusion, none of the eight hematological variables were found to differ significantly between the Ranch Hand and Comparison groups. In fact, group equivalence was more apparent at the followup examination than at the Baseline examination. The classical effects of age, race, and smoking were demonstrated with most of the hematological variables. The longitudinal analyses also suggested that neither group manifested an impairment of the hematopoietic system. Exposure index analyses did not support a plausible dose-response relationship for any of the hematological variables. RENAL A summary of all renal variables, including unadjusted and adjusted analyses, is displayed in Table 11. A historical assessment of kidney disease/kidney stones by a reviewof-systems questionnaire showed no significant differences between the Ranch Hand and Comparison groups (see Table 12). An adjusted analysis did not alter this conclusion as an adjusted relative risk of 0.95 (95$ C.I.: [0.71,1.25], p-0.693) was demonstrated. These statistics appeared to be'in marked contrast to the Baseline historical findings. Differences vis-a-'vis the Baseline were most likely due to a difference in questionnaire techniques. Current renal function was evaluated by five laboratory variables: urine protein, occult blood, urine, white blood cell counts (WBC's), blood urea nitrogen (BUN), and urine specific gravity. Invasive procedures were not used. The unadjusted analysis of proteinuria showed no group differences (Est. RR: 1.18, 95% C.I.: [0.75,1.86], p-0.485), but the adjusted analysis showed an interaction of group and diabetic class; appropriate stratified analyses revealed that the prevalence of proteinuria was lower in the Ranch Hands than in the Comparisons in the diabetic and impaired strata, but higher in the normal strata for the Ranch Hands. These results were in contrast to the 28 �TABLE 11. Overall Summary Results of Unadjusted and Adjusted Analyses for Renal Variables Variable Unadjusted Adjusted Reported Kidney Disease NS **** Urinary Occult Blood NS **** Urinary Leukocytosis NS **** BUN NS **** Urine Specific Gravity NS*: NS Urinary P r o t e i n NS: NS NS* **** Not s i g n i f i c a n t (p>0.10). Borderline s i g n i f i c a n t (0.05<p<0.10). ****Group-by-covariate interaction. TABLE 12. Unadjusted Analysis of History of Kidney Disease/Kidney Stones by Group History of Kidney Disease/Stones No Yes Group Ranch Hand Number 94 Percent 9.3 Percent Total 920 90.7 1,163 90.1 1,291 Est. Relative Risk (95% C.I.) p-Value 1,014 Number 0.93 (0.70,1.23) 0.619 Comparison 128 9.9 29. �Baseline findings, which showed a marginally significant proteinuria in the Comparison group (p-0.055), and overall, lower prevalence rates of proteinuria. The unadjusted prevalence rates for hematuria were similar for both groups (Est. RR: 1.14, 95$ C.I.: [0.91,1.42], p-0.239). Three significant interactions involving group membership and covariates precluded a direct adjusted comparison of the estimated prevalence rates. Covariate analyses indicated increased hematuria in Blacks and among enlisted personnel. Ultimately via a series of stratified analyses, statistical equivalence was determined for the Black enlisted strata of both groups. Of particular note was the approximate tenfold increase in hematuria in both groups over that observed at Baseline, a finding most likely due to different laboratory techniques (reagent-strip testing versus microscopic observation). Similar results were found for leukocyturia, i.e., a nonsignificant unadjusted analysis (Est. RR: 1.24, 95% C.I.: [0.93,1.64], p-0.145), and a. significant three-way interaction (group, age, race) in the adjusted analysis. Significant covariate effects were noted for diabetic class and occupation for nonblack participants, whereas age was a significant adjusting variable for Blacks. A significant group difference was found only for the younger, nonblack Ranch Hands. The overall results were consistent with the Baseline findings. BUN levels did not vary significantly by group (p-0.554, unadjusted). Adjusted analyses showed significant covariate effects for age and occupation and interactions for group and race and for race and diabetic class. An analysis stratified by race revealed no significant group differences for nonblacks, but a significantly higher adjusted mean BUN level in Black Comparisons than in Black Ranch Hands. Overall, the BUN results were similar to those observed at the Baseline examination. Urine specific gravity levels manifested marginally significant group differences (p-0.082, unadjusted). The adjusted analysis disclosed significant covariate effects of diabetic class and the interactions of group and race and group and occupation. Analyses by race showed no strata with significantly lower mean levels for Ranch hands. In contrast to the Baseline values, the followup urine specific gravities were lower, a finding most likely attributable to differences in laboratory methodology (falling drop method versus multistick procedure). Exposure index relationship at the prevalence rates or within occupational analyses showed very little evidence of a dose-response followup examination. No patterns in the relationship of mean levels among the exposure index levels were seen strata. The longitudinal analysis was based solely upon a contrast of BUN levels between the two examinations. The unadjusted mean BUN value increased slightly from the Baseline to the followup examination, but the increases were symmetrical in the two groups and nonsignificant (p-0.48). In conclusion, none of the six renal assessment variables showed a significant difference between the Ranch Hand and Comparison groups by unadjusted tests. However, in the adjusted analyses, all renal measurements except 30 �reported kidney disease revealed group-by-covariate interactions. These interactions were often complex, making it impossible to reach a firm conclusion as to the presence of an herbicide effect. ENDOCRINOLOGY Questionnaire and review-of-systems data for past thyroid disease were essentially equivalent in both the Ranch Hand and Comparison groups. These historical data were confirmed by medical record reviews. Physical examination findings were necessarily limited to data from palpation of thyroid glands and testicles; the unadjusted results showed no significant group differences. TABLE 13. Medical Record Verification Results of Reported Thyroid Disease by Group Group Verification Status Ranch Hand Comparison Number vith Reported Thyroid Conditions 7 21 Medical Records Reviewed 7 21 Medical Records Pending 0 0 100 100 Percent Thyroid Conditions Verified The physical examination and laboratory testing results of all endocrino-logical variables are summarized in Table HJ. Evaluation of the endocrine system was conducted primarily by laboratory testing of hormone levels. The thyroid test battery consisted of T^ % Uptake and TSH assays. The T3 $ Uptake data showed no group differences for either mean values or frequency of abnormally low or high values. Occupation was a significant covariate. TSH results revealed a significantly higher mean level in the Ranch Hand group, but this difference was not found by categorical testing of proportions of abnormally high TSH results. Mean levels of testosterone were significantly elevated among Ranch Hands as contrasted with Comparisons in the 10 to 25 percent body fat category, but this was not reflected by the categorical tests. For the few participants 31 �TABLE 14. Overall Summary Results of Unadjusted and Adjusted Continuous and Categorical Analyses of Bndocrinological Variables Unadjusted Test Mean Categorical Mean Questionnaire and Physical Examination * Past Thyroid Disease (SelfAdministered) • Past Thyroid Disease (Interviewer Administered) • Thyroid Abnormalities • Testicular Abnormalities Laboratory Testing T, Z Uptake NS TSH Testosterone Initial Cortisol Adjusted Categorical • b • :b NS • b NS • b NS NS* Overall: NS Lov vs. Normal: NS High vs. Normal: NS 0.019 NS 0.03S Overall: NS Lov vs. Normal: NS High vs. Normal: NS NS Overall: NS Lov vs. Normal: NS High vs. Normal: NS NS NS « NS —b Overall: NS Lov vs. Normal: NS High vs. Normal: NS 0.025 NS **** Overall: NS Lov vs. Normal: NS High vs. Normal: NS NS b b 2-Hour Cortisol * * * * —* Differential Cortisol Overall: 0.034 2-Hour Postprandial NS Overall: 0.038 NS Glucose Impaired vs. Normal: 0.024 Impaired vs. Normal: 0.022 Diabetic vs. Normal: NS Diabetic vs. Normal: NS • NS Diabetes (Composite NS Indicator) —'Analysis not feasible. NS: Not significant (p>0.10). — Analysis not performed. NS*: Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. 32 �with less than 10 percent body fat (six Ranch Hands, four Comparisons), mean testosterone levels were lower for Ranch hands than for Comparisons. Age, occupation, and percent body fat were significant adjusting variables. Two timed cortisol specimens showed no significant group differences in mean values and percent abnormalities. The difference between the timed cortisol results, termed the differential cortisol, showed no significant group differences for nonblacks or Blacks born before 1942, but Black Ranch Hands born in or after 1942 had a lower mean differential cortisol level than Comparisons. Age, percent body fat, and personality type were significant . covariates in these analyses. Group means of 2-hpur postprandial glucose levels were not statistically different, but categorical testing revealed that there was a significantly higher frequency of glucose-impaired (at least 140 but less than 200 mg/dl) Comparisons than Ranch Hands. A constructed variable comprised of known diabetics and individuals classified as diabetic by the glucose tolerance test, showed no difference between the Ranch Hand and Comparison groups. As expected, past and current diabetes were highly influenced by the covariates age, race, and percent body fat. Exposure index analyses did not reveal any pattern consistent with a dose^response relationship. Enlisted flyers in the medium exposure level were significantly different from those in the low exposure level for 2-hour cortisol, differential cortisol, and 2«hour postprandial glucose. However, the corresponding high versus low contrasts were not statistically significant. Longitudinal analyses of T, % Uptake, TSH, and testosterone levels on all individuals attending both the Baseline and followup examinations revealed only symmetrical and nonsignificant changes in the Ranch Hand and Comparison groups in the interval between examinations. In conclusion, both limited historical and physical examination data, seven endocrinological laboratory variables, and a composite indicator of diabetes did not demonstrate consistent patterns indicating an herbicide effect. However, there was a significant interaction between group and percent body fat for testosterone that could be interpreted as an herbicide effect. TSH and testosterone means tests were statistically significant, and in the expected direction of an herbicide effect, but these results were not confirmed by categorical testing. Also significant was the impaired category of the glucose tolerance test, which showed an excess in the Comparison group. The consistent demonstration of the classical effects of the covariates age, ' race, occupation, and percent body fat on appropriate endocrine variables provided support for these conclusions. Overall, the endocrine health status of both groups was reasonably comparable. IMMUNOLOGY Immunologic competence was measured by cell surface marker (phenotypic) studies and cell stimulation studies on 47 percent of the study population, and by a four antigen series of skin tests in 76 percent of participants to assess the delayed hypersensitivity response. Table 15 summarizes the results of all unadjusted and adjusted analyses on 11 primary variables spanning the first two of these three functional areas. 33 �TABLE 15. Overall Suaaary Results of Unadjusted and Adjusted Analyses of Isaunologlcal Variables Variable Unadjusted Adjusted Total T Cells (T ) NS A•^ ^ WwWn Helper T Cells ( , T) Suppressor T Cells ( , T) B Cells Monocytes HLA-DR Cells T4/T, Ratio NS NS NS NS NS NS NS NS **** **** **** NS Unst initiated Response (PHA) PHA Net Response Pokeveed Net Response HLC Net Response NS NS NS NS NS NS NS **** NS:Not significant (p>0.10). ****Significant group-by-covariate interaction. Cell surface marker studies were conducted for total T cells (T^), helper T cells (T^), suppressor T cells (Tg), B cells, monocytes, and HLA-DR cells; the ratio of TVTg cells was included in the analysis. Because of inherent significant day«to»day and batch*to-batch variation, all results (including functional stimulation studies) were adjusted for blood-draw day variation. Statistical testing of the seven phenotypic cell markers did not reveal any significant group differences (interactions excepted), either unadjusted or adjusted for the covariates of age, race, occupation, current smoking, lifetime smoking history (pack^-years), current alcohol use, or lifetime alcohol use (drink-years). Similarly, none of the unadjusted or adjusted analyses of the functional stimulation studies (for phytohemagglutinin, pokeweed mitogen, or mixed lymphocyte culture) showed any statistically significant group differences. However, the adjusted analyses for total T cells, B cells, monocytes, HLA^DR cells, pokeweed mitogen, and net mixed lymphocyte culture stimulation showed some significant group-by-covariate interactions, precluding direct adjusted group contrasts. Overall, no discernible pattern was identified to suggest a detriment in any subgroup of either the Ranch Hands or Comparisons. Results were similar between the analyses of the total Comparison group and'the analyses of the Original Comparisons. The covariate effects of age, race, smoking, and alcohol use were generally profound on most variables in the phenotypic and stimulation studies. Consistently decreasing values of all cell markers and stimulated cells were associated with increasing age, whereas Increased levels of smoking were usually associated with increases in the values of those variables. Blacks �had consistently higher stimulated cell counts than nonblacks, but this effect was not observed for counts of T cells, B cells, or HLA-DR cells. Enlisted personnel generally had higher cell surface marker counts than officers. Exposure index analyses of cell surface markers revealed no pattern consistent with a dose-response relationship. For enlisted groundcrew, the mean total T cell and suppressor T cell counts for the medium exposure level were significantly lower than those of the low exposure level, but were slightly lower than those of the high exposure level. The exposure index analyses of the functional stimulation tests revealed no consistent significant dose;-response patterns for net PHA counts or net MLC counts. For net pokeweed counts, enlisted flyers in the high exposure level had a significantly lower adjusted count than enlisted flyers in the low exposure level, and a decreasing trend was apparent. The delayed hypersensitivity response was assessed by the skin test antigens of mumps, Candida alb leans, Trichophyton, and staph-'phage-lysate. The JtSrhour measurements of skin induration and erythema for the four test's showed marked inter-reader variation. Analyses showed that one of the three skin test readers too often measured induration larger than erythema (a clinically unacceptable finding), in an average of 30 percent of the readings, and did not yield measurements that detected a case of possible or overt anergy, whereas the other two readers found this condition in 5.6 percent of the participants. Remaining data from Readers 1 and 3, however, were found to vary significantly in clinical interpretation over duration of the examination. Consequently, all skin test data were declared invalid, and were not used'in the assessment of group differences. The skin test reading problems led to the use of additional clinical quality control procedures for the AFHS followup examination begun in May 1987. In conclusion, no significant group differences were judged present for the comprehensive cell surface marker or functional stimulation studies. The profound effects of age, smoking, and alcohol use were observed in these immunologic tests. The assessment of delayed hypersensitivity skin responses was precluded by poor data quality and excluded from further analysis. Overall, there was no indication of impaired immunologic competence in either group. RESPIRATORY SYSTEM A summary of the results on the analyses of reported history of respiratory illness and of radiological and clinical findings is given in Table 16. Based on the 31 December 1986 mortality data, there were seven deaths from respiratory conditions in the Comparison group and none in the Ranch Hand group. 35 �TABLE 16. Overall Summary Results of Unadjusted and Adjusted Analyses of Pulmonary Disease Pulmonary Disease Unadjusted Adjusted NS NS NS NS NS NS NS Reported History of Respiratory Illness Asthma Bronchitis Pleurisy Pneumonia Tuberculosis **** NS **** Radiological and Clinical Findings Thorax and Lungs Asymmetrical Expiration Hyperresonance Dullness Wheezes Rales X Ray NS NS NS NS NS NS NS NS NS NS NS NS **** NS NS: Not significant (p>0.10) ****Group-by-covariate interaction. There were no group differences found for reported history of asthma bronchitis, pleurisy, or tuberculosis based on the unadjusted analyses. Adjustments for age and lifetime smoking did not alter the findings of group similarity, although there was a significant group-bypack-year interaction for pleurisy and for tuberculosis. Similarly, there were no significant group differences in the unadjusted analyses for the radiological and clinical respiratory findings of thorax and lungs, asymmetrical expiration, hyperresonance, dullness, wheezes, rales and xrray interpretations. These findings were supported by the adjusted analyses, although there was a group-by-age interaction for rales. The exposure index analyses revealed no consistent dose^response pattern. Analyses of past history of respiratory illness and the clinical and radiological examination of the chest and lungs did not reveal any statistically significant differences between the Ranch Hand and Comparison groups suggestive of herbicide related disease. Several group^byrcovariate interactions did exhibit statistical significance, but these findings did not indicate any consistent patterns suggesting different disease experience in the two groups. 36 �CONCLUSIONS This chapter summarizes the conclusions drawn from the statistical analyses that have been conducted on the Air Force Health Study data base. The followup study, which began in 1985, was the logical extension of the 1982 Baseline, building upon the strengths of the Baseline study and utilizing the data collected at both the Baseline and the followup. The high level of Government support and outstanding participation of the study subjects that characterized the Baseline study were maintained through this first followup. STUDY PERFORMANCE ASPECTS Of the living Baseline study participants, 99.2 percent were located and asked to participate in the followup. Participation in the followup physical examination and questionnaire was very high. Of the fully compliant Baseline participants, 971 of the 1,045 Ranch Hands (92.9$) and 1,139 of the 1,224 . Comparisons (93.1$).participated in the followup. Thus, there was no group difference in compliance of the Baseline participants at the followup. Overall, the 2,309 participants in the followup (1,016 Ranch Hands and 1,293 Comparisons) represented a loss of 159 individuals and .a gain of 199 since Baseline. One percent of the fully compliant Baseline population died between 1982 and the 1985 followup examination. The bias/compliance analyses suggested that there had been no change between Baseline and the followup in the way replacements volunteered for entry into the study, and that no additional bias had been introduced at the followup due to scheduling differences. Although replacements were not health-matched at Baseline as they were at the followup, they were similar to refusals with respect to reported health, medication use, and income level. This result supported the conclusion that there has been little, if any, selection bias due to differential participation in the Comparison group and supported the use of the total Comparison group in the main analyses presented in this report. POPULATION CHARACTERISTICS Overall, the Ranch Hands and Comparisons reported similar social and behavioral characteristics. No significant differences were found in age, educational background, religious preference, current military status, and income level. Significantly more Ranch Hands smoked cigarettes at the time of the followup'examination than did Comparisons, but there was no significant difference between groups on past cigarette, cigar, and pipe use and on recent^ and past use of marijuana. A much higher percentage of participants reported past marijuana use at the followup than at Baseline. This difference was most likely due to a greater level of confidentiality afforded by the questionnaire technique. Risk taking behavior, assessed by questions on potentially dangerous recreational activities, revealed borderline significance. Slightly more Comparisons were scuba divers and more Ranch Hands raced motor vehicles. The difference in scuba diving was also significant at Baseline. 37 �Patterns of Results Both the chapter conclusions and the final conclusions of this report have been predicated upon concepts of consistency, specificity, coherence, strength, and plausibility as they apply to the interpretation of group differences. In particular, careful consideration has been given to a variety of data and patterns of results that have emerged from the clinical evaluations. Specifically, there were few differences in the proportions of abnormalities between groups; the positive associations have not aggregated in the clinical areas of prime dioxin concern, nor have they been of serious clinical importance; the unadjusted results have been remarkably concordant with the adjusted results, both in. terms of relative risk and p value; the analyses using the Original Comparison set have largely mirrored the results found with the total Comparison group; many of the group differences noted at Baseline have disappeared at the followup examination, and only a few new associations have emerged; almost all of the covariates have acted as expected in the adjusted analyses; and the exposure index analyses and the group-by-covariate interactions have not demonstrated biological patterns of concern and appeared to'be more likely due to chance than not. Due to the acknowledged limitations of the exposure index used in this report (and considering the potential use of dioxin body burden levels at the next followup), dose-response relationships have not been emphasized in reaching final conclusions. The overall pattern of these findings indicates that this followup study cannot be viewed as alarming from the traditional perspectives of clinical medicine or epidemiology. This study, in fact, demonstrates similarity in current health status between the Ranch Hand and Comparison groups. CLINICAL ASPECTS General Health The nonspecific assessment of general health showed relatively close similarity between the two groups. Ranch Hands rated their health as fair or poor more frequently, but this difference was found only in the enlisted groundcrew and not in the officers nor enlisted flyers. The perception of health in both groups had improved since Baseline. Physician-rated appearance of relative age was not found to be significantly different at the followup in contrast to the Baseline finding that a higher percent of Ranch Hands than Comparisons looked younger than their stated age. The categorical analysis of sedimentation rate showed that the Ranch Hands had more abnormalities than the Comparisons. These results were not supported by the continuous analysis of mean sedimentation rates and were opposite to the Baseline results, which showed that younger Comparisons had elevated sedimentation rates. The categorical analysis of percent body fat showed no significant differences between the two groups, which was consistent with Baseline. However, the continuous analysis found that the Ranch Hands had a significantly lower mean percent body fat using age, race, and occupation as covariates. The detailed exposure analyses revealed no consistent exposure effects, and this result was consistent with the Baseline analysis. No longitudinal difference was found on perception of health. A significant group difference was found over time for the longitudinal analysis of sedimentation rate due to the change in the findings between the two examinations, possibly related to a change in laboratory methodology. 38 �Malignancy Skin and systemic cancers, both suspected and verified by medical records, showed no significant group differences for the Baselinerfollowup interval (1982-1985). However, for all neoplasms combined (malignant, benign, and uncertain), a borderline significant excess in the Ranch Hand group was noted in an unadjusted analysis. The analyses of interval cancers revealed group interactions for verified and verified plus suspected basal cell carcinoma and verified plus suspected systemic cancers. Nonsignificant findings were observed for verified and verified plus suspected sun exposure-related cancers. Verified systemic cancers did not differ significantly between groups. The analyses of lifetime cancer found significant results for verified basal cell carcinoma and verified sun exposures-related skin cancers. Group interactions were noted for systemic cancer categories and for verified plus suspected basal cell carcinoma. The higher rate of basal cell carcinoma in the -Ranch Hands versus the Comparisons found at Baseline was nonsignificant for the followup interval, but due to the effect of the larger number of Baseline cases and the significant confounding of average residential latitude, the adjusted analysis of lifetime basal cell carcinoma emerged as statistically significant. There were several disparities in the distribution of testicular, colon, and smoking-related tumors in the groups. Further, one case of soft tissue sarcoma and one possible lymphoma (both in Ranch Hands) were diagnosed in the interval, balancing the two similar cases found in the Comparison group at Baseline. Considering that the systemic cancer curves are in their early stages for both groups, with perhaps insufficient latency, the cancer results of the followup examination should not be viewed as disturbing, but as cause for continued monitoring. Neurological Assessment None of the 27 neurological variables demonstrated a significant group difference, although several variables had relative risks which were greater than one. There was no group difference in reported neurological illnesses for the interval or for a lifetime history. Of the cranial nerve variables, speech and tongue position were marginally significant, with the Ranch Hands at a slight detriment. The analyses of peripheral nerve function showed no significant differences between the Ranch Hands and the Comparisons. In the analysis of central nervous system function, hand tremor was found to be of borderline significance, with the Ranch Hands faring slightly worse than the Comparisons. A borderline significant group interaction (Ranch Hand hand tremor by insecticide exposure) may have had biological and operational significance. Overall, substantially fewer neurological abnormalities were detected at the followup examination than at the Baseline examination. The exposure analyses showed only occasional statistically significant results, although no consistent pattern with increasing exposure was evident. In the longitudinal analysis of the Babinski reflex, a significant change over time was observed. This was due to a nonsignificant finding in the Ranch Hands at the followup, which differed from the significant adverse finding at Baseline. The covariates of age, alcohol history, and diabetes showed classical effects with many of the neurological measurements. Overall, the followup examination results were quite similar to the Baseline findings. 39 �Psychological Assessment The reported and verified data on lifetime psychological illnesses showed no significant differences between groups. Distributional tests of the 14 Minnesota Multiphasic Personality Inventory (MMPI) scales, stratified by occupation, revealed that only 2 of the 42 results approached significance. For the total Cornell Medical Index (CMI), separate distributional tests were conducted with stratification by age, race, occupation, education, and current drinking status; a significant difference was found for one statum of each of the covariates. In all cases, the mean of the Ranch Hand distribution was greater than the mean of the Comparisons. The analysis of the 14 MMPI scales showed that there was a significant difference between the two groups for denial and masculinity/femininity, with more abnormalities in the Comparisons than the Ranch Hands. The results of the analyses for hysteria were of borderline significance, with more abnormalities in the Ranch Hands. There were more abnormalities in the Ranch Hands than the Comparisons for social introversion, which was of borderline significance. Differences in the to.tal CMI 'and A.-H area subscore were found to be significant, with more abnormalities in the Ranch Hands. There was no significant difference between the two groups on the Halstead-^Reitan Battery impairment index, a measure of the functional integrity of the CNS. The exposure index analyses did not reveal any pattern consistent with a dose-response relationship. As expected, the effects of age, educational level, and alcoholic history showed profound effects on many of the psychological measurements. Gastrointestinal Assessment Although the followup gastrointestinal assessment disclosed more statistically significant findings than the Baseline examination, the abnormalities were distributed equally between the two groups, and there was no clinical, statistical, or exposure pattern consistent with an herbicide-related effect on health. No historical or biochemical evidence was found to suggest an increased likelihood of porphyria cutanea tarda (PCT) in the Ranch Hand group. Only sparse and nonsignificant liver disorders were reported for the interval between Baseline and followup. Also, for the lifetime history of liver disorders, there were no significant differences between groups. Further, there were no significant group differences in reported lifetime peptic ulcer disease. A review of digestive system mortality showed a relative excess in the Ranch Hands but a relative lack of malignant neoplasms. The results of the physical examination showed a borderline increase'of hepatomegaly in the Ranch Hand group. There was a significantly lower mean serum glutamic-pyruvio transminase (SGPT) level, a greater mean alkaline phosphatase level, and a lower mean uroporphyrin level in the Ranch Hand group. The analysis of coproporphyrin was of borderline significance, with the mean of the Ranch Hands in excess of the mean of the Comparisons. No group differences were found for serum glutamic-oxaloacetic transminase (SCOT), gamma^glutamyl transpeptidase (GGTP), total and direct bilirubin, lactic dehydrogenase (LDH), cholesterol, or triglycerides. The numerous grouprbycovariate interactions did not disclose any consistent subgroup patterns detrimental to the Ranch Hands. These findings were generally consistent with the results of the 1982 assessment. The longitudinal analyses for SCOT, SGPT, and GGTP showed no significant differences between results by group over time. 40 �Dermatologies! Evaluation No significant group differences were identified in the dermatological evaluation. None of the questionnaire data showed an increased likelihood of past chloracne, as determined by anatomic patterns of acne, and no cases were diagnosed in the physical examination. Analyses were conducted on six dermatologic disorders (comedones, acneiform lesions, acneiform scars, inclusion cysts, depigmentation, and hyperplgmentation) and on a composite variable of 16 other minor conditions (the latter not generally associated with chloracne). Exposure index analyses did not reveal consistent patterns suggestive of a'dose-response relationship. The longitudinal analysis, based on a composite dermatology index, showed no significant differences between the results over time. Substantially more dermatologic abnormalities were detected at the followup examination than at the Baseline examination. In general, however, the followup results were consistent with the findings at Baseline. Cardiovascular Evaluation Overall there was remarkable similarity in the cardiovascular health of the Ranch Hands and the Comparisons. Of the 27 cardiovascular variables, there was a significant difference for only one, verified heart disease, with an excess in the Ranch Hand group. This finding was largely unsupported by other cardiac measurements. The cardiovascular assessment was based on reported and verified heart disease; the measurement of central cardiac function by systolic blood pressure, abnormal heart sounds, and EGG findings; and the evaluation of peripheral vascular function by diastolic blood pressure, funduscopic examination, presence of carotid bruits, and detailed manual and Doppler measurements of five peripheral pulses. Doppler recordings of five peripheral pulses were similar in both groups,, a finding which was in marked contrast to the Baseline examination that found significant pulse deficits in the Ranch Hand group. This change was most likely due to a required 4-hour abstinence from tobacco prior to the pulse measurements. Overall, the exposure analyses were unsupportive of any meaningful doseiresponse relationship. The longitudinal analyses confirmed the change in pulse abnormalities in the Ranch Hand group over time, but showed no significant group change in overall EGG findings between the examinations. Hematological Evaluation The hematological evaluation found that neither group manifested an impairment of the hematopoietic system, consistent with similar findings at the Baseline. The evaluation was based on eight peripheral blood variables: red blood cells (RBC), white blood cells (WBC), hemoglobin (HGB), hematocrit concentration (HCT), corpuscular volume (MCV), corpuscular hemoglobin (MCH), corpuscular hemoglobin concentration (MCHC), and platelet count (PLT). Both the discrete and categorical analyses revealed no significant group differences. The covariate effects of age, race, occupation, and smoking history were highly significant for many of the variables. Two group-iby-covariate interactions in the analyses of mean differences did not appear to have a meaningful interpretation. The exposure index analyses did not support any plausible dose'-response relationship. The longitudinal �analyses of MCV, MCH, and PLT found significant differences only for PLT between the Baseline and the followup, with the Ranch Hands exhibiting a slight decline in mean level from Baseline and the Comparisons showing an opposite change. Renal Assessment None of the six renal variables of reported kidney disease, urine protein, occult blood, urine white blood cell count, blood urea nitrogen, and urine specific gravity showed a significant difference between the two groups based on the unadjusted analyses. In the adjusted analyses of the laboratory variables, however, there were significant group-by-covariate interactions that did not yield a consistent pattern to suggest a renal detriment to either group. The finding of group equivalence for past kidney disease was in contrast to the Baseline examination, which found significantly more reported disease in the Ranch Hand group. The difference in findings is more likely due to a change in questionnaire wording than to a true change in renal health. Like the Baseline findings, the exposure index analyses showed very little evidence of a dose-response relationship. In the longitudinal analyses of blood urea nitrogen, there were no significant changes between the examinations by group. Endocrine Assessment In general, the endocrine health status of the Ranch Hands and the Comparisons was reasonably comparable. The examination found no significant differences between the two groups for past thyroid disease, or thyroid and testicular abnormalities determined by palpation. In the analyses of the seven laboratory values (To % Uptake; thyroid stimulating hormone (TSH); testosterone; initial, second, and differential cortisol; and postprandial glucose), significant differences were found for TSH and testosterone, with higher mean levels in the Ranch Hands. These analyses were not supported by the categorical analyses. The thyroid test results were conflicting with respect to an assertion of hypothyroidism in the Ranch Hands (a possible dioxin effect). Mean levels of testosterone were significantly elevated in the Ranch Hand group as contrasted with the Comparisons in the 10-25 percent body fat category. The effects of personality score and percent body fat on the differential cortisol levels were not fully expected. Although tests of 2-hour postprandial mean values showed no significant group differences, comparable categorical tests revealed that significantly fewer Ranch Hands had impaired glucose levels, but conversely, had more (nonsignificant) diabetic levels of glucose. Analyses of the composite diabetes indicator (history plus 2*hour postprandial results) did not disclose significant group differences. The exposure index analyses suggested that the enlisted flyers in the medium exposure level were significantly different from those in the low exposure level for differential cortisol, postprandial glucose, and testosterone. The corresponding high to low contrasts were not significant. The longitudinal analyses were based on T, % Uptake, TSH, and testosterone, and revealed only symmetrical and nonsignificant changes in the Ranch Hand and Comparison groups over the time interval. �Immunological Evaluation Overall, there were no significant group differences or any indication of impaired immunological competence in either group based on comprehensive cell surface marker and functional stimulation studies. Six cell surface markers (total T cells, helper T cells, suppressor T cells, B cells, monocytes, HLA-DR cells, and a constructed helper/suppressor ratio variable) and three functional stimulation studies (PHA, pokeweed, and mixed lymphocyte culture) were conducted on 4? percent of the study population. No significant differences were revealed for five of these variables, in the analyses of the other five variables, there were significant group-by^covariate interactions, but no discernible pattern was identified to suggest a detriment in any subgroup of either group. Skin test assessments of delayed hypersensitivity were characterized by interpleader variation and shifting diagnostic criteria for anergy. The skin test data were judged invalid and were not subjected to statistical testing for group differences. No consistent pattern of immunological deficits could be associated with increasing levels of herbicide exposure in the Ranch Hand group. Pulmonary Disease The pulmonary assessment did not reveal any statistically significant differences between the Ranch Hand and Comparison groups that were suggestive of an herbicide-related disease. The analyses consisted of group assessments of respiratory disease incidence, physical examination abnormalities, and the current prevalence of x-ray abnormalities. There were no significant differences between the Ranch Hands and Comparisons for history of asthma, bronchitis, pneumonia, or for six of seven clinical variables (excluding rales) determined by x-ray or auscultation. Analyses of history of pleurisy, history of tuberculosis, and rales showed significant but inconsistent group-by-covariate interactions. These findings did not indicate any patterns suggesting a different disease experience in the two groups. The exposure index analyses did not reveal any consistent pattern suggestive of an increasing dose response. CONCLUSION The results of the first followup study in 1985 have shown a subtle but consistent narrowing of medical differences between the Ranch Hands and Comparisons since the Baseline Study in 1982. The 1985 examination results provide reassuring evidence that the current state of health of the Ranch Hand participants is unrelated to herbicide exposure in Vietnam. Continued close medical surveillance of these military populations is strongly indicated. This followup report concludes that there is not sufficient plausible or consistent scientific evidence at this time to implicate a causal relationship between herbicide exposure and adverse health in the Ranch Hand group. FUTURE DIRECTIONS The scope and complexity of the AFHS has required gradual refinement and correction to meet the challenges of changing technology and scientific direction, and to ensure continued participation of all enrolled members. �This chapter outlines some of the changes incorporated in the fifths-year followup examination and identifies several areas of future work expected to significantly augment the study. FIFTHr-YEAR FOLLOWUP EXAMINATION Since the fifth-year followup examination was initiated prior to the full analysis of the data from the third-year examination, most modifications were founded upon quality control issues and the desire to make the clinical content of the examination more responsive to the medical needs of the participants. Clinical quality control enhancements were made to improve measurement techniques. The digit preference noted in systolic and diastolic blood pressure readings led to the use of automated blood pressure recording; all other parameters of the blood pressure readings (e.g., sitting position, three recordings, nondominant arm a.t heart level) were not changed. The problem in skin test reading was met by a rigorous quality control plan that included the following elements: refresher training for readers; a required reading of the four skin tests of all participants by both readers, each blind to the results of the other; a required reread of 10 percent of all tests by each of the readers, each blind to the previous reading; and a required weekly report citing numbers and proportions of participants with possible anergy, reversal of induration"erythema measurements, and untoward skin reactions or other reading problems (e.g., participant refusal). In addition, new skin test forms were developed to facilitate accurate recording and transcription; specific clinical criteria were formulated to require consultation by an allergist; and the skin test measurement criterion for possible anergy, consistent with current World Health Organization guidelines, was adopted for the clinical interpretation of all skin test readings. It is anticipated that this clinical quality control program will standardize both readings and interpretations, and will produce a uniformly superior data set. EXPOSURE INDEX REFINEMENTS Since the development of the Study Protocol and the analysis of the 1982 Baseline data, there has been concern among some scientists and the principal investigators over the accuracy and validity of the exposure estimates. It is unclear whether statistically significant differences in some variables between the Ranch Hand and Comparison groups, unsupported by dose?-response estimates, have been due to chance, or whether true differences are obscured by an inadequate exposure index or group misclassification. In mid-1986, strong correlations between dioxin levels in fat tissue and serum were demonstrated by the CDC and other institutions. Because of these results, the Air Force is currently engaged in a collaborative study with CDC to determine whether serum dioxin levels vary significantly in the Ranch Hand population. Approximately 200 AFHS volunteers have supplied a pint of blood to be analyzed for dioxin at the CDC laboratories. If clear and meaningful �exposure findings are evident from this study, several additional studies are feasible: testing can be expanded to the entire study population and a meaningful exposure index based on total current TCDD body burden may be developed; and by means of archived AFHS serum samples from the Baseline study, it may be possible to calculate a reasonably precise half-life of TCDD in humans. These expanded studies will allow the estimation of body burdens of TCDD at the time of departure from SEA (assuming the absence of intervening vocational and recreational exposures). If, in fact, these potential studies become reality within the next 2 years, the fifths-year followup study data will be statistically analyzed using a more appropriate exposure index. In anticipation of this advance, the AFHS is currently collecting 280-^350 ml of blood from all volunteers attending the fifth-year followup study. ADDITIONAL ANALYSES AND STUDIES As in the 1 98M Baseline Report, not all of the measured dependent variables were subjected to statistical analysis (e.g., prothrombin, leutinizing hormone, follicle stimulating hormone), largely because they were not within the bounds of the Air Force^prescribed analyses. Exploration of many of the unanalyzed variables is contemplated as time and resources permit. Similarly, many analytic opportunities to define possible symptom-clinical sign clusters or syndromes by multivariate analysis of variance techniques were passed over due to time and charter. Particularly challenging as an area of future work may be the changing relationships of some immunological variables over time and the biological impact of these changes on the induction of diseases such as cancer. Likewise, future efforts to define shifting cardiovascular disease patterns are a logical extension of the rich longitudinal data base of the AFHS. Such efforts await future analysis and publication. The assessment of possible selection and participation bias has been addressed in a comprehensive manner in this report (see Chapter 5). The analyses and discussion suggest that statistical use of the total Comparison group (versus the Original Comparison group) is justified in this report, and that the impact of selection and participation biases have been minimal. As the followup studies continue, it is anticipated that a wealth of data on compliance-participation factors will be available for continued comprehensive bias analyses. In particular, it is hoped that more complete data will exist to examine the true differences in current health status between refusals and their replacements. As the data set grows over time, the bias analyses will become more complex and will have to deal with changing motivations of the participants to continue in this study. Such bias analyses and assessments will always be of great importance to this study as they ultimately set the bounds for an inference on herbicide causality. � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 059 Folder The folder containing the original item. 1584 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Lathrop, George D. William H. Wolfe Joel E. Michalek Judson C. Miner Michael R. Peterson Stella G. Machado Theodore G. Karrison William D. Grubbs Wanda F. Thomas Date A point or period of time associated with an event in the lifecycle of the resource 1987-10-01 Title A name given to the resource Air Force Health Study: An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides, Summary, First Followup Examination Results, January 1985-September 1987 Subject The topic of the resource Air Force Health Study morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/281d8a128a8712e513d8bba495ef4193.pdf 2917c78f2f135bda34d4006615f46b59 PDF Text Text Item ID Number 01535 Author Lathrop, George D. Corporate Author Report/Article TltlB Air Force Health Study: An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides, Volume I, First Follow/up Examination Results, January 1985-September 1987 Journal/Book Title Year 1987 Month/Day October Color n Number of Images 632 nnsnrlntnn NntHg * Contract no. F41689-85-D-0010 and SAIC Project no. 2816-XX-195/254-XX . Wednesday, May 23,2001 Page 1586 of 1608 �Air Force Health Study An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides SAIC Team Air Force Team George D. Lathrop, M.D., M.P.H., Ph.D. Stella G. Machado, Ph.D. Theodore G. Karrison, Ph.D. William D. Grubbs, Ph.D. Wanda F. Thomas, M.S. COL William H. Wolfe, M.D., M.P.H. Joel E. Michalek, Ph.D. LTC Judson C. Miner, D.V.M., M.P.H. LTC Michael R. Peterson, D.V.M., M.P.H., Dr.P.H. Project Manager: Program Manager: R.W. Ogershok W.F. Thomas SCIENCE APPLICATIONS INTERNATIONAL CORPORATION 8400 Westoark Drive McLean, Virginia 22102 EPIDEMIOLOGY DIVISION USAF School of Aerospace Medicine Human Systems Division (AFSC) Brooks Air Force Base, Texas 78235 October 1987 VOLUME I First Followup Examination Results January 1985 to September 1987 Contract Number F41689-85-D-0010 SAIC Project Number: 2-816-XX-195/254-XX (Distribution Unlimited) �SECURITY llAttif ICATI6N OE TMiTPAGT Form Approved OMB No. 0704-0188 REPORT DOCUMENTATION PAGE 1a. REPORT SECURITY CLASSIFICATION 1b. RESTRICTIVE MARKINGS Unclassified 3 . DISTRIBUTION/AVAILABILITY OF-REPORT . SECURITY CLASSIFICATION AUTHORITY 2b. DECLASSIFICATION/DOWNGRADING SCHEDULE Approved for public release; distribution unlimited 4. PERFORMING ORGANIZATION REPORT NUMBER{S) 5. MONITORING ORGANIZATION REPORT NUMBER(S) 6*. NAME OF PERFORMING ORGANIZATION 7a. NAME OF MONITORING ORGANIZATION Science Applications International Corp. Life Sciences & Systems Dept 6b, OFFICE SYMBOL (If applicable) Human Systems Division 6c ADDRESS (C/ty, Start, and ZIP Cock) (HSD) 7b. ADDRESS (C/ty, State, and ZIP Code) McLean. Virginia 22102 Brocks Air Force Base. Texas 78235-5000 8a. NAME OF FUNDING/SPONSORING ORGANIZATION 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER 8b. OFFICE SYMBOL (If applicable) F41689-S5-D-0010 Be. ADDRESS (C/ty, Staf«, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS PROGRAM ELEMENT NO. PROJECT NO. TASK NO WORK UNIT ACCESSION NO. 6530F 0003 2767 1. TITLE (Include Security Classification) An EpidearLologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides. First Follcwup Examination Results 2. PERSONAL AUTHOR(S) G.D. Lathrop. SAIC; W.H. Wolfe, USAF; S.G. Machado, SAIC; J.E. MLchalek, USAF; T.G. Karri con. TT. of C: J.C. MLt»r. USAF: W.D. Grtibbs. SAICi M.R. Peterson. USAFt W.F. Thntiaa. SATH. a. TYPE OF REPORT 15. PAGE COUNT !4. DATE OF REPORT (Year, Month. Day) 13b. TIME COVERED J Interim 1982-1985 FROM 1/85 TO 9/87 1987 Oct 1 16 11 r 16. SUPPLEMENTARY NOTATION 17. COSATI CODES FIELD 06 GROUP 1 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number) EpidearLologic Investigation Fnenoxy Herbicides Herbicide Orange SUB-GROUP 05 Dicodn Ranch Hand Air Force Health Study Morbidity 19. ABSTRACT (Continue on reverse if necessary and identify by block number) Bos report presents the results of the health assessment of the 1,016 Ranch Hands and the 1,293 Comparisons who participated in the 1985 f ollowup examination of the Air Force Health Study. The purpose of the study is to determine whether Iccg-tenn health effects exist and can be attributed to occupational exposure to herbicides. The result showed a subtle but consistent narrowing of medical differences between the two groups since the Baseline study in 1982; however, the Ranch Hands continue to manifest slightly more minor adverse health conditions than the Comparisons. Continued surveillance of these two groups is indicated. The report concludes that there is not sufficient evidence to implicate a causal relationship between herbicide exposure and adverse health in the Ranch Hand group. 20. DISTRIBUTION / A VAll ABILITY OF ABSTRACT 0 UNCLASSIFIED-UN'tiMlTEO 21 ABSTRACT SECURITY CLASSIFICATION D SAME AS RPT. Q 01-. NAME OF RESPONSIBLE INDIVIDUAL 'OL WILLIAM H. WOLFE, USAF, MC 00 Form 1473, JUN 86 ;..»$ 22b TELEPHONCr/nc/ude Area Code) 512-530-2604 Previous edir i>f obsolete 22c. OFFICE SYMBOL USAFSAM/EK SECURITY CLASSIFICATION Of THIS PAGE �EXECUTIVE SUMMARY FIRST FOLLOWUP MORBIDITY STUDY The Air Force Health Study is an epidemiological study conducted to determine whether adverse health effects exist and can be attributed to occupational exposure to Herbicide Orange. The study consists of mortality and morbidity components, based on a matched cohort design in a nonconcurrent prospective setting with followup studies. The Baseline study was conducted in 1982, and the first followup morbidity study was performed in 1985. The purpose of this report is to present the results of the first followup study. In the Baseline morbidity study, each living Ranch Hand was matched to the first living and compliant member of a randomly selected Comparison mortality set based on age, race, and military occupation, producing an approximate 1:1 contrast. The Comparisons had served in numerous flying organizations that transported cargo to, from, and within Vietnam but were not involved in the aerial spray operations of Herbicide Orange. Recruitment for the first followup was in accordance with the Study Protocol: All previous participants and refusals, newly located study members, and replacements (matched to noncompliant Comparisons on self-perception of health) were invited. Of the living Baseline study participants, 99.2 percent were contacted to enroll in the followup on a strictly voluntary basis. Participation was very high, with 93 percent of both the Ranch Hands and the Comparisons fully compliant at Baseline also participating in the followup. Overall, the 2,309 followup participants (1,016 Ranch Hands and 1,293 Comparisons) represented a loss to the study of 159 individuals but a gain of 199 new participants since Baseline. Statistical analyses of selection and participation bias supported the use of the total Comparison group for the main analyses presented in this report. The followup study was conducted under contract to the Air Force by Science Applications International Corporation, in conjunction with the Scripps Clinic and Research Foundation and the National Opinion Research Center. Most of the data were collected through face-to-face interviews and physical examinations conducted at the Scripps Clinic in La Jolla, California. Other data sources included medical and military records and the 1982 Baseline data base. As a contract requirement, all data collection personnel were blind to exposure status, and all phases of the study were monitored by stringent quality control. The statistical analyses were based on analysis of variance and covariance, chi-square tests, Fisher's exact tests, general linear models, Kolmogorov-Smirnov tests, logistic regression, proportional odds models, t-tests, and log-linear models. The questionnaire and physical examination data were analyzed by major organ system. The primary focus was on the assessment of differences between the Ranch Hand and Comparison groups based on data from the first followup. Additionally, dose-response relationships within the Ranch Hand group were examined, and longitudinal assessments of differences in the changes of the two groups between the examinations were conducted for selected variables. iii �In terms of general health, Ranch Hand enlisted groundcrew rated their health as fair or poor more frequently than their enlisted Comparisons; differences were not observed for the enlisted flyers or the officers. Physician examiners detected no differences for appearance of illness or distress or for the appearance of relative age. The Ranch Hands had significantly lower percent body fat. They also had a higher proportion of sedimentation rate abnormalities than the Comparisons, but mean sedimentation rates were not statistically different between the two groups. No significant differences between the Ranch Hand and Comparison groups were seen in the 1982-1985 interval for skin or systemic cancers. However, when overall lifetime basal cell carcinoma rates were adjusted for risk factors involved in the cause of such cancers (e.g., sun exposure, skin color, skin reaction to sun), Ranch Hands had a significantly higher proportion of basal cell carcinoma than Comparisons. No group differences were observed for systemic cancer, although two cases of possible dioxin-related cancer were noted in Ranch Hands, bringing the lifetime total to two of these cancers in each group. Overall, the cancer findings were not viewed as disturbing but as reason for continued medical surveillance. The neurological assessment of cranial nerve function, peripheral nerve function, and central nervous system coordination did not reveal any consistently significant group differences, although abnormalities tended to aggregate in the Ranch Hands. The Babinski reflex (found adverse in the Ranch Hands at the 1982 Baseline examination) was equal in both groups at the 1985 followup. Age, alcohol, and diabetes showed classical effects with many neurological measures. In the psychological evaluation based on the Minnesota Multiphasic Personality Inventory, the Comparisons had significantly more abnormalities for the denial and masculinity/femininity scales, whereas the Ranch Hands manifested marginally more abnormalities in the hysteria and social introversion scales. The Ranch Hands showed more abnormalities on the Cornell Medical Index scales than did-the Comparisons, but no differences were detected between the two groups on the functionally oriented Halstead Reitan Battery. There were no group differences for current or past neuroses or psychoses. Age, educational level, and alcohol history showed strong and expected effects on the psychological measures. Both the interval and the lifetime history of liver disease were equal in both groups, as was a lifetime history of peptic ulcer disease. Of nine liver function and two porphyrin laboratory tests, the Comparisons had significantly higher serum glutamic pyruvic transaminase and uroporphyrin means, whereas the Ranch Hands had a significantly higher mean alkaline phosphate level and a borderline elevated coproporphyrin value. There was no evidence to suggest an increased likelihood of porphyria cutanea tarda in the Ranch Hand group. In the dermatological assessment, not one case of chloracne was diagnosed on examination, nor was historical acne anatomically distributed in a pattern that suggested past chloracne in the Ranch Hand group. Exposure and longitudinal analyses were also essentially negative. The cardiovascular evaluation showed no significant group differences for reported or verified hypertension, reported heart disease, or reported or iv �verified heart attacks. However, the frequency of verified heart disease was significantly greater in the Ranch Hands than the Comparisons. The assessment of the central cardiac function by systolic blood pressure and electrocardiogram did not reveal any meaningful group differences. Evaluation of peripheral pulses by the Doppler technique revealed group equivalence in marked contrast to the Baseline examination, which found significant pulse deficits in the Ranch Hands. This change was likely due to required tobacco abstinence before the'pulse measurements. Overall, the groups were remarkably similar in cardiovascular health. The assessment of eight hematological measures showed no significant group differences. In fact, the groups were more similar at the followup examination than at the Baseline examination. Age, race, and smoking were significant risk factors for most hematological measures. The groups did not differ significantly in reported past kidney disease, although the Baseline questionnaire noted such in the Ranch Hands. Five laboratory measures of renal function were similar between groups in the unadjusted analyses. No pattern of results suggested a detriment to either group in the adjusted analyses. For the endocrine function, TSH and testosterone means were significantly higher in the Ranch Hands, but these results were not supported by the categorical tests. The impaired category of the glucose tolerance test revealed an excess in the Comparison group. Examination results for past thyroid disease, thyroid and testicular abnormalities, and additional tests for cortisol level and T3 % Uptake were similar in both groups. Age, race, occupation, percent body fat, and personality type were often significant adjusting variables. Overall, the endocrine health status was comparable in both groups. Comprehensive immunological tests composed of six cell surface marker studies and three functional stimulation studies showed no significant group differences in the unadjusted analyses. Age, smoking, and alcohol usage were generally strong covariates. The assessment of delayed hypersensitivity by skin testing was declared invalid because of excessive reader variation and shifting diagnostic criteria. The pulmonary assessment, consisting of past history, physical examination, and x-ray results did not indicate any consistently different disease patterns in the two groups. Age and lifetime smoking history were important risk factors for most pulmonary measures. The exposure index analyses, which were stratified by occupation, revealed sporadic differences between exposure levels; however, there were no consistent dose-response relationships that supported an herbicide effect for any clinical area. Longitudinal analyses were conducted for 19 variables, and 5 showed significant differences in the changes of the groups between the Baseline and followup examinations. Of these 5 variables, 1 (sedimentation rate) was believed to be related to a change in laboratory methods, and the other 4 (Babinski reflex, depression, platelet count, and manual all pulse index) were attributed to true changes over time for the groups. In comparing all results between the examinations as well as the formal longitudinal analyses, �a subtle, but consistent, decrease in group differences over the 3-year period has been observed. The process of inferring causality is complex and must be based on careful consideration of many factors. Any interpretations of the data must consider the biological plausibility, clinical significance, specificity and consistency of the findings, and a host of statistical factors, such as strength of the association, lack of independence of the measurements, and multiple testing. By direct and indirect evidence, it is concluded that this study is free of overt bias and that the measurement systems used to obtain the data were accurate and valid. By an overall pattern assessment, it is further concluded that the Ranch Hand and Comparison populations are similar. Finally, this first followup examination report concludes that there is insufficient evidence to support a cause and effect relationship between herbicide exposure and adverse health in the Ranch Hand group at this time. The study has revealed a number of minor medical findings that require continued surveillance. In full context, the results of this study must be viewed as additional reassuring evidence that, at this time, the current state of health of the Ranch Hand participants is unrelated to herbicide exposure in Vietnam. vi �PREFACE The release of this 1987 followup Morbidity Report marks more than 8-1/2 years of intensive Air Force research into the herbicide question. Since the commitment to Congress in October 1978 to conduct an epidemiologic investigation of Air Force personnel who aerially disseminated herbicides in the Vietnam War (code-named Operation Ranch Hand), the United States Air Force Surgeon General has issued the following publications: a Study Protocol, four annual mortality reports, the Baseline Morbidity Report, and this first followup morbidity report. Within the next 2 years, the second followup morbidity report, other annual mortality reports, and an expanded birth defects study are expected for publication. This level of commitment has used approximately $40 million of contract research funds, excluding significant Air Force in-house expenditures. Nearly 100 Government, academic, and industry scientists have guided and contributed to the Air Force Health Study (AFHS) since its inception. The Air Force's current advisory committee, chaired by Dr. Robert W. Miller of the National Cancer Institute, is responsible for providing assistance on all scientific and medical matters pertaining to the AFHS. The distinguished panelists are listed in Appendix A. There are numerous scientific strengths in the AFHS, beginning with the unequivocal exposure status of the Ranch Hand population, estimated to have been, on the average, 1,000 times that experienced by an unclothed man directly beneath a spraying aircraft. In the other direction, the Ranch Hand population was probably less exposed to dioxin than many studied industrial populations (based upon a lack of chloracne), and may not develop adverse health consequences because of a possible threshold mechanism. However, the participants of the AFHS have a more defined exposure than the ground troops and constitute a larger population under study than industrial cohorts. The chief strength of the AFHS is its design. The interwoven study elements of multiple mortality assessments, a Baseline morbidity study, and five followup morbidity studies over 20 years provide a comprehensive approach to the detection of attributable adverse health effects. The weakest feature of the design is the mortality assessment which, in the absence of significant case clustering, cannot detect group differences for very rare conditions (e.g., soft tissue sarcoma) because of the inherent constraints of the limited size of the Ranch Hand population. To some extent, this problem may be offset for the more prevalent cancers by combining both living and fatal cancers for future analyses. The strength of the mortality studies should increase with the aging of the study population and the concomitant increase in death with the passage of time. All four mortality assessments have shown that the Ranch Hand population is faring about the same as the Comparison group, with no unusual causes of death, increased frequency of death, or evidence suggesting death at younger ages. Because of the healthy veteran effect, both groups are surviving significantly longer than similarly aged civilians. The morbidity assessment, released in 1984, disclosed only minor differences between the Ranch vii �Hands and the Comparisons, and these differences were not traditional indicators of dioxin-related disease. Both the content and the progress of the AFHS has been presented on many occasions to Congress, to the media, and to scientific meetings around the world. On the whole, the AFHS has been very well received in these circles, giving additional strength and credence to this work. This report of the first followup study is important as it marks the sustained commitment of Congress and the Air Force to pursue the Agent Orange question to its logical scientific conclusion. From the medical and scientific perspectives, this followup examination gives the first opportunity to confirm or refute some of the Baseline findings, and to explore subtle longitudinal changes,while controlling for confounding factors. The fifth-year followup examination, which will have been initiated when this report is released, will be conducted at an average time of 20 years postexposure for the Ranch Hands, a critical period for the emergence of attributable cancer. Followup studies such as these provide the most powerful scientific means of detecting emerging herbicide effects. This report differs slightly from the Baseline Morbidity Report in several ways. The populations under study have changed slightly (see Chapter 2), since some Ranch Hands and Comparisons have voluntarily dropped out of the study, and additional study participants have joined (via the Comparison replacement strategy, or the addition of formerly noncompliant participants). Further, a greater variety of statistical techniques are used to explore bias considerations, subgroup categorical differences (see Chapter 7), and "best" model fitting via the use of two- and three-way interactions. In addition, specific medical tests were included in this examination to clarify whether less specific Baseline findings were relevant (e.g., Doppler measurement of arterial pulses). Early in both the examination and analysis phases of this followup examination, it became clear that a joint Air Force-contractor approach to the analysis of the data was required. The Air Force elected to perform much of the analytical work of this report (e.g., bias, compliance, longitudinal, and pulmonary analyses). Thus, this study has transitioned from "independent" contract work to a genuine team effort between the Science Applications International Corporation (SAIC) and the Air Force scientific staffs. In the spirit of this enriching teamwork, SAIC has listed the Air Force scientific staff co-equally on the cover page of this report. Because of the highly professional scientific interchanges on many challenging aspects of the analytical work, it is believed that this report represents a scientific product unattainable by either team independent of the other. A brief explanation of this report to the reader is in order. This report is written primarily for clinical epidemiologists, clinicians, and biostatisticians so that they may fully evaluate the data and analytic techniques herein. There are segments of this report that will be difficult for even the most experienced of these specialists to understand. Complete familiarity with the Study Protocol and prior mprtality and morbidity reports is essential in the full understanding of this report. Thus, this report is not intended for rapid distillation by the layman or by media representatives. It should be noted that the intent of the introductions of the clinical chapters is to provide only a broad overview of the literature with respect to dioxin endpoints. In addition, the statistical analyses in this report were generally prescribed by the Air.Force (based primarily upon viii �analyses performed for the Baseline Morbidity Report) and are not ad hoc analyses. The report format has been established to be complete, rigorous, and straightforward on all issues so that maximum scientific credibility will be maintained. As with the Baseline Report, the contractor, with Air Force authority, or the Air Force itself, will respond to telephone or written inquiries about the content of this report. This report, prepared by Science Applications International Corporation, is submitted as partial fulfillment of Contract No. F41689-85-D-0010. ix �ACKNOWLEDGMENTS The authors of the report gratefully acknowledge the outstanding support of all the contributors to this project. To all the individuals, named and unnamed, whose dedication and hard work over the past 2-1/2 years have made this report possible, the authors wish to express their sincere appreciation. U.S. Air Force Coinvestigators: Lt. Col. F. Page Armstrong, USAF (Ret.), Nurse Epidemiologist Vincent V. Elequin, Medical Record Librarian Alton Rahe, Mathematical Statistician Lt. Col. John Silva, Consultant, Immunology Support in conducting the statistical analysis: Michael B. Lustik, SAIC Paul Meier, Ph.D., University of Chicago Dung B. Phan, SAIC Wai-Kouk W. Yu, SAIC Data processing and management support: Cristina E. Buchholz, SAIC Melody Darby, USAF Christie L. Dyer, SAIC Steven C. Fullerton, SAIC Task Manager Dawnelle Gonzenbach, USAF George Sacerich, USAF Conduct of the medical records coding: Calvin E. Hollman, USAF Maricella Luna, USAF Earl A. Metts, USAF Janie E. Ridgill, SAIC Consultant Marion B. Yonce, SAIC Consultant Edward E. Zimmerman, USAF xi �Conduct of the physical examinations: Maung H. Aung, M.D., SCRF Dianna M. Cooper, SCRF Roger C. Cornell, M.D., SCRF Karen Curd, M.D., SCRF Donald J. Oalessio, M.D., SCRF Roberta M. Davidson, R.N., SCRF William R. Dito, M.D., SCRF Betty Greene, SCRF Gene T. Izuno, M.D., SCRF L. Dee Jacobsen, Ph.D., SCRF Sharon Lav, SIRL Tony P. Lopez, M.D., SCRF David A. Mathison, M.D., SCRF Anthony P. Moore, M.D., SCRF Robert M. Nakamura, M.D., SCRF Shirley M. Otis, M.D., SCRF Roy F. Perkins, M.D., SCRF John S. Romine, M.D., SCRF Kathleen Rooney, SCRF Stephen K. Sargeant, M.D., SCRF Stanley G. Seat, M.D., SCRF Abbas Sedaghat, M.D., SCRF Marjorie E. Seybold, M.D., SCRF Robert B. Sigafoes, M.D., SCRF Jack C. Sipe, M.D., SCRF Ernest S. Tucker, M.D., SIRL Tonia Vyenielo, M.D., SCRF David E. Williams, M.D., SCRF, Medical Project Director Questionnaire administration and scheduling: Mary Catherine Burich, NORC Terrence D. Callier, NORC Ellwood Carter, NORC Charlene Harris, NORC Celia E. Romans, NORC Suzanne Turner, NORC Logistical arrangements: Joyce A. Douglass, SAIC, Task Manager Jacqueline P. Kirk, SAIC Martha Jean Perkins, SAIC Editorial support and report production: Thelma M. Bailey, SAIC Bernadette A. Bannister, SAIC L. Jean Massie, SAIC Consultant Grace Verchek, SAIC Lenore C. Wagner, SAIC Anna B. Wittig, SAIC Consultant xii �Management and quality review: Patrick A. Bannister, SAIC Leon B. Ellwein, Ph.D., SAIC Consultant Charles Pricker, SAIC Consultant Michael J. Higgins, SAIC Laurence C. Novotney, SAIC Carole J. O'Toole, SAIC James F. Striegel, Ph.D., SAIC Contractual and administrative support: Marie H. Manber, SAIC Lloyd E. Payne, Jr., USAF Joyce C. Standish, SAIC Support and Encouragement: Ranch Hand Association Members And, for making this study possible: All Study Participants xiii �TABLE OF CONTENTS EXECUTIVE ill PREFACE vii ACKNOWLEDGEMENTS xi 1. BACKGROUND STUDY DESIGN PURPOSE REFERENCES 1-1 1-2 1-3 1-4 2. 2-1 2-1 2-2 POPULATION BASELINE CANDIDATE IDENTIFICATION FOLLOWUP CANDIDATE IDENTIFICATION PARTICIPANT SELECTION 2-2 ENROLLMENT ; PERSONAL CHARACTERISTICS AND HABITS OF FOLLOWUP POPULATION LONGITUDINAL LOSSES AND GAINS SUMMARY REFERENCES ' 2-2 2-4 2-10 2-13 2-14 3. QUESTIONNAIRE METHODOLOGY QUESTIONNAIRE DEVELOPMENT INTERVIEWER TRAINING TELEPHONE SURVEY SCHEDULING OF PARTICIPANTS DATA COLLECTION DATA PROCESSING REFERENCES 3-1 3-1 3-2 3-2 3-3 3-3 3-5 3-6 4. PHYSICAL EXAMINATION METHODOLOGY EXAMINATION CONTENT CONDUCT OF EXAMINATIONS 4-1 4-2 4-2 5. 5-1 5-1 STUDY SELECTION AND PARTICIPATION BIAS INTRODUCTION AND BASELINE SUMMARY The Protocol 5-1 The Baseline Replacement Operation 5-1 The Baseline Selection Bias Analyses 5-2 The Baseline Compliance Bias Analyses THE FIRST FOLLOWUP SCHEDULING AND REPLACEMENT OPERATION FIRST FOLLOWUP COMPLIANCE^ FACTORS KNOWN OR SUSPECTED TO INFLUENCE STUDY PARTICIPATION THE TELEPHONE SURVEY. xv 5-3 5-3 5-4 5-7 5-7 �TABLE OF CONTENTS (continued) REPLACEMENT COMPARISONS VERSUS THE NONCOMPLIANT COMPARISONS THEY REPLACED Baseline Replacement First Followup Replacement .0 SCHEDULING AT FIRST FOLLOWUP NEW REPLACEMENTS VERSUS OLD REPLACEMENTS ORIGINAL COMPARISONS VERSUS SHIFTED ORIGINAL COMPARISONS PARTIALLY COMPLIANT VERSUS FULLY COMPLIANT PARTICIPANTS CONCLUSIONS 6. QUALITY CONTROL ADMINISTRATIVE QUALITY ASSURANCE QUESTIONNAIRE QUALITY CONTROL PHYSICAL EXAMINATION QUALITY CONTROL LABORATORY QUALITY CONTROL Quality Control Procedures for the Immunology Laboratory.... DATA MANAGEMENT QUALITY CONTROL Overview of Quality Control Procedures Data Processing System Design Design and Administration of Physical and Psychological Examination Forms. Data Completeness Checks Data Validation Techniques Medical Records Coding Quality Control STATISTICAL ANALYSIS QUALITY CONTROL REFERENCES 7. STATISTICAL METHODS STATISTICAL STUDY DESIGN STATISTICAL ISSUES Intervening Variables Power Multiple Endpoints and Comparisons Paired Versus Unpaired Analyses Mortality and Morbidity Data Cutpoints Exclusions OVERVIEW OF STATISTICAL PROCEDURES Preliminary Analysis « Core Analysis . Continuous Dependent Variables. Categorical Dependent Variables Modeling Strategy. EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES General Continuous Data. Categorical Data REFERENCES xv i 5-12 5-12 5-15 5-16 5-16 5-23 5-30 5-33 6-1 6-1 6-1 6-3 6-4 6-6 6-7 6-7 6-8 6-8 6-10 6-11 6-11 6-12 6-13 7-1 7-1 7-2 7-3 7-4 7-4 7-7 7-8 7-8 7-8 7-8 7-14 7-14 7-14 7-15 7-15 7-16 7-17 7-17 7-17 7-18 7-19 �TABLE OF CONTENTS (continued) 8. EXPOSURE INDEX REFERENCES 8-1 8-5 9. GENERAL HEALTH INTRODUCTION Baseline Summary Results Parameters of the 1985 General Health Assessment RESULTS AND DISCUSSION Subjective Assessments Self-Perception of Health Appearance of Illness or Distress Appearance of Relative Age Objective Assessments Erythrocyte Sedimentation Rate Percent Body Fat EXPOSURE INDEX ANALYSES Self-Perception of Health Appearance of Relative Age., Erythrocyte Sedimentation Rate Percent Body Fat LONGITUDINAL ANALYSES SUMMARY AND CONCLUSIONS REFERENCES 9-1 9-1 9-1 9-1 9-2 9-2 9-2 9-7 9-7 9-10 9-10 9-12 9-14 9-14 9-16 9-17 9-19 9-20 9-21 9-24 10. MALIGNANCY INTRODUCTION Baseline Summary Results Parameters of the 1985 Cancer Assessment RESULTS AND DISCUSSION General Questionnaire Data Physical Examination Data Statistical Analysis Baseline-Followup Interval Interval Skin Neoplasms Interal Systemic Neoplasms Lifetime (Baseline and Interval) Lifetime Skin Neoplasms Lifetime Systemic Neoplasms Comparison of Baseline, Interval, and Lifetime Results... Malignant Skin Neoplasms Malignant Systemic Neoplasms Baseline Participants EXPOSURE INDEX ANALYSES DISCUSSION Skin Cancer Systemic Cancer All Cancer SUMMARY AND CONCLUSIONS REFERENCES 10-1 10-1 10-5 10-6 10-6 10-6 10-6 10-7 10-7 10-8 10-8 10-24 10-30 10-33 10-43 10-51 10-51 10-51 10-53 10-55 10-63 10-64 10-66 10-68 10-68 10-73 xvii �TABLE OF CONTENTS (continued) 11. NEUROLOGICAL ASSESSMENT INTRODUCTION. Baseline Summary Results Parameters of the 1985 Neurological Assessment. RESULTS AND DISCUSSION General Questionnaire Data Physical Examination Data Cranial Nerve Function Peripheral Nerve Status Central Nervous System Coordination EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES SUMMARY AND CONCLUSIONS REFERENCES 11-1 11-1 ' 11-2 11-3 11-3 11-3 11-5 11-6 11-9 11-14 11-14 11-18 11-25 11-25 11-30 12. PSYCHOLOGICAL ASSESSMENT INTRODUCTION Baseline Summary Results Parameters of the 1985 Psychological Assessment RESULTS AND DISCUSSION Questionnaire Data Psychological Examination Data Statistical Analysis Minnesota Multiphasic Personality Inventory (MMPI) Cornell Medical Index (CMI) Halstead-Reitan Battery (HRB) EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES DISCUSSION SUMMARY AND CONCLUSIONS REFERENCES 12-1 12-1 12-3 12-4 12-4 12-4 12-5 12-7 12-7 12-19 12-24 12-25 12-38 12-40 12-42 12-45 13. GASTROINTESTINAL ASSESSMENT INTRODUCTION Baseline Summary Results Parameters of the 1985 Gastrointestinal Assessment... RESULTS AND DISCUSSION Questionnaire Data, Liver Disorders Peptic Ulcer Diseases. Mortality Count Data Physical Examination Data. General Laboratory Examination Data Statistical Analyses. Serum Glutamic-Oxaloacetic Transaminase (SCOT) Serum Glutamic-Pyruvic Transaminase (SGPT) Gamma-Glutamyl Transpeptidase (GGTP) Alkaline Phosphatase 13-1 13-1 13-2 13-3 13-3 13-4 13-8 13-9 13-11 13-11 13-14 13-14 13-22 13-22 13-23 xviii �TABLE OF CONTENTS (continued) Total Bilirubin Direct Bilirubin Lactic Dehydrogenase (LDH) Cholesterol Triglycerides Uroporphyr in Coproporphyrin Discussion Questionnaire-Laboratory Correlations: Porphyria Cutanea Tarda EXPOSURE INDEX ANALYSES SCOT SGPT GGTP Alkaline Phosphatase Total Bilirubin Direct Bilirubin. LDH Cholesterol Triglycerides Uroporphyrins and Coproporphyrin EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES SUMMARY AND CONCLUSIONS REFERENCES 14. DERMATOLOGICAL EVALUATION INTRODUCTION Baseline Summary Resul ts Parameters of the 1985 Dermatological Evaluation RESULTS AND DISCUSSION General Questionnaire Data Occurrence of Acne Duration of Acne Location of Acne Physical Examination Data Preliminary Dependent Variables and Covariate Relationships Analyses of Individual Dependent Variables Biopsy Results EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES DISCUSSION SUMMARY AND CONCLUSIONS REFERENCES xix 13-24 13-24 .. 13-25 13-25 13-26 13-26 13-27 13-27 13-28 13-30 13-31 13-31 13-31 13-42 13-42 13-42 13-42 13-42 13-43 13-43 13-43 13-43 13-44 13-48 14-1 14-1 14-2 14-2 14-3 14-3 14-3 14-5 14-6 14-7 14-8 14-8 14-11 14-25 14-27 14-33 14-34 14-34 14-37 �TABLE OF CONTENTS (continued) 15. CARDIOVASCULAR EVALUATION INTRODUCTION 15-1 15-1 Baseline Summary Results .. Parameters of the 1985 Cardiovascular Examination RESULTS AND DISCUSSION. 15-2 15-3 15-4 Questionnaire Data: Reported and Verified Heart Disease.... 15-4 Morbidity-Mortality Analysis 15-5 Physical Examination Data 15-11 Central Cardiac Function 15-11 Peripheral Vascular Function 15-20 Diastolic Blood Pressure 15-20 EXPOSURE INDEX ANALYSES 15-36 Reported and Verified Heart Disease Central Cardiac Function Peripheral Vascular System Association of Cardiovascular Examination Findings with Verified Heart Disease LONGITUDINAL ANALYSES DISCUSSION SUMMARY AND CONCLUSIONS REFERENCES 15-36 15-36 15-42 15-42 15-49 15-50 15-51 15-56 16. HEMATOLOGICAL EVALUATION INTRODUCTION....'.. 16-1 16-1 Baseline Summary Results Parameters of the 1985 Hematological Evaluation RESULTS AND DISCUSSION 16-2 16-3 16-3 General Unadjusted Categorical Analyses Unadjusted Analyses of Continuous Data Dependent Variable and Covariate Relationships Adjusted Categorical Analyses. Adjusted Analyses of Continuous Data. Discussion Red Blood Cell Count (RBC) White Blood Cell Count (WBC) Hemoglobin Concentration (HGB) 16-3 16-4 16-6 16-7 16-8 16-9 16-9 16-9 ,.. 16-11 16-11 Hematocrit (HCT). 16-12 Mean Corpuscular. Volume (MCV) Mean Corpuscular Hemoglobin (MCH) 16-12 16-13 Mean Corpuscular Hemoglobin Concentration (MCHC) Platelet Count (PLT) EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES ... .. SUMMARY AND CONCLUSIONS.. REFERENCES ... ....',....... xx 16-13 16-14 16-14 16-20 16-21 16-24 �TABLE OF CONTENTS (continued) 17. RENAL ASSESSMENT INTRODUCTION Baseline Summary Results Parameters of the 1985 Renal Assessment RESULTS AND DISCUSSION Questionnaire Data Physical Examination Data Laboratory Data Urinary Protein Urinary Occult Blood Urinary White Blood Cell Count Blood Urea Nitrogen (BUN) Urinary Specific Gravity EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES SUMMARY AND CONCLUSIONS REFERENCES 17-1 17-1 17-1 17-2 17-3 17-3 17-5 17-5 17-6 17-9 17-12 17-15 17-16 17-17 17-23 17-23 17-27 18. ENDOCRINE ASSESSMENT INTRODUCTION Baseline Summary Results Parameters of the 1985 Endocrine Assessment RESULTS AND DISCUSSION Questionnaire Data Physical Examination Data Laboratory Test Data General Thyroid Function: T3 % Uptake and Thyroid Stimulating Hormone (TSH) Testosterone Cortisol: Initial, 2-Hour, and Differential Glucose Metabolism: 2-Hour Postprandial Glucose and Composite Diabetes Indicator EXPOSURE INDEX ANALYSES LONGITUDINAL ANALYSES SUMMARY AND CONCLUSIONS REFERENCES 18-1 18-1 18-2 18-2 18-3 18-3 18-3 18-6 18-6 19. IMMUNOLOGICAL EVALUATION INTRODUCTION Baseline Summary Results Parameters of the 1985 Immunologic Profile Rationale of the Immunologic Measurements Immunology Methodologies Cell Surface Marker. Analysis Phytohemagglutinin (PHA) and Pokeveed Mitogen Stimulation Assays Mixed Lymphocyte Reaction Natural Killer Cell Assays Interpretive Considerations 19-1 19-1 19-2 19-3 19-3 19-6 19-6 xx i 18-10 18-16 18-17 18-18 18-24 18-25 18-30 18-34 19-6 19-6 19-7 19-7 �TABLE OF CONTENTS (continued) RESULTS AND DISCUSSION 19-8 Cell Surface Marker (Phenotypic) Studies 19-8 Total T Cells (T ) 19-14 Helper T Cells (T4) 19-14 Suppressor T Cells ( . T) 19-15 B Cells Monocy tes 19-15 19-16 HLA-DR Cells T4/T8 Ratio 19-17 19-18 Functional Stimulation Studies Unstimulated Response (PHA) 19-18 19-20 PHA Net Response 19-23 Pokeweed Net Response Net Response to MLC Stimulation 19-24 19-24 Discussion EXPOSURE INDEX ANALYSES 19-25 19-25 Cell Surface Markers Functional Stimulation Tests. 19-32 19-32 SKIN TESTING RESULTS. 19-33 General Statistical Analyses and Interpretations 19-33 19-34 SUMMARY AND CONCLUSIONS REFERENCES 19-42 19-45 20. PULMONARY DISEASE; INTRODUCTION 20-1 20-1 Baseline Summary Results 20-1 Parameters of the 1985 Pulmonary Examination RESULTS AND DISCUSSION 20-2 20-2 Mortality Experience Unadjusted Morbidity Analyses. 20-2 20-2 Adjusted Morbidity Analyses EXPOSURE ANALYSES SUMMARY AND CONCLUSIONS REFERENCES 20-3 20-3 20-11 20-13 21. INTERPRETIVE CONSIDERATIONS DIOXIN ENDPOINTS EXPOSURE TYPES OF MEASUREMENTS BASELINE-FOLLOWUP EXAMINATION DIFFERENCES STUDY BIASES. GROUP INTERACTIONS: PATTERN RECOGNITION CLASSICAL COVARIATES .„ MULTIPLE COMPARISONS CAUSALITY. . 21-1 21-1 21-2 21-3 21-3 21-4 21-5 21-9 . xxii 21-9 21-10 �TABLE OF CONTENTS (continued) 22. CONCLUSIONS INTRODUCTION STUDY PERFORMANCE ASPECTS POPULATION CHARACTERISTICS Patterns of Results CLINICAL ASPECTS General Health Malignancy ' Neurological Assessment Psychological Assessment Gastrointestinal Assessment Dermatological Evaluation Cardiovascular Evaluation. Hematological Evaluation Renal Assessment Endocrine Assessment Immunological Evaluation Pulmonary Disease CONCLUSION 23. FUTURE DIRECTIONS FIFTH-YEAR FOLLOWUP EXAMINATION EXPOSURE INDEX REFINEMENTS ADDITIONAL ANALYSES AND STUDIES xxiii .. . .. 22-1 22-1 22-1 22-1 22-2 22-2 22-2 22-3 22-3 22-4 22-4 22-5 22-5 22-5 22-6 22-6 22-7 22-7 22-7 23-1 23-1 23-1 23-2 �LIST OF APPENDICES Appendix A Advisory Committee on Special Studies Relating to the Possible Long-Term Health Effects of Phenoxy Herbicides and Contaminants B Questionnaire Methodology. C Physical Examination Methodology D Study Selection and Participation Bias. E Statistical Methods F Exposure Index G General Health H Neoplasia I Neurological Assessment J Psychological Assessment K Gastrointestinal Assessment L Dermatological Evaluation M Cardiovascular Assessment N Hematological Evaluation 0 Renal Assessment P Endocrine Assessment Q Immunological Evaluation. R Pulmonary Disease. .. .... .... S Glossary of Abbreviations xxiv , A-l B-l C-l D-l E-l F-l ,. G-l H-l . 1-1 J-l K-l L-l M-l . N-l . 0-1 P-l Q-l R-l S-l �LIST OF TABLES Table 2-1 2-2 2-3 2-4 2-5 2-6 2-7 2-8 2-9 2-10 4-1 4-2 5-1 5-2 5-3 5-4 5-5 5-6 5-7 5-8 5-9 5-10 5-11 Candidate Followup Participants by Group and Baseline Compliance Status Participants Enrolled in the Followup Study by Group and Baseline Compliance Status Age (in 1985) of Participants of the Followup Examination by Group History of Tobacco and Alcohol Use of Participants of the Followup Examination by Group Average Use of Tobacco Products and Alcohol for Those Reporting Use of These Substances: Participants of the Followup Examination by Group Educational Background of Participants of the Followup Examination by Group Religious Preference of Participants of the Followup Examination by Group Military Status of Participants of the Followup Examination by Group Risk-Taking Behavior of Participants of the Followup Examination by Group Losses/Gains of Participants Between the Baseline and Followup Examinations 2-3 2-5 2-6 2-7 2-8 2-9 2-9 2-10 2-11 2-12 Elements of the Followup Physical Examination Laboratory Test Procedures of the Followup Physical • Examination 4-3 Baseline Versus First Followup Sample Sizes Reasons for Nonparticipation in the First Followup of 56 Ranch Hands and 50 Comparisons Who Were Fully Compliant at Baseline Reported Health Status of 35 Ranch Hands and 42 Comparisons Fully Compliant at Baseline and Noncompliant at First Followup Baseline Status of Newly Examined Participants Summary of Reasons for Noncompleted Telephone Interviews... Summary of Results to the Telephone Questionnaire Contrast of Interviewer's Remark from Telephone Interviews and Reported Health Status Self-Reported Health of Previously Uncontacted Comparisons, in 1986, Versus Self-Reported Health Status of Original Comparisons at Baseline Noncompliant Original Comparisons and Replacement Comparisons Versus Their Baseline Replacements: Reported Health Status at Baseline Noncompliant Original Comparisons and Replacement Comparisons Versus Their Baseline Replacements: Medication Use at Baseline Noncompliant Original Comparisons and Replacement Comparisons Versus Their Baseline Replacements: Income at Baseline 5-4 xxv 4-4 5-5 5-5 5-6 5-8 5-9 5-11 5-12 5-13 5-13 5-14 �LIST OF TABLES (Continued) Table 19-5 19-6 19-7 19-8 19-9 19-10 19-11 19-12 19-13 19-14 20-1 20-2 20-3 20-4 20-5 20-6 20-7 20-8 21-1 21-2 Page Adjusted Analyses for Cell Surface Markers by Group Unadjusted Analyses for Functional Stimulation Tests by Group Association Between Functional Stimulation Test Variables and the Covariates in the Combined Ranch Hand and Comparison Groups „ ...... Adjusted (Directionality Shown) for Functional Stimulation Tests by Group Adjusted Exposure Index Analyses for Cell Surface Markers by Occupation Adjusted Exposure Index Analysis for Functional Stimulation Tests by Occupation Summary of Exposure Index by Covariate Interactions for Functional Stimulation Tests Clinical Interpretation Categories of Skin Test Results by Specific Measurement Criteria at SCRF Induration Erythema Relationships in Average Percentage Over Four Skin Tests, by Reader.. Overall Sumary Results of Unadjusted and Adjusted Analyses of Immunologic Variables Unadjusted Analyses of Reported History of Respiratory Illness by Group Unadjsted Analyses of Radiological and Clinical Respiratory System Findings by Group Adjustd Analyses of Respiratory Variables by Group.. Summary of Group-by-Covariate Interactions for Respiratory Variables Exposure Index Analysis Results for Officers p-Values of Dependent Variable-by-Covariate Association Exposure Index Analysis Results for Enlisted Flyers p-Values of Dependent Variable-by-Covariate Association.. Exposure Index Analysis Results for Enlisted Groundcrew: p-Values of Dependent Variable by Covariate Association.. Overall Summary Results of Unadjusted and Adjusted Analyses of Pulmonary Disease Summary Associations of Adverse Health Effects to TCDD Exposure Reported in the Literature Summary of Significant Covariate Strata (or Covariate Level Difference) Found Within Significant Two- and Three-Factor Group-by-Covariate Interactions by Clinical Chapter and Dependent Variable (Group Direction and p-Value) xxxvi 19-12 19-20 19-21 19-22 19-26 19-30 19-33 19-34 19-41 19-43 20-4 20-5 20-6 20-7 20-10 20-10 20-11 20-12 21-1 21-6 �LIST OF TABLES (Continued) Table 9-1 9-2 9-3 9-4 9-5 9-6 9-7 9-8 9-9 9-10 9-11 9-12 9-13 9-14 9-15 9-16 9-17 9-18 9-19 9-20 9-21 10-1 10-2 Page Unadjusted Analysis for Self-Perception of Health by Group Association Between Self-Perception of Health and Age, Race, Occupation, and Personality Score in the Combined Ranch Hand and Comparison Groups Adjusted Relative Risks of Self-Perception of Health by Occupation Frequency of Self-Perception of Health by Occupation and Group Unadjusted Analysis for Appearance of Acute Illness or Distress by Group Unadjusted Analysis for Appearance of Relative Age by Group Association Between Appearance of Relative Age and Age, Race, and Occupation in the Combined Ranch Hand and Comparison Groups Adjusted Relative Risks of Appearance of Relative Age by Occupation Association Between Sedimentation Rate and Age, Race, Occupation, and Personality Score in the Combined Ranch Hand and Comparison Groups Unadjusted Analysis for Sedimentation Rate by Group Association Between Percent Body Fat and Age, Race, and Occupation in the Combined Ranch Hand and Comparison Groups Unadjusted Exposure Index Analysis of Self-Perception of Health by Occupation Adjusted Relative Risk of Self-Perception of Health by Occupation and Exposure Contrast Unadjusted Exposure Index Analysis of Appearance of Relative Age by Occupation Adjusted Mean Sedimentation Rates by Occupation Unadjusted Exposure Index Analysis of Sedimentation Rate by Occupation Adjusted Relative Risk of Sedimentation Rate by Occupation and Exposure Contrast Unadjusted Means of Percent Body Fat by Occupation Unadjusted Exposure Index Analysis of Percent Body Fat by Occupation Longitudinal Analysis of Self-Perception of Health and Sedimentation Rate: A Contrast of Baseline and First Followup Examination Abnormalities Overall Summary Results of Unadjusted and Adjusted Analyses of General Health Variables Unadjusted Analyses of Followup Participants with Verified and Suspected Neoplasms in the Baseline-Followup Interval by Group (Nonblacks and Blacks) Unadjusted Analyses of Nonblack Followup Participants with Verified and Suspected Malignant Skin Neoplasms in the Baseline-Followup Interval by Cell Type and Group xxvii 9-3 9-4 9-5 9-6 9-7 9-8 9-9 9-9 9-11 9-12 9-13 9-15 9-15 9-16 9-17 9-18 9-18 9-19 9-20 9-21 9-22 10-9 10-10 �LIST OF TABLES (Continued) Table 10-3 10-4 10-5 10-6 10-7 10-8 10-9 10-10 10-11 10-12 10-13 10-14 10-15 10-16 10-17 10-18 10-19 Unadjusted Analyses of Nonblack Followup Participants with Verified and Suspected Malignant Skin Neoplasms in the Baseline-Followup Interval by Anotomic Site and Group Unadjusted Analyses of Nonblack Followup Participants with Nonmelanoma Skin Neoplasms and Sun Exposure-Related Skin Malignancies in the Baseline-Followup Interval Occurring on the Face, Head, or Neck, by Occupation .. Covariates for Analysis of Malignant Skin Neoplasms Summary of Associations Between Incidence Rates of Basal Cell Carcinoma and Sun Exposure-Related Skin Malignancies and the Covariates, in Baseline-Followup Interval for Combined Followup Ranch Hand and Comparison Nonblack Participants Adjusted Analyses of Nonblack Followup Participants for Malignant Skin Neoplasm Incidence During BaselineFollowup Interval Summary of Followup Participants with Verified Malignant Systemic Neoplasms in Baseline-Followup Interval by Group Summary of Followup Participants with Suspected Malignant Systemic Neoplasms at Physical Examination by Group Unadjusted Analyses of Followup Participants with Verified and Suspected Malignant Systemic Neoplasms in the Baseline-Followup Interval by Group Summary of Associations Between Incidence Rates of All Malignant Systemic Neoplasms and the Covariates in the Baseline-Followup Interval for Combined Followup Ranch Hand and Comparison Groups Adjusted Analyses of Followup Participants for the Incidence of All Malignant Systemic Neoplasms During the Baseline-Followup Interval Unadjusted Analyses of Followup Participants with Lifetime Occurrence of Verified or Suspected Neoplasms by Group (Blacks Nonblacks) .Unadjusted Analyses of Nonblack Followup Participants with Lifetime Occurrence of Verified and Suspected Malignant Skin Neoplasms by Cell Type and Group Association Between Lifetime Incidence of Suspected Basal Cell Carcinoma and the Covariates for the Combined Followup Ranch Hand and Comparison Nonblack Participants. Adjusted Analyses of Nonblack Followup Participants for Lifetime Occurrence of Malignant Skin Neoplasms Incidence Summary of Followup Participants with Lifetime Incidence of Verified Malignant Systemic Neoplasms by Group Summary of Followup Participants with Lifetime Soft Tissue Sarcoma, Leukemia or Lymphoma by Group Unadjusted Analyses of Lifetime Incidence Rates of All Malignant Systemic Neoplasms Combined, by Group xxviii 10-12 10-13 10-15 10-19 10-23 10-25 10-26 10-28 10-29 10-31 10-32 10-34 10-36 10-42 10-44 10-45 10-46 �LIST OF TABLES (Continued) Table 10-20 Association Between Lifetime Incidence of All Malignant Systemic Neoplasms and the Covariates for Combined Followup Ranch Hand and Comparison Participants 10-21 Adjusted Analyses for Lifetime Incidence of All Malignant Systemic Neoplasms Combined 10-22 Unadjusted and Adjusted Analyses of the Incidence of All Verified Malignant Skin and Systemic Neoplasms and Basal Cell Carcinoma: Baseline-Followup Interval, and Lifetime Occurrence 10-23 Fully Compliant Baseline Participants by Status at Followup Examination and Group 10-24 Fully Compliant Baseline Participants Who Did Not Participate in Followup Examination by Status and Group.. 10-25 Fully Compliant Baseline Participants Also in Followup Examination by Malignant Neoplasm Status 10-26 Adjusted Exposure Index Analysis for Followup Participants for Occurrence of Malignant Neoplasms in the BaselineFollowup Interval 10-27 Adjusted Exposure Index Analysis for Followup Participants for Occurrence of Malignant Neoplasms 10-28 Computed Risks of Basal Cell Carcinoma by Group at Varying Levels of Four Risk Factors, Relative to Comparisons at Low Risk 10-29 Overall Summary Table: Unadjusted and Adjusted Analysis of Interval and Lifetime Skin and Systemic Cancer Incidence 11-1 11-2 11-3 11-4 11-5 11-6 11-7 11-8 11-9 11-10 11-11 Exclusions and Missing Data for Neurological Assessment by Group Unadjusted Analysis for Verified Neurological Disease by Group—1982-1985 Unadjusted Analysis for Verified Neurological Disease by Group—Baseline and First Followup Studies Combined Association Between Seven Neurological Variables and Three Summary Indices and the Covariates in the Combined Ranch Hand and Comparison Groups Unadjusted Analyses for Cranial Nerve Function by Group.... Adjusted Analyses for Selected Variables of Cranial Nerve Function by Group Summary Table of Group-by-Covariate Interactions for Neurological Variables '. Unadjusted Analyses for Peripheral Nerve Function by Group Adjusted Analyses for Selected Variables of Peripheral Nerve Function by Group Unadjusted Analyses for CNS Coordination Variables by Group '. Adjusted Analyses for Selected Variables of CNS Coordination by Group xxix 10-47 10-51 10-52 10-53 10-54 10-56 10-57 10-60 10-67 10-69 11-4 11-5 11-6 11-8 11-10 11-12 11-13 11-15 11-16 11-17 11-18 �LIST OF TABLES (Continued) Table Page 11-12 Adjusted Exposure Index Analyses for Neurological Variables by Occupation 11-13 Summary of Exposure Index-by-Covariate Interactions for Neurological Variables 11-14 Longitudinal Analysis of Romberg Sign and Babinski Reflex; A Contrast of Baseline and First Followup Examination Abnormalities 11-15 Overall Summary Results of Unadjusted and Adjusted Analyses of Neurological Variables 12-1 12-2 12-3 12-4 12-5 12-6 12-7 12-8 12-9 12-10 12-11 12-12 13-1 13-2 13-3 13-4 13-5 13-6 13-7 13-8 Unadjusted Analyses for Reported Psychological Illnesses by Group: Baseline and First Followup Studies Combined.. Unadjusted Analyses for MMPI by Group Association Between MMPI Variables and the Covariates in the Combined Ranch Hand and Comparison Groups Adjusted Analyses for MMPI by Group Unadjusted Analyses for the Cornell Medical Index (CMI) by Group. Association Between CMI Variables and the Covariates in the Combined Ranch Hand and Comparison Groups Adjusted Analyses for CMI Variables by Group Summary Results for the Halstead-Reitan Battery Impairment Index Analyses Adjusted Exposure Index Analyses for Psychological Variables by Occupation. Summary of Exposure Index-by-Covariate Interactions in Adjusted Analyses of Psychological Variables Longitudinal Analysis of Depression and Denial: A Contrast of Baseline and First Followup Examination Abnormalities Overall Summary Results of Adjusted and Unadjusted Analyses of Psychological Variables Number of Other Liver Conditions Reported by Study Participants at Followup by Group (Verified by Medical Record Review) Unadjusted Analyses for Baseline and Interval History of Liver Disease by Group (Verified by Medical Record Review) Medical Record Verification of Reported Liver Symptoms and Diseases by Group (Baseline and Interval Questionnaires Combined) Unadjusted Analysis of Blood Type by Group » Frequency of Diagnosed and Reported Ulcer Disease by Group Unadjusted Analyses of Peptic Ulcer Disease by Blood Type by Group :.............. Frequency of Digestive System Mortality by Group Unadjusted Analysis of Enlarged Livers Diagnosed at Physical Examination by Group xxx 11-20 11-25 11-26 11-27 12-5 12-9 12-11 12-12 12-20 12-21 12-22 12-26 12-27 12-37 12-39 12-43 13-5 13-6 13-7 13-8 13-9 13-10 13-11 13-12 �LIST OF TABLES (Continued) Table 13-9 13-10 13-11 13-12 13-13 13-14 13-15 13-16 13-17 13-18 13-19 14-1 14-2 14-3 14-4 14-5 14-6 14-7 14-8 14-9 14-10 14-11 Page Laboratory Norms for Nine Hepatic Function Variables and Two Porphyrin Determinations Unadjusted Continuous 'and Categorical Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Group. Assocation Between Nine Hepatic Function Variables and Two Porphyrin Determinations and Six Covariates in the Combined Ranch Hand and Comparison Groups Adjusted Continuous and Categorical Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Group Unadjusted Analysis for Interval History of Skin Bruises, Skin Patches, and Skin Sensitivity by Group Unadjusted Analysis for Interval Porphyrin Abnormalities by Group and Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire Unadjusted Analysis for Interval Uroporphyrin Abnormalities by Group and Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire Adjusted Categorical Exposure Index Analyses (Main Effects Model) Results for Hepatic Function Variables by Occupation Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Longitudinal Analyses for SCOT, SGPT, and GGTP: A Contract of Baseline and First Followup Examination Test Means Overall Summary Results of Unadjusted and Adjusted Analyses of Nine Hepatic Function Variables and Two Porphyrin Metabolite Tests Unadjusted Analysis for Reported Historical Occurrence of Acne by Group Unadjusted Analysis for Reported Historical Occurrence of Acne Relative to 1961 by Group Unadjusted Analysis for Reported Historical Occurrence of Acne Relative to SEA Tour of Duty for Post-1961 Acne by Group Adjusted Analysis for Duration of Acne (in Years) for Post-1961 Acne by Group Association Between Dermatological Variables and Age, Race, Occupation, and Pre-SEA Acne in the Combined Ranch Hand and Comparison Groups Unadjusted Analysis for Comedones by Group Adjusted Analysis for Comedones by Group Unadjusted Analysis for Acneiform Lesions by Group Adjusted Analysis for Acneiform Lesions by Group Unadjusted Analysis for Acneiform Scars by Group Adjusted Analysis for Acneiform Scars by Group xxx i 13-13 13-15 13-17 13-18 13-28 13-29 13-30 13-32 13-36 13-44 13-46 14-5 14-6 14-7 14-7 14-11 14-12 14-12 14-13 14-14 14-14 14-15 �LIST OP TABLES (Continued) Table Page 14-12 14-13 14-14 14-15 14-16 14-17 14-18 14-19 14-20 14-21 14-16 14-16 14-17 14-17 14-18 14-19 14-20 14-21 14-22 Unadjusted Analysis for Depigmentation by Group Adjusted Analysis for Depigmentation by Group Unadjusted Analysis for Inclusion Cysts by Group.... Adjusted Analysis for Inclusion Cysts by Group..... Unadjusted Analysis for Hyperpigraentation by Group Adjusted Analysis for Hyperpigmentation by Group..... Unadjusted Analysis for Other Abnormalities by Group Adjusted Analysis for Other Abnormalities by Group Unadjusted Analysis for the Dermatology Index by Group Association Between the Dermatology Index and Age, Race, Occupation, and Presence of Pre-SEA Acne in the Combined Ranch Hand and Comparison Groups 14-22 Adjusted Analysis for the Dermatology Index by SEA Acne Class and Group 14-23 Adjusted Relative Risks for Contrasts of Dermatology Index by Pre-SEA Class 14-24 Summary of Histologic Descriptions of Skin Biopsy by Group 14-25 Adjusted Exposure Index Analysis for Dermatological Variables by Occupation 14-26 Summary of Exposure Index by Covariate Interactions Encountered in Adjusted Analysis of Dermatological Variables 14-27 Longitudinal Analysis of the Dermatology Index: A Contrast of Baseline and First Followup Examination Abnormalities 14-28 Overall Summary Results of Unadjusted and Adjusted Analyses of Questionnaire and Physical Examination Dermatological Variables 15-1 15-2 15-3 15-4 15-5 15-6 15-7 15-8 Unadjusted Analyses for Reported and Verified Heart Disease by Group Association Between Verified Essential Hypertension and the Covariates in the Combined Ranch Hand and Comparison Groups Association Between Verified Heart Disease and the Covariates in the Combined Ranch Hand and Comparison Groups Association Between Verified Myocardial Infarctions and the Covariates in the Combined Ranch Hand and Comparison Groups „. Adjusted Analyses for Reported and Verified Heart Disease.. Unadjusted Analyses for Central Cardiac Function By Group (Diabetics Excluded) Association Between Central Cardiac Function Variables and the Covariates in the Combined Ranch Hand and Comparison Groups (Diabetics Excluded) Adjusted Analyses for Central Cardiac Function (Diabetics Excluded) xxx ii 14-23 14-24 14-25 14-26 14-28 14-32 14-33 14-35 15-6 15-7 15-8 15-9 15-10 15-13 15-15 15-16 �LIST OP TABLES (Continued) Table 15-9 15-10 15-11 15-12 15-13 15-14 15-15 15-16 15-17 15-18 15-19 15-20 16-1 16-2 16-3 16-4 16-5 16-6 16-7 16-8 16-9 Page Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) Assocation Between Peripheral Vascular Function Variables and the Covariates in the Combined Ranch Hand and Comparison Groups (Diabetics Excluded) Adjusted Analysis for Peripheral Vascular Function by Group (Diabetics Excluded) Agreement Between Manual and Doppler Pulse Assessments (McNemar's r Test) Summary Statistics for Cardiovascular Covariates by Group.. Adjusted Exposure Index Analyses for Reported and Verified Heart Disease by Occupation Adjusted Exposure Index Analyses for Central Cardiac Function Variables by Occupation Adjusted Exposure Index Analyses for Diastolic Blood Pressure Funduscopic Abnormalities and Carotid Bruits by Occupation Adjusted Exposure Index Analyses for Peripheral Vascular System Manual Pulse Readings by Occupation Adjusted Exposure Index Analyses for Peripheral Vascular System Doppler Pulse Reading by Occupation Longitudinal Analyses of All Pulses Index and Overall ECG's: A Contrast of Baseline and First Followup Examination Abnormalities Overall Summary Results of Unadjusted and Adjusted Analyses Cardiovascular Variables Laboratory Parameters for Hematological Test Variables Unadjusted Categorical Analyses for Hematological Variables by Group Unadjusted Continuous Analyses for Hematological Variables (Contrast of Group Means) Association Between Hematological Variables and Age, Race, Occupation, and Smoking History in the Combined Ranch Hand and Comparison Groups Adjusted Categorical Analyses for Hematologic Variables (Abnormal versus Normal), Adjusted for Age, Race, Occupation, and Smoking) Adjusted Continuous Analyses for Hematological Variables, (Ranch Hand-Comparison Group Differences) Unadjusted Categorical Exposure Index Analyses for Hematological Variables by Occupation Adjusted Categorical Exposure Index Analyses (Log-Linear Models) for Hematological Variables by Occupation (p-Values) Unadjusted Continuous Exposure Index Analyses for Hematological Variables by Occupation (Analysis of Variance) xxxiii 15-21 15-24 15-28 15-31 15-34 15-37 15-39 15-43 15-44 15-47 15-49 15-52 16-4 16-5 16-6 16-7 16-8 16-10 16-15 16-17 16-18 �LIST OF TABLES (Continued) Table 16-10 16-11 16-12 17-1 17-2 17-3 17-4 17-5 17-6 17-7 17-8 17-9 17-10 Summary of Exposure Index-by-Covariate Interactions Encountered in Adjusted Continuous Analyses of Hematological Variables (General Linear Models) Longitudinal Analyses for MCV, MCH, and PLT: A Contrast of Baseline and First Followup Examination Test Means.... Overall Summary Results of Unadjusted and Adjusted Analyses of Hematological Variables Unadjusted Analysis of History of Kidney Disease/Kidney Stones by Group Association Between Kidney Disease/Kidney Stones and Age, Race, Occupation, and Diabetic Class in the Combined Ranch Hand and Comparison Groups Laboratory Norms for Five Renal Variables Summary of Renal Laboratory Variables by Group Association Between Urinary Protein and Age, Race, Occupation, and Diabetic Class in the Combined Ranch Hand and Comparison Groups Frequency of Urinary Protein by Diabetic Class and Group... Adjusted Relative Risks for Urinary Protein by Diabetic Class Association Between Urinary Occult Blood and Age, Race, Occupation, and Diabetic Class in the Combined Ranch Hand and Comparison Groups Adjusted Analysis for Urinary Occult Blood for Nonblacks by Group Frequency of Urinary Occult Blood for Blacks by Group Frequency of Urinary Occult Blood for Blacks by Occupation and Group 17-12 Frequency of Urinary WBC/HPF for Nonblacks by Group........ 17-13 Adjusted Analyses for Urinary WBC/HPF for Nonblacks by Age Category and Group. 17-14 Frequency of Urinary WBC for Blacks by Occupational Category and Group 17-15 Adjusted Analyses for Urinary WBC/HPF for Black Enlisted Flyers and Groundcrew by Age and Group 17-16 Adjusted Analysis of BUN by Race and Group 17-17 Adjusted Analysis of Urine Specific Gravity by Race, Occupation, and Group 17-18 Adjusted Categorical Exposure Index Analyses for Renal Variables by Occupation 17-19 Adjusted Continuous Exposure Index Analyses for Renal Variables. ,. 17-20 Summary of Exposure Index^by-Covariate Interactions for Renal Variables 17-21 Longitudinal Analysis of BUNs A Contrast of Baseline and First Followup Examination Laboratory Means 16-19 16-20 16-22 17-3 17-4 17-5 17-6 17-7 17-8 17-8 17-9 17-10 17-11 17-11 xxx iv 17-11 17-12 17-13 17-14 17-14 17-16 17-17 17-19 17-21 17-23 17-24 �LIST OF TABLES (Continued) Table 17-22 Page Overall Summary Results of Unadjusted and Adjusted Analyses for Renal Variables Unadjusted Analysis for Reporting of Thyroid Symptoms/Disease by Questionnaire Method by Group 18-2 Medical Record Verification Results of Reported Thyroid Disease by Group 18-3 Unadjusted Analysis for Thyroid and Testicular Conditions by Group 18-4 Laboratory Endocrinological Variables: SCRF Normal and Abnormal Ranges 18-5 Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group 18-6 Adjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group 18-7 Association Between T3 % Uptake and Age, Race, Occupation, and Personality Type in the Combined Ranch Hand and Comparison Groups . 18-8 Adjusted Categorical Analysis for T, % Uptake 18-9 Adjusted Categorical Analysis for TSH 18-10 Adjusted Continuous Analysis for TSH by Group 18-11 Adjusted Categorical Analysis for Testosterone... 18-12 Adjusted Continuous Analysis for Initial Cortisol by Group. 18-13 Association Between Differential Cortisol and Age, Race, Occupation, Percent Body Pat, and Personality Score in the Combined Ranch Hand and Comparison Groups 18-14 Adjusted Categorical Analysis for 2-Hour Postprandial Glucose 18-15 Adjusted Continuous Analysis for 2-Hour Postprandial Glucose by Group 18-16 Adjusted Analysis for Diabetes (Composite Indicator) 18-17 Adjusted Exposure Index Analyses for Endocrinological Variables by Occupation 18-18 Summary of Exposure Index-by-Covariate Interactions Encountered in Analyses of Endocrinological Variables.... 18-19 Longitudinal Analysis for Testosterone, T. % Uptake, and TSH: A Contrast of Baseline and First Followup Examination Test Means 18-20 Overall Summary Results of Unadjusted and Adjusted Continuous and Categorical Analyses of Endocrinological Variables 17-24 18-1 19-1 19-2 19-3 19-4 Medical Significance of the Immunologic Data Frequencies of Participants Who Took the Immunological Tests and the Skin Tests, by Group Unadjusted Analyses for Cell Surface Markers by Group Association Between Cell Surface Markers Variables and the Covariates in the Combined Ranch Hand and Comparison Groups (Directionality Shown) xxxv 18-4 18-5 18-5 18-6 18-7 18-11 18-13 18-14 18-15 18-15 18-16 18-19 18-19 18-23 18-23 18-24 18-26 18-30 18-31 18-32 19-4 19-8 19-10 19-11 �LIST OF TABLES (Continued) Table 19-5 19-6 19-7 19-8 19-9 19-10 19-11 19-12 19-13 19-14 20-1 20-2 20-3 20-4 20-5 20-6 20-7 20-8 21-1 21-2 Page Adjusted Analyses for Cell Surface Markers by Group Unadjusted Analyses for Functional Stimulation Tests by Group Association Between Functional Stimulation Test Variables and the Covariates in the Combined Ranch Hand and Comparison Groups Adjusted (Directionality Shown) for Functional Stimulation Tests by Group Adjusted Exposure Index Analyses for Cell Surface Markers by Occupation Adjusted Exposure Index Analysis for Functional Stimulation Tests by Occupation Summary of Exposure Index by Covariate Interactions for Functional Stimulation Tests Clinical Interpretation Categories of Skin Test Results by Specific Measurement Criteria at SCRF Induration Erythema Relationships in Average Percentage Over Four Skin Tests, by Reader..... Overall Sumary Results of Unadjusted and Adjusted Analyses of Immunologic Variables Unadjusted Analyses of Reported History of Respiratory Illness by Group Unadjsted Analyses of Radiological and Clinical Respiratory System Findings by Group Adjustd Analyses of Respiratory Variables by Group......... Summary of Group-by-Covariate Interactions for Respiratory Variables Exposure Index Analysis Results for Officers p-Values of Dependent Variable-by-Covariate Association Exposure Index Analysis Results for Enlisted Flyers p-Values of Dependent Variable-by-Covariate Association.. Exposure Index Analysis Results for Enlisted Groundcrew: p-Values of Dependent Variable by Covariate Association.. Overall Summary Results of Unadjusted and Adjusted Analyses of Pulmonary Disease Summary Associations of Adverse Health Effects to TCDD Exposure Reported in the Literature Summary of Significant Covariate Strata Found Within Significant Two- and Three-Factor Group-by-Covariate Interactions by Clinical Chapter and Dependent Variable (Group Direction and p-Value) xxx vi 19-12 19-20 19-21 19-22 19-26 19-30 19-33 19-34 19-41 19-43 20-4 20-5 20-6 20-7 20-10 20-10 20-11 20-12 21-1 21-6 �LIST OF FIGURES Figure 3-1 Page Selection Procedure for the Questionnaire, Physical Examination, and Followup Study 3-4 Flow Diagram of Day One Followup Interview and Physical Examination 4-6 Flow Diagram of Day Two Followup Interviews and Physical Examinat ion 4-7 4-3 Mark-Sense Form for Followup Neurological Examination 4-8 4-4 Mark-Sense Form for Followup Dermatological Examination and Biopsy 4-9 Percent Completed Physical Examination by Calendar Date for All Comparisons 5-17 5-2 Percent Completed Physical by Calendar Date 5-17 5-3 Percent Completed Physical Examination By Calendar Date for Unrestricted New and Old Replacement Comparisons 5-18 Two Levels of Quality Control Applied to All Collected Data Prior to Statistical Analysis 6-9 14-1 Occurrence of Acne by Time for First Followup Participants 14-4 14-2 Location of Post-SEA and Pre- and Post-SEA Acne by Group 14-9 14-3 Location of Post-SEA Acne by Group 14-10 18-1 Mean Cortisol Levels by Personality Type, Adjusted for Age and Percent Body Fat, by Time of. Specimen Collection 18-20 Mean Cortisol Levels by Percent Body Fat, Adjusted for Age and Personality Type, by Time of Specimen Collection 18-21 Relationship of Induration Measurements to Erythema Measurements for the Mumps Skin Test Reader 1 Results 19-35 4-1 4-2 5-1 6-1 18-2 19-1 19-2 Relationship of Induration Measurements to Erythema Measurements for the Trichophyton Skin Test Reader 1 Results... 19-36 19-3 Relationship of Induration Measurements to Erythema Measurements for the Mumps Skin Test Reader 2 Results 19-4 Relationship of Induration Measurements to Erythema Measurements for the Trichophyton Skin Test Reader 2 Results... xxxvii 19-37 19-38 �LIST OF FIGURES (continued) Figure Page 19-5 Relationship of Induration Measurements to Erythema Measurements for the Mumps Skin Test Reader 3 Results 19-39 19-6 Relationship of Induration Measurements to Erythema Measurements of the Trichophyton Skin Test Reader 3 Results.... 19-40 xxxviii �UNCLASSIFIED TH>t MCVAlTV REPORT DOCUMENTATION PAOf U A«*OAf iCCUAIt* CIAUI'ICATIOM Unclassified ~f«> MCUAltt ClAMl'«CATlON AvTMOAltV 9. O<«1Ai«wTiOM'AVA(tA,«iuTV 0» AttfOAT Distribution ~~~" authorized to U.S. Government agencies only By reason of administrative use (review), 15 July 1987. Other requests for this document must be referred to HQ HSD/YAC, Brooks Air Force Base, Texas. ft* O«UAMI'*CATICN/OOWNOAAO4NO tCxCOvtl 4. M A'OAMIKO OAOAHHATIOK At *>AT MUMMAtftl •* MAMi 0» f»A*OAM«l«O OAOANUATIO* Science Applications International Corporation, Life Sciences and Systems Department •4. AOOMM «Cfly. <«*» *« «/ ft*J McLean, Virginia "1 1 MOMlTOAIMO OAOAHKA1IOM A«»OAt *WM»t Alfl - ww*»w Human Systems Division (BSD) * AOOAttt (Crfy. «MW «*« Itf t»4*l . • 22102 Brooks Air Force Base, Texas •>.0f )CttVM*Ok to. NAMC or »g*oi*»0/VO*»OA«M« OAOAMlXATlOM 78235-5000 •. f AOCUACMINT IN»TAuM«Mf iOCNTlVlCATlOH HVMtCA <r/v/««**it> F41689-85-D-0010 1*. tOUACC O* »VNOtNO MOt. MOOAAM »A«J«<T •*. AOOACtt (C«* <MM *»4 *W C«*f •UMCMTMO. 65306F »*4. 2767 TA«« WOA< v*nt — 0003 no. MO M. tiTvt ifMHi* *M«ni» «IM««A««I»«»* An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides. First Pollowup Examination Results ta.MA*OMAl AvTHOAWI G.D- Lathrop, SAIC; S.G. Machado, SAIC; T,G. Karrison, D.of C; H.D. Grubbs, SAIC; W.F. Thomas, SAIC; W.H. Wolfe,' OSAF; J.E. Michalek, OSAF; J.C. Miner, USAF; M.R. Peterson, DSAF. 14. OAK Of At*OAT >V» . »*.. 9«v( 1 «ft >*C« COuiT Vi* IVM o* M^OAT I «}* TIMC eovo»to 1987 July 15 J^ 1,000 Interim 1982-1985 | »MOM 1/85 TO 9/87 M, (WMvtMlNtAAT NOTATION »». JlffkO 06 * • It >W*^tCt TCAMl NT««MiM •• «TM«W if ******* •** **»ltft >| MM* ««««l«r« Epidemiologic Investigation Dioxin Pnenory Herbicides Ranch Hand Herbicidti Orange Air Force Health Study Morbidity CO4ATI COO«* 0»OW* ' «Ut. OA, 05 A A t T A A ^ T I^A«f«AM« Amft Mm^rmtf if •JJJiajri AA^ ***ltf» ft* itaife ««j»U»J This report presents the results of the health assessment of the 1,016 Ranch Hands and the 1,293 Comparisons who participated in the 1985 follovup examination of the Air Force Health Study. The purpose of the study is to determine whether long-term health effects exist and can be attributed to occupational exposure to herbicides. The result showed a subtle but consistent narrowing of medical differences between the two groups since the Baseline study in 1982; however, the Ranch Hands continue to manifest slightly more minor adverse health conditions than the Comparisons. Continued surveillance of these two groups is indicated. The report concludes that there is not sufficient evidence to implicate a causal relationship between herbicide exposure and adverse health in the Ranch Hand group. 1. AMTAACT SCCUAtTV UMClASllfllO/VNllMtTf O 0 lift. N A M C Of ACS»OHSl«l( R.W. Ogershok W.H. Wolfe 00 FOAM 1473. 83 APR *AM( AS A»T. O OTKvUAS O unclassified NVM«A m (512) 536-2274 (5121 536-2604 «0«TiON O' 1 MM n H Ol$0llt(. UNCLASSIFIED IICVAIT* 0» . �CHAPTER 1 BACKGROUND This chapter briefly describes the background of the Air Force Health Study (AFHS) and provides an overview of the study design and purpose of this report. Portions of this chapter have been paraphrased from the Baseline Morbidity Report, 24 February 1984. In January 1962, President John F. Kennedy approved a program of aerial herbicide dissemination, for the purpose of defoliation and crop destruction, in support of tactical military operations in the Republic of Vietnam (RVN). Under this program, code-named Operation Ranch Hand and in operation from 1962 to 1971, approximately 19 million gallons of herbicides were dispersed on an estimated 10 to.20 percent of South Vietnam.1'2 Approximately 11 million gallons of Herbicide Orange, the primary defoliant of the six herbicides utilized in the program, were disseminated. Operation Ranch Hand was the subject of intense scrutiny from the start due to the controversial nature of the program and political sensitivity to chemical warfare charges contained in enemy propaganda. The concerns, which were initially based on military, political, and ecological issues, shifted during 1977 to health issues. Numerous claims of exposure to herbicides, particularly Herbicide Orange and its dioxin contaminant, and subsequent adverse health effects among U.S. military service personnel have resulted in class action litigation and substantial controversy. Social concern for the Herbicide Orange issue continues to be manifest by continuing scientific research, media presentations, congressional hearings, and legal action. The U.S. Air Force Medical Service's concern for the health of Air Force personnel exposed to herbicides was demonstrated in October 1978 when the Air Force Deputy Surgeon General made a commitment to Congress and to the White House to conduct a health study on the Ranch Hand population, the aviators who disseminated the majority of the defoliants in the RVN. The prevailing reasons for the study commitment included the availability of a definitive occupational exposure to herbicides, a sufficient sample size for survey and clinical research, the ability to ascertain the population at risk, and an opportunity for the Air Force Medical Corps to fulfill its adage "we care" to the Air Force community. the Air Force School of Aerospace Medicine, Brooks Air Force Base, Texas, was tasked by the Surgeon General to develop the Study Protocol. In 1982, after extensive peer review, the epidemiologic study began, and the Protocol was published. Since 1978, numerous animal and human studies of dioxin effects have been planned or initiated by governmental agencies, universities, and industrial firms. The key scientific issue in these studies was the extent of exposure, e.g., who was exposed and how much each individual was exposed. Unfortunately, population identification and exposure estimation, which are critical for a valid study of ground troops, have been scientifically elusive. 1-1 �It is believed that of all the military personnel who served in the RVN, the Ranch Hand population was the most highly exposed to herbicides. Exposure estimates indicate that the average Ranch Hand received 1,000 times more exposure to Herbicide Orange and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) during his tour in the RVN than an average male would receive standing unclothed under a spraying aircraft in an open field. Based on the principle of doseresponse, the Ranch Hands should manifest more and/or earlier evidence of adverse health. Thus, the results of the AFHS should serve as an indicator of herbicide effects in ground personnel. STUDY DESIGN The purpose of the study is to determine whether adverse health effects exist and can be attributed to occupational exposure to Herbicide Orange. The study, consisting of mortality and morbidity components, is based on a matched cohort design in a nonconcurrent prospective setting with followup studies. Complete details on the design are provided in the Study Protocol. The nonconcurrent aspect of the design results from the fact that the Ranch Hands were exposed over time between 1962 and 1971. This staggered exposure is accounted for in the design of the studies to address latency considerations. For the Baseline study, the population ascertainment process identified 1,264 Ranch Hand personnel who served in the RVN between 1962 and 1971. By the time the first followup began in 1985, an additional 11 Ranch Hands had been identified, bringing the total Ranch Hand population to 1,275. A Comparison group was formed, consisting of individuals assigned to selected Air Force units with missions of flying cargo to, from, and within the RVN during the same period. Using a computerized nearest neighbor selection procedure, a maximum of 10 Comparisons was selected for each Ranch Hand, matching on age, race, and military occupation. After personnel record reviews, each Ranch Hand who was determined to be eligible and fully suitable for study had an average of 8.2 Comparison subjects. The mortality component addresses mortality from the time of the RVN assignment. A Baseline mortality study was conducted in 1982, and the mortality followup consists of annual mortality updates for 20 years. For the Baseline study and the first four updates, five individuals were randomly selected from the matched Comparison set for a 1:5 design. Subsequent to 1986, the design will be expanded to include all of the individuals in the Comparison set. The Baseline morbidity component, begun in 1982, reconstructed the medical history of each participant by reviewing and coding past medical records. A cross-sectional element, designed to assess the participant's current state of mental and physical health, was based on comprehensive questionnaires and physical examinations given to the participants. For this component of the study, each living Ranch Hand and the first living member of his Comparison set were selected to participate in the examination. Sequential questionnaires, medical record reviews, and physical examinations in 1985, 1987, 1992, 1997, and 2002 comprise the morbidity study followup. 1-2 �PURPOSE The 1985 morbidity followup is the subject of this report. The objective of the morbidity followup is to continue the investigation of the possible long-term health effects following exposure to TCDD-containing herbicides. This report describes the procedures and results of the first morbidity followup of the AFHS. Analysis of reproductive and fertility data will be conducted by the U.S. Air Force and is not part of this report. 1-3 �CHAPTER 1 \ REFERENCES 1. Young, A.L., J.A. Calcagni, C.E. Thalken, and J.W. Tremblay. 1978. The toxicology, environmental fate, and human risk of herbicide orange and its associated dioxin. Technical report OEHL-TR-78-92, USAF Occupational and Environmental Health Laboratory, Brooks AFB, Texas. 247 pp. 2. Buckingham, W.A., Jr. 1982. Operation Ranch Hand: The Air Force and herbicides in Southeast Asia, 1961-1971. Office of Air Force History, United States Air Force, Washington, D,C. pp. 9-69, 199-201. 3. Lathrop, G.D., W.H. Wolfe, R.A. Albanese, and P.M. Moynahan. 1982. Epidemiologic investigation of health effects in Air Force personnel following exposure to herbicides: Study protocol. Technical report 82-44, USAF School of Aerospace Medicine. 172 pp. Available from NTIS, Springfield, Virginia. 1-4 �CHAPTER 2 POPULATION This chapter provides a description of participant selection, the enrollment process, and the demographic characteristics of the population that participated in the clinical and questionnaire portions of the first followup morbidity study in 1985. BASELINE CANDIDATE IDENTIFICATION The study population for the first followup was defined by the Air Force investigators as part of the Baseline study design. Using detailed searches through Air Force and other Government record systems, a total of 1,264 personnel who had participated in Operation Ranch Hand was identified. Using the same historical data sources, a Comparison population of 24,971 individuals that had been assigned to a variety of military cargo missions in Southeast Asia during the same time period was identified. The Ranch Hand and the Comparison populations were matched after all individuals who had been killed in the Vietnam conflict were removed. The matching process was conducted using a computer program employing iterative nearest-neighbor statistical techniques in order to associate each Ranch Hand with 10 Comparisons by race (Black/nonblack), closest date of birth, and occupational category during Vietnam service (officer-pilot, officernavigator, officer-nonflying, enlisted flyer, and enlisted groundcrew). For each Ranch Hand, 1 of the 10 matched Comparisons was selected at random and designated the Original Comparison. The resulting exposed and multiple matched Comparison study design was used for the Baseline effort. During the questionnaire administration of the Baseline study, it was discovered that 18 percent of the Comparison population had been misselected with respect to their Southeast Asia military experience. After eliminating these ineligible Comparisons, the remaining Comparison set was collapsed to a 1:8 study design, which was used for all subsequent eligibility determinations. During the course of the Baseline morbidity study, five new Ranch Hands were verified as eligible for the study and were added to the exposed group. In addition, two Ranch Hands who had been misclassified as-Comparisons were identified during the questionnaire administration. These individuals were reclassified as exposed and new Comparisons were assigned appropriately. Following the completion of the Baseline morbidity study, 10 additional Operation Ranch Hand participants were located and added to the study population for the followup phases. 2-1 �FOLLOVUP CANDIDATE IDENTIFICATION One of the preliminary tasks associated with the followup study was to conduct a telephone survey of uncontacted replacement candidates. The purpose of the survey was to obtain new information on the candidate's general health, economic situation, and willingness to participate in the study. The Air Force address file, assembled and maintained since 1981, provided the basis for the telephone survey contact list, A location algorithm described in Chapter 3 was developed in order to find those individuals no longer at the address and telephone number indicated in the Baseline file. A total of 7,411 candidate replacements out of the candidate file of 7,963 was located, interviewed using computer-aided telephone interview (CATI) techniques, and confirmed as eligible candidate study participants. Of the 552 candidates who could not be interviewed, 26 were deceased, 335 refused, 190 were unlocatable, and 1 respondent had not served in Southeast Asia and was therefore ineligible for the study. Table 2-1 provides the number of candidate participants by Baseline compliance category for the Ranch Hand and Comparison groups. PARTICIPANT SELECTION The participant selection protocol used for the followup was similar to that used at Baseline with one important exception. If the Original Comparison declined to participate, the next randomly ordered candidate for the corresponding Ranch Hand with the same self-perception of health was contacted and recruited for the study. This matching process was not feasible at Baseline because the addresses of the Comparison pool were not fully ascertained. Perception of health was subjectively determined by the candidate during the telephone interview. The rationale for matching replacement Comparisons on self-perceived health status was an attempt to minimize any bias that might result from differential compliance. All candidates who had been contacted and invited to participate during the Baseline, including those who were refusals and partial compilers, were contacted and invited to the followup along with newly verified or located Ranch Hands and their Comparisons. ENROLLMENT The enrollment of candidates was based on the Baseline lists and health status information from the telephone survey. Recruitment was conducted for questionnaire interviews and clinical examinations that began in May 1985 and ended in March 1986. Approximately 70 individuals were examined each week in two groups of 35. A total of 2,309 Ranch Hands and Comparisons participated in both the questionnaire and clinical examination portions of the AFHS followup. Since the followup questionnaire was administered at the physical examination site, there were no "partially compliant" participants at followup. 2-2 �TABLE 2-1. Candidate Followup Participants by Group and Baseline Compliance Status Number Category Candidate Ranch Hands (by Baseline Status) 1,045 Ranch Hands Who Completed Both Baseline Questionnaire and Physical Examination (Fully Compliant) 129 Ranch Hands Who Completed Only Baseline Questionnaire (Partially Compliant) 32 Ranch Hands Who Declined to Take Part in Baseline (Noncompliant) 10 Newly Verified or Located Ranch Hands 1,216 Total Candidate Comparisons (by Baseline Status) 936 Original Comparisons Who Completed Both Baseline Questionnaire and Physical Examination (Fully Compliant) 220 Original Comparisons Who Completed Only Baseline Questionnaire (Partially Compliant) 79 288 Original Comparisons Who Declined to Take Part in Baseline (Noncompliant) Replacement Comparisons Who Completed Both Baseline Questionnaire and Physical Examination (Fully Compliant) 88 Replacement Comparisons Who Completed Only Baseline Questionnaire (Partially Compliant) 49 Replacement Comparisons Who Declined to Take Part in the Study (Noncompliant) 7,411 Replacement Comparisons Who Had Not Been Contacted Previously 9,071 Total 2-3 �Enrollment was managed using an automated scheduling and tracking system to maintain and record all candidate recruitment contacts, actions, and status; clinical examination group scheduling; schedule modifications, cancellations, and completions; and a comprehensive set of logistic management reports. An effort was made to successfully recruit every individual eligible for the study. The number of participants who participated in the physical examination and questionnaire of the first followup is provided in Table 2-2. Of the 1,016 Ranch Hands, all but 53 had matched Comparisons who also participated in the study. Due to the selection strategy used and the recruitment of previous noncompliants, several of the Ranch Hands had multiple Comparisons. -The selection strategy resulted in 79 Ranch Hands having 2 Comparisons, 9 having 3 Comparisons, and 1 Ranch Hand having a total of 5 Comparisons completing the followup. In accordance with the Study Protocol, eligible Comparisons were enrolled without regard to the compliance status of the corresponding Ranch Hand. There were 229 Comparisons in the followup study whose matched Ranch Hand did not participate. PERSONAL CHARACTERISTICS AND HABITS OF FOLLOWUP POPULATION The data on personal characteristics of the Ranch Hand and Comparison individuals were obtained from the followup questionnaire. The areas of tobacco, alcohol, and marijuana use; personal and family income; education; religious preference; active duty/retired/separated status; and risk-taking behavior received particular attention. These variables were examined to assess the similarity of the two groups in social and behavioral characteristics, which were not included in the statistical matching process. The participants in the study were matched on age. The age characteristics of the study population are shown in Table 2-3. The mean and median ages of the Ranch Hand and Comparison groups were nearly identical. The smoking and alcohol-use habits of the study subjects are displayed in Table 2-4. More participating Ranch Hands smoked cigarettes at the time of the followup physical examination than did the Comparisons (40.1% versus 35.OX). This difference in current smoking behavior was statistically significant (p=0.01). In the intervening years since the Baseline examination, 5.6 percent of the Ranch Hands and 4.6 percent of the Comparisons had stopped smoking. The proportions of participants who ever smoked cigarettes, pipes, or cigars were not significantly different in the two groups. Similarly, the number of participants who drank alcohol in the years since 1982 was not statistically different between groups. Data concerning the use of marijuana were gathered by different methods in the two interviews. In the Baseline questionnaire in 1982, confidentiality of response was given to all participants, but answers were identifiable for each participant. At the 1985 followup, random response techniques1 were used on the marijuana questions to overcome the problem of participants either refusing to respond or giving misleading replies to these highly sensitive and personal questions. With this technique, a coin was flipped by the respondent, who then answered either the marijuana question or a neutral unrelated question, which had an answer of known probability. The outcome of 2-4 �TABLE 2-2. Participants Enrolled in the Followup Study by Group and Baseline Compliance Status Number Category Enrolled Ranch Hands (by Baseline Status) 971 Ranch Hands Who Completed Both Baseline Questionnaire and Physical Examination (Fully Compliant) 39 Ranch Hands Who Completed Only Baseline Questionnaire (Partially Compliant) 0 Ranch Hands Who Declined to Take Part in Baseline (Noncompliant) 6 Newly Verified or Located Ranch Hands 1,016 Total Enrolled Comparisons (by Baseline Status) 872 Original Comparisons Who Completed Both Baseline Questionnaire and Physical Examination (Fully Compliant) 61 Original Comparisons Who Completed Only Baseline Questionnaire (Partially Compliant) 10 Original Comparisons Who Declined to Take Part in Baseline (Noncompliant) 12 New Original Comparisons 267 Replacement Comparisons Who Completed Both Baseline Questionnaire and Physical Examination (Fully Compliant) 32 Replacement Comparisons Who Completed Only Baseline Questionnaire (Partially Compliant) 11 Replacement Comparisons Who Declined to Take Part in Baseline (Noncompliant) 28 New Replacement Comparisons 1,293 Total 2-5 �TABLE 2-3. Age (in 1985) of Participants of the Follovup Examination by Group Group Ranch Hand Age Category Number Comparison Percent Number Percent 43 or Less 412 40.6 549 42.5 44 to 62 568 55.9 693 53.6 63 or More Total 1,016 3.9 51 3.5 36 100.0 1,293 100.0 Group Ranch Hand Comparison Range 35-72 Years 35-77 Years Mean 46.9 Years 46.8 Years Median 47 Years 46 Years the coin flip was unknown to the interviewer. Thus, no given reply could be traced, although the proportion of the population that had the sensitive characteristic (marijuana use) could be estimated. There were no statistically significant differences between the Ranch Hand and Comparison groups in the reported use of marijuana in the 30 days preceding the examination (7.8% and 9.2%, respectively). A much higher percentage, 26.3 percent of the Ranch Hands and 31.0 percent of the Comparisons, reported smoking marijuana at some time in the past. At Baseline, only 5.1 percent of each group reported ever using marijuana. These differences over time were most likely due to a greater sense of confidentiality generated by the random response techniques used in the 1985 questionnaire. The mean usage levels of tobacco and alcohol among those participants who did indulge in these habits are shown in Table 2-5 as pack-years, cigaryears, pipe-years, or drink-years. Mean alcohol use per day was 6.26 drinks per day for the Ranch Hands and 6.42 for the Comparisons. In most of the cumulative measurements, the median level of use was lower than the mean level, indicating that the heavy users of these substances skewed the distributions. Eighty-nine percent of both groups reported having consumed alcohol since the last physical examination. Differences in these calculated variables might have been due to either actual changes in behavior or to differences in the questionnaires used to collect the basic data. 2-6 �TABLE 2-4. History of Tobacco and Alcohol Use of Participants of the Followup Examination by Group Group Comparison Ranch Hand Habit Yes Percent No Percent Yes Percent No Percent p-Value Current Use of Cigarettes 407 40.1 609 59.9 453 - 35.0 840 65.0 0.01 Past History of Cigarettes 752 74.0 264 26.0 944 73.0 349 27.0 0.58 Past History of Cigar Use 249 24.5 767 75.5 345 26.7 948 73.3 0.24 Past History of Pipe Use 265 26.1 751 73.9 340 26.3 953 73.7 0.92 Past History of Marijuana Use* 26.3 73.7 31.0 69.0 0.15 Marijuana Use* within Past 30 Days 7.8 92.2 9.2 90.8 0.52 88.7 146 11.3 0.98 Use of Alcohol since Last Interview 901 88.7 115 11.3 1,147 ^Estimates based on random response technique. 2-7 �TABLE 2-5. Average Use of Tobacco Products and Alcohol for Those Reporting Use of These Substances: Participants of the Followup Examination by Group Group Ranch Hand Comparison Mean Median Mean Median 26.54 25.00 25.77 25.00 Cigarettes, Pack- Years (Cumulative) 17.69 13.00 17.61 13.00 Cigar-Years (Cumulative) 11.25 1.30 10.96 1.00 Pipe-Years (Cumulative) 20.03 6.10 16.90 4.00 Alcohol Drinks per Day (Current Use) 6.26 6.00 6.42 5.00 Drink-Years (Since Last Interview) 1.81 0.80 1.89 0.74 26 .59 12 .80 25 .04 13.00 Substance Cigarettes per Day (Current Use) Drink-Years (Cumulative) Educational background and religious preference for the two groups are presented in Tables 2-6 and 2-7. The current military status of each individual was classified as active duty, retired, separated, reserve duty, or deceased. There were no significant differences between the two groups. These data are presented in Table 2-8 and showed equivalence of the two groups in these social variables. Data on income were collected in a categorical form, and the median income levels of the Ranch Hand and Comparison groups were comparable. The median personal income in both groups was in the $25,000 to $30,000 range, and the median total family income ranged from $40,000 to $45,000 in each group. Risk-taking behavior patterns of the study population were assessed by a series of questions that emphasized participation in potentially dangerous recreational activities. These data are summarized in Table 2-9. In motorvehicle racing (automobiles, boats, and motorcycles) and scuba diving, there were group differences of borderline significance (p=0.07 and p=0.09, respectively). Slightly more Comparisons were scuba divers (12.4% versus 10.1%), and more Ranch Hands raced motor vehicles (12.9% versus 10.4%). There was a significant difference in scuba diving at Baseline (p=0.04), when more Comparisons were scuba divers (12.7% versus 9.9%). 2-8 �TABLE 2-6. Educational Background of Participants of the Follovup Examination by Group Group Ranch Hand Comparison Number Percent Number High School/GED 522 51.4 655 Associate Degree 84 8.3 114 8.8 BA/BS Degree 194 19.1 271 21.0 Graduate Degree 203 20.0 239 18.5 13 1.3 14 1.1 Educational Level Unknown Percent 50.7 p=0.64 TABLE 2-7. Religious Preference of Participants of the Follovup Examination by Group Group Comparison Ranch Hand Religious Preference Number Percent Number Protestant 671 66.0 856 66.2 Catholic 215 21.2 281 21.7 9 0.9 15 1.2 Other 37 3.6 54 4.2 None 84 8.3 87 6.7 Jewish p=0.60 2-9 Percent �TABLE 2-8. Military Status of Participants of the Followup Examination by Group Group Ranch Hand Military Status Comparison Number Percent Number Percent 89 8.8 118 9.1 Retired 553 54.4 683 52.8 Separated 313 30.8 420 32.5 55 5.4 65 5.0 6 0.6 7 0.5 Active Duty Reserve Forces Deceased* p=0.90 "Died after the followup examination. These data reflected the overall equivalence of the two groups in social and behavioral characteristics. The differences observed when these data were contrasted to similar data at Baseline might have reflected differences in data collection methods or slight changes in the cohorts rather than changes in behavior among group members. LONGITUDINAL LOSSES AND GAINS A total of 2,269 Ranch Hands and Comparisons was fully compliant with the Baseline study. The study population of 2,309 for the followup included a loss of 159 participants and the addition of 199 individuals. Loss to the followup occurred either because the participant was deceased, refused to participate, or was unlocatable. The loss to followup was 7 percent in both the Ranch Hand and Comparison groups. Of the 69 Comparisons lost to the followup study due to refusal or inability to locate, 17 were replaced. For the remaining 52, no replacement who satisfied the health status matching criterion and was willing to participate was identified from the candidate replacements. The categories of these individuals are provided in Table 2-10. A total of 199 new participants were recruited into the study based on the selection methodology used. Information on the new participants is provided in Table 2-10. 2-10 �TABLE 2-9. Risk-Taking Behavior of Participants of the Follovup Examination by Group Group Ranch Hand Activity Yes Percent Comparison No Percent Yes Percent No Percent p-Value Scuba Diving 103 10.1 913 89 .9 160 12.4 1,133 87.6 0.09 Auto, Boat, or Motorcycle Racing 131 12.9 885 87 .1 135 10.4 1,158 89.6 0.07 Acrobatic Flying 43 4.2 973 95 .8 43 3.3 1,250 96.7 0.25 Sky Diving 22 2.2 994 97 .8 32 2.5 1,261 97.5 0.62 Hang Gliding 11 1.1 1,005 98 .9 14 1.1 1,279 98.9 1.00 Mountain Climbing 82 8.1 934 91 .9 102 1,191 92.1 0.86 Surfboard Riding 81 8.0 935 92 .0 91 7.0 1,202 93.0 0.40 Long-Di stance Sailing 54 5.3 962 94 .7 55 4.3 1,238 95.7 0.23 Fast Downhill Skiing* 170 16.7 846 83.3 184 14.2 1,108 85.8 0.10 p=0.10 *0ne Comparison was unwilling to respond. 2-11 7.9 �TABLE 2-10. Losses/Gains of Participants Between the Baseline and Followup Examinations Losses Category Number 5 Ranch Hands Deceased Ranch Hand Refusals Ranch Hands Unlocatable 74 Total Ranch Hands Lost 16 Comparisons Deceased 55 14 Comparison Refusals Comparisons Unlocatable 85 Total Comparisons Lost 10 59 Gains Category Number 39 6 Ranch Hands Partially Compliant at Baseline Newly Verified or Located Ranch Hands 45 Total Ranch Hands Added to Study 61 Partially Compliant Original Comparisons at Baseline Partially Compliant Replacement Comparisons at Baseline Newly Selected Original Comparisons (For Newly Verified Ranch Hands) Replacements for Compliant Comparisons Who Refused Followup Noncompliant Original Comparisons Who Agreed to Attend Followup Noncompliant Replacement Comparisons Who Agreed to Attend Followup Original Comparison Not Locatable at Baseline but Found at Followup Replacement Comparisons Not Locatable at Baseline but Found at Followup Replacement Comparisons Not Contacted at Baseline 32 11 16 10 11 1 3 9 154 Total Comparisons Added to Study 2-12 �SUMMARY Participants were recruited for the first followup in accordance with the Study Protocol. All participants (Ranch Hands and Comparisons) who were contacted for enrollment at Baseline were recruited for this phase of the study. Newly verified and located Ranch Hands, since Baseline, and their respective Comparisons were invited to join the study. Due to refusals among the Comparisons,'replacements from the previously uncontacted Comparisons were selected for enrollment. The replacements were matched to the refusing Comparisons on self-perception of health; health status data were obtained in the telephone survey. Personal characteristics of the two groups were compared, based on data obtained from the followup questionnaire. Contrasts of age, educational background, religious preference, current military status, and income revealed no significant differences between the Ranch Hand and Comparison groups. Significantly more Ranch Hands smoked cigarettes at the time of the followup examination than did Comparisons, although there were no significant differences found for past history of cigarettes, cigars, or pipe use or for recent or past use of marijuana. A much higher percentage of both groups reported smoking marijuana at some time in the past at the followup than at Baseline. This difference was most likely due to a greater sense of confidentiality generated by the random response techniques used in 1985. The use of alcohol since the Baseline examination was not significantly different between the two groups. The difference in the risk-taking behavior patterns of the Ranch Hands and the Comparisons was marginally significant. Slightly more Ranch Hands than Comparisons raced motor vehicles, and more Comparisons were scuba divers. The followup study population included the loss of 159 participants (74 Ranch Hands and 85 Comparisons) who were fully compliant at Baseline and the addition of 199 participants (45 Ranch Hands and 154 Comparisons). The 199 newly examined study subjects consisted of 132 participants (39 Ranch Hands, 61 Original Comparisons, and 32 replacement Comparisons) who were partially compliant at Baseline, 21 participants (10 Originals and 11 replacements) who refused at Baseline, and 46 participants (6 Ranch Hands, 12 Originals, and 28 replacements) who were new to the study. Thus, the study population for the first followup of the AFHS consisted of 2,309 individuals: 1,016 who had been associated with Operation Ranch Hand and 1,293 Comparisons. 2-13 �CHAPTER 2 REFERENCES 1. Greenberg, B.C., A-L.A. Abdul-Ela, W.R. Simmons, and D.G. Horvitz. 1969. The unrelated question randomized response model: Theoretical framework. J. Am. Stat. Assoc. 64(326)-.520-539. 2-14 �CHAPTER 3 QUESTIONNAIRE METHODOLOGY This chapter discusses the development and the implementation of the questionnaires used in the study: the participant interval questionnaire, the spouse interval questionnaire, the Baseline participant and spouse questionnaires, and the telephone survey of previously uncontacted Comparisons. The participant interval questionnaire was designed to capture the study participant's health history in the 3 years since his participation in the Baseline study. Data collection was comparable to the Baseline effort: The questionnaire was very similar, and it was administered using the same faceto-face methodology to virtually the same population. In the Baseline study, interviews were conducted in the participants' homes and the followup interview was conducted at the physical examination site. The revised methodology was more efficient and better subject to quality control. The spouse interval questionnaire collected reproductive data similar to those collected at Baseline from spouses for the interval since Baseline. The spouse interval questionnaires were mailed to the spouses to be selfadministered, or were completed in La Jolla, California, if the spouse accompanied the participant to the physical examination site. Analysis of the spouse data is not included in this report. Since some study subjects refused to participate in 1982 and other participants were new to the study, Baseline questionnaires were administered to these new participants and their spouses. The same procedures used at « Baseline were used to administer the Baseline questionnaires in the homes of these individuals. The elements of each questionnaire are identified in Table B-l of Appendix B. Questionnaire development and administration and scheduling of participants were conducted by the National Opinion Research Center (NORC), a social science research center at the University of Chicago. QUESTIONNAIRE DEVELOPMENT The goal of questionnaire development was to maintain to the maximum extent possible the question wordings, context, and procedures that were used in the 1982 Baseline study. The largest task of questionnaire development was asking for interval histories on crucial questionnaire items to update the information provided by the 1982 Baseline questionnaires. For the participant interval questionnaire, new questions were also developed on risk factors for skin cancer, since the Baseline Morbidity Report found Ranch Hands to have an excess of nonmelanoma skin cancer.1'3 Because the chemical constituents of Herbicide Orange had not previously been associated with skin cancer in the literature, no questions had been included in the Baseline participant questionnaire to collect information on risk factors for this condition. 3-1 �New questions were added to determine personality type, since Type A behavior is associated with coronary heart disease. The Jenkins Activity Scale was administered to collect these data. Enhancements were also made to improve data collection for birth defects, smoking habits, and drinking habits. A copy of the participant interval questionnaire is provided in Appendix B. An information sheet containing a computer-generated summary of key respondent answers to the Baseline survey was used to provide bounded recall for participants. Even when given a precise "starting date," respondents frequently repeat information given earlier, neglect to report new information because they thought they had previously reported it, and otherwise misplace events in time or forget them completely. The best means of preventing such errors is through the use of bounded recall, in which the respondent is reminded of information he has already reported and new information is sought with reference to an updated information sheet. Among the data elements included were date of birth, highest educational degree, military status at last interview, marital status at last interview, and name of spouse. The questionnaire was pretested on 8 ineligible individuals who had been interviewed during Baseline, and on 10 men who participated in the pretest examination. INTERVIEWER TRAINING Twelve interviewers were recruited and trained by NORC's field management and Chicago office staffs in May 1985 to administer the interval questionnaires. The onsite NORC interview staff was not informed of the exposure status of any study participant either before or after contract completion. The site supervisor reported to the Project Director in Chicago on a weekly basis, and quarterly visits were made to the site by the Director. The site supervisor observed a sample of interviews, at least one per interviewer per week, and reviewed and edited interview questionnaires before shipping them to Chicago for further processing. In early 1985, personal interviewers were recruited to conduct Baseline interviews for new participants in their homes. The interviewers were trained in the Chicago NORC office, using questionnaires and procedures established for the Baseline survey. They were supervised by an assistant survey director in the NORC office, who edited each completed questionnaire and talked with each interviewer weekly. TELEPHONE SURVEY The telephone survey of uncontacted Comparisons was intended to gather data on the general health status of the 7,963 replacement candidates for the active Comparison group. The sample consisted of men who served in C-130 units in Southeast Asia between 1962 and 1971, but who did not participate actively in the Baseline phase of the study. A total of 7,411 cases (93%) was completed by NORC computer-assisted telephone interviewers. The telephone survey was conducted prior to the scheduling of the physical examinations. 3-2 �The key question asked was, "Compared to other people your age, would you say that your health is...excellent, good, fair, poor?" Other questions asked about current medications, severity of illness or injury during the last 6 months, and income. Locating and refusal conversion algorithms similar to the Baseline data collection efforts were used. The data from the telephone survey of uncontacted Comparisons were used to select a replacement whose self-reported health status matched that of the noncompliant Comparison. If a willing replacement was not found by this method, the perception of health status variable was dichotomized into excellent/good versus fair/poor, and a new replacement was selected from the Comparison set. If this second attempt at identifying a suitable replacement failed, no replacement was made. The selection procedure is provided in Figure 3-1. In this example, the first randomly ordered Comparison was contacted but refused to participate. In the second attempt, the Comparison was deceased. The third Comparison volunteered to participate in the morbidity study. SCHEDULING OF PARTICIPANTS NORC recruited and trained four schedulers to perform the initial contacts with study subjects. Their training included background information on the details and purpose of the study, simulation of the actual scheduling of calls, documentation of results, and conversion of refusals. An initial letter was sent by the Air Force to each study subject, informing him of the upcoming interval physical examination. The NORC scheduler then followed this letter with a call to attempt to schedule the participant. Refusals occurred at a number of steps in the scheduling process. A team of conversion specialists was assigned to contact refusing study subjects and attempt conversions. Help in conversion was also received from individuals in the U.S. Air Force School of Aerospace Medicine and the Ranch Hand Association. Many more participants were scheduled, but due to "no-shows" at the examination site, and passive refusals who rescheduled frequently, the final figure stood at 2,309. The Baseline interviewer contacted the potential study participant by telephone for scheduling the in-home interview. Toward the end of the physical examination phase, the Baseline questionnaire was administered at the examination site by one of the interviewers who had been trained in administering that questionnaire. Of the 106 participant Baseline questionnaires administered during the first followup, 21 had to be conducted at the examination site. The supervisor of the Baseline interviewers conducted the locating efforts for new and interval participants. Procedures similar to those used in 1982 were followed: a postal search, followed by a local telephone directory search, a motor vehicle registration search, and personal locating efforts in the area of last known residence when appropriate. The Air Force also provided locating support through its records. DATA COLLECTION Upon arrival at the Scripps Clinic and Research Foundation (SCRF), the participant received a schedule including the time and place for the interval interview, and a race-matched interviewer was appointed to conduct the 3-3 �Comparison Individuals (Randomly Ordered) Randomly Selected Mortality Controls Matched Ranch Hand t — + * ** 111 1:1 Unwilling Deceased Volunteered Replacement Candidates Figure 3-1. Selection Procedure for the Questionnaire, Physical Examination, and Followup Study 3-4 �interview. Because of scheduling problems and the unavailability of a Black interviewer, 65 of the 143 Black study participants were interviewed by whites. As in all of the personal interviews for the AFHS, interviewers were required to ask questions exactly as written, were not allowed to interpret questions or inject personal commentary, and were not allowed to skip between sections of the questionnaire. They were also instructed to probe "don't know" answers at least once. During the interview, medical record release forms were signed. The respondent was also asked to give the current name and address for each former spouse listed in the questionnaire, so that spouse questionnaires could be mailed to these individuals. The spouse interval survey was mailed to current spouses at the time the study subject was at the SCRF. Two NORC Chicago telephone interviewers were trained to prompt refusing spouses to return the questionnaire, or to administer the spouse interview by telephone as part of the refusal conversion effort. If the spouse also traveled to La Jolla, the questionnaire was completed under the supervision of a site interviewer. Of the 1,898 completed spouse interval questionnaires, 1,066 were returned by mail, 348 were completed by telephone, and 484 were completed in La Jolla. DATA PROCESSING All completed interviews were sent to the NORC Chicago office following editing by the site supervisor, who retrieved missing data from study subjects while they were still onsite; any further retrieval of critical items was conducted from the Chicago office through telephone contacts. Critical items were those for which missing data were unacceptable. The questionnaires were coded for data entry by a staff of five coders who received a week of training on the various AFHS instruments. Data entry was programmed to provide value and range checks as the data were being entered, to perform logic checks and arithmetic checks, to flag important missing items, and to verify the key entry of 10 percent of each questionnaire. Then the data were run through an automated cleaning program to detect a wide range of data errors that were corrected by pulling the hard copy questionnaires and reviewing each situation on a case-by-case basis. No changes were ever made in the hard copy data; corrections were entered into the data tape, and the tape was run against the cleaning program until no errors were detected. 3-5 �CHAPTER 3 REFERENCES 1. Vitaliano, P., and F. Urbach. 1980. The relative importance of risk factors in nonmelanoma carcinoma. Arch. Dermatol. 116:454-456. 2. Stern, R.S., and K. Momtaz. 1984. Skin typing for assessment of skin cancer risk and acute response to UV-B and oral methoxalen photochemotherapy. Arch. Dermatpi. 120:869-873. 3. Scotto, J., and T.R. Fears. 1978. Skin cancer epidemiology: Research needs. Natl. Cancer Inst. Monogr. 50:169-177. 4. Jenkins, C.D., R.H. Rosenman, and M. Friedman. 1967. Development of an objective psychological test for the determination of the coronary-prone behavior pattern in employed men. J. Chronic Pis. 20:371-379. 3-6 �CHAPTER 4 PHYSICAL EXAMINATION METHODOLOGY The first followup examination was provided to four categories of individuals: those who had taken the Baseline questionnaire and Baseline physical examination; those who had been invited to the Baseline events but chose not to participate, only took the questionnaire, or were unlocatable; those Comparisons who had not been invited previously, but who were selected as replacements for Baseline Comparisons noncompliant to this followup examination; and the six newly identified Ranch Hands. . As noted in the Baseline Report, all potential study participants were verified as eligible for the AFHS following a detailed review of military personnel records. Replacement individuals were carefully selected, by matching data on the self-perception of health from the noncompliant Comparison (obtained from the telephone survey) with those of the replacement candidate (see Chapter 3 for details). The followup examination differed logistically from the Baseline examination in one significant way: All structured interval questionnaires were administered at the examination site as contrasted to the in-home interviews conducted at Baseline. The followup examination consisted of the following major elements: • Interval Questionnaire • Combat Experience Questionnaire • Review-of-Systerns Questionnaire • Psychological Testing • Physical Examination • Specialized Testing, e.g., Doppler Arterial Studies • Laboratory Testing • Psychological and Medical Outbriefings. Details of the above examination elements were carefully prescribed by the Air Force and set forth as contractual requirements. Clinical innovations or variations were neither desired nor authorized; all proposed examination procedural changes were reviewed in detail by Air Force technical and contractual personnel. An important objective of the technical review was to ensure that bias was not created, by any procedural change. The requirement to maintain blind examinations was particularly stringent: The clinical staff was prohibited from knowing or seeking information as to the group identity (Ranch Hand, Comparison) of any participant. At the end of the examination, each participant was asked to note on the critique form whether such information was sought by any member of the clinical or paramedical staff. 4-1 �EXAMINATION CONTENT Examination content was designed by the Air Force to emphasize detection of medical endpoints suspected of being associated with exposure to phenoxy herbicides, chlorophenols, or dioxin. In addition, findings in the Baseline examination were used by the Air Force to direct changes in the followup examination (e.g., abnormal pulses at Baseline suggested the need for Doppler measurements at the followup). The general'content of the physical examination and psychological test battery is shown in Table 4-1, and the complete laboratory test series is displayed in Table 4-2. Quality control requirements for both laboratory testing and clinical procedures were extensive. Although details are provided in Chapter 6, the following categories provide an overview of the extent of the quality emphasis. For laboratory testing, single reagent lots and control standards were used when practical, duplicate specimens were routinely and blindly retested, testing overlaps were mandatory when test reagents required change, and fast initial response cumulative statistical techniques (FIR CUSUM) were used to detect rapidly any subtle test drift over time. In addition, 50 specimens from the Baseline serum bank were retested to assess the comparability of laboratory methods. The SCRF clinical team was carefully instructed to assure clinical quality. The quality control elements included: a pretest of the examination process; detailed clinical inspection techniques by SCRF, Science Applications International Corporation (SAIC), and Air Force physicians and personnel; preprinted mark-sense examination forms; clinical quality assurance meetings to detect and correct problems; and blindness of exposure status at the examination. In addition, participant rapportbuilding techniques were added to boost participation in future followup studies, such as participant critique forms and recreational opportunities afforded to the accompanying family members. CONDUCT OF EXAMINATIONS All examinations were conducted at SCRF, La Jolla, California, from May 1985 to March 1986. Except for weeks with national holidays, two groups of participants, averaging about 32 per group, were examined weekly. Midway through the study, NORC recruiters noted that a number of participants refused the examination because of weekday business commitments or because of single-parent responsibilities. Consequently, two special weekend examinations were arranged late in the examination cycle, and many of the former refusals were then able to attend. The examination was identical to the regular 2 1/2-day process, except that it was compressed into 2 days by reducing the number of participants in a group. The logistics effort required in contacting, transporting, and examining 2,309 study members was formidable. Preexamination contacts consisted of the telephone health survey, telephone recruitment to the examination if necessary, and calls by either the NORC scheduling specialists or by the travel agent to arrange transportation and determine whether special requirements existed (e.g., wheelchair assistance, weekend examination schedule). Once scheduling was reasonably firm, the SAIC logistics coordinator sent each participant a detailed information package outlining dietary requirements, inbriefing schedules, important telephone numbers, a request for medical records, and local maps designating examination-site eating and recreational facilities. 4-2 �TABLE 4-1. Elements of the Followup Physical Examination Elements Remarks General Physical Examination Internist Neurological Examination Neurologist Dermatological Examination Dermatologist Electrocardiogram Resting, 4-Hour Fasting and Nicotine Abstinence Doppler Peripheral Arterial Blood Flow Studies 4-Hour Nicotine Abstinence Chest X Ray Immunological Studies 50% Random Sample Skin Test Studies 75% Sample Psychological Evaluation: Minnesota Multiphasic Personality Inventory (MMPI) Cornell Medical Index Halstead-Reitan Battery Patient Outbriefing and Discussion of Individual Results 4-3 Medical Diagnostician, Internist, and Ph.D. Psychologist �TABLE 4-2. Laboratory Test Procedures of the Follovup Physical Examination Clinical Laboratory Fasting Glucose Blood Urea Nitrogen (BUN) Cholesterol HDL Cholesterol Triglyceride Serum Glutamic-Oxaloacetic Transarainase (SCOT) Serum Glutamic-Pyruvic Transaminase (SGPT) Gamma-Glutamyl Transpeptidase (GGTP) Alkaline Phosphatase Lactic Dehydrogenase (LDH) Thyroid Stimulating Hormone (TSH) Initial Cortisol 2-Hour Cortisol Prothrombin Time Quantitative Immunoglobulins Complete Blood Count (CBC) Leuteinizing Hormone (LH) 2-Hour Postprandial Glucose Creative Phosphokinase (CPK) Total Bilirubin Direct Bilirubin Total Protein Protein Electrophoresis Routine Urinalysis T3 % Uptake T 4 Testosterone Hepatitis B Surface Antigen Hepatitis B Surface Antibody Follicle Stimulating Hormone (FSH) Rapid Plasma Reagin (RPR) Porphyrins (Mayo Clinic) Sedimentation Rate Immunological Laboratory Cell Surface (Phenotype) Analyses Lymphocyte Mitogen Stimulation Assays Mixed Lymphocyte Culture (MLC) Natural Killer Cell Assay by Specific Cellular Cytotoxicity Using K-562 Target Cells Natural Killer Cell Assay (Using Interferon) by Specific Cellular Cytotoxicity Using K-562 Target Cells 4-4 �The logistical flow of the entire examination process was complex. Figures 4-1 and 4-2 outline participant flow for the first 2 examination days. As depicted in these figures, each group of participants (generally containing equal numbers of Ranch Hands and Comparisons) was transported early in the morning to SCRF on the first 2 days in a fasting state; tobacco, alcohol, and coffee abstinence were also required. Following initial inbriefing and blood draw on the first day, each participant was randomly assigned to the examination group or to the psychological testing group. On the second day, these groups were reversed. After randomization, each member was given an individualized 3-day schedule outlining his medical, interviewing, and laboratory appointments. The schedule carefully noted the specific required periods of fasting and tobacco abstinence (see Figures 4-1 and 4-2 for generalized periods in relation to ECG and Doppler testing). Each individual was reminded of the fact that all aspects of the examination were strictly voluntary, and that refusals would be honored without question. Both general and specific consent forms (e.g., skin biopsy), approved by the Air Force, were explained in detail. In contrast to the Baseline examination, great reliance was placed upon each individual to find the appropriate clinic area at his scheduled time. This approach had great appeal to this self-reliant population as evidenced by critique feedback. Throughout the examination day, generous time was provided for waiting-room activities, i.e., renewal of past friendships, discussions of the Vietnam War, consumption of refreshments when permitted, and completion of paperwork. Day 3 of the examination was largely spent in finishing up the specialty examinations and receiving the outbriefings from a psychologist and medical diagnostician. Only upon completion of these important debriefings were the participants paid their stipend, reimbursed for travel expenses, and transported to the airport. As noted previously, the SCRF clinical team was hand-picked for participation in this project. In total, 15 board-certified physicians in internal medicine, neurology, and dermatology participated in the general, specialty, and diagnostic examination. To reduce observer variability, turnover in the clinical or paramedical staffs was minimized during the 9 months of examinations. One SCRF physician served as the Project Medical Director, responsible for the scheduling, conduct, and quality control of the examinations. All examining physicians were introduced to the mark-sense examination forms during the pretest examination. The layout of the form was designed to parallel the flow of the clinical examination so as to minimize recording errors. Because data transcription was not permitted, each physician was responsible for filling in the bubbled form. To a large extent, these marksense forms and subsequent quality control were the primary reason for a remarkably clean data set. Two examples of the mark-sense forms are presented as Figures 4-3 and 4-4; a complete set of forms is provided in Appendix C. For the first followup, the special testing included Doppler tests, delayed hypersensitivity skin tests, and immunological tests. Doppler measurements were obtained on all participants by highly experienced technicians; results were recorded and Polaroid photographs were taken of representative oscilloscope displays. As previously noted, considerable emphasis was placed upon tobacco abstinence prior to Doppler evaluations. Skin tests for four antigens were administered in a standardized manner: Candida (1:1,000 weight/volume, 0.1 ml intradermal), mumps (2 complementfixing units), Trichophyton (1:1,000 weight/volume, 0.1 ml intradermal), and 4-5 �Day One PM AM Group A: Physical Examination tKIKttdl I !*=!• 1 1 o.» 1 £«», 1 k r I NOHC lillefvww. xHdv 1® I1 N< Wailing Room (Group Rappon) 1 1 HdhhMt) r-"*" 1 MMPl 1 NOHC XHdv Waiting Room \/ J^ Kay: Slaggeiad G*uup Movement Schectuled Movanwni by Indnndual 4 Hours Post PiattchaJ and Tobacco Abstinence Figure 4-1. Flow Diagram of Day One Followup Interview and Physical Examination 5 • Medical Videos a Hi o I 1 1 111< Group B: Psychological Testing Group Movement •„ • Refreshments Ail L*> 1 1 1 CtXIMH /\ it co..- ir ,^, , , II , >*s. �CHAPTER 5 STUDY SELECTION AND PARTICIPATION BIAS INTRODUCTION AND BASELINE SUMMARY The Protocol During the design phase, the authors of the Protocol anticipated that loss to followup would pose the greatest threat to the validity of the study. In particular, they expected differential compliance with relatively more Ranch Hands self-selecting themselves into the study than Comparisons and with health differences of unknown character between noncompliant Ranch Hands and noncompliant Comparisons. As a partial correction, the study design specified that noncompliant Comparisons would be replaced by Comparisons having the same values of the matching variables and the same health perception. In this way, the replacement Comparisons would serve as surrogates for those Comparisons who refused to participate. This, in turn, would tend to reduce the bias due to noncompliance in the Comparison group and would have the added advantage of maintaining this group's sample size. The Comparison in each randomized matched set who happened to be first asked to participate in the Baseline questionnaire and physical examination was identified as the Original Comparison for his respective Ranch Hand (in accordance with the Protocol). If the Original Comparison was noncompliant, that is, he refused to take the Baseline questionnaire or physical examination, his replacement was called a replacement Comparison. Replacement Comparisons were so distinguished to satisfy the Protocol requirement.that they be contrasted with the noncompliant Comparisons, also called refusals, they replaced. No corresponding replacement strategy for the Ranch Hands was possible since all Ranch Hands had been identified and invited to participate. The Protocol further specified that the replacements would be statistically compared with the noncompliant Original Comparisons to determine the extent to which the replacement strategy was being realized. The statistical contrast of replacements and refusals was to be based on responses to a noncompliance telephone questionnaire administered to refusals and to their potential replacements. This questionnaire assessed self-perception of health, days lost from work due to illness, and medication use, and was to serve as the basis for the health matching called for in the Protocol. Although the Protocol was not explicit on this point, it implied that the decision to include or exclude the replacements from the study would be based only on this contrast. The Baseline Replacement Operation The health-matching questions (identical to the noncompliance questionnaire) were, in fact, hot administered to any potential replacement 5-1 �Comparison before selection at Baseline, although questions regarding self-perception of health, medication use, and work loss were asked as part of the Baseline questionnaire after entry into the study. The noncompliance telephone questionnaire was offered to nohcompliant study participants, but only 79 completed the telephone questionnaire, and of these only 57 were actually replaced. Replacements were, therefore, not health matched to refusals at Baseline. Rather, they were matched only on the basic matching variables: date of birth, race, and occupation. The statistical contrast of refusals and their replacements was not performed at Baseline. During the scheduling operation at Baseline, two untoward events occurred that led to the identification of two additional categories of Comparisons, shifted Comparisons and Air Force-interviewed replacements. First, 212 of the Original Comparisons were discovered to be ineligible for participation in the study due to errors in the data base regarding their unit of assignment in Southeast Asia. These men had not served in Southeast Asia but, due to a duplication of codes, were mistakenly included in the Comparison population. They were deleted from the study. This resulted in another Comparison in each previously randomized match set being first asked to participate in the study. These new Original Comparisons were figuratively called "shifted" Comparisons, labeled S in the Baseline Report, to describe the effective movement of these Comparisons- in each matched set to fill the space left by the removed ineligible Original Comparison. The eligible Original Comparisons were labeled 0 in the Baseline report. Shifted Comparisons are more accurately referred to here as shifted Original Comparisons to emphasize that they are not replacement Comparisons and that they are the legitimate Original Comparisons for their respective Ranch Hands. Shifted Original Comparisons are not replacement Comparisons because their invitation to participate in the study was not the result of a previous refusal of another Comparison in their respective matched sets. Shifted Original Comparisons were identified to reflect concern that the process by which Comparisons were determined ineligible may not have distributed ineligible Comparisons uniformly. Second, 30 replacement Comparisons were interviewed by Air Force staff rather than by the contractor. These replacements were labeled A. All other replacement Comparisons, labeled R, were simply called "replacements." The removal of ineligible Comparisons from the study caused a pause in the scheduling operation that delayed the scheduling of the shifted Original and replacement Comparisons relative to that of the Original Comparisons. This scheduling delay is apparent in Figures V-3 and V-4 in the Baseline Report. Some study investigators speculated that this scheduling slip might cause shifted Original Comparisons and replacement Comparisons to self-select differently from Original Comparisons. Statistical analyses in Chapter V of the Baseline Report and further unpublished analyses following the release of the Baseline Report investigated the effect of this scheduling problem. The Baseline Selection Bias Analyses Since replacements were not health matched at Baseline to their corresponding noncompliant Comparisons and since differential scheduling opportunity may have created self-selection biases, statistical contrasts of the various Comparison groups were done at Baseline. In particular, the Comparisons labeled 0, S, R, and A were contrasted on the basis of self-perception of health, medication use, work loss, and five clinical variables. 5-2 �The results of these analyses suggested to some investigators that shifted Original Comparisons were not statistically distinguishable from Original Comparisons and that shifted Original Comparisons were not statistically different from replacements, but that replacement Comparisons appeared to be statistically different from Original Comparisons. The 30 Air Force-interviewed replacement Comparisons were not statistically distinguishable from other replacement Comparisons and were not investigated further as a group. Since opinions differed among Air Force principal investigators and statisticians, a management decision was reached to use only the Original Comparisons in the primary analyses and to contrast Ranch Hands with all Comparisons in the appendix of the Baseline Report. The reader is referred to Chapter V of the Baseline Report for additional detail. In retrospect, the concern with statistical distinguishability between replacement Comparisons and Original Comparisons is difficult to justify, since the only valid question regarding the replacements is their similarity to the refusals whom they replaced. The Baseline Compliance Bias Analyses Telephone questionnaire data obtained from the 57 noncompliant Comparisons, who were replaced, were not analyzed in the Baseline Report. Instead, compliance bias was analyzed by contrasting partially compliant with fully compliant participants, with adjustment for group (Ranch Hands, 0, S, R, A). These analyses were based on data from the Baseline questionnaire regarding self-perception of health, medication use, work loss, anger, anxiety, erosion, depression, liver ailments, miscarriages, and acne. Results suggested that partially compliant participants were statistically different from fully compliant participants for some of these variables. Based on these results, calculations were presented to suggest that the noncompliance bias could produce an error in relative risk of 25 percent, either overestimating or underestimating the risk, and a spurious mean shift of up to 8 percent in either direction. THE FIRST FOLLOVUP SCHEDULING AND REPLACEMENT OPERATION Matching of replacements to noncompliant Comparisons on the basis of health status was initiated with the first followup scheduling operation. This was accomplished by administering a short telephone questionnaire to all previously uncontacted Comparisons and then using their health status responses to select from among the Comparisons in a matched set the first one who was similar to the refusal regarding self-perception of health. In addition, NORC was required to schedule replacements within 5 working days of a confirmed refusal. These features were intended to correct the described Baseline scheduling deficiencies and to bring the study into Protocol compliance regarding health matching of replacements. To further minimize the possibility of scheduling bias, the entire study population was partitioned into 79 groups; these groups were then randomly scheduled for an examination time. In this way, no single group would be favored a priori for a certain scheduling period. The groupings, consisting of approximately 32 participants, corresponded to the examination groups established at Baseline. Group integrity was maintained to enhance study compliance and comradery. Study participants were given the option to remain with their group or to reschedule their examination at a time more convenient to them. 5-3 �FIRST FOLLOWUP COMPLIANCE Eighty-five percent (1,016/1,191) of the Ranch Hands and 81 percent (955/1,176) of the Original Comparisons participated in the first followup examination and questionnaire process. Of 288 replacements, 267, or 93 percent, chose to attend the first followup examination; additionally, 71 new replacements participated in the followup, yielding a total sample size of 338 replacements at followup. These counts and others are summarized in Table 5-1. In Table 5-1 and subsequently in this report, the shifted Original Comparisons were combined with the Original Comparisons, and the Air Force replacements were combined with the replacement Comparisons. TABLE 5-1. Baseline Versus First Followup Sample Sizes Group Comparison Participation Baseline Only Baseline and Followup Followup Only Ranch Hand Original Replacement 74 971 45 64 872 83 21 267 71 Although fully compliant at Baseline, 74 Ranch Hands, 64 Original Comparisons, and 21 replacement Comparisons chose not to participate in the first followup examination. In the interim, 10 of the 74 Ranch Hands and 16 of the 85 Comparisons died. An additional 5 of the 74 Ranch Hands and 14 of the 85 Comparisons were unlocatable during the scheduling operation. There were 56 of 59 remaining Ranch Hands and 50 of 55 remaining Comparisons who refused to participate in the first followup, although they were alive and locatable during scheduling, and responded to the noncompliance telephone questionnaire, giving their reported health status and reason for nonparticipation. The 3 remaining Ranch Hands and 5 Comparisons refused to participate in the telephone survey. Reasons for nonparticipation given in the telephone survey are summarized in Table 5-2. The totals in Table 5-2 do not correspond to Table 5-1 because some participants gave more than one reason for nonparticipation. Of the 56 living locatable Ranch Hands and the 50 Comparisons who took the noncompliance telephone questionnaire, only 35 Ranch Hands and 42 Comparisons responded to the question regarding health status. The reported health status of these 77 nonparticipants is summarized in Table 5-3. 5-4 �TABLE 5-2. Reasons for Nonparticipation in the First Follovup of 56 Ranch Hands and 50 Comparisons Who Were Fully Compliant at Baseline* Group Comparison Ranch Hand Reason Number Fear of Physical Job Commitment Dissatisfaction with USAF No Time or Interest Travel Distance, Family Confidentiality Health Reasons Passive Refusal Dissatisfaction with Baseline Financial Hardship Other Total Number Percent Percent 0 13 10 7 13 0 8 11 5 0 17 13 9 17 0 11 15 7 2 9 9 6 12 1 3 6 2 4 16 16 11 21 2 5 11 4 3 5 4 7 0 7 0 12 75 57 *Some participants gave more than one reason for nonparticipation. TABLE 5-3. Reported Health Status of 35 Ranch Hands and 42 Comparisons Fully Compliant at Baseline and Noncompliant at First Follovup Group Reported Health Ranch Hand Status Number Percent Excellent Good Fair Poor Total 5 22 6 2 14 63 17 6 Number Percent 10 22 8 2 24 52 19 5 42 35 p=0.72 5-5 Comparison �Among the individuals responding to the health status question, there was no statistically significant difference between noncompllant Ranch Hands and Comparisons regarding reported health (p=0.72). Further detail regarding the 45 Ranch Hands, 83 Originals, and 71 replacements newly examined at followup is shown in Table 5-4, which gives the Baseline status of these participants. Taking the questionnaire but not the physical examination at Baseline were 39 of the 45 Ranch Hands newly examined at followup. Five of the 45 Ranch Hands who were identified too late to be invited at Baseline were simply described as having had "no action" taken. TABLE 5-4. Baseline Status of Newly Examined Participants Group Comparisons Baseline Status Interview Only, Refused Physical Examination No Interview, No Physical Examination Unlocatable No Action Proxy New to Study Total Ranch Hand Original Replacement 39 61 32 0 10 11 0 5 1 0 1 11 0 0 3 16 0 9 45 83 71 Of the 71 newly examined replacements, 43 (32+11) were either partially compliant at Baseline or were at least contacted at Baseline and, therefore, identified as replacements, although not health matched to a noncompliant Comparison. The remaining 28 newly examined replacements were not previously contacted. Of these, 14 were health-matched replacements and 2 were replacements added to the study in August 1985 after completion of the Baseline physical examination. Thus, of the 71 replacements who took the physical examination for the first time at followup, only 14 were new health-matched replacements. All 71 replacements may be regarded as new to the study, even though 43 had been previously contacted at Baseline and knew that they were potential study participants. The 28 replacements who had not been previously contacted may be regarded as new in a more restrictive sense since they did not know of their potential involvement in this study before they were recruited for the first followup examination. This set of 71 replacement Comparisons and the subset of 28 are distinguished from each other using 5-6 �the unrestricted and restricted definitions of "new" to provide data regarding changes in replacement self-selection, an issue explored later in this chapter. FACTORS KNOWN OR SUSPECTED TO INFLUENCE STUDY PARTICIPATION A multitude of factors may be considered to influence self-selection. These may be broadly classified as health, logistic, operational, publicity, or demographic factors. The Baseline Report contains a list of specific factors within each of these categories. For example, health factors are thought to include self-perception of health as well as demonstrable health indicators, such as medication use and work days lost due to illness or injury. Logistic factors are thought to include distance to the examination site, reluctance to spend time away from family or job, income, and occupation. Demographic factors might include flying status, age, race, or military duty status (active, retired, separated). Operational factors include any aspect of study operation that may cause differential compliance, such as differential treatment of participants during scheduling, physical examination, interview, or debriefing. Publicity factors have to do with national attitudes and media presentations regarding the Agent Orange issue, the Vietnam war, veteran health care, or health care in general. Additionally, these considerations may affect people differently and, in particular, may influence Ranch Hands differently than Comparisons. The decision to volunteer for this study is admittedly complex, making statistical assessment of compliance bias difficult and necessarily crude in that many of the factors contributing to self-selection cannot be measured directly. Instead, compliance bias was investigated at first followup, as in the Baseline Report. Specifically, it was investigated with respect to selfperception of health, medication use, daily aspirin use, work days lost due to illness or injury, and income in comparing partially compliant with fully compliant participants. In other selection bias assessments, such as statistical contrasts of Original and shifted Original Comparisons, these same factors and 26 variables taken from the physical examination and psychometric testing were analyzed. THE TELEPHONE SURVEY In April 1985, all previously uncontacted living Comparisons were identified for telephone contact -to assess their current health. This health status information was necessary for the matching of replacements to noncompliant Comparisons. From a total of 9,982 available Comparisons, 7,963 were included in the telephone survey. The 2,019 nonselected Comparisons included 488 deceased, as of 1 August 1985, and 1,531 who had been previously contacted. The group of 1,531 previously contacted Comparisons comprised all Comparisons who were fully compliant, partially compliant, or noncompliant at Baseline. The survey questionnaire is, shown in Appendix D. In brief, it queried the respondent regarding self-perception of health (excellent, good, fair, poor), current prescribed medication use (yes, no), work days lost due to illness or injury, special health care needs (wheelchair, nurse, or other special equipment), and income (less than $20,000, $20,000 to $40,000, or more than $40,000). If the respondent indicated that he was taking 5-7 �prescribed medication, he was asked to identify the illness for which the medication was prescribed. If work days were lost due to illness or injury, the respondent was asked to identify the causing illness or injury. If special health care or equipment was needed, he was asked to specify the illness or condition requiring the special care. He was further asked to distinguish conditions requiring special care from those that were previously identified in response to the medication and days lost from work questions. The telephone interview was accomplished via CATI. Of the 7,963 cases fielded, 7,411 telephone surveys were actually completed. The nature of the 552 noncompletions is summarized in Table 5-5. TABLE 5-5. Summary of Reasons for Noncompleted Telephone Interviews Reason Number Percent of 7,963 Deceased Active Refusal Passive Refusal Unlocatable Ineligible 26 93 242 190 1 0.3 1.2 3.0 2.4 0.0 Total 552 6.9 Several questionnaires that could not be administered by telephone were accomplished by mail; these numbered 540 out of the 7,411 completed. Summaries of the responses to each of the five questions are shown in Table 5-6. Of the 1,271 respondents who reported that they had lost work days due to illness or injury, 550 (43%) lost 1 to 5 days, 197 (15%) lost between 6 and 10 days, and 524 (41%) lost more than 10 days. The maximum number of days reported lost was 965. The 56 respondents who reported more than 180 days lost misinterpreted the question; it referred only to the past 6 months. The telephone interviewer reported whether the respondent was friendly, cooperative but not interested, impatient, or hostile. The association between the interviewer's remark and the self-reported health of the respondent was investigated. The results are displayed in Table 5-7. The association between the interviewer's remark and reported health status is statistically significant (p=0.02), with hostile repondents reporting poorer health than friendly, cooperative, or impatient respondents. Other analyses of these data, not shown here, demonstrated significant associations between health perception and income (p=0.001), rank (p=0.001), age (p=0.001), medication use (p=0.001), and need for special health care (p=0.001). Positive health perception increased with income and rank and 5-8 �TABLE 5-6. Summary of Results to the Telephone Questionnaire Self-Assessment of Health Compared to Others Same Age Response Excellent Good Fair Poor Do Not Know Missing Total Number Percent 2,882 3,306 972 245 3 3 38.89 44.61 13.11 3.31 0.04 0.04 7,411 100.00 Taking Medication for Current Illness Response Yes No Refused Missing Total Number Percent 2,129 5,277 1 4 28.73 71.20 0.01 0.05 7,411 100.00 Illness or Injury Absence From Job During Last 6 Months Response Yes No Refused Missing Total Number Percent 1,271 6,135 ' 3 2 17.15 82.78 0.04 0.03 7,411 100.00 5-9 �TABLE 5-6. (continued) Summary of Results to the Telephone Questionnaire Need Assistance in Daily Activities Response Number Percent Yes No Refused Missing 114 7,291 4 2 1.54 98.38 0.05 0.03 7,411 100.00 Total Earned Income From Any Job During 1984 Response Number Percent Yes No Refused Missing 6,636 755 17 3 89.54 10.19 0.23 0.04 7,411 100.00 Total Income Level Response Less than $20,000 $20,000-$40,000 More than $40,000 Not Applicable Refused Do Not Know Missing Total Number Percent 2,015 3,034 1,411 774 161 9 7 27.19 40.94 19.04 10.44 2.17 0.12 0.10 7,411 100.00 5-10 �TABLE 5-7. Contrast of Interviewer's Remark from Telephone Interviews and Reported Health Status Reported Health Status Remark Excellent Good Fair PerPerNumber cent Number cent Friendly 2,209 39 622 38 Cooperative Impatient 42 40 9 21 Hostile Total 2,882 39 Number Total Poor PerPerPercent Number cent Number cent 44 46 45 63 730 229 10 3 13 14 9 7 191 44 6 4 3 3 6 9 5,606 1,650 106 43 3,306 45 972 13 245 3 7,405 2,476 755 48 27 76 22 1 0 p=0.02 decreased with age, medication use, and special health care. Further, there was no significant association between health perception and the duration of the telephone interview (p=0.17) or the time of day of the interview (p=0.98). There was no significant health-by-duration-by-time of day interaction (p=0.77). These data were also used to assess the self-reported health of 773 Original Comparisons (excluding shifted Original Comparisons) fully compliant at Baseline relative to the reported health of the 7,411 previously uncontacted Comparisons who completed the telephone survey. The selfreported health status of the Original Comparisons from the Baseline questionnaire was contrasted with that of the previously uncontacted Comparisons on a three-category scale (excellent, good, fair/poor) with an adjustment for date of birth (born during or before 1942, born after 1942). The results are displayed in Table 5-8. Previously uncontacted Comparisons who completed the survey are indicated by T (telephone); Original Comparisons are labeled 0. Data are missing for 12 Original Comparisons and 16 telephone-surveyed Comparisons. There was no statistically significant difference between these groups regarding health perception after adjustment for age (p=0.14), and this equivalence did not change with age (p=0.80). Additionally, there was a statistically significant age effect (p=0.001), as expected. These results suggested that the Original Comparisons were representative of the entire Comparison cohort with respect to health perception. 5-11 �TABLE 5-8. Self-Reported Health of Previously Uncontacted Comparisons, in 1986, Versus Self-Reported Health Status of Original Comparisons at Baseline Health Perception Excellent Age Fair/Poor Good Group* Number Percent Number Percent Number Percent Total Born >1942 T 0 1,847 203 39 39 2,003 239 43 46 837 83 18 16 4,687 525 Born <1942 T 0 1,034 91 38 39 1,298 120 48 51 376 25 14 11 2,708 236 *T = previously uncontacted Comparisons 0 = Original Comparisons. REPLACEMENT COMPARISONS VERSUS THE NONCOMPLIANT COMPARISONS THEY REPLACED Baseline Replacement These analyses are refinements of the analyses in Chapter V of the Baseline Report. Of 288 Comparisons replaced at Baseline, only 57 responded to the short noncompliance telephone questionnaire shown in the appendix. These 57 comprised 38 Original Comparisons and 19 replacements. As in the followup telephone survey, the short noncompliance telephone questionnaire queried respondents on health status, work days lost due to illness, medication use, and income level. In accordance with the Protocol, replacements were statistically contrasted with the noncompliant Comparisons they replaced based on their reported health status (excellent, good, fair, poor), medication use (yes, no), and income level (less than $20,000, $20,000 to $40,000, more than $40,000). This contrast, with adjustment for group membership (Original, replacement) of the noncompliant Comparison, is shown in Table 5-9. There was no significant difference between the reported health patterns in the upper and lower panels of Table 5-9. When these two tables were merged, no statistically significant difference was found between the health status of noncompliant Comparisons and their non-health-matched replacements (p=0.99). It is noteworthy that 53 percent of Original and replacement noncompliant Comparisons were matched, by chance, perfectly to their replacements on the basis of reported health status. Only 7 percent (4/57) were mismatched by two categories and one replacement was mismatched by three categories. These same groups were contrasted on medication use; the results are shown in Table 5-10. 5-12 �TABLE 5-9. Noncompliant Original Comparisons and Replacement Comparisons Versus Their Baseline Replacements: Reported Health Status at Baseline Health Status of Replacements Group Noncompliant Original Comparison Health Status Excellent Good Fair Poor Total Noncompliant Replacement Excellent Good 1 3 9 1 1 24 4 7 1 0 12 Fair Poor Total 2 0 0 0 0 0 0 0 19 16 2 1 2 0 38 7 3 1 0 5 3 0 0 0 0 0 0 0 0 0 0 12 6 1 0 1 1 8 0 0 19 Excellent Good Fair Poor Total TABLE 5-10. Noncompliant Original Comparisons and Replacement Comparisons Versus Their Baseline Replacements: Medication Use at Baseline Medication Use of Replacements Group Noncompliant Original Comparison Medication Use Yes .. No Total Yes No 0 3 4 31 4 34 3 35 38 0 1 1 17 1 18 18 19 Total Noncompliant Replacement Yes No Total 5-13 �Due to sparseness these data were not analyzed. It is interesting to note, however, that there was 82 percent agreement in the upper panel of Table 5-9 (31/38) and 89 percent in the lower panel (17/19), with 84 percent agreement in the combined table (48/57), close to expected within group percentages of 83 and 90 percent, respectively, due purely to chance. Work loss was not analyzed due to slight differences between the way the work loss question was worded in the noncompliance telephone and telephone survey questionnaires. The contrast regarding income level is shown in Table 5-11. TABLE 5-11. Noncompliant Original Comparisons and Replacement Comparisons Versus Their Baseline Replacements: Income at Baseline Income Level of Replacements (in thousands) Group Income Level Noncompliant Original Comparison <$20 $20-$40 >$40 <$20 1 6 0 Total Noncompliant Replacement $20-$40 3 6 7 >$40 Total 0 3 6 4 15 13 32* 16 <$20 $20-$40 >$40 0 1 1 0 7 3 2 0 5 10 Total 2 8 9 19 *Six noncompliant Original Comparisons were unwilling to respond. The patterns of income matching in the first and second panels of Table 5-11 were not significantly different (p>0.10). In the combined table, replacements reported significantly lower income than the Comparisons they replaced (p<0.05) although 49 percent (25/51) were perfectly categorically matched. These analyses suggested that the Baseline replacements were very similar to the noncompliant Comparisons they replaced regarding reported health status, medication use, and income. These analyses were also pertinent to the question of whether there was selection bias due to noncompliance in the Comparison group. The predominantly negative findings suggested that there was little or no Comparison selection bias. These 5-14 �results suggested that the upper-bound bias calculations reported in Chapter V of the Baseline Report are overestimates of reality. However, lack of clinical data for the noncompliant Comparisons precluded refining those Baseline bias calculations at this time. Accordingly, the Baseline selection bias calculations may be viewed as crude bounds to an unknown bias that must await future data for proper recalculation. First Follovup Replacement Replacements were matched to noncompliant Comparisons at first followup on the basis of the matching variables—date of birth, race, and occupation— and self-reported health status (excellent, good, fair, poor), as recorded in the telephone survey. This was accomplished by recording the self-reported health status of the noncompliant Comparison during the attempt to schedule and matching that status against those of the other Comparisons in the same matched set. A Comparison in a matched set was considered to replace a noncompliant Comparison if he had the same health status as that recorded for the noncompliant Comparison during the attempt to schedule him. If no willing Comparison reporting the same health status could be found in the matched set, health status was dichotomized to excellent or good versus fair or poor. A willing Comparison with the same health status as the refusal on the dichotomized scale was then accepted as a replacement, if no willing Comparison could be found using the dichotomized scale, attempts to find a replacement were terminated. During this process, 14 Comparisons were health matched to noncompliant Comparisons. The results are summarized in Table 5-12. TABLE 5-12. Health Status of Refusals and Their Matched Replacements Refusal's Health Replacement's Health Excellent Good Fair Poor Total Excellent Good Fair Poor Total 1 5 0 0 2 6 0 0 0 0 0 0 0 0 0 0 3 11 0 0 6 8 0 0 14 All refusals reported good or excellent health. This implied that bias due to noncompliance in the Comparison group could possibly bias the study away from finding an herbicide effect. The inclusion of health-matched replacements tended to correct for this by replacing healthy noncompliant Comparisons with healthy replacement Comparisons. The relatively small number of new health-matched replacements minimized the actual effect of this bias "correction," however. 5-15 �SCHEDULING AT FIRST FOLLOWUP The schedulers were required to find and schedule a willing healthmatched replacement within 5 working days of a confirmed refusal to correct scheduling differences experienced at Baseline. This constraint proved impractical to implement since Comparisons would vacillate, forcing a series of repeated telephone calls. Rather than terminate the process at 5 days, as required by the contract, the schedulers continued their recruiting attempts, sometimes for several months. Hence, new health-matched replacements were brought into the study much later than other participants. The percent completing the physical examination by calendar date is plotted in Figure 5-1 for all Ranch Hands, Original Comparisons, and all Comparisons. The corresponding plot for Ranch Hands, Original Comparisons, old replacements, and the 28 restricted new replacement Comparisons is shown in Figure 5-2. Additionally, schedulers experienced reticence and vacillation with other Comparisons being scheduled for the first time. In particular, as a group, the 71 unrestricted new replacement Comparisons were also scheduled later than other participants. Figure 5-3 shows the percent of Ranch Hands, Original Comparisons, "old" Comparisons, and the 71 unrestricted newly examined replacement Comparisons completing the physical examination by calendar date. During the scheduling for the 1987 followup examination, schedulers will attempt to schedule health-matched replacements within 15 working days of a refusal. NEW REPLACEMENTS VERSUS OLD REPLACEMENTS Another statistical issue of concern is the homogeneity of the replacement Comparisons. The validity of the study might be compromised if, for example, newly admitted replacements had self-selected themselves into the study differently than previously admitted replacements. This kind of difference may occur due to changes in public opinion regarding the Agent Orange issue, the national political climate, changes in national opinion regarding health care, changes in the location of the examination site, or a combination of these and other factors. This issue was addressed by comparing new with old replacements on a variety of endpoints with adjustment for the matching variables. Blacks were deleted from the analyses. Two separate series of analyses were performed, one for each of the two kinds of new replacements (unrestricted and restricted) defined earlier. First, unrestricted new replacements were identified as the 71 replacements who were examined for the first time at first followup, regardless of their compliance at Baseline. Second, analyses were restricted to the 28 replacements who were examined for the first time and who had never been contacted before the first followup; these were called restricted new Comparisons. In each of the two series of new replacement analyses, all replacements not satisfying the definition of "new" are included by referring to them as "old" replacements. All "old" replacements were at least contacted at Baseline and were fully compliant at first followup. 5-16 �10090807060- Cumulative Percent 50 • 40 - 30- 20100M r J J A S 1985 IDD O 0 N D J Date at Clinic 0 0 o All Comparisons F M 1986 o o o Original Comparisons * * * Ranch Hands Figure 5-1. Percent Completed Physical Examination by Calendar Date for All Comparisons 100 90 80 70 60 Cumulative Percent 50 • 40 30 20 10- 0M r J i J r A 1985 i S i 0 i N Date at Clinic. Restricted New Replacement Comparisons Original Comparisons \ D F 1986 o o o Old Replacements & a-a Ranch Hands Figure 5-2. Percent Completed Physical by Calendar Date 5-17 M �100 90 8070- Cumulative 60Percent 50. 40302010- 0 M T J ~r J A 1985 IDD 1 0 —r S 1 N ~r J Date at Clinic 0 Q"$ Unrestricted New Replacement Comparisons * * * Original Comparisons M —r A 1986 €> o o Old Replacements Ranch Hands Figure 5-3. Percent Completed Physical Examination by Calendar Date for Unrestricted New and Old Replacement Comparisons In each of the tvo series of analyses, new and old replacement Comparisons were contrasted on health perception (excellent, good, fair, or poor), medication use (yes, no), work loss (yes, no), and daily use of aspirin (yes, no). Blacks were deleted from all analyses. New and old replacements were then contrasted on 20 clinical determinations from the first follovup examination. Table 5-13 shows tvo cross-classifications of 313 nonblack replacements, from a total of 338 replacements fully compliant at first follovup, by group (old, nev) and reported health status. In the unrestricted sense, the reported health status of nev and old replacements differed significantly (p-0.04), vith nev replacements reporting more fair or poor health than old replacements. In the restricted sense, the difference betveen nev and old replacements vas statistically significant (p=0.001), vith nev replacements tending to declare themselves of fair or poor health more often than old replacements. 5-18 �The same groups were contrasted on medication use; the results are shown in Table 5-14. The difference between old and new Comparisons under the unrestricted definition was not statistically significant (p=0.16) as regards medication use. The difference between old and new Comparisons under the restricted definition was, however, statistically significant (p»0.003). This difference was due to the higher reported medication use of the 26 nonblack new replacements not previously contacted. New and old replacements were contrasted on work loss due to illness; the results are shown in Table 5-15. TABLE 5-13. Reported Health Status of Nonblack New and Old Replacements, According to Two Definitions of "New" Unrestricted Old Restricted New New Old Number Percent Number Percent Number Percent Number Percent Health Excellent Good Fair/Poor 142 91 19 Total 56 36 8 252 30 20 11 49 33 18 161 103 23 56 36 8 26 287 61 p=0.04 42 31 27 11 8 7 p=0.001 TABLE 5-14. Reported Medication Use of Nonblack New and Old Replacements, According to Two Definitions of "New" Restricted Unrestricted Old Medication Yes No Total New New Old Number Percent Number Percent Number Percent Number Percent 30 222 12 88 . 252 12 49 61 p=0.16 20 80 33 254 11 89 287 26 p=0.003 5-19 9 17 35 65 �TABLE 5-15. Reported Work Loss of Nonblack New and Old Replacements, According to Two Definitions of "New" Unrestricted Old Work Loss Yes No Total Restricted New Old New Number Percent Number Percent Number Percent Number Percent 47 205 19 81 252 12 49 61 p=0.99 20 80 54 233 19 81 287 5 21 19 81 26 p=0.99 The difference between new and old replacements regarding work loss under the unrestricted or restricted definition was not statistically significant (p=0.99 and p=0.99, respectively). Results of a similar contrast on daily aspirin usage are shown in Table 5-16. The difference between new and old replacements regarding daily use of aspirin under the unrestricted or the restricted definition was not statistically significant (p=0.99 and p=0.75, respectively). It is noteworthy that the differences for general health and medication use did not occur for work loss and daily aspirin usage, suggesting that some participants may have over-reported when asked less specific questions about their health. New and old replacement Comparisons were also compared on 20 clinical and psychometric variables measured during the physical examination and psychological testing. These 20 variables are a subset from 26 selected from among an entire collection of nearly 200 endpoints in this study by requiring near statistical independence within and between organ systems. Variables selection was accomplished by screening the correlation matrices of variables as an entire set and separately within each organ system, including examining partial correlations between single variables and linear combinations of other variables within organ systems. Identified first were 10 variables with pairwise correlations less than 0.10 in absolute value. This was followed by identification of 16 additional variables with pairwise correlations between 0.10 and 0.20 in absolute value, making a total of 26 variables. These variable selection screens were accomplished on Baseline data for 1,154 nonblack fully compliant Comparisons subsequent to publication of the Baseline Report. The complete set of 26 dependent variables selected as 5-20 �TABLE 5-16. Reported Daily Aspirin Usage of Nonblack New and Old Replacements, According to Two Definitions of "New" Unrestricted Old Restricted New Old New Aspirin Usage Number Percent Number Percent Number Percent Number Percent Yes No 182 69 Total 73 27 251 44 17 61 p=0.99 72 28 206 80 72 28 286 20 6 77 23 26 p-0.75 nearly statistically independent is shown in Table 5-17. correlation matrix of these 26 variables as determined on Comparison data set is shown in Table 0-1 of Appendix D. that relative statistical independence of these variables biological independence of these variables. The Baseline the entire It is recognized does not imply These 26 variables were intended to serve as the basis for statistical contrasts of Original Comparisons, shifted Original Comparisons, and replacement Comparisons in the decision regarding the inclusion of shifted Original Comparisons and replacement Comparisons in the primary analyses. Generically, the analyses first compared two groups on each of the 26 variables with adjustment for rank (officer, enlisted), age at Baseline (40 or under, over 40), occupation (officer flyer, officer nonflying, enlisted flyer, enlisted groundcrew), and race (Black, nonblack). Blacks were deleted from the analysis. The total number of significant differences on the first set of 10 dependent variables was used as the basis for a decision regarding group difference. These 10 analyses were assumed to be 10 independent repetitions of a Bernoulli trial with probability of 0.05 of success under the null hypothesis that there were no group differences for any of the 10 variables. The probability of observing three or more successes in 10 independent repetitions of a Bernoulli trial, with probability of 0.05 of success, is 0.012. The entire set of 26 analyses was then assessed to test the hypothesis of group equality. The probability of 4 or more successes in 26 independent repetitions of a Bernoulli trial, with probability of 0.05 of success, is 0.039. These 2 critical values, both probabilities below 0.05, were used to assess the analyses on the 10 and on the 26 selected variables. 5-21 �TABLE 5-17. Twenty-Six Dependent Variables Selected as Nearly Statistically Independent With the Use of Baseline Data Variables Having Pairwise Absolute Correlations Less Than 0.10 Total Bilirubin (TBILI) Diastolic Blood Pressure (DBP) White Blood Cell Count (WBC) Skin Index (SKIN) MMPI Depression Scale (MMPID) Blood Urea Nitrogen (BUN) Urine Specific Gravity (USG) Pulse Index (PULSE) Nerve Conduction Velocity Above the Elbow (NCVE) Semen Count (SEMEN) Variables Having Pairwise Absolute Correlations Greater Than 0.10 and Less Than 0.20 Red Blood Cell Count (RBC) FEV1/FVC (PULM) Glucose (GLUC) Electrocardiogram (ECG) Platelet Count (PLAT) Full IQ (IQ) Central Nervous System Index (CNS) Nerve Conduction Velocity Above the Ankle (NCVA) Cholesterol (CHOL) Alkaline Phosphatase (ALKPHOS) Coproporphyrins (COPRO) Delta-Aminolevulinic Acid (ALA) Thyroid T4 (T4) Testosterone (TEST) Sedimentation Rate (SED) Gamma-Glutamyl Transpeptidase (GGTP) 5-22 �The statistical issue of how to account for the many interactions in the 26 separate analyses was not resolved during or since the first application of this method. Only the group main effect was regarded as the basis for determining whether a particular analysis was a success. At first followup, only 20 of the 26 variables were measured. The six variables not measured were the two-nerve conduction velocities (NCVE, NCVA), semen count (SEMEN), FEV1/FVC (PULM), full IQ (IQ), and delta-aminolevulinic acid (ALA). New and old replacements were contrasted on each of the remaining 20 variables via the general linear model and log-linear model. The variables—skin index (SKIN), pulse index (PULSE), electrocardiogram (ECG), and central nervous system index (CNS)—were analyzed as dichotomous variables, with each being scored abnormal if any of its components were abnormal. All others were analyzed as continuous variables. The correlation matrix of the 20 variables, based on 1,210 nonblack Comparisons fully compliant at first followup, on first followup data is shown in Table D-2 of Appendix D. The results of these analyses contrasting new versus old replacements with "new" following the unrestrictive definition and Blacks removed from the analyses are shown in Table 5-18. There were 61 nonblack new replacements and 251 nonblack old replacements. In some analyses, the dependent variable was transformed to better approximate normality. Unadjusted means are presented when there is a significant interaction involving group. The probability of observing 2 or more successes in 8 independent repetitions of a Bernoulli trial, with probability of 0.05 of success, is 0.057. In view of the results for the first 8 dependent variables in Table 5-18, new and old replacements appeared to be statistically indistinguishable. The probability of observing 3 or more successes in 20 independent repetitions of a Bernoulli trial, with probability 0.05 of success, is 0.075; the probability of 4 or more is 0.016. Recognizing the slight correlations between the dependent variables in the lower panel of Table 5-18, and the results of the analyses, new and old replacements again appeared to be statistically indistinguishable. The same analyses were conducted to contrast new and old replacement Comparisons, with "new" defined in the restrictive sense. The results are shown in Table 5-19, with the same notations as Table 5-18. The same binominal critical values, 2 for the first panel and 4 for the entire set of 20 analyses, and the results shown in Table 5-18 indicated that there was no statistical difference between the 26 nonblack new replacements and the 287 nonblack old replacements. The negative findings shown in Tables 5-18 and 5-19 suggested very strongly that there has been no change in the way replacements self-select for entry into this study. ORIGINAL COMPARISONS VERSUS SHIFTED ORIGINAL COMPARISONS The removal of ineligible Comparisons early in the Baseline scheduling operation resulted in the exclusion of approximately 18 percent of all Comparisons from the study. Since some of these ineligibles had been randomized as Original Comparisons, some previously randomized Comparisons were allocated to the positions vacated by the removed original Comparisons and, thus, were referred to as shifted Original Comparisons. 5-23 �TABLE 5-18. Summary Results of Unrestricted New Versus Old Nonblack Replacements Contrasted on 20 Variables Replacement Group Means* (Percent Abnormal) Variable (Transformation) Old New p-Value Significant Interactions Variables With Absolute Pairwise Correlations Less Than 0.10 TBILI (LOG) DBF (SORT) WBC (LOG) SKIN MMPID (LOG) BUN (SORT) USG PULSE 0.76 79.17 7.06 (54.0) 56.21 14.15 1.014 (16.7) 0.76 79.51 7.13 (49.2) 57.19 13.79 1.014 (11.5) NS NS NS NS NS NS NS . GRP*OCC, GRP*AGE Variables With Absolute Pairwise Correlation Between 0.10 and 0.20 RBC GLUC (LOG) ECG PLAT (SQRT) CNS CHOL (SQRT) ALKPHOS (LOG) COPRO (SQRT) T4 TEST (SQRT) SED (LOG) GGTP (LOG) 5.00 109.31 (15.5) 269.5 (2.8) 212.7 87.9 116.9 7.51 601.4 4.17 31.06 GRP*OCC*AGE 5.00 101.33 (13.1) 275.0 (5.0) 208.8 87.10 122.6 7,94 605.3 4.93 29.77 NS NS NS NS NS GRP*OCC 0.03 NS NS GRP*OCC*AGE GRP*AGE *A11 means are expressed in original units. NS: LOG: SQRT: Not significant (p>0.05) Analysis performed on logarithmic scale. Analysis performed on square root scale. GRP: Group OCC: Occupation AGE: Birth year (Age) 5-24 �TABLE 5-19. Summary Results of Restricted New Versus Old Nonblack Replacements Contrasted on 20 Variables Replacement.Group Means* (Percent Abnormal) Variable (Transformation) Old New p-Value Significant Interactions Variables With Absolute Pairwise Correlations Less Than 0.10 TBILI (LOG) DBF (SORT) WBC (LOG) SKIN MMPID (LOG) BUN (SORT) USG PULSE 0.76 79.44 7.01 (52.3) 56.11 14.02 1.014 (15.3) 0.75 76.98 7.91 (61.5) 59.73 14.75 1.013 (19.2) NS NS NS NS NS NS NS NS Dependent Variables With Absolute Pairvise Correlation Between 0.10 and 0.20 RBC GLUC (LOG) ECG PLAT (SQRT) CNS CHOL (SQRT) ALKPHOS (LOG) COPRO (SQRT) T4 TEST (SQRT) SED (LOG) GGTP (LOG) 5.01 108.8 (14.3) 270.5 (2.8) 212.5 87.75 117.8 7.56 601.2 4.15 31.23 4.90 95.86 (23.1) 271.56 (7.7) 205.6 87.72 120.5 8.00 612.6 6.37 26.41 NS 0.007 GRP*AGE NS NS NS NS NS NS NS 0.03 NS *A11 means are expressed in original units. NS: Not significant (p>0.05). LOG: Analysis performed on logarithmic scale. SQRT: Analysis performed on square root scale. 5-25 �Fully compliant Original and shifted Original Comparisons were compared in the Baseline Report with respect to reported health status, medication use, and work loss. Group differences for health status were significant (p=0.001) but were not so for medication use or for work loss; the shifted Original Comparisons tended to report themselves in poorer health than the Original Comparisons but were statistically equivalent to the Originals regarding medication use and work loss. Fully compliant Original and shifted Original Comparisons were contrasted at first followup on reported health status, work loss, medication use, and daily use of aspirin. As in the Baseline Report, these analyses were done for only nonblack Comparisons. The results of the contrast of Original and shifted Original Comparisons on reported health status are shown in Table 5-20. Here, health status is evaluated on a three-category scale (excellent, good, fair/poor). The group difference between Original and shifted Original nonblack Comparisons regarding reported health status was not significant (p=0.30). The results of the contrast of Original versus shifted Original Comparisons on medication use are shown in Table 5-21. The group difference between Original and shifted Original nonblack Comparisons regarding medication use was not significant (p=0.68). The results of the contrast on work loss are shown in Table 5-22. The group difference between nonblack Original and shifted Original Comparisons regarding work loss was not significant (p=0.82). The results of the contrast on daily aspirin usage are shown in Table 5-23. The group difference between Original and shifted Original nonblack Comparisons regarding daily aspirin usage was not significant (p=0.98). Fully compliant Original and shifted Original nonblack Comparisons were also contrasted on each of the full set of 26 nearly uncorrelated variables shown in Table 5-17 on Baseline data. The results are shown in Table 5-24. Sedimentation rate (SED) was analyzed as a categorical variable with values low (0-1), medium (2-3), and high (3-4). The percents of Original Comparisons within these categories were 35.8, 33.1, and 31.1 percent, respectively; the shifted Original Comparison percents were 30.8, 36.3, and 32.9, respectively. The probability of observing 3 or more successes in 10 independent repetitions of a Bernoulli trial, with a probability of 0.05 of success, is 0.0115. The probability of observing 2 or more is 0.0861. Based on these critical values and the results shown in the upper panel of Table 5-24, there appeared to be no statistical difference between Original Comparisons and shifted Original Comparisons. The probability of observing 4 or more successes in 26 independent repetitions of a Bernoulli trial is 0.039. The probability of observing at most 2 successes in 26 independent repetitions of a Bernoulli trial, with probability 0.05 of success, is 0.86. Based on these critical values and the known slight correlation of the 16 dependent variables in the second panel of Table 5-19, these results suggested that Original and shifted Original Comparisons are not statistically distinguishable. 5-26 �TABLE 5-20. Reported Health Status of Fully Compliant Original and Shifted Original Nonblack Comparisons: First Followup Original Comparison Group Original Reported Health Shifted Original Number Percent Number Percent Excellent Good Fair/Poor 387 307 53 Total 747 52 41 7 51 45 4 76 68 6 Total 463 375 59 p-Value 0.30 897 150 TABLE 5-21. Medication Use of Fully Compliant Original and Shifted Original Nonblack Comparisons: First Followup Original Comparison Group Shifted Original Original Medication Use Yes No Number Percent Number Percent 102 645 Total 14 86 747 23 127 150 5-27 15 85 Total 125 772 897 p-Value 0.68 �TABLE 5-22. Work Loss of Fully Compliant Original and Shifted Original Nohblack Comparisons: First Followup Original Comparison Group Shifted Original Original ¥ork Loss Number Percent Number Percent 631 125 No Yes Total 116 25 83 17 756 82 18 Total 747 150 p-Value 0.82 897 141 TABLE 5-23. Daily Aspirin Use of Fully Compliant Original and Shifted Original Nonblack Comparisons: First Followup Original Comparison Group Shifted Original Original Daily Aspirin Use Yes No Number Percent 529 218 Total 71 29 747 Number Percent 107 43 150 5-28 71 29 Total 636 261 897 p-Value 0.98 �TABLE 5-24. Summary Results of Original Versus Shifted Original Nonblack Comparisons on 26 Variables at Baseline Original Comparison Group Means* (Percent Abnormal) Variable (Transformation) Original Shifted Original p-Value Significant Interactions Variables With Absolute Pairvise Correlations Less Than 0.10 TBILI DBF WBC SKIN MMPID BUN USG PULSE NCVE SEMEN (LOG) 0.61 80.46 7.52 (37.5) 56.25 14.26 1.0209 (10.7) 56.26 77.4 0.61 78.95 7.18 (43.8) 58.40 13.76 1.0205 (.) 89 55.88 72.8 GRP*OCC*AGE NS NS NS NS NS NS NS NS NS Variables With Absolute Pairvise Correlation Between 0.10 and 0.20 RBC PULM GLUC (LOG) ECG PLAT IQ CNS NCVA CHOL ALKPHOS COPRO (LOG) ALA T4 TEST SED GGTP (LOG) 5.18 5.20 0.80 0.81 97.4 94.5 (26.7) (27.6) 269.9 270.6 108.4 108.6 (31.5) (23.7) 47.59 48.17 213.1 220.7 7.84 7.60 30.4 31.1 2,497.0 2,505.3 8.35 8.42 634.6 634.3 given in text 35.53 38.43 NS NS NS NS NS NS 0.02 0.01 NS NS NS NS NS NS NS . NS *A11 means are expressed in original units. 5-29 �Taken together, the results displayed in Table 5-24 very strongly suggested that Original and shifted Original Comparisons did not differ statistically at Baseline. These analyses were repeated on the 20 available variables at the first followup. The results are shown in Table 5-25. The results in the first and second panels of Table 5-25 and the binomial critical values given above suggested that no statistical difference was present between the Original and shifted Original Comparisons. A single multivariate linear regression analysis was done on the 20 dependent variables shown in Table 5-25; no significant interactions involving group (Original, shifted Original) were noted and the group effect was not significant (p=0.28). Taken together, these analyses strongly suggested that there was also no statistical difference between Original and shifted Original Comparisons at first followup. PARTIALLY COMPLIANT VERSUS FULLY COMPLIANT PARTICIPANTS Ideally, compliance bias should be assessed by comparing the health of noncompliant and fully compliant participants with adjustment for group (Ranch Hand, Comparison) and the matching variables. The only information available on the noncompliant participants, however, is their responses to the health status questions, if they were willing to answer them, during the telephone conversation in which they refused to participate in the study. Noncompliant Comparisons were contrasted with their Baseline replacements (see noncompliance telephone questionnaire data, Tables 5-9 to 5-12). In addition, as in the Baseline Report, selection bias was studied by contrasting partially compliant with fully compliant participants with adjustment for group (Ranch Hand, Comparison). Taking the Baseline questionnaire at followup but refusing to take the physical examination or followup questionnaire were 9 Ranch Hands and 30 Comparisons who were either nonlocatable or noncompliant at Baseline. These 39 men were the only partially compliant participants at first followup. Their Baseline compliance is summarized in Table 5-26. One of these individuals, a Ranch Hand with no interview, no physical, and no telephone interview, was Black. The label "no action" indicates that these individuals were not contacted because the Baseline contract expired. Individuals labeled "new Comparisons" were added to the study after the Baseline examination but before start of the first followup. Data from these 39 partially compliant participants were statistically compared with similar data from fully compliant participants with adjustment for group (Ranch Hand, Comparison). This is shown in Table 5-27. Endpoints evaluated were reported health, medication use, and work loss. These analyses are similar to those reported in Table V-15 of the Baseline Report. Reported health status was collapsed to two categories (excellent, good/fair/poor) due to sparse data. One Black participant, a Ranch Hand, was deleted from these analyses. The health versus compliance association in these data was of borderline statistical significance (p=0.08), with partially compliant participants tending to report themselves in better health than fully compliant 5-30 �TABLE 5-25. Summary Results of Original Versus Shifted Original Nonblack Comparisons on 20 Variables: First Followup Original Comparison Group Means* (Percent Abnormal) Variable (Transformation) Original Shifted Original p-Value Significant Interactions Variables With Absolute Pairwise Correlations Less Than 0.10 TBILI (LOG) DBF (SORT) WBC (LOG) SKIN MMPID (LOG) BUN (SORT) USG PULSE 0.75 80.0 6.88 (49.7) 56.2 14.8 1.015 (16.7) GRP*OCC*AGE 0.73 79.60 6.92 (42.1) 55.1 14.04 1.015 (16.4) NS GRP*AGE NS NS NS NS NS Variables With Absolute Pairwise Correlation Between 0.10 and 0.20 RBC GLUC (LOG) ECG PLAT (SORT) CNS CHOL (SORT) ALKPHOS (LOG) COPRO (SORT) T4 TEST (SORT) SED (LOG) GGTP (LOG) 4.97 111.8 (15.3) 263.2 (2.6) 219.5 89.76 115.4 7.58 576.6 5.11 32.39 4.95 111.6 (11.9) 271.9 (2.3) 214.1 85.53 114.9 7.58 559.0 4.91 29.77 NS NS NS NS NS NS NS NS NS GRP*OCC, GRP*AGE NS NS *A11 means are expressed in original units. 5-31 �TABLE 5-26. Baseline Compliance Status of 39 Partially Compliant Participants: First Follovup Group Ranch Hand Baseline Compliance Comparison 23 No Interview, No Physical, No Telephone Interview No Interview, No Physical, Telephone Interview New Comparison 0 3 No Action 4 3 Total 9 30 TABLE 5-27. Reported Health of Partially Compliant Versus Fully Compliant Nonblack Participants Group Ranch Hands Comparisons Compliance Status Reported Health Number Percent Number Percent Total Full Excellent Good/Fair/Poor 473 482 Total Partial 43 46 955 5 3 Excellent Good/Fair/Poor Total 635 575 20 23 20 10 1,108 1,057 2,165 1,210 30 5-32 57 54 80 77 25 13 38 �participants; 66 percent of partially compliant participants reported excellent health while only 51 percent of fully compliant participants reported excellent health. This association did not change with group (p-0.91). The data on medication use and compliance status demonstrated no association (p=0.57), and this equivalence did not change with group (p=0.79). These data are shown in Table 5-28. As shown in Table 5-29, the work loss-by-compliance association in these data was significant (p=0.03), with 84 percent of fully compliant participants reporting work loss and 95 percent of partially compliant participants reporting work loss. These data are sparse and are not considered supportive or nonsupportive of the compliance bias calculations presented in the Baseline Report. The conclusions of the Baseline Report regarding the potential effects of compliance bias should be regarded as conservative overestimates, but worthy of consideration in inference formulations until more data become available. CONCLUSIONS These predominantly negative findings suggest that there has been no change in the way replacements self-select for entry into this study and, due to the obvious scheduling differences between new and old replacements, that no additional bias has been introduced at followup by scheduling differences. These data also strongly suggest that shifted Original Comparisons are not statistically distinguishable from Original Comparisons, either at Baseline or at first followup. This interpretation is also equivalent to the conclusion that no additional bias was introduced by scheduling differences between Original Comparisons and shifted Original Comparisons at Baseline. Available data on noncompliant Comparisons and their replacements suggest that, although replacements were not health-matched to refusals at Baseline, they are r.emarkably similar to refusals with respect to reported health, medication use, and income level. This result also supports a conclusion that there has been little, if any, selection bias due to nonparticipation in the Comparison group. This conclusion supports the use of the total Comparison group for all of the main analyses in the body of this report. Data regarding the few partially compliant participants at first followup are not sufficient to confirm or deny compliance bias calculations published in the Baseline Report. 5-33 �TABLE 5-28. Medication Use of Partially Compliant Versus Fully Compliant Nonblack Participants Group Comparison Ranch Hand Compliance Status Medication Use Full Number Percent Number Percent Total Yes No 123 832 Total Partial 955 1 7 Yes No Total 167 1,043 58 290 56 1,875 1,210 42 44 2,165 25 21 3 27 8 TABLE 75 79 30 4 34 38 5-29. Work Loss of Partially Compliant Versus Fully Compliant Nonblack Participants Group Ranch Hand Compliance Status Work Loss Full Yes No Number Percent Number Percent Total 796 155 Total Partial Comparison 44 44 951 Yes No 8 0 Total 1,010 200 1,210 22 0 28 2 30 5-34 56 56 1,806 355 2,161 78 100 36 2 38 �CHAPTER 6 QUALITY CONTROL During the first AFHS followup, stringent adherence to quality assurance (QA) was planned for and upheld throughout the study, from project initiation to final product delivery and acceptance by the Air Force. A quality program plan was developed for this study cycle, outlining all contract activities requiring periodic and/or systematic QA and quality control (QC) monitoring. The purpose of this chapter is to provide an overview of the specific QA measures developed and used by the project team, specifically in the areas of administrative QC; questionnaire, physical, and psychological examination QC; laboratory QC measures; data base management QA; and statistical QC. ADMINISTRATIVE QUALITY ASSURANCE In recognition of the magnitude, complexity, and importance of the AFHS, a Quality Review Committee (QRC) was established at the initiation of the third-year followup for the purpose of providing general oversight to the AFHS QA Program and advice on the appropriateness of program management and QC actions. The QRC was composed of senior corporate personnel from the prime contractor. These independent reviewers remained separate from the project management staff. The QRC met formally each quarter to review recent study progress and any issues that either had an impact on study quality or were perceived as a potential problem. Assisting the QRC in day-to-day oversight responsibilities was a QA officer responsible for reviewing procedures, performance, and work products from all task managers and key project staff. As part of the monitoring function, the QA officer received exception reports from project task managers whenever an incident occurred that appeared to affect study quality. Monthly reports were also prepared for the Air Force, documenting project compliance with project QA criteria and noting any instances of noncompliance. An additional measure of corporate QC was implemented through independent QA audits of individual project tasks. Members of the QRC determined first-hand whether QA procedures for a particular task were being conducted, whether procedures were appropriate for the task, and whether QA was complete for all aspects of each task. The remainder of this chapter comprises specific QA procedures followed for the individual tasks. QUESTIONNAIRE QUALITY CONTROL NORC used both onsite and home-office QA procedures to produce a comprehensive data set. All AFHS questionnaires were pretested to evaluate 6-1 �their completion time and participant acceptability before they were used at the SCRP. Onsite QC procedures included weekly observation and rating of each interviewer, editing of every questionnaire at the completion of the interview, and monitoring of participant evaluations. The Air Force also continuously conducted QA observations of all onsite activities. QC of data processing included manually editing each questionnaire, including a 100-percent verification of critical items for each questionnaire, computerized cleaning (with both single item and inter!tern review for range and consistency), identifying outliers, and reviewing the actual questionnaire copy to reconcile or correct detected errors. All telephone surveys were monitored for quality and accuracy of interviewer performance by NORC supervisors. The telephone survey supervisor monitored 3 percent of each interviewer's calls to assure an appropriate presentation and an accurate transcription of responses. An additional 5 percent of the participants were recontacted after the interview to evaluate interviewer performance and validate that the correct respondent had been contacted. NORC recruited and trained interviewers according to the detailed procedures described in Chapter 3. A minimum number of interviewers was selected to reduce interviewer variability. Additionally, these individuals were blinded to the participants' exposure status to avoid any bias. Interviewers were required to ask questions exactly as recorded, and in the order in which they appeared. No personal interpretation was allowed. An onsite field manager closely supervised each interviewer's work regularly, observing individual interviews weekly during the examination schedule. The field manager reported directly to the NORC Project Director weekly, and was reviewed by the Project Director during quarterly site visits, to ensure direct accountability by the home office and the field manager for promptly resolving any issues. Specifically, interviewers were checked for accuracy in questionnaire skip patterns, probing, circling of the correct code, control of the interview, voice quality, reading, and use of associated documents. When called for, the onsite manager gave immediate retraining after each observation and documented the content of this training. At weekly meetings, held with all interviewers, the field manager used generalizations from individual interviewer performance observations to train an entire group of interviewers. The NORC field manager also monitored participant evaluations of the study closely and used the information gathered to plan and implement retraining. The manager and staff edited each completed questionnaire before it was shipped to Chicago, attempting to retrieve missing data while the study participant was at the physical examination site. Missing or ambiguous data were also retrieved by telephone when necessary. Spouse fertility data were obtained independently of the participant interview by sending the mail questionnaire while the study participant was at the examination site, and by having a group meeting for wives who accompanied their spouses to the clinic site, where they could complete their questionnaires in private. The Assistant Survey Director in Chicago supervised and edited all interviews conducted at home with participants and spouses. 6-2 �Once the participant and spouse questionnaires were received in Chicago, they vere edited for completeness by a coding supervisor and staff dedicated to the AFHS for the entire project. Resolution of inconsistencies was accomplished by staff members, who standardized all responses prior to keypunching. Questionnaires were then coded, and a 10-percent recede was done on open-ended items. When a batch failed the 10-percent recede, the entire batch was receded and the coding staff was retrained. One hundred percent quality control was accomplished by the Air Force. During data entry, range validity checks were performed and 10 percent of the most important items in each questionnaire was verified. Data were then passed through a computer program that checked for inter- and intracolumn errors. When errors were detected, the-questionnaires were reviewed and the errors corrected. The process continued until no errors were detected by the cleaning program. Then, frequencies were reviewed and any anomalies or errors previously undetected were corrected by reviewing the questionnaires on a case-by-case basis. All corrections were entered into the data tape, but no changes were made to the data recorded in the questionnaires. QA reports were generated monthly, detailing the summary statistics on the number of questionnaires reviewed, the number and types of transcriptions failing QC checks, and the average number of coding errors per batch processed. PHYSICAL EXAMINATION QUALITY CONTROL QC was emphasized in the physical examination, as this data source provided most of the medical information for clinical and epidemiological analyses. Initial concern for a high-quality physical examination was addressed by a stringent SCRF selection process for all personnel who were to directly interact with the participants. Each staff member was hand-selected for the AFHS on the basis of expertise, experience, and a commitment to remain with the study throughout the examination cycle. Further, the Air Force Technical Team reviewed the credentials of all key staff members and approved their participation in the study. A complete pretest physical examination, interview, psychological test, and laboratory workup was done for 10 volunteers several weeks before the scheduled start of the study. Refresher training was given to the dermatologists to enhance their skill in diagnosing chloracne, techniques for detecting specific heart sounds were reviewed with the internists, and diagnosticians were reminded of the need to review Baseline examination data as they formulated all diagnoses. Further, all aspects of patient contact were reviewed: the initial inbriefing of the participants, the logistics of transportation and patient flow within the clinic, and the final outbriefing by the diagnostician. During the examinations, refinements continued whenever operational problems were detected by the SCRF staff and the Air Force onsite monitor, or when participants identified areas requiring improvement. Both of these types of information were addressed during the weekly clinical QA meeting of key SCRF staff, chaired by the SCRF Medical Project Director and attended by an Air Force representative. In addition, written critique forms submitted by all participants were reviewed in detail at the SCRF weekly meetings, 6-3 �providing additional insight to both temporary shortcomings of the entire logistic process as well as the numerous strong points of the programs. Following examination of each participant group, all physical examination forms were reviewed by the SCRF staff for omissions, incomplete examinations, and inconsistencies. The examiners or technicians were quickly contacted to correct the data. Special effort was made to complete this review while the participants were at the examination site. In all cases of data correction, a complete audit trail was maintained. Finally, all mark-sense physical examination forms were read by an optical scanner to ensure total continuity and sensibility of the final examination contents. (This subject is discussed in more detail in the Data Management Quality Control section of this chapter.) Compliance with all aspects of the physical examination was monitored daily by the Air Force onsite monitor and the SCRF Medical Project Director. Additional periodic inspections were conducted by the SCRF Chief of Medicine and the SAIC Principal Investigator. All such clinical reviews were done unobtrusively, and with the full consent of the participant; suggestions or corrections to the examination procedure were always discussed privately with the attending physician. These inspections emphasized aspects of clinical techniques, sequencing and completeness of the clinical data with respect to the examination forms, and the total blindness of the examinations. Of particular note were the detailed daily log entries of the five Air Force monitors. These entries ensured continuity of knowledge (the monitors rotated approximately every 2 weeks) by documenting examination procedural changes and recording events requiring followup by either the Air Force or the prime contractor. Establishment of rapport with each study participant was a primary goal of all organizations involved in this study. Although "rapport building" may not be a traditional QA parameter in most research studies, it is paramount in the AFHS because maintaining the satisfaction of participants encourages them to continue in the study, and thus a significant reduction in future statistical power or bias, or both, is avoided. Every staff member, therefore, from the initial telephone recruiter to the nurse coordinator and the Project Manager, emphasized courtesy, empathy, assistance, and personalized treatment of each participant. LABORATORY QUALITY CONTROL Before the study was begun, specific QC laboratory procedures were designed, developed, and implemented to rapidly detect problems related to test/assay performance, validity of reagents, analysis of data, and reporting of results. All laboratory assays for the study were done with state-of-theart laboratory equipment and techniques. Laboratory facilities all had the equivalent of National Institutes of Health Biosafety Level 2 (BSL-2) approval ratings and were certified by the College of American Pathology (CAP). Hematology assays were performed on Coulter S Plus® equipment; sedimentation rate determinations were performed using the large-tube Westergren method. The Dupont Automated Chemical Analyzer® (ACA) was used to perform the biochemical assays; radioimmunoassays (RIA) were done with standard test kits; and porphyrin was assayed by high-performance liquid 6-4 �chromatography at the Mayo Clinic in Rochester, Minnesota. Hepatitis B tests were performed using Abbott kits, and manually performed electrophoresis and monospecific antibodies were used for immunologlobulin assays. Blood-cell counts were performed with standard microscopy, and Clinitek, a reflectance spectometry urinalysis, was used for all urinalyses. All other assays were done using industry-approved equipment and techniques. All laboratory operations were controlled with the use of an integrated medical laboratory management information system that incorporated direct device to data base interfaces for automated testing equipment, and data entry for manual tests was performed by the laboratory technologists. An automated audit trail and a set of comments for technologist entries were kept for each test so that any QC results could be retraced. Procedural QC included using instrumentation and reagents from one lot number throughout the study. Strict standards of calibration for all automated laboratory equipment were maintained at all times. Trilevel or bilevel controls were used as the primary means for monitoring the quality of all tests. On every group of participant samples, one control (low, medium, or high) was run at the start, after every ninth sample, and at the end of each test run. Each trilevel control was used before repeating it in the run, when more than 18 experimental samples were analyzed. In addition, split aliquots were made from every tenth patient sample and were analyzed separately to measure test reproducibility. All QC data were analyzed and summarized in formal QC reports generated weekly. QC data were subjected to independent statistical analysis to produce and analyze time-dependent trends. For all equipment malfunctions or other exceptions, a formal QC exception report was prepared by the responsible individual and forwarded to the QA officer and the project management team. An additional measure of quality control introduced during the study was the CUSUM tests run with trilevel controls. In particular, the fast initial response cumulative sum (FIR CUSUM) QC technique was used. It has an advantage in detecting long-term 2 subtle drift that could have substantial adverse analytical consequences. FIR is a special case of the CUSUM QC scheme that increases the overall effectiveness of the QC procedure. Unlike QC procedures using standard control charts, which compare each observation to designated limits, these tests utilize the cumulative sum of deviations from a target value. CUSUM statistics were accumulated for each of the trilevels to quickly detect instrument calibration problems as identified by excessive drift. If an out-of-control situation was indicated, the graph showed when the change first occurred. Coefficient of variation (CV) standards were established before the study for each test. All adjacent patient samples were reanalyzed after the equipment was thoroughly checked and fresh controls were run. FIR CUSUM generally has been applied to QC in industry, particularly in high-volume, high-precision applications. To our knowledge, FIR CUSUM has not generally been applied in a biomedical setting. According to SCRF laboratory personnel, this procedure proved so successful in the AFHS that most of the SCRF clinical laboratory will begin using it in the near future. 6-5 �As the examination portion of this study ended, all laboratory outliers were analyzed for logical validity by an independent clinician. All out-ofrange test results were examined and scored as clinically explainable, clinically possible, or clinically unexplained. Quality Control Procedures for the Immunology Laboratory The QC procedures for the Cellular Immunology section of the AFHS were structured to rapidly detect any problems in four major test parameters: (1) assay performance, (2) reagent validity, (3) data analysis, and (4) results reporting. The QC measures were detailed in the Quality Procedures Plan and documented before testing started. Compliance was monitored daily by the Cellular Immunology laboratory supervisor. Key aspects of the program included instrument and equipment calibration and maintenance, assay controls, accuracy and precision determination, and system failure checks. QC measures followed in all Cellular Immunology assays included: • Blood sample from a normal, healthy control individual with each group of AFHS patient samples • Duplicate testing of one random patient sample in each assay • Quadruplicate testing of each patient sample for each variable in each of the functional assays (e.g., PHA stimulation, natural killer cell effector/target ratios) • Parallel testing and monitoring reactivity of various lots of reagents when appropriate • Verification of patient and specimen identification by at least two individuals before final reporting to the data base • Note codes attached to any data point with a detected deviation from normal due to procedural setup error, assay malfunction, equipment malfunction, or assay technical error • • Review of all final assay reports by the Cellular Immunology laboratory supervisor prior to entry into the data base. QC for each functional assay including phytohemagglutinin (PHA), pokeweed, mixed lymphocyte culture (MLC), and natural killer cell consisted of monitoring assay controls, duplicate sample reproducibility, and any trends in reagent reactivity. Assay precision was determined by calculating the CV of the quadruplicates for each variable tested. Also, a mean value of the CV for each assay was calculated. Individual CV's of 15 percent or less were the target values for the stimulated samples in the mitogen and natural killer cell assays. The Student's t-test was applied to duplicates to determine if there was a significant difference in sampling for the functional assays. Critical t-values at the 0.05 significance level were used to determin^ if duplicate sample results varied significantly. Grubbs' statistical test was used to identify any statistically significant outlier. This test was applied only to samples whose CV's were greater than 20 percent at a p-value of 0.01. The mitogen stimulation (PHA, pokeweed) effect was 6-6 �followed by daily evaluation of the radioactive counts in counts per minute (cpm) for each mitogen. When counts fell below expected values, suggesting that reagent deterioration had occurred, new aliquots were used. QC measures for the cell surface marker assays were calculation of T4+T8/T11 cell ratios, evaluation of flow cytometer computer outputs (cytograms and histograms), and duplicate sample testing. T +T8/T cellular ratios should approximate the value 1.0 for a normal population, validity of cytogram and histogram distributions generated by the flow cytometer was confirmed by the Cellular Immunology laboratory supervisor for each sample analyzed. The percent positive cells for each surface marker was determined in .the duplicates and viewed graphically using a microcomputer program. Any significant differences between duplicates were noted and follgwed for abnormal trends. On completion of this followup effort, the entire cellular immunology data base was reviewed by the Air Force team, laboratory staff, and consultants. Comments attached to the data points were also reviewed. Any data point that appeared unusual was reviewed and identified as an unexplained outlier. Unexplained outliers were deleted from the data base as errors of an unknown nature. This review was conducted without knowledge of exposure status. DATA MANAGEMENT QUALITY CONTROL Overviev of Quality Control Procedures The QC program for the data management activity consisted of multiple checks at all steps of the examination, data collection, and data processing cycle. Data QC procedures for data collection, conversion, and integration were developed before the clinical examinations began. Pretesting of all forms, procedures, and logistic arrangements was conducted 3 weeks before the examinations actually began. Additionally, during the first 2 months of the clinical examinations, all data collection activities were intensely scrutinized to detect and correct procedural deficiencies. QC activities also included automated QC techniques applied to laboratory data, clinical evaluations of all laboratory outliers, review of all physical examination findings by an independent diagnostician, and automated and manual data quality checking of hard copy against transcribed computer files for all questionnaire, physical examination, and medical coding data streams. Five interwoven layers of QC were instituted to ensure data integrity. Efforts focused on (1) data processing system design, (2) design and administration of all exams or questionnaires, (3) data completeness checks, (4) data validation techniques, and (5) quality control of medical records coding. In some cases, the QC procedures about to be described were implemented throughout the data management task rather than assigned to a particular activity. These comprehensive QC procedures will be mentioned where appropriate throughout the remainder of this section. 6-7 �Data Processing System Design For each data stream, standards were set to establish data element format (character or numeric), data element naming conventions, data element text labels, numeric codes for qualitative responses and results, QC range checks for continuous data elements, and QC validity checks for categorical data. A data dictionary provided detailed information on each data element. A systems integration approach was applied to the design and implementation of data collection procedures and techniques so that data emanating from the various study sources (physical examination, questionnaire, laboratory) were consistent in file format and structure. This was necessary to ensure that all data could be integrated into a single data base management system for analysis. Figure 6-1 provides an overview of the QC activities used in the data base management process. Forms and questionnaires were carefully designed to ensure that all required data elements would be collected according to the Study Protocol. The design of these instruments was such that they reflected the order in which the examination itself would be administered and provided for the sequential receding of information to streamline remaining data management activities. Completed medical records and questionnaires were converted from hard copy to machine-readable images using customized data-entry systems or state-of-the-art optical mark reading equipment. Verification procedures were performed to ensure that a uniquely identified participant record existed within each data file, and that the appropriate number of responses for each applicable field was provided. Data files were then verified against original data sheets and corrected as necessary. Data files were then subjected to validity checks. Any potentially conflicting results as well as any data values falling at the extremes of expected ranges were manually reviewed. Extreme values were reverified against the original raw data copies and either corrected or documented as valid results. Potentially conflicting results were returned to the examiners for review. These results were then documented as correctly recorded, corrected, or flagged for exclusion from analysis because of unresolvable examiner errors or omissions. Once the edits were completed and the data reverified, the "cleaned" files or tapes were transferred to the data analysis center for final inspection and integration into the study data base. For this QC measure, each data file was loaded into a Statistical Analysis System (SAS®) data set, and descriptive analyses were run. The validation, correction, transmission, and analysis QC procedures were repeated as necessary to ensure that all extreme or suspicious values had been validated. Design and Administration of Physical and Psychological Examination Forms As mentioned, the examination forms were designed to solicit all required data such that recording time was minimized, comprehension was enhanced, and data input could occur with a minimum of transcription errors. Optical Mark Recognition (OMR) technologies were selected to eliminate the risk of transcription errors and were applied to all psychological tests. 6-8 �Data QC Flow Chart Physical and Psychological Examinations / Laboratory Analysis 1I Onsite Review 1 Level 1 QC Contract Report ^^/~~^ / / / / Interview/ Questionnaire / Contract Report FIR Onsite Review Form QA Check Contract Report Data Cleaning Optical Scan Microprocessor Scan Univariate Analysis Contract Report Univariate Analysis Contract Report i_i Level 2 QC Merge/ Load Automated QC Final Contract Report Figure 6-1. Two Levels of Quality Control Applied to All Collected Data Prior to Statistical Analysis Univariate Analysis / / �Customized mark-sense forms were also developed and OMR technology was used to achieve these same objectives for segments of the physical examination and the self-administered questionnaires. The use of mark-sense forms allowed the creation of computerized data files directly from the raw data recorded on these forms. QC procedures for all data collection instruments began with a review of all forms immediately as they were completed. Any forms containing missing examination results were returned to the examining physician for completion before the participants left the site. Any questionable results or "hard-to-diagnose" conditions (such as heart sounds or peripheral pulses) were verified by the diagnostician at the outbriefing. All examination forms were signed by the examining physician, and the examiner identification number was coded in the data base so that interexaminer variation could be analyzed. Detailed QC records were maintained, which indicated the examining physician and the type of deficiency detected. Deficiency reports were reviewed by the study coordinator to detect any patterns of physician data entry error. A final level of QC audit was accomplished by Air Force statisticians, who conducted a detailed screening of the data and checked for errors. Data Completeness Checks Customized programming of the OMR allowed for the identification of those forms (and their corresponding data records) with missing responses, as well as those with multiple responses to questions that required a single response. The OMR scanner was programmed to reject forms that failed completeness and multiple response checks and to output a control code for each rejected form. The control code identified the location of the first three verification checks failed for a given form. Vhen a raw data form was rejected, the reason for the rejection was determined and the exact data element was corrected by comparing the rejected raw data form to the values recorded in the data record created by the scanner. A customized set of rejection and resolution codes was developed for the study to describe all the reasons for a form's rejection and any subsequent reasons for changing a data value. Various codes identified values recovered from light marks, missing marks explained by examiner comments, and missing comment flags resolved by the presence or absence of text in the comment areas. These codes ensured data completeness by accounting for all questionable or missing responses. (See examples of marksense forms in Figures 4-3 and 4-4.) Some of the rejected forms did not contain actual data errors but rather anomalies created in using mark-sense cards for data collection. For instance, incompletely erased responses and responses marked with too little carbon or graphite were incorrectly counted or missed, respectively, by the scanner. Examiners also tended to clearly mark responses for abnormal findings while bypassing or lightly marking responses for expected or desired findings. Failure of the form tp provide the correct number of expected responses always resulted in rejection. These technology-based errors were resolved, as were the anticipated, more traditional errors. The rejection code, data location code, resolution code, data inspector's initials, and correct data value were directly posted to a 6-10 �participant's data record. This innovative technique not only effectively maintained a comprehensive audit trail of all record manipulations, it also provided a mechanism for measuring the frequency of specific errors. Careful monitoring identified trends where individual data values were missed as well as the frequency with which individual examiners incorrectly marked their examination forms. Statistics were compiled on out-of-range results and data omissions that had been accepted in the previous QC audits. The results were monitored to detect trends, possible bias situations, and other data quality problems. This information was reviewed and relayed to examiners and internal auditors to assist in preventing or correcting chronic, but avoidable, problems. Data Validation Techniques QC activities also included data validation techniques. As mentioned earlier, data files were examined in a series of verification and validation procedures developed to check the results within each participant's record for logical consistency and abnormal findings. Any records noted to have ambiguous findings, incongruent observations, extreme results, or nonobvious errors or omissions were listed.and submitted for review to a physician. Again, clinical judgments were made by the auditing physician in assigning a validation code for each extreme or questionable data result. The validation codes allowed for indicating that data were deciphered from examiner comments or from related findings from another specialty area, or were accurately recorded and logically consistent with other findings for the participant. Data points that could not be definitively validated or recovered through clinical judgment and consultation with the original examiner were assigned codes noting missing or invalid data values. These unrecoverable data points were excluded from subsequent analysis. Medical Records Coding Quality Control Upon completion of the NORC data processing, all AFHS questionnaires were forwarded to SAIC for the medical coding of reported conditions. The International Classification of Diseases, 9th Revision, Clinical Modification (morbidity); International Classification of Diseases, 9th Revision (mortality); Systematized Nomenclature of Medicine (anatomic site); and American Hospital Formulary Service (medications) coding schemes were used, suitably modified. Each questionnaire was coded by two coders working independently. The results of the two coders were forwarded to the USAF for 100-percent QA/QC and final adjudication. The information from the physical examination was coded similarly. After the coding data were adjudicated, they were returned to SAIC for data entry. The coding sheets were batched, key entered, verified, and corrected. The corrections were also verified. The .key entry and verification functions were performed by.various operators. Five percent, or 100 records of each batch (whichever was larger), was randomly selected and subjected to manual reverification. An error rate of greater than 1 percent of this sample mandated reverification of the entire batch. In this final QA/QC check, the automated files were reviewed and compared to the hard copy by trained medical record coders, all of whom satisfied the minimum requirement of Accredited Record Technician or Registered Record Administrator eligibility. 6-11 �A manual tracking system was used to retrieve medical records. A chronological log was maintained to track participant requests for authorization to obtain medical record(s), receipt of the authorizations, requests for records from the provider, and receipt of the records from the provider. Identifying information in these logs included participant name, case number, date of action, condition(s) to be verified, dependent name (if appropriate), and type of medical provider (Federal/non-Federal). Due to the intricacies of obtaining medical records from Federal facilities, this task ultimately became the responsibility of the Air Force. STATISTICAL ANALYSIS QUALITY CONTROL Specific QC measures vere developed for activities falling within the statistical analysis task: construction of data bases for the statistical analysis of each clinical chapter, the statistical analysis itself, and the production of statistical reports to serve as the basis for the clinical chapters. Each specialized statistical data base was constructed by defining and locating each variable within the many subparts of the composite followup data base. Lists of variables and their data sources were submitted to the Air Force for approval. Although the data had been subjected to QC procedures during collection, statistical checks for outliers and other improbable values were conducted? anomalies identified by the statisticians were discussed with those responsible for the data collection, i.e., either NORC or SCRF. QA largely depended on regular communication and general agreement among statisticians. Several meetings and consultations among the Air Force team, the Principal Investigator, the SAIC statisticians, and the University of Chicago staff members were held in conjunction with the development of the data analysis plan. During the course of the analysis there were frequent telephone conversations. Any problems arising in the statistical analysis were resolved by team discussion. The software was checked by comparing results from analyses on the same variable by different programs (for example, BMDP*-LR [logistic regression] and BMDP®-4F [log-linear model] will give the same results for dichotomous variables when the program options are chosen properly). The statisticians frequently checked that the number of observations used in an analysis was correct, and peer review ensured that the program code was appropriate for the chosen procedure. The analyses were conducted in accordance with the data analysis plan which was reviewed extensively. Throughout the study, duplicate data bases were maintained by the USAF and SAIC. Upon completion of the analyses, SAIC delivered all analysis software and SAS data sets for each clinical area to the USAF for final review and archiving. All tables and statistical results were checked against the computer output from which they were derived, and all statistical statements in the text were checked for consistency with the results given in the tables. Additionally, drafts of chapters in the report were reviewed by the USAF and SAIC investigators, and the QRC. 6-12 �CHAPTER 6 REFERENCES 1. Bissell, A.F. 1969. CUSUM techniques for quality control. Appl. Stat. Vol. 18. 2. Lucas, J.M., and R. Crosier. 1982. Fast initial response for CUSUM quality control schemes: Give your CUSUM a headstart. Technometrics Vol. 24. 3. Grubbs, F.E. 1969. Procedures for detecting outlying observations in samples. Technometrics XI:1-21. 6-13 �CHAPTER 7 STATISTICAL METHODS . This chapter summarizes the key statistical elements of the study design, the statistical analysis issues, and the specific statistical methods used in the analysis. Additional details may be found in the USAF Study Protocol. The primary focus of the statistical analysis was a contrast of health status of the Ranch Hand and Comparison groups. Assessments were made of the proportions of participants with abnormal findings and of mean levels of key laboratory measurements. The analyses encompassed both simple contrasts between the two groups and more complex methods, in which adjustment was made for important covariates. In addition to these analyses, the possibility of an increasing response of medical problems with herbicide dose was explored, since if indeed there were an effect, more problems would be expected among the more heavily exposed. Although exact dosage information is not available, an exposure index was developed for the exposed population (the Ranch Hands) that approximates the potential herbicide exposure of each individual, incorporating information such as the occupation of the individual, his period of duty in the spraying operation, and the numbers of barrels per day of herbicide used during that period. Details on the exposure index are given in Chapter 8. Dose-response analyses were conducted for the Ranch Hands only, using this exposure index as a surrogate measure of dose. Interpretation of the results of the exposure index analyses, however, depends critically on the accuracy of the exposure index, which presently can be regarded as only fair. (Improved dosage information will be obtained for future studies from recently developed serum dioxin assay techniques.) Thus, the analyses of overall group differences between the Ranch Hands and the Comparisons are given primary emphasis, and the exposure index analyses merely supplement them. STATISTICAL STUDY DESIGN An overt herbicide effect would be characterized by more symptoms, signs, abnormal laboratory tests, syndromes, or diseases in the Ranch Hand group than in the Comparison group. If the disease(s) were fatal, increased mortality might also be observed. A subclinical herbicide effect would be detected as an increase in abnormal findings on the physical examination (particularly laboratory tests) that may or may not also be associated with symptom reporting or increased mortality. Thus, the basic objective of the statistical analysis is to test for differences between the Ranch Hand (exposed) group and the Comparison (nonexposed) group. 7-1 �In general, two types of data are used in the analysis. First, there are subjective data on symptoms reported by the participant in the questionnaire and in the review-of-systems section of the physical examination. Second, there are objective data, which include medical findings or signs identified during the physical examination, or by reviews of laboratory results, medical records, and death certificates. Symptoms reported by the study participants are subjective by definition, and are subject to influences that could result in erroneous conclusions. An association found between reported symptoms and herbicide exposure must be subjected to further confirmation, as the observations may result from over- or under-reporting bias and may not be indicative of a true herbicide effect. On the other hand, the medical findings data do not suffer from the same degree of participant influence. The medical findings and medical records review were conducted by highly trained individuals employed for the duration of the data collection and assessment phases of the study. They were held to stringent QC standards, as described in Chapter 6, to ensure that these data were as objective and accurate as possible. Incorporated in the study design is a feature that attempts to check for and correct symptom-reporting errors. A key component is a reported symptom verification process conducted by reviewing participant medical records and findings from the physical examination. In the retrospective morbidity portion of the study, the participant is questioned on past illnesses and medical conditions. With the participant's consent, an effort is made to obtain the medical records to verify the reported condition and, thus, to substantiate any unverified conditions. In addition, the study design includes verification of negative responses to determine unreported conditions. The medical record review process is time intensive and only a portion of the data was available for analysis in this study. Over-reporting was assessed by comparing the reported illness rates with the results of the physical examination and medical record review. Similarly, the assessment and correction of under-reporting requires the review of medical records to identify unreported illnesses. Obviously, this under-reporting assessment is restricted to conditions for which medical care was obtained or that were identifiable at the physical examination. STATISTICAL ISSUES In conducting the statistical analysis of the data in this study, there are a number of underlying issues. Except for bias, which is the topic of Chapter 5, these issues are discussed in this section. However, based upon the results of the bias analysis presented in Chapter 5, all statistical analyses in the clinical chapters use the contrast of Ranch Hands versus the total Comparison group. For the purposes of completeness and cross-reference to the Baseline report, identical analyses using the contrast of the Ranch Hands versus the Original Comparisons have been conducted, and these results are presented in the form of summary tables in each chapter appendix. 7-2 �Intervening Variables When comparing any two groups of individuals, the exact proportion of diseased individuals in each group is usually found to differ. The purpose of classical statistical hypothesis testing is to determine whether the observed difference in disease rates could be due to chance alone. If the observed difference is not attributable to chance, the two groups are considered representative of two truly different populations. If a statistically significant difference is found between the Ranch Hand group and the Comparison group, results from more rigorous statistical procedures must be examined and the medical context considered before the possibility of a causal relationship between disease and group (exposure) can be entertained. Alternatively, the absence of a statistically significant difference between groups does not exclude the possibility of a true causal relationship between exposure and disease. Thus, group associations, whether significant or not, should be examined with adjustment for other variables called intervening variables (explanatory variables, risk, factors, or covariates) that may account for, or mask, a true effect. For example, the two groups might differ with respect to age or racial composition, each of which may affect the outcome of the study. To protect against this, the technique of matching was used: The Ranch Hands and Comparisons were matched on age, race, and military occupation. Since it is not feasible to perfectly match a Comparison to an exposed individual with respect to all important explanatory variables, statistical procedures may be used to adjust for such explanatory variables so that valid interpretations can be made of apparent group differences. Thus, it was necessary to identify and collect data on suspected explanatory variables. Unfortunately, there is no way to ensure that all important intervening variables are taken into account. The best method that can be achieved is to incorporate all known covariates in the data collection and analysis. In most studies, covariates are variables measured prior to exposure. However, in the AFHS, except for the matching variables and historical data related to events prior to service in Southeast Asia, most covariate values were obtained at the Baseline or first followup interview and physical examination, which occurred 10 to 20 years following exposure. These covariates can generally be referred to as time-dependent covariates. They can elucidate the causal path between exposure and a particular disease; however, they are in a sense both dependent and independent variables, and therefore, analyses involving such covariates require careful interpretation. Besides covariates, both confounding variables and interactions must also be considered. A confounding variable is an intervening variable associated with both disease and exposure. (This is in contrast with a covariate that is associated only with disease.) Adjustments must be made for confounding variables to avoid a biased estimate of the group-disease relationship. An interaction exists when the effect of one variable varies across the levels of another variable. For example, the group difference might be large in one occupation, group and negligible in another. Incorporating interactions in the analysis allows for the identification of subpopulations at increased or decreased risk. 7-3 �Pover Conducting a statistical test using a Type I error, also called alpha level, of 0.05 (a =0.05) means that, on the average, in 5 cases out of 100, a false conclusion that an association (herbicide effect) exists would be made when in reality, there is no association. The other possible inference error (called a Type II error) is that of failing to detect an association when it actually exists. The probability of a Type II error (3) for a statistical test is 1 minus the power of the test. The power of the test is the probability that the test will reject the hypothesis of no herbicide effect when an effect does in fact exist. The power of a test depends on the group sample sizes, the disease prevalence rate, and the true group difference measured in terms of relative risk. Table 7-1 contains the approximate sample size required to detect specific relative risks with an approximate power of 0.8 (3 =0.2) using an alpha level of 0.05 for a two-sided test and assuming equal Ranch Hand and Comparison group sizes and unpaired analyses. Relative risk is the ratio of the disease prevalence rate of the Ranch Hand and Comparison groups. Conditions or diseases with comparison population prevalence rates and exposed group relative risks corresponding to those below the heavy black line on the table can be detected with an approximate 0.8 probability with the sample sizes used in this study. Table 7-2 provides the same information for continuous variables in terms of percentage mean shift and variability, assuming unpaired testing of a normally distributed variable and equal sample sizes. In the first followup of the AFHS, 1,016 Ranch Hands participated in the physical examination. In this size group, the chance of identifying zero cases of a disease with a prevalence of 1/500 or less is greater than 10 percent. Table 7-3 contains the probability of encountering no cases of disease states for cumulative prevalence rates of 1/200, 1/500, 1/1,000, 1/2,000, 1/5,000, and 1/10,000. Multiple Endpoints and Comparisons In developing the Protocol for the AFHS, previous animal and epidemiclogic studies, case reports, and veterans' concerns were reviewed to delineate the possible effects of exposure. The conclusion was reached that a comprehensive evaluation was needed due to the lack of an easily identifiable symptom complex in individual patients. Consequently, the morbidity study is very broad in scope, involving the collection and analysis of data related to general health indices as well as specific organ systems and clinical disease categories. The large number of endpoints under consideration presents a difficult problem in the assessment of Type I error rates. More than 150 dependent variables were tested, not to mention tests for interaction and multiple contrasts among the low, medium, and high exposure-level categories in the exposure index analyses. Furthermore, the dependent variables were correlated to varying degrees, and this makes it even more difficult to assess the attained significance levels. To allow for multiple endpoints, Bonferroni's inequality, which requires significance at the a /K, level where K. is the number of endpoints considered, may be used, but this procedure 7-4 �TABLE 7-1. Required SanpLe Sizes To Detect Group Differences in Two-Sanple Testing Assuming Equal Sanple Sizes* (Relative Risk Calculations) Occurrence Rate of Disease in Control Population Relative Risk (Multiplicative Factor of Occurrence Rate for Exposed Group) 1.25 1.50 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 ID75SO 2,822,082 783,901 235,164 78,384 43,544 29,391 21,944 17,415 14,393 12,243 10,640 B75QO 1,410,882 391,901 117,564 39,184 21,766 14,690 10,968 8,703 7,193 6,118 5,317 1 T^OO 281,922 78,301 23,484 7,824 4,344 2,930 2,187 1,735 1,433 1,218 1,058 1 300 140,802 39,101 11,724 3,904 2,166 1,460 1,089 863 713 606 526 1 KB 27,906 7,741 2,316 768 424 284 211 167 137 116 ICO 1 30 13,794 3,821 1,140 376 206 137 101 79 65 54 47 *This study has unequal sample sizes; therefore, the tabled values are understated. The similar table in the Baseline Morbidity Report, 24 February 1984, is in error because tabulated sample sizes were only one-half of their correct values. 7-5 �TABLE 7-2. Required Sample Sizes To Detect Group Differences in Two-Sample Testing Assuming Equal Sample Sizes* (Mean Shift Calculations) Mean shift Variability (CT/M) 0.05 0.10 0.25 0.50 0.75 0.5* 1,568 6,272 39,200 156,800 352,800 1.0* 392 1,568 9,800 39,200 88,200 1.5% 175 697 4,356 17,423 39,200 2.0% 98 392 2,450 9,800 22,050 2.5% 63 251 1,568 6,272 14,112 5.0% 16 63 392 1,568 3,528 7.5% 7 28 175 697 1,568 10.0% 4 16 98 392 882 *This study has unequal sample sizes; therefore, the tabled values are understated. The similar table in the Baseline Morbidity Report, 24 February 1984, is in error because tabulated sample sizes were only one-half of their correct values. 7-6 �TABLE 7-3. Probability of Zero Cases as a Function of Prevalence Disease Prevalence Probability of Finding Zero Cases in a Group of 1,016 Participants 1/10,000 1/5,000 1/2,000 1/1,000 1/500 1/200 0.903 0.816 0.602 0.362 0.131 0.006 becomes increasingly more conservative as the correlation among the endpoints increases. For the analysis results in this report, an alpha level of 0.05 was used for each dependent variable. In addition, group contrasts in strata defined by levels of a covariate involving in a group-by-covariate interaction were assessed by an alpha level of 0.05. The same was true for exposure level strata. In light of the multiple-endpoints problem, extreme caution in the interpretation of statistical results was required. A first consideration was the strength of the association in terms of the significance of the relative risk or difference in group means. All associations with p-values of 0.10 or less were examined and are described in this report. Then, careful consideration was given to the pattern of statistically significant results. Were only a few sporadic endpoints statistically significant, or was significance achieved on a number of endpoints indicating the same organ system failure? Were the significant results all in the same direction, and did they make biological and clinical sense? Did they confirm previous studies, or were they new findings? Paired Versus Unpaired Analyses Matching subjects in a study design on selected variables improves the comparability of the groups to be compared and, depending on the relationship of the matching variables to the study objective, the matching can be used explicitly in the analysis. In this study, the Comparison group was matched to the exposed group on age (to the nearest month of birth), race (Black, nonblack), and occupational category (officer-pilot, officer-navigator, officer-nonflyer, enlisted flyer, enlisted groundcrew). The matching was exact for occupational category, nearly exact for race (three mismatches occurred because of recording errors), and very close with respect to age (69% of the mortality population was matched to the nearest month of birth and more than 95% to the nearest year of birth). The general approach in this report, however, was to conduct unpaired analyses using all available data, based on stratification and/or covariate adjustment. In an unpaired analysis, the matching still serves to improve 7-7 �the comparability of the two groups, and precision is usually gained from the stratification and covariate adjustment. Mortality and Morbidity Data The AFHS incorporated both mortality and morbidity endpoints. The mortality data have been, and will continue to be, subjected to separate analysis. Interpretation of the morbidity analyses must be made in the light of the mortality results, particularly as the study continues and the number of deaths increases. Differential mortality in the two groups could obviously have an important impact on contrasts of physical examination findings in the surviving cohorts. This issue was examined in the analysis of selected diseases, for example, cancer. Outpoints The variables in this study were discrete, categorical, or continuous. Many served primarily as dependent variables, and when in the continuous form, powerful analyses were possible. In other settings, particularly when log-linear or logistic regression models were fitted, it is often necessary to dichotomize or discretize the continuous variables. Discretization, by establishing suitable nonoverlapping intervals or cutpoints, was often the result of a judgment requiring both statistical and clinical input. In general, cutpoint decisions considered the form of the variable, distribution of the variable, established values (e.g., cholesterol, normalabnormal, as specified by a given technique in a given laboratory), scientific values set by precedence (e.g., systolic and diastolic normal threshold 140/90), and error induction by another variable (e.g., use of the blood pressure threshold in obese-armed individuals). The approach to the selection of appropriate cutpoints was to select all cutpoints on a case-by-case basis and, where indicated, use the norms of the SCRF laboratory. Exclusions Due to medical considerations, certain subjects were excluded from the analyses of selected clinical categories. The exclusions were generally defined in the Baseline study and are identified in the clinical chapters of this report. Other exclusions were the result of missing data. OVERVIEW OF STATISTICAL PROCEDURES This section summarizes the basic statistical approach used in the data analysis of the first followup of the AFHS. The approach consisted of four parts: (1) preliminary analysis of the dependent variables and covariates to check for data anomalies and to obtain a general overview of the data, (2) core analyses to carefully determine any- possible effect of herbicide exposure, (3) analysis of the exposure index to investigate the dose-response relationship for the Ranch Hand group only, and (4) longitudinal analysis to examine changes over time. A summary of the statistical techniques utilized is provided in Table 7-4. This basic approach was utilized in the analyses for each clinical category. 7-8 �TABLE 7-4. Summary of Statistical Procedures Chi-Square Contingency Table Test The chi-square test of independence2 is calculated for a contingency table by the following formula: X2 = Z(f0-f9)2/f9 where the sum is taken over all cells of the contingency table and fo=observed frequency in a cell fa=expected frequency under the hypothesis of independence. Large values indicate deviations from the null 2 hypothesis and are tested for significance by comparing the calculated X to the tables of the chi-square distribution. Fisher's Exact Test Fisher's exact test is a randomization test of the hypothesis of independence for a 2x2 contingency table. This technique is useful for small samples and sparse cells. This is a permutation test based on the exact probability of observing the particular set of frequencies. General Linear Model Analysis The form of the general linear model1 for two independent variables is: Y - a + 01X1 + 32X2 + fl^Xj + e where Y = dependent variable (continuous) a = level of Y at Xx = 0 and X2 = 0, i.e., the intercept X1,X2 = measured value of the first and second independent variables, respectively, which may be continuous or discrete P^jSj = coefficient indicating linear association between Y and X a , Y and X2, respectively P12 = coefficient reflecting the linear interaction of Xt and X2 e = error term. This model assumes that the error terms are independent and normally distributed with a mean of 0 and a constant variance. Extension to multiple independent variables and interaction terms is immediate. 7-9 �TABLE 7-4. (continued) Summary of Statistical Procedures Linear regression, multiple regression, analysis of variance, and analysis of covariance are all examples of general linear model analysis. Kolmogorov-Smirnov Distribution Test The Kolmogorov-Smirnov (K-S) test is a nonparametric procedure which assesses differences between the distribution of two samples. Specifically, the K-S procedure tests the hypothesis that populations n, and Ji2 are identical and is designed to detect all possible deviations from this hypothesis. The assumptions of the K-S test are that the observations from the two samples are mutually independent and that both sets of observations are samples from the same distribution. Logistic Regression Analysis The logistic regression model2'4 enables a dichotomous dependent variable to be modeled in a regression framework with continuous and/or discrete independent variables. For two risk factors, such as group and age, the logistic regression model would be: logit P = <x+ (31X1 + 02X2 + 13^X^2 + e where P = probability of disease for an individual with risk factors X and X2 logit P = In (P/l-P), i.e., the log odds for disease X1 = first risk factor, e.g., group X2 = second risk factor, e.g., age. The parameters are interpreted as follows: a = log odds for the disease when both factors are at a 0 level 31 = coefficient indicating the group effect adjusted for age P2 = coefficient indicating the age effect adjusted for group 312 = coefficient indicating the interaction between group and age e = error term. In the absence of an interaction (|31? = 0), exp(p1) reflects the adjusted odds ratio for individuals in Group 1 (Xt = 1) relative to 7-10 �TABLE 7-4. (continued) Summary of Statistical Procedures Group 0 (Xx =0). If the probability of disease is small, the odds ratio will be approximately equal to the relative risk. Homogeneity of the odds ratios5 across different strata was assessed by the method of Breslow and Day. Throughout this report the adjusted odds ratios are referred to as adjusted relative risks. Correspondingly, in the absence of covariates (i.e., unadjusted analysis) the odds ratios are referred to as estimated relative risks. Proportional Odds Model The proportional odds model6 allows for the analysis of an ordered outcome variable. The model assumes that the odds of falling below a certain level rather than above it for individuals at different levels of an independent variable X are in constant ratio. For example, if the response takes one of the four values "excellent," "good," "fair," or "poor," and X is a simple indicator variable designating group (Ranch Hand versus Comparison), then the proportional odds model states that the odds for responding "excellent" versus "good," "fair," or "poor" in the Ranch Hand group are a multiple, exp(fJ), of the corresponding odds in the Comparison group. Likewise, the odds for responding "excellent" or "good" versus "fair" or "poor" in the Ranch Hand group are the same multiple, exp(P), of the corresponding odds in the Comparison group, as are the odds for responding "excellent," "good," or "fair" versus "poor" in the two groups. Thus, the model is appropriate whenever one frequency distribution is "shifted left" relative to another distribution. Incorporation of other variables into X allows the estimation of proportional odds ratios adjusted for covariates. Let the ordered response Y take values in the range 1 to K, and let JlA(X), i=l,...,K, denote the probability of responding at level i for an individual with covariate vector X. Let K.(X) be the odds that Y< j given X, i.e. , ~ n,(X) + rc (X) + ... + Ji.(X) K,(X) The proportional odds model specifies that ^(X) =• K.J exp(|3'X), for constant K^ 7-11 �TABLE 7-4. (continued) Summary of Statistical Procedures Thus the ratio of odds for individuals at covariate levels X1 and X2 is exp{p'(X1 - X2)} ^ ~2 and depends only on Xx - X2 and not on j . Log-linear Analysis Log-linear analysis2 is a statistical technique for analyzing crossclassified data or contingency tables. A saturated log-linear model for a three-way table is: 10 <Zijk> = U o U + U l(i) 13(ik) + U + U 2(j) + U 3(k, + U 12(lj, + U 23(jfk) + 123(ijk) where Zi Ut = expected cell count A . = specific one-factor effect U 12(i .. U = specific two-factor effect or interaction i23<ijk) = tnr ee-factor effect or interaction. The simplest models are obtained by including only the significant U-terms. Adjusted relative risks are derived from the estimated U-terms from an adequately fitting model. McNemar's Test McNemar's test4 effectively considers discordant pairs in which only the Ranch Hand or only the Comparison member in each pair experiences the abnormality. Using a chi-square approximation with continuity correction, the following statistic is used to test whether the off-diagonal entries are evenly divided: , (|b-c|-D2 b+c Where b and c are the number of pairs in which only the Ranch Hand is abnormal or only the Comparison is abnormal, respectively. This test is compared to a chi-squared distribution with one degree of freedom. 7-12 �TABLE 7-4. (continued) Summary of Statistical Procedures Test for Linear Trend For a kx2 contingency table in which the k groups fall into a natural order, Armitage developed a test for a linear trend in the proportions. Let PA denote the proportion of individuals in the ith row possessing some attribute (e.g., proportion of individuals with abnormal values at each of the three exposure level categories). A score, X., is assigned to each of the k levels of the row variable, and the regression coefficient, "& of Pi on XA is estimated. The regression coefficient is estimated in the usual way except that P1 is weighted by the sample size, n.. , in each row. provides a normal deviate for testing the null hypotheses of 0 = 0. 7-13 �Preliminary Analysis The preliminary analysis included the calculation of basic descriptive measures for the dependent and independent variables (covariates), for each group (Ranch Hand and Comparison). The descriptive measures included frequency distributions, histograms, mean, median, standard deviation, and range. These analyses provided an overview of each variable and the relationship of the Ranch Hand group to the Comparison group. In addition, the preliminary analysis provided insight for the construction of composite variables, the plausibility of normal/abnormal limits and cutpoints, and the choice of possible transformations to enhance the normality of the distribution of continuous dependent variables. Another purpose of the preliminary analysis was to examine the relationship between the covariates and the dependent variables and the relationships between and among the covariates. To accomplish this, cross tabulations of discrete variables were constructed and analyzed by the chi-square, or Fisher's exact test. For continuous variables, simple t-tests of group differences were done and product-moment correlation coefficients were computed. The preliminary analyses were accomplished with the use of the SAS*. Selected covariate tables are presented in the clinical chapters for illustration. Core Analysis The core analysis consisted of a series of steps taken to ascertain whether or not the data indicated a significant difference between the Ranch Hand and Comparison groups for each dependent variable. Both unadjusted and adjusted analyses were performed and are presented for each clinical chapter. Unadjusted analyses are simple contrasts between the Ranch Hand and Comparison groups of the mean values, or proportion with abnormal values, of each dependent variable, by t-tests, one-way analysis of variance, Fisher's exact test, or chi-square tests, as appropriate. Adjusted analyses take into account important covariates in the assessment of possible group differences, i.e., the covariates are included in the general linear, logistic regression, proportional odds models, or log-linear models. Continuous Dependent Variables When the dependent variable was continuous, the general linear models (GLM) procedure of SAS® was used to fit a model of the dependent variable in terms of the group indicator (Ranch Hand or Comparison) and appropriate covariates, and interactions between covariates. The covariates could be continuous or categorical variables. If necessary, the dependent variable was transformed prior to analysis by a transformation (e.g., logarithm) to enhance normality of its distribution. When a "best" model was fitted, according to the strategy outlined below, the test for significance of the group difference was then done on the adjusted group means, provided there were no significant interactions between the group indicator and any of the covariates. Group differences in the presence of interactions were assessed using stratification by different levels of the covariate(s) involved in the interaction or estimation of group differences at selected covariate levels using the best model identified. 7-14 �For some non-normally distributed dependent variables, the KolmogorovSmirnov (K-S) test of significant differences betveen the distributions of the variables in the two study groups was conducted. The K-S test is a nonparametrie test for the equality of two distributions designed to detect broad classes of alternatives. Categorical Dependent Variables Discrete dependent variables were analyzed by methods parallel to those used for continuous variables. For dichotomous variables, logistic regression was carried out by the program BMDP*-LR; for this analysis, the covariates could be either continuous or discrete. For polychotomous dependent variables, where the number of categories was three or more, log-linear modeling was performed by the use of the program BMDP®-AF, by incorporating the full (k)-factor interaction term involving the (k) covariates used in the model. For this type of analysis, all covariates had to be categorized. The models were fitted by the method of maximum likelihood. To make the results parallel to those obtained by logistic regression, i.e., because of the distinction between dependent and independent variables, the marginals were fixed in the model, effectively converting the log-linear model into a logit model. The significance of the relative risk for group was determined by examination of the appropriate model, as determined by the study, that includes all statistically significant effects and the group indicator or by examination of the significant interactions. Adjusted relative risks were derived from the coefficients of the appropriate model. Modeling Strategy In each clinical category, many covariates were considered for inclusion in the statistical models for adjusted group contrasts. The large number of such covariates and consequent interaction terms and the resulting difficulties of interpretation forced the adoption of a strategy for identifying a moderately simple model involving only significant effects. Interpretation of possible group differences was then made in the context of this simple model. A schematic representation of the generalized modeling strategy is provided in Appendix E. An initial model including all two-factor interactions and all threefactor interactions involving group was examined. Global tests at the 0.15 level, or individual tests at the 0.05 level, were used to screen out unnecessary three-factor interactions. A hierarchical stepwise deletion strategy was then used, eliminating effects with p>0.05 (except the main group effect) and retaining lower order effects if involved in higher order interactions, to result in the simplest model. Interactions between covariates, if significant, were retained as effects. The analysis was carried out by different statisticians, and there are necessarily subtle differences between them in presentation and approach. This, however, should not affect any of the final conclusions as to group differences. In some chapters, for instance, adjusted group means are presented, and in others the differences between the adjusted group means are 7-15 �presented. In each case, the same conclusion may be drawn since the statistic of relevance is the difference between the adjusted group mean and the associated p-value. Further, if an interaction of group with a continuous covariate was found, two equally valid methods were used to illustrate how the interaction was arising. One method was to categorize the continuous covariate and describe the group differences within each (covariate-defined) stratum. Another technique was to present group differences for several selected values of the covariatel Further, in the presence of small frequencies of abnormalities, exposure index analyses were occasionally carried out using only the main effects model (i.e., using group and all the covariates but not including interaction terms). It is recognized that, due to the large number of group-by-covariate interactions examined (up to 7 per dependent variable) for some 150 variables, some of the group-by-covariate interactions judged significant at the 0.05 level may be spurious, i.e., chance occurrences and not of biological relevance. This is analogous to the concept of Type I error for a two-sample adjusted contrast. When several covariates are used in an adjusted analysis of the group contrast for a single dependent variable, and each group-by-covariate interaction is tested at the 0.05 level, the chance of finding at least one that is statistically significant is, of course, greater than 0.05; this is assuming that there is no group effect or group-by-covariate interaction. How much greater depends on the interrelatedness of the covariates and their association with the dependent variable. For a study of this size, with many interrelated dependent variables being examined, it is not known how to estimate the number of group-bycovariate interactions that may be due to chance alone. However, this frequency clearly will be more than 5 percent. It is noted that this concept should be considered when significant group-by-covariate interactions are interpreted. Further, it is important that the size of the p-value associated with each group-by-covariate interaction be carefully weighed, as should be the pattern of the interaction findings for related dependent variables. EXPOSURE INDEX ANALYSES As described in Chapter 8, the exposure index was constructed to portray the level of dose of the herbicide for the Ranch Hand or exposed group only. Exposure index analyses were conducted on all dependent variables. The objective of the analyses was to determine if there was a difference in the levels of the dependent variable corresponding to the levels of the exposure index. The exposure index was trichotomized as high, medium, and low, separately, for each of the three occupational groups: officer, enlisted flyer, enlisted groundcrew. Thus, separate analyses were conducted for each occupational cohort. Discrete dependent variables were evaluated using log-linear and logistic regression models, treating exposure level as a categorical variable (by means of two indicator variables) and adjusting for covariates. For continuous dependent variables, a general linear model was fit, adjusting for covariates and using two indicator variables to designate exposure level. Contrasts between medium and low, and between high and low exposure levels, were also examined. 7-16 �LONGITUDINAL ANALYSES General Another objective of the AFHS is to observe the Ranch Hand population and the Comparison group carefully over time for the emergence, or deleterious rate change, of symptoms, signs, laboratory parameters, or frank disease. This followup objective is not without scientific and logistic challenge, considering mobile populations, problems of loss to study, changing laboratory methods and diagnostic criteria, and the diversity of many changing factors over a period encompassing numerous follovup examinations. The following sections describe the statistical procedures used for both continuous and categorical longitudinal data. Continuous Data A repeated measurements analysis of variance procedure10 was used to analyze the variables measured on a continuous scale. The model for the dependent variable (Y) measurement on the kth participant (i^) in the ith group (o^) at the jth time ( 3 ) is as follows: |. Y ijk = u + a£ + i^,., + 0j + 0 ^ + eijk 0 The sources of variation and associated degrees of freedom are given below: Source Degrees of Freedom* Group (Ranch Hand vs. Comparison) Subject/Group Time (Baseline vs. Followup) Group-by-Time (Subject-by-Time)/Group 1 2,108 1 1 2,108 *Based on 971 Ranch Hands and 1,139 Comparisons. The primary source of interest is the group-by-time interaction (o3ij). Vith measurements on each participant at only two times (Baseline and followup), a test on this interaction is equivalent to a test on the equality of mean differences (Baseline minus followup) between the Ranch Hand and Comparison groups. Care must be taken in the interpretation of the main effect, time ((3.) (i.e., overall Baseline mean versus overall followup mean). This effect is totally confounded with laboratory differences, and with over 2,000 participants, "significant differences".come easily. The source of variation due to group (a.) reflects a difference between the overall Ranch Hand and Comparison means (averaged over both times). This source should complement the group difference findings at Baseline and at 7-17 �followup, provided the group changes were consistent (no significant groupby-time interaction). All available participants were used at each Baseline and followup analysis, while only the participants with both measurements are included in the repeated measurement analysis. Covariates were not used in these analyses. Generally, time-independent (e.g., year of birth) and time-dependent (e.g., smoking) covariates can be used. Only the time-dependent covariates would affect the primary source of interest, namely the group-by-time interaction. Hence, all of the previously considered time-independent covariates would affect only the main group effect, a source not of primary interest since it is being considered in the separate cross-sectional analyses. Categorical Data Frequently, data were collected as normal-abnormal, or continuous measurements were discretized into this binomial response. For each Ranch Hand and Comparison group, a Baseline versus followup 2x2 (normal-abnormal) table of frequencies was prepared (paired data): Followup Ranch Hand Comparison Abnormal Normal Abnormal Normal Abnormal Abnormal Normal Normal Baseline As with the McNemar test, only the Abnormal-formal and Normal->Abnormal off-diagonal data were used in further contrasts. A conventional ) test was ( used to test the null hypothesis of a comparable change pattern for the two groups (unpaired data). Change Pattern Normal- AbnormalAbnormal Normal Ranch Hand Group Comparison This test is equivalent to 1 testing no group-by-time-by-endpoint interaction in a matched pair analysis. 7-18 �CHAPTER 7 REFERENCES 1. Neter, J., and W. Wasserman. 1974. Applied linear statistical models. Homewood, Illinois: Richard D. Irvin, Inc. 2. Bishop, Y.M.M., S.E. Feinberg, and P.W. Holland. 1975. Discrete multivariate analysis; Theory and practice. Cambridge: MIT Press. 3. Hollander, M., and D. Wolfe. 1973. Nonparametric statistical methods. New York: John Wiley & Sons. 4. Fleiss, J.L. 1981. Statistical methods for rates and proportions. 2d ed. New York: John Wiley & Sons. 5. Breslow, N.E., and N.E. Day. 1980. Statistical methods in cancer research. Volume I, The Analysis of Case-Control Studies, Lyon, International Agency for Research on Cancer (IARC Scientific Publications No. 32). 6. McCullagh, P. 1980. Regression models for ordinal data. J. Royal Stat. Soc. B42(2):109-142. 7. Armitage, P. 1955. Tests for linear trends in proportions and frequencies. Biometrics 11, 375-386. 8. Box, G.E.P., and D.R. Cox. 1964. An analysis of transformations. J. Royal Stat. Soc. 826:211. 9. Feinberg, S.E. 1981. The analysis of cross-classified data. Cambridge: MIT Press. 10. Winer, B.J. 1971. Statistical principles in experimental design. 2d ed. New York: McGraw Hill. 11. Breslow, N.E. 1982. Covariance adjustment of relative risk in matched studies. Biometrics 38:661-672. 7-19 �CHAPTER 8 EXPOSURE INDEX This chapter describes the development of the exposure index of the AFHS. Portions of this chapter are paraphrased from the Baseline Morbidity Report, 24 February 1984. An increased incidence of adverse health effects at higher levels of exposure represents a classic increasing dose-response relationship. The potential relationship of clinical endpoints with herbicide exposure can be tested using an estimate of exposure, hereinafter called an exposure index, for each member of the Ranch Hand cohort of the AFHS. However, due to a variety of biomedical mechanisms, there can be exceptions to the hypothesis of a consistently increasing dose-response relationship. An index of potential exposure to any of four TCDD-containing herbicides from fixed-wing spray missions was constructed for each Ranch Hand from the available historical data. The index serves as an estimate only, since the actual concentration of TCDD in the herbicides varied from lot to lot and individual assessments of actual body burden cannot be made. The four TCDDcontaining herbicides used in the development of the index are Herbicide Orange, Herbicide Purple, Herbicide Pink, and Herbicide Green. The exposure index was designed to correlate as closely as possible with exposure and is not an exact measure of actual individual exposures. Although the index contains errors when used to assess the exposure of a specific individual, it provides some degree of useful inference for groups of similarly exposed individuals. In summary, the exposure index in the AFHS is a surrogate indicator of TCDD exposure. The exposure index developed for the Baseline study and used in this report is defined in Table 8-1. The exposure index for the ith subject, denoted E , is defined as the product of the TCDD weighting factor, the gallons of TCDD-containing herbicide sprayed in the Republic of Vietnam theater during the tour of the ith subject, and the inverse of the number of men sharing the subject's duties during the tour of the ith subject. Each of these factors is described below. The TCDD weighting factor reflects the estimated relative concentration of TCDD in the herbicides sprayed. The estimated mean concentrations of TCDD in Herbicide Orange, Herbicide Purple, Herbicide Pink, and Herbicide Green are 2 parts per million (ppm), 33 ppm, 66 ppm, and 66 ppm, respectively. Archived samples of Herbicide Purple indicate a mean concentration of approximately 33 ppm, and samples of Herbicide Orange had a mean concentration of about 2 ppm. Since Herbicide Pink and Herbicide Green contained twice as much 2,4,5-T as Herbicide Purple, the mean concentration- of TCDD in these two herbicides was approximately 66 ppm. Based on procurement records and dissemination information, a combination of Herbicide Green, Herbicide 8-1 �TABLE 8-1. Algorithm for Exposure Index (Gallons of TCDDContaining Herbicide Weighting^ x Sprayed in the RVN Theater During the (Factor J Tour of the ith Subject, [TCDD E.. = ] Number of Men with Subject's Duties in the RVN Theater During the Tour of the ith Subject where E. = Exposure Index for the ith Subject TCDD Weighting Factor ( 24.0 if before 1 July 1965 I 1.0 if on or after 1 July 1965 Since prior to 1 July 1965 a combination of Herbicides Green, Pink, and Purple with a mean concentration of 48.0 ppm was sprayed, and after 1 July 1965 only Herbicide Orange with a mean concentration of 2 ppm was sprayed, the ratio is then 48?2 or 24:1. Gallons of TCDD-Containing'l Herbicide Sprayed in the RVN Theater During the Tour of the ith Subject [Number of Gallons of Herbicides Orange, ^Green, Pink, and Purple Expressed in Herbicide Orange Equivalent Gallons teased on Mean Concentration of TCDD Using the following: Herbicide Mean Concentration (ppm) of TCDD Green Orange Pink Purple 66 2 66 33 Number of Men with Subject's "| Duties in the RVN Theater During^ the Tour of the ith Subject ) Source: [Number of Personnel ^in the Same (Occupational Category Baseline Morbidity Report, 24 February 1984. 8-2 �Pink, and Herbicide Purple was sprayed between January 1962 and 1965. The estimated mean concentration of TCDD for this time was 48.0 ppm, using available data on the number of gallons procured and sprayed. The Herbs Tape and other data sources1 indicate that only Herbicide Orange was disseminated after 1 July 1965. Normalizing to Herbicide Orange, the weighting factor becomes 24.0 before 1 July 1965 and 1.0 after 1 July 1965. Using the Herbs Tape, Contemporary Historical Evaluation and Combat Operations (CHECO) Reports, and quarterly operations reports, a table of gallons of TCDD-containing herbicide sprayed for each month of the operation was constructed. Gallons of Herbicides Purple, Pink, and Green were converted to Herbicide Orange equivalent gallons based on the TCDD weighting factor of 24.0. This information is provided in Table F-l of Appendix F. The dates and occupational category of each Ranch Hand's tour(s) in the Republic of Vietnam were obtained by a manual review of military records. The study design specified five occupational categories: (1) officer-pilot, (2) officer-navigator, (3) officer-nonflying, (4) enlisted flyer, and (5) enlisted groundcrew. Based on the review of the records, the Ranch Hand manning for each occupational category by month was compiled. This information is also presented in Table F-l of Appendix F. A numeric exposure index reflecting the effective number of gallons of Herbicide Orange to which each individual was potentially exposed was computed. For the purpose of analysis, the values were categorized as high, medium, or low for each occupational category. Only three occupational categories were used. The three officer categories were combined into one since .pilots and navigators were exposed in the same manner and the officernonflying category, which included a relatively small number of participants, consisted of administrators whose exposure was considered to be essentially zero. The overall group of "nonexposed" Ranch Hands, estimated at approximately 2 percent of the Ranch Hand group, was analyzed in the low exposure category (see Table 8-2), conceivably leading to dilution of the exposure analyses and group contrasts. The exposure index categorizations developed for the Baseline study and used in this report are provided in Table 8-2, along with the frequencies of Ranch Hand participants by occupation and exposure level. The current exposure index is not specific to job and, therefore, may underestimate exposure for those individuals whose jobs required routine handling of herbicide. For example, maintenance schedules for the aircraft herbicide spray tank required that an emergency dump valve be periodically greased, requiring entry into the tank. The current exposure index cannot distinguish between men who received such exposure and men who did not. The extent to which individuals are misclassified by the current exposure index is not known, precluding bias calculations at this time. Because of the acknowledged imprecision of the exposure index, Air Force efforts are under way to develop.new perspectives of exposure. One effort is the construction of a new questionnaire for the 459 enlisted groundcrew personnel that may permit more accurate exposure analyses within this category. Another approach is the measurement of serum dioxin levels. 8-3 �TABLE 8-2. Exposure Index Categorization of 1,016 Compliant Ranch Hands Occupational Group Exposure Category Effective Herbicide Orange Gallons Corresponding to Exposure Category Number of Ranch Hand Participants in Exposure Category Officer Low Medium High <35,000 35,500-70,000 >70,000 127 130 123 Enlisted Flyer Low Medium High <50,000 50,500-85,000 >85,000 55 65 57 Enlisted Groundcrew Low Medium High <20,000 20,500-27,000 >27,000 154 163 142 Total 1,016 The Air Force currently is conducting a pilot study in conjunction with the laboratories of the Centers for Disease Control, Atlanta, Georgia, to determine levels of TCDD in serum and to establish the validity of exposure differential within the Ranch Hand and Comparison groups. This study is in accordance with the Study Protocol commitment to estimate dosage of TCDD as accurately as current technology permits. If successful, use of timeadjusted TCDD levels would permit more accurate exposure analyses within the Ranch Hand group. Perhaps of most importance, accurate TCDD levels within the Ranch Hand group could standardize exposure to a comparable baseline for all participants. Thus, the use of adjusted TCDD levels will place the exposure concepts on a firm scientific basis, and if herbicide effects exist, they can be discerned more accurately. 8-4 �CHAPTER 8 REFERENCES Young, A.L., J.A. Calcagni, C.E. Thalken, and J.W. Tremblay. 1978. The toxicology, environmental fate, and human risk of herbicide orange and its associated dioxin. Technical report OEHL-TR-78-92, USAF Occupational and Environmental Health Laboratory, Brooks AFB, Texas. 247 pp. 8-5 �CHAFFER 9 GENERAL HEALTH INTRODUCTION The effects of heavy, acute exposure to TCDD have been demonstrated in a number of different organ systems. It is plausible, therefore, that chronic low-dose exposure to TCDD might induce subtle, interrelated effects that are not organ-system specific, but are manifest only in general terms, or affect the state of "well-being." However, it is difficult to measure overall health objectively, and for this reason general health outcomes, as defined by this study, should be judged in context with other more specific clinical endpoints. (It should be noted that "general health" outcomes have not traditionally been considered in other dioxin morbidity studies.) Baseline Summary Results Five general health variables were included in the Baseline examination: self-perception of health, appearance of illness or distress, relative age, sedimentation rate, and percent body fat. In the analysis of the 1982 Baseline examination data, a statistically significant difference was found between the Ranch Hand and Comparison groups in self-perception of health, with a greater percentage of Ranch Hands reporting their health as fair or poor than Comparisons. This was true in both the younger and older age groups (p=0.017 and p=0.025 for individuals 40 or less and more than 40 years of age, respectively). The relative risk of the Ranch Hand group was also somewhat greater in the younger subgroup than in the older subgroup (1.8 and 1.4, respectively). Since only 9 of 1,811 individuals were reported by the examining physician as appearing ill or distressed, this designation was apparently reserved for only very ill or distressed individuals. Nevertheless, 8 of the 9 individuals were Ranch Hands, the difference being of borderline significance (p=0.056). Conversely, more Ranch Hands than Comparisons were reported by the examiners as appearing younger than their actual ages (4.9% versus 2.5%, p=0.029). No overall differences in percent body fat or sedimentation rate were found, although a significant interaction between age, group, and sedimentation rate was noted; younger exposed group members had fewer sedimentation rate abnormalities than did their Comparisons, whereas no difference was found in participants more than 40 years old. No statistically significant dose-response relationships were detected in the Ranch Hand group. Parameters of the 1985 General Health Assessment Variables of the Baseline examination (self-perception of health, appearance of illness or distress, relative age, sedimentation rate, and percent body fat) were analyzed for the third year followup effort. 9-1 �As an assessment of the general health status of each individual, three subjective measures were made as well as two more objective measures. During the health interview each study participant was asked, "Compared to other people your age, would you say that your health is excellent, good, fair, or poor?" This self-assessment of health is susceptible to varying degrees of conscious and subconscious bias. The examiner recorded the appearance of illness or distress (yes/no) and noted the appearance of the subject as younger than, older than, or the same as his stated age. To the degree that the examining physicians were kept blind to the study subject's group membership (Ranch Hand, Comparison), their assessments were less subject to bias. The two objective measures were percent body fat, calculated from the body mass index, and the erythrocyte sedimentation rate. Although both variables are rather indirect measures of the general state of health, they are accepted indicators of poor health. The adjusted statistical analyses below accounted for differences associated with age, race, and occupation. In the analysis of self-perception of health and sedimentation rate, adjustment was also made for personality score, determined from the Jenkins Activity Survey. This is a continuous variable derived by means of a discriminant-function equation based on items that best discriminate men judged to be Type A from those judged as Type B. Positive scores reflected the Type A direction and negative scores the Type B direction. Table G-l of Appendix G gives the distribution of the covariates in the Ranch Hand and Comparison groups. Age, race, and occupation were distributed similarly in the two groups (due to matching), and personality scores were also not significantly different. Aside from the subjective nature and potential bias in the self-reported perception of health, no specific issues related to assessment methodology require further comment. No individuals were excluded from analysis, except those with missing data. Chi-square tests and logistic regression models were applied to the categorical data. The sedimentation rate was normalized by logarithmic transformation. The proportional odds model was also used for ordinal data provided by the self-perception of health and relative age variables. Fisher's exact test was applied to the reporting of illness or distress by the examining physician because of the small number of cases who were classified as "ill." A two-sample t-test was used to assess differences in unadjusted group means, followed by multiple regression analysis to incorporate covariates, for percent body fat and sedimentation rate. RESULTS AND DISCUSSION Subjective Assessments Self-Perception of Health Each participant was asked to designate his health as excellent, good, fair, or poor. The frequency distributions of self-perception of health for each cohort are given in Table 9-1. 9-2 �TABLE 9-1, Unadjusted Analysis for Self-Perception of Health by Group Self-Perception of Health Excellent Group Good Fair Poor Number Percent Number Percent Number Percent Number Percent Total Ranch Hand 490 48 .2 434 42.7 74 7.3 18 1.8 1,016 Comparison 674 52 .1 525 40.6 81 6.3 13 1.0 1,293 P=0.14 The summarized data in Table 9-1 show that a higher percentage of Ranch Hands perceived their health to be fair or poor (9.1%) than the Comparisons (7.3%), although this difference was not statistically significant (Est. RR: 1.25, 95% C.I.: [0.95,1.64], p=0.14). Of considerable interest is that the percentage of both groups perceiving their health as only fair or poor was lower than that reported at the Baseline examination 3 years earlier (20.4% and 15.9% for Ranch Hands and Comparisons, respectively). This shift was the opposite of that expected from an aging effect. The data collection technique was an in-home interview in 1982 versus an onsite clinic interview in 1985, but this was not judged to be the likely cause of the improvement in health perceptions for the 3-year period. Whatever the cause, the effects were similar in both groups. A test of association between health perception (dichotomized as excellent/good and fair/poor) was performed with the covariates of age (born in or after 1942, born before 1942), race, occupation, and personality score (Jenkins score, trichotomized as low [less than -5], medium [between -5 and 5], and high [greater than 5]). These associations were examined both within the Ranch Hand and Comparison groups and pooled over the two groups. The findings were similar, and Table 9-2 shows the results after pooling. These results indicated a significant effect of age, with a higher percentage of the older cohort than the younger cohort reporting their health as fair or poor, as well as a significant effect of occupation, with the percentage of enlisted personnel reporting fair or poor health nearly twice that of the officers. No significant associations were noted for race or personality score. 9-3 �TABLE 9-2. Association Between Self-Perception of Health and Age, Race, Occupation, and Personality Score in the Combined Ranch Hand and Comparison Groups Self-Perception of Health Covariate Excellent/Good Fair/Poor Covariate Category Number Percent Number Percent Total p-Value Age Born >1942 903 Born <1942 1,220 94.0 90.5 58 128 6.0 9.5 961 1,348 0.003 Race Black Nonblack 130 1,993 90.9 92.0 13 173 9.1 8.0 143 2,166 0.76 819 94.8 45 5.2 864 Occupation Officer Enlisted Flyer Enlisted Groundcrew 347 89.7 40 10.3 387 957 90.4 101 9.6 1,058 Personality Score 827 716 573 92.2 91.2 92.6 70 69 46 7.8 8.8 7.4 897 785 619 Low Medium High 9-4 <0.001 0.61 �Adjusted analyses of self-perception of health were done by logistic regression using the covariates of age, race, occupation, and personality type. (Self-perception of health was dichotomized and the covariates categorized as in Table 9-2.) These analyses revealed statistically significant age and occupation effects, as well as a significant group-byoccupation interaction (p=0.015). Exponentiation of linear combinations of relevant regression coefficients generated adjusted relative risks for each occupational stratum. These summary data are presented in Table 9-3. TABLE 9-3. Adjusted Relative Risks of Self-Perception of Health by Occupation Adj. Relative Risk (95% C.I.) p-Value Officer 0.78 (0.42,1.46) 0.441 Enlisted Flyer 0.75 (0.38,1.46) 0.395 Enlisted Groundcrew 1.90 (1.25,2.88) 0.003 Occupation These analyses showed significant group differences in the selfperception of health for the enlisted groundcrew category but not for the officers or enlisted flyers. This is perhaps more clearly seen in Table 9-4, which gives the frequency distribution of self-perception of health stratified by occupation. Among officers and enlisted flyers, a lower percentage of Ranch Hands than Comparisons perceived their health as fair or poor. (These same Ranch Hands were also less likely to view their health as excellent.) In the enlisted groundcrew cohort, 12.7 percent of the Ranch Hands reported their health as fair or poor versus 7.2 percent of the Comparisons. Because the logistic model does not account for the ordinal nature of the self-perception of health variable, a proportional odds model for ordinal responses was also fit to the data in Tables 9-1 and 9-4. For the ordinal responses in Table 9-1, the proportional odds model yielded a statistically significant result (p=0.037), with poorer health estimated to be 1.18 times greater in the Ranch Hand group than in the Comparison group (952 C.I.: [1.01,1.39]). For the data in Table 9-4, a proportional,odds model fit to each occupational stratum (adjusting for age) yielded p-values of 0.65 for officers, 0.43 for enlisted flyers, and 0.031 for enlisted groundcrew. Thus, only the enlisted groundcrew category reached statistical significance, with adjusted proportional odds of 1.30 (95% C.I.: [1.02,1.64]). 9-5 �TABLE 9-4. Frequency of Self-Perception of Health by Occupation and Group Self-Perception of Health Excellent Occupation Good Fair Poor Number Percent Number Percent Number Percent Number Percent Total Officer Ranch Hand 238 Comparison 314 62.6 64.9 124 143 32.6 29.6 13 23 3.4 4.8 5 4 1.3 0.8 380 484 Enlisted Flyer Ranch Hand Comparison 67 94 37.8 44.8 94 92 53.1 43.8 13 19 7.3 9.0 3 5 1.7 2.4 177 210 Enlisted Groundcrew Ranch Hand Comparison 185 266 40.3 44.4 216 290 47.1 48.4 48 39 10.5 6.5 10 4 2.2 0.7 459 599 Similar results were obtained when the analyses were performed on the 1,016 Ranch Hands and 955 Original Comparisons completing the third-year health interview. These results are provided in Table G-2 of Appendix G. In the unadjusted analysis, the estimated relative risk for fair or poor health versus excellent or good health reached statistical significance (Est. RR: 1.43, 95% C.I.: [1.03,2.00], p=0.042). In the adjusted analysis, group membership, age, and occupation effects were all statistically significant with an adjusted relative risk of 1.48 (95% C.I.: [1.05,2.07]). The groupby-occupation interaction, however, did not reach statistical significance (p=0.23). Nevertheless, little difference was seen in the officers and enlisted flyers, whereas among the enlisted groundcrew, 12.7 percent of the Ranch Hands versus 7.4 percent of the Original Comparisons reported their health as fair or poor. Contrasts of the Ranch Hand and Original Comparison groups using the proportional odds model yielded only borderline significant results. For the unadjusted analysis applied to the overall data, the estimated proportional odds were 1.17 (95%, C.I.: [0,99,1.39], p«0.073). Stratifying by occupation and adjusting for age gave p-values of 0.76, 0.11, and 0.078 for the officers, enlisted flyers, and enlisted groundcrew, respectively. The adjusted proportional odds in the enlisted groundcrew cohort were 1.26 (95% C.I.: [0.97,1.62]). 9-6 �Appearance of Illness or Distress The recording of the appearance of acute ill health or physical distress at the examination was intended .to capture significant subjective health data that might (though not likely) escape corroboration by other physical examination or laboratory data. In particular, examining physicians were requested to affirm the presence of acute distress when the sign of hippocratic facies was present, a sign not easily feigned by participants. Very few participants were diagnosed as being acutely ill; these data are summarized in Table 9-5. TABLE 9-5. Unadjusted Analysis for Appearance of Acute Illness or Distress by Group Acute Illness or Distress Yes Group No Number Percent Number Percent Total Ranch Hand 4 0.4 1,010 99.6 1,014 Comparison 6 0.5 1,287 99.5 p-Value* 1,293 0.53 *Fisher's exact test, 1-sided. These data were too sparse to permit further meaningful analyses. Descriptively, it was noted that 9 of the 10 ill individuals were in the older age group; 9 of 10 were nonblack; and 2 were officers, 4 were enlisted flyers, and 4 were enlisted groundcrew. The 6 ill Comparison individuals were all Original Comparisons, as can be seen in Table G-3 of Appendix G. Further, these results were in substantial contrast to the Baseline findings that revealed a marginally significant excess (p=0.056) of acute distress among the Ranch Hands. Appearance of Relative Age The examining physicians scored each participant as appearing younger, older, or the same as his chronological age. These data are presented in Table 9-6. 9-7 �TABLE 9-6. Unadjusted Analysis for Appearance of Relative Age by Group Appearance of Relative Age Younger Group Number Percent Ranch Hand 16 1.6 Comparison 9 0.7 Same Number Older Percent Number Percent Total 957 94.3 42 4.1 1,015 1,233 95.4 51 3.9 p-Value 1,293 0.12 These frequency distributions showed that a slightly higher percentage of Ranch Hands than Comparisons appeared younger than their stated age, and almost equivalent percentages in both groups appeared older. Overall, there was no significant difference in the two distributions. The unadjusted findings in Table 9-6, however, did not confirm the significant tendency (p=0.029) at the 1982 Baseline examination for a higher percentage of the Ranch Hands than Comparisons to appear younger than their stated ages. Table 9-7 presents the association between each of the covariates and relative age (dichotomized as older looking versus the same or younger looking) after combining the Ranch Hand and Comparison groups. As noted from this table, age and race were not significantly associated with the appearance of relative age, whereas occupation did reveal a significant association, with about 6 percent of the enlisted personnel appearing older than their stated ages compared to 1 percent of the officers. An adjusted analysis using logistic regression with the covariates age, race, and occupation showed a significant effect due to occupation as well as a significant group-by-occupation interaction (p-0.038). Adjusted relative risks for each occupational stratum are given in Table 9-8. The adjusted relative risk was greater than 1 for the officers, i.e., the odds of appearing older were greater in the Ranch Hand group than in the Comparison group, but the relative risk was less than 1 for the enlisted flyers. However, the associated confidence intervals were rather broad and did not rule out a relative risk of 1 in each case. Again, because the logistic regression model does not account for the ordinal nature of the dependent variable, a proportional odds model was applied to the enlisted flyer cohort (data in the officer and enlisted groundcrew strata did not fit the model properly). The estimated proportional odds for the enlisted flyer cohort were nonsignificant (estimated odds: 0.49, 95% C.I.: [0.22,1.11], p=0.087). 9-8 �TABLE 9-7. Association Between Appearance of Relative Age and Age, Race, and Occupation in the Combined Ranch Hand and Comparison Groups Appearance of Relative Age Younger /Same Older Covariate Category Number Age Born >1942 Born <1942 914 1,301 95.2 96.5 46 47 4.8 3.5 960 1,348 0.14 Race Black Nonblack 138 2,077 96.5 95.9 5 88 3.5 4.1 143 2,165 0.91 Occupation Officer Enlisted Flyer Enlisted Groundcrew 855 362 99.0 93.5 9 25 1.0 6.5 864 387 998 94.4 59 5.6 Covariate Percent Number Percent Total p-Value 1,057 TABLE 9-8. Adjusted Relative Risks of Appearance of Relative Age by Occupation Adj. Relative Risk (95% C.I.) p-Value Officer 4.52 (0.94,21.9) 0.060 Enlisted Flyer 0.44 (0.23,1.27) 0.159 Enlisted Groundcrew 1.05 (0.62,1.78) 0.849 Occupation 9-9 <0.001 �A contrast of the Ranch Hand group with the Original Comparisons gave similar results, as shown in Table G-4 of Appendix G. Overall, there was little difference, but the group-by-occupation interaction was of borderline significance in the adjusted analysis (p=0.052). Differences were largely confined to the enlisted flyers, where fewer Ranch Hands than Comparisons appeared older than their stated ages (Adj. RR: 0.47, 95% C.I.: [0.20,1.12], p=0.089) (see Table G-5 of Appendix G). A proportional odds model applied to the enlisted flyer stratum gave adjusted proportional odds of 0.45 (95% C.I.: [0.20,1.02], p-0.055). Objective Assessments Two objective but nonspecific indicators of general health, the erythrocyte sedimentation rate and percent body fat, were analyzed in both discrete and continuous forms. Because the sedimentation rate was a highly skewed variable, it was normalized by logarithmic transformation for the continuous analyses. The sedimentation rate dichotomy was set at 20 mm/hr or less (normal) and greater than 20 mm/hr (abnormal) by the large-tube Westergren method. Percent body fat was based on height and weight obtained during the examination and was calculated according to the following formula: Percent Body Fat = (Weight[kg]/Height[m])(1.264) - 13.305. It is recognized that this formula will overstate the percent body fat for very muscular, large-boned men. Percent body fat was trichotomized into less than 10 percent (lean), 10 to 25 percent (normal), and greater than 25 percent (obese), consistent with the Baseline Report. Because of the sparseness of the lean category, it was often necessary to use a dichotomous variable of lean-normal versus obese. Erythrocyte Sedimentation Rate The unadjusted contrast of log sedimentation rate means revealed no significant group differences (mean±SE=l.620+0.026 in the Ranch Hand group versus 1.595±0.021 in the Comparison group, t=0.73, p=0.47). The geometric mean values were 5.05 and 4.93 for the Ranch Hand and Comparison groups, respectively. Tests of association of dichotomized sedimentation rate, with the covariates age, race, occupation, and personality score, pooled over both groups, were conducted; these summarized data are shown in Table 9-9. These results showed significant effects of age, with older individuals having a higher frequency of abnormal sedimentation rates than younger individuals, and a significant effect of personality score, with Type B individuals (low personality score) having more sedimentation rate abnormalities. The effect of occupation was of borderline significance (p=0.080), with a slightly higher percentage of abnormal values among the enlisted flyers than among officers or enlisted groundcrew. There was no evidence of any association between race and abnormal sedimentation rate. An analysis of the log sedimentation rate, adjusting for age, race, occupation, and personality score, detected significant effects for all of the covariates except race, as well as a significant age-by-personality score interaction. As in the unadjusted analysis, the adjusted analysis did not reveal any significant difference between the Ranch Hand and Comparison groups (p=0.412). 9-10 �TABLE 9-9. Association Between Sedimentation Rate and Age, Race, Occupation, and Personality Score in the Combined Ranch Hand and Comparison Groups Sedimentation Rate Abnormal >20mm/hr Normal <20mm/hr Covariate Number Percent Covariate Category Number Percent Total p-Value Age 941 Born >1942 Born <1942 1,263 97.9 93.7 20 85 2.1 6.3 961 1,348 <0.001 Race Black Nonblack 136 2,068 95.1 95.5 7 98 4.9 4.5 143 2,166 0.999 Occupation Officer 828 Enlisted 361 Flyer 1,015 Enlisted Groundcrew 95.8 93.3 36 26 4.2 6.7 864 387 0.080 95.9 43 4.1 1,058 94.0 96.6 96.1 54 27 24 6.0 3.4 3.9 897 785 619 Personality Score Low Medium High 843 758 595 9-11 0.026 �However, in the dichotoraous form, sedimentation rate abnormalities were significantly more prevalent in the Ranch Hands than Comparisons (Est. RR: 1.63, 95% C.I.i [1.12,2.38], p=0.013); these results are given in Table 9-10. Logistic regression analysis found significant effects for age and personality score, and the adjusted relative risk of 1.68 (95% C.I.: [1.13,2.49], p=0.011), was very similar to the estimated relative risk of 1.63. TABLE 9-10. Unadjusted Analysis for Sedimentation Rate by Group Sedimentation Rate Normal <20 mm/hr Group Number Percent Abnormal >20 mm/hr Number Percent Total Ranch Hand 957 94.2 59 5.8 1,016 Comparison 1,247 96.4 46 3.6 p-Value 1,293 0.013 The mean log sedimentation rate in the Original Comparisons was 1.636 plus or minus 0.025, not significantly different from the Ranch Hand mean (t=-0.45, p=0.65). The regression analysis yielded results very similar to those reported above, with little difference in the adjusted group means. Logistic regression analyses also gave similar results, with significantly more abnormalities in the Ranch Hand group (p=0.037). In summary, there was no difference between groups based upon mean values of the sedimentation rate, unadjusted or adjusted, but both unadjusted and adjusted discrete analyses snowed a significantly higher prevalence of sedimentation rate abnormalities in the Ranch Hand group. This finding was opposite to the Baseline findings in which Ranch Hands age 40 or less had significantly fewer sedimentation rate abnormalities than Comparisons, with no group difference in individuals over the age of 40. Percent Body Fat The mean percent body fat of Ranch Hands was significantly lower than that of Comparisons (21.10%±0.15.versus 21.54%±0.14, respectively; p=0.037). Because there were only a few values in the lean category (6 in the Ranch Hand group and 4 in the Comparison group), percent body fat was dichotomized into at most 25 percent (lean and normal) and more than 25 percent (obese) for tests of association between percent body fat and the covariates age, race, and occupation. The results are given in Table 9-11. 9-12 �TABLE 9-11. Association Between Percent Body Fat and Age, Race, and Occupation in the Combined Ranch Hand and Comparison Groups Percent Body Fat Lean/Normal <25% Obese >25% Covariate Covariate Category Age Born XL942 802 Born <1942 1,060 83.4 78.7 159 287 Race Black Nonblack 110 1,752 76.9 80.9 719 314 829 Occupation Officer Enlisted Flyer Enlisted Groundcrew Number Total p-Value 16.6 21.3 961 1,347 0.005 33 413 23.1 19.1 143 2,165 0.29 83.3 81.1 144 73 16.7 18.9 863 387 78.4 229 21.6 Percent Number Percent 0.023 1,058 These data demonstrated the significant effects of age, with a higher percentage of obesity in older men, and occupation, with a higher prevalence of obesity in enlisted personnel than in officers. Race was a noncontributory covariate. The covariate of smoking was unexplored. An adjusted analysis of percent body fat, with the same covariates, also showed the significant effects of age, occupation, and an age-by-occupation interaction. The adjusted results showed a small, but significantly lower mean level of body fat in the Ranch Hand group (adjusted difference=-0.443± 0.210, p=0.035). With percent body fat dichotomized into obese versus normal or lean, the percent obese was lower in the Ranch Hands than in the Comparisons (18.2% versus 20.2%), but the difference was not significant (Est. RR: 0.90, 95% C.I.: [0.71,1.08], p=0.25). Logistic regression analysis also failed to detect a significant group difference (Adj. RR: 0.87, 95% C.I.: [0.71,1.08], p-0.204). Analysis of percent body fat in the Ranch Hands and Original Comparisons gave somewhat different results. The overall difference in means was significant as before: 21.10 plus or minus 0.15 in the Ranch Hand group versus 21.58 plus or minus 0.16 in the Original Comparison group (t=-2.15, p=0.032). However, the regression analysis detected a statistically significant groupby-race interaction (p=0.041). The adjusted difference in mean percent "body fat (Ranch Hand versus Comparison) was greater in Black participants (-2.26%) 9-13 �than in nonblack participants (-0.34%). Of the Original Comparisons (Table G-7 of Appendix G), 20.4 percent were obese, greater than, but not significantly different from, the percent obese in the Ranch Hand group (p=0.230). Logistic regression analyses again detected significant age and occupation effects, but it detected no significant interaction between these variables. There was no strong evidence of a group-by-race interaction (models including all two-factor interactions gave a Z-value of 1.19 for the group-by-race interaction). The group effect was not statistically significant (Adj. RRs 0.87, 95% C.I.: [0.70,1.09], p=0.242). In summary, the unadjusted and adjusted tests of mean percent body fat showed a significantly lower value for Ranch Hands; correspondingly fewer Ranch Hands than Comparisons were obese, although this difference was not statistically significant. Few individuals were lean (less than 10 percent body fat). The 1982 Baseline examination found no difference in group means (p=0.67), or proportion of abnormalities (p=0.89). Further, analyses based solely upon the Original Comparison cohort found the difference in mean percent body fat between the Ranch Hand and Comparison groups to be greater in Blacks than nonblacks. EXPOSURE INDEX ANALYSES The exposure index, expressed in equivalent gallons of dioxin-containing herbicide potentially encountered by each Ranch Hand during his tour of duty in Vietnam, was categorized as low, medium, and high. Because it is not possible to assess the relative exposure between occupational groups, and since different cutoff values were used in the three occupational categories, separate analyses were performed within each occupational cohort. A detailed description of the exposure index is found in Chapter 8. Exposure analyses were performed on four of the five general health variables. Only four Ranch Hands were recorded as appearing ill or distressed (two were officers, both in the low-exposure category, and two were enlisted flyers, both in the high-exposure category). Further analysis was not done on this variable. Self-Perception of Health Table 9-12 presents dichotomized self-perception of health data by exposure level for the 1,016 Ranch Hands. While these unadjusted contrasts did not reach statistical significance within any of the occupational strata, the linear trend from low to high exposure in the officer cohort of the fair/poor category was of interest, and was subjected to further testing. Although the numbers were small at each exposure level, a test for linear trend led to a borderline significant increase of 2.5 plus or minus 1.3 percent per unit (step) increase in the exposure level category (p=0.064). Logistic regression analyses adjusted for age (dichotomized), race, and personality score (trichotomized) did not detect any significant exposure level effects. The only significant covariate effect found was for age in the enlisted groundcrew cohort. The adjusted relative risk for each occupational stratum is given in Table 9-13. 9-14 �TABLE 9-12. Unadjusted Exposure Index Analysis of Self-Perception of Health by Occupation Self-Perception of Health Excellent/Good Exposure Index Occupation Officer Number Percent Fair/Poor Number Percent Total p-Value* Low Medium High 124 124 114 97.6 95.4 92.7 3 6 9 2.4 4.6 7.3 127 130 123 0.17 Enlisted Flyer Low Medium High 51 59 51 92.7 90.8 89.5 4 6 6 7.3 9.2 10.5 55 65 57 0.83 Enlisted Groundcrew Low Medium High 134 146 121 87.0 89.6 85.2 20 17 21 13.0 10.4 14.8 154 163 142 0.51 *Chi-square tests, 2 d.f, TABLE 9-13. Adjusted Relative Risk of Self-Perception of Health by Occupation and Exposure Contrast Exposure Contrast Adj. Relative Risk (95% C.I.) p-Value Officer Medium vs. Low High vs. Low 2.00 (0.49,8.15) 2.93 (0.76,11.3) 0.334 0.119 Enlisted Flyer Medium vs. Low High vs. Low 1.30 (0.35,4.86) 1.50 (0.40,5.64) 0.700 0.549 Medium vs. Low High vs. Low 0.95 (0.47,1.92) 1.21 (0.62,2.35) 0.882 0.580 Occupation Enlisted Groundcrew 9-15 �Appearance of Relative Age The dichotomy of appearance of relative age was assessed for exposure effects in each occupational cohort. These unadjusted analyses, shown in Table 9-14, provided no evidence of a dose-response effect. As can be seen, the number of participants within each stratum appearing older than their stated ages was quite small. The adjusted analyses by logistic regression did not detect any significant exposure or covariate effects. TABLE 9-14. Unadjusted Exposure Index Analysis of Appearance of Relative Age by Occupation Relative Age Younger/Same Occupation Exposure Index Number Percent Older Number Percent Total p-Value* Officer Low Medium High 125 127 121 98.4 97.7 98.4 2 3 2 1.6 2.3 1.6 127 130 0.89 123 Enlisted Flyer Low Medium High 52 62 55 94.6 95.4 96.5 3 3 2 5.4 4.6 3.5 55 65 0.88 57 Enlisted Groundcrew Low Medium High 146 151 134 94.8 93.2 94.4 8 11 8 5.2 6.8 5.6 154 162 0.82 142 *Chi-square tests, 2 d.f. 9-16 �Erythrocyte Sedimentation Rate The sedimentation rate was analyzed both continuously on a logarithmic scale and dichotomously (normal, abnormal). One-way analyses of variance were performed on the sedimentation rate means categorized by occupation and exposure level. These tests showed no significant differences in the officer and the enlisted flyer strata (p=0.76, p=0.64, respectively). In the enlisted groundcrew stratum the means were marginally different, with the mean sedimentation rate increasing with increasing exposure level, but the differences were not statistically significant (p=0.12). When these data were adjusted by an analysis of covariance for age, the difference in mean sedimentation rates in the enlisted groundcrew was less noteworthy (p=0.33). Age was positively associated with the mean sedimentation rate in all three occupational strata (p<0.001, p=0.009, and p<0.001, respectively). The adjusted tests are reflected in Table 9-15 (means and confidence limits have been transformed back to the original scale). A categorical analysis of the sedimentation rate by exposure level for each occupational stratum was also conducted. Differing from the previous continuous analyses, the categorical contrasts revealed a significant exposure effect (p=0.027) in the enlisted flyer stratum, albeit with small numbers. These summarized data are shown in Table 9-16. Adjustment for age, race, and personality score revealed a significant high versus low exposure contrast in the enlisted flyer stratum. The adjusted analysis is fully shown in Table 9-17. TABLE 9-15. Adjusted Mean Sedimentation Rates by Occupation Exposure Index, Adjusted Mean, mm/hr (95% C.I.) Occupation Low Medium High p-Value Officer 5.40 (4.71 ,6.19) 4.78 (4.17 ,5.47) 4.69 (4.09,5.37) 0.31 Enlisted Flyer 5.10 (4.11 ,6.33) 6.00 (4.91 ,7.32) 5.00 (4.04,6. 19) 0.41 Enlisted Groundcrew 4.66 (4.10,5.29) 5.09 (4.49,5.77) 5.35 (4.69,6.12) 0.33 9-17 �TABLE 9-16. Unadjusted Exposure Index Analysis of Sedimentation Rate by Occupation Sedimentation Rate Abnormal >20mm/hr Normal <20mm/hr Exposure Index Number Percent Number Percent Total Low Medium High 117 125 119 92.1 96.2 95.9 10 5 5 7.9 3.8 4.1 127 130 123 Enlisted Flyer Low Medium High 53 62 48 96.4 95.4 84.2 2 3 9 3.6 4.6 15.8 55 65 57 0.027 Enlisted Groundcrew Low Medium High 142 156 136 92.2 95.7 95.8 12 7 6 7.8 4.3 4.2 154 163 142 0.290 Occupation Officer *Chi-square tests, 2 d.f. TABLE 9-17. Adjusted Relative Risk of Sedimentation Rate by Occupation and Exposure Contrast Occupation Exposure Contrast Adj. Relative Risk (95% C.I.) p-Value Officer Medium vs. Low High vs. Low 0.47 (0.16,1.41) 0.50 (0.17,1.52) 0.177 0.226 Enlisted Flyer Medium vs. Low . High vs. Low 1.28 (0.21,7.96) 4.97 (1.02,24.2) 0.790 0.047 Enlisted Groundcrew Medium vs. Low High vs. Low 0.76 (0.28,2.06) 0.54 (0.19,1.49) 0.592 0.234 9-18 p-Value* 0.27 �Percent Body Fat Exposure analyses of percent body fat were done using both linear models and logistic regression. One-way analyses of variance for means found no statistically significant exposure differences in the occupational cohorts. These statistics are presented in Table 9-18. TABLE 9-18. Unadjusted Means of Percent Body Fat by Occupation Exposure Index, Mean±SE Occupation Low Medium High p-Value Officer 20.99±0.36 21.11±0.41 21.26±0.36 0.88 Enlisted Flyer 20.65±0.55 21.26±0.77 21.59±0.77 0.65 Enlisted Groundcrew 20.91±0.42 21.43±0.41 20.79±0.44 0.53 Linear models including age, race, and two-factor exposure level-bycovariate interactions found no significant difference in the adjusted exposure level means for percent body fat. The effect of age was significant in the officer cohort (p=0.003), and of borderline significance in the enlisted groundcrew stratum (p=0.064). Race was nonsignificant throughout all the tests. The unadjusted categorical assessment of percent body fat, shown in Table 9-19, revealed no significant exposure effects. However, in the enlisted flyer stratum, a test for linear trend in the proportions gave a borderline significant result (p=0.054), with an estimated step increase of 6.8 plus or minus 3.6 percent per unit increase in exposure-level category. An adjusted analysis by logistic regression did not reveal significant exposure level effects but did detect significant effects of age in the officer and enlisted groundcrew categories. In summary, detailed exposure analyses were performed on four of five dependent variables used to assess general health status. Only a very few of the tests approached statistical significance (multiple comparisons notwithstanding); of these, three associations suggested a trend of adverse effects from low to high exposure; but only one was statistically significant, and there was no consistency across occupational strata (health perception in officers, p=0.064; sedimentation rate in enlisted flyers, p=0.027; and percent body fat in enlisted flyers, p=0.054). These results were relatively comparable to the negative exposure findings in the Baseline Report. 9-19 �TABLE 9-19. Unadjusted Exposure Index Analysis of Percent Body Fat by Occupation Percent Body Fat Obese <25% Lean/Normal <25% Occupation Officer Enlisted Flyer Enlisted Groundcrew Exposure Level Number Percent Number Percent Total p-Value* 104 110 100 81.9 84.6 81.3 23 20 23 18.1 15.4 18.7 127 130 123 0.76 Medium High 50 53 44 90.9 81.5 77.2 5 12 13 9.1 18.5 22.8 55 65 57 0.14 Low Medium High 126 131 113 81.8 80.4 79.6 28 32 29 18.2 19.6 20.4 154 163 142 0.88 Low Medium High L0¥ *Chi-square tests, 2 d.f. LONGITUDINAL ANALYSES Two variables, self-perception of health and sedimentation rate, were prescribed to assess the longitudinal differences between the 1982 Baseline examination and the 1985 followup examination. Both variables were analyzed in the discrete form. The four categories of perception of health were reduced to normal (excellent/good) and abnormal (fair/poor). The respective laboratory norms of 12 or less mm/hr and more than 12 mm/hr for the Baseline sedimentation rates, and 20 or less mm/hr and more than 20 mm/hr for the followup examination were used to categorize the sedimentation rate data into normal and abnormal groups. The off-diagonal data (normal to abnormal, abnormal to normal) from the two examinations were contrasted by group membership, a process equivalent to testing for a group-by-time-by-clinical endpoint interaction. The results of these tests, unadjusted for covariates, are given in Table 9-20. These analyses showed an equivalence of the change in self-perception of health in the two groups between examinations, but a highly significant group difference in the change in sedimentation rate abnormalities. The latter was explained by the fact that the Baseline examination determined a significant excess of sedimentation rate abnormalities in the Comparisons whereas at the followup examination, the Ranch Hands had a significantly higher proportion of abnormalities. Perhaps as a related fact, it is recognized that the sedimentation rate laboratory test procedure changed to a more sensitive one at the followup examination. 9-20 �TABLE 9-20. Longitudinal Analysis of Self-Perception of Health and Sedimentation Rate: A Contrast of Baseline and First Followup Examination Abnormalities Followup Examination Variable Group Ranch Hand Odds p-Value Baseline Examination Abnormal Normal Ratio (OR*) (ORRH vs. ORC) 0.21 62 27 127 750 Abnormal Normal 49 28 124 936 0.23 Ranch Hand Abnormal Normal 17 39 16 899 2.44 Comparison Sedimentation Rate Abnormal Normal Comparison SelfPerception of Health Abnormal Normal 14 27 37 1,061 0.73 0.84 0.002 Number Normal Baseline, Abnormal Followup *0dds Ratio: Number Abnormal Baseline, Normal Followup. SUMMARY AND CONCLUSIONS General physical health was evaluated by five measures, three of which were subjective (self-perception of health, appearance of distress, and appearance of relative age), and two of which were objective (percent body fat and sedimentation rate). Table 9-21 presents a summary of all the unadjusted and adjusted analyses of these five variables. The Ranch Hands rated their health as fair or poor more often than the Comparisons (9.1% versus 7.3%, respectively), but this difference was not significant by categorical testing. However, further analysis revealed a significant group-by-occupation interaction; differences were largely confined to the enlisted groundcrew category. Both the Ranch Hand and Comparison groups noticeably improved their perceptions of health from the 1982 Baseline examination. Only 10 individuals were reported as appearing acutely ill or distressed at the followup examination, 4 were Ranch Hands and 6 were Comparisons. This difference was not statistically significant and the data were insufficient for adjusted analyses. 9-21 �TABLE 9-21. Overall Summary Results of Unadjusted and Adjusted Analyses of General Health Variables Unadjusted Variable Categorical Self-Perception of Health NS Appearance of Illness/Disstress Mean Categorical Mean NS Appearance of Relative Age Adjusted — **** NS — **** Sedimentation Rate 0.013 NS 0.011 NS Percent Body Fat NS 0.037 NS 0.035 —Analysis not performed. ****Group-by-covariate interaction. "Analysis not possible due to sparse data. Appearance of relative age, as determined by the examining physician, showed 1.6 percent of the Ranch Hands appearing younger than their stated age, 94.3 percent appearing the same, and 4.1 percent appearing older (as contrasted to 0.7%, 95.4%, and 3.9%, respectively, in the Comparison group). There was a significant group-by-occupation interaction, but none of the estimated relative risks for the occupational categories was significant. This observation at the followup examination contrasted with the significant tendency at the Baseline for a higher percentage of Ranch Hands than Comparisons to appear younger than their stated ages. The geometric mean sedimentation rates (5.05 mm/hr Ranch Hand versus 4.93 mm/hr Comparison) did not differ significantly by group, either unadjusted or after adjustment for age, race, occupation, personality score, and an age-by-personality score interaction. However, in the dichotomous form, 5.8 percent of the Ranch Hands had sedimentation rate abnormalities as contrasted to 3.6 percent in the Comparison group. This difference was significant by both unadjusted and adjusted tests. Also, this finding was opposite to that of the Baseline examination, where it was noted that younger Comparisons had significantly elevated sedimentation rates. 9-22 �The mean percent body fat of the Ranch Hands was significantly lower than the Comparisons (21.10%±0.15, 21.54%±0.14, respectively, p=0.037), and was of nearly the same magnitude after adjustment for age, race, and occupation. However, both unadjusted and adjusted categorical tests did not reveal significant group differences, although the percent obese was lower in the Ranch Hands than in the Comparisons. No group differences in percent body fat were noted at the Baseline examination. Detailed exposure analyses were done on four general health variables (appearance of acute distress was too sparse for testing). Only one analysis demonstrated statistical significance, i.e., a positive association of sedimentation rate abnormalities with increasing exposure in the enlisted flyer cohort. Overall, no consistent pattern of exposure effects was discernible, and the exposure findings at the third-year followup were similar to the findings at Baseline. Longitudinal differences between the 1982 Baseline and the 1985 followup examination were assessed by analyses of two discrete variables, selfperception of health and sedimentation rate. Perceived health showed no significant group differences over time, but both the Ranch Hand and Comparison groups paradoxically reported symmetrical improvements in their perceptions over the 3-year period. The sedimentation rate analysis revealed a highly significant group difference (p=0.002), due to a reversal of findings between examinations, i.e., a significant detriment in the younger Comparisons at the Baseline versus a significant detriment in the Ranch Hands at the followup. The cause(s) and biological relevance of this observation are unclear. In conclusion, a nonspecific assessment of general physical health has shown relatively close similarity between the Ranch Hand and Comparison groups, with the Ranch Hands continuing to perceive their health more negatively than the Comparisons, having a slightly more favorable percent body-fat proportion, but a higher proportion of abnormal sedimentation rates that reflects a marked change since the Baseline examination. These findings must be placed in context with the organ and system-specific evaluations found in the succeeding chapters. 9-23 �CHAPTER 9 REFERENCES 1. Jenkins, C.D., R.H. Rosenman, and S.J. Zyzanski. 1974. Prediction of clinical coronary heart disease by a test for the coronary-prone behavior pattern. New Eng. J. Med. 290(23):1271-1275. 9-24 �CHAPTER 10 MALIGNANCY INTRODUCTION Cancer is a major suspect disease following exposure to chlorophenols, phenoxy herbicides, and dioxin. Both systemic cancer and skin cancer are key focal points of this study. The issue of military service related cancer in Vietnam veterans first arose in 1978-1979. Media presentations emphasized several early cancer deaths in several Army veterans, which were allegedly caused by exposure to Agent Orange. The media reinforced the causal allegations by citing animal studies, which demonstrated a carcinogenic effect, and a few human studies, which showed excessive cancer in specific occupational groups. So effective and sustained were the media presentations that today the public equates dioxin and Agent Orange exposure to cancer. In the larger context of environmental controversies, Young aptly described the Agent Orange issue as being at the crossroads of science and social concern. The scientific community has responded to the dioxin question by a massive research effort, which in concert with class action lawsuits, is expected to cost more than a billion dollars in the near future. The core of the overall research effort is basic and applied cancer research. Traditional animal-to-man extrapolation difficulties and interspecies variability have limited the direct applicability of much of the experimental work to the controversy. Major epidemiologic challenges have included: the ability to control/characterize bias; selection of suitable controls or reference groups; quality/quantity of exposure; misclassification of exposure; confounding exposure to known injurious chemicals; sample size and statistical power; number and selection of relevant risk factors; lack of antecedent disease or syndromes (other than chloracne); time to event (latency); rarity of the endpoint; and tumor type (carcinoma, sarcoma) differences found in many studies. For these reasons, there is no scientific consensus on the dioxin-cancer question. There is, however, a common thread, raising concern over soft tissue sarcomas (STS) and non-Hodgkin's lymphoma (NHL). Pertinent animal and human studies underscore the concern.over cancer. Numerous animal studies have been conducted to delineate the role of TCDD on tumor initiation, tumor promotion, mutagenesis, cocarcinogenesis, and DNA reactivity. The consensus of most research is that TCDD is only weakly mutagenic, does not covalently bind to DNA or cause it to initiate repair synthe|is, and behaves as a strong tumor promoter in already initiated cells. 10-1 �The oncogenic response to TCDD in animals has been repeatedly shown to depend upon animal species and strain, dose, age, sex, and route of administration. Conventional skin bioassays in mice produced mixed results in some studies ' but caused significant dermal fibrosarcomas in other studies using different strains of animals. In the presence of a strong carcinogen, TCDD induced skin papillomas in homozygous hairless mice (but not in the heterozygous strain), clearly supporting the promoter role of TCDD, a nongenetic mechanism judged to be related to receptor binding. Ingestion studies in several rat strains at doses of 0.07-0.1 yg/kg/day produced hepatocellular carcinomas, squamous cell carcinomas of the oropharynx and lung, and follicular cell thyroid adenomas. ' In two mouse strains, gavage doses of 0.07-0.3 yg/kg/day produced hepatocellular carcinomas and thyroid tumors. In the presence of partial hepatectomy and diethylnitrosamine, subcutaneous TCDD administration to rats resulted in hepatocellular carcinomas, demonstrating the promoter mechanism of TCDD. Based upon these and other studies, the International Agency for Research on Cancer (IARC) designated TCDD as carcinogenic in 1982. There are insufficient data to implicate 2,4-D and 2,4,5-T as carcinogens. The majority of animal studies have shown carcinomas rather than sarcomas, the tumor cited in some human studies. If TCDD oncogenicity in humans is to be supported, the differences in tumor types between animals and man requires explanation. In a series of publications beginning in 1974, commonly known as the "Swedish studies," extensive inquiry was made into occupational cancer following exposure to a variety of herbicides. Four related efforts using Swedish railroad workers found an increased cancer incidence mostly associated with non-TCDD herbicides. However, a case-control analysis of these data by other investigators suggested cancer promotion following phenoxy acid exposure. Prompted by a slight increase in STS in the railroad workers and clinical experience with a case series of STS, Hardell and coworkers launched an extensive second round of studies. ~ These efforts showed statistically significant increased risks for STS, Hodgkin's Disease (HD), and NHL. For exposure to phenoxy acids alone, the risk ratio ranged from 5.3 to 6.8 for STS in northern and southern Sweden, respectively, while a range of 3.3 to 6.6 was noted for exposure to chlorophenol alone. For malignant lymphoma (HD plus NHL), risk ratios of 8.4 and 4.8 were respectively demonstrated for chlorophenol and phenoxy acid exposures. An association of nasal and nasopharyngeal cancer to chlorophenol exposure (risk ratio, 6.7) was also detected, but other specifically focused studies of primary liver cancer and colon cancer were negative with respect to phenoxy acid or chlorophenol exposure. '2 The colon cancer study was conducted specifically to demonstrate a lack of respondent bias to "validate" previous questionnaire and interview methods used in the STS studies. From the outset, the Swedish studies have been criticized on methodologic issues, ~ prompting the primary authors, Axelson and Hardell, to respond with clarifications, new calculations, amplifying studies on additional cohorts, and studies on other cancers. 2'2 '2 ~ 1 The chief criticisms centered upon possible respondent and observational biases, 10-2 �selection of controls, confounding exposures, and degree of true exposure to phenoxy acids and chlorophenols. The authors answered these criticisms within the inherent constraints of the case-control methodology. Their efforts have been characterized as careful, clever, and properly stated, and have received favorable reviews. ' 3 Four small industrial mortality studies were conducted in the late 1970's and early 1980's. NIOSH investigators pooled the data from these studies and noted that three of the 105 deaths (2.9%) in these studies were due to STS as contrasted to an expected 0.07 percent in the U.S. general population. This study has been criticized for the hasty addition of possibly noncomparable industrial cohorts, and the lack of histologic confirmation of the STS cases. A subsequent case report added another STS case to the industrial studies, and two other reports revealed three unrelated STS cases also arising from the industrial sector. ' However, upon closer inspection, only two of the first four cases were confirmed as STS by an independent histologic review. Other review findings of the seven total cases were noteworthy: there was poor agreement on the histologic subtype of the soft tissue tumors, and because of a quirk in the International Classification of Diseases (ICD) System, wherein organ-specific sarcomas are coded separately from soft and connective tissue tumors (ICD 171), deathcertificaje based studies will underascertain STS by approximately 40 percent. ' This latter problem did not affect the Swedish studies. Other cancer studies throughout the world showed mixed support for the Swedish findings. An Italian case-control effort showed a weak association between ovarian mesothelial tumors and herbicide exposure, whereas a Finnish study of a small number of pesticide sprayers understandably did not detect any cases of STS or malignant lymphomas (ML).4 A study of more than 4,000 Danish phenoxy herbicide workers noted five STS cases (versus 1.8 expected) and seven ML cases (versus 5.4 expected). The author concluded that the STS observation supported the Swedish work and that the ML rate did not. One New Zealand case-control study showed a nonsignificant relative risk of 1.3 for STS among occupations consistent with phenoxy herbicide exposure, although a risk of 7.2 was noted for STS and potential chlorophenol exposure in tanneries. A related second cancer registry-based case-control study revealed significant excesses of agricultural and forestry occupations from ML cases and multiple myeloma cases (odds ratio 1.25). 8 In a similar but larger cancer registry study in.Sweden, there was no increased risk of STS (relative risk 0.9) in agricultural or forestry workers as contrasted to other industrial workers. Further, the STS risk was constant over time in spite of increased usage of phenoxy acid herbicides from 1947 to 1970. This Swedish study did not confirm or show a trend consistent with the earlier Hardell Swedish studies. A recent U.S. case-control study from the Kansas cancer registry has provided partial support for Hardell's observations. The Kansas study was very similar in methodology to the early Swedish studies and tried to avoid bias and misclassification. An overall risk of 1.6 was found for NHL in men exposed to herbicides, particularly 2,4-D. As the frequency of herbicide exposure increased to more than 20 days per year, the risk of NHL increased to sixfold vis-a-vis nonfarmers. For herbicide applicators, the risk for NHL 10-3 �was 8.0. A simultaneously published review of the Kansas work noted that this should shift scientific concern from STS to NHL. A population-based case-control study of STS and NHL in western Washington found no overall increased risk of these diseases associated with an occupational history of exposure to chlorophenols or phenoxy herbicides. However, risks of NHL were significantly elevated in the specific occupational categories of farmers, forestry herbicide applicators, and those individuals potentially exposed to phenoxy herbicides in any occupation for 15 years or more. An increased risk of NHL was also noted among those with occupational exposure to insecticides, organic solvents, lead, and welding fumes. A number of Vietnam veteran studies has attempted to determine whether veterans have experienced excessive mortality, particularly from cancer. Most of the studies used proportionate mortality ratio (PMR) methodology and equated Vietnam service with potential exposure to Agent Orange, a procedure of considerable imprecision (misclassification). These exposure allocation difficulties, coupled with the inherent methodological weaknesses of the PMR technique, have minimized the contribution of these studies to the overall cancer issue. As might be predicted by these problems, almost all of the veteran studies were negative for generic cancer associations, as well as for STS, HD, and NHL associations. As an example of the veteran studies, the Australian retrospective cohort mortality effort revealed an overall relative mortality ratio of 0.99, an overall cancer mortality ratio of 0.95, and nonsignificant statistical differences for STS, NHL, and HD. In a recent Vietnam experience study of STS using the case-control method, no significant association was found between military service in Vietnam and the subsequent occurrence of STS. No consistent pattern for other cancer types has emerged from the entire body of herbicide literature. None of the leukemias has been associated with exposure to Herbicide Orange nor any of its constituents. Two studies noted slight increases in gastric cancer ' and two others cited modest risks for lung cancer. ' A recent Swedish study reported slight excesses of rectal cancer in male workers and increased cervical cancer from the exposed female cohort. Overall, these and other observations have not been consistent with the expectation that dioxin, as a cancer promoter, should increase the occurrence of common "background" cancers. From another perspective, if clear-cut exposure to 2,4-D or dioxin is shown to cause an immunological deficiency (see Chapter 19), an expectation would be an excessive representation of B-cell tumors from the population of NHL cases. " An excess of B-cell neoplasms has, in fact, not been described in NHL cases from industrial or veteran cohorts to date. It is unlikely that the cancer question will be clearly resolved in the near future. Dioxin exposure in industry and agriculture has fallen precipitously since the 1970's, while exposures to 2,4-D and non-TCDD containing herbicides have continued. Veteran studies characterized by low or undocumented exposure to Agent Orange, and/or of small cohort size are unlikely to contribute substantive data for the evaluation of type-specific cancers although they may contribute to the resolution of the generic cancer issue. 10-4 �In summary, Swedish studies first noted an approximate sixfold risk of soft tissue sarcoma and malignant lymphoma in forestry workers exposed to both phenoxy acid herbicides (not containing the dioxin contaminant) and chlorophenols (containing dioxin). A large number of international studies were predominantly nonsupportive of the Swedish observations. Recent U.S. research on agricultural workers, however, provided some support for a nonHodgkin's lymphoma-phenoxy acid exposure association. The future scientific focus is expected to shift from dioxin herbicides to nondioxin herbicides and from soft tissue sarcomas to malignant lymphomas. Studies of other veteran populations will not likely contribute to the new emphasis, largely because of exposure uncertainties. Baseline Summary Results Cancer received major emphasis during the 1982 AFHS. The assessment of malignancy used data from both the in-home questionnaire and the review-ofsystems questionnaire obtained during the physical examination as well as data from the examination itself. All subjective data were verified by medical record reviews. In addition, tabulation of mortality count data from the Baseline Mortality Report was used in conjunction with cancer morbidity information. The overall results showed an equivalence of systemic cancer (p=0.46) in the two groups but significantly more nonmelanotic skin cancer (p=0.03) in the Ranch Hands. Of 50 reported systemic cancers from the Ranch Hand and Comparison groups, 28 (14 in each group) were verified by medical records and pathology reports. A visual inspection of anatomic sites showed a slight excess of genitourinary cancer and oropharyngeal cancer but a relative deficit of digestive system neoplasms in the Ranch Hands. A combined morbidity-mortality analysis derived from the initial 1:1 match (Ranch Hand to the C-l Comparison member) disclosed similar distributions. One case of soft tissue sarcoma and one case of Hodgkin's Disease were confirmed, both in the Comparison group. Exposure analyses for industrial chemicals and x rays were negative as were most of the herbicide exposure analyses in the Ranch Hand group. All of the exposure analyses were based upon very small numbers, and interactions were noted in several strata. Questionnaire data verified by medical record reviews revealed significantly more skin cancer in the Ranch Hands (relative odds 2.35). Basal cell carcinoma accounted for 83.9 percent of the reported skin cancers in both groups and was concentrated anatomically on the face, head, and neck. The few melanoma and squamous cell cancers were evenly distributed between the Ranch Hand and Comparison groups. All skin cancers occurred in nonblacks. Adjustments for occupational exposures (e.g., asbestos, degreasing chemicals) did not alter the increased rate of skin cancer in .the Ranch Hand group. Skin cancer in both groups was associated with exposure to industrial chemicals (p=0.03). Herbicide exposure analyses in the Ranch Hand group were essentially negative, although confounding was noted in many of the analyses. Outdoor occupations subsequent to military service as a covariate did not account for the significant skin cancer association. 10-5 �Parameters of the 1985 Malignancy Assessment The emphasis on cancer was increased during the first followup study in 1985. With the Baseline finding of excessive skin cancer in the Ranch Hands, and the lack of covariate data to refine that association, considerable attention was devoted to skin cancer. The questionnaire was altered to collect information on each geographic location in which a participant lived for more than 12 months in order to calculate a cumulative "lifetime" sun exposure index based on geographic latitude, since ultraviolet light exposure has been acknowledged as the primary cause of basal cell carcinoma. Detailed data on skin tannability, eye, skin, and hair color, and parental ethnicity were also obtained. In addition, emphasis at the dermatologic examination was shifted from acne/chloracne to skin cancer, and punch biopsies were sought for all suspected malignant lesions. The participants were asked to bring copies of their medical records to facilitate the verification of reported malignancies. Highly structured smoking data were collected for more detailed covariate adjustments, and Baseline questions on exposure to other carcinogens were repeated to gather interval data. No invasive procedures were used at the followup physical examination to detect evidence of systemic cancer. Thus, the dependent variables of the analyses below are Baseline analyses, but covariate analyses have been expanded and systemic cancers. The lifetime occurrence of cancer, as interval occurrence of skin and systemic cancers between the followup examinations, is analyzed. similar to the for both skin well as the Baseline and Minor numeric differences in various tables that follpw reflect missing data from the covariates. The statistical methods used throughout this chapter are Fisher's exact test, chi-square tests of association, and logistic regression models (BMDP®-LR) for adjusted group contrasts of neoplasm incidence rates. RESULTS AND DISCUSSION General Malignant and benign neoplasms, carcinomas in situ, and neoplasms of uncertain behavior or unspecified nature are studied in this chapter. The term "systemic" is used throughout to denote a nonskin neoplasm. The term "unspecified" is used to denote a neoplasm of uncertain behavior or unspecified nature. Neoplasm refers to any new and abnormal growth which may or may not be malignant. Malignant neoplasms (malignancies, cancer) are those neoplasms that are capable of invasion and metastasis. Questionnaire Data At the followup examination, participants provided information on cancer during the interval between examinations and participants who were new to the study gave their lifetime history. All reported neoplasms entered the medical records review process for verification. Only 11 Ranch Hands (1.1%) and 12 Comparisons (0.9%) reported neoplastic conditions which could not be 10-6 �substantiated (all of the skin); the group difference was nonsignificant (p-0.833). Physical Examination Data Some possible neoplastic conditions were discovered by the physicians at the physical examination. Many suspicious skin lesions were biopsied and the pathology determined. However, for some suspected skin neoplasms and all suspected systemic neoplasms, verification was not complete at the time of writing this report, and thus both verified and suspected neoplasms are described and analyzed. The term suspected is used throughout to denote those possible neoplastic conditions noted by the physicians at the followup examination for which the results of verification are not yet available. Consideration of suspected neoplasms was justifiable in particular for skin neoplasms, for which the biopsy confirmation rate is high. Statistical Analysis The statistical analysis is described in three sections. The first section presents unadjusted and adjusted analyses of skin and systemic neoplasm incidence in the Baseline-followup interval, and is referred to as interval analysis. In the second section, unadjusted and adjusted analyses of lifetime skin and systemic neoplasm incidence are analyzed for the followup participants, incorporating Baseline information. Since there were very few neoplasm occurrences before the SEA tours, this combined interval and Baseline analysis is referred to throughout as lifetime analysis. Lastly, the neoplasm history and mortality of the fully compliant Baseline participants subsequent to Baseline are described. All analyses are of the numbers of participants with (one or more) neoplasms, and not of the total number of neoplasms. The purpose of these three analyses is to present a comprehensive picture of the neoplasia history of the followup participants, and to provide some additional information on the neoplasia status of the Baseline participants subsequent to Baseline. There was a slight difference between the Baseline and followup cohorts. The interval and lifetime analyses pertain to neoplasm incidence among followup participants only. The third section pertains to Baseline participants only, describing their history of neoplasm incidence and mortality since Baseline. A fully combined morbidity-mortality analysis was not feasible for this report. Assuming a (two-sided) a -level of 0.05 and power 0.8, the sample sizes were sufficient to detect a relative risk of 2.56 when the Comparison neoplasm incidence rate is 1 percent, and a relative risk of 1.63 when the Comparison neoplasm incidence rate is 5 percent. For nonblacks only, the corresponding detectable relative risks are 2.63 and 1.65, respectively. All analyses of data from Ranch Hands and the Original Comparisons only are given in Appendix H. This appendix also contains other tabulations, such as covariate and interaction tables. 10-7 �Baseline-Follovup Interval Table 10-1 shows the Baseline-follovup interval neoplasm history for the followup participants. The interval began in January 1982 for participants new to the study, i.e., the 45 new Ranch Hands, the 71 new replacement Comparisons, and 83 newly compliant Original Comparisons. The total numbers of participants with verified neoplasms were 161/1,016 (15.8%) Ranch Hands and 170/1,293 (13.1%) Comparisons; the group difference was marginally significant (p=0.073). The relative frequencies of participants with verified plus suspected neoplasms, 17.4 percent of Ranch Hands and 16.2 percent of Comparisons, did not differ significantly between groups (p=0.466). Appendix Table H-l gives the numbers of participants with verified or suspected neoplasms and unadjusted analyses for the Ranch Hands and Original Comparisons in the Baseline-followup interval. Interval Skin Neoplasms Of Ranch Hands with verified neoplasms of all types (malignant, benign, and uncertain) 70.8 percent (114/161) had skin neoplasms; the corresponding percentage for the Comparisons was 68.2 percent (116/170). The difference in these proportions was not significant (p=0.634). When suspected neoplasms were included, the contrast was 70.1 percent (124/177) versus 67.6 percent (142/210), again not significant (p=0.660). No Blacks were found to have skin cancer, as anticipated since Blacks have a lower susceptibility to sun-induced skin cancer. Therefore, analysis of skin cancer was limited to nonblacks. Of Ranch Hands with skin neoplasms, 32.5 percent (37/114) had malignant neoplasms, as contrasted to 34.5 percent (40/116) of the Comparisons (p=0.781). When suspected malignant skin neoplasms were included, the contrast was 37.9 percent (47/124) versus 42.3 percent (60/142), and was not significant (p=0.531). For the remainder of this section, only malignant skin neoplasms are analyzed. The dependent variables examined were basal cell carcinomas, melanomas, squamous cell carcinomas, all skin cancers combined, and a group of skin cancers called sun exposure-related skin malignancies. The sun exposure-related skin malignancies were defined as basal cell carcinomas, melanomas, and malignant epithelial neoplasms not otherwise specified (NOS). The latter were included because they are frequently misdiagnosed basal cell carcinomas; three Ranch Hands had this diagnosis. Interval Malignant Skin Neoplasms Table 10-2 presents the numbers of participants with verified and suspected malignant skin neoplasms by cell type: basal cell carcinomas, squamous cell carcinomas, melanomas, all skin malignancies combined, and the sun exposure-related skin malignancies, together with the results of unadjusted group contrasts. For the sake of completeness, the total numbers of malignancies of each type are also given. The majority of the 10-8 �TABLE 10-1. Unadjusted Analyses of Followup Participants with Verified (Nonblacks and Blades) Group* Ranch Hand Site Skin Systanic All Neoplasm Behavior and Status Comparison , Number** Percent Number** Percent Malignant Verified Verified and Suspected Benign Verified Verified and Suspected Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected Any Skin Neoplasm* Verified Verified and Suspected Total** p-Value*** 37 47 3.6 4.6 40 60 3.1 4.6 77 107 0.485 0.999 76 78 7.5 7.7 77 83 6.0 6.4 153 161 0.152 0.250 1 1 0.1 0.1 1 0.1 1 • 0.1 2 2 099 .9 0.999 114 124 11.2 12.2 116 142 9.0 11.0 230 266 0.030 0.393 8 12 08 . 1.2 7 12 0.5 0.9 15 24 0.603 0.680 42 48 4.1 4.7 50 61 3.9 4.7 92 109 0.749 0.999 6 6 0.6 06 . 7 11 0.5 0.9 13 17 0.999 0.625 55 65 5.4 6.4 61 80 4.7 6.2 116 145 0.445 0.863 Malignant, Benign, Uncertain Behavior, Unspecified Nature Verified 161 Verified and Suspected 177 15.8 17.4 170 210 13.1 16.2 331 387 0.073 046 .6 Malignant Verified ' Verified and Suspected Benign Verified Verified and Suspected Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected Any Systemic Neoplasm3 Verified Verified and Suspected *Sample sizes: 1,016 Ranch Hands and 1,293 Comparisons. **Number of participants. ***Fisher's exact test. ^Participant has one or more malignant, benign, or unspecified skin neoplasms. Participant has one or more malignant, benign, or unspecified systemic neoplasms. °Participant has one or more malignant or benign skin or systemic neoplasms. 10-9 �TABLE 10-2. Unadjusted Analyses of Nonblack Follovup Participants vith Verified and Suspected Malignant Skin Neoplasms in the Baseline-Followup Interval by Cell Type and Group Group* Cell Type Status Statistic** Ranch Hand Comparison Est. Relative Risk (95% C.I.) p-Value Verified Number/% Total Neoplasms 29 42 3.0% 30 40 2.5% 1.23 (0.73,2.07) 0.429 Verified & Suspected Number/% Total Neoplasms 36 53 3.8% 48 63 4.0% 0.95 (0.61, 1.47) 0.824 Verified Number/% Total Neoplasms 4 6 0.4% 4 4 0.3% 1.27 (0.32,5.08) 0.738 Verified & Suspected Number/% Total Neoplasms 4 6 0.4% 5 5 0.4% 1.01 (0.27,3.78) Verified Number/% Total Neoplasms I 2 0.1% 3 3 0.3% 00, 0.42 ( . 4 4.06) 0.635 Verified & Suspected Number/% Total Neoplasms I 2 0.1% 6 7 0.5% 0.21 ( . 3 1.75) 0.142 00, All Malignant Verified Skin Neoplasms Verified & Suspected Number /% Total Neoplasms 37 56 3.9% 40 52 3.3% 1. 18 (0.75, 1.86) 0.486 Number/% Total Neoplasms 47 70 4.9% 60 81 5.0% 0.99 (0.67, 1.47) 0.999 Sun-Exposure Verified Related Malignant Neoplasms3 Verified & Suspected Number/% Total Neoplasms 32 47 3.4% 33 43 2.7% 1.24 (0.75,2.02) 0.447 Number/% Total Neoplasms 39 58 4.1% 53 71 4.4% 0.93 (0.61,1.42) 0.749 Basal Cell Carcinoma Squamous Cell Carcinoma o Melanoma *Number of participants—956 Ranch Hands and 1,210 Comparisons. **Number and percent of participants; total number of malignant neoplasms of specified cell type. a Basal cell carcinoma, melanoma, and malignant epithelial neoplasms NOS. 0.999 �participants with verified skin malignancies had basal cell carcinomas: 78.4 percent (29/37) Ranch Hands versus 75.0 percent (30/40) Comparisons; the difference between the groups was not significant (p=0.792). Unadjusted Analyses Table 10-2 shows that no significant group differences were found in the incidence rates of either verified or verified plus suspected malignant skin neoplasms. For verified basal cell carcinomas, the estimated relative risk of Ranch Hands versus Comparisons was 1.23 (95% C.I.: [0.73,2.07]) and was not significant (p=0.429). The estimated relative risk for verified squamous cell carcinoma, 1.27 (95% C.I.: [0.32,5.08]), was also not significant (p=0.738). The estimated relative risk for verified melanoma, 0.42 (95% C.I.: [0.04,4.06]), was also not significant (p=0.635). There were very few occurrences of melanoma (one Ranch Hand and three Comparisons) since this is a much rarer condition than other kinds of skin cancer. There were no significant differences between the groups for all verified malignant skin cancers combined (Est. RR: 1.18, 95% C.I.: [0.75,1.86], p=0.486) or for the category of sun exposure-related skin malignancies (Est. RR: 1.24, 95% C.I.: [0.75,2.02], p=0.447). When both verified and suspected malignant skin neoplasms were analyzed, the conclusions were similar, namely, there were no significant differences between the groups, and moreover, the estimated relative risks were closer to 1. No group differences were found in the parallel contrasts of Ranch Hands versus Original Comparisons (see Table H-2 of Appendix H). As shown in Table 10-3, additional analyses contrasted group differences in the anatomic location of basal cell carcinomas, melanomas, and sun exposure-related skin malignancies. Most occurrences of basal cell carcinoma and sun exposure-related skin malignancies were on the face, head, or neck, or the upper extremities. The relative frequency of occurrences of verified basal cell carcinomas at these combined sites was 89.7 percent for Ranch Hands and 80.0 percent for Comparisons of the total number of occurrences in each group, respectively. The group contrast (26/29 versus 24/30) was not significant (p=0.472). These combined sites accounted for 90.6 percent (29/32) of the sun exposure-related malignancies for Ranch Hands versus 72.7 percent (24/33) for Comparisons; this contrast was also not significant (p=0.108). The corresponding contrasts, when suspected malignant neoplasms were included with the verified malignant neoplasms, were also not significant. One Ranch Hand had verified melanoma of the face, and three Comparisons had verified melanoma on the trunk. Two other Comparisons had suspected melanoma, also on the trunk. The group contrast for melanomas on the trunk was not significant for verified conditions (p=0.260), but was marginally significant for verified plus suspected conditions (p=0.071), the detriment being in the Comparison group. Table 10-4 gives the frequencies of participants with face, head, and neck skin malignancies by group and occupation. Specifically, nonmelanoma malignant skin neoplasms and the sun exposure-related malignant skin neoplasms are listed by occupational category. For officers and enlisted groundcrew, the frequencies of participants with face, head, and neck malignant skin neoplasms (both malignant nonmelanoma and the malignant sun exposure-related skin neoplasms) did not differ significantly by group. However, the Ranch Hand enlisted flyers had a significantly higher frequency 10-11 �10-3. Malignant Skin Neoplasms in the Ra«*>liT*>-PnlLowup Interval by Anatomic Site and Group Basal (jell Carcinoma Sun-Exposure Related Malignancies Group* Group* Site Face, Head, Neck K Status** Ranch Hand Comparison p-Value Number/% Verified Verified & Suspected 24 2.5% 29 3.0% Upper extreai- Number/% ities Verified Verified & Suspected Melanoma Group* Ranch Hand Comparison p-Value Ranch Hand Comparison p-Value 034 .7 089 .9 27 2.8% 32 3.4% 23 1.9% 36 3.0% 0.194 0.622 1 0.1% 1 0.1% 0 0.0% 0 4 1 .4 .9 1 0.1% 0 9 9 5 0.5% 5 0.5% 3 0.3% 0.313 4 0.3% 0 5 0 .2 5 05 .% 5 0.5% 3 0.3% 5 04 .% 0.313 0.757 0 00 .% 0 00 .% 0 00 .% 0 00 .% a a Number/% Verified Verified & Suspected 2 0.2% 4 04 .% 6 0.5% 0 4 9 .7 11 0 9 0 2 0 .% .0 2 0.2% 4 04 .% 9 0.7% 14 1.2% 0.126 003 .9 0 0.0% 0 00 .% 3 0.3% 5 04 .% 020 .6 0.071 Lower Extrem- Number/% ities Verified Verified & Suspected 0 0.0% 0 0.0% 0 00 .% 0 00 .% 0 00 .% 0 00 .% 0 00 .% 0 00 .% 0 00 .% 0 0.0% 0 00 .% 0 0.0% a a Other Sites Number/% and Sites NOS Verified Verified & Suspected I 0.1% I 0.1% 2 0.2% 0 9 9 .9 2 0.2% 0 9 9 .9 1 0.1% 1 0.1% 2 0.2% 2 0.2% 099 .9 099 .9 0 0.0% 0 0.0% 0 00 .% 0 00 .% a a All Locations Number/% Verified Verified & Suspected 29 3.0% 36 3.8% 32 3.4% 39 4.1% 33 2.7% 53 4 4 .% 047 .4 079 .4 1 0.1% 1 0.1% 3 0.3% 6 0.5% 0.635 0.142 Trunk 23 1.9% 35 2.9% 30 2.5% 48 4 0 .% a a 049 .2 084 .2 *Number of participants — 956 Ranch Hands, 1,210 Comparisons. **Number and percent of participants. a No occurrences in either group. a a �TABLE 10-4. Unadjusted Analyses of Hnriblark FolloHup Rarticipants with Nomelanana Maligpant Skin Neoplasms and Sun-Exposure Related Skin Malignancies in the Basel Ine-RJloHUp Interval Occurring on the Face, Head, or Neck by Occupation Occupation Status Statistic n Group* Cotppyision Ranch H^ryi p-Value Number Percent Number Percent 477 373 Group* Conparison Ranch Hand p-Value Number Percent Number Percent 477 373 Verified Face, Head, Neck 14 Other Site 5 No Cancer 354 3.8 1.3 94.9 17 4 456 3.6 08 . 95.6 099 .9* 056 .1" 12 3 358 3.2 08 . 9. 60 13 5 459 2.7 1.1 96.2 068 .8* 099 .9" Verified & Suspected Face, Head, Neck 20 7 Other Site No Cancer 346 5.4 1.9 92.8 23 8 446 48 . 1.7 93.5 074 .5* 099 .9" 17 4 352 4.6 1.1 9. 44 20 9 448 4.2 1.9 93.9 086 .6* 048 .0" Officer n 167 167 193 193 Verified 8 Face, Head, Neck Other Site 1 No Cancer 158 Enlisted Flyer 4.8 06 . 94.6 3 0 190 1.6 00 . 98.5 0.121* 046 .5" 8 1 158 48 . 06 . 94.6 2 1 190 1.0 0.5 98.5 009 .4* 099 .9" 8 Verified & Suspected Face, Head, Neck 1 Other Site No Cancer 158 48 . 06 . 9. 46 5 1 187 2.6 06 . 96.9 036 .9* 099 .9" 8 1 158 48 . 06 . 94.6 4 2 187 2.1 10 . 96.9 028 .3* 099 .9" n 416 416 540 540 Verified 7 Face, Head, Neck Other Site 1 No Cancer 408 1.7 02 . 98.1 9 4 527 1.7 0.7 97.6 099 .9* 035 .9" 7 1 408 1.7 0.2 98.1 8 4 528 1.5 0.7 97.8 080 .0* 0.395" Enlisted Groundcrew Verified & Suspected Face, Head, Neck 7 Other Site 3 No Cancer 406 1.7 07 . 97.6 D 6 521 2.4 1.1 96.5 050 .0* 079 .3" 7 2 407 1.7 0.5 9. 78 12 6 522 2.2 1.1 96.7 064 .4* 047 .7" *Number and percent of participants. *Fisher's exact test for face, head, or neck versus no malignancy. Fisher's exact test for other site versus no malignancy. �of malignant sun exposure-related skin neoplasms than the corresponding Comparisons, 4.8 percent versus 1.0 percent (p«0.049). For nonmelanoma malignant skin neoplasms, the contrast was 4.8 percent versus 1.6 percent, but the difference was not significant (p=0.121). Inclusion of suspected malignant neoplasms with the verified malignant neoplasms reduced the significance of the difference between the groups for both the sun exposure-related skin malignancies and the nonmelanoma malignant skin neoplasms. Adjusted group contrasts of the incidence rate of basal cell carcinomas and malignant sun exposure-related skin neoplasms were done for verified and verified plus suspected conditions. Adjusted analyses were not carried out, however, for melanomas or squamous cell carcinomas because of the small frequencies. Covariates The covariates considered for the adjusted analyses of malignant skin neoplasm incidence, listed in Table 10-5, were the matching variables age and occupation; history of alcohol and cigarette use; host factors, comprising skin color, eye color, hair color, and ethnic background; reaction of skin to sun exposure; average lifetime residential latitude; and exposure to recognized carcinogens. Age was used as a continuous variable in the adjusted analyses, but was categorized for ease of presentation in the report. Eye color, hair color, and skin color were coded by the dermatologist at the physical examination. Hair color was determined by comparing the hair at the back of the neck with 17 numbered standardized hair samples and selecting the most closely matching hair sample. Similarly, skin color groupings from dark brown to pale peach were determined by comparing standardized flesh-colored squares against the skin of the inside upper arm. For the analysis, hair and skin colors were grouped as shown in Table 10-5. Each participant was assigned to one of four ethnic groups according to his responses to questions on race, as given in Table 10-5. (Blacks were omitted from the table because the analysis of malignant skin neoplasia was restricted to nonblacks.) These ethnic categories are approximate groupings in terms of susceptibility to sun-induced skin damage. The ethnic categories also generally correlate to skin color, a commonly known important risk factor for skin cancer. A lifetime residential history was obtained from all participants by a questionnaire. Residential history, relative to the equator, is a surrogate measure of sun exposure (but does not account for altitude or average sun-days at each location), an important risk factor for skin cancer. Each participant was asked to list all residences chronologically, citing both the city (or military installation) and the years of residence at each location since birth. Residences of less than 1 year were not sought because of the frequent short-term military travels of these cohorts. By standardized geographic atlases, the latitude (in degrees and minutes) of each residence was recorded. The Air Force subsequently checked all of the latitude determinations for accuracy. The average lifetime residential latitude of each participant was calculated by dividing the total degree-years (i.e., sum of latitude [degrees] times number of years lived there) from all residences by the total number of residential years listed. 10-14 �TABLE 10-5. Covariates for Analyses of Malignant Skin Neoplasms Category Covariate Age Born XL942, 1923-1941, <1922a Occupation Officer, Enlisted Flyer, Enlisted Groundcrew Lifetime Cigarette Smoking Pack-years: 0, >0-20, >20-40, >40 Lifetime Alcohol Consumption Drink-years: 0, >0-5, >5-30, >30-100, >100 Ethnic Background A, B, C, Db Skin Color Dark, medium, pale, dark peach, pale peach Hair Color Black, dark brown, light brown, blond, red Eye Color Brown, hazel, green, gray, bluec Reaction of Skin to Sun Exposure : (A.I) After first 30 minutes of summer sun (A.2) After >2 hours, after first exposure (A.3) After repeated sun exposures Burns, usually burns, burns mildly, rarely burns Burns painfully, burns, becomes red, no reaction Freckles with no tan, tans mildly, tans moderately, tans deep brown Sun-Reaction Index (Composite)*3 (1) Burns painfully (A.2) and/or freckles with no tan (A.3) (2) Burns (A.2) and/or tans mildly (A.3) (3) All other reactions Residential History (Average Latitude) Average latitude <37°, >37° Exposure to Carcinogens/Groups of Carcinogens Set I3 Asbestos Nonmedical X Rays Industrial Chemicals Herbicides Insecticides Degreasing Chemicals Yes, Yes, Yes, Yes, Yes, Yes, No No No No No No 10-15 �TABLE 10-5. (continued) Covariates for Analyses of Malignant Skin Neoplasms Covariate Set 2e Anthracene Arsenic Benzene Benzidene Chromates Coal Tar Creosote Aminodiphenyl Chlororaethyl Ether Mustard Gas Naphthylamine Cutting Oils Trichloroethylene Ultraviolet Light (not sun) Vinyl Chloride Composite Carcinogen Exposure Category Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, Yes, No No No No No No No No No No No No No No No Yes, if yes for exposure to any carcinogen in set 2, otherwise no. a Used as a continuous variable in adjusted analysis. b A B C D - English, Welsh, Scottish, Irish. Scandinavian, German, Polish, Russian, other Slavic, Jewish, French. Spanish, Italian, Greek. Mexican, American Indian, Asian. c Participant with one green eye and one brown eye is coded as green. Questionnaire data (see Appendix B). *AFHS Form 2 (see Appendix C). 10-16 �Recognizing that both total degree-years and average lifetime latitude could be covariates for malignant sun exposure-related skin neoplasms, average latitude was selected because of the high correlation of degree-years with chronological age, a separate risk factor already used in the analyses. Further, average residential latitude was believed to be a more stable measure in the presence of some lack of precision in the source data. In all analyses, the average residential latitude was used as a dichotomous variable (less than 37° N latitude, greater than or equal to 37° N latitude). A line across the United States at 37° N approximates a line from San Francisco, California, to Richmond, Virginia. Examination of the group distributions of the latitude variable suggest that it is a significant confounding variable. Specifically, 56.7 percent of the nonblack Ranch Hands had an average lifetime residential latitude greater than or equal to 37° N latitude versus 49.4 percent of the nonblack Comparisons (p=0.001). Although the average lifetime group residential latitudes appear similar (37.21° N latitude for the Ranch Hands, and 36.74° N latitude for the Comparisons), this difference is also highly significant (p=0.003), reflecting the substantial power of the analysis of continuous data. Participants reported their susceptibility to the effects of sunexposure damage by answering three questions about their skin reaction to sun: the reaction after the first 30 minutes of exposure to summer sun, the reaction after 2 or more hours of sun exposure after the first 30-minute exposure, and the reaction after repeated exposures (see questions 10-12 on page 71 of the questionnaire provided in Appendix B). Since these three responses are highly correlated, a composite sun-reaction variable for use in the adjusted analysis, called the sun-reaction index, was constructed from the last two questions (2-hour and repeated exposure reactions) after examination of the association between basal cell carcinoma incidence and the three skin reaction variables. The sun-reaction index had three categories. The first category corresponded to the most sensitive reaction on the last two questions, the second category corresponded to the next less sensitive .reaction on these two questions, and the third category comprised the remaining responses. Detailed questionnaire information on exposure to asbestos, nonmedical x rays, industrial chemicals, herbicides, insecticides, and degreasing chemicals was obtained from each participant. Self-reported information on exposure to 15 individual carcinogens was obtained at the physical examination. A composite carcinogen exposure variable was constructed from these responses on individual carcinogens: A participant had a positive score for this variable if he reported exposure to one or more of the 15 carcinogens, otherwise he had a negative score. Self-reported information on asbestos and radiation exposure was not used because this information was obtained in more detail from the questionnaire. The nonblack Ranch Hands differed significantly from the nonblack Comparisons in their exposure (yes/no) to nonmedical x rays (19.3% versus 25.6%, p<0.001). They also differed significantly from the Comparisons in their exposure to herbicides (94.1% versus 29.8%, p<0.001) and insecticides (70.2% versus 53.1%, p<0.001), possibly reflecting Vietnam experience. These variables were not used in the adjusted analysis. Further, there were significant or marginally significant group differences in the self-reported exposures to several individual carcinogens, in each instance relatively more (nonblack) Ranch Hands than Comparisons reported exposure: arsenic (2.7% 10-17 �versus 1.2%, p-0.016), naphthylamine (3.3% versus 1.7%, p=0.024), cutting oils (12.7% versus 8.7%, p=0.003), benzene (4.3% versus 2.7%, p=0.056), and benzidine (0.8% versus 0.3%, p=0.070). Results were similar when Blacks were included in the analysis. Covariate Associations Table 10-6 gives a summary of the chi-square tests of association between all covariates and the incidence of basal cell carcinomas and sun exposure-related malignancies. Details of these tests of association are provided in Appendix H, Table H-3. There was a significant increase in the incidence rate of verified basal cell carcinomas with increasing age (p=0.001). There was a significant difference in the incidence rate of basal cell carcinomas among occupation groups, with enlisted groundcrew having a lower incidence rate (1.8%) than officers (3.7%) and enlisted flyers (3.1%) (p=0.047). Since officers are, on the average, 5 years older than enlisted participants, this occupation effect may be due to some confounding with age. There was a higher incidence rate for average lifetime residential latitude less than 37° N versus greater than or equal to 37° N latitude (p=0.008). Furthermore, there was a strong difference for different levels of the sun-reaction index (p<0.001), and the three skin-reaction-to-sun variables (p<0.001 for all). Participants who tended to burn most had a lower rate (1.4%) than those with a milder reaction (6.0%), and a similar rate to those who tended to tan (1.9%) (an unexpected finding). There was a significant relationship between the incidence rate of basal cell carcinoma and total pack-years of lifetime smoking (p=0.023 for verifieds). This effect may also be due to confounding with age rather than to a primary smoking effect (see Table H-5 of Appendix H). No significant association was found between the incidence rate of verified basal cell carcinoma and lifetime drink-years. No significant associations were found with ethnic group, skin color, eye color, and hair color. However, when the ethnic group categories were dichotomized as Celtic or English versus other ethnic groups, the association was marginally significant (p=0.093). Skin color was dichotomized as dark peach or light peach versus other colors, and the association was significant (Est. RR: 3.00, 95% C.I.: [1.08,8.33], p=0.024). Hair color was dichotomized as blond or red versus other colors. The association of hair color with basal cell carcinoma incidence was not significant (p=0.384). Furthermore, no significant relationship was found between basal cell carcinoma incidence and the composite carcinogen-exposure variable (p=0.523) or the grouped or individual carcinogens. The associations between the covariates and the incidence of verified plus suspected basal cell carcinomas paralleled those for the verified basal cell carcinomas only, except that the difference in rates among ethnic groups was significant (p=0.046), hair color was significant (p=0.040), and a marginally significant positive relationship was found with nonmedical x-ray exposure (p=0.084) and herbicide.exposure (p=0.072). The difference among occupation groups, however, was more significant (p=0.003). 10-18 �TABLE 10-6. Summary of Associations Between Incidence Rates of Basal Cell Carcinoma and Sun Exposure-Related Skin Malignancies and the Covariates, in the Baseline-Followup Interval for Combined Follovup Ranch Hand and Comparison Nonblack Participants Covariate Sun Exposure-Related Skin Malignancies Basal Cell Carcinoma Verified & Verified & Suspected Verified Suspected Verified p-Value p-Value p-Value p-Value Age 0.001 <0.001 0.004 <0.001 Occupation 0,047 0.003 NS* 0.006 Lifetime Cigarette Smoking 0.023 0.005 0.012 0.007 Lifetime Alcohol Consumption NS NS NS NS Ethnic Background NS 0.046 NS 0.036 Skin Color NS** NS NS NS** Hair Color NS 0.040 NS NS* Eye Color NS NS NS NS Reaction of Skin to Sun Exposure: (Q.I) After first 30 minutes of summer sun (Q.2) After >2 hours, after first exposure (Q.3) After repeated sun exposures Sun-Reaction Index (Composite) Residential History (Average Latitude) Exposure to Carcinogens/Groups of Carcinogens Set la Asbestos Non-medical X Rays Industrial Chemicals Herbicides Insecticides Degreasing Chemicals 0.001 <0.001 <0.001 <0.001 <0.001 0.027 0.001 0.016 <0.001 0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.008 0.004 0.011 0.003 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS 10-19 NS NS* NS NS* NS NS �TABLE 10-6. (continued) Summary of Associations Between Incidence Rates of Basal Cell Carcinoma and Sun Exposure-Related Skin Malignancies and the Covariates, in the Baseline-Followup Interval for Combined Follovup Ranch Hand and Comparison Nonblack Participants Covariate Set 2° Anthracene Arsenic Benzene Benzidene Chromates Coal Tar Creosote Aminodiphenyl Chloromethyl Ether Mustard Gas Naphthylamine Cutting Oils Trichloroethylene Ultraviolet Light (not sun) Vinyl Chloride Composite Carcinogen Exposure NS: Sun Exposure-Related Skin Malignancies Basal Cell Carcinoma Verified & Verified & Verified Suspected Verified Suspected p-Value p-Value p-Value p-Value NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS Not significant (p<0.10). NS*: Borderline significant (0.05<p<0.010). **Not significant when five categories of skin color examined; however, when dichotomized, p=0.024 for verified basal cell carcinoma and p=0.036 for verified and suspected sun exposure-related skin malignancies. a Questionnaire data. b AFHS Form 2. 10-20 �As expected, the relationships between the incidence of verified sun exposure-related skin malignancies and the covariates were similar to those just described for basal cell carcinomas (Table 10-6 and Table H-4 of Appendix H). For verified conditions, there was a strong increase in incidence rate with age (p=0.004), total lifetime smoking (p=0.012), average lifetime residential latitude (p=0.011), the reaction-to-sun exposure variables (p<0.001 for all), and the sun-reaction index (p<0.001), with similar strong associations for the verified plus suspected conditions. The difference among occupation groups was marginally significant (p=0.077) for verified conditions; this difference was significant (p=0.006) for verified plus suspected sun exposure-related skin malignancies (officers 5.9%, enlisted flyers 4.2%, enlisted groundcrew 2.8%). There was no association with the composite carcinogen-exposure variable, either for verified (p=0.879) or for verified plus suspected conditions (p=0.608). Table 10-6 shows no significant association between the incidence rate of verified sun exposure-related skin malignancies and ethnic group, hair color, skin color, or eye color. When suspected conditions were included, the ethnic group association was significant (p=0.036), and the association with hair color became borderline significant (p=0,051). There were higher incidence rates among those of Celtic or English background as opposed to other ethnic backgrounds, and among participants with blond or red hair as opposed to other colors (see Table H-4 of Appendix H). As in the analysis of basal cell carcinomas, the ethnic group, hair color, and skin color categories were collapsed, resulting in (for verified conditions): p«0.054 for those of Celtic or English backgrounds versus other ethnic backgrounds (Est. RR: 2.04, 95% C.I.: [1.00,4.17]) and p=0.031 for skin color peach versus not-peach (Est. RR: 2.61, 95% C.I.: [1.04,6.58]), but no significant association with hair color grouped as blond or red versus other (p=0.268) was found. Adjusted Analyses Because of the obvious interrelatedness among the host factors of hair color, skin color, eye color, ethnic background, and reaction of skin to sun, and because a smaller set of covariates was required for the adjusted analyses, a "main-effects" statistical model of basal cell carcinoma with the following covariates was used: age, occupation, total pack-years, lifetime drinking, ethnic background (dichotomized), hair color (blond or red versus other), eye color, skin color (peach tones versus other), the three skin-reaction-to-sun variables, average lifetime residential latitude (less than 37° N versus greater than or equal to 37° N), and the composite carcinogen exposure variable. The results of this analysis are given in Appendix H, Table H-5. The results showed that ethnic background, hair color, and the 30-minute skin-reaction-to-sun variable, while individually associated with basal cell carcinoma incidence, are relatively less important than the other host factors, namely skin color, and the 2-hour and repeatedexposure skin-reaction-to-sun variables, and were thus not included in the adjusted analyses. Total drink-years and the composite carcinogen exposure variable were not significant and thus were not used in the adjusted analyses. A parallel analysis was conducted in which the composite sunreaction index replaced all three skin-reaction-to-sun variables, and it was found that this substitution could be made without altering the relative contributions of the other covariates. For further reduction of the number of covariates, pack-years of smoking, although of interest (p=0.096), was 10-21 �also omitted. Thus, a reduced set of covariates for further analysis of the group contrasts was identified as age, occupation, skin color, average lifetime residential latitude, and the sun-reaction index. The results of adjusted analyses of group contrasts in the incidence rate of basal cell carcinoma and sun exposure-related skin malignancies are presented in Table 10-7. Parallel results for Ranch Hands contrasted with the Original Comparisons are given in Appendix H, Table H-6. A significant group-by-occupation interaction was found for verified interval basal cell carcinoma (p=0.044). Significant covariates were age (p=0.003), average residential latitude (p=0.003) and the sun-reaction index (p<0.001). The interaction was due to a significant difference in rates for enlisted flyers but not for officers or enlisted groundcrew: Ranch Hand enlisted flyers had a significantly (p=0.019) greater incidence rate of basal cell carcinomas than the corresponding Comparisons, 5.4 percent versus 1.0 percent (Adj. RR: 6.50, 95% C.I.: [1.36,31.01]) (see Appendix H, Table H-7). There was a significant group-by-sun-reaction index interaction in the analysis of verified plus suspected basal cell carcinomas (p=0.024); this was in part attributable to the absence of Ranch Hands who reported burning easily. The group frequencies for the three levels of this variable (burn easily, intermediate reaction, tan easily) were: Ranch Hands 0 (0%), 17 (8.9%), and 19 (2.7%), respectively, and Comparisons 4 (5.2%), 15 (5.7%), and 28 (3.2%), respectively. The incidence rate for Ranch Hands who had a moderate reaction to sun was (nonsignificantly) greater than that of the Comparisons. The details of this interaction are given in Appendix H, Table H-7. A skin color-by-age interaction (p=0.044) and average latitude (p=0.003) made significant contributions to the model. Results of the analyses for Original Comparisons were nonsignificant for verified conditions, although a marginally significant group-by-sun reaction interaction was found (p=0.051). The results for verified plus suspected conditions revealed a significant group-by-sun reaction index interaction (p=0.007) (see Table H-6 of Appendix H). Ranch Hands who had a moderate skin reaction to sun revealed a significantly greater incidence rate of verified basal cell neoplasms than corresponding Original Comparisons (Adj. RR: 2.81, 95% C.I.i [1.05,7.55], p=0.040) (Table H-8). This finding was marginally significant with the inclusion of suspected carcinomas (Adj. RR: 2.38, 95% C.I.: [0.98,5.76], p=0.055). The adjusted relative risk for the incidence rate of verified sun exposure-related skin malignancies was 1.37 (95% C.I.: [0.83,2.28]) and was not significant (p=0.221) (Table 10-7). Age (p<0.001), the sun-reaction index (p<0.001), and average lifetime residential latitude (p=0.008) contributed to the adjustment. No group difference was apparent when suspected malignancies were included. The adjusted relative risk was 1.05 (95% C.I.: [0.68,1.62], p=0.825), and the significant covariates were a skin color-bysun-reaction index interaction (p=0.028), a skin color-by-age interaction (p=0.028), and a skin color-by-residential latitude interaction (p=0.041). 10-22 �TABLE 10-7. Adjusted Analyses of Nonblack Follbwup Participants for Malignant Skin Neoplasm Incidence During the Baseline-Followup Interval Adj. Relative Risk (95% C.I.) Variable Status Basal Cell Carcinoma Verified **** **** AGE (p=0.003) LAT (p=0.003) SUNREAC (p<0.001) GRP*OCC (p=0.044) Verified & Suspected **** **** LAT (p=0.003) GRP*SUNREAC (p=0.024) SKIN*AGE (p=0.044) Sun-Exposure Verified Malignant Skin Neoplasms Verified & Suspected p-Value Covariate Remarks* 1.37 (0.83,2.28) 0.221 AGE (p<0.001) SUNREAC (p<0.001) LAT (p=0.008) 1.05 (0.68,1.62) 0.825 SKIN*SUNREAC (p=0.028) SKIN*AGE (p=0.028) SKIN*LAT (p=0.041) ^Abbreviations; LAT: average lifetime residential latitude SUNREAC: sun reaction index GRP: group OCC: occupation SKIN: skin color ****Group-by-covariate interaction—adjusted relative risk, confidence interval, and p-value not presented. 10-23 �Analysis of the Ranch Hands versus Original Comparisons contrasts found a significant group-by-skin color interaction for verified sun exposurerelated malignancies (p=0.036), and a significant group-by-sun reaction index interaction (p=0.030), similar to that found for basal cell carcinoma, for the verified plus suspected malignant neoplasms (see Appendix H, Tables H-6 and H-8, for details). The group-by-skin color interaction was due to a lower incidence rate for nonpeach Ranch Hands than Original Comparisons (Adj. RR: 0.20, 95% C.I.: [0.02,1.80], p=0.150), but a higher incidence rate for peach toned Ranch Hands than Original Comparisons (Adj. RR: 1.70, 95% C.I.: [0.95,3.04], p=0.073). The group-by-sun reaction index interaction (verified and suspected) was again due to Ranch Hands who react moderately to the sun having a higher incidence rate than similar Original Comparisons (Adj. RR: 2.74, 95% C.I.: [1.14,6.63], p=0.025). Interval Systemic Neoplasms As shown in Table 10-1, eight Ranch Hands (0.8%) and seven Comparisons (0.5%) had verified malignant systemic neoplasms in the interval between the Baseline and followup examinations. When suspected malignant systemic neoplasms were included, the numbers were 12 Ranch Hands (1.2%) and 12 Comparisons ( . % . The proportions of malignancies among the systemic neoplasms 09) of all types (malignant, benign, uncertain) were similar in the two groups: 14.5 percent (8/55) for Ranch Hands and 11.5 percent (7/61) for Comparisons (p=0.783). Inclusion of suspected conditions did not change the conclusion from this contrast: 18.5 percent (12/65) Ranch Hands versus 15.0 percent (12/80) Comparisons (p=0.656). For the remainder of this section, only malignant (verified and suspected) systemic neoplasms occurring in the Baseline to followup interval are analyzed. These occurrences were distinct from those reported at Baseline. No new metastatic systemic neoplasms were reported in the interval. Interval Malignant Systemic Neoplasms Table 10-8 shows the sites of the new malignant neoplasms reported by the eight Ranch Hands and seven Comparisons. Classification of malignancies was based on I CD-9 with special coding for tumor type as well as site, thus avoiding problems of underreporting of STS. Six Ranch Hands and five Comparisons had suspected systemic neoplasms in this interval (Table 10-9), making a total of 12 in each group, since 2 Ranch Hands with verified systemic neoplasms also had suspected systemic neoplasms. The frequencies were too small for indepth analysis of individual sites. Table 10-8 shows that two Ranch Hands had malignant neoplasms of the oral cavity and pharynx versus no Comparisons, and three Comparisons but no Ranch Hands had malignant neoplasms of the colon. For all digestive system malignancies (esophagus plus colon), there were four occurrences among Comparisons but none among Ranch Hands. The analyses that follow are based on the combination of all interval malignant systemic neoplasms regardless of specific site, both verified and verified plus suspected. Table H-9 of Appendix H lists the malignancy sites for the eight Ranch Hands and the six Original Comparisons in the Baseline-followup interval. 10-24 �TABLE 10-8. Summary of Follovup Participants with Verified Malignant Systemic Neoplasms in Baseline-Follovup Interval by Group Group Site Ranch Hand Comparison Total Oral Cavity and Pharynx 2a'b 0 2 Thyroid Gland 0 1 1 Esophagus 0 lc 1 Bronchus and Lung 1 0 1 Colon 0 3d'8 3 Kidney and Bladder 2 1 3 Prostate 1 1 2 Testicles 1 0 1 Connective and Other Soft Tissue 1 0 1 8 7 15 Total a lncludes one Ranch Hand with separate malignancies of tongue and epiglottis and also malignant neoplasm of bone. b lncludes one Ranch Hand with separate malignant neoplasms of tongue and oropharynx and secondary malignant neoplasm of other site. c Also has malignant neoplasm of bone. d lncludes one Comparison with secondary malignant neoplasms of liver and bone and bone marrow. 8 Includes one Comparison with secondary malignant neoplasm of liver. 10-25 �TABLE 10-9. Summary of Follovup Participants with Suspected Malignant Systemic Neoplasms at Physical Examination by Group Group Site Bronchus and Lung Rectum 4 a,b 0 c Comparison Total 2 6 1 Ranch Hand 1 Liver l 0 1 Prostate 0 1 1 Lymphatic and Hematopoietic Tissue ld 0 1 Unspecified Site 0 1 1 6 5 11 Total "includes one Ranch Hand with a suspected maglignant neoplasm of either lung, mediastinum, esophagus, or ill-defined site within digestive organs and peritoneum. Includes one Ranch Hand with a suspected secondary malignant neoplasm of lung. c Not specified as primary or secondary. Suspected as either Hodgkins disease, leukemia, or lymphoma. 10-26 �There is no parallel table for suspected malignant systemic neoplasms since the five Comparisons with suspected conditions in Table 10-9 are Original Comparisons. Unadjusted Analyses As shown in Table 10-10, the unadjusted group contrast for all verified malignant systemic neoplasms was not significant (p=0.603), with an estimated relative risk of 1.46 (95% C.I.: [0.53,4.03]). When suspected malignant neoplasms were included with the verified malignancies, the estimated relative risk was 1.28 (95% C.I.: [0.57,2.85]), and was also not significant (p=0.680). A parallel unadjusted analysis for Ranch Hands versus Original Comparisons gave similar nonsignificant results (Appendix Table H-10). Covariates The covariates considered for the adjusted analysis of all interval malignant systemic neoplasms combined were age, race, occupation, smoking and drinking history, exposure to the groups of carcinogens, exposure to the individual carcinogens, and the composite carcinogen exposure variable as listed in Table 10-5. The categories used for age, pack-years, and drinkyears were the same. Age was used as a continuous variable in the adjusted analyses but was categorized for ease of presentation in the report. No Blacks had verified systemic neoplasms, but in contrast to the skin cancer analysis, Blacks were retained in the analysis. Covaciate Associations Table 10-11 summarizes the results of chi-square tests of association between the incidence rate of all malignant systemic neoplasms combined and the covariates considered for use in the adjusted analyses. Details of the covariate relationships are given in Appendix H, Table H-ll. There was a significant increase in the incidence rate of all verified interval malignant systemic neoplasms with increasing age (p<0.001) and a marginally significant difference among occupations (p-0.056). The incidence rates for officers, enlisted flyers,, and enlisted groundcrew were 1.2 percent, 0.5 percent, and 0.3 percent, respectively. There was a marginally significant association with total lifetime alcohol consumption (p=0.082). The test for differences in incidence rates among pack-year levels of smoking was not significant (p=0.220), although an increasing trend was apparent. Some of the occupation effect may be attributable to confounding with age. There was a significant negative association with insecticide exposure for verified malignant systemic neoplasms (p=0.014). Table H-ll of Appendix H shows that there were a few significant or marginally significant positive associations with individual carcinogens: e.g., with naphthylamine (p=0.050), benzidine (p=0.088), and coal tar (p=0.079). However, in many instances the self-reported exposure frequencies were very small. 10-27 �TAKE 10-10. Unadjusted Analyses of Followup Participants with Verified and Suspected Malignant Systemic Neoplasms in the Baseline-Followup Interval by Group Status Verified Verified & Suspected Statistic Number of Participants/% Total Neoplasms Number of Participants/% Total Neoplasms Group Ranch Hand Comparison 8 0.8% 7 0.5% 12 12 0 9 .% 23 16 p-Value 1.46 ( . 3 4 0 ) 05,.3 0.603 1.28 ( . 7 2 8 ) 05,.5 060 .8 10 12 1.2% Est. Relative Risk ( 5 C.I.) 9% 10-28 �TABLE 10-11. Summary of Associations Between Incidence Rates of All Malignant Systemic Neoplasms Combined and the Covariates in the Baseline-Followup Interval for Combined Followup Ranch Hand and Comparison Groups Covariate Age Verified p-Value <0.001 Verified & Suspected p-Value 0.001 Race NS NS Occupation NS* NS Lifetime Cigarette Smoking NS NS Lifetime Alcohol Consumption NS* NS Exposure to Carcinogens/Groups of Carcinogens: Set la Asbestos Non-medical X Rays Industrial Chemicals Herbicides Insecticides Degreasing Chemicals Set 2b Anthracene Arsenic Benzene Benzidene Chromates Coal Tar Creosote Aminodiphenyl Chloromethyl Ether Mustard Gas Naphthylamine Cutting Oils Trichloroethylene Ultraviolet Light (not sun) Vinyl Chloride Composite Carcinogen Exposure NS NS NS NS 0.014 NS NS 0.049 NS NS NS* NS NS NS NS NS* NS NS* NS NS NS NS 0.050 NS NS NS NS NS NS* NS NS NS NS NS NS* 0.023 NS* 0.019 NS NS NS NS NS NS NS*; Borderline significant (0.05<p<0.10). NS: Not significant (p>0.10) a Questionnaire data. b AFHS Form 2. 10-29 �The covariate associations for verified plus suspected malignant systemic neoplasms were similar to those for verified only. The association with occupation was no longer significant (p=0.193), and there was a significant positive association with nonmedical x-ray exposure (p=0.049). There were some significant and marginally significant positive associations with individual carcinogens: with naphthylamine (p=0.019), chloromethyl ether (p=0.023), arsenic (p=0.069), mustard gas (p=0.090), and aminodiphenyl (p=0.061) (see Appendix H, Table H-ll). The covariates used for the adjusted group contrast of the incidence rate of all malignant systemic neoplasms were race, age (continuous), occupation, and pack-years. Adjusted Analyses The adjusted relative risks for all verified and verified plus suspected malignant systemic neoplasms are presented in Table 10-12. For verified malignant systemic neoplasms, there was no significant difference between groups (Adj. RR: 1.51, 95% C.I.: [0.54,4.22], p=0.434). Age made a significant contribution to the adjustment (p<0.001). Parallel results for Ranch Hands contrasted with Original Comparisons are given in Table H-12 of Appendix H. A significant group-by-occupation interaction was found in the adjusted analysis of verified plus suspected malignant systemic neoplasms (p=0.027). This was due to significantly more cases of malignant systemic neoplasms among Ranch Hand enlisted flyers than among corresponding Comparisons (4/175 [2 verified, 2 suspected] versus 0/209, Fisher's exact test=0.042), whereas the incidence rate for officers was lower (but not significantly) for Ranch Hands than for the corresponding Comparisons, and equivalent for the enlisted groundcrew (see Table H-13 of Appendix H). Age (p<0.001) and a race-by-packyear interaction (p=0.035) made significant contributions to the adjustment. Comparable results were found for the contrast of Ranch Hands with the Original Comparisons (see Tables H-12 and H-14 of Appendix H). Lifetime (Baseline and Interval) Data from the Baseline and followup examinations were merged to obtain records of the lifetime history of neoplasm incidence for those followup participants who participated at Baseline. New participants provided lifetime information at the followup examination. Neoplasms prior to service in Southeast Asia were excluded from all analyses. All data from the Baseline study have been verified, but as described in the previous section, the status of some suspected interval neoplasms remains unclear, and thus both verified and verified plus suspected neoplasms are described and analyzed in this section. Table 10-13 shows that 21.3 percent (216/1,016) of Ranch Hands and 16.2 percent (209/1,293) of Comparisons had skin or systemic neoplasms of some type (malignant, benign, and uncertain). The group difference in incidence rates was significant (p=0.002), with an estimated relative risk of 1.40 (95% C.I.i [1.13,1.73]). When suspected neoplasms were included, the contrast was less marked (22.7% [231] of Ranch Hands versus 19.3% [249] of Comparisons) but still statistically significant (p=0.044), with an estimated relative risk of 1.23 (95% C.I;: [1.01,1.51]). 10-30 �TABLE 10-12. Adjusted Analyses of Followup Participants for the Incidence of All Malignant Systemic Neoplasms During the Baseline-Followup Interval Variable Malignant Systemic Neoplasms Adj. Relative Risk (95% C.I.) 1.51 (0.54,4.22) p-Value Covariate Remarks 0.434 AGE (p<0.001) **** GRP*OCC (p=0.027) AGE (p<0.001) RACE*PACKYR (p=0.035) (Verified) Malignant Systemic Neoplasms (Verified & Suspected) **** ****Group-by-covariate interaction—adjusted relative risk, confidence interval, and p-value not presented. 10-31 �TABLE 10-13. Unadjusted Analyses of Follovup Participants with lifetime Occurrence of Verified and Suspected Neoplasms by Group (Nonblacks and Blacks) Group* Banch Hand Site Skin Systemic All Neoplasm Behavior and Status Comparison Number** Percent Number** Percent Malignant Verified 66 Verified and Suspected 75 Benign Verified 84 Verified and Suspected 86 Uncertain Behavior and Unspecified Nature: Verified 1 Verified and Suspected 1 Any Skin Neoplasm* Verified 150 Verified and Suspected 159 Total** p-Value*** 6.5 7.4 66 85 5.1 6.6 132 160 0.175 048 .5 8.3 8.5 79 85 6.1 6.6 163 171 009 .4 0.093 0.1 0.1 1 1 0.1 0.1 2 2 099 .9 099 .9 14.8 15.7 140 165 10.8 12.8 290 324 0.005 0.053 17 21 1.7 2.1 17 22 1.3 1.7 34 43 0.491 .0.538 51 57 5.0 5.6 64 75 5.0 5.8 115 132 0.999 0.857 15 15 1.5 1.5 14 18 1.1 1.4 29 33 0.453 0.862 81 91 8.0 9.0 87 106 6.7 8.2 168 197 0.259 0.548 Malignant, Benign, Uncertain Behavior, Unspecified Nature Verified 216 Verified and Suspected 231 21.3 22.7 209 249 16.2 19.3 425 480 0.002 004 .4 Malignant Verified Verified and Suspected Benign Verified Verified and Suspected Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected Any Systemic Neoplasm13 Verified Verified and Suspected *Sample sizes: 1,016 Ranch Hands, 1,293 Comparisons. **Number of participants. ***Fisher's exact test. ^Participant has one or more malignant, benign, or unspecified skin neoplasm. Participant has one or more malignant, benign, or unspecified systemic neoplasm. c Participant has one or more malignant or benign skin or systemic neoplasm. 10-32 �Table H-15 of Appendix H is parallel to Table 10-13 for Ranch Hands and Original Comparisons only, Lifetime Skin Neoplasms As seen in Table 10-13, 69.4 percent (150/216) of Ranch Hands with neoplasms had skin cancer; the corresponding percentage for Comparisons was 67.0 percent (140/209). The group difference in these proportions was not significant (p=*0.604). This contrast, when suspected neoplasms were included, was 68.8 percent (159/231) versus 66.3 percent (165/249), which again was not significant (p=0.560). The overall percentage of Black and nonblack Ranch Hands with verified lifetime skin neoplasms of any type was 14.8 percent (150/1,016), versus 10.8 percent (140/1,293) for Comparisons. No Black followup participants had ever had skin neoplasms, nor did any Baseline Black participants. The overall percentage of nonblack Ranch Hands with skin neoplasms of any type was 15.7 percent (150/956) and was significantly (p=0.006) greater than that of the Comparisons with 11.6 percent (140/1,210). The estimated relative risk was 1.42 95% C.I.: [1.11,1.82]). When both verified and suspected neoplasms were in the analysis, the contrast was marginally significant (p=0.060): Ranch Hands 16.6 percent (159/956) versus Comparisons with 13.6 percent (165/1,210) (Estimated RR: 1.26, 95% C.I.: [1.00,1.60]). For the remainder of this subsection, only malignant skin neoplasms are examined. Furthermore, the analysis was restricted to nonblacks. The dependent variables examined were the same as those of the previous section (basal cell carcinoma, melanoma, squamous cell carcinoma, all malignant skin neoplasms combined and sun exposure-related skin malignancies). Lifetime Malignant Skin Neoplasms Table 10-14 presents the unadjusted analyses of the frequencies of nonblack participants in each group with lifetime occurrences of basal cell carcinoma, squamous cell carcinoma, melanoma, all malignant skin neoplasms, and the sun exposure-related skin malignancies. For completeness, the total number of malignancies of each type is also given. Table H-16 of Appendix H presents parallel analyses for Ranch Hands and Original Comparisons. Unadjusted Analyses There was a higher relative frequency (5.5%) of Ranch Hands who had basal cell carcinomas than of Comparisons (4.1%), but the difference was not significant (p=0.128). The estimated relative risk was 1.36 (95% C.I.: [0.92,2.02]). With the inclusion of suspected basal cell carcinoma, the estimated relative risk was also not significant (p=0.579). Of the 53 Ranch Hands with verified basal cell carcinomas, 17 (32.1%) had 2 or more occurrences. The corresponding number for the Comparisons was 14/50 (28.0%). The group contrast of the percentages with multiple basal cell carcinomas versus no basal cell carcinomas was not significant (17/920 versus 14/1,174, p=0.274), nor was the corresponding contrast when suspected basal cell carcinomas were included (19/916 versus 16/1,159, p=0.234). 10-33 �TABLE 10-14. Unadjusted Analyses of Nonblack Followup Participants vith Lifetime Occurrence of Verified and Suspected Malignant Skin Neoplasms by Cell Type and Group Group* Cell Type Status Statistic** Ranch Hand Comparison Est. Relative Risk (95% C.I.) p-Value Verified Number/% Total Neoplasms 53 77 5.5% 50 76 4.1% 1.36 ( .92,2. 02) 0.128 0 Verified & Suspected Number/% Total Neoplasms 59 88 6.2% 67 99 5.5% 1.12 ( .78,1. 61) 0.579 0 Verified Number/% Total Neoplasms 4 6 0.4% 6 7 0.5% 0.84 ( .24,3. 00) 0.999 0 Verified & Suspected Number/% Total Neoplasms 4 6 0.4% 7 8 0.6% 0.72 ( .21,2. 47) 0.764 0 Verified Number/% Total Neoplasms 5 6 0.5% 5 6 0.4% 1.27 ( .37,4. 39) 0.757 0 Verified & Suspected Number/% Total Neoplasms 5 6 0.5% 8 10 0.7% 0.79 ( .26,2. 42) 0.784 0 All Malignant Verified Skin Neoplasms Verified & Suspected Number/% Total Neoplasms 66 100 6.9% 66 100 5.5% 1.29 ( .90,1.83) 0.175 0 Number/% Total Neoplasms 75 114 7.9% 85 129 7.0% 0 1.13 ( .82,1. 56) 0.508 Sun-Exposure Verified Related Malignant Neoplasms3 Verified & Suspected Number/% Total Neoplasms 59 87 6.2% 55 83 4.6% 1.38 ( . 0 ,95,2. 02) 0.100 Number/% Total Neoplasms 65 98 6.8% 74 111 6.1% 1.12 ( . 0 ,79,1.58) 0.537 Basal Cell Carcinoma Squamous Cell Carcinoma Melanoma *Number of participants—956 Ranch Hands and 1,210 Comparisons. **Number and percent of participants; total number of malignant neoplasms of specified cell type. a Basal cell carcinoma, melanoma, and malignant epithelial neoplasms NOS. �The frequencies of participants who had squamous cell carcinoma were very small: 4 Ranch Hands (0.4%) and 6 Comparisons ( . % . The estimated 05) relative risk was 0.84 (95% C.I.: [0.24,3.00]), and the contrast was far from significant (p=0.999). Inclusion of suspected squamous cell carcinoma did not change this finding. The frequency of Ranch Hands who had melanoma, 5 (0.5%), was slightly greater than that of the Comparisons, 5 (0.4%), but the contrast was not significant (p=0.757); the estimated relative risk was 1.27 (95% C.I.: [0.37,4.39]). Inclusion of suspected melanoma inverted the relative risk to 0.79, which was again not significant. This analysis had little power due to small frequencies. For sun exposure-related skin malignancies, there was a higher percentage of Ranch Hands than Comparisons (6.2% versus 4.6%), but the contrast was only of borderline significance (p=0.100); the estimated relative risk was 1.38 (95% C.I.: [0.95,2.02]). When suspected sun exposure-related skin malignancies were included, the group difference was not significant (p=0.537), with estimated relative risk 1.12 (95% C.I.: [0.79,1.58]). As in the previous section, adjusted analyses were only carried out for basal cell carcinoma and the sun exposure-related skin malignancies. Covariates The same covariates as for the interval analysis (Table 10-5) were considered for the adjusted analysis of the lifetime incidence rates of basal cell carcinoma and sun exposure-related skin malignancies: age, occupation, history of cigarette smoking and alcohol consumption, the same host factors and average latitude, and exposure to the same recognized carcinogens. The covariates used for the adjusted analyses were the same as in the interval analysis, namely age, occupation, sun reaction index, average lifetime residential latitude, and skin color. Covariate Associations Table 10-15 presents details of the associations between the incidence rate of basal cell carcinoma and the following covariates: age; occupation; pack-years of smoking, lifetime drink-years; ethnic background, hair color, skin color, eye color; skin-reaction-to-sun variables, sun-reaction index; average residential latitude, and exposure to individual carcinogens and groups of carcinogens. For the incidence of verified basal cell carcinoma, the same associations were found as in the interval analysis, namely, an increasing incidence rate with increasing age (p<0.001), a significant difference among occupations (p=0.017; officers 6.4%, enlisted flyers 4.2%, enlisted groundcrew 3.6%), and significant associations with average lifetime residential latitude (p=0.026), all the skinr-react ion-to-sun variables (p<0.001 for all), the sun-reaction index (p<0.001), and increasing total pack-years (p=0.024). There was evidence of a higher incidence rate of basal cell carcinomas among the heavy drinkers, although the test for the difference among drinking categories was not significant. 10-35 �TABLE 10-15. Association Between Lifetime Incidence of Basal Cell Carcinoma and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent p-Value 21 75 7 2.4 <0.001 6.3 8.1 24 91 11 2.7 7.6 12.6 <0.001 Officer Enlisted Flyer Enlisted Groundcrew 850 360 956 54 15 34 6.4 4.2 3.6 0.017 68 18 40 8.0 5.0 4.2 0.002 Total Lifetime Smoking (Pack- Years) 0 >0-20 >20-40 >40 616 998 391 157 32 36 21 14 5.2 3.6 5.4 8.9 0.024 37 43 31 15 6.0 4.3 7.9 9.6 0.010 Total Lifetime Alcohol Consumption (Drink- Years) 0 >0-5 >5-30 >30-100 >100 141 717 655 479 104 7 37 29 19 8 5.0 5.2 4.4 4.0 7.7 0.548 8 43 34 30 8 5.7 6.0 5.2 6.3 7.7 0.855 Ethnic Background* A 1,582 424 63 42 85 16 ,1 0 5.4 3.8 1.6 0.0 0.132 107 16 1 0 6.8 3.8 1.6 0.0 0.016 B C D Skin Color Dark Medium Pale Dark Peach Pale Peach 1 73 308 1,262 520 0 2 9 69 23 0.0 2.7 2.9 5.5 4.4 0.339 0 2 11 82 31 0.0 2.7 3.6 6.5 6.0 0.263 Born XL942 Born 1923-41 Born <1922 Occupation 03 Total Participants Number* Percent p-Value 882 1,197 87 Age o Verified and Suspected �TABLE 10-15. (continued) Association Between Lifetime Incidence of Basal Cell Carcinoma and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category -Eye Color Burns Painfully Burns Becomes Red No Reaction Skin Reaction Freckles, No Tan After Repeated Tans Mildly Exposure to Sun Tans Moderately Tans Deep Brown . 4.7 6.4 2.5 5.4 4.2 0.338 35 30 6 7 48 5.4 6.6 5.0 7.5 5.7 0.853 20 42 32 2 7 4.6 4.1 5.7 12.5 6.5 0.278 24 53 38 3 8 5.5 5.1 6.8 18.8 7.4 0.120 43 60 3.8 5.9 0.026 51 75 247 429 805 681 Skin Reaction Burns to First 30 Min. Usually Burns of Sun Exposure Burns Mildly Rarely Burns 30 29 3 5 36 1,136 1,022 >37° Residential History (Average <37° Latitude) Skin Reaction to >2 Hrs of Sun After First Exposure Number* Percent p-Value 439 1,038 563 16 108 Black Dark Brown Light Brown Red Blond o Number* Percent p-Value 645 455 119 93 850 Brown Hazel Green Grey Blue Hair Color Total Participants Verified and Suspected 21 36 29 16 8.5 <0.001 8.4 3.6 2.4 25 44 32 24 10.1 10.3 4.0 3.5 <0.001 120 338 1,043 663 9 31 42 21 7.5 <0.001 9.2 4.0 3.2 11 33 54 28 9.2 9.8 5.2 4.2 0.001 45 314 1,019 783 4 31 37 30 8.9 9.9 3.6 3.8 5 36 47 37 11.1 11.5 4.6 4.7 <0.001 <0.001 4.5 ' 7.3 0.006 �TABLE 10-15. (continued) Association Between Lifetime Incidence of Basal Cell Carcinoma and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Sun Reaction Index Tends to Burn Mild Reaction Tends to Tan Exposures to Carcinogens Asbestos Nonmedical X Rays Total Participants Verified and Suspected Number* Percent p-Value 145 454 1,562 10 41 51 6.9 <0.001 9.0 3.3 Yes No 458 1,708 18 85 3.9 5.0 Yes No 494 1,672 29 74 Industrial Chemicals Yes No 1,126 1,040 Herbicides Yes No 1,261 905 Insecticides Yes No 1,313 853 Degreasing Chemicals Yes No Yes No Number* Percent p-Value 12 46 67 8.3 10.1 4.3 <0.001 0.389 25 101 5.5 5.9 0.822 5.9 4.4 0.187 37 89 7.5 5.3 0.080 49 54 4.4 5.2 0.365 60 66 5.3 6.4 0.314 65 38 5.2 4.2 0.357 81 45 6.4 5.0 0.164 69 34 5.3 4.0 0.181 82 44 6.3 5.2 0.303 1,261 905 60 43 4.8 4.8 0.999 72 54 5.7 6.0 0.852 489 1,653 21 80 4.3 4.8 0.716 24 100 4.9 6.1 0.379 CO 00 Composite Carcinogen Exposure ' �TABLE 10-15. (continued) Association Between Lifetime Incidence of Basal Cell Carcinoma and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Verified and Suspected o i u> Total Participants Anthracene Yes No 2 2,161 0 103 0.0 4.8 0.999 0 126 0.0 5.8 0.999 Yes No 41 2,124 4 98 9.8 4.6 0.124 5 120 12.2 5.7 0.084 Benzene Exposure to Individual Carcinogens Covariate Category Arsenic Covariate Yes No 74 2,091 6 97 8.1 4.6 0.162 7 119 9.5 5.7 0.198 Benzidine Yes No 11 2,154 1 102 9.1 4.7 0.416 1 125 9.1 5.8 0.484 Chroma tes Yes No 84 2,079 4 97 4.8 4.7 0.999 5 119 6.0 5.7 0.812 Coal Tar Yes No 68 2,097 2 101 2.9 4.8 0.770 3 123 4.4 5.9 0.795 Creosote Yes No 159 2,007 9 94 5.7 4.7 0.560 10 116 6.3 5.8 0.726 Aminodiphenyl Yes No 6 2,157 0 102 0.0 4.7 0.999 0 125 0.0 5.8 0.999 Chloromethyl Ether Yes No 21 2,142 2 101 9.5 4.7 0.264 2 124 9.5 5.8 0.348 Mustard Gas Yes No 6 2,159 0 103 0.0 4.8 0.999 0 126 0.0 5.8 0.999 Number* Percent p-Value Number* Percent p-Value �TABLE 10-15. (continued) Association Between Lifetime Incidence of Basal Cell Carcinoma and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified and Suspected Verified Total Participants Naphthylamine Yes No 52 2,112 3 99 5.8 4.7 0.734 4 121 7.7 5.7 0.540 Cutting Oils Yes No 226 1,939 12 91 5.3 4.7 0.622 15 111 6.6 5.7 0.549 Yes No 184 1,979 5 98 2.7 5.0 0.207 7 119 3.8 6.0 0.253 Ultraviolet Light Yes No 44 2,120 5 98 11.4 4.6 0.055 5 121 11.4 5.7 0.179 Vinyl Chloride Exposure to Individual Carcinogens (continued) Covariate Category Trichloroethylene Covariate 31 Yes No 2,133 0 103 0.0 4.8 0.399 1 125 3.2 5.9 0.999 o Number* Percent p-Value *Number of participants with basal cell carcinomas * a Ethnic Background: A B C D - English, Welsh, Scottish, Irish Scandinavian, German, Polish, Russian, Other Slavic, Jewish, French Spanish, Italian, Greek. Mexican, American Indian, Asian • Number* Percent p-Value �There was a significant (p<0.001) association with the sun-reaction index. Participants with the most sensitive skin had a somewhat lower rate (6.9%) of verified basal cell carcinoma lifetime than the participants in the next most sensitive category (9.0%), although the difference was not as marked as in the Baseline-followup interval. However, the rate for those who tanned easily was much lower (3.3%) than for those who did not. A marginally significant positive association was found with self-reported exposure to non-sun ultraviolet light (p=0.055). The results were similar for the verified plus suspected basal cell carcinomas. There was a significant (p=0.016) difference among ethnic backgrounds, with participants with Celtic or English backgrounds having higher incidence rates than those with other backgrounds. Further, there were marginally significant positive associations in incidence rates with nonmedical x-ray exposure (p=0.080) and arsenic (p=0.084), a recognized skin carcinogen, but the association with ultraviolet light was not significant. The details of associations between the incidence rates of verified and suspected sun exposure-related skin malignancies and the covariates are given in Appendix H, Table H-17. The significant covariates for verified conditions were age (p<0.001), occupation (p=»0.009), total pack-years (p=0.021), average latitude (p=0.026), and sun-reaction index (p<0.001). The same pattern held for verified plus suspected sun exposure-related skin malignancies. There was a marginally significant positive association with ultraviolet light exposure (p=0.078) for the verified conditions only, and with herbicide exposure (p=0.076) for the verified plus suspected conditions. The covariates chosen for the adjusted analysis were age, occupation, skin color, average lifetime residential latitude and the sun-reaction index. Adjusted Analysis The results of adjusted analyses of group contrasts for lifetime skin malignancies are given in Table 10-16, There was significant evidence of a higher incidence rate of verified basal cell carcinoma in the Ranch Hand group as contrasted with the Comparisons (p=0.035). The adjusted relative risk was 1.56 (95% C.I.: [1.03,2.37]). A sun-reaction index-by-average latitude interaction (p=0.026), a skin color-by-sun-reaction index interaction (p<0.001), and an occupation-by-age interaction (p=0.047) made significant contributions to the model. The adjustment by average residential latitude, which is greater for Ranch Hands than Comparisons, contributed to a higher relative risk resulting from the adjusted analysis than from the unadjusted (see Table 10-14). When suspected basal cell carcinomas were . included in the analysis, a significant group-by-sun-reaction index interaction (p=0.040) was found. Age (p<0.001), a skin color-by-average residential latitude (p=0.024), and a skin color-by-sun-reaction index interaction (p<0.001) made significant contributions to the adjustment. This was due to a significant increase in basal cell carcinoma incidence for Ranch Hands with an intermediate skin reaction to sun over similar Comparisons (Adj. RR: 1.97, 95% C.I.: [1.04,3.73], p=0.038) (Appendix H, Table H-18). Similar results were found in the contrast of Ranch Hand versus Original Comparisons (Table H-19). Namely, for verified basal cell carcinoma, and for verified plus suspected basal cell carcinomas, significant group-by-sunreaction index interactions were found (p=0.010 and p=0.003, respectively [see Table H-20 for additional details on the interactions]). 10-41 �TABLE 10-16. Adjusted Analyses of Nonblack Followup Participants for Lifetime Malignant Skin Neoplasm Incidence Variable Status Adj. Relative Risk (95% C.I.) Basal Cell Carcinoma Verified 1.56 (1.03,2.37) 0.035 SKIN*SUNREAC (p<0.001) OCC*AGE (p=0.047) SUNREAC*LAT (p=0.026) **** **** AGE (p<0.001) GRP*SUNREAC (p=0.040) SKIN*LAT (p=0.024) SKIN*SUNREAC (P<0.001) 1.54 (1.04,2.29) 0.030 AGE (p<0.001) SKIN*LAT (p=0.016) SKIN*SUNREAC (p<0.001) 1.23 (0.86,1.77) 0.252 AGE (p<0.001) SKIN*LAT (p=0.013) SKIN*SUNREAC (p<0.001) Verified & Suspected Malignant Verified Sun-Exposure Skin Neoplasms Verified & Suspected p-Value Covariate Remarks ****Group-by-covariate interaction—adjusted relative risk, confidence interval, and p-value not presented. 10-42 �As shown in Table 10-16, there was a significantly higher incidence rate of sun exposure-related skin malignancies among Ranch Hands as contrasted with Comparisons (Adj. RR: 1.54, 95% C.I.: [1.04,2.29], p=0.030). Significant contributions were noted for age (p<0.001), a skin color-by-sunreaction index interaction (p<0.001), and an average latitude-by-skin color interaction (p=0.016). When suspected sun exposure-related skin malignancies were included in the analysis, the adjusted relative risk became 1.23 (95% C.I.: [0.86,1.77]) and was no longer significant (p=0.252). Age (p<0.001), a skin color-by-sun-reaction index interaction (p<0.001), and average latitude-by-skin color interaction (p=0.013) contributed significantly to the adjustment. When Ranch Hands were contrasted to Original Comparisons, significant group-by-sun reaction index interactions were found for verified, and verified plus suspected, sun-exposure related skin neoplasms (p=0.045,p=0.016, respectively). These interactions were due to significant relative risks for those participants with intermediate reactions of skin to sun, as was also found for basal cell carcinomas only (see Appendix Tables H-19 and H-20 for details). Lifetime Systemic Neoplasms Table 10-13 shows that 81 (8.0%) Ranch Hands and 87 (6.7%) Comparisons had a verified history of systemic neoplasms of any type (malignant, benign, or uncertain). The estimated relative risk was 1.20 (95% C.I.: [0.88,1.65]), and was not significant (p=0.259). With the inclusion of suspected systemic neoplasms, the frequencies were 9.0 percent (91/1,016) for Ranch Hands and 8.2 percent (106/1,293) for Comparisons, with an estimated relative risk of 1.10 (95% C.I.: [0.82,1.48]), and the contrast was also not significant (p=0.548). For Ranch Hands with systemic neoplasms of any type, the percentage with malignant neoplasms was 21.0 percent (17/81) and the corresponding rate for Comparisons was 19.5 percent (17/87), a nonsignificant group difference (p=0.849). Including suspected systemic malignancies, these frequencies were 23.1 percent (21/91) for Ranch Hands and 20.8 percent (22/106) for Comparisons. Again, the group difference was not significant (p=0.731). For the remainder of this section, only malignant systemic neoplasms are discussed. Lifetime Malignant Systemic Neoplasms Table 10-17 presents the frequencies of verified lifetime malignant systemic neoplasms by site. Three Ranch Hands versus no Comparisons had malignant neoplasms of the oral cavity and pharynx; these occurred at ages 45, 52, and 57. The group difference in incidence rate was marginally significant (p=0.085). No Ranch Hands but 5 Comparisons had malignant neoplasms of the colon; the group difference in incidence rate was also marginally significant (p=0.072). Three Ranch Hands but no Comparisons had testicular malignancies, but the.group difference in incidence rates was only marginally significant (p=0.085). These occurred at ages 35, 38, and 54. The suspected malignant neoplasms are listed in Table 10-9. Table H-21 of Appendix H gives a list of verified lifetime malignant systemic neoplasms for Ranch Hands and Original Comparisons. 10-43 �TABLE 10-17. Summary of Follovup Participants With Lifetime Incidence of Verified Malignant Systemic Neoplasms by Group Group Site Eye Oral Cavity and Pharynx Ranch Hand Comparison Total 1 0 1 0 3 3 a,b Larynx 0 1 1 Thyroid Gland 0 2 2 Esophagus 0 lc 1 Bronchus and Lung 2 0 1 Colon 0 Kidney and Bladder 4 3 7 Prostate 2 2 4 Testicles 3 0 3 Connective and Other Soft Tissue 1 1 2 Hodgkin's Disease 0 1 1 Ill-Defined Sites 1£ lg 2 Total 5d.e 17 17 5 34 a Includes one Ranch Hand with separate malignancies of tongue and epiglottis and also malignant neoplasm of bone. Includes one Ranch Hand with separate malignant neoplasms of tongue and oropharynx and secondary malignant neoplasm of other site. °Also has malignant neoplasm of bone. d lncudes one Comparison with secondary malignant neoplasms of liver and bone and bone marrow. 8 Includes one Comparison with secondary malignant neoplasm of liver. £ Malignant neoplasm of thorax. Malignant neoplasm of face, head, or neck. 9 10-44 �One Ranch Hand and one Comparison had neoplasms of connective and other soft tissue. The Comparison had a fibrosarcoma at age 28 (reported at Baseline) and the Ranch Hand participant had malignant fibrous histiocytoma at age 63 (reported at followup). Both of these conditions are classified as soft tissue sarcoma. Since soft tissue sarcoma and malignant neoplasms of the lymphatic system are of concern in this study, the occurrences of these malignancies are shown by group in Table 10-18. The occurrences of these four malignancies are too small to support further statistical analysis. TABLE 10-18. Summary of Followup Participants with Lifetime Soft Tissue Sarcoma, Leukemia or Lymphoma by Group Group Site Ranch Hand Comparison Verified Soft Tissue Sarcoma 1 1 Verified Hodgkin's Disease 0 1 Suspected Leukemia, Hodgkin's Disease, or non-Hodgkin's Lymphoma 1 0 Unadjusted Analysis Table 10-19 shows the results of unadjusted analyses of the frequencies of participants in each group with verified or verified plus suspected malignant systemic neoplasms combined. The estimated relative risk for all malignant systemic neoplasms was 1.28 (95% C.I.: 0.65,2.51) and was not significant (p=0.491). With the inclusion of suspected malignant, neoplasms, the estimated relative risk was 1.22 (952 C.I.: 0.67,2.23) and was also not significant (p=0.538). Similar nonsignificant results were found for Ranch Hands contrasted with Original Comparisons (see Table H-22 of Appendix H). Covariates The same covariates used for the interval history of malignant systemic neoplasms were used for the adjusted analysis of lifetime malignant systemic neoplasms, namely, age, race, occupation, history of cigarette smoking and alcohol consumption, and exposure to carcinogens. Total smoking and alcohol consumption were estimated up to the followup examination, and may be different if estimated only up to the year of diagnosis of a neoplasm (if any). Further, age at followup rather than age at diagnosis was used in the analysis. 10-45 �TAKE 10-19. Unadjusted Analyses of lifetime Incidence Bates of ALL Malignant Systemic Neoplasms Combined, by Group Group Status Statistic Verified Number of Participants/% Total Neoplasms Verified & Suspected Participants/% Total Neoplasms Ranch Hand Comparison Est. Relative Risk (95%C.I.) p-Value 17 1.7* 17 1.3* 1.28 (0.65, 2.51) 0 4 1 .9 25 22 Number of 21 2.1% 36 22. 1 7 .* 27 1.22 ( . 7 2 2 ) 06,.3 0.538 Covariate Associations Associations between the incidence rate of all malignant systemic neoplasms combined and the covariates are presented in Table 10-20. For verified malignant systemic neoplasms, strong associations were found with increasing age (p<0.001) and occupation (officers 2.3%, enlisted flyers 1.3%, and enlisted groundcrew 0.9%, p=0.028). These same associations were also found for verified plus suspected systemic malignancies. The association with smoking history was not significant, either for verified or for verified plus suspected malignancies. The incidence rate of all malignant systemic neoplasms increased marginally significantly (p=0.073) with increasing levels of total lifetime alcohol consumption. For verified plus suspected malignancies, the difference among drink-year categories was- also marginally significant (p=0.080). No significant association was found with the composite carcinogen exposure variable. A significant association was found between the incidence of verified malignant systemic neoplasms and naphthylamine (p=0.048). There was a significant positive association between the verified plus suspected conditions and naphthylamine (p=0.019), and a marginally significant association with chloromethyl ether (p=0.067). The covariates used for the adjusted analysis of the incidence of malignant systemic neoplasms were race, age (continuous), occupation, pack-years, drink-years, and the composite carcinogen-exposure variable, Adjusted Analysis Table 10-21 shows that, in the adjusted analysis of the group contrast in incidence of all systemic malignancies combined, there was a significant group-by-occupation interaction (p=0.023). This was due to a difference in rates for the enlisted flyers, 5.Ranch Hands versus 0 Comparisons (unadjusted p-value=0.019), whereas the incidence rates for officers and enlisted groundcrew did not differ significantly between groups (p=0.698 and 0.922, respectively) (Table H-23). Age made a significant'contribution to the adjustment. When suspected systemic malignancies were combined with the verified systemic malignancies, a group-by-occupation interaction (p=0.002) 10-46 �TABLE 1 - 0 02. Association Between Lifetime Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for Combined Followup Ranch Hand and Comparison Participants Verified Total Participants Number* Percent Verified and Suspected Covariate Category p-Value Number* Percent p-Value Age Born >1942 Born 1923-41 Born <1922 961 1,261 87 4 24 6 0.4 1.9 6.9 <0.001 7 30 6 0.7 2.4 6.9 <0.001 Race . Nonblack Black 2,166 143 34 0 1.6 0.0 0.267 42 1 1.9 0.7 0.517 Occupation Officer Enlisted Flyer Enlisted Groundcrew 864 387 1,058 20 5 9 2.3 1.3 0.9 0.028 23 7 13 2.7 1.8 1.2 0.069 Total Lifetime Smoking (Pack- Years) 0 >0-20 >20-40 >40 658 1,081 406 158 6 15 9 4 0.9 1.4 2.2 2.5 0.237 8 20 11 4 1.2 1.9 2.7 2.5 0.324 Total Lifetime Alcohol Consumption (Drink- Years) 0 >0-5 >5-30 >30-100 >100 151 760 703 508 108 1 7 8 11 4 0.7 0.9 1.1 2.2 3.7 0.073 2 10 10 13 5 1.3 1.3 1.4 2.6 4.6 0.080 �TABLE 10-20. (continued) Association Between Lifetime Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for Combined Follovup Ranch Hand and Comparison Participants Verified Total Participants Number* Percent Verified and Suspected p-Value Covariate Exposures to Carcinogens Asbestos 499 1,810 Yes No 5 29 1.0 1.6 0.405 7 36 1.4 2.0 0.459 Nonmedical X Rays 541 1,768 Yes No 9 25 1.7 1.4 0.684 14 29 2.6 1.6 0.150 Industrial Chemicals o Category 1,199 1,110 Yes No 14 20 1.2 1.8 0.229 20 23 1.7 2.1 0.539 Herbicides 1,339 970 Yes No 18 16 1.3 1.7 0.601 23 20 1.7 2.1 0.538 Insecticides 1,389 920 Yes No 17 17 1.2 1.9 0.223 23 20 1.7 2.2 0.432 Degreasing Chemicals 1,343 966 Yes No 18 16 1.3 1.7 0.600 26 17 1.9 1.8 0.876 519 1,762 Yes No 7 27 1.4 1.5 0.999 8 34 1.5 1.9 0.711 00 Composite Carcinogen Exposure Number* Percent p-Value �TABLE 10-20. (continued) Association Between Lifetime Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for Combined Followup Ranch Band and Comparison Participants Verified and Suspected Verified Total Participants Number* Percent p-Value Number* Percent p-Value Covariate Exposure to Individual Carcinogens Anthracene 2 2,303 Yes No 0 34 0.0 1.5 0.999 0 43 0.0 1.9 0.999 Arsenic 42 2,266 Yes No 0 34 0.0 1.5 0.999 2 41 4.8 1.8 0.183 Benzene o i Category 83 2,225 Yes No 2 32 2.4 1.4 0.348 2 41 2.4 1.8 0.666 Benzidine 14 2,293 Yes No 1 33 7.1 1.4 0.188 1 41 7.1 1.8 0.227 Chroma tes 88 2,218 Yes No 2 32 2.3 1.4 0.375 2 41 2.3 1.9 0.679 Coal Tar 73 2,235 Yes No 2 32 2.7 1.4 0.292 2 41 2.7 1.8 0.397 Creosote 164 2,145 Yes No 2 32 1.2 1.5 0.999 4 39 2.4 1.8 0.543 Aminodiphenyl 6 2,300 Yes No 0 34 0.0 1.5 0.999 1 42 16.7 1.8 0.107 Chloromethyl Ether 23 2,282 Yes No 1 33 4.4 1.5 0.291 2 41 8.7 1.8 0.067 Mustard Gas 9 2,299 Yes No 0 34 0.0 1.5 0.999 1 42 11.1 1.8 0.156 �TABLE 10-20. (continued) Association Between Lifetime Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for Combined Follovup Ranch Hand and Comparison Participants Verified Total Participants Covariate Category Exposure to Individual Carcinogens (continued) Naphthylamine 56 2,251 Yes No 3 31 5.4 1.4 0.048 4 39 7.1 1.7 0.019 Cutting Oils 243 2,065 Yes No 5 29 2.1 1.4 0.396 7 36 2.9 1.7 0.209 Trichloroethylene o I iji o Number* 200 2,106 Yes No 5 29 2.5 1.4 0.211 6 37 3.0 1.8 0.264 Ultraviolet Light 51 2,256 Yes No 1 33 2.0 1.5 0.535 1 42 2.0 1.9 0.621 Vinyl Chloride 33 2,273 Yes No 0 34 0.0 1.5 0.999 1 42 3.0 1.9 0.465 *Number of participants with malignant systemic neoplasms. Percent Verified and Suspected p-Value Number* Percent p-Value �TABLE 10-21. Adjusted Analyses for Lifetime Incidence of All Malignant Systemic Neoplasms Combined Variable Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks Systemic Malignancies (Verified) **** **** GRP*OCC (p=0.023) AGE (p<0.001) Systemic Malignancies (Verified & Suspected) **** **** GRP*OCC (p=0.002) AGE (p<0.001) RACE*PACKYR (p=0.032) ****Group-by-covariate interaction—adjusted relative risk, confidence interval, and p-value not presented. was also found; this was also due to the high rates for the Ranch Hand enlisted flyers. Comparison of Baseline, Interval, and Lifetime Results Table 10-22 compares the unadjusted and adjusted contrasts from the Baseline report with those from the Baseline-followup interval and the whole post-SEA period, for the incidence of all verified malignant skin neoplasms combined, verified basal cell carcinomas, and all verified malignant systemic neoplasms combined. There were, of course, differences in the Baseline and followup cohorts, but there was a sufficiently large overlap to make such a comparative tabulation useful. Malignant Skin Neoplasms The significant relative risks for all malignant skin neoplasms seen at Baseline were not evident for the Baseline-followup interval. However, for lifetime basal cell carcinoma, a significant adjusted group contrast was found (p=0.035). The difference in the incidence rates of all skin neoplasms and in basal cell carcinomas only between the Ranch Hands and the Comparisons appears to have decreased over time, as evidenced by the fact that the interval estimated and adjusted relative risks were closer to 1 than those for the lifetime, i.e., interval plus Baseline period. Malignant Systemic Neoplasms The unadjusted group contrasts in incidence rates of all malignant systemic neoplasms combined were not significant for Baseline, for the Baseline-followup interval, or for lifetime (Baseline plus interval), nor was the adjusted group contrast for the Baseline-followup interval. The 10-51 �TABLE 10-22. Uhadjusted and Adjusted Analyses of the Incidence of ML Verified Malignant Skin and Systemic Neoplasms and Basal Cell Carcinoma: Baseline, Baseline-Followup Interval, and lifetime Occurrence Site Statistic All Malignant Number of Participants Skin with Neoplasms/Percent:0 Ranch Hand Neoplasms Comparison Baseline-Followup Interval Baseline* 35 25 Est. RR/p-Value 3.3% 2.0% (.7d 00) 1.62 Adj. RR/p-Value Basal Cell Carcinoma 31 21 Est. RR/p-Value Est. RR/p-Value Adj. RR/p-Value * 29 30 * 3.0% 2.5% 66 66 1.29 —* 53 50 6.9% 5.4% (.7) 0159 —* 5.5% 4.1% (.4) 007d 1.71 1.23 (.2) 0499 1.36 (.2) 0189 * **** **** 1.56 (.3) 005 8 7 0.8% 0.5% 17 17 * Adj. RR/p-Value All Malignant Number of Participants Systemic with Neoplasms/Percent :f Ranch Hand Neoplasms Comparison 3.0% 1.7% 3.9% 3.3% (.8) 0466 1.18 * —* Number of Participants with Neoplasms/Percent:0 Ranch Hand Comparison 37 40 Lifetime Occurrence 1.2% 0.9% 13 11 (.6d 04) 1.35 1.46 (.0) 0639 * 1.51 (.3) 044 * 1.28 **** 1.7% 1.3% (.9) 0419 **** 'Analysis not done a Baseline participants: 1 0 5 Ranch Hands, 1,224 Comparisons. ,4 b Followup participants: 1 0 6 Ranch Hands, 1,293 Comparisons. ,1 c Nonblacks only for followup participants ( 5 Ranch Hands, 1,210 Comparisons), both nonblacks and 96 Blacks for Baseline participants. d Chi-square test. "Fisher's exact test. f All participants. ****Grc>up-by-covariate interaction. 10-52 �estimated lifetime relative risk appears closer to 1 than for the two intervals separately, but the small number of occurrences and intervening mortality preclude more definitive statements. Baseline Participants This brief section summarizes the mortality and malignant neoplasm history of the fully compliant Baseline participants in the interval up to the followup examination. Mortality information up through the end of 1985 was considered. This discussion is directed to the question of whether competing mortality affected the preceding analysis of incident cancers among living participants. • Of the 1,045 Ranch Hands and 1,224 Comparisons who were fully compliant at Baseline, 971 Ranch Hands and 1,139 Comparisons returned to the followup examination. Table 10-23 presents the numbers of Baseline participants according to whether they completed the followup examination and whether they were alive at the end of 1985. TABLE 10-23. Fully Compliant Baseline Participants by Status at Followup Examination and Group Participated in Followup Examination Group Status Ranch Hand Comparison Total 2 1,137 5 2,105 Y e s Dead" Alive 3 968 No Dead Alive 9 65 15 70 24 135 1,045 1,224 2,269 Total a Died in 1985, but subsequent to participation in the examination. For the participants who did not return for the followup examination, Table 10-24 shows that 2 of the 9 deaths among Ranch Hands were due to malignant neoplasms, compared with 5 of the 15 deaths among the Comparisons. One Ranch Hand who died had a malignant skin neoplasm, but this was not the primary cause of death. Among the 65 Ranch Hands who did not return for the followup examination, 5 had verified malignant neoplasms at Baseline, including 1 systemic neoplasm (of the kidney), as contrasted with 2 among 70 Comparisons .who had verified malignant (skin) neoplasms. Thus, among the 74 Ranch Hands not returning for followup, there were 8 with incident or fatal neoplasms, as compared to 7 of 85 Comparisons; the group difference was not significant (p=0.788). 10-53 �TABLE 10-24. Fully Compliant Baseline Participants Who Did Not Participate in Followup Examination by Status and Group Group Status Ranch Hand Comparison Total Dead—Primary Cause of Death: Malignant Neoplasm 2a Other Causes 7C 5 10 7 17 Lost to Followup: Verified Malignant Neoplasm at Baseline No Malignant Neoplasm at Baseline 5d 60 2e 68 7 128 "Both with lung cancer. b Three with lung cancer, one with malignant neoplasm of intestine (location unspecified), one with malignant neoplasm of an ill-defined site (face, head, or neck). c lncludes one Ranch Hand with malignant skin neoplasm. Four with malignant skin neoplasms, one with malignant systemic neoplasm (kidney). 8 Two with malignant skin neoplasms. 10-54 �For the participants who did return for the followup examination, Table 10-25 gives the frequencies and percentages of the respective group totals according to neoplasm status at Baseline and at followup. Analysis showed that there was no significant group difference (p=0.115) in the pattern of neoplasm incidence at Baseline and/or at followup. The results of this section show approximate equivalence between the groups for the disease of cancer (fatal or nonfatal) since Baseline, and in the proportions of participants with malignancies at Baseline, followup, or both. EXPOSURE INDEX ANALYSES Unadjusted and adjusted exposure index analyses were conducted within each occupational cohort of the Ranch Hand group (see Chapter 8 for details on the exposure index). Interval and lifetime occurrences of basal cell carcinomas, sun-exposure related malignant skin neoplasms, and malignant systemic neoplasms were examined. As was done in the core analyses, verified conditions and verified plus suspected malignancies were each investigated. Blacks were excluded from all malignant skin neoplasm analyses. Group contrasts in incidence rates of malignant skin neoplasms were adjusted for the covariates of age, sun reaction index, and average residential latitude. Adjusted analyses for malignant systemic neoplasms accounted for the effects of age and race. For each dependent variable, exposure level frequencies and percentages are presented in Appendix Tables H-26 and H-27, for interval and lifetime, respectively, along with the results of the unadjusted analyses. Pearson's chi-square test was used to reflect overall exposure index differences, and Fisher's exact test was used to investigate medium versus low and high versus low exposure level contrasts. Results of the adjusted analyses are presented in Tables 10-26 and 10-27, for interval and lifetime, respectively. These results are presented in the context of a main effects model containing exposure index and all adjusting covariates. Several significant or marginally significant overall results were found. None was suggestive of a strictly increasing dose response effect; in fact, most showed decreasing incidence rates with increasing exposure. Among officers, in the unadjusted interval analysis, significant or marginally significant results were found among nonblacks for verified and suspected basal cell carcinomas (overall p=0.042), sun-exposure related malignant skin neoplasms (verified: overall p=0.096, verified plus suspected: overall p=0.021), and among Blacks and nonblacks for verified plus suspected malignant systemic neoplasms (overall p=0.081). These findings were primarily due to higher percentages of malignancies in the medium exposure level than in the high or low categories for each variable (see Appendix Table H-26 for frequencies). The corresponding adjusted analyses were nonsignificant for basal cell carcinoma (overall p=0.156), verified sun-exposure malignancies (overall p=0.272), and systemic malignant neoplasms (overall p=0.109). The adjusted results were marginally significant for verified plus suspected sun-exposure malignancies (overall p=0.095). 10-55 �TABLE 10-25. Fully Compliant Baseline Participants Also in Followup Examination by Malignant Neoplasm Status Group Malignant Neoplasm at Baseline Malignant Neoplasm at Followup Ranch Hand Number Percent Comparison Number Percent Total Yes 10 1.0 15 1.3 25 No 37 3.8 28 2.5 65 Yes 36 3.7 31 2.7 67 No 888* 93.5 1,953 Yes No Total 971 a 91.5 l,065a 1,139 Includes three Ranch Hands and two Comparisons who died after followup. 10-56 2,110 �TABLE 10-26. Adjusted Exposure Index Analysis for Followup Participants for occurrence of Malignant Neoplasms in the Baseline-Followup Interval Variable Occupation Officer Basal Cell3 Carcinoma (Verified Only) Enlisted Flyer Exposure Index Low Medium High Total* Total* Total* 124 54 138 Officer Basal Cella Carcinoma (Verified and Suspected) 61 121 51 Adj. Relative Risk ( 5 C.I.) 9% p-Value Overall M vs. L H vs. L 2.02 ( . 0 8 1 ) 05,.0 0.91 ( . 8 4 6 ) 01,.8 0.415 0.320 0.908 Overall H vs. L H vs. L 0.35 (0.05,2.20) 0.11 (0.01,1.10) 0.080 0.261 0.061 i Enlisted Groundcrew o i Ln 127 Contrast 124 Enlisted Flyer Enlisted Groundcrew 54 138 149 127 61 149 129 121 51 129 Overall M vs. L H vs. L 0.51 (0.07,3.53) 0.19 (0.02,2.14) 0.346 0.496 0.179 Overall M vs. L H vs. L 2.40 (0.73,7.88) 0.91 (0.22,3.76) 0.156 0.149 0.892 Overall M vs. L H vs. L 0.35 (0.05,2.20) 0.11 (0.01,1.10) 0.080 0.261 0.061 Overall M vs. L H vs. L 0.36 (0.06,2.25) 0.14 (0.01,1.44) 0.165 0.274 0.098 �TABLE 10-26. (continued) Adjusted Exposure Index Analysis for Follovup Participants for Occurrence of Malignant Neoplasms in the Baseline-Follovup Interval Occupation Variable Officer Low Total* 124 Exposure Index Medium High Total* Total* 127 121 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L 2.38 (0.61,9.30) 0.95 (0.18,4.88) 0.272 0.214 0.949 Y a Sun-Exposure Related Malignancies (Verified Only) Enlisted Flyer 54 Enlisted Groundcrew Officer o m 00 Sun-Exposurea Related Malignancies (Verified and Suspected) 138 124 Enlisted Flyer Enlisted Groundcrew 54 138 61 149 127 60 149 51 129 121 51 129 Overall M vs. L H vs. L 0.35 ( . 5 2 2 ) 00,.0 0.11 (0.01,1.10) 0.080 0.261 0.061 Overall M vs. L H vs. L 0.83 (0.15,4.55) 0.50 (0.07,3.39) 0.767 0.826 0.481 Overall M vs. L H vs. L 2.68 (0.83,8.67) 0.93 (0.22,3.86) 0.095 0.100 0.921 Overall M vs. L H vs. L 0.35 ( . 5 2 2 ) 00,.0 0.11 (0.01,1.10) 0.080 0.261 0.061 Overall M vs. L H vs. L 0.59 (0.12,2.94) 0.36 (0.06,2.20) 0.514 0.519 0.268 �TABLE 10-26. (continued) Adjusted Exposure Index Analysis for Followup Participants for Occurrence of Malignant Neoplasms in the Baseline-Pollovup Interval Variable Occupation Officer Systemic15 Malignancies (Verified Only) Enlisted Flyer Exposure Index Low Medium High Total* Total* Total* 127 55 Enlisted Groundcrev 154 Officer 127 130 65 163 123 57 142 o Systemicb Malignancies (Verified and Suspected) Enlisted Flyer Enlisted Groundcrew 55 154 *Total number of participants. a Nonblacks only. b Blacks and nonblacks. —Analyses not done due to sparse cells. 130 65 163 123 57 142 Contrast Overall M vs. L H vs. L Overall M vs. L H vs. L Overall M vs. L H vs. L Overall M vs. L H vs. L Overall M vs. L H vs. L Overall M vs. L H vs. L Adj. Relative Risk (95% C.I.) p-Value 1.60 (0.15,17.22) — 0.365 0.696 — _—. — — — — — — — — 2.95 (0.31,27.73) — 0.25 (0.02,3.90) 0.38 ( . 3 4 9 ) 00,.0 — — 0.109 0.344 0.557 0.326 0.458 — — — �TABLE 10-27. Adjusted Exposure Index Analysis for Follovup Participants for Lifetime Occurrence of Malignant Neoplasms Variable Occupation Officer Basal Cell Carcinoma (Verified Only)a Enlisted Flyer Exposure Index Low Medium High Total* Total* Total* 124 54 Enlisted Groundcrew Officer o i Basal Cell Carcinoma (Verified and Suspected)3 138 124 Enlisted Flyer Enlisted Groundcrew 54 138 127 61 149 127 60 149 121 51 129 121 51 129 Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Overall M vs. L H vs. L 1.33 (0.48,3.66) 1.27 (0.45,3.60) 0.841 0.580 0.647 Overall M vs. L H vs. L 0.23 (0.03,1.61) 0.08 ( . 1 0 7 ) 00,.8 0.024 0.141 0.030 Overall M vs. L H vs. L 1.10 (0.31,3.86) 0.87 (0.24,3.20) 0.937 0.881 0.832 Overall M vs. L H vs. L 1.49 (0.59,3.78) 1.22 (0.46,3.24) 0.699 0.404 0.694 Overall M vs. L H vs. L 0.23 (0.03,1.61) 0.08 ( . 1 0 7 ) 00,.8 0.024 0.141 0.030 Overall M vs. L H vs. L 0.89 (0.27,2.97) 0.71 (0.20,2,48) 0.860 0.849 0.589 �TABLE 10-27. (continued) Adjusted Exposure Index Analysis for Followup Participants for Lifetime Occurrence of Malignant Neoplasms Variable Occupation Officer Sun-Exposure Related Malignancies (Verified Only)3 Enlisted Flyer Low Total* 124 54 Enlisted Groundcrew 138 Officer 124 Exposure Index Medium High Total* Total* 127 61 149 121 51 129 o Sun-Exposure Related Malignancies (Verified and Suspected)3 Enlisted Flyer Enlisted Groundcrew 54 138 127 60 149 121 51 129 Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Overall M vs. L H vs. L 1.19 (0.47,3.00) 0.99 (0.37,2.64) 0.906 0.717 0.980 Overall M vs. L H vs. L 0.42 (0.08,2.19) 0.09 (0.01,0.89) 0.045 0.300 0.039 Overall M vs. L H vs. L 1.35 (0.40,4.58) 0.88 (0.24,3.25) 0.785 0.627 0.850 Overall M vs. L H vs. L 1.33 (0.56,3.16) 0.97 (0.38,2.47) 0.722 0.518 0.952 Overall M vs. L H vs. L 0.42 (0.08,2.19) 0.09 (0.01,0.89) 0.045 0.300 0.039 Overall M vs. L H vs. L 1.10 (0.34,3.52) 0.72 (0.20,2.52) 0.785 0.879 0.603 �TABLE 10-27. (continued) Adjusted Exposure Index Analysis for Follovup Participants for Lifetime Occurrence of Malignant Neoplasms Variable - Occupation Officer Systemic Malignancies (Verified Only) Enlisted Flyer Exposure Index Low Medium High Total* Total* Total* 127 55 Enlisted Groundcrew 154 Officer 127 130 65 163 123 57 142 o t to Systemic Malignancies (Verified and Suspected) Enlisted Flyer Enlisted Groundcrew 55 154 *Total number of participants. a Nonblacks only. Blacks and nonblacks. —Analyses not done due to sparse cells. 130 65 163 123 57 142 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L 1.11 (0.18,7.01) 1.49 (0.24,9.16) 0.902 0.911 0.669 Overall M vs. L H vs. L 0.86 (0.11,7.08) 0.46 (0.04,5.46) 0.806 0.892 0.540 Overall M vs. L H vs. L — — 0.073 — — Overall M vs. L H vs. L 1.69 (0.30,9.65) 1.47 (0.24,8.95) 0.829 0.554 0.679 Overall M vs. L H vs. L 0.51 (0.08,3.47) 0.54 (0.08,3.57) 0.741 0.494 0.527 Overall M vs. L H vs. L — —— 0.087 — — �For the interval analysis, enlisted flyers exhibited a marginally significant decreasing dose-response effect for verified basal cell carcinomas in both the unadjusted (p=0.073) and adjusted analyses (p=0.080). (All Ranch Hand enlisted flyer interval malignant skin neoplasms were verified basal cell carcinomas; thus, interval results for verified and verified plus suspected basal cell carcinoma and the corresponding sun-exposure related neoplasms were identical. Similarly, for lifetime analyses, verified and verified plus suspected analyses were the same). The percentages of participants with interval basal cell neoplasms were 11.1 percent, 3.3 percent, and 1.9 percent for the low, medium, and high exposure categories, respectively. The enlisted groundcrew exhibited a nonsignificant decreasing dose-response effect for basal cell carcinomas and sun-exposure related malignant neoplasms. In the adjusted lifetime analysis for enlisted flyers (Table 10-27), there were significant findings, similar to the interval analysis, namely a decreasing dose-response effect for basal cell carcinomas (overall p=0.024; Adj. RR (medium versus low]: 0.23, 95% C.I.: [0.03, 1.61], Adj. RR [high versus low]: 0.08, 95% C.I.: [0.01, 0.78]), and for sun-exposure related skin malignancies (overall p=»0.045; Adj. RR [medium versus low]: 0.42, 95% C.I.: [0.08, 2.19], Adj. RR [high versus low]: 0.09, 95% C.I.: [0.01, 0.89]). The percentages of participants with lifetime basal cell carcinomas were 13.0 percent, 3.3 percent, and 1.9 percent for the low, medium, and high exposure categories, respectively. The corresponding percentages for lifetime sun-exposure related skin malignancies were 13.0 percent, 4.9 percent, and 1.9 percent. For the enlisted groundcrew cohort, a marginally significant result was found for all systemic malignancies combined in the adjusted analyses (verified only: overall p=0.073; verified plus suspected: overall p=0.087). Of the four verified systemic malignancies, three were in the medium exposure category and one was from the high category. There was one additional suspected malignant neoplasm in the high exposure category. No significant results were found for officers in the lifetime analysis. DISCUSSION The statistical analyses of cancer endpoints in this chapter have carefully followed the prescribed boundaries of the SAIC analytic plan approved by the Air Force. Specific latency analyses of certain cancers associated with environmental exposures were not performed, nor were contrasts of cancer-specific incidence rates to SEER data judged appropriate. Further, embedded case control studies on selected cancers were not performed due to concern for bias. The statistical analyses focused on neoplasms occurring during the time interval between 1982 and 1985 (Baseline to followup). However, because these relatively young and healthy cohorts yielded small numbers of cancers in this short interval, and because of the intense scientific interest in malignant disease, the analysis went beyond the assessment of the incidence of malignant neoplasms in this interval. Lifetime (Baseline and followup data combined) analyses of malignant incident neoplasms were conducted. Cancers occurring prior to military duty in SEA were excluded. A full cancer mortality-morbidity analysis was not attempted but simple tabulations of cancer incidence and mortality of Baseline participants were made. Interval and lifetime analyses were expanded to include suspected cancers noted at followup. Further, grouped cancers that were not likely related were 10-63 �analyzed (all systemic cancers and malignant sun exposure-related skin neoplasms). These efforts, however, have introduced several subtle interpretive issues that should be noted, e.g., skin cancer rates are for nonblacks only, whereas systemic cancer rates are for all races; lifetime group rates are on only those attending the followup examination; and verified and suspected cancer categories included more cases but the data are less reliable. Further, contrasts of cancer rates, particularly skin cancer, between the Baseline results and followup results, or lifetime results, must account for the slight differences in the Baseline and followup cohorts, racial adjustment (Blacks were not omitted from skin cancer analyses at Baseline), skin cancer classification, the change in focus from the Original Comparisons to the total Comparison group, and whether the data were adjusted for covariates. Skin Cancer The emphasis on skin cancer at the followup examination was predicated upon the finding of a significant excess of such cancers at the Baseline examination, and the lack of risk factor data to conduct appropriate adjusted analyses. Because of shifting factors (cited above) between the examinations, a "direct look" at the skin cancer association is not straightforward. Figure 10-1 is presented as an aid to clarify the skin cancer observations over the two examinations. This diagram compares the Baseline and followup analyses. So that the unadjusted Baseline results could be contrasted to the followup results, the estimated relative risk of basal cell carcinoma among nonblack Ranch Hands versus all nonblack Comparisons (not just Originals) was calculated, using data in the Baseline Report. This unadjusted analysis gave a significant relative risk of 1.77 (p=0.049). These results could then be directly contrasted to the unadjusted followup results, which showed a narrowing of group differences over the 3-year interval (Est. RR: 1.23, p=0.429). (It is noted that this contrast compares skin cancer rates of approximately 23 years to 3 years at different levels of age risk.) The adjusted analysis revealed a significant group-by-occupation interaction, due to a significantly higher rate of basal cell carcinomas among Ranch Hand enlisted flyers than the corresponding Comparisons (Adj. RR; 6.50, p=0.019), but very similar rates in the two groups for officers and enlisted groundcrew were seen. The Baseline data were carefully merged (to avoid duplicate counts) with the followup data to assess the total lifetime incidence of basal cell carcinomas between groups. The addition of the nonsignificant followup results to the significant Baseline results produced a nonsignificant lifetime assessment (Est. RR: 1.36, p=0.128), as expected. However, when the lifetime data were adjusted for covariate effects, a significant result emerged (Adj. RR: 1.56, p=0.035), with Ranch Hands having a significant excess of lifetime basal cell carcinoma. A careful examination of the covariates showed that the variable of average residential lifetime latitude was most likely responsible for the significant adjusted results. The latitude variable was a significant confounding variable since it was associated with basal cell carcinomas and with average lifetime latitude which varied significantly by group. 10-64 �Baseline Results Followup Results Lifetime Results "Skin Cancer" Basal Cell Cancers Basal Cell Cancers (Unadj., Original Comparisons, All Races) NS (Unadj., Total Comparisons, Nonblacks) r [RR=1.23, p=0.429] New Analysis NS* (Unadj., Total Comparisons, All Races) I 1 Basal Cell Cancers NS (Unadj. Total "*" Comparisons, Nonblacks) [RR-1.36, p-0.128] **** (Adjusted for all Covariates, Total Comparisons, Nonblacks) (Adjusted for all Covariates, Total Comparisons, Nonblacks) [Group-by-Occ, p=0.044] [RR=1.56, p=0.035] (Unadj., Total Comparisons, Nonblacks) [RR=1.77, p=0.049] S: NS: NS*: »***. Significant (p < 0.05). Not significant (p > 0.10). Borderline significant (0.05 < p <, 0.10). Group-by-covariate interaction. Figure 10-1. Schematic Diagram of Unadjusted and Adjusted Skin Cancer Results, by Significance and Relative Risk, and by Examination Period (Time). 10-65 �Because of the significant confounding effect of the latitude variable, it was examined closely for misclassification or bias. An initial review of the residential history forms showed occasional discrepancies between total residential years and chronologic age. This was generally due to sporadic underreporting, and to the data collection instructions which required the citation only of residences of one year or longer. However, analyses showed fairly good concordance between reported residential years and chronologic age. No significant group difference was found for the inaccuracy of residential reporting (p=0.684), validating the use of all residential histories even though some were slightly imprecise. In the course of reviewing the covariate effects on basal cell carcinoma, the data suggested some unexpected associations. To sharpen these contrasts, adjusted risks were estimated at set levels of skin reaction to sun, skin color, average lifetime residential latitude, and age, relative to the lowest risk observed, i.e., Comparisons 40 years old (at Baseline) who have lived on average in northern latitudes and tan easily were arbitrarily assigned a risk of 1.00. These computed risks are given in Table 10-28. These results show uniform increased risks in the Ranch Hands over both the base level of one and the Comparisons in the same covariate strata. Further, in all strata, age, latitude, and skin color behave as expected. However, the sun-reaction index does not behave as expected since those who burn easily have lower relative risks than those who have an intermediate reaction to sun, although they do have higher relative risks than those who tan easily. This may represent avoidance of sun exposure or the use of sunblock by those individuals. Skin cancer, and particularly basal cell carcinoma, has been emphasized in this report because of the significant group differences detected at Baseline (and the theoretical link to TCDD causation), and the borderline significant adjusted results found for the lifetime rates. The results of the third-year followup analysis suggest that if group differences continue to narrow (where pX).15) at the fifth-year followup examination, the lifetime results would likely not be significant even with full adjustment. Systemic Cancer The analyses of systemic cancer for both the interval and lifetime periods have necessarily been limited by scant data. Cancer specific analyses, in particular, have not provided meaningful results because of low counts. However, some variation in tumor type was noted in the two groups: colon cancer (5 Comparisons, 0 Ranch Hands), testicular cancer (3 Ranch Hands, 0 Comparisons), and smoking related tumors of the oral cavity, pharynx, bronchus, and lung (5 Ranch Hands, 0 Comparisons). Testicular and smoking related tumors have not been associated with exposure to herbicides or TCDD. Table 10-18 cited counts of malignancies that have been associated to herbicides and dioxin exposure. Because of the relative rareness of the diseases soft tissue sarcoma (STS), Hodgkin's disease, and non-Hodgkin's lymphoma, lifetime rates were expected to be exceptionally low. Most of the covariate associations with systemic cancer were anticipated, but the change in significance for smoking (significant at Baseline, borderline significant for lifetime cancers) was not expected, particularly as the cancer cases increased during the interval. 10-66 �TABLE 10-28. Confuted Rides of Basal Cell Carcinoma by Group at Varying Levels of Four Risk Factors, Relative to Comparisons at law Risk* Covariate Categories Skin Reaction Average Lifetime to Sun Residential latitude Skin Color: Not Peach Age at Baseline Comparison Ranch Hand Skin Color; Peach Ranch Comparison Hand 10* .0* 2.99 1.48 4.43 1.55 4,62 2.30 6.85 40 60 1.63 4.87 2.42 7.23 2.52 7.53 3.74 11.18 >37°N 40 60 3.04 9.09 4.52 13.50 4.71 14.06 6.99 2.7 08 40 60 4.97 14.83 7.37 22.02 7.68 22.93 11.40 34.04 >37°N 40 60 2.02 6.04 3.00 8% , 3.13 9.33 4.64 13.86 <37°N Bums Easily 40 60 <37°N Intermediate Reaction >37°N <37°N Tans Easily 40 60 3.30 9.85 4.90 14.62 5.10 15.22 7.57 22.60 ^Computed from main effects model with latitude, skin reaction to sun, and skin color as covariates. **Base Category (Lowest Risk). 10-67 �All Cancers As previously noted, the interrelatedness of many of the analyzed cancer variables has created a compounding of statistical significance, and care should be taken in making inferences and final conclusions. An almost uniform dilutional effect was created by adding "suspected" cancers to the analyses, as there were more of this category in the Comparisons than in the Ranch Hands. The use of suspected neoplasms was deemed necessary in order to best describe the complete cancer findings, recognizing that confirmation of all suspected cases was difficult. Two patterns emerged from the analyses. All relative risks exceeded the value of one, except that of lifetime verified melanoma and verified or verified plus suspected squamous cell carcinoma. Some of the elevated risks were due to the relatedness of the variables as stated, but the relative risks for the unrelated variables skin cancer and systemic cancer both exceeded one. The joint consideration of both yielded a significant relative risk. The second pattern was of the group-by-covariate interactions observed for seven of the analyses; 3 of them involved the covariate of occupation and 4 involved skin reaction to sun. The three group-by-occupation interactions all showed a significant detriment to the Ranch Hand enlisted flying cohort. Further analyses of air crewmembers versus noncrewmembers revealed a significant risk of basal cell carcinoma for the Ranch Hand air crewmembers (RRs 1.94, p.O.049). Since enlisted Ranch Hand flyers in the interval exhibited more basal cell carcinomas (RR: 6.5, p=0.019) and more verified and suspected systemic cancers (4/175 RH with systemic neoplasms versus 0/209 Comparisons, p=0.042), there may be more reason to assume a biologic foundation than chance, although the reason is obscure. The four group-by-sun reaction index interactions all revealed a significant or marginally significant detriment to Ranch Hands who reacted mildly to the sun. In full context, the cancer observations cannot be viewed as disturbing at this time. The skin cancer group differences have narrowed over a 3-year period. An additional analytic observation on skin cancer is that inclusion or exclusion of only one or two cases was shown to alter the choice of the best statistical model, affecting the presence or absence of both covariates and group-by-covariate interactions, and also change the p-value of the adjusted group difference above or below the alpha level of 0.05. For systemic cancer, both groups are at the lower end of the expected ascending cancer curves, where numeric and tumor type fluctuations are expected. A recognized bench-mark for the latency of many cancers is 20 years, and this will not be achieved by most participants until the 5-year followup examination, 2 years from now. Cancer findings at that time will be the basis upon which firm conclusions can be made. SUMMARY AND CONCLUSIONS The cancer analysis focused on cancer occurrences in the Baselinefollowup interval, and also included analyses of the Baseline plus interval cancer history. A summary of the cancer findings is given in Table 10-29'. No significant unadjusted differences were found between nonblack Ranch Hands and Comparisons in the Interval (Baseline-Followup) incidence rates of basal cell carcinoma, melanoma, squamous cell carcinoma, all malignant skin cancers, sun-exposure related malignant neoplasms (comprising basal cell 10-68 �TABLE 10-29. Overall Summary Table: Unadjusted and Adjusted Analysis of Interval and Lifetime Skin and Systemic Cancer Incidence Cancer Type Baseline-Followup Interval UnadjustedAdjusted Lifetime (Baseline & Followup) UnadjustedAdjusted Malignant Skin Cancer (Nonblacks only) Verified Basal Cell Carcinoma NS **** NS Verified plus Suspected Basal Cell Carcinoma NS **** NS Verified Melanoma NS a NS Verified plus Suspected Melanoma NS a NS Verified Squamous Cell Carcinoma NS a NS Verified plus Suspected Squamous Cell Carcinoma NS a NS Verified Sun Exposure Skin Cancers NS NS NS* Verified plus Suspected Sun Exposure Skin Cancers NS NS NS All Verified Malignant Skin Cancers NS ~a NS Verified plus Suspected Malignant Skin Cancers NS NS Verified Skin Cancers of Any Type NS* S Verified plus Suspected Skin Cancers of Any Type NS NS* 10-69 **** NS �TABLE 10-29. Overall Summary Table: Unadjusted and Adjusted Analysis of Interval and Lifetime Skin and Systemic Cancer Incidence (continued) Cancer Type Baseline-Followup Interval UnadjustedAdjusted Lifetime (Baseline & Follovup) UnadjustedAdjusted Malignant Systemic Cancer (Blacks and Nonblacks) Verified Systemic Cancer NS NS NS **** Verified plus Suspected Systemic Cancer NS **** NS **** NS* —a S —a All Neoplasms (Blacks and Nonblacks) Any Type, Any Location" Verified NS: Not significant (p>0.10). ****Group-by-covariate Interaction. —aAnalysis not done. NS*: Borderline significant (0.05<p<0.10). Comprises malignant, benign, uncertain behavior. S: Significant (p<0.05). 10-70 �carcinoma, melanoma, and epithelial neoplasms NOS) or all malignant skin cancers as a group. The unadjusted group contrast of all skin neoplasms (comprising malignant and benign neoplasms, and neoplasms of uncertain behavior or unspecified nature) was marginally significant, with a higher rate among .Ranch Hands. When suspected malignant skin cancers (noted at Followup but not verified at the time of writing) were included in the analyses with the verified conditions, all the unadjusted group contrasts were nonsignificant. The covariates used for the adjusted analyses of basal cell carcinoma and the sun exposure related skin malignancies were age, occupation, skin color, reaction of skin to sun, and average latitude, all of which were highly associated with skin cancer incidence. Other host factors were related to skin cancer incidence, but not as strongly as those included in the analysis. A borderline association with smoking history was noted, and was determined to be partly an age effect. Analysis of the incidence of interval basal cell carcinoma revealed a significant group-by-occupation interaction, due to a significant group difference for enlisted flyers, but not for officers or enlisted groundcrew. Inclusion, of suspected basal cell carcinoma resulted in a group-by-sun reaction index interaction. This was due to Ranch Hands with an intermediate reaction to sun having a higher relative risk than the corresponding Comparisons. The adjusted group contrast of the incidence rates of verified sun-exposure related skin cancers was not significant; inclusion of suspected conditions did not alter this lack of significance. There was no significant group difference for Blacks and nonblacks in the unadjusted incidence rates of all interval verified malignant systemic neoplasms combined, nor was there a significant difference in the adjusted group rates. Analysis of the verified plus suspected interval systemic cancers showed a nonsignificant unadjusted group difference, but a group by occupation interaction was found in the adjusted analysis. This was due to a significant group difference of verified plus suspected systemic malignancies among the enlisted flyers with five occurrences among the Ranch Hands, but none among the Comparisons. Age and a race-by-packyear interaction were important adjusting factors. The Baseline and Followup data were combined for the assessment of lifetime incidence of cancer; occurrences of cancer prior to Vietnam were excluded. There were no significant unadjusted group differences in lifetime incidence rates among nonblacks for basal cell carcinoma, melanoma, squamous cell carcinoma, the sun exposure related skin cancers, or all malignant skin cancers combined. The unadjusted group contrast of all lifetime skin malignancies was significant, with a higher rate among Ranch Hands. Inclusion of suspected cancers with the verified cancers reduced the difference between the groups for all these malignant skin contrasts, except for the sun exposure related skin cancers, for which a marginally significant group difference was found. However, the contrast of all skin malignancies remained close to significance. Adjusted analysis of the incidence rates of lifetime basal cell carcinoma revealed a significantly higher incidence rate among Ranch Hands 10-71 �(Adj. RR: 1.56, p=0.035). Significant effects of an occupation-by-age interaction, a skin color-by-sun reaction index interaction, and a sun reaction index-by-average residential latitude interaction were seen. The adjustment resulted in a significant relative risk that, moreover, was higher than the unadjusted relative risk. Average residential latitude, associated with both group and skin cancer, and skin color, which was associated with the disease and marginally associated with group, played a major part in the change from the unadjusted analysis due to confounding. Inclusion of suspected basal cell carcinoma in the adjusted analysis resulted in a group by sun reaction index interaction, as was noted for the interval analysis. The adjusted group contrast in incidence rates of the sun-exposure related skin cancers was also significant (Adj. RR: 1.54, p=0.030), which is not surprising since the majority are basal cell carcinoma. Inclusion of the suspected conditions resulted in a non-significant group contrast. The unadjusted group contrasts of the incidence rates of all systemic cancers combined were not significant, both for verified and verified plus supected conditions. There was one new occurrence of a soft tissue sarcoma (Ranch Hand) and one suspected cancer of the lymphatic system (Ranch Hand), in addition to the one previously reported soft tissue sarcoma and one Hodgkin's disease in the Comparison group. Adjusted analysis of all lifetime malignant systemic neoplasms as a group, however, revealed a group by occupation interaction, due to a significantly higher rate for Ranch Hand enlisted flyers as contrasted to Comparisons. The same result was found for verified plus suspected systemic cancers. In conclusion, there were no adjusted or unadjusted differences between groups in basal cell carcinoma incidence in the Baseline-followup interval. At Baseline, a significantly higher rate of basal cell carcinoma was found for Ranch Hands when contrasted with Original Comparisons. When the Baseline data were combined with the interval data, adjusted analysis, but not the unadjusted analysis, revealed a significantly higher rate of basal cell carcinoma among the Ranch Hands than among all Comparisons. The relative risk of basal cell carcinoma appears to be declining over time. Relative risks of basal cell carcinoma and systemic cancer were found to be consistently larger than 1. Most of the skin cancers were basal cell carcinomas, upon which most of the skin cancer analysis focused, thus relative risks for sun-exposure related skin neoplasms and all malignant skin cancers as a group were very similar to those for basal cell carcinoma. The number of occurrences of systemic cancer was small, in part because the cohort is relatively young, and although the relative risks (lifetime and interval) are greater than 1, the difference between groups is not significant. Sufficient time may not have elapsed since Vietnam to enable a group difference in systemic neoplasms, if one exists, to be apparent. 10-72 �CHAPTER 10 REFERENCES 1. Young, A.L. 1981. Agent orange at the crossroads of science and social concern. Research report 2750-81, Air Command and Staff College, Air University, Maxwell AFB, Alabama. 2. Tschirley, F.H. 1986. Dioxin. Sci. Amer. 254:29-35. 3. Young, A.L., H.K. Rang, and B.M. Shepard. 1985. Rationale and descriptions of the federally sponsored epidemiologic research in the United States on the phenoxy herbicides and chlorinated dioxin contaminants. In Chlorinated dioxins and dibenzofurans in the total environment II; ed. L.H. Keith, C. Rappe, and G. Choudhary pp. 155-166.Stoneham, Massachusetts: Butterworth Publishers. 4. Poland, A. 1984. Reflections on the mechanism of action of halogenated aromatic hydrocarbons. In Banbury report 18; Biological mechanisms of dioxin action, ed. A. Poland and R.D. Kimbrough, pp. 109-117.Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 5. DiGiovanni, J.A., A. Viaje, D.L. Berry, et al. 1977. Tumor initiating ability of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and Aroclor 1254 in a two-stage system of mouse skin carcinogenesis. Bull. Environ. Contain. Toxicol. 18:522-557. 6. Berry, D.L., T.J. Slaga, J. DiGiovanni, et al. 1979. Studies with chlorinated dibenzo-p-dioxins, polybrominated biphenyls and polychlorinated biphenyls in a two-stage system of mouse skin tumorigenesis: Potent anticarcinogenic effects. Ann. N.Y. Acad. Sci. 320:405-414. 7. National Toxicology Program (NTP). 1982. Carcinogenesis bioassay of 2,3,7,8-tetrachlorodibenzo-p-dioxin (CAS no. 1746-01-6) in SwissWebster mice (dermal study). (Report 80-32, technical report series no. 201, NIH publication no. 82-1757, Research Triangle Park, North . Carolina. 8. Poland, A., D. Palen, and E. Glover. 1982. Tumour promotion by TCDD in skin of HRS/J hairless mice. Nature 300:271-273. 9. Kociba, R.J., D.G. Reyes, J.E. Beyer, R.M. Carreon, C.E. Wade, D.A. Dittenber, R.P. Kalnins, L.E. Frauson, C.N. Park, S.D. Barnard, R.A. Hummel, and C.G. Humiston. 1978. Results of a two-year chronic toxicity and oncogenicity study of 2,3,7,8-tetrachlorodibenzo-p-dioxin in rats. Toxicol. Appl. Pharmacol. 46:279-303. 10-73 �10. National Toxicology Program (NTP). 1982. Carcinogenesis bioassay of 2,3,7,8-tetrachlorodibenzo-p-dioxin (CAS no. 1746-01-6) in OsborneMendel rats and B6C3F1 mice (gavage study). (Report 80-31, technical report series no. 209, NIH publication no. 82-1765, Research Triangle Park, North Carolina. 11. Pitot, H.C., T. Goldsworthy, H.A. Campbell, and A. Poland. 1980. Quantitative evaluation of the promotion by 2,3,7,8-tetrachlorodibenzo-pdioxin of hepatocarcinogenesis from diethylnitrosamine. Cancer Res. 40:3616-3620. 12. Sundell, L., M. Rehn, and 0. Axelson. 1974. Exposure to herbicides— mortality and tumor incidence: An epidemiological investigation in Swedish railway workers. Lakartidningen 71:2466-2470. 13. Axelson, 0., and L. Sundell. 1974. Herbicide exposure, mortality, and tumor incidence. An epidemiological investigation on Swedish railroad workers. Work Environ. Health 11:21-28. 14. Axelson, 0., and L. Sundell. 1977. Phenoxy acids and cancer. Lakartidningen 74:2887-2888. 15. Axelson, 0., L. Sundell, K. Andersson, C. Edling, C. Hogstedt, and H. Kling. 1980. Herbicide exposure and tumor mortality; An updated epidemiological investigation on Swedish railroad workers. Scand. J. Work Environ. Health 6:73-79. 16. Huff, J.E., J.A. Moore, R. Saracci, and L. Tomatis. 1980. Long-term hazards of polychlorinated dibenzodioxins and polychlorinated dibenzofurans. Environ. Health Perspect. 36:221-240. 17. Hardell, L. 1977. Malignant mesenchymal tumors and exposure to phenoxy acids: A clinical observation. Lakartidningen 74:542-546. 18. Hardell, L. 1977. Soft-tissue sarcomas and exposure to phenoxyacetic acids and cancer. Lakartidningen 74:2735. 19. Hardell, L. 1979. Malignant lymphoma of histiocytic type and exposure to phenoxyacetate or chlorophenols. Lancet 1:55-56. 20. Hardell, L., M. Eriksson, P. Lenner, et al. 1981. Malignant lymphoma and exposure to chemicals, especially organic solvents, chlorophenols and phenoxy acids: A case control study. Br. J. Cancer 43:169-176. 21. Eriksson, M., L. Hardell, N.O. Berg, T. Moller, and 0. Axelson. 1981. Soft tissue sarcomas and exposure to chemical substances: A casereferent study. Br. J. Ind. Med. 38:27-33. 22. Hardell, L. 1981. Relation of soft-tissue sarcoma, malignant lymphoma and colon cancer to phenoxy acids, chlorophenols and other agents. Scand. J. Work Environ. Health. 7:119-130. 23. Hardell, L., B. Johansson, and 0. Axelson. 1982. Epidemiological study of nasal and nasopharyngeal cancer and their relation to phenoxy acid or chlorophenol exposure. Am. J. Indus. Med. 3:247-257. 10-74 �24. Hardell, L., N.O. Bengtsson, U. Jonsson, S. Eriksson, and L.G. Larsson. 1984. Aetiological aspects on primary liver cancer with special regard to alcohol, organic solvents and acute intermittent porphyria—an epidemiological investigation. Br. J. Cancer 50:389-397. 25. Hardell, L., and 0. Axelson. 1984. Phenoxyherbicides and other pesticides in the etiology of cancer: Some comments on the Swedish experience. Presented at Cancer Prevention—Strategies in the Workplace. University of California, San Francisco, December 1984. 26. Jannerfeldt, E. 1980. Epidemiological methodology and pesticide studies. Lakartidningen 77(12):1096. 27. (Editorial.) 1982. Phenoxy herbicides, trichlorophenols, and softtissue sarcomas. Lancet 1(8278):1051-1052. 28. Coggon, D., and E.D. Acheson. 1982. Do phenoxy herbicides cause cancer in man? Lancet 1(8278):1057-1059. 29. Axelson, 0. 1978. Aspects on confounding in occupational health epidemiology, Letter. Scand. J. Work Environ. Health 4:85-89. 30. Axelson, 0. 1980. Views on criticism of pesticide studies. Lakartidningen 77(12):1096-1099. 31. Axelson, 0. 1980. A note on observational bias in case-referent studies in occupational health epidemiology. Scand. J. Work Environ. Health 6:80-82. 32. Remington, R.D. 1980. Specific summary critique of five investigations related to concerns about agent orange. Congressional Record, August 6, 1980, pp. S 10911, S 10912. 33. U.S. Environmental Protection Agency. 1985. Health assessment document for polychlorinated dibenzo-p-dioxins. Final report, September 1985. 34. Ott, M.G., B.B. Holder, and R.D. Olson. 1980. A mortality analysis of employees engaged in the manufacture of 2,4,5-trichlorophenoxyacetic acid. J. Occup. Med. 22:47-50. 35. Cook, R.R., J.C. Townsend, M.G. Ott, and L.G. Silverstein. 1980. Mortality experience of employees exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). J. Occup. Med. 22:530-532. 36. Zack, J.A., and R.R. Suskind. 1980. The mortality experience of workers exposed to tetrachlorodibenzo-dioxin in a trichlorophenol process accident. J. Occup. Med. 22:11-44. 37. Zack, J.A., and W.R. Gaffey. 1983. A mortality study of workers employed at the Monsanto.Company Plant in Nitro, West Virginia. Environ. Sci. Res. 26:575. 38. Honchar, P.A., and W.E. Halperin. 1981. 2,4,5-T, trichlorophenol, and soft-tissue sarcoma. Lancet l(8214):268-269. 10-75 �39. Cook, R.R. 1981. Dioxin, chloracne, and soft-tissue sarcoma. 1(8220):618-619. Lancet 40. Moses, M., and I.J. Selikoff. 1981. Soft-tissue sarcomas, phenoxy herbicides, and chlorinated phenols. Lancet 1(8234):1370. 41. Johnson, F.E., M.A. Kugler, and S.M. Brown, 1981. Soft tissue sarcomas and chlorinated phenols. Lancet 2(8236):40. 42. Fingerhut, M.A., W.E. Halperin, P.A. Honchar, A.B. Smith, D.H. Groth, and W.O. Russell. 1984. An evaluation of reports of dioxin exposure and soft tissue sarcoma pathology in U.S. chemical workers. In Banbury report 18; Biological mechanisms of dioxin action, ed. A. Poland and R. D. Kimbrough, pp. 461-470.Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 43. Percy, C., E. Stanek, and L. Gloeckler. 1981. Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am. J. Public Health 71(3):242-250. 44. Donna, A., P.G. Betta, F. Robutti, P. Crosignani, F. Berrino, D. Bellinger!. 1984. Ovarian mesothelial tumors and herbicides: case-control study. Carcinogenesis 5:941-942. A 45. Riihimaeki, V., S. Asp, E. Pukklala, and S. Hernberg. 1983. Mortality and cancer morbidity among chlorinated phenoxy acid applicators in Finland. Chemophere 12:779-784. 46. Lynge, E. 1985. A follow-up study of cancer incidence among workers in manufacture of phenoxy herbicides in Denmark. Br. J. Cancer 52:259-270. 47. Smith, A.H., N.E. Pearce, D.O. Fisher, et al. 1984. Soft-tissue sarcoma and exposure to phenoxyherbicides and chlorophenols in New Zealand. JNCI 73:1111-1117. 48. Pearce, N.E., A.H. Smith, and D.O. Fisher. 1985. Malignant lymphoma and multiple myeloma linked with agricultural occupations in a New Zealand cancer registry-based study. Am. J. Epidemic!. 121:225-237. 49. Wiklund, K., and L.E. Holm. 1986. Soft tissue sarcoma risk in Swedish agricultural and forestry workers. JNCI 76(2):229-234. 50. Hoar, S.K., A. Blair, F.F. Holmes, et al. 1986. Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA 256:1141-1147. 51. Colton, T. 1986. Herbicide exposure and cancer. 256:1176-1178. Editorial. JAMA 52. Woods, J.S., L. Polissar, R.K. Severson, L.S. Heuser, and B.C. Kulander. 1987. Soft tissue sarcoma and non-Hodgkin's lymphoma in relating to phenoxy herbicide and chlorinated phenol exposure in western Washington. Preprint. To appear in JNCI, May 1987. 10-76 �53. Sarma, P.R., and G. Jacobs. 1981. Thoracic soft-tissue sarcoma in Vietnam veterans exposed to agent orange. New Engl. J. Med. 306(18):1109. 54. Greenwald, P., B. Kovasznay, D.N. Collins, and G. Therriault. 1984. Sarcomas of soft tissue after Vietnam service. JNCI 73:1107-1109. 55. Kogan, M.D., and R.W. Clapp. 1985. Mortality among Vietnam veterans in Massachusetts, 1972-1983. Boston, Massachusetts"Division of Health Statistics and Research, Massachusetts Department of Public Health. 56. Royal Commission on the Use and Effects of Chemical Agents on Australian Personnel in Vietnam. 1985. Cancer. Vol. 4 in Final Report, pp. VIII-129. Canberra, Australia: Australian Government Publishing Service. 57. Lawrence, C.E., A.A. Reilly, P. Quickenton, P. Greenwald, W.F. Page, A.J. Kuntz. 1985. Mortality patterns of New York State Vietnam veterans. Am. J. Public Health 75:277-279. 58. Wendt, A.S. 1985. Agent orange Iowa survey of Vietnam veterans. Report, Iowa State Department of Health, July 1985. Final 59. Holmes, A.P. 1986. West Virginia Vietnam-era veterans mortality study. Preliminary report, Health Statistics Center, West Virginia Department of Health, January 1986., 60. Anderson, H.A., L.P. Hanrahan, M. Jensen, D. Laurin, W. Yick, and P. Wirgman. 1986. Wisconsin Vietnam veteran mortality study. Final report, Wisconsin Department of Health and Social Services, March 1986. 61. Rang et al. 1986. Soft tissue sarcomas and military service in Vietnam: A case comparison group analysis of hospital patients. J_^ Occup. Med. 28:1215-1218. 62. Thiess, A.M., R. Frentzel-Beyme, and R. Link. 1982. Mortality study of persons exposed to dioxins in a trichlorophenol process accident that occurred in the BASF AG on November 17, 1953. Am. J. Ind. Med. 3:179-189. 63. Blair, A., D.J. Grauman, J.H. Lubin, et al. 1983. Lung cancer and other causes of death among licensed pesticide applicators. JNCI 71:31-37. 64. Barthel, E. 1981. Cancer risk in agricultural workers exposed to pesticides. Arch. Geschwulstforsch. 51(7):579-585. 65. Gatti, R.A., and R.A. Good. 1971. Occurrence of malignancy in immunodeficiency diseases: A literature review. Cancer 28:89-98. 66. Louie, S. and R.S. Schwartz. 1978. Immunodeficiency and the pathogenesis of lymphoma and leukemia. Semin. Hematol. 15:117-138. 10-77 �67. Di Carlo, E.F., J.B. Amberson, C.E. Metroka, P. Ballard, A. Moore, J.A. Mouradian. 1986. Malignant lymphomas and the acquired immunodeficiency syndrome. Arch. Pathol. Lab. Med. 110:1012-1016. 68. Lathrop, G.D., P.M. Moynahan, R.A. Albanese, and W.H. Wolfe. epidemiologic investigation of health effects in Air Force following exposure to herbicides—Baseline mortality study Brooks Air Force Base, Texas;Epidemiology Division, Data Division, USAF School of Aerospace Medicine. 1983. An personnel results. Sciences 69. Holman, C.D.J., and Armstrong, B.K. "Pigmentary Traits, Ethnic Origin, Benigh Nevi, and Family History as Risk Factors for Cutaneous Malignant Melanoma." Journal of the National Cancer Institute 72;257-266, 1984. 70. Scot to, J. and Fears, T.R. "Skin Cancer Epidemiology: Research Needs," National Cancer Institute Monograph 50:169-177, 1978. 10-78 �CHAPTER 11 NEUROLOGICAL ASSESSMENT INTRODUCTION Neurological signs and symptoms, as distinguished from overt diagnosable neurological disease, have been consistently associated with industrial exposure to chlorophenols, phenoxy herbicides, and TCDD. Thus, the neurological system comprises a major examination focal point in all dioxin morbidity studies. This report carefully separates central and peripheral neurological status from "neurobehavioral" parameters, which are discussed in Chapter 12, Psychological Assessment. Based on animal experiments, neurotoxicity can be attributed to the compounds 2,4-D and TCDD. For low to moderate doses, both central and peripheral acute effects occur but appear to be reversible. " The effects of 2,4-D are presumably due to disruption in the neuromuscular transport system of organic acid anions. A variety of 2,4-D experiments in several animal species generally shows a wide range of neural pathology including electroencephalographic (EEC) desynchronization, demyelination, myotonia, loss of coordination, and uncontrolled motor activity. No substantive data support the isolated neurotoxicity of 2,4,5-T. Numerous case reports following accidental human exposures or suicide attempts with 2,4-D have shown a remarkable neurologic parallel to the animal studies. ~ In particular, 2,4-D and TCDD have been implicated in a wide array of central neurological signs and symptoms, including headache, vomiting, dizziness, disorientation, sleep disturbance, stupor, memory loss, loss of coordination, and EEC abnormalities or alterations from a baseline tracing. ' ' Peripheral abnormalities have included demyelination, acute degeneration of ganglion cells, temporary paralysis, anesthesia, hyperesthesia, paresthesia, neuralgic pain, numbness, tingling, muscle pain, muscle fasciculations, depressed or absent deep tendon reflexes, weakness, decreased nerve conduction velocities, "polyneuritis," and limb fatigue. " These peripheral signs and symptoms in industrial workers have received the generic diagnostic label "neurasthenia." Both the number and severity of symptoms tended to aggregate in individuals with chloracne as contrasted to those without chloracne.11'16'17 In general, there is consistency between the various case reports of neurasthenia and results from uncontrolled clinical studies. Of particular relevance is the consistency in findings from studies of both industrial manufacturing and industrial accidents. This literature provides the clearcut conclusion that neurological.impairment is caused directly by exposure to 2,4-D and TCDD. Not answered satisfactorily in the literature, however, are the issues of complete reversibilty of observed signs and symptoms and the long-term impact on health and quality of life. 11-1 �Because of the conclusive evidence that two of three Agent Orange ingredients cause neurological "disease," it follows that significant exposure to Agent Orange could manifest neurologic signs, symptoms, or sequelae. In fact, over 10 percent of Vietnam veterans who enlisted in the VA^gent Orange Registry cited one or more symptoms of the neurasthenic complex. The VA Registry is a comprehensive listing, predominantly of veterans alleging health impairments due to Agent Orange exposure. The Registry does not purport to be a scientific effort upon which cause-and-effect relationships can be established. Nonetheless, some individuals believe that the symptom array in the VA Registry is so compatible with case reports and numerator-oriented clinical studies that the veterans must, in fact, have suffered adverse health effects from their Vietnam service and presumed exposure to Agent Orange. Others point to the intense media attention to "Agent Orange symptoms" during the formation of the Registry, and presume that the veterans' complaints are largely due to an "over-reporting" or compensation bias. Clearly, only well-controlled, well-conducted epidemiologic studies of veterans known to have been exposed to Agent Orange can answer the question of cause and effect for illnesses, including the specific question of whether single or multiple neurologic signs and symptoms are also attributable to these exposures. Baseline Summary Results The 1982 AFHS neurological assessment consisted of questionnaire, physical examination, and electromyographic data obtained by examiners and technicians who were blinded to the group identity of each participant. The physical examination required an average of 30 minutes to complete. Those few individuals with positive RPR tests, a screening serological test for syphilis, and those with peripheral edema were deleted from the statistical analyses. Covariates of reported alcohol usage, exposure to insecticides and industrial chemicals, and glucose intolerance (diabetes) were analyzed. Results of the questionnaire disclosed no significant group differences in reported neurological diseases. The physical examination did not reveal any statistically significant group differences in the function of all 12 cranial nerves, nor any effects due to the covariates of alcohol or diabetes. Peripheral nerve function was assessed by the quality of four reflexes (patellar, Achilles, biceps, and Babinski), muscle strength/bulk, and reaction to the stimuli of pin prick, light touch, and vibration. Other than a statistically significant increase (p»0.03) in Ranch Hand Babinski reflexes, significant group differences were not detected. The alcohol covariate demonstrated a marginal effect (p=0.07) on pin-prick reaction, while glucose intolerance showed a profound effect on the patellar and Achilles reflexes and reactions to light touch and vibration. Nerve conduction velocities were obtained on the ulnar nerve, above and below the elbow, and the peroneal nerve by highly standardized methods. The results for each segmental measurement were nearly identical in the Ranch Hand and Comparison groups. Conduction velocity showed highly significant inverse relationships to both alcohol (measured in drink-years) and glucose intolerance in almost all of the anatomic measurements. No group associations or interactions were detected with the covariates of industrial and degreasing chemicals and insecticides. 11-2 �No significant group differences were detected in four measures of central neurological function (tremor, finger-nose coordination, modified positive Romberg's sign, or abnormal gait). Alcohol usage was significantly associated with the presence of tremor, and glucose intolerance was highly correlated to abnormal balance and the presence of tremor. Of a total of 84 exposure index analyses on all of the dependent variables, 3 were statistically significant but were either nonlinear or biologically implausible. In summary, the detailed neurological examination and assessment did not reveal statistically significant increases in abnormalities in the Ranch Hands, nor were consistent dose-response relationships noted for herbicide exposure. The classical neurological effects of alcohol ingestion and diabetes were repeatedly observed in the neurological evaluations. Parameters of the 1985 Neurological Assessment The 1985 AFHS neurological examination deleted the measurements of nerve conduction velocities but otherwise repeated the format of the Baseline examination. The questionnaire maintained a historical focus of neurasthenia via five questions for the 1982-1985 interval. With this similarity in examination and questionnaire, the dependent variables of the analyses were almost identical to those of the Baseline study, however, the number of covariates was slightly increased. Diabetic status was trichotomized: Individuals reporting a history of diabetes (unverified) and individuals exhibiting glucose intolerance with postprandial glucose levels greater than or equal to 200 mg/dl were classified as diabetic, participants with glucose levels of at least 140 mg/dl but less than 200 mg/dl were classified as impaired, and participants with glucose levels less than 140 mg/dl were classified as normal. Race was included as a covariate, and lifetime alcohol use was updated on the basis of enhanced information from the 1985 questionnaire. The analyses were based on 1,016 Ranch Hands and 1,293 Comparisons. Individuals confirmed to be positive for syphilis by fluorescent treponemal antibody (FTA) testing were excluded from all analyses. Individuals with peripheral pitting or nonpitting edema were excluded only for the analyses of pin prick, light touch, and vibration. Numeric differences in the following tables are due to missing dependent variables or covariate data. The exclusions and missing covariate data are summarized in Table 11-1. The unadjusted analyses used chi-square or Fisher's exact test for frequency table analyses. Adjusted analyses were not performed where only sparse numbers of abnormalities were found. Logistic regression models were used in all adjusted analyses. Parallel analyses using Original Comparisons can be found in Appendix I, Tables 1-3 through 1-13. RESULTS AND DISCUSSION General Detailed neurological data were obtained on all participants by standard physical examination techniques. Four board-certified SCRF neurologists, all 11-3 �TABLE 11-1. Exclusions and Hissing Data for Neurological Assessment by Group Group Data Category Ranch Hand Comparison Total Lifetime Alcohol History (Drink-Years); Missing Data 39 40 79 Peripheral Edema (Exclusion Category for Pin Prick, Light Touch, and Ankle Vibration) 13 16 29 Diabetic Class (Missing Data) 0 4 4 Positive Syphilis Serology (RPR and FTA) Exclusion Category 0 1 1 blinded to the exposure status of the participants, conducted the examinations. Data were collected to assess three specific clinical areas: cranial nerve function, peripheral nerve function, and central nervous system (CNS) function. The analyses in this chapter are presented in the order of these functional areas. The unadjusted statistical analyses presented in this chapter are straightforward group contrasts of dichotomous (normal/abnormal) dependent variables using Fisher's exact test. Logistic regression models for adjusted analyses used the covariates of age (born in or after 1942, born between 1923 and 1941, born in or before 1922), race (Black, nonblack), occupation (OCC) (officer, enlisted flyer, enlisted groundcrew), diabetic class (DIAB) (normal, less than 140 mg/dl glucose? impaired, at least 140 mg/dl but less than 200 mg/dl glucose; diabetic, greater than or equal to 200 mg/dl glucose or past diabetic history), lifetime alcohol use (DRKYR) (total drink-years: 0, greater than 0 to 50, greater than 50), and unprotected exposure to insecticides (INS) (recorded as yes/no, excluding herbicide exposure). The models are "best-fit" following a step-down strategy beginning with all two-way interactions among the six covariates. Only variables with a substantial number of abnormalities were analyzed. Several summary indices were constructed for functionally related variables with low counts of abnormalities. A summary index was created for the cranial nerve function by combining the 15 cranial nerve parameters into a single index, which was classified as normal if all parameters were normal. Another cranial nerve function was created in a similar fashion, excluding neck range of motion due to the much higher percentage of abnormalities found for this variable relative to the other parameters. The four coordination parameters of the central nervous 11-4 �system were similarly combined to form a summary index. These constructed indices are presented more for the purpose of inspection than for inference making. Since the corneal reflex (as one measure of the trigeminal nerve function) contained no abnormalities for either group, no table is presented with this variable. The statistical power to detect a given relative risk in many of the subsequent analyses was somewhat limited. With the use of a two-sided cfr-level of 0.05 and power of 0.80, the sample sizes were sufficient to detect a 49 percent increase in the frequency of abnormal values for neck range of motion, a 69 percent increase for light touch but only a doubling for tremor, and an elevenfold increase for gag reflex. Power was generally poor in these analyses because of the extremely small number of abnormalities observed in both the Ranch Hand and Comparison groups. Questionnaire Data For the interval questionnaire, each participant was asked to update his health history for neurologic conditions occurring between 1982 and 1985. All affirmative histories were subjected to medical record verification, and appropriate ICD-9-CM coding. All verified neurological diseases were placed into six broad disease categories. These data are summarized in Table 11-2. TABLE 11-2. Unadjusted Analysis for Verified Neurological Disease by Group*—1982-1985 Group Abnormalities Ranch Hand Disease Category Inflammatory Diseases Hereditary and Degenerative Diseases Peripheral Disorders Disorders of the Eye Disorders of the Ear Other Disorders Comparison Number Percent Number Percent Total p-Value** 0 2 0.0 0.2 0 0 0.0 0.0 0 2 — 0.194 18 5 6 8 1.8 0.5 0.6 0.8 27 7 7 3 2.1 0.5 0.5 0.2 45 12 13 11 0.651 0.999 0.999 0.069 *Based on 1,016 Ranch Hands and 1,293 Comparisons; some participants may be classified in more than one category. **Fisher's exact test. 11-5 �All of these analyses were based on very small numbers of abnormalities, but none of the six general disease categories showed statistically significant differences between groups, although the marginal significance of the Other Disorders category is of interest. To determine whether lifetime differences in neurologic disease exist between the Ranch Hand and Comparison groups, verified followup data were combined with verified Baseline historical data. This tabulation is presented in Table 11-3. TABLE 11-3. Unadjusted Analysis for Verified Neurological Disease by Group*—Baseline and First Followup Studies Combined Group Abnormalities Ranch Hands Disease Category Comparisons Number Percent Number Percent Inflammatory Diseases Hereditary and Degenerative Diseases Peripheral Disorders Disorders of the Eye Disorders of the Ear Other Disorders Total p-Value** 0.660 0.999 3 2 0.3 0.2 2 3 0.2 0.2 5 5 23 2.3 1.6 2.4 1.5 2.9 1.8 2.2 1.1 61 39 0.361 . 16 24 15 38 23 53 29 0.889 29 14 0.747 0.453 *Based on 1,016 Ranch Hands and 1,293 Comparisons; some participants may be classified in more than one category. **Fisher's exact test. Like the followup data, the combined data revealed no statistically significant differences in any disease category. Also, there was no significant difference in patterns of disease for each group (p=0.721). Physical Examination Data Dependent Variable and Covariate Relationships: Cranial Nerve Function, Peripheral Nerve Status, and Central Nervous System Coordination Responses from both groups were combined and analyzed with the six covariates. In addition, current drinking (yes/no) and lifetime history of 11-6 �unprotected exposure to industrial and degreasing chemicals (yes/no) were also evaluated. Indices constructed from dependent variables from the cranial nerve function and central nervous system coordination processes were also included. A summary tabulation of covariate associations is shown in Table 11-4. The 10 variables in this table include variables from the peripheral nerve status and CNS process as well as the cranial nerve function and constitute the subset of variables for which adjusted analyses were performed. These results generally showed the profound association of classical risk factors for neurological deficits. Increases in the percentages of abnormalities for Achilles reflex, muscle status, neck range of motion, and the cranial nerve function index (which included neck range of motion) were associated with increases in age. Increasing percentages of abnormalities for pin prick and light touch were noted for increasing age from the young category (3.4% and 2.7% for pin prick and light touch, respectively) to the middle-aged category (8.1% and 4.7%, respectively), but a declining proportion of abnormalities was observed from the middle- to older-age categories (7.3% and 1.2%, respectively). No age effect was noted for gait, the CNS index, the cranial nerve index (neck range of motion excluded), and, surprisingly, for tremor. Race was not a significant covariate for any dependent variable. A significant occupational effect was observed for the CNS summary index (p=0.021, with both enlisted categories having a higher frequency of abnormalities [5.7% and 4.1% for enlisted flyers and enlisted groundcrew, respectively] than the officer category [2.6%]) and for the neck range of motion variable (p=0.010, with increasing proportions of abnormalities from the enlisted groundcrew [4.6%] to officers [7.5%] to enlisted flyers[8.0%]). Abnormalities in the Achilles tendon reflex were related to a graduated increase in drink-years of alcohol. For the variables of pin prick, light touch, muscle status, neck range of motion, and cranial nerve index (with neck range of motion included), the 0 drink-year category was related to a higher frequency of abnormalities than the greater than 0 to 50 drink-year category, which in turn was associated with a lower frequency of abnormalities than the greater than 50 drink-year category. For the current drinker (which was not used for modeling), the percentage of abnormalities for Achilles reflex and gait was significantly greater (p=0.007 and p=0.001 for Achilles reflex and gait, respectively) for current nondrinkers than for current drinkers. This relationship was reversed for the CNS summary index. For both the Achilles tendon reflex and the response to pin prick, the frequencies of abnormalities significantly increased from the diabetic classes of normal to impaired to diabetic (p<0.001 for both variables). For the variables of light touch, muscle status, gait, and CNS summary index, the associations with diabetic status were mixed: The normal diabetic class had a higher proportion of abnormalities than the impaired stratum which, in turn, had a lower proportion of abnormalities than the overtly diabetic class. Unexpectedly, the proportion of tremor abnormalities was highest for the normal diabetic class and became, successively lower in the impaired and diabetic strata (2.48%, 0.45%, and 0%, respectively). A higher proportion of pin prick abnormalities was associated with a history of unprotected exposure to insecticides (p=0.040; 6.94% for exposed versus 4.8% for unexposed). The other dependent variables were not 11-7 �TABLE Three Summary TiKyippg 3nd the Govariates in the Combined Ranch Band and Comparison Groups Covariate Exposure Degreasing Chemicals* NS 000 .5 NS <0.001 000 .4 NS NS 006 .2 NS** NS NS Race Occupation Total Drink-years Achilles Reflex O01 .0 NS NS 002 .2 007 .0 <0.001 Pin Prick 4.0 001 NS NS 004 .0 NS 007 .2 NS NS 006 .0 NS O01 .0 NS NS 001 .0 NS** O01 .0 NS 005 .2 Gait NS NS NS NS 001 .0 003 .3 NS NS NS CNS Index NS NS 001 .2 NS 002 .1 006 .1 NS NS NS Tremor NS NS NS NS NS 001 .1 NS NS NS Neck Range of Motion O01 .0 NS 000 .1 004 .1 NS NS** NS 009 .3 NS Cranial Nerve Function Index <0.001 NS NS** 0.032 NS NS NS NS** NS NS NS NS NS NS NS light Touch Muscle Status Cranial Nerve Function Index (Neck Range of Motion Excluded) NS: Not significant ( > . 0 . p01) * Variable not used in adjusted analyses. NS**: Borderline significant ( . 5 < p <0.10). 00 NS** Current Drinking* Industrial Chemicals* Age Dependent Variable Diabetic Class Insecticides NS NS** NS �significantly affected by the insecticide covariate. For most dependent variables, both Ranch Hands and Comparisons exposed to degreasing or industrial chemicals exhibited a smaller percentage of abnormalities than participants without exposure. Because the biologic basis of these findings is not readily apparent, these two variables were not used as adjusting covariates. Cranial Nerve Function All 12 cranial nerves were assessed as unilateral or bilateral; these unadjusted data are presented in Table 11-5. All bilateral assessments (e.g., right visual field, left visual field) were combined for the analyses; an abnormality consisted of a right and/or a left abnormality. The analysis of the 12 variables and two cranial nerve function summary indices did not reveal statistically significant group differences. Since no abnormalities are present for the variables of speech and tongue position in the Comparison group, the estimated relative risk for these variables was approximated by adding 0.5 to each cell. The low frequency of abnormal counts in all variables, except neck range of motion, contrasts with the 1982 Baseline findings, which found substantially more abnormalities. For example, ocular movement was recorded as abnormal in more than 30 percent of the participants at Baseline while only 0.7 percent of participants were found to be abnormal at followup. Because of the few abnormalities for all variables except neck range of motion, two summary indices of cranial nerve function were constructed. One indicated whether or not a participant is abnormal for any of the 15 variables, while the other was a composite for all except neck range of motion. The analyses of these indices are reflected in Table 11-5, and showed no statistically significant group differences, although the index excluding neck range of motion is of borderline significance. Speech and tongue position relative to midline were also of borderline significance, although the analysis was affected by sparse numbers of abnormalities. The constructed indices are presented more for the purpose of inspection than for inference making. Because of sparse numbers of abnormalities, adjusted analyses were performed only on the variable neck range of motion and the cranial nerve function summary indices, with and without neck range of motion data. The results of these analyses are given in Table 11-6. None of the results were statistically significant, although the cranial nerve function index, without neck range of motion, was marginally significant (p=0.061) when participants with missing drink-years were included. In the primary adjusted analysis for this variable, drink-years was included in a significant covariate interaction. However, an alternative model was also examined that included participants with missing drink-years due to the disparity in group response for these participants (4 out of 39 Ranch Hands abnormal, 0 out of 40 Comparisons abnormal). The results of these adjusted analyses are nearly identical to the unadjusted analyses (see Table 11-5). A borderline significant result of a group (GRP)-by-age interaction (p=0.0501) for neck range of motion existed, and an additional analysis stratifying by age is provided in Table 11-7. This table presents the results of interaction analyses from variables assessing the peripheral nerve status and central nervous system coordination process as well. 11-9 �TSBLE11-5. unadjusted Analyses for Cranial Nerve Function by (koup Group Ranch Hand Comparison Statistic Number Percent Number Percent I Olfactory n Abnormal Normal 1,016 10 1,006 1.0 99.0 1,292 10 1,282 0.8 99.2 1.27 ( . 3 3 0 ) 0.654 05,.7 n n Abnormal Normal 1,016 6 1,010 0.6 99.4 1,292 6 1,286 0.5 99.5 1.27 ( . 1 3 % 0.774 04,.) Light Reaction m n Oculomotor Abnormal Normal 1,015 8 1,007 0.8 99.2 1,289 9 1,280 0.7 99.3 1.13 ( . 3 2 9 ) 0.811 04,.4 1,016 6 1,010 0.6 99.4 1,292 10 1,282 0.8 99.2 0.76 ( . 8 2 1 ) 0 8 1 02,.0 .0 Ocular Movements m n Oculomotor Abnormal IV Normal Trochlear VI Abducens Facial Sensation V n Trigeminal Abnormal Normal 1,014 4 1,010 04 . 99.6 1,290 2 1,288 0.2 99.8 2.55 ( . 7 1 . 5 0 4 5 04,39) .1 Jaw Clench V n Trigeminal Abnormal Normal 1,016 2 1,014 0.2 99.8 1,292 2 1,290 0.2 99.8 1.27 ( . 8 9 0 ) 0.999 01,.5 n Abnormal Normal 1,016 7 1,009 0.7 99.3 1,292 4 1,288 0.3 99.7 2.23 ( . 7 7 4 ) 0.230 06,.1 n Abnormal Normal 1,015 7 1,008 0.7 99.3 1,292 7 1,285 0.5 99.5 1.28 ( . 5 3 6 ) 0.789 04,.5 n vm Acoustic . Abnormal Normal 1,015 2 1,013 0.2 99.8 1,292 1 1,291 0.1 99.9 2.55 ( . 3 2 . 5 0.586 02,81) Variable Cranial Nerve Smell Visual Fields Smile Optic vn Facial Palpebral Fissures Balance vn Facial 11-10 Est. Relative Risk ( 5 C.I.) p-Value 9% �TABLE 11-5. (continued) Unadjusted Analyses for Cranial Nerve Ruction by Group Group Ranch Hand Comparison Number Percent Number Percent Est. Relative Risk ( 5 C.I.) p-Value 9% Variable Cranial Nerve Gag Reflex IX n Abnormal Glossopharyngeal Normal 1,014 1 1,013 0.1 99.9 1,291 1 1,290 0.1 99.9 1.27 ( . 8 2 . 8 0.999 00,03) Speech X Vagus n Abnormal Normal 1,016 3 1,013 0.3 99.7 1,291 0 1,291 00 . 100.0 8.92 ( . 6 1 2 8 ) 0.085 04,7.9° Tongue Position Relative to Midline X Vagus n Abnormal Normal 1,015 3 1,012 0.3 99.7 1,292 0 1,292 00 . 100.0 8.94 ( . 6 173. 19)a 0.085 04, Palate and Uvula Movement XI Spinal Accessory n Abnormal Normal 1,014 2 1,012 0.2 99.8 1,291 1 1,290 0.1 99.9 2.55 ( . 3 2 . 6 0.586 02,81) Neck Range of Motion XH n Hypoglossal Abnormal Normal 1,016 61 955 6.0 94.0 1,292 84 1,208 6.5 93.5 0.92 ( . 5 1 2 ) 06,.9 0.666 Cranial Nerve Rnction Index n Abnormal Normal 1,003 96 907 9.6 90.4 1,275 115 1,160 9.0 91.0 1.07 ( . 0 1 4 ) 08,.2 0.663 Cranial Nerve Rjnction Index (Neck Range of Motion Excluded) n Abnormal Normal 1,003 42 961 4.2 95.8 1,275 35 1,240 2.7 97.3 1.55 ( . 8 2 4 ) 09,.4 0.062 Statistic "Estimated relative risk and 95% confidence interval calculated after adding 0.5 to each cell. 11-11 �TABLE 11-6. Adjusted Analyses for Selected Variables of Cranial Nerve Function by Group Ranch Band Variable Est. Relative .. Statistic Njnber Percent Number Percent Risk<95* C I ) p-Value Neck Range of Motion n 1,016 Abnormal 61 Normal 955 1,292 84 6.0 94.0 1,208 6.5 93.5 Cranial Nerve Function Index n 1,003 Abnormal % Normal 907 9.6 90.4 1,275 115 1,160 Covariate Remarks* 9.0 91.0 0.90 ( . 3 1 2 ) 0 5 1 06,.7 . 3 AGE(pO.OOl) GRP*AGE (nBrginal:pM).0501) 1.07 ( . 0 1 4 ) 0.666 08,.2 AGE(pO.OOl) Cranial n 964 1,232 08,.0 . 5 1.42 ( . 8 2 3 ) 0 1 3 DIAB*INS(p=0.022) OCC*EROR(p=0.011) 34 Nerve 38 3.9 Abnormal 2.8 Rnction Normal 926 96.1 1,198 97.2 OCC*DIAB(p=0.015) Index (Neck Alternative Model—Includes Missing Drink-Year Participants*fb Range of Motion n 1,271 1.56 ( . 8 2 4 ) 0 0 1 DIAB*INS(p=0.017) 09,.9 . 6 1,003 Excluded) Abnormal 42 4.2 34 2.7 OCC*DIAB(p=0.016) Normal 961 95.8 1,237 97.3 ^Abbreviations: GBP: group DIAB: diabetic class INS: insecticide exposure OCC: occupation EFKXR: drink-years 'lifetime alcohol consunption (total drink-years) not used as a covariate. b 79 missing drink-year participants: 4/39 Ranch Hands abnormal; 0/40 Comparisons abnormal. 11-12 �TftBLE 11-7. Sunmary "teble of Gtoup-by-Cbvariate Interactions for Neurological Variables Group Ranch Hands Variable Interaction Comparisons Adj. Relative Stratification Statistic Number Percent Number Percent Risk ( 5 C.I.) p-Value 9% n Abnormal Normal 412 10 402 2.4 97.6 549 5 0.9 3.03 ( . 2 9 0 ) 0.045 10,.0 544 99.1 n Abnormal Nornal 568 47 521 8.3 91.7 693 70 1 . 0.82 (0.55,1-21) 0.319 01 623 8 . 99 Bom <$1922 n Abnormal Normal 36 4 32 11.1 88.9 50 9 18.0 41 82.0 ( . 5 ( . 6 1.97) 0.361 05 01, Abnormal n Abnormal Normal 76 13 63 17.1 82.9 94 10 84 1.74 ( . 1 4.24) 0.223 07, n Abnormal Normal 105 1 104 1.0 99.0 n Abnormal Norual 822 45 777 5.5 9. 45 n Abnormal Normal 703 22 681 3.1 9. 69 683 8 1.2 2.60 ( . 5 5 9 ) 0.022 11,.0 675 9 . 88 Group-byInsecticides Exposure Not Exposed , n to Insecticide Abnormal Nonral 313 4 309 1.3 98.7 605 11 1.8 0 6 ( . 2 2 1 ) 0.532 .9 02,.9 594 98.2 Bom > 1942 Neck Range Group-byof Motion Age Pin Prick Group-byDiabetic Class Bom 1923-1941 Impaired Normal Exposed to Insecticides Tremor 11-13 10.6 89.4 174 16 9.2 0.09 ( . 1 0.69) 0.021 00, 158 90.8 1,005 53 5.3 952 94.7 1.02 ( . 8 1 5 ) 0.920 06,.4 �The stratified analysis for neck range of motion showed a higher proportion of younger Ranch Hands with neck range of motion abnormalities than younger Comparisons (p=0.045). Although not statistically significant, middle-aged and older Comparisons had higher proportions of abnormalities than did the Ranch Hands. Peripheral Nerve Status Peripheral nerve integrity was assessed by light pin prick, light touch (cotton sticks), visual inspection (and palpation, if indicated) of muscle mass, vibratory sensation as measured at the ankle with a tuning fork of 128 Hz, three deep tendon reflexes (patellar, Achilles, and biceps), and the Babinski reflex. The unadjusted analyses are given in Table 11-8. As noted previously, the analyses of pin prick, light touch, and vibratory sensation excluded the 29 participants with peripheral edema. These results showed that peripheral nerve function did not vary significantly by group. Adjusted analyses were performed by logistic regression on four peripheral nerve variables. The other variables had relatively sparse numbers of abnormalities. The covariates were age, race, occupation, drink-years of alcohol, diabetic class, and exposure to insecticides. These statistics are displayed in Table 11-9. For the variables light touch, muscle status, and the Achilles reflex, group differences were nonsignificant; the results were nearly identical to the unadjusted analyses. For the variable pin prick, however, a significant group-by-diabetic class interaction (p=0.003) was observed. This interaction was explored and the results are depicted in Table 11-7. As shown, the interaction suggests a difference, due to a lower proportion of abnormal pin-prick results in Ranch Hand impaired diabetics than in Comparisons (Adj. RRs 0.09,95% C.I.J [0.01,0.69], p=0.021), whereas both the abnormal and normal diabetic classes showed no significant group differences. Central Nervous System Coordination CNS coordination was evaluated clinically with four variables: hand tremor, rapid finger-to-nose coordination, one-foot standing balance (modified Romberg sign), and observation of gait for at least 10 steps. In addition, a constructed variable, the CNS summary index, was derived by summarizing abnormalities from all four CNS variables. The unadjusted analyses of these five variables are shown in Table 11-10. These results revealed no statistically significant group differences for the four primary CNS variables, although the borderline significance of tremor, with a higher proportion of abnormalities in the Ranch Hands, is interesting. The statistical power to detect a given relative risk was poor because of the small percentages of abnormalities. The CNS summary index was statistically significant, with Ranch Hands manifesting a higher proportion of abnormalities; this result should be interpreted with caution, however, si'nce this index was constructed after the data were examined. Three of the five variables with sufficient proportions of abnormalities were adjusted by six covariates, and these results are summarized in Table 11-11. 11-14 �TABLE 11-8. Unadjusted Analyses for Peripheral Nerve Function by Group Group Ranch Hand Variable Comparison Statistic Number Percent Number Percent Est . Relative Risk (95% C.I.) p-Value n Abnormal Normal 1,003 59 944 5.9 94. 1 1,276 80 1,196 6.3 93.7 0.93 (0.66,1 .32) 0.725 Light Touch n Abnormal Normal 1,003 38 965 3.8 96.2 1,276 47 1,229 3.7 96.3 1.03 (0.67,1 .59) 0.912 Muscle Status n Abnormal Normal 1,016 26 990 2.6 97.4 1,292 33 1,259 2.6 97.4 1.00 (0.60,1 .69) 0.999 Vibratory Sensation n 1,003 11 Abnormal Normal 992 1. 1 98.9 1,276 10 1,266 0.8 99.2 1.40 (0.59,3 .32) 0.510 Patellar Reflex n Abnormal Normal 1,016 1. 1 98.9 1,290 16 1,274 1.2 98.8 0.87 (0.40,1 .89) 0.846 Achilles Reflex n Abnormal Normal 1,009 5.7 94.3 1,284 75 1,209 5.8 94.2 0.98 (0.69,1 .40) 0.999 Biceps Reflex n Abnormal Normal 1,016 0.9 99.1 1,292 10 1,282 0.8 99.2 1.15 (0.46,2 .83) 0.819 n Abnormal Normal 1,011 0.4 99.6 1,287 5 1,282 0.4 99.6 1.02 (0.27,3 .80) 0.999 Pin Prick Babinski Reflex 11 1,005 58 951 9 1,007 4 1,007 11-15 �TfiBUBll-9. Adjusted Analyses for Selected Variables of Peripheral Nerve Ruction by Group Group Ranch Hand Comparison Statistic Number Percent ftnfcer Percent Pin Prick n Abnormal Normal 1,003 59 944 5.9 94.1 1,273 79 1,194 n Abnormal Normal 964 37 927 3.8 96.2 1,236 46 1,190 n Abnormal Normal 977 25 952 2.6 97.4 1,248 31 1,217 2.5 97.5 Achilles Reflex n Abnormal Normal 971 56 915 5.8 94.2 1,240 71 1,169 5.7 94.3 **** 3.7 96.3 Muscle Status p-Value 6.2 93.8 Light Touch Adj. Relative Risk (95XC.I.) **** Variable presented. Covariate Remarks GRP*DIAB(pd0.003) AGE(p<D.001) 1.02 (0.65,1.60) 0.921 OCC*RACE(p=0.013) AGE(p=0.043) EROR(p=0.031) 1.00 (0.57,1.75) 0.999 EROR*AGE(p=0.009) DIAB*INS(p4).039) 1.00 (0.69,1.45) 0.999 DRKXR*OCC(p=0.016) AGE(p<0.001) DIAB(p<D.001) interaction—adjusted relative risk, confidence interval, and p-value are not 11-16 �TABLE 11-10. Unadjusted Analyses for CNS Coordination Variables by Group Group Ranch Hand Variable Tremor Comparison Statistic Number Percent Number Percent Est. Relative Risk (95% C.I.) p- Value n Abnormal Normal 1,016 26 990 2.6 97.4 1,292 19 1,273 1.5 98.5 1.76 (0.97 ,3.20) 0.069 n Abnormal Normal 1,015 9 1,006 0.9 99.1 1,292 7 1,285 0.5 99.5 1.64 (0.61 ,4.43) 0.327 Romberg Sign n 1,015 2 Abnormal Normal 1,013 0.2 99.8 1,292 1 1,291 0.1 99.9 2.55 (0.23 ,28.15) 0.586 Gait n Abnormal Normal 1,016 20 996 2.0 98.0 1,290 16 1,274 1.2 98.8 1.60 (0.82 ,3.10) 0.178 n Abnormal Normal 1,015 48 967 4.7 95.3 1,290 39 1,251 3.0 97.0 1.59 (1.04 ,2.45) 0.036 Coordination CNS Summary Index 11-17 �TfiHE 11-11. Adjusted Analyses for Selected Variables of CMS Coordination by Group Group Ranch Hand "Variable Tremor Gait CNS Summary Index Comparison Adj. Relative Statistic Number Percent Number Percent Risk (95% C.I.) p-Value Covariate Remarks* n 1,016 Abnormal 26 Normal 990 2.6 97.4 1,288 19 1,269 1.5 98.5 1.70 ( . 3 3 0 ) 0 0 0 09,.9 .8 GRP*INS (marginal:p=0.055) DIAB(p=0.001) 977 20 957 2.0 98.0 1,246 15 1,231 1.2 98.8 1.74 ( . 8, . 7 0.110 08 3 4 ) DIAB(p=0.030) ERKXR*INS(p=0.047) n 1,015 Abnormal 48 967 Normal 4.7 95.3 1,286 38 1,248 3.0 97.0 1.57 (1.01,2.43) 0.042 DIAB(p=0.003) OCC(P=0.018) n Abnormal Normal These statistics were quite similar to the unadjusted tests, and showed borderline significance for tremor, nonsignificance for gait, and significance for the CNS summary index. The unexpected inverse relationship of tremor abnormalities to diabetic classification is again noted. The borderline group-by-insecticide interaction was investigated, and the results are given in Table 11-7. As shown, the relative risk for Ranch Hands exposed to insecticides was statistically significant (RR: 2.60, 95% C.I.: [1.15,2.90], p=0.022), whereas the relative risk for unexposed Ranch Hands was nonsignificant. This finding may have both an operational and biologic foundation, because records indicate that some Ranch Hands were exposed to the insecticide Malathion®, a cholinesterase inhibitor, during insecticide missions for malaria prevention. Comparisons, by definition, did not fly these missions. EXPOSURE INDEX ANALYSES Exposure index analyses were conducted within each occupation cohort of the Ranch Hand group to search fpr dose-response relationships (see Chapter 8 for details on the exposure index). All 27 variables and three summary indices were explored (unadjusted for any covariates) as with the unadjusted tests for group differences discussed previously in this chapter. These variables were investigated using Pearson's chi-square test and Fisher's exact 11-18 �test. Adjusted analyses were performed by logistic regression for the 10 variables (7 neurological parameters and 3 summary indices) for which adjusted analyses of group differences were previously examined. These analyses were accomplished, adjusted for age, diabetic class, insecticide exposure, and drink-years (all discretized), and any significant pairwise interactions between the exposure index and these covariates. Race was not included in adjusted analyses because of the absence of any race effect in the previous group difference analyses. Overall significance in the proportion of abnormalities among the exposure index levels of low, medium, and high was determined, as well as contrasts in the proportion of abnormalities between the medium and low exposure levels, and between the high and low exposure levels. Exclusions were made as described previously. Results of the adjusted analysis are presented in Table 11-12, and results for unadjusted analyses appear in Table 1-1 of Appendix I. Results from further study of exposure index-by-covariate interactions are given in Table 1-2 of Appendix I. Unadjusted analyses revealed borderline significant differences among exposure index levels for pin prick in enlisted groundcrew (p=0.052) and Achilles reflex in enlisted flyers (p=0.059). The data did not support an increase in the proportion of abnormalities with increasing exposure levels, however. Adjusted analyses yielded similar conclusions, in that significant or borderline significant results did not support an increase in the proportion of abnormalities with increasing exposure, and that very few significant results were observed. The pattern of abnormalities with the 10 variables was studied, and in no occupational strata was an increasing dose-response relationship evident. In fact, the high exposure level often had a smaller (although nonsignificant) proportion of abnormalities than the low and medium levels. Interactions were present for 5 of the 10 variables, and occurred primarily in the enlisted groundcrew stratum. A summary of these interactions is presented in Table 11-13. Meaningful interpretation of the interactions was difficult, due to the small numbers of abnormalities within a covariate strata. No significant adverse effects to participants with higher exposure levels were evident, however, in this analysis. In summary, no evidence of an increasing dose-response relationship at the followup examination was observed. No increase in prevalence rates was seen as exposure levels increased. These results essentially were in agreement with the findings of the Baseline Study. 11-19 �TABLE 11-12. Adjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj. Relative Risk ( 5 C.I.) p-Value 9% Officer 127 120 Overall M vs. L H vs. L 096 .0 0.82 (0.31,2.18) 0.686 0.97 (0.37,2.56) 0.955 Enlisted Flyer 51 61 53 Overall M vs. L H vs. L 0.940 0.79 (0.20,3.20) 0.744 0.83 (0.21,3.31) 0.786 Enlisted Groundcrew 148 160 132 Overall M vs. L H vs. L 0.299 0.93 (0.27,3.21) 0.908 0.36 (0.09,1.51) 0.163 Officer Neck Range of Motion 125 120 127 119 Overall M vs. L H vs. L 0.551 0.63 (0.28,1.44) 0.277 0.78 (0.35,1.78) 0.560 Enlisted Flyer 51 60 53 Overall M vs. L H vs. L 0.808 1.00 (0.29,3.43) 0.999 0.68 (0.18,2.59) 0.569 145 158 131 Overall M vs. L H vs. L **() **! **() **! H» g Cranial Nerve Function Index Enlisted Groundcrew **() **! **() **! **() **! �TABLE 11-12. (continued) Adjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj. Relative Risk (95% C.I.) p-Value 120 127 119 Overall M vs. L H vs. L 0.148 0.30 (0.08,1.22) 0.093 0.36 (0.09,1.45) 0.150 51 60 53 Overall M vs. L H vs. L 0.860 1.04 (0.13,8.27) 0.969 0.56 (0.05,6.58) 0.642 145 158 131 Overall M vs. L H vs. L 0.894 0.75 (0.23,2.45) 0.639 0.84 (0.25,2.76) 6.773 Officer 124 124 119 Overall M vs. L H vs. L 0.277 0.43 (0.13,1.38) 0.156 0.49 (0.17,1.43) 0.191 Enlisted Flyer 51 60 53 Overall M vs. L H vs. L 0.399 0.33 (0.05,2.35) 0.267 1.02 (0.23,4.60) 0.979 146 159 128 Overall M vs. L H vs. L 0.108 0.86 (0.32,2.34) 0.765 0.28 ( . 7 1 0 ) 0.062 00,.7 Officer Cranial Nerve Enlisted Function Flyer (Neck Range of Motion Excluded) Enlisted Groundcrew Pin Prick Enlisted Groundcrev �TABLE 11-12. (continued) Adjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Officer Light Touch 124 124 119 Overall M vs. L H vs. L 0.047 0.39 (0.11,1.40) 0.148 0.20 ( . 5 0 8 ) 0.027 00,.3 Enlisted Flyer 51 60 53 Overall M vs. L H vs. L ****(2) ****(2) ****(2) ****(2) ****(2) Enlisted Groundcrew 146 159 128 Overall M vs. L H vs. L 0.777 1.27 (0.34,4.80) 0.725 0.74 (0.16,3.35) 0.699 Officer ,_, i-1 125 127 120 Overall M vs. L H vs. L 0.105 0.15 (0.02,1.01) 0.051 0.57 (0.14,2.30) 0.433 Enlisted Flyer 51 61 53 Overall M vs. L H vs. L 0.979 0.90 (0.04,22.10) 0.946 0.74 (0.04,14.77) 0.841 148 160 132 Overall M vs. L H vs. L ****(3) ****(3) to Muscle Status Enlisted Groundcrew ****(3) ****(3) ****(3) �TABLE 11-12. (continued) Adjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj . Relative Risk (95% C.I.) p-Value Officer Achilles Reflex 122 126 120 Overall M vs. L H vs. L 0.384 0.43 (0.13,1-46) 0.175 0.65 (0.21,1-99) 0.448 Enlisted Flyer 51 60 53 Overall M vs. L H vs. L 0.021 0.65 (0.16,2.76) 0.564 Overall M vs. L H vs. L **() **3 **() **3 ****(3) ****(3) **() **3 Enlisted Groundcrew 147 160 132 Officer 125 127 120 Overall M vs. L H vs. L 0.219 0.19 ( . 2 1 6 ) 0.132 00,.6 0.63 ( . 4 2 8 ) 0.548 01,.9 Enlisted Flyer 51 61 53 Overall M vs. L H vs. L 0.625 2.11 (0.19,23.39) 0.542 2.95 (0.29,30.43) 0.364 148 160 132 Overall M vs. L H vs. L 0.396 0.91 (0.22,3.66) 0 8 9 .8 0.28 (0.03,2.44) 0.248 to u> Tremor Enlisted Groundcrew �TABLE 11-12. (continued) Adjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj . Relative Risk ( 5 C.I.) 9% p-Value Officer Gait 125 127 120 Overall M vs. L H vs. L 0.483 0.26 (0.02,3.25) 0.298 0.89 (0.12,6.76) 0.912 Enlisted Flyer 51 61 53 Overall M vs. L H vs. L 0.188 0.64 ( . 7 6 0 ) 0.693 00,.5 Enlisted Groundcrew ^ 148 160 132 Overall M vs. L fl vs. L 0.576 0.42 (0.07,2.51) 0.343 0.88 (0.19,3.99) 0.868 Officer M 125 127 120 Overall M vs. L H vs. L 0.123 0.22 (0.04,1.10) 0.066 0.57 (0.15,2.10) 0.399 Enlisted Flyer 51 60 53 Overall M vs. L H vs. L 0.930 1.21 (0.25,5.92) 0.818 0.90 (0.17,4.80) 0.899 148 160 132 Overall M vs. L H vs. L ****(2) ****(2) to -JS CNS Summary Index Enlisted Groundcrew **() **2 ****(2) ****(2) —No abnormal participants present in medium exposure index level for Achilles reflex (or high level for gait) in enlisted flyers. ****(l)Exposure index-by-diabetic class interaction—relative risk and p-value not presented. ****(2)Exposure index-by-insecticide exposure interaction—relative risk, confidence interval, and p-value not presented. ****(3)Exposure index-by-age interaction—relative risk, confidence interval, and p-value not presented. �TABLE 11-13. Summary of Exposure Index-by-Covariate Interactions for Neurological Variables p-Value Variable Occupation Covariate CNF Summary Index Enlisted Groundcrew Diabetic Class 0.045 Light Touch Enlisted Flyer Insecticide Exposure 0.026 Muscle Status Enlisted Groundcrew Age 0.026 Achilles Reflex Enlisted Groundcrew Age 0.014 CNS Summary Index Enlisted Groundcrew Insecticide Exposure 0.010 LONGITUDINAL ANALYSES Two variables, the modified Romberg sign and the Babinski reflex, were investigated to assess longitudinal differences between the 1982 Baseline examination and the 1985 followup examination. Both variables were classified as abnormal or normal. As shown in Table 11-14, 2x2 tables were constructed for each group for each variable. This table shows the number of participants who were abnormal at the Baseline examination and abnormal at the followup examination, abnormal at Baseline and normal at the followup, normal at Baseline and abnormal at the followup, and normal at both Baseline and the followup. The odds ratio is the ratio of the number of participants who were normal at Baseline and abnormal at the followup to the number of participants who were abnormal at Baseline and normal at the followup (the "off-diagonal" elements). The p-value was derived from Pearson's chi-square test of the hypothesis that there was comparable change in the two groups over time. These data showed no longitudinal difference in the change pattern in the Romberg sign in the two groups, but they did show a significant change in the Babinski reflex. In the Baseline examination, the Ranch Hands had a significantly greater proportion of reflex abnormalities than the Comparisons, but the followup examination showed approximately the same percentage of abnormality in both groups (Est. RR: 1.02, 95% C.I.s [0.27,3.80, p=0.999]). SUMMARY AND CONCLUSIONS Interval questionnaire data (1982 through 1985) on neurological illnesses, verified by medical records, revealed no significant group differences. These data were added to verified Baseline historical information to assess possible differences in the lifetime experience of neurological disease. Again, there was no significant difference between the Ranch Hand and Comparison groups. 11-25 �TABLE 11-14. Longitudinal Analysis of Romberg Sign and Babinski Reflex: A Contrast of Baseline and First Followup Examination Abnormalities Group Variable 1982 Baseline Exam 1985 Followup Exam Odds p-Value Abnormal Normal Ratio (OR)* (ORPH vs. OR^ K it C Ranch Hand Abnormal Normal 2 0 188 777 0 Comparison Abnormal Normal 0 1 250 886 0.004 Ranch Babinski Hand Reflex Comparison Abnormal Normal 1 3 7 953 0.43 Abnormal Normal 0 1 1,129 5.00 Romberg Sign 0.38 *0dds Ratio: 0.04 5 Number Normal Baseline, Abnormal Followup Number Abnormal Baseline, Normal Followup. A detailed neurological examination evaluated neurological integrity in three broad areas: cranial nerve function, peripheral nerve function, and central nervous system (CNS) coordination. The summary analytic results for all measurement variables comprising these three functional areas are presented in Table 11-15. Assessment of the 12 cranial nerves was based on the measurement of 14 variables. Two summary indices were constructed. Both the unadjusted and adjusted analyses did not disclose any statistically significant group differences, although two variables, speech and tongue position, were of borderline significance, with Ranch Hands faring worse than Comparisons. One of the two cranial nerve summary indices was marginally significant, again with the Ranch Hands at a slight detriment. The unadjusted and adjusted analyses of peripheral nerve function, as measured by eight variables (four reflexes, three sensory determinations, and muscle mass), did not reveal significant group differences. CNS coordination was evaluated by four measurements and a constructed summary variable. Hand tremor was found to be of borderline significance, with the Ranch Hands faring slightly worse than the Comparisons. The CNS summary index showed a significant detriment to the Ranch Hands. The exposure analyses for neurological variables with reasonable counts of abnormalities showed only occasional statistically significant results. No consistent pattern with increasing exposure was evident for any occupational category of the Ranch Hand group. 11-26 �TABLE 11-15. Overall Summary Results of Unadjusted and Adjusted Analyses of Neurological Variables Variable Unadjusted Adjusted Direction of Results** Questionnaire" Physical Examination Neurological Disease (Interval) Neurological Disease (History) NSb NS Cranial Nerve Function Smell Visual Fields NS NS Light Reaction Ocular Movements Facial Sensation Corneal Reflex NS NS NSc — Jaw Clench Smile NS NS Palpebral Fissures NS Balance Gag Reflex NS NS Speech Tongue Position Relative to Midline Palate and Uvula Movement Neck Range of Motion Cranial Nerve Function Index Cranial Nerve Function Index (excluding Neck Range of Motion) NS* RH>C NS* NS NS NS RH>C NS NS NS* NS* Peripheral Nerve Function Pin Prick Light Touch Muscle Status Vibratory Sensation Patellar Reflex Achilles Reflex Biceps Reflex Babinski Reflex NS NS NS NS NS NS NS NS 11-27 **** NS NS NS RH>C �TABLE 11-15. (continued) Overall Summary Results of Unadjusted and Adjusted Analyses of Neurological Variables Variable Unadjusted Adjusted Direction of Results** Central Nervous System Coordination Tremor Coordination Romberg Sign Gait CNS Summary Index NS* NS NS .NS 0.036 NS* RH>C — — NS 0.042 RH>C **RH>C: More abnormalities in Ranch Hand group than in Comparison group. "Disease categories include: inflammatory diseases, heriditary and degenerative diseases, peripheral disorders, disorders of the eye, disorders of the ear, and other disorders. NS:Not significant (p>0.10). No inflammatory diseases noted; borderline significant (p=0.069, RH>C) for other disorders; not significant for remaining categories. —Analysis not performed because of sparse number of abnormalities. c No abnormalities present. NS*Borderline significant (0.05<p<0.10)< Constructed variable. ****Group-by-covariate interaction. 11-28 �In a longitudinal analysis of the Romberg sign and the Babinski reflex, only the Babinski reflex revealed a significant difference between the Baseline and followup examination, with the Ranch Hands converting from significant adverse findings at Baseline to favorable nonsignificant findings at the followup examination. Overall, the followup examination findings are quite similar to the Baseline findings. However, several distinct patterns were evident from the analyses: (1) The followup examination detected substantially fewer abnormalities for almost all measurement variables, (2) the decrease in abnormalities was equivalent in both groups, (3) most of the covariate effects were classical, although exceptions were evident, (4) the adjusted analyses were uniformly similar to the unadjusted analyses, (5) the constructed summary variables were generally statistically significant, or of borderline significance (however some indices were created after the data were examined), and (6) although statistical significance at the pre-assigned a -level of 0.05 was not achieved for any of the measurement variables, abnormalities tended to cluster in the Ranch Hand group. Of the three group-by-covariate interactions in the adjusted analyses, only one, a borderline group-by-insecticide exposure interaction for hand tremor, where Ranch Hands exposed to insecticides had a marginally significant adverse effect, was of probable biologic (and operational) significance. In conclusion, none of the 27 neurological variables demonstrated a significant group difference, although several showed an aggregation of abnormalities in the Ranch Hand group, which merits continued surveillance. Historical reporting of neurologic disease was equal in both groups. The clinical sensitivity in detecting neurological deficits varied substantially between the Baseline and the followup examinations, but the number of statistically significant variables remained about the same. None of the exposure analyses revealed dose-response patterns in the Ranch Hand occupational categories. The longitudinal analyses disclosed a favorable reversal of significant Babinski reflex abnormalities at Baseline to nonsignificant findings at the followup examination for the Ranch Hands. The similarity in results between unadjusted and adjusted statistical tests is evidence of group equality for the traditionally important neurological covariates of age, alcohol, and diabetes. Of three group-by-covariate interactions in the adjusted analyses, only the Ranch Hand insecticide interaction with hand tremor was biologically plausible. 11-29 �CHAPTER 11 REFERENCES 1. Dougherty, J.A., G.E. Schulze, R.T. Taylor, and J. Blake. 1984. Behavioral toxicity of an agent orange component: 2,4-D. Oral presentation to the Veterans Administration Advisory Committee on Health-Related Effects of Herbicides, Washington, D.C., December 11, 1984. 2. Squibb, R.E., H.A. Tilson, and C.L. Mitchell. 1983. Neurobehavioral assessment of 2,4-dichlorophenoxyacetic acid (2,4-D) in rats. Neurobeh. Toxicol. Teratol. 5:331-335. 3. Desi, I., J. Sos, and I. Nikolits. 1962. New evidence concerning the nervous site of action of a chemical herbicide causing intoxication. Acta Physiol. 22:73-80. 4. Kim, C.S., L.A. O'Tuama, D. Mann, and C.R. Roe. 1983. Saturable cumulation of the anionic herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) by rabbit choroid plexus: Early developmental origin and interaction vith salicylates. J. Pharmacol. Exp. Ther. 225:699-704. 5. Goldstein, N.P., P.H. Jones, and J.R. Brown. 1959. Peripheral neuropathy after exposure to an ester of dichlorophenoxyacetic acid. JAMA 171(10):1306-1309. 6. Todd, R.L. 1962. A case of 2,4-D intoxication. J. Iowa Med. Soc. 52:663-664. 7. Berkley, M.C., and K.R. Magee. 1963. Neuropathy following exposure to a dimethylamine salt of 2,4-D. Arch. Int. Med. 111:133-134. 8. Berwick, P. 1970. 2,4-Dichlorophenoxyacetic acid poisoning in man. JAMA 214(6):1114-1117. 9. Wallis, W.E., A. Van Poznak, and F. Plum. 1970. Generalized muscular stiffness, fasciculations, and myokymia of peripheral nerve origin. Arch. Neurol. 22:430-439. 10. Park, J., I. Darrien, and L.F. Prescott. 1977. Pharmacokinetic studies in severe intoxication with 2,4-D and Mecoprop. Clin. Toxicol. 18:154-155. 11. Bauer, H., K.H. Schulz, and.U. Spiegelberg. 1961. Berufliche Vergif tungen bei der Herstellung von Chlorphenol-Verbindungen (Long-term hazards of polychlorinated dibenzodioxins and polychlorinated dibenzofurans). Arch. Gewerbephathol. Gewerbehyg. 18:538-555. Reported in IRAC (1978). 11-30 �12. Pazderova-Vejlupkova, J., M. Nemcova, J. Pickova, L. Jirasek, and E. Lukas. 1981. The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men. Arch. Environ. Health 36:5-11. 13. Oliver, R.M. 1975. Toxic effects of 2,3,7,8-tetrachloro-dibenzo1,4-dioxin in laboratory workers. Br. J. Ind. Med. 32:49-53. 14. Boeri, E., B. Bordo, P. Crenna, et al. 1978. Preliminary results of a neurological investigaton of the population exposed to TCDD in the Seveso region. Riv. Pat. Nerv. Ment. 99:111-128. 15. Singer, R., M. Moses, J. Valciukas, R. Lilis, and I.J. Selikoff. 1982. Nerve conduction velocity studies of workers employed in the manufacture of phenoxy herbicides. Environ. Res. 29:297-311. 16. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichloro-phenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 17. Filippini, G., B. Bordo, P. Crenna, N. Massetto, M. Musicco, and R. Boeri. 1981. Relationship between clinical and electrophysiological findings and indicators of heavy exposure to 2,3,7,8-tetrachlorodibenzo-dioxin. Scand. J. Work Environ. Health 7:257-262. 18. Flicker, M.R., and A.L. Young. 1983. Evaluation of veterans for agent orange exposure. Presented at the Symposium on Chlorinated Dioxins and Dibenzofurans in the Total Environment given before the Division of Environmental Chemist-ry, American Chemical Society, Washington, D.C., September 1983. 11-31 �CHAPTER 12 PSYCHOLOGICAL ASSESSMENT INTRODUCTION Emotional illnesses or psychological abnormalities are not recognized as primary clinical endpoints following exposure to chlorophenols, phenoxy herbicides, and dioxin. "Neurobehavioral effects" occasionally ascribed to such exposures have been, in fact, predominantly neurological symptoms for which causation is not disputed (see Chapter 11). Higher CNS functioning, in terms of cognitive skills, personality, and reactivity, may be temporarily or permanently impaired depending on the exposure and the ability to measure accurately the psychological changes. Animal studies provide little insight into possible human psychological problems. Animal signs of lethargy, stupor, poor coordination, lack of feeding, and agitation have been observed in multiple studies involving many species. These signs have generally been attributed to the "wasting syndrome" or multi-organ toxicity, rather than primary CNS toxicity. A study of "behavioral" effects in rats following single and weekly doses of 2,4-D showed that the central effects of decreased coordination and lever-pressing behavior were transient and reversible. Further, no latent CNS impairment was detected after a d-amphetamine challenge. Human studies and case reports have occasionally noted psychological disorders or symptom complexes following exposure to herbicides and TCDD. Complaints included headache, anxiety, malaise, depression, abnormal anger, mood changes, sleep disturbances, decreased libido, and impotence. Scientific confirmation of these symptoms by psychological testing is difficult and exclusion of other plausible causes such as age, preexisting psychological abnormalities, or even motivation for compensation is often impossible. Most studies have merely recorded complaints and have not pursued their validation by indepth functional testing. Early studies of industrial chemical workers first provided the suggestion of psychological effects. Followup studies from the Nitro, West Virginia, accident in 1949, showed "nervousness," fatigue, irritability, cold intolerance, and decreased libido in many of the workers with chloracne, but most of these symptoms subsided over a 4-year period. ' Two followup studies in 1979, by different investigators of expanded (but slightly different) plant cohorts, noted reports of sexual dysfunction and decreased libido. ' One of these studies noted that these observations (and insomnia) were significantly increased in individuals with chloracne. Neither of these followup efforts conducted, neurobehavioral tests to validate the reported symptoms. Other industrially based studies reported symptoms of fatigue, decreased libido, impotence, sleep disturbances, ' 1reduced emotional responses, sensory deficits of smell, taste, and hearing, reading 12-1 �difficulties,9 memory loss,11 and emotional disorders.12'13 Symptoms of depression and anxiety have been associated with disfiguring chloracne. One study found a relationship between chloracne and hypomania as determined from the MMPI,1 and another noted that two of three chemists involved in the synthesis of TCDD developed marked personality changes. Although data interpretation problems exist, the Czechoslovakian 10-year followup study cited eight cases of severe dementia in exposed workers and reported that symptoms of anxiety and depression decreased over the followup period. A contemporary cross-sectional morbidity study of a mobile-home park, environmentally contaminated with dioxin, showed subclinical hepatic, hematologic, immunologic, and psychological changes in exposed residents. Significant abnormalities were recorded in the exposed group for the tension/ anxiety and anger/hostility scales of the profile of mood states (POMS) inventory, as well as the vocabulary subtest of the Wechsler adult intelligence scale (WAIS). However, functional testing by the Halstead-Reitan battery (HRB) did not reveal significant group differences. There was no way to differentiate between the primary effects of exposure and the secondary effects of media attention. In contrast to industrial cohorts, the study of chemically related psychological problems in veterans has proved more difficult because of the confounding effects of combat stress and the post-traumatic stress disorder (PTSD), and the uncertainty of exposure. Of almost 100,000 Vietnam veterans registered in the VA's Agent Orange Registry in 1983, 18 percent complained of "nervousness" and 10 percent cited personality disorders. A psychiatric review of 132 veterans included in the Registry, most of whom had been referred for treatment, disclosed a symptom hierarchy of sleep disorders (53%), mood depression (36%), suicidal thoughts (35%), and irritability (31%). Fifty-three percent of these veterans received the PTSD diagnosis. In 1980, the American Psychiatric Association established the term "post-traumatic stress disorder" to define a neurosis caused by extreme psychic trauma, e.g., natural disaster, war, imprisonment, or torture. PTSD comprises the symptoms of anxiety, "powder keg" anger, depression, irritability, restlessness, recurrent intrusive dreams, flashbacks, and sleeplessness. Quiescent PTSD may be acutely reactivated in some individuals by specific triggering events (e.g., visiting the Vietnam Memorial). The disorder is equally applicable to civilians following emotionally traumatic experiences. The onset of PTSD may immediately follow the traumatic event or it may occur years afterward. The older war terms shell shock, combat fatigue, and anxiety reaction generally referred to the more immediate symptoms following the trauma although components of PTSD are now recognized in veterans of earlier wars. The prevalence of PTSD in Vietnam veterans is unknown, and even the qualitative assessments of "common" or "rare" are debatable. 1 > 2 A 7-month incidence of legal and emotional maladjustments in returning Vietnam veterans occurred at the rate of 23 percent and did not differ significantly from comparable rates in nonveterans. Though a concise definition of PTSD exists, there is controversy as to the best means of diagnosis. Some workers prefer a full and thorough clinical interview while others favor empiric symptom scales. Clearly, each method serves a different, but highly related, purpose: clinical diagnosis in individuals versus an epidemiological/ statistical diagnosis in groups. 12-2 �Risk factors for the development of PTSD may include emotional predisposition, social/ethnic background, parental factors, race, and combat intensity ranging from slight involvement to atrocity behavior. ' ' Parallel conditions to PTSD (or perhaps unrecognized components of PTSD) encompass alcoholism, drug abuse, lawlessnes^ (arrests/felony convictions), personality disorders, and frank psychosis. ' ~ This chapter attempts to isolate any psychological disorders attributable to herbicide exposure. Baseline Summary Results Extensive psychological parameters were assessed on all participants during the 1982 Baseline questionnaire and physical examination. The expected high degree of concordance between education (college, high school) and military status (officer, enlisted) was observed and validated the sole use of education as a covariate representing socioeconomic status for most analyses. There were no questionnaire differences for past history of emotional or psychological illnesses between the Ranch Hand and Comparison groups. For the psychological indices of fatigue, anger, erosion, anxiety, and severity of depression (as determined by a modification of the Diagnostic Interview Schedule ), no group differences were detected among the college-educated Ranch Hands. However, for the high school-educated stratum, Ranch Hands demonstrated highly significant pathology for fatigue, anger, erosion, and anxiety. An unadjusted analysis of reported depression showed significantly more depression in the Ranch Hands, as did the isolation index adjusted for educational level. Exposure index analyses from the Ranch Hand questionnaire data did not suggest a relationship between exposure and psychological abnormality. At the time of the physical examination, additional self-reported data were collected with the Cornell Index and the MMPI. The CNS functional testing was conducted by a modified HRB, and intelligence was measured by the WAIS. The Cornell Index showed a significant increase in psychophysiologic symptoms in the high school-educated Ranch Hands. Six of 10 parameters of the Cornell Index were abnormal in the Ranch Hands (e.g., fear, startle, psychosomatic) as contrasted to the Original Comparisons, and all abnormal responses/parameters were inversely related to education to a statistically significant degree. MMPI results in the high school-educated participants, showed differences in the scales of denial, hypochondria, masculinity/ femininity, and mania/hypomania as contrasted to the college-educated group. Only the social introversion scale was significant in the college-educated participants. The effect of. education was influential (p<0.01) in all scales of the MMPI. Race was not a significant covariate. All self-reported data, including those from the in-home questionnaire, were not adjusted for possible group differences in PTSD or combat experience/intensity. Performance testing by the HRB showed no neuropsychiatric impairment in the Ranch Hands as contrasted to their overall self-administered MMPI and Cornell Index. In fact, Ranch Hand over-reporting was suggested in several parameters, but was not proved. The effect of education on the HalsteadReitan testing was profound (p<0.0001). WAIS intelligence scores revealed very close group similarities in the full-scale and verbal and performance 12-3 �scales. As expected, the intelligence quotient (IQ) of college graduates was significantly higher than the IQ of high-school graduates. Exposure index analyses of the HRB and WAIS data were negative and disclosed no patterns that suggested an herbicide effect. Parameters of the 1985 Psychological Assessment Two of the psychological tests (MMPI, HRB) conducted at the 1982 Baseline examination were repeated at the first followup examination in 1985. Repetitive testing was accomplished for purposes of clinical validation, establishment of comparable longitudinal parameters, and comparable covariate adjustments by concurrently derived PTSD and combat experience indices. Questions from the Diagnostic Interview Schedule were deleted from the followup questionnaire and were replaced by questions on combat experience in Vietnam. An updated history of mental and emotional disorders was obtained on all participants. A PTSD indicator was derived from a new MMPI subscale and was used for covariate adjustments of non-MMPI psychological data. The WAIS IQ assessment was deleted, but all parameters of the MMPI and HRB were retained. The Cornell Medical Index (CMI) was substituted for the Cornell Index in the 1985 psychological assessment. The dependent variables and covariates of the followup examination are similar to those analyzed at the Baseline. Longitudinal analyses of the MMPI scales of denial and depression consider the change of psychological test indices between groups. All statistical analyses are based on 1,016 Ranch Hands and 1,293 Comparisons. No individuals were excluded from the analysis of the psychological data for medical reasons. Sample size differences in the tables below reflect missing data from scale or battery test results, or from relevant covariates. The statistical tests use log-linear models, logistic regression models, Kolmogorov-Smirnov nonparametric tests, Fisher's exact test, and Pearson's chi-square test. Parallel analyses using Original Comparisons are in Tables J-8 through J-18 of Appendix J. RESULTS AND DISCUSSION Questionnaire Data At the followup interview, each participant was asked whether he had ever had a mental or emotional disorder. Whenever possible, the conditions were coded using ICD-9-CM. Reported disorders for which treatment was obtained were subsequently verified by reviews of medical records. Table 12-1 contains a tabulation of the distribution of these psychological illnesses, with information from the Baseline and followup studies combined. None of the types of illness categories showed statistically significant differences between groups; however, the "other neuroses" category is significant (p=0.037), with the Ranch Hands showing more adverse effects, when only Original Comparisons are used (see Table J-8 of Appendix J). 12-4 �TABLE 12-1. Unadjusted Analyses for Reported Psychological Illnesses by Group: Baseline and First Followup Studies Combined* Group Abnormalities Ranch Hand Type of Illness Psychoses Alcohol Dependence Anxiety Other Neuroses Comparison Number Percent 14 9 7 72 1.4 0.9 0.7 7.1 Number 9 8 13 74 Percent 0.7 0.6 1.0 5.7 Total 23 17 20 146 p-Value** 0.138 0.473 0.501 0.197 *Analyses based on 1,016 Ranch Hands and 1,293 Comparisons; some participants may have had more than one illness. **Fisher's exact test. Psychological Examination Data The MMPI is a self-administered test consisting of 566 questions on various aspects of behavior and personality. The results of the MMPI are numerical scores for 14 scales. The scales are anxiety (psychasthenia), consistency (F-scale), defensiveness (L-scale), denial (K-scale), depression, hypochondria, hysteria, mania/hypomania, masculinity/femininity, paranoia, psychopathic/deviate, schizophrenia, social introversion, and validity. The normal range of scores from 30 to 70 was used to categorize the results as normal or abnormal for all scales except validity. For validity (the number of unanswered questions) categories of 0 or greater than 0 were used. The test was administered to all 2,309 participants. A participant was considered nonresponsive in the MMPI if more than 30 questions (approximately 5%) were unanswered. Due to nonresponse, data on six participants, (two Ranch Hands and four Comparisons) were omitted from the analysis of all variables except validity. Thus, the MMPI analyses were based on 1,014 Ranch Hands and 1,289 Comparisons. The CMI is a self-administered instrument used to collect a substantial amount of medical and psychiatric data. The 195 questions of the CMI are partitioned into 18 sections (A to R) with the number of questions within a section ranging from 6 to 23. The analysis of the CMI was based on three scores: the total CMI score, an M-R subscore, and an A-H area subscore. The total CMI score is the number of affirmative responses on the entire questionnaire and is analyzed as a continuous variable. The M-R subscore, which deals with mood and feeling patterns, is a useful indicator of 12-5 �emotional ill-health. This subscore is the total number of affirmative responses to the 51 questions in sections M-R and is trichotomized as 0, 1 to 10, or greater than 10 for the analysis. The A-H area subscore is a measure of the scatter of complaints, indicating a diffuse medical problem, although other interpretations are possible. An abnormal A-H area subscore is defined as the number of sections (of A-H) with three or more affirmative responses. The A-H area subscore, which ranges from 0 to 8, is trichotomized as 0, 1 to 3, or 4 to 8 for the analysis. Consistent with the 5 percent nonresponse exclusion used for the MMPI, analysis of the total CMI score is based on scores with at least a 95 percent response rate or no more than 10 unanswered items from the total 195. M-R subscores are deleted from the analyses if three or more questions were unanswered from the 51 questions. For the A-H area subscore, participants who failed to answer all items were excluded from the analyses. Using these response criteria, analyses of the total CMI score are based on the scores of 1,000 Ranch Hands (16 deleted) and 1,268 Comparisons (25 deleted); the M-R subscore analyses use the results of 998 Ranch Hands (18 deleted) and 1,267 Comparisons (26 deleted); and the A-H area subscore analyses use 914 Ranch Hands (102 deleted) and 1,148 Comparisons (145 deleted). The HRB is a neuropsychological test that was administered to all participants to assess the functional integrity of the CNS. The battery consists of seven subtests: category (abstract recognition and analysis), total-time tactile performance, memory tactile performance, localization tactile performance, rhythm, speech, and finger tapping. In addition, other tests were performed (e.g., trailmaking, tests of recent memory) but do not contribute to the impairment index. For each participant who completed all seven s.ubtests, an impairment index, equal to the number of subtests in which the participant scored abnormally, is computed. This variable is dichotomized as normal (impairment index <3) or abnormal (impairment index X3). Twenty participants (10 in each group) refused or did not complete one or more of the seven subtests. Thus, the analyses of the HRB impairment index are based on data from 1,006 Ranch Hands and 1,283 Comparisons. Fisher's exact test was used to contrast the number of excluded participants between groups. A significant difference was not observed (p=0.654). The analyses of the psychological variables were adjusted for age (born in 1942 or after, born between 1923 and 1941, born in 1922 or before), race (Black, nonblack), education (high school, college), and drink-years (0, greater than 0 to 50, greater than 50). Education was dichotomized into high school and college categories, for purposes of analysis, from the classifications of (1) no high school diploma, (2) high school diploma, (3) attended college, and (4) college diploma. This variable was based on Baseline education levels, and participants with incomplete information were classified as high school educated. In addition, the analyses of the MMPI scales were adjusted for the combat index, a surrogate measure for PTSD. This index was constructed from 15 self-administered questions on combat experiences (see Appendix C, page C-15, AFHS Form 8). Associations of these 15 variables with PTSD, as measured from a subset of the MMPI questions, were examined, and responses to four questions showed statistically significant or marginally significant associations with PTSD. The four questions were (1) flew in aircraft that received battle damage, (2) had a close friend killed in action, (3) encountered mines or booby traps, and (4) wounded. An index, equal to the number of affirmative responses to these four questions, was computed and used as a trichotomized covariate (low, [0; n=708 (30.7%)], 12-6 �medium [1; n=814 (35.4%)], high [2-4; n= 781 (33.9%)], 6 missing participants, as with MMPI scales) for the analyses of the MMPI scales. While this index was associated with PTSD, it does not necessarily measure stress but does measure combat experience. The analyses of the CMI and HRB tests were adjusted for PTSD, based on the number of affirmative responses to a subset of 49 questions of the MMPI. For these analyses, PTSD was dichotomized as yes/no using greater than 30 affirmative responses2 as a positive indicator of PTSD. Sixteen participants (10 Ranch Hands, 6 Comparisons) were classified as having PTSD under this guideline. (Note that this indicator of PTSD was not used as a covariate for the analyses of MMPI scales, because the variable was based on the responses used in the calculation of the MMPI scores.) Current alcohol use (yes/no) and occupation were examined as potential covariates and are provided in the summary tables for inspection. Current alcohol use was highly correlated with drink-years, which better explained the dependent variables under study. Similarly, occupation was highly correlated with education (p<0.001). In this case, education was selected. Statistical Analysis Minnesota Multiphasic Personality Inventory (MMPI) The distributions of the Ranch Hand and Comparison groups for the 14 MMPI variables were contrasted using the Kolmogorov-Smirnov nonparametric tests and stratified by occupation (officer, enlisted flyer, enlisted groundcrew), for a total of 42 tests. Unadjusted analyses were performed using Fisher's exact test. Covariate analyses, using Fisher's exact or Pearson's chi-square test, were conducted for age, race, 'education, drinkyears, combat index, current alcohol use, and occupation. Logistic regression techniques were used to conduct the adjusted analyses. In the adjusted analyses, all covariates were used as discrete variables with the exception of age, which was used as a continuous variable. Current alcohol use and occupation were not used in the adjusted analysis. Using a two-sided a-level of 0.05, and with power of 0.80, the sample sizes are sufficient to detect a 38 percent increase in the rate of abnormal scores for depression, a 61 percent increase in the rate of abnormal scores for denial, and a 119 percent increase in the rate of abnormal scores for social introversion. Distributional Analyses The Kolmogorov-Smirnov tests identified no statistically significant differences between the Ranch Hand and Comparison distributions for the 14 MMPI variables at the 0.05 significance level for each occupational category. Only 2 of the 42 tests even approached significance, mania/ hypomania (Ranch Hand and Comparison officers, p=0.092) and psychopathic/ deviate (Ranch Hand and Comparison enlisted flyers, p=0.088). Results of the Kolmogorov-Smirnov tests are provided in Tables J-l to J-3 of Appendix J. It is noted that stratification by occupation reduced the sample size for each test and consequently decreased the power; that is, a larger maximum difference between the Ranch Hand and Comparison distributions is needed to show significance when the sample size is decreased, as is the case when stratification by occupation is performed. 12-7 �Unadjusted and Adjusted Analyses The unadjusted results, covariate tests of association, and adjusted results of the analyses for the 14 MMPI variables are summarized in Tables 12-2 to 12-4, respectively. Summary tables, which investigate interactions involving group, are provided in Table J-4 of Appendix J. The results of the tests of association for current alcohol use and occupation are presented in Table 12-3 for inspection, but are not discussed in the text since the measure of total drink-years was more appropriate for use in the analyses. Anxiety The unadjusted analysis showed no statistically significant difference in the anxiety scale between the Ranch Hands and the Comparisons (p=0.311). The tests of association with the covariates, using the pooled group categorical data, revealed statistically significant effects for age (p=0.010) and education (p<0.001). For age, 8.4 percent of the participants born in or after 1942 were scored as abnormal, as were 5.3 percent of those born from 1923 to 1941, and 4.6 percent of those born in or before 1922. The high school subgroup had a higher percentage (8.5%) of abnormalities than the college subgroup (4.4£). For the test of association, drink-years was marginally significant (p=0.058), based on the percent of abnormalities for 0, greater than 0 to 50, and greater than 50 drink-years: 10.0 percent, 5.9 percent, and 8.2 percent, respectively. In the adjusted analysis, there was no statistically significant difference between groups (p=0.512). In this analysis, education (EDUC) showed a statistically significant effect (p<0.001). The interaction, age-bycombat-index (CI), was also statistically significant (p=0.008). A group(GRP)-by-education interaction was marginally significant (p=0.057). Further investigation of this"interaction revealed an adjusted relative risk of 1.39 for the high school stratum and 0.68 for the college stratum. However, these relative risks were not significantly different from 1.00 (p=0.114, p=0.233, respectively). The exploration of this interaction is shown in Table J-4 of Appendix J. Consistency The unadjusted test of the MMPI consistency scale revealed no statistically significant difference between the Ranch Hand and Comparison groups (p-0.222). Based on the tests of association, education was statistically significant (p-0.010) with 3.9 percent abnormalities in the high school category and 2.0 percent abnormalities in the college category. In addition, the test of association with drink-years was statistically significant (p=0.021); the categories 0 and greater than 0 to 50 drink-years each had a percent abnormal frequency of 2.7, whereas there were 5.6 percent abnormalities in the greater than 50 drink-years category. In the adjusted analysis of the consistency scale, a group-by-education interaction was statistically significant (p=0.013). Further analysis of the interaction (shown in Table J-4 of Appendix J) revealed that the high school 12-8 �TABLE 12-2. Unadjusted Analyses for MMPI by Group Group Variable Anxiety Consistency to Defensiveness Denial Depression Hypochondria Hysteria Ranch Hand Statistic Number Percent Comparison Est. Relative Number Percent Risk (95% C.I.) p-Value n Abnormal Normal 1,014 73 941 7.2 92.8 1,289 79 1,210 6.1 93.9 1.19 (0.86,1.65) 0.311 n Abnormal Normal 1,014 36 3.6 978 . 96.4 1,289 34 1,255 2.6 97.4 1.36 ( . 4 2.19) 08, 0.222 n Abnormal Normal 1,014 23 991 2.3 97.7 1,289 35 1,254 2.7 97.3 0.83 (0.49,1.42) 0.592 n Abnormal Normal 1,014 17 997 1.7 98.3 1,289 58 1,231 4.5 95.5 0.36 (0.21,0.63) <0.001 n Abnormal Normal 1,014 114 900 11.2 88.8 1,289 126 1,163 9.8 90.2 1.17 (0.89,1.53) 0.272 n Abnormal Normal 1,014 119 895 11.7 88.3 1,289 129 1,160 10.0 90.0 1.20 (0.92,1.56) 0.198 n Abnormal Normal 1,014 123 891 12.1 87.9 1,289 125 1,164 9.7 90.3 1.29 (0.99,1.67) 0.067 �TABLE 12-2. (continued) Unadjusted Analyses for MHPI by Group Group Variable Ranch Hand Comparison Est. Relative Statistic Number Percent Number Percent Risk (95* C.I.) p-Value Mania/Hypomania n Abnormal Normal 1,014 63 951 6.2 93.8 1,289 88 1,201 6.8 93.2 0.90 (0.65,1.26) 0.611 Masculinity/ Femininity n Abnormal Normal 1,014 66 948 6.5 93.5 1,289 120 1,169 9.3 90.7 0.68 ( . 0 0.93) 05, 0.017 Paranoia n Abnormal Normal 1,014 31 983 3.1 96.9 1,289 28 1,261 2.2 97.8 1.42 (0.85,2.38) 0.187 Psychopathic/ Deviate n Abnormal Normal 1,014 120 894 11.8 88.2 1,289 149 1,140 11.6 88.4 1.03 ( . 0 1.33) 08, 0.845 Schizophrenia n Abnormal Normal 1,014 94 920 9.3 90.7 1,289 101 1,188 7.8 92.2 1.20 (0.90,1.61) 0.228 Social Introversion n Abnormal Normal 1,014 26 988 2.6 97.4 1,289 19 1,270 1.5 98.5 1.76 (0.97,3.20) 0.069 Validity n >0 0 1,016 224 792 22.0 78.0 1,293 271 1,022 21.0 79.0 1.07 (0.87,1.30) 0.540 NJ I �TABLE 12-3. Association Between MMPI Variables and the Covariates in the Combined Ranch Hand and Comparison Groups MMPI Scale Age Race Drink.- Combat Education Years Index Current** Alcohol Use Occupation** Anxiety 0.010 NS <0.001 NS* NS 0.001 <0.001 Consistency NS NS 0.010 0.021 NS NS <0.001 Defensiveness 0.028 0.025 0.001 <0.001 Denial 0.037 NS Depression NS NS Hypochondria 0.031 Hysteria <0.001 <0.001 NS* NS NS <0.001 0.002 NS NS <0.001 0.025 <0.001 0.041 0.044 <0.001 0.044 NS <0.001 0.006 NS 0.027 <0.001 Mania/Hypomania NS NS 0.011 0.001 NS 0.022 Masculinity/ Femininity 0.005 NS NS NS NS 0.005 Paranoia 0.022 NS NS NS NS NS* 0.014 Psychopathic/ Deviate NS 0.001 0.001 <0.001 NS NS* <0.001 Schizophrenia NS NS <0.001 NS Social Introversion 0.003 Validity NS <0.001 NS NS <0.001 NS 0.027 NS 0.014 NS NS* <0.001 NS* NS NS NS* <0.001 NS NS NS* NS NS - Not significant (p>0.10). *Borderline significant (0.05<p<0.10). **Not used in adjusted analyses. 12-11 NS �TABLE 12-4. Adjusted Analyses for MMPI by Group Group Adj. Relative Risk ( 5 C.I.) 9% p-Value Covariate Remarks* 1,285 1.12 (0.80,1.57) 0.512 EDUC (p<0.001) AGE*CI (p=0.008) GRP*EDUC (marginal: p=0.057) 974 1,246 **** **** 976 1,250 0.77 (0.45,1.33) 0.347 1,012 1,285 0.37 (0.21,0.66) <0.001 974 1,246 1.10 (0.84,1.45) 0.497 Hypochondria 1,012 1,285 1.12 (0.85,1.47) 0.431 AGE (p=0.002) RACE (p=0.026) EDUC (p<0.001) CI (p=0.043) Hysteria 1,014 1,289 1.27 (0.97,1.66) 0.077 AGE (p=0.003) EDUC (p<0.001) 974 1,246 0.80 (0.56,1.13) 0.203 DRKYR (p=0.006) AGE*CI (p=0.046) Variable Ranch Hand Total Comparison Total Anxiety 1,012 Consistency Defensiveness to h-1 NJ Denial Depression Mania/Hyporaania AGE (p=0.007) DRKYR (p=0.026) CI (p=0.041) GRP*EDUC (p=0.013) EDUC (p<0.001) DRKYR (p<0.001) EDUC*CI (p=0.044) EDUC (p<0.001) DRKYR (p=0.013) GRP*CI (marginal: p=0.055) �TABLE 12-4. (continued) Adjusted Analyses for MMPI by Group Group Variable Comparison Total Masculinity/ Femininity 1,014 Paranoia N5 Ranch Hand Total Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks* 1,289 0.69 ( . 0 0 9 ) 05,.5 0.020 EDUC (p<0.001) RACE*AGE (p=0.008) 1,012 1,285 **** **** AGE*CI (p=0.003) GRP*AGE (p=0.036) Psychopathic/ Deviate 974 1,246 (0.79,1.36) 0.780 EDUC (p=0.011) AGE*CI (p=0.003) RACE*DRKYR (p=0.015) Schizophrenia 976 1,250 **** **** RACE*DRKYR (p=0.017) GRP*EDUC (p=0.010) Social Introversion 1,012 1,285 **** **** AGE (p=0.004) GRP*CI (p=0.037) Validity 1,014 1,289 **** **** AGE*CI (p=0.030) GRP*RACE (p=0.012) 1.04 *Abbreviations: EDUC: education Cl: combat index GRP: group DRKYR: drink-years of alcohol ****Group-by-covariate interaction — adjusted relative risk, confidence interval, and p-value are not presented. �Ranch Hand category had a marginally significantly higher percentage of abnormal participants (5.6%) than the high school Comparisons (2.9%) (p=0.051). The adjusted relative risk for the high school classification was 1.81 with 95 percent confidence bounds of 1.00 and 3.28. In contrast, the percentage of abnormalities in the Comparison college-educated stratum was higher than the corresponding Ranch Hand subgroup (2.6 percent, 1.4 percent, respectively), but the difference was not statistically significant (p=0.110). Age, drink-years (DRKYR), and combat index were also statistically significant (p=0.007, p«0.026, p=0.041, respectively) in the adjusted analyses. Defensiveness For the MMPI defensiveness scale, there was no significant difference between groups, based on the unadjusted analysis (p=0.592). The tests of association showed statistically significant differences for all variables except combat index, which was marginally different statistically. The percentage of abnormalities for the age categories (born in or after 1942, born between 1923 and 1941, and born in or before 1922) were 3.3, 1.8, and 4.6, respectively (p=0.028). There were 2.3 percent abnormalities for nonblacks as compared to 5.6 percent for Blacks (p=0.025). The percent abnormalities for the high school- and college-educated categories were 3.8 and 1.0, respectively (p<0.001). For the 0 drink-years category, there were 10.0 percent abnormalities; the percent abnormalities for the greater than 0 to 50 and greater than 50 drink-years were 2.4 and 0.6, respectively (p<0.001). For combat index, which was only marginally statistically significant (p=0.093), the percent abnormalities were 3.5 for the low, 2.1 for the medium, and 1.9 for the high categorizations. In the adjusted analysis, there was no significant difference between the Ranch Hand and Comparison groups (p=0.347). In this analysis, the covariates of education (p<0.001) and drink-years (p<0,001) were statistically significant. Denial Based on the unadjusted analysis, .there was a statistically significant difference between the two groups on the MMPI denial scale (p<0.001), with 4.5 percent abnormalities in the Comparison group as contrasted to only 1.7 percent in the Ranch Hand group. The estimated relative risk was 0.36 with a 95 percent confidence interval of 0.21 to 0.63. The tests of association found only age as a statistically significant covariate (p=0.037). Men born in or after 1942 and those born between 1923 and 1941 had 3.0 percent and 3.1 percent abnormalities, respectively, as compared to 8.0 percent abnormalities for those born in or before 1922. The adjusted analysis showed a statistically significant difference between groups (p<0.001). The adjusted relative risk estimate was 0.37 with 95 percent confidence bounds of 0.21 and 0.66. For this analysis, the education-by-combat index interaction was also statistically significant (p=0.044). 12-14 �Depression The unadjusted analysis of the depression scale revealed no statistically significant difference between the two groups (p=0.272). In the covariate tests of association, education and drink-years showed statistically significant effects (p<0.001, p=0.002, respectively). There was a higher percentage of abnormalities in the high school-educated category (13.1%) than in the college-educated category (7.2%). For drink-years, the highest rate of abnormality was in the highest category of alcohol use (15.8%), followed by the nondrinker with 10.7 percent abnormalities and the moderate category with 9.4 percent. In the adjusted analysis, there was no statistically significant difference between groups (p=0.497), but there was a marginally significant group-by-combat index interaction (p=0.055). This interaction was explored further and is shown in Table J-4 of Appendix J. The analysis of the groupby-combat index interaction revealed a marginal difference within the low (0) category of the combat index (p=0.055), but not within the medium and high categories. In contrasting the 192 Ranch Hands and the 490 Comparisons in the 0 category, there were 14.6 percent abnormalities in the Ranch Hand group versus 8.2 percent in the Comparisons (p=0.039). The adjusted relative risk for the 0 category of the combat index was 1.73 with a 95 percent confidence interval of 1.03 to 2.91. Education (p<0.001) and drink-years (p=0.013) also exhibited statistically significant effects in the adjusted analysis. Hypochondria There was no statistically significant difference for the MMPI hypochondria scale between the Ranch Hand and Comparison groups (p=0.198). In the covariate tests of association, all five variables were statistically significant. Of men born in or after 1942, 8.8 percent had abnormalities as compared to 12.2 percent and 12.6 percent of those born between 1923 and 1941 and in or before 1922, respectively (p=0.031). The rates of abnormalities for Blacks and nonblacks were 16.8 percent and 10.4 percent, respectively (p=0.025). There was a highly statistically significant difference for education (p<0.001) with the high school-educated category having 13.9 percent abnormalities and the college-educated category having 7.0 percent. There was also a statistically significant difference for drink-years (p-0.041). The lowest rate of abnormalities was in the greater than 0 to 50 drink-years category with 9.9 percent; the corresponding percentages for the 0 drink-year and greater than 50 drink-year categories were 12.7 and 14.3, respectively. The percent abnormalities in the low, medium, and high combat index categories were 9.8, 9.4, and 13.2, respectively (p=0.027). The adjusted analysis showed no significant difference between the Ranch Hand and Comparison groups (p=0.431). In this analysis, age (p=0.002), race (p=0.026), education (p<0.001), and combat index (p=0.043) were statistically significant covariates. 12-15 �Hysteria Based on the unadjusted analysis of the MMPI hysteria scale, the difference between the two groups approached statistical significance (p=0.067). The percent abnormalities were 12.1 and 9.7 for the Ranch Hand and Comparison groups, respectively. The estimated relative risk, was 1.29 with a 95 percent confidence interval of 0.99 to 1.67. The covariate tests of association showed that there were statistically significant differences for age (p=0.044), education (p<0.001), and drinkyears (p=0.006). There were 12.6 percent, 12.1 percent, and 8.9 percent abnormalities in the age categories born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. The high school-educated category had a higher percentage of abnormalities (12.9%) than the collegeeducated category (8.2%). The drink-years category with the lowest percentage of abnormalities was greater than 0 to 50 with 9.6 percent? the 0 drink-years and the greater than 50 drink-years categories had 14.0 and 14.9 percent abnormalities, respectively. The adjusted analysis also approached significance (p=0.077). The adjusted relative risk was 1.27 with 95 percent confidence bounds of 0.97 and 1.66. Age and education were statistically significant covariates in the adjusted model (p=0.003, p<0.001, respectively). Drink-years was marginally significant (p=0.068) in the presence of other covariates, but was not included in the final adjusted model. Mania/Hypomania For the unadjusted analysis of the mania/hypomania scale of the MMPI, there was no statistical difference between the Ranch Hand and the Comparison groups (p=0.611). In the covariate tests of association, there were statistically significant differences for drink-years and combat index (p=0.011, and p=0.001, respectively). For the mania/hypomania scale, the 0 drink-years category had 6.7 percent abnormalities, the greater than 0 to 50 drink-years category had 5.8 percent, and the greater than 50 drink-years category contained 10.2 percent. The frequencies of abnormalities increased from the low to the high level of the combat index; the percentages were 5.0, 5.3, and 9.4, respectively. Based on the adjusted analysis, there was no statistically significant difference between the two groups (p=0.203). Drink-years was a significant covariate (p«0.006), as was the age-by-combat index interaction (p=0.046). Masculini ty/Feminini ty The masculinity/femininity scale of the MMPI measures the stereotype "macho" attitudes of the test subjects. There was a statistically significant group difference for this scale of the MMPI, unadjusted for covariates (p=0.017). There was a higher percentage of abnormalities in the Comparison group (9.3%) than in the Ranch Hand group (6.5%). The estimated relative risk was 0.68, and the 95 percent confidence interval was 0.50 to 0.93. 12-16 �There was a statistically significant difference detected for age (p=0.005) and for education (p<0.001), based on the pooled group data in the covariate tests of association. The highest rate of abnormalities was found in men born in or after 1942 (10.2%); whereas those born between 1923 and 1941 had 6.4 percent, and those born in or before 1922 had 8.0 percent. For education, the college-educated category showed an abnormal rate of 10.3 percent versus the high school category with 6.2 percent abnormalities. The adjusted analysis also showed a statistically significant difference between the two groups (p«0.020), with an adjusted relative risk of 0.69 (95% C.I.: [0.50,0.95]). Education and a race-by-age interaction were statistically significant in the adjusted analysis (p<0.001, p=0.008, respectively). These covariate associations follow expectations. Paranoia The unadjusted analysis of the MMPI paranoia scale did not reveal a statistically significant group difference (p=0.187). Based on the pooled group data, the covariate test of association for age was statistically significant (p=0.022). There was 3.6 percent abnormalities for men born in or after 1942, 2.0 percent for those born between 1923 and 1941, and no abnormalities for men born in or before 1922. The adjusted analysis revealed a significant group-by-age interaction (p=0.036). The age-by-combat index interaction was also statistically significant (p=0.003). The group interaction was examined by combining the participants born between 1923 and 1941 with those born in or before 1922, and basing the test on two age categories (born in or after 1942 and born before 1942), due to problems with 0 counts (see Table J-4 of Appendix J). The analysis showed a higher percentage of abnormal Ranch Hands than abnormal Comparisons for participants born before 1942 (2.7% and 1.2%, respectively; p=0.027). The relative risk estimate for this age category was 2.63 (95% C.I.: [1.11,6.20]). In contrast, for the stratum born in or after 1942, the frequencies of abnormalities were nearly the same in each group (3.7% for Ranch Hands, 3.5% for Comparisons; p=0.712). Psychopathic/Deviate No significant difference between the two groups was identified in the unadjusted analysis of this MMPI scale (p=0.845). In the covariate tests of association, there were statistically significant differences for race, education, and drink-years. There were 21.0 percent abnormalities for Blacks as compared to 11.1 percent for nonblacks (p=0.001). For education, there were 13.8 percent abnormalities in the high school-educated category and 9.1 percent in the college-educated category (p=0.001). The highest rate of abnormalities in the drink-year categories was 20.2 percent for the category of greater than 50 drink-years; the percent abnormalities for the 0 and greater than 0 to 50 categories were 11.3 and 10.1, respectively (p<0.001). Based on the adjusted analysis, there was no significant difference between the Ranch Hand and Comparison groups (p=0.780). In this analysis, education (p=0.011), the age-by-combat index interaction (p=0.003), and the 12-17 �race-by-drink-year interaction (p=0.015) were statistically significant adjusting variables. Schizophrenia The unadjusted tests showed no significant difference between the Ranch Hand and Comparison groups for the MMPI schizophrenia scale (p=0.228). Based on the pooled group data, the covariate tests of association revealed that education (p<0.001) and drink-years (p=0.014) had statistically significant effects. The high school-educated category had a statistically significant higher rate of abnormalities (11.0%) than the college-educated category (5.4%). For drink-years, the highest percent of abnormalities was in the greater than 50 drink-year category (12.6%), followed by the 0 drinkyear category with 8.7 percent, and the greater than 0 to 50 drink-year category, which had 7.7 percent abnormalities. In the adjusted analysis, the group-by-education interaction was significant (p=0.010) (see Table J-4 of Appendix J). The race-by-drink-year interaction was also statistically significant (p=0.017). Analysis of the high school and college strata showed a higher percentage of abnormal Ranch Hands than abnormal Comparisons in the high school classification (13.4% versus 9.5%, respectively; p=0.033). The relative risk estimate for high school participants was 1.51, with 95 percent confidence bounds of 1.05 and 2.16. The college-educated stratum revealed a nonsignificant group difference, but the Ranch Hands had a lower rate of schizophrenia abnormalities than the Comparison group (4.1% and 6.3%, respectively). Social Introversion Based on the unadjusted analysis, the difference between the two groups approached significance (p=0.069). The Ranch Hand group had 2.6 percent abnormalities as contrasted to 1.5 percent abnormalities in the Comparison group. The 95 percent confidence bounds on the estimated relative risk of 1.76 were 0.97 and 3.20. Age was the only statistically significant covariate (p=0.003). The participants who were born in or after 1942 had a higher percentage of abnormalities (3.1%) than either those born between 1923 and 1941 or those born in or before 1922; both of these latter age categories had a 1.1 percent frequency of abnormalities. Education was of marginal significance (p=0.099) with 2.4 percent of the high school-educated participants scored as abnormal as compared to 1.4 percent of the college-educated participants. The groupby-combat index interaction was statistically significant in the adjusted analysis (p=0.037) (see Table J-4 of Appendix J). The analysis of the group-by-combat index interaction showed a difference within the low (0) combat index category with the Ranch Hands having a significantly higher percentage of abnormalities than the Comparisons (5.6% and 1.2%, respectively; p=0.002). The adjusted relative risk for this combat index category was 4.86, with a 95 percent confidence interval of 1.77 to 13.36. The medium and high combat index strata showed no statistically significant group differences (p=0.478, p=0.677, respectively). In this adjusted model, age also had a significant effect (p=0.004). 12-18 �Validity For the MMPI validity scale, the unadjusted tests showed no significant difference between the Ranch Hand and Comparison groups (p=0.540). The covariate tests of association showed that Blacks had a significantly higher frequency of abnormalities (35.0%) than nonblacks (20.5%) (p<0.001). The adjusted analysis revealed a statistically significant groupby-race interaction (p=0.012). A covariate interaction, age-by-combat index, was also found to be statistically significant (p=0.030). Further investigation of the group interaction disclosed a higher percentage of Black Comparisons with scores greater than 0 than Black Ranch Hands (42.2%, 25.0%, respectively), with an adjusted relative risk of 0.46 (p=0.038, 95% C.I.: [0.22,0.96]). In contrast, the nonblack stratum revealed a slightly higher proportion of abnormalities in the Ranch Hands, with an adjusted relative risk of 1.20 (95% C.I.: [0.97,1.49], p=0.095) (see Table J-4 of Appendix J). Cornell Medical Index (CMI) Three variables derived from the CMI were analyzed: the total CMI, M-R subscore, and the A-H. area subscore. The total CMI was analyzed as a continuous variable, using a log (X+l) transformation, where X was the number of affirmative answers. Based on the Kolmogorov-Smirnov test, the distributions of the Ranch Hand and Comparison total CMI scores were contrasted. For this set of analyses, the data were stratified separately by the covariates of age, race, education, current alcohol use, and occupation. The unadjusted analysis of total CMI was based on the two-sample t-test. Analysis of variance and two-sample t-tests were used to analyze the covariates, and the adjusted analysis on the total CMI was based on analysis of covariance techniques, using SAS®-GLM. Age was analyzed as a continuous variable in the adjusted analysis. Using a two-sided ot-level of 0.05, and with power of 0.80, the sample sizes were sufficient to detect a 10.2 percent mean shift in the total CMI score relative to the mean observed in the Comparison group. Pearson's chi-square test was used to conduct the unadjusted analyses and the covariate tests of association of the M-R subscore and the A-H area subscore, which were trichotomized into low, medium, and high classes. The adjusted analyses of these two variables were conducted by log-linear techniques using BMDP*-4F. In all three CMI variables, a higher score is associated with a higher degree of abnormality. The results of the unadjusted analysis, covariate tests of association, and the adjusted analyses on the three CMI variables are summarized in Tables 12-5 to 12-7, respectively. As discussed for the MMPI variables, the results of the covariate tests of association for current alcohol use and for occupation are provided in the summary table for information only. 12-19 �TABLE 12-5. Unadjusted Analyses for the Cornell Medical Index (CHI) by Group Group Ranch Hand Comparison Variable Statistic Total CMI n Mean* 95% C.I.a 1,000 11.74 (11.17,12.35) 1,268 10.42 (9.95,10.90) n Number/% -0 (Low) 1-10 (Medium) >10 (High) 998 1,267 Est. Relative Risk (95% C.I.) M-R Subscore 538 408 52 53.9% 40.9% 5.2% 726 57 .3% 484 38 .2% 57 4.5% to I p-Value <0.001 Overall 0.252 Medium vs. Low 1.14 (0.96,1.35) High vs. Low 1.23 (0.83,1.82) 0.146 Overall 0.003 Medium vs. Low 1.33 (1.11, 1.60) High vs. Low 1.46 (1.08,1.98) 0.003 0.314 NJ O A-H Area Subscore n Number/% -0 (Low) 1-3 (Medium) 4-8 (High) 1,148 914 360 • 449 105 39.4% 49.1% 11.5% 537 46 .8% 504 43 .9% 107 9.3% 0.013 transformed from log (X+l) scale, where x was the number of questions answered "yes." —No relative risk given for Total CMI, which was analyzed as a continuous variable. �TABLE 12-6, Association Betveen CMI Variables and the Covariates in the Combined Ranch Hand and Comparison Groups CMI Variable Age Race Education DrinkYears PTSD Current* Alcohol Use Occupation* Total CMI <0.001 NS <0.001 <0.001 <0.001 <0.001 <0.001 M-R Subscore <0.001 0.022 <0.001 NS* <0.001 0.043 <0.001 A-H Area Subscore <0.001 NS <0.001 <0.001 <0.001 0.010 <0.001 NS: Not significant (p>0.10). NS*: Borderline significant (0.05<p<0.10). **Not used in adjusted analyses. 12-21 �TABLE 12-7. Adjusted Analyses for CMI Variables by Group Group Variable Statistic Ranch Hand Total CHI n Adj. Mean 95% C.I. 962 **** **** 1,229 **** **** Comparison M-R Subscore n 998 1,265 A-H Area Score n 881 1,113 to I to Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks* **** PTSD (p<0.001) RACE*DRKYR (p=0.039) AGE*EDUC (p=0.005) GRP*EDUC (p=0.003) Overall Medium vs. Low: 1.14 (0.95,1.35) High vs. Low: 1.12 (0.74,1.70) 0.339 0.152 0.598 AGE (p<0.001) EDUC (p<0.001) PTSD (p<0.001) GRP*EDUC (marginal: p=0.067) Overall Medium vs. Low: 1.27 (1.06,1.53) High vs. Low: 1.24 (0.90,1.71) 0.040 0.011 0.190 AGE (p<0.001) EDUC (p<0.001) PTSD (p<0.001) DRKYR (p=0.014) *Additional Abbreviations; PTSD: Post-Traumatic Stress Disorder ****Group-by-covariate interaction—adjusted mean, confidence interval, and p-value not presented. No relative risk given for total CMI, which was analyzed as a continuous variable. �Distributional Analyses The Kolmogorov-Smirnov tests showed statistically significant differences between the Ranch Hand and Comparison distributions for the total CMI for one category for each of the covariates. For age, the distribution of Ranch Hands born in or after 1942 was statistically different from the corresponding distribution for the Comparisons (p<0.001). The distributions of the nonblack Ranch Hand and Comparison responses also differed significantly (p=0.003). The contrast of the high school-educated Ranch Hand and Comparison distributions revealed a statistically significant difference (p<0.001). The distributions for Ranch Hand and Comparison current drinkers were also statistically different (p=0.024). For occupation, the enlisted groundcrew distributions for Ranch Hands and Comparisons were statistically different (p=0.007). Except for the covariate age, all significant differences in distributions for each covariate were found in the category having the largest sample size. The results of the 12 Kolmogorov-Smirnov tests are summarized in Table J-5 of Appendix J. Unadjusted and Adjusted Analyses Total Cornell Medical Index Based on the unadjusted analysis, as depicted in Table 12-5, the total CMI means of the Ranch Hand and Comparison groups were statistically different (p<0.001). The mean, as transformed from the log (X+l) scale, of the 1,000 Ranch Hands was 11.74 as compared to 10.42 for the Comparisons. The covariate tests of association identified that age, education, drink-years, and PTSD were highly significant (p<0.001 for all). For age, the (transformed) means of the categories showed an increase; the means of those born in or after 1942, between 1923 and 1941, and in or before 1922 were 10.08, 11.49, and 14.53, respectively. The mean of the high schooleducated category (12.97) was statistically higher than the mean of the college-educated category (8.99). The mean of the greater than 50 drinkyears was 14.49 as compared to means of 10.37 and 10.34 for the 0 and greater than 0 to 50 drink-years, respectively. The mean of the participants with a positive measure of PTSD was 71.77, whereas 10.83 was the mean of those without a positive measure of PTSD. In the adjusted analysis, there was a significant group-by-education interaction (p-0.003). Further analysis of the interaction (see Table J-4 of Appendix J) snowed that the high school-educated Ranch Hands had a higher adjusted mean total CMI than the high school-educated Comparisons (p<0.001). No significant difference was seen in the college stratum. PTSD was a significant covariate (p<0.001). The covariate interactions, race-by-drinkyears and age-by-education, were also significant in the adjusted model (p=0.039, p=0.005, respectively). M-R Subscore The results of the unadjusted analysis on the M-R subscore, an indicator of emotional health, revealed no significant difference between groups (p-0.252). 12-23 �The covariate tests of association on the pooled group data showed that age (p<0.001), race (p=0.022), education (p<0.001), and PTSD (p<0.001) were statistically significant covariates. For age, participants born in or after 1942 had a higher percentage of scores greater than 0 when compared to the other categories. Blacks had a higher percentage of scores greater than 0 than nonblacks. For education, the college-educated category had a higher percentage of 0 scores. The M-R subscores were distributed differently for participants with and without PTSD. For example, 15 of 16 participants with PTSD had an M-R subscore greater than 10, whereas only 4.2 percent of the participants without PTSD had a similar score. Drink-years showed a marginally significant effect (p=0.054); the greater than 50 drink-year category exhibited the largest percentage of participants with scores greater than 0. No significant difference between the two groups was identified in the adjusted analysis. There was a marginally significant group-by-education interaction (p=0.067). Further investigation of this interaction (see Table J-4 of Appendix J) showed a significant difference for the high schooleducated stratum (p=0.030) but not for the college-educated stratum. This difference results from the contrast of the medium (1 to 10) and low (0) categories, with the Ranch Hands having a higher percentage of participants in the medium category for the M-R subscore than in the low category (Adj. RR: 1.37, 95% C.I.: [1.07,1.75], p»0.014). In this analysis, age, education, and PTSD were highly significant adjusting variables (p<0.001 for all). A-H Area Subscore Based on the unadjusted results, the A-H area subscore—an indicator of diffuse medical problems—revealed a significant difference between the Ranch Hand and Comparison groups (p=0.003). This was due to the increased percentage of Ranch Hands over Comparisons in both the medium (1 to 3) and the high (4 to 8) categories (p=0.003, p=0.013, respectively). The covariate tests on the A-H area subscore showed that age, education, drink-years, and PTSD were highly significant covariates (p<0.001 for all). Older participants (born in or before 1922) had the lowest percentage of 0 scores. The college-educated category had a higher percentage of 0 scores than the high school-educated category. For drink-years, the lowest percentage of 0 scores was in the greater than 50 drink-years category. Twelve of 16 participants with PTSD had scores of 4 to 8, as compared to. 9.7 percent of participants without PTSD. Results of the adjusted analysis were similar to the unadjusted analysis and indicated that the two groups were statistically different (p=0.040). The overall group difference was predominately due to an increased adjusted percentage of Ranch Hands over Comparisons in the medium (1 to 3) versus low (0) contrast (p=0.011). The adjusted relative risk for this contrast was 1.27 with 95 percent confidence bounds of 1.06 and 1.53. In the adjusted model, age, education, and PTSD were significant covariates (p<0.001 for all); drink-years was also statistically significant (p=0.014). Halstead-Reitan Battery (HUB) The unadjusted analysis of the impairment index, the one variable from the HRB, was performed by using Fisher's exact test. Fisher's exact test and 12-24 �Pearson's chi-square test were used to conduct the covariate tests of association. The adjusted analysis was based on logistic regression techniques using BMDP*-LR. The results of the analyses of the HRB impairment index are summarized in Table 12-8. The unadjusted contrast of the 1,006 Ranch Hand scores and the 1,283 Comparison scores for the HRB impairment index revealed no statistically significant group differences (p=0.533). The covariate tests of association showed that age, race, and education were highly significant covariates (p<0.001 for all), and drink-years also was statistically significant (p=0.002). For age, the highest percent frequency of abnormalities was in the category of participants born in or before 1922 (66.3%); the corresponding frequencies for the participants born between 1923 and 1941 and for those born in or after 1942 were 38.3 percent and 25.1 percent, respectively. Blacks had a significantly higher percentage of abnormal scores, with 57.1 percent as compared to 32.3 percent for nonblacks. The college-educated category had a 22.3 percent frequency of abnormalities versus 43.5 percent for the high school-educated category. With respect to drink-years, the highest percentage of abnormalities (41.2%) was for greater than 50 drink-years; the 0 drink-year and greater than 0 to 50 drink-year categories had 38.0 percent and 32.0 percent, respectively. There was no significant difference identified between the two groups based on the adjusted analysis (p=0.697). Age, race, and education were statistically significant covariates (p<0.001 for all). EXPOSURE INDEX ANALYSES Exposure index analyses were conducted within each occupational cohort of the Ranch Hand group (see Chapter 8 for details on the exposure index). All variables, except the total CMI, were investigated, (unadjusted for any covariates), using Pearson's chi-square test and Fisher's exact test. Analyses of the total CMI were accomplished by t-tests and analysis of variance and covariance techniques. A log transformation was used in both adjusted and unadjusted analyses, and participants with PTSD were deleted. Adjusted analyses were performed using logistic regression, incorporating the covariates of race, age, education, and drink-years, as well as any significant pairwise interactions between the exposure index and these covariates. Age was treated as a continuous variable in the analyses. For the MMPI variables, combat index was also included as a covariate. For the HRB impairment index, participants classified as having PTSD were deleted from the analysis. The M-R subscore and the A-H area subscore were collapsed into 2 categories for analysis: 0 and greater than 0. Participants with PTSD were also deleted from this analysis. Overall significance in the proportion of abnormalities among the exposure index levels of low, medium, and high was determined, as well as contrasts in the proportion of abnormalities between the medium and low exposure levels, and between the high and low exposure levels. Results of the adjusted analyses are presented in Table 12-9, and parallel results for unadjusted analyses are presented in Table J-6 of Appendix J. Results from further study of exposure index-by-covariate interactions are given in Table J-7 of Appendix J. 12-25 �TABLE 12-8. Summary Results for the Halstead-Reitan Battery Impairment Index Analyses Group Analysis Comparison Number Percent Est./Adj. Relative Risk (95% C.I.) p-Value Unadjusted Analysis to 1 eo Ranch Hand Statistic Number Percent n Abnormal Normal 1,283 427 856 1.06 (0.89,1.26) 1,006 348 658 34.6 65.4 33.3 66.7 0.533 N/A AGE (p<0.001) RACE (p<0.001) EDUC (p<0.001) DRKYR (p=0.002) PTSD (p=0.431) ALC (p=0.004) OCC (p<0.001) Covariate Tests of Association3 cr> Adjusted Analysis Covariate Remarks* n 1,006 1,283 1.04 (0.86,1.25) 0.697 AGE (p<0.001) RACE (p<0.001) EDUC (p<0.001) * Additional Abbreviations: ALC: current alcohol use (yes/no) OCC: occupation a Based on pooled group data; current alcohol use (ALC) and occupation (OCC) provided for information only. �TABLE 12-9. Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer Anxiety Statistic* n Low 125 Exposure Index Medium 126 61 Adj. Relative Risk (95* C.I.) p-Value High Contrast 120 Overall H vs. L H vs. L 2.46 (0.36,16.82) 2.43 (0.35,16.81) 0.562 0.358 0.367 Overall M vs. L H vs. L 0.44 (0.12,1.70) 0.28 ( . 05, 1.44) 0 0.215 0.235 0.127 Enlisted Flyer n ; 50 53 Enlisted Groundcrev n 148 160 131 Overall H vs. L H vs. L **() **! **() **! Officer n 125 126 120 Overall M vs. L H vs. L 1.10 (0.14,8.59) 0.274 0.925 Overall M vs. L H vs. L 0.39 0.30 (0.06,2.37) (0.03,2.93) 0.425 0.304 0.303 Overall M vs. L H vs. L 0.87 (0.32,2.34) 0.56 (0.18,1.67) 0.550 0.781 0.296 to Consistency Enlisted Flyer n Enlisted Groundcrev n 50 148 61 160 53 131 **() **! **() **! **() **! �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Statistic* Low Exposure Index Medium High Contrast 125 126 120 Overall H vs. L H vs. L n 50 61 53 Overall M vs. L H vs. L Enlisted Groundcrew n 148 160 131 Overall M vs. L ° H vs. L Officer n Officer Defensiveness NJ CO Denial Adj. Relative Risk (95% C.I.) p-Value n' Enlisted Flyer Enlisted Flyer n Enlisted Groundcrew n 125 50 148 126 61 160 120 53 131 0.518 0.613 0.17 (0.001,29.09) 0.503 1.37 (0.02,77.86) 0.878 0.79 (0.23,2.78) 1.31 (0.40,4.23) Overall M vs. L H vs. L ****(2) ****(2) Overall M vs. L H vs. L 1.03 (0.09,11.69) Overall M vs. L H vs. L 0.737 0.719 0.656 ****(2) ****(2) ****(2) 0.234 0.984 0.109 1.41 (0.18,11.09) 0.747 �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer Depression Statistic* n n Enlisted Groundcrev n Officer Hypochondria Enlisted Flyer n Enlisted Flyer n Enlisted Groundcrew n Low 125 50 148 125 50 148 Exposure Index Medium 126 61 160 126 61 160 High 120 53 131 120 53 131 Contrast Adj. Relative Risk ( 5 C.I.) p-Value 9% Overall H vs. L H vs. L 0.62 (0.20,1-88) 1.24 (0.46,3.33) 0.411 0.393 0.669 Overall H vs. L H vs. L 0.55 (0.18,1.67) 0.31 (0.09,1.10) 0.160 0.295 0.070 Overall M vs. L H vs. L **() **! **() **! **() **! **() **! **() **! Overall M vs. L H vs. L ****(3) ****(3) ****(3) ****(3) ****(3) Overall M vs. L H vs. L 0.33 (0.09,1.18) 0.74 (0.26,2.14) Overall M vs. L H vs. L **() **! **() **! 0.195 0.087 0.581 **() **! **() **! **() **! �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer Hysteria Statistic* n n Enlisted Groundcrew n Officer Mania/ Hypomania Enlisted Flyer n Enlisted Flyer n Enlisted Groundcrev n Low 125 50 148 125 50 148 Exposure Index Medium 126 61 160 126 61 160 High 120 53 131 120 53 131 Contrast Adjo Relative Risk ( 5 C.I.) p-Value 9% Overall M vs. L H vs. L ****(3) ****(3) Overall M vs. L H vs. L 0.55 (0.18,1.74) 0.41 (0.12,1.37) Overall M vs. L H vs. L **() **! **() **! **() **! **() **! **() **! Overall M vs. L H vs. L ****(4) ****(4) ****(4) ****(4) ****(4) Overall M vs. L H vs. L 2.51 (0.55,11.53) 1.66 (0.35,7.89) 0.474 0.236 0.527 Overall M vs. L H vs. L 0.97 (0.38,2.45) 0.61 (0.21,1.75) 0.597 0.945 0.356 ****(3) ****(3) ****(3) 0.306 0.312 0.148 �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer Statistic* n Low 125 Exposure Index Medium 126 High 120 Contrast Overall M vs. L H vs. L Adj. Relative Risk (95% C.I.) p-Value ****(3) ****(3) ****(3) ****(3) ****(3) 0.045 n 50 61 53 Overall M vs. L H vs. L n 148 160 131 Overall M vs. L H vs. L 0.50 (0.16,1.57) 0.75 (0.25,2.24) Officer Paranoia Enlisted Flyer Enlisted Groundcrew Masculinity/ Femininity n Overall M vs. L H vs. L ****(2) ****(2) ****(2) ****(2) ****(2) Overall M vs. L H vs. L ****(2) ****(2) ****(2) ****(2) ****(2) Enlisted Flyer n Enlisted n Groundcrew (a) 125 50 148 126 61 160 120 53 131 Overall M vs. L H vs. L 1.06 (0.31,3.66) 1.47 (0.44,4.92) 0.479 0.234 0.604 0.789 0.922 0.530 �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer Psychopathic/ Deviate Statistic* n n n Officer Schizophrenia Enlisted Flyer Enlisted Groundcrev to I OJ to Low n Enlisted Flyer n Enlisted Groundcrew n 125 . 50 148 125 50 148 Exposure Index Medium 126 61 160 126 61 160 High 120 53 131 120 53 131 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L 1.01 (0.34,2.98) 1.78 (0.65,4.83) 0.427 0.985 0.259 Overall H vs. L H vs. L 1.20 (0.42,3.41) 0.79 (0.24,2.54) 0.759 0.731 0.689 Overall M vs. L H vs. L ****(3) ****(3) Overall M vs. L H vs. L 0.72 (0.18,2.97) 0.38 (0.07,2.12) 0.511 0.654 0.269 Overall M vs. L H vs. L 0.70 (0.21,2.35) 0.52 (0.14,1.97) 0.615 0.559 0.338 Overall M vs. L H vs. L 1.32 (0.66,2.61) 1.30 (0.64,2.64) 0.682 0.429 0.471 ****(3) ****(3) ****(3) �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer Statistic* n n n Officer Validity Enlisted Flyer Enlisted Groundcrev Social Introversion n Enlisted Flyer n Enlisted Groundcrew n Low 125 50 148 125 51 148 Exposure Index Medium 126 61 160 126 61 160 High 120 53 131 120 53 131 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L 1.86 (0.16,21.91) 0.247 0.620 Overall M vs. L H vs. L 0.20 0.30 (0.01,4.85) (0.02,5.61) 0.521 0.321 0.418 Overall M vs. L H vs. L 0.47 (0.15,1.49) 0.87 (0.28,2.67) 0.394 0.199 0.805 Overall M vs. L H vs. L 0.97 (0.53,1.76) 0.48 (0.24,0.93) 0.049 0.920 0.031 Overall M vs. L H vs. L 0.67 (0.23,1.94) 1.26 (0.47,3.40) 0.479 0.459 0.649 Overall M vs. L H vs. L 1.22 (0.71,2.11) 1.22 (0.69,2.14) 0.718 0.470 0.499 �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Statistic* Exposure Index Medium Low High Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L ****(4) ****(4) ****(4) 51 48 61 n Adj. Mean ****(3,4) ****(3,4) ****(3,4) 95* C.I. ****(3,4) ****(3,4) ****(3,4) Overall M vs. L H vs. L ****(3,4) ****(3,4) ****(3,4) Enlisted Groundcrev 145 154 n 12.48 Adj. Mean(b) 13.67 95Z C.I.(b) (11.33, (10.30, 16.45) 15.09) 125 13.09 (10.81, 15.82) Overall M vs. L H vs. L Officer n 119 Overall M vs. L H vs. L 0.72 (0.41,1.28) 1.11 (0.64,1.93) 0.301 0.265 0.715 Overall M vs. L H vs. L ****(4) ****(4) **() **4 ****(4) ****(4) Officer Total CHI H-R Subscore 124 n Adj. Mean ****(4) 95X C.I. ****(4) Enlisted Flyer 123 Enlisted Flyer n 48 Enlisted Groundcrev n 146 124 ****(4) ****(4) 124 120 ****(4) ****(4) 61 51 152 127 Overall M vs. L H vs. L 0.608 0.319 0.655 0.82 (0.51,1.31) 0.73 (0.44,1.19) 0.427 0.403 0.201 �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Officer A-H Area Subscore Statistic* n Enlisted Flyer n Enlisted Groundcrev n Officer n Low 111 45 129 Exposure Index Medium 109 5? 145 High Contrast 112 Overall M vs. L H vs. L 45 118 i-» NJ 1 U> 01 HRB Impairmerit Index Enlisted Flyer n Enlisted Groundcrev n 124 47 145 126 61 158 118 52 127 Overall M vs. L H vs. L Adj. Relative Risk (95* C.I.) p-Value 0.78 (0.44,1.37) 0.75 (0.43,1.31) ****(3,4) ****(3,4) 0.546 0.383 0.311 ****(3,4) ****(3,4) ****(3,4) Overall M vs. L H vs. L 0.84 ( . 0 1 4 ) 05,.0 0.92 (0.53,1.59) Overall M vs. L H vs. L 0.81 (0.43,1.53) 0.57 (0.29,1.12) 0.255 0.512 0.103 Overall H vs. L H vs. L 2.28 (0.96,5.44) 1.39 (0.58,3.37) 0.159 0.063 0.461 Overall M vs. L H vs. L * * ( ). **! **() **! 0.427 0.499 0.767 **() **! ****(!) ****(!) �TABLE 12-9. (continued) Adjusted Exposure Index Analyses for Psychological Variables by Occupation *n: represents total sample size for variable in given occupational stratum. (a): marginal exposure index by race interaction (p=0.055) — relative risk, confidence interval, and p-value presented, and additional information provided in interaction summaries. (b): converted from log (X-t-1) scale, where X was the number of questions answered yes. * * ( ) exposure index-by-race interaction — relative risk, confidence interval, and p-value not presented. **!: ****(2): exposure index-by-age interaction — relative risk, confidence interval, and p-value not presented. i-* tsj ^ ****(3): exposure index-by-education interaction — relative risk, confidence interval, and p-value not presented, a* **() * * 4 : exposure index-by-drink-year interaction — relative risk/adjusted mean, confidence interval, and p-value not presented. ****(3,4): exposure index-by-education and exposure index-by-drink-year interaction — relative risk/adjusted mean, confidence interval, and p-value not presented. : no relative risk given for Total CMI, which was analyzed as a continuous variable. �Unadjusted analyses revealed a borderline significant difference between the high and low exposure levels for masculinity/femininity in officers (Est. RR: 2.38, 95% C.I.: [0.94,6.06], p=0.075), and for the total CMI in officers (low means 7.99, high mean: 10.04, p=0.018; overall p-value: 0.049). These data supported an increase in the proportion of abnormalities with increasing exposure levels. Other significant or marginally significant results were associated with a decrease in the proportion of abnormalities with an increase in exposure level. The frequency of abnormalities for the different exposure index levels exhibited no graduated pattern across exposure levels. Within the officer stratum, five variables demonstrated an increasing dose-response relationship, although usually nonsignificant; however, four variables showed the opposite pattern, that is, a decreasing proportion of abnormalities with increasing exposure levels. Few significant results were observed in the adjusted analysis, as in the unadjusted analysis. The medium level of the HRB impairment index for enlisted flyers showed an increased relative risk over the low level (Adj. RR: 2.28, 95% C.I.: [0.96,5.44], p=0.063). Many exposure index-by-covariate interactions were present, however, which prevented a direct comparison. Interactions were present for 13 of the 18 variables, but no occupational stratum was predominant. A summary of these interactions is presented in Table 12-10. TABLE 12-10. Summary of Exposure Index-by-Covariate Interactions in Adjusted Analyses of Psychological Variables Variable Occupation Anxiety Denial Depression Hypochondria Hypochondria Hysteria Hysteria Mania/Hypomania Masculini ty/Femini ty Paranoia Paranoia Paranoia Psychopathic/Deviate Total CMI Total CMI Total CMI M-R Subscore A-H Area Subscore A-H Area Subscore HRB Impairment Index Enlisted Officer Enlisted Officer Enlisted Officer Enlisted Officer Officer Officer Enlisted Enlisted Enlisted Officer Enlisted Enlisted Enlisted Enlisted Enlisted Enlisted Covariate Groundcrew Groundcrew Groundcrew Groundcrew Flyer Groundcrew Groundcrew Flyer Flyer Flyer Flyer Flyer Groundcrew 12-37 Race Age Race Education Race Education Race Drink-Years Education Age Age Race Education Drink- Years Education Drink-Years Drink-Years Education Drink-Years . Race p-Value 0.020 0.048 0.050 0.005 0.033 0.018 0.007 0.015 0.018 0.044 0.004 0.055 (marginal) 0.040 0.034 0.027 0.021 0.042 0.009 0.004 0.031 �Significant or borderline significant results in these interactions, suggestive of a dose-response relationship (i.e., increasing abnormalities or more abnormal means as exposure increases), were as follows: (1) Hysteria in college-educated officers, overall p-value = 0.025; high versus low contrast (Adj. RR: 3.49, 95% C.I.: [1.17,10.32], p=0.024); increase in the proportion of abnormalities with increasing exposure levels. (2) Mania/Hypomania in officers with greater than 50 drink-years, high versus low contrast, p»0.067; analysis affected by sparse cell sizes, however. (3) Masculinity/Femininity in college-educated officers, medium versus low contrast (Adj. RR: 3.05, 95% C.I.: [1.01,9.08], p=0.048); increase in the proportion of abnormalities with increasing exposure levels. (4) Total CMI in high school-educated, nondrinking, enlisted flyers, medium versus low contrast, p=0.018. (5) Total CMI in college-educated, nondrinking, enlisted flyers, overall p-value =0.060; analysis affected by sparse cell sizes, however. (6) M-R subscore in nondrinking, enlisted flyers, overall p-value = 0.060; analysis affected by sparse cell sizes, however. (7) A-H area subscore in high school-educated, nondrinking, enlisted flyers, overall p-value = 0.007; analysis affected by sparse cell sizes, however. ( ) HRB impairment index in nonblack enlisted groundcrew, medium versus 8 low contrast (Adj. RR: 1.88, 95% C.I.: [1.09,3.25], p-0.024). All other significant interaction results were not consistent with a dose-response relationship. In summary, no consistent or strong patterns of increasing dose-response relationship were evident throughout the psychological exposure index analyses. LONGITUDINAL ANALYSES Two scales for the MMPI, depression and denial, were significantly different by group at Baseline and were investigated to assess the longitudinal differences between the 1982 Baseline examination and the 1985 followup examination. Both variables are scores and were classified as abnormal or normal according to criteria given previously. These variables have been stratified by education level. As shown in Table 12-11, 2x2 tables were constructed for each group for each variable. These tables show the number of participants who were abnormal at Baseline and abnormal at followup, abnormal at Baseline and normal at followup, normal at Baseline and abnormal at followup, and normal at both Baseline and followup examinations. 12-38 �TABLE 12-11. Longitudinal Analysis of Depression and Denial: A Contrast of Baseline and First Followup Examination Abnormalities Variable Education Group 1982 Baseline Exam 1985 Followup Exam Odds p-Value Ratio (OR)* (OR^ vs/ORc) Abnormal Ranch Hand Abnormal Normal 59 31 48 570 0.65 Comparison Abnormal Depression High School Normal 44 52 43 695 1.21 11 10 9 227 1.11 7 1.36 5 2.20 0.04 Normal College Ranch Band Abnormal Normal 0.73 Comparison Abnormal Normal High School Ranch Hand Abnormal Normal 2 11 690 Comparison Abnormal Normal Denial 15 . 11 276 6 32 10 786 Ranch Hand Abnormal 0 5 3 249 0.56 College Normal 3.20 1.67 0.32 Comparison Abnormal Normal 5 13 -WVM- THHn- Nurober Normal Baseline, Abnormal Followup Number Abnormal Baseline, Normal Followup 12-39 3 288 4.33 �The odds ratio given is the ratio of the number of participants who were normal at the Baseline and abnormal at the followup to the number of participants who were abnormal at the Baseline and normal at the followup (the "off-diagonal" elements). The changes in normal/abnormal status within each group are contrasted between the Ranch Hand and Comparison groups, and the p-value is derived-from Pearson's chi-square test of the hypothesis that the pattern of change in the two groups is the same. The data showed a significant difference (p=0.04) in the depression scores in the two groups between examinations for the high school-educated stratum: significantly more Comparisons developed depression in the interval. The percentage of Ranch Hands with abnormalities for depression decreased from the Baseline examination to the followup examination, in contrast to the Comparison group, which showed an increase in depression abnormalities. No significant difference in the pattern of change for depression was found in the college-educated stratum, nor were any significant differences observed for denial. DISCUSSION The MMPI is a comprehensive, self-administered questionnaire containing 566 questions that broadly assess behavior, personality, and validity and consistency indicators of the responses. The MMPI data are divided into 14 scales that are not mutually exclusive for specific questions. In this study, an additional MMPI scale for the characterization of PTSD is used to identify highly correlated combat experiences of the participants. Four combat questions were selected as a surrogate measure of PTSD, and an index of these questions is used as a covariate in all of the adjusted analyses of the MMPI subscales. Distributional testing for the 14 scales of the MMPI, stratified by occupation, yielded no significant differences or discernible patterns between the two groups. In contrast, both unadjusted and adjusted analyses showed significant group differences for the denial and masculinity/ femininity scales, with the Comparisons having higher proportions of abnormalities than the Ranch Hands. Also, borderline significant associations (0.05<p<0.10) were observed for the hysteria and social introversion scales, with the Ranch Hands having slightly higher proportions of abnormalities than the Comparisons. The discrepancy in results between Kolmogorov-Smirnov distributional testing and the refined statistical models was also noted in the 1984 Baseline Report. The unadjusted and adjusted results were completely comparable with respect to group differences when direct contrast was possible, i.e., when no group-by-covariate interactions were present. Of the seven group interactions noted in the adjusted analyses, three involved the covariate of education, with the high school-educated Ranch Hands faring worse than high school-educated Comparisons. Further, the high school strata usually exhibited a higher frequency of abnormalities than the college-educated strata. Overall education showed a profound effect either as a main effect or by an interaction with another covariate. The strong influence of education was also detected in the Baseline data. Analyses using only the Original Comparisons often showed stronger group differences than the analyses based upon the total Comparison group (see Tables J-13 to J-18 of Appendix J). 12-40 �A direct comparison of the MMPI results between the Baseline and followup examinations is hampered by the small change in cohorts and the difference in statistical models. In general, at the followup the Ranch Hands manifested more MMPI scale abnormalities than the Comparisons, as judged by the number of relative risks greater than one. However, the highly significant results for the denial scale, with the Comparisons having a higher proportion of abnormalities than the Ranch Hands, suggested that the Comparisons may be underreporting on all of the MMPI scales, and consequently more relative risks greater than one would be expected. A contrast of the adjusted Baseline MMPI results to the adjusted (and unadjusted results where interactions are noted in the adjusted tests) results of the followup suggest a relatively consistent pattern of narrowing group differences over time (e.g., hypochrondria, depression, hysteria, schizophrenia scales), either by a decrease in Ranch Hand abnormalities or an increase of Comparison abnormalities. This trend was also suggested in the longitudinal analysis of two scales (depression and denial) although only the "favorable" Ranch Hand change in depression for the high school stratum reached statistical significance. Overall, the followup MMPI data suggested a subtle, but consistent, decrease in reporting of concerns (or strength of concerns) in the Ranch Hands. Only 16 participants were identified as possibly having PTSD by the MMPI subscale. Further, only 4 of 15 combat experience questions manifested strong correlation to these possible PTSD cases. Most PTSD surveys have focused on U.S. Army ground personnel, obscuring direct comparisons to U.S. Air Force personnel because of inherent differences in combat experience, education, proportion of officers, and career motivation. The CMI revealed a significant group difference for the total score and the A-H area subscore, with the Ranch Hands exhibiting higher mean scores or higher frequencies of abnormal scores. There was no group difference for the M-R subscore. These results differed slightly from the distributional tests which showed one statistically significant stratum, where the Ranch Hand mean was greater than the Comparison mean, for each covariate (see Table J-5 of Appendix J). Because the Baseline CMI was in a different format, direct comparison of each psychological parameter to the followup CMI is not feasible. However, the Baseline CMI noted statistically significant group differences for 5 of 10 parameters, which is in approximate accord with the magnitude and direction of the results found at the followup examination. This analysis of the total CMI analyzed at followup has sufficient statistical power to detect a mean difference of one response out of 195 questions (0.5% difference, at power=0.8) between the groups. Education showed the same profound effect on the adjusted analyses as was noted at Baseline. The functional integrity of the CNS, as measured by the HRB impairment index, showed no significant group differences. There was similarity (Adj. RR: 1.04, 95% C.I.: [0.86,1.25], p«0.697) in results of the impairment index. As in the Baseline analysis, education was a major covariate in the followup examination; the additionally strong effects of age and race were also noted at the followup examination. Although valid differences exist between groups for some measures, there is no indication that these differences are manifest or confirmed by impaired CNS function, a reasonable medical expectation for chemically induced neurobahavioral pathology. Adjustment of the HRB results for PTSD (not feasible at the Baseline analysis) suggests that some group differences lack organic basis. 12-41 �SUMMARY AND CONCLUSIONS Questionnaire data (verified by medical record reviews) for the lifetime events of psychotic illness, alcohol dependence, anxiety, or other neuroses disclosed no significant differences between groups for these conditions. Analyses of the followup psychological examination emphasized 14 scales from the Minnesota Multiphasic Personality Inventory (MMPI), 3 parameters of the Cornell Medical Index (CMI), and the Halstead-Reitan Battery (HRB) impairment index. The similarity of the group distribution for the 14 MMPI variables, each stratified by the 3 occupational categories, was examined, and only 2 of the 42 tests approached statistical significance. The group distributions of the total CMI score were similarly contrasted, with separate analyses performed with stratification by the five covariates of age, race, occupation, education, and current drinking status. For one stratum of each of these covariates, a significant difference in the distribution of the Ranch Hand and Comparison scores was found. In all cases for the CMI, the Ranch Hand mean was greater than the Comparison mean. Distributional analyses using Original Comparisons generally reflected the same results as those involving the total Comparison group. Results of unadjusted and adjusted analyses on all of the 18 psychological variables are given in Table 12-12. The unadjusted analyses showed a significant difference for the MMPI scales of denial (p<0.001) and masculinity/femininity (p=0.017), the total CMI (p<0.001), and the Section A-H area subscore (p=0.003). A borderline significant difference was observed for the MMPI scales of hysteria (p=0.067) and social introversion (p=0.069). Comparisons had a greater percentage of abnormal scores for the denial and masculinity/femininity scales, whereas Ranch Hands showed adverse findings for the other four variables. The overall MMPI results have been interpreted in light of the significant increased denial in the Comparison group. The covariates age, education, drink-years, current alcohol use, and occupation had pronounced effects on the psychological variables, with a significant association or a borderline significant association with at least two-thirds of the 18 psychological variables. Many dependent variables in this chapter were affected by age in an expected pattern. Very few variables exhibited this pattern of consistency with drink-years. The intermediate category of greater than 0 to 50 drink-years often had the smallest proportion of abnormalities. The post-traumatic stress disorder (PTSD) variable, derived from a subset of the MMPI, was strongly associated with the CMI measures, but not with the HRB Impairment Index. Race and the Vietnam combat index (used for the MMPI subscales) had significant associations with a lesser amount of the psychological variables (6 of 18 variables and 3 of 14 variables, for race and combat index, respectively). The adjusted analyses were generally quite similar to the unadjusted analyses with respect to group differences, although a direct comparison of these analyses was often clouded by the presence of a substantial number of interactions (six group-by-covariate interactions were significant, and three interactions approached significance [0.05<p<0.10]). The MMPI scales of denial and masculinity/femininity were statistically significant in both the 12-42 �TABLE 12-12. Overall Summary Results of Adjusted and Unadjusted Analyses of Psychological Variables Variable Questionnaire; Psychological Illness Unadjusted Adjusted Direction of Results* NS Psychological Examination; MMPI Anxiety Consistency Defensiveness Denial Depression Hypochondria Hysteria Mania/Hypomania Masculini ty/Feminini ty Paranoia Psychopathic/Deviate Schizophrenia Social Introversion Validity NS NS NS <0.001 NS NS b NS*b NS 0.017 NS NS NS NS*b NS NS **** NS <0.001 NS NS NS*b NS 0.020 **** NS **** **** **** CMI Total CMI M-R Subscore A-H Area Subscore <0.001 NS 0.003 **** NS 0.040 HRB Impairment Index NS ORH RH>C ORH RH>C RH>C RH>C NS *RH>C - more abnormalities in Ranch Hands; ORH - more abnormalities in Comparisons. b lllnesses include psychosis, alcohol dependence, anxiety, and other neuroses. —Analysis not performed. NS: Not significant. NS*: Borderline significant (0.05<p<0.10). ****Interaction involving group. 12-43 �adjusted and unadjusted analyses, where Comparisons showed an adverse effect over Ranch Hands. The A-H area subscore of the CMI (suggesting diffuse medical problems) was also significant, where the Ranch Hands had higher mean scores than the Comparisons, suggesting the Ranch Hands had more illness. Education was often involved in significant group interactions with high school-educated Ranch Hands demonstrating a higher percentage of abnormal scores than high school-educated Comparisons. No group differences were observed in the college-educated stratum. The M-R subscore of the CMI, a broad indicator of emotional health, was not statistically different between the two groups. The HRB impairment index, a measure of central nervous system (CNS) functional integrity, did not differ significantly between the Ranch Hand and Comparison groups. Strong covariates in the adjusted analysis were age, race, and education. Because of alternate statistical models and slightly different psychological testing parameters, a direct contrast between the psychological results of the Baseline and followup examinations was not always possible. However, several broad patterns were observed: (1) the discordance between distributional tests and results from traditional statistical models of the MMPI variables was noted with data from both examinations; (2) there was a narrowing of group differences at the followup examination for most subjective variables, either by a decrease in Ranch Hand reporting, or by an increase in Comparison reporting; and (3) as at the Baseline, functional CNS testing, as measured by the HRB impairment index, showed no group differences, and did not support an organic basis for differences in self-reported symptomatology. The longitudinal analysis of two MMPI scales, depression and denial, showed a significant reversal of depression seen at Baseline in the high school-educated Ranch Hands. The determination of PTSD in both Air Force cohorts by a relatively new MMPI scale showed a prevalence rate of less than 1 percent. This low rate is strongly influenced by characteristics of the study population (e.g., age, education, and officer ratio). Unadjusted exposure index analyses did not reveal any patterns consistent with a dose-response relationship. For the adjusted exposure analyses, approximately one-third presented exposure interactions with the covariates of race, education, and age, but no consistent pattern could be identified. In conclusion, some test measures of psychological health (MMPI and CMI) did not show substantial adverse effects for either group. Significant test results were present in both groups or were noted in specific subgroups of a covariate.' Educational level, age, and alcohol use showed strong effects on the psychological scales and scores in this psychological assessment. There was a subtle but consistent trend for more favorable subjective test results at the followup examination for the Ranch Hands relative to the Comparisons. Testing of the CNS by the HRB demonstrated an almost identical prevalence of pathology in both groups. 12-44 �CHAPTER 12 REFERENCES 1. Peterson, R.E., M.D. Seefeld, B.J. Christian, C.L. Potter, C.K. Kelling, and R.E. Keesey. 1984. The wasting syndrome in 2,3,7,8-tetrachlorodibenzo-p-dioxin toxicity: Basic features and their interpretation. In Banbury report 18: Biological mechanisms of dioxin action, ed. A. Poland and R.D. Kimbrough, pp. 291-308. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory, 2. Dougherty, J.A., G.E. Schulze, R.T. Taylor, and J. Blake. 1984. Behavioral toxicity of an agent orange component: 2,4-D. Oral presentation to the Veterans Administration Advisory Committee on Health-Related Effects of Herbicides, Washington, D.C., December 11, 1984. 3. Ashe, W.F., and R.R. Suskind. 1949, 1950. Reports on chloracne cases, Monsanto Chemical Company, Nitro, West Virginia. In Report of the Kettering Laboratory, December 1949 and April 1950. 4. Suskind, R.R. July 1953. A clinical and environmental survey, Monsanto Chemical Company, Nitro, West Virginia. In Report of the Kettering Laboratory, July 1953. 5. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichloro-phenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 6. Suskind, R.R., and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 7. Baader, E.W., and A.J. Bauer. 1951. Industrial intoxication due to pentachlorophenol. Ind. Med. Surg. 20:289-290. 8. Suskind, R.R. 1977. Chloracne and associated health problems in the manufacture of 2,4,5-T. Report to the Joint Conference, National Institute of Environmental Health Sciences and International Agency for Research on Cancer, World Health Organization, Lyon, France, January 1977. 9. Goldman, P.J. 1973. Schweist akute Chlorakne, eine Massenintoxikation durch 2,3,7,8-Tetrachlorpdibenzodioxin (Severe, acute chloracne, a mass intoxication due to 2,3,7,8-tetrachloridbenzo-dioxin). Per Hautarzt. 24(4):149-152. 12-45 �10. Vos, J.G., T.J. Sterringa, D. Zellenrath, H.J. Docter, and L.M. Daldkerup. 1977. TCDD accident at a chemical factory in the Netherlands. Report to the Joint Conference, National Institute of Environmental Health Sciences and International Agency for Research on Cancer, World Health Organization, Lyon, France, January 1977. 11. Telegina, K.A., and L.J. Bikbulatova. 1970. Affection of the follicular apparatus of the skin in workers employed in the production of the butyl ester of 2,4,5-T. Vestnik. Derm. Ven. 44:35-39. 12. Jirasek, L., J. Kalensky, K. Kubec, et al. 1974. Acne chlorina, porphyria cutanea tarda and other manifestations of general intoxication during the manufacture of herbicides, part 2. Czech. Dermatol. 49(3):145-157. 13. Pazderova-Vejlupkova, J., M. Nemcova, J. Pickova, L. Jirasek, and E. Lukas. 1981. The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men. Arch. Environ. Health 36:5-11. 14. Poland, A.P., D. Smith, G. Metter, and P. Fossick. 1971. A health survey of workers in a 2,4-D and 2,4,5-T plant, with special attention to chloracne, porphyria cutanea tardas, and psychologic parameters. Arch. Environ. Health 22(3):316-327. 15. Oliver, R.M. 1975. Toxic effects of 2,3,7,8-tetrachloro-dibenzo-l, 4-dioxin in laboratory workers. Br. J. Ind. Med. 32:46-53. 16. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, W.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. JAMA 255:2031-2038. 17. Flicker, M.R., and A.L. Young. 1983. Evaluation of veterans for agent orange exposure. Presented at the Symposium on Chlorinated Dioxins and Dibenzofurans in the Total Environment, given before the Division of Environmental Chemistry, American Chemical Society, Washington, D.C, September 1983. 18. Blackburn, A.B. 1983. Review of the effects of agent orange: A psychiatric perspective on the controversy. Military Hed. 148:333-340. 19. Diagnostic and statistical manual of mental disorders, 3rd. ed. 1980. Washington, D.C.:AmericanPsychiatry Association. 20. Faltus, F.J., A.O. Sirota, J. Parsons, M. Daamen, and M.L. Schare. 1986. Exacerbation of post-traumatic stress disorder symptomatology in Vietnam veterans. Military Med. 151:648-649. 21. Van Putten, T., and J. Yager. 1984. Posttraumatic stress disorder. Arch. Gen. Psychiatry 41:411-413. 12-46 �22. Atkinson, R.M., R.G. Henderson, L.F. Sparr, and S. Deale. 1982. Assessment of Vietnam veterans for posttraumatic stress disorder in veterans disability claims. Am. J. Psychiatry 129:1118-1121. 23. Borus, J.F. 1974. Incidence of maladjustment in Vietnam returnees. Arch. Gen. Psychiatry 30:554-557. 24. Keane, T.M., R.F. Malloy, and J.A. Fairbank. 19B4. Empirical development of an MMPI subscale for the assessment of combat-related posttraumatic stress disorder. J. Consulting and Clinical Psychology 52:888-891. 25. Yager, T., R. Laufer, and M. Gallops. 1984. Some problems associated with war experience in men of the Vietnam generation. Arch. Gen. Psychiatry 41:327-333. 26. Laufer, R.S., M.S. Gallops, and E. Frey-Wouters. 1984. War stress and trauma: The Vietnam veteran experience. J. Health and Social Behavior 25:65-85. 27. Sierles, F.S., J.J. Chen, R.E. McFarland, and M.A. Taylor. 1983. Post-traumatic stress disorder and concurrent psychiatric illness: preliminary report. Am. J. Psychiatry 140:1177-1179. A 28. Robins, L.N., J.E. Helzer, K.S. Ratcliff, and W. Seyfried. 1982. Validity of the diagnostic interview schedule, version II: DSM-III diagnoses. Psychol. Med. 12:1855-1870. 29. Cornell University Medical College. 1949. Cornell medical index health questionnaire. Ithaca, New York: Cornell University. 12-47 �CHAPTER 13 GASTROINTESTINAL ASSESSMENT INTRODUCTION This system assessment centers on reported peptic ulcer and liver disease, and current hepatic function and porphyria as determined by comprehensive laboratory testing. The liver is a major target organ for single high-dose and continued low-dose exposure to chlorophenols and TCDD. Peptic/stomach ulcer disease and porphyria cutanea tarda (PCT) are suspected clinical endpoints following moderate- to high-level exposures. A variety of experimental animal studies ~ have demonstrated hepatic dysfunction and porphyria following a wide range of exposures to TCDD. The effects of exposure, as measured by enzymatic change, however, generally appear to be more related to species than to dose and route of administration. Gross organ pathology in the digestive system and associated clinical symptoms have been observed following TCDD oral administration to (or accidental ingestion by) animals. Pathological lesions have included gastric ulcers, metaplasia of the gastric mucosa, ileitus, hepatic hypertrophy and degeneration, hepatic parenchymal cell necrosis, and hepatic lipid accumulation. Scientific interest has centered on changes in hepatic enzymes following TCDD administration. Clearly, TCDD has proved to be an exceptional inducer of hepatic enzymes and mixed function oxidases, and a powerful inhibitor of other enzymes. Specifically, the induction of cytochrome P-450, a ferrocytochrome enzyme, by TCDD has been demonstrated in many species and most of their tissues. Further, marked increases in cytochrome P-450 have been implicated in the mechanism of hepatotoxicity, although other factors, such as genetic susceptibility via the Ah locus, iron levels, and lipid peroxidation (but not vitamin A), are also contributory. TCDD has also been shown to produce hepatic porphyria in animals by a reduction in uroporphyrinogen decarboxylase, possibly due to the activation of the P-450 enzyme. ' The porphyriogenic effect of TCDD has also been influencedl fby genetic susceptibility, iron levels, sex, and ambient tem2 perature. In correlation with some human studies, hexachlorobenzene was found to be more porphyriogenic than TCDD. Numerous morbidity studies, predominantly from the industrial sector, have noted significant abnormal liver function in exposed workers, with and without the presence of clinical hepatic disease. Abnormal liver function test results have been found for direct bilirubin, alkaline phosphatase, triglycerides, cholesterol, serum glutamic-oxaloacetic transaminase (SCOT), gamma-glutamyl transpeptidase (GGTP), urine d-glucaric acid, etc. ~ The consistent finding of elevated cholesterol levels may have predictive significance with respect to future heart disease (see Chapter 15), but at present there is no evidence for this. 13-1 �Contemporary studies have focused on two indirect measures of hepatic microsomal activity, GGTP and urine d-glucaric acid. In the study of the English industrial incident, several Seveso investigations, and the two studies of the Monsanto plant in Nitro, West Virginia, there was modest agreement in observing elevated GGTP and urine d-glucaric acid levels in exposed individuals.1 '19'2 ' Common to all studies was the observation that individuals with chloracne manifested significantly more abnormal liver function tests than exposed individuals without chloracne or unexposed individuals, suggesting a link to TCDD exposure. Several industrial studies have shown altered porphyrin excretion patterns (predominantly an increase in uroporphyrin) or clinical evidence of PCT, particularly in chronically exposed workers.2 " Individuals with low chronic exposure or high acute exposure (Seveso) have not shown these signs. Further, detailed reviews of the suspected association have identified the following scientific study design and interpretive problems: (1) multiple etiologies of PCT or abnormal porphyrin excretion patterns (chemical exposure, genetic makeup, alcohol consumption), (2) misdiagnosis of PCT, and (3) confounding of chemical exposures for the industrial cohorts. Some investigators believe that the PCT cases found in the early U.S. and European studies were more likely caused by exposure to chlorobenzenes than to TCDD. Overall, the evidence at present is inconclusive to establish a causal association between PCT and TCDD exposure. A recent industrial study based on questionnaire data has suggested an association of stomach/peptic ulcers with exposure to TCDD. This finding at the Monsanto plant differs from similar research using a slightly different cohort at the same plant which produced a negative conclusion on peptic ulcer disease. The gastric ulcer-TCDD association has not been reported in other cohort dioxin morbidity studies, but ulcer disease has generally not been a major research focus. The preliminary gastric ulcerTCDD association is fortified somewhat by studies that have shown significant gastric mucosal damage in monkeys following oral administration of TCDD. Baseline Summary Results The 1982 AFHS examination conducted an extensive evaluation of hepatic status by questionnaire, physical examination, and laboratory testing. The questionnaire elicited data on liver conditions, liver disease, and symptoms compatible with PCT, as well as detailed information on PCT risk factors (e.g., alcohol consumption, chemical exposures). The physical examination measured hepatomegaly when present and determined liver function and porphyrin patterns by a comprehensive battery of 12 laboratory tests. The questionnaire showed that Ranch Hands reported more miscellaneous liver conditions (verified by medical record reviews) and more skin changes compatible with PCT than their Comparisons. Although the PCT-reported data were statistically significant, no cases of PCT were diagnosed at examination in either cohort. The physical examination detected a twofold increase in hepatomegaly in the Ranch Hands, but the numbers were small and not statistically significant. Many of the laboratory test results demonstrated statistical interactions with the covariates. These interactions can be interpreted as being 13-2 �suggestive of an herbicide effect. Ranch Hands had slightly higher GGTP and lactic dehydrogenase (LDH) results and lower cholesterol levels; no differences were found for bilirubin or alkaline phosphatase levels. SCOT, serum glutamic-pyruvic transaminase (SGPT), and LDH results in the Ranch Hands interacted with the covariates alcohol, degreasing chemicals, and industrial chemicals differently than they did in the Comparisons. All of these two-factor interactions were statistically significant (p<0.05). There were no significant group differences in uroporphyrin, coproporphyrin, or d-aminolevulinic acid levels, nor did any test set support a diagnosis of PCT. Exposure analyses were essentially negative. The comprehensive hepatic evaluation did not reveal any consistent pattern of significant liver damage in the Ranch Hand group. Nevertheless, because of subtle profile differences in conjunction with questionnaire results and recent literature citations, the gastrointestinal system continues to be targeted for intensive examination throughout all phases of the followup effort. Parameters of the 1985 Gastrointestinal Assessment The 1985 AFHS examination continued the emphasis on hepatic function and expanded the porphyrin test battery to six assays. In addition, new components were added to the questionnaire to assess past and current diagnosed peptic ulcer disease, along with a series of screening questions to assess possible undiagnosed disease. Covariate data on aspirin usage, blood group, and family history of peptic ulcer were likewise obtained. Additional probes on intestinal parasites, gallbladder disease, and other liver conditions were also added. Because of the known profound effects of alcohol ingestion on hepatic function, a detailed alcohol consumption history was obtained by questionnaire. Thus, the dependent variables and covariates in the analyses below reflect a substantial enhancement over those assessed in the 1982 Baseline examination. Because of the effects of increased body temperature and past/current hepatitis B on some liver function tests, participants with a fever of 100 or more degrees Fahrenheit and/or a positive hepatitis B surface antigen (HBgAg) test were excluded from the analyses. Categorization of continuous clinical variables to dichotomous variables was largely accomplished by use of normal test values from the SCRF laboratory. Minor numeric differences in the tables that follow are due to an occasionally missing value. The analyses are generally based on 1,009 Ranch Hands and 1,289 total Comparisons after removal of the febrile and positive HBgAg participants. The statistical analyses relied largely on general linear models (SAS®-GLM), logistic regression techniques (BMDP®-LR), and log-linear models (BMDP®-4F). Parallel analyses using Original Comparisons are found in Tables K-7 to K-16 of Appendix K. RESULTS AND DISCUSSION This chapter, entitled "Evaluation of Hepatic Status" in the Baseline Report, incorporates the new elements of peptic ulcer disease and mortality from diseases of the digestive system; hence, the chapter name change to "Gastrointestinal Assessment." 13-3 �Because of the importance of gastrointestinal disorders, numerous historical and laboratory variables were chosen for evaluation. The analyses are reported in the following order: questionnaire data, mortality data, physical examination findings, laboratory results, exposure index analyses, and representative longitudinal analyses. Questionnaire Data: Liver Disorders At the followup examination, each participant was asked whether he had developed hepatitis, jaundice, cirrhosis, or other liver disorders during the interval 1982 to 1985. Affirmative responses were subsequently subject to verification by medical record reviews. Since the Baseline interview, eight Ranch Hands and five Comparisons cited a verified history of hepatitis (p=0.264); four Ranch Hands and five Comparisons reported a subsequently verified history of enlarged liver (p=0.999); one from each group noted a verified symptom of jaundice; one Ranch Hand cited a confirmed interval history of cirrhosis; and six Ranch Hands and six Comparisons gave verified histories of seven miscellaneous liver disorders (p=0.774). Table 13-1 presents the ICD code and descriptive diagnosis of the miscellaneous liver disorders by group. Because the number of respondents with new liver disorders was small and precluded meaningful analyses, the verified interval history was added to the verified Baseline history to assess possible lifetime differences for liver disease. These combined results are presented in Table 13-2. On the basis of combined data, the verified questionnaire responses for historic hepatitis, jaundice, cirrhosis, enlarged liver, and miscellaneous liver disorders did not vary significantly between the Ranch Hand and Comparison groups. The results for miscellaneous liver disorders differed from the Baseline findings. At Baseline, significantly more Ranch Hands than Original Comparisons had a verified liver disorder other than jaundice, hepatitis, or cirrhosis (13/1,045 versus 1/773; p=0.006). Subsequent to Baseline, the status of one additional Ranch Hand disorder and one more Original Comparison disorder was verified. Including these two new verified conditions with the data from replacement and shifted Comparisons, the group contrast at Baseline would have been of borderline significance (14/1,045 versus 7/1,224; p=0.077). Combining these Baseline data with the followup data resulted in nonsignificant lifetime results. However, the combined Baseline and interval analysis contrasting the Ranch Hands and the Original Comparisons was marginally significant (p=0.065) due to the contribution of the significant Baseline results. The verification status of reported liver symptoms and diseases is presented in Table 13-3. The data reflect the proportions of historic reporting that were verified by medical record reviews, and are contrasted by group for each variable. These data showed that the proportion of verified disease was not statistically significant between groups except for the category of enlarged liver which,showed a higher confirmation rate in the Comparison group. Thus, over-reporting or symptom/disease misclassification by the participants was not a function of group membership. 13-4 �TABLE 13-1. Number of Other Liver Conditions Reported by Study Participants at Followup by Group (Verified by Medical Record Review) ICD* Code (Meaning) Group Ranch H a n d C o m p a r i s o n 5713 (Alcoholic Liver Damage) 1 57420 (Calculus of Gallbladder without Mention of Cholecystitis) 0 7891 (Hepatomegaly) 1 1 7904/7905 (Enzyme Elevation) 3 0 7948 (Abnormal Liver Scan) 1 1 E9426 (Adverse Effect of Drug) 0 1 M81406 (Adenocarcinoma) 0 1 6 6 Total *ICD = International Classification of Diseases. 13-5 �TABLE 13-2. Unadjusted Analyses for Baseline and Interval History of Liver Disease by Group (Verified by Medical Record Review) Group Ranch Hand Disease Statistic Number Comparison Percent Number Est. Relative Percent Risk (95% C.I.) p-Value Hepatitis (Viral and Alcoholic) n Yes No 1,016 37 979 3.6 96.4 1,293 43 1,250 3.3 96.7 1.10 (0.70 ,1.72) 0.731 Jaundice n Yes No 1,016 20 996 2.0 98.0 1,293 28 1,265 2.2 97.8 0.91 (0.51 ,1.62) 0.771 n Yes No 1,016 3 1,013 0.5 99.5 1,293 . 2 1,291 0.2 99.8 1.91 (0.32 ,11.46) 0.660 Enlarged Liver n Yes No 1,016 17 999 1.7 98.3 1,293 24 1,269 1.9 98.1 0.90 (0.48 ,1.68) 0.874 Miscellaneous Liver Disorders n Yes No 1,016 17 999 1.7 98.3 1,293 13 1,280 1.0 99.0 1.68 Cirrhosis 13-6 (0.81 ,3.47) 0.195 �TABLE 13-3. Medical Record Verification of Reported Liver Symptoms and Diseases by Group (Baseline and Interval Questionnaires Combined) Group Variable Verification Status Ranch Hand Comparison 47 44 3 53 48 5 37 78.7 43 81.1 43 23 20 59 35 24 20 46.5 28 47.5 7 5 2 3 3 0 3 42.9 2 66.7 30 29 I 29 29 0 11 56.7 24 82.8 21 20 1 14 14 0 17 94.4 13 92.9 Hepatitis Number Reported Medical Records Reviewed Medical Records Pending or Not Released Number Verified Percent Verified Jaundice Cirrhosis Enlarged Liver Miscellaneous Liver Disorders Number Reported Medical Records Reviewed Medical Records Pending or Not Released Number Verified Percent Verified Number Reported Medical Records Reviewed Medical Records Pending or Not Released Number Verified Percent Verified Number Reported Medical Records Reviewed Medical Records Pending or Not Released Number Verified Percent Verified Number Reported Medical Records Reviewed Medical Records Pending or Not Released Number Verified Percent Verified 13-7 p-Value 0.806 0.999 0.999 0.047 0.627 �Peptic Ulcer Diseases The primary purpose of these analyses was to compare the ulcer disease experience of the Ranch Hand and Comparison groups. Since blood type has been reported to affect the incidence of peptic ulcer disease, blood type was used as a covariate in these analyses. The military medical and personnel records of the 2,309 study participants were reviewed to determine the blood type as recorded in these sources. The distribution of blood types in the two groups is shown in Table 13-4. TABLE 13-4. Unadjusted Analysis of Blood Type by Group Blood Type 0 Group A B AB Number Percent Number Percent Number Percent Number Percent Ranch Hand 378 45.4 334 40.1 87 10.5 33 4.0 Comparison 504 46.4 425 39.1 125 11.5 33 3.0 Total* 832 1,087 p=0. 60 *184 Ranch Hands and 206 Comparisons missing from blood type analysis. The blood type distribution was not significantly different in the two groups (p=0.60), and was similar to the distribution of blood types in the general U.S. white male population (p=0.57). Both physical examination diagnoses and questionnare responses to questions concerning ulcers were used as sources of data on the-occurrence of ulcer disease. A total of 58 participants was diagnosed as having ulcer disease at the time of the examination; however, 13 had to be deleted from the analyses of physical examination data and 15 from the analyses of questionnaires due to missing data on blood type. On questionnaires, 42 reported currently having ulcers and an additional 126 reported having had ulcers in the past. These data are summarized in Table 13-5. A three-factor log-linear analysis (group, ulcer, blood type) of data from the physical examination showed a significant three-factor interaction, with the Ranch Hand rate being higher in blood types AB and 0, and lower for types A and B (p=0.03). Stratified analyses of each blood type were conducted and did not reveal any statistically significant group differences. These data are shown in Table 13-6. 13-8 �TABLE 13-5. Frequency of Diagnosed and Reported Ulcer Disease by Group Group Ranch Hand Variable Statistic Number Percent Comparison Number Percent Total Diagnosed Disease (Physical Examination Data) n Yes No 832 19 813 2.3 97.7 1,087 26 1,061 2.4 97.6 1,919 45 1,874 Reported Disease (Questionnaire Data) n Current Past None 832 22 53 757 2.6 6.4 91.0 1,085 20 73 992 1.8 6.7 91.4 1,917 44 126 1,749 A three-factor log-linear analysis of questionnaire data was also performed. This analysis looked at current and past history of ulcer disease. No significant group differences or multifactor interactions were seen, with all p-values being greater than 0.10. These analyses demonstrated overall group equivalence within the Ranch Hand and Comparison groups with respect to blood type and present and past ulcer disease. Mortality Count Data Linkage of digestive system mortality to observed historic or examination morbidity has not been explored in this report; the linkage process, with the use of the Comparison replacement strategy, remains an open research issue. From a broader perspective, however, review of mortality count data in conjunction with current morbidity data may be useful in identifying disease pattern(s) with respect to group membership, organ-specific disease, and important covariates. For these purposes, the latest mortality count data (as of 31 December 1985) are summarized in Table 13-7. These data showed a large mortality contribution (approximately 50%)from liver disease in both groups and.a relative excess in Ranch Hands as contrasted to Comparisons. For malignant neoplasms, there was a relative excess in the Comparison group. There is also the suggestion that alcohol is an important risk factor. The relative excess of malignant neoplasms in the Comparison group is also striking. Overall, the slight excess of digestive system mortality in the Ranch Hands and the differences in distribution of 13-9 �TABLE 13-6. Unadjusted Analyses of Peptic Ulcer Disease by Blood Type by Group Group Ranch Hand Comparison Blood Type Statistic Number Percent Number Percent Est. Relative Risk (95% C.I.) p-Value 0 n Yes No 378 13 365 3.4 96.6 504 11 493 2.2 97.8 1.60 (0.70,3.60) 0.37 A n Yes No 334 4 330 1.2 98.8 425 12 413 2.8 97.2 0.42 (0.14,1.29) 0.21 B n Yes No 87 0 0.0 87 100.0 125 3 122 2.4 97.6 0.27a AB n Yes No 33 2 31 33 0 33 0.0 100.0 0.49a 6.1 93.9 —Estimated relative risk and confidence interval not calculated due to zero count in a cell. "Fisher's exact test. 13-10 �TABLE 13-7. Frequency of Digestive System Mortality by Group Deaths, by Group ICD Code Ranch Hand 1:5 Comparison 1 0 3 0 1 1 0 2 Total 2 6 5 1 1 0 1 15 8 Pancreatitis (5770) Alcoholic cirrhosis (5712) Nonalcoholic cirrhosis (5715) Nonalcoholic fatty liver (5718) Chronic liver disease (5728) Alcoholic liver disease (5711) Duodenal ulcer (5325) Malignant neoplasm (150-159) 31 deaths by cause in the two groups raise the issue of competing mortality. Interpretation of the analyses in this report of hepatic function and liver disease, with alcohol consumption taken into account, should be reviewed in the light of these mortality data. Physical Examination Data Gastrointestinal dysfunction was not a major focus of the physical examination except for a comprehensive biochemical profile of the liver. Consequently, only data on hepatomegaly were analyzed, and results of the analysis are shown in Table 13-8. The analysis showed a marginally significant excess (eight cases versus three) of hepatomegaly in the Ranch Hands (p=0.069). These results were in relative contrast to the Baseline examination findings of 1.56 percent and 0.78 percent in the Ranch Hand and Comparison groups, respectively (p=0.138), in the sense that fewer abnormalities were detected at the followup, although at both examinations the difference favored the Comparisons. The group data for hepatomegaly were pooled and compared to the covariates of age, race, occupation, current alcohol use (one or less drinks per day, more than one to four drinks per day, and more than four drinks per day), lifetime exposure to industrial chemicals, and lifetime exposure to degreasing chemicals. Only age and occupation showed significant associations with hepatomegaly (p=0.018, p=0.026, respectively). Because of sparse data, an adjusted analysis was not conducted. General Laboratory Examination Data As in the Baseline Report, the followup examination emphasized evaluation of laboratory data, particularly for hepatic function. Thus, this 13-11 �TABLE 13-8. Unadjusted Analysis of Enlarged Livers Diagnosed at Physical Examination by Group* Enlarged Liver Yes Group Number No Percent Number Percent Total p-Value 0.069 Ranch Hand 8 0.8 1,002 99.2 1,010 Comparison 3 0.2 1,287 99.8 1,290 *Excludes participants with positive HB a Ag. section reports on nine laboratory tests of hepatic function and on two tests reflecting porphyrin metabolism. Normal ranges for these 11 variables as determined by the SCRF and the Mayo Clinic Laboratories are presented in Table 13-9. Only values greater than the normal range were considered important in the assessment of dysfunction. Analyses of the nine hepatic variables were adjusted for the covariates of age, race, occupation (OCC), current alcohol use (ALC), days of exposure to industrial chemicals (1C), and days of exposure to degreasing chemicals (DC). For the two porphyrin analyses, blood urea nitrogen was used as a covariate. Because the hepatic test variables encompass acute to chronic effects, there was no "ideal" alcohol covariate (e.g., drink-years, current alcohol consumption in drinks per day). The covariate alcohol use was obtained from questionnaire data, centering on daily alcohol consumption (beer, wine, liquor) for those participants who reported drinking at least one drink in the 2 weeks preceding the examination. Thus, the alcohol covariate measures recent drinking intensity and may be more useful in" adjustment of acute variables (e.g., GGTP, SGPT) than variables related to chronic liver dysfunction (e.g., bilirubin determinations, alkaline phosphatase). Exposure to industrial chemicals and degreasing chemicals was measured in cumulative days of unprotected exposure, and was derived from the 1982 and 1985 questionnaires. These data, therefore, represent lifetime exposure. Exclusion categories consisted of fever (over 100 degrees Fahrenheit) and positive HBgAg tests, because of the known effects of these conditions on liver function tests. Three participants (two Ranch Hands, one Comparison) were excluded because of fever, and eight (five Ranch Hands, three Comparisons) because of a positive HBgAg test (seven positive, one missing). In addition, due to missing alcohol data, nine other individuals (six Ranch Hands, three Comparisons) were deleted from the analyses when current alcohol use was found to be a significant covariate. 13-12 �TABLE 13-9. Laboratory Norms for Nine Hepatic Function Variables and Two Porphyrin Determinations Variable SCOT SGPT GGTP Alkaline Phosphatase Total Bilirubin Direct Bilirubin LDH Cholesterol8 Triglycerides* Uroporphyrinb Coproporphyrin Unit SCRF Normal SCRF Abnormal U/L U/L U/L U/L rag/dl mg/dl U/L mg/dl mg/dl mg/24 hrs mg/24 hrs 27-47 3-36 15-85 50-136 <1.5 <0.36 100-190 <260 <320 <46 <96 >48 >37 >86 XL37 >1.5 20.37 >191 >261 >321 >47 >97 a SCRF provides age-dependent normal ranges; these values represent the maximum normal limits for those older than 40. b Performed at the Mayo Clinic. 13-13 �Statistical Analyses The nine dependent variables from the hepatic battery were subjected to three types of basic analyses: (1) a continuous dependent variable adjusted by continuous covariates (CC), (2) a continuous dependent variable adjusted by discrete covariates (CD), and (3) a discrete (categorical) dependent variable adjusted by discrete covariates (DD), except for current alcohol use, which was left as a continuous variable for model-fitting and power purposes. General linear models (SAS®) were used for the CC and CD analyses, and BMDP®-LR was used for the DD analyses. As noted in Chapter 7, Statistical Methods, all adjustments were carried out with the simplest model, including all significant covariates and twoand three-way interactions. The log transformation was used for the nine hepatic variables and for uroporphyrin, while a square root transformation was employed for the coproporphyrin variable. Since some direct bilirubin values were 0, the value 0.10 was added prior to log transformation. The sample sizes were sufficient to detect a 1.93-fold increase in the frequency of abnormal values for alkaline phosphatase and a 1.42-fold increase in the frequency of abnormal values for SGPT, using a (two-sided) a -level of .0.05 and power 0.80, Further, the sample sizes were sufficient to detect a 0.7 percent mean shift in alkaline phosphatase, a 1.8 percent mean shift in SGPT, and a 2.8 percent mean shift in uroporphyrin values. The results of the analyses on the 11 dependent variables are presented in the following summary tables (Tables 13-10 through 13-12), followed by descriptive narrative text. The summary tables are in the following logical order: unadjusted results, covariate tests of association, and adjusted results. Tables K-l and K-2 of Appendix K summarize interactions from the statistical analyses. All analytic information on any given variable can be obtained by scanning the summary tables. The following discussion condenses the key information on each dependent variable. Group-by-covariate interactions are narratively presented. The variables are organized in the same order as given in the tables. Serum Glutamic-Oxaloacetic Transaminase (SCOT) The unadjusted continuous (group means) and categorical (percent abnormalities) tests showed no statistically significant differences between groups (p=0.298 and p=0.999, respectively). Tests of association with the covariates using pooled group categorical data demonstrated the significant effect of race (a higher percentage of abnormalities in Blacks than nonblacks, 13.5% versus 7.6%; p<0.022) and current alcohol use (21.2% abnormal values associated with more than four drinks per day, 9.0% abnormals for more than one to four drinks per day, and 5.8% for one or less drinks per day; p<0.001). Similarly, the mean SGOT levels differed significantly between races (p<0.001) and by current alcohol use (p<0.001). The CC adjusted model showed no significant group differences (p=0.309). Significant covariates were race, an interaction of current alcohol use-bydegreasing chemicals, and an interaction of current alcohol use-by-age (all 13-14 �TABLE 13-10. Unadjusted Continuous and Categorical Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Group Group Variable Statistic Ranch Hand Comparison SCOT n Mean 95% C.I. Number/% Normal High 1,009 33.5 (32.8,34.1) 1,289 33.0 (32.5,33.5) n Mean 95% C.I. Number /% Normal High 1,009 21.6 (20.9, 22.3) n Mean 95% C.I. Number/% Normal High 1,009 32.8 (31.4,34.3) n Mean 95% C.I. Number/% Normal High 1,009 91.8 (90.4, 93.3) n Mean 95% C.I. Number /% Normal High 1,009 0.74 (0.73,0.76) SGPT GGTP Alkaline Phosphatase Total Bilirubin 929 92.1% 80 7.9% 872 86.4% 137 13.6% 919 91.1% 90 8.9% 953 94.5% 56 5.6% 982 97.3% 27 2.7% 13-15 Est. Relative Risk (95% C.I.) p-Value 0.298 1,187 92.1% 1.00 (0.74,1.36) 0.999 102 7.9% 1,289 22.5 (21.9,23.1) 0.051 1,102 85.5% 0.93 (0.73,1.17) 0.546 187 14.5% 1,289 32.4 (31.2,33.6) 0.632 1,172 90.9% 0.98 (0.74,1.31) 0.942 117 9.1% 1,289 89.3 (88.1,90.6) 0.009 1,236 95.9% 1.37 (0.93,2.01) 0.114 53 4.1% 1,289 0.75 (0.74,0.76) 0.576 1,250 97.0% 0.88 (0.54,1.45) 0.706 39 3.0% �TABLE 13-10. (continued) Unadjusted Continuous and Categorical Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Group Group Variable Statistic Ranch Hand Comparison Direct Bilirubin n Mean 95% C.I. Number/% Normal High 1,009 0.18 (0.17,0.18) 1,289 0.18 (0.17,0.18) n Mean 95% C.I. Number /% Normal High 1,009 123.5 (122.2,124.8) n Mean 95% C.I. Number /% Normal High 1,009 214.3 (211.8,216.8) Est. Relative Risk (95% C.I.) p-Value LDH Cholesterol 971 38 999 10 96.2% 3.8% 99.0% 1.0% 863 85.5% 146 14.5% 0.981 1,246 96.7% 1.13 (0.73,1.77) 0.649 43 3.3% 1,289 123.9 (122.7 ,125.2) 0.655 1,272 98.7% 0.75 (0.34,1.64) 0.560 17 1.3% 1,289 215.0 (212.8 ,217.2) 0.688 1,082 83.9% 0.88 (0.70,1.11) 0.322 207 16.1% Triglycerides n Mean 95% C.I. Number/% Normal High 1,009 1,289 118.5 117.3 (113.8,123.3) (113.4 ,121.4) Uroporphyrin n Mean 95% C.I. 1,006 16.9 (16.2,17.7) 1,286 17.9 (17.3,18.6) 0.048 1,008 119.1 (116.2,122.0) 1,287 115.6 (113.0,118.2) 0.081 Coproporphyrin n Mean 95% C.I. 941 93.3% 68 6.7% 13-16 0.719 1,210 93.9% 1.11 (0.79,1.55) 0.549 79 6.1% �TABLE 13-11. Association Between Nine Hepatic Function Variables and Two Forphyrin Determinations and Six Covariates in the Combined Ranch Hand and Comparison Groups Analysis* Age Race SCOT C D NS NS SGPT C D <0.001 0.001 GGTP C D 0.012 NS Alkaline Phosphatase C D NS NS Total Bilirubin C D NS NS* NS <0.001 Direct Bilirubin C D NS NS LDH C D Cholesterol Industrial Degreasing Occupation Alcohol Chemicals Chemicals <0.001 0.022 Variable NS NS <0.001 <0.001 NS NS NS NS <0.001 <0.001 NS NS 0.017 NS 0.032 NS <0.001 <0.001 NS NS NS NS <0.001 0.003 <0.001a NS*a 0.011 NS 0.008 NS NS NS NS NS NS* 0.015 NS NS <0.001 NS NS NS NS NS <0.001 NS 0.006 NS NS NS NS NS NS NS C D <0.001 0.010 NS NS 0.002 0.008 <0.001 0.018 NS NS NS NS Triglycerides C D <0.001 <0.001 0.031 NS 0.013 NS 0.030 NS NS* NS* 0.019 NS Uroporphyrins C NS NS NS NS NS Coproporphyrins C NS NS <0.001 0.021 NS NS NS NS NS 0.003 NS NS Continuous (C)/Discrete (D). NS: Not significant (p>0.10) NS*: Borderline significant (p.05<p<0.10). a Wine consumption. 13-17 NS NS <0.001 <0.001 0.030 NS NS 0.010 NS* �TAHTK 13-12. Adjusted Continuous and Categorical Analyses for Hepatic Function Variables and T\*> Borphyrin Determinatdons by Group Group Variable Analysis Statistic Ranch Band Comparison CC SOOT CD CD OC p s CD CD CC QGIP CD CD Covariate Ranarks* n 1,003 Adj. Mean 34.8 95* C.I. (33.8,35.7) n 1,003 Adj. Mean **** 95% C.I. **** n 103 ,0 1,286 34.3 — 039 .0 (343.) 3.,53 1,286 **** _ **** **** ' 1,286 1 0 ( . 5 1 4 )0 8 8 . 3 07,.1 .6 n 1,286 ALC*DC(p=0.008), RAGE*DC(p=Q.015) 22.2 — 0.048 AGE*AKXP=0-001),r RACE*IC(p=0.017) (21.3, 23.3) ALD*DC(p=0.032), AG£*ALC(p=0.022) 1,286 — 0.029 OCC*AGE(p=0.026),, IC(p=0.049) 22.9 (21.1,24.8) 1,286 0.93 (0.73,1.18) 0.531 A(£*ALC(p=Q.004) 1,003 21.4 95% C.I. (20.4,22 •4) n 1,003 Adj. Mean 21.9 95% C.I. (20.2,23.8) n 1,003 Adj. Mean SGPT Adj. Relative Risk (95% C.I.) p-Value 1,286 n 103 ,0 Adj. Mean 37.5 37.0 0.575 95% C I ( 5 2 4 . ) ( 4 7 3 . ) . . 3.,01 3.,93 n 103 ,0 1,286 Adj. Mean 4 . 41 068 .6 4. 36 95% C I ( 0 0 4 . ) ( 9 6 4 . ) . . 4.,86 3.^79 n 1,286 1 0 ( . 4 1 3 )0 9 1 . 0 07,.4 . 7 1,003 ALC*DC(D<0.001) AGE*AlC(p<D.001) RACE(p«0.001) GRP*ALC(p=0.048) ALC*IC(p=0.008) DC(p=0.019), RACE(p<0.001) AGE*AI£(p<D.001) OCC*AIJC(D<0.001) RACE ( = . 2 ) p006 AGE*ALC(p<0.001) ,RACE*IC(p=0.011) ALC*DC(p<O.C01), ACE*TC(p=0.009) ACE*ALC(p=0.023), OCC*ALC(p=0.044) RACE(pO.COl) AGE*AIC(p<0.001), RACE(p=0.016) �TSHIE 13-12. Continued) feriahles and Ttao Porphyrin Determinations by Ckoup Group Variable Analysis Statistic Ranch Hand Comparison CO Alkaline Phosphatase CD ED OC Total Bilirubin CD DD OC Direct Bilirubin CD DD n 1,003 Adj. Mean 91.6 95% C.I. (89.4,93.9) n Adj. Mean 95% C.I. n 1,003 **** **** 1,003 1,285 89.1 (87.0,91.2) 1,285 Adj. Relative Risk (95% C.I.) p-Value — Covariate Remarks* AG£*IC(p=0.010), RACE*IC(p=0.007) 0.008 OCC(p<0.001), W3NE(p<0.001) GRP*IC(p=Q.011), #£*IC(p4).019) RACE*IC(p=0.002), OCC(p<0.001) WINE (p<0.001) 1.44 (0.97,2.13) 0.070 WINE*DC(p4).006), AGE*IC(p=0.005) RACE*IC(p=0.004), OCC*IC(p=0.016) _ 1,285 **** — A3G*DC(p=0.039) 0.599 RACE*ALC(p=0.007) RAC£*OCC(p=0.001) n 1,003 Adj. Mean 0.78 95% C.I. (0.75,0.81) 1,286 0.78 (0.75,0.81) n 1,003 Adj. Mean 0.83 95% C.I. (0.79,0.87) n 1,009 1,286 RAGE*ALC(p=0.004) — 0.598 OCC*ALC(p=0.034) 0.83 OOC*RACE(p=0.002) (0.80,0.87) 1,289 0.89 (0.54,1.47) 0.648 RACE(pO.OOl) n 1,003 Adj. Mean 0.18 95% C.I. (0.17,0.20) 1,286 0.18 (0.17,0.19) — 0.972 RACE*ALC(p=0.025) n 1,003 Adj. Mean 0.21 95% C.I. (0.19,0.22) n 1,003 1,286 0.20 (0.19,0.22) 1,286 — DC*IC(p=0.025), ALCMX:(p=0.012) 0.830 RACE*ALC(p=0.019), OCC*ALC(p=0.002) **** **** GRP*IC(p=0.012), RACE(p=0.014) ALC(p=0.026) �TAWR 13-12. (continued) Adjusted Continuous and Categorical Analyses for Hepatic Rncti.cn Variables and Two Borphyrin Deteodnaticns by Group Group Variable Analysis Statistic Ranch Hand Comparison CC LDH CD DD CC 8 Cholesterol CD ED CC Triglycerides CD ED n Adj. Mean 95% C I .. 1,003 Adj. Realtive Risk ( 5 C I ) p-Value 9% . . 1,286 «M Covariate Remarks* GRP*AGE(p=Q.018), OCC*IC(p=Q.014) RACE*IC(p=Q.024) AAAA n 1,003 1,286 RACE(p<D.001), AGE(p<D.001) Mj. Mean 130.0 130.5 — 0.671 DC(p=0.016) 95% C.I. (127.3,132.8) (127.8,133.1) n 1,009 1,289 0.75 (0.34,1.64) 0.560 n 103 ,0 126 ,8 Adj. Mean 295 1. 220.4 — 064 .0 95% C I ( 1 . , 2 . ) ( 1 . , 2 . ) .. 245247 255254 n 103 ,0 126 ,8 Adj. Mean 223.8 249 2. — 058 .4 95% C I ( 1 . , 3 . ) ( 1 . , 3 . ) . . 277201 289210 n 103 ,0 1 2 6 0 8 ( . 8 1 0 ) 0.1S1 ,8 . 5 06,.8 RACE*DC(p=O.021), RACE*OCC(p=0.005) I ( = . 4 ) ALC(p<0.001) Cp003, AG£(p<D.001) RACE*OCC(p=0.027), Al£(p<0.001) AGE(p<0.001) n 103 ,0 126 ,8 Adj. Mean **** **** — 95% C I .. **** **** n 103 ,0 126 ,8 Adj. Mean 125 1. 121 1. — 95% C I ( 0 . , 2 . ) ( 0 . , 2 . ) .. 137119 137112 n 109 ,0 129 ,8 **** GRP*AGE(p=0.015), AIC*DC(p=0.005) RACE*ALC(i>=0.031), OCC(p<D.001) **** RACE*AlJC(p=0.012), AGE(p=d0.029) OCC(p=0.039) OCC(p<D.001), RACE(p<0.001) 0 9 5 AGE(p<0.001), ALC(p=0.038) .0 **** GRP*OCC(p=0.027), RACE ( = . 2 ) p006 IC(p=0.038) �TABLE 13-12. (continued) Adjusted Continuous and Categorical Analyses for Hepatic Flncti.cn Variables and Two Borphyrin Determinations by Group Group Variable Analysis Statistic Ranch Band Comparison CC n Adj. Mean 1,000 **** rtCV Uroporphyrin **A* f t T y;X5 C.I. Coproporphyrin CC Jnlrrlr^Ij Adj. Relative Risk (95%C.I.) p-Value 1,283 **** ^«_ -* _t^t_l- * * * * n 1 0 2 , 0 1 2 4 , 8 Adj. Mean 1 9 3 1 5 7 1. 1 . 95% C I ( 1 . , 2 . ) ( 1 . , 1 . ) . . 164122 132182 "XfCfCX Covariate Remarks* GRP*EUN(p=0.015) DC*OOC(p=O.005) ALC(p=0.026) 0 0 5 BUN(p<0.001) .6 *Abbreviations; GRP: group OCC: occupation ALC: currait alcohol use WINE: vine consumption DC: exposure to degreasing chemicals 1C: exposure to industrial chemicals BIN: blood urea nitrogen — No relative risk or confidence interval given for continuous analyses. **** Group-by-covariate interaction—adjusted mean/relative risk, confidence interval, and p-value are not presented. �with p<0.001). The CD analysis revealed a significant group (GRP)-by-current alcohol use interaction (p=0.048), precluding a direct group contrast. Exploration of the interaction disclosed that the Ranch Hands had a significantly higher (p=0.010) mean SCOT for the more than one to four drinks per day category, whereas there were no significant group differences for the one or less drinks per day or more than four drinks per day categories (see Table K-l of Appendix K). Other significant covariate effects included d.egreasing chemicals (p=0.019), race.(p<0.001), and a. current alcohol use-by-industrial chemical (1C) interaction (p=0,008). The DD SCOT analysis showed no significant group differences (p=0.868). Covariates making significant contributions were race (p=0.026), an age-by-current alcohol use interaction (p<0.001), and an occupation (OCC)-by-current alcohol use interaction (p<0.001). Serum Glutamic-Pyruvic Transaminase (SGPT) The unadjusted categorical analysis was not significant (p=0.546), but the comparison of group means showed a borderline significant result, with the Comparisons having a higher mean SGPT than the Ranch Hands (p=0.051). Covariate associations with the pooled categorical Ranch Hand and Comparison group data showed an inverse relationship (p=0.001) between SGPT levels and age, with 17.1 percent abnormalities for those born in or after 1942, 12.3 percent for those born between 1923 and 1941, and 8.1 percent for those born in or before 1922.' The relationship with current alcohol use was also profound (p<0.001), with 23.4 percent abnormals noted for more than four drinks per day, 15.3 percent abnormals for more than one to four drinks per day, and 12.4 percent for one or less drinks per day. The direction and magnitude of the covariate effects of age and alcohol were quite similar for the tests of association with the mean SGPT level of both groups (p<0.001 for both covariates). No significant group interactions were detected in either the discrete or the continuous analyses. The CC-adjusted analysis yielded a significant group difference, with the Comparisons having a higher group mean than the Ranch Hands (p=0.048). The model was adjusted by the interactions of current alcohol use-by-degreasing chemicals (p=0.008), current alcohol use-by-age (p=0.001), race-by-degreasing chemicals (p=0.015), and race-by-industrial chemicals (p=0.017). The CD model also showed a significantly elevated mean SGPT in the Comparison group (p=0.029). The analysis was adjusted for exposure to industrial chemicals (p=0.049), and the interactions of age-byoccupation (p=0.026), age-by-current alcohol use (p=0.022), and current alcohol use-by-degreasing chemicals (p=0.032). A borderline significant interaction (p=0.0505) between group and current alcohol use was found, but because of modeling strategy, this interaction was not included in the final model. (This interaction is explored further in Table K-l in Appendix K, however.) The DD-adjusted analysis, like the unadjusted discrete analysis, disclosed a nonsignificant group difference (p=0.531). The model was adjusted for an age-by-current alcohol use interaction (p=0.004). Gamma-Glutamyl Transpeptidase (GGTP) The unadjusted contrasts of both mean levels of GGTP and the frequency of abnormalities showed no significant differences between the Ranch Hand and Comparison groups (p=0.632 and p=0.942, respectively). 13-22 �For discrete covariate associations, significance was noted for race, with 14.9 percent abnormals in Blacks and 8.6 percent for nonblacks (p=0.021), and current alcohol use, with 26.1 percent abnormals for more than four drinks per day, 10.5 percent for more than one to four drinks per day, and 6.2 percent for one or less drinks per day use (p<0.001). While the mean level of GGTP was similarly affected by race and current alcohol (p<0.001 for both covariates), it was also influenced by age (30.3 U/L for those born in or before 1922, 33.9 U/L for those born between 1923 and 1941, and 31.1 U/L for those born in or after 1942; p=0.012) and occupation (31.5 U/L for officers, 35.2 U/L for enlisted flyers, and 32.5 U/L for enlisted groundcrew; p=0.032). Each of the three adjusted analyses consistently produced nonsignificant group differences (CC: p=0.575; CD: p=0.668; DD: p=0.971). None of the three models was affected by a group-by-covariate interaction. The CC analysis was adjusted by four covariate interactions: age-by-current alcohol use (p<0.001), race-by-industrial chemicals (p=0.011), current alcohol useby-degreasing chemicals (p<0.001), and age-by-degreasing chemicals (p=0.009). The CD model was adjusted by race (p<0.001), by an age-by-current alcohol use interaction (p=0.023), and by an occupation-by-current alcohol use interaction (p=0.044). The DD analysis was adjusted by race (p=0.016) and by an age-by-current alcohol use interaction (p<0.001). Alkaline Phosphatase The analysis of group mean values showed a significantly higher (p=0.009) Ranch Hand mean (91.8 U/L) than that observed in the Comparison group (89.3 U/L). The unadjusted categorical analysis revealed a higher percentage of Ranch Hand abnormalities (5.6%) than Comparison abnormalities (4.1%), but this difference was not significant (Est. RR=1.37, 95% C.I.: [0.93,2.01], p=0.114). With pooled group data, significant covariate associations were found between the proportion of abnormal values and occupation (p=0.003), industrial chemicals (p=0.030), and marginally significant associations with wine consumption (p=0.056) and degreasing chemicals (p=0.091). The mean value of alkaline phosphatase depended significantly on all four of these covariates. The CC-adjusted analysis also showed a significantly higher mean value of alkaline phosphatase in the Ranch Hand group (p=0.008). The model was adjusted by the significant covariates of wine consumption (WINE) (p<0.001), occupation (p<0.001), and the interactions of age-by-industrial chemicals (p=0.010) and race-by-industrial chemicals (p=0.007). Wine consumption was used as a covariate instead of alcohol intensity since wine showed a very strong negative association with alkaline phosphatase. This effect masked a very weak positive association between beer or liquor consumption and alkaline phosphatase. In the CD model a significant group-by-industrial chemicals interaction was found (p=0.011). Specifically, in those individuals exposed to industrial chemicals, the Ranch Hands had a significantly higher mean value than the Comparisons (p<0.001), whereas in the unexposed stratum, the mean values were not significantly different between groups (p=0.973; see Table K-l of Appendix K). The CD analysis was also adjusted by wine consumption (p<0.001), occupation (p<0.001), and the interactions of age-by-industrial chemicals (p=0.019) and race-by-industrial chemicals (p=0.002). 13-23 �The DD model revealed a marginally significant group difference (Adj. RR: 1.44, 95% C.I.: [0.97,2.13], p=0.070) following adjustment by four significant interactions of wine-by-degreasing chemicals (p=0.006), age-byindustrial chemicals (p=0.005), race-by-ihdustrial chemicals (p=0.004), and occupation-by-industrial chemicals (p=0.016). Total Bilirubin Both the continuous and categorical unadjusted analyses found no significant differences in total bilirubin values between groups (p=0.576 and p=0.706, respectively). The covariate associations for both groups showed a significant effect of race (8.5% abnormal in Blacks versus 2.5% in nonblacks; p<0.001). Significant differences in mean total bilirubin levels were found between occupational groups (0.76 mg/dl for officers, 0.72 mg/dl for enlisted flyers, and 0.75 mg/dl for enlisted groundcrew; p=0.011), and with increasing levels of current alcohol use (0.80 for more than four drinks per day, 0.75 for more than one to four drinks per day, and 0.74 for one or less drinks per day; p=0.008). Further, increasing levels of total bilirubin were marginally associated with age (p=0,093). The CC model, adjusted for the interactions of age-by-degreasing chemicals (p=0.039), race-by-current alcohol use (p=0.007), and race-byoccupation (p=0.001), revealed no significant differences in total bilirubin means between groups (p=0.599). Similarly, the CD analysis found no difference between group means (p=0.598) after adjustment for the interactions of race-by-current alcohol use (p=0.004), occupation-by-current alcohol use (p=0.034), and occupation by race (p=0.002). The DD model, adjusted for race (p<0.001), also failed to detect significant group differences in the proportion of total bilirubin abnormalities (p=0.648). Direct Bilirubin Neither the continuous nor the categorical unadjusted tests disclosed significant differences between the Ranch Hand and Comparison groups (p=0.981 and p=0.649, respectively). A covariate association with the categorical data combined from both groups was noted for race, with 7.8 percent abnormalities found in Blacks as contrasted to 3.3 percent in nonblacks (p=0.015). There was a significant association between mean values of direct bilirubin and current alcohol use (0.21 mg/dl, 0.17 mg/dl, and 0.17 mg/dl for more than four drinks per day, more than one to four drinks per day, and one or less drinks per day, respectively; p<0.001) and a marginally significant difference due to race (0.20 mg/dl for Blacks versus 0.18 mg/dl for nonblacks; p=0.059). For both the CC and CD analyses, no significant group differences were found (p=0.972 and p=0.830, respectively). The CC model was adjusted for a race-by-current alcohol use interaction (p=0.025), and the CD model was adjusted for the significant interactions of race-by-current alcohol use (p=0.019), occupation-by-current alcohol use (p=0.002), current alcohol use-by-degreasing chemicals (p=0.012), and degreasing chemicals-by-industrial chemicals (p=0.025). The DD analysis revealed a group-by-industrial chemical 13-24 �exposure interaction (p=0.012). For participants exposed to industrial chemicals, the Ranch Hands had a higher proportion with abnormal values than the Comparisons (5.3% abnormal versus 2.9%, respectively; p=0.035), whereas there was no group difference for participants not exposed to industrial chemicals (p=0.144). Each stratum of the interaction was adjusted for race (p=0.014) and current alcohol use (p=0.026). The biological relevance of this interaction is unclear at this time. Lactic Dehydrogenase (LDH) No significant differences were found between the groups, either in the proportion of abnormal values (p=0.560) or in the mean levels of LDH (p=0.655). Significant effects for age (121.6 U/L, 124.6 U/L, 135.3 U/L for those born in or after 1942, between 1923 and 1941, and in or before 1922, respectively; p<0.001) and race (129.5 U/L for Blacks versus 123.4 U/L for nonblacks; p=0.006) were found in the tests of mean LDH levels. The CC analysis revealed a group-by-age interaction (p=0.018), although no significant adjusted group differences were found for any of the three age strata. The model was also adjusted for the significant interactions of occupation-by-exposure to industrial chemicals (p=0.014) and race by exposure to industrial chemicals (p=0.024). The CD model revealed no significant group differences after adjustment by age (p<0.001), race (p<0.001), and degreasing chemicals (p=0.016). Similarly, the DD analysis found no significant group differences, and no covariates made a significant contribution to the model. Cholesterol No significant differences were found between groups, either in the proportion of abnormal cholesterol levels (p=0.322) or in mean values of cholesterol (p=0.688) by unadjusted tests. However, in contrast, analysis of the Ranch Hand group versus the Original Comparisons (see Table K-9 of Appendix K) showed that the Comparisons had a significantly higher proportion of abnormal levels than the Ranch Hands (18.3% versus 14.5%, respectively; Est. RR: 0.76, 95% C.I.: [0.60,0.96], p=0.023). This observation was also found at Baseline. Significant covariate associations were noted between the proportion of participants with abnormally high cholesterol levels and age (12.7% for those born in or after 1942, 17.2% for those born between 1923 and 1941, and 18.4% for those born in or before 1922; p=0.010), occupation (14.9% for officers, 20.5% for enlisted flyers, and 13.9% for enlisted groundcrew; p=0.008), and current alcohol use (14.1% for one or less drinks per day, 16.4% for more than one to four drinks per day, and 21.7% for more than four drinks per day; p=0.018). For the associations between mean cholesterol levels and age, occupation, and current alcohol use, the significance of the covariate effects was greater than for the discrete analyses (p<0.001, p=0.002, and p<0.001, respectively). The CC results showed no significant group difference (p=0.604). The model was adjusted by age (p<0.001), current alcohol use (p<0.001), industrial chemical exposure (p=0.043), and the race-by-degreasing chemicals (p=0.021) and race-by-occupation (p=0.005) interactions. The CD analysis was negative for significant group differences (p=0.548). The analysis included the covariate contributions made by age (p<0.001), current alcohol use 13-25 �(p<0.001), and a race-by-occupation interaction (p=0.027). The DD analysis also showed no significant difference between groups for adjusted proportions of participants with abnormal cholesterol levels (p=0.181). Contributing covariates included age (p=0.029), occupation (p=0.039), and a race-bycurrent alcohol use interaction (p=0.012). In all of the discrete cholesterol analyses, the cutpoint of 260 mg/dl was used. Triglycerides In the unadjusted analyses, no significant differences in the proportion of participants with abnormal triglyceride levels or in mean values were found between the Ranch Hand and Comparison groups (p=0.549 and p=0.719, respectively). The covariate tests of association for percent abnormal triglycerides disclosed the significant effect of race (2.1% for Blacks and 6.7% for nonblacks; p=0.031) and a marginally significant association for industrial chemical exposure (p=0.073). For mean triglyceride levels, significant associations for age (p<0.001), race (p<0.001), occupation (p=0.013), current alcohol use (p=0.030), and degreasing chemicals (p=>0.019) were noted, in addition to a marginally significant association with exposure to industrial chemicals (p=0.077). The CC analysis revealed a significant group-by-age interaction (p=0.015), which showed a significantly elevated mean triglyceride level in Ranch Hands (p=0.039) born in or before 1922 as compared to similarly aged Comparisons (see Table K-l of Appendix K). There were no significant differences for the other two age strata. A significant adjusting covariate was occupation (p<0.001); in addition, the current alcohol use-by-degreasing chemicals (p=0.005) and race-by-current alcohol use (p»0.031) interactions were used for adjustment. The CD-adjusted analysis found no significant group differences (p=0.905). The model was adjusted by age (p<0.001), race (p<0.001), occupation (p<0.001), and current alcohol use (p=0.038). The DD analysis found a significant group-by-occupation interaction (p=0.027). Stratification by occupation revealed a significant increase in the proportion of abnormal triglyceride levels for Ranch Hand officers (Adj. RR? 1,77, 95% C.I.: [1.04,3.01], p=0.035) but no significant group differences were discerned for the enlisted flyer or enlisted groundcrew strata. The models were adjusted by race (p=0.026) and industrial chemical exposure (p=0.038). A cutpoint of 320 mg/dl was used to distinguish abnormal from normal. Uroporphyrin The uroporphyrin variable was analyzed only in the continuous form. The unadjusted analysis revealed a significant difference between group means (Comparisons 17.9 mg/24 hrs, Ranch Hands 16.9 mg/24 hrs; p=0.048). A CC model found a significant group-by-blood urea nitrogen (BUN) interaction (p=0.015; see Table K-l of Appendix K). To interpret the interaction, BUN was dichotomized at the median value of 14 mg/dl. Stratifying by BUN levels revealed a significantly greater (p<0.001) uroporphyrin mean for Comparisons than for Ranch Hands for BUN levels of 14 or less mg/dl and a nonsignificant but greater Ranch Hand mean for participants with BUN levels of 13-26 �more than 14 mg/dl. The stratified model was adjusted for current alcohol use (p=0.026) and the occupation-by-degreasing chemicals (p=0.005) interaction. Coproporphyrin As with the uroporphyrin variable, coproporphyrin was analyzed only as a continuous variable. The unadjusted analysis revealed a borderline significant difference in the mean coproporphyrin levels (119.1 mg/24 hrs for Ranch Hands and 115.6 mg/24 hrs for Comparisons; p=0.081). The covariate tests of association detected the significant effects of age (p»0.003) and current alcohol use (p=<0.001). A CC model, adjusted by BUN (p<0.001) and an age-by-current alcohol use interaction (p=0.003) revealed a borderline significant group difference (p=0.065) similar to the unadjusted analysis. The adjusted coproporphyrin means were 119.3 mg/24 hrs and 115.7 mg/24 hrs for the Ranch Hands and Comparisons, respectively. Discussion The results from the nine hepatic and two porphyrin analyses were not totally consistent with the Baseline findings. Several analytical reasons may possibly explain some of these differences, i.e., the adjusted analyses herein used the additional covariates of age, race, and occupation (the matching variables), and all two-way covariate interactions. However, as the Baseline data were not reanalyzed with the model process and total Comparison group used in this report, the contribution of analytic technique versus a true change in hepatic status is unknown. The Baseline Report noted a significantly lower mean cholesterol level in the Ranch Hands (opposite of an expected dioxin effect) and slight tendencies for higher GGTP and LDH values in the Ranch Hands. In this chapter, the analyses have shown equivalent group cholesterol levels, an increased SGFT mean in the Comparisons, an increased mean alkaline phosphatase in the Ranch Hands, an increased uroporphyrin mean in the Comparisons, and a borderline increased coproporphyrin mean in the Ranch Hands. The individual hepatic assay results were not suggestive of a pattern of significant hepatic damage in the Ranch Hands that might be supportive of an herbicide effect. Further, there was no consistent group-by-covariate interaction that suggests a detriment to a specific subcategory of the Ranch Hands. For those covariates used in both the Baseline study and this followup study, the direction and magnitude of their effects were relatively consistent between the studies. However, an unexpected association between wine drinking and alkaline phosphatase lacks a plausible explanation, particularly considering the inverse relationship, i.e., increasing alkaline phosphatase levels with decreasing wine consumption. These findings suggested the association between wine and alkaline phosphatase may be due to imprecision in data collection. None of the categorical (normal/abnormal categories) analyses was statistically significant, whereas all of the significant results were 13-27 �generated by the more powerful contrasts of continuously distributed hepatic data. Both porphyrin analyses showed group associations and are in distinct contrast to the otherwise largely negative hepatic findings. The significantly elevated uroporphyrin mean value in the Comparisons was directly opposite to that expected if dioxin-induced PCT were prevalent in the Ranch Hands. The primary biochemical defect in PCT is the reduced activity of uroporphyrinogen decarboxylase, an enzyme that metabolizes uroporphyrin. This defect leads to increased levels of uroporphyrin and coproporphyrin. Questionnaire-Laboratory Correlations; Porphyria Cutanea Tarda In the interval questionnaire all participants were asked whether their skin manifested "patches," excessive bruises, or sensitivity. These questions were deemed important in order to bound the maximum prevalence of cutaneous disorders compatible with a diagnosis of PCT. These historical data are given in Table 13-13. TABLE 13-13. Unadjusted Analysis for Interval History of Skin Bruises, Skin Patches, and Skin Sensitivity by Group Bruises, Patches, or Sensitivity Yes No Group Number Percent Ranch Hand 265 26.2 Comparison 260 20.2 Number Percent 746 1,029 Total p-Value 73.8 1,011 0.001 79.8 1,289 These data revealed that the Ranch Hands reported significantly more cutaneous symptoms (26.2%) than the Comparisons (20.2%). However, these data were substantially less than those reported at the Baseline in-home questionnaire, which also showed a statistically significant excess in the Ranch Hands. To determine if the skin histories might be related to PCT, the historic data were compared to the porphyrin test results. The abnormal/normal cutpoint of the coproporphyrin assays was reset to the 95th percentile because the normal range of the laboratory overclassified the proportion of abnormals. Table 13-14 gives the tabular display of both porphyrin test results by the reporting history of skin disorders. 13-28 �TABLE 13-14. Unadjusted Analyses for Porphyrin Abnormalities by Group and Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire Group Abnormal Porphyrin Findings for a Participant Skin Patch, Bruise, or 0 1 2 Sensitivity Reported Number Percent Number Percent Number Percent Total p-Value* Both Groups Yes No 472 1,593 90.1 90.2 48 165 9.2 9.3 4 9 0.8 0.5 524 0.789 1,767 Ranch Hand Yes No 239 670 90.5 90.3 24 70 9.1 9.4 1 2 0.4 0.3 264 0.950 742 Comparison Yes No 233 923 89.6 90.1 24 95 9.2 9.3 3 7 1.2 0.7 260 0.742 1,025 *Chi-square test, 2 d.f. The data from both groups combined suggest that there is no relationship between a history of cutaneous disorders and porphyrin test positivity. The group-specific data in the table also show a lack of a statistically significant association between the reporting of skin patches, bruises, or sensitivity and the presence of an abnormal porphyrin test result. However, in both the Ranch Hand and Comparison groups, participants who had abnormal tests for both uroporphyrins and coproporphyrins were more likely to have reported cutaneous disorders than participants with normal findings for both tests. Consequently, the data were retabulated, focusing only upon uroporphyrin abnormalities (absolutely required for a diagnosis of PCT) and reporting of cutaneous disorders. These data are summarized in Table 13-15. These data suggest that the relative risk of increased uroporphyrin abnormalities for Ranch Hands is independent of whether or not a study participant reported skin patches, bruises, or sensitivities at the followup questionnaire (Breslow-Day test of homogeneity of odds ratio, p=0.791). In each instance (reported/not reported), the estimated relative risk was nonsignificant and less than 1, and in both the Ranch Hand group and the Comparison group there was a higher percentage of uroporphyrin abnormalities for participants who did not report skin patches, bruises, or sensitivity than for participants who did report these conditions. Thus, the sequential displays of Tables 13-13 through 13-15 show excessive reporting of PCT-like cutaneous symptoms in the Ranch Hand group that was not related to test abnormalities for both uroporphyrin and coproporphyrin abnormal test results, or for uroporphyrin abnormalities alone. These analyses were consistent with the clinical observation that 13-29 �TABLE 13-15. Unadjusted Analyses for Uroporphyrin Abnormalities by Group and Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire Group Ranch Hand Comparison StratifiEst. Relative Variable cation Statistic Number Percent Number Percent Risk (95% C.I.) p-Value n Skin Patch, Abnormal Bruise, or Normal Sensitivity Reported 264 12 252 4.5 95 .5 260 12 248 4.6 95 .4 0.98 (0.43,2.23) 0.999 742 42 700 5.7 94.3 1,025 66 959 6.4 93.6 0.89 (0.62,1.28) 0.547 Uroporphyrin n Skin Patch, Abnormal Bruise, or Normal Sensitivity Not Reported only one differential diagnosis at the examination entertained the diagnosis of PCT. Based on all of these observations, PCT was a rare, if not nonexistent, condition in the Ranch Hands. EXPOSURE INDEX ANALYSES Both unadjusted and adjusted exposure index analyses were carried out for the nine laboratory tests of hepatic function and the two porphyrin metabolite tests. The porphyrin variables were analyzed only as continuous variables, while the others were analyzed both as continuous variables and discretized variables. Five covariates were included in the adjusted analyses: age, race, current alcohol use, exposure to degreasing chemicals (yes/no), and exposure to industrial chemicals (yes/no). Current alcohol use was treated as a continuous variable for all adjusted analyses, and age was treated as a continuous variable for the continuous adjusted analyses. Age was trichotomized (born in or after 1942, born between 1923 and 1941, and born in or before 1922) for the discrete adjusted analyses. In addition, the covariate BUN was used in the porphyrin analyses. For each variable, exposure level frequencies and percents are presented in Table K-3 of Appendix K along with the results of the unadjusted discrete 13-30 �analyses using Pearson's chi-square test to reflect overall exposure index differences and Fisher's exact test to investigate medium versus low and high versus low exposure level contrasts. Unadjusted means for each exposure level are presented in Table K-4 of Appendix K, along with the results of the unadjusted continuous analyses (using an F-test for an overall group assessment) and t-tests to examine medium versus low and high versus low exposure index contrasts. Results of the adjusted categorical and adjusted continuous analyses are presented in Tables 13-16 and 13-17, respectively. These results are presented in the context of a main effects model containing exposure index and all five covariates. Additional adjusted continuous analyses were conducted to examine pairwise interactions involving the exposure index and the covariates. Unadjusted and adjusted results for each variable are discussed in sequence. SGOT Within each occupation cohort, the low exposure level had the lowest percentage of abnormalities and the lowest mean. A marginally significant overall exposure level relationship (p=0.065) was found in the unadjusted discrete analysis for the enlisted groundcrew. This association was statistically significant in the adjusted analysis (p=0.023), exhibiting a doseresponse effect; medium versus low (Adj. RR: 2.14, 95% C.I.: [0.77,5.99], p=0,147) and high versus low (Adj. RRs 3.64, 95% C.I.: [1.36,9.72], p=0,010). A nonsignificant dose-response relationship was observed in the corresponding unadjusted and adjusted continuous analyses (p=0.418 and p=0.409, respectively), with unadjusted means of 32.9 U/L, 33.2 U/L, and 34.4 U/L for the low, medium, and high exposure levels, respectively. No significant results were found for enlisted flyers and officers. SGPT Within the enlisted groundcrew and enlisted flyer cohorts the low exposure level had the lowest percentage of abnormalities and the lowest mean value. This situation was reversed for the officers who exhibited the highest percentage of abnormal measurements and highest mean value in the low exposure categories. A significant overall result was found for enlisted flyers in the adjusted discrete analysis (p=0.036; medium versus low, Adj. RR: 6.55, 95% C.I.s [1.25,34.43], p=0.026); high versus low, Adj. RR: 4.29, 95% C.I.: [0.75,24.35], p=0.101). In the corresponding adjusted continuous analyses, a marginally significant dose-response relationship was observed (p=0.058) with adjusted means 18.1 U/L, 21.4 U/L, and 21.8 U/L for the low, medium, and high exposure levels, respectively. No significant results were found for officers or enlisted groundcrew. GGTP No significant or marginally significant results were found. A nonsignificant dose-response relationship was seen for enlisted flyers and officers in the continuous analyses but, conversely, a nonsignificant decreasing dose-response relationship was seen in the enlisted groundcrew. 13-31 �TABLE 13-16. Adjusted Categorical Exposure Index Analyses (Main Effects Model) Results for Hepatic Function Variables by Occupation Exposure Index Occupation Low Total Medium Total High Total Officer Variable 125 129 120 to 55 152 Officer i u> Enlisted Flyer Enlisted Groundcrew SCOT 125 Enlisted SGPT 65 160 129 55 57 140 120 65 Flyer Enlisted Groundcrew 152 160 140 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L 1.60 (0.64,3.98) 1.02 (0.38,2.77) 0.508 0.312 0.963 Overall M vs. L H vs. L 7.79 (0.77,79.20) 5.38 (0.49,59.50) 0.108 0.083 0.170 Overall M vs. L H vs. L 2.14 (0.77,5.99) 3.64 (1.36,9.72) 0.023 0.147 0.010 Overall M vs. L H vs. L 0.97 (0.48,1.97) 0.77 (0.37,1.64) 0.768 0.933 0.504 57 M vs. L H vs. L Overall 6.55 (1.25,34.43) 4.29 (0.75,24.35) 0.026 0.101 Overall M vs. L H vs. L 1.53 (0.77,3.01) 1.18 (0.57,2.48) 0.457 0.223 0.655 0.036 �TABLE 13-16. (continued) Adjusted Categorical Exposure Index Analyses (Main Effects Model) Results for Hepatic Function Variables by Occupation Exposure Index Occupation Low Total Medium Total High Total Officer Variable 125 129 120 Enlisted Flyer 55 Enlisted Groundcrew 152 Officer GGTP 126 65 57 Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Overall M vs. L H vs. L 1.02 ( . 8 2 7 ) 03,.2. 0.94 (0.35,2.54) 0.987 0.968 0.906 Overall M vs. L H vs. L 1.51 (0.41,5.65) 1.46 (0.37,5.78) 0.798 0.536 0.586 Overall M vs. L H vs. L 0.74 (0.34,1.64) 0.89 (0.40,1.97) 0.760 0.462 0.776 Overall M vs. L H vs. L 2.44 (0.65,9.05) 0.91 (0.19,4.36) 0.272 0.184 0.926 Overall M vs. L H vs. L 4.84 (0.52,44.80) 5.34 (0.58,49.06) 0.191 0.165 0.139 Overall M vs. L H vs. L 1.35 (0.50,3.59) 1.82 (0.72,4.59) 0.431 0.552 0.202 CO u> w Alkaline Phosphatase Enlisted Flyer 54 Enlisted Groundcrew 153 160 129 64 160 140 120 56 141 �TABLE 13-16. (continued) Adjusted Categorical Exposure Index Analyses (Main Effects Model) Results for Hepatic Function Variables by Occupation Exposure Index Total Bilirubin Occupation Low Total Medium Total High Total Officer Variable 125 129 120 65 57 152 160 140 Officer •Direct Bilirubin 54 Enlisted Groundcrew u> i Enlisted Flyer3 125 Enlisted Flyer 55 Enlisted Groundcrew 152 129 65 160 120 57 140 Contrast Adj . Relative Risk ( 5 C.I.) 9% p-Value Overall M vs. L H vs. L 0.67 (0.10,4.51) 1.10 (0.18,6.61) 0.851 0.677 0.915 — — — Overall M vs. L H vs. L 0.41 (0.10,1.65) 1.02 (0.32,3.23) 0.332 0.208 0.971 Overall M vs. L H vs. L 2.69 (0.46,15.82) 3.10 (0.56,17.25) 0.354 0.274 0.196 Overall M vs. L H vs. L 2.97 (0.48,18.38) 1.79 (0.24,13.43) 0.466 0.241 0.571 Overall M vs. L H vs. L 1.61 (0.43,6.06) 1.40 (0.36,5.51) 0.767 0.480 0.628 �TABLE 13-16. (continued) Adjusted Categorical Exposure Index Analyses (Main Effects Model) Results for Hepatic Function Variables by Occupation Exposure Index Occupation Low Total Medium Total High Total Officer Variable 125 129 120 55 152 Officer en Enlisted Flyer Enlisted Groundcrew Cholesterol 125 Enlisted Flyer 55 Enlisted Groundcrew Triglycerides 152 65 160 129 65 160 57 140 120 57 140 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall M vs. L H vs. L 0.54 (0.27,1-09) 0.50 (0.25,1.03) 0.107 0.085 0.060 Overall M vs. L H vs. L 1.02 (0.38,2.73) 1.12 ( . 2 3 0 ) 04,.2 0.972 0.962 0.822 Overall M vs. L H vs. L 1.20 (0.57,2.55) 1.61 (0.78,3.30) 0.417 0.630 0.194 Overall M vs. L . H vs. L 0.97 ( . 8 2 4 ) 03,.5 1.35 (0.55,3.32) 0.721 0.946 0.514 Overall M vs. L H vs. L 2.66 (0.62,11.39) 2.06 ( . 4 9 6 ) 04,.0 0.379 0.189 0.358 Overall M vs. L H vs. L 0.44 (0.14,1.42) 0.60 (0.19,1.86) 0.363 0.173 0.375 *No analysis done since there were only three abnormal (one medium, two high)% �TABLE 13-17. Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index SGPT Low Medium High Contrast p-Value n Adj . Mean 125 33.6 129 34.7 120 33.8 Overall M vs. L H vs. L 0.718 0.450 0.904 Enlisted Flyer n Adj . Mean 55 30.3 65 32.8 57 32.7 Overall M vs. L H vs. L 0.276 0.144 0.184 n Adj . Mean 152 33.5 160 34.1 140 35.0 Overall M vs. L H vs. L 0.409 0.595 0.183 Officer ^ 0 0 i OO Statistic Enlisted Groundcrew SCOT Occupation Officer Variable n Adj . Mean 125 20.1 129 20.0 120 19.1 Overall M vs. L H vs. L 0.695 0.969 0.451 Enlisted Flyer n Adj, Mean 55 18.1 65 21.4 57 21.8 Overall M vs. L H vs. L 0.058 0.047 0.030 Enlisted Groundcrew n Adj . Mean 152 20.2 160 21.4 140 21.0 Overall M vs. L H vs. L 0.581 0.309 0.492 �TABLE 13-17. (continued) Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Low Medium High Contrast p-Value n Ad j . Mean 125 30.9 129 32.2 120 32.4 Overall M vs. L H vs. L 0.828 0.611 0.580 Enlisted Flyer n Adj . Mean 55 36.6 65 42.6 57 44.6 Overall M vs. L H vs. L 0.286 0.230 0.132 n Adj . Mean 152 36.9 160 36.6 140 33.1 Overall M vs. L H vs. L 0.299 0.914 0.159 Officer ^ Statistic Enlisted Groundcrev GGTP Occupation Officer Variable n Adj . Mean 126 82.3 129 82.9 120 83.8 Overall M vs. L H vs. L 0.843 0.808 0.561 Enlisted Flyer n Adj . Mean 54 90.7 64 99.3 56 97.7 Overall M vs. L H vs. L 0.127 0.053 0.122 Enlisted Groundcrew n Adj . Mean 153 91.5 160 94.0 141 93.5 Overall M vs. L H vs. L 0.576 0.318 0.444 u> i u> Alkaline Phosphatase �TABLE 13-17. (continued) Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Statistic Low Medium High Contrast p-Value n Adj . Mean 125 0.77 129 0.75 120 0.79 Overall M vs. L H vs. L 0.439 0.504 0.545 Enlisted Flyer n Adj . Mean 55 0.69 65 0.76 57 0.79 Overall M vs. L H vs. L 0.070 0.128 0.023 Enlisted Groundcrew n Adj . Mean 152 0.73 160 0.74 140 0.78 Overall M vs. L H vs. L 0.240 0.838 0.117 Officer. Total Bilirubin Occupation Officer Variable n Adj . Mean 125 0.20 129 0.19 120 0.21 Overall M vs. L H vs. L 0.567 0.517 0.689 Enlisted Flyer n Adj . Mean 55 0.18 65 0.19 57 0.19 Overall M vs. L H vs. L 0.724 0.471 0.498 Enlisted Groundcrew n Adj . Mean 152 0.17 160 0.19 140 0.18 Overall M vs. L H vs. L 0.550 0.277 0.670 u> Ui 00 Direct Bilirubin �TABLE 13-17. (continued) Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Statistic Low Medium High Contrast p-Value n Ad j . Mean 125 134.0 129 131.3 120 128.9 Overall M vs. L H vs. L 0.232 0.373 008 .8 Enlisted Flyer n Adj . Mean 55 114.4 65 112.4 57 120.9 Overall M vs. L H vs. L 0.101 0.619 0.129 Enlisted Groundcrew n Adj . Mean 152 125.3 160 125.3 140 129.7 •Overall M vs. L H vs. L 0.092 0.997 0.055 Officer LDH Occupation Officer Variable n Adj . Mean 125 236.7 129 225.0 120 224.2 Overall M vs. L H vs. L 0.049 0.039 0.029 Enlisted Flyer n Adj . Mean 55 213.2 65 208.1 57 220.6 Overall M vs. L H vs. L 0.214 0.492 0.343 Enlisted Groundcrew n Adj . Mean 152 209.6 160 211.1 140 210.1 Overall M vs. L H vs. L 0.945 0.742 0.927 w u> Cholesterol �TABLE 13-17. (continued) Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Statistic Low Medium High Contrast p-Value n Ad j . Mean 125 110.0 129 108.5 120 116.3 Overall M vs. L H vs. L 0.739 0.886 0.558 Enlisted Flyer n Ad j . Mean 55 112.5 65 111.2 57 113.7 Overall M vs. L H vs. L 0.981 0.919 0.927 Enlisted Groundcrew n Ad j . Mean 152 110.9 160 109.9 140 107.9 Overall M vs. L H vs. L 0.922 0.890 0.690 Officer Triglycerides Occupation Officer Variable n Ad j . Mean 125 17.49 129 16.69 120 17.45 Overall M vs. L H vs. L 0.856 0.621 0.977 Enlisted Flyer n Ad j . Mean 54 18.58 65 16.96 57 18.27 Overall M vs. L H vs. L 0.703 0.438 0.890 Enlisted Groundcrew n Adj . Mean 151 16.39 160 16.54 139 15.45 Overall M vs. L H vs. L 0.644 0.903 0.451 w i .• £» --- Uroporphyrin �TABLE 13-17. (continued) Adjusted Continuous Exposure Index Analyses (Main Effects Model) for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Coproporphyrin Occupation Statistic Low Medium High Contrast p-Value Officer Variable n Ad j . Mean 125 127.65 129 128.84 120 130.26 Overall M vs. L H vs. L 0.901 0.833 0.649 Enlisted Flyer n Ad j . Mean 55 108.67 65 115.31 57 109.81 Overall M vs. L H vs. L 0.669 0.408 0.890 Enlisted Groundcrew n Ad j . Mean 151 115.28 160 115.71 140 122.88 Overall M vs. L H vs. L 0.325 0.935 0.177 �Alkaline Phosphatase For the enlisted groundcrew and the enlisted flyers, the lowest abnormal prevalence rate and lowest mean value were found in the low exposure category. A nonsignificant increasing dose-response relationship was seen within these occupations for the discrete analyses. In both unadjusted and adjusted continuous analyses, a marginally significant medium versus low contrast was found for enlisted flyers (p=0.086 and p=0.053, respectively), with unadjusted means of 88.9 U/L, 96.3 U/L, and 95.2 U/L for the low, medium, and high exposure levels, respectively. Total Bilirubin Discrete analyses revealed no significant findings; adjusted discrete analyses for enlisted flyers were not done due to sparse data. Continuous analyses revealed a significant overall effect (p=0.045, unadjusted) for enlisted flyers, which was marginally significant after adjustment (p=0.070). In both unadjusted and adjusted analyses, the high versus low mean contrast was significant (p=0.014 and p=0.023, respectively), with unadjusted means of 0.66 mg/dl, 0.73 mg/dl, and 0.76 mg/dl for the low, medium, and high exposure levels, respectively. Direct Bilirubin There were no significant exposure findings in either the continuous or discrete analyses, although within each occupational cohort, the lowest abnormal prevalence rate was found in the low exposure group. LDH The unadjusted discrete analyses revealed no significant or marginally significant results. No adjusted discrete analyses were done due to sparse data. The unadjusted continuous analyses for the enlisted groundcrew showed a significant overall relationship with the exposure index (p*0.031), with mean values of 123.1 U/L, 122.3 U/L, 127.9 U/L for the low, medium, and high exposure levels; the high versus low contrast was significant (p=0.037). After adjustment, the continuous analyses for enlisted groundcrew revealed marginally significant results (p=0.092, overall; p=0.055, high versus low). No significant or marginally significant results were seen for enlisted flyers or officers. Enlisted flyers and enlisted groundcrew had the largest mean values for their highest exposure category, which is reversed in the officers, who exhibited a nonsignificant decreasing dose-response relationship with exposure level. Cholesterol Significant or marginally significant results were found for officers in the direction of a decreasing dose-response relationship in both the adjusted continuous (overall p=0.049, medium versus low p=0.039, high versus low p=0.029) and adjusted discrete (medium versus low p=0.085, high versus low p=0.060) analyses. Neither of the enlisted cohorts demonstrated a similar decreasing response. 13-42 �Triglycerides No significant or marginally significant results were found. Uroporphyrins and Coproporphyrins No significant or marginally significant results were found. EXPOSURE INDEX ANALYSES Additional continuous analyses were done to examine pairwise interactions involving exposure level and the covariates. Ten exposure group-bycovariate interactions were found at p<0.05. All interactions were found in the enlisted flyer and enlisted groundcrew occupations. Eight of the interactions involved current alcohol consumption, one involved age, and one involved race. The interactions are summarized in Tables K-5 and K-6 of Appendix K. In Table K-5 of Appendix K, the slope of the continuous covariate with respect to the dependent variable is provided for each of the three exposure levels. Table K-6 of Appendix K presents the mean level of direct bilirubin for each of the three exposure levels by race. The interactions involving current alcohol consumption are mainly due to a nonsignificant dependent variable response to increasing alcohol consumption in the low exposure group in contrast to a significant positive response for the medium and high groups. The SCOT, SGPT, and GGTP interaction results for the enlisted groundcrew provide support for an interpretation of herbicide effect. In summary, the nine hepatic function variables and two porphyrin metabolite variables showed no conclusive evidence of a dose-response relationship at the followup examination. Five overall exposure group differences were found. Only two of these (SCOT for enlisted groundcrew, and total bilirubin for enlisted flyers) supported a dose-response relationship. LONGITUDINAL ANALYSES Three hepatic enzyme variables, SCOT, SGPT, and GGTP, were chosen for longitudinal analysis, spanning the spectrum of intermediate to acute effects. These test variables were chosen because both the Baseline and the followup assays were performed by the high-precision ACA 500® DuPont technology. The data from these three hepatic variables are arrayed in Table 13-18. The SGOT and SGPT data showed slight but uniform increases from the Baseline examination. These increases were proportionately the same for both the Ranch Hand and Comparison groups. These changes may reflect an aging effect or are due to laboratory variation. As indicated by the equality-ofdifference p-values, none of the three hepatic variables showed a statistically significant difference in the changes from Baseline to followup between groups. 13-43 �TABLE 13-18. Longitudinal Analyses for SCOT, SGPT, and GGTP: A Contrast of Baseline and First Follovup Examination Test Means Variable Means 1982 1985 Baseline Followup p-Value* (Equality of Difference) Group Total SCOT Ranch Hand Comparison 971 1,139 32.91 32.97 33.73 33.73 0.61 SGPT Ranch Hand Comparison 971 1,139 20.08 20.51 21.82 22.44 0.72 GGTP Ranch Hand Comparison 971 1,139 39.26 38.64 33.16 32.35 0.63 *Analyzed in log units. SUMMARY AND CONCLUSIONS The interval questionnaire revealed sparse reporting of liver disorders from 1982 to 1985 that was not significantly different between groups. Historical liver disease was verified by medical records, and these data were added to the verified Baseline history to assess possible lifetime differences. No significant differences were found. The medical record verification process showed that the historical data were generally correctly reported and classified between groups, except for the category of enlarged liver which showed a higher verification rate in the Comparison group. Digestive system mortality showed an overall nonsignificant excess in the Ranch Hands, but a relative nonsignificant excess of malignant neoplasms in the Comparisons. No differences were found for past or current peptic ulcer disease for the Ranch Hand and Comparison groups, adjusted for standard covariates as well as blood type. The physical examination disclosed a borderline significant increase of hepatomegaly in the Ranch Hand group. Emphasis was placed on nine laboratory test variables measuring liver function, i.e., serum glutamic-oxaloacetic transaminase (SCOT), serum glutamic-pyruvic transaminase (SGPT), gammaglutamyl transpeptidase (GGTP), alkaline phosphatase, total and direct bilirubin, lactic dehydrogenase (LDH), cholesterol, and triglycerides. In addition, uroporphyrin and coproporphyrin measurements were obtained to assess liver function and the likelihood of porphyria cutanea tarda (PCT). The nine hepatic variables were subjected to continuous and categorical statistical tests, and were adjusted for the covariates age, race, occupation, current alcohol consumption, and unprotected exposure to both industrial chemicals and degreasing chemicals. Final statistical models used only 13-44 �the significant covariates and two-way interactions for adjustment. The two porphyrin measurements were analyzed only in the continuous form. The overall summary results of the analyses of these 11 variables are given in Table 13-19. The results showed a significantly lower mean SGPT level, a greater mean alkaline phosphatase level, a lower mean uroporphyrin level for Ranch Hands as contrasted with Comparisons, and a marginally significant greater mean coproporphyrin level. Only in the instance of alkaline phosphatase did the discrete analysis approach statistical significance. No group differences were noted for SCOT, GGTP, total and direct bilirubin, LDH, cholesterol, or triglycerides. However, an analysis using only the Original Comparisons revealed a significantly greater mean cholesterol level in the Comparison group. A review of the covariate effects in the adjusted statistical models revealed that all covariates behaved as expected with the exception of alcohol consumption for the alkaline phosphatase analysis, which showed an inverse relationship with wine consumption. Exploration of group-by-group covariate interactions for alkaline phosphatase, direct bilirubin, triglycerides, SCOT, and uroporphyrins revealed significant group differences within specific covariate strata. In particular, Ranch Hands exposed to industrial chemicals had a significantly higher adjusted mean level of alkaline phosphatase and a significantly higher abnormal prevalence rate of direct bilirubin than similarly exposed Comparisons. For triglycerides, Ranch Hands born in or before 1922 had a significantly higher adjusted mean level than similar aged Comparisons, while Ranch Hand officers exhibited a significantly higher abnormal prevalence rate than Comparison officers. For SCOT, Ranch Hand moderate current drinkers (more than one to four drinks per day) had a significantly higher mean level than corresponding Comparisons. In the opposite direction, Comparisons with a mean BUN level less than or equal to 14 (median for all participants) were found to have a significantly higher adjusted mean uroporphyrin level than similar Ranch Hands. These results did not disclose any common pattern detrimental to the Ranch Hand group. These findings were generally consistent with the 1982 Baseline data, which disclosed a significantly increased mean cholesterol level in the Comparisons and nonsignificant Ranch Hand mean elevations for GGTP and LDH. Slight differences in analytic results are probably due to the use of more fully adjusted models used for the followup examination data. Overall, the followup examination laboratory data showed no adverse clinical or exposure patterns in either group. Further, the detection of significant mean shifts (still within normal range) by the continuous statistical tests, not mirrored by the categorical tests, suggests a circumstance of statistical power rather than findings of biological relevance. Of the five significant or marginally significant results that were found in the adjusted exposure index analyses, four exhibited a trend suggestive of an increasing dose-response relationship. In the enlisted flyer cohort, the percentages of SGPT abormalities were significantly different and increased from the low to the high exposure categories. The corresponding mean values were marginally significantly different among exposure levels. Also,'the mean levels of total bilirubin were marginally significantly different among exposure levels, increasing with exposure level. For enlisted groundcrew, the percentage of SCOT abnormalities significantly differed among 13-45 �TABLE 13-19. Overall Summary Results of Unadjusted and Adjusted Analyses of Nine Hepatic Function Variables and Two Porphyrin Metabolite Tests Unadjusted Mean Adjusted* Categorical Variable Mean CC Questionnaire Liver Disease (Lifetime History) Hepatitis Jaundice Cirrhosis Enlarged Liver Miscellaneous Liver Disorders Peptic Ulcer Disease NS Physical Examination Hepatomegaly Categorical CD DD Direction of Results** NS* NS NS NS NS NS NSa RH>C Laboratory Testing .SCOT SGPT GGTP Alkaline Phosphate Total Bilirubin Direct Bilirubin LDH Cholesterol Triglycerides Uroporphyrin Coproporphyrin NS NS* NS 0.009 NS NS NS NS NS 0.048 NS* NS NS NS NS NS NS NS NS NS NS 0.048 NS 0.008 NS NS **** NS **** **** NS* **** 0.029 NS **** NS NS NS NS NS NS NS NS NS* NS **** NS NS **** Questionnaire-Laboratory Correlation Skin Bruises, Patches,— 0.001 and Sensitivity *C: Continuous D: Discrete **RH>Cj more abnormalities, or higher mean value, in Ranch Hands, ORH: more abnormalities, or.higher mean value, in Comparisons. a Adjusted for blood type. NS: Not significant (p>0.10). NS*: Borderline significant (0.05<p<0.10). —Analysis not performed. ****Group-by-covariate interaction. 13-46 ORH RH>C ORH RH>C RH>C �exposure levels. Within the enlisted flyer cohort, all nine laboratory tests of hepatic function had the lowest percentage of abnormalities in the low exposure category; correspondingly, six of the nine mean levels were lowest for the low exposure category. Of the ten group-by-covariate interactions that were found, three (SCOT, SGPT, and GGTP) supported a dose-response relationship in the enlisted groundcrew cohort. Exploration of these interactions revealed a trend that showed an increasing association between current alcohol consumption and the dependent variables for increasing exposure levels. Longitudinal analyses for SCOT, SGPT, and GGTP disclosed no statistically significant group differences in the mean shifts from the Baseline to the followup examination. Interval reporting of PCT-like symptoms of skin patches, bruises, and sensitivity was significantly increased in the Ranch Hands (p=0.001). However, when these historic data were contrasted to both uroporphyrin and coproporphyrin abnormalities, no correlation was apparent, nor were there any significant group differences. Since an elevation in the uroporphyrin level is required for a diagnosis of PCT, the historic data were retabulated with only uroporphyrin abnormalities; again, no group differences were apparent, .and, in fact, uroporphyrin abnormalities in both groups were higher in those participants without a history of skin disorders than in those participants with such a history. The likelihood of bona fide PCT among study participants, and particularly among the Ranch Hands, appears to be remote. In conclusion, the followup examination disclosed more statistically significant findings for tests of liver function than the Baseline examination, but they were equally divided between the two groups and did not demonstrate clinical, statistical, or exposure patterns consistent with an herbicide-related effect on health. No evidence was found to suggest an increased likelihood of PCT among the Ranch Hand group. 13-47 �CHAPTER 13 REFERENCES 1. Kimbrough, R.D., C.D. Carter, J.A. Liddle, R.E. Cline, and P.E. Phillips. 1977. Epidemiology and pathology of a tetrachlorodibenzodioxin poisoning episode. Arch. Environ. Health 32(2):7-86. 2. McNulty, W.P. 1977. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin for Rhesus monkeys: Brief report. Bull. Environ. Contam. Toxicol. 18(1):108-109. 3. Olson, J.R., M.A. Holscher, and R.A. Neal. 1980. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in the golden Syrian hamster. Toxicol. Appl. Pharmacol. 55:67-78. 4. Palmer, J.S., and R.D. Radeleff. 1964. The toxicologic effects of ' certain fungicides and herbicides on sheep and cattle. Ann. N.Y. Acad. Sci. 11:729-736. 5. Goldstein, J.A., P. Hickman, H. Bergman, and J.G. Vos. 1973. Hepatic porphyria induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin in the mouse. Res. Commun. Chem. Pathoi. Pharmacol. 6:919. 6. Madhukar, B.V., and F. Matsumura. 1981, Difference in the nature of induction of mixed-function oxidase systems of the rat liver among phenobarbital, DDT, 3-methylcholanthrene, and TCDD. Toxicol. Appl. Pharmacol. 61:110-118. 7. Kohli, K.K., and J.A. Goldstein. 1981. Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on hepatic and renal prostaglandin synthetase. Life Sci. 19:299-305. 8. Thunberg, T., and H. Hakansson. 1983. Vitamin A (retinol) status in the Gunn rat: The effect of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Arch. Toxicol. 53:225-234. 9. Goldstein, J.A., P. Hickman, and D.L. Jue. 1974. Experimental hepatic porphyria induced by polychlorinated biphenyls. Toxicol. Appl. Pharmacol. 27:437. 10. Sassa, S., H. De Verneuil, and A. Kappas. 1984. Inhibition of uroporphyrinogen decarboxylase activity in polyhalogenated aromatic hydrocarbon poisoning. In Banbury report 18: Biological mechanisms of dioxin action, ed. A.. Poland and R.D. Kimbrough, pp. 215-222. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 13-48 �11. Sweeney, G., D. Basford, B. Rowley, and G. Goddard. 1984. Mechanisms underlying the hepatotoxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin. In Banbury report 18: Biological mechanisms of dioxin action, ed. A. Poland and R.D. Kimbrough, pp. 255-239. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 12. Greig, J. 1984. Differences between skin and liver toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in mice. In Banbury report 18: Biological mechanisms of dioxin action, ed. A. Poland and R.D. Kimbrough, pp. 391-397. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 13. Goldmann, P.J. 1973. Schweist akute Chlorakne, eine Massenintoxikation durch 2,3,7,8-Tetrachlorodibenzodioxin (Severe, acute chloracne, a mass intoxication due to 2,3,7,8-tetrachlorodibenzo-dioxin). Per Hautarzt. 24(4):149-152. 14. Oliver, R.M. 1975. Toxic effects of 2,3,7,8-tetrachlorodibenzo 1,4-dioxin in laboratory workers. Br. J. Ind. Med. 32:49-53. 15. Reggiani, G. 1980. Acute human exposure to TCDD in Seveso, Italy. J^ Toxicol. Environ. Health 6:27-43. 16. Reggiani, G. 1979. Estimation of the TCDD toxic potential in the light of the Seveso accident. Arch. Toxicol. 2:291-302. 17. Suskind, R.R. 1978. Chloracne and associated health problems in the manufacture of 2,4,5-T. Report to the Joint Conference, National Institute of Environmental Health Sciences and International Agency for Research on Cancer, World Health Organization, Lyon, France, January 11, 1978. 7 pp. 18. May, G. 1982. Tetrachlorodibenzodioxin: A survey of subjects ten years after exposure. Br. J. Ind. Med. 39:128-135. 19. Ideo, G., G. Bellati, A. Bellobuono, A. Mocarelli, P. Marocchi, A. and P. Brambilla. 1982. Increased urinary d-glucaric acid excretion by children living in an area polluted with tetrachlorodibenzodioxin (TCDD). Clin. Chem. Acta. 120:273-283. 20. May, G. 1973. Chlorance from the accidental production of tetrachlorodibenzodioxin. Br. J. Ind. Med. 30:276-283. 21. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5trichloro-phenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 22. Suskind, R.R., and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 13-49 �23. Pazderova-Vejlupkova, J., M. Nemcova, J. Pickova, L. Jirasek, and E. Lukas. 1981. The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men. Arch. Environ. Health 36:5-11. 24. Martin, J.V. 1984. Lipid abnormalities in workers exposed to dioxin. Br. J. Ind. Med. 41:254-256. 25. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, tf.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. JAMA 255:2031-2038. 26. Oliver, R.M. 1975. Toxic effects of 2,3,7,8-tetrachrloro-dibenzo-l, 4-dioxin in laboratory workers. Br. J. Ind. Med. 32:46-53. 27. Bleiberg, J., M. Wallen, R. Brodkin, and I.L. Applebaum. 1964. Industrially acquired porphyria. Arch. Dermatol. 89:793-797. 28. Jirasek, L., J. Kalensky, K. Kubec, et al. 1974. Acne chlorina, porphyria cutanea tarda and other manifestations of general intoxication during the manufacture of herbicides, part 2. Czech Dermatol. 49(3):145-157. 29. Poland, A.P., D. Smith, G. Metter, and P. Fossick. 1971. A health survey of workers in a 2,4-D and 2,4,5-T plant, with special attention to chloracne, porphyria cutanea tarda, and psychologic parameters. Arch. Environ. Health 22(3):316-327. 30. Peters, H.A., A. Gocmen, D.J. Cripps, G.T. Bryan, and I. Dogramaci. 1982. Epidemiology of hexachlorobenzene-induced porphyria in Turkey. Arch. Neurol. 39:744-749. 31. Rubenstein, E., and D.D. Federman, eds. 1986. Metabolism: The porphyrias. Chap. 9 in Scientific American Medicine. New York: Scientific American, Inc. 13-50 �CHAPTER 14 DERMATOLOGICAL EVALUATION INTRODUCTION The skin is a major target organ following heavy exposure to chlorophenols and dioxin and, therefore, is a primary focus of the AFHS clinical examination. Since the association between chlorinated chemicals and chloracne was first noted in 1957,1'2 a variety of animal experiments have shown the dermal sensitivity of rabbits, monkeys, and hairless mice to TCDD, 2,4,5-T (contaminated with TCDD), and other chlorinated dibenzo compounds, furans, or their brominated analogs.1"7 Chloracne is not associated with exposure to 2,4-D. Accidental exposure to waste oils containing TCDD has caused significant dermal symptoms, including loss of hair, ulceratiye dermatitis, and inflamed mucous membranes in horses, dogs, cats, and mice. ' Studies have suggested that the chloracnegens induce a series of pathological skin changes in target cells of the epithelial lining of sebaceous glands via the Ah receptor. Hyperkeratinization of these cells eventually leads to the formation of the comedone characteristic of aerie. In humans, development of the hallmark rash, chloracne, is generally acknowledged to represent substantial topical or systemic exposure to one or more chloracnegens. ' ' ' Acute fulminant chloracne is characterized by a maculopapular rash of active comedones, conforming to an eyeglass or facial butterfly(1distribution, often accompanied by chest, back, or eyelid lesions. The severity of the chloracne appears to be generally dose related, but may also depend on the route of administration, age, genetic predisposition, and/or the existence of acne vulgaris or other skin disorders. ' ' Occasionally, exposure, via contaminated clothes of an9 industrial worker, has been associated with chloracne in family members. Sequelae from severe chloracne include actinic elastosis, afpe scars, disfigurement, excessive hair growth, and Peyronie's disease. ' Severe chloracne is often accompanied by acute effects in other organ systems. In contrast, low to moderate exposure to chloracnegens generally produces mild chloracne with few, if any, attendant systemic signs and symptoms. The clinical diagnosis of acute chloracne is easier than the diagnosis of subacute and chronic chloracne. In the latter instances, a history of exposure to chloracnegens is essential in the diagnosis, particularly if the individual has experienced adolescent acne. Chronic chloracne has been clinically observed more than 30 years after onset, but a biopsy is often necessary to confirm these cases. Mild or transient cases of chloracne may be confused with persistent adolescent acne or other skin conditions. 14-1 �As noted in the AFHS Baseline Morbidity Report, over one-half of the veteran complaints in the Veterans Administration Herbicide Registry involved dermatological conditions, a fact sometimes alluded to as "evidence" of exposure to Agent Orange. In actuality, skin disease was a major medical problem among American troops serving in Vietnam. Forty-seven percent of the combat-days lost in the 9th Infantry Division from July 1968 to June 1969 were due to dermatological conditions. These diseases were directly related to the tropical climate and terrain. Only in rare cases has the Veterans Administration made a diagnosis of chloracne in a Vietnam combat veteran. The natural history of chloracne suggests that most cases should have been diagnosed while in Vietnam, but a dermatological survey failed to reveal any cases. Most recognized chloracne cases have been diagnosed in chemical plant workers or in victims of industrial accidents. Thousands of cases were recorded in the 1930-1940 era, and earlier descriptions of chloracne-like disease were found in 1897 to 1901. Industrial exposure to chloracnegens has been generally characterized as moderate-prolonged or severe-acute. In the setting of casual-sporadic exposure, as in the typical cases of the contaminated housing areas in Times Beach, Missouri, and the Quail Run Trailer Park, chloracne is virtually unknown. A number of dioxin morbidity studies have shown a clustering of abnormal laboratory tests in individuals with chloracne. ' ,24-27 Thig hag some investigators to believe that long-term sequelae to dioxin exposure will be found only in people with chloracne. Other investigators feel that this belief is not consistent with normal spectrum-of-illness concepts and that effects may occur in the absence of chloracne. Baseline Summary Results The 1982 Baseline clinical examination revealed an unexpected significant excess (p=0.03) of basal cell carcinoma in the Ranch Hand group. Risk factor data (e.g., sun exposure, host factors of tannability, complexion) were not collected in 1982. The 1982 examination focused on the diagnosis of chloracne both in historical terms by a detailed questionnaire and in contemporary terms via a comprehensive clinical assessment. The questionnaire data did not demonstrate anatomic, incidence, or onset-time patterns of acne in the Ranch Hand group that might support an inference of past chloracne, nor did the physical examination detect a single case. Fourteen biopsies from 11 participants also failed to document a chloracne diagnosis. A dermatology index (the number of clinically detected skin abnormalities per individual) was virtually identical between the Ranch Hand and Comparison groups, and was associated with the history of past acne in both groups. No exposure level associations were noted in any occupational category of the Ranch Hand group. The comprehensive dermatological assessment did not reveal evidence of past or current chloracne in the Ranch Hand group. Parameters of the 1985 Dermatological Evaluation Questionnaire data recaptured many of the acne parameters of the 1982 questionnaire, and the physical examination parameters were similar to the 14-2 �1982 Baseline examination. Particular emphasis was given to the diagnosis of basal cell carcinoma and to the collection of risk factor data, e.g., skin color, reaction to sun, ethnicity (see Chapter 10, Malignancy). Thus, the dependent variables and covariates of the analyses below closely approximated those previously conducted on the Baseline examination and questionnaire data. The adjusted statistical analyses were based on logistic regression (BMDP®-LR) and log-linear models (BMDP®-4F), and the unadjusted analyses primarily use Pearson's chi-square test and Fisher's exact test. In addition, an empiric Venn diagram was used to explore the potential of historic chloracne. Parallel analyses using only Original Comparisons are presented in Tables L-3 through L-ll of Appendix L. RESULTS AND DISCUSSION General Detailed dermatological data were obtained by standard physical examination techniques. Numeric differences in summary tables reflect missing dependent variable and undeterminable covariate information. One participant refused the dermatology examination; consequently, all skin disorder analyses were based on 2,308 participants. Data were collected on 22 skin disorders, which were in turn reduced to eight variables for analysis: comedones, acneiform lesions, acneiform scars, depigmentation, inclusion cysts, hyperpigmentation, other abnormalities, and the dermatology index. Descriptions of skin biopsies, which were also conducted at the physical examination, are given in this chapter. Followup questionnaire information regarding the presence, time, and location of acne was also analyzed. The analyses in this chapter first investigate questionnaire information on acne, and subsequent analyses center upon the eight skin disorder variables and the skin biopsies. Four covariates were included in this analysis: age, race, occupation, and presence of acne before duty in Southeast Asia. Age is used in its continuous form for all adjusted logistic regression analyses, but age is trichotomized (born in 1942 or after, born between 1923 and 1941, and born in 1922 or before) for presentation in summary tables and for use in dependent variable and covariate association analyses and log-linear models. Participants were categorized as either Black or nonblack. Occupation was divided into the three classifications of officer, enlisted flyer, and enlisted groundcrew. Sample size differences in subsequent adjusted analyses reflect missing dependent variable data or missing data on the presence of acne before duty in Southeast Asia. Questionnaire Data The acne status of each participant was determined by Baseline and followup questionnaire information. In particular, the occurrence of acne and the dates for acne occurrence have been determined and analyzed. Additionally, the analysis of the location of acne is presented for a subset of the participants who have had acne. Figure 14-1 below is a diagram explaining the occurrence of acne by time determination for the 2,309 participants, along with frequencies and an explanation of terms. 14-3 �Determination Presence of Acne All Acne in 1961 or Before (for Participants with Acne) Acne Reference to Beginning of First SEA Tour of Duty (for Participants with Acne Sometime after 1961) (138) (205) Yes to Acne — Reported acne on both/either Baseline and/or follow/up study. No to Acne — Never had acne. Pre-1961 Acne — Participants with acne who had last occurrence of acne in or before 1961. Post-1961 Acne — Participants with acne who had an occurrence of acne sometime after 1961. Undetermined — Time reference not determinable from date information available. Pre-SEA Acne — Participants with post-1961 acne who had all occurrences of acne before the start of first Southeast Asia (SEA) tour (as determined from military records). Post-SEA Acne — Participants with post-1961 acne who had all occurrences of acne after the start of first SEA tour. Pre- and Post-SEA Acne — Participants with post-1961 acne who had multiple occurrences, both before and after the start of first SEA tour, or a case of acne that began before the start of first SEA tour and-that ended after starting SEA tour. *: Analysis of location of acne performed for these participants. Figure 14-1. Occurrence of Acne by Time for First Followup Participants 14-4 �The distinction was made between pre-1961 and post-1961, since herbicide missions in Vietnam commenced in 1962. Responses of 2,309 participants indicated that 1,415 individuals never had acne, 379 had acne before 1961, 138 had acne after 1961 but before duty in SEA, 205 had acne both before and after duty in SEA, 146 had acne only after SEA duty, and 26 participants could not be specifically classified. Occurrence of Acne The reported occurrence of acne, as determined by Baseline and followup questionnaires, is displayed in Table 14-1. The analysis showed that the Ranch Hand group reported slightly more acne than the Comparison group, although the difference is nonsignificant (p=0.111). Analyses using Original Comparisons only showed a borderline significance (p=0.071) found in Table L-3 of Appendix L. The participants who responded "yes" to acne were categorized according to whether their acne occurred before or after 1961. The distribution of pre-1961 versus post-1961 acne is given in Table 14-2. TABLE 14-1. Unadjusted Analysis for Reported Historical Occurrence of Acne by Group Acne Yes Group Number Ranch Hand Comparison 412 482 Total 894 No Percent Number 40.6 37.3 604 811 1,415 14-5 Percent Total 59.4 62.7 1,016 1,293 2,309 Summary Statistics Est. RR: 1.15 952 C.I.: (0.97,1.36) p-Value: 0.111 �TABLE 14-2. Unadjusted Analysis for Reported Historical Occurrence of Acne Relative to 1961 by Group* Occurrence of Acne Post-1961 Group Number Ranch Hand Comparison 239 271 Total 510 Pre-1961 Percent Number 58.3 56.6 171 208 379 Percent Total 41.7 43.4 410 479 Summary Statistics Est. RR: (for post1961 cases): 1.07 95% C.I,: (0.82, 1.04) p-Value: 0.634 889 *Five participants deleted due to missing data at time of occurrence. As shown, no significant difference in the distribution of post-1961 versus pre-1961 acne existed between Ranch Hands and Comparisons (p=0.634). Cases of post-1961 acne were classified to SEA tour(s) of duty, as determined by military records. The distribution of post-1961 acne cases relative to SEA is shown in Table 14-3. This marginal significance (p=0.058) was due primarily to a larger percentage of Ranch Hands in the post-SEA category, as contrasted with the Comparisons (35.1% versus 25.3%). Duration of Acne The approximate duration of acne was examined among the three SEA categories by group using a two-factor analysis of variance. The calculation of acne duration for participants with multiple occurrences in overlapping time periods counted time periods only once. A square root transformation was used to normalize the duration data. Results from duration of acne analyses are given in Table 14-4. 14-6 �TABLE 14-3. Unadjusted Analysis for Reported Historical Occurrence of Acne Relative to SEA Tour of Duty for Post-1961 Acne by Group* Post-1961 Acne Pre-SEA Group Ranch Hand Comparison Total Post-SEA Pre- and Post-SEA Number Percent Number Percent Number 58 80 138 25.4 30.7 80 66 146 35.1 25.3 Percent Total 90 115 205 39.5 44.1 228 261 489 p-Value 0.058 *Twenty-one post-1961 participants with acne deleted due to missing data on time of occurrence. TABLE 14-4. Adjusted Analysis for Duration of Acne (in Years) for Post-1961 Acne by Group* Group Ranch Hand Comparison Total Total 219 252 471 Adjusted Mean** 95% C.I.** p-Value 8.18 7.49 (7.43,8.96) (6.82,8.19) 0.189 Covariate Remarks Time Reference to SEA (p<0.001) *Eighteen participants deleted due to missing data on time of occurrence. **Converted from square root scale. This adjusted analysis showed no significant effect due to group (p=0.189), but a highly significant effect due to SEA category (p<0.001), with the pre- and post-SEA category having higher mean durations than the pre-SEA or post-SEA categories, which were nearly identical. No interaction was present between group and SEA category (p=0.314). A categorical analysis was performed, in which duration was categorized into 5-year increments (five duration categories, the last being greater than 20 years). There was no significant difference between groups (pre-SEA, p=0.520; post-SEA, p=0.776; pre- and post-SEA', p=0.880). Location of Acne The location of acne for participants classified as post-SEA or pre- and post-SEA (351 participants) was analyzed. Spatial distribution of acne with 14-7 �primary emphasis on acne on the temples, around the eyes, or on the.ears was determined from the questionnaire; these data are presented in Figures 14-2 and 14-3. Figure 14-2 shows the distribution of acne for Ranch Hands and Comparisons, for post-SEA and pre- and post-SEA participants combined, whereas Figure 14-3 represents a similar distribution for only post-SEA participants. If more than one episode of acne occurred, cases involving the temples, eyes, or ears took precedence. Also, multiple-site involvement took precedence over single-site involvement. The Ranch Hand and Comparison Venn diagrams were contrasted by chisquare analysis of a 2x8 table, and no difference in the spatial distribution was noted for the combination of pre- and post-SEA and post-SEA groups (p=0.706), or for the analysis of only the post-SEA group (p=0.699). Sparse data cells were present in the analysis of both figures. Differences in spatial distributions were also not evident when the "other sites" classification was deleted (p=0.770 and p=0.664, respectively). If the intersection of the circles in these figures (i.e., temples, ears, and eyes) is contrasted with the rest of the locations of acne, no significant difference is seen (p=0.189 and p=0.627 for the combination of post-SEA and pre- and post-SEA groups and for only the post-SEA group, respectively). Physical Examination Data Twenty-two skin disorders were assessed at the dermatological examination (page C-9 of Appendix C). These disorders were combined into eight variables for analytic purposes. Comedones, acneiform lesions, acneiform scars, depigmentation, inclusion cysts, and hyperpigmentation were analyzed separately. The remaining 16 conditions were grouped to form a broad variable called "other abnormalities." Analysis of skin cancer is included in the malignancy chapter and will not be discussed here. Additionally, comedones, acneiform lesions, acneiform scars, and inclusion cysts were grouped to construct a dermatology index, which summed the number of abnormalities for these four conditions for each participant. Logistic regression techniques, with the use of BMDP®-LR, were utilized for adjusted analysis of all these variables except the dermatology index, which used BMDP®-4F. The sample sizes were sufficient to detect a 27-percent increase in the prevalence rate for comedones, a 30 percent increase in the prevalence rate for acneiform scars, and a 12 percent increase in the prevalence of at least one abnormality for the dermatology index, using a two-sided a -level of 0.05 with a power of 0.80. No cases of chloracne were chemically diagnosed. Preliminary Dependent Variables and Covariate Relationships The association of the eight skin disorder variables in both groups and the covariates of age (born in or after 1942, born between 1923 and 1941, born in or before 1922), race (Black or nonblack), occupation, and presence of preSEA acne (yes/no) was assessed using Pearson's Chi-square test and Fisher's exact .test. Table 14-5 is a summary of the associations of the dependent variables with these four covariates. Seven additional participants, who were initially classified as "undetermined," were reclassified as having acne before duty in SEA, based on data gathered by telephone. Nineteen participants were omitted from analyses involving presence of pre-SEA acne, because historical information on the date of onset of acne was not available. 14-8 �Ranch Hand 103 Other Sites n=169 Comparison 121 Other Sites n=181 Figure 14-2. Location of Post-SEA and Preand Post-SEA Acne by Group 14-9 �Ranch Hand 53 Other Sites n=80 Comparison 48 Other Sites n=66 Figure 14-3. Location of Post-SEA and Acne by Group 14-10 �TABLE 14-5. Association Between Dermatological Variables and Age, Race, Occupation, and Fre-SEA Acne in the Combined Ranch Hand and Comparison Groups Variable Comedones Acneiform Lesions Acneiform Scars Depigmentation Inclusion Cysts Hyperpigmentation Other Abnormalities Dermatology Index Age <0.001 <0.001 <0.001 NS NS NS <0.001 NS Race <0.001 NS* <0.001 0.009 NS <0.001 <0.001 NS Occupation <0.001 NS* <0.001 NS 0.036 <0.001 <0.001 0.010 Pre-SEA Acne NS <0.001 <0.001 NS NS 0.003 NS* <0.001 NS: Not significant (p>0.10) NS*; Borderline significant (0.05<p<0.10) effect with variable. Age had a significant effect on four of the variables. Prevalence rates for comedones and other abnormalities were highest for older participants (born in or before 1922). On the other hand, the prevalence of acneiform lesions and acneiform scars was higher in the younger participants (born in or after 1942). Nonblacks had a significantly higher prevalence of comedones and other abnormalities and a marginally significant increase (p=0.055) in acneiform lesions. Blacks had a significantly higher prevalence rate for acneiform scars, depigmentation, and hyperpigmentation. Occupation had a significant or marginally significant effect on seven of the eight variables, with either enlisted flyers or enlisted groundcrew generally having a higher percentage of abnormalities. Participants with pre-SEA acne had a significantly higher prevalence rate for acneiform lesions and acneiform scars, and a higher percentage with at least one abnormality in the dermatology index. Participants without acne pre-SEA had a significantly higher prevalence rate for hyperpigmentation, and a marginally significantly higher prevalence rate (p=0.084) for other abnormalities. Analyses of Individual Dependent Variables Comedones As reflected in Table 14-6, there was not a significant difference (p=0.361) between the proportion of participants with comedones in the Ranch Hand and Comparison groups, unadjusted for any covariates. 14-11 �TABLE 14-6. Unadjusted Analysis for Comedones by Group Comedones Absent Present Group Number Ranch Hand Comparison Percent Number 250 340 24.6 26.3 766 952 Percent 75.4 73.7 Total 1,016 1,292 Summary Statistics Est. RR: 0.91 95% C.I.: (0.76,1.10) p-Value: 0.361 Tests of association between the presence of comedones in both groups and the four covariates indicated that there was not a significant effect due to the presence of pre-SEA acne (p=0.355), but that there were significant effects due to occupation (p<0.001), age (p<0.001), and race (p<0.001). The proportion of participants with comedones increased with age (18.9% for participants born in or after 1942, 29.8% for participants born between 1923 and 1941, and 37.9% for participants born in or before 1922). Significantly more nonblacks had comedones than Blacks (26.5% versus 11.9%), and enlisted flyers had more than enlisted groundcrew or officers (34.4%, 24.8%, and 22.6%, respectively). An adjusted analysis of the proportion of participants with comedones was performed using logistic regression techniques. Results are presented in Table 14-7. TABLE 14-7. Adjusted Analysis for Comedones by Group Ranch Hand Total 1,007 Comparison Total Adjusted Relative Risk (95% C.I.) p-Value 1,282 0.89 (0.74,1.09) 0.260 14-12 Covariate Remarks Occupation (p<0.001) Presence of Pre-SEA Acne (p=0.038) Race-by-Age (p=0.046) �Again, no significant differences were found between groups (p=0.260). Occupation, pre-SEA acne, and a race-by-age interaction were significant (p<0.001, p=0.038, and p=0.046, respectively). Compared to Baseline findings, the percentage of participants with comedones increased in the Comparison group but decreased in the Ranch Hand group. Estimated and adjusted relative risks were both less than 1.0 in the followup study, while the estimated relative risk in the Baseline study was slightly greater than 1.0 (RR=1.05, with Original Comparisons used), but statistically nonsignificant. Acneiform Lesions As shown in Table 14-8, there was not a significant difference between the proportion of participants with acneiform lesions in the Ranch Hand and Comparison groups, unadjusted for any covariates (p=0.624). TABLE 14-8. Unadjusted Analysis for Acneiform Lesions by Group Acneiform Lesions Present Group Ranch Hand Comparison Number 188 228 Absent Percent Number 18.5 17.6 828 1,064 Percent 81.5 82.4 Total 1,016 1,292 Summary Statistics Est. RR: 1.06 95* C.I.i (0.86,1.31) p-Value: 0.624 Test.? of association between the presence of acneiform lesions in both groups and the four covariates revealed marginally significant effects due to race (p=0.055) and occupation (p=0.064), and significant effects for age (p<0.001) and presence of pre-SEA acne (p<0.001). Nonblacks had a marginally significantly higher proportion of participants with acneiform lesions than Blacks (18.4% versus 11.9%). The proportion of participants with lesions was greatest for enlisted groundcrew (20.1%), as compared to the other occupations (officers, 16.4%; enlisted flyers, 16.0%). The proportion of participants with acneiform lesions decreased with age (born in or after 1942, 23.0%; born between 1923 and 1941, 14.8%; born in or before 1922, 10.3%). A significantly higher proportion of participants with acne present before SEA had lesions (22.4%), as compared.with those not having acne before SEA (16.0%). An adjusted analysis of the proportion of participants with acneiform lesions was performed using logistic regression techniques. Results of this analysis are summarized in Table 14-9. 14-13 �TABLE 14-9. Adjusted Analysis for Acneiform Lesions by Group Ranch Hand Total 1,007 Comparison Total 1,282 Adjusted Relative Risk (95% C.I.) Covariate Remarks p-Value 1.08 (0.87,1.34) 0.512 Age (p<0.001) Race (p=0.014) Presence of Pre-SEA Acne (p=0.008) The results showed no significant differences between groups (p=0.512). Age (p<0.001), race (p=0.014), and presence of pre-SEA acne (p=0.008) were significant adjusting variables in this analysis. The Baseline and followup results for acneiform lesions were nearly identical with respect to group differences. Acneiform Scars Table 14-10 shows no significant difference between the proportion of participants with acneiform scars in the Ranch Hand and Comparison groups, unadjusted for any covariates (p«0.720). TABLE 14-10. Unadjusted Analysis for Acneiform Scars by Group Acneiform Scars Present Group Number Ranch Hand Comparison 150 183 Absent Percent Number 14.8 14.2 866 1,109 Percent 85.2 85.8 Total 1,016 1,292 Summary Statistics Est. RR: 1.05 95% C.I.: (0.83,1.33) p-Value: 0.720 Tests of association between the presence of acneiform scars in both groups and the covariates disclosed significant effects due to the four variables (p<0.001). As age increased, the proportion of participants with 14-14 �acneiform scars decreased (17.9% for participants born in or after 1942, 12.4% for participants born between 1923 and 1941, and 5.7% for participants born in or before 1922). Significantly more Blacks had scars than nonblacks (28.0% and 13.5%, respectively), and enlisted personnel had more than officers (enlisted groundcrew, 16.9%; enlisted flyers, 16.5%; and officers, 10.4%). The pre-SEA acne classification had a significantly higher proportion of participants with acneiform scars than the non pre-SEA acne classification. An adjusted analysis of the proportion of participants with acneiform scars was performed using logistic regression techniques. Results are given in Table 14-11. TABLE 14-11. Adjusted Analysis for Acneiform Scars by Group Ranch Hand Total 1,007 Comparison Total Adjusted Relative Risk. (95% C.I.) p-Value 1,282 1.07 (0.84,1.36) 0.584 Covariate Remarks Age (p=0.006) Race (p<0.001) Occupation (p=0.016) Presence of Pre-SEA Acne (p<0.001) No significant group differences were found (p=0.584). As in the covariate analysis with acneiform scars, significant effects in the adjusted analysis were observed due to all four covariates (age, p=0.006; race, p<0.001; occupation, p=0.0i6; presence of pre-SEA acne, p<0.001). The results for acneiform scars, as with the acneiform lesions, were quite similar in the followup and Baseline studies. Depigmentation Table 14-12 shows the contrast between the proportion of participants with depigmentation in the Ranch Hand and Comparison groups, unadjusted for any covariates. The proportion of participants with depigmentation was greater in the Comparison than in the Ranch Hand group; however, the difference between groups was nonsignificant (p=0.143). 14-15 �TABLE 14-12. Unadjusted Analysis for Depigmentation by Group Depigmentation Absent Present Group Ranch Hand Comparison Number Percent Number 102 155 10.0 12.0 Percent 914 1,137 90.0 88.0 Summary Statistics Total 1,016 1,292 Est. RR: 0.82 95% C.I.: (0.63,1.07) p- Value: 0.143 Tests of association between the presence of depigmentation in both groups and the four covariates determined a significant effect due to race (p=0.009), but shoved nonsignificant effects for age, occupation, and presence of pre-SEA acne. An adjusted analysis of the proportion of participants with depigmentation was performed using logistic regression techniques. The statistics are presented in Table 14-13. TABLE 14-13. Adjusted Analysis for Depigmentation by Group Ranch Hand Total 1,016 Comparison Total 1,292 Adjusted Relative Risk (95* C.I.) 0.82 (0.63,1.07) p-Value 0.144 Covariate Remarks Race (p=0.010) No significant difference was observed between groups (p=0.144). Race was the only significant covariate in this adjusted analysis (p=0.010). Depigmentation was not analyzed in the Baseline study. Inclusion Cysts As reflected in Table 14-14, there was not a significant difference between the proportion of participants with inclusion cysts in the Ranch Hand and Comparison groups, unadjusted for any covariates (p=0.303). 14-16 �TABLE 14-14. Unadjusted Analysis for Inclusion Cysts by Group Inclusion Cysts Present Group Ranch Hand Comparison Number Absent Percent Number 114 164 11.2 12.7 Percent 902 1,128 88.8 87.3 Total 1,016 1,292 Summary Statistics Est. RR: 0.87 95% C.I.: (0.67,1.12) p-Value: 0.303 Tests of association between the presence of inclusion cysts in both groups and the covariates of age, race, occupation, and presence of pre-SEA acne showed no significant effects due to age (p=0.437), race (p«0.506), or presence of pre-SEA acne (p-0.449). Occupation, however, exhibited a significant effect (p=0.036), with the enlisted flyer category having the highest proportion of participants with inclusion cysts (15.8% versus 11.9% and 10.8% for officers and enlisted groundcrew, respectively. An adjusted analysis of the proportion of participants with inclusion cysts was performed using logistic regression techniques. Results are presented in Table 14-15. TABLE 14-15. Adjusted Analysis for Inclusion Cysts by Group Ranch Hand Total 1,016 Comparison Total Adjusted Relative Risk (95% C.I.) p-Value Covariate Remarks 1,292 0.86 (0.67,1.12) 0.260 Occupation (p=0.041) No significant differences for inclusion cysts were found between the Ranch Hand and the Comparison groups (p=0.260). Occupation was the only significant covariate in this analysis (p=0.041). With reference to the Baseline study, the percentage of participants with inclusion cysts at the followup increased in the Comparison group, and 14-17 �decreased slightly in the Ranch Hand group. These differences could be due to changes in disease over time, different examiners, or changes in the cohorts examined. Both estimated and adjusted relative risks were less than one in the followup, while the estimated relative risk at the Baseline was slightly greater than one (RR=1.10 for Original Comparisons) but was not statistically significant. Hyperpigmentation Table 14-16 shows there was not a significant difference between the proportion of participants with hyperpigmentation in the Ranch Hand and Comparison groups, unadjusted for any covariates (p=0.762). TABLE 14-16. Unadjusted Analysis for Hyperpigmentation by Group Hyperpigmentation Present Group Ranch Hand Comparison Absent Number Percent Number 228 283 22.4 21.9 788 1,009 Percent 77.6 78.1 Total 1,016 1,292 Summary Statistics Est. RR: 1.03 95% C.I.: (0.85,1.26) p-Value: 0.762 Tests of association between the presence of hyperpigmentation in both groups and the four covariates revealed there was not a significant effect due to age (p=0.833), but that significant effects were due to race (p<0.001), occupation (p<0.001), and presence of pre-SEA acne (p=0.003). Blacks had a much higher prevalence of hyperpigmentation than nonblacks (53.1% for Blacks, 20.1% for nonblacks), and enlisted personnel had a higher prevalence of hyperpigmentation than officers (enlisted groundcrew, 25.5%; enlisted flyers, 23.5%; officers, 17.4%). The proportion of participants with hyperpigmentation was greater in the absence of pre-SEA acne (23.8%) than in the presence of pre-SEA acne (18.2%). An adjusted analysis of the proportion of participants with hyperpigmentation was performed using logistic regression techniques. Results are given in Table 14-17. 14-18 �TABLE 14-17. Adjusted Analysis for Hyperpigmentation by Group Ranch Hand Total 1,007 Comparison Total 1,282 Adjusted Relative Risk (95% C.I.) 1.04 (0.85,1.27) p-Value 0.720 Covariate Remarks Race (p<0.001) Occupation (p-0.009) Presence of Pre-SEA Acne (p=0.009) No significant group differences (p=0.720) were noted, although significant effects of race (p<0.001), occupation (p=0.009), and presence of pre-SEA acne (p=0.009) were evident. The proportion of participants with hyperpigmentation has increased since the Baseline study. Almost three times as many abnormalities were , found at the followup study (approximately 22% versus 8%). The relative risk estimate was closer to 1 in the followup study, but relative risks from both the Baseline and followup studies were not significantly different from 1. These differences could be due to disease or examination techniques. Other Abnormalities The study of other abnormalities encompassed a wide range of dermatological disorders. Included in this variable were the following abnormalities: (1) (2) (3) (4) (5) (6) (7) (8) Jaundice Spider Angiomata Palmar Erythema Suspected Melanoma Palmar Keratoses Actinic Keratoses Petechiae Ecchymoses (9) (10) (11) (12) (13) (14) (15) (16) Conjunctival Abnormality Oral Mucosal Abnormality Fingernail Abnormality Toenail Abnormality Dermatographia Cutis Rhomboidalis Suspected Basal Cell Carcinoma Suspected Squamous Cell Carcinoma With respect to the category "Other Abnormalities," a participant was considered normal only if he was.negative for all of these conditions. If one or more abnormalities existed, then the participant was considered abnormal. 14-19 �As reflected in Table 14-18, there was not a significant difference between the proportion of participants with other abnormalities in the Ranch Hand and Comparison groups, unadjusted for any covariates (p=0.349). TABLE 14-18. Unadjusted Analysis for Other Abnormalities by Group Other Abnormalities Abnormal Group Ranch Hand Comparison Number 608 748 Normal Percent Number 59.8 57.9 408 544 Percent 40.2 42.1 Total 1,016 1,292 Summary Statistics Est. RR: 1.08 95% C.I.: (0.92,1.28) p-Values 0.349 Tests of association between the presence of other abnormalities in both groups and the four covariates found a marginally significant effect due to the presence of pre-SEA acne (p=0.084), and significant effects due to age (p<0.001), occupation (p<0.001), and race (p<0.001). The proportion of participants with other abnormalities in the absence of pre-SEA acne (59.9%) was marginally significantly larger than the proportion of participants with other abnormalities who also had pre-SEA acne (56.1%). The proportion of participants with other abnormalities increased with age (with a low of 43.3% for participants born in or after 1942 to a high of 82.8% for participants born in or before 1922). Nonblacks had a significantly and substantially higher percentage of other abnormalities than Blacks (60.3% and 35.7%, respectively). Enlisted groundcrew had a lower proportion of abnormalities than officers or enlisted flyers (53.3%, 63.2%, and 63.8%, respectively). An adjusted analysis of the proportion of participants with other abnormalities was performed using logistic regression techniques. Results are presented in Table 14-19. Again, no significant difference was observed between groups (p=0.432). Age and race were significant covariates in this analysis (p<0.001 for both). In reference to the Baseline study, the percentage of participants with other abnormalities has increased in both the Comparison and the Ranch Hand groups. In the Baseline study, the estimated relative risk for Ranch Hands versus Original Comparisons was 0.77, significantly less than 1.00. The estimate of the relative risk has increased in the followup study to 1.08. The percentage of other abnormalities has increased from approximately 14 percent in the Baseline study to nearly 59 percent in the followup study. 14-20 �TABLE 14-19. Adjusted Analysis for Other Abnormalities by Group Ranch Hand Total 1,016 Comparison Total 1,292 Adjusted Relative Risk (95% C.I.) 1.07 (0.90,1.28) p-Value 0.432 Covariate Remarks Age (p<0.001) Race (p<0.001) Dermatology Index Four of the previously analyzed conditions (comedones, acneiform lesions, acneiform scars, and inclusion cysts) were used to construct a dermatology index. All four conditions are indicators of possible chloracne. The index was formulated by counting the number of abnormalities present in a participant for the four conditions. Consequently, the dermatology index ranged from 0 to 4, where 0 indicated that the participant had none of these abnormalities and 4 indicated that the participant had all of these abnormalities. Table 14-20 presents the number and the percent of participants with each of these five scores by group. A significant difference between the Ranch Hand and Comparison groups was not observed for this dermatology index, unadjusted for any covariates (p=0.576, 4 d.f.). Covariate main effect analyses found nonsignificant effects due to age (p=0.407) and race (p=0.558), but significant effects for occupation (p-0.010) and the presence of acne pre-SEA (p<0.001). These data are summarized in Table 14-21. By occupation, 55.8 percent of the officers had no abnormalities, whereas 50.8 percent of the enlisted groundcrew and 44.4 percent of the enlisted flyers had no abnormalities. The stratum corresponding to participants with pre-SEA acne present had a larger percentage of participants with at least one abnormality (see Table 14-21). An adjusted analysis of the five scores of the dermatology index was performed using log-linear modeling techniques. Significant effects were noted for occupation and an interaction between group and presence of pre-SEA acne (p=0.005, p=>0.041, respectively). Consequently, an analysis, stratifying by pre-SEA acne status, was performed, and the results are shown in Table 14-22. The adjusted relative risk for each of the index scores (1 to 4, separately, versus the 0 score), the 95 percent confidence interval, and the p-value for each contrast for each pre-SEA acne class are given in Table 14-23. 14-21 �TABLE 14-20. Unadjusted Analysis for the Dermatology Index by Group Dermatology Index Score 0 Group Ranch Hand Comparison 1 2 3 Number Percent Number Percent Number Percent Number 533 658 52.5 50.9 318 420 31.3 32.5 121 154 11.9 11.9 34 53 Overall p-Value (4 d.f.)=0.576 Contrast Est. Relative Risk (95% C.I.) p-Value* N9 to 1 vs. 0 2 vs. 0 3 vs. 0 4 vs. 0 *Fisher's exact test. 0.94 0.97 0.79 1.76 ( .78 ,1 .13) 0 ( .75 ,1 .26) 0 ( .51 ,1 .24) 0 ( .67 ,4.66) 0 0.480 0.840 0.317 0.327 Percent 3.3 4.1 Number Percent 10 7 1.0 0.5 Total 1,016 1,292 �TABLE 14-21. . o v - - r and Presence of Pre-SE&i^mt^ in the Combined Ranch —_ _ j vw^—jw 1 rbwi ani O"*nrwm<vn GCOUDS •»—-»~y-i» Dermatology Index Score* 0 1 2 3 4 Covariate Covariate Category Age Born XL942 501 52.2 2% 30.8 114 11.9 40 4.2 9 0.9 960 Bom 1923-1941 Born <L922 647 43 51.3 49.4 408 34 32.4 39.1 156 5 12.4 5.7 42 5 3.3 5.7 '8 0 0.6 0.0 1,261 87 83 1,108 58.0 51.2 38 700 26.6 32.3 17 258 11.9 11.9 4 83 2.8 3.8 1 16 0.7 0.8 143 2165 0.558 482 172 55.8 44.4 265 135 30.7 34.9 91 58 10.5 21 19 2.4 4.9 5 3 0.6 0.8 864 387 0.010 15.0 537 50.8 338 32.0 126 11.9 47 4.4 9 0.9 1,057 842 337 54.0 46.2 514 220 32.9 30.1 153 121 9.8 16.6 44 42 2.8 5.8 7 9 0.4 1.2 1,560 729 Race Number Percent Number Percent Number Percent Number Percent Number Percent Black Nonblack Total" p-VaW 0.407 Occupation Officer Bilisted Flyer Enlisted Groundcrev Presence of Pre-Sea Acne** No Yes 0.001 * Score denotes the number of abnormalities (for comedones, acneiform lesions, acneiform scars, and inclusion cysts) diagnosed. **Nineteen participants could not be classified. * One participant refused to take the dermatology examination. b Pearson's chi-square test. �TABLE 14-22. Adjusted Analysis for the Dermatology Index by SEA Acne Class and Group Dermatology Index Score* 1 0 Pre-SEA Acne Class ^ No pre-SEA -t- acne i Group Number Percent Number 2 Percent Number 3 ' Percent Number Percent 4 Number . Percent Total Ranch Hand Comparison 360 482 52.6 55.0 234 280 34.2 32.0 69 84 10.1 9.6 16 28 2.3 3.2 5 2 0.7 0.2 684 876 Ranch Hand Comparison 167 170 51.7 41.9 82 138 25.4 34.0 51 70 15.8 17.2 18 24 5.6 5.9 5 4 1.5 1.0 323 406 Pre-SEA acne - *Score denotes the number of abnormalities (comedones, acneiform lesions, acneiform scars, and inclusion cysts) diagnosed. �TABLE 14-23. Adjusted Relative Risks for Contrasts of Dermatology Index by Pre-SEA Class Pre-SEA Acne No Yes Contrast Adjusted Relative Risk 95% C.I. p-Value 1 abnormality vs. 0 abnormalities 2 vs. 0 3 vs. 0 4 vs. 0 1.12 1.10 0.77 3.09 (0.90,1-39) (0.77,1.55) (0.41,1.44) (0.65,14.62) 0.315 0.605 0.411 0.155 1 2 3 4 0.60 0.73 0.75 1.19 (0.42,0.85) (0.48,1.12) . (0.39,1.43) (0.33,4.38) 0.004 0.148 0.380 0.788 vs. 0 vs. 0 vs. 0 vs. 0 This analysis showed a significant difference between groups only when contrasting the proportion of participants with one abnormality (out of four) to the proportion of participants with no abnormalities for participants with pre-SEA acne (p=0.004). However, Comparisons were more likely to have one abnormality than the Ranch Hands, as is evidenced by the relative risk and confidence interval being less than 1. In contrast to the Baseline study, the percentage of participants with a score of 1 or more has increased at the followup examination for both the Ranch Hand and Comparison groups (8.1% for Ranch Hands, 12.1% for Comparisons). The estimated relative risks, when the dermatology index is condensed into two categories, were 1.11 for the Baseline examination and 0.94 for the followup examination. Biopsy Results Dermatologists were instructed to perform skin biopsies on any lesions they suspected of being malignant. Of the 40 biopsies collected from 35 participants, none was suggestive of chloracne. Histologic descriptions of these biopsies are presented in Table 14-24. With the exception of confirmed basal cell carcinoma, no single diagnostic category predominated. 14-25 �TABLE 14-24. Summary of Histologic Descriptions of Skin Biopsy by Group Group Ranch Hand Basal Cell Carcinoma Suspected Basal Cell Carcinoma Suspected Unspecified Carcinoma Unspecified Carcinoma Dermatofibroma Pigmented Nevus Dyschromia Keratoderma, Acquired Melanoacanthoma (Papilloma) Intradermal Nevus Junctional Nevus Cavernous Hemangioma Degenerative Skin Disorder Other Specified Disorders of Skin Local Infection of Skin Other Dermatoses Total Comparison 7 0 0 1 3 1 1 1 0 1 0 0 0 0 1 5 Histologic Description 4 3 1 0 0 2 0 1 1 0 1 1 1 1 0 2 21 Comments a,b b c d a c c 18 *0ne participant had a basal cell carcinoma at one site and an acquired keratoderma at another site. 'One participant had a basal cell carcinoma at one site and a suspected basal cell carcinoma at another site. : 0ne participant had a local infection of the skin, a suspected unspecified carcinoma, and a dermatosis at the same site. J 0ne participant had two cases of dyschromia at two different sites. 14-26 �EXPOSURE INDEX ANALYSES Exposure index analyses were conducted within each occupational cohort of the Ranch Hand group to search for dose-response relationships (see Chapter 8 for details on the exposure index). The dermatology index was collapsed into two categories, 0 and greater than 0. All eight dermatological variables were explored, unadjusted for any covariates, using Pearson's chi-square test and Fisher's exact test. Adjusted analyses were performed by logistic regression for these variables, using age, race, presence of pre-SEA acne, and any significant pairwise interactions between the exposure index and these covariates. Overall significance in the proportion of abnormalities among the exposure index levels of low, medium, and high was determined, as well as contrasts in the proportion of abnormalities between the medium and low exposure levels, and between the high and low exposure levels. Age was used as a continuous variable in the adjusted analyses. Results of the adjusted analyses for these eight variables are presented in Table 14-25, and counterpart results for unadjusted analyses are presented in Table L-l of Appendix L. Results from further investigation of exposure index by covariate interactions are given in Table L-2 of Appendix L. Significant or marginally significant results were present for some of these variables based on unadjusted analyses. A borderline significantly higher prevalence of comedones (Est. RR: 1.78, 95% C.I.: [0.95,3.35], p=0.084) for the contrast of medium exposure to low exposure was seen for officers. Marginally significant results for the contrast of high exposure to low exposure were also present for acneiform scars for officers (Est. RR: 2.38, 95% C.I.i [0.94,6.06], p=»0.075) and enlisted groundcrew (Est. RR: 1.82, 95% C.I.: [1.00,3.30], p-0.053), as well as for other abnormalities for officers (Est. RR: 1.66, 95% C.I.: [0.98,2.78], p=0.067). The data for these last three variable-occupation combinations supported an increase in the proportion of abnormalities from low to high exposure. Significant or marginally significant results were also observed for medium exposure versus low exposure in officers and enlisted groundcrew for depigmentation, and for high exposure versus low exposure in other abnormalities with enlisted flyers, but prevalence decreased as the exposure level increased in these cases. The frequency of abnormalities for the different exposure index levels exhibited no consistent pattern across occupations. However, within the officer and enlisted groundcrew occupations, most variables showed the low exposure level to have the lowest prevalence of abnormalities or the high exposure level to have the highest prevalence, whereas very few variables showed this pattern for enlisted flyers. Adjusted analyses revealed patterns similar to those of the unadjusted analyses. Results of the counterpart adjusted analyses to the situations described above are detailed below. (1) Comedones in officers, medium versus low: Adj. RR: 1.62, 95% C.I.: (0.83,3.15), p=0.154. 14-27 �TABLE 14-25. Adjusted Exposure Index Analysis for Dermatological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj . Relative Risk ( 5 C.I.) 9% p-Value Officer 122 Overall M vs. L H vs. L 0.334 1.62 (0.83,3.15) 0.154 1.44 (0.74,2.83) 0.283 Enlisted Flyer 55 65 56 Overall M vs. L H vs. L 0.413 0.65 (0.30,1.41) 0.276 0.61 (0.27,1.37) 0.234 152 162 140 Overall M vs. L H vs. L 0.878 0.94 (0.55,1.60) 0.808 1.08 (0.63,1.83) 0.782 Officer ^ *• 129 Enlisted Groundcrew Comedones 126 126 129 122 Overall M vs. L H vs. L 0.669 1.06 (0.52,2.15) 0.880 1.34 (0.67,2.66) 0.409 Enlisted Flyer 55 65 56 Overall M vs. L H vs. L 0.917 0.91 (0.32,2.60) 0.856 1.14 (0.39,3.35) 0.814 152 162 140 Overall M vs. L H vs. L 0.674 1.01 (0.58,1.75) 0.973 1.25 (0.71,2.20) 0.431 tsj OO Acneiform Lesions Enlisted Groundcrew �TABLE 14-25. (continued) Adjusted Exposure Index Analysis for Dermatological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value » Officer 126 129 122 Overall M vs. L H vs. L **() **! **() **! **() **! **() **! **() **! 55 65 56 Overall M vs. L H vs. L 0.363 0.82 (0.31,2.13) 0.682 0.47 (0.16,1.39) 0.174 152 162 140 Overall M vs. L H vs. L 0.068 1.22 (0.66,2.27) 0.519 2.00 (1.08,3.67) 0.026 Officer ^ *• " Enlisted Flyer Enlisted Groundcrewa Acneiform Scars 126 129 122 Overall M vs. L H vs. L 0.006 0.33 (0.11,0.98) 0.045 1.50 (0.69,3.25) 0.302 Enlisted Flyer 55 65 56 Overall M vs. L H vs. L 0.493 0.53 ( . 8 1 5 ) 0.245 01,.4 0.67 (0.24,1.90) 0.450 152 162 140 Overall M vs. L H vs. L ****(2) ****(2) to VO Depigmentation Enlisted Groundcrew **() **2 **() **2 **() **2 �TABLE 14-25. (continued) Adjusted Exposure Index Analysis for Dermatological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj. Relative Risk (95% C.I.) p-Value Officer13 Hyperpigmentation 122 Overall M vs. L H vs. L 0.221 2.05 ( .91,4.60) 0.082 0 1.32 ( .56,3.11) 0.532 0 Enlisted Flyer 55 65 56 Overall M vs. L H vs. L 0.881 1.24 ( .41,3.78) 0.707 0 1.33 ( .42,4.17) 0.630 0 152 162 140 Overall M vs. L H vs. L 0.916 0.91 ( .43,1.93) 0.806 0 1.07 ( .51,2.24) 0.856 0 Officer _, *u> o 129 Enlisted Groundcrew Inclusion Cysts 126 126 129 122 Overall M vs. L H vs. L 0.813 0.92 ( .47,1.79) 0.807 0 0. 0 ( .41,1.58) 0.525 8 0 Enlisted Flyer 55 65 56 Overall M vs. L H vs. L 0.656 0.71 ( .29,1.76) 0.465 0 1.04 ( .43,2.53) 0.930 0 152 162 140 Overall M vs. L H vs. L 0.365 1.20 (0.71,2.01) 0.494 0.81 ( . 6 1.41) 0.450 04, Enlisted Groundcrew �TABLE 14-25. (continued) Adjusted Exposure Index Analysis for Dermatological Variables by Occupation Exposure Index Variable Occupation Low Total Medium Total High Total Contrast Adj. Relative Risk ( 5 C.I.) p-Value 9% Officer 122 Overall M vs. L H vs. L 0.309 1.30 (0.75,2.24) 0.346 1.53 (0.88,2.65) 0.129 Enlisted Flyer 55 65 56 Overall M vs. L H vs. L 009 .4 0.66 (0.28,1.56) 0.341 0.35 ( . 4 0 8 ) 0.018 01,.3 152 162 140 Overall M vs. L H vs. L 0.764 0.85 (0.52,1.36) 0.489 0.87 (0.52,1.43) 0.580 Officer Dermatology Index 129 Enlisted Groundcrew Other Abnormalities 126 126 129 122 Overall M vs. L H vs. L **() **! **() **! Enlisted Flyer Enlisted Groundcrew **() **! **() **! **() **! 55 65 56 Overall M vs. L H vs. L 0.618 0.74 (0.36,1.54) 0.423 0.71 (0.33,1.51) 0.368 152 162 140 Overall M vs. L H vs. L 0.469 1.01 (0.65,1.59) 0.955 1.30 (0.82,2.06) 0.270 "Marginal exposure index-by-presence of pre-SEA acne interaction (p=0.056)—relative risk, confidence interval and p-value presented, and additional information provided in interaction summaries. * * ( ) Exposure index-by-presence of pre-SEA acne and exposure index-by-race interaction—relative risk, **!: confidence interval, and p-value not presented. * * ( ) Exposure index-by-presence of pre-SEA acne interaction—relative risk, confidence interval and **2: p-value not presented. �(2) Acneiforra scars in officers, high versus low: interaction present; direct contrast of adjusted and unadjusted analyses not possible. (3) Acneiform scars in enlisted groundcrew, high versus lows Adj. RR: 2.00, 95% C.I.'. (1.08,3.67), p=0.026; overall p-value=0.068, increase in the proportion of abnormalities with increasing exposure levels supported. (4) Other abnormalities in officers, high versus lows Adj. RR: 1.53, 95% C.l.t (0.88,2.65), p=0.129. Other adjusted analyses that showed significance or marginal significance exhibited a decreasing prevalence with increasing exposure level. All other adjusted analyses showed an interaction with covariates (described below) or nonsignificant results. Interactions were present for three of the eight variables and were observed for officers and enlisted groundcrew. A summary of these interactions is presented below in Table 14-26. TABLE 14-26. Summary of Exposure Index by Covariate Interactions Encountered in Adjusted Analysis of Dermatological Variables Occupation Covariate p-Value Acneiform Scars Officer Race 0.003 Acneiform Scars Officer Presence of Pre-SEA acne 0.003 Variable Acneiform Scars Enlisted Groundcrew Presence of Pre-SEA acne (marginal) 0.056 Depigmentation Enlisted Groundcrew Presence of Pre-SEA acne 0.035 Dermatology Index* Officer Race 0.026 Dermatology Index* Officer Presence of Pre-SEA acne 0.029 *Variable was collapsed into two categories, 0 and >0. 14-32 �As can be seen, all variables and occupations with interactions had a significant exposure index-by-presence of pre-SEA acne interaction or significant exposure index-by-race and exposure index-by-presence of pre-SEA acne interactions. Meaningful interpretation of many of the subsequent stratified analyses was hindered by small sample sizes, but two situations were of particular interest. For acneiform scars on officers, nonblack personnel without pre-SEA acne at low exposure had no participants with scars, whereas nonblack personnel exposed at the medium and high levels had 7.8 percent and 10.5 percent of participants with scars, respectively. Also, with acneiform scars for enlisted groundcrew, an increase in the prevalence of abnormalities for increasing levels of exposure was present for participants with pre-SEA acne, with an adjusted relative risk of 5.38 (95% C.I.: [1.45,19.96], p=0.012) for the contrast of high exposure versus low exposure. In summary, the results suggested the presence of an increasing doseresponse relationship in certain occupations for a few of the dermatological variables or within substrata of these variables, but no consistent pattern was evident throughout the dermatological exposure index evaluation. LONGITUDINAL ANALYSES The dermatology index was chosen to assess longitudinal differences between the 1982 Baseline examination and the 1985 followup examination. In testing for this difference, the dermatology index scores were collapsed into two categories: normal (dermatology index score of 0) and abnormal (dermatology index score greater than 0). As shown in Table 14-27, 2x2 tables were constructed for each group. These tables show the number of participants who were abnormal at the Baseline examination and abnormal at the followup, abnormal at Baseline and normal at followup, normal at Baseline and abnormal at followup, and normal at both Baseline and followup. The odds ratios given are the ratios of the number of participants who were normal at the Baseline and abnormal at the followup to the number of participants who were abnormal at the Baseline and normal at the followup (the "off-diagonal" elements). TABLE 14-27. Longitudinal Analysis of the Dermatology Index: A Contrast of Baseline and First Followup Examination Abnormalities 1982 1985 Followup Exam Group Baseline Exam Abnormal Normal Ranch Hand - Abnormal Normal 241 228 136 366 1.68 Comparison Abnormal Normal 283 283 136 437 2.08 Odds Ratio (OR)* p-Value (ORRH vs ORC) 0.15 *0dds Ratio: Number Normal Baseline, Abnormal Followup Number Abnormal Baseline, Normal Followup 14-33 �The changes in normal/abnormal status within each group were compared, and the p-value given was derived from Pearson's chi-square test of the hypothesis that the pattern of change in the two groups was the same. These results showed that the difference in the pattern is not significant (p=0.15). DISCUSSION The relative risks for all eight dermatological variables approached unity (none was statistically significant), an observation previously noted at the Baseline examination (except for the category "Other Abnormalities," which predominated in the Comparisons). More dermatological abnormalities were recorded at the followup (for six of the seven variables shared between the examinations) than at the Baseline—the increase in detection was slightly stronger in the Comparison group than in the Ranch Hand group. For example, in the category "Other Abnormalities," the reporting of skin lesions generally increased from about 14 percent to 59 percent. The overall difference between the two examinations probably reflects a combination of factors, e.g., changes in disease, chance, the addition of new participants, and possible differences in clinical practice between the two groups of dermatologists. The histologic categories of skin cancer (confirmed or suspected, any type), as examined by biopsies, showed a similarity between both groups. SUMMARY AND CONCLUSIONS Interval questionnaire data on the occurrence, time, and location of acne were analyzed to assess the possible historical diagnosis of chloracne. No significant difference was observed between groups for reported occurrence of acne, although the Ranch Hand cohort reported slightly more acne. The occurrence of acne relative to 1961 was comparable between groups. A marginally significant difference in the occurrence of post-1961 acne was found, with more Ranch Hands than Comparisons reporting acne strictly post-SEA. The duration of post-1961 acne was not significantly different between the two groups. For participants with post-SEA acne, the spatial eyeglass distribution of acne (suggesting chloracne) was observed to be similar for the Ranch Hand and Comparison groups, both for individual sites and the combination of acne on the eyelids, ears, and temples. This analysis suggested that the occurrence of skin disease compatible with chloracne was not different in the two groups. Analyses of the followup physical examination data, as with the Baseline examination, placed primary emphasis on six dermatologic disorders: comedones, acneiform lesions, acneiform scars, inclusion cysts, depigmentation, and hyperpigmentation. Secondary emphasis was given to 16 other minor conditions (generally not associated with chloracne) recorded at the physical examination. No significant findings occurred in any variable, as reflected in Table 14-28. 14-34 �TABLE 14-28. Overall Summary Results of Unadjusted and Adjusted Analyses of Questionnaire and Physical Examination Dermatological Variables Variable Unadjusted Adjusted Questionnaire Incidence of Acne Occurrence NS Relative to 1961 NS Relative to SEA (Post-1961 Cases) NS* Duration of Acne NS Location, of Acne NS — — — NS — Physical Examination Comedones NS NS Acneiforra Lesions NS NS Acneiform Scars NS NS Depigmentation NS NS Inclusion Cysts NS NS Hyperpigmentation NS NS NS NS NS **** Other Abnormalities . Dermatology Index NS: Not significant (p>0.10). — Analyses not performed. NS*: Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. 14-35 �No significant difference vas found for any of these variables in the unadjusted analyses. The variable consisting of the 16 secondary conditions, labeled "other abnormalities," had the largest difference in the prevalence of abnormalities for the Ranch Hand cohort over the Comparison group (Est. RR: 1.08, 95% C.I.: [0.92,1.28], p=0.349), but the difference was clearly nonsignificant. The covariate effects of age, race, occupation, and the presence of pre-SEA acne were often profound with respect to the recorded dermatologic conditions. The adjusted analyses closely mirrored the unadjusted analyses, with no significance noted between groups for any variable. Only one group-bycovariate interaction was observed in the adjusted analysis of the dermatology index, with a group-by-presence of pre-SEA acne interaction noted. However, further analysis of this interaction did not show an adverse effect in the Ranch Hand group. Exposure index analyses did support dose-response relationships for some of the variables in certain occupational strata, but did not reveal a strong pattern of results suggesting a relationship between skin disease and herbicide exposure. Overall, the followup examination results paralleled the Baseline findings. Although the followup examination detected more dermatologic abnormalities than those present at Baseline, slightly more abnormalities were found in the Comparisons, and most relative risks approached unity. The longitudinal analysis for the dermatology index showed no statistically significant differences between groups in the change in results from the Baseline to the followup examination. In conclusion, none of the questionnaire results disclosed an increased likelihood of past chloracne in the Ranch Hands. The physical examination did not diagnose a current case of chloracne. The dermatological data were similar between the Ranch Hand and Comparison groups, and the longitudinal analysis of the dermatology index suggested equivalence between the Baseline and followup examination results. 14-36 �CHAPTER 14 REFERENCES 1. Kimmig, J., and K.H. Schulz. 1957. Occupational acne due to chlorinated aromatic cyclic esters. Dermatologica 115:540. 2. Kimmig, J., and K.H. Schulz. 1957. Chlorinated aromatic cyclic ethers as the cause of so-called chloracne. Naturvissenschaften 44:337-338. \ 3. Jones, E.L., and H. Kizek. 1962. A technique for testing acnegenic potency in rabbits, aplied to potent acnegen, 2,3,7,8 tetrachlorodibenzo-p-dioxin. J. Invest. Dermatol. 9:511-517. 4. McConnell, E.E., J.A. Moore, and D.W. Dalgard. 1978. Toxicity of 2,3,7,8-tetrachlorodibenozo-p-dioxin in Rhesus monkeys (Macaca mulatta) following a single oral dose. Toxicol. Appl. Pharmacol. 43(1):175-187. 5. Kimbrough, R.D. 1980. Occupational exposure. No. 4 in Halogenated biphenyls, terphenyls, naphthalenes, dibenzodioxins, and related" roducts, ed. R.D. Kimbrough, p. 373. Topics in Environ. Health", Isevier/North Holland, Amsterdam. f 6. Young, A.L. 1980. The chlorinated dibenzo-p-dioxins. Chap. 5 in The science of 2,4,5-T and associated phenoxy herbicides, ed. R.L. Metcalf and W. Stumm, pp. 133-205. New York:Wiley-Interscience. 7. Knutson, J.C., and A. Poland. 1982. Response of murine epidermis to 2,3,7,8-tetrachlorodibenzo-p-dioxin: Interaction of the Ah and hr loci. Cell 30: 225-234. 8. Kay, J.H., R.J. Palazzolo, and J.C. Calandra. 1965. Subacute dermal toxicity of 2,4-D. Arch. Environ. Health 11:648-651. 9. Carter, C.D., R.D. Kimbrough, J.A. Liddle, R.E. Cline, M.M. Zack, W.F. Barthel, R.E. Koehler, and P.E. Phillips. 1975. Tetrachlorodibenzodioxin: An accidental poisoning episode in horse arenas. Science 188(4189):738-740. 10. Case, A.A. 1976. Tetrachlorodibenzodioxin (TCDD)—clinical aspects of poisoning. Clin. Toxicol. 9(6):963-967. 11. Greenlee, W.F., R. Osborne, L.G. Hudson, and W.A. Toscano. 1984. Studies on the mechanisms of toxicity of TCDD to human epidermis. In Banbury report 18: Biological mechanisms of dioxin action; ed. A. Poland and R.D. Kimbrough, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory, pp. 365-372. 14-37 �12. Reggiani, G. 1980. Acute human exposure to TCDD in Seveso, Italy. Toxicol. Environ. Health 6:27-43. J^ 13. May, G. 1973. Chloracne from the accidental production of tetrachlorodibenzodioxin. Br. J. Ind. Med. 30:276-283. 14. Jirasek, L., J. Kalensky, and K. Kubec. 1973. Acne chlorina and porphyria cutanea tarda during the manufacture of herbicides, part 1. Czech. Dermatol. 48(5):306-315. 15. Bleiberg, J., M. Wallen, R. Brodkin, and I.L. Applebaum. 1964. Industrially acquired porphyria. Arch. Dermatol. 89:793-797. 16. Suskind, R.R,, and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 17. Oliver, R.M. 1975. Toxic effect of 2,3,7,8-tetrachloro-dibenzo-l, 4-dioxin in laboratory workers. Br. J. Ind. Med. 32:46-53. 18. Crow, K.D. 1983. Significance of cutaneous lesions in the symptomatology of exposure to dioxins and other chloracnegens. In human and environmental risks of chlorinated dipxins and related compounds, R.E. Tucker, et al., eds.pp. 605-612, New York:Plenum Press. 19. Allen, A.M. 1977. Skin diseases in Vietnam, 1965-1972. Internal Medicine in Vietnam, Vol. 1, ed. A.J. Ognibene, p. 42. Washington, D.C.: Center of Military History, Government Printing Office. \ 20. Halprin, K.M. 1980. Chloracne recognition and its significance. Presented at the Second Continuing Education Conference on Herbicide Orange, Washington, D.C., May 28-30. 21. Crow, K.D. 1970. Chloracne. 56:79-90. Trans. St. John's Hosp. Dermatol. Soc. 22. Hoffman, R.E., P.A. Stehr-Green, K.B, Webb, G. Evans, A.P. Knutsen, W.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzo-p-dixoin. JAMA 255:2031-2038. 23. Stehr, P.A., G. Stein, H. Falk, et al. 1986. A pilot epidemiologic study of possible health effects associated with 2,3,7,8-tetrachlorodibenzo-p-dioxin contamination in Missouri. Arch. Environ. Health 41:16-22. 24. Jirasek, L., J. Kalensky, K. Kubec, et al. 1974. Acne chlorina, porphyria cutanea tarda and other manifestations of general intoxication during the manufacture of herbicides, part 2. Czech. Dermatol. 49(3):145-157. 25. Goldmann, P.J. 1973. Schweistakute Chlor-akne, eine Massenintoxikation durch 2,3,7,8-Tetrachlorodibenzodioxin (Severe acute chloracne, a mass intoxication due to 2,3,6,7-tetrachlorodibenzodioxin). Hautarzt 24(4):149-152. 14-38 �26. Okumura, H., and S. Katauki. 1969. A clinical study of oil disease (chlorinated biphenyl poisoning), particularly the internal medical signs. Fukuoka Acta Med. 60:440-446. 27. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichlorophenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 28. Lathrop, G.D. 1985. Assessments of a controversy: Agent orange and its association with dioxin, science assessment, toxicology forum. Given at the Institute of Pathobiology; Aspen, Colorado, July 1985. 14-39 �CHAPTER 15 CARDIOVASCULAR EVALUATION INTRODUCTION Cardiac disease and peripheral vascular disease are not classically recognized sequelae of exposure to phenoxy herbicides, chlorophenols, or dioxin. Most observational and experimental animal studies using 2,4-D, 2,4,5-T, or TCDD have not extensively commented on resulting cardiac abnormalities or dysfunction. The studies described below viewed the cardiac abnormalities as expected consequences of a moribund*state, and not as an indicator of primary cardiac toxicity to the putative chemical. Following oral administration of 2,4-D and 2,4,5-T, sheep and cattle developed cardiac hemorrhages. A lethal oral dose of TCDD in young Rhesus monkeys produced increased heart weights in another experiment.2 Horses and cats showed generalized vascular degeneration following exposure to soil contaminated with TCDD, and mice and guinea pigs fed high amounts of TCDD manifested low heart weights. A teratogenic experiment using 2,4,5-T in developing fish eggs showed graduated lethality and cardiovascular anomalies, which included enlarged veins and heart chambers. Another study using ventricular muscle strips from chick embryos exposed to PCB's (including TCDD) showed a marked decrease in contractility. This primary cardiotoxic response was presumably mediated by the Ah receptor, and was associated with increased prostaglandin synthesis. Human case reports, case series of individuals with chloracne, and epidemiological studies also confirmed that cardiac function is not a sensitive indicator of exposure to herbicides or TCDD. In three case reports of acute 2,4-D poisoning, cardiac dilation and cardiac arrest were observed in the one fatal case, while only transient nodal tachycardia was observed in one of the two nonfatal cases. ' Three laboratory technicians with chloracne, neurological symptoms, and hypercholesterolemia following sig- 10 nificant direct exposure to TCDD did not manifest any cardiac dysfunction, however, of 10 industrial workers with chloracne, 4 complained of heart palpitations and shortness of breath. In another two studies totaling 128 industrial workers, no excesses of cardiac complaints or findings were noted. 1? Furthermore, in two contemporary epidemiological studies using similar cohorts from the Nitro, West Virginia, plant, no significant cardiac impairments were detected in exposed workers. f l However, one study found significantly lower levels of high density lipoprotein (HDL) cholesterol in 1 individuals with chloracne as contrasted to individuals without chloracne. Two recent clinical-epidemiological pilot studies of residential areas in Missouri contaminated by TCDD did not disclose any significant cardiac disease in exposed residents,1 ' 8 although the Times Beach study noted a borderline association of diminished peripheral pulses in the exposed group (as did the AFHS Baseline study). 15-1 �Because the herbicide literature has not identified consistent cardiovascular findings that merited a specific clinical focus, this study has collected generalized data on past cardiac events by questionnaire and medical record reviews. Current cardiac and peripheral vascular status were measured by physical examination and laboratory procedures. Coronary heart disease (CHD) has been of general concern in this study because both male cohorts are largely within the high risk ages of 40 to 65. Since TCDD probably does not directly and permanently affect cardiovascular function, a theoretical question that arises is whether TCDD might have altered a cardiovascular disease risk factor that will exert a future adverse impact. There may be indirect evidence for such a possibility. Risk factors for CHD include age, sex, race, family history, past personal history, diabetes (all types), smoking, cholesterol (and cholesterol-HDL ratio), diet, blood pressure, body weight, exercise pattern, stress (personality type), and alcohol. ~ Of these risk factors, hypertension and cholesterol have received consistent attention in clinical and epidemiological evaluations. Hypertension, either at routine examination or via specific study, has*not been related to phenoxy herbicide or TCDD exposure. However, hypercholesterolemia has been 13 6 24 2 associated with repeatedly acute exposure to chlorophenols and dioxin. °'12' ' ' ' Baseline Summary Results The 1982 Baseline examination found no statistically significant differences between the Ranch Hand and Comparison groups in systolic or diastolic blood pressure, the frequency of abnormal electrocardiographs (ECG's), heart sound abnormalities, abnormal funduscopic findings, or carotid bruits. However, a statistically significant difference emerged in the frequency of abnormal peripheral pulses: 12.8 percent of the nonblack Ranch Hands exhibited absent or diminished peripheral pulses compared to 9.4 percent of the nonblack Original Comparisons (p=0.05). This difference was consistent across various pulse combinations and remained statistically significant when all Ranch Hands were contrasted with all Comparisons, adjusting for age, past smoking history, and cholesterol level. No statistically significant differences were found between the two groups-in the occurrence of reported or verified heart disease or heart attacks, although a significant group-by-heart disease-by-smoking interaction was noted in the older (40 or more years of age) subgroup, i.e., older Ranch Hands smoking more than 10 pack-years developed more heart disease than their Comparisons, whereas older Ranch Hands smoking less than 10 pack-years exhibited less heart disease. No significant dose-response relationships of any of the cardiovascular response variables with the exposure index were noted. Over 80 percent of reported cardiac conditions obtained from the study questionnaire were verified by a detailed review of medical records. There was also strong correlation between the past medical history of cardiac disease and the Baseline cardiovascular examination findings. However, the differences in peripheral pulse abnormalities primarily occurred in older individuals without a history of cardiovascular disease. These abnormalities, therefore, may be a precursor to more serious arterial disease or central dysfunction. 15-2 �. Finally, the veil-known risk factors of age, smoking, and cholesterol were found to be highly correlated with each other and with several of the cardiovascular response variables. Parameters of the 1985 Cardiovascular Examination The 1985 cardiovascular examination was very similar to the 1982 Baseline examination. Data collection was divided into three major categories: heart disease history, central cardiac function, and peripheral vascular function. Historical data were collected by a questionnaire administered at the examination site, covering the interval from 1982 through 1985. In addition, the review-of-systerns portion of the physical examination recorded the overall history of heart trouble and other serious illnesses. Medical records were sought on all individuals to verify the reported conditions and to determine the time of occurrence of major cardiac events. Each participant was classified as to whether or not he developed essential hypertension, and whether he developed heart disease or had an acute myocardial infarction since his tour of duty in Southeast Asia (SEA). These endpoints were analyzed along with all other dependent variables to assess the degree of correlation between the history of cardiovascular disease and present medical findings. In addition, mortality findings were combined with the cardiovascular disease histories to form additional endpoints. . Central cardiac function was assessed by the measurements of systolic blood pressure, heart sounds (by auscultation), and an ECG. Blood pressure was determined in a standardized manner (see section on Physical Examination Data), and all examiners.and diagnosticians were retrained on the detection of fourth heart sounds and the notation of innocent murmurs without recording them as abnormal heart sounds. ECG's were obtained after adherence to a 4-hour fast and abstinence from tobacco. Twelve-lead ECG's were recorded with a rhythm strip, and the following items were considered to be abnormal: right bundle branch block (RBBB), left bundle branch block (LBBB), nonspecific T-wave changes, bradycardia, tachycardia, arrhythmia, and other diagnoses (e.g., A-V block, evidence of a prior myocardial infarction). Evaluation of the peripheral vascular system was based on diastolic blood pressure, funduscopic examination, auscultation of the carotid arteries, and determination of the quality of five peripheral pulses. The presence of carotid bruits was recorded in both carotid arteries. The femoral, popliteal, dorsalis pedis, posterior tibial, and radial pulses were assessed both by manual palpation and Doppler techniques because of the significant group differences discovered at the Baseline examination. Doppler results were considered the "gold standard" for the pulse measurements, although sensitivity correlations were established with palpation results. Rate changes of abnormal pulses occurring since the Baseline examination were also examined. In addition to the above dependent variables, considerable analytical attention was directed to the cardiovascular risk factors of age, race, occupation (OCC), and updated values for smoking history (pack-years [PACKYR], and current smoking level [CSMOK]), alcohol history (drink-years [DRKYR], and current drinking level [ALC]), cholesterol (CHOL), HDL, cholesterol-HDL ratio (CHOL/HDL), percent body fat (%BFAT), personality score (PS), and differential cortisol response (DIFCORT). 15-3 �Individuals with a verified history of diabetes (or those with an elevated 2-hour postprandial glucose level) were excluded from all analyses except the morbidity-mortality analysis. In addition, individuals with peripheral edema were excluded from analyses of the manual peripheral pulses because of the difficulty of measuring the pulse in the presence of edema. Logistic regression models were used for dichotomous variables, and general linear models for continuous variables. All covariates except race and occupation were treated as continuous variables. Due to the large number of covariates, analyses were carried out as follows. Models adjusting only for age, race, and occupation were examined first, followed by models incorporating group (GRP)-by-age, group-by-race, and group-by-occupation interactions. Analyses were then performed, adjusting for (1) all covariates and (2) all covariates, but with only one variable selected from among each of the sets: pack-years of smoking, current smoking; cholesterol, HDL, cholesterol-HDL ratio; and drink-years of alcohol, current alcohol intensity. Selection of the covariate from each set was based on examination of the pairwise covariate-by-dependent variable associations and the coefficient from the fully adjusted model. Stepwise modeling was then conducted using all covariates, but with only one variable selected from each of the sets described above. Only group-bycovariate interactions were examined, as were the three-factor interactions of group-by-age-by-race, group-by-age-by-occupation, and group-by-race-byoccupation. "Best models" refer to the models including only the statistically significant covariate and interaction terms. Minor numeric disparities in the tables that follow reflect missing dependent variable or covariate data. Parallel analyses using Original Comparisons can be found in Tables M-12 through M-20 of Appendix M. Morbidity and mortality data on the full Ranch Hand cohort and an appropriate Comparison cohort were tabulated for four endpoints: (1) death (any cause) or verified nonfatal heart disease, (2) death (any cause) or verified nonfatal myocardial infarction, (3) fatal or nonfatal verified heart disease, and (4) fatal or nonfatal verified myocardial infarction or fatal heart disease. This analysis involved a number of assumptions, particularly with respect to missing histories in the noncompliant study subjects. RESULTS AND DISCUSSION Questionnaire Data; Reported and Verified Heart Disease For each participant, a cardiovascular disease history was obtained from both the questionnaire and physical examination review of systems history. The baseline and third-year followup data were merged to determine, for each participant completing the third-year followup examination, whether there was ever a reported history of cardiovascular disease following service in Vietnam. Reported conditions were verified by medical record reviews and classified according to the ICD-9-CM. The following three variables were analyzed in terms of both reported and verified events: Variables ICD-9CM Codes Essential Hypertension 401 Heart Disease (Excluding Essential 391, 393-398, 402, 404 Hypertension) 410-414, 415-417, 420-429 Acute Myocardial Infarction 410 15-4 �Table 15-1 gives the unadjusted analysis of reported and verified cardiovascular disease in the Ranch Hand and Comparison groups and the results of unadjusted group contrasts. Essential hypertension was reported in slightly over 25 percent of the participants, with rates not significantly different in the two groups (p=0.596). About 80 percent of these cases were verified, leaving similar rates of 20.7 and 20.2 percent in the Ranch Hand and Comparison groups, respectively, for verified essential hypertension. Reported heart disease was a little higher in the Ranch Hand group (28.1% vs. 26.1%) but the difference in the percentage of verified heart disease was of borderline significance (23.8% vs. 20.3%, p=0.054). The rates of reported and verified myocardial infarctions were about 2 percent and 1 percent, respectively, and not significantly different in the two groups. The associations between each of the covariates and the three verified cardiovascular endpoints are presented in Tables 15-2, 15-3, and 15-4. The tables containing the covariate associations with the reported cardiovascular diseases are included in Tables M-l through M-3 of Appendix M. All reported cardiac illnesses (verified and unverified) are included in these tables. Many of the classic risk factors were identified. Age, smoking, cholesterol and/or cholesterol-HDL ratio, percent body fat, differential cortisol, and alcohol use were significantly associated with reported and verified essential hypertension, although the smoking effect was in the opposite direction of that expected. Age, occupation, and the cholesterol-HDL ratio were significantly associated with reported and verified heart disease, with more disease found in officers than in enlisted personnel. Age, pack-years of smoking, cholesterol-HDL ratio, and drink-years of alcohol were significantly associated with reported and/or verified myocardial infarction (the smoking effect being in the expected direction). The results of logistic regression analyses adjusting for these variables are presented in Table 15-5. The results were similar to the unadjusted results, but the adjusted ralative risk for verified heart disease reached statistical significance (p=0.036). No significant group-bycovariate interactions were noted. Nearly identical results were obtained in the analysis of the Ranch Hands and Original Comparisons (see Tables M-12 and M-13 of Appendix M). Morbidity-Mortality Analysis Differential mortality in the two groups could introduce bias in the analysis of morbidity data. For the cardiovascular evaluation, morbidity and mortality data on all Ranch Hands (diabetics included) and the first Comparison of the randomly ordered set matched to the Ranch Hands were combined to estimate the frequency of four hierarchical cardiovascular endpoints. Because of competing mortality and possible misclassification of the cause of death, the endpoints of death (any cause) or verified nonfatal heart disease, and death (any cause) or verified nonfatal myocardial infarction were examined to assess group differences in the most extreme case (i.e., all deaths being associated with cardiovascular disease). The other two endpoints were limited to fatal or nonfatal verified heart disease, and fatal or nonfatal verified myocardial infarction or fatal heart disease. The analysis was based on 1,257 Ranch Hands and 1,253 Comparisons. The history of each individual from the end of his tour of duty in SEA to the present was reviewed. Histories of verified heart disease and myocardial 15-5 �TABLE 15-1. Unadjusted Analyses for Reported and Verified Heart Disease by Group Group Ranch Hand Variable Statistic Number Percent Comparison Number Percent Est. Relative Risk ( 5 C.I.) 9% p-Value Reported Essential Hypertension n Yes No 942 247 695 26.2 73.8 1,206 304 902 25.2 74.8 1.05 (0.87,1.28) 0.596 Verified Essential Hypertension n Yes No 942 195 747 20.7 79.3 1,206 244 962 20.2 79.8 1.03 (0.83,1.27) 0.787 Reported Heart Disease (Excluding Hypertension) n Yes No 942 265 677 28.1 71.9 1,206 315 891 26.1 73.9 1.11 (0.91,1.34) 0.298 Verified Heart Disease (Excluding Hypertension) n Yes No 942 224 718 23.8 76.2 1,206 245 961 20.3 79.7 1.22 ( . 0 1 5 ) 10,.0 0.054 Reported Myocardial Infarction n Yes No 942 20 922 2.1 97.9 1,206 22 1,184 1.8 98.2 1.17 (0.63,2.15) 0.617 Verified Myocardial Infarction n Yes No 942 9 933 1.0 99.0 1,206 13 1,193 1.1 98.9 0.88 (0.38,2.08) 0.779 Ul a, �TABLE 15-2. Association Between Verified Essential Hypertension and the Covariates in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Percent Abnormal p-Value Age Born >1942 Born <1942 934 1,214 17.2 22.9 0.001 Race Black Nonblack 126 2,022 25.4 20.1 0.191 Occupation Officer Enlisted Flyer Enlisted Groundcrev 807 354 987 21.2 20.1 20.0 0.798 Current Smoking 0 >0 - 20 >20 1,262 463 422 22.8 16.6 17.5 0.005 Pack-Years Smoking 0 >0 - 10 >10 512 760 869 24.8 17.9 20.0 0.010 Cholesterol <200 >200 - 230 >230 766 650 732 15.5 21.1 25.0 <0.001 <40 >40 - 50 >50 719 754 675 21.6 20.6 19.1 0.524 717 743 688 16.2 21.3 24.0 0.001 10 1,758 379 0.0 16.7 38.5 <0.001 HDL Cholesterol-HDL <4.2 Ratio >4.2 - <5.5 >5.5 Percent Body Fat <10 10 - 25 >25 Personality Score <-5 -5-5 >5 829 731 580 22.3 20.5 17.8 0.113 Differential CortjLsol <0.6 >0.6 - 4.0 >4.0 704 745 683 23.6 19.1 18.4 0.033 Current Alcohol Use (Drinks/Day) 0 >0 - 1 >1 592 809 738 21.4 17.2 23.2 0.011 Drink-Years Alcohol <1.25 >1.25 - 25 >25 691 719 666 21.1 18.4 22.8 0.116 15-7 �TABLE 15-3. Association Between Verified Heart Disease and the Covariates in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Percent Abnormal p-Value Age Born >1942 Born <1942 934 1,214 17.9 24.9 <0.001 Race Black Nonblack 126 2,022 23.0 21.8 0.826 Occupation Officer Enlisted Flyer Enlisted Groundcrev 807 354 987 24.8 20.9 19.8 0.034 Current Smoking 0 >0 - 20 >20 1,262 463 422 22.7 21.2 19.9 0.461 Pack-Years Smoking 0 >0 - 10 >10 512 760 869 23.2 20.9 21.9 0.617 Cholesterol <200 >200 - 230 >230 766 650 732 21.2 21.1 23.2 0.533 <40 >40 - 50 >50 719 754 675 21.7 20.4 23.6 0.357 CholesterolHDL Ratio <4.2 >4.2 - <5.5 >5.5 717 743 688 24.1 18.8 22.7 0.041 Percent Body Pat <10 10 - 25 >25 10 1,758 379 30.0 22.1 20.3 0.619 Personality Score <-5 -5-5 >5 829 731 580 21.4 20.9 24.0 0.369 Differential Cortisol <0.6 >0.6 - 4.0 >4.0 704 745 683 19.2 22.4 24.0 0.084 Current Alcohol Use Drinks/Day 0 <1 >0 - 1 592 809 738 23.0 22.5 20.3 0.441 Drink-Years Alcohol <1.25 >1.25 - 25 >25 691 719 666 21.4 22.0 21.8 0.968 HDL 15-8 �TABLE 15-4. Association Between Verified Myocardial Infarction and the Covariates in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Percent Abnormal Age Born XL942 Born <1942 934 • 1,214 0.2 1.6 0.002 Race Black Nonblack 126 2,022 0.0 1.1 0.471 Occupation Officer Enlisted Flyer Enlisted Groundcrev 807 354 987 0.9 1.4 1.0 0.697 p-Value Current Smoking 0 >0 - 20 >20 1,262 463 422 0.8 1.7 1.0 0.228 Pack-years Smoking 0 >0 - 10 >10 512 760 869 0.2 0.8 1.7 0.018 Cholesterol <200 >200 - 230 >230 766 650 732 0.5 0.9 1.6 0.095 <40 >40 - 50 >50 719 754 675 1.5 0.9 0.6 0.210 CholesterolHDL Ratio <4.2 >4.2 - <5.5 >5.5 717 743 688 0.4 0.9 1.7 0.046 Percent Body Fat <10 10 - 25 >25 10 1,758 379 0.0 1.0 0.8 0.872 Personality Score <-5 -5-5 >5 829 731 580 0.8 1.5 0.7 0.278 Differential Cortisol <0.6 >0.6 - 4.0 >4.0 704 745 683 1.0 1.1 1.0 0.989 Current Alcohol Use (Drinks/Day) 0 >0 - 1 >1 592 809 738 1.5 0.9 0.8 0.376 Drink-Years Alcohol SI. 25 >1.25 - 25 >25 691 719 666 1.3 0.4 1.4 0.143 HDL 15-9 �TABLE 15-5. Adjusted Analyses for Reported and Verified Heart Disease Adj. Relative Risk (95% C.I.) 'p-Value Reported Essential Hypertension 1.14 (0.93,1.41) 0.211 AGE (p<0.001), CSMOK (p=0.001),CHOL (p<0.001), %BFAT (p<0.001), ALC (p<0.001) Verified Essential Hypertension 1.11 (0.89,1-39) 0.347 AGE (p=0.021), CSMOK (p=0.021), CHOL (p<0.001), %BFAT (p<0.001), PS (p=0.039) Reported Heart Disease 1.12 (0.92,1.36) 0.258 AGE (p<0.001) Verified Heart Disease 1.25 (1.02,1.54) 0.036 AGE (p<0.001) Reported Myocardial Infarction 1.16 (0.60,2.23) 0.667 AGE (p<0.001),OCC (p=0.014), CHOL/HDL (p=.0.016) Verified Myocardial Infarction 0.93 (0.38,2.23) 0.865 AGE (p<0.001), CHOL/HDL (p=0.025) Variable ^Abbreviations; CSMOK: CHOLs %BFAT: ALC: PS: OCC: CHOL/HDL: Current smoking Cholesterol Percent body fat Current alcohol use (drinks/day) Personality score Occupation Cholesterol-HDL ratio 15-10 Covariate Remarks* �infarction for living individuals who were noncompliant at Baseline and at the followup were missing. For the living noncompliant individuals, the observed rate in the compliant individuals was used to estimate the number of nonfatal events among the noncompliant individuals for each cohort. It was assumed that there were no nonfatal cardiovascular events in the noncompliant individuals who died due to a cause other than cardiovascular system failure. The results are shown in Table M-4 of Appendix M. There was a total of 66 deaths in the Ranch Hand group and 77 in the group of Comparisons. The estimated percentage of Ranch Hands who died (any cause) or had a verified nonfatal history of heart disease was 27.4 as contrasted to 24.5 in the Comparisons. The rate of verified nonfatal myocardial infarctions was approximately 1 percent in each group. The estimated percentage of deaths (any cause) or verified nonfatal myocardial infarction was 6.4 percent in the Ranch Hands and 7.0 percent in the Comparisons. Only 5 of the 66 deaths in the Ranch Hands and 3 of the 77 deaths in the Comparisons either were from heart disease, or were individuals who had verified heart disease histories. The estimated percentage of fatal and nonfatal verified heart disease was 22.5 percent in the Ranch Hands and 18.6 percent in the Comparisons. Of the 66 deaths in the Ranch Hands only 1 individual died from cardiovascular disease or had a verified history of myocardial infarction as compared to 2 of the 77 deaths in the Comparisons. The estimated percentage of fatal or nonfatal verified myocardial infarction or fatal heart disease was 1.2 percent in the Ranch Hands and 1.0 percent in the Comparisons. These contrasts must be interpreted guardedly since they involve some unverifiable assumptions. Nevertheless, they are consistent with the morbidity findings presented in the chapter, and tend to show that the clinical cardiovascular disease spectrum is approximately equal in both groups. Physical Examination Data Central Cardiac Function Central cardiac function was assessed by the measurement of systolic blood pressure, heart sounds, and an ECG. Systolic blood pressure was determined by a standardized sphygmometer, at the appearance of the first sound with the ndndominant arm placed at heart level; the lowest value of three readings was recorded. Detection of abnormal heart sounds was conducted by standard auscultation with the participant placed in sitting, supine, and left lateral supine positions. Fourth heart sounds were assessed; murmurs were graded in intensity and location and were judged to be functional (normal) or organic (abnormal) in nature. Fourth heart sounds were scored as abnormal. ECG data were collected by a standardized 12-lead machine; approximately 95 percent of the clinical interpretations were performed by one cardiologist. All participants were asked to abstain from smoking for at least 4 hours prior to their ECG. 15-11 �Systolic Blood Pressure Systolic blood pressure was analyzed both as continuous and dichotomized variables (normal, 140 or less mm Hg; abnormal, more than 140 mm Hg). Combined distributional data from both groups revealed significant digit preference for values ending in zero (p<0.0001 for both systolic and diastolic readings), but standard statistical analyses were performed since the zero-digit peaks (e.g., 130, 140, 150 mm Hg) were relatively uniform and did not visually differ between the Ranch Hand and Comparison groups. Zero digit readings were recorded for 59.4 percent of the systolic blood pressures and 55.0 percent of the diastolic blood pressures. Table 15-6 gives the percentage of participants with abnormally high systolic values. The percent of abnormals was not significantly different from each other (p=0.529). Systolic blood pressure, analyzed as a continuous variable, had a mean of 118.96 mm Hg (95% C.I.: [118.06,119.86]) for the Ranch Hand group and a mean of 119.55 mm Hg (95% C.I.: [118.71,120.39]) for the Comparison group. These means were not significantly different (p=0.349). The means were also not significantly different when Original Comparisons were used (p=0.182). The association between each of the covariates (categorized into either two or three levels) and dichotomized systolic blood pressure in the combined Ranch Hand and Comparison groups is shown in Table 15-7. Age, cholesterol, percent body fat, personality score, and alcohol use (both current use and drink-years) were significantly associated with increased systolic pressure. These covariate effects were in the direction typically found in other studies,2 ' except for personality score where those participants with low scores (in the Type B direction) had the highest percentage of abnormal values. Adjustment of the categorical systolic blood pressure by the above covariates was performed by logistic regression analysis, and these results are presented in Table 15-8. As shown, there were no significant differences between the Ranch Hand and Comparison groups (p=0.920). Age, cholesterol, percent body fat, personality score, and current alcohol use all had statistically significant effects. An adjusted analysis of systolic blood pressure in the continuous form revealed a significant group (GRP)-by-ageby-race interaction (p=0.012) along with the significant main effects of current smoking (p<0.001), cholesterol (p<0.001), percent body fat (p<0.001), personality score (p<0.001), and current alcohol use (p=0.002). Exploration of the interaction revealed that among Blacks there was a group-by-age interaction (p=0.007), with a mean systolic pressure greater in the Ranch Hand group than in the Comparison group at the younger age levels, but lower at the older age levels. The estimated Ranch Hand-Comparison difference was 4.56 (± 3.30) mm Hg at the Baseline age of 35 and -16.01 (± 5.87) mm Hg at the Baseline age of 53 (see Table M-5 of Appendix M). In the nonblack cohort the group-by-age interaction was not significant (p=0.338), nor was there evidence of any overall group effect (p=0.356). In the analysis of the Ranch Hands and Original Comparisons, there were no statistically significant group differences, either unadjusted or adjusted for covariate effects (see Tables M-14 and M-15 of Appendix M). 15-12 �TABLE 15-6. Unadjusted Analyses for Central Cardiac Function By Group (Diabetics Excluded) Group Ranch Hand Statistic Systolic Blood Pressure n Abnormal Normal 942 60 882 6.4 93.6 1,205 85 1,120 7.1 92.9 0.90 (0.64, 1.26) 0.529 Heart Sounds n Abnormal Normal 941 31 910 3.3 96.7 1,206 32 1174 2.7 97.3 1.25 (0.76,2.06) 0.384 ECG (Overall) n Abnormal Normal 947 121 821 12.8 87.2 1,206 169 1,037 14.0 86.0 0.90 (0.70,1.16) 0.430 ECG: n' Abnormal Normal 942 5 937 0.5 99.5 1,206 9 1,197 0.7 99.3 0.71 (0.24,2.13) 0.542 n Abnormal Normal 942 0 942 0.0 100.0 1,206 0 1,206 0.0 100.0 n Abnormal Normal 942 85 857 9.0 91.0 1,206 107 1,099 8.9 91.1 i Percent Number Est. Relative Risk (95% C.I.) Variable Ul Number Comparison Percent p-Value u> ECG: RBBB LBBB ECG: Nonspecific T-¥ave Changes — 1.02 (0.76,1.37) 0.904 �TABLE 15-6. (continued) Unadjusted Analyses for Central Cardiac Function By Group (Diabetics Excluded) Group Ranch Hand n Abnormal Normal 942 45 897 4.8 95.2 1,206 56 1,150 4.6 95.4 n Abnormal Normal 942 0 942 0.0 100.0 1,206 0 1,206 0.0 100.0 n Abnormal Normal 942 31 911 3.3 96.7 1,206 41 1,165 3.4 96.6 0.97 (0.60,1.55) 0.889 n Abnormal Normal 942 97 845 10.3 89.7 1,206 132 1,074 11.0 89.0 0.93 (0.71,1.23) 0.631 ECG: Tachycardia ECG: Arrhythmia ECG: Other Diagnoses —No relative risk given, since no abnormals are present. Percent Est. Relative Risk (95% C.I.) Statistic ECG: Bradycardia I Percent Number Variable Ui Number Comparison 1.03 (0.69,1.54) p-Value 0.889 — �T6BUS15-7. Association Between Central Cardiac Flncticn Variables and the Govariates in the Gombinad Ranch Bawl and Comparison GEOUIIS (Diabetics Excluded) Current Pack-Years Race Occupation Snaking Smoking Percent Cholesterol- Body Personality Cholesterol HDL HDL Ratio Fat Score Differential Cortisol Current Alcohol Use (Drinks Drink-Years per Day) Alcohol Variable Systolic Blood Pressure 00.001 NS NS NS NS 00.001 NS NS 0.001 0.002 NS 0.018 00.001 Heart Sounds 0.005 NS NS* NS NS NS NS* 0.003 NS NS NS NS NS BOG (Overall) f t I K Age NS NS* NS 0.010 NS* NS* 0.016 00.001 NS NS NS NS N S NS NS NS BOG: RBBB 00.001 N S N S N S N S N S N S N S N S N S BOG: 00.001 Nonspecific T-Wave Changes NS NS NS 0.038 0.002 NS <D.001 00.001 NS NS NS 0.006 BOG: Bradycardia NS* NS 0.010 NS 0.007 NS* 0.002 00.001 NS NS NS NS, NS BOG: Arrhythmia NS NS 0.023 NS 0.028 NS NS NS NS NS NS NS NS BOG: Other Diagnoses 00.001 NS 0.011 NS* 0.023 NS NS NS NS NS NS 0.019 NS NS: Not significant (pX).10). NS*: Borderline significant (0.05 <p<D.10). �TABLE 15-8. Adjusted Analyses for Central Cardiac Function (Diabetics Excluded)* Variable Systolic Blood Pressure (Discrete) Systolic Blood Pressure (Continuous) Heart Sounds Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks* 0.98 (0.69,1.40) 0.920 AGE (p<0.001) CHOL (p=0.004) 2BFAT (p<0.001) PS (p=0.002) ALC (p=0.020) **** **** GRP*RACE*AGE (p=0.012) CSMOK (p<0.001) CHOL (p<0.001) XBFAT (p<0.001) PS (p<0.001) ALC (p=0.002) 1.33 (0.80,2.24) 0.276 ECG (Overall) **** **** ECG: RBBB 0.72 (0.24,2.15) 0.555 AGE (p»0.008) ECG: Nonspecific ST-T-Wave Changes 1.12 (0.81,1.53) 0.497 AGE (p<0.001) RACE (p=0.005) CHOL (p=0.007) %BFAT (p<0.001) 15-16 AGE (p<0.001) RACE (p=0.003) CHOL/HDL (p=0.002) AGE (p<0.001) RACE (p=0.005) %BFAT (p<0.001) GRP*PACKYR (p=0.008) �TABLE 15-8. (continued) Adjusted Analyses for Central Cardiac Function (Diabetics Excluded)* Variable Adj. Relative Risk (95% C.I.) p-Value ECG: Bradycardia 1.08 (0.72,1-62) 0.726 ECG: Arrhythmia **** **** ECG: Other Diagnoses 0.92 (0.69,1.23) 0.575 Covariate Remarks** OCC (p=0.047) CHOL/HDL (p<0.001) AGE (p=0.001) OCC (p<0.001) GRP*PACKYR (p=0.018) GRP*%BFAT (p=0.038) AGE (p<0.001) RACE (p-0.015) CSMOK (p=0.039) *Some adjusted analyses did not explore effects of all covariates due to sparse number of abnormalities (see text). **Additional Abbreviations; GRP: group PACKYR: pack-years smoking ****Group-by-covariate interaction, relative risk/difference in group means, 95% confidence interval, and p-value not presented (see Table M-5 of Appendix M). 15-17 �Heart Sounds As shown in Table 15-6, the unadjusted frequency of abnormal heart sounds in the two groups was not significantly different (p=0.384). The covariate tests of association (Table 15-7) showed significant effects for age (p=0.005), cholesterol-HDL ratio (p=0.003), and a borderline association with occupation (p=0.069). Increased age (born before 1942)'had a frequency of 3.9 percent heart sound abnormalities as contrasted to 1.6 percent abnormalities in the younger age group (born in or after 1942). The cholesterol-HDL ratio (less than or equal to 4.2, between 4.2 and 5.5, and greater than or equal to 5.5) was positively associated with increasing frequencies of abnormal heart sounds (1.7, 2.6, and 4.7 percent, respectively). The observed frequencies of abnormal heart sounds were 3.8, 1.4, and 2.7 percent in the officers, enlisted flyers, and the enlisted groundcrew, respectively. The adjusted analysis (Table 15-8) did not detect any significant group differences (p=0.276). Age, race, and the cholesterol-HDL ratio were significant covariates (p<0.001, p=0.003, and p=0.002, respectively). No two- or three-way group interactions were noted. Similarly, nonsignificant results were found in the analyses of the Original Comparisons versus the Ranch Hands (see Table M-15 of Appendix M). Electrocardiograph Findings All EGG tracings were scored as normal or abnormal; specific abnormalities included RBBB, LBBB, nonspecific T-wave changes, bradycardia, tachycardia, arrhythmia, and other diagnoses. The unadjusted analysis of these variables (Table 15-6) showed no statistically significant differences in the overall ECG results, or any of the specific subcategories, between the Ranch Hand and Comparison groups. Two additional findings in the analysis were of interest: (1) the Ranch Hands had a uniformly lower number of ECG abnormalities than the Comparisons (though not statistically significant), and (2) the sum of the specific ECG findings exceeds the proportion of abnormalities scored on the overall ECG because some individuals accounted for two or more abnormalities. The associations between the covariates and the various ECG findings are presented in Table 15-7. Age was significantly associated with the overall ECG findings (p<0.001), nonspecific T-wave changes (p<0.001), and other ECG diagnoses (p<0.001), with more abnormalities found in the older age group. Occupation was significantly associated with bradycardia (p=0.010), arrhythmia (p=0.023), and other ECG findings (p=0.011). A higher percentage of officers than enlisted flyers or groundcrew had bradycardia, whereas enlisted flyers had the lowest proportion of arrhythmias, and enlisted groundcrew had the highest percentage. Officers and enlisted flyers had a higher percentage than the enlisted groundcrew cohort of other ECG findings. Pack-years of smoking was significantly associated with the overall ECG findings (p=0.010), T-wave changes (p=0.038), bradycardia (p=0.007), arrhythmia (p=0.028), and other ECG diagnoses (p=0.023). For the overall ECG findings, nonspecific T-wave changes, and arrhythmias, the moderate smoking group (greater than 0 to 10 pack-years) had the fewest abnormalities. 15-18 �Bradycardia was negatively associated with pack-years of smoking, with the highest frequency of abnormalities (7.2%) found in the 0 pack-years category versus the lowest proportion (3.6%) of abnormalities in the greater than 10 pack-years category. Cholesterol levels and/or the cholesterol-HDL ratio were positively associated with abnormalities in the overall ECG (p=0.016) and T-wave findings (p<0.001), but were negatively associated with bradycardia (p<0.001). Increased percent body fat was significantly associated with overall ECG abnormalities (p<0.001) and nonspecific T-wave changes (p<0.001). Drinkyears of alcohol was only associated with T-wave changes (p=0.006), with more abnormalities in the greater than 25 drink-years category than in the less than or equal to 1.25 drink-years category, but relatively fewer abnormalities in the more than 1.25 to 25 drink-years category. The covariate of current alcohol use was associated only with the category of other ECG diagnoses (p=0.019), but not in a consistent manner (individuals averaging less than one drink per day had more abnormalities than nondrinkers, but those averaging more than one drink per day had the lowest percentage of abnormalities). The covariates of race, current smoking, personality score, and differential cortisol level, however, did not significantly affect the variables of central cardiac function. Results from the adjusted logistic regression analyses are shown in Table 15-8, No significant group differences were detected for categorical RBBB, T-wave changes, bradycardia, and other ECG diagnoses. The covariates of age, race, percent body fat, pack-years of smoking, current smoking, cholesterol, and cholesterol-HDL ratio were significantly associated with one or more of the ECG variables. RBBB was adjusted only for age due to the small number of abnormalities. The adjusted analysis of the overall ECG findings revealed a significant group-by-pack-year interaction (p=0.008), and the analysis of the arrhythmia variable disclosed two significant interactions: a group-by-pack-year association (p=0.018) and a group-by-percent body fat association (p=0.038). All of these interactions are displayed in Table M-5 of Appendix M. In the case of the overall ECG findings, the adjusted relative risk among nonsmokers was significantly less than one (p=0.038), i.e., a lower risk for Ranch Hands than Comparisons. For heavy smokers (30 pack-years), the adjusted relative risk was 1.25 (95% C.I.: [0.89,1.76], p=0.197). For cardiac arrhythmias, exploration of the group-by-pack-year interaction at the approximate mean percent body fat of 21 percent showed a borderline significant relationship favoring the nonsmoking Ranch Hands (Adj. RR: 0.58, 95% C.I.: [0.30,1.10], p=0.093); heavy smoking Ranch Hands had a higher proportion of arrhythmias than heavy smoking Comparisons, but this association was not statistically significant (p=0.162). For the group-by-percent body fat interaction, 10 percent and 30 percent body fat levels were analyzed at the approximate median of 7 pack-years of smoking. The adjusted relative risk of "0.23 (95% C.I.: [0.07,0.78]) was statistically significant for the 10 percent body fat category (p=0.018), indicating a lower adjusted frequency of cardiac arrhythmias for nonobese Ranch Hands than for nonobese Comparisons, This situation was reversed for obese Ranch Hands, but the association was not statistically significant (RR: 1.88, 95% C.I.: [0.66,5.34], p=0.234). The adjusted analyses using the Original Comparisons were nearly identical to the analyses of the total Comparison group, including the three 15-19 �group interactions for overall EGG findings and cardiac arrhythmias described above. The analyses of the Original Comparison group are found in Tables M-15 and M-16 of Appendix M. Peripheral Vascular Function Peripheral vascular function was assessed by the diastolic blood pressure, funduscopic examination of small vessels, the presence or absence of carotid bruits,, and both manual palpation and Doppler bilateral measurements of the radial, femoral, popliteal, dorsalis pedis, and posterior tibial pulses. Individual peripheral pulses were combined to form overall indices of peripheral vascular status. Diastolic blood pressure was measured by the standard auscultatory technique, and was recorded at the pressure level corresponding to the disappearance of sound. The funduscopic examination was conducted with undilated pupils in a standard manner, with emphasis placed upon the detection of arterio-venous nicking, hemorrhages, exudate, and papilledema. Carotid bruits were assessed by standard bilateral auscultation; confirmation of bruits was not attempted by the Doppler technique. Manual pulse determinations were performed by the examining physician, independent of the Doppler measurements performed by qualified technicians. Tobacco abstinence for at least four hours was required for the Doppler examination, but not for the manual palpation. Only the physician diagnostician had access to both sets of pulse data. Diastolic Blood Pressure Diastolic blood pressure was analyzed as a continuous variable and as a dichotomized variable (normal value less than or equal to 90 mm Hg; abnormal value greater than 90 mm Hg). As with the systolic readings, a significant zero digit preference was noted for the diastolic blood pressure values. Table 15-9 arrays the results of the unadjusted categorical analyses. As shown, there are no statistically significant group differences for the proportions of diastolic abnormalities (p=0.999). Diastolic blood pressure, analyzed as a continuous variable, had a mean of 79.76 mm Hg (95% C.I.: [71.97, 80.35]) for the Ranch Hand group and a mean of 79.77 mm Hg (95% C.I.: [79.24, 80.30]) for the Comparison Group. These means were not significantly different (p=0.986). The means were also not significantly different when Original Comparisons were used (p=0.555). The tests of covariate association with diastolic blood pressure are given in Table 15-10. Cholesterol, cholesterol-HDL ratio, percent body fat, differential cortisol, and current alcohol use were significantly related to diastolic blood pressure (p<0.001, p=0.006, p<0.001, p=0.041, and p=O.OU, respectively). For increasing cholesterol, cholesterol-HDL ratio, and percent body fat, increases in proportions of abnormal diastolic blood pressure were obtained, whereas for increasing differential cortisol values, a decline in blood pressure abnormalities was found. Current alcohol use (drinks per day) revealed an inconsistent association with diastolic blood pressure abnormalities, with nondrinkers having a higher proportion of abnormalities than low-level drinkers, but a lower proportion of abnormalities than moderate drinkers (8.3, 6.4, and 10.6 percent abnormalities, respectively). The covariates of age, race, occupation, current smoking, pack-years of smoking, HDL, personality score, and drink-years of alcohol were not associated with diastolic blood pressure abnormalities. 15-20 �TABLE 15-9. Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) Group Ranch Hand Variable Statistic Number Percent Comparison Number Percent Est. Relative Risk (95% C.I.) p-Value Diastolic Blood Pressure n Abnormal Normal 942 79 863 8.4 91.6 1,204 101 1,103 8.4 91.6 1.00 (0.74,1.36) 0.999 Funduscopic Examination n Abnormal Normal 941 7 934 0.7 99.3 1,206 6 1,200 0.5 99.5 1.50 (0.50,4.47) 0.472 Carotid Bruits n Abnormal Normal 941 7 934 0.7 99.3 1,205 7 1,198 0.6 99.4 1.28 (0.45,3.66) 0.646 Radial Pulses (Manual) n Abnormal Normal 929 4 925 0.4 99.6 1,191 8 1,183 0.7 99.3 0.64 (0.19,2.13) 0.465 Radial Pulses (Doppler) n Abnormal Normal 942 3 939 0.3 99.7 1,203 4 1,199 0.3 99.7 0.96 (0.21,4.30) 0.952 Femoral Pulses (Manual) n Abnormal Normal 929 20 909 2.2 97.8 1,191 25 1,166 2.1 97.9 1.03 (0.57,1.86) 0.932 ,_, In 1 £ • �TABLE 15-9. (continued) Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) Group Ranch Hand Variable Statistic Comparison Number Percent Number Percent Est. Relative Risk (95% C.I.) p-Value Femoral Pulses (Doppler) Ln I n Abnormal Normal 942 6 936 0.6 99.4 1,205 4 1,201 0.3 99 .7 1.92 (0.54,6.82) 0.312 Popliteal Pulses (Manual) n Abnormal Normal 929 16 913 1.7 98.3 1,191 28 163 1, 2.4 97 .6 0.73 (0.39,1.35) 0.317 Popliteal Pulses (Doppler) n Abnormal Normal 942 10 932 1. 1 98.9 1,204 8 1,196 0.7 99 •3 1.60 (0.63,4.08) 0.322 Dorsalis n Pedis Pulses Abnormal (Manual) Normal 929 102 827 11.0 89.0 1,191 127 1,064 10 ,7 89 .3 1.03 (0.78,1.36) 0.818 n Dorsalis Pedis Pulses Abnormal (Doppler) Normal 938 228 710 24.3 75.7 1,202 274 928 22 .8 77 .2 1.09 (0.89,1.33) 0.412 Posterior n Tibial Pulses Abnormal (Manual) Normal 929 27 902 2.9 97.1 1,191 31 160 1, 2.6 97 .4 1.12 (0.66,1.89) 0.674 Posterior n Tibial Pulses Abnormal (Doppler) Normal 939 19 920 2.0 98.0 1,202 25 1,177 2.1 97 .9 0.97 (0.53,1.78) 0.928 �TABLE 15-9. (continued) Unadjusted Analyses for Peripheral Vascular Function by Group, (Diabetics Excluded) Group Ranch Hand Variable Statistic Number Percent Comparison Number Percent Est. Relative Risk (95% C.I.) p-Value Leg Pulses (Manual) 1_l n Abnormal Normal 929 131 798 14.1 85.9 1,191 177 1,014 14.9 85.1 0.94 (0.74,1.20) 0.624 Leg Pulses (Doppler) n Abnormal Normal 938 237 701 25.3 74.7 1,202 288 914 24.0 76.0 1.07 (0.88,1.31) 0.490 Peripheral Pulses (Manual) n Abnormal Normal 929 133 796 14.3 85.7 1,191 181 1,010 15.2 84.8 0.93 (0.73,1.18) 0.575 Peripheral Pulses (Doppler) n Abnormal Normal 938 239 699 25.5 74.5 1,202 290 912 24.1 75.9 1.08 (0.88,1.31) 0.472 All Pulses (Manual) n Abnormal Normal 929 133 796 14.3 85.7 1,191 182 1,009 15.3 84.7 0.93 (0.73,1.18) 0.535 All Pulses (Doppler) n Abnormal Normal 938 239 699 25.5 74.5 1,201 291 910 24.2 75.8 1.07 (0.88,1.30) 0.509 Ol i £ �TABLE 15-10. in the Combined Ranch Band and Comparison Groups (Diabetics Excluded) Variable Diastolic Blood Pressure NS NS Funduscopic Examination Carotid Bruits Ln Age Current Pack-Years Race Occupation inking Snaking Percent Cholesterol- Body Personality Differential Cortisol Cholesterol HDL HDL Ratio Fat Score Current Alcohol Use (Drinks Drink-Years per Day) Alcohol NS 0.006 O.C01 NS 0.041 0.014 NS NS 0.016 0.026 NS NS 0.004 NS NS* NS NS NS NS NS NS 0.021 NS NS NS 0.033 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS* 0.002 O.001 0.001 NS NS NS <D.001 NS NS 0.001 0.006 NS* NS 0.019 NS NS NS NS 0.013 NS NS NS 0.001 NS NS 0.002 NS NS NS NS NS* 0.002 NS NS O.001 0.010 NS NS NS* NS NS NS NS NS NS NS NS 0.004 0.040 NS NS NS MS NS* NS NS NS* NS Radial Pulses NS (Manual) NS NS NS Radial Pulses NS (Doppler) NS NS Femoral Pulses (Manual) 40.001 NS Femoral Pulses (Doppler) NS* Popliteal Pulses (Manual) NS* Popliteal Pulses (Doppler) O.001 �TABLE 15-10. (contiiued) Association Between Peripheral Vascular Function feriables and the Covariates in tie Combined Band) Hand and Comparison Groins (Diabetics Excluded) Current Pack-Years Race Occupation Soaking Snaking Percent Current Alcohol Cholesterol- Body Personality Differential Use (Drinks Drink-Years Cholesterol BDL BEL Ratio Fat Score Cortisol per Day) Alcohol Variable Dorsal is Pedis Pulses (Manual) ^_t Age 0.018 NS NS* NS NS NS NS NS NS NS NS MS NS Dorsalis Pedis Pulses (Doppler) US* 0.007 NS NS MS MS MS MS MS MS NS NS Posterior 0.034 <0.001 NS <0.001 O.001 NS* NS NS* NS NS NS NS NS Posterior Tibial Pulses (Doppler) 0.003 NS* NS 0.021 NS NS NS 0.035 NS 0.028 NS* NS NS* Leg Pulses (Manual) 0.001 NS NS NS 0.031 NS 0.013 0.013 NS NS NS NS NS NS NS NS NS 0.001 Oi 1 KJ Ui Tibial Pulses (tonal) L e g Pulses (Doppler) Peripheral Pulses (Manual) 0.020 0.009 O.001 NS 0.012 NS N NS S N S N 0.028 NS S N S 0.013 N S 0.010 N N5 S N S NS NS NS �TWE 35-10. (ontiiuBd) in the Ocpihin^ R^vh Hand and Gwpiri.sm Grtmps (Diabetics RKrhpM) Current Pack-Years Race Occupation Sacking Smoking Percent Cholesterol- Body Personality Cholesterol HDL HDL Ratio Fat Score Variable Peripheral Pulses (Doppler) 0.037 0.015 0.019 NS NS N All Pulses (Manual) i Age O.C01NS NS NS 0.023 NS 0.013 All Pulses (Doppler) 0.032 0.014 0.023 NS NS NS NS 1 0 NS: Not significant (pX).10). NS*: Borderline significant (0.05<p<D.10). S N S N Differential Cortisol Current Alcohol Use (Drinks Drink-Years per Day) Alcohol S NS NS NS NS 0.012 NS NS NS NS NS NS NS NS NS NS NS S N �The adjusted categorical and continuous analyses are shown in Table 15-11. No significant group differences were found in the proportions of diastolic abnormalities (p=0.653), or in the difference of group mean values (p=0.299). The covariates of current smoking, cholesterol, percent body fat, and current alcohol use were statistically significant in both the categorical and continuous analyses; age was significant only in the analysis of group mean differences (p=0.005). No significant group-by-covariate interactions were found in either the logistic regression or general linear models. The adjusted analyses for the Original Comparisons were very similar to those described on the total Comparison group (see Table M-18 of Appendix M). Funduscopic Examination The funduscopic examination detected only 13 individuals with arteriovenous nicking (a sign of chronic blood pressure elevation) or vessel hemorrhages, 7 from the Ranch Hands and 6 from the Comparisons (p=0.472, Table 15-9). The covariate tests of association are given in Table 15-10.. Age, race, cholesterol/HDL ratio, percent body fat, and current drinking were statistically significant (p=0.004, 0.040, 0.016, 0.026, and 0.004, respectively). All funduscopic abnormalities were found in the older age group (born in or before 1942). Blacks had a higher proportion of abnormalities than nonblacks (2.4 percent versus 0.5 percent, respectively). The highest cholesterol-HDL category contained the highest proportion of funduscopic abnormalities; and increasing levels of percent body fat were associated with increases in proportions of abnormalities. Current alcohol consumption showed that nondrinkers had the highest proportion of abnormalities. The covariates of occupation, current smoking, pack-years of smoking, cholesterol, HDL, personality score, differential cortisol and drink-years of alcohol did not show significant effects. In the adjusted analysis by logistic regression (Table 15-11), there were no significant differences in funduscopic abnormalities between the Ranch Hand and Comparison groups (Adj. RR: 1.78; 95% C.I.: [0.56,5.62], p=0.322). Due to sparse data the model was adjusted only for the covariates of age, race, cholesterol-HDL ratio, percent body fat, and current alcohol consumption; and all were significant in the model. No group interactions were detected, and the results of the contrast of the Ranch Hand with the Original Comparison group were also nonsignificant (Table M-18 of Appendix M). Carotid Bruits The unadjusted group contrast of carotid bruits is displayed in Table 15-9. The proportions of bruits in both groups were similar (Est. RR: 1.28, 95% C.I.: [0.45,3.66], p=0.646). Overall, only 14 bruits were detected, 7 from each group, limiting the scope of the adjusted analyses. The covariate effects are given in Table 15-10. Age, current smoking, and cholesterol were of borderline statistical significance, whereas drinkyears of alcohol was significantly correlated with carotid bruits (p=0.021), with the greater than 25 drink-years category having the highest proportion. 15-27 �TABLE 15-11. Adjusted Analysis for Peripheral Vascular Function by Group (Diabetics Excluded)* Statistical/ Clinical Analysis Adj. Relative Risk (95* C.I.) p-Value Discrete 1.08 (0.78,1-48) 0.653 CSMOK (p-0. 040) CHOL (p<0.001) XBFAT (p<0.001) ALC (p-0. 008) Continuous 0.38 (-0.34, 1.11)' 0.299* AGE (p.O.005) CSMOK (p<0.001) CHOL (p<0.001) %BFAT (p<0.001) ALC (p.O.002) Funduscopic Examination 1.78 (0.56,5.62) 0.322 AGE (p<0.001) RACE (p<0.001) CHOL/HDL (p.O.017) XBFAT (p.O.037) ALC (p.O.038) Carotid Bruits 1.05 (0.35,3.16) 0.928 AGE.(p.O.024) DRKYR (p<0.001) Variable Diastolic Blood Pressure Covariate Remarks** Radial Pulses Manual Ooppler 0.64 (0.19,2.14) 0.96 (0.21,4.30)* 0.472h 0.952 AGE (p.O.040) Femoral Pulses Manual 1.21 (0.63,2.31) 0.562 AGE (p<0.001) CHOL/HDL (p-0.010) *BFAT (p<0.001) DIFCORT (p.O.002) Doppler 1.74 (0.48,6.31) 0.401 AGE (p.O.001) CSMOK (p.O.001) CHOL/HDL (p-0.042) Manual **** **** AGE (p-0.003) PACKYR (p«0.005) CHOL/HDL (p-0.Oil) GRP*RACE (p.O.038) Doppler 1.50 (0.58,3.91) 0.401 AGE (p<0.001) RACE (p.O.023) CSMOK (p<0.001) Manual **** **** AGE (p-0.004) DRKYR (p-0.038) GRP*OCC (p.O.046) Doppler 1.07 (0.87,1.31) 0.535 AGE (p.O.004) RACE (p.O,006) ZBFAT (p-0.003) Popliteal Pulses Dorsalis Pedis Pulses 15-28 �TABLE 15-11. (continued) Adjusted Analysis for Peripheral Vascular Function by Group (Diabetics Excluded)* Variable Statistical/ Clinical Analysis Adj. Relative Risk (95* C.I.) p-Value Covariate Remarks** AGE (p<0.001) RACE (p<0.001) PACKYR (p-0. 007) GRP*OCC (p-0. 01 7) 0.94 (0.50,1.77) 0.849 AGE (p<0.001) RACE (p.O.002) CSMOK (p.O.007) CHOL/HDL (p.O.015) Manual **** **** AGE (p<0.001) GRP*OCC (p.O.016) GRP**BFAT (p.O.034) 1.06 (0.87,1-30) 0.549 AGE (p-0.001) RACE (p.O.029) XBFAT (p-0.006) Manual **** **** AGE (p<0.001) XBFAT (p.O.018) GRP*OCC (p.O.033) 1.06 (0.87,1.30) 0.562 AGE (p.O.001) ZBFAT (p.O.006) Manual **** **** AGE (p<0.001) ZBFAT (p.O.022) GRP*OCC (p.O.036) Doppler All Pulses **** Ooppler Peripheral Pulses **** Doppler Leg Pulses Manual Doppler Posterior Tibial Pulses 1.06 (0.86,1.29) 0.603 AGE (p.O.001) ZBFAT (p.O.006) *Sonte adjusted analyses did not explore effects of all covariates due to sparse number of abnormalities (see text). **Additional Abbreviations: DRKYRsdrink-years of alcohol aDIFCORT: differential cortisol. "Difference in group means (Ranch Hand-Comparison) and associated p-value given, rather than relative risk, for continuous analysis of dependent variables. Unadjusted for any covariates—same results as for unadjusted analysis. ****Group-by-covariate interaction—relative risk, confidence interval, and p-value not presented (see Table M-6 of Appendix M). 15-29 �The adjusted analysis was performed with only the covariates of age and drink-years of alcohol due to the small number of detected bruits. The results (Table 15-11) demonstrate a lack of significant group differences (Adj. RR: 1.05, 95% C.I.: 0.35, 3.16], p=0.928). Both age and drink-years of alcohol were significant adjusting variables, but no significant group interactions were noted. The results of the Ranch Hand, Original Comparison group contrast was also nonsignificant (see Table M-18 of Appendix M). Peripheral Pulses Five peripheral pulses (radial, femoral, popliteal, dorsalis pedis, and posterior tibial) were analyzed using data assessments from both manual palpation and Doppler recordings. Palpation data from the examining physician were judged abnormal if the pulse was diminished or absent on either side. Assessment of the Doppler data was more complex and involved visual examination of the waveform morphology (pulsatility, systolic forward flow, and diastolic reverse flow) on analog strips and Polaroid® photographs, with careful comparison of the laterality of results. Confirmatory functional data (e.g., treadmill, segmental pressure readings) of abnormal pulses were not performed. The interpretation of each pulse was scored as normal, mild impairment, moderate impairment, severe impairment, or total occlusion (for the purpose of this analysis, all interpretations other than normal were considered abnormal). All Doppler measurements were conducted with a minimum of a 4-hour abstinence from smoking; compliance to the nonsmoking requirement was recorded by the Doppler technician. Besides analysis of each pulse as a distinct dependent variable, three pulse aggregates were prescribed for analysis in order to maintain continuity with the Baseline analysis. The rationale of the pulse aggregates was to localize pulse abnormalities in broad anatomic categories. The aggregates were: leg pulses (femoral, popliteal, dorsalis pedis, and posterior tibial); peripheral pulses (radial, femoral, popliteal, dorsalis pedis, and posterio tibial); and all pulses (peripheral pulses plus carotid pulses, the latter assessed by only manual techniques). Any one abnormal pulse in an aggregate constituted an abnormality for the overall category. The agreement of manual and Doppler assessments was-tested by McNemar's chi-square test using paired data when an individual was compliant to both examination procedures. The paired analyses for the radial, femoral, popliteal, dorsalis pedis, posterior tibial, leg, peripheral, and all pulses are displayed in Table 15-12. As shown, the two methods of pulse assessment differed profoundly (p<0.001) for the femoral, popliteal, dorsalis pedis, leg pulses, peripheral pulses, and all pulses, but only mildly (p=0.044) for the posterior tibial pulse; the methodology differences for the radial pulse were not significantly discordant (p=0.149). Further, as shown by the offdiagonal elements in the specific pulse tables, the manual palpation method classified more cases as abnormal for the femoral, popliteal, and posterior tibial pulses, whereas the Doppler technique detected more abnormalities for the dorsalis pedis pulse, and consequently, the three pulse aggregates. Overall, more credence is given to the Doppler results due to the more "objective" means of determining a pulse abnormality. The unadjusted analyses of all the pulses and pulse aggregates by manual and Doppler techniques (Table 15-9) showed that no statistically significant 15-30 �TABLE 15-12. Agreement Between Manual and Doppler Pulse Assessments (McNemar's )C Test) Radial Femoral DOPPLER 1*lormal DOPPLER Abnormal 2,102 3 9 Normal MANUAL Abnormal 3 Normal Abnormal Normal MANUAL Abnormal X2 = 2,08 p=0.149 Popliteal 2,072 3 38 6 X2 = 28.2 p<0.001 Dorsalis Pedis DOPPLER DOPPLER Normal Abnormal Normal Abnormal 2,067 8 35 Normal MANUAL Abnormal 8 Normal MANUAL Abnormal 341 71 155 X2 = 175.6 X2 - 15.7 p<0.001 Posterior Tibial 1,546 p<0.001 Leg DOPPLER Normal Abnormal Normal MANUAL Abnormal DOPPLER Normal Abnormal 2,035 23 40 16 Normal MANUAL Abnormal 1,462 346 135 170 X2 = 91.7 X2 = 4.1 p=0.044 p<0.001 All Peripheral DOPPLER Normal Abnormal Normal MANUAL Abnormal DOPPLER Normal Abnormal 1,455 347 139 172 Normal MANUAL Abnormal X2 = 88.2 p<0.001 1,454 347 138 173 X2 = 89.2 15-31 p<0.001 �group differences were detected for any pulse or pulse combination by either technique. The covariate tests of association for each pulse and pulse combination by technique are listed in Table 15-10. The following paragraphs describe the results shown in this table. Increased age (born before 1942) was significantly associated with a" higher proportion of pulse abnormalities for the femoral pulses (manual; p<0.001), popliteal pulses (Doppler; p=0.002), dorsalis pedis pulses (manual; p=0.018), posterior tibial pulses (manual, p=0.034; Doppler, p=0.003), leg pulses (manual, p=0.001; Doppler, p=0.020), peripheral pulses (manual, p<0.001; Doppler, p=0.037), and all pulses (manual, p<0.001, Doppler, p=0.032). Age was of borderline significance (0.050<p<0.100) for femoral pulses (Doppler), and for popliteal pulses (manual) and dorsalis pedis pulses (Doppler). Race was associated with dorsalis pedis pulses (Doppler, p=0.001), posterior tibial pulses (manual, p<0.001), leg pulses (Doppler, p=0.009), peripheral pulses (Doppler, p=0.015), and all pulses (Doppler, p»0.014), with Blacks having a lower proportion of abnormalities for the dorsalis pedis, leg, peripheral, and all pulses than nonblacks, but a higher proportion of abnormalities for the posterior tibial pulse. Race was of borderline significance for the Doppler-determined posterior tibial pulses. Occupation was significantly associated with abnormalities of the dorsalis pedis pulses (Doppler, p=0.007), leg pulses (Doppler, p=0.012), peripheral pulses (Doppler, p=0.019), and all pulses (Doppler, p=0.023), with officers uniformly having more abnormalities than enlisted flyers, who had more abnormalities than enlisted groundcrew. Current smoking (cigarettes per day) was significantly associated with increased abnormalities for the posterior tibial pulses (manual, p<0.001; Doppler, p=0.021), femoral pulses (Doppler, p=0.001), and the popliteal pulses (Doppler, p<0.001), despite the 4-hour abstinence prior to the Doppler examination. A relationship of increased smoking and increased abnormalities was only observed for the Doppler determination of the femoral and popliteal pulses. Pack-years of smoking was significantly related to increased abnormalities with popliteal pulses (manual, p=0.001; Doppler, p=0.010), posterior tibial pulses (manual, p<0.001), femoral pulses (Doppler, p=0.006), leg pulses (manual, p=0.031), peripheral pulses (manual, p=0.028), and all pulses (manual, p=0.023). Classical increasing associations were noted for the popliteal pulses (manual and Doppler), the posterior tibial pulses (manual), and the femoral pulses (Doppler). For the related variables involving cholesterol, the cholesterol-HDL ratio showed the most numerous and strongest associations with pulse abnormalities. The cholesterol-HDL ratio was significantly and positively associated with increases in manually determined radial, femoral, and popliteal pulse abnormalities (p=0.033, p<0.001, and p=0.002, respectively); however, other significant associations with all pulses and the leg and .peripheral pulse indices revealed an inconsistent pattern (p=0.012, p=0.013, and p=0.010, respectively). In addition, the ratio was significantly related to femoral and posterior tibial pulse abnormalities, as detected by the Doppler technique (p=0.019, p=0.035, respectively), but the relationships were not uniform from low to high values of the ratio. HDL was significantly associated with manually determined pulse abnormalities for femoral, leg, 15-32 �peripheral, and all pulses (p=0.002, p=0.013, p=0.013, p=0.013, respectively), but in all four cases, the mid-level category of HDL (greater than 40 to 50) was associated with the lowest proportion of abnormalities. Cholesterol showed only marginally significant associations with increased abnormalities of femoral pulse (manual and Doppler) and posterior tibial pulses (manual). Percent body fat was significantly associated with increases of femoral pulse abnormalities (manual, p=0.001); personality score was associated with posterior tibial deficits (Doppler; p=0.028; nonlinear pattern); and drinkyears of alcohol was related to femoral pulse abnormalities detected by both methods (manual, p<0.001; Doppler, p=0.013). Finally, in addition to numerous other marginally significant associations (e.g., drink-years and posterior tibial abnormalities, Doppler, p=0.083; drink-years and popliteal abnormalities, manual, p=0.085), differential cortisol showed a nonlinear association with posterior tibial pulse abnormalities (Doppler, p=0.074). The distribution of each of the covariates in the Ranch Hand and Comparison groups is presented in Table 15-13. As noted, the distributions of the three matching variables, age, race, and occupation, are nearly identical (p=0.987, p=0.745, and p=0.661, respectively). For current smoking, however, Ranch Hands smoke significantly more cigarettes per day (higher mean level) than the Comparisons (p=0.043) a finding also observed at Baseline. Additionally, the difference in mean percent body fat was of borderline significance (p=0.074), with a slightly higher average level in the Comparison group. The results of the adjusted analyses for the manual and Doppler pulse determinations are presented in Table 15-11. Due to the small number of abnormalities, manual radial pulses were adjusted only for age and the cholesterol-HDL ratio, and Doppler radial pulses were not adjusted for any covariates. Similarly, femoral Doppler pulses were adjusted only for age, current smoking, and the cholesterol-HDL ratio. Doppler popliteal pulses were adjusted only for main covariate effects, i.e., interactions were not examined. The adjusted analyses of all Doppler-determined pulse and pulse aggregate abnormalities did not disclose any significant differences between the Ranch Hand and Comparison groups. Age showed a consistent and profound effect in all of the adjusted Doppler analyses, whereas race, percent body • fat, and smoking were significantly influential in about half of the analyses, and the cholesterol-HDL ratio was significant for only two of the pulse variables. The effects of these four covariates were all in the expected (classical) direction. For the manual pulse readings, the adjusted results (Table 15-11) were decidedly different from the Doppler analyses, with all but the radial and femoral pulses involved in significant group-by-covariate interactions. There were no significant group differences for the radial and femoral pulses (p=0.472, p=0.562, respectively). For manually determined popliteal pulses, there was a significant group-by-race interaction (p=0.038), with Blacks having an adjusted relative risk of 6.74 (95% C.I.: [0.72,63.40], p=0.095) in contrast to nonblacks, who had an adjusted relative risk of 0.55 (95% C.I.: 0.28,1.12] p=0.099]). All significant group-by-covariate interactions are shown in Table M-6 of Appendix M. 15-33 �TABLE 15-13. Summary Statistics for Cardiovascular Covariates by Group Covariate Covariate Category Group Ranch Hand Comparison p-Value Percent Percent Black Nonblack 5.6 94.4 6.0 94.0 0.745 Officer Enlisted Flyer Enlisted Groundcrew 37.2 17.3 45.5 37.9 15.8 46.3 0.661 Mean ± SE Mean ± SE Age (At Baseline) 43.57±0.25 43.57±0.22 0.987 Current Smoking* 10.50±0.50 9.19±0.42 0.043 Pack-years Smoking 12.62±0.52 12.51±0.48 0.883 216.8±1.3 218.1±1.2 0.463 46.32±0.42 46.90±0.35 0.288 4.99±0.05 4.92±0.04 0.303 Percent Body Fat 20.85±0.16 21.23±0.14 0.074 Personality Score -1.11±0.30 -1.50±0.26 0.322 Differential Cortisol 2.31±0.13 2.46±0.12 0.398 Current Alcohol Use (Drinks per Day) 1.23±0.07 1.28±0.07 0.611 Drink-years Alcohol 25.62±1.44 22.91±0.96 0.117 Race Occupation Cholesterol HDL Cholesterol-HDL Ratio "Equivalent cigarettes/day. —Covariate not categorized for these results. 15-34 �For the dorsalis pedis, posterior tibial, leg, peripheral, and all pulses, significant group interactions with occupation were detected (p=0.046, p=0.017, p=0.016, p=0.033, and p=0.036, respectively). In all cases, the adjusted relative risk was less than one for the officers and greater than one for the enlisted flyers and groundcrew. In addition, the adjusted relative risk for enlisted flyers was consistently greater than the risk for the enlisted groundcrew. Statistically significant associations by pulse, by occupational category, were as follows: Posterior tibial pulses in enlisted flyer, p=0.032; leg pulses in officers (21% body fat level), p=0.026; peripheral pulses in officers, p=0.030; all pulses in officers, p=0.030. All other pulse-occupational strata contrasts were not statistically significant. As there was also a significant group by percent body fat interaction for leg pulses (p=0.034), each occupational category was analyzed by level of obesity (obese, percent body fat greater than 25 percent; nonobese, percent body fat equal to or less than 25 percent). For officers, the adjusted relative risks were less than one for both the obese (Adj. RR: 0.44, 95% C.I.: [0.17, 1.12], p=0.084) and the nonobese (Adj. RR: 0.66, 95% C.I.: [0.42,1.04], p=0.072). For enlisted flyer personnel, the adjusted relative risks were greater than one for both body fat categories, but were not statistically significant. The enlisted groundcrew manifested an adjusted relative risk of less than 1 for obese individuals (Adj. RR; 0.91, 95% C.I.: [0.39,2.10], p=0.818), and greater than 1 for nonobese individuals (Adj. RR: 1.20, 95% C.I.: [0.79,1.83], p«0.390), but also not statistically significant. The unadjusted analyses of the manual and Doppler pulse assessments (shown in Table M-17 of Appendix M), using the Original Comparisons, did not disclose any significant group differences. For the Doppler adjusted analyses, the results for the Ranch Hand versus Original Comparison contrasts were similar to those found in the Ranch Hand versus total Comparison group, i.e., no statistically significant group differences or group-by-covariate interactions. For the adjusted manual pulse determinations, however, the results differed somewhat from the contrast of the Ranch Hand versus total Comparison group in terms of the significant group-by-covariate interactions detected (see Tables M-18 and M-19 of Appendix M). As before, there were no statistically significant group differences for radial and femoral pulses. For popliteal pulses, however, there was a significant (p=0.048) group-byoccupation interaction, with an adjusted relative risk of less than one for the officers (p=0.219) and greater than one for the enlisted flyers, although not significantly so (p=0.165). For dorsalis pedis pulses, there were no significant group effects or interactions, but for posterior tibial pulses the results were similar to those found in the contrast of the Ranch Hands versus the total Comparison group analysis, i.e., a significant group-byoccupation interaction. For the three pulse aggregates, there were significant group-by-occupation and group-by-percent body fat interactions for the leg pulses (officers having a risk less than one; enlisted flyers and enlisted groundcrew having risks greater than one) and significant group-bypercent body fat interactions for peripheral pulses and all pulses (individuals with low percent body fat having adjusted relative risks greater than one, and obese individuals having an adjusted risk less than one). 15-35 �EXPOSURE INDEX ANALYSES Exposure index analyses were conducted for the Ranch Hand officer, enlisted flyer, and enlisted groundcrew cohorts separately to determine if any dose-response relationships could be identified. In many cases, the data were too sparse to permit statistical comparisons. Adjusted analyses included the exposure level and.only the main effects of age, race, packyears of smoking, cholesterol-HDL ratio, percent body fat, personality score, differential cortisol, and current drinks per day, whenever appropriate. (In several instances, the stepwise logistic modeling did not detect any statistically significant covariate effects. However, adjusted best model results may differ slightly from the unadjusted results due to the omission of individuals with missing covariate information from the adjusted analysis.) Reported and Verified Heart Disease Tabular results of adjusted exposure index analyses for reported and verified heart disease are presented in Table 15-14 (unadjusted exposure index analyses are in Table M-7 of Appendix M). There were no statistically significant differences for reported or verified essential hypertension or reported or verified myocardial infarction by exposure level. (The data on myocardial infarctions were quite sparse.) Results were also negative for reported and verified heart disease, except for the enlisted groundcrew cohort, where the percentage of individuals with reported or verified disease was lowest in the medium exposure category. Central Cardiac Function Table 15-15 gives the adjusted exposure results for systolic blood pressure (dichotomized), heart sounds, and ECG findings. The unadjusted exposure analyses are given in Table M-8 of Appendix M. The only exposure level effect reaching statistical significance was the medium versus low contrast for bradycardia in the enlisted groundcrew (p=0.048), where the adjusted relative risk was significantly less than one. There were borderline significant effects, with adjusted relative risks greater than one for systolic blood pressure (enlisted groundcrew, medium versus low exposure) and T-wave findings (enlisted flyers, medium versus low contrast). There were borderline significant effects, with relative risks less than one for T-wave findings in the enlisted groundcrew cohort, medium versus low exposure (unadjusted only), and high versus low contrast (adjusted only). The results for systolic blood pressure analyzed as a continuous variable showed no statistically significant exposure level effects, either unadjusted or adjusted for covariates. The adjusted medium versus low exposure level contrast was of borderline significance in the enlisted groundcrew (p=0.069). Age, percent body fat, and personality score were significant covariates in one or.more occupational strata. 15-36 �TABLE 15-14. Adjusted Exposure Index Analyses for Reported and Verified Heart Disease by Occupation Significant Covariates Officer Medium vs. Low High vs. Low 0.92 (0.50,1-70) 1.08 (0.58, 2.01) 0.795 0.810 AGE (p=0.016) %BFAT (p<0.001) Enlisted Flyer Medium vs. Low High vs. Low 0.84 (0.34,2.05) 1.34 (0.57,3.16) 0.704 0.509 %BFAT (p=0.002) Enlisted Groundcrew . i Occupation Reported Essential Hypertension OJ Medium vs. Low High vs. Low 1.37 (0.79,2.43) 1.26 (0.68,2.33) 0.289 0.459 ZBFAT (p<0.001) Officer Medium vs. Low High vs. Low 0.94 ( . 8 1 8 ) 04,.4 1.36 (0.70,2.65) 0.849 0.363 XBFAT (p<0.001) Enlisted Flyer Medium vs. Low High vs. Low 0.45 (0.15,1.33) 0.92 ( . 5 2 4 ) 03,.0 0.150 0.865 DIFCORT (p=0.026) Enlisted Groundcrew Medium vs. Low High vs. Low 1.47 (0.82,2.66) 1.33 (0.71,2.49) 0.201 0.379 %BFAT (p<0.001) Verified Essential Hypertension Contrast Adj. Relative Risk (95% C.I.) Variable p-Value 4 Reported Heart Disease Officer Medium vs. Low High vs. Low 0.79 (0.45,1.38) 0.69 (0.39,1.23) 0.407 0.204 AGE (p=0.011) Enlisted Flyer Medium vs. Low High vs. Low 1.30 ( . 8 2 9 ) 05,.4 0.66 (0.27,1.62) 0.529 0.368 NONE Enlisted Groundcrev Medium vs. Low High vs. Low 0.51 ( . 9 0 9 ) 02,.0 1.10 ( . 5 1 8 ) 06,.6 0.020 0.711 AGE (p=0.046) �TABLE 15-14. (continued) Adjusted Exposure Index Analyses for Reported and Verified Heart Disease by Occupation Adj. Relative Risk (95% C.I.) p-Value Medium vs. Low High vs. Low 0.75 (0.42,1.34) 0.73 (0.40,1.32) 0.332 0.298 AGE (p=0.007) Enlisted Flyer Medium vs. Low High vs. Low 1.11 (0.48,2.59) 0.57 (0.22,1.46) 0.803 0.242 NONE Enlisted Groundcrew Medium vs.. Low High vs. Low 0.38 (0.20,0.73) 0.95 (0.55,1.66) 004 .0 0.865 AGE (p=0.024) Officer Medium vs. Low High vs. Low 4.01 (0.43,37.2) 1.20 ( . 7 19.9) 00, 0.222 O.897 ALC (p=0.044) Enlisted Flyer Medium vs. Low High vs. Low Enlisted Groundcrew Medium vs. Low High vs. Low 0.86 (0.13,5.47) 0.79 (0.14,4.35) 0.873 0.787 AGE (p<0.001) Officer Medium vs. Low High vs. Low Variable Occupation Verified Heart Disease Officer Contrast Significant Covariates Oi i CO 00 Reported Myocardial Infarction • Verified Myocardial Infarction Enlisted Flyer Medium vs. Low High vs. Low — Medium vs. Low High vs. Low Enlisted Groundcrev — —Analysis not performed due to sparse cells. — — — — �TABLE 15-15. Adjusted Exposure Index Analyses for Central Cardiac Function Variables by Occupation Variable Occupation Contrast Adj. Relative Risk ( 5 C I ) 9% .. p-Value Significant Covariates Officer AGE (p=0.004) PS (p=0.010) Enlisted Flyer Medium vs. Low High vs. Low 0.94 (0.22,3.98) 0.74 (0.16,3.46) 0.936 0.697 NONE Medium vs. Low High vs. Low 2.76 (0.93,8.24) 1.97 (0.61,6.32) 0.069 0.254 AGE (p=0.041) %BFAT (p=0.006) Medium vs. Low High vs. Low 0.71 (0.16,3.16) 1.33 (0.33,5.40) 0.660 0.689 AGE (p=0.004) DIFCORT (p=0.009) Enlisted Flyer — — Enlisted Groundcrev Medium vs. Low High vs. Low 0.25 (0.05,1.41) 1.26 ( . 0 4 0 ) 04,.0 0.116 0.689 CHOL/HDL (p<0.001) Officer ECG 0.638 0.952 Officer Heart Sounds 0.78 (0.27,2.26) 1.03 ( . 5 3 0 ) 03,.8 Enlisted Groundcrew Systolic Blood Pressure Medium vs. Low High vs. Low Medium vs. Low High vs. Low 1.36 (0.63,2.97) 1.15 (0.50,2.62) 0.435 0.741 AGE (p=0.007) 2BFAT (p=0.009) Enlisted Flyer Medium vs. Low High vs. Low 1.54 (0.53,4.32) 0.86 (0.29,2.49) 0.412 0.779 AGE (p=0.007) %BFAT (p<0.001) Enlisted Groundcrew Medium vs. Low High vs. Low 0.66 (0.29,1.54) 0.76 (0.34,1.71) 0.342 0.516 AGE (p=0.001) PACKYR (p=0.036) — DIFCORT (p=0.038) �TABLE 15-15. (continued) Adjusted Exposure Index Analyses for Central Cardiac Function Variables by Occupation Variable Occupation Contrast Adj. Relative Risk (95% C.I.) p-Value Significant Covariates Officer Enlisted Flyer Medium vs. Low High vs. Low 3.10 (0.85,11.28) 1.64 (0.43,6.28) 0.085 0.472 AGE (p=0.032) £BFAT (p<0.001) Medium vs. Low High vs. Low 0.50 (0.19,1.28) 0.37 (0.14,1.02) 0.150 0.055 AGE (p=0.003) Medium vs. Low High vs. Low 1.06 (0.37,3.06) 1.10 (0.37,3.27) 0.912 0.865 RACE (p=0.035) Enlisted Flyer Medium vs. Low High vs. Low 5.09 (0.57,45.1) 2.04 (0.18,23.1) 0.144 0.569 NONE Medium vs. Low High vs. Low 0.21 ( . 4 0 9 ) 00,.8 0.50 (0.15,1.65) 0.048 0.254 NONE Officer Arrhythmia AGE (p=0.027) RACE (p=0.027) %BFAT (p=0.002) Enlisted Groundcrew Bradycardia 0.289 0.441 Officer Ln I 1.65 ( . 5 4 2 ) 06,.0 1.47 (0.55,3.92) Enlisted Groundcrew Nonspecific T-tfave Changes Medium vs. Low High vs. Low Medium vs. Low High vs. Low 0.21 ( . 4 1 1 ) 00,.4 0.17 (0.02,1.44) 0.070 0.105 AGE (p=0.011) Enlisted Flyer Medium vs. Low High vs. Low Enlisted Groundcrew Medium vs. Low High vs. Low — 0.63 (0.18,2.26) 0.99 (0.32,3.02) — 0.484 0.984 — PACKYR (p<0.001) �TABLE 15-15. (continued) Adjusted Exposure Index Analyses for Central Cardiac Function Variables by Occupation Variable Occupation Contrast Adj. Relative Risk (95% C.I.) p-Value Significant Covariates Officer 1.29 ( .61,2.72) 0 0.86 ( .37,1.96) 0 0.704 0.711 AGE (p=0.003) Enlisted Flyer Medium vs . Low High vs. Low 0.54 ( .17,1.78) 0 0 0.37 ( .11,1.32) 0.312 0.522 AGE (p=0.034) Enlisted Groundcrev Other Diagnoses Medium vs . Low High vs. Low Medium vs . Low High vs. Low 0 0.91 ( .36,2.29) 0.76 (0- 30,1.92) 0.841 0.562 AGE (p<0.001) RACE (p=0.030) —Analysis not performed due to sparse cells. �Peripheral Vascular System There were no significant dose-response effects for diastolic blood pressure (dichotomized), funduscopic abnormalities, or carotid bruits (Table 15-16). Analysis of diastolic blood pressure as a continuous variable also did not reveal any statistically significant exposure level effects. Significant covariates were percent body fat, personality type, cholesterol-HDL ratio, and current alcohol use. Exposure index analyses of the peripheral pulses did not detect any statistically significant exposure effects, either unadjusted (Tables M-9 and M-10 of Appendix M for the manual and Doppler pulse readings) or adjusted (Tables 15-17 and 15-18). Main-effect exposure analyses of 6 historical and verified heart disease variables, 10 central cardiac function variables, and 11 peripheral cardiac function variables (with both manual and Doppler results), showed no evidence of a dose-response relationship at the followup examination. Two statist;ically significant and several borderline significant exposure associations lacked a pattern of dose-response consistency, and appeared to be random in nature. Association of Cardiovascular Examination Findings With Verified Heart Disease The central and peripheral cardiovascular examination findings were analyzed together with the verified cardiovascular disease endpoints to determine the degree of correlation between the third-year followup examination and the past medical history. The results are shown in Table M-ll of Appendix M. There were highly significant associations between verified essential hypertension and systolic and diastolic blood pressures, ECG abnormalities, and abnormal fundi (p<0.001, <0.001, <0.001, 0.008, respectively). There was also a significant association between essential hypertension and abnormal heart sounds (p=0.036), as well as a borderline significant association between hypertension and carotid bruits (p=0.080). The frequency of verified essential hypertension, however, was not significantly different in those with and without peripheral pulse abnormalities (as determined by either the manual or Doppler technique). For verified heart disease, there was a negative association with diastolic blood pressure (p=0.043) and positive associations with ECG abnormalities, heart sounds, abnormal fundi, and abnormal peripheral pulses as determined by the Doppler technique (p<0.001, p=0.017, p=0.014, and p=0.007, respectively). Finally, there were significant positive associations between ECG and heart sound abnormalities (p<0.001 for both) and the occurrence of a verified myocardial infarction. The consistency between the examination findings and the past medical history provides support for the overall validity of the cardiovascular measurement systems, whether by selfreport, medical records, physician assessments, or objective determinations (e.g., ECG). 15-42 �TABLE 15-16. Adjusted Exposure Index Analyses for Diastolic Blood Pressure Funduscopic Abnormalities and Carotid Bruits by Occupation Variable Occupation Contrast Adj. Relative Risk (95% C.I.) p-Value Significant Covariates Officer 0.667 0.465 ZBFAT (p=0 .002) Enlisted Flyer Medium vs . Low 0.76 (0.19,2 .98) High vs . Low 1.06 (0.29,3 .84) 0.697 0.928 ZBFAT (P=0 . 0 ) 06 Enlisted Groundcrev Diastolic Blood Pressure Medium vs . Low 0.79 (0.27,2 .30) High vs . Low 1.44 (0.54,3 .86) Medium vs. Low High vs . Low 1.50 (0.63,3 .59) 1.24 (0.53,2 .86) 0.363 0.667 CHOL/HDL (p=0.037) ZBFAT (p<0.001) Officer u> 0.337a Funduscopic Enlisted Abnormalities Flyer Enlisted Groundcrew Officer Carotid Bruits 0.388* Enlisted Flyer Enlisted Groundcrew 'Overall analysis; sparse cells, chi-square test may not be valid. —Analysis not performed due to sparse cells. �TABLE 15-17. Adjusted Exposure Index Analyses for Peripheral Vascular System Manual Pulse Readings by Occupation Variable Occupation Contrast Officer Enlisted Flyer Enlisted Groundcrew Medium vs. Low High vs. Low 0.928 0.952 AGE (p=0.030) RACE (p=0.048) DIFCORT (p=0.048) Medium vs. Low High vs. Low Enlisted Groundcrew 0.91 (0.14,5.90) 1.05 (0.20,5.52) Medium vs. Low High vs. Low Enlisted Flyer RACE (p=0.006) XBFAT (p=0.032) PS (p=0.041) Medium vs. Low High vs. Low Officer 0.509 0.238 Medium vs. Low High vs. Low Medium vs. Low High vs. Low 0.36 (0.02,7.42) 3.20 (0.31,32.6) Medium vs. Low High vs. Low Enlisted Flyer Significant Covariates Medium vs. Low High vs. Low Officer p-Value Medium vs. Low High vs. Low Enlisted Groundcrew Radial Pulses Adj. Relative Risk ( 5 C.I.) 9% Ul Femoral Pulses Popliteal Pulses �TABLE 15-17. (continued) Adjusted Exposure Index Analyses for Peripheral Vascular System Manual Pulse Readings by Occupation Variable Occupation Contrast Adj. Relative Risk (95% C.I.) p-Value Significant Covariates Officer NONE Enlisted Flyer Medium vs. Low High vs . Low 1.53 (0.51,4.65) 1.39 (0.45,4.33) 0.447 0.569 NONE Medium vs. Low High vs. Low 0.61 (0.27,1.35) 0.95 ( . 5 2 0 ) 04,.2 0.222 0.897 NONE Medium vs. Low High vs. Low Enlisted Flyer Medium vs. Low High vs. Low 0.75 1.26 Enlisted Groundcrev Ul 0.764 0.171 Officer *- 0.88 (0.40,1-97) 0.52 (0.20,1-33) Enlisted Groundcrew Dorsalis Pedis Pulses Medium vs. Low High vs. Low Leg Pulses — — (0.13,4.20) (0.26,6.10) 0.741 0.772 RACE (p=0.027) Medium vs. Low High vs. Low 2.00 (0.48,8.33) 1.51 (0.37,6.17) 0.337 0.569 AGE (p=0.003) RACE (p<0.001) Officer Posterior Tibial Pulses — Medium vs. Low High vs. Low 0.96 (0.44,2.12) 0.84 (0.37,1.95) 0.920 0.697 NONE Enlisted Flyer Medium vs. Low High vs. Low 1.01 (0.37,2.75) 1.21 (0.45,3.27) 0.984 0.711 PS (p=0.034) Enlisted Groundcrew Medium vs. Low High vs. Low 0.69 (0.35,1.36) 0.89 (0.46,1.75) 0.285 0.741 NONE �TABLE 15-17. (continued) Adjusted Exposure Index Analyses for Peripheral Vascular System Manual Pulse Readings by Occupation Variable Occupation Contrast Adj. Relative Risk (95% C.I.) p-Value Significant Covariates Officer All Pulses NONE Enlisted Flyer Medium vs. Low High vs. Low 1.01 (0.37,2.75) 1.21 (0.45,3.27) 0.984 0.711 PS (p=0.034) Medium vs. Low High vs. Low 0.69 (0.35,1.36) 0.89 (0.46,1.75) 0.285 0.741 NONE Medium vs. Low High vs. Low 0.89 (0.41,1,94) 0.86 (0.42,1.75) 0.764 0.719 NONE Enlisted Flyer Medium vs. Low High vs. Low 1.01 (0.37,2.75) 1.21 (0.45,3.27) 0.984 0.711 PS (p=0.034) Enlisted Groundcrev I 0.764 0.719 Officer Ui 0.89 (0.41,1.94) 0.86 (0.42,1.75) Enlisted Groundcrev Peripheral Pulses Medium vs. Low High vs. Low Medium vs. Low High vs. Low 0.69 (0.35,1.36) 0.89 (0.46,1.75) 0.285 0.741 NONE —Analysis not performed due to sparse cells. �TABLE 15-18. Adjusted Exposure Index Analyses for Peripheral Vascular System Doppler Pulse Reading by Occupation Variable Occupation Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Significant Covariates Officer 1.12 (0.63,1.97) 1.08 (0.60,1-96) 0.704 0.787 NONE Enlisted Flyer Medium vs. Low High vs. Low 1.30 (0.51,3.28) 1.43 (0.56,3.64) 0.575 0.447 NONE Enlisted Groundcrew Medium vs. Low High vs. Low 0.94 (0.53,1.65) 1.04 (0.58,1.87) 0.772 0.764 RACE (p=0.034) Officer Medium vs. Low High vs. Low 2.09 (0.38,11.6) 1.01 (0.13,7.58) 0.401 0.992 AGE (p=0.025) DIFCORT (p=0.026) Enlisted Flyer Medium vs. Low High vs. Low Enlisted Groundcrev Medium vs. Low High vs. Low Officer Dorsal is Pedis Pulses Medium vs. Low High vs. Low Medium vs. Low High vs. Low 1.26 (0.71,2.21) 1.19 (0.66,2.13) 0.430 0.562 NONE Enlisted Flyer Medium vs. Low High vs. Low 1.57 (0.63,3.90) 1.58 (0.63,3.98) 0.327 0.327 NONE Enlisted Groundcrew Medium vs. Low High vs. Low 0.94 (0.54,1.87) 1.01 (0.57,1.80) 0.818 0.772 NONE tn Posterior Tibial Pulses Leg Pulses — — — — — — �TABLE 15-18. (continued) Adjusted Exposure Index Analyses for Peripheral Vascular System Doppler Pulse Reading by Occupation Variable Occupation Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Significant Covariates Officer NONE Enlisted Flyer Medium vs. Low High vs. Low 1.57 (0.63,3.90) 1.58 (0.63,3.98) 0.327 0.327 NONE Medium vs. Low High vs. Low 0 9 (0.52,1.57) .0 1.02 (0.58,1.79) 0.704 0.960 NONE Medium vs. Low High vs. Low 1.26 (0.71,2.21) 1.19 (0.66,2.13) 0.430 0.562 NONE Enlisted Flyer Medium vs. Low High vs. Low 1.57 (0.63,3.90) 1.58 (0.63,3.98) 0.327 0.327 NONE Enlisted Groundcrev *- 0.430 0.562 Officer Ul 1.26 (0.71,2.21) 1.19 (0.66,2.13) Enlisted Groundcrev Peripheral Pulses Medium vs. Low High vs. Low Medium vs. Low High vs. Low 0.90 (0.52,1.57) 1.02 (0.58,1.79) 0.704 0.960 NONE. 00 All Pulses —Analysis not performed due to sparse cells. �LONGITUDINAL ANALYSES Two cardiovascular variables, the index of all pulses (by palpation) and the overall ECG interpretation, were investigated to assess the longitudinal differences between the 1982 Baseline examination and the 1985 followup examination. Both variables are classified as abnormal or normal. As shown in Table 15-19, 2x2 tables were constructed for each group for each variable. These tables show the number of participants who were abnormal at Baseline and abnormal at followup, abnormal at Baseline and normal at followup, normal at Baseline and abnormal at followup, and normal at both Baseline and followup examinations. The odds ratio given is the ratio of the number of participants who were normal at the Baseline and abnormal at the followup to the number of participants who were abnormal at the Baseline and normal at the followup (the "off-diagonal" elements). The changes in normal/abnormal status within each group are contrasted between the Ranch Hand and Comparison groups, and the p-value is derived from Pearson's chi-square test of the hypothesis that the pattern of change in the two groups is the same. TABLE 15-19. Longitudinal Analyses of All Pulses Index and Overall ECG's: A Contrast of Baseline and First Followup Examination Abnormalities Variable Group 1982 Baseline Exam 1985 Followup Exam Odds* Ratio (OR) p-Value (ORRH vs. ORC) Abnormal Normal All Pulses Ranch Hand Abnormal (Manual) Normal 50 104 72 743 1.44 Comparison Abnormal Normal 40 153 63 880 2.43 Ranch Hand Abnormal Normal 86 43 192 650 0.22 Comparison Abnormal Normal 112 56 208 763 0.27 0.01 ECG (Overall) 0.42 Number Normal Baseline, Abnormal Followup *0dds Ratio: Number Abnormal Baseline, Normal Followup 15-49 • �The data showed a significant difference (p=0.01) in the pulse index in the two groups between examinations. The percentage of Ranch Hands and Comparisons with abnormalities for the pulse index increased from the Baseline examination to the followup examination; however, the Comparison group showed a larger increase in the proportion of pulse index abnormalities. The greater relative increase in the Comparisons caused the significant result. No significant group differences were detected between examinations for overall ECG abnormalities (p=0.42). DISCUSSION In general, the foregoing analyses on a wide range of cardiovascular variables, have shown a lack of significant differences between the Ranch Hands and the Comparisons. The sole exception was the finding of increased verified heart disease in the Ranch Hands versus the Comparisons (24% and 20%, respectively, p=0.054, unadjusted; p=0.036, adjusted). These results were not noted in the Baseline examination (p=0.982, unadjusted). A review of the relative risk patterns, whether or not statistically significant, for all of the other cardiovascular variables showed general equality, with about half of the risks below unity and half above. This rough equivalence suggests that, although the Ranch Hands have slightly more reported heart disease, the finding is not mirrored by substantial and consistent clinical cardiovascular defects at this time. This observation should not be lightly dismissed, and is cause for continued close surveillance. The most notable cardiovascular finding at the followup examination was the lack of significant peripheral pulse abnormalities, which were unexpectedly found at the 1982 Baseline examination (p=0.05).. The primary contributory cause of the change in pulse significance from Baseline to followup was probably the rigid 4-hour tobacco abstinence required prior to Doppler testing (due to the known vasoconstriction effects of nicotine). Tobacco abstinence, however, was not a requirement for the Baseline manual pulse readings. Although tobacco abstinence was not a requirement prior to manual readings at the followup examination, there was general compliance to the smoking prohibition, particularly if a participant's general physical examination preceded the Doppler testing. Therefore it might be expected that the manual readings would show more pulse abnormalities than Doppler testing; in fact, this was the case (see section on Peripheral Pulses). Whatever the true cause(s), the prevailing fact is that there are no longer significant group differences in pulse abnormalities, as noted by both manual and Doppler techniques, regardless of the poor agreement between the two methods. The close approximation of the estimated relative risks to unity for practically all of the cardiovascular variables is clearly indicative of equivalent cardiovascular health between the two groups. Furthermore, the general similarity of the unadjusted and adjusted results was suggestive of near equivalence of the important cardiovascular risk factors in the Ranch Hands and Comparisons (see Table 15-13), as well as a balance for unanalyzed or hidden covariates of importance. These health assessments of the two groups are considerably strengthened by the almost consistent, classical effects of the covariates in this chapter. In particular, the age effect was uniformly profound, affecting 15-50 �almost all of the dependent variables in the functional categories of reported-verified heart diseases, and central and peripheral vascular function. The covariates of race, percent body fat, and cholesterol (particularly the cholesterol-HDL ratio), and smoking were also generally strong and consistent in their effects. Statistically significant, positive associations were seen between the current level of smoking and posterior tibial, popliteal, and femoral pulses, as well as borderline significant associations between current smoking and other ECG diagnoses, carotid bruits, and reported myocardial infarctions. However, significant negative associations were observed between current smoking and reported and verified essential hypertension. Pack-years of smoking was significantly positively associated with several ECG variables and pulse assessments, although not always in a consistently increasing manner. There was a statistically significant and consistently increasing effect of pack-years of smoking on reported and verified myocardial infarctions, but there was a negative association between pack-years of smoking and verified essential hypertension, with the greatest number of abnormalities in the zero pack-year category. Alcohol was infrequently interactive with the dependent variables, but covariate tests of association generally revealed the classical pattern of more cardiovascular abnormalities in the nondrinking category than in the low drinking category. Personality score, however, usually failed to demonstrate the "expected" aggregation of cardiovascular abnormalities in the Type A direction. In fact, most associations were in the Type B direction. Generally, only cardiovascular studies ascertaining personality type by the Structured Interview technique have shown an association of Type A personality (Type A-l, in particular) to heart disease endpoints, and conversely, studies using questionnaire techniques to measure personality type have not demonstrated the association. Lastly, the strong association between historical-verified cardiovascular events and the specific dependent variables provides assurance that the overall cardiovascular measurements have been accurate and valid. SUMMARY AND CONCLUSIONS The cardiovascular health of both cohorts was assessed by collection of reported and record-verified heart disease events; measurement of central cardiac function by systolic blood pressure, abnormal heart sounds, and electrocardiograph (ECG) findings; and evaluation of peripheral vascular function by diastolic blood pressure, funduscopic examination, presence of carotid bruits, and detailed manual and Doppler measurements of five peripheral pulses. Table 15-20 presents the overall summary of the unadjusted and adjusted results. Where possible, the analyses used the covariates of age, race, occupation, percent body fat, cholesterol, high density lipoprotein (HDL) cholesterol, cholesterol-HDL ratio, smoking history (pack-years and current smoking level), alcohol history (drink-years and current drinking level), personality score, and differential cortisol. The cardiovascular variables did not reveal significant group differences, with the exception of verified heart disease, for which the proportions of recorded cardiac events were 24 and 20 percent in the Ranch Hand and Comparison groups, respectively, (p=0.054 unadjusted, p=0.036 adjusted). This finding was not reinforced by results of individual questionnaire or examination variables showing impairment in the Ranch Hands. There was a remarkable balance in relative risks above and below unity between the groups. 15-51 �TABLE 15-20. Overall Summary Results of Unadjusted and Adjusted Analyses Cardiovascular Variables Variable Statistical/ Clinical Analysis Unadjusted Adjusted Historical and Verified Heart Disease Reported Verified Reported Verified Reported Verified Hypertension Hypertension Heart Disease8 Heart Disease* Heart Attack Heart Attack NS NS NS NS* NS NS " NS NS NS S NS NS Central Cardiac Function Systolic Blood Pressure Discrete Continuous Heart Sounds Electrocardiogram (Overall) ECG: RBBB ECG: LBBB ECG: Nonspecific T-Wave Changes ECG: Bradycardia ECG: Tachycardia ECG: Arrhythmia ECG: Other Diagnoses NS NS NS NS NS NS NS NS NS 15-52 NS **** NS **** NS N/A NS NS N/A **** NS �TABLE 15-20. (continued) Overall Summary Results of Unadjusted and Adjusted Analyses Cardiovascular Variables Statistical/ Clinical Analysis Unadjusted Adjusted Funduscopic Examination NS NS NS NS NS NS Carotid Bruits NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS **** Variable Peripheral Vascular Function Diastolic Blood Pressure Radial Pulses Femoral Pulses Popliteal Pulses Dorsalis Pedis Pulses Posterior Tibial Pulses Leg Pulses Peripheral Pulses All Pulses Discrete Continuous Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler Manual Doppler NS NS:Not significant (p>0.10). NS*:Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. a Excluding hypertension. b RH>C (Adj. RR: 1.25;'95% C.I.: [1.02, 1.54], p=0.036). 15-53 NS **** NS **** NS **** NS **** NS **** NS �Other related analyses showed an absence of significant group differences in reported or verified hypertension, reported or verified heart attacks, and reported heart disease. There was good correlation between the verified cardiovascular history and the central and peripheral cardiovascular abnormalities detected at the physical examination, supporting accuracy and validity of the cardiovascular measurements. The adjusted analyses of central cardiac function disclosed a significant group-by-age interaction involving systolic blood pressure in the Black cohort, with a mean systolic blood pressure greater in the Ranch Hands than the Comparisons at younger age levels, but a lower mean pressure at the older ages; the group-by-age interaction was not significant in the nonblack cohort. Additionally, there was a significant group-by-pack-years of smoking interaction for the overall EGG findings, and significant group-by-pack-years of smoking and group-by-percent body fat interactions for arrhythmia, but they all generally pointed to lower adjusted relative risks in the Ranch Hands. In the analysis of peripheral vascular function, no significant group differences were observed for abnormalities involving radial, femoral, popliteal, posterior tibial, dorsalis pedis, or three anatomic aggregates of these pulses, either by manual palpation or Doppler techniques. This overall finding was in distinct contrast to the 1982 Baseline examination, which by the manual palpation method, showed significant peripheral pulse deficits in the Ranch Hands. This favorable pulse reversal over the two examinations is primarily attributed to the rigid 4-hour tobacco abstinence applied prior to Doppler testing, although other factors may be related. The lack of group differences for pulse abnormalities was noted even though the manual and Doppler techniques differed significantly (p<0.05, p<0.001 for most) in the detection of abnormalities for all but one of the pulses or pulse combinations. For manually-determined pulse abnormalities, there was a significant group-by-race interaction for the popliteal pulses, a significant group-bypercent body fat interaction for the leg pulses, and significant group-byoccupation interactions for the posterior tibial, dosalis pedis, and the three pulse aggregates (leg, peripheral, and all pulses). No interactions were encountered in the adjusted analyses of the Doppler results, and none showed significant group differences. Statistical analyses involving the Original Comparisons also showed no significant differences in the cardiovascular measurements between groups, although slightly different interactions were detected in some of the adjusted analyses. For the exposure analyses, the only statistically significant effects were those pointing to less bradycardia and less reported and verified heart disease in the medium exposure level category, as contrasted to the low exposure category, among the enlisted groundcrew. In many cases there were too few abnormalities within the occupational categories to permit formal statistical tests. Overall, the.exposure analyses were deemed as unsupportive of any meaningful dose-response relationships. The longitudinal analysis of the pulse index confirmed the significant difference in the change in the pattern of results from the Baseline examination to the followup examination, largely due to a relatively greater 15-54 �increase of pulse abnormalities in the Comparison group than in the Ranch Hand group. There was no significant change in pattern between the two groups in overall ECG findings between examinations. There was a similar distribution of the covariates between groups, except for a slightly higher level of current Ranch Hand smoking (also observed at Baseline), and a corresponding slightly lower mean percent body fat. The general covariate effects were strong and showed expected, classical associations with the cardiovascular measurements. However, unexpected effects were consistently noted for personality score, with higher proportions of various cardiovascular abnormalities associated with scores in the Type B direction, a finding possibly attributable to the method of personality determination. Nonetheless,^ the repeated demonstration of classical covariate associations with cardiovascular pathology lends considerable credence to the quality of the data. Although smoking was positively associated with many of the cardiovascular measurements, negative associations were seen between current smoking and reported and verified essential hypertension and between pack-years of smoking and verified hypertension. In conclusion, of 27 cardiovascular variables, only one, verified heart disease, showed a significant excess in the Ranch Hands, but this finding was largely unsupported by other cardiac measurements. Both manual palpation and Doppler recordings of five peripheral pulses were similar in both groups, in marked contrast to the 1982 Baseline examination which found significant pulse deficits in the Ranch Hand group. This change at the followup examination was most likely due to required tobacco abstinence prior to the pulse measurements. Exposure index analyses did not support a consistent doseresponse relationship for any variable. Overall, there was remarkable similarity in the cardiovascular health between the Ranch Hand and Comparison groups. 15-55 �CHAPTER 15 REFERENCES 1. Palmer, J.S., and R.D. Radeleff. 1964. The toxicologic effects of certain fungicides and herbicides on sheep and cattle. Ann. N.Y. Acad. Sci. 111:729-736. 2. McConnell, E.E., J.A. Moore, and D.W. Dalgard. 1978. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in Rhesus monkeys (Macaca mulatta) following a single oral dose. Toxicol. Appl. Pharmacol. 43(1):175-187. 3. Kimbrough, R.D., C.D. Carter, J.A. Liddle, R.E. Cline, and P.E. Phillips. 1977. Epidemiology and pathology of a tetrachlorodibenzodioxin poisoning episode. Arch. Environ. Health 32(2):77-86. 4. McConnell, E.E., J.A. Moore, J.K. Baseman, and M.W. Harris. 1978. The comparative toxicity of chlorinated dibenzo-p-dioxins in mice and guinea pigs. Toxicol. Appl. Pharmacol. 44(2):335-356. 5. Schreiweis, D.O., and G.J. Murray. 1976. Cardiovascular malformations in Oryzias latipes embryos treated with 2,4,5-trichlorophenoxyacetic acid (2,4,5-T). Teratology 14(3):287-290. 6. Rifkind, A.B., Y. Hattori, R. Levi, M.J. Hughes, C. Quilley, and D.R. Alonso. The chick embryo as a model for PCB and dioxin toxicity: Evidence of cardiotoxicity and increased prostaglandin synthesis. In Banbury report 18; Biological mechanisms of dioxin action ed. A. Poland and R.D. Kimbrough, pp. 255-266.Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 7. Dudley, A.W., and N.T. Thapar. 1972. Fatal human ingestion of 2,4-D, a common herbicide. Arch. Path. 94:270-275. 8. Paggiaro, P.L., E. Martino, and S. Mariotti. 1974. A case of 2,4-dichlorophenoxyacetic acid (2,4-D) poisoning. Med. Lavoro 65(3-4):128-135. 9. Berwick, P. 1970. 2,4-Dichlorophenoxyacetic acid poisoning in man. JAMA 214(6):1114-1117. 10. Oliver, R.M. 1975. Toxic effects of 2,3,7,8-tetrachloro-dibenzo-l, 4-dioxin in laboratory workers. Br. J. Ind. Med. 32:46-53. 11. Baader, E.W., and A.J. Bauer. 1951. Industrial intoxication due to pentachlorophenol. Ind. Med. Surg. 20:289-290. 12. Jirasek, L., J. Kalensky, K. Kubec, et al. 1974. Acne chlorina, porphyria cutanea tarda and other manifestations of general intoxication during the manufacture of herbicides, part 2. Czech. Dermatol. 49(3):145-157. 15-56 �13. Pazderova-Vejlupkova, J., M. Nemcova, J. Pickova, L. Jirasek, and E. Lukas. 1981. The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men. Arch. Environ. Health 36:5-11. 14. Poland, A.P., D. Smith, G. Metter, and P. Possick. 1971. A health survey of workers in a 2,4-D and 2,4,5-T plant, with special attention to chloracne, porphyria cutanea tarda, and psychologic parameters. Arch. Environ. Health 22(3):316-327. 15. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichloro-phenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 16. Suskind, R.R., and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 17. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, W.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. JAMA 255:2031-2038. 18. Stehr, P.A., G. Stein, H. Falk, et al. 1986. A pilot epidemiologic study of possible health effects associated with 2,3,7,8-tetrachlorodibenzo-p-dioxin contamination in Missouri. Arch. Environ. Health 41:16-22. 19. Troxler, R.G., and H.A. Schwertner. 1985. Cholesterol, stress, lifestyle, and coronary heart disease. Aviat. Space Environ. Med. 56:660-665. 20. American Heart Association Steering Committee for Medical and Community Program. 1980. Risk factors and coronary disease: A statement for physicians. Circulation 62:449A-455A. 21. Castelli, W.P. 1984. Epidemiology of coronary heart disease: Framingham study. Am. J. Med. 76:4-12. The 22. Multiple Risk Factor Intervention Trial Research Group (Neaton, J.D., L.H. Kuller, D. Wentworth, et al.). 1984. Total and cardiovascular mortality in relation to cigarette smoking, serum cholesterol concentration, and diastolic blood pressure among black and white males followed for five years. Am. Heart J. 108:759-769. 23. Morton, W.E., E.D. Crawford, R.A. Maricle, D.D. Douglas, and V.H. Freed. 1975. Hypertension in Oregon pesticide-formulating workers. J. Occ. Med. 17(3):182-185. 24. Martin, J.V. 1984. Lipid abnormalities in workers exposed to dioxin. Br. J. Ind. Med. 41:254-256. 25. Ashe, W.F., and R.R. Suskind. 1949, 1950. Reports on chloracne cases, Monsanto Chemical Company, Nitro, West Virginia. In Report of the Kettering Laboratory, December 1949 and April 1950. 15-57 �26. Lipid Research Clinic Program: The Lipid Research Clinic's Coronary Prevention Trial Results: II. The relationship of reduction in incidence of coronary heart disease to cholesterol lowering. 1984. JAMA 251:365-374. 27. Stamler, J., D. Wentvorth, and J.D. Neaton. 1986. Is relationship between serum cholesterol and risk of premature death from coronary heart disease continuous and graded? JAMA 256:2823-2828. 15-58 �CHAPTER 16 HEMATOLOGICAL EVALUATION INTRODUCTION Although direct impairment of the hematopoietic system may result from exposure to chlorophenols or dioxin, marked abnormalities in many of the circulating hematological elements may also be due to the severe and often endstage toxicity observed in other organs or organ systems. Animal experiments have confirmed both direct and indirect hematopoietic effects of TCDD. In a chronic low-dose feeding study of TCDD in eight monkeys, decreased hemoglobin and hematocrit values were noted at the 6-month mark in all animals. Four of these monkeys expired in 7 to 11 months and all had anemia, leukopenia, and thrombocytopenia. Necropsy of three sacrificed animals at 1 year showed multi-organ pathology including bone marrow degeneration, atrophy of lymphopoietic tissue, and numerous hemorrhages in a variety of organs. In another monkey experiment, using single low and high doses of TCDD, early hematological effects included increased neutrophil counts in the low-dose group and lymphopenia and thrombocytopenia in the high-dose group.2 At the end of the experiment, half the sternal bone-marrow samples revealed a decrease in overall cellular!ty and an increase in the myeloid-erythroid cell ratio. Rat experiments with TCDD demonstrated relatively consistent results. One study revealed elevated erythrocyte, reticulocyte, and neutrophil counts with depressed values for the mean corpuscular volume, mean corpuscular hemoglobin, platelet counts, and clot retraction times. The authors attributed most of these effects to terminal dehydration and nonspecific toxicity. Another rat study using gavage doses of TCDD varying from 0.001 to 1.0 wg/kg demonstrated depressed red blood cell counts and packed cell volumes in the high-dose group.4 In a mixed-dose regimen using rats, mice, and guinea pigs, dose-related decreases in lymphocyte and leukocyte numbers were observed in mice and guinea pigs within 1 week following TCDD administration. Thrombocytopenia and hemoconcentration were found in rats. Because of the lymphopenia in mice and guinea pigs, TCDD was judged to be immunosuppressive. In general, human observational studies showed fewer and less consistent hematological findings than the structured animal experiments. A case report of 2,4-D intoxication with marked neurological findings described transient bone marrow depression with peripheral leukopenia and granulocytopenia. In two industrial accidents involving significant contamination with TCDD and resulting cases of chloracne, only temporary depression of peripheral leukocyte and lymphocyte formation was observed. ' Two contemporary indepth morbidity studies9'10 of the Nitro, Vest Virginia, accident included routine clinical complete blood counts and differential counts, and hemoglobin and hematocrit determinations. Though these studies shared overlapping study cohorts, they did not report any of the 16-1 �hematological results in their publications; presumably, there were no significant differences in any of the parameters between the exposed and the unexposed cohorts. The two pilot studies of TCDD-contaminated residential areas in Missouri also included routine hematological assays of peripheral blood. ' One study paradoxically noted a significantly increased mean platelet count in the high-risk group, although the data were not adjusted for smoking.11 The Quail Run study, predominantly emphasizing cell-mediated immunity, found significant group differences in the mean leukocyte count, mean absolute granulpcyte count, and the mean percentage of monocytes in the differential count. Unfortunately, the authors neglected to identify the group (exposed or unexposed) that had the abnormal hematological findings. However, the finding of a significantly higher proportion of individuals with white blood cell counts exceeding 10,000/mm was in the exposed group. Baseline Summary Results A number of statistically significant group differences and interactions emerged in the analysis of the 1982 Baseline examination. The Ranch Hand group had a significantly higher adjusted mean red blood cell corpuscular volume and corpuscular hemoglobin value than the Comparison group (p»0.05, p=0.04, respectively), although the magnitude of the difference was small in each case. The Ranch Hand adjusted mean values for six other parameters, i.e., red blood cell count, white blood cell count, hemoglobin, hematocrit, mean corpuscular hemoglobin concentration, and platelet count, were nearly identical to the adjusted means of the Comparison group, and all were well within normal range. Similarly, the percent of abnormal values for these eight variables, as established by the upper and lower limits of normal, did not vary by group. Linear models demonstrated the profound effect of smoking, as measured in pack-years. With increased smoking, white blood cell, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, and platelet values increased, whereas the mean corpuscular hemoglobin concentration showed a significant negative association with smoking. The red blood cell count revealed a borderline significant negative relationship to smoking. No statistically significant group-by-smoking interactions were detected. The exposure index analyses conducted within the Ranch Hand group disclosed two statistically significant exposure-level effects as well as seven significant or borderline-significant exposure-level-by-smoking interactions. In the officer cohort, the percentage of mean corpuscular hemoglobin abnormalities increased with increasing exposure level. The high-exposure group also had the highest percentage of mean corpuscular hemoglobin concentration abnormalities. No significant associations were found, however, in the enlisted flyer or enlisted groundcrew cohort. Five interactions involved a decreasing association (gradient of slopes) between the hematological measure and pack-years of smoking with increasing exposure level, one showed an increasing association with increasing exposure level, and one was uninterpretable. The report concluded that the overall statistical findings were somewhat consistent among themselves, and that medical morbidity was not significant. 16-2 �Parameters of the 1985 Hematological Evaluation The 1985 hematological assessment was identical to the 1982 Baseline evaluation. The eight hematological variables were red blood cell count (RBC), white blood cell count (WBC), hemoglobin (HGB), hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and platelet count (PLT); these variables were determined by routine hematological procedures. The normal ranges of the SCRF-determined values differed somewhat from those employed in 1982 by the Kelsey-Seybold Clinic. As before, the analysis of the hematological data included the covariates of age, race, occupation, and smoking. Updated and more comprehensive smoking data, in terms of pack-years and current smoking (including cigar- and pipe-smoking), were used in most analyses. Excluded were three individuals with fever at the time of examination (two Ranch Hands and one Comparison). Hematological variables in the continuous form were analyzed by general linear models adjusting for age, race, occupation, and smoking. The hematological data, trichotomized as abnormally low, normal, or abnormally high, were subjected to log-linear (logit) analysis, adjusted for the same covariates. Minor differences in the table totals within this chapter reflect rare missing data for either the dependent variables or the covariates. Parallel analyses using Original Comparisons can be found in Tables N-4 through N-9 of Appendix N. RESULTS AND DISCUSSION General Eight hematological assays were performed on peripheral blood specimens obtained from all participants on the first day of the physical examination. Table 16-1 lists the assays, the abbreviations used in this chapter, the SCRF laboratory normal range for each assay, and the required laboratory coefficient of variation for each assay. The SCRF laboratory norms varied to some extent from the values used at the Baseline examination (see pages XVI-3-1, Baseline Report). The SCRF laboratory coefficients of variation met or exceeded contract requirements and were uniformly achieved due to the precision of the Coulter 5-Plus automated instrument, in conjunction with rigorous FIR CUSUM quality control techniques (see Chapter 6). The overall precision in the laboratory aspects of the hematological assays, is reflected in the analytic ability to discern minute mean shifts . between groups. Representative statistical power statements are as follows. Sample sizes were sufficiently large to detect a 0.87 percent mean shift in RBC and a 2.5 percent mean shift in PLT values using an a -level of 0.05 (twosided) and a power of 0.80. Further, the sample sizes were sufficient to detect a 1.66-fold increase in the frequency of abnormal values for RBC, and a 1.96-fold increase in the frequency of abnormal values for PLT, with 80 percent certainty. 16-3 �TABLE 16-1. Laboratory Parameters for Hematological Test Variables Hematological Test Abbreviation SCRF Laboratory Normal Range Contract Required Coefficient of Variation (in percent) Red Blood Cell Count RBC 4.3-5.9 million/cubic mm 2.0 White Blood Cell Count WBC A.5-11.0 thousand/cubic mm 2.5 Hemoglobin HGB 13.9-16.3 grams/100 ml 1.1 Hematocrit HCT 39.0-55.0 ml/100 ml 3.0 Mean Corpuscular Volume MCV 80.0-97.0 cubic micra 2.0 Mean Corpuscular Hemoglobin MCH 26.0-34.0 micromicrogram 2.0 Mean Corpuscular Hemoglobin Concentration MCHC 31.0-37.0 percent 2.0 Platelet Count PLT 130-400 thousand/cubic mm 3.5 The statistical analyses in this chapter are presented in the following order: unadjusted tests, covariate tests of association, adjusted analyses, exposure analyses, and longitudinal contrasts. A variable-by-variable discussion summarizes all of the analyses, and representative exposure analyses are also presented. Group-by-covariate interactions are narratively presented, and illustrated by calculating Ranch Hand-Comparison differences at selected covariate levels. The interaction data tables are found in Tables N-2 and N-3 of Appendix N. Unadjusted Categorical Analyses Data from the eight hematological variables were categorized as abnormally low, normal, or abnormally high according to the SCRF laboratory norms cited in Table 16-1. The frequency distribution of these discretized data is presented by group in Table 16-2. As shown, there were no statistically significant, or even marginally significant, differences between the groups. Only one abnormal MCHC value was found among all study participants. 16-4 �TABLE 16-2. Unadjusted Categorical Analyses for Hematological Variables by Group Abnormally Low Variable Group Normal Abnormally High Number Percent Number Percent Number Percent Total p-Value* RBC Ranch Hand Comparison 30 42 3.0 3.2 976 1,239 96.2 95.9 8 11 0.8 0.8 1,014 1,292 0.910 WBC Ranch Hand Comparison 45 63 4.4 4.9 906 1,149 89.4 88.9 62 80 6.1 6.2 1,013 1,292 0.883 HGB Ranch Hand Comparison 39 44 3.8 3.4 752 960 74.2 74.3 223 288 22.0 22.3 1,014 1,292 0.848 HCT Ranch Hand Comparison 11 15 1.1 1.2 1,001 1,274 98.7 98.6 2 3 0.2 0.2 1,014 1,292 0.999 MCV Ranch Hand Comparison 10 13 1.0 1.0 857 1,094 84.5 84.7 147 185 14.5 14.3 1,014 1,292 0.992 MCH Ranch Hand Comparison 7 7 0.7 0.5 943 1,211 93.0 93.7 64 74 6.3 5.7 1,014 1,292 0.755 MCHC Ranch Hand Comparison 1 0 0.1 0.0 1,013 1,292 99.9 100.0 0 0 0.0 0.0 1,014 1,292 — Ranch Hand Comparison 5 3 0.5 0.2 987 1,264 97.4 97.9 21 24 2.1 1.9 1,013 1,291 0.828 PLT *Chi-square test, 2 d.f., except for HCT and PLT which were obtained from continuity adjusted chi-square tests on 1 d.f. (Abnormally high category pooled with normal, and abnormally low category pooled with normal for HCT and PLT, respectively.) —Only one abnormal MCHC value; p-value not given. 16-5 �Unadjusted Analyses of Continuous Data The unadjusted tests of group means from the continuous data for the eight variables are displayed in Table 16-3. The variables WBC and PLT were analyzed in logarithmic units because of their right-skewed original distributions. Antilog values of the means are given for ease of interpretation but their standard error or variance terms are consequently omitted since the relevance of these terms pertains only to the logarithmic scale. The sample sizes were 1,014 for the Ranch Hand group and 1,292 for the Comparisons, except for WBC (Ranch Hands, 1,013; Comparisons, 1,292) and PLT (Ranch Hands, 1,013; Comparisons, 1,291). As shown in Table 16-3, there were no statistically significant group differences between the unadjusted means of each variable. TABLE 16-3. Unadjusted Continuous Analyses for Heoatological Variables (Contrast of Group Means) Group Mean±SE Variable Ranch Hand Difference ±SE Comparison RBC 4.964±0.012 4.982±0.010 WBC8 7.003 6,891 -0.019±0.016 — t-Statistic p-Value -1.19 0.233 1.34 0.182 HGB 15.624±0.033 15.626±0.029 -0.002±0.044 -0.05 0.958 HCT 45.904±0.097 45.952±0.083 -0.048±0.127 -0.38 0.703 MCV 92.596±0.150 92.346±0.132 0.250±0.200 1.25 0.210 MCH 31.544±0.055 31.431±0.049 0.113±0.074 1.53 0.125 MCHC 34.040±0.021 34.009±0.017 0.031±0.027 1.17 0.243 0.96 0.337 PLT* 265.2 263.0 — 'Means transformed from log scale. —Difference and standard errors (SE) not presented, since variables were analyzed on logarithmic scale.. 16-6 �Dependent Variable and Covariate Relationships The data from the Ranch Hand and Comparison groups were pooled for each of the eight hematological variables and analyzed independently with the covariates of age (born in or after 1942, born before 1942), race (Black, nonblack), occupation (officer, enlisted flyer, enlisted groundcrew), and smoking history (0 pack-years; greater than 0 to 10 pack-years; and greater than 10 pack-years). These analyses are summarized in terms of statistical significance (p-values) in Table 16-4. As noted, each of the dependent variables was substantially affected by one or more of the covariates. The exact nature of the covariate influence, e.g., directionality, significance, consistency across related variables, is presented in the variable-byvariable discussion section. Covariate effects were also analyzed in continuous form with the use of linear regression models (see Table 16-6 and discussion following). In addition, covariate distributions were examined between groups (see Table N-l of Appendix N). TABLE 16-4. Association Betveen Hematological Variables and Age, Race, Occupation, and Smoking History in the Combined Ranch Hand and Comparison Groups Variable Age Race Occupation Smoking History RBC 0.010 <0.001 NS NS WBC NS* <0.001 0.001 <0.001 HGB NS 0.002 <0.001 0.003 HCT NS <0.001 NS NS MCV <0.001 <0.001 0.004 <0.001 MCH <0.001 <0.001 0.003 <0.001 — NS — MCHC PLT — 'NS NS NS: Not significant (p>0.10). NS*: Borderline significant (0.05<p<0.10). —Not analyzed due to sparse data. 16-7 — 0.004 �Adjusted Categorical Analyses Log-linear (logit) models for each of the hematological variables were fit to adjust for age, race, occupation, and smoking history. In addition, all significant group-by-covariate interactions were examined. The covariate of current level of smoking (used in the adjusted continuous analyses described below) was not included in the categorical analyses to avoid problems with sparse cells. Adjusted relative risks for Ranch Hand-Comparison contrasts were calculated for the categories of abnormally low values versus normal values and for abnormally high values versus normal values. Adjusted relative risks were not computed for the abnormally high versus normal categories for HCT, or for the abnormally low versus normal categories for PLT, due to sparse data. The results of these analyses are given in Table 16-5 and were quite similar to the unadjusted results, with no statistically significant or borderline significant associations found. TABLE 16-5. Adjusted Categorical Analyses for Hematological Variables (Abnormal Versus Normal), Adjusted for Age, Race, Occupation, and Smoking Abnormally Low vs. Normal Variable Abnormally High vs. Normal Adj. Relative Risk (95% C.I.) Adj. Relative Risk (95% C.I.) p-Value p-Value RBC 0.93 (0.59,1.47). 0.762 1.04 (0.48,2.28) 0.920 WBC 0.96 (0.66,1.42) 0.854 0.97 (0.69,1.36) 0.852 HGB 1.12 (0.74,1.80) 0.522 0.98 (0.80,1.19) 0.824 HCT 1.02 (0.51,2.06)' 0.954" MCV 1.08 (0.52,2.26) 0.787 0.99 (0.78,1.26) 0.960 MCH 1.33 (0.56,3.17) 0.525 1.10 (0.79,1.54) 0.574 1.14 (0.66,1.98)" 0.638b PLT "Abnormally low versus normal/abnormally high. —Not analyzed due to sparse data. b Abnormally high versus normal/abnormally low. 16-8 �Adjusted Analyses of Continuous Data General linear regression models were performed, adjusting for age (at the Baseline examination), race, occupation (OCC), smoking history (packyears (PACKYRJ), and current level of smoking (cigarettes per day [CSMOK]). The linear models were fit to examine the main effects of group (GRP) membership, the covariates, and two- and three-factor interactions among these variables (only three-factor interactions involving group were considered). The hierarchical modeling approach as described in Chapter 7, Statistical Methods, was performed to arrive at a "best model" containing the group effect and all statistically significant covariate main effects and interactions. The results of the adjusted analyses for the hematological variables, along with the significance of the adjusting covariates and covariate interactions are summarized in Table 16-6. These results indicated a lack of significant group differences for RBC, HGB, HCT, MCV, MCH, and MCHC after adjustment for five covariates. Two analyses, WBC and PLT, showed significant group-by-covariate interactions; the statistics of these interactions (along with borderline interactions for RBC) are given in Table N-2 of Appendix N, and the narrative descriptions of these interactions are included in the following variable-by-variable summary presentations. Discussion The following variable-by-variable discussion presents the findings for the unadjusted and adjusted results, main covariate effects, group-associated interactions, and when appropriate, Ranch Hand versus Original Comparison contrasts, and comparisons to Baseline results. The results of the covariate effects and covariate interactions (not involving group) for the adjusted analyses are found in Table 16-6; group-by-covariate interactions are given in Table N-2 of Appendix N. Red Blood Cell Count (RBC) Both the categorical and continuous unadjusted analyses found no statistically significant differences in RBC values between groups. The covariate associations for both groups combined showed a significant effect of age (RBC abnormally low in 4.0% of the older cohort versus 1.9% of the younger; p=0.010) and race (Blacks having 6.3% and 4.2% in the abnormally low and high categories versus 2.9% and 0.6% in nonblacks, respectively; p<0.001). Continuous regression analyses also detected significant effects of current smoking (p=0.004) and an age-by-occupation interaction (p=0.013). The adjusted categorical analysis showed no significant group difference, but the adjusted continuous analysis revealed a borderline significant (p=0.086) three-factor interaction of group-by-occupation-by-smoking history. Estimated Ranch Hand-Comparison contrasts revealed a significant difference (p=0.010) for enlisted groundcrew, 30 pack-years with Ranch Hands exhibiting a slightly lower RBC count than the Comparisons (see Table N-2 of Appendix N). 16-9 �TABLE 16-6. Adjusted Continuous Analyses for Hematological Variables, (Ranch Hand-Comparison Group Differences) Variable RBC WBC Ranch Hand-Comparison Group Difference ± SE p-Value 0.172 **** AGE*OCC (p=0.013) CSMOK (p=0.004) **** -0.021±0.015a Covariate Remarks* GRP*RACE*AGE (p=0.005) GRP*AGE*PACKYR (p=0.004) GRP*RACE*OCC (p=0.004) HGB -0.034±0.042 0.410 AGE*OCC (p»0.002) RACE*OCC (p=0.013) CSMOK (p<0.001) HCT -0.151±0.121 0.210 AGE*OCC (p=0.004) RACE*OCC (p=0.003) OCC*PACKYR (p»0.035) CSMOK (p<0.001) MCV 0.108±0.188 0.565 RACE*AGE (p<0.001) RACE*OCC (p-0.015) RACE*CSMOK (p=0.025) MCH 0.062±0.070 0.378 RACE*AGE (p=0.015) CSMOK (p<0.001) OCC (p<0.001) MCHC 0.032+0.026 0.226 RACE (p=0.001) CSMOK (p»0.042) PLT **** **** GRP*RACE*PACKYR (p<0.001) GRP*RACE*CSMOK (p-0.024) OCC (p=0.039) AGE (p-0.006) *Abbreviations OCC; Occupation CSMOK: Current level of smoking (cigarettes per day) GRP: Group PACKYR: Smoking history (pack-years) "Also, borderline significant three-factor interaction (see text). ****Group-by-covariate interaction; group difference, standard error (SE) and p-value not presented. 16-10 �A similar, but slightly weaker interaction was observed in the analysis of the Original Comparisons versus the Ranch Hands. The general finding of insignificant group differences supported the Baseline observations (despite the use of different statistical procedures), but the followup results differed by the mild three-factor interaction. White Blood Cell Count (WBC) The categorical and unadjusted continuous analyses did not disclose any significant differences in WBC levels between the Ranch Hand and Comparison groups. Covariate tests showed a borderline effect of age (with the older cohort having a slightly lower proportion of abnormally low WBC levels—4.2% versus 5.4% in the younger cohort), and the highly significant effects of race (p<0.001), occupation (p=0.001), and smoking history (p<0.001). Blacks had a much higher proportion of abnormally low WBC counts (15.42) versus nonblacks (4.0%); higher proportions of enlisted flyers and enlisted groundcrew personnel (9.1% and 7.2%, respectively) had abnormally high WBC counts versus officers (3.6%). Increasing frequencies of leukocytosis were associated with increasing levels of smoking. The adjusted categorical analysis was nonrevealing with respect to group differences, but the adjusted continuous analysis disclosed three significant three-factor interactions involving group membership: group-by-race-by-age (p=0.005), group-by-age-by-smoking history (pack-years; p=0.004), and group-by-race-by-occupation (p=0.004). Further analyses were conducted stratifying by race (see Table N-2 of Appendix N). Among Blacks, the best model revealed significant group-byoccupation and group-by-age interactions (p-0.045, p=0.024, respectively). Group differences for covariate levels corresponding to young officers and young enlisted flyers were statistically significant, with the adjusted mean WBC value considerably lower in the Ranch Hand group than in the Comparison group. Conversely, the adjusted difference for the older enlisted groundcrew was in the opposite direction. The results for nonblacks were more precise: The group-by-age-by-smoking history interaction was highly significant (p=0.002), with young heavy smokers having a WBC level approximately 12 percent greater in the Ranch Hands than the Comparisons. Other differences were small in magnitude and not statistically significant. Ranch Hand and Original Comparison contrasts were similar for nonblacks, but for Blacks, the group-by-occupation and group-by-age interactions did not reach statistical significance (p=0.077, p=0.134, respectively). The nonsignificance of the unadjusted and categorical adjusted analyses was equivalent to the findings at the Baseline examination. However, possibly due to different model selections, no interactions were noted at Baseline. Race and occupation were not used as covariates at Baseline. Hemoglobin (HGB) None of the four analyses, unadjusted and adjusted categorical tests and unadjusted and adjusted tests of mean differences, detected a significant difference between groups. 16-11 �Covariate tests of association revealed the profound effects of race (8.4% abnormally low in Blacks versus 3.3% in nonblacks; p=0.002), occupation (25.1% and 25.6% abnormally high in enlisted flyers and groundcrew, respectively, versus 16.7% in officers; p<0.001), and smoking history (with proportions of abnormally high HGB levels associated with increases in pack-years of smoking; p=0.003). Continuous analyses detected significant effects of current smoking (p<0.001), occupation-by-age (p=0.002), and occupationby-race (p=0.013) interactions. No significant group-by-covariate interactions were noted. Analysis of the Ranch Hands and Original Comparisons, however, found significant three-factor interactions of group-by-race-by-age (p=0.030) and group-by-race-by-occupation (p=0.020) (see Tables N-7 and N-8 of Appendix N). For equivalent analyses, the followup results were quite analogous to the Baseline study results. Hematocrit (HCT) All of the unadjusted and adjusted categorical tests and analyses of mean differences failed to detect any group differences. Since there were only five abnormally high values, this category was combined with the normal category in the categorical analyses. The association of race to HCT was highly significant, with 4.9 percent abnormally low values noted in Blacks versus 0.9 percent in nonblacks (p<0.001). Regression analyses also detected significant effects of current smoking (p<0.001) as well as age-by-occupation (p-0.004), race-by-occupation (p-0.003), and occupation-by-smoking history (p»0.035) interactions. In both categorical and continuous adjusted analyses, no significant group-bycovariate interactions were detected. Analyses of data from the Ranch Hands and Original Comparisons, however, detected significant three-factor interactions of group-by-race-by-age (p=0.026) and group-by-race-by-occupation (p=0.011) (see Tables N-7 and N-8 of Appendix N). Mean Corpuscular Volume (HCV) No significant group differences were detected for MCV abnormalities or mean values by any of the unadjusted or adjusted analyses. Main covariate effects were profound for age (p<0.001), race (p<0.001), occupation (p«0.004), and 'smoking history (p<0.001). The older cohort had a greater frequency of abnormally high MCV values than did the younger age group (18.0% vs. 9.4%, respectively), and Blacks had a far greater frequency of abnormally low MCV values than nonblacks (7.7% vs. 0.6%, respectively). Enlisted groundcrew personnel had a lower percentage of abnormally high values than officers or enlisted flyers (12.5%, 15.5%, and 17.0%, respectively), and increases in pack-years of smoking were associated with increasing percentages of abnormally high levels (0 pack-years: 4.7%; greater than 0 to 10 pack-years: 13.1%; and greater than 10 pack-years: 21.0%). Continuous analyses detected significant interactions of race-by-age (p<0.001), race-by-occupation (p=0.015), and race-by-current.smoking (p=0.025). The analysis of the Ranch Hand and Original Comparisons revealed a significant group-by-race interaction (p=0.031) for the categorical analyses and significant group-by-age-by-smoking history (p=0.041) and group-by-age-by-current smoking (p=0.012) interactions in the continuous 16-12 �analyses. Various contrasts are given in Table N-8 of Appendix N. No explanations are apparent for these interactions except chance. The followup examination results of MCV (i.e., significant interactions) differed from the Baseline results, which showed a significantly larger adjusted mean MCV value in the Ranch Hands. Mean Corpuscular Hemoglobin (MCH) MCH abnormalities and mean values did not differ significantly by group in any of the unadjusted or adjusted analyses. Main effects were very significant for all of the covariates. The older cohort had a greater frequency of abnormally high MCH values than the younger group (7.9% vs. 3.2%, respectively; p<0.001), while Blacks had a greater frequency of low abnormalities than nonblacks (4.9% vs. 0.3%, respectively; p<0.001). Enlisted groundcrew had a higher proportion of abnormalities in the lower range than enlisted flyers and officers (1.0%, 0.3%, 0.2%, respectively), but they had a lower proportion of high-range abnormalities compared to the other occupations (4.3%, 7.8%, and 7.3%, respectively). The overall p-value was 0.003. Increasing pack-years of smoking was associated with increasing frequencies of high abnormal MCH results (0 pack-years: 2.1%; greater than 0 to 10 pack-years: 6.0%; and greater than 10 pack-years: 8.3%; p<0.001). Continuous analyses detected a significant race-by-age interaction (p=0.015), as well as significant effects of current smoking (p<0.001) and occupation (p<0.001). The followup findings did not support the Baseline observation of significantly increased MCH in the Ranch Hands, although the mean was still higher (both unadjusted and adjusted) in the Ranch Hand group. In the analysis of the Ranch Hands and the Original Comparisons, a significant three-factor interaction of group-by-age-by-current smoking emerged (p=0.026). Table N-8 of Appendix N presents Ranch Hand-Comparison differences for selected covariate levels corresponding to 35- and 53-yearold nonsmokers, one-pack-per-day current smokers, and two-packs-per-day current smokers. The differences were positive for all contrasts except the 53-year-old smokers, when the differences became increasingly more negative with increasing levels of smoking. Mean Corpuscular Hemoglobin Concentration (MCHC) In both groups, only one abnormal MCHC count was recorded for either the abnormally low or abnormally high categories, precluding unadjusted or adjusted categorical tests, and exploration of main covariate effects. No significant group differences were detected by the unadjusted or adjusted tests of MCHC means, although race (p=0.001) and current smoking (p=0.042) were significantly associated with MCHC (higher MCHC in nonblacks and decreasing MCHC associated with increasing current levels of smoking). Similar findings were noted in the analysis of Ranch Hand and Original Comparisons, and overall, the followup findings were comparable to the 1982 Baseline MCHC results. 16-13 �Platelet Count (PLT) Neither the unadjusted nor the adjusted categorical analysis showed statistically significant group differences. Analysis of continuous data disclosed significant effects due to occupation (p=0.039), age (p=0.006), group-by-race-by-smoking history (p<0.001), and group-by-race-by-current smoking (p=0.024) interactions, with higher PLT values in the heavily smoking Ranch Hands but similar values for nonsmokers (see Table N-2 of Appendix N). The significant interactions of group-by-race-by-smoking history (p=0.011) and group-by-age (p=0.040) were also noted for the analyses involving the Original Comparisons (see Table N-8 of Appendix N). The percentages of abnormally high PLT counts increased with increasing packyears of smoking (0 pack-years: 0.8%; greater than 0 to 10 pack-years: 2.0%; and greater than 10 pack-years: 2.6%). Other than the interactions encountered in the adjusted analyses, the overall findings at the followup were comparable to the Baseline PLT results. EXPOSURE INDEX ANALYSES Exposure index analyses were conducted within each occupational cohort of the Ranch Hand group to search for dose-response relationships (see Chapter 8 for details on the exposure index). Log-linear models were fit to the categorical data to examine the effects of exposure and pack-years of . smoking, as well as the interaction between these variables. The normal and abnormally high categories were pooled for the RBC count, and the abnormally low and normal response categories were pooled for MCV, MCH, and PLT due to empty cells in some strata. Because of the small numbers of abnormal values,' analyses were not conducted for HCT or MCHC. The results of the unadjusted categorical analyses are presented in Table 16-7, and the counterpart adjusted analyses are given in Table 16-8. The unadjusted analyses showed only a statistically significant result for the WBC count in the enlisted flyer category, due primarily to an excess of abnormally low values in the high exposure category. The very sparse data support a trend from low to high exposure, and the finding of abnormally low WBC counts associated with exposure is in the direction expected for an herbicide effect. However, the exposure association with abnormally low WBC counts converted to borderline significance (p=0.082) in the adjusted analysis. There were no statistically significant exposure level-by-smoking history interactions. Similar analyses in the other occupational strata (with much larger sample sizes) did not produce this pattern. The unadjusted analysis of means for all eight hematological variables was carried out by a one-way analysis of variance. The results are arrayed in Table 16-9. These analyses revealed only one statistically significant result (p=0.038), the RBC count in the enlisted groundcrew stratum where individuals in the medium exposure category had a higher mean RBC level than those in the low or high exposure categories. Thus, these significant RBC findings did not demonstrate a dose-response relationship. The results for HCT in the enlisted groundcrew stratum were of borderline significance (p=0.052) with the highest mean HCT level in the medium exposure category. In contrast to the categorical analyses, mean WBC levels in the enlisted flyers were not significantly different among the three exposure levels. 16-14 �IMBUE 16-7. Unadjusted Categorical Exposure Index Analyses for Hanatological Variables by Occupation Abnormally Low Normal Abnormally High Vari- Occu- Exposure able pation Index Number Percent Number Percent Number Percent Total p-Value 123 125 114 96.8 96.2 93.4 1 1 2 0.8 0.8 1.6 127 130 122 0.522* 54 64 54 98.2 98.5 94.7 0 0 0 0.0 0.0 0.0 55 65 57 0.401a 3.9 1.2 2.8 148 158 136 96.1 97.5 95.8 0 2 2 0.0 1.2 1.4 154 162 142 0.329* 7 7 5 5.5 5.4 4.1 115 118 110 90.6 90.8 90.2 5 5 7 3.9 3.8 5.7 127 130 122 0.919 Low Medium High 0 1 6 00 . 1.6 10.5 51 59 47 92.7 92.2 82.5 4 4 4 7.3 6.2 7.0 55 64 57 0.045 Enlisted Low Groundcrev Medium High 4 8 7 2.6 4.9 4.9 139 142 125 90.3 87.6 88.0 11 12 10 7.1 7.4 7.0 154 162 142 0.839 Officer Low Medium High 7 2 6 5.5 1.5 4.9 100 106 92 78.7 81.5 75.4 20 22 24 15.8 16.9 19.7 127 130 122 0.425 Enlisted Flyer Low Medium High 3 3 5 5.4 4.6 8.8 36 51 36 65.4 78.5 63.2 16 11 16 29.1 16.9 28.1 55 65 57 0.350 Enlisted Low Groundcrew Medium High 5 4 4 3.2 2.5 2.8 119 110 102 77.3 67.9 71.8 30 48 36 19.5 29.6 25.4 154 162 142 0.352 3 4 6 2.4 3.1 4.9 Low Medium High 1 1 3 1.8 1.5 . 5.3 6 2 4 Officer Low Medium High Enlisted Flyer HGB Lov Median High Enlisted Low Groundcrev Medium High WBC Officer Enlisted Flyer FBC 16-15 �TMttE 16-7. (continued) Unadjusted Categorical Exposure Index Analyses for Banatological Variables by Occupation Abnormally Low Normal Abnormally High Vari- Occu- Exposure able pation Index Number Percent Number Percent Number Percent MCV Officer Low Medium High 1 1 0 Low Medium High 0 0 Enlisted Low Groundcrew Medium High 2 Enlisted Flyer 0 p-Value 87.4 85.4 83.6 15 18 20 11.8 13.8 16.4 127 130 122 050 .8" 43 54 47 78.2 83.1 12 11 10 21.8 16.9 17.5 55 0.764b 65 57 139 8.4 16.0 15.5 154 162 142 0.091b 127 0.916b 0.8 0.8 0.0 111 111 00 . 0.0 0.0 Total 102 82.5 133 90.3 82.1 3 2.1 117 82.4 13 26 22 1 0 0 08 . 00 . 00 . 117 121 112 92.1 93.1 91.8 9 9 10 7.1 6.9 8.2 130 122 00 . 00 . 00 . 51 60 54 92.7 92.3 94.7 4 5 3 7.3 7.7 5.3 55 65 57 0.855b Median High 0 0 0 Enlisted Low Groundcrew Medium 1 2 3 0.6 1.2 2.1 147 151 130 95.4 93.2 91.6 6 154 162 142 0.626b 9 9 3.9 5.6 6.3 2 1 0 1.6 08 . 00 . 120 126 119 94.5 97.7 97.5 5 2 3 3.9 1.6 2.5 127 129 122 047 .8" High Low Medium High 1 0 0 1.8 00 . 00 . 51 64 57 92.7 5.4 1.5 00 . 55 65 57 0.135b 98.5 100 0. 3 1 0 Enlisted Low Groundcrew Medium High MCH 3 1.3 1.8 0 1 0 0.0 152 158 140 98.7 97.5 9. 86 2 3 2 1.3 1.8 1.4 154 162 142 0.914b Officer Low Medium High Enlisted Flyer Low High PLT Officer Enlisted Flyer Low Medium 0.6 00 . 'Normal pooled with.abnormally high. "Abnormally low pooled with normal. 16-16 �TABLE 16-8. Adjusted Categorical Exposure Index Analyses (Log-Linear Models) for Hematological Variables by Occupation (p-Values) Variable Occupation Exposure Index Effect* Smoking History Effect** Exposure Index-bySmoking History RBC Officer Enlisted Flyer Enlisted Groundcrev 0.593 0.552 0.310 0.246 0.364 0.515 0.472 0.981 0.717 WBC Officer Enlisted Flyer Enlisted Groundcrew 0.928 0.082 0.761 0.001 0.121 0.009 0.616 0.971 0.104 HGB Officer Enlisted Flyer Enlisted Groundcrew 0.444. 0.413 0.299 0.393 0.647 0.104 0.424 0.980 0.143 MCV Officer Enlisted Flyer Enlisted Groundcrew 0.718 0.619 0.101 <0.001 0.020 0.028 0.334 0.490 0.574 MCH Officer Enlisted Flyer Enlisted Groundcrew 0.852 0.800 0.681 0.002 0.168 0.288 0.777 0.514 0.530 PLT Officer Enlisted Flyer Enlisted Groundcrew 0.410 0.178 0.910 0.099 0.816 0.363 0.708 0.976 0.996 *Adjusted for smoking history (no interaction), **Adjusted for exposure index (no interaction), 16-17 �TABLE 16-9. Unadjusted Continuous Exposure Index Analyses for Hematological Variables by Occupation (Analysis of Variance) Exposure Index Mean ± SE Occupation Variable Low (n-127) Officer RBC WBC* HGB HCT MCV MCH MCHC PLT* 4.904±0.030 6.488 15.468±0.084 45.379+0.243 92.648±0.430 31.606±0.161 34.090±0.060 253.66 (n-55) Enlisted Flyer RBC WBC* HGB HCT MCV MCH MCHC PLTa 4.972±0.048 7.531 15.785±0.149 46.345±0.425 93.269±0.618 31.782±0.222 34.065+0.075 272.87 (n=154) Enlisted RBC Groundcrew WBC* HGB HCT MCV MCH MCHC PLT* 4.990+0.032 7.185 15.566±0.099 45.740±0.284 91.737±0.399 31.251±0.154 34.032±0.059 270.97 Medium (n»130) 4.861±0.029 6.553 15.463+0.087 45.313+0.255 93.260+0.365 31.851+0.134 34.123+0.060 255.70* (n-65) 4.942±0.037 7.236 15.629±0.110 45.908±0.315 92.923±0.501 31.675±0.187 34.058±0.072 275.34° (n=162) 5.094±0.031 7.236 15.807±0.075 46.580±0.210 91.672±0.404 31.138±0.151 33.941±0.051 273.42 High (n=122) 4.899+0.034 6.753 15.593+0.094 45.791±0.284 93.548±0.367 31.884±0.123 34.067±0.059 256.72 n=129. c n=64. 16-18 0.560 0.512 0.507 0.380 0.252 0.314 0.801 0.799 (n-57) 4.957±0.053 6.966 15.721±0.180 46.300±0.535 93.400±0.572 31.735+0.208 33.956±0.072 261.13 0.894 0.378 0.744 0.717 0.817 0.933 0.508 0.382 (n»142) 4.999±0.033 7.389 15.685±0.090 46.086±0.252 92.376±0.458 31.468+0.166 34.031±0.057 268.27 'Standard errors (SE) not presented, since variables were analyzed on logarithmic scale. b p-Value 0.038 0.686 0.147 0.052 0.436 0.325 0.409 0.748 �SUMMARY AND CONCLUSIONS The functional integrity of the hematopoietic system was assessed by the measurement of eight peripheral blood variables: red blood cell count (RBC), white blood cell count (WBC), hemoglobin (HGB), hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and platelet count (PLT). These variables were analyzed in the discrete form to detect differences in the percentages of values outside the designated laboratory range, as well as in the continuous form to detect shifts in mean values between the two groups. A summary of all of these analyses, unadjusted and adjusted for the covariates of age, race, occupation, and smoking, is presented in Table 16-12. The unadjusted discrete analysis of the percent abnormal values, both low and high, showed no statistically significant differences between the Ranch Hand and Comparison groups for any of the hematological variables. Similarly, the adjusted categorical analysis disclosed that none of the adjusted relative risks was significant for either group, and that no significant group-by-covariate interactions were present. The unadjusted continuous analysis did not detect any significant differences in group means for any of the eight variables. The adjusted continuous analysis found no significant group differences for HGB, HCT, MCV, MCH, and MCHC, but encountered significant three-factor interactions for WBC (group-by-race-by-age, group-by-age-by-smoking history, and group-byrace-by-occupation), for PLT (group-by-race-by-smoking history and group-byrace-by-current level of smoking), and a borderline interaction for RBC (group-by-occupation-by-smoking history). Ranch Hand versus Original Comparison analyses revealed further significant interactions for HGB, HCT, MCV, and MCH. As no group strata demonstrated consistent patterns of hematologic impairment, biologic relevance was not assigned to the interactions. The covariate effects of age, race, occupation, and smoking history were highly significant for many of the hematological variables. The effect of race was particularly profound for all variables except PLT. There was fair consistency in the covariate effects upon the RBCrelated variables. Generally, decreasing hematologic values were associated with increasing age and the Black race, and increasing hematologic values were associated with increasing smoking. The detection of these classical covariate effects lends credence to the overall finding of nonsignificant group differences for all of the hematological variables. Significant group differences found for MCV and MCH at the Baseline examination were not significant at the first followup. Other differences (e.g., covariate effects, interactions) between the Baseline and followup examinations may be due to small numeric shifts in the cohorts under study (see Chapter 2) and the selection of alternate statistical models, or due to chance. Unadjusted continuous exposure analyses in the Ranch Hand group revealed only one significant effect (RBC in enlisted groundcrew) and one borderline effect (HCT in enlisted groundcrew), but neither was consistent with a plausible dose-response relationship. The adjusted continuous exposure analyses found only one significant contrast (HCT, medium exposure versus low exposure, enlisted groundcrew). However, seven exposure level-by-covariate interactions were noted for four of the hematological variables. Discrete outcome analyses of the exposure level index revealed a significant result only for WBC in the enlisted flyers. 16-21 �TABLE 16-12. Overall Summary Results of Unadjusted and Adjusted Analyses of Hematological Variables Unadjusted Adjusted Mean Categorical Mean Categorical RBC NS NS NS* NS WBC NS NS **** NS HGB NS NS NS NS HCT NS NS NS NS MCV NS NS NS NS MCH NS NS NS NS MCHC NS PLT NS NS — NS **** — NS NS: Not sgnificant (p>0.10). NS*: Borderline -significant group-by-covariate interaction (0.05<p<0.10). —Analysis not performed due to sparse data. ****Group-by-covariate interaction. Note: Significant group-by-covariate interaction, Ranch Hands versus Original Comparisons only, for HGB, HCT, MCV, and MCH. 16-22 �The longitudinal analyses of MCV, MCH, and PLT found significant differences only for PLT values between the Baseline and followup examinations, with the Baseline group difference in mean values closing to near equivalence at the followup examination. In conclusion, none of the eight hematological variables were found to differ significantly between the Ranch Hand and Comparison groups. In fact, group equivalence was more apparent at the followup examination than at the Baseline examination. The classical effects of age, race, and smoking were demonstrated with most of the hematological variables. The longitudinal analyses also suggested that neither group manifested an impairment of the hematopoietic system. Exposure index analyses did not support a plausible dose-response relationship for any of the hematological variables. 16-23 �CHAPTER 16 REFERENCES 1. Allen, J.R., D.A. Barsotti, J.P. Van Miller, L.J. Abrahamson, and J.J. Lalich. 1977. Morphological changes in monkeys consuming a diet containing low levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Fd. Cosmet. Toxicol. 15:401-410. 2. McConnell, E.E., J.A. Moore, and D.W. Dalgard. 1978. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in Rhesus monkeys (Macaca mulatta) following a single oral dose. Toxicol. Appl. Pharmacol. 43(1):175-187. 3. Weissberg, J.B., and J.G. Zinkl. 1973. Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin upon hemostasis and hematologic function in the rat. Environ. Health Perspect. 5:119-123. 4. Kociba, R.J., P.A. Keeler, C.N. Park, and P.J. Gehring. 1976. 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD): results of a 13-week oral toxicity study in rats. Toxicol. Appl. Pharmacol. 35:553-574. 5. Zinkl, J.G., J.G. Vos, J.A. Moore, and B.N. Gupta. 1973. Hematologic and clinical chemistry effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in laboratory animals. Environ. Health Perspect. 5:111-118. 6. Todd, R.L. 1962. A case of 2,4-D intoxication. J. Iowa Med. Soc. 52:663-664. 7. May, G. 1973. Chloracne from the accidental production of tetrachlorodibenzodioxin. Br. J. Ind. Med. 30:276-283. 8. Pocchiari, F., V. Silano, and A. Zampieri. 1979. Human health effects from accidental release of tetrachlorodibenzo-p-dioxin (TCDD) at Seveso, Italy. Ann. N.Y. Acad. Sci. 320:311-320. 9. Suskind, R.R., and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 10. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5trichloro-phenoxyacetic acid: comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 11. Stehr, P.A., G. Stein, H. Falk, et al. 1986. A pilot epidemiologic study of possible health effects associated with 2,3,7,8-tetrachlorodibenzo-p-dioxin contamination in Missouri. Arch. Environ. Health 41:16-22. 16-24 �12. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, W.F. Schramra, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-terra exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. JAMA 255:2031-2038. 16-25 �CHAPTER 17 RENAL ASSESSMENT INTRODUCTION Renal dysfunction and overt renal disease are not considered to be important clinical sequelae of exposure to phenoxy acids, chlorophenols, or TCDD. In man and animals, 2,4-D, 2,4,5-T, and TCDD are excreted by the kidney, largely in the unmetabolized state via a first-order kinetic process. Excretion of these compounds appears to be a function of the proximal convoluted tubules. " In experimental animals, renal damage is generally noted only when very high or lethal doses of TCDD have been administered, an observation that reflects the severe systemic toxicity of TCDD as contrasted to a doubtful role of primary nephrotoxicity. A variety of experimental pharmacokinetjc studies have been conducted in man using both ingested 2,4-D and 2,4,5-T. (1 ' Most of these studies suggested an unconjugated excretion of these compounds by first-order kinetics. No acute deleterious effects, as detected by urinalysis or blood chemistries, were either noted or recorded for the volunteer subjects. In contrast, following significant exposure to a horse arena filled with TCDD-contaminated waste products, a 6-year-old girl developed hemorrhagic cystitis, pyelonephritis, and proteinuria. Horses exposed to this arena and other contaminated arenas also frequently manifested hematuria. A thorough 5-year followup examination of the young girl was essentially normal and did not reveal any renal sequalae. Most dioxin morbidity studies have only briefly mentioned renal disease and function, and then in the context of routine data collected at physical examination rather than as a specific clinical focus. Some studies of significant occupational exposure have been almost devoid of commentary on renal dysfunction.1 ~1 A contemporary study of a residentially exposed cohort showed negative renal findings. The Times Beach, Missouri, pilot study demonstrated historical "trends" of increased urinary tract disease by questionnaire, along with a compatible pattern of leukocyturia and hematuria manifest at physical examination, but none of the observations was statistically significant. The Monsanto industrial morbidity studies reported essentially negative urinalysis findings, although data were not presented. Baseline Summary Results The 1982 Baseline examination assessed renal disease and function by questionnaire and basic urinalysis testing. 17-1 �Based on questionnaire information, the Ranch Hand group reported significantly more kidney disease than the Comparisons (p=0.039), but this finding was not substantiated by laboratory test results, even when all abnormalities were summed over the five tests of BUN, creatinine clearance, presence of occult blood, five or more urine WBC's per high-power field (HPF), and the presence of urine protein. The Comparison group manifested a twofold increase in proteinuria (p-0.055). The distributions of creatinine clearance levels were similar in both groups, as were the means of the BUN, urine specific gravity, and WBC's/HPF. Difficulty in assessing the degree and significance of hidden noncompliance to the full 24-hour urine collection made the interpretation of the creatinine clearance test results somewhat problematic. Of some interest, known noncompliance to urine collection was observed much more frequently (p<0.001) in the elderly participants. Of 18 herbicide exposure analyses, only 1 (enlisted flyer category) was statistically significant vis-a-vis a history of kidney disease, and it did not demonstrate a linear increase from low to high exposure. The validity of the renal assessment was reinforced by the demonstrated effects of the covariates of age (born in or after 1942, born before 1942) and 2-hour status after postprandial glucose levels (less than 120 mg/dl, greater than or equal to 120 mg/dl). Blood urea nitrogen increased with age and specific gravity decreased (p<0.001 for both), while an abnormally high postprandial glucose level indicative of diabetes was associated only with an increasing urine specific gravity, as expected. Overall, the Baseline renal assessment suggested an excess of historical kidney disease in the Ranch Hand group that was not corroborated by laboratory urinalysis testing. Parameters of the 1985 Renal Assessment Because of the essentially negative Baseline results, the fact that kidney disease is not a prime clinical endpoint, and the manifest compliance problems with a 24-hour urine collection, the 1985 examination process did not emphasize further inquiry into renal disease and function. The onsite NORC questionnaire did not specifically probe for a 1982-1985 interval history of kidney disease, although severe cases may be captured by the generic question, "any other major condition?" or by a detailed extraction of review-of-systerns data obtained at the physical examination. Laboratory testing parameters included all the Baseline dependent variables except the creatinine clearance level (omitted because the plasma creatinine assay was deleted from the test battery). Also, the analysis of composite renal abnormalities was deleted. In addition, the 24-hour urine collection was reduced to a 12-hour collection (5:30 a.m. to 5:30 p.m.) to ease participant burden while still maintaining validity for the porphyrin analyses (see Chapter 13). The accuracy of the 12-hour urine collection was not assessed during the 1985 examination. Renal data analyses paralleled the Baseline analysis except for deleting one of the dependent variables and a composite analysis, adding the covariate of race, and defining the covariate of diabetic class as diabetic, impaired, or normal. No clinical exclusion categories applied to the renal analysis. Minor numerical differences in the tables are due to rare missing dependent 17-2 �variable or covariate data. Adjusted statistical analyses using the above covariates were based on 1,016 Ranch Hands and 1,293 Comparisons and used logistic regression and analysis of covariance methods. When age was used as a covariate in the logistic regression models, the continuous form was used mathematically, but for summary table purposes, age is displayed as a dichotomy. Parallel analyses using the Original Comparisons can be found in Appendix 0 (see Tables 0-3 through 0-5). Tests of association between dependent variables and covariates emphasized Fisher's exact test and Pearson's chi-square test for discrete dependent variables and t-tests and analysis of variance techniques for continuous dependent variables. RESULTS AND DISCUSSION Questionnaire Data History of renal disease was assessed by a self-administered review-ofsystems question list at the physical examination. Specific structured questions on renal disease were not incorporated in the NORC questionnaire. The review-of-systems questions, i.e., "kidney trouble?" "kidney stones?" were open-ended with respect to time, and reflected conditions that arose at any time in the past. These questionnaire data did not show a significant difference between the Ranch Hand and Comparison groups, as reflected by the analysis in Table 17-1. Tests of association between the historical presence of kidney disease in both groups and the covariates of race, occupation, diabetes, and age are given in Table 17-2. TABLE 17-1. Unadjusted Analysis of History of Kidney Disease/Kidney Stones by Group History of Kidney Disease/Stones Yes Group Number No Percent Number Percent Est. Relative Total Risk (95% C.I.) p-Value Ranch Hand 94 9.3 920 90.7 1,014 Comparison 128 9.9 1,163 90.1 1,291 0.93 (0.70,1.23) 0.619 17-3 �TABLE 17-2. Association Between Kidney Disease/Kidney Stones and Age, Race, Occupation, and Diabetic Class in the Combined Ranch Hand and Comparison Groups History of Kidney Disease/Stones Yes Covariate Covariate Category Age Born XL942 Born <1942 66 156 Race Nonblack Black Number Percent Occupation Officer Enlisted Flyer Enlisted Groundcrew Diabetic* Class No Diabetic Impaired Normal Number Percent Total p-Value 6.9 11 .6 894 1,189 93.1 88.4 960 1,345 <0.001a 214 8 9.9 5.6 1,949 134 90.1 94.4 2,163 142 0.106 83 9.6 781 90.4 864 36 9.3 350 90.7 386 103 9.8 952 90.2 1,055 14 41 166 8.0 14.5 9.0 161 242 1,677 92.0 85.5 91.0 175 283 1,843 0.969* 0.011 Fisher's exact test. b Pearson's chi-square test. *Unable to classify four participants, due to missing 2-hour postprandial glucose level and no historical evidence of diabetes. These results showed that there was no significant effect due to race or occupation. In contrast, there was a significant effect due to diabetic class (p=0.011), with participants in the impaired diabetic class having a significantly higher proportion of past kidney disease than those in the normal or diabetic classes-. Older participants also had a significantly higher history of past renal events than younger participants (p<0.001). A logistic regression analysis of the history of kidney disease and kidney stones using the above four covariates gave a result very similar to the unadjusted analysis (Adj. RR: 0.95, 95% C.I.: [0.71,1.25], p=0.693). Race and occupation were not significant covariates. However, diabetic class and age were significant covariates (p=0.041 and p<0.001, respectively). These analyses showed that there was no difference in the history of renal disease between the Ranch Hand and Comparison groups, and that the 17-4 �proportions of past kidney disease and kidney stones were significantly influenced by age and diabetic class. While these findings are consistent with traditional expectations in renal disease, they were in direct contrast to the findings of the 1982 Baseline examination, which revealed a significant excess of historical kidney disease in the Ranch Hand group, and group data that were not influenced by age or glucose levels. It is concluded that there were no significant group differences in past renal disease. Physical Examination Data No physical examination procedures were used to evaluate the renal system as most procedures are invasive and beyond the scope of this voluntary examination. Accordingly, the renal system was evaluated primarily by laboratory data. Laboratory Data Five renal variables were quantitated by general laboratory procedures to assess nonspecific renal system function. The presence or absence of urine protein was determined by standard reagent strip testing. Hematuria and leukocyturia were measured by high-power microscopic examination after centrifugation for 5 minutes. Urine specific gravities were measured by Ames' Multisticks; those urines exceeding normal limits were remeasured by standardized refractometers. BUN levels were assayed by a DuPont Automated Chemical Analyzer, model 500. The SCRF laboratory normal values from these variables are given in Table 17-3. TABLE 17-3. Laboratory Norms for Five Renal Variables Renal Variable Normal Abnormal Urine Protein Occult Blood WBC/HPF BUN (mg/dl) Specific Gravity Absent Absent <2 .7-22 1.005-1.03 Present >1 RBC/HPF >2 >23 <1.004 In this section, urinary protein, hematuria, and leukocyturia were analyzed as discrete variables, whereas BUN and urine specific gravity were analyzed as continuous variables. The number and percent of subjects with abnormal values for the discrete variables are displayed in the summary Table 17-4, along with the number of participants, the unadjusted means, and standard errors of the continuous variables. 17-5 �TABLE 17-4. Summary of Renal Laboratory Variables by Group Group Comparison Ranch Hand Renal Variable Number Abnormal Urine Protein Occult Blood WBC/HPF 37 182 102 Percent Abnormal 3.6 17.9 10.0 Number Abnormal 40 208 107 Percent Abnormal 3.1 16.1 8.3 Unadjusted pr-Value 0.485 0.239 0.145 Renal Variable Unadjusted Mean Standard (Sample Size) Error Unadjusted Mean Standard (Sample Size) Error Unadjusted p-Value BUN (mg/dl) Specific Gravity 14.21* (1,016) 1.0157 (1,016) 0.0002 14.30* (1,293) 1.0152 (1,292) 0.0002 0.554 0.082 *Arithmetic mean calculated on square root scale and transformed to original units. —Standard error not given, since analysis performed on square root scale. The following statistical power statements apply to several variables displayed in Table 17-4. At a standard a -level of 0.05 and a power of 0.80, the sample sizes were sufficient to detect a 1.28-fold increase in the frequency of percent abnormal values for urinary occult blood, and a 1.43-fold increase in the percentage of leukocyturia, both over that observed in the Comparison group. Further, the sample sizes were adequate to reveal a 2.9 percent mean shift in the BUN value relative to the mean observed in the Comparison group. Urinary Protein As displayed in Table 17-4, the Ranch Hand group had a prevalence rate of urinary protein of 3.6 percent versus 3.1 percent in the Comparison group (Est. RR: 1.18, 95* C.I.s [0.75,1.86], p=0.485). This difference was not significant. Tests of association were conducted with pooled participant data using the covariates of race, occupation, diabetic class, and age. These tests are presented in Table 17-5. 17-6 �TABLE 17-5. Association Betveen Urinary Protein and Age, Race, Occupation, and Diabetic Class in the Combined Ranch Hand and Comparison Groups Presence of Urinary Protein Yes No Covariate Covariate Category Number Percent Number Percent Age Born XL942 Born <1942 34 43 3.5 3.2 927 1,304 96.5 96.8 961 1,347 0.641* Race Nonblack Black 65 12 3.0 8.4 2,100 131 97.0 91.6 2,165 143 0.002" Occupation Officer 18 Enlisted Flyer 11 Enlisted Groundcrew 48 2.1 845 97.9 863 2.8 376 97.2 387 4.5 1,010 95.5 1,058 20 18 39 11.4 6.4 2.1 155 264 1,808 88.6 93.6 97.9 175 282 1,847 Diabetic* Class Diabetic Impaired Normal Total p-Value 0.010b <0.001b Fisher's exact test. b Pearson's chi-square test. *Unable to classify four participants, due to missing 2-hour postprandial glucose level and no historical evidence of diabetes. These results suggested no age effect, but significant associations for the covariates of race (p*0.002), occupation (p=0.010), and diabetic class (p<0.001) were noted. The significant covariate effects were attributable to higher percentages of urinary protein abnormalities in Blacks versus nonblacks, enlisted groundcrew versus officers or enlisted flyers, and diabetes (past history [unverified] or greater than or equal to 200 mg/dl glucose) versus impaired glucose tolerance (at least 140 but less than 200 mg/dl glucose) versus normal glucose tolerance (less than 140 mg/dl glucose). The prevalence rates of urinary protein abnormalities were adjusted by logistic regression models using the above four covariates. Race and occupation demonstrated significant effects (p=0.023 and p=0.023, respectively), while age did not (p=0.294). Because of a significant interaction between group and diabetic class (p=0.047), stratified analyses were conducted to provide further clarification. The results are shown in Table 17-6. 17-7 �The adjusted relative risk, 95 percent confidence interval, and group p-value for each diabetic class are shown in Table 17-7. TABLE 17-6. Frequency of Urinary Protein by Diabetic Class and Group Presence of Urinary Protein No Yes Diabetic Class Group Number Percent Number Percent Total Diabetic Ranch Hand Comparison 7 13 9.0 13.4 71 84 91.0 86.6 78 97 Impaired Ranch Hand Comparison 5 13 4.7 7.4 101 163 95.3 92.6 106 176 Normal Ranch Hand Comparison 25 14 3.0 1.4 807 1,001 97.0 98.6 832 1,015 TABLE 17-7. Adjusted Relative Risks for Urinary Protein by Diabetic Class Diabetic Class Diabetic Impaired Normal Adjusted Relative Risk 0.66 0.66 2.23 95% C.I. (0.25, 1.77) (0.23, 1.93) (1.15, 4.32) p-Value 0.414 0.453 0.018 This analysis showed that the estimated prevalence of urinary protein is lower in the Ranch Hand group than in the Comparison group for the diabetic and glucose-impaired strata. Conversely, for the normal diabetic class, the Ranch Hand group manifested a significant increased prevalence of positive urinary protein as contrasted with the Comparison group. These followup examination results were different from the 1982 Baseline examination, which showed significantly more proteinuria in the Comparison group. The prevalence of proteinuria in the followup examination was about 75 percent higher than the prevalence observed in the Baseline study. The interaction of group and diabetic class suggested Ranch Hand increases in proteinuria for normal glucose tolerance participants. 17-8 �Urinary Occult Blood Hematuria was determined by microscopic examination. For both groups combined, the frequency distribution of RBC count data was: 0 RBC/HPF, 82.15 percent; 1-2 RBC/HPF, 15.13 percent; 3-5 RBC/HPF, 2.03 percent; and greater than 5 RBC/HPF, 0.69 percent. As noted in Table 17-4, the prevalence of urinary occult blood in the Ranch Hand group (17.9%) was slightly higher than the rate observed for the Comparison group (16.1%). The unadjusted analysis showed no significant group differences for occult blood (Est. RR: 1.14, 95% C.I.: [0.91,1.42], p=0.239). Tests of association with the covariates of race, occupation, diabetic class, and age were conducted using combined group data for urinary occult blood, and these results are given in Table 17-8. TABLE 17-8. Association Between Urinary Occult Blood and Age, Race, Occupation, and Diabetic Class in the Combined Ranch Hand and Comparison Groups Presence of Urinary Occult Blood Yes Covariate Covariate Category Number No Percent Number Percent Total p-Value Age Born XL942 Born <1942 148 242 15.4 18.0 812 1,105 84.6 82.0 960 1 ,347 0.115a Race Nonblack Black 355 35 16.4 24.5 1,809 108 83.6 75.5 2 ,164 143 0.016* Occupation Officer Enlisted Flyer Enlisted Groundcrew 118 13.7 745 86.3 863 .76 19.6 311 80.4 387 196 18.5 861 81.5 1 ,057 Diabetic Class* 33 52 305 18.9 18.5 16.5 142 229 1,542 81.1 81.5 83.5 175 281 1 ,847 Diabetic Impaired Normal 0.005b 0.296b "Fisher's exact test. Pearson's chi-square test. *Unable to classify four participants, due to missing 2-hour postprandial glucose level and no historical evidence of diabetes. 17-9 �As reflected in Table 17-8, there was no significant effect due to diabetic class or age. However, Blacks had a significantly higher prevalence of urinary occult blood than nonblacks (p=0.016), and significant effects were also due to occupation (p=0.005), with officers having a lower proportion of positive occult blood determinations than enlisted personnel. An adjusted analysis of urinary occult blood proportions was conducted by logistic regression techniques. Multiple significant three-factor interactions were noted, e.g., group-by-occupation-by-race (p=0.008), groupby-age-by-diabetic class (p=0.045), and group-by-occupation-by-diabetic class (p=0.017). Consequently, a series of analyses stratified by race were performed to determine adjusted relative risks for nonblacks and Blacks separately. The adjusted results for nonblack participants are given in Table 17-9. TABLE 17-9. Adjusted Analysis for Urinary Occult Blood for Nonblacks by Group Presence of Urinary Occult Blood Yes Group Ranch Hand Comparison Number Percent 166 189 17.4 15.6 No Number Percent Total 789 1,020 82.6 84.4 955 1,209 Summary Statistics Adj. RR: 1 .13 95% C.I.: (0.91,1.42) p- Value: 0.291 The covariates of occupation and age contributed significant effects (p<0.001 and p=0.002, respectively) to this analysis. Diabetic class was not significant (p=0.863), and was consequently not included in the final model. No significant group differences were found (p=0.291). Table 17-10 shows the frequencies for Black participants. The adjusted analysis of the data on Blacks showed a significant interaction of group and occupation (p=0.003). Table 17-11 presents frequencies and percents for the presence of urinary occult blood for each group, stratified by occupation. This table demonstrates that the group-by-occupation interaction for Blacks was due to the Ranch Hand officers having a lesser prevalence of occult blood abnormalities than Comparison officers, while conversely, Ranch Hand enlisted personnel showed a.higher prevalence of abnormalities than enlisted Comparisons. Because of the absence of hematuria in Black Ranch Hand officers, no relative risk was calculated. Consequently, the Black enlisted occupational categories were combined and investigated further through logistic regression techniques. This analysis did not show a difference of urinary occult blood percentages in the Ranch Hand Black 17-10 �TABLE 17-10. Frequency of Urinary Occult Blood for Blacks by Group Presence of Urinary Occult Blood Yes No Number Percent Group Ranch Hand Comparison 16 19 26.7 22.9 Number Percent 44 64 73.3 77.1 Total 60 83 TABLE 17-11. Frequency of Urinary Occult Blood for Blacks by Occupation and Group Presence of Urinary Occult Blood Yes Occupation Officer Enlisted Flyer Enlisted Groundcrew Group Number No Percent Number Percent Total Ranch Hand Comparison 0 3 0.0 42.9 7 4 100.0 57.1 7 7 Ranch Hand Comparison 3 1 30.0 5.9 7 16 70.0 94.1 10 17 Ranch Hand Comparison 13 15 30.2 25.4 30 44 69.8 74.6 43 59 enlisted and the Comparison Black enlisted strata (Est. RR: 1.62, 95% C.I.: [0.73,3.63], (p«0.239). The effects of age (p=0.817), occupation (p-0.171), and diabetic class (p=0.145) were not statistically significant, and were not included in the final adjusted analysis. In conclusion, both unadjusted and adjusted stratified analyses (by race) did not reveal a consistent and plausible excess of hematuria in the Ranch Hand group. The tenfold or greater increase in the cross-sectional prevalence of hematuria compared to the Baseline examination (1.3% of both groups) to this followup examination may be due to a different sensitivity of the laboratory techniques of reagent-strip testing versus microscopic observation. Nonetheless, an approximate prevalence of 17 percent hematuria merits reevaluation at the next followup examination. 17-11 �Urinary White Blood Cell Count Leukocyturia was assessed by microscopic examination. As noted in Table 17-3, more than two white blood cells per high-power field (WBC/HPF) were considered abnormal by the SCRF laboratory. This is in distinct contrast to the cutpoint of five WBC/HPF used at the Baseline examination. Table 17-4 shows the group frequencies of abnormal urine WBC's. The unadjusted analysis revealed a nonsignificant group effect (Est. RR: 1.24, 95% C.I.i [0.93,1.64], p=0.145). Tests of association were conducted between the frequency of abnormal WBC counts in both groups and the covariates of race, occupation, diabetic class, and age. The results revealed a significantly higher prevalence of abnormal counts for Blacks than nonblacks (p<0.001), an effect due to occupation (p=0.023), with a lower prevalence of abnormalities for officers than enlisted personnel and an effect due to diabetic class (p=0.046), with a lower prevalence of abnormal WBC counts in the normal diabetic class than in either the impaired or diabetic classifications. Age was noncontributory (p«0.508). Adjusted analyses of leukocyturia by group were performed by logistic regression techniques. A significant three-way interaction for group, age, and race was detected (p=0.004), requiring further stratified analyses. A summary of the frequencies £or nonblacks is presented in Table 17-12. TABLE 17-12. Frequency of Urinary WBC/HPF for Nonblacks by Group Urinary WBC/HPF Abnormal Group Ranch Hand Comparison Number 92 88 Normal Percent Number 9.6 7.3 864 1,121 Percent 90.4 92.7 Total 956 1,209 The logistic regression adjustment of the data for nonblacks showed significant covariate effects for occupation (p»0.046) and diabetic class (p=0.031), and a significant interaction between group and age (p=0.018). Consequently, additional analyses were conducted stratifying by age (born in or after 1942, born before 1942), and are shown in Table 17-13. 17-12 �TABLE 17-13. Adjusted Analyses for Urinary WBC/HPF for Nonblacks by Age Category and Group Urinary WBC/HPF Abnormal Age Group Normal Number Percent Number Percent Total Summary Statistics Born XL942 Ranch Hand Comparison 41 24 10.8 4.8 339 478 89.2 95.2 380 502 Adj. RR: 2.42 95* C.I.: (1.43,4.09) p-Value: 0.001 Born <1942 Ranch Hand Comparison 51 64 8.9 9.1 525 643 91.1 90.9 576 707 Adj. RR: 0.99 95* C.I.: (0.67,1-46) p-Value: 0.956 As depicted by the above table, the adjusted rate of nonblack young Ranch Hands with abnormal urinary white blood cell counts was significantly greater than that for nonblack Comparisons (p=0.001 adjusted for occupation and diabetic class). Demonstrating the interaction involving age and group, the adjusted rate of nonblack older Ranch Hands with abnormal urinary WBC counts was nonsignificant and less than older nonblack Comparisons (p=0.956 adjusted for occupation and diabetic class). Similar analyses were conducted for Black participants. Rates of abnormal urinary white blood cell count levels were 16.7 percent and 22.9 percent (n=60 and 83) for Black Ranch Hands and Black Comparisons, respectively. Significant interactions involving group and occupation (p=0.002) and group and age (p^O.OOl) were found. Additional analyses stratified by occupation were performed. Frequencies stratified by occupation are shown in Table 17-14. This table clearly shows how the proportions of WBC abnormalities vary by group within the various occupational categories. However, because of the lack of abnormalities in the Black Ranch Hand officer stratum, an adjusted relative risk was not calculated for this occupation. Thus, Black enlisted categories were combined and subjected to further logistic regression techniques. The analysis showed yet another interaction, between group and age (p=0.026), requiring an additional stratification by age. Results of these analyses are presented in Table 17-15. 17-13 �TABLE 17-14. Frequency of Urinary WBC for Blacks by Occupational Category and Group Urinary WBC/HPF Count Abnormal Occupation Number Percent Group Normal Number Percent Total Officer Ranch Hand Comparison 0 2 0.0 28.6 7 5 100.0 71.4 7 7 Enlisted Flyer Ranch Hand Comparison 3 3 30.0 17.6 7 14 70.0 82.4 10 17 Enlisted Groundcrew Ranch Hand Comparison 7 14 16.3 23.7 36 45 83.7 76.3 43 59 TABLE 17-15. Adjusted Analyses for Urinary WBC/HPF for Black Enlisted Flyers and Groundcrew by Age and Group Urinary WBC/HPF Count Abnormal Age Group Normal Number Percent Number Percent Total Summary Statistics Born XL942 Ranch Hand Comparison 4 13 13.8 28.3 25 33 86.2 71.7 29 Adj. RR: 0.41 46 95% C.I.: (0.12,1.40) p-Value: 0.153 Born <1942 Ranch Hand Comparison 6 4 25 .0 13 ,3 18 26 75 .0 86 .7 24 Adj. RR: 2.17 79) 30 95% C.I. : (0.53,8. p-Value: 0.279 17-14 �In the presence of relatively small sample sizes, these results demonstrated that the prevalence of abnormal urinary white cell counts in Black enlisted personnel did not vary significantly by group for either age category, although the reversal of group proportions for different ages was prominent and fully reflective of the group-by-age interaction. It is noted that the Black group-by-age interaction is opposite the nonblack group-by-age interaction (see Table 17-13), explaining the significant three-way interaction involving group, age, and race. In summary, the unadjusted analysis of urinary WBC/HPF abnormalities showed no group differences, but the adjusted analyses showed significant effects for diabetic class and occupation for nonblack enlisted participants, and a group-by-age interaction for both Black and nonblack enlisted participants. Only for younger nonblack participants was a significant group effect seen (Ranch Hands>Comparisons). The observations from this examination were consistent with the negative Baseline findings. Blood Urea Nitrogen (BUN) BUN was analyzed as a continuous variable using two sample t-tests, analysis of variance, and analysis of covariance techniques. The data were transformed to the square root scale for analysis. Adjusted analyses used the covariates of race, occupation, diabetic class, and age, as in analysis of discrete dependent variables. As noted in Table 17-4, unadjusted group summary statistics revealed no significant differences in mean BUN levels (p=0.554). The groups were combined and contrasted to the covariates, and results are presented below. These tests of covariate association showed a significant racial effect (p=0.007), with a higher mean BUN level for nonblacks than Blacks; a significant effect for occupation (p<0.001), with officers having a higher mean level than both enlisted categories; a significant age effect (p<0.001), with a higher mean BUN level for older than for younger participants; and a marginally significant (p=0.059) difference due to diabetic class, with participants in the impaired category having the highest mean BUN level. An analysis of covariance using the above four covariates demonstrated the significant effects of age (p<0.001), occupation (p=0.015), and significant group-by-race (p=0.022) and race-by-diabetic class (p=0.024) interactions. Table 17-16 presents mean BUN values, adjusted by the covariates and covariate interactions, stratified by race. Test results for the equality of adjusted means between groups are given in the p-value column. As noted from this table, Black Comparisons had a significantly higher adjusted mean BUN level than Black Ranch Hands (p=*0.017), and there was no group difference for nonblacks. These results were analogous to the findings at the Baseline examination (although race was not used as a covariate), i.e., no detriment to the Ranch Hand group and a significant covariate effect of age. 17-15 �TABLE 17-16. Adjusted Analysis of BUN by Race and Group Race Group Total Adjusted Mean* p-Value Nonblack Ranch Hand Comparison 956 1,206 14.15 14.17 0.907 Black Ranch Hand Comparison 60 83 12.40 13.75 0.017 *Converted from square root scale. Urinary Specific Gravity The unadjusted means of the urine specific gravity disclosed a marginally significant difference between the Ranch Hand and Comparison groups (p=0.082). The summary statistics of the unadjusted analysis are given in Table 17-4. By t-tests and analysis of variance, tests of association were performed on the combined groups using the covariates of race, occupation, diabetic class, and age. These tests showed a significant effect of occupation (p<0.001), with officers having the lowest mean urine specific gravity and the enlisted groundcrew category having the highest, and a significant effect (p-0.018) due to diabetic class, with the diabetic category having the highest specific gravity and the normal (nondiabetic) class having the lowest mean value. The effects of age and race were not statistically significant (p=0.382 and p=0.065, respectively). An analysis of covariance with these four covariates showed significant effects due to diabetic class (p=0.019), and significant group-by-race (p=0.017) and group-by-occupation (p=0.034) interactions/ Adjusted group mean specific gravities were stratified by race and by occupation. The results are presented in the summary Table 17-17. These stratified group data showed a difference for nonblack enlisted groundcrew, but Comparisons had a lower adjusted mean urine specific gravity level than Ranch Hands (low specific gravity representing renal dysfunction). Noteworthy is the contrast of results between this followup examination and the Baseline examination in 1982. The urine specific gravities of the followup examination appeared to be very substantially lower than those of the Baseline. A probable explanation was the difference in methods of assessing specific gravity. At the Baseline, the Ames' Clinilab automated procedure (falling drop) was used, as contrasted to the Ames' Multistick procedure at the followup. Both examinations used specimens obtained early on the second examination day, and did not use aliquots of 12- or 24-hour urine collections that were used for the porphyrin analyses. Although the 17-16 �TABLE 17-17. Adjusted Analysis of Urine Specific Gravity by Race, Occupation, and Group Group Total Adjusted Mean p-Value Ranch Hand Comparison 373 474 1.0153 1.0151 0.734 Ranch Hand Comparison 167 193 1.0158 1.0161 0.631 Ranch Hand Enlisted Groundcrew Comparison Race 416 538 1.0174 1.0157 <0.001 Occupation Nonblack Officer Enlisted Flyer Officer Ranch Hand Comparison 7 7 1.0158 1.0186 0.462 Enlisted Flyer Ranch Hand Comparison 10 17 1.0144 1.0158 0.624 Enlisted Groundcrew Black Ranch Hand Comparison 43 59 1.0162 1.0183 0.157 covariate effect of age upon specific gravity was not observed at the followup as it had been at the Baseline, both examinations demonstrated the marked effect of diabetes upon specific gravity, i.e., a higher specific gravity was detected in diabetics than in nondiabetics. EXPOSURE INDEX ANALYSES Exposure index analyses were conducted within each occupational cohort of the Ranch Hand group to search for dose-response relationships (see Chapter 8 for details on the exposure index). The variables of kidney disease, urinary protein, urinary occult blood, and urinary white blood cell count were investigated (unadjusted for any covariates) using Pearson's chi-square test and Fisher's exact test. Adjusted analyses were performed by logistic regression for these variables, using age, race, diabetic class, and any significant pairwise interactions between the exposure index and these covariates. Overall significance in the proportion of abnormalities among the exposure index levels of low, medium, and high was determined, as well as contrasts of the proportion of abnormalities between medium and low exposure levels, and between the high and low exposure levels. Age was used as a continuous variable in the adjusted analyses, and dichotomized (born in or after 1942, born before 1942) when age was involved in an interaction with the exposure index. 17-17 �Analyses of mean blood urea nitrogen and urine specific gravity (continuous variables) were performed, unadjusted for any covariates or interactions, using analysis of variance techniques and t-tests. Analysis of covariance models were used in adjusted analyses. Contrasts of medium versus low exposure and high versus low exposure were also studied. A square root transformation was applied to the 'blood urea nitrogen data. Results of the adjusted analyses for these six variables are presented in Tables 17-18 and 17-19, and counterpart results for unadjusted analyses are presented in Table 0-1 of Appendix 0. Results from further investigation of exposure index-by-covariate interactions are given in Table 0-2 of Appendix 0. Unadjusted analyses revealed no significant differences among exposure index levels for any occupation. Further investigation of these variables, for which the medium versus low and the high versus low contrasts were also examined, revealed only two variables having borderline significance: kidney disease in enlisted flyers, high versus low (Est. RR: 0.25, 95% C.I.: [0.05,1.26], psO.091), and urinary occult blood in enlisted groundcrew, high versus low (Est. RR: 1.77, 95% C.I.: [1.00,3.13], p=0.061). The results for urinary occult blood in enlisted groundcrew supported an increase in the proportion of abnormalities from low to high exposure, whereas the kidney disease data showed the opposite effect. The frequency of abnormalities (or mean levels closer to the abnormal range for continuous variables) for the different exposure index levels exhibited no graduated pattern across exposure levels. The number of combinations for which the medium exposure level had the smallest proportion of abnormalities (or more abnormal mean level) was greater than the other exposure levels. Adjusted analyses revealed no significant differences among exposure index levels for any occupational stratum. Interactions were present for four of the six variables, however, and were observed in all occupations. A summary of these interactions is presented in Table 17-20. No interaction patterns in either the covariates or occupations were observed. The only contrast observed approaching significance for an adverse effect at higher exposure levels was observed for urinary protein (officers in normal diabetic class, high versus low, p=0.097), but this contrast was highly affected by sparse cell sizes (see Table 0-2 of Appendix 0). In summary, six renal variables showed no evidence of an increasing dose-response relationship at the followup examination. No patterns in the relationship of prevalence rates among the exposure index levels were seen within occupational strata. The exposure index level patterns observed at the Baseline examination for kidney disease in the enlisted flyer stratum were not seen at the first followup examination. Overall, both the Baseline and followup examinations showed very little evidence of a dose-response relationship. 17-18 �TABLE 17-18. Adjusted Categorical Exposure Index Analyses for Renal Variables by Occupation Exposure Index Variable Occupation Officer Low Total 127 Kidney Disease Enlisted Flyer • Enlisted Groundcrew 153 Officer 127 Urinary Protein Enlisted Flyer Enlisted Groundcrew 55 55 154 Medium Total 130 65 163 130 65 163 High Total 123 57 141 123 57 142 Contrast Adj. Relative Risk ( 5 C.I.) 9% p-Value Overall M vs. L H vs. L 0.93 (0.37,2.34) 1.67 (0.72,3.88) 0.314 0.878 0.236 Overall M vs. L H vs. L 1.05 (0.34,3.22) 0.26 ( . 5 1 3 ) 00,.1 0.124 0.935 0.102 Overall M vs. L H vs. L 0.57 (0.26,1.26) 0.58 (0.25,1.31) 0.269 0.163 0.189 Overall M vs. L H vs. L **() **! **() **! **() **! **() **! **() **! Overall M vs. L H vs. L 0.34 (0.03,4.61) 0.41 (0.03,4.99) 0.657 0.420 0.486 Overall M vs. L H vs. L ****(2) ****(2) ****(2) ****(2) ****(2) �TABLE 17-18. (continued) Adjusted Categorical Exposure Index Analyses for Renal Variables by Occupation Exposure Index Variable Occupation Officer Urinary Occult Blood Enlisted Flyer Low Total 127 55 Medium Total 130 65 High Total 123 57 Contrast Adj. Relative Risk (95% C.I.) p-Value Overall H vs. L B vs. L 0.80 (0.38,1.71) 1.40 ( . 0 2 8 ) 07,.0 0.299 0.566 0.345 Overall M vs. L H vs. L 0.97 ( . 9 2 4 ) 03,.3 0.43 (0.15,1.24) 0.187 0.950 0.118 Enlisted Groundcrew Urinary Vhite Blood Cell 154 163 141 Overall M vs. L H vs. L ****(3) ****(3) ****(3) ****(3) ****(3) Officer i o NJ 127 130 123 Overall H vs. L H vs. L 0.55 (0.20,1.51) 0.85 (0.34,2.10) 0.488 0.247 0.718 Overall M vs. L H vs. L * * ( , 3) **! * * ( , 3) **! * * ( , 3) **! * * ( , 3) **! * * ( , 3) **! Overall M vs. L B vs. L 0.68 (0.33,1.38) 1.05 (0.53,2.08) 0.424 0.284 0.886 Enlisted Flyer Count Enlisted Groundcrew 55 154 65 163 57 142 * * ( ) exposure index-by-diabetic class interaction — relative risk, confidence interval, and p-value not **!: presented. ****(2): exposure index-by-race interaction — relative risk, confidence interval, and p-value not presented. ****(3): exposure index-by-age interaction — relative risk, confidence interval, and p-value not presented. ****(!,3): exposure index-by-diabetic class and exposure index-by-age interaction — relative risk, confidence interval, and p-value not presented. �TABLE 17-19. Adjusted Continuous Exposure Index Analyses for Renal Variables Exposure Index Blood Urea Nitrogen i to Occupation Statistic Low Medium High Contrast p-Value Officer Variable n Adj . Mean 95% C.I. 127 ****(2) ****(2) 130 ****(2) ****(2) 123 ****(2) ****(2) Overall M vs. L H vs. L ****(2) ****(2) ****(2) Enlisted Flyer n Adj , Mean 95% C.I. 55 13.59 (12.02, 15.26) 65 13.76 (12.32, 15.27) 57 13.77 (12.30, 15.32) Overall M vs. L H vs. L 0.961 0.808 0.804 Enlisted Groundcrev n Adj . Mean 95% C.I. 154 13.31 (12.50, 14.15) 163 13.18 (12.41, 13.98) 142 13.08 (12.30, 13.88) Overall M vs. L H vs. L 0.829 0.722 0.544 �TABLE 17-19. (continued) Adjusted Continuous Exposure Index Analyses for Renal Variables Exposure Index l-» 1 Statistic Low Medium High Contrast p-Value n Ad j . Mean 95% C.I. 127 1.0161 (1.0131, 1.0191) 130 1.0167 (1.0138, 1.0197) 123 1.0165 (1.0136, 1.0194) Overall M vs. L H vs. L 0.755 0.457 0.647 Enlisted Flyer n Adj. Mean 95% C.I. 55 1.0159 (1.0128, 1.0191) 65 1.0157 (1.0129, 1.0185) 57 1.0142 (1.0113, 1.0171) Overall M vs. L H vs. L 0.378 0.861 0.205 Enlisted Groundcrev Urine Specific Gravity Occupation Officer Variable n Adj . Mean 95% C.I. 154 1.0166 (1.0148, 1.0184) 163 1.0166 (1.0149, 1.0183) 142 1.0164 (1.0147, 1.0182) Overall M vs. L H vs. L 0.974 0.976 0.854 NJ NJ * * ( ) exposure index-by-race interaction — adjusted mean, confidence interval, and p-value not presented. **2: �TABLE 17-20. Summary of Exposure Index-by-Covariate Interactions for Renal Variables Variable Occupation Covariate p-Value Urinary Protein Urinary Protein Urinary Occult Blood Urinary White Blood Cell Count . Urinary White Blood Cell Count Blood Urea Nitrogen Officer Enlisted Groundcrew • Enlisted Groundcrew Diabetic Class Race Age 0.004 0.023 0.032 Enlisted Flyer Age 0.015 Enlisted Flyer Officer Diabetic Class Race 0.029 0.009 LONGITUDINAL ANALYSES One variable, the BUN level, was used to assess longitudinal differences between the 1982 Baseline examination and the 1985 followup examination. This variable was selected from the five renal assays because it was judged that serial BUN levels would be more indicative of long-term renal health than the others; further, both examination measurements were made by the same high-precision automated analyzer, permitting a more valid comparison. Other commentary, contrasting general results of the other four renal variables to the Baseline, has been made for each variable above. BUN was analyzed as a continuous variable by repeated measurements analysis of variance (see Chapter 7, Statistical Methods). A square root transformation was used. The data were not adjusted by covariates. The sample base for this analysis was the number of participants who attended both examinations; the results are given in Table 17-21. These data indicated a slight and relatively symmetrical increase in the BUN level in both groups. Based upon longitudinal analyses of BUN, there was no evidence to assert a detriment in the renal health of the Ranch Hand group. SUMMARY AND CONCLUSIONS A summary of all renal variables, including unadjusted and adjusted analyses, is displayed in Table 17-22. 17-23 �TABLE 17-21. Longitudinal Analysis of BUN: A Contrast of Baseline and First Follovup Examination Laboratory Means BUN Means Group 1982 Baseline 1985 Followup Total Ranch Hand 13.72 14.21 971 Comparison 13.93 14.30 1,139 p-Value (Equality of Difference) 0.48 TABLE 17-22. Overall Summary Results of Unadjusted and Adjusted Analyses for Renal Variables Variable Unadjusted Adjusted Reported Kidney Disease NS NS Urinary Protein NS **** Urinary Occult Blood NS **** Urinary Leukocytosis NS **** BUN NS **** Urine Specific Gravity NS* **** NS: Not significant (p>0.10). NS*: Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. 17-24 �A historical assessment of kidney disease/kidney stones by a review-ofsysteras questionnaire showed no significant differences between the Ranch Hand and Comparison groups. An adjusted analysis did not alter this conclusion as an adjusted relative risk of 0.95 (95* C.I.: [0.71,1.25], p=0.693) was demonstrated. These statistics appeared to be in marked contrast to the Baseline historical findings. Differences vis-a-vis the Baseline were most likely due to a difference in questionnaire techniques. Current renal function was evaluated by five laboratory variables: urine protein, occult blood, urine, white blood cell counts (WBC's), blood urea nitrogen (BUN), and urine specific gravity. Invasive procedures were not used. The unadjusted analysis of proteinuria showed no group differences (Est. RR: 1.18, 95% C.I.: [0.75,1.86], p=0.485), but the adjusted analysis showed an interaction of group and diabetic class; appropriate stratified analyses revealed that the prevalence of proteinuria was lower in the Ranch Hands than in the Comparisons in the diabetic and impaired strata, but higher in the normal strata for the Ranch Hands. These results were in contrast to the Baseline findings, which showed a marginally significant proteinuria in the Comparison group (p=0.055), and overall, lower prevalence rates of proteinuria. The unadjusted prevalence rates for hematuria were similar for both groups (Est. RR: 1.14, 95% C.I.: [0.91,1.42], p=0.239). Three significant interactions involving group membership and covariates precluded a direct adjusted comparison of the estimated prevalence rates. Covariate analyses indicated increased hematuria in Blacks and among enlisted personnel. Ultimately via a series of stratified analyses, statistical equivalence was determined for the Black enlisted strata of both groups. Of particular note was the approximate tenfold increase in hematuria in both groups over that observed at Baseline, a finding most likely due to different laboratory techniques (reagent-strip testing versus microscopic observation). Similar results were found for leukocyturia, i.e., a nonsignificant unadjusted analysis (Est. RR: 1.24, 95% C.I.: [0.93,1.64], p=0.145), and a significant three-way interaction (group, age, race) in the adjusted analysis. Significant covariate effects were noted for diabetic class and occupation for nonblack participants, whereas age was a significant adjusting variable for Blacks. A significant group difference was found only for the younger, nonblack Ranch Hands. The overall results were consistent with the Baseline findings. BUN levels did not vary significantly by group (p=0.554, unadjusted). Adjusted analyses showed significant covariate effects for age and occupation and interactions for group and race and for race and diabetic class. An analysis stratified by race revealed no significant group differences for nonblacks, but a significantly higher adjusted mean BUN level in Black Comparisons than in Black Ranch Hands. Overall, the BUN results were similar to those observed at the Baseline examination. Urine specific gravity levels manifested marginally significant group differences (p=0.082, unadjusted). The adjusted analysis disclosed significant covariate effects of diabetic class and the interactions of group and race and group and occupation. < Analyses by race showed no strata with significantly lower mean levels for Ranch Hands. In contrast to the Baseline 17-25 �values, the followup urine specific gravities were lower, a finding most likely attributable to differences in laboratory methodology (falling drop method versus multistick procedure). Exposure index relationship at the prevalence rates or within occupational analyses showed very little evidence of a dose-response followup examination. No patterns in the relationship of mean levels among the exposure index levels were seen strata. The longitudinal analysis was based solely upon a contrast of BUN levels between the two examinations. The unadjusted mean BUN value increased slightly from the Baseline to the followup examination, but the increases were symmetrical in the two groups and nonsignificant (p=0.48). In conclusion, none of the six renal assessment variables showed a significant difference between the Ranch Hand and Comparison groups by unadjusted tests. However, in the adjusted analyses, all renal measurements except reported kidney disease revealed group-by-covariate interactions. These interactions were often complex, making it impossible to reach a firm conclusion as to the presence of an herbicide effect. 17-26 �CHAPTER 17 REFERENCES 1. St. John, L.E., D.G. Wagner, and D.J. Lisk. 1964. kuron, silvex, and 2,4,5-T in the dairy cow. 47:1267-1270. 2. Erne, K. 1966. Fate of atrazine, J. Dairy Sci. Studies on the animal metabolism of phenoxyacetic herbicides. Acta Vet. Scand. 7:264-271. 3. Matsumura, A. 1970. man. The fate of 2,4,5-trichlorophenoxyacetic acid in Jap. J. Environ. Health 12:20-25. 4. Gehring, P.J., C.G. Kramer, B.A. Schwetz, J.Q. Rose, and V.K. Rove. 1973. The fate of 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) following oral administration to man. Toxicol. Appl. Pharmacol. 26:352-361. 5. Kohli, J.D., R.N. Khanna, B.N. Gupta, M.M. Dhar, J.S. Tandon, and K.P. Sircar. 1974. Absorption and excretion of 2,4,5-trichlorophenoxyacetic acid in man. Arch. Int. Pharmacodyn. 210:250-255. 6. Bjorklund, N.E., and K. Erne. 1971. Phenoxy-acid-induced renal changes in the chicken: I. Ultrastructure. Acta Vet. Scand. 12:243-256. 7. Fowler, B.A., G.E.R Hook, and G.tf. Lucier. 1977. Tetrachloro-dibenzo-pdioxin induction of renal microsomal enzyme systems: Ultrastructural effects on pars recta (S,) proximal tubule cells of the rat kidney. J. Pharmacol. Exp. Ther. 203(3):712-721. 8. Koschier, F.J., and M. Acara. 1979. Transport of 2,4,5-trichlorophenoxyacetate in the isolated, perfused rat kidney. J. Pharmacol. Exp. Ther. 208:287-293. 9. Gupta, B.N., J.G. Vos, J.A. Moore, J.G. Zinkl, and B.C. Bullock. 1973. Pathological effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in laboratory animals. Environ. Health Persp. 5:125-140. 10. Pegg, D.G., W.R. Hewitt, K.M. McCormack, and J.B. Hook. 1976. Effect of 2,3,7,8-tetrachlorodibenzo-p-dioxin on renal function in the rat. J^ Toxicol. Environ. Health 2:55-65. 11. Courtney, K.D., J.P. Putnam, and J.E. Andres. 1978. Metabolic studies with TCDD (dioxin) treated rats. Arch. Environ. Contain. Toxicol. 7(4):385-396. 12. Hook, J.B., K.M. McCormack, and W.M Kluwe. 1978. Renal effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin. In Pentachlorophenol; Chemistry, pharmacology and environmental toxicology^ed. K.R. Rao, pp. 381-388.New York:Plenum Press. 17-27 �13. Kohli, J.D., R.N. Khanna, B.N. Gupta, M.M. Dhar, J.S. Tandon, and K.P. Sircar. 1974. Absorption and excretion of 2,4-dichlorophenoxyacetic acid in man. Xenobiotica 4(2):97-100. 14. Sauerhoff, M.W., W.H. Braun, G.E. Blau, and P.J. Gehring. 1977. The fate of 2,4-dichlorophenoxyacetic acid (2,4-D) following oral administration to man. Toxicology 8:3-11. 15. Carter, C.D., R.D. Kimbrough, J.A. Liddle, R.E. Cline, M.M. Zack, W.F. Barthel, R.E. Koehler, and P.E. Phillips. 1975. Tetrachlorodibenzodioxin: An accidental poisoning episode in horse arenas. Science 188(4189):738-740. 16. Beale, M.G., W.T. Shearer, M.M. Karl, and A.M. Robson. effects of dioxin exposure. Lancet 1(8014):748. 1977. Long-term 17. Poland, A.P., D. Smith, G. Metter, and P. Fossick. 1971. A health survey of workers in a 2,4-D and 2,4,5-T plant, with special attention to chloracne, porphyria cutanea tarda, and psychologic parameters. Arch. Environ. Health 22(3):316-327. 18. Oliver, R.M, 1975. Toxic effects of 2,3,7,8-tetrachloro-dibenzo-l, 4-dioxin in laboratory workers. Br. J. Ind. Med. 32:46-53. 19. Pazderova-Vejlupkova, J., M. Nemcova, J. Pickova, L. Jirasek, and E. Lukas. 1981. The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men. Arch. Environ. Health 36:5-11. 20. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, W.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. JAMA 255:2031-2038. 21. Stehr, P.A., G. Stein, H. Falk, et al. 1986. A pilot epidemiologic study of possible health effects associated with 2,3,7,8-tetrachlorodibenzo-p-dioxin contamination in Missouri. Arch. Environ. Health 41:16-22. 22. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichloro-phenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 23. Suskind, R.R., and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 17-28 �CHAPTER 18 ENDOCRINE ASSESSMENT INTRODUCTION The human endocrine system is generally not thought to be influenced by chlorophenol or TCDD exposure. This is not so in animals, however. A wide range of endocrine abnormalities in many animal species has been induced experimentally by TCDD, and includes hypoglycemia, hypothyroxinemia, ' reduced progesterone levels,3 and increased testosterone levels, the latter presumably reflecting decreased liver catabolism due to parenchymal liver damage or an inhibition of the cytochrome P-450 system.4 Further, thymic atrophy, one of the most sensitive indicators of TCDD toxicity, has been shown not to be mediated by the pituitary-adrenal axis. Comparable animal data for the isolated effects of 2,4-D and 2,4,5-T have been noticeably meager. Other animal studies have emphasized the endocrine system, and thyroid function in particular, as important in causing or ameliorating TCDD toxicity, and not simply as an endpoint response. ' Mounting experimental evidence suggests that both natural and radiation-induced hypothyroidism protect against TCDD lethality and that this favorable process can be quickly reversed by treatments with T4. ' If the protective reaction of hypothyroidism in animals can be extrapolated to humans, it suggests that cases of hypothyroidism or altered patterns of thyroid hormones may aggregate in groups of highly exposed workers (particularly in those with chloracne) and/or, alternatively, that severe sequelae of TCDD exposure may be associated with hyperthyroidism. In fact, such thyroid findings have not been commonly reported in dioxin morbidity studies. Occasional cases of hypothyroidism and thyromegaly have been linked to exposures to polybrominated biphenyls and hexachlorobenzene, but the data were too sparse and oblique to support a causal relationship for hypothyroidism and TCDD exposure. °' An assessment of the Times Beach, Missouri, residents, whose community was contaminated with TCDD, did not reveal TSH or T4 differences between the high- and low-risk groups. Temporary glycosuria and impaired glucose tolerance tests were noted in two studies of industrial workers exposed to TCDD. ' However, neither abnormal glucose metabolism nor frank diabetes was specifically noted in other comparable studies. -1 Overall, dioxin morbidity studies have not rigorously assessed the clinical or biochemical parameters of the endocrine system. A detailed description of endocrine function following TCDD exposure was the 1984 AFHS Baseline Morbidity Report, summarized below. 18-1 �Baseline Summary Results The 1982 Baseline examination did not explore historical endocrinological disorders by questionnaire sufficiently to merit analysis. Hence, a comprehensive biochemical assessment of the endocrine system was used for analysis. Five measures of endocrine status were assessed, T3 % Uptake, T4, free thyroxine index (FTI), testosterone, and 2-hour postprandial glucose. Three hormones, follicle stimulating hormone, leutinizing hormone, and cortisol, and correlations of all hormones to various fertility measurements remain for future analysis. Results showed significant group differences for T, % Uptake, predominantly in Ranch Hands 40 years old or less, and abnormally low T % Uptake values, highest for those with high percent body fat. No group difference was noted for elevated 2-hour postprandial glucose values, and as classically expected, the prevalence of abnormal values was associated with older ages and higher percent body fat. Similarly, low testosterone levels were identical in both groups and were associated with increasing age and increasing percent body fat. Higher mean testosterone values (although still within "normal range") were significantly more prevalent in the Ranch Hand group. Significant mean shifts were not noted for the T3 % Uptake, T4, or FTI variable, although the T3 % Uptake was associated with a group-by-age interaction. The exposure index analyses were essentially negative for the T3 % Uptake and T4 variables. FTI, postprandial glucose, and testosterone analyses were marked by a series of covariate interactions in varying occupational categories. Of some note were the significant percent body fat-by-exposure interactions in two occupational strata in the glucose determination. In summary, the endocrine system, as measured by five biochemical assays, did not reveal clinically apparent abnormalities that could be attributed to Herbicide Orange exposure. However, significant mean shifts in several values (although still in normal range) presented trends that were both consistent and conflicting vis-a-vis an herbicide etiology. These data, coupled with the emerging animal literature on the profound influence of the endocrine system on lethality and body fat metabolism following TCDD exposure, clearly underscore the importance of evaluating the endocrine system more comprehensively, as was done in the third-year followup study in 1985. Parameters of the 1985 Endocrine Assessment The 1985 AFHS endocrine test battery was slightly altered from Baseline and included T3 % Uptake, TSH, testosterone, 2-hour postprandial glucose, and timed paired cortisols. The 100 gram glucose load was standardized by a Glucola* challenge (as contrasted to an estimated 100 gram carbohydrate breakfast at Baseline) in preparation for a more definitive assessment of diabetes. Specific questionnaire data on past diabetes and thyroid disease were collected for assessment. 18-2 �Thus, the analyses of endocrine function were comparable to those conducted on Baseline data. Additional refinements included adding diabetes (past and current) as a dependent variable, and the covariates race and personality type, when appropriate. Continuous dependent variables were dichotomized into normal/abnormal categories when necessary using the SCRF values of normal range. Numerous exclusion criteria, e.g., thyroidectomy, orchiectomy, supplemental steroid medication, and diabetes, were used for specific dependent variables. Variations in the numbers of observations in the tables, therefore, reflect these exclusions in addition to rare missing data from the dependent or adjusting variables. Comparable analyses using the Original Comparisons are found in Tables P-4 to P-6 of Appendix P. Log-linear models (BMDP®-4F), general linear models (SAS«-GLM), and logistic regression models (BMDP*-LR) formed the core of the statistical approach. RESULTS AND DISCUSSION Questionnaire Data General screening questions on thyroid function and disease were posed to each participant. Two instruments were used: a self-administered review-of-systerns form containing five questions (e.g., goiter or thyroid trouble, use of thyroid medication?) and the interval health questionnaire with the single question, "thyroid problems?" administered by a trained interviewer. These data are summarized in Table 18-1. Table 18-1 shows that past and current thyroid problems vary according to the interview technique; the group difference in the self-administered questionnaire response was not significant, but the group difference in the interviewer-obtained response was borderline significant. The higher proportion of thyroid disease with the review-of-systerns questionnaire was most likely due to the broader range of prompting questions or interpretation of the questions by the study participant. Since the interviewer-administered questionnaire contained medical provider information for each positive response, verification by medical record review was possible. These data are summarized in Table 18-2 and demonstrated equivalent verification findings in the Ranch Hand and Comparison groups. Thus, the relative absence of reported thyroid disease in the Ranch Hand group appears valid. Physical Examination Data Physical examination of the endocrine system was necessarily limited to manual palpation of the thyroid gland and the testes. Thyroid abnormalities consisted of an enlarged gland with or without nodules or tenderness, while abnormal testes were noted for atrophied glands. The overall palpation results are summarized in Table 18-3. The physical examination data for thyroid abnormalities were clearly supportive of the findings of the questionnaire/review of systems analysis. The proportion of testicular abnormalities (only atrophy represented in the above analysis) was essentially equivalent in both groups. 18-3 �TABLE 18-1. Unadjusted Analysis for Reporting of Thyroid Symptoms/Disease by Questionnaire Method by Group Group Ranch Hand Questionnaire Method Statistic Self-Administered n Diseased3 Not Diseased n Diseased Not Diseased Interviewer Administered Comparison Percent Number Percent Est. Relative Risk (95% C.I.) p-Value* 1,016 48 968 4.7 95.3 1,293 57 1,236 4.4 95 .6 1.08 (0.73,1.59) 0.763 1,016 7 1,009 0.7 99.3 1,293 21 1,272 1.6 98 .4 0.42 (0.18,0.99) 0.054 Number 00 I *Fisher's exact test. a Participants answered positively to having thyroid or goiter trouble, high thyroid level, low thyroid level, lump in throat, or taking thyroid medication. 'Participants answered positively to having thyroid problems since last interviewed. �TABLE 18-2. Medical Record Verification Results of Reported Thyroid Disease by Group Group Ranch Hand Verification Status Comparison Number with Reported Thyroid Conditions 7 21 Medical Records Reviewed 7 21 Medical Records Pending 0 0 100 100 Percent Thyroid Conditions Verified TABLE 18-3. Unadjusted Analysis for Thyroid and Testicular Conditions by Group Group Ranch Hand Comparison Variable Statistic Number Percent Number Percent Thyroid* Est. Relative Risk (95% C.I.) p-Value n 1,015 Abnormal 342 Normal 673 33.7 66.3 1,293 431 862 33.3 1.02 (0.85,1.21) 66.7 0.860 Testicular" n 1,002 Abnormal 26 Normal 976 2.6 97.4 1,289 41 1,248 3.2 0.81 (0.49,1.34) 96.8 0.454 *Thyroidectomies omitted; thyroid abnormal if palpably tender or enlarged, or if nodules present. Orchiectomies omitted; testes abnormal if atrophied (compared to normal). 18-5 �Laboratory Test Data General The collection of relatively scant endocrinological data by questionnaire and physical examination techniques was due to competing priorities of the examination and to the primary reliance upon laboratory testing as established by the 1982 Baseline examination. With research-grade laboratory quality control and reasonably large sample sizes, it was judged that even small mean shifts could be discerned in the test variables. In the presence of corroborating data, these shifts may be ascribed to an herbicide effect if, in fact, one exists. The endocrinological assessment centered upon analysis of laboratory data for T3 % Uptake, TSH, testosterone, timed paired cortisol specimens (the latter three assays conducted by radioimmunoassay [RIA]), 2-hour postprandial glucose, and a composite indicator of past and current diabetes. Normal values of these measurements, as determined by the SCRF Laboratory, are categorized in Table 18-4. It is noted that some of these variables have associated "cutpoints" that differ considerably from those used by the 1982 examining laboratory. Based upon the SCRP laboratory norms, the endocrinological variables distributed into normal and abnormal proportions as displayed in Table 18-5. Unadjusted Ranch Hand and Comparison group means are also provided for quick contrast. TABLE 18-4. Laboratory Endocrinological Variables: SCRF Normal and Abnormal Ranges Variable T3 % Uptake <24% TSH Testosterone Abnormally High 24-32% <7.5 uU/ml <7 ug/dl >7.5 uU/ml 270-1,100 mg/dl >1,100 mg/dl <140 mg/dl <270 mg/dl 2-Hour Postprandial Glucose Cortisol Normal Abnormally Low >140-<200 mg/dl (impaired) >200 mg/dl (diabetic) 7-25 ug/dl >25 ug/dl 18-6 �TABLE 18-5. Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group Group Variable Statistic T3 % Uptake n Mean 95% C.I. Number/% Low Normal High 1,003 27.79 (27.67,27.91) 1,270 27.73 (27.62,27.84) 7 969 27 0.7% 96.6% 2.7% 18 1.4% 1,221 96.1% 31 2.4% n Mean 95% C.I. Number/% Normal High 1,003 1.158 (1.13,1.19) 99.3% 0.7% 1,264 99.5% 6 0.5% n Mean 95% C.I. Number/% Lov Normal High 1,000 597.3 (584.0,610.8) oo i Testosterone 996 7 38 949 13 3.8% 94.9% 1.3% Comparison Est. Relative Risk (95% C.I.) 1,270 1.107 (1.08,1.13) TSH Ranch Hand Contrast — Overall Low vs. Normal High vs. Normal 0.457a 0.49 ( .20,1.18) 0 1.10 (0.65,1.85) 0.248b 0.110C 0.789C 0.019a 1.48 (0.50,4.42) 1,288 578.3 (566.9,589.9) 49 3.8% 1,225 95.1% 14 1.1% p-Value 0.579° 0.035a Overall Lov vs. Normal High vs. Normal 1.00 (0.65,1.54) 1.20 (0.56,2.56) b 0.896C 0.999 0.698C �TABLE 18-5. (continued) Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group Group Ranch Hand Comparison Variable Statistic Initial Cortisol n Mean 95% C.I. Number/% Low Normal High 1,009 11.62 (11.39,11.85) 5.2% 94.2% 0.7% 64 5.0% 1,207 94.0% 13 1.0% n Mean 95% C.I. Number/% Low Normal High 1,009 9.30 (9.10 ,9.51) 0 1,281 3 n Mean 95% C.I. 1,009 2.30 (2.05 ,2.55) 1,284 2.46 (2.24 ,2.69) Est. Relative Risk (95% C.I.) p-Value 1,284 9.27 ( . 0,9.44) 91 0 0.0% 1,005 99.6% 4 0.4% Contrast 2-Hour Cortisol 03 00 Differential Cortisol 52 950 7 1,284 11.68 (11.48,11.89) 0.0% 99.8% 0.2% 0.668a Overall Low vs. Normal High vs. Normal 1.03 (0.71,1.50) 0.68 (0.27,1.72) 0.708b 0.924C 0.501C 0.793a 1.70 (0.38,7.61) 0.706C 0.349a �TABLE 18-5. (continued) Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group Group Variable 2-Hour Postprandial Glucose £ i VO Statistic n Mean 95% C.I. Number/% Normal Impaired Diabetic Diabetes (Composite Indicator) n Number/% Yes No Ranch Hand 976 107.9 (105.9 ,110.0) 836 106 . 34 85.7% 10.9% 3.5% 1,016 74 942 Comparison Contrast Est. Relative Risk (95% C.I.) p-Value 1,235 109.0 (107.3 ,110.7) 0.435a 1,026 83.1% Overall 176 14.3% Impaired vs. Normal 33 2.7% Diabetic vs. Normal 0.74 (0.57,0.96) 1.26 ( . 8 2 0 ) 07,.6 0.038b 0.024° 0.382C 1.09 (0.79,1.50) 0.622° 1,293 7.3% 92.7% 87 6.7% 1,206 93.3% —Relative risk not given for continuous analyses of variables. a t-test. b Chi-square test. c Fisher's exact test. �The following representative statistical power statements (for power 0.8, 2-sided a =0.05) may be applied to parameters of several variables listed in Table 18-5. The sample sizes were sufficient to detect a 1.9-fold increase in the frequency of percent abnormal high values for T3 % Uptake and a 2.5-fold increase in percent abnormal high values for testosterone, relative to that observed in the Comparison group. In addition, the sample sizes were sufficient to detect a 2.7 percent mean shift in TSH and a 1.5 percent mean shift in the first cortisol specimen, over those means observed in the Comparison group. Table 18-5 shows remarkably comparable unadjusted group means and distributional parameters for Ranch Hands and Comparisons in T3 % Uptake, initial cortisol, and 2-hour cortisol. For TSH, testosterone, and 2-hour postprandial glucose, however, there was disparity between the statistical results of the means test and the distributional chi-square test, suggesting that significant differences may exist between the Ranch Hand and Comparison groups. Since all endocrinological variables were known to depend upon classical covariates such as age and race, each variable was reanalyzed by general linear models (using transformations when necessary), logistic regression analyses, or log-linear models adjusted for these covariates. The results of these adjusted analyses are presented in a series of functional endocrine groups below. Table 18-6 presents complete details on the adjusted analyses for all the endocrinological variables. Thyroid Function: T3 % Uptake and Thyroid Stimulating Hormone (TSH) Assessment of both thyroid assays excluded all participants on thyroid medication (as determined by both the self-administered questionnaire and the structured NORC questionnaire) as well as participants with partial or total thyroidectomies. Thus, 13 Ranch Hands and 20 Comparisons were omitted from the following analyses. T3 % Uptake The T3 % Uptake categorical data, as summarized in Table 18-5, were reanalyzed controlling for the covariate effects of occupation, race, age, and personality type. Group data were pooled to reveal the marginal effects of the four covariates. These data are summarized in Table 18-7. The analysis of these data showed a significant effect of occupation (p=0.024) on the percentage of participants with abnormal T3 % Uptake results. Specifically, this was mostly attributable to a relatively high percentage of officers with high T3 % Uptake levels (31 observed versus 21.5 expected, see Table 18-7) and a low percentage of enlisted flyers with high T3 % Uptake results (5 observed versus 9.8 expected). Table 18-7 also shows a marginal effect of personality type on T3 % Uptake results (however, this effect was significant [p=0.035] when analysis was restricted to Ranch Hands and Original Comparisons). Most of the personality-type effect was due to larger numbers than expected of Type A 18-10 �T6HEE 18-6. Adjusted Continuous and Categorical Analyses for Group Variable Statistic T3 %Uj>take n TSH n Adj. Mean 95% C.I. 998 1.158 (1.13,1.19) 1,267 1.109 (1.08,1.14) n 1,003 1,270 n Adj. Mean 95% C.I. 1,000 AArArA' A A AA 1,270 Contrast p-Value Covariate Remarks* 0.250 0.117 0.809 OCC (p=0.025) 0.025 AGE*PERSTYPE(p=O.037) icJcJck 1,003 Comparison Adj. Relative Risk (95% C.I.) 1,287 A1 A"A A Testosterone Ranch Hand Overall Low vs. Normal 0.50 (0.21,1.19) High vs. Normal 1.10 (0.50,2.44) — High vs. Normal 1.48 (0.50,4.42) GRP*BFAT (p=0.024) __ Overall Low vs. Normal 1.00 (0.64,1.55) High vs. Normal 1.13 (0.48,2.64) Initial Cortisol n Adj. Mean 95% C.I. 1,004 11.42 (10.59,12.31) 1,280 11.49 (10.66,12.38) 0.579 — -l-l-l-lr XXXVT A3E*BFAT (p=0.024) RACE (p=0.004) 0.949 0.986 0'.774 ACS (pO.OOl) %BEAT (pO.OOl) 0.659 PGE (p<0.001) %BFAT (p<0.001) PERSTYPE (p=0.002) RACE*OCC (p=0.009) �TABLE 18-6. (continued) Adjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group Group^ Variable Statistic Differential n Adj. Mean Ranch Hand Comparison 1,004 ycMcfc Adj. Relative Risk (95% C.I.) p-Value 1,280 "kXTCK Contrast ™ KTCfCX Covariate Remarks* GRP*AGE*RACE Cortisol (p=0.032) PERSTYPE (p=0.005) %BFAT (p<0.001) Diabetes (Composite Indicator) n Adj. Mean 95%C.I. n 0.487 %BFAT (p<0.001) OCC (p<D.001) AGE*RAOE (p=0.002) Overall Impaired vs. Normal 0.73 (0.56,0.%) Diabetic vs. Normal 1.26 (0.72,2.22) 2-Hour Postprandial Glucose 0.034 0:022 0.421 AGE (pO.OOl) RACE (p=<0.016) JffiFAT (p<0.001) Diseased vs. Normal 1.12 (0.80,1.56) 0.500 %BFAT (p<0.001) AGE*RACE (p=0.005) 976 1,234 115.3 114.4 (107.3,122.0) (108.2,123.0) 1,016 — 1 292 *Abbreviations: GKP: Group OCC: Occupation PEBSTYPE: Personality type (A or B) • %BFAT: Percent body fat —No relative risk or confidence interval given for continuous analyses. ****Group-by-covariate interaction-^Adjusted mean/relative risk, confidence interval, and p-value are not presented. �TABLE 18-7. Association Betveen T % Uptake and Age, Race, Occupation, ana Personality Type in the Combined Ranch Hand and Comparison Groups Percent Abnormal Covariate Covariate Category Total Low High p-Value Age Born XL942 Born <1942 953 1,320 1.05 1.14 2.52 2.58 0.977 Race Black Nonblack 143 2,130 1.40 1.08 1.40 2.63 0.628 Occupation Officer Enlisted Flyer Enlisted Groundcrew 842 383 0.59 1.04 3.68 1.31 0.024 1,048 1.53 2.10 997 1,268 1.60 0.71 2.91 2.21 Personality Type A Direction B Direction 0.071 participants with lower T % Uptake levels. The covariates age and race were not correlated with T, % uptake abnormalities. Log-linear models were then used to assess possible group differences in T3 % Uptake abnormalities, adjusting for occupation (OCC), race, age, and personality type (PERSTYPE). The covariates age, race, and personality type did not contribute significantly to the fit of the adjusted model and were deleted to yield the simplest model, which included occupation. This analysis was summarized in terms of adjusted relative risks and is displayed in Table 18-8. There were no significant differences in percent abnormalities of T3 % Uptake between the Ranch Hand and the Comparison groups. Occupation demonstrated a significant effect (p=0.025). Personality type, although marginally significant (p-0.068), did not affect the assessment of group differences. Thyroid Stimulating Hormone (TSH) TSH laboratory values were analyzed in both discrete and continuous forms. As noted in Table 18-5, an unadjusted t-test of group means showed a statistically significant elevation of TSH in the Ranch Hand group, whereas the categorical analysis did not reveal a statistically significant group difference in the percentage of abnormalities. Exclusion categories and the number of participants were identical to the T3 % Uptake analyses. 18-13 �TABLE 18-8. Adjusted Categorical Analysis for T % Uptake Analysis Contrast Adjusted Relative Risk. 95% C.I. Overall p-Value 0.250 Abnormally Low vs. Normal 0.50 (0.21,1.19) 1.10 (0.50,2.44) Occupation(p=0.025) 0.117 Abnormally High vs. Normal Covariate Remarks 0.809 *Chi-square test (2 d.f.) for group difference. Unadjusted covariate analyses of discrete TSH data from the combined Ranch Hand and Comparison groups showed a borderline significant difference (p=0.071) among occupational groups, with a higher proportion of enlisted flyers with abnormally high TSH levels than observed in the officer or enlisted groundcrew population. The covariates age (born in or after 1942, born before 1942), race, and personality type were nonsignificant. A stepwise logistic regression analysis was performed. The final model was identical to the unadjusted analysis as none of the covariates were significantly associated with TSH. The adjusted percent TSH abnormalities by group were expressed as relative risks. For completeness this summary analysis is shown again in Table 18-9. TSH was subsequently analyzed as a continuous variable. The unadjusted group contrast (determined by a t-test following transformation of TSH values to an inverse square root scale) showed a statistically significant (p=0.019) increase in the mean TSH of the Ranch Hand group, as depicted in Table 18-5. After suitable model fitting, group mean data were adjusted for age (continuous), personality type, and an age-by-personality type interaction. Adjusted results are shown in Table 18-10. As shown, the Ranch Hand TSH mean was significantly elevated over the Comparison group mean after covariate adjustment. However, the group mean values were well below the observed cutoff value of 7.5 uU/ml. The herbicide literature suggests a possibility of primary or secondary hypothyroidism as an endpoint following TCDD exposure. Hypothyroidism, as manifest by the test parameters in this study, should produce a 1 tendency toward depressed T3 % Uptake levels and increased levels of TSH. In the Ranch Hand group, the T3 % Uptake did not indicate hypothyroidism, whereas the TSH mean value showed an increase consistent with hypothyroidism. Questionnaire, physical examination, and laboratory data on thyroid function 18-14 �TABLE 18-9. Adjusted Categorical Analysis for TSH Adjusted Relative Risk 95% C.I. p-Value 1.48 (0.50,4.42) 0.579 TABLE 18-10. Adjusted Continuous Analysis for TSH by Group Group Ranch Hand Total* Adjusted Mean 95% C.I. 998 1.158 (1.13,1.19) p-Value 0.025 Comparison 1,267 1.109 Covariate Remarks Age-by-Personality Type (p=0.037) (1.08,1.14) *Eight participants excluded because of missing data on personality type; 35 participants excluded because of thyroid medication. 18-15 �and disease led to the conclusion that there were no essential differences indicating thyroid disease between the Ranch Hand and the Comparison groups. Testosterone Serum testosterone levels were measured by RIA on all participants. Normal range values from the SCRF Laboratory were used to categorize all data into abnormally low, normal, abnormally high determinations (see Table 18-4). All analyses omitted participants with unilateral or bilateral orchiectomies, and those participants on supplemental testosterone medication. The unadjusted categorical analysis (see Table 18-5) showed no significant differences (p=0.896) in the proportions of abnormalities between the Ranch Hand group and the total Comparison group. The groups were combined and the relationships between categorized testosterone levels and the covariates occupation, race, age, percent body fat (%BFAT), and personality type were examined. Significant statistical differences were noted for occupation (p-0.012), increasing age (p<0.001), and increasing percent body fat (p<0.001). No effect was found due to race or personality type. An adjusted analysis was done to determine the simplest model using the significant covariates, and relative risks were calculated. This analysis is depicted in Table 18-11. These results showed that neither percent low testosterone abnormalities nor percent high testosterone abnormalities were excessive in the Ranch Hand group, as the confidence interval of the adjusted relative risks included the value 1.00. TABLE 18-11. Adjusted Categorical Analysis for Testosterone Analysis Contrast Adjusted Relative Risk 95% C.I. Overall* p-Value Covariate Remarks 0.949 Abnormally Low vs. Normal 1.00 (0.64,1.55) 0.986 Abnormally High vs. Normal 1.13 (0.48,2.64) 0.774 *Chi-square test (2 d.f.) for group difference. 18-16 Age(p<0.001) Percent Body Fat (p<0.001) �In contrast to the negative categorical analyses, the unadjusted test of testosterone means showed a significant elevation in the Ranch Hand group (see Table 18-5). Using similar covariates as in the adjusted categorical analyses, the group means were contrasted by an analysis of covariance. A significant group-by-percent body fat interaction was found (p=0.024). This was due to Ranch Hands having a significantly lower mean than Comparisons (654.4 mg/dl versus 1042.8 mg/dl, p=0.012), for the less than 10 percent body fat category, but a significantly higher mean for the 10 to 25 percent body fat category (603.3 mg/dl versus 582.4 mg/dl), and a nonsignificantly higher mean for the greater than 25 percent body fat category (463.0 mg/dl versus 456.7 mg/dl). However, the number of participants in the less than 10 percent body fat category was very small: six Ranch Hands and four Comparisons, and without these, the overall Ranch Hand mean testosterone level was higher than that for Comparisons. An age-by-percent body fat interaction (p=0.024) and race (p=0.004) were significant covariates. The group interaction is summarized in Table P-l of Appendix P. The adjusted analysis showed a significantly elevated mean testosterone level in the Ranch Hand group for the 10 to 25 percent body fat category, which comprised 80 percent of the Ranch Hand and 79 percent of the Comparison participants, whereas the categorical analyses did not reveal any group differences. These findings might be viewed as supportive of an herbicide effect. Cortisol: Initial, 2-Hour, and Differential Cortisol measurements were obtained in the AFHS for two reasons: as a general indicator of the integrity of the endocrine system (and specifically as a functional measure of the pituitary-adrenal circuit), and as an important secondary risk factor in coronary artery disease (CAD). As cholesterol is a metabolic precursor to cortisol, there has been longstanding scientific interest on cause-and-effect relationships between these substances. Clearly, steroid and ACTH treatments have been implicated in induced hypercholesterolemia and possibly resulting CAD. " Cholesterol elevations have been consistently .noted following exposure to TCDD (see Chapter 15) and, therefore, are of prime interest in this study. Consequently, exploration of the cholesterol-TCDD or cholesterol-CAD relationship must also account for cortisol differences, if any. Timed serum specimens were obtained from all participants at a 2-hour interval early on the second day of the examination. The difference between the timed paired specimens was termed the "differential cortisol." The value of the first specimen was generally higher than the value of the second specimen (due to liver catabolism). The mean values of the two cortisol determinations (initial and 2-hour) for the Ranch Hand and the Comparison groups (as reflected in Table 18-5) did not differ by unadjusted t-tests (p=0.668, p=0.793, respectively), Further, the unadjusted categorical analyses for both specimens based on the normal values of the SCRF Laboratory also did not demonstrate significant group differences (p=0.708, p=0.706, respectively). 18-17 �By an analysis of covariance, using the covariates age, occupation, race, percent body fat, and personality type, the mean value of the initial cortisol specimen was adjusted and contrasted by group. These results are given in Table 18-12, and as indicated, there was no statistically significant group difference. Tests of association between the differential cortisol and the covariates (Table 18-13) disclosed significant effects by percent body fat and personality type (p=0.002, p=0.006, respectively). Age was only slightly suggestive of an effect. An adjusted analysis was performed using the above covariates. A groupby-age-by-race interaction was found (p=0.032). Personality type (p=0.005) and percent body fat (p<0.001) were significant covariates. The interaction found a significantly lower mean differential cortisol level for Black Ranch Hands (p»0.003) born in or after 1942 (unadjusted mean 0.17 ug/dl, adjusted mean -0.46 ug/dl) versus corresponding Comparisons (unadjusted mean 2.78 ug/dl, adjusted mean 2.33 ug/dl); no significant differences were found for older Blacks or nonblacks. The interaction is summarized in Table P-l of Appendix P. The analyses discussed above showed that the Ranch Hand and Comparison groups did not differ with regard to both paired cortisol specimens, and the differential cortisol of those specimens for all nonblacks and Blacks born before 1942. For Blacks born in or after 1942 (32 Ranch Hands, 47 Comparisons) the mean differential cortisol level was lower for Ranch Hands than Comparisons. The mean cortisol levels for each personality type and percent body fat category were plotted over time. Figure 18-1 shows the rate of decrease in cortisol for Type A and Type B personalities, adjusted for percent body fat and age. Similarly, Figure 18-2 shows the rate of decrease in cortisol in three categories of percent body fat, adjusted by personality type and age. The effect of personality type and percent body fat upon the levels of cortisol and the rate of change of cortisol over the 2-hour period are noteworthy. Age was also a significant covariate. Type A personalities began with slightly lower cortisol levels but had a lower rate of decrease of cortisol over the next 2 hours as contrasted to Type B personalities. This analysis demonstrated the ability of the Jenkins Activity Scale to differentiate personality type in this cohort, as measured by differential cortisol levels. The strong effect of percent body fat upon cortisol was not expected. Glucose Metabolism: 2-Hour Postprandial Glucose and Composite Diabetes Indicator The 1985 examination at SCRF presented two major changes in the assessment of glucose metabolism as contrasted to the 1982 Baseline examination: (1) the accepted laboratory criteria by which to diagnose diabetes shifted from the standard of 120 mg/dl or more at 2 hours to a designation of "impaired" glucose tolerance (at least 140 but less than 200 mg/dl) and "diabetic" glucose tolerance (at least 200 mg/dl), and (2) participants were given a standardized 100 gram Glucola* challenge rather than an estimated 100 gram carbohydrate breakfast. Further, most known diabetics were encouraged not to take the Glucola* challenge. 18-18 �TABLE 18-12. Adjusted Continuous Analysis for Initial Cortisol by Group Group Total* Ranch Hand Adjusted Mean 1,004 11.42 0.659 Comparison 1,280 Covariate Remarks p-Value 11.49 Age (p<0.001) Personality Type (p=0.002) Percent Body Fat (p<0.001) Occupation-by-Race (p=0.009) * Nine participants omitted due to missing data on personality type and body fat. TABLE 18-13. Association Betveen Differential Cortisol and Age, Race, Occupation, Percent Body Fat, and Personality Score in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Mean Differential Cortisol Level p-Value Age Born > 1942 Born < 1942 955 1,338 2.24 2.51 0.122s Race Black Nonblack 143 2,150 2.16 2.41 0.575* Occupation Officer Enlisted Flyer Enlisted Groundcrev 852 385 1,056 2.55 2.48 2.23 0.203b <10% 10-25% >25% 10 1,846 436 1.80 2.54 1.79 0.002b 1,002 1,283 2.12 2.60 0.006* Percent Body Fat Personality A Direction Type B Direction "By t-test. b 18-19 L �12- 11.511- 10.5Cortisol Units/ml 10- 9.5- 98.5- 87:30 a.m. 9:30 a.m. Time of Serum Collection D Type A * Type B Figure 18-1. Mean Cortisol Levels by Personality Type, Adjusted for Age and Percent Body Fat, by Time of Specimen Collection 18-20 �12- 11.511- 10.5Cortisol Units/ml 10- 9.598.5- 8- I I 7:30 a.m. 9:30 a.m. Time of Serum Collection D <10% Body Fat x 10-15% Body Fat 0 >25% Body Fat Figure 18-2. Mean Cortisol Levels by Percent Body Fat, Adjusted for Age and Personality Type, by Time of Specimen Collection 18-21 �All participants were provided high carbohydrate menus preceding the examination, and were encouraged to consume high calorie meals for 3 days immediately before their examination to improve the diagnostic efficiency of the glucose tolerance test. At the examination site, compliance or noncompliance to the carbohydrate diet was recorded but reported compliance was not analyzed. These data, however, were not used to exclude participants from the analyses, as the 1984 Baseline Report showed that compliance to the diet was inconsequential to the analyses. All known diabetics, as determined by the Baseline history and the 1982-1985 interval questionnaire, were excluded from the glucose tolerance analyses. However, the 43 Ranch Hands and the 59 Comparisons comprising the exclusion group were included in the composite diabetes analysis. 2-Hour Postprandial Glucose As noted in Table 18-5, a trichotomized contrast of the 2-hour postprandial glucose showed a statistically significant difference (p=0.038) between the Ranch Hand and the Comparison groups. This was due to a slightly higher percentage of Ranch Hands in the diabetic category, and a lower percentage of them in the impaired category relative to the Comparison group. Both study groups were pooled to assess the covariate main effects of age, race, occupation, and personality type. The results showed a significant effect for occupation (p=0.030), largely due to a higher proportion of enlisted flyers having impaired glucose levels. Race, age, and percent body fat were significant covariates (p=0.037, p<0.001, p<0.001, respectively), with Blacks, older ages, and high body fat categories having many more observed abnormalities than nonblack, younger age, and normal body fat categories. Personality type showed no effect (p=0.562). Using the three covariates age, race, and percent body fat, the percent impaired and percent high glucose categories were adjusted and relative risks were calculated. These data are summarized in Table 18-14 and revealed that ' significantly fewer Ranch Hands had impaired glucose levels (at least 140 but less than 200 mg/dl) than did Comparison members, as demonstrated by the fact that the relative risk was bracketed by a confidence interval with upper limit less than 1.00. Conversely, more Ranch Hands had diabetic levels of glucose (at least 200 mg/dl) on the 2-hour postprandial test than did the Comparisons, but this excess was not statistically significant. The 2-hour postprandial glucose level was also analyzed as a continuous variable. Group data were transformed to a logarithmic scale and were adjusted by a general linear model using the covariates age, race, occupation, and percent body fat. This analysis is reflected in Table 18-15. These results showed no group difference for the 2-hour postprandial glucose variable. Significant covariate effects are noted for percent body fat (p<0.001), occupation (p<0.001), and the age-by-race interaction (p*0.002). 18-22 �TABLE 18-14. Adjusted Categorical Analysis for 2-Hour Postprandial Glucose Analysis Contrast Adjusted Relative Risk. 95% C.I. Covariate Remarks p-Value Overall* 0.034 Impaired vs. Normal 0.73 (0.56,0.96) 0.022 Diabetic vs. % Normal 1.26 (0.72,2.22) Age (p<0.001) Race (p=0.016) Percent Body Fat (p<0.001) 0.421 "Chi-square test (2 d.f.) for group difference. TABLE 18-15. Adjusted Continuous Analysis for 2-Hour Postprandial Glucose by Group Group Total Ranch Hand 976 Comparison 1,234 Adjusted Mean 95% C.I. 114.4 (107.3,122.0) p-Value Covariate Remarks Age-by-Race(p*0.002) 0.487 Occupation(p<0.001) 115.3 (108.2,123.0) Percent Body Fat(p<0.001) 18-23 �Composite Diabetes Indicator This variable was constructed by selecting participants with a known history of diabetes via the Baseline or interval (1982-1985) questionnaire, and adding them to the group whose 2-hour postprandial glucose level was at least 200 mg/dl at the 1985 examination. Thus, this pool represents all "true diabetics," past and present. These data were contrasted to the "nondiabetics," recognizing the mild degree of misclassification introduced by considering glucose-impaired individuals as normal. The unadjusted frequencies (Table 18-5) were 7.3 percent diabetics in the Ranch Hand group and 6.7 percent diabetics in the Comparison group (p=0.622). A series of analyses were conducted to determine the best adjusting model for these data, using stepdown procedures from a model containing all main effects and two- or three-way interactions. The final adjustment used the significant covariates of percent body fat and an age-by-race interaction to adjust the proportions of diabetes in each group. These results, formulated as a relative risk, are presented in Table 18-16. The adjusted results indicated no significant difference in the frequency of past and current diabetes in the Ranch Hand and Comparison groups. The analyses above provide a firm platform to conclude that both study groups were essentially equal with respect to glucose metabolism, and past and current diabetes. Although the herbicide literature suggests a possible endpoint of diabetes, this followup study provides no support for that notion. The slight discrepancies between the categorical tests of glucose abnormalities and the assessment of mean values are probably explained on distributional grounds. EXPOSURE INDEX ANALYSES Within each occupational category, exposure index analyses were carried out to assess possible dose-response relationships (see details in Chapter 8). The variables T3 % Uptake, TSH, testosterone, initial and 2-hour cortisol, differential cortisol, and 2-hour postprandial glucose were analyzed as continuous variables by t-tests and analysis of variance (unadjusted by any of the covariates). Adjusted analyses were performed using general linear models; adjusting covariates were age, race, occupation, and as appropriate, percent body fat and personality type. Group-by-covariate interactions were explored for each analysis, and tests were made of differences in means among the three exposure levels as well as contrasts of means between the medium and low exposure levels, and between the high and low exposure levels. The dependent variables were transformed prior to analysis as described earlier in this chapter. TABLE 18-16. Adjusted Analysis for Diabetes (Composite Indicator) Adjusted Relative Risk 95% C.I. p-Value 1.12 (0.80,1.56) 0.500 18-24 Covariate Remarks Age-by-Race ( p»0 . 005 ) Percent Body Fat (p<0.001) �Results of the adjusted analyses are presented in Table 18-17 and parallel results for unadjusted analyses are given in Table P-2 of Appendix P. Results of investigation of any exposure index by covariate interactions are given in Table P-3 of Appendix P. Unadjusted analyses showed significant differences either among exposure levels or in the high versus lov or medium versus low exposure level contrasts for testosterone for officers, and initial cortisol, differential cortisol, and 2-hour postprandial glucose for enlisted flyers. For officers, a significantly lower mean testosterone level was seen for the medium exposure level as contrasted to the low exposure level (547.4 mg/dl versus 599.4 mg/dl, p=0.041). Enlisted flyers had significantly lower mean initial cortisol in the medium as contrasted with low exposure level (11.08 ug/dl versus 11.97 ug/dl, p=0.001); participants in the high exposure level also had a much lower mean, 11.13 ug/dl, as contrasted with the low exposure level but the difference was not significant. Enlisted flyers had a significant difference in'differential cortisol among exposure index levels (p=0.003). The mean differential cortisol levels were 3.43 ug/dl, 1.20 yg/dl, 2.30 ug/dl for the low, medium, and high exposure levels, respectively; the medium versus low contrast was very significant (p<0.001), and the high versus low contrast was marginally significant (p=0.092). Mean 2-hour postprandial glucose for enlisted flyers in the medium exposure category was much higher than in the low category: 118.0 mg/dl versus 100.9 mg/dl (p-0.015). However, the mean glucose level for the high exposure category was not as high as that for the medium level, 110.9 mg/dl. The difference among all the exposure levels was close to significance (p=0.051). Adjusted analyses (Table 18-17) showed patterns very similar to unadjusted analyses. A summary of exposure index by covariate interactions found are listed in Table 18-18. The adjusted mean TSH level for enlisted flyers was significantly higher in the high exposure level as contrasted with the low exposure level (p=0.045); moreover, there was a steady trend upwards with low, medium, and high exposure levels. Enlisted flyers in the medium exposure level had a higher adjusted mean 2-hour cortisol level than the low exposure level (p=0.034), but no trend was apparent. There was a significant difference in differential cortisol among the exposure levels of enlisted flyers (p=0.008) and the medium exposure level had a much lower adjusted mean than the low exposure level (p=0.002). No clear trend with increasing exposure was apparent. Further, enlisted flyers in the medium exposure level had a higher mean postprandial glucose than the lower level (p=0.012), and the overall test for differences among the three levels was significant (p=0.042). In summary, the emergent pattern was that the enlisted flyers in the medium exposure level were significantly different from those in the low exposure level for 2-hour cortisol, differential cortisol, 2-hour postprandial glucose and marginally significantly different (p=0.098) for testosterone. However, the corresponding high versus low contrasts were not statistically significant. LONGITUDINAL ANALYSES Three endocrine variables were chosen for longitudinal analysis: testosterone, T3 % Uptake, and TSH. Only participants attending both examinations were eligible. The three variables were measured by relatively comparable laboratory techniques at the Kelsey-Seybold Laboratory in 1982 and 18-25 �TABLE 18-17. Adjusted Exposure Index Analyses for Endocrinological Variables' by Occupation Exposure Index Variable Medium High Low Officer n Adj . mean 95% C.I. 124 120 Overall 126 28.15 28.17 28.58 M vs. L (27.36,29.01) ( 7 7 , 9 4 ) (27.35,28.98) H vs. L 2.82.1 0.180 0.120 0.928 n Adj . mean 95% C.I. 55 55 65 Overall 27.45 27.62 27.95 M vs. L (26.69,28.24) (26.90,28.36) (27.24,28.68) H vs. L 0.388 0.639 0.178 Enlisted Groundcrew »-» 00 Statistic Enlisted Flyer T3 % Uptake Occupation n Adj . mean 95% C.I. 140 Overall 153 160 27.87 28.00 27.96 M vs. L (27.60,28.41) (27.56,27.96) (27.45,28.30) H vs. L 0.853 0.857 0.579 Officer n Adj . mean 95% C.I. Overall 124 120 126 1.212 1.343 M vs. L 1.263 (1.045,1.555) (1.011,1.479) (1.107, 1.664) H vs. L 0.262 0.513 0.332 Enlisted Flyer n Adj . mean 95% C.I. 55 55 65 Overall 1.058 M vs. L 0.899 1.005 (0.768,1.067) (0.860,1.191) (0.904,1.254) H vs. L 0.120 0.155 0.045 Enlisted Groundcrew n Adj . mean 95% C.I. 140 Overall 153 160 M vs. L 1.135 1.174 1.151 (1.041,1.243) (1.054,1.263) (1.070,1.294) H vs. L 0.807 0.775 0.513 1 N} TSH Contrast p-Value �TABLE 18-17. (continued) Adjusted Exposure Index Analyses for Endocrinological Variables by Occupation Exposure Index 00 Initial Cortisol Statistic Low Medium Officer n Adj. mean 95% C.I. 125 482.6 (414.7,555.5) 128 116 Overall 464.5 461.0 M vs. L (395.8,531.2) (397.9,536.2) H vs. L 0.560 0.312 0.405 n Adj. mean 95% C.I. 55 57 63 Overall M vs. L 507.7 571.3 536.1 (427.6,594.7) (492.7,655.7) (462.6,614.9) H vs. L 0.251 0.098 0.454 Enlisted Groundcrev Testosterone Occupation Enlisted Flyer Variable n Adj. mean 95% C.I. 153 **** **** 161 **** **** 137 **** **** Overall M vs. L H vs. L ****a **** **** Officer n Adj. mean 95% C.I. 124 **** **** 130 **** **** 119 **** **** Overall H vs. L H vs. L ***** **** Enlisted Flyer n Adj. mean 95% C.I. 55 65 11.69 11.11 (10.38,13.17) (9.96,12.39) 57 11.08 (9.97,12.32) Overall M vs. L H vs. L 0.533 0.335 0.320 Enlisted Groundcrew n Adj. mean 95% C.I. 154 11.11 (9.96,12.40) 140 11.01 (9.88,12.27) Overall M vs. L H vs. L 0.948 0.757 0.809 160 10.98 (9.87,12.23) High Contrast p-Value �Adjusted Exposure Index Analyses for Endocrinological Variables by Occupation Exposure Index Variable 2-Hour Cortisol Occupation Officer Enlisted Flyer Enlisted Groundcrev Differential Cortisol Officer Enlisted Flyer Enlisted Groundcrew Statistic Low Medium High Contrast p-Value �TABLE 18-17. (continued) Adjusted Exposure Index Analyses for Endocrinological Variables by Occupation Exposure Index Medium 00 I Statistic Low Officer n Ad j . mean 95% C.I. 121 124 111.1 106.8 (100.8,122.4) (97.1,117.5) n Ad j . mean 95% C.I. Enlisted Groundcrew 2-Hour Postprandial Glucose Occupation Enlisted Flyer Variable n Ad j . mean 95% C.I. High Contrast p-Value 111 108.1 (98.1,119.3) Overall M vs. L H vs. L 0.411 0.191 0.383 54 113.7 (97.9,132.0) 62 56 Overall 134.4 121.9 M vs. L (117.2,154.2) (106.5,139.6) H vs. L 0.042 0.012 0.286 150 107.1 (96.2,119.3) 155 110.4 (99.2,122.7) 138 109.2 (98.2,121.6) Overall M vs. L H vs. L ****Group-by-covariate interaction—adjusted mean, confidence interval, and p-value not given. "Exposure index-by-percent body fat interaction. b Exposure index-by-race interaction. °Exposure index-by-race and exposure index-by-personality type interactions. 0.706 0.413 0.597 �TABLE 18-18. Summary of Exposure Index-by-Covariate Interactions Encountered in Analyses of Endocrinological Variables Variable Occupation Covariate p-Value Testosterone Enlisted Groundcrew Percent Body Fat 0.001 Initial Cortisol Officer Percent Body Fat 0.037 2-Hour Cortisol Officer Percent Body Fat 0.011 2-Hour Cortisol Enlisted Groundcrew Differential Cortisol Enlisted Groundcrew Race Race Personality Type 0.006 0.007 0.021 at the SCRF Laboratory in 1985. As described in Chapter 7, "Statistical Methods," each variable was analyzed continuously by a repeated measurements analysis of variance. Testosterone data were subjected to a square root transformation, and TSH values received a logarithmic transformation. Results of the analysis are shown in Table 18-19. As shown in Table 18-19, all three variables declined from their Baseline values, but the reductions over time were relatively proportional for each group by variable. It is concluded that significant differences between groups did not exist for the change in levels between the Baseline examination and the first followup examination. The symmetrical changes in the testosterone and T3 % Uptake variables are speculatively attributed to a 3-year aging effect, but the change in TSH values is suggestive of a change in laboratory methods. There is no suggestion of an adverse rate change in either the Ranch Hand or Comparison group. SUMMARY AND CONCLUSIONS The physical examination and laboratory testing results of all endocrinological variables are summarized in Table 18-20. Questionnaire and review-of-systerns data for past thyroid disease were essentially equivalent in both the Ranch Hand and Comparison groups. These historical data were confirmed by medical record reviews. Physical examination findings were necessarily limited to data from palpation of thyroid glands and testicles; the unadjusted results showed no significant group differences. 18-30 �Longitudinal Analysis for Testosterone, T3 % Uptake, and TSH: A Contrast of Baseline and First Followup Examination Test Means Means Variable Group Testosterone p-Value (Equality of Difference) Ranch Hand Comparison TSH 1985 Followup Ranch Hand Comparison T3 % Uptake Total 1982 Baseline Ranch Hand Comparison Evaluation of the endocrine system was conducted primarily by laboratory testing of hormone levels. The thyroid test battery consisted of T3 % Uptake and TSH assays. The T3 % Uptake data showed no group differences for either mean values or frequency of abnormally low or high values. Occupation was a significant covariate. TSH results revealed a significantly higher mean level in the Ranch Hand group, but this difference was not found by categorical testing of proportions of abnormally high TSH results. Mean levels of testosterone were significantly elevated' among Ranch Hands as contrasted with Comparisons in the 10 to 25 percent body fat category, but this was not reflected by the categorical tests. For the few participants with less than 10 percent body fat (six Ranch Hands, four Comparisons), mean testosterone levels were lower for Ranch Hands than for Comparisons. Age, occupation, and percent body fat were significant adjusting variables. Two timed cortisol specimens showed no significant group differences in mean values and percent abnormalities. The difference between the timed cortisol results, termed the differential cortisol, showed no significant group differences for nonblacks or Blacks born before 1942, but Black Ranch Hands born in or after 1942 had a lower mean differential cortisol level than Comparisons. Age, percent body fat, and personality type were significant covariates in these analyses. Group means of 2-hour postprandial glucose levels were not statistically different, but categorical testing revealed that there was a significantly higher frequency of glucose-impaired (at least 140 but less than 200 mg/dl) Comparisons than Ranch Hands. A constructed variable comprised of known diabetics and individuals classified as diabetic by the glucose tolerance test, showed no difference between the Ranch Hand and Comparison groups. As expected, past and current diabetes were highly influenced by the covariates age, race, and percent body fat. �Overall Summary Results of Unadjusted and Adjusted Continuous and Categorical Analyses of Endocrinological Variables Unadjusted Test Mean Categorical Questionnaire and Physical Examination Past Thyroid Disease (SelfAdministered) Past Thyroid Disease (Interviewer Administered) Thyroid Abnormalities Testicular Abnormalities Laboratory Testing T, % Uptake TSH Testosterone Initial Cortisol 2-Hour Cortisol Differential Cortisol 2-Hour Postprandial Glucose Diabetes (Composite Indicator) ~aAnalysis not feasible. NS^ Not significant (p>0.10). — Analysis not performed. NS*j Borderline significant (0.05<p<0.10). ****Group-by-covariate interaction. Ad j us t ed Mean Categorical �Exposure index analyses did not reveal any pattern consistent with a dose-response relationship. Enlisted flyers in the medium exposure level were significantly different from those in the low exposure level for 2-hour cortisol, differential cortisol, and 2-hour postprandial glucose. However, the corresponding high versus low contrasts were not statistically significant. Longitudinal analyses of T3 % Uptake, TSH, and testosterone levels on all individuals attending both the Baseline and followup examinations revealed only symmetrical and nonsignificant changes in the Ranch Hand and Comparison groups in the interval between examinations. In conclusion, both limited historical and physical examination data, seven endocrinological laboratory variables, and a composite indicator of diabetes did not demonstrate consistent patterns indicating an herbicide effect. However, there was a significant interaction between group and percent body fat for testosterone that could be interpreted as an herbicide effect. TSH and testosterone means tests were statistically significant, and in the expected direction of an herbicide effect, but these results were not confirmed by categorical testing. Also significant was the impaired category of the glucose tolerance test, which showed an excess in the Comparison group. The consistent demonstration of the classical effects of the covariates age, race, occupation, and percent body fat on appropriate endocrine variables provided support for these conclusions. Overall, the endocrine health status of both groups was reasonably comparable. 18-33 �Chapter 18 REFERENCES 1. Potter, C.L., I.G. Sipes, D.H. Russell. 1983. Hypothyroxinemia and . ... , hyothermia in rats in response to 2,3,7,8-tetrachlorodibenzo-p-dioxin ' ' ' ion. Tnxirnl. Annl. Pharmarnl. 69:89-95. administration. Toxicol. Appl. Pharmacol. fiQiflQ-QS. 2. Bastomsky, C.H. 1977. Enhanced thyroxine metabolism and high uptake goiters in rats after a single dose of 2,3,7,8-tetrachlorodibenzop-dioxin. Endocrinology 101:292-296. 3. Barsotti, D.A., L.J. Abrahamson, and J.R. Allen. 1979. Hormonal alterations in female Rhesus monkeys fed a diet containing 2,3,7,8-tetrachlorodibenzo-p-dioxin. Bull. Environ. Contam. Toxicol. 21(4-5):463-469. " 4. Nienstedt, ¥., M. Parkki, P. Uotila, and A. Aitio. 1979. Effect of 2,3,7,8-tetrachlorodibenzo-p-dioxin on the hepatic metabolism of testosterone in the rat. Toxicology 13:233-236. 5. Van Logten, M.J., B.N. Gupta, E.E. McConnell, and J.A. Moore. 1980. Role of the endocrine system in the action of 2,3,7,8-TCDD on the thymus. Toxicology 15(2):135-144. 6. Neal, R.A., P.W. Beatty, and T.A. Gasiewicz. 1979. Studies on the mechanism of toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). In Health effects of halogenated aromatic hydrocarbons, ed. W.J. Nicholson and J.A. Moore, 320:204.New York:The New York Academy of Sciences. 7. Rozman, K., T. Rozman, and H. Greim. 1984. Effect of thyroidectomy and thyroxine on 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induced toxicity. Toxicol. Appl. Pharmacol. 72:372. 8. Rozman, K., E. Scheufler, T. Rozman, T. Pazdernick, and H. Greim. 1984. Effect of thyroxine (T4) and triiodothyronine (T3) on TCDD toxicity in thyroidectomized rats. Toxicologist 4:189. 9. Rozman, K.K. 1984. Role of thyroid hormones and brown adipose tissue in the toxicity of TCDD. In Banbury Report 18; Biological mechanisms of dioxin action, ed. A. Poland and R.O. Kimbrough, ppT 345-354.Cold Springs Harbor, New York: Cold Spring Harbor Laboratory. 10. Bahn, A.K., J.L. Mills, P.J. Snyder, P.H. Gann, L. Houten, 0. Bialik, L. Hollmann, and R.D. Utiger. 1980. Hypothyroidism in workers exposed to polybrominated biphenyls. N. Engl. J. Med. 302:31-33. 18-34 �11. Peters, H.A., A. Gocmen, D.J. Cripps, G.T. Bryan, and I. Dogramaci. 1982. Epidemiology of hexachlorobenzene-induced porphyria in Turkey. Arch. Neurol. 39:744-749. 12. Stehr, P.A., G. Stein, H. Falk, et al. 1986. A pilot epidemiologic study of possible health effects associated with 2,3,7,8tetrachlorodibenzo-p-dioxin contamination in Missiouri. Arch. Environ. Health 41:16-22. 13. May, G. 1973. Chloracne from the accidental production of tetrachlorodibenzodioxin. Br. J. Inds. Med. 30:276-283. 14. Pazderova-Vejlupkova, J., M. Nemcova; J. Pickova, L. Jirasek, and E. Lukas. 1981. The development and prognosis of chronic intoxication by tetrachlorodibenzo-p-dioxin in men. Arch. Environ. Health 36:5-11. 15. Poland, A.P., D. Smith, G. Metter, and P. Possick. 1971. A health survey of workers in a 2,4-D and 2,4,5-T plant, with special attention to chloracne, porphyria cutanea tarda, and psychologic parameters. Arch. Environ. Health 22(3):316-327. 16. Suskind, R.R., and V.S. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 17. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, V.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzo-pdioxin. JAMA 255:2031-2038. 18. Moses, M., R. Lilis, K.D. Crow, J. Thornton, A. Fischbein, H.A. Anderson, and I.J. Selikoff. 1984. Health status of workers with past exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5trichloro-phenoxyacetic acid: Comparison of findings with and without chloracne. Am. J. Ind. Med. 5:161-182. 19. Gruhn, J.G., C.P. Barsano, and Y. Kumar. 1987. The development of tests of thyroid function. Arch. Pathol. Lab. Med. 111:84-100. 20. Schwertner, H.A., R.G. Troxler, G.S. Uhl, and W.G. Jackson. 1984. Relationship between cortisol and cholesterol in men with coronary artery disease and type A behaviour. Arteriosclerosis 4:59-64. 21. Troxler, R.G. and H.A. Schwertner. 1985. , lifestyle, and coronary heart disease. 56:660-665. 22. Aldersberg D., L. Schaefer, S.R. Drachman. 1950. Development of hypercholesterolemia during cortisone and ACTH therapy. JAMA 144:909-914. 23. Stern, M.P., O.G. Kolterman, J.F. Fries, H.O. McDevitt, G.M. Reaven. 1973. Adrenocortical steroid treatment of rheumatic diseases: Effects on lipid metabolism. Arch. Intern. Med. 132:97-100. 18-35 Cholesterol, stress, Aviat. Space Environ. Med. �24. Friedman, M., S.O. Byers, and R.H. Rosenman. 1972. Plasma ACTH and cortisol concentration of coronary-prone subjects. Proc. Soc. Exp. Bio. Med. 140:681-684. 25. Rubenstein, E., and D.D. Federman, eds. 1986. Metabolism: Diabetes Mellitus. Chap. 9. in Scientific American Medicine. New York: Scientific American, Inc. 18-36 �CHAPTER 19 IMMUNOLOGICAL EVALUATION INTRODUCTION Overt damage to organs of the immune system and depressed immunologic function have been noted in a variety of animals exposed to TCDD. As the fields of immunology and immunotoxicology have grown within the past 10 years, a significant spectrum of subtle immunotoxic effects has also been described in animals, but for many possible reasons, comparable adverse effects have not been consistently recorded in exposed human individuals or cohorts. Thus, an intensive search is underway to ascertain the effects of TCDD on the human immune system, particularly with respect to the development of cancer. Every major ongoing dioxin morbidity study in the United States has now incorporated comprehensive laboratory assessments of the immune system. Numerous animal studies have demonstrated significant immunotoxicity following the administration of TCDD. The relatively consistent observations of decreased thymus weight (with cortical atrophy and a depletion of lymphocytes), atrophy of other lymphoid tissue, depressed cellular bone marrow, and decreased humoral and cell-mediated immunity and increased susceptibility to infection have been noted in a variety of animals, including monkeys, rabbits, guinea pigs, rats, and mice. ~ The immune-response effects varied by species, species strain, age, integrity of the endocrine system, dose, and route of administration. Generally, the immunologic parameters returned to normal or approximate normal values over time, even following moderate to high doses of TCDD. While experiments have been conducted to assess the immunotoxicity of TCDD, little has been published on the immunotoxicity of 2,4-D or 2,4,5-T. The. immune system is so sensitive to TCDD that immune function has frequently been used as a marker of toxicity in the absence of other biologic effects. The mechanism of TCDD immunotoxicity is under intensive investigation by molecular biologists, pathologists, and geneticists. In general, TCDD toxicity is probably linked to the Ah receptor, and specifically to the Ah allele which governs microsomal enzyme induction as reflected by aryl hydrocarbon hydroxylase and cytochrome P-448/450 levels.1 This premise underscores the questions of the degree to which the human response to TCDD is dependent upon the Ah locus or other genetic receptors, and how this response is manifested in the immune system. Animal studies and several observational studies in humans have shown variable results. Data from the Times Beach, Missouri, episode disclosed no group differences for various T lymphocyte populations, proliferative responses to PHA, concanavalin A, pokeweed or tetanus toxoid stimulation, and in skin testing with seven antigens.1 A report of the assessment of the immune system of men exposed to TCDD in an industrial accident in Britain 19-1 �did not discuss the results of the measurement of the immunoglobulin profile, lymphocytes, T and B cells, response to PHA, and three hematologic variables. A prior publication on the same cohort cited unpublished findings of Ward, suggesting a reduction in the capacity of the "primary" immune system. A longitudinal evaluation of 48 highly exposed children (one-half with chloracne) from the Seveso incident showed significantly elevated complement hemolytic activity over six measurements during the period 1976 to 1979 (although the biologic significance of this is unknown) and an increased Pf9liferative response to PHA and pokeweed during the first three screenings. This study (as others) was characterized by shifting a study population over the observation period and by excessive laboratory variation that may have masked other true group differences. Nonetheless, the Seveso data may be interpreted as indicative of a stimulated immune system, particularly cellmediated immunity, differing substantially from the bulk of animal studies, which showed decreased activity. A recent study of residents of the TCDD-contaminated Quail Run Mobile Home Park in Missouri also revealed data that conflicted with the Seveso experience. A statistically significant amount of anergy and relative anergy was detected in the TCDD-exposed group, as determined by the multitest applicator (seven-antigen test system). Inter-reader variation presented major interpretive difficulties. Nevertheless, findings suggestive of decreased cell-mediated immunity were provided by decreased T , T4, and T X1 cell percentages. Also noted was an increased lymphoproliferative response to pokeweed mitogen (PWM). The overall depression of immunologic response was not correlated with an increase in clinical disease. Baseline Summary Results Immunologic function and phenotypic marker studies were performed on 592 participants (297 Ranch Hands, 295 Comparisons) randomly selected by the terminal digit of their case number. Because of laboratory problems, e.g., fluctuating quality control and lack of simultaneous differential counts on the peripheral mononuclear cells, a special Immunology Review Committee was convened to determine which data were relevant for analysis. Such decisions were made on a case-by-case basis without knowledge of Ranch Hand or Comparison group membership. The Committee concluded that the data could be analyzed on a group basis, but interpretation of data on an individual basis was inappropriate. Analyses of the cell surface markers (T11, T3, T4, T8, B, the T4/Tg ratio, and the total lymphocyte count [TLC]) showed no significant group differences. However, the increased smoking was significantly associated with increases in cell counts but not with the T4/T8 ratio and B cells, whereas increasing age was significantly associated with decreasing TLC and T. cells. o Functional studies of T and B cells via reaction to antigenic (tetanus toxoid) or mitogen (phytohemagglutinin, concanavalin A, and pokeweed) stimulation showed no group differences. Similarly, unadjusted and adjusted mean values of the four assays were not significantly different between groups, but one unstimulated control value (reflecting Baseline thymidine uptake by T cells) was significantly decreased in the Ranch Hands. The biologic relevance of this finding was unclear. 19-2 �Further, in the covariate analysis of the functional studies, group-bysmoking and group-by-alcohol interactions were noted. Of greater importance, however, was the finding that lymphocytic response increased with increased smoking, but was depressed in association with increasing age. In summary, both immunologic function and cell marker studies did not show significant impairment in the Ranch Hand group, or patterns supportive of an herbicide effect. Smoking, for the first time to the knowledge of the authors, was associated with a significant increase in the marker cells T l l f T3, T4, and T8, and in the total lymphocyte count, with a concomitant increase in lymphocytic response to PWM. Parameters of the 1985 Immunologic Profile The format for the 1985 AFHS physical examination placed more emphasis on the immunologic assessment than did the 1982 Baseline profile. The random sampling scheme was expanded to produce an approximate 50-percent sample of the cohorts, and included the same terminal digits of the case number used at Baseline in order to include all individuals evaluated in 1982 and thus establish a longitudinal data base. All immunologic tests were performed at the Scripps Immunology Reference Laboratory (SIRL). The battery of phenotypic marker assays and functional tests was slightly modified from the Baseline profile. The assay for HLA-DR cells was added to the battery of marker studies, and a functional test for natural killer cells (with and without interferon) was substituted for the concanavalin A stimulation assay. A comprehensive set of skin tests for the antigens Candida, mumps, Trichophyton, and staph-phage-lysate was added to evaluate the integrity of the delayed hypersensitivity response. The dependent variables of the analyses in this chapter arise from three distinct measurement systems: phenotypic marker studies, functional studies, and skin testing. There were more covariates than in the Baseline study, namely age, race, occupational category, exposure index, and new smoking and alcohol data from the 1985 questionnaire. Participants deleted from the immunological analyses included those with recent radiation or chemotherapy, and those individuals on immunosuppressive or systemic steroid medication. Marginal totals in the tables below vary somewhat due to missing covariate data. Thus, numbers in the table also vary according to which immunologic data sources were used in the analysis. In general, most analyses are based upon data from 465 Ranch Hands and 585 Comparisons. Analytic tests included t-tests, general linear models (SAS®-GLM), logistic regression (BMDP®-LR), and Fisher's exact test. Parallel analyses using Original Comparisons are in Tables Q-5 through Q-10 of Appendix Q. Rationale of the Immunologic Measurements Because of rapid changes in,our knowledge of the immune system, Table 19-1 is provided as an aid in interpreting the medical significance of the immunological data. 19-3 �TABLE 19-1. Medical Significance of the Immunological Data Immunologic Measure Rationale of the Measurement Disease/Syndrome/Condition Endpoint MARKER STUDIES Measures total T cells coincident with sheep rosette receptor on cell surface (most are T4 and Tg cells). Decreased in immune deficiency/ increased with lymphoproliferative disorders. Measures peripheral blood B cells, no reaction with T cells, granulocytes, or monocytes. Decreased in immunodeficiency/ increased in lymphoproliferative disorders. Measures T cells which exhibit helper/ inducer phenotype. Decreased in AIDS/increased in autoimmune diseases. OKT Measures T cells which exhibit suppressor/ cytotoxic functions. Variable in autoimmune diseases. Increased in some viral illnesses and immunodeficiencies. 4 Leu M3 Measures mature monocytes in peripheral blood. Increases with inflammation. HLA-DR Measures cells expressing HLA-DR antigen; includes B cells and monocytes. B cell deficiency/ Agamraaglobulinemia. Mixed Leukocyte Culture (MLC) Measures reactivity of T cells to foreign histocompatiblity antigens on unrelated lymphocytes. HLA sensitization/transplantation. PHA Measures functional capability of T cells to become activated by mitogen and undergo proliferation. T cell deficiency. NKC (with interferon) NKC (without interferon) Measures natural killer cell lytic activity with and without interferon treatment of the natural killer cells. Decreased natural defenses. Leu 12 vO FUNCTIONAL STUDIES �TABLE 19-1. (continued) Medical Significance of the Immunological Data Immunologic Measure Rationale of the Measurement Disease/Syndrome/Condition Endpoint FUNCTIONAL STUDIES (continued) PUM Measures functional capability of T cells to become activated by mitogen and undergo proliferation. T cell deficiencies. Each measures skin reactivity induced by specific antigen injected intradermally and correlates with recall T cell sensitivity to the antigen. Antigen reactivity or sensitivity/ Anergy. SKIN TESTS Candida Mumps Tricophyton Staph-phage-lysate vO I �Immunology Methodologies The isolation of peripheral blood mononuclear (PBM) cells was the first step to prepare for testing immune competence and enumeration of phenotypic markers. Heparinized whole blood was obtained from each patient. PBM's were isolated by Ficoll-Hypaque density gradient centrifugation. The PBM's were then washed and resuspended in HB101 media (HANA Biologies, Inc.) supplemented with 10M units/ml penicillin, 10,000 meg/ml streptomycin, 1 percent sodium pyruvate (100 mM), and 1 percent L-glutamine (200 mM). To determine percent monocyte and granulocyte contamination of the PBM cell preparations, an aliquot of the cells was stained with a nonspecific esterase stain. PBM concentration was adjusted for each individual assay, Cell Surface Marker Analysis Mouse monoclonal antibodies directed against specific surface markers were used to identify and quantitate different cell populations in the peripheral blood of the participants. Mononuclear cell concentrations adjusted to 1.0 x 10 cells/ml were incubated with the following fluorescein isothiocyanate conjugated monoclonal antibodies: CD2(OKT11*), CD4(OKT4*), CD8(OKT8*), CD19(Leul2**), CD14(LeuM3**), and HLA-DR(OKDR*). These cell surface antibodies measure total numbers of T and B lymphocytes, monocytes, helper T lymphocytes, suppressor T lymphocytes, and those cells carrying the HLA-DR antigen. A flow cytometer (Spectrum III, Ortho Diagnostic Systems, Raritan, New Jersey) was used to measure percent positive for each specific surface marker and absolute numbers were calculated. Phytohemagglutinin (PHA) and Pokeveed Mitogen Stimulation Assays Mitogens were used to stimulate the proliferation of lymphocytes in vitro. During the proliferative response, the lymphocytes undergo blast transformation and incorporate radioactive thymidine into their DNA. Participant lymphocyte concentrations were adjusted to 2.0 x 10 cells/ml in supplemented HB101 media. Samples were cultured in quadruplicate. Individual cultures consisted of 0.1 ml of cell suspension and 0.1 ml of mitogen solution in microtiter plates. The cultures were incubated in an atmosphere of 5 percent CO, at 37 degrees Centigrade. Participant cells were cultured with PHA (12 ug/ml, Sigma Chemical Co., St. Louis, Missouri) for a total of 4 days and pokeweed mitogen (0.05 ug/ml, Sigma Chemical Co., St. Louis, Missouri) for a total of 5 days. The cultures were pulsed with tritiated thymidine (1.0 uCi/ microtiter well) for 4 hours and then harvested on a multiple automated harvester. Cellular proliferation was assessed by tritiated thymidine uptake measured by liquid scintillation counting. Mixed Lymphocyte Reaction Histocompatibility antigens (HLA) can also stimulate lymphocytes causing blast transformation. Donor lymphocytes were used to stimulate the proliferation of participants' lymphocytes in vitro. A pool of donor lymphocytes *0rtho Diagnostic Systems, Raritan, New Jersey **Becton Dickinson Monoclonal Center, Inc., Mountain View, California 19-6 �was frozen and used as a stimulator pool throughout the course of the study. An aliquot of this pool was thawed daily. Viability of this pool was assessed by trypan blue exclusion. A pool of freshly isolated lymphocytes was prepared daily and also used as stimulator cells. Both pools of stimulator cells were inactivated by irradiation ( , 0 rad). 300 Stimulator pools and participant lymphocyte concentrations were adjusted to 1.0 x 10 cells/ml in supplemented HB101 media. Samples were cultured in quadruplicate. Individual cultures consisted of 0.1 ml of participant cell suspension and 0.1 ml of stimulator cell suspension, in microtiter plates. The cultures were incubated in an atmosphere of 5 percent CO, at 37 degrees Centigrade for 6 days. The cultures were pulsed with tritiated thymidine (1.0 uCi/ microtiter well) for 16 hours and then harvested on a multiple automated cell harvester. Cellular proliferation was assessed by tritiated thymidine uptake measured by liquid scintillation counting. Natural Killer Cell Assays Mononuclear cells from the participant were evaluated to assess the ability of 'certain peripheral blood cells to kill target cells from a K-562 leukemia cell line. The K-562 target cells were preincubated with radioactive chromium ( Cr) at 37 degrees Centigrade in 5 percent C02 for 1 hour, washed, and the cell concentration adjusted to 1.6 x 10 cells/ml. A 50 ul aliquot of radioactive K-562 cells was added to each microtiter well. Participant lymphocytes were adjusted to three different concentrations: 0.53, 1.6, and 2.7 x 10 cells/ml. One ml of each of these concentrations was incubated with 20 units of recombinant y-interferon (Genentech, Inc., San Francisco, California) for 1 hour at about 37 degrees Centigrade. Quadruplicate 150 ul aliquots of each concentration, with and without interferon preincubation, were dispensed in a microtiter plate. Four wells contained media alone to determine the spontaneous release of radioactivity from the K-562 cells. Four wells contained 1 percent Triton X-100 to determine the maximal release of radioactivity. The final effector to target ratios were 50;1, 30:1, and 10:1. The microtiter plates were centrifuged briefly at low speed and incubated at 37 degrees Centigrade in 5 percent C02 for 3 hours. A 100 pi aliquot of the supernatant was removed from each well and counted on a gamma counter. Percent chromium release from the K-562 target cells was determined for each effector:target cell ratio. Interpretive Considerations The values of the results of assays of immunologic status are more variable than those found in routine single reactant clinical chemistry assays. Often there are numerous biochemical factors/metabolites that affect the immunologic assay results so that interpretations of normalcy must be in the context of those 'obtained concurrently in a normal control cohort group. Such controls allow for proper adjustments of the raw assay data in order to minimize the broad range of technical and reagent effects in the various immunologic assays. These adjustments in the raw assay data results will correct for such variability and allow for the detection of any significant biologic abnormalities. Because of the need for these control adjustments, the immunologic assay results cannot be meaningfully compared to existing normal ranges determined on different groups of individuals at other times. 19-7 �RESULTS AND DISCUSSION Cell Surface Marker (Phenotypic) Studies Immunological tests were carried out on 47 percent (1,085) of the participants because of the complexity of the assay and the expense of these tests. The participants were randomly selected so that approximately 50 percent of each group of participants arriving for the physical examination had blood drawn for the immunological tests. Logistical delay during the initial weeks of the examination reduced the number to less than 50 percent. Within each group, blood was drawn for the immunological tests from about one-half of the selected participants on the first day of the physical examination, and from the remainder on the following day. Skin tests, which were scheduled for the first day, were therefore carried out after the blood draw on the first day for the first half of the immuno-tested participants. Skin tests were not done for those participants selected for immunological testing on day two in order to avoid any effect the skin test antigens might have on the cell counts and functions. Thus, 553 participants received both the iinmunological tests and the skin tests, 532 received the immunological tests but not the skin tests, 1,206 received the skin tests but not the immunological tests, and 18 received neither. Table 19-2 gives the frequencies of the participants in each exposure group who had the tests. Participants who were taking anti-inflammatory or immunosuppression medication or who had recently received x-ray treatment or chemotherapy for cancer were excluded from all the analyses. Participants taking aspirin, however, were not excluded. Frequencies of Participants Who Took the Immunological Tests and the Skin Tests, by Group Skin Tests Group Ranch Hand Immunology Tests No Yes Total Comparison No Yes Total No Yes Total �For those participants who were given the immunological tests, following dependent variables were examined: total T cells, helper suppressor T cells, B cells, monocytes, HLA-DR cells, and the T4/Tg suppressor cell) ratio. These variables were treated as continuous analysis. the T cells, (helper/ in the The covariates considered in the analysis were the matching variables (age, race, occupation), smoking history (current cigarettes/day and total pack-years of smoking), and alcohol consumption (average number of drinks per day during the 2 weeks prior to the physical examination and total drinkyears). The covariates age and the smoking history and alcohol consumption variables were used as continuous variables in the analyses since the relationships between the dependent variables and the covariates were generally monotonic. Considerable day-to-day variation exists in the results of immunological tests due to a number of extraneous factors, including temperature, humidity, and sensitivity of the instrumentation. Significant batch-to-batch variation (among examination groups) was apparent for total T cells, suppressor T cells, B cells, and the T4/T8 ratio, and significant blood-draw day variation was apparent for helper T cells, monocytes, and HLA-DR cells. Adjustments in the analyses were made for these sources of variation by using batch or blooddraw day indicators. Throughout this section, appropriate adjustment was carried out in the assessment of group differences of the dependent variables; this analysis was unadjusted for the covariates listed above and is referred to as the "unadjusted" analysis. Adjustment was also made for batch-to-batch or blood-draw day variation in the analyses of the associations of the dependent variables with the covariates. Further, this adjustment was also used in the fitting of general linear models to assess the group differences, adjusted for the covariates. Prior to analysis, group data were pooled for each continuous variable and were examined to determine whether transformation would enhance normality or distributional symmetry. The following transformations were used in the analyses: Variable Transformation Total T cells Square root Helper T cells Square root Suppressor T cells Logarithm B cells Square root Monocytes Logarithm HLA-DR cells Square root T4/T8 ratio , Logarithm The results of the analyses in this section are summarized in Tables 19-3 through 19-5. Table 19-3 presents the unadjusted analyses for the cell surface markers, Table 19-4 displays the covariate associations, and Table 19-5 gives the adjusted results. These tables are accompanied by 19-9 �Unadjusted Analyses for Cell Surface Markers by Group Group Variable Total T Cells Helper T Cells (T4) Suppressor T Cells (T8) B Cells Monocytes HLA-DR Cells T4/T8 Ratio Statistic Ranch Hand Comparison p-Value �Association Between Cell Surface Marker Variables and the Covariates in the (Combined Ranch Hand and Comparison Groups (Directionality Shown) Variable Race Occupation Age Current Alcohol (Drinks/Day) DrinkYears Current Smoking (Cigarettes/Day) Lifetime Smoking (Pack-years) Total T Cells Helper T Cells Suppressor T " Cells B Cells Monocytes HLA-DR Cells T./T. Ratio 4 o "Monotone decreasing. NS: Not significant (p>0.10) b N: Nonblack °Monotone increasing. B: Black Generally decreasing trend. 0: Officer e E: Enlisted personnel (flyer and lncreases, drop-off at highest category. lncreases from 0 category, then decreases, but not back to same level. Increases from 0 category, then steady decrease with increasing levels. 9 groundcrew) Generally increasing trend. F: Enlisted flyer Flat for 0 and first few categories, then increases. G: Enlisted groundcrew �TABLE 19-5. Adjusted Analyses for Cell Surface Markers by Group Group Variable Statistic Ranch Hand Comparison p-Value Total TCells n Adj. Mean 95% C.I. 442 567 iLJ.iti .!• TvTOwC TCnrCn Helper TCells n Adj. Mean 95% C I .. 439 566 894 6. 885 7. 062 .6 (3., 9 3 2 (4., 984 861 0 . ) 892 0.) BATCH (p=0.021) nAY(BATCH) (p=0.014) AGE (p<0.001) ALOOCC (p=0.008) CSMCK*OCC (p=0.023) ALC*CSMCK (p=0.006) (p=o.oi2) Suppressor TCells n Adj. Mean 95% C I .. 463 50 8 530.8 537.9 0.640 (506.8, 556.0) (516.1, 560.5) BATCH (p<0.001) OCC (p=0.014) AGE (p=0.004) ALC (p=0.020) CSMCK (p<0.001) BCells n Adj. Mean 95% C I .. 435 561 BATCH (p<0.001) ALC (p=0.006) AGE*CSMCK (p=0.025) E8KHl*RACE (p=0.026) GRP*PACm (p=0.018) GRP*RACE*OCC (p=0.046) Monocytes n Adj. Mean 95% C.I. 440 568 BATCH (pO.OOl) DAY(BATCH) (p<0.001) PACE (p=0.032) EBKXR (p=0.013) CSMOK (p<0.001) PACKYR (p=0.006) GRP*OCC (p=0.044) GRP*ALC (p=0.010) 19-12 Covariate Remarks* BATCH (p=0.029) AGE (p=0.009) ALC (p=0.001) CSMCK (p<0.001) GRP*RACE (p=0.033) (p=0.015) �TABLE 1 - . (continued) 95 Adjusted Analyses for Gall Surface Markers by Group Group Variable Statistic Ranch Hand Comparison p-Value HLA-CR Cells n Adj. Mean 95% C I .. 49 5 50 8 T4/T 8 Ratio n Adj. Mean 95% C I .. 461 577 150 .7 152 .5 068 .7 (1.501, 1.643) (1.491, 1.616) Covariate Remarks* MKH (p<0.001) DAY(MDCH) (p=0.004) OCC (p=0.035) CSMCK (p<0.001) GRP*ALC (p=0.045) AG£*PACKYR (p=0.005) B^TCH (p<0.001) OCC (p=0.020) CSMOK (p<0.001) *Abbreyiations BATCH: batch-to-batch variation among examination groups DAY(MICH): blood-draw day variation ALC: current alcohol use CSMOK: current smoking OCC: occupation GRP: group ERKXR: lifetime alcohol use (drink-years) PACKYR: lifetime smoking (pack-years) ****Significant group-by-covariate interaction—adjusted mean, confidence interval, and p-value not presented. 19-13 �discussion of each variable. The results of adjusted analyses with group-bycovariate interactions are found in Table Q-l of Appendix Q. Total T Cells ( ^ T ) No significant difference was found between groups in the mean values of total T cells (p=0.736). These data were analyzed without adjustment for any covariates except batch-to-batch variation. The data were pooled for the two groups, and the relationship with the covariates was examined. Significant associations were found with occupation (p=0.005), age (p=0.002), current smoking (p<0.001), and pack-years (p<0.001). A marginal association (p=0.069) was found with current alcohol use due to a steady decrease in mean counts with higher drinking levels. Officers had a lower mean count (1,539 cells/mm ) | a enlisted flyers (1,668 cells/mm ), or hn enlisted groundcrew (1,647 cells/mm ) . The mean count decreased with age: 1,663 cells/mm , 1,582 cells/mm , and 1,404 cells/mm for those born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. The mean count increased with increasing current smoking and increasing lifetime smoking history (pack-years). A general linear model was fitted to assess the group difference in mean count of total T cells with adjustment for each covariate and any interactions that made significant contributions to the model. Batch-to-batch variation was a significant covariate (p=0.029). A significant group-by-race interaction was found (p=0.033); Black Ranch Hands had a significantly lower adjusted mean count than Black Comparisons (1,566 cells/mm versus 1,888 cells/mm ; p=0.039), but the group difference for nonblacks was not significant (p=0.619) (see Table Q-l of Appendix Q). The following covariates were significant: age (p=0.009), current alcohol use (p=0.001), current smoking (p<0.001), and a drink-year-by-pack-year interaction (p=0.015). Analyses using only Original Comparisons showed the same results as when using the total Comparison group (see Tables Q-6 and Q-7 of Appendix Q), with a group-by-race interaction present (p=0.028). Helper T Cells ( 4 T) No significant difference was found between groups in the mean values of helper T cells (p=0.610). This contrast was analyzed without adjustment for any covariates except blood-draw day variation. The data were pooled for the two groups, and the relationship with the covariates was examined. Significant associations were found with occupation (p=0.024), age (p<0.001), current smoking (p<^).001), and pack-years (p<0.001). Officers had gt lower mean count (831 cells/mm ) than enlisted flyers ( 8 cells/mm ) or enlisted groundcrew ( 9 cells/mm ). There was a decrease 85 84 in the mean count with increasing age: 907 cells/mm , 850 cells/mm , and 713 cells/mm for those born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. The mean count increased with increasing levels of current smoking and with increasing pack-years of lifetime smoking. 19-14 �Adjusted analyses assessed the group difference in mean count of helper T cells with adjustment for each covariate and any significant interactions. Adjustment for the blood-draw day variation was included. Age made a significant contribution to the model (p<0.001)1 The following interactions between covariates were significant: current alcohol use-by-occupation (p=0.008), current smoking-by-occupation (p=0.023), current alcohol use-bycurrent smoking (p=0.006), and drink-years-by-pack-years (p=0.012). The adjusted group difference in mean count was not significant (p=0.662): 869 cells/mm for the Ranch Hand group versus 879 cells/mm for the Comparison group. Adjusted analyses using Original Comparisons (Table Q-6 of Appendix Q) also revealed a nonsignificant group difference (p=0.835). Suppressor T Cells (TB) No significant difference was found between groups in the mean values of suppressor T cells (p=0.671). This contrast was analyzed without adjustment for any covariates except batch-to-batch variation. The data were pooled for the two groups, and the relationship with the covariates was examined. Significant associations were found with occupation (p<0.001), age (p<0.001), and current smoking (p<0.001). The mean count for officers was less than the mean count for enlisted groundcrew, which was in turn less than the mean count for enlisted flyers; the means were 492 cells/ mm , 540 cells/mm , and.575 cells/mm , respectively. The mean counts decreased with increasing age: 557 cells/mm , 512 cells/mm , and 439 cells/ mm for participants born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. The mean counts increased with increasing levels of current smoking. Marginally significant associations were found with current alcohol use (p=0.058, mean T8 counts decreased with increasing current levels of drinking) and pack-years (p=0.076, mean counts increased with increasing pack-years). The adjusted analysis of group differences in mean count of suppressor T cells was made with adjustment for each covariate and any interactions that made significant contributions, including significant batch-to-batch variation (p<0.001). Significant adjusting covariates were occupation (p=0.014), age (p=0.004), current alcohol use (p=0.020), and current smoking (p<0.001). The adjusted group difference was not significant (p=0.640). A marginal (p=0.063) group-by-race interaction was not retained in the final model, but was explored. Black Ranch Hands had a lower adjusted mean count than Black Comparisons (512 cells/mm versus 649 cells/mm , p=0.056), whereas the difference between nonblack groups was negligible (531 cells/mm3 for Ranch Hands and 532 cells/mm for Comparisons, p=0.974). Analyses involving the Original Comparisons showed a significant interaction between group and race (p=0.010) (see Tables Q-6 and Q-7 of Appendix Q), with the same pattern seen for the group-by-race interaction for the total Comparison group. B Cells No significant difference was found between groups in the mean values of B cells (p=0.594). This contrast was analyzed without adjustment for any covariates except batch-to-batch variation. 19-15 �Significant associations using pooled data were found between B cells and occupation (p<0.001), age (p<0.001), current alcohol use (p=0.001), drink-years (p=0.047), current smoking (p<0.001), and pack-years (p=0.005). Officers had a lower mean count than enlisted flyers and groundcrew (166 cells/mm , 205 cells/mm , and 201 cells/mm , respectively). The mean count decreased with increasing age: 206 cells/mm , 177 cells/mm , and 146 cells/mm for those born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. The mean counts decreased with an increasing number of drinks per day, and also with higher levels of total lifetime drinking, except for the "never-drinkers," whose level was lower than the greater than 30 to 100 drink-year group; the means for the drinkyear categories weres 0, 180 cells/mm ; greater than 0 to 5, 199 cells/mm ; greater than 5 to 30, 189 cells/mm ; greater than 30 to 100, 183 cells/mm3; and greater than 100, 150 cells/mm . The nonsmokers had a lower mean count than current smokers, whereas among the smokers the mean counts decreased with higher current smoking levels. The means for the different currentsmoking (cigarettes/day) categories were: 0, 166 cells/mm ; greater than 0 to 20, 237 cells/mm 5 greater than 20 to 40, 222 cells/mm ; and greater than 40, 202 cells/mm3. Lifetime smokers had a higher mean count than "neversmokers"; otherwise the pattern was not clear. Adjusted analyses, including adjustment for the significant (p<0.001) batch-to-batch variation, were used to investigate the mean count of B cells. Adjustment was made for each covariate and any interactions that made significant contributions. A significant group-by-race-by-occupation interaction was found (p=0.046), along with a group-by-pack-year interaction (p=0.018). Significant contributions were made by current alcohol use (p=0.006), an age-by-current smoking interaction (p=0.025), and a drink-years-by-race interaction (p=0.026). The analysis consequently was performed separately for nonblacks and Blacks. For nonblacks, the group-by-pack-year interaction persisted (p=0.021) (see Table Q-l of Appendix Q). Ranch Hands who had never smoked had a much lower adjusted mean count than the corresponding Comparisons, 154 cells/mm versus 190 cells/mm (p=0.004). Among smokers, the adjusted mean count for the greater than 0 to 20 pack-year category was less for Ranch Hands than for Comparisons. For both the greater than 20 to 40 and the greater than 40 pack-year categories, the adjusted mean count was higher for Ranch Hands than for Comparisons. The p-values for these three contrasts were greater than 0.10. For Blacks, the unadjusted group difference was not significant (p=0.808; Ranch Hands, 186 cells/mm , versus Comparisons, 194 cells/mm ). Adjusted means were not calculated because no covariates made any significant contribution to an adjusted model, and moreover, adjustment for batch-to-batch variation was not possible because of the small number of Black participants. Other significant covariates and interactions in the adjustment for nonblacks included occupation (p=0.047), drink-years (p<0.001), and an age-by-current smoking interaction (p=0.039). Monocytes No significant difference was found between groups in the mean value of monocytes (p=0.427). This contrast was analyzed without adjustment for any covariates except blood-draw day variation. 19-16 �The data were pooled for the two groups, and the relationship with the cbvariates was examined. Significant associations were found with race (p«0.027), occupation (p=0.019), current drinking (p=0.031), drink-years (p<0.001), current smoking (p<0.001), and pack-years (p<0.001). Blacks had a lower mean count than nonblacks (37.1 cells/mm3 versus 45.7 cells/mm , respectively). Officers had a lower mean count (42.3 cells/mm ) than enlisted flyers ( 4 4 cells/mm ), who had a lower mean count than enlisted 4. groundcrew ( 8 2 cells/mm3). Higher mean counts were associated with higher 4. current drinking levels. There were increases in mean counts with higher drink-years and with increasing amounts of both current and lifetime smoking. Assessment of the group difference in mean count of monocytes was done with adjustment for each covariate and any interactions that made significant contributions, including blood-draw day variation. A significant group-by-occupation interaction (p=0.044) and a significant group-by-current alcohol use interaction (p=0.010) were found. For interpretation, these were explored in a model including the group-byoccupation-by-current alcohol use interaction, with the alcohol variable discretized (see Table Q-l of Appendix Q). Except for those men consuming more than two to four drinks per day, Ranch Hand officers had a higher adjusted mean count than Comparison officers, the difference being large ( 4 2 cells/mm versus 32.3 cells/mm ) for nondrinkers, (p=0.060). For 4. enlisted flyers, except those in the greater than four drinks per day category, Ranch Hands had a lower adjusted mean count than corresponding Comparisons. For the greater than two to four drinks per day category, a large difference between adjusted means (32.7 cells/mm for Ranch Hands, 56.2 cells/mm for Comparisons) was observed (p=0.097). Further, it was found that for enlisted groundcrew not currently drinking, Ranch Hands had a lower adjusted mean count than the corresponding Comparisons, whereas the Ranch Hand current drinkers had higher adjusted mean counts than the corresponding Comparisons. The difference was large ( 8 9 cells/mm versus 35.3 6. cells/mm ) for the greater than four drinks per day category (p=0.003). Significant effects on the monocyte counts were also seen for race (p=0.032), drink-years (p=0.013), current smoking (p<0.001), and pack-years (p=0.006). Analyses using Original Comparisons revealed a significant (p=0.040) group-by-age interaction (see Tables Q-6 and Q-7 of Appendix Q). This was due to a lower count for Ranch Hands than Comparisons for those born in or after 1942 ( 1 4 cells/mm versus 48.0 cells/mm , p=0.048), a higher 4. count for Ranch Hands than Comparisons for those born between 1923 and 1941 ( 8 2 cells/mm versus 42.8 cells/mm , p=0.058), and very little difference 4. for those born in or before 1922 (p=0.924). HLA-DR Cells No significant difference was found between groups in the mean values of HLA-DR cells (p=0.842). This contrast was analyzed without adjustment for any covariates except blood-draw day variation. Significant associations were found using pooled data with occupation (p<0.001), age (p=0.010), current smoking (p<0.001), and pack-years (p<0.001). Officers had a lower mean count than enlisted participants (526 cells/mm versus 597 cells/mm for flyers and 598 cells/mm for groundcrew). The average mean count was higher for younger participants than for 19-17 �older participants: 588 cells/mm3, 557 cells/mm3, and 555 cells/mm3 for those born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. There was a significant increase in average mean counts with increasing levels of both current and lifetime smoking. There was a marginally significant increase in mean cell counts with drink-years (p=0,083). Analyses, with adjustment for blood-draw day, each covariate, and any interactions, were carried out to assess the group difference in mean count of HLA-DR cells. A significant group-by-current alcohol use interaction was found (p=0.045); for Ranch Hands drinking more than four drinks per day, the adjusted mean count was greater, 564 cells/mm versus 473 cells/mm , than for Comparisons (p=0.052), whereas no appreciable group differences were apparent for the participants drinking four or fewer drinks per day (see Table Q-l of Appendix Q). Significant effects were seen with occupation (p=0.035), current smoking (p<0.001), and an age-by-pack-year interaction (p=0.005). Analyses using Original Comparisons (Table Q-6 of Appendix Q) did not show a significant group-by-current alcohol use interaction (p=0.152), and no significant difference between groups was observed (psO.887). T4/T8 Ratio No significant difference was found between groups in the mean value of the T4/T8 ratio (p=0.499). This contrast was analyzed without adjustment for any covariates except batch-to-batch variation. The data were pooled for the two groups, and the relationship with the covariates was examined. Significant associations were found with drinkyears (p=0.049), current smoking (p<0.001), and pack-years (p=0.001). The mean T4/T8 ratio generally increased with increasing drink-years, and increased with increasing amounts of current smoking and total pack-years. There was a marginally significant association with occupation (p=0.063). Enlisted flyers had a lower average ratio than officers and enlisted groundcrew ( . 8 versus 1.62 and 1.60, respectively). 14 The adjusted group difference in the T4/T8 ratio was not significant (p=0.678; Ranch Hands 1.57 versus Comparisons 1.55). Significant effects were seen for occupation (p=0.020), current smoking (p<0.001), and batch-tobatch variation (p<0.001). In the analysis of the Original Comparisons, a significant group-bycurrent smoking interaction was found (p=0.016). Further analysis snowed a significant difference between groups (Ranch Hand mean ratio of 1.84 versus Original Comparison mean ratio of 1.51, p=0.004), for the greater than 20 to 40 cigarettes per day category (see Tables Q-6 and Q-7 of Appendix Q). Functional Stimulation Studies Statistical analyses were performed on cell function responses to PHA, PWM, and MLC. For each stimulated cell population, autologous controls were also studied. The measurements resulting from each test were the average counts over four samples for the stimulated cell population and for the autologous controls. The net average count, defined as the difference 19-18 �between the average counts per minute (CPM) for the stimulated and the control cells, was also calculated. Cell response data were obtained from the 1,085 immunologically tested participants. The exclusion conditions were the same as in the previous section, namely, participants who were taking anti-inflammatory or immunosuppressant medication, or who had recently experienced radiation therapy or chemotherapy for cancer. Review of the immunological data base by the Air Force and SIRL resulted in certain test exclusions due to technical error, and equipment malfunction, those identified by quality control procedures (Grubbs' test2 ), and unexplained outliers. For the mean cell counts per minute (CPM) analyzed for this section, a total of 17 data points were excluded as unexplained outliers from the data base, involving eight participants (three Ranch Hands and five Comparisons): 1 point was invalid due to technical error in the assay for MLC stimulated cells (Ranch Hand) and the remainder were outliers in the PWM controls or stimulated cells (two Ranch Hand controls, 13 Comparison controls, and one Ranch Hand stimulated cells). This meant that, for one participant, the PWM control mean was omitted from the analysis and, for the other seven participants, the means were calculated from fewer than four points. No unexplained data points were found for the PHA-stimulated cells or corresponding controls. All analyses were adjusted for significant blood-draw day variation, and the same covariates were used as in the adjusted analyses of the cell surface markers. The covariates age, current smoking, pack-years, current alcohol use, and drink-years were discretized because marginal examination showed generally nonlinear responses of the cell function variables with these covariates. Thus, the p-values given in this section for the marginal association of the variables with each covariate indicate the significance of the differences among the categories defined by the levels of the covariate. Prior to analysis, the data were transformed to enhance normality or at least distributional symmetry. The following transformations were used: Variable Transformation Unstimulated Response (PHA) logarithm PHA Net Response none Pokeweed Net Response square root MLC Net Response square root The summarized results of this section are given in Tables 19-6 through 19-8 (see Table Q-l of Appendix Q for results involving group-by-covariate interactions). Only results for the unstimulated controls for the PHA assay are presented as an assessment of the function of the immune system in the unchallenged state. However, separate controls were run for each assay since incubation periods vary for each1test procedure. In the analysis of data on the net response for each assay, the appropriate control was used. Analysis of each control assay was performed, and no significant group differences were noted. 19-19 �Unadjusted Analyses for Functional Stimulation Tests by Group Variable Statistic* Group Ranch H a n d C o m p a r i s o n p - V a l u e Unstimulated Response (PHA) PHA Net Response Pokeweed Net Response MLC Net Response Unstimulated Response (PHA) No significant difference was found between groups in the mean values of PHA Unstimulated responses (p=0.979). These control values were derived from Unstimulated cells and reflect baseline cell function. This contrast was analyzed without adjustment for any covariates except blood-draw day variation. The data were pooled for the two groups, and the relationship with the covariates was examined. Significant associations"were found with race (p<0.001), age (p<0.001), and drink-years (p=0.048). The average mean count for nonblacks was lower than for Blacks: 1,629 CPM and 2,210 CPM, respectively. There was a strong decrease in mean count with increasing age. For those born in or after 1942, the mean was 1,770 CPM; for those born between 1923 and 1941, the mean was 1,606 CPM; and for those born in or before 1922, the mean count' was 1,238 CPM. The mean count generally decreased with increasing drink-years, with a maximum mean count of 1,726 CPM for nondrinkers and a minimum mean count of 1,414 CPM for participants with greater than 100 drink-years. A marginally significant association was found with occupation (p=0.086). The average mean count for officers was lower than that for enlisted flyers, which was in turn lower than that for enlisted groundcrew: the means were, respectively, 1,592 CPM, 1,662 CPM, and 1,713 CPM. This relationship with occupation was not seen with the PWM or MLC Unstimulated responses. Since these values were derived from the same blood specimens and were Unstimulated, this observation may represent a chance occurrence. There was a marginally significant association with pack-years (p=0.051); the response generally increased with increasing pack-years. �Association Between Functional Stimulation Test Variables and the Covariates in the Combined Ranch Hand and Comparison Groups (Directionality Shown) Variable Race Occupation Age Current Alcohol (Drinks/Day) DrinkYears Unstimulated Response (PHA) PHA Net Response Pokeweed Net Response MLC Net Response a Monotone decreasing. Generally decreasing trend. c Flat for 0 and first few categories, then increases. d lncreases,- drop-off at highest category. e Generally increasing trend. f Monotone increasing. 9 Increases from 0 category, then decreases, but not back to same level. NS: N: B: 0: F: G: Not significant ( > . 0 . p01) Nonblack Black Officer Enlisted flyer Enlisted groundcrew Current Smoking (Cigarettes/Day) Lifetime Smoking Pack-years �TABLE 19-8. Adjusted Analyses for fractional Stimulation Tests by Group Group Variable Statistic* Ranch Hand Unstimulated n Response Mean (PHA) 95% C I .. 44 6 1,741 (,9, 191 155 ,0) Covariate Comparison p-Value 54 8 1,731 (,9, 182 153 ,8) 085 .5 Remarks BATCH ( < . 0 ) p001 DAY (BATCH) ( < . 0 ) p001 RACE ( < . 0 ) p001 AGE ( < . 0 ) p001 BATCH (p<0.001) DAY (BATCH) (p<0.001) RACE (p=0.011) AGE*CSMOK (p=0.007) AUXSMOK (p=0.008) PHANet Response n Adj. Mean 95% C I .. 461 581 023 .3 1892 8,5 1320 9,8 (7,3, 2059 (7,8, 2604 1602 1,2) 1189 0,1) Net Pokeweed Response n Mean 95% C I .. 463 9,6 157 (219 1 1 4 1 8,8 0 , 5 ) 52 8 9,9 007 (108 9 , 6 ) 8,0, 9 6 9 059 .7 BATCH (p<0.001) DAY (BATCH) (p<0.001) RACE*OCC (p=0.024) ALC*OCC (p=O.036) ALC*CSMCK (p=0.009) Net MLC Response n Adj. Mean 95% C I .. 40 3 50 5 ** ** BATCH (p<0.001) DAY (BATCH) (p=0.001) EMOR (p<0.001) ALC (p=0.001) GRP*PACKSR (p=0.046) RACE*CSMOK (p=0.043) *Group means and confidence intervals expressed as counts per minute (CPM). ****Group-by-covariate interaction—adjusted mean, confidence interval, and p-value not presented. 19-22 �(The means were 1,657 CPM, 1,696 CPM, 1,519 CPM, and 1,712 CPM for 0, greater than 0 to 20, greater than 20 to 40, and greater than 40 pack-years, respectively.) Adjusted analyses to assess the group difference in mean counts of PHA controls were performed with adjustment for each covariate and any interactions that made significant contributions, including significant blood-draw day variation. The group difference in adjusted mean count was not significant (p=0.855; Ranch Hand group mean of 1,741 CPM versus Comparison group mean of 1,731 CPM). Race and age were significant covariates (p<0.001 for both). Adjusted and unadjusted analyses using the Original Comparisons (see Tables Q-8 and Q-9 of Appendix Q) showed similar results; i.e., no significant group difference (p=0.608, unadjusted; p=0.613, adjusted). PHA Net Response No significant difference was found between groups in the mean values of net response to PHA (p=0.339). This contrast was analyzed without adjustment for any covariates except blood-draw day variation. The data were pooled for the two groups and the relationship with the covariates was examined. Significant associations were found for race (p=0.002), age (p<0.001), current alcohol use (p<0.001), and drink-years (p=0.002). Nonblacks had a lower net count than Blacks (208,953 CPM, 233,622 CPM, respectively). There was a steady decrease in net count with increasing age: the means were 217,003 CPM, 206,901 CPM, and 184,419 CPM for those born in or after 1942, born between 1923 and 1941, and born in or before 1922, respectively. Those currently drinking more than four drinks per day had a lower mean net count than those drinking less. The participants in the greater than 100 drink-year category had a lower mean net count than those with fewer drink-years. Using a general linear model with adjustment for each covariate and any significant interactions including blood-draw day variation, the adjusted group difference was found to be not significant (p=0.233; Ranch Hand count of 193,280 CPM versus Comparison count of 188,952 CPM). Significant contributions were made b'y race (p=0.011), an age-by-current smoking interaction (p=0.007), and a current alcohol use-by-current smoking interaction (p=0.008). A marginally significant (p=0.057) group-by-occupation interaction was excluded from the final model. However, this interaction was explored, and was found to be due to a group difference among enlisted flyers (p=0.014); the adjusted mean Ranch Hand net stimulated count was greater than that of the Comparisons (207,050 CPM and 185,344 CPM, respectively). Analyses using the Original Comparisons (see Tables Q-9 and Q-10 of Appendix Q) revealed a significant group-by-occupation interaction (p=0.017), with results similar to the Ranch Hand versus total Comparison contrast of net counts; namely, enlisted flyer Ranch Hands had an adjusted mean count greater than enlisted flyer Original Comparisons (p=0.003). 19-23 �Pokeweed Net Response No significant difference was found between groups in the mean values of net response to pokeweed (p=0.317). These data were analyzed without adjustment for any covariates except blood-draw day variation. Significant associations were found using the pooled group data with drink-years (p=0.038), current smoking (p<0.001), and pack-years (p=0.001). The mean count was higher for those with greater than 100 drink-years and lower for never-drinkers, but with no pattern for the in between categories. For both current and lifetime smoking (pack-years), there was a steady upward trend in mean counts with increasing levels of smoking. The difference in adjusted group means was not significant: Ranch Hands, 91,567 CPM, and Comparisons, 90,097 CPM (p=0.579). The following interactions were significant: race-by-occupation (p=0.024), current alcohol use-by-occupation (p=0.036), and current alcohol use-by-current smoking (p=0.009). Net Response to MLC Stimulation No significant difference was found between groups in the mean response to MLC stimulation (p=0.185). These data were analyzed without adjustment for any covariates except blood-draw day variation. The data were pooled for the two groups, and the relationship with the covariates was examined. Significant associations were found for age (p=0.035), drink-years (p=0.008), current smoking (p<0.001), and pack-years (p=0.015). The net mean count generally decreased with increasing age: 84,543 CPM, 72,408 CPM, and 79,081 CPM for those born in or after 1942, between 1923 and 1941, and in or before 1922, respectively. The net mean count was lowest for never-drinkers, with no clear pattern among the drinkers: 66,933 CPM, 78,555 CPM, 80,713 CPM, 84,236 CPM, and 80,416 CPM for the 0, greater than 0 to 5, greater than 5 to 30, greater than 30 to 100, and greater than 100 drink-year categories, respectively. There was a monotonically increasing trend in net average count with current smoking, the nonsmokers having a much lower value than the smokers. An equivalent pattern was found for lifetime smoking (pack-years). Adjusted analyses were carried out to assess the group difference in mean counts of MLC net response, including adjustment for the significant blood-draw day variation. A significant group by pack-year interaction was found (p=0.046). Never-smoking Ranch Hands had a lower adjusted mean count (68,921 CPM) than the corresponding Comparisons (77,232 CPM) (p=0.053). Ranch Hands in the greater than 0 to 20 pack-year category had a lower adjusted mean count (67,976 CPM) than the corresponding Comparisons (74,333 CPM) (p=0,057). The adjusted means for the Comparisons decreased with increasing pack-years, whereas those of the Ranch Hands generally increased (see Table Q-l of Appendix Q). Significant contributions were made to the model by drink-years (p<0,001), current alcohol use (p=0.001), and a race-by-current smoking interaction (p=0.043). 19-24 �Discussion The performance of the phenotypic and cell stimulation studies was monitored daily by highly structured quality assurance techniques (see Chapter 6). This resulted in a remarkably error-free data set, in contrast to the immunologic tests at the Baseline study that required the assistance of a review group to determine which data were appropriate for analysis. The finding of significant blood-draw day and batch-to-batch variation at the followup examination was judged to be totally normal and inherent within the test procedures; only a few data points within specific variables were omitted because of outlying values. The unique use of a "batch" variable for adjustment of all the phenotypic and stimulation studies permitted unadjusted and covariate-adjusted group contrasts while controlling for inherent laboratory variation. All unadjusted and adjusted analyses (without group interactions) showed no significant group differences. Analysis of MLC revealed a group-by-packyear of smoking interaction with lower counts in the Ranch Hand group than in the Comparison group for 0 and greater than 0 to 20 pack-year categories. Despite differences in the quality of Baseline and followup results, slight changes in cohort numbers, and different mathematical models, there was remarkable concordance in the immunologic results of both examinations, both for the dependent variables and for the effects of the covariates. No judgment of adverse immunologic competence was made for any variable, or sets of variables, or in substrata examined because of group-by-covariate interactions for the cell surface marker and cell stimulation studies. EXPOSURE INDEX ANALYSES Within each occupational category, exposure index analyses were conducted to assess possible dose-response relationships (see details in Chapter 8). Analyses were performed for the cell surface marker variables (total T cells, helper T cells, suppressor T cells, B cells, monocytes, HLA-DR cells, and the T4/T8 ratio) and for the functional stimulation tests (the control counts per minute for the PHA test, and the net PHA, PWM, and MLC counts per minute). Analyses were not done for the skin test responses. Unadjusted and adjusted analyses were performed using general linear models. Exposure index-by-covariate interactions were explored in the adjusted analyses. Covariates were age, race, current and lifetime alcohol use (drink-years), and current and lifetime cigarette smoking (pack-years). For each analysis, an overall test was made of the differences among the means corresponding to the low, medium, and high exposure index levels. Medium versus low and high versus low contrasts of means were also made. Results of the adjusted analyses are .presented in Table 19-9 for cell surface markers and 19-10 for functional stimulation tests. Parallel results of unadjusted analyses are given in Tables Q-2 and Q-3, Appendix Q. Results of exposure index-by-covariate interactions are also given in Table Q-4 of Appendix Q. 19-25 �Adjusted Exposure Index Analyses for Cell Surface Markers by Occupation Exposure Index Variable Total T Cells Occupation Officer Enlisted Flyer Enlisted Groundcrew Helper T Cells Officer Enlisted Flyer Enlisted Groundcrew Statistic Low Medium High Contrast p-Value �Adjusted Exposure Index Analyses for Cell Surface Markers by Occupation Exposure Index Variable Suppressor T Cells Occupation Officer Enlisted Flyer Enlisted Groundcrew B Cells Officer Enlisted Flyer Enlisted Groundcrew Statistic Low Medium High Contrast p-Value �Adjusted Exposure Index Analyses for Cell Surface Markers by Occupation Exposure Index Variable Monocytes Occupation Officer Enlisted Flyer Enlisted Groundcrew HLA-DR Cells Officer Enlisted Flyer Enlisted Groundcrew Statistic Low Medium High Contrast p-Value �Adjusted Exposure Index Analyses for Cell Surface Markers by Occupation Exposure Index Variable T4/T8 Ratio Occupation Officer Enlisted Flyer Enlisted Groundcrew Statistic Low Medium High Contrast p-Value �Adjusted Exposure Index Analyses for Functional Stimulation Tests by Occupation Exposure Index Variable Unstimulated Response (PHA) Occupation Officer Enlisted Flyer Enlisted Groundcrew PHA Net Response Officer Enlisted Flyer Enlisted Groundcrew Statistic Low Medium High Contrast p-Value �Adjusted Exposure Index Analyses for Functional Stimulation Tests by Occupation Exposure Index Variable Pokeweed Net Response Occupation Statistic Low Medium High Contrast Officer Enlisted Flyer Enlisted Groundcrew MLC Net Response Officer Enlisted Flyer Enlisted Groundcrew *Group-by-covariate interaction—adjusted mean, confidence interval, and p-value not presented. p-Value �Cell Surface Markers Unadjusted analyses revealed very few significant results. Among enlisted groundcrew, the medium exposure level had a significantly lower mean total T cell count than the low exposure level (1,555 cells/mm versus 1,759 cells/mm , p=0.032), and the high exposure level mean was marginally significantly (p=0.091) lower than the low exposure level mean (1,586 cells/ mm versus 1,759 cells/mm ) . Suppressor T cells, for enlisted groundcrew in the low exposure level, were marginally significantly higher than in the medium or high exposure levels (575.5, 502.3, 505.5 cells/mm , respectively? medium versus low, p=0.063, high versus low, p=0.097). For enlisted flyers, the trends with exposure level were steadily downwards for total T cells, helper T cells, B cells, and the T4/T8 ratio, and upwards for suppressor T cells and monocytes, but no contrasts were significant. Adjusted analyses revealed marginally significant differences among exposure levels of enlisted groundcrew for total T cells (p=0.068) and suppressor T cells (p=0.088). For both total and suppressor T cells, the means for the medium and high exposure levels were much lower than those of the low exposure level. For total T cells, the adjusted means were: low, 1,737 cells/mm ; medium, 1,533 cells/mm 5 and high, 1,558 cells/mm (medium versus low: p=0.029, high versus low: p-0.085). For suppressor T ce].ls, the adjusted means were: low, 558.6 cells/mm ; medium, 4 0 8 cells/mm ; and 8. high, 483.7 cells/mm (medium versus low p=0.044, high versus low p=0.081). A similar but less marked pattern was seen for helper T cells. In summary, there was no consistent evidence of any significant doseresponse pattern in an occupational category. For the enlisted flyer cohort, six of the seven variables revealed nonsignificant dose-response trends in the unadjusted analyses, but only two trends persisted after adjustment by the covariates. Functional Stimulation Tests Exposure index analyses were performed on PHA unstimulated responses, and net PHA, PWM, and MLC counts. For officers, the unadjusted mean PHA unstimulated response counts varied significantly among exposure levels (p=0.047). The means were 1,705 CPM, 1,428 CPM, and 1,809 CPM, respectively, for low, medium, and high exposure levels (medium versus low: p«0.071, high versus low: p=0.557). The PWM net count for enlisted flyers was significantly lower for the high versus the low exposure levels (55,480 CPM versus 92,847 CPM, p=0.011). The PHA net count had a downward trend with increasing exposure level for enlisted groundcrew. The PWM net count for officers had an increasing trend but there was no statistically significant difference among exposure levels. In the adjusted analyses, officers had a significant exposure index-bydrink-year interaction (p=0.011) for PHA controls, and an exposure index-byage interaction for the PHA net count (p=0.003) (see Table 19-11 for a summary of these interactions). .Although the numbers were small, the high 19-32 �Summary of Exposure Index by Covariate Interactions for Functional Stimulation Tests Variable Occupation Covariate Unstimulated Response (PHA) Officer Drink-years PHA Net Response Officer p-Value Age exposure-level nondrinking officers had a lower mean PHA control count than the low exposure level (1,557 CPM versus 3,273 CPM, p=0.031), and the high exposure level officers with more than 100 drink-years had a higher PHA control count than the corresponding low exposure group ( , 0 CPM versus 670 1,983 CPM, p=0.049). Officers born in or after 1942 had a lower PHA net count in the medium exposure level as contrasted with the low exposure level (153,534 CPM versus 261,397 CPM, p=0.002). Other adjusted analyses revealed that enlisted flyers had a lower PWM net count in the high exposure level as compared to the low exposure level (111,772 CPM versus 173,897 CPM, p=0.014), as in the unadjusted analysis. Enlisted groundcrew in the medium exposure level also had a marginally significantly higher MLC net count as compared to the low exposure level (78,259 CPM versus 61,403 CPM, p=0.097). In summary, there was no evidence for a strong dose-response relationship, but there was a trend for declining PWM and MLC net counts for enlisted flyers with increasing exposure level. SKIN TESTING RESULTS General Four skin test antigens, mumps, Candida albicans, Trichophyton, and staph-phage-lysate, were intradermally administered to 76 percent (1,759) of the participants on the first day of the examination. Skin tests were not given to the remaining 24 percent of the population because they had been selected to give blood for the immunological tests on the second day of their examination. Candida albicans and Trichophyton tests were administered ( . ml) at a 1:1000 weight/volume dilution because of clinical concern that 01 a 1:100 or higher concentration might induce significant skin reactions and cause morbidity in the active pilot population. Mumps was given at a dose of 2 complement-fixing units, and staph-phage-lysate was administered at a dose of 6-9 x 10 colony-forming units of Staph. aureus and 0.5 - 5 x 10 bacteriaphage plaque-forming units. �Three experienced technicians from the SCRF Allergy Division measured the size of both induration and skin erythema by the "pen method" at 24 and 48 hours after administration. Each reader was required to measure the skin reactions by a millimeter ruler and record length and breadth measurements at each of the four sites, refer exaggerated reactions to an allergist, collect medication use data, and sign the data form. The skin test data were interpreted by defined criteria, as given in Table 19-12. Other categories included: impairment noted, clinical correlation required; normal (versus abnormal) results; with medications noted; and refusal. Of the 1,759 participants with skin tests, 269 were excluded from the analyses for the following reasons; 205 due to missing reader signature or failure of the participant to report for the 48-hour reading; 58 because of immunosuppressive medication, cancer chemotherapy, or x-ray therapy; 3 for impaired hypersensitivity requiring more tests; and 3 due to refusal. Readings at 24 hours were not analyzed since these readings occurred prior to peak response to the antigens. Statistical Analyses and Interpretations The initial analytical intent was to test Ranch Hand-Comparison group differences in skin test response by standard models, using both discrete and continuous data. In the preanalysis of the continuously distributed data (length by width measurement of the skin reactions), there was a suggestion of profound reader variation. This observation generated a series of contrasts between the readers prior to group testing. Figures 19-1 through 19-6 show contrasts of the three skin test readers from the tests of mumps and Trichophyton. Each graph shows the individual plots of the 48-hour induration square area measurement versus the 48-hour TABLE 19-12. Clinical Interpretation Categories of Skin Test Results by Specific Measurement Criteria at SCRF Clinical Interpretation Category Measurement Criteria Normal (Delayed Cutaneous Hypersensitivity Intact) Length (L) induration @ 48 Hrs >10 mm Width (W) induration @ 48 Hrs >~10 mm on any one of four skin tests Probably Normal (Probably Intact Delayed Cutaneous Hypersensitivity) L induration @ 48 Hrs > 5-<10 mm W induration @ 48 Hrs > 5-<10 mm on any one of four skin tests Possibly Anergic L, W induration o_r erythema @ 48 Hrs >0-<5 mm on any one or more of four skin tests Anergic L and W, induration at 48 hrs = 0 on all skin tests. 19-34 �10 - v£> I U> Ul 4 5 Log (Area of 48-Hour ERYTH Meas + 1) Figure 19-1. Relationship of Induration Measurements to Erythema Measurements for the Mumps Skin Test Reader 1 Results �10 - 4 5 6 Log (Area of 48-Hour ERYTH Meas + 1) Figure 19-2. Relationship of Induration Measurements to Erythema Measurements for the Trichophyton Skin Test. Reader 1 Results i 9 1 10 �104 Log (Area of 48-Hour ERYTH Meas + 1) Figure 19-3. Relationship of Induration Measurements to Erythema Measurements for the Mumps Skin Test Reader 2 Results �0 10 �Log (Area of 48-Hour INDUR Meas + 1) i 3 4 5 6 Log (Area of 48-Hour ERYTH Meas + 1) Figure 19-5. Relationship of Induration Measurements to Erythema Measurements for the Mumps Skin Test Reader 3 Results I 10 �' o (Area of 48-Hour INDUR Meas + 1) I 0 I 2. I I I I 3i 4 5 6 Log (Area of 48-Hour ERYTH Meas + 1) Figure 19-6. Relationship of Induration Measurements to Erythema Measurements for the Trichophyton Skin Test Reader 3 Results i 8 10 �square area erythema measurement by specific skin test and reader. These measurements are presented in log units to centralize the outlying values. These analyses were done because the size of induration rarely exceeds the size of the erythema reaction. Thus, each of the depicted graphs shows a line of values with the sizes of erythema equal to the size of induration; this line, and all values on, or to the lower right of the line, are labeled "clinically acceptable" values. All values above and to the left of the line, deemed "clinically unacceptable," are probably due to hurried measurements by inspection (rather than the pen method) or recording errors. These figures demonstrated a marked difference in the occurrence of clinically unacceptable results between readers for comparable tests. Specifically, Reader 2's measurements revealed a higher proportion of clinically unacceptable results than those observed with Readers 1 and 3. Further, the graphs supported some variation in the clinically acceptable measurement values between Reader 1 and Reader 3. Because of these discordances, further analyses of the continuously distributed data were abandoned in favor of discretized analyses. Categorical analyses were conducted on two parameters of the skin testing results, the area measurement relationship of induration to erythema, and the clinical interpretation of the skin test readings. Each of the three readers was compared for 48-hour measurements on the same skin test, categorizing the induration-erythema relationship as (1) Induration (I) equals Erythema (E) (both values equal to zero), (2) E greater than I, (3) I equals E, and (4) I greater than E. As previously noted, only the category of I greater than E was judged clinically unacceptable. An analysis of these four categories, by reader, for each of the four skin tests, showed a profound statistical difference (p<0.001) between the readers for all four skin tests. An average of the percentages for each category by reader is shown in Table 19-13, exemplifying the marked differences (a p-value is inappropriate due to the averaging). Induration Erythema Relationships in Average Percentage Over Four Skin Tests, by Reader In Percent Reader *I:Induration E:Erythema I=E=Zero* E>I I=E I>E �These data show marked reader differences for the category I greater than E. The magnitude of clinically unacceptable results ( 0 0 on the 3.% average for four skin tests) for Reader 2 (visually shown in Figures 19-3 and 19-4) strongly suggested that this entire data set was invalid. Further, the data pattern from Reader 2 was shown to be uniform over time, confirming the existence of a consistent bias. In this light, the existence and magnitude of a reverse error for Reader 2, i.e., misreadings of I equals E equals 0, E greater than I, and I equals E, seem plausible, but unestimable. Of these three categories, I equals E equals 0 is the most clinically important (suggesting anergy), and Table 19-13 provides clear evidence of a negative bias, with Readers 1 and 3 showing over three times more average detection of anergy than Reader 2. Also of interest in Table 19-13 are the substantial differences in the categories E greater than I and I equals E for Readers 1 and 3. Analyses of the four skin tests by erythema-induration relationships showed statistically significant differences beween Readers 1 and 3 for all four tests (p<0.001 for mumps, Candida albicans, and Trichophyton, and p=0.036 for staph-phage-lysate). The decision to remove Reader 2 data from subsequent analysis was agreed to by all the Principal Investigators, recognizing the minimal role of erythema as a contemporary indicator of anergy. This decision was based on the concern that an error in erythema measurement likely indicated an error in measurement of induration (the predominant indicator of anergy). In preparation for the analysis of group data remaining from Readers 1 and 3, it was noted that the clinical interpretations (see Table 19-12) from these valid readings were inconsistent over time of the study. Specifically, 80 percent of relative anergy and anergy occurred in the first 10 of 81 groups of participants (or 2 1/2 months of the 9-month examination period). Further, the proportion of diagnoses of anergy between the allergists was disproportionate. The value of these analyses was therefore reduced. SUMMARY AND CONCLUSIONS Immunologic competence was measured by cell surface marker (phenotypic) studies and cell stimulation studies on 47 percent of the study population, and by a four antigen series of skin tests in 76 percent of participants to assess the delayed hypersensitivity response. Table 19-14 summarizes the results of all unadjusted and adjusted analyses on 11 primary variables spanning the first two of these three functional areas. Cell surface marker studies were conducted for total T cells (T-1), helper T cells (T4), suppressor T cells (Tg), B cells, monocytes, and HLA-DR cells; the ratio of T /T8 cells was included in the analysis. Because of inherent significant day-to-day and batch-to-batch variation, all results (including functional stimulation studies) were adjusted for blood-draw day variation. Statistical testing of the seven phenotypic cell markers did not reveal any significant group differences (interactions excepted), either unadjusted or adjusted for the covariates of age, race, occupation, current smoking, lifetime smoking history (pack-years), current alcohol use, or lifetime alcohol use (drink-years). Similarly, none of the unadjusted or adjusted analyses of the functional stimulation studies (for phytohemagglutinin, pokeweed mitogen, or mixed lymphocyte culture) showed any 19-42 �Overall Summary Results of Unadjusted and Adjusted Analyses of Immunological Variables Variable Unadjusted Adjusted Total T Cells (T ) Helper T Cells (T4) Suppressor T Cells (T ) B Cells Monocytes HLA-DR Cells T4/T8 Ratio Unstiraulated Response (PHA) PHA Net Response Pokeweed Net Response MLC Net Response NS:Not significant ( > . 0 . p01) ****Significant group-by-covariate interaction. statistically significant group differences. However, the adjusted analyses for total T cells, B cells, raonocytes, HLA-DR cells, pokeweed mitogen, and net mixed lymphocyte culture stimulation showed some significant 'group-bycovariate interactions, precluding direct adjusted group contrasts. Overall, no discernible pattern was identified to suggest a detriment in any subgroup of either the Ranch Hands or Comparisons. Results were similar between the analyses of the total Comparison group and the analyses of the Original Comparisons. The covariate effects of age, race, smoking, and alcohol use were generally profound on most variables in the phenotypic and stimulation studies. Consistently decreasing values of all cell markers and stimulated cells were associated with increasing age, whereas increased levels of smoking were usually associated with increases in the values of those variables. Blacks had consistently higher stimulated cell counts than nonblacks, but this effect was not observed for counts of T cells, B cells, or HLA-DR cells. Enlisted personnel generally had higher cell surface marker counts than officers. Exposure index analyses of cell surface markers revealed no pattern consistent with a dose-response relationship. For enlisted groundcrew, the mean total T cell and suppressor T cell counts for the medium exposure level were significantly lower than those of the low exposure level, but were slightly lower than those of the.high exposure level. The exposure index analyses of the functional stimulation tests revealed no consistent significant dose-response patterns for net PHA counts or net MLC counts. For net pokeweed counts, enlisted flyers in the high exposure level had a significantly lower adjusted count than enlisted flyers in the low exposure level, and a decreasing trend was apparent. �The delayed hypersensitivity response was assessed by the skin test antigens of mumps, Candida albicans, Trichophyton, and staph-phage-lysate. The 48-hour measurements of skin induration and erythema for the four tests showed marked inter-reader variation. Analyses showed that one of the three skin test readers too often measured induration larger than erythema (a clinically unacceptable finding), in an average of 30 percent of the readings, and did not yield measurements that detected a case of possible or overt anergy, whereas the other two readers found this condition in 5.6 percent of the participants. Remaining data from Readers 1 and 3, however, were found to vary significantly in clinical interpretation over duration of the examination. Consequently, all skin test data were declared invalid, and were not used in the assessment of group differences. The skin test reading problems led to the use of additional clinical quality control procedures for the AFHS followup examination begun in May 1987. In conclusion, no significant group differences were judged present for the comprehensive cell surface marker or functional stimulation studies. The profound effects of age, smoking, and alcohol use were observed in these immunologic tests. The assessment of delayed hypersensitivity skin responses was precluded by poor data quality and excluded from further analysis. Overall, there was no indication of impaired immunologic competence in either group. 19-44 �CHAPTER 19 REFERENCES 1. Vos, J.G., J.A. Moore, and J.G. Zinkl. 1973. Effect of 2,3,7,8tetrachlorodibenzo-p-dioxin on the immune system of laboratory animals. Environ. Health Perspec. 5:149-162. 2. Zinkl, J.G., J.G. Vos, J.A. Moore, and B.N. Gupta. 1973. Hematologic and clinical chemistry effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in laboratory animals. Environ. Health Perspec. 5;111-118. 3. Vos, J.G., and J.A. Moore. 1974. Supression of cellular immunity in rats and mice by maternal treatment with 2,3,7,8-tetrachlorodibenzop-dioxin. Int. Arch. Allerg. Appl. Immunol. 47:777-794. 4. Thigpen, J.E., R.E. Faith, K.E. McConnell, and J.A. Moore. 1975. Increased susceptibility to bacterial infection as a sequela of exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Infect, and Immun. 12(6):1319-1324. 5. Faith, R.E., and J.A. Moore. 1977. Impairment of thymus-dependent immune functions by exposure of the developing immune system to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). J. Toxicol. Environ. Health 3:451-465. 6. McNulty, W.P. 1977. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin for Rhesus monkeys: Brief report. Bull. Environ. Contam. Toxicol. 1()1819 81:0-0. 7. Faith, R.E., M.I. Luster, and J.A. Moore. 1978. Chemical separation of helper cell functions and delayed hypersensitivity responses. Cellular Immunol. 40:275-284. 8. Vos, J.G., J.G. Kreeftenberg, H.V.B. Engel, A. Minderhoud, and L.M. Van Noorle Jansen. 1978. Studies on 2,3,7,8-tetrachlorodibenzo-p-dioxin induced immune suppression and decreased resistance to infection: Endotoxin hypersensitivity, serum zinc concentrations and effect of thymosin treatment. Toxicology 9:75-86. 9. McConnell, E.E., J.A. Moore, and D.W. Dalgard. 1978. Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in Rhesus monkeys (Macaca mulatta) following a single oral dose. Toxicol. Appl. Pharmacol. 43(10):175-187. 10. Sharma, R.P., and P.J. Gehring. 1979. Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on splenic lymphocyte transformation in mice after single and repeated exposures. Ann. N.Y. Acad. Sci. 320:487-497. ~~ 11. Faith, R.E., and M.I. Luster. 1979. Investigations on the effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on parameters of various immune functions. Ann. N.Y. Acad. Sci. 320:564-571. 19-45 �CHAPTER 19 REFERENCES (Continued) 12. Dean, J.H., M.I. Luster, G.A. Boorman, K. Chae, L.D. Lauer, R.W. Luebke, L.D. Lawson, and R.E. Wilson. 1981. Assessment of immunotoxicity induced by the environmental chemicals 2,3,7,8-tetrachlorodibenzop-dioxin, diethylstilbestrol and benzo(a)pyrene. In Advances in Immunopharinacology, ed. J. Hadden, L. Chedid, P. Mullen, and F. Spreafico, pp. 37-50. New York: Pergamon Press. 13. Clark, D.A., J. Gauldie, M.R. Szevczuk, and G. Sweeney. 1981. Enhanced suppressor cell activity as a mechanism of immunosuppression by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Proc. Soc. Exp. Biol. Med. 168:290-299. 14. Poland, A. 1984. Reflections on the mechanism of action of halogenated aromatic hydrocarbons. In Banbury report 18: Biological mechanisms of dioxin action, ed. A. Poland and R.D. Kimbrough, pp. 109-117. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory. 15. Knutsen, A.P. 1984. Immunologic effects of TCDD exposure in humans. Bull. Environ. Contam. Toxicol. 33:673-681. 16. May, G. 1982. Tetrachlorodibenzodioxin: A survey of subjects ten years after exposure. Br. J. Ind. Med. 39:128-135. 17. Hay, A. 18. Sirchia, G.G. 1982. Exposure to TCDD: Immunologic effects. In Plans for clinical and epidemiologic follovup after area-wide chemical contamination; proceedings of an international workshop, Washington, D.C., March 1980. Washington,. D.C.: National Academy Press. 19. Hoffman, R.E., P.A. Stehr-Green, K.B. Webb, G. Evans, A.P. Knutsen, W.F. Schramm, J.L. Staake, B.B. Gibson, and K.K. Steinberg. 1986. Health effects of long-term exposure to 2,3,7,8-tetrachlorodibenzop-dioxin. JAMA 255:2031-2038. 20. Grubbs, F.E. 1969. Procedures for detecting outlying observations in samples. Technometrics XI:1-21. 1981. Dioxin hazards: Secrecy at Coalite. 19-46 Nature 290:729. �CHAPTER 20 PULMONARY DISEASE INTRODUCTION Pulmonary dysfunction and overt pulmonary disease are not recognized clinical entities resulting from exposure to chlorophenols or TCDD. Acute exposure to chlorophenols, phenoxy herbicides, and TCDD, have caused the traditional acute symptoms of cough, nasal/lung irritation, shortness of breath, and, occasionally, bronchitis. These symptoms have been noted almost exclusively in industrial workers and not in individuals experiencing casual contact. Long-term sequelae arising from the acute symptom stage in ill individuals have not been generally known because of minimal followup and surveillance of the pulmonary symptoms. Only one contemporary morbidity study has attributed pulmonary dysfunction to phenoxy herbicide and TCDD exposure. The percent abnormal pulmonary parameters of forced expiratory volume (FEV), forced vital capacity (FVC), forced expiratory volume in one second (FEV^/FVC ratio, and forced midexpiratory flow rate (FEF25_75) were significantly higher in exposed workers who currently smoke, than in nonexposed workers who smoke. In considerable contrast, these test parameters were essentially equal in nonsmokers and former smokers of both the exposed and nonexposed groups. The effect of current smoking persisted after a logistic regression analysis adjusting for pack-years of cigarette smoking. Adjusted means of the test parameters FEV, FVC, and FEV1/FVC also showed significant differences for current smokers but not for nonsmokers or former smokers. As with other nonclassical clinical endpoints, prior investigators perhaps undervalued the incorporation of pulmonary disease and function into their study protocols. Further, due to the profound effect of smoking on pulmonary function, great emphasis must be placed in the collection of highly accurate, detailed, and validated smoking data as an adjustment variable, a process that is not straightforward in today's environment of antismoking. The only recent data comparable to this study are found in the 1984 AFHS Baseline Morbidity Report, which is reviewed below. Baseline Summary Results The 1982 Baseline examination explored historical pulmonary disease by questionnaire and active pulmonary function by standardized spirometric technique at the physical examination. These areas were of significant interest because of routine operational inhalation of Herbicide Orange by all Ranch Hand flying crewmen as well as ground maintenance personnel (Baseline Report Chapter 1, Buckingham). 20-1 �The questionnaire revealed no group differences for historical diagnoses of tuberculosis and fungal infections, pneumonia, cancer, or chronic sinusitis and upper respiratory disease. At the physical examination the unadjusted means for FEV1 (percent predicted), FVC, and the FEV /FVC ratio were almost identical between the Ranch Hands and Comparisons. Adjusted mean values were not calculated due to significant interactions (age, group, and pulmonary function for FEV1 and FVC; smoking with FEV1/FVC). Detailed exposure analyses showed two significant associations in the enlisted flyer and enlisted groundcrew strata, but neither was indicative of linear dose response. Attempts to adjust the means of the pulmonary function values for age and smoking revealed several interactions, but essentially negative results. Overall, there were no pulmonary disease or pulmonary function data or associations of concern. Parameters of the 1985 Pulmonary Examination Because of the essentially negative pulmonary analyses from the Baseline examination, pulmonary function (spirometric) studies were not performed during the first followup examination. Collection of pulmonary data was limited to a questionnaire history of respiratory disease, physical examination of the thorax and lungs, and pulmonary abnormalities detected on a routine chest x ray. Thus, t,he data analyses consist of group assessments of respiratory disease incidence, physical examination abnormalities, and the current prevalence of x-ray abnormalities. Covariate adjustments are made for age and smoking (yes, no, former, and pack-years). Minor numeric differences in the tables are due to rare missing dependent variable or covariable data. The analyses are based on 1,016 Ranch Hands and 1,293 Comparisons. No exclusions based on clinical conditions were made. Mortality due to respiratory disease, as of 31 December 1985, in the Ranch Hand and the 1:5 matched Comparison cohort is summarized. Morbidity data are analyzed using linear and loglinear models. RESULTS AND DISCUSSION Mortality Experience The mortality of the Ranch Hand and Comparison groups through 31 December 1985 was evaluated. There were seven deaths from respiratory system conditions in the Comparison group and none in the Ranch Hand group. This analysis was based on the 1:5 Ranch Hand to Comparison mortality study cohorts. Two of these deaths were Comparison flying officers, three were enlisted flyers, and the remaining two were enlisted groundcrew. Unadjusted Morbidity Analyses Analyses were performed on the history of respiratory illnesses as provided by the participants during the physical examination. The results of the 20-2 �radiological and clinical examination of the lungs and chest were also analyzed. These unadjusted analyses are summarized in Tables 20-1 and 20-2. As shown, no significant group differences were observed for history of asthma, bronchitis, pleurisy, pneumonia, or tuberculosis. Similar nonsignificant results were found in the evaluation of the clinical variables. Parallel analyses were conducted using data from the Original Comparisons, with comparable results (Appendix R, Table R-l). Adjusted Morbidity Analyses Statistical adjustment for the effects of age and lifetime smoking did not alter the findings of group similarity seen in the unadjusted analyses. Lifetime smoking was categorized as nonsmoking (0 pack-years), moderate (greater than 0 to 10 pack-years) and heavy (greater than 10 pack-years). These results are shown in Table 20-3. Lifetime smoking consistently exerts significant effects on nearly all historical illness and clinical examination variables, and age was an important factor for the history of pneumonia and the clinical assessment of thorax and lungs (representing an overall clinical assessment of normality/ abnormality in the respiratory system), chest asymmetry, the presence of hyperresonance, rales, and the presence of x-ray abnormality. There were significant or borderline significant group-by-pack-year interactions in analyses of a history of pleurisy and tuberculosis, for the presence of rales on examination, and for x-ray abnormality. There was also an interaction for asthma of borderline significance (p=0.068). A significant group-by-age interaction was seen for the presence of rales. The results of analyses stratified to clarify these interactions are shown in Table 20-4. Nonsmoking Ranch Hands had significantly more asthma (p=0.050) than their nonsmoking Comparisons, while the history of asthma was not significantly different in either category of smokers. Pleurisy was significantly more frequent in moderately smoking Ranch Hands (p=0.0001), but bordered on being significantly-increased in heavily smoking Comparisons (p=0.060). Analyses of a history of tuberculosis and the presence of rales was hampered by small numbers of cases in both groups (a total of 13 cases). The presence of several cells containing zeros makes interpretation of these analyses extremely difficult. Except in those strata with zero cells, no statistical significance was noted. In the analysis of x-ray abnormalities, the nonsmoking Ranch Hands had significantly less abnormality (p=0.030) than the nonsmoking Comparisons. Analyses of other strata did not reveal any significant group differences. These adjusted analyses were performed on data from the Original Comparisons, with similar results (see Tables 20-2 and 20-3). EXPOSURE ANALYSES The pulmonary data from the Ranch Hands were analyzed using the exposure index as a covariate (categorized as high, medium, or low within each occupational stratum). The percent abnormality at each level of exposure for each clinical or historical variable is presented in Tables 20-5, 20-6, and 20-7. 20-3 �Unadjusted Analyses of Reported History of Respiratory Illness by Group Group Ranch Hand Variable Asthma Bronchitis Pleurisy Pneumonia Tuberculosis Statistic Comparison Number Percent Number Percent Est. Relative Risk ( 5 C.I.) 9% p-Value �Unadjusted Analyses of Radiological and Clinical Respiratory System Findings by Group Group Ranch Hand Variable Thorax and Lungs Asymmetrical Expiration Hyperresonance Dullness Wheezes Rales X Ray Statistic Comparison Number Percent Number Percent Est. Relative Risk (95% C.I.) p-Value �Adjusted Analyses of Respiratory Variables by Group* Group Variable Ranch Hand Total Comparison Total Adj. Relative Risk. (95% C.I.) p-Value Covariate Remarks** As thma PACKYR (p=0.023) GRP*PACKYR (Borderline: p=0.068) Bronchitis None Pleurisy GRP*PACKYR (p^O.0026) Pneumonia AGE (p»0.0001) Tuberculosis GRP*PACKYR (p=0.034) Thorax and Lungs AGE (p<0.0001) PACKYR (p<0.001) Asymmetrical Expiration AGE*PACKYR (p=0.036) Hyperresonance AGE (p<0.0001) PACKYR (p<0.0001) Dullness None Wheezes PACKYR (p<0.0001) Rales GRP*AGE (p=0.046) GRP*PACKYR (Borderline: p=0.070) AGE*PACKYR (Borderline: p=0.090) AGE (p<0.0001) PACKYR (p»0.0019) GRP*PACKYR (Borderline: p=0.060) *Group-by-covariate interactions are described in Table 20-4. **Abbreviations PACKYR: Lifetime smoking history (pack-years) GRP: Group ****Group-by-covariate interaction—relative risk, confidence interval, and p-value not presented. �Suaaary of Group-by-Covariate Interactions for Respiratory Variables Group Ranch Hand Variable Interaction Asthma Group-byPack-Year Pleurisy Group-byPack-Year Stratification Statistic Number Percent Comparison Number Percent Adj. Relative Risk (95% C.I.) p-Value �Summary of Group-by-Covariate Interactions for Respiratory Variables Group Ranch Hand Variable Interaction Tuberculosis Group-byPack-Year Rales Group-byAge Stratification Statistic Number Percent Comparison Number Percent Adj. Relative Risk (95% C.I.) p-Value �Siuuary of Group-by-Covariate Interactions for Respiratory Variables Group Ranch Hand Variable Interaction Rales Group-byPack-Year X Ray Group-byPack-Year Stratification Statistic Number Percent Comparison Number Percent Adj. Relative Risk (95% C.I.) p-Value �TABLE 20-5. Exposure Index Analysis Results for Officers p-Values of Dependent Variable-by-Covariate Association*'b Variable D*EXP D*AGE D*PACKffi D*EXP *AGE D*EXP D*AGE D*EXP* *PAOOR *PACm AGE*PAOQR Abnormal Total Percent Asthma Bronchitis Pleurisy Pneumonia Tuberculosis Thorax and Lungs Asynrnetrical Exp. Hyperresonance Dullness Wheezes Pales XRay 'Dependent variable indicated by 0 in coluon headings. b Abbreviations: EXP: Exposure index. PACKYR: Pack-years. TABLE 20-6. Exposure Index Analysis Results for Enlisted Flyers p-Values of Dependent Variable-by-Covariate Association* Variable Asthma Bronchitis Pleurisy Pneumonia Tuberculosis Thorax and Lungs Asymmetrical Exp. Hyperresonance Dullness Wheezes Bales XRay D*EXP D*AGE D*PAOOR D*EXP *AGE D*EXP D*AGE D*EXP* *PAOOR *PAOCZR AGE*PAOOR Abnornal Total Percent �TABLE 20-7. Exposure Index Analysis Results for Bilisted Gcoundcrew: p-Values of Dependent Variable by Covariate Association* Overall Variable D*EXP D*AGE D*PACm D*EXP* *AG£ D*EXP D*AGE D*EXP* *PACKXR *PAOOR AG£*PACm Abnormal Total Percent Asthma Bronchitis Pleurisy Pneumonia Tuberculosis Thorax and Lungs Asynmetrical Exp. Hyperresonance Dullness Vheezes Rales XRay Two sets of analyses were performed on enlisted groundcrew data. In the first set of analyses, all three year-of-birth categories (born after 1942, born between 1922 and 1942, born before 1922) were used. In the second set of analyses, only those born between 1922 and 1942 and after 1942 were used, since only one enlisted groundcrew Ranch Hand was born before 1922. All testing results in the two sets of analyses were the same, except for the asthma-by-age interaction shown in Table 20-6. Each of the dependent variable-by-exposure category interactions are noted by occupation category in Appendix R, Tables R-4 through R-18. These data are considered too sparse for meaningful interpretation. SUMMARY AND CONCLUSIONS A summary of the results on the analyses of reported history of respiratory illness and of radiological and clinical findings is given in Table 20-8. Based on the 31 December 1986 mortality data, there were seven deaths from respiratory conditions in the Comparison group and none in the Ranch Hand group. �TABLE 2 - . 08 Overall Summary Results of Unadjusted and Adjusted Analyses of Pulmonary Disease Pulmonary Disease Unadjusted Adjusted Reported History of Respiratory Illness Asthma Bronchitis Pleurisy Pneumonia Tuberculosis Radiological and Clinical Findings Thorax and Lungs Asymmetrical Expiration Hyperresonance Dullness Wheezes Rales X Ray NS: Not significant (p>0.10) ****Group-by-covariate interaction. There were no group differences found for reported history of asthma, bronchitis, pleurisy, or tuberculosis based on the unadjusted analyses. Adjustments for age and lifetime smoking did not alter the findings of group similarity, although there was a significant group-by-pack-year interaction for pleurisy and for tuberculosis. Similarly, there were no significant group differences in the unadjusted analyses for the radiological and clinical respiratory findings of thorax and lungs, asymmetrical expiration, hyperresonance, dullness, wheezes, rales, and x-ray interpretations. These findings were supported by the adjusted analyses, although there was a group-by-age interaction for rales. The exposure index analyses revealed no consistent dose-response pattern. Analyses of past history of respiratory illness and the clinical and radiological examination of the chest and lungs did not reveal any statistically significant differences between the Ranch Hand and Comparison groups suggestive of herbicide related disease. Several group-by-covariate interactions did exhibit statistical significance, but these findings did not indicate any consistent patterns suggesting different disease experience in the two groups. �REFERENCES CHAPTER 20 1. Suskind, R.R., and V.H. Hertzberg. 1984. Human health effects of 2,4,5-T and its toxic contaminants. JAMA 251:2372-2380. 2. Lathrop, G.D., P.M. Moynahan, R.A. Albanese, and W.H. Wolfe. Epidemiologic Investigation of Health Effects in Air Force Following Exposure to Herbicides—Baseline Mortality Study Epidemiology Division, Data Sciences Division, USAF School Aerospace Mecicine, Brooks Air Force Base, Texas. 20-13 1983. An Personnel Results. of �CHAPTER 21 INTERPRETIVE CONSIDERATIONS This chapter reviews several scientific issues that should be considered when attempting to reach conclusions on a study of this size and complexity. These issues are critical to the interpretation of the data analyses in this report. Data patterns observed in many clinical chapters of this report are also summarized so that hypothesis testing of group differences may be placed in better perspective. DIOXIN ENDPOINTS Based upon data in this report, final conclusions on herbicide causality must consider results of the various clinical areas, reflected in the separate chapters. Each chapter introduction has attempted to highlight the major organ systems that are known or suspected to be significantly affected by the ingredients of Agent Orange with particular emphasis on the effects of dioxin. Categories of clinical endpoints and their generally accepted degree of association with dioxin are presented in Table 21-1. These associations are based on the scientific literature. TABLE 21-1. Summary Associations of Adverse Health Effects to TCDD Exposure Reported in the Literature Degree of Association by Clinical Chapter Confirmed Highly Suspected Moderately Suspected Dermatology Neurology Hepatic Malignancy General Health Immunology Negative or Weakly Suspected Psychology Cardiovascular Hematology Endocrine Renal Pulmonary It is recognized that alternative conclusions based on these patterns of association are possible within the framework of current knowledge, particularly for the highly and moderately suspected areas (malignancy, general health, immunology). However, for illustrative purposes, two extremes are presented: multiple adverse findings in the Ranch Hand group for the areas 21-1 �of dermatology, neurology, hepatic (discussed in Chapter 13), and cancer would suggest a case for TCDD causality, whereas multiple adverse findings in the weakly suspected areas, and not in any of the confirmed areas, would be difficult to ascribe to an overall TCDD causation. The aspects of biological plausibility and specificity require balanced interpretation across clinical chapters, with careful attention placed on nonsignificant findings as well as significant findings. The chapters in this report should be viewed as artificial boundaries for convenience of presentation, and should not discourage consideration of their relatedness, or of the individual variables within them. EXPOSURE Approximately 600 exposure index analyses have been conducted in this study, underscoring attempts to associate increasing proportions of various abnormalities to estimates of increasing exposure. To determine whether the results of the exposure analyses varied by chance, several perspectives were taken. Of the 255 adjusted exposure analyses (excluding 39 with interactions), 13 were statistically significant, a figure which is the expected number (based on a =0.05). It is recognized that this contrast is a crude yardstick, considering the relatedness of the dependent variables, statistical power, disproportionate representation of chapter variables, and the presence of interactions. The six possible patterns of exposure response (increasing, decreasing, V-shaped with fewer abnormalities at the low exposure level than the high exposure level, V-shaped with more abnormalities at the low exposure level than at the high exposure level, inverted V-shaped with fewer abnormalities at the low exposure level than the high exposure level, and inverted V-shaped with more abnormalities at the low exposure level than at the high exposure level) were tabulated (regardless of statistical significance) for the clinical chapters of dermatology, neurology, psychology, and renal. As noted in Table 21-1, two of these chapters contain clinical variables that have had confirmed associations to TCDD exposure, and two chapters have had negative or weakly suspected associations to TCDD. Of the 126 exposure analyses in these four chapters, 21 (or one-sixth) showed the primary pattern of interest, an increase—exactly the number expected. Taken together, these analyses suggest that statistically significant exposure analyses may have occurred due to chance among the data set, and that the pattern of dose-response may also have been random. These inferences, or that the exposure index was unrelated to actual exposure, together with the acknowledged limitations of the exposure index, indicate that estimated exposure may only be weakly relied upon to assert a causal relationship. Based upon the current exposure index calculations, either of the above inferential alternatives is possible. The use of serum dioxin levels (see Chapter 23, Future Directions) in the next report will clarify the exposure calculations of this report and the Baseline Report. Thus, from an interpretive context, final conclusions on dose-response, and the implications to herbicide causation are based on current knowledge available for this report. These conclusions could change with future analyses using a factual exposure concept. 21-2 �TYPES OF MEASUREMENTS This report includes all types of measures traditionally used in morbidity followup epidemiologic studies, e.g., self-reports, structured interview responses, medical record data, physician findings, scalar measurements, biopsy results, laboratory determinations, morbidity indices, and mortality results. At many points in this report, various terms have been used to qualitatively describe the data and analyses arising from the measurement processes. In particular, the terms "subjective," "objective," "continuous," and "categorical," and "constructed indices" have been used to connote differences in data or data sets that are important in making statements of inference. From the perspective of the Study Protocol, significant group differences for subjective historical variables, not mirrored by significant group differences in medical record findings or physician/laboratory testing, may be viewed as preliminary evidence of over-reporting by a group. The opposite finding of significant group differences for physical examination variables in the absence of reported symptoms may support the primary conclusion of significant subclinical group differences. Either of these alternatives may greatly affect an overall inference of herbicide causality. Hence, the descriptive phrases "subjective data" and "objective data" have not been used as value judgments of the worth of the data, but simply as inferential qualifiers. This report contains numerous comments on the differences in results between analyses of continuous versus categorical data from the same variable (exclusively laboratory data). Because the statistical power is stronger for detecting mean shifts than categorical differences, it was anticipated that very small mean shifts might be more easily discerned.than differences in proportions of abnormalities between the two groups. Both methods of examining the data reveal important aspects of the distribution. Inferentially, when both types of analyses were done, greater weight has been given to significant group differences when analyses of both data forms agree. Lesser weight was given to significant differences seen in only one analysis, and least weight to significant shifts in means if both group means were within normal range, and the mean difference was not supported by other statistical findings in related variables (e.g., hepatic test battery). Consistent patterns of findings within an organ system, or between related organ systems, is required to strongly suggest an inference of causality. Several summary indices were constructed in this report, e.g., dermatology index, cranial nerve function index, and anatomic categories of abnormal peripheral pulses, and are similar to some indices in the 1984 Baseline Report. They were formed by summing or grouping related abnormalities for the purposes of assessing increased numbers and/or showing group directionality of overall results. They should not be strongly considered in final inferences because they are artificially derived. BASELINE-FOLLOVUP EXAMINATION DIFFERENCES A common difficulty of followup studies is the inherent variation in measurement systems from one observation period to the next. To the maximum extent possible, the USAF has restricted clinical variation by requiring the use of identical laboratory equipment for most clinical chemistries, by the 21-3 �use of 50 samples from the Baseline serum bank to evaluate interexamination laboratory differences, and by the use of carefully prescribed written clinical procedures that allow little room for variation. Nonetheless, some interexamination variability must be expected, but in the presence of blindness to group membership, there is no reason to expect biases in the results with respect to either the Ranch Hand or Comparison groups. This report has cited classical longitudinal analyses to assess changes in variables between the examinations by group. Of 21 variables examined, 5 showed statistically significant group differences in the changes between examinations. Four of these significant results were attributed to actual changes over time, while the other (e.g., sedimentation rate) was believed due to a change in laboratory methodology. Other less refined longitudinal contrasts consisting of narrative discussions of Baseline results versus followup results have been presented in all chapters. Interpretive caution is required in assessing examination similarities or differences because of the slight changes in cohort composition between the examinations (see Chapter 2, Population), the use of slightly different statistical models and modeling strategy (see Chapter 7, Statistical Methods), and sometimes the use of the Original Comparison group. The relative contribution of these changes was not explored mathematically, but is believed to have played a minimal role in accounting for any large group shifts between examinations. In the context of comparing results between examinations, there has been a subtle but consistent observation that group differences have substantially narrowed over the 3-year period, either by decreased findings in the Ranch Hands, increased findings in the Comparisons, or a combination of both mechanisms. In general, several broad interpretations are possible: any bona fide herbicide effect decreases over time, that the convergence is largely attributable to unquantifiable factors, that both examinations have produced chance results, or that these observations have been affected by the slight shifts in cohort composition and modeling strategy. Several segments of this report have noted marked differences in the prevalence rates of abnormalities found at the Baseline and followup specialty examinations, e.g., the dermatology and neurology clinical assessments. The followup dermatological examination detected substantially more abnormalities than the Baseline examination, whereas far greater numbers of neurological abnormalities were noted at the Baseline examination than at the followup for some variables. These examination variances were affected by differences in "clinical sensitivity" between the examining teams, although clearly other factors (such as a true change in disease-abnormality status or slight cohort differences) contributed. The phrase "clinical sensitivity" refers to the inherent differences in clinical styles and interpretations of possible abnormalities that often prevail. Because of examiner blindness to exposure status, and because of the judgment that the interexamination variation was within the artful bounds of accepted medical practice, no bias was thought to have resulted from this inherent variation. STUDY BIASES Each reviewer of this report must reach a conclusion on whether the results of this study have been seriously flawed by the design, the operation 21-4 �of significant biases, or both. The Protocol authors believe that the comprehensive multifaceted design is the chief strength of this study, although it is recognized that each and every published phase of the study must invite renewed inspection of fundamental scientific aspects of the study design. It is believed that, with the exception of skin test readings, all data in this study were collected accurately and validly, and that blindness to group membership was well maintained throughout the collection process. This opinion is important from an inferential perspective in that both misclassification of data (tending to dilute true group differences) and bias in data (creating a false group effect) most likely did not occur appreciably in this study. Thus, it is believed that both the magnitude and direction of the group results found in this study reflect truth to the maximum degree possible, within the inherent boundaries of statistical models to account for all important adjusting variables. GROUP INTERACTIONS: PATTERN RECOGNITION Many of the adjusted analyses in this report have demonstrated significant group-by-covariate interactions, requiring stratified analyses to determine the nature of significant group differences. All significant twoand three-factor interactions have been included in the main text or in appendices. The analysis of followup data has found substantially more interactions than the analysis of Baseline data, due primarily to the larger number of covariates used in the followup analyses. Several related viewpoints have aided in the overall interpretation of group-by-covariate interaction in the report. In the presence of a significant interaction, a direct conclusion on main group effects cannot be made, and the focal point of interpretation resides with the covariate stratum containing the significant group effect (or a reversal in nonsignificant group effects across strata). Past this point, however, there appears to be little consensus in how to best place the interaction into inferential context. Further interpretations appear to be largely individualistic. No consistent pattern has emerged to support a finding of impairment in the Ranch Hands for any specific stratum of one or more covariates. In fact, of all the two- and three-factor interactions encountered, only one was thought to have possible biologic relevance. Other interactions may have such relevance, but the reason was not apparent. As with tests of group differences, significant interactions may occur by chance, but the method to calculate an expected number of group-by-covariate interactions, unfortunately, remains an open research question. Because of the possible diverse interpretations of" interactions, all significant two- and three-factor interactions involving group with statistically significant strata are presented in Table 21-2 for detailed inspection. No particular covariate or group pattern is noted, although the variables in psychology and gastrointestinal showed Ranch Hands at a relative detriment, while the interactions in the cardiovascular .chapter indicated detrimental findings in the Comparisons. Most variables without interactions in this report have shown remarkable concordance between unadjusted and adjusted results, both in terms of absolute value of relative risk and of statistical significance. 21-5 �TABLE 21-2. Summary of Significant Covariate Strata (or Covariate Level Difference) Found Within Significant Two- and Three-Factor Group-by-Covariate Interactions by Clinical Chapter and Dependent Variable (Group Direction and p-Value) Clinical Chapter Dependent Variable Covariate Stratum RH>C ORH p-Value General Health Self-Perception of Health Enlisted Groundcrew * 0.003 Malignancy Basal Cell Carcinoma (Verified Interval) Systemic Cancer (Verified plus Suspected, Interval) Basal Cell Carcinoma (Verified plus Suspected, Lifetime) Systemic Cancers (Verified, Lifetime) Systemic Cancer (Verified plus Suspected, Lifetime) Enlisted Flyer * 0.019 Enlisted Flyer * Neurology Pin Prick Impaired (Diabetic Class) Psychology Paranoia Schizophrenia Social Introversion Validity Total CMI Born Before 1942 High School Combat Index—Low Black High School * * * 0.027 0.033 002 .0 0.038 <0.001 Gastrointestinal SCOT Alkaline Phosphatase Direct Bilirubin Triglycerides (cont.) Triglycerides (disc.) Uroporphyrins 1-4 Drinks Exposed to Exposed to Born In or Officer BUN<14 * * * * * 0.010 <0.001 0.035 0.039 0.035 <0.001 to (-» I 0.038 Intermediate Skin Reaction to Sun Enlisted Flyer * Enlisted Flyer * per Day Ind. Chems. Ind. Chems. Before 1922 0.042 0.019 0.004 0.021 �Summary of Significant Covariate Strata (or Covariate Level Difference) Found ¥ithin Significant Two- and Three-Factor Group-by-Covariate Interactions by Clinical Chapter and Dependent Variable (Group Direction and p-Value) Clinical Chapter Dependent Variable Covariate Stratum RH>C* Dermatology Dermatology Index Pre-SEA Acne: Cardiovascular Systolic Blood Pressure ECG (Overall) ECG (Arrhythmia) Posterior Pulses (Manual) Leg Pulses (Manual) Peripheral Pulses (Manual) Black/53 Yrs Old 0 Pack-years 7 Pack-years/10% Body Fat Enlisted Flyer Officer/21% Body Fat Officer Hematology WBC Nonblack/30 Pack-years/ 35 Yrs Old Black/Officer/35 Yrs Old Black/EFL/35 Yrs Old Nonblack/30 Pack-years and 1 pack/day Black/30 Pack-years and 1 pack/day WBC WBC PLT PLT Renal BUN Urine Specific Gravity Endocrinology Urinary Protein Urinary WBC 1 vs. 0 Normal (Diabetic Class) Nonblack/Born In or After 1942 Black Nonblack/Enlisted Groundcrew Testosterone Testosterone Differential Cortisol <10% Body Fat 10-25% Body Fat Black/Born In or After 1942 ORE p-Value �TABLE 21-2. .(continued) Summary of Significant Covariate Strata (or Covariate Level Difference) Found Within Significant Two- and Three-Factor Group-by-Covariate Interactions by Clinical Chapter and Dependent Variable (Group Direction and p-Value) Clinical Chapter Dependent Variable Covariate Stratum Immunology Total T Cells B Cells Monocytes Black Nonblack/0 Pack-years Enlisted Groundcrew/ 4 Drinks/Day Pulmonary Pleurisy Tuberculosis X-ray RH>C* 1-10 Pack-years 1-10 Pack-years 0 Pack-years Total Interactions: 43 *Relative risk greater than one, or Ranch Hand mean greater than Comparison mean. ORH p-Value �CLASSICAL COVARIATES Many of the dependent variables in this report are known to be significantly affected by risk factors also measured in this study. The use of these covariates in the adjusted analyses has served to clarify Ranch HandComparison group differences in the presence of significant covariate group differences. Such adjustments, whether by a single covariate, multiple covariates, or covariate interactions, have given results on group differences generally quite similar to the unadjusted analyses both in terms of relative risk and statistical signficance. In fact, in only one instance in this report has an unadjusted result of pX).10 changed to a value of p<0.05 in the adjusted analysis. The covariates used in this study were not effect modifiers (which may be synergistic with exposure and also be equally distributed between groups). Consistent effects were observed for almost all of the classical covariates of age, race, occupation, education, alcohol, smoking, percent body fat, and glucose tolerance. In only a few instances were unexpected effects noted, e.g., personality type, wine consumption, and a few smoking and alcohol "inversions." The overall covariate effects observed in this study indeed reflect the mainstream of results found in well-conducted epidemiologic studies, and lend credence to the validity of the clinical endpoints and covariate values in this report. MULTIPLE COMPARISONS As noted in Chapter 7, Statistical Methods, the problem of multiple comparisons is complex and not easily adjudicated because of the total number of statistical tests, the number of tests performed on each dependent variable, and the biologic relatedness of many of the variables. A conscious effort has been made to expand inferential interest to borderline group associations (0.05<p<0.10) thereby increasing the probability of the acceptance of a false association. Each chapter summary has carefully flagged all borderline associations to provide expanded summary statements for possible inclusion in deriving final conclusions. Additional confidence in the final acceptance or rejection of an overall herbicide effect would be warranted if the majority of borderline associations were in the same consistent direction as the significant associations. Multiple analyses on the same variable have been conducted in this report. Continuous and categorical data have been subjected to both unadjusted and adjusted analyses, and multiple adjusted analyses were sometimes conducted with different covariates or slightly different covariate sets. The question arises as to which results best reflect the truth when different results are found. In general, the following approach has been followed: the statistical significance of both continuous and categorical analyses is convincing, while significance for only the continuous analysis must be viewed in terms of the biologic relevance of the mean shift detected. Overall,-the multiple comparison issue is due to repeated hypothesis testing for group, exposure, and interaction strata differences. The calculation of expected numbers of significant associations for these tests is difficult (if not impossible) because of the relatedness of the dependent variables, the relatedness of the covariates, and the often difficult analytic decisions that arise in a "step-down, best model" strategy. Thus, 21-9 �the final assessment of whether the frequency of significant associations does not meet, or exceeds expectation, must remain an interpretive judgment of each reader. CAUSALITY The AFHS is an inferential assessment of observed group differences. The inference of herbicide causality will be determined by a balanced judgment of the following factors; biological plausibility, consistency, specificity, coherence, time relationships, and strength of association. Except for aspects of association strength, most of these causality factors have been discussed in the preceding sections of this chapter. Nearly every statistically significant group difference in this report has only been of moderate to weak strength. Highly significant p-values (p<0.001) were not found for main group associations, but were observed for covariate tests. A few strata in the group interactions were highly significant. Most of .the statistically significant estimated relative risks were below the value of 2.0 (a traditional boundary of interest in epidemiology). The few relative risks above 2.0 generally had very wide confidence intervals due to low proportions of detected abnormalities. Weakly significant associations, in particular, are cause to reassess the element of chance and the possible presence of other causality factors before a final conclusion of cause and effect is determined. 21-10 �CHAPTER 22 CONCLUSIONS INTRODUCTION This chapter summarizes the conclusions drawn from the statistical analyses that have been conducted on the Air Force Health Study data base. The followup study, which began in 1985, was the logical extension of the 1982 Baseline study, building upon the strengths of the Baseline study and utilizing the data collected at both the Baseline and the followup. The high level of Government support and outstanding participation of the study subjects that characterized the Baseline study were maintained through this first followup. STUDY PERFORMANCE ASPECTS Of the living Baseline study participants, 99.2 percent were located and asked to participate in the followup. Participation in the followup physical examination and questionnaire was very high. Of the fully compliant Baseline participants, 971 of the 1,045 Ranch Hands (92.9%) and 1,139 of the 1,224 Comparisons (93,1%) participated in the followup. Thus, there was no group difference in compliance of the Baseline participants at the followup. Overall, the 2,309 participants in the followup (1,016 Ranch Hands and 1,293 Comparisons) represented a loss of 159 individuals and a gain of 199 since Baseline. One percent of the fully compliant Baseline population died between 1982 and the 1985 followup examination. The bias/compliance analyses suggested that there had been no change between Baseline and the followup in the way replacements volunteered for entry into the study, and that no additional bias had been introduced at the followup due to scheduling differences. Although replacements were not health-matched at Baseline as they were at the followup, they were similar to refusals with respect to reported health, medication use, and income level. The results supported the use of the total Comparison group in the main analyses presented in this report. POPULATION CHARACTERISTICS Overall, the Ranch Hands and Comparisons reported similar social and behavioral characteristics. No significant differences were found in age, educational background, religious preference, current military status, and income level. Significantly more Ranch Hands smoked cigarettes at the time of the followup examination than did Comparisons, but there was no significant difference between groups on past cigarette, cigar, and pipe use and on recent and past use of marijuana. A much higher percentage of participants 22-1 �reported past marijuana use at the followup than at Baseline. This difference was most likely due to a greater level of confidentiality afforded by the questionnaire technique. Risk taking behavior, assessed by questions on potentially dangerous recreational activities, revealed borderline significance. Slightly more Comparisons were scuba divers and more Ranch Hands raced motor vehicles. The difference in scuba diving was also significant at Baseline. Patterns of Results Both the chapter conclusions and the final conclusions of this report have been predicated upon concepts of consistency, specificity, coherence, strength, and plausibility as they apply to the interpretation of group differences. In particular, careful consideration has been given to a variety of data and patterns of results that have emerged from the clinical evaluations. Specifically, there were few differences in the proportions of abnormalities between groups; the positive associations have not aggregated in the clinical areas of prime dioxin concern, nor have they been of serious clinical importance; the unadjusted results have been remarkably concordant with the adjusted results, both in terms of relative risk and p value; the analyses using the Original Comparison set have largely mirrored the results found with the total Comparison group; many of the group differences noted at Baseline have disappeared at the followup examination, and only a few new associations have emerged; almost all of the covariates have acted as expected in the adjusted analyses; and the exposure index analyses and the group-by-covariate interactions have not demonstrated biological patterns of concern and appeared to be more likely due to chance than not. Due to the acknowledged limitations of the exposure index used in this report (and considering the potential use of dioxin body burden levels at the next followup), dose-response relationships have not been emphasized in reaching final conclusions. The overall pattern of these findings indicates that this followup study cannot be viewed as alarming from the traditional perspectives of clinical medicine or epidemiology. This study, in fact, demonstrates similarity in current health status between the Ranch Hand and Comparison groups. CLINICAL ASPECTS General Health The nonspecific assessment of general health showed relatively close similarity between the two groups. ' Ranch Hands rated their health as fair or poor more frequently, but this difference was found only in the enlisted groundcrew and not in the officers nor enlisted flyers. The perception of health in both groups had improved since Baseline. Physician-rated appearance of relative age was not found to be significantly different at the followup in contrast to the Baseline finding that a higher percent of Ranch Hands than Comparisons looked younger than their stated age. The categorical analysis of sedimentation rate showed that the Ranch Hands had more abnormalities than the Comparisons. These results were not supported by the continuous analysis of mean sedimentation rates and were opposite to the 22-2 �Baseline results, which showed that younger Comparisons had elevated sedimentation rates. The categorical analysis of percent body fat showed no significant differences between the two groups, which was consistent with Baseline. However, the continuous analysis found that the Ranch Hands had a significantly lower mean percent body fat using age, race, and occupation as covariates. The detailed exposure analyses revealed no consistent exposure effects, and this result was consistent with the Baseline analysis. No longitudinal difference was found on perception of health. A significant group difference was found over time for the longitudinal analysis of sedimentation rate due to the change in the findings between the two examinations, possibly related to a change in laboratory methodology. Malignancy Skin and systemic cancers, both suspected and verified by medical records, showed no significant group differences for the Baseline-followup interval (1982-1985). However, for all neoplasms combined (malignant, benign, and uncertain), a borderline significant excess in the Ranch Hand group was noted in an unadjusted analysis. The analyses of interval cancers revealed group interactions for verified and verified plus suspected basal cell carcinoma and verified plus suspected systemic cancers. Nonsignificant findings were observed for verified and verified plus suspected sun exposurerelated cancers. Verified systemic cancers did not differ significantly between groups. The analyses of lifetime cancer found significant results for verified basal cell carcinoma and verified sun exposure-related skin cancers. Group interactions were noted for systemic cancer categories and for verified plus suspected basal cell carcinoma. The higher rate of basal cell carcinoma in the Ranch Hands versus the Comparisons found at Baseline was nonsignificant for the followup interval, but due to the effect of the larger number of Baseline cases and the significant confounding of average residential latitude, the adjusted analysis of lifetime basal cell carcinoma emerged as statistically significant. There were several disparities in the distribution of testicular, colon, and smoking-related tumors in the groups. Further, one case of soft tissue sarcoma and one possible lymphoma (both in Ranch Hands) were diagnosed in the interval, balancing the two similar cases found in the Comparison group at Baseline. Considering that the systemic cancer curves are in their early stages for both groups, with perhaps insufficient latency, the cancer results of the followup examination should not be viewed as disturbing, but as cause for continued monitoring. Neurological Assessment None of the 27 neurological variables demonstrated a significant group difference, although several variables had relative risks which were greater than one. There was no group difference in reported neurological illnesses for the interval or for a lifetime history. Of the cranial nerve variables, speech and tongue position were marginally significant, with the Ranch Hands at a slight detriment. The analyses of peripheral nerve function showed no significant differences between the Ranch Hands and the Comparisons. In the analysis of central nervous system function, hand tremor was found to be of borderline significance, with the Ranch Hands faring slightly worse than the Comparisons. A borderline significant group interaction (Ranch Hand hand tremor by insecticide exposure) may have had biological and operational 22-3 �significance. Overall, substantially fewer neurological abnormalities were detected at the followup examination than at the Baseline examination. The exposure analyses showed only occasional statistically significant results, although no consistent pattern with increasing exposure was evident. In the longitudinal analysis of the Babinski reflex, a significant change over time was observed. This was due to a nonsignificant finding in the Ranch Hands at the followup, which differed from the significant adverse finding at Baseline. The covariates of age, alcohol history, and diabetes showed classical effects with many of the neurological measurements. Overall, the followup examination results were quite similar to the Baseline findings. Psychological Assessment The reported and verified data on lifetime psychological illnesses showed no significant differences between groups. Distributional tests of the 14 Minnesota Multiphasic Personality Inventory (MMPI) scales, stratified by occupation, revealed that only 2 of the 42 results approached significance. For the total Cornell Medical Index (CMI), separate distributional tests were conducted with stratification by age, race, occupation, education, and current drinking status; a significant difference was found for one statum of each of the covariates. In all cases, the mean of the Ranch Hand distribution was greater than the mean of the Comparisons. The analysis of the 14 MMPI scales showed that there was a significant difference between the two groups for denial and masculinity/femininity, with more abnormalities in the Comparisons than the Ranch Hands. The results of the analyses for hysteria were of borderline significance, with more abnormalities in the Ranch Hands. There were more abnormalities in the Ranch Hands than the Comparisons for social introversion, which was of borderline significance. Differences in the total CMI and A-H area subscore were found to be significant, with more abnormalities in the Ranch Hands. There was no significant difference between the two groups on the Halstead-Reitan Battery impairment index, a measure of the functional integrity of the CNS. The exposure index analyses did not reveal any pattern consistent with a dose-response relationship. As expected, the effects of age, educational level, and alcoholic history showed profound effects on many of the psychological measurements. Gastrointestinal Assessment Although the followup gastrointestinal assessment disclosed more statistically significant findings than the Baseline examination, the abnormalities were distributed equally between the two groups, and there was no clinical, statistical, or exposure pattern consistent with an herbicide-related effect on health. No historical or biochemical evidence was found to suggest an increased likelihood of porphyria cutanea tarda (PCT) in the Ranch Hand group. Only sparse and nonsignificant liver disorders were reported for the interval between Baseline and followup. Also, for the lifetime history of liver disorders, there were no significant differences between groups. Further, there were no significant group differences in reported lifetime peptic ulcer disease. A review of digestive system mortality showed a relative excess in the Ranch Hands but a relative lack of malignant neoplasms. The results of the physical examination showed a borderline increase of hepatomegaly in the Ranch Hand group. There was a significantly lower mean serum glutamic-pyruvic transminase (SGPT) level, a greater mean alkaline phosphatase level, and a lower mean uroporphyrin level in the Ranch 22-4 �Hand group. The analysis of coproporphyrin was of borderline significance, with the mean of the Ranch Hands in excess of the mean of the Comparisons. No group differences were found for serum glutamic-oxaloacetic transminase (SCOT), gamma-glutamyl transpeptidase (GGTP), total and direct bilirubin, lactic dehydrogenase (LDH), cholesterol, or triglycerides. The numerous group-by-covariate interactions did not disclose any consistent subgroup patterns detrimental to the Ranch Hands. These findings were generally consistent with the results of the 1982 assessment. The longitudinal analyses for SCOT, SGPT, and GGTP showed no significant differences between results by group over time. Dermatological Evaluation No significant group differences were identified in the dermatological evaluation. None of the questionnaire data showed an increased likelihood of past chloracne, as determined by anatomic patterns of acne, and no cases were diagnosed in the physical examination. Analyses were conducted on six dermatologic disorders (comedones, acneiform lesions, acneiform scars, inclusion cysts, depigmentation, and hyperpigmentation) and on a composite variable of 16 other minor conditions (the latter not generally associated with chloracne). Exposure index analyses did not reveal consistent patterns suggestive of a dose-response relationship. The longitudinal analysis, based on a composite dermatology index, showed no significant differences between the results over time. Substantially more dermatologic abnormalities were detected at the followup examination than at the Baseline examination. In general, however, the followup results were consistent with the findings at Baseline. Cardiovascular Evaluation Overall there was general similarity in the cardiovascular health of the Ranch Hands and the Comparisons. Of the 27 cardiovascular variables, there was a. significant difference for only one, verified heart disease, with an excess in the Ranch Hand group. This finding was largely unsupported by other cardiac measurements. The cardiovascular assessment was based on reported and verified heart disease; the measurement of central cardiac function by systolic blood pressure, abnormal heart sounds, and ECG findings; and the evaluation of peripheral vascula'r function by diastolic blood pressure, funduscopic examination, presence of carotid bruits, and detailed manual and Doppler measurements of five peripheral pulses. Doppler recordings of five peripheral pulses were similar in both groups, a finding which was in marked contrast to the Baseline examination that found significant pulse deficits in the Ranch Hand group. This change was most likely due to a required 4-hour abstinence from tobacco prior to the pulse measurements. Overall, the exposure analyses were unsupportive of any meaningful doseresponse relationship. The longitudinal analyses confirmed the change in pulse abnormalities in the Ranch Hand group over time, but showed no significant group change in overall ECG findings between the examinations. Hematological Evaluation The hematological evaluation found that neither group manifested an impairment of the hematopoietic system, consistent with similar findings at 22-5 �the Baseline. The evaluation was based on eight peripheral blood variables: red blood cells (RBC), white blood cells (WBC), hemoglobin (HGB), hematocrit concentration (HCT), corpuscular volume (MCV), corpuscular hemoglobin (MCH), corpuscular hemoglobin concentration (MCHC), and platelet count (PLT). Both the discrete and categorical analyses revealed no significant group differences. The covariate effects of age, race, occupation, and smoking history were highly significant for many of the variables. Two group-by-covariate interactions in the analyses of mean differences did not appear to have a meaningful interpretation. The exposure index analyses did not support any plausible dose-response relationship. The longitudinal analyses of MCV, MCH, and PLT found significant differences only for PLT between the Baseline and the followup, with the Ranch Hands exhibiting a slight decline in mean level from Baseline and the Comparisons showing an opposite change. Renal Assessment None of the six renal variables of reported kidney disease, urine protein, occult blood, urine white blood cell count, blood urea nitrogen, and urine specific gravity showed a significant difference between the two groups based on the unadjusted analyses. In the adjusted analyses of the laboratory variables, however, there were significant group-by-covariate interactions that did not yield a consistent pattern to suggest a renal detriment to either group. The finding of group equivalence for past kidney disease was in contrast to the Baseline examination, which found significantly more reported disease in the Ranch Hand group. The difference in findings is more likely due to a change in questionnaire wording than to a true change in renal health. Like the Baseline findings, the exposure index analyses showed very little evidence of a dose-response relationship. In the longitudinal analyses of blood urea nitrogen, there was no significant group difference in the change between the examinations. Endocrine Assessment In general, the endocrine health status of the Ranch Hands and the Comparisons was reasonably comparable. The examination found no significant differences between the two groups for past thyroid disease, or thyroid and testicular abnormalities determined by palpation. In the analyses of the seven laboratory values (T3 % Uptake; thyroid stimulating hormone [TSH]; testosterone; initial, second, and differential cortisol; and postprandial glucose), significant differences were found for TSH and testosterone, with higher mean levels in the Ranch Hands. These analyses were not supported by the categorical analyses. The thyroid test results were conflicting with respect to an assertion of hypothyroidism in the Ranch Hands (a possible dioxin effect). Mean levels of testosterone were significantly elevated in the Ranch Hand group as contrasted with the Comparisons in the 10-25 percent body fat category. The effects of personality score and percent body fat on the differential cortisol levels were not fully expected. Although tests of 2-hour postprandial mean values showed no significant group differences, comparable categorical tests revealed that significantly fewer Ranch Hands had impaired glucose levels, but conversely, had more (nonsignificant) diabetic levels of glucose. Analyses of the composite diabetes indicator (history plus 2-hour postprandial results) did not disclose significant group differences. The exposure index analyses suggested that the enlisted flyers in the medium exposure level were significantly different from those in the 22-6 �low exposure level for differential cortisol, postprandial glucose, and testosterone. The corresponding high to low contrasts were not significant. The longitudinal analyses were based on T^ % Uptake, TSH, and testosterone, and revealed only symmetrical and nonsignificant changes in the Ranch Hand and Comparison groups over the time interval. Immunological Evaluation Overall, there were no significant group differences or any indication of impaired immunological competence in either group based on comprehensive cell surface marker and functional stimulation studies. Six cell surface markers (total T cells, helper T cells, suppressor T cells, B cells, monocytes, HLA-DR cells, and a constructed helper/suppressor ratio variable) and three functional stimulation studies (PHA, pokeweed, and mixed lymphocyte culture) were conducted on 47 percent of the study population. No significant differences were revealed for five of these variables. In the analyses of the other five variables, there were significant group-by-covariate interactions, but no discernible pattern was identified to suggest a detriment in any subgroup of either group. Skin test assessments of delayed hypersensitivity were characterized by inter-reader variation and shifting diagnostic criteria for anergy. The skin test data were judged invalid and were not subjected to statistical testing for group differences. No consistent pattern of immunological deficits could be associated with increasing levels of herbicide exposure in the Ranch Hand group. Pulmonary Disease The pulmonary assessment did not reveal any statistically significant differences between the Ranch Hand and Comparison groups that were suggestive of an herbicide-related disease. The analyses consisted of group assessments of respiratory disease incidence, physical examination abnormalities, and the current prevalence of x-ray abnormalities. There were no significant differences between the Ranch Hands and Comparisons for history of asthma, bronchitis, pneumonia, or for six of seven clinical variables (excluding rales) determined by x-ray or auscultation. Analyses of history of pleurisy, history of tuberculosis, and rales showed significant but inconsistent groupby-covariate interactions. These findings did not indicate any patterns suggesting a different disease experience in the two groups. The exposure index analyses did not reveal any consistent pattern suggestive of an increasing dose response. CONCLUSION The results of the first followup study in 1985 have shown a, subtle but consistent narrowing of medical differences between the Ranch Hands and Comparisons since the Baseline Study in 1982. The 1985 examination results provide reassuring evidence that the current state of health of the Ranch Hand participants is unrelated to herbicide exposure in Vietnam. Continued close medical surveillance of these military populations is strongly indicated. This followup report concludes that there is not sufficient plausible or consistent scientific evidence at this time to implicate a causal relationship between herbicide exposure and adverse health in the Ranch Hand group. 22-7 �CHAPTER 23 FUTURE DIRECTIONS The scope and complexity of the AFHS has required gradual refinement and correction to meet the challenges of changing technology and scientific direction, and to ensure continued participation of all enrolled members. This chapter outlines some of the changes incorporated in the fifth-year followup examination and identifies several areas of future work expected to significantly augment the study. FIFTH-YEAR FOLLOWUP EXAMINATION Since the fifth-year followup examination was initiated prior to the full analysis of the data from the third-year examination, most modifications were founded upon quality control issues and the desire to make the clinical content of the examination more responsive to the medical needs of the participants. Clinical quality control enhancements were made to improve measurement techniques. The digit preference noted in systolic and diastolic blood pressure readings led to the use of automated blood pressure recording; all other parameters of the blood pressure readings (e.g., sitting position, three recordings, nondominant arm at heart level) were not changed. The problem in skin test reading was met by a rigorous quality control plan that included the following elements: refresher training for readers; a required reading of the four skin tests of all participants by both readers, each blind to the results of the other; a required reread of 10 percent of all tests by each of the readers, each blind to the previous reading; and a required weekly report citing numbers and proportions of participants with possible anergy, reversal of induration-erythema measurements, and untoward skin reactions or other reading problems (e.g., participant refusal). In addition, new skin test forms were developed to facilitate accurate recording and transcription; specific clinical criteria were formulated to require consultation by an allergist; and the skin test measurement criterion for possible anergy, consistent with current World Health Organization guidelines, was adopted for the clinical interpretation of all skin test readings. It is anticipated that this clinical quality control program will standardize both readings and interpretations, and will produce a uniformly superior data set. EXPOSURE INDEX REFINEMENTS Since the development of the Study Protocol and the analysis of the 1982 Baseline data, there has been concern among some scientists and the principal 23-1 �investigators over the accuracy and validity of the exposure estimates. It is unclear whether statistically significant differences in some variables between the Ranch Hand and Comparison groups, unsupported by dose-response estimates, have been due to chance, or whether true differences are obscured by an inadequate exposure index or group misclassification. In mid-1986, strong correlations between dioxin levels in fat tissue and serum were demonstrated by the CDC and other institutions. Because of these results, the Air Force is currently engaged in a collaborative study with CDC to determine whether serum dioxin levels vary significantly in the Ranch Hand population. Approximately 200 AFHS volunteers have supplied a pint of blood to be analyzed for dioxin at the CDC laboratories. If clear and meaningful exposure findings are evident from this study, several additional studies are feasible: testing can be expanded to the entire study population and a meaningful exposure index based on total current TCDD body burden may be developed; and by means of archived AFHS serum samples from the Baseline study, it may be possible to calculate a reasonably precise half-life of TCDD in humans. These expanded studies will allow the estimation of body burdens of TCDD at the time of departure from SEA (assuming the absence of intervening vocational and recreational exposures). If, in fact, these potential studies become reality within the next 2 years, the fifth-year followup study data will be statistically analyzed using a more appropriate exposure index. In anticipation of this advance, the AFHS is currently collecting 280 to 350 ml of blood from all volunteers attending the fifth-year followup study. ADDITIONAL ANALYSES AND STUDIES As in the 1984 Baseline Report, not all of the measured dependent variables were subjected to statistical analysis (e.g., prothrombin, leutinizing hormone, follicle stimulating hormone), largely because they were not within the bounds of the Air Force-prescribed analyses. Exploration of many of the unanalyzed variables is contemplated as time and resources permit. Similarly, many analytic opportunities to define possible symptom-clinical sign clusters or syndromes by multivariate analysis of variance techniques were passed over due to time and charter. Particularly challenging as an area of future work may be the changing relationships of some immunological variables over time and the biological impact of these changes on the induction of diseases such as cancer. Likewise, future efforts to define shifting cardiovascular disease patterns are a logical extension of the rich longitudinal data base of the AFHS. Such efforts await future analysis and publication. The assessment of possible selection and participation bias has been addressed in a comprehensive manner in this report (see Chapter 5). The analyses and discussion suggest that statistical use of the total Comparison group (versus the Original Comparison group) is justified in this report, and that the impact of selection and participation biases have been minimal. As the followup studies continue, it is anticipated that a wealth of data on 23-2 �compliance-participation factors will be available for continued comprehensive bias analyses. In particular, it is hoped that more complete data will exist to examine the true differences in current health status between refusals and their replacements. As the data set grows over time, the bias analyses will become more complex and will have to deal with changing motivations of the participants to continue in this study. Such bias analyses and assessments will always be of great importance to this study as they ultimately set the bounds for an inference on herbicide causality. 23-3 � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 059 Folder The folder containing the original item. 1585 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Lathrop, George D. William H. Wolfe Joel E. Michalek Judson C. Miner Michael R. Peterson Stella G. Machado Theodore G. Karrison William D. Grubbs Wanda F. Thomas Date A point or period of time associated with an event in the lifecycle of the resource 1987-10-01 Title A name given to the resource Air Force Health Study: An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides, Volume I, First Followup Examination Results, January 1985-September 1987 Subject The topic of the resource Air Force Health Study dioxin morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/fe93542fe9f1ceee027678eef4d5f7bb.pdf fba26b8edca7a530691dc9d16c535ee4 PDF Text Text Item ID Number 01586 Author Lathrop, George D. Corporate Author RODOrt/ArtlClO TitlO Air Force Health Study: An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides, Volume II, First Followup Examination Results, January 1985-September 1987 Journal/Book Title Yoar Month/Day Color 1987 October n Number of limps 496 DOSOrlDton NOtOS Contract no. F41689-85-D-0010 and SAIC Project no. 2816-XX-195/254-XX. Wednesday, May 23, 2001 Page 1587 of 1608 �Air Force Health Study An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides SAIC Team Air Force Team George D. Lathrop, M.D., M.P.H., Ph.D. Stella G. Machado, Ph.D. Theodore G. Karrison, Ph.D. William D. Grubbs, Ph.D. Wanda F. Thomas, M.S. COL William H. Wolfe, M.D., M.P.H. Joel E. Michalek, Ph.D. LTC Judson C. Miner, D.V.M., M.P.H. LTC Michael R. Peterson, D.V.M., M.P.H., Dr.P.H. Project Manager: Program Manager: W.F. Thomas SCIENCE APPLICATIONS INTERNATIONAL CORPORATION 8400 Westpatk Drive McLean, Virginia 22102 EPIDEMIOLOGY DIVISION US AF School of Aerospace Medicine Human Systems Division (AFSC) Brooks Air Force Base, Texas 78235 October 1987 VOLUME II First Followup Examination Results January 1985 to September 1987 Contract Number F41689-85-D-0010 SAIC Project Number: 2-816-XX-195/254-XX (Distribution Unlimited) R.W. Ogershok �APPENDIX A Advisory Committee on Special Studies Relating to the Possible Long-Term Health Effects of Phenoxy Herbicides and Contaminants �Advisory Committee on Special Studies Relating to the Possible Long-Term Health Effects of Phenoxy Herbicides and Contaminants Committee Members Robert W. Miller, M.D., M.P.H., Dr. P.H. Chairman Chief, Clinical Epidemiology Branch 8C41 Landow Building National Cancer Institute National Institutes of Health Bethesda, Maryland 20892 Tel: (301) 496-5785 Charles C. Brown, Ph.D. Division of Cancer Prevention and Control Blair Building 3A07A National Cancer Institute National Institutes of Health Bethesda, Maryland 20892 Tel: (301) 427-8720 Jan M. Friedman, M.D., Ph.D. Director, Division of Clinical Genetics University of Texas Health Science Center 5323 Harry Hines Blvd. Dallas, Texas 75235 Tel: (214) 688-2143 Julianne Byrne, Ph.D. Executive Secretary Clinical Epidemiology Branch 8C41 Landow Building National Cancer Institute National Institutes of Health Bethesda, Maryland 20892 Tel: (301) 496-4947 Kathleen Kreiss, M.D. National Jewish Hospital and Research Center 3800 E. Coifax Avenue Denver, Colorado 80206 Tel: (303) 398-1525 Leonard T. Kurland, M.D., Dr. P.H. Professor and Chairman Dept. of Medical Statistics and Epidemiology Mayo Clinic 200 First Street, S.W. Rochester, Minnesota 55905 Tel: (507) 284-3216 George W. Comstock, M.D., M.P.H., Dr. P.H. Professor of Epidemiology Johns Hopkins University Johns Hopkins Research Center Box 2067 Hagerstown, Maryland 21740 Tel: (301) 791-3230 A-l �Richard R. Monson, M.D., Sc.D. Department of Epidemiology Harvard School of Public Health 677 Huntington Avenue Boston, Massachusetts 02115 Tel: (617) 732-1050 Herbert Pardes, M.D. Professor and Chairman Department of Psychiatry N.Y. State Psychiatric Institute 722 West 168th Street, Room 1411 New York, New York 10032 Tel: (212) 960-2500 Norton Nelson, Ph.D. Department of Environmental Medicine New York University School of Medicine New York, New York 10016 Tel: (914) 351-2566 Craig T. Ramey, Ph.D. Director of Research Frank Porter Graham Child Development Center University of North Carolina 54 Bypass Chapel Hill, North Carolina 27514 Tel: (919) 966-4121 A-2 �APPENDIX B Questionnaire Methodology �TABLE B-l. Elements of the Interval Questionnaires Type Method Elements Participant Interval Questionnaire In Person Demographic, educational, occupational, toxic exposures, reproductive experience, medical Spouse* Interval Questionnaire By Mail Comprehensive reproductive history Baseline Participant Questionnaire In Person Demographic, educational, occupational, medical, compliance, toxic exposures, and reproductive experience Baseline Spouse* Questionnaire In Person Comprehensive reproductive history Telephone Survey By Telephone Self-perception of health: current medications, severity of recent illness, absenteeism, income level *Present, former, or partner. B-l �B-2 �Participant Interval Questionnaire BKUIH UEC:K 01 DECKS OI-O2 442BSec-1 TIME BEGAN SECTION 1i AM | 1O-13X PM 3.INTERVIEHERI INTRODUCTION This part of the study asks about the health of current and foraer Air Force MeBbera and their faBillea. At various points in the questionnaire, we w i l l be using the terB "biological" to describe faally relationships. For exaaple, we Bight ask about your 'biological* children. When we uae this terB, we do not Bean your step-children or step-parent* or people related to you through adoption. He Bean people related to you by blood. You Bay refuae to answer any question you chooee. However, we and the Air Force aak that you anawer as Bany of the queatlona aa you can, so the results will accurately and fully tell your atory. I'd alao like to eBphaeise that we need aa accurate a picture aa you can reBBBber. So when we ask you about the date* of event* In your life, please think carefully. Hy records Indicate that your data of birth la (DATE or BIRTH. FROM mroRMATION t, ITEM I). la that correct? Ho (CORRECT 1NIURIIATIOR 2. B. C. Was It In 1981 or 1982? (RECORD ON INFORMATION SHEET. Naa It in the Spring, Busier, rail or Winter? (INTERVIEW! IF SPRING, RECORD ON INFO SHEET AS MARCH IF SUMMER, RECORD ON INFO SHEET AS JUNE IF FALL, RECORD 'ON INFO SHEET AS SEPTEMBER IF WINTER, RECORD ON INFO SHEET AS DECEMBER COHTIMUB TO FIND DOT FROM TBB MISSING nrroRiiAnoH FROM TMB INFORMATION INtUDMnlKJBj SU&I* I5-I6/ O2 17-18/ .OS 21-22/ 1b which of the following racial or ethnic groups does your biological father belong? (CODE ALL THAT APPLY) (PROBEi what others?) English/Heleh....01 SI-S2/ Engliah/Welsh....o1 Scottish 02 S3-54/ Scottish .03 5S-S6/ 19-2O/ .04 Irish 16-17/ 02 18-19/ .03 20-21/ Irish O4 S7-58/ Irish O4 22-23/ OS 23- 24/ Scandinavian OS S9-6O/ Scandinavian.....OS 24-2S/ 06 25-26/ Polish O6 61-62/ Polish O6 2S-27/ Russian O7 27-2B/ Ruaaian O7 Ruasian 07 28-J9/ Other Slavic OS 29-3O/ Other Slavic O8 6S-66/ Other S l a v i c . . . . . O H 30-31/ .....09 3I-32/ Jewish 09 67-68/ Jewish O9 32-33/ French. .10 33-34/ French 10 69-7O/ French 10 34-3S/ Italian it 3S-36/ Italian it 71-72/ Italian II J6-17/ Spanish 1} 37-3B/ Spanish 12 73-74X Spanish II 38-39/ 13 7S-76/ Mexican 13 40-41/ 14 77-7B/ Greek 14 42-43/ Jewlah (RECORD ON INFORMATION SHEET) What Booth did the Interview take place? CAN'T REMEMBER EXACT MONTH ASK C) INTERVIENERi 1b which of the following racial or ethnic groups doea your biological Bother belong? (CODE ALL THAT APPLY) (PROBE• What others?) Polish ADD GO ID Q.2)...2 Do you reateBber the Bonth and year you were interviewed the last tiae for thia study? We are talking about the interview in your hoaw before you went for your physical exaB. A. Englleh/Helah....O1 C. FATHER Scandinavian Yea w B. MOTHER To which of the following racial or ethnic qroupa do you belong? (CODE ALL THAT APPLY) (PROBEi What others?) Scottish First 1 have a few background queationa to aak you* 1. ASK A POR RESPONDENT, THEN ASK B AND C FOR PARENTS A. RESPONDENT IF R Mexican 13 39-40/ Mexican Greek 14 41-42/ Greek American Indian..is 43-44X taerlcan Asian 16 4S-46/ «»lan .17 47-4S/ African African Other (SPECIFY) 63-64/ Indian..15 79-8OX BEGIN DECK O2 16 10-11/ 17 12-1 3/ Other (SPECIFY) .18 49-SO/ American Indian..IS 44-4S/ Asian 16 46-47/ African 17 4B-49/ ..IB SO-S1/ Other (SPECIFY) ..18 14--15/ �DECK 03 URCKS 02-03 SECTION 2l 1. 2. EDUCATION My records show that when you were lent interviewed you had received a (DEGREE LAST OBTAINED FROM ITEM 2. IMFOMUTIOD SHBKT). Have you received any additional regular school certificates, diplomas or degrees since that Lime, that is, since (DATE OF LAST INTERVIEW)? Since (DATE OF LAST INTERVIEW) have you participated in any civilian job training program (other than the for ml schooling that we discussed), that prepared you for a Major change in your occupation? Tes (ASK No Yes (ASK A AND B) A) 1 (SKIP TO Q.I) 2 I 52/ Ho (SKIP TO Q.2) 2 1ST PROGRAM INTERVIEHERt 2ND PROGRAM 3RD PROGRAM FOR EACH DEGREE CODED IN A, ASK B. B. A. What certificates, diplomas, and/or degrees did you qet? (CODE ALL THAT APPLT) High school diploMa For what kind of work was your first civilian training program preparing you? In what year did you receive (DEGREE IN A. RECORD TEAR O1 SJ-S4/ "LJLJ E. for what kind of work was your second civilian training prograsi preparing you? For what kind of work waa your third civilian training prograst preparing you? 55-5C/ Tear High school equivalency diploam........02 57-SS/ Associate of Arts (A.A.I. 61-6J/ .01 "LJLJ Tear "LJLJ 59.-60/ 63-64/ IJ-IS/ 03 *- B. Bachelor of Arts (B.A.) or Bachelor of Science (B.S.I O4 6S-M/ "LJLJ Tear 67-C8/ Masters (H.A. or M.S.) OS 69-70/ «L_LJ Tear 7S-76/ In what Month and year did you start this training? 71-72/ "LJLJ Tear 25-I7/ Doctorate (Ph.D., M.D., M.O., Sc.D.)..O6 T3-74/ Month «»L_LJ Month K. L_L_L_I_J Month Tear Month 20- 2V In what Month and year did you start this training? I_J__I_J I Month Tear 40-4V Tear 28-31/ In what Month and year did you complete this training? 79-BO/ Tear 77-7S/ BEGIN DECK OJ 08 IO-I1/ J. l_L_J_LJ Tear I8-I9/ year did you complete this training? . .07 No certificate, dlploM, or degree (volunteered) In what month and year did you a tart this training? C. In what Month and Others (SPECIFY) 37-19/ In what Month and year did you cosplete this training? I__I_J_J Tear J2-JV Month I Tear 44-47/ CURRENTLT IN TRAINING...1 CURRENTLT IN TRAINING...1 CURRENTLT IN TRAINING...1 O. H. L. Have you participated In any other civilian job training prograM that prepared you for a Major change In your occupation? Tes..(ASK r) I No.(SKIP TO 0.1)..2 24/| Have you participated In any other civilian job training prograM that prepared you for a Major change in your occupation? Yes..(ASK I) 1 Ho.(SKIP TO p. 31..2 36/ Have you participated in any other civilian job training prograM that prepared you for a Major change In your occupation? Tes(GO TO HEH OUEX).t Mo..(GO TO p.J) 2 48, �5. 3. Hava you served in the military (at a l l ) since (DATE OF LAST INTER V I E W ) Tea No 4. I (SKIP TO SECTION 3> 49/ BHCIN OFXK O4 (Continual! A. 2 3RD PROGRAM 2NU PROGRAM 1ST PROGRAM For what kind of wi»rk was your third M i l i t a r y training prograM preparing you? For what kind of work w«« your second M i l i t a r y training prograM preparing you? For what kind of work was vmir f i r s t M i l i t a r y training prograM preparing you? Are you currently serving in the Military? so/ If.*.., No*••• 5. Now, let'a talk about any military and specialised training program* that prepared you for a Major chanqe in your occupation. Since (DATE OF LAST INTERVIEW), (and besides the forMl schooling, and job training program you've told M shoot), have you participated in any Military technical or specialized training prograa* ttiat prepared you for a Major chanqe in your carssr? Tea....(ASK A-EI 1 B. Hhat is the AFSC for that Job? G. Hhat Is the AFSC for that Job? L. SI/ C. CO In what Month and year did you atart this training? I I J I Hontfi Year H. In what Month and year did you start this training? I I 57-60/ l>. I. In what Month and year did you coMptrt** this train!nq? L.LJ _J _I_J I 4/ 10- LJL_UL_LJ 66-70Y _ _ _ _ 52-S6/ Ho.(SKIP TO SUCTION 31..2 H* In what, month and y«?ar. did you start thin L r a t n i n i ? I__I__J_J Month Yenr 7I-74/ iS-in/ N. In what Month und year did you conplPte thin training? In whal -aonlh yf*<ir did you cowpl^l" this LJ_I__I__I Honlh Y*-ar UL_LJ_I Month Year l_LJLJLJ Month Year What is the AFSC for that Jol>? SI-64/ ?S-7e/ CURRENTLY IN TRAINING..I CURRENTLY IN TRAINING..I E. Hive you •participated In any other M i l i t a r y Job training prograM that prepared you for a Major change in your occupation? Y e a . . I ASK F ) . . I 6S/ Have you participated in any other M i l i t a r y Jon training prograM that prepared you for a Major change in your occupation? Yes..(ASK K ) . . 1 ?9/ CUItRraiTLY IN TRAIHIII'i. .» O. Hav-s you participated in any oti'*r M t U l . i r y Job tmlninq prtxirAM that prepared you for a major chan<i* In your occupation? Yes<NEH 24-27/B �DECKS 04-05 RUCTION 3: KHPI.OYHFJIT 1. (Continued) E. How I have now* questions about working. Please tell me ahoiiL any jobs you've had that lasted for 3 Months or longer sine*. (DATE OF LAST INTERVIEW). If you had nore than one Job at the SAM« tine* please tell Me about each joh separately. Count chanqes of jobs for the saMe employer as separate jobs. Do not include jobs in the Military. Let's start with the s*>st recent regular job you've had and work back in tiMe to (DATE OF LAST INTERVIEW), 1. Please look at this card and tell Me which niuber best describes the kind of industry you (work/worked) in? ( W R I T E IN NUMBER) ENTER NUMBER, In what Month and year did you start your Most recent job that lasted 3 Months or longer? P. I Year A. I I Month HO CIVILIAN JOBS (SKIP TO SECTION 4) I J2/ CURRENT JOB What (is/was) the n«Me of your employer? G. | 65-66/ | In what Month and year did this job end or is Job* 2B-3I/ Month | I this your current I 67-70/ Tear (SKIP TO O.2> » 1 1V What was the Main reason you stooped working on your job? 33-S7/ T B. 7Z-73/ (Is/Was) this a full-ti»e or part-tisn job? i. Ful 1-tiMe Pa r t- ttMe C. I 2 S8/ What kind of business (is/was) that—what (do/did) they Make or do there? , ^ FOR EACH SUBSTANCE CODED IN p. 2, ASK A. Mhils working at (EMPLOYER) (do/did) you COMB in contact with any of the substances on this card? By contact I Mean that you inhaled, tasted, had skin contact with these fibers snd chemicals or were exposed to ionising or nuclear radiation. CODE ALL THAT APPLY A. In general, how nany days a Month did you co«e in contact with (SUBSTANCE)? Less than once a Month 59-61/ Asbestos 01 I 74-7S/ | BEGIN DECK OS Ionizing or nuclear radiation O2 10-11/ I D. What (do/did) you actually do on the job—what (are/were) sane of your Main duties? 62-6V 1 Industrial cheoicals | Insecticides or pesticides Degreaslng che*lcel Defoliants or herbicides O3 14-1S/ I O4 OS O« None of the above (SKIP TO O.S)...07 95 7S-77/ I days 95 12-13/ | days 95 16-I7/ »8-1»/ | 22-21/ | 26-J7/ | 30-3I/ day | | | days | days | | days 95 2O-2V 95 J4-25/ 95 I8-29/ �III'.Ill HI.' I 3. All of the tUe C. 1 Hhat kind of business was that—what did they «ake or do there? What did you actually do on the Job—what were some of your Bain duties? 2 Never (Continued) D. t Same of the time 4. 6. Hlille you "ere on that Job, how often (do/did) you wash to renove the (SUBSTANCES) or use protective gear — would you say all of the tlxe, ao«ie of the tine, or never? (SKIP TO p.5) 12/ Which of the following (do/did) you use on that job? (CODE ALL THAT APPLI) 13-IS/ » Air filter Ol 3J-34/ Goggles 02 3S-36/ Face .hleld 0) 37-J8/ Special clothing 04 39-4O/ Waahlng facilities...... OS 4I-42/ Self-contained or supplied air breathing apparatus 0« 4J-44/ Moo. 07 4S-4«/ E. HAND CARD E HAND ENTER NUMBER I | F. 1 I. (SKIP TO Q.21) L_J Hontli 16-1 47/ L_J I8-21/ Year CURRENT JOBi....(SUP TO 0.7) OOO1 2 G* 6. | In what Bonth and year did this Job end? I Did you have another job before the Job with CHAME IN O. IA) but. since (DATE or LAST INTERVIEW) that lasted ) mntha or longer? No | CARD C 00 T, Please look at this card and tell He which number beat describes the kind of Industry you worked In? (WRITE IN NUMBER) Hhat was the nain reaaon you stopped working on your Job? In what snath and year did you start that Job? 22-2V I A. I I I Honth Tear I 4B-S./ Hhat was the naM of your employer? _______ B. Has this a full-tin* or part-tin* Job? rull-tlM (Continued) I Part-tlM 6. S2-76/ .2 77/ III, �I>H( K (H, UECKS 06-07 7. FOR EACH SUBSTANCE While workinq al (EMPLOYER) did you come in contact with any of th<* substances on this card? By contact I mean that you inhaled! tasted* had skin contact with these fibers and chemcials, or were exposed to ionlzlnq or nuclear radiation. (CODE ALL THAT APPLY) Asbestos Ol .03 Industrial chemicals A. In general how many days a month did you come in contact with (SUBSTANCE)? Less than once a month 24-2V | Ionizing or nuclear radiation....O2 IO. conni IN p. 7. ARK A. | 12-11/ | | 6S/ No II. I | day* 95 O4 36-37/ ( oegreaslng chemicals OS Defoliants or herbicide* O6 44-4S/ | J None of the above (SKIP TO p.tO).O7 8. 4O-4I/ | | | | days Some of the time. 95 46-47/ Ha* this a full-time or part-time job? rul 1- time 1 Part-time 2 Never BEGIN DECK O7 «O-34/ B. 1 66-69/ 42-41/ 95 4B-49/ All of the time L_l Year What was the name of your employer? While you were on that job, how often did you wash to remove the (SUBSTANCES) or use protective gear — would you say all of the time, some of the time, or never? CO I 39-19/ | | days 2 14-35/ A. Insecticides or pesticides I Month 30-3I/ | (SKIP TO 0.21) In what month and year did you start that job? 26-27/ |__| 2B-29/ | Did you have another job before the job with (NAME IN Q.6A) but, since (DATE OF LAST INTF.RVIEH)? SO/ 3 C* Miat kind of business was that what did they make or do there? J6-38/ 9. (SKIP TO O..1O) Which of the following did you use on that job? " ottle." ^ 'CtU*Ulr *» - «• Job-what -ere some of your main ' (CODE ALL THAT APPLY) Air filter 01 S1-S2/ Goggles O2 53-S4/ Face shield Ol S5-S6/ Special clothing O4 S7-58/ Hashing facilities OS 59-60/ Self-contained or supplied air breathing apparatus O6 6I-62/ None 07 6J-64/ 39-41/ HAND I CARD E I Please look at thia card and tell me which number best describes tn* kind of Industry you worked In? ENTER NUMBED, | II. (Continued) | | 42-43/ �II. 14. (Continued) F. Mhlch of the following did you use on that Job? In what Month and year did thla Job end? (CODE ALL THAT APPLY) I I I I O1 ID-It/ Goggles I Air f i l t e r O2 12-1 3/ 44-47/ CURRENT JOB G. HAND CARD E what was the naln reason you stopped working on your Job? 4B-49/ 12. While Hording at (EMPLOYER) did you cone In contact >.wltli any of the •ubatancea on thla card? By contact I >ean that you Inhaled, tasted, had skin contact with these fibers and cbeMlcals, or were exposed to Ionizing or nuclear radiation. CODE ALL THAT APPLY A. Asbestos 01 SO-51/ | Ionizing or nuclear radiation....O2 HAND CARD O Industrial chemicals | 54-5S/ | O3 | days | 5S-S9/ | | days | 16-I7/ O5 18-19/ O6 2O-21/ O7 22-2J/ (DATE Or LAST INTERVIEH 17 1 Yes Ho 95 52-SV 16. | days I 04, 62-63/ I I I days 95 «4-«S/ OS «6-«7/ I I I days ......95 OS I I days 95 2 Year I ZS-28/ «•-«»/ Defoliants or herbicides I__LJ Honth 95 «O-««/ Insecticide* or pesticides (SKIP TO 0.21) 24/ In what Month and year did you start that Job? 95 56-S7/ Degreastng cheMlcals 72-7J/ None of the above (SKIP TO Q.1S).O7 13. 14-IV O4 Old you have another Job before Job with (NAME IN g.HA), but since In general how Many days a Month did you COMS tn contact with (SUBSTANCE)? Leas than once a Month W vO 15. rOft EACH SUBSTANCE COOED IN O.I2, ASK A. O3 Self-contained or supplied air breathing apparatus tOf.lt TO O.12)....OOOI Face shield Hashing f a c i l i t i e s Year Special clothing None Month 7O-7I/ I A. what was the nans of your employer? 29-S3/ 74-7S/ B. Mas this a futl-tlMe or part-tlM Job? While you were on that Job, how often did you wash to remove the (SUBSTANCES) or uee protective gear — would you say all of the Han, mourn of the tls*, or never? All of the tlM 1 SOM of the tlMe..... 1 1 Part-tlMe 2 54/ 2 Never Ful 1- tiMe (SKIP TO Q.1S) 76/ C. what kind of business was that—what did they Make or do there? S5-57/ 16. (Continued) �16. (Continued) D. IB. what did you actually do on the Job—what were duties? of your Ha In While you were that Job, how often did you wash lo remove the (SUBSTANCES) or use protective qear — would you say All of the tine, some of the time, or never? All of the time I Some of the til.-..; 2 Never 24/ 3 58-60/ E. Please look at this card and toll me which number best describes the kind of Industry you worked In? (WRITE IN NUMBER) 19. ( S K I P TO Q.2O) Which of the following did you use on that Job? CODE ALL THAT APPLY Air f i l t e r ENTER NUMBER I | F. | | I I I Month I I 20. 67-6B/ i—* O 29-JO/ O4 J1-J2/ O5 JJ-34/ Of. JS-36/ None G. , What was the Mln reason you stopped working on your Job? Cd O3 Self-contained or supplied air breathing apparatus CURRENT JOBI....ISKIF TO Q.1D....OOOI 27- 28/ Mashing facilities 63-6S/ O2 Special clothing Tear 25-2V Face shield HAND I CARD E \ In uhat month and year did this Job and? 01 Goggles 61-62/ 07 37- 38/ Did you have another job before the Job (NAME IN O. ISA), but since (DATE OF LAST INTERVIEW)? . Yes 17. Hhlle working at (EMPLOYER) did you cone In contact with any of the substances on this card? By contact I aiean that you Inhaled, tasted, had skin contact with these fiber* and chemicals, or were exposed to tootling or nuclear radiation. (CODE ALL THAT APPLY) Asbestos tOK EACH SUBSTANCE CODED IN 0.17, ASK A. No 21. In general, how Many days a Month did you COM In contact with (SUBSTANCE)? Less than onth 1 During the past six Months, did Illness or injury keep you from, work, not counting work around the house? 1 No 69-7O/ days .95 71-72/ 7J-74/ days .95 79-60/ days .95 (SKIP TO SECTION 4 . . ).) 7S-76/ .95 40/ (SKIP TO SECTION 4>...2 Retired Ol )9/ 2 Yes Ionizing or nuclear radiation....Ol 12-I3/ Unemployed..(SKIP TO SECTION 4 . . ).4 22. HAND CARD D (USE HEM QUEX) Industrial chemicals Insecticide's or pesticides OJ 77-7B/ BEGIN DECK O9 .04 to-tt/ Altogether, how tuny days did Illness or Injury keep you from work during the past six months? (REFERS TO -WORKING DAYS* OHLY) 41-43/ Degreaslng chemicals.... ...OS 14-1S/ days .95 16-I7/ Defoliants or herbicides OC 18-19/ days .95 2O-21/ None of the above (SUIT TO Q.2OI.O7 22-23/ ENTER NUMBER OF DAYSi 21. What illnesses or Injuries caused you to alas work? 44/ �DECK O9 SECTION 4i 1. 3. HII.ITMtY Now I urn qoinq to ask you about aone of your experiences in the M i l i t a r y . First, which of the followinq statements best describes your asaiqnoenl durinq the vietaan Mar? (Continued) C. Followinq your (retireMent/separatlon/dlacharqe) in (DATE IN B.), did you reenter the at Bed forces? Yes. A crew Member in Vietnam who was on • flying status 1 45/ No.. Not a crew Member, hut flew one or Bore Missions in Vietnam. .2 A crew Member, but did not log flyinq tim» in VietnaM 3 4. Not a crew nearer. . . . . . . . . • • • • • • . . • • • . • * • . . . . . . . . . . • • . . . . • • • .4 2. INTERVIEWER! HAS R SERVED IN MILITARY SINCE LAST IHTERV1CM7 IN SECTION 2, COOED "»ES"?) YES (GO TO p.3) (IS O.3 I would like to ask you the na>ea of all the countries. Including the United States, you have been stationed in since (DATE OF LAST INTERVIEW) When last Interviewed you were stationed In (COUNTRY FROH mrORMATIOH ITEH 3), and your assignment began In (DATE OF ASSIGNMENT FROM ITEH 3). Is that correct? 46/ 1 Yes....(ASK • THROUGH K> 3. (SKIP TO SECTION *) No A. 47/ (ASK A THROUGH C). (SKIP TO Q 4 . . .).. Were you retired, discharged or separated? Retired t Discharged/Separated B. 48/ 2 In what aonth and year were you (retlrad/discharged/separated) fro the (BRANCH OF SERVICE FROM IHKNMKnoil SMCT ITEM 3)? 4»-52/ Month Year 2 2 Since (OATE OF LAST INTERVIEW) have you retired, been discharged or separated fro. the (BRANCH OF SERVICE FROM INFORMATION 8HEKT ITEM 3)7 Yes 1 No...(CORRECT nrOMMTIOM SHUT, THEN ASK B THROUGH K) NO s«/ �IIIOCKS 09-11) 4. HECKS 10-11 BEGIN DECK IO (Continued) 4. Since (DATE OF LAST INTERVIEW), In what country were you next ntatloned while on active duty? Include temporary duties of greater than 9O days. Since (DATE OF LAST INTERVIEW), tn what country were you next stationed while on active duty? Include temporary duties of greater than 9O days. (Continued) O. (Do/Old) your duties In (COUNTRY) Include flying* 60/ BEGIX DECK I I O. 1 Yes COUNTRY B. H. In what stonth and year did you begin and end active duty In (COUNTRY)? E. 3S-36/ IO-II/ In what month and year did you begin and end active duty in (COUNTRY)? X. How Many flight hours did you log while In {COUNTRY)? In what Month and year did you begin and end active duty In (COUNTRY)? I Yes No...(SKIP TO GI....2 55-S6/R (Do/Did) your duties In (COUNTRY) Include flying? IO/ P. No..(SKIP TO CO....2 How aiany flight hours did you log while in AA. How Many flight hours did you log while In (COUNTRY)? 11-M/ l__l__l I Honth Year I I I_JL_J Honth Year I r. 37-40/ 57-60/ END EMO L_I_J_J I L_I_J_J_J Honth Year Year 4I-44/ 6I-64/ Current Cur r n . . . . . . . et......I I C. What specific Job assignments (do/did) you have in (COUNTRY)? Can you give me the Air Force Speciality Code? (PROBEt Nhat others?) N. What specific job assignments (do/did) you have in (COUNTRY)? Can you give a* the Mr Force Speciality Code? (PROBEi What others?) V. What specific job assignments (do/did) you have in (COUNTRY)? Can you give SM the Air Force Specialty Code? (PROBEs Nhat others?) 1. I '• 1. I I I_J 6S-69/ I_I_J_J_J_J 20-I4/ I I I I I | 7S-79/ 3 . !_J_J_J_J_J 30-34/ I I I I I 4S-49/ 2. 3. 3I-34/ DDDD Hour* l_l I Yes BEGIN BEGIN Honth (Do/Did) your duties In (COUNTRY) Include flying? 30/ No...(SKIP TO RI....2 (COUNTRY)? 6I-64/ Z. I_J_J I_JL_| 50-S4/ what specific letter and numerical designation^) did each aircraft have? Nhat specific letter and nunerical designation^a) did each aircraft have? BB. What specific letter and numerical designation^) did each aircraft have? �DFCKS 11-12 4. H. (Continued) R. in your joH assignin (COUNTRY), (do/did) you come in contact with any of the substances on thin card? By contact 1 mean that you inhaled. tasted, had skin contact with these fiber a and chemicals. or were exposed to ionizing or nuclear radiation. CODE ALL THAT APPLY Ionising or nuclear radiation.... 02 66-«7/ degreasing chesiicaia.OS 58- 59/ degreaaing chesdcala.OS 72-T3/ defoliant* or herbicides 74-7S/ defoliants or O6 60-6 1/ none of the above 62-6 3/ none of the above 7«-77/ O« 20-21 / none of the above 22-2J/ asbestos | | | 50- b . less than less than ionizing or nuclear radiation ionictnq or nuclear radiation | | | 26- 27/ | I | 39- 40/ less than less than industrial chemicals | || 28- 29/ | | | S2-53/ | | | 41-42/ industrial chemicals f leas than inaectlclde* or pe*Ucide* | | I 4 3- 4V lea* than insecticides or pesticide* | degreaaing cheMlcal* decreasing chemicals | I 54-5S/ less than degress! nq che»ical* | I| 32-S3/ | | | 4S-46/ | | 56-57/ less than | I | 58- S9/ less than I. leas than less than defoliant* or herbicide* | || 34-3S/ lea* than degreasing chemical*. OS !•-!»/ defoliants or herbicides In general, how nany days a iftonth (do/did) you come in contact w i t h (SUBSTANCE)? insecticides or pesticides | | | 30-H/ less than inaectlclde* or pe*ticide*. . . ......O4 IS-17/ 70-7 I/ | | | J7-18/ ID. lea* than induatrlal cheMical*.O3 14-IS/ insecticide* or asbestos iesa than Ionizing or nuclear radiation. .. .02 12-1 I/ Insecticides or In general, how Many days a Month (do/did) you COMC in contact with (SUBSTANCE)? ionizing or nuclear radiation BEGIN DECK 12 industrial chenlcal*.O3 6B-69/ 56- 5 7/ | || 24- 2V industrial chemical* industrial chemicals. O3 54-S5/ S. less titan 10-1 1/ 64-6V ionizing or nuclear radiation. ..»O2 52-5V asbestos contact with these fibers and cheMicala, or were exposed to ionising or nuclear radiation. CODE ALL HAND CARD 0 D 50-5I/ CC. In your job aaaiqnMents while stationed in (COUNTRY), (do/did) you COMA in contact with any of the aubntancea on thla card? By contact I Mean that you inhaled. tasted, had skin THAT APPLY HAND CARD D I you coMe in contact with any of the substances on this card? By contact I Mean that you inhaled. tasted, had akin contact with theae fibera and cheMicala, or were exposed to ionicing or nuclear radiation. CODE ALL THAT APPLY HAND CARD I—' OJ In your job assignaients whi le stationed in (COUNTRY), (do/did) In general, how Many days * Month (do/did) you cove in contact with (SUBSTANCE)? defoliants or herblcidea | || 47-4B/ le*a than defoliants or herbicides | | | Did you waah to re*ove the (SUBSTANCES) or did you u*e protective clothing or gear when stationed in (COUNTRY) all the tine, BOMB of the tine, or never? T. Did you wash to reMove the (SUBSTANCES) or did you use protective clothing or gear when stationed in (COUNTRY) all the tiae. BOMS of the tine, or never? SOMC of the ti«...l SOMB of Never. (SKIP TO K ) . . 3 Never. (SKIP TO V K . J * the U.MC...2 * 60-6 l/ less than EE. Did you wash to remove the (SUBSTANCES) or did you use protective clothing or gear when stationed in (COUNTRY) All the time, some of the tine, or never? Some of the time ... 2 Never. ( S K I P TO G G > . 3 * �DECKS 12-1) OFCi442asec-5 -24- DBCK I BRGIN DECK 1 ] J. Which of the following did you use on that Job? CODE ALL THAT APPLY Air f i l l e r CARD 1 E | II. Which of the following did you use on that job? CODE ALL THAT APPLY 01 63-A4/ Air filter Ol 67-6D/ Face Shield. ...O3 14-IS/ SECTION 5: 1. Yen Face Shield. ...0) Special Special Washing facilities Hashing facilities 29- JO/ No A. 7»-72/ Self contained or supplied air breathing 05 18-19/ Self contained or supplied air breathing INTERVIEWER I A) 1 (CORRECT IHrO MEET, THEN ASK A) 40/ 2 HAS STATUS 'MARRIED" AT LAST INTERVIEW? YES Self contained or supplied air breathing (SKIP TO C) t NO OS 33-34/ B. (ASK B) 2 INTERVIEWER) SEE INFORMATION SBEBT ITEM S. WITH PARTNER' AT LAST INTERVIEW? YES W (ASK Special Hashing facilities. ....05 K. MARITAL AND FERTILITY HISTORY Now 1 would like to ask you about your personal relationships. when we lalk»rt w i t h you last, you said you were (READ MARITAL STATUS FROM INFORMATION SHEET ITEM 4). is that correct? 1O-1I/ Face Shield.... O3 FF. Which of the following did you use on that Job? CODE A1.L THAT APPLY Are there any other countries that you have been stationed In since (DATE OF LAST INTERVIEW)? Yea. (GO TO L)...» NO. (SKIP TO SECTION 5J..2 V. 77/ Are there any other countries that you have been stationed In since (DATE OP LAST INTERVIEW)? Yes. (GO TO NK..I NO. (SKIP TO SECTION S)..J HO CG. Are there any other countries that you have been stationed In since (DATE OF LAST INTERVIEW)? M/ Tea. ..(USE HEW OUEX)....l C. 4I/ HAS RESPONDENT "LIVING (ASK C> 42/ ....(SKIP TO Q What Is the current f u l l nane of the person you were living with in (DATE OF LAST INTERVIEW)? J9/ HO. (GO TO MCTION 5)., 2 FIRST NAME INTERVIEWER I D. MIDDLE NAME RECORD NAME ON INFORMATION SHEET, ITEM OS In what Month and year did you start living with (NAME)? ENTER MONTH AND YEAR |_ | MO | |_ | _ | IK 4J-46/ (GO TO 0.2) E° According to our records, you were married to (NAME AT INFORMATION SHEET ITEM 6). Is that correct? Yes .................................. . ...... 1 No. .(CORRECT INFO SHEET, THEN GO TO V - 2 K . . . 2 47/ �OFC:442BSec-5 1. DECKS 11-16 (Continued) F. 2. In what month and year did you get married to (NAHEl? ENTER MONTH MO YEAR | | | | NO | (Continued) E. | 4B-SI/ Durinq this (Barriaqe/relalionahip), (aincr the (DATK OF LAST INTERVIEW)|, 4id you ever have a problem conc*lvinq a child because of prolonged Reparation? YR Yea. >2. Have you stopped living Kith (NAME)? Tea (ASK No.. At 1 F. Ho A. (SKIP TO C) How did this (marriage/relationship) and? • HAND CARD And what is (NAME OF SPOUSE/PARTNER)'a present street address? 1 " STREET ADOS ESS Separation*••••*«••••••••••••••••••«•••••••.1 BEGIN DECK 14 10- 34/ S3/ G. And what city, state and sip coda does (SPOUSE/PARTNER) live in? Dlvorce....•••••••••••••••••••••••••••••••••2 G Death of •poiiM or partner.. (ASK f-mt 0.3) I CITT i 3S-54/ H. I I I I I STATE S5-S6/ I ZIP I And what is (NAME of SPOUSE/PARTNER) 's present telephone number? I B. I CODE S7-61/ I In what month and year did till* occur? ENTER MONTH AMD YEA*! | | | | | HO I S4-S7/ I I AREA CODE I I I I - I NUMBER I I I I 62-7I/ TK H-» Ul C. During thia (marriage/relationship), how many times war* you living apart tram (NAME) (or 3 months or mra alnca (DATE OF LAST INTERVIEW)? ENTER MUHBE* OP TIMES I Nona D. ............ |_ !• S8-S9/ (_ | Thinking of all the people you know, who would be the one person who would be most likely to know where (NAME OF SPOUSE/PARTNER) is? ENTER FULL NAME OF PERSON BELOW AND ASK J-H. BEGIN DECK IS 1O-29/ (LAST), 3O-44/ (SKIP TO F) For how Mny montha did yon »!»• »f*rt the C first/next) (MIDDLE) time? J. Flr.t tlM, | | I 60-61/ Second time. | | | 62-63X Third UMI | | | 64-6S/ Fourth time I | | I 66-6T/ What Is (PERSON'S) relationship to (NAME OF SPOUSE/PARTNER)? L_L_L_LJ_L_LJ_J_LJ_J_LJ_J K. Where does (PERSON) live? 47-71/ APT! STREET ADDRESS BEGIN DECK 16 L.LJ STATE CITT 10-29/ 30-3I/ I I IZIP I CODEI I 32-16/ �DECKS 2. (Continued) L. 4. (Continued) Hhal ia (PERSON'S) telephone nunber? A. AREA CODE What la the nil name of (thia person/your w i f e ) ? LJ_I PHONE NUMBER NO PHONE H. 16-17 ID I ..< IF (PERSON! HAS PHONE. In whoae naan la the phone Hated? (PERSON'S) name (SKIP TO H) OTHER 77-7B/ 1 (SPECIFY BELOM) (LAST) 47/ 2 DON'T KNOH.. (FIRST) B INTERVIEWER! 48-7 2/ _^__^ (LAST), (NIDDLE) RECORD FULL NAME OF (SPOUSE/PARTNER) ON INFORMATION SHEET, ITEM O7 AND RECORD IDI ABOVE. BEGIN DECK 17 What waa her full Balden nane? (FIRST! (MIDDLE) J N. INTERVIEHER> HAS F STOPPED LIVING WITH SPOUSE OR PARTNER? IIS "TES* CODED AT Q.2T) what waa her blrthdate? .,-t Ye* RECORD DATEi | I | I I | | ( HO 73/ DA I IO-29/ | | JO-35/ YR W No (SKIP TO p.lO) 2 B. 3. In what aontn and year did you (reconcile/get aarrled to/start living with) (NAHE)? Since (DATE OF LAST INTERVIEW), have you reconciled, lurried (again) or have you lived together •• • partner (or 3 Bonthe or aiore with aomone to whoa you Keren't Mrrled? Tea No (ASK A) 1 (SKIP TO Q.IO) ENTER HONTH AND YCARi HO Co t Have you •topped living with (NAHE)? How Bany tlewa did you get Harried or live aa a partner wlUi aowone (or 3 auntha or Bore el nee (DATE OF LAST INTERVIEW)? RECORD NUMBER OF TIHESi | | No 75/ D. (ASK D-P)... 40/ (SKIP TO 0.4F). Haw did Uils (Harrlage/relatlonehlp) end? Separation 4. HAND CARD Thjnklnq of (that/the tlrat) relatlonablp alnce (DATE OF LAST INTERVIEW), did you a*rry thla peraoB? Tea. Ho RECONCILED. 36-19/ YR 74/ Yea A. | | ( | | i Dt vorca 2 6 76/ Death of apouae or partner. . I ASK «-B, THEN •KIP TO O.5I...3 417 �DECKS 18-JO 4. (Continued) E. 4* In what month and year did this occur? ENTER MONTH MID YEAR. (Continued) L. | | | _| HO | 42-4S/ YR Thinking of all the people you know, either around here or elsewher*. who would be the one person who would be moat l i k e l y to know where (NAME OF SPOUSE/PARTNER) is? ENTER FULL NAME OF PERSON BELCH AND ASK H-P. 47-6V F. (LAST). During this (marriage/relationship), how riany timea were you living apart from (NAME) for 3 months or more since (DATE OF LAST INTERVIEW)? 66-8O/ (FIRST) (MIDDLE) BEGIN DECK ENTER NUMBER OF TIHESl | | | 19 46-47/ M. What Is (PERSON'S) relationship to (NAME OF SPOUSE/PARTNER)? OR 10-II/ Hone (GO TO I) OO N. G. "here does (PERSON) live? For how many months did you live apart the (first/next) time? I2-36/ First time, | | | 48-49/ Second time, | | | SO-SI/ Third time, l__l__l S2-S3/ Fourth time, I_J__J 54-SS/ STREET ADDRESS APTI CITY H. 17-56/ O. I i I I ZIP CODE S9-63/ - I. I I I I 1 In whose name la the phone listed? (SKIP TO Q.S) 1 OTHER (SPECIFY BELOW) 2 DON'T KHOM And what is IF (PERSON) HAS PHONE. (PERSON'S) name I. P. 64-73/ I PHONE NUMBER NO PHONE S6/ I (PERSON'S) telephone number? I AREA CODE L-JLJ During this (marriage/relationship), (since the (DATE OF LAST INTERVIEW)!, did you aver have a problem conceiving a child because of prolonged separation? Yes Hhat is L_LJ I STATE S7-58/ (SKIP TO O.5) 8 7V (NAME or SPOUSE/PARTNER)'* present street address? S7-80/ BEGIN DECK 2O STREET ADDRESS BEGIN DECK 18 J. L_L_I STATE CITY 10-29/ K. L_L_L_L_J_I XIP CODE 30-1I/ 32-36/ And what is (NAME at SPOUSE/FARTHER I • s present telephone number? LJLJLJ - I_J__LJ_J "-«•/ IO-34/ (LASTI, And what city, state and sip coda does (SPOUSE/FARTHER) live In? (FIRST) S. INTERVIEWER, (MIDDLE) IS THERE A SECOND RELATIONSHIP SINCE THE DATE LAST INTERVIEW? (1C NUMBER OP TIMES RECORDED Hf O.3A EODAI. TO 2 OR MORE?) YES NO (GO TO O.6) (SKIP TO Q.1O) I 2 3V �DEI'K 6. Thinking of the next relationship since (DATE or LAST INTERVIEW), did you »arry this person? Yes. 6. (Continued) E. In what month and year did this occur? 36/ ENTER MONTH AND YEAR: | No.. A. Mhat is the current full MUM | | | HO of this person? 2D-2I F. 7I-74/ | YR During this relationship, how many times were you l i v i n g apart from (NAME) for 3 months or more since (DATE or LAST INTERVIEW)? ENTER NUHRER OF TIHESt |_ 7S-76/ |_ | (LAST) 37-38/ (FIRST) INTERVIEWED (HIDOLE) None ...... (GO TO I) RECORD FULL NAME OF (SPOUSE/PARTNER) ON INFORMATION SHEET, ITEM O7 AND IDf ABOVE. G. First tlmei.........|_ | _ | I I I I I I I LJLJLJ Hhat was her blrthdate? OO For how many months did you live apart the (first/next) time? Hhat was her full Maiden nans? I .................. RECORD DATEt I I I | ) [ | HO DA | | »-»/ | 77-7B/ Second time i........|_ |_ | 79-BO/ BEGIN DECK 21 Third timei.........|_ |_ | IO-II/ S9-64/ Fourth tlmei........|_ |_ | YH I2-13/ W *—• B. oo In what month and year did you (get married to/start living with) (NAME)? EHTER HOHTH AND YEAR I | | | NO | | N. During this (marriage/relationship), (since the (DATE OF LAST INTERVIEW)), did you ever have • problem conceiving a child because of prolonged separation? 6S-68/ Yes No ..... . C. ......................... I M/ YH .................... 2 Have you stopped living with (HAME)? I. Yes , (ASK And what is (NAME OF SPOUSE/PARTNER) "a present street address? I O-P) 15-39/ No D. (SKIP TO r) 2 STREET ADDRESS ' ' ~ How did thia (marriage/relationship) end? Separation..................I HAND CARD G to/ 3, And what city, state and zip code does (SPOUSE/PARTNER) live in? |_ I_ I STATE Divorce.....................2 40-59/ Death of spouse or partner (ASK «-o, TB» SKIP TO B.71 3 1_ I_ I_ I_ I _ ZIP CODE 62-66/ �OWK 21-iJ -34- 6. DECKS 21-24 (Continued) 7. K. And what l« (NAME Or SPOUSE/PARTNER) *s present telephone IMTEKVlEWEKi r? IS THERE A THIRD RELATIONSHIP SINCE THE DATE LAST INTERVIEW? lie Hunan or TIMES UCOHDCD n O.JA BOOAI. TO I OH HORE7I LI—I—I AREA COOS LJ 67-7*/ I NUMBER IES .....(CO TO g.8» HO PHONE. L. .1 I n / HO.....(SKIP TO 0.10) Thinking of all the people you know, ulther around here or elsewhere, who would be th« on* person who would be mom*, llfcsly to know where (NAME Or SPOUSE/PARTNER) Is? ENTER fUU. NAME OT PERSON BELOW AND ASK H-P. BEGIN DECK 22 *. Thinking of the n«xt relationship since (DATE Or LAST 1MTERV1EM), did you starry this person? Yes. (LAST). Ho.. 30-44/ (MIDDLE) M. *. Hhat Is the current full oaM of this per person} What Is (PERSON'S) relationship to (HAHC OT SPOUSE/PARTNER)} 7S-76/ ..a .rn H. Cd I Mhere does (PERSON) Ural INTERVIEHERi (MIDDLE) RECX)RD rOLL HAHE Of (SPOUSE/FARTHER) ON IHFORMATIOH SHEET. ITEM O7 AMD ID I ABOVE. 47-7»/ M»t IMS her full svldsn naa*7 KPTl STREET ADDRESS BEGIN DECK 24 •COIN DECK 33 CITY Hhat was her blrthdate} IO-2»/ RECORD DATEl | | |__l_| HO O. DA I 3O-3S/ IR What Is (PERSON'S) telephone nwber? B. IAREAICODE LJ l__l_l__l - I L_L_LJ PHONE HUHBER HO PHONE P. I 32-3C/ 30-3I/ IF (PERSON) HAS PtlOHEi ENTER MONTH AND »EA»« I 1 OTHER (SPECirt MELON) I DON'T KHOf (SKIP TO Q.7) | | MO In whoee nan* Is the phone listed} (PERSON'S) MM....(SKIP TO P-7) • 47/ C. | | | J6-3S/ n Have you stopped living with (NAME)} Yes Ho 4B-72/ (LAST). In what aonths and year did you (get Married to/»tart living with). <MHE)} (ASK D-K....I (SKIP TO D . . . 2 40/ �DKCKS 24-2S -15- < Coun 11 ntied) O. B. How did this (.tarrlaae/relatloMhlp) end? I 4I/ I I I I AREA O»E I I_J I - I NUMBER I I I_J «-4«/ HO PHONE Death of spouse or partner •AM m-m. no L. •Kir to g-») 1 In what «ionth and year did title occur? ENTER HOHTH MO TEAR I LJ_LJ__I HO T. And xhal is (HAHR OF SPOIIsr./PARTNeR) *a present tvlnphone nuohnr? Divorce**......*...*.******.*2 G f. (Oantlniind) r. Separation HAND CARD DECKS «-27 Thinking of all the people you know, either around here or elnewher<s, who would be the one peraon wno would be neat likely to know where (HAHC OF SPOUSE/PARTNERS la? ENTER FULL MAKE OF PERSON BEUW AND ASK H-P. «*-«*/ 47-7I/ (LAST*, m During thle (»arrlage/relattonet>tp). how siany tinea were you living apart f t am (NAME) (or 1 Matha or «wre a luce (DATE or LAST (MIDDLE) (FIRST) INTERV1EM)? ENTER NUMBER OP TIMES. | | | 4S-47/ H. I OH Hone (GO TO Nhat t> (PERSON'S) relationship to (NAHE Or SPOUSE/PARTNER)? ft.9) I I I ! I I I I I I I I I I BEGIN DECK 26 OO H. Mhere does (PERSON) live? 00 ts} O G. IO-J4/ For hov nany souths did you live apart the ttlrat/nemtl tine? STREET ADDRESS First tlM | | | 4B-49/ Second tlMi | | | 50-5I/ I I S2-SV L_l_J *«-**/ third tie* fourth tine H. ....| During thle (•arrteae/relatlof.ehlp). etnce the (DATE or LAST INTERVIEW}, did you ever have a problen conceiving a child becauee of prolonged eeparationT Tea Ho I. I L_LJ STATE 3S-44/ O. L_L_LJLJLJ ZIP CODE SS-56/ Mhat Is (PERSON'S) telephone number? LJ_LJ AREA CODE LJL-LJ -NUMBER L_LJLJ_J PHONE HO PHONE P. 5?-tt/ IP (PERSON) HAS PHOHEi 2 S6/ In whoss naiM ia 62-7(/ I the phone lilted? (PERSON'S) »«•«....(SKIP TO 0.9) OTHER And what la (HAKE OP SPOUSE/FARTHER)*e pceaent street eddreea7 I (SPEC I FT BELOM) 2 (SKIP TO 0.9) H DON'T KNOW 72/ S7-8O/ BEGIN DECK 27 STREET ADDRESS •EGIH DECK 25 J. (LAST). 10-J4/ And vhat city, state end sip code does (SrOUSE/PARTHER) live In? L_LJ STATE C1TT IO-29/ 3O-1I/ (FIRST) SIP CODE 12-I6/ (MIDDLE) �-J7- IUX'ICS 2 J - 2 K 9. INTERVIEWER I ~ IS THERE A FOURTH RF.I.ATIONSHIP SINCE TIIE DATE LAST INTERVIEW? (IS MMBBIt Of TINES M9CORDED IH Q. 3A EQUAL TO 4 Oil HDRB?) Y E S . . . . ( G O TO HEM QUESTIONNAIRE) 1 NO II. INTERVIEWS; ASK THIS p"F.STION FOR EACH CHILD LISTED OM CRILmUM'S RECORD FORM FOR WHOM THERE IS NO DEATH DATE. What is (NAME OF 1ST CHILD/NAME OF 2ND CHILD. ET<:.»'a current »qe? RECORD OH CHILDREN'S RECORD FORM. 35/ 2 IF DECEASED SINCE LAST INTERVIEW. ASK A-C. OTHERS CO TO O. 12. A. When did (CHILD) die? RECORD FORM. RECORD DAT, MONTH. AND YEAR ON CHILORJ9TS B. What was the cauae of death? C. Where la (CHILD)'a death registered? VERIFICATION OP BIOLOGICAL CHILDREN USING caiLMors RECORD row 10. INTERVlEWERi ARE CHILDREN LISTED OH CBILDMM'S RECORD FORM? CHILD ID ' I YES (ASK A) 1 NO (ASK Bt 2 36/ I I 39-457 RECORD BELOW. CHILD IDI I BEGIN DECK 28 In what city and atate? | | 60^-617 CHILD ID, I I I io~-TT7 CAUSEi RBGISTRATIOHt A. W I'd like to read information about your (child/children) froa our last interview lo check our records. Aa of {DATE or LAST INTERVIEW), our record* show that you have had (NUMBER OP CHILDREN). . .(READ EACH CHILD'S PULL NAMJE, SEX. AND BIRTHDATE AND MOTHER'S NAME). la that correct} IF INFORMATION IS CORRECT (GOTOQ.11> 1 IF INFORMATION It INCORRECT, HAKE CORRECTIONS ON CHILDREN'S KBCORD FORM (THEN GO TO Q.I1) (CITT) IF INFORMATION IS INCORRECT. ASK FOR (CHILD/CHILDREN) *S FULL NAME, SEX, BIRTHDATE AND MOTHER'S HAHP.. RECORD BEGINNING AT LINE 01 OM CBIUHIBt'8 ....(THEN GO TO 0.11) 63-7B/ (CITI) IJ-28/ SB-59/ (STATE) 79-8O/ (STATE) 29-JO/ INTEBVlrWEHi HAS R BEEN HARRIED OR HAD A PARTNER FOR J MONTHS OR MORE SINCE (DATE OF LAST INTERVIEW)? 1CS HO Our recorda show that you had not had any children of your of (DATE OP LAST INTERview). la that correct? IF INFORMATION IS CORRECT (GO TO 0.12) 12. (CITt) (STATE) 2 3V 4Z-S7/ 1 A. Ho B. (SKIP TO SECTION 6) SI/ 2 Haa/Have (your wife/any of your partners) bee since (DATE OF LAST INTERVIEW)? lea 3B/ , (ASK A) pregnant by you (ASK B . . ).. 32/ (SKIP TO p.25). How K«ny preqnancies (has your wife/have your partners) had with you since (DATE OF LAST INTERVIEW)? ENTER NUMBER OF PREGNANCIES! (CO TO 0.11) I I I 33-J4/ �DECKS 2B-29 13. Mien did ( Ihat/lhr f i r s t , etc.) pregnancy begin? RHTER MONTH AND YEAR. A. IHTKRVIEMKR; What Month and year? |_ |_ |_ |_ | HO >H II. F.. 3S-38/ HAS R HAD MORE THAN ONE RELATIONSHIP SINCE DATE OF LAST INTER VI EX? (SEE INFORMATION SHEET. ITEMS O5, O6 AND O7) (Continued) Please look at this card and tell *e all the lumbers of the ty|>es of birth control you or (NAME) were using when she hecane pregnant, cone ALL THAT APPLY. I (GO TO C) ......... 2 Jelly, cream, suppository. 39/ Hhich (spouse/partner) had thi» pregnancy? RECORD 40-64/ (LAST) 74- 75/ 76- 7 7/ 78- 79/ O4 BBC IN DECK 29 IO-I1/ S. 6. 7. 8. 9. 10. II. Condom Diaphr Diaphr RhythM Rhyth* Hi thdr 12. HO 8. .......... Douche 4. H (ASK B) 7J-73/ 1. 2. HAND CARD Other (SPECIFY) 1 2- 1 3/ I4-I5/ 16-I7/ lfl-19/ 20-21/ 22-23/ oa 24- 25/ 12 TFIRST) (MIDDLE) 26- 27/ DON'T INTERVIEWER. RECORD ID I FROM INFORMATION SHEET, ITEM OS, O6, OR 07 6S-66/ L_J_I 00 C. N) N) How many Months did it pregnant (this tine)? RECORD HONTHS AND/OR YEARS | | HOS | AND/OH | Hain't trying | YRS | Still pregnant ..... (SKIP TO 0.25) OO 69-70/ A. «a Hera either you or (NAME or SPOUSE/PARTNER) oalnq birth control at the tlM ahe becaa* pregnant? tea...........(ASK E> COO TO Q. 14) 1 28/ live birth... ......... (ASK A-J) ............. I Miscarriage ........ (SKIP TO p. 16) ........... 2 Stillbirth ....... ..(SKIP TO 0.16)...........3 Abortion...........(SKIP TO Q.I6)........ . 4 .. 67-ea/ OH Ho Did that pregnancy result in a live birth, or in a Miscarriage, stillbirth, or abortion, lor is (NAME) still pregnant)? take (NAME OF SPOUSE/PARTNER) to becoM DON'T KNOW. D. 14. ........... S What is the f i r s t and last naM of the child as it appears on the birth certificate? RECORD OH CBIU>KE*)'S RBCOW) PORH OH MirrUMEMTART CHILDREN'S RECORD POWf. IRTERVIEHERl RECORD ID FROM OHLDRBI'SftBOOHDPORH |_ |_ | 29- JO/ When Has (CHILD) born? ENTER BIRTHDATE OH CHIUHIEH'S RECORD PORN Oft MPPLEHKNTART CHILDRBI'S MKCOPD FORM. 71/ 2 C. Was (CHILD) nale or fenale? RECORD OH CnUDRKH'S HECORD PORH OR SOPPLEMBTTMT CHILDREN'S RECORD POXH. How Much did (CHILD) weigh at birth? ENTER POUHDSt AND OUNCES! OR Don't know | | j 3I-32/ | | | 33-34/ 98 �14. 14. (Conlinued) E. (Conlinup.1l K. When did (CHILD) die? RECORD DAY. HOMTH. AM) YEAR OH CHILDREN'S RI9CORD POKH OR SUPPLEMENTARY CHILDREN'S RECORD PORH. L. Has (CHILD) a twin? What was the cause of death? Ye*. Mo.. RECORD BELOW. 41/ F. Has (CHILD) preMture, full te», or overdue? PreMture Pull tera Overdue Don't know t 2 1 ..8 J6/ N. Where is (CHILD'S) death registered? In what city and state? (CITY) 42-S7/ G. How old (STATE) SB-59/ was (MANE OF MOTHER) when (CHILD) was born? 15. RECORD AGEi |___L_J Don't know IHTERVlrWERi 37-J8/ IS THERE A SECOND PREGNANCY SINCE THE DATE OF LAST INTERVIEW (IS NUHBER OF PREGNANCIES IN 0.1 2B EQUAL TO 2 OR HORE? I 98 YES H. 16. DOCTOR'S HUME -BTT I NO What is the MM and address of the doctor or Medical facility who has (CHILD) •• birth registration records? RECORD BELOM (SKIP TO 0.17) (SKIP TO 0.25) 60/ 2 When did that pregnancy end? FACILITY HAKE RECORD DATEl to | | | | HO | | DA | 6I-66/ YR STREET ADDRESS A. J_L_I IMTERVIEMERi I. ENTER NUMBER OF WEEKS I (STATE) (CITY) Don't know RECORD NAHE MD ADDRESS OH MEDICAL CONSEHT rORH | | | 67-68/ 9g IF MISCARRIAGE OR STILLBIRTH, ASK B-C| OTHERS GO TO D. what is the nave and address of the doctor or.Mdlcal facility who has (CHILD) •• current awdlcal records? RECORD BELOW. B. DOCTOR'S HAHE How Mny weeks had (NAHE) been pregnant when that happened? FACILITY NAME Did a doctor tell you why this (olscarriaqe/stillhirth) .iqht have occurred? Yes (ASK C) t No STREET ADDRESS (GO TO D) J 69/ JLJLJI (CITY) IHTERVIEHERi (STATE) C. What did the doctor say caused the (Miscarriage/stillbirth)? RECORD VERBATIM. RECORD HAHB AMD ADDRESS OH MEDICAL CONSENT FORM 70/ J. Hhat is (CHILD)'a current aoe? RECORD IH CHILDREN'S RECORD FORM OR SOPrLEHENTARY CHIUHUOI'I RBOOfD tORN. IP DECEASED. ASK K-H. OTHERS GO TO Q. 15. �DECKS 24-3(1 16. (Continued) D. 17. INTERVIEWERl (Continued) IS THERE A SECOND PREGNANCY SINCE DATE OF LAST INTERVIEW} (IS NUMBER OT PREGNANCIES IN Q. »2B EQUAL TO 3 OK MORE?) YES (GO TO 0.17) 1 NO 17. OF.CK ] 0 (SKIP TO Q.2M 2 When did the next pregnancy begin? types of birth control you or (NAME) were using when she became pregnant. CODE AU. THAT APPLY. 7I/ 1. 2. 5. 4. S. 6. 7. HAND CARD H What Month «nd year? 42-437 44-457 46-47/ Douche Jelly, 72-7S/ (_JH_I I HO YR e. 9. A. INTERVIEWERl HAS K HAD MORE THAN ONE RELATIONSHIP SINCE DATE OF LAST INTERVIEW? (SEE INTO SHEET, ITEMS OT. O* AND Oil YES (ASK B) I (GO TO C l . . . NO.. Which (apouse/partner) had this pregnancy? 18« Did that pregnancy result In • live birth) or In a piscarriage, stillbirth, or abortion, (or is (NAME) still pregnant)} Live birth Hiscarrlage Stillbirth Abortion... Still pregnant rvJ (MIDDLE) INTERVIEWER) RECORD ID I FROM INFORMATION SHEET. ITEM. OS. O6, OX O7 A. L_L_I RECORD HOHTHS AND/OR YEARSt 35-36/ take (NAME OT SPOUSE/PARTNER) to bscom | | MOS | AND/OR I Hasn't trying OR DON'T NKM D. I I YRS I 2 3 4 5 687 What Is the first and last naa* of the child as It appears on the birth certificate? RECORD ON CHILDREN'S RECORD FORM OR SUPPLEMENTARY CHILDREN'S RBCORD PORN. INTERVIEWERi RECORD ID FROM CHILDREN'S RECORD FORM OR SOTPLEMIMTARY CHILDREN'S RBCORD FORM. | | | 69-70/ B. Mien vas (CHILD) born? ENTER BIRTHOATE ON CHILDREN'S RBCORD PORN OH SUPPLEMENTARY CHILDREN'S RBCORD PORN. C. Has (CHILD) Bale or fenale? RECORD ON CULDRnTS RECORD PORN OR BOrPLEMBITARY CHILDREN'S RBCORD FORM. D. How mich did (CHILD) weigh at Mrth? 39-40/ 98 were either you or (NAME OP SPOUSE/PARTNER) using birth control at the tlaie ah* became pregnant? No (ASK A - J ) . . . (SKIP TO Q.2O) (SKIP TO p.20) (SKIP TO O.2O) (SKIP TO O. 25) 37-3S/ OO Yee 64-6S/ 66-67/ 12 IO-34/. How Many snaths did it pregnant (thla tljw>? <tg 58- 59/ 60-61/ 62-63/ 2 RECORD NAHEt C. O9 S2-53/ 54-557 S6-57/ 07 It/ BEGIN DECK 3O B. 10. tl. 11. Condos) Diaphr Diaphr Rhythm Rhythai Nlthdr awal Other (SPECIFY) DON'T ENTER MONTH AND YEARi 48- 49/ 50-5 1/ OS (ASK E) (SKIP TO p. 18) 1 2 41/ ENTER POUNDS I AND OUNCES I OR Don't know | _ |_ | 71-727 73-747 98 �BEGIN DECK IB. (Continued) E. 18. Mas (CHILD) a twin? 31 (Continued) K. When did (CHILD) (lie? RECORD DAY, MONTH. AND YEAR ON CHILDREN'S RECORD FORM OR SUPPLEMENTARY CHILDREN'S RECORD FORM. L. Hhal was the cause of death? 7S/ RECORD BELOW. 10/ F. Has (CHILD) preMature, full tern, or overdue? PreMature. Full terM Overdue Don't know G. How old I 2 3 8 76/ | Don't knov | 19. INTERVIEHERl | I1-26/ I I (STATE) 27-2B/ IS THERE A THIRD PRPGNANCY SINCE THE DATE OF LAST INTERVIEW? (IS NUMBER OF PREGNANCIES IN O. 128 EQUAL TO 3 OR MORE?) YES 20. (SKIP TO p. 21) I NO What t» the tame and address of the doctor or Medical facility oho ha» (CHILD)*• birth registration records? RECORD MOON NJ (Jl In what city and state? I 98 DOCTOR'S MAHE Where is (CHILD'S) death registered? (CITY) was (NAME OF (WINER) when (CHILD) wa» born? RECORD AGE I H. (SKIP TO O.25) 29/ 2 Hhen did that pregnancy end? FACILITY NAME RECORD DATEi ' * ' MO ' • DA ' I JO-JS/ YR STREET ADDRESS J (CITT) INTERVIEHERl I. I I A. (STATE) F.NTER NUMBER OF WEEKS: RECORD MMUt AMD ADDRESS OH HEDICAL CONSENT FORH Don't know Hhat is the msM and address of the doctor or Medical facility who haa (CHILD) 'a currant Medical records? RECORD BELCH OR | | | 16-J7/ 98 IF MISCARRIAGE OR STILLBIRTH, ASK B-C| OTHERS GO TO D. B. DOCTOR'S NAME How many weeks had (NAME) been pregnant when that happened? FACILITY NAME Did a doctor tell why this (Miscarriage/stillbirth) Might have occurred? BO/ Yes (ASK C) 1 No (GO TO D) 2 STREET ADORES* (CITT) INTERVIEHERl J. L_I_J (STATE) RECORD NAME AND ADDRESS ON HEDICAL CONSENT FORH Hhat is (CHILD) "8 current age? RECORD IH CnUMBN'B RECORD FORM OR SUFFLCMEHTARf CBILDRBT* RBOORD FORH. IF DECEASED. ARK K-H. OTHERS GO TO 0.19. C. Hhat did the doctor say caused the (Miscarriage/stillbirth)? RECORD VERBATIM. IS/ �HECIN UEX'K 32 2O. 21. (Continued) INTERVIEWER: YES 21. IGO TO 0.21) I (SKIP TO 0.2S) 2. 3 4 5. 6 . 7 8 9 HAND Mhat Month and year? ENTER MONTH AND YEARI INTERVIEWER; 4O/ Please look at thla card and tell M all the numbers of the types of birth control you or (NAME) were tialng when she becaae pregnant. CODE ALL THAT APPLY. 2 When diet the next pregnancy begin? A. E. IS THERE A THIRD PREGNANCY SI NOB DATE OF LAST INTERVIEW? (IS HUHBKR OF PREGNANCIES IN p. 12B EQUAL TO 3?) NO (Continued) | | || | HO VR 4J-44/ to. I NO 12. 12 22. RECORD NAHEt Live birth 46-7O/ A. INTERVIEMERt RECORD ID « FROM INFORMATION SHEET, ITEM OS, O6, OR O7 L_L_J 71-72/ How nany Months did It tak* (NAME OF SPOUSE/PARTNER) to becoM pregnant (this tlrM)7 AND.OR 7S-74/ .00 OR 75-76/ Tea (ASK E) No...(SKIP TO p.22) I 2 77/ 4 5 (SKIP TO Q.24) (SKIP TO O.2S) RECORD ID FROH CUJUMIEH'S RJX3ORD FORM OR SUPPLBHENTUII OHLDRJM'l |UKXNU> FORM. |_ | 37-3B/ B. When %>aa (CHILD) born? C. Nat (CHILD) Bale or fe»ale? RECORD ON CHILDREN'S RBCORD FORM OR SOmaMENTARI CHILDREN'S MCORD FORM. D. flow such did (CHILD) weigh at birth? .8 9 Were either you or (NAME OF SPOUSE/PARTNEK) using birth control at the tine >he becaiw pregnant? 36/ 2 3 What is the f i r s t and last naa* of the child as it appears on the birth certificate? RECORD ON CHILDREN'S RBCORD FORM OR SOPFLKMEHTAMY CHILDREN'S RBCORD FORM. INTER VIEWER i D. t Abortion Still pregnant (MIDDLE) DON'T KNOW.... (ASK A-J) Miscarriage (SKIP TO O..24) Stillbirth.........(SKIP TO O.24) rv> wasn't trying. 32-3J/ 34-3S/ Did that pregnancy result in a live birth) or in a Biscarrlage, stillbirth, or abortion, lor ia (NAME) still pregnant!? iLAST) RECORD MONTHS AND/OR TEARSt Other (SPECIFY) 45/ Which (spouse/partner) had this pregnancy? C. 12-S3/ 14- IS/ 16-17/ 18-19/ 20-21/ 22-Z3/ 24-2S/ 26-27 / 20-29 / 3O- 3 1 / 2 (ASK Cl (FIRST) 1O 11, HAS R HAD MORE THAN ONE RELATIONSHIP SINCE DATE OF LAST INTERVIEW? (SEB IHPO SHEET, ITEMS OS, O6 AND O7) IBS ,O4 OS ENTER 8IRTHDATE ON CHILDREN'S RECORD FORM OR (OPPLaMBirTART CBUDRKN'8 RECORD FORM. ENTER POUNDS I | | | 39-40/ | | | 41-42/ AND OUNCES, OR Don't k n o w . . . . . . . . . . . . 9 8 �22. (Continued) 22. E. Was (CHILD) a twin? L. Where is (CHILD'S) death registered? How old INTERVIEUERt " Don't know | | J 90 (CO TO HEH QUESTIONNAIRE) NO (SKIP TO 0.2i) 1 68/ 2 Mhen did that preqnancy end? RECORD DATE, DOCTOR'S NAME 66-67X IS THERE A FOURTH PREGHAHCY SINCE THE DATE Of LAST INTERVIEW? (IS BUMPER OF PREGNANCIES IN O.I 26 EpUAL TO 4 Oil MORE?) »ES 24. I 5O-65/ 45-46/ Hhat is the name and address of the doctor or medical facility who has (CHILD!*s birth registration records? RECORD MELON I (STATE) (CITY) was (NAME OF MOTHER) When (CHILD) was born? RECORD AGEi T In what c i t y and slate? I I 2 3 8 23 H. RECORD BELCH. Has (CHILD) premature, full term, or overdue? Premature roll term Overdue Don't kno« G. What was Ihe cause of dealh? M. Yes. No.. F. (Continued) I L_LJ_LJ ' HO DA YR 69-74/ FACILITY HAME 47Y H. How many weeks had (HAMFJ been pregnant when that happened? STREET ADDRESS ENTER NUMBER OF WEEKS. JLJLJ (STATE) (CITY) INTERVHMERI I. Don't know | 7b-76/ 98 B. Did • doctor tell why this miscarriage mitlht have occurred? ' DOCTOR'S NAME 1 IF MISCARRIAGE. ASK B-C| OTHERS GO TO D. RECORD HAKE AMD ADDRESS OH HEDICAL COHSEHT FDRH Mhat is the name end address of the doctor or medical facility who has (CHILD) <s current medical records? RECORD BELCH | Yes (ASK C> » (GO TO D) 2 "/ FACILITY MAKE Ho STREET ADDRESS C. (CITY I INTERVIEHERi What did the doctor say caused the miscarriage? RECORD VERBATIM. to/ (STATE) RECORD NAME MO ADDRESS OH MEDICAL CON8EMT FORM J. Hhat Is (CHILD)'s current aq«7 RECORD IM CBOOKEM'* RBCORD FORM OH SUPPLEMENTARY CTILORJ9VB RBCCMD nMM. IF DECEASED, ASK r-H. OTHERS 00 TO p. 23. *.. When did (CHILD) die? RECORD DAY. MONTH, ANT) YEAR OH CHILDREN'S D. INTERVIEWER: ' IS THERE A FOURTH PREGNANCY SINCE DATE OF LAST INTERVIEW? (IS NUMBER OF PREGNANCIES IN 0.12B FOUAI. TO 4 OR MORE?) jES NO (GO TO NEW QUESTIONNAIRE) .. .I 2 79/ �BEGIN DECK 2S. )} since (DATE OF LAST INTERVIEW) have you ever tried for a period of one year or More, to conceive m child and were not able to do eo? Yea 32. INTERVIEWER- IS THERE A SECOND PERIOD OF INFERTILITY SINCE DATE OF LAST INTERVIEW? (IS O.Z6 CODED "THO" OR HORE7) IO/ No YES (SKIP TO SECTION 6>. BO 26. (SHIP TO SECTION 6). For how many periods of one year or Bore did this happen? 33. One... Two... Three. • I/ Since (DATE OF LAST INTERVIEW), in what aonth and year did the aecond period begin? And in what annth and year did It end? Begin Since (DATE OF LAST INTERVIEW), in what exmth and year did the flrat period begin? And in what Month and year did it and? Begin 12-IS/ LJL_L_LJ MO 28. CO End L_LJ_LJ YR MO 34. I6-19/ ««•«»> U_I_U MO YR During thla aecond period what waa your wife or partner'a flrat naa«? RECORD «LON. <a 6</ a*"0 '-WT-' YR 3S. During (PERIOD) what waa your wife or partner*a Mrat naa«7 RECORD 8ELCM. Mow old waa (MANE) in (BEGINNING DATE OF PERIOD)? JO-33/ 34-3S/ M CO RECORD AGEi M. 29. How old waa (NAME) In (BEGIMHING DATE OF PKRIODI7 RECORD KG* 30. ( | | 37. 64-6S/ 66/ CODE 'PERIOD 2" AND HAND SELF-ADMINISTERED FORM 1. There are many reaaona that eone couple* find it difficult or l«r>oa*lble to conceive a child. Pleaae read thia card and circle the nunber on Side A for each reaaon which applied to you for thla period. Side B provide* reaaona appropriate for your epouee. Circle aa Mny reaponaea aa appropriate. CODE •PERIOD I" AND HAND SBLr-AOHIMISTERED FOMi 1. Nov pleaae f i l l out thla card and place it in Ute envelope when you are finlahed. | Ho.. During (PERIOD), did either of you aea a doctor to dlacuaa any difficultly in conceiving children? There are awny reaaona that eo«e couplea find it difficult or inpoaalble to conceive a child, pleaae read thla card and circle the nuMber on side A tor each reaaon which applied to you for thla period, side B provldea reaaona appropriate for your apouae. circle aa eumy reaponaea aa appropriate. | Yea. No.. 31. | During tbta aecond period dl'd either of you aee a doctor to dlacuaa any difficulties In conceiving children? 36-J7/ Yea. 44-47X End 40-41/ L-LJU-J HO YR. 27. 19 / (GO 10 0 . 3 ) ) . . . . Now pleaae fill out thla card and place it in the envelope when you are finlahed. 38. INTERVIEW HI. IS THERE A THIRD PERIOD OF INFERTILITY SINCE DATE OF LAST INTERVIEW? (IS Q.26 CODED "THREE"?) YES NO (<X> TO Q.19) (SHIP TO SECTION 6) 1 2 67/ �DECKS 33-34 39. Begin 68-71/ End OR HAS NOT L_I_LJ_J HO YR U_l_l_l MO YR 4O. 0000 During (PERIOD) what waa your wife or partner 'a flrat MM? '-A- I L-LJLJLJLJL | | | INTERVIEWER; 1O-23/ 24-2S/ HAS RESPONDENT HAD ANY BIOLOGICAL CHILDREN? YES NO (ASK A) 1 (SKIP TO 0.22) 2 A. 26-I1/ ME CHILDREN RECORDED OH CHILDREN'S RECORD FORM? • YES 42. (ASK B) I NO (SKIP TO 0.2) 2 30/ During (PERIOD), did either of you see a doctor to dlacuea any dlfflcultlea In conceiving children? B. Tea. FOR EACH CHILD LISTED ON CHILDREN'S RECORD FORM ASK: Our records Indicate that (CHILD)(had/did not have) a birth defect in 1982, when you were laat interviewed. Is this information correct? No.. IF INFORMATION IS CORRECT..(GO TO O.2). ro vo 43. 2t First, I would like to verify whether (any of) your biological child(ren) have |had| a birth defect. HOH old waa (NAME) In (BEGINNING DAT* OF PERIOD)? RECORD AGBl Child and Family Health How I would like to ask you some questions about birth defects in your family. By birth defects I mean a physical abnormality present (though not necessarily noticed) at the time of birth. Birth defects ranqe in severity from unusual birthmarks to a missing or mishappen limb. Birth defects can affect any part of the body, including bones, body organs such as kidneys or the heart, reproductive and respiratory systems, blood, and the skin. I. RECORD BELOH. 41. Section 6: Since (DATE OF LAST INTERVIEW), In what Month and year did the next) period begin? And In what Booth and year did it end? .1 IF INFORMATION IS INCORRECT, MAKE CORRECTIONS OH CHILDREN'S RECORD FORM (Then go to O.2) There are Mny reaaona that some conplca find It difficult or impossible to conceive • child. Pleaae read thla card and circle the number on Side A for each reaeon which applied to you for thla period. Side B provide* reaeone appropriate foe your apouae. Circle a* Many reeponaea aa appropriate. 2 Don* t know. CODE 'PERIOD 3* AND HAND SELF-ADMINISTERED FORM 1. B .(GO TO Q.2) nut EACH CHILD (EXCEPT CHILDREN WHO DIED BEFORE 1982) ASKi Has a(n additional) defect been identified in (CHILD)(since 1982)? RBCORD OH CHILDREN'S RECORD FORM OK SUPPLEMENTARY CHILDREN'S RECORD FORM. Now pleaaa fill out this card and place It In the envelope when you are finished. 3. INTERVIEWERi HAVE DIED. ASK QUESTIONS 4-2O FOR EACH CHILD, INCLUDING ALL WHO MAY (Now I would like to ask about (NEXT CHILD)). �-56- orCi442BSec-6 2«> CHIIO 1ST CHI 10 CHIIO'S WNfi CHIIO'S lot HOIHCR'S 10* 4. | | 1 32- 5 V J4-3V OM (CHIIOI «v«r k*v« • r»> 49- 50/ SI-52/ No »•» No ..i .. 2 m ..1 .. 2 M/ «...! 2 5V CATftOMtl... pky»lc*l or Kitor k»p«tr»«iit.l 3. 2 3>/ MM (CHIlOt »»«r <l»gno<«d ..(A PLEASE GO OH TO NEXT PAGE &B . I I. M/ ...(0.6I...2 40-4V •. «tat Ho< of Don't kno. DtFCCI m WPAIIHEHT KIN iKNTirifo m iCHiioit CHtCK CHI LOOT'S MOOR) ram «MD QS. t,i, wo 4. ». IMIHWIiKilli IS THfRE WIOfHdl CHIIOI ns. .(».?) NO.. TIS ....... ASK 0.4 MO » I 2 ....... (SKIT ID 9.221 46/ i ..1*1 imTOo.j». NO ?. R «;/ 2 .« AM> 3)..I 10.22 2 i «/ («.»» I 2 <0.ll 2 o« tblrth 4*l«ct<l>/_ 4I/ ICHIIOI 014 you (or MMVOA* •!»•> *UcM» ICHItD'I) Cklrth «*t*ct<»l/l»*«lriw«tl «ltk • doctor* r.» No (so ID «.«> <SXM> TOQ.IIt 6V �-57- DFX-KS 34-35 CHHD'S WW(: 3RD CHltD 4TH CHILD J 66-6T/ ' 6B-69/ No 2 70/ »«S I 1 2 2 I4-I3/ 16- 1?/ t.» I Ho 2 IB/ I 72/ I 1 6TH CHILD 1 3I-32/ 1 33-34, 1 | 1 | CHUO'S ID « MOTHER'S Ip I 4B-4./ 30-3I/ ». 2 20/ N> ».« Y« No 7I/ 1 . 3TH CHILD ..LI J_L I L t JL.L IMTER»IE.ntR! IF DCriCIISI. DISABILITY. OR HPAIFMEN1 FOUND SINCE IW2 ASK 0.»; OTHERS 00 TO 9.10. Klxt It tko ••«• mkd nddrot* of tk* voctor «fco nlM|iK»«d CHILD •• ko»l»n * tlrtk «of oet/lBnalrmMit t DOCTOR'S HME DOCTOR'S NME CD OR FACILITY MM! FACIIITT NME STREET ADDRESS ....(A I B)....l 7V ....(A I I . . B..I 2I/ ....(A t B)....l 3B/ STDfET AOORESS ....(A 1 B)....l 3V .....<9.A>.....2 10. 74- 7 7/ 22- 2V L_l__l_l_l HO m HD TR MO MD TR DM tko doctor >*y th.l (CHILD) nootf(l/o«) My toitlnn, •odlentloo, trootnonf. t«ro*ryv or 3*-3*/ 3V-42/ Yd koenno of • Iblrtli ••loct/leoolnmtlt (By 2*/ Don't mo-.....» Don't kM».....B 60/ 4V • wlwolctiolr, ••Ikor. •rtltlcUl Itak, body lt or crwtckos). Don't kno.....A Don't kno».....B II. BESIM DEOK 33 (9.7) 1 IQ/ ....(ASK A). ...2 (9.> * »»/ (9.7 1 44/ ....(ASK AI....2 ....(ASK AI....2 9.7 1 «!/ ....(ASK AJ....2 ..19.4 AND 3 . I I. 4V ..19.4 AND 31. . 1 DU (CHILD) *vnr r*e«l«o •»y t*ttln||, M4lcotlon. tr**l*on*B ••rnory or •ooelnl onvlniw of • (klrlk d«f< Don't lwo> 12. ..(9.4 AND 3>..t ll/ ..(9.4 AMD 31.. 1 2B/ Don't Knov.................B *2/ • •••(9.221.... .2 At »y tl<M. did (CHILD'S) (klrtk «ofoet(i>/loo«lrMiit> Intorloro In pay «ny «ltk ICMILO'*) ckyilol or KKUI >o»*lo*M«tf For OKMinIo, I Yot (GO TO 9.13) No I ASK A) I 6»/ 2 .(O.III .(ASK A) , Jo. or -HI*, tv I2/ 46/ Don't kno> •V A. IMTEI«lt»€Rl DM TMtM * TtS COO60 AT g.10 or t (00 TO 0.1» ...... I »•/ .(0.13) O.I It (9.*l (O.I It 1 2 IV (».«) 1 (9.IH...J 30/ I9.*> 1 47/ 1 . 1 ... 01 1 . 2 I9.*> 19.11 2 «....( ASK Bl.. ...... .2 1 ««/ •. l»I[Xlt«t»l IS THEHC ANOTHEH CHILD1 TES....IOO BACK TO O.4 FOR NEXT CHILD! . . I .. T00.« 1 .IASK «> .(0.4) .(0.221 »!/ �lit* t :. JHD CHILD 4IH CHILD » • • - lit 61 H CHILD MH CHILD 2ND CHILD ISf CHIID H H I 1 DOCTOR'S NAME OR DOCTOR'S NAME ?)/ 1 1 1 IJ/ CHIID'S NMC: CHILD'S ID 1 NDIICR'S ID 1 OR FACIlIt NAHE 1 1 DOCTOR'S NAHE OR DOCTOR'S NAHE OR FACILITY HAMC I It/ FACILITY NAME 25/ 1. 1 FACILITY NAME OH (CHIID'SI doctor lay ttot U«r o(l (CHIID'S) (klrtk «*r«ct(«)/liv«lr««iiti ««t/«w* llf»-tlir««t«nlfig If Utt ••trfltftt m, It/ )J/ SIREET M»RESS 1 1 STATE CITY STREET ADDRESS 1 1 CU< 1 1 STATE SIREET ADDRESS 1 CUT 1 1 STATE 1 CITY 1 MtrMl<4 1 mtrnn II (CHILD) iH Ml r.c.lv. xiroM-r. mftlettlan, » >p«cl«l «l.t. or >o»« o«hw Mdlcil Inter rwvtloo.) 1 StATl A. INTEMIEICRi (ASK A) 1 Ho (ASK A) J (ASK M.....2 Don't kno»...IA9( Al * (ASK A>.....« !»/ (ASK A) IS CHILD WOER DIE MFORC HE/SHE MS I HO fEAKS OlOt 14. t.t JZ/ STREET ADDRESS OK (CHIIO) •«•!• >••< k»lp • Itk Mtlng. «r»»lng. k»tkl«t> or Mlxg tk« tolut HO J 3V )9/ )4/ <0/ JV «!/ J6/ 42/ l*V*lr«*attMH»l» locltido w tsJ rtlkM- tkM >i«t itxxtl.g kr t» «»llt l« KM««d.l .2 ... ..J IS, ••CWM ol • (klrtk tf«Uct/ l«V*lr«mtlc 4H (CHILOI «»«r K6IH DECK it ««• or iio*4 *ny ••chMtlctl •hMlcMlr. nllwr. bo«y (ASK A) 2 (ASK Al J cr«tck»* to corry out •<twf <«r ecti'iti.tt 1. * (9.l» 1 M/ (9.111 * 1 " (9.111 1 21/ <9.l)> M» (CHIlOt ovwr un.bl. <o tcto »«rt «t >ll In 1 2»/ cklloron k*c«i*>« of • (ASK » J (ASK •) J CASK 01 J» A. (9.4) ..1 I2/ (9.41 1 I*/ (9.4 1 24/ *<M 1 IkM (CHUO) ovw ll»lt«d SO/ •toy k«/»k* co«l4 do k»c»wt* fiiwlraw>t>l (9.22) ..2 (9.22 a <Q.22> a �2HO CHIIO CHIIO'S MAKE: 3RD CHIIO 4TH CHIIO SIH CHIIO •IH CHILD CHIIO'S ID I itoiMtH't ibT" H 1 11 11i III 1 1 1 U. I 1 014 (CHILD'S) (birth 4.toctl»l/ l*o«lr».nt> ovor k..p (*!•/ <»/ (ASK A) (ASK A) (ASK Al 1 I Jt (ASK Al SV 1 • •••••(A9C Ala,***,*! • •••••(ASK Ai v > C ( 4 B 3 (ASK Al (ASK Al J J» s*/ ...l»* A) 1 6V (0.111 ..... 1 'V 1 014 (CHILD) ov.r KKV.J to 90 to o cwtaln typo ol school, or t>« In o Yo*. (GO TO 0.1*1 ...... 1 M/ (ASK •) of (kll/kwl(blrtk ••••••CASK Al* •••**2 IASK AI....2 .2 ........ ...<o.ie> i '«/ ...(ASK 81...J J ••••••(AW Al •••*••• vu A. • I/ ASK A .(GO TO 0.191 No.. 43/ t.». korl Iron going to school! Ml (CHIIO) .v.r ll.lt.4 •7/ .I . »/ .J III . . . (Q.20I.......I ... ••....•..•.•.. SCkOOl •ttOAdOIICO OT !• tol«9 >kl* to loom kocow* ol Ikli/horKklrtk 4«/ Jt I . Bocovw ol IkU/horl (klrtk I 4.f»ctUI/l«*«lrB.iit> 414 SI/ 43/ SI/ (CHItO) ov»r HM« • lot inro hot* tkM othor Ckll4ro« (CHILD'S) ••> In oolag oiittloo. •V 7 fO*tl«f «o (ckool. 901119 to «ko »toro. Mi4 ottMr ovoryvoy oclKltl.l Ilko tK.ll It. >•/ SI/ 4«/ •«/ of • (klrtk tfofoct/ I, 414 ICHIIOI •vor (woo1 tko kolo of •wlkor pwioo for ovory4«y 4ctl«l*lol »ck «« tckliit cwo ol tho kooto T.I. .I . HO.. .J 4 1/ ..••••(Q.I7>.......t (ASK Al J <«/ SV IO.ITI (ASK Al S9/ 1 J 9V I *v *.....<Q.I'I.*** . 1 . IASK AI a BEGIN DECK 31 Tot (80 TO 0.2 II Ill (klB/kor fro> -jrkl«9 on • Jok for »oyt A. IASK Al*..,..2 tot MIII/Ho.14 ICHIIO'S) klrtk «ofoct(tl (koop/b*.* lt.pt) 6V B. «IM/»>«I4 (CHItOI (b»/h«y. kOMl Ibltod In IM »l"4 ol «ock (h./th.) co.14 Uo/h»>o «o»o) fcocouM ol (hl«/korl klrtk 4«f«ct(ill Mll/lto«l4 (CHILD) (b>/kovo k*M> Ih(to4 In IM Moo*t of >ork lko/lh*l ctx!4 (4o/M«o 4oo.) k*c«u>. ol (kli/korl klrtk 4*loctltil 7»/ J or »r*Mrliif iwoltl 1 . 0 I »!/ .(O.III ID/ .(ASK AI....J It, (COTOQ.2I) I IV HD........IASK Bi...........J V.«. ..I No., .J .(g.2i) i .(ASK BI....J >V ii/ �OETK 17 WO CHILD 4IM CMItO MM CHIID «H CHIIO 2ND CHILD 1ST CHILD 1 1 I I ..I .B ..9I I 1 I)/ ....(ASK A>. ...2 9.iB> i 2i/ IASK AI....2 ..1 . * .. 91 1 ( CHILD'S MM4EI 1 1 2V IO.IBI ....IASK AI...J IASK Al . . I .* ..9l > 1 (O.I* I/ I CHILD'S ID* NOIHtlt'S 10 * 21. J INItmiiWH: ODES RtSPOHDCNI MA»t 'ANOtHtR CHILOI .I.. «S....(6O BMX IO 9 I . . I 4V a ....iQ.iat i M/ ....(ASK 6>...J . . I9.l*> .. 1 JJ/ ....IASK Bl... J 1 JO/ ....(ASK Bl... J O 9J21 2 19.22 22. e«****|A5K BI»« ( ***I Did you ever have • birth detect? tern (ASK It) Mo ft. What kind of birth defect « • it? • T 21. US t 2 Any otheri? Do yon have any biological brother* or eictere? Include any brothers or BlBtern who Bay have died before the age of 1. Tee ........... (ASK 0.24) .......... 1 No .......... (SKIP TO P. 25)........2 Don't Know.. (SHIP TO O.25)........8 »*•* •*» •• •• • •! 4I/ . 1 4Z/ ... ..J ....49.211 1 IV ....(9.211 1 MS ....IASK AI....2 ....<ASK AI....2 ..1 . 1 .. 92 ) ..I . I ..»2I 1 I*/ IASK BI....2 1 2J/ ....IASK Bl... J ....19.2*1 1 W *SK A)...J ..1 . 1 .. 02 1 IASK At 1 3V ....(ASK BI....2 fQ.Zt> ; 2 M/ 2 <Q*2I> 1 ••••••(ASK •)••••••! 4V 49/ �24. 4IH CHIlo 3RD CHI ID 6TH CHIIO STH CHIIO Did any of your hioloqical brothers or sisters ever have a hirlh defeeL? (ASK A) t (SKIP TO p.25) 2 »«s 1 1 1 1 1 I I 1 1 1 Ho Don't know...(SKIP TO Q.2S) (Q.JI 1 *>/ <«.» 1 >!/ 9.) IQ.22I..-..I II 1 «/ A. (Q.2J1....2 S4/ (Q.22I....2 .k... .19.111 Mho had a defect, brothers, sisters, or both? J Brothers I Sisters 2 Both 55/ 3 FOB RACH SIBLING WITH A BIRTH DEFECT, ASKt What kind of birth defect did your (brother/sister) have? Has this sibling a half (brother/sister) or a full (brother/sister)? RECORD BELOW. Siblin9 I Slblinq 2 DEFECT! 56/ DEFECT t SB/ Half (brother/sister)...! Full (brother/sister)...} S7/ Half (brother/sister)..! Full (brother/sister) . .2 S9/ u> 01 Sibling ] Sibling 4 DEFECTi DEFECT i 62/ Half (brother/sister)...! Full (brother/sister)...I 25. 6O/ 61/ Half (brother/sister)..) Full (brother/sister)..2 63/ How I would like to ask you aoM questions about your biological parents. Did either your biological pother or biological father ever have • birth defect? US (GO TO 0.261 1 Mo ( S K I P TO 0.20) 64,' 2 Don't know..(SKIP TO O.2S)..B 26. Which parent had a birth defect? Mather only 1 Father o n l y . . . . . . . . . .2 Both parents 3 6S/ �DECKS 37-3B 27. t*1 !442ns<><--7 SF.CTIOtl 7l What kind of hirlh defect did your (PARENT) have? Hothen 66/ Father. HEALTH 67/ I. Now let's t a l k about health. Conpared to other people your aqe, would you say that your health is ... (READ CHOICES)? Now I have Bone different kind of questions. 20. Excellent Mas anyone near to you died In the last 12 Months? Good ........................ Fair ........................ 3 Poor ........................ 24/ 2 4 Yes..(ASK A AND B . . ). HAND CARD 1 Ho 2 68/ ................... » I A. What was the person's relationship to you? as apply. CODE ALL THAT APPLY. Please choose as Many 2. Since (DATE OF LAST INTERVIEW) have you had acne on your face, chest or back? A. Child Ol 69-70/ Ye B. Parent O2 71-72/ No c. Spouse/partner 03 7J-74/ D. Brother or sister...................................O4 7S-76/ E. Other near relative of you or your spouse/partner 3. F. Other (SPECIFY) ...06 79-BO/ BEQIH DECK 39 Ourinq what year, between (DATE OF LAST INTERVIEW) and now, did you last have acne on your face, chest or hack? 4. 26-27/ Think about the (first/next) tine you had acne on your face, chest or back between (DATE OF LAST INTERVIEW) and now. Mien did it start and until when did it last? (PROBE FOR ALL PERIODS OF TIME). IO-I1/ Third First Mist (was the date/Mrs the dates) of the death(s)? year? ENTE* MONTH AND YEARi 2 RECORD YEARi _O7 B. ( S K I P TO 0'8> 2V ...OS 77-7B/ Friend G. 1 I—I HO Mhst month and l_J_J I—I—LI—I Mo Yr I2-I5/ I EHTEM MONTH AND YEARi I ' LJ_J HO YR I6-I9/ L_l_l 20-J3/ MO YR I L-LJ—I—I Mo Ho Yr to to YR LJ_LJ_I ENTER MONTH AMD YEARi LJ_L_LJ Yr LJ_L_LJ U—l—LJ Mo Yr �IIKC KS JH- )•) 7. HAND CARD p. J A. 4 IHTPRVIFMRRi ASK A FOB FACH TIHR IN p. 4 What was Lhe name of the doctor or medical facility you consulted at the tine? Plea se show Me on this diaqran where the acne (is/was) located (the ( f i r s t / n e x t ) time). CIRCLE "ITS' OR -MOSECOND TIME Ho Yes FIRST TIHR Yes NO THIRD TIME Yes No First Nane Temples 1 2 52/ 1 2 «!/ 1 Eyes or eyelids 1 2 53/ 1 2 62/ 1 2 7»/ Ears t 2 54/ 1 2 6V 1 2 72/ Cheeks 1 2 55/ 1 2 64/ t 2 7V Nose 1 2 S6/ 1 2 65/ 1 2 7«/ 2 70/ OR Facility Nane A. What is the address of that (doctor/medical facility)? 14/ STRERT ADDRRSS Forehead 1 2 57/ 1 2 66/ 1 2 «/ Jaw, Chin other 1 2 58/ 1 2 «7/ 1 2 76/ Chest 1 2 59/ 1 2 6B/ 1 2 77/ 69/ 1 2 78/ Back 1 2 60/ 1 2 J I I (STATE) B. What Is your blood type? IS/ IF ANY "YES" TO TEMPLES, BYES, EYELIDS OH EARS IN A ABOVE. ASK 0.5 ALL OTHERS SKlr TO p. 8 S. Between (DATE or LAST INTERVIEW) and now, did you »»er consult a, doctor or Medical facility about tha acne on roar lt«ples/eyes/eyellds/ears)7 Yea Ho (SKIP TO p.») fi. 1 2 79/ What Month and year did you last consult a doctor about tha acne on your Itemples/eyes/eysllds/earsl? BEGIN DECK 39 LJLJLJLJ 10-iv A. 9. Is that positive or negative? During the last year, how often, on averaqe. would you say you use aspirin? Hare than 4 aspirin a day I 4 aspirin a day (2 dosss a day) 2 2 aspirin a day (1 dose a day) 3 6-e aspirin a week (I dose, 3-4 days/week)..4 4 aspirin a week or less S Hone 6 I 7/ �ASK OF ALL RESPONDENTS 10. IS. The next questions deal with your reaction to the sun without the uee of guntAn lotions. If your akin wa* exposed to the sunlight for the f i r s t tine in the snaMi-r for a period of 3O Minutes would you Alwaya burn, never tan I Usually burn, tan with difficulty LULU no 2 2J-26/ Tr IB/ 16. HAND I CARD H Nhat is the full name of the doctor who Made the diagnosis or the Medical facility where the dlaqnoels was Made? SoMetlsies burn Mildly, tan above avaraqe....! Rarely burn, tan with •••• 11. During whal Month and vear did a doctor first tell you that you had a peptic or RloMnrh ulcer? LAST 4 In the oiimmrr. once you have already been In the aun several tipiea. what reaction will your akin have the next tie* .you qo out la the sun for two or More hours on • brlqht day? Mould you say you qet A painful burn I A burn 2 SOMB redneas only 3 No Reaction FIRST I9/ 4 fACIUTl MMIE Ik. Hhat i« the addresa ot that (doctor/Medical f a c i l i t y ) ? 27/ STREET ADDRESS U) 00 12. After repeated sun exposures. Cor exaMple, a two wesk vacation outdoors, will your skin becos* . . . Only freckled or no suntan at all HAND I CARD N t Only Mildly tanned due to a tendency to peel 2 Moderately tanned 3 Very brown and deeply tanned.... 11. 14. l CITT , 2O/ Durlnq sny period in your life, did s doctor ever tell you that you had a peptic or stoMSch ulcer? fas (GO TO ff.15) 1 22/ .(SKIP TO Q.24) 2 I Da you have a peptic or atomach ulcer now? lea ...4 HAND SELr-Af*UNISTntn> FORM 2. He would like you to tell us sll the places youeve lived sines you ware born. Please list all the places you've lived for More than 12 Months starting with the flrat place 2I/ since birth, 'lease take your tine. It will probably take you to Minutes or so to f i l l out this fon. Please begin. Ho. 17« I (STATE) 28/ B.. o. IB. Mhat aionth and year did you laat conault a doctor for your peptic or atoMach ulcer? LJJLLJ Ho Tr 29-J2/ Have you ever during any period in your l i f e had a bleeding ulcer? 1 Tea No (SKIP TO p.21) 2 IV �2O. 24. During what nooth(a) and year(a) did you have a bleeding ulcer? FROM FROM FBOM Ir Tr TO Please indicate which of the following »e»lx!rs of your bioloqical f a n i l y have ever had a peptic or «to«ach ulcer? TO (to TO 2. Father Full Brother 4. Yr Mather 1. Ho I. ............... Half Brother ..O1 ................. O2 72-7I/ ........... O3 74-7S/ ........... O4 76-77/ OS 78- 79/ 21. 5. SB/ Tea. ............ Hone 8. Here you ever (during any period In your life) hospitalized for your peptic or atoMch ulcer? ............ Half Slater 7. Tr Full Slater 6. Ha 7U-7I/ Don't Knot. BEG III DECK 40 O6 1O-1I/ ................... O7 ............. 12-I3/ 98 14-1S/ Mo.. 2S. 22. Do you have or have you recently had aharp upper etomach pain? Tea Have you ever (during any period in your lifel had eurqery for your peptic or etoMdi ulcer? 26. ...... Tea 21. Are you currently taking any prescribed •edlclnee for your peptic or • toaiach ulcer? tee ......................... Ho ..... (SKIP TO 0.24) A. 1 ....... I6/ 2 Haa this pain relieved by food, Milk, or antacids? 7 UJ I ......................... MO ...... (SKIP TO Q. 28) 17/ Mo 60/ 27. Haa thle atoiuch pain awakened you fron aleep? 2 Tee IB/ Mo What are the naam ot the Medicine* you are taking? (PROBE I WHAT OTHEH87) 1) 6I-6V 2) 64-66X 3) 67-S9/ 20. Have you vonited blood recently? Tea. 19/ No.. 29. Have you recently experienced dark tar colored atoola or bowel •nvenanta? Tea. Ho.. 20/ �-76- Now I would like to ask you mome questions that deal only with the period of tl»e between (DATE OF LAST INTERVIEW) and now. rai ASK A rntouGB o ron EACH CONDITION CODED TES. A IIMK OF IASI • B»t.OMI IDAIt Of lltat l> tto lull IASI INKNVItH) tot ••>• »M titXttt ot MB, I* mat wxitk ••• y«*r «I4 • doctor tlr.t *•)! To. ttet rou tad <OJNDIIIUH)I > • doctor tol« (m ttat »ou ta< ... tt» ooclor «ho •»«• tt* «l*9no>l« or tta «.dle.l t.clllt, .l»r» t L_LJ LJU SKIP n PLEASE GO OH TO NEXT PAGE at City H«r»l« treklMl i t n/ SKIP nnciro 10 l_U * rir»t J_L_I Clf» It.t. �DECK 4O C 0 I Ara you currant ly taking any araccrlbad n»dlclaa» tor your ICOMDITIOHIt Oo you hava ICOHOIIIOH) oo.I Ta> 1 Ho Ta> Ho I 3V 1 2 SKIP TO F •kat ara tka aanaa ot aadlclnaa you ara taklaat tay otkaraf Nkan did you last concult a doctor lor (CONDIIION) batwaaa IIMTC OF 1 Alt INTCRV|[H) 0*4 ^o^f •> 1 1 11 1 1 SKIP TO F 3V t f Ha 21 31 Tr i Sine* tOA't Of l*SI M>at It tka full *a»a and addraM ot tka doctor or a*dlcal laclllty ««« |Bst conftaltad toM »<X1 ttot you ta4 .. 32. taamlat latt Mm Htot l> tte lull Mn* oiitf cd«r«ss of tte doctor •!» ~*« tlu) dlognosl* or tta Mdlol luclllry •teru tta *l I 4*/ I w H/ i sMI/ bt^on (DATE OF IASI INUmitV! «n« •c«. !• Btot «onth ••« y«ar «!• • doctor tlrtt toll you that you •»« CCONOIIIOHH iMnmitm «•• • «octor '^ 2 6V SKIP TO l_J_| LLJ M> Ir II/ »I-70/ Q.33 Flr«t Na-a 4£4*y OR Facility Hana Facility K«~ Straat Mara» Straat AddraM 1 City 1 I W 1 2 »!/ VIP TO F •1 1 1 II City 33. 1 1 A towt condltlo*! 2 SKIP TO F LJU 1 1 Stata SKIP (SPfCIFyi 21 ii-if/ 31 _ IV 10 0.34 «/ L_U N> Stat. LU IT 14-TI/ Flnt *wa H-aOV OR Facility Facility Ha»a Struot Addr*» Straat Mdraai 1 City 1 1 Stata J_LJ City �PFCIH nrx'K 41 MaiNOCCK 41 f Slaca (OAR OF LAS I INttmitX) kat a doctor tola you tfcat you ha« ... Do you hay* Ara you currant ly lading any arascrlbaa* (CONDIIION) aaalclnai for •Hat ara tka (CtfOIIICH) «<nt your (caomomi •aM si n4lclM« yew «r» Olilout k»«»M« IDATE Of lASf INKIt- lh« •actor or Mdlc*! iKlllty •tat oth»r>t »li«» w4 «o«l totum* IODK OF LAS1 IMICmlEK) >itd MM M* aaara» ol you l«t coo»»lt»< T.I No No 2 tO/ SKIP I 2 tut consult • voclor lor Wwt l» t»a lull •au. In >tat aoatk •»4 yuar ultf • uoctor tint tail you ttet you »M (COWIIKMII no«l» M» 34. 11/ Da, aa.larga« llnrl I TO F F 2 4J/ LL' LL' »r 4J-46/ SKIP 10 SKIP IO Hlu>t U thu lull •WMJ and •aorusf ol tta doctor ubo >m<le tl* diagnosis or tM «.0lc>l Uclllty • h»r« tto 4J/ O.M JI-JV Facility FKltlty I 01 ItrMt «»•»» -LLI U I 2 M/ SKIP TO F 1 2 IV SKIP TO F ,5«/ • LLJ LJLJ fc »r l«t 41/ 3?-4O/ St.t. City Cttr * w SKIP TO 9.1* LLI M> LLI «r 4H-JI/ Flrat I Facility Facility I Straat Adarasi Itraat M«ra» L.LJ City Stata .LLI City Stata �DECK 42 Do you kav* (CONDITION) Ar» you currently t.kliH) uny proscribed •uXIctiwi lor •out your ICONDIIIONIt to* Ho T.» Z M/ SKIP I Ho 2 3V SKIP TO 10 F F MMM did you iMt consult Sl.o. «MH Of l»S I d doctor lor Mwt or. IK. MM* ol •*4lclnu> you dr« tdklnit ICOKDIIION) B>tx« IOAH Of LAS I IM limit*) ..d now. In »h»t wMlh nnuu d«d •ddrots ol tatuuM (DATE OF lASf IMTER- Any otk»r*l Ik* doctor or •»dlc*l Uclllty you ld»t c •Ixt It Ite lull •»• y*v 4I« • doctor tlr«t toll you tk»t IHURVUHI kit • doctor tol« you that you KM ... Mwt I. tkn lull Ih. dl»gi>o>U or tl» ••dlul facility uh>r» tto dl««- you kM lOUNOiriunit ••«• MM] •dorusf ol tta doctor .to Mdu •osl* u»l pdd*? L_U •to Utt I Ir SKIP (to ID •!-*•/ Tr I*- It/ First I FKlllty I Fdclllty w Slrd*t Addroi Str..t Addrdts _LU 41 I JO/ SKIP TO F I 2 7I/ SKIP TO F Stdt* City LJU LJ_J »r City If. L_U l_U OrrkMll «l tk» llv«rt l..t SKIP 4-SIH-O-JIJ-l 10-IV n> Mo IT L»t n-n/ Flr.t I FKlllty I FdClllty ' Str«*t AddrMl Strxt Addruti _LU City J_LJ City Stutu �-84- Do you heve (CONDITION) eon! Mkee 414 yo* let! CeeMlt • 4octer lor Are you currently teklng eny ere»erlbe4 W>et ere tee Ho keteeee UMTf OF LAST INTEP> film eee «eet 2 26/ II 2 2V SKIP TO T». I btmii (OAK OF l«T IHIUWKm •«« K», I* «l»t •oom iw«l» ••• e***t 1 '»/ SKIP F 52-M/ tli* «l«9m>»l* or tto M4lc*l Uclllty •tor* tte *l*g- »o« lut co«>.lt*4 TO F •tat It tta lull •••• «B4 Mlar*»s ol <!• «octor »ho «M* e»4 y*er 414 • doctor flrtt t*ll you tt»t yea k>« (OWCIIIOMU rM> at LL.ILU ee4lcliiet lor yoer (CONDITION)! No Wxt U tlw t»ll (CONDITION) •em el e*4leleet yen ere teMefi Met etkeret Since (OAK UF IASI tHIunitli) h» e eoctor tele yen thet you he4 •• L_l_l LU He TO/ Ji-41/ Vr «0-*)/ O.S» Ftclllty N*« FKlllty I D3 Strwt M*-*u LJU I I 1 2 4V 1 2 «4/ SKIP TO F TO F I I SKIP City S»»*« Clff If. tell kIMMr ereblcuT 1 1 11 1 1 t..t H ~ M/ I 2 »V SKIP N> | | « TO 21 9.4* Jl I I I ( St.t. »»-*»/ Flr.t Hene FMlllty I Feclllty I Street Mereu Street A«er«i» City City _L_U �MM.KS •»4lcllM» lor your (CONOIIIOHX lM»t y*» mm 4I4 yo* Ho To» I Tl/ SKIP 10 No Wwt «•• tko ••Ml ol •*4lcl* you at* t*hl»«l «k*t otkM-ll F F iHHmum k» • «octor «kal l» tk* lull KONDITIOW MM «M •»*•>• of kotMM (DATE OF tAJI INTiMtlt»> M4 »o«» tk* •actor or •.4lol locltlty yo* lot comilt.4 LULU II SKIP TO SUo* IOAK OF I.HSI lut convlt • doctor for fr« you currootly taking My proscrlbof Do you bavo (CONDITION) 42-4) Ho » Tr AW» •!!« •«<»•••» of tl» ooctor >ho ••«• •no y«or 414 • 4octor tint toll you tftot tta <l«gno>l> or tM (Xim iKlllty you KM ICDMOIIIUMIl •tor* tr» «logM»|« ••• **4of A* otter (SHCIFyi • Itloul I 2 JV SKIP L_U L_U Ho Tr J6-H/ 13-ltV H«IN HOt 41 •l»t )• tM.lull M». Ill >k>t •ooth toU you tk>t you ta4 .. 40. la»t I ••turn* IDAIE Of 0.41 Flr.t I Facility Facility I T ttroot MOTIM Straot Mdro«» LLI LJU Cl ty City 2 It/ SKIP TO F I 2 IV LU LU II SKIP r ». 31 •• 23-»V 1*1 » Tr SKIP l*-32/ ID «.42 «•/ SI at* LJU LJU Ho Tr 42-4V Flr.t I Ficlllty Facility I Stroot M*r*l> -LU City Stoto City �BET.1H DECK Oo you navo I CONDI I ION) •ooJclm lor MOM o ooctor for 1COHDITIOH) •hot oro tko No 2 4I/ SKIP Of o«4lcl«0« Wwl It tko (oil »0t>000 IDATf O_ 1 TO F OF IAST INK* ! • • 9*4 ••*!••*• Ol !» L_U L.LJ I NOIIT tM ooctor •!» »o> tl« dlognoslt or tho •»•!€•! ffoclllty >N>ro tta AOffll ••• 42. tv X roiolrotory contltloa •tkor tkoo omi«io«l»l I 21 LLI 2 ID/ SKIP TO 1r 12-IV ».43 11/ 31 Ntat It tta lull A*M OA4 adtfross ol M0ve III »Mt nonth •M yo*r «l« • •octur Mrit toll you ttot you kM lOOHDIllUHII tko toctor or M4lcol Uclllty yov lost con«olto4 you wo tlklngt Wwt otkorol Voi euwnn (OAK Of LASI IHItKVIEKt •nd Sl»o> tDAH OF L«SI INHUXEII) MI • Doctor toN you thot you Iwd .. you yo«r (CGN0ltlON>t «o«J r«« Aro you corronfly toking ofty •roscrlboo' Flr.t Foclllty I Stroot Moro» Stroot MJron Stoto «J. My ofkor oojor coXltlont 1 1 2 oV SKIP TO f 1 1 «4/ SKIP TO F II f 1 II 1 1 31 M-tOf t SKIP re/ " St.t. City U I City " "' . K> 9.44 ^ U/ L,L_! L.L.I to If lf-22/ '•/ Flr.t Mow Foci 11 ty No»o FocllKy I Stroot Moroti Stroot M«ro» _LLJ City 44 City �0 Ar« you cur- •ton did you iMt comult • doctor lor •Mt It tb> lull Do you h*v«j rently taking t>ny pr«»erlb*i*} (CONDITION) (KdlclnM lor Vh«t «•• ti* your ICOMOIIION)! »«M* Ol M«JlcU Utnan «MIE W I AST INTtKmn> Md aoul •to., during «l»l Bonth tM doctor or •»dlc<l (Klllty »«• «nd M«r«ti ol tto doctor Mteo »«d« !<• dl»g«o>l> or •nd f««r did • doctor Mr it toll you «r»t CCONOIIIOm •onf tn> Mdlcal laclllty •fwro th» dlAgnosli you l«*t consHlto you *r« takliif t Wwt oth.»r«? «Mt l« tto (.11 TOO tod tOOMOIIIUNII 44. 2 24/ I i »/ SKIP SKIP 10 TO LU LU I) tr I. . .. »*»• you «.«r U toor •Mir. IM. kton tr..t.d f«r • »»«t«l or Motional I 2 SKIP ID I I I «—«—' "o 21 r ririt i SI facility FKlllty I Cd Str..t Addr*si Strxt Mdr«» _L_U CMy 2 40/ SKIP TO F 1 2 41/ LULU O »ir r ». SI Ha JV ir 0-4V flr.t Facility itoM StrMt _LLJ City J_U City �PFIGIN I1PCK 45 45. At any t(«e since (DATE OF LAST INTERVIEW) has a doctor told you that you had cancer? Ye* 1 IO/ 46. Did the itoclor tell you that thia wai a akin cancer or a ayatenic (body) cancer? No...(SKIP TO p.49) AT* you cor- Mm) 414 you last con»lt • 4octor lor Do you h«v* •CONDITION! r wit I f taking any pr««crlb«4 M4lclM< lor Kb** W« lOnOIHOW Mat l> tl» l»l I ••» m4 •44r«« of iN»I your (CONDITION)! • •••» of »«4lcla*s kolDMll (OAlt Of IkSI INliM- tin *oclor or a^lcol (acllltf TOW *re T«» No 2 6V SKIP TO F r«t I No 2 «V 91 IP llfcet OthMTSl < «Mt Skin cancer only..................................1 Syatemtc cancer only (SKIP TO p.48) BOTH SKIN MO SYSTEMIC CANCER ymi l«»t co«»lt«4 It/ 2 3 6KIH CANCER ONLY LULU I) 2 U.H 47. 2) rlcaae look at this chart and tall •» utiara each of your akin cancer* lla/waa) located. tt-W 5) INTERVIEWER! flrtt ! INDICATE THE AHATOttlCM. COOK FOR EACH SITE BEING REPORTED. SITE NUMBER RITE CODE 73-TT/ *-00 COOESi llr**t AMrm LU *OI) (14) 4O2) eye Lid (01) Ear Scalp or Forehead Aral or Hand, Not Otherwise Specified (15) Genitals 104) (05) (ICI (17) Le<j Foot (18) 1*9 or root. Not Otherwise Specified Skin. Not Otherwise Specified Upperlip, Not Otherwise Specified Lowerllp, Nat Otherwise Specified tip. Not Otherwise Specified Nose Head or Neck, Not Otherwise Specified (06) Cheek, chin or Jaw (07) M*dt or Supraclavlcular (08) Veraillion (09) Trunk, front (10) Trunk, Back (til Trunk, Not Otherwise Specified (12) Arm (11) Hand (19) (2O> (21) (22) �INTERVIEHERl DECKS 45-46 DECK 45 -93FOR EACH SITE REPORTED ASK A THROUGH F SKIM CANCER ONLY 47. SITE 1 A. SITE 2 In «hat Month and year wa* cancar or the (SITE) Heat dlaqnoaed? 1 1 I 1 1 A. In vhat aonth and year vaa cancer of the (SITE) flret dlagnoeed? A. L_l 1 L_LJ •*•»/ Ho. Tr. I ie-2./ D. In nhat Month and year nee cancer of the (SITE! firat diaonoeed? L_L_I L_LJ »>-»/ SITE 3 SITE 2 SITE I SITE 3 D. What treatMente or Medicine* (do/did) you take for cancer of the(SITE>7 CODE ALL THAT APPLY What treatments or Medlclnea (do/did) you take for cancer of thetSITEl? CODE ALL THAT APPLY Hhat treauwnta or Medlclnea (do/did) you take for cancer of the)SITE)7 CODE ALL THAT APPLY Tr. HAND CARD MAUD CARD B. Miat kind of akin cancer vaa thia7 Tr. HAND CARD » P B. Ho. Hhat kind of akin cancer vae thie7 B. 36/ Radiation 1 S6/ Radiation 2 37/ CheMOtherapy 2 57/ CheMOtherapy*....2 77/ 3 7B/ 79/ Surgery 3 Other.(SPECIFY)..4 Hhat kind of akin cancer Man thlal Baaal call SquaMoue cell SqiunotM cell 1 38/ Surgery.... X SB/ Surgery 39/ Other.(SPBCirr)..4 S9/ 76/ Other.(SPECIFY)..4 Baael cell SquaMoua cell Cancer Mtaetatlc to the akin 4 Cancar Metaatatlc to the akin 4 During vhat Month and year did you firat receive (EACH nCATHEHT COOED IN O) for cancer of the (SITE)7 Cancar Mataatatlc to the akin 4 C. what ! the full • naeie and addreaa of the doctor or the Medical facility where the dlegnoeie vae Made? Laat Maaw (that la the toll naM and addreaa of the doctor or live Medical facility vhere the dlaanoeia •aa Made? Uat Nee* IV rlrat Name C. "hat la the fall nee* and addreaa of the doctor or the Medical facility vhere the dlagnoeie vaa Made* nrat MM HO ULI YR 40-41/ L_L_I L.I-1 HO YR Facility HUM Facility NaM Street Addreaa Street Addraaa YR 48-S1/ OTHER L.L-1 HO L-L-I YR S2-SS/ ir SECOND SITE CODED IN Q.47 CO TO SITE 2 CUT city 1 1 1 1 1 State * 1 1 State * 1 1 State * HO L.L-1 L_ I-1 MO YR 60-63/ L_L_I L_l_l HO YR YR 10-13/ CHEMOTHERAPY | SURGERY | |_ | HO |_ L I YR I4-I7/ | | «4-67/ SURGERY |_LJ HO Street Addreaa City RADIA- 44-47/ OR raellltr Naaie E. O BEGIN DECK 46 During vhat Month and year did you f i r a t receive (EACH TREATMENT CODED IN D) for cancar of the (SITE)? TION Firat MM OR During vhat Month and year did you firat receive (EACH TREATMENT CODED IN O) for cancer of the (SITEI7 RADIA- I V OR t. RADIA- teat NaM W/ NODE NONE. HONE.. w C. t r Basal cell v£> Radiation CheMotherapy Ho. ir Q.46 CODED •!-, SKIP TO Q.49 L.L-1 YR •-I •7/ ««» LLI L-LJ HO YR | HO IF Q.46 CODED "1", SKIP TO 0.49 | 1S-2I/ OTHER L.LJ HO L.L-1 YR 2I-2S/ 72-7S/ IF THIRD SITE CODED IN Q.47 GO TO SITE 3 | YR IF Q.46 CODED SKIP TO Q.49 �DECK DEIK.S 46-47 47 48. (Continued) SYSTEMIC (HOOT) CANCER ONLY BODY PART I A. In what part of your body (is/was) cancer loca ted (RECORD VERBATIM) 26-4S/ BODY FART 2 in what part at your body (is/was) career located? 47 A* (RECORD VERBATIM) S1-70/ In what part of your body (Is/was) cancer located? (RECORD VERBATIM) IO-29/ Last Name Last Mas* Last "we B. OR Oil Oil B. Hhat king of cancer First Nane First first Nan* 4V 40/ 35/ Hhat kind of cancer was it? Hhat is the full na»e and address of the doctor or the •edlcal facility where the diagnosis was nade? Hhat is the full name and address of the doctor or the • •edlcal facility where the diagnosis was nade? D. What U the full name and address of the doctor or the •edtcsl facility where the diagnosis was *ada? BOOT PART 3 BODY PART 3 BODY PART 2 BODY PART I BEGIN DECK Hhat kind of cancer facility Facility Nane Street Address Street Address City •as It? Facility Ms Street Address was it? City City 30/ W C. In what Konth and year wac cancer of the (BODY PART) first dlaqnoaed? U_l LJU «>-*»/ Mr. Yr. In what Booth and year waa cancer of the (BOOT PART) first diagnosed? _l U_J I I_J LJ_I State In what awnth and year was cancer of the I BODY PART) first diagnosed? Hhat treatments or •edlclnes (do/did) you take for cancer of the (BODY PART)? CODE ALL THAT APPLY what treatments or •edlclnes (do/did) you take for cancer of the (BODY PART)? CODE ALL THAT APPLY Hhat treatments or medicines (do/did) you take for cancer of the (BODY PART)? CODE MX THAT APPLY LJU U_l Mo. Tr. State Radiation HAND CARD I 36/ Radiation t 4t/ Radiation I 46/ Chemotherapy 2 37/ Cbenother spy 2 42/ Chemotherapy 2 47/ Surgery 3 38/ Surgery 3 4V Surgery 3 48/ Other.(SPECIFY)..4 39, Other.(SPECIFY)..4 44/ Other.(SPECIFY)..4 49/ HOHE.. �4 . (Continued) 8 49. F. During what nonth and year did you first receive (EACH TREATMENT CODEO IN E) for cancer of the (BODY PART)? F. During what »onth and year did you flrat receive (EACH TREATMENT CODED IN E) for cancer of the (BODY PART)T RADIA- RAD1A- TIOH BODY PART 3 BODY PART 2 BODY PART 1 |_LJ I_U HO YR TION LJU I_LJ HO HO YR During what awnth and year did you flrat receive (EACH TREATMENT CODED IN E) for cancer of the (BODY PART)? RADIATION TR 47-70/ 50-53/ CHEMOTHERAPY LJLJ l_l_l F. CHEHO- ™ERA« LJ_J m 54-S7/ U_l Tt 7 i-;*/ |_J_J HO CHEHOT««APY LJ-J LJU 1R I5-I8/ M any lioe ainr* (DATP OF LAST INTERVIEW) has a doctor lolit you that you had leukevia? Yea CASK A-F) 1 32/ No. A. .(CO TO P.SU 2 Thinkinq about the period between (DATE OF (.AST INTERVIEW) and now, in what wonth and year was your leukeaia diaqnoaed? 33-S6/ LJU LJU Ho. Yr. B. Nhat la the name and addreaa of the doctor or the oedlcal facility where the diaqnoaia waa nade? LJU HO YR I9-22/ 37/ SURGERY LJU HO LJU- SURGERY LJU HO YR 58-6 1/ OTHE« LJU' LJU HO YR OTHER 62-6S/ C. IS THERE ANOTHER BODY PART AFFECTED? Yea.. (CO TO 48A-Body Part 2 . . l ).. LJLJ SURGERY LJU HO M 7S-7«/ BEGIN DECK 48 L J J U_l HO YR 10-IJ/ LJU YR rirat Haaie 2J-26/ OTHER LJ_J LJU HO YR 27-30/ Facility C. IS THERE ANOTHER BODY FART AFFECTED! Yea. . G TO (O 48A-Body Part 3 . . l ).. G. IS THERE ANOTHER BODY PART AFFECTED! Yea..(GO TO NEW QUO) street Mdreaa I I City No. .(SKIP TO Q 4q\ No..(SKIP TO Q.4f>... H . (SKIP TO o. } 66/ I4/ .2 Jl/ C. Mhat treatjnenta or Mediclnea have you taken for (DATE OF LAST INTERVIEW)? I) 21 I State leukevia aince 38-40/ 41-43/ 1) D. For the period between (DATE OF LAST INTERVIEW) and now, during what Booth and year did you flrat receive (EACH TREATMENT OR MEDICINE IN O? HJ. Yr. TREATMENT I TREATMENT 2 TREATMENT 3 LJU LJU LJU LJU l_l_l LJU 47-SO/ 51-54/ SS-58/ �INKKIIEMR: FOH IACH KS, ASK A IHUUGHB: A. 49. (Continued) ASK C-C FOR €ACH "IIS" Al 8. CARD E. What is Ihe none and nddreaa of the ftocLor or Medical facility you last consulted about your leukemia? J 90. 59/ Sl.c. WAIF. Of LAS1 IHUH»lfH) 6 , ,o. .. On .tat fffl at your body did ro. tev. (COHDIIIONII Any otter MTtl Laat Hade c. B. Did ,00 4llo>» ICOWITIQM) .It* • doctorl o. Mh*t ».s th. dl«gnosl«l Khar Is Ite docl-*•«* tot .*»..»! First I. Patctoi of row tkl. LJ_I Sit* CM* M-64/ raciittr I 2 M/ Flr«t N*«* Sit* Co*. LJ_J Street Mdreae Sit* Co*. F.clllf, A**r*ii 6».)0/ I J_J_J rttr F. w <J1 KJ Clt, » | I St.I. State During what Month and year did you laat conault (HJME IN El? I_U I_U Ho. Tr. 2. i*>l*r kTMlna el tk> (III* tkMi m**ll | || I 2 IV Sit* COM >)-?•/ i J io/ 11/ Sit. Co*. fsell It, Addr.o sit* CM* >*-ao/ at, Ist.t, I i CO fO n.M COHIIHUCU OODfSi (Oil Sc*l> ar For* I»M (»i« Er> tu l«>l CM(041 NOH <05I I*M or Mick, Hot OtterelM S»*el tl*« 10*1 Clmk. CM* or Jw <•!} Ntcft «r SDCTKlnloilcr 114) (13) (1*1 (III (Itl Off 174) 10*1 »mk. Fro«t 4101 **»k. IK* (2I> (III Oil (211 »»iMi. Hot Utter. IH SpoclllM »r- *-. or Mod, Mot Ottor.U. Steeltl*4 Qiill.ll l*« foot 1*9 or Foot, Hot Ottor.lt* S»*cltl*d fkl*. Not Utter.It* Sp.clll*4 UMorli*. Not Otter.lt. S»*clllM toiwll*. Ho* Otter.lM S»*cMU« llr. No' Otter.I w S»«clll.4 �99. ASK C-6 FUN EACH ICO.tl.Wdl •H5" AI 8. (WO CM» F. i . J Durl*f mint font* and y**r ••• tk* MM ***• «l tk* doctor tfc*t? (Mr I** (tot «o»tk ••*• y*or 414 *oa lo»t coiiMlt JO. Sl.c. WAR OF l«SI B. Ck> .k»t r*rt ol yoir ta» rou Ul« (On <H»cu»» txxV •!« you HIM you lost coa»vlt*d *ko»t (COMMIIOHIf tdMHIIUNI Nl»t ••» tta ICUMOIIIONIT otter »«rtt CH lit or aMIc*! facility .11 » • doctor? dl«gno>l«t >tet >•» L U L U Ha. LULU •>-,./ Tr. rir.t I •te. J. No Ski* tkot *M •«tr* Moiltlm | or too*** to tr. MM *iur*9il Sit* Cod* 1 da/ 1 Jl-JV kwrt for ** room! I I 3I/ | | | Sit* Cod* 14-1V F*ellltv I Addr*» Sit* Cod* J4-JI/ I Strwt «ddr*M City I I St.!. Clt, CUOiS FOR ».> LU Ha. LU »r. GO TO 0-50 COHIINUlO - LLI LU 22-2V flrtf i •a. Tr. COOtS t Soil or ror*k*M Ar» or Kn«. Hat OthM-.U* Onltcli Er» lid 10)1 Facility I 1141 MM OH 1011 IU» 1041 S»*cllU« UC MOM City t»»d tr 104) CIMk, cM« or J*. Faat U or Faat. Hot Otter. I >. | 1011 Mick or S**r*cl*vl«l«r llfl Skin. Hat Otteolw &p«cltl*d 101) IOt» Mr*4lla« k-Mik. fr*«t 1201 (JOMrllr. Hit otlw>l»* (101 »Mik. B*ck 121) low ll». Hat Otk>r>lu lilt »Mk. Hat Otter«lM Sp*clll*d 111! AT. II M MM H*ck. Mat Otlwuli* t»f III! (III Str**t Mdr**> 10)1 if*clll*d S**cltl*« III, Hat Otk*r«lw Sa*cltl*d �DECK 49 10. ICoo»ll»4l ASK A THROUGH J FOR iACH »IS F. I. *• Owing «tof iwM •tat yW Ml tMIt It tlM ••»• M4 III A>l«* fro* Injury. • Inc* tOHt OF IASI *••« »o« »" LU LU LL . J L U Na. Mo. *r. B. 10. During <HMt •44r*» of tlw doctor or n»4lc*l facility yo> lot eonult*4 •bout ICOMMTIONIt IT. Thinning about tko »•'•«"' «•»••« (OAli OF IAST IHIeWICKI nn4 no.. .k*n 414 >.. 4. LU LU I 2 »/ Ho. »r. JI-JV Fcclllty I M4TMI Clt» J_U In an, of Ooyouitlll h . (OMIITIDN) .. rM Mrtt notice (COOIIIONII *• * °; llblch Il«b> or matin w. •t«*cl*4l C. LU LU No. Tr. •*.«,*/ No �DF.I'KS 49-SI) JO. V). (Co«<lMw4) J. H. Fro» (DATE OF IASI Dlo you »oo • doctor lor wring >k>t Moth! ICONDI IIOHIt at Ml tko «l«g*o>l«t Owrlnij vh«t «onth and During >Mt ml y««r 4ld you l*tt consult •<lik-«» ol IK. doctor | T M M l •fto *mt* t •rid yo«rl»l us tk» (OOMDITIOH) -o»t _ Mo. Mol It th* MM m* ••*•«» al tko doct (NAME FROM III or th»BM-|c*l you Islt con*ult«tf •bout (COtOIIIOMK _rr, L U l_U KIP Ho. Flr.t H«U TO L.L-1 Vr. Jl/ •—'—' '—!—' 22-2V Ho. 1r. >€/ ,,-20 I—I—' I—I—' Mo. Tr. Vr. on ~U7 Fcclllty I Stroot M4r«< Street AMr«M U1 LLJ LLI "-•* Mo. Yr. TO K6IN OtOC M LULU "•» Mo. Yr. I 2 M/ _l L_U Kir Mo. »r. •"• FKlllly I FociIIty I ttroot M«r*» City J L L I City L U $t.t. JI-34/ �ASK A THROUGH I FOR EACH IE S A. e. Thinking .bout tk* period Utx»» (DATE SI. All** lro» Injury. >lK* (DATE or IAST IHIEIWIEMI tov* you "Mel. I Irti or *»>cl*s Mr* C. Do rou »tlll h*<* (COtOIIIOH) OF tASI yo« Mr»t nolle* <COMDI1KM>t I. P*r*l(l«it I_U LU •«n*nt lone In Mr el 2 IV Mo. Tr. J*-JV PLEASE GO OH TO HEXT PAGE 2. r*r«l.«»nt KkM •»« L.U LU I 2 «J/ «o. »r. 44-4W * 2 »0/ �DECKS B.IM*. (OAIE Of OK f ou M* • I I mtlmrlK) ••. U doctor lor »Mr*» ol t»» doctor •». «rlnf »t»t (OMJIIION)! »hoa»4» tto <!•(•<»I• Ma* *•• tta «l«jil»lU Mvt U •*• ••-• «H4 Dvrl*« -Ml ««>ll> •»« •o-tte ••* ,»«rl»l or tto •»«tc*l •MTMi ol tlx doctor y.«r 4U ro. l.tt co»».l> n tto (OmOIIION) l Ixllltf «h^. tto or M4lc»l (.cllllr INAK IKOM III you l«tt con«ult*4 .bout (CONDI IIOHIt IfOIHOCCK 51 L_U L_U Ho. Ir. 2 »/ SKIP IO U.t I 1 / 1 2 •to. TKT •I/ LL.I LULU __ Ir, Facility JU at, 17* City JLJU Hot* U l LJLJ L_L_I No. «r. 1 Ult I It/ II/ SKIP m L_LJ L.L.I •*-*" •to. II/ iMt I LULU LL« LLI •*.••/ Hr»t MM* Oil Ho. Ho. »r. >r. fed HIT J_U it*t« "Tut* Ir. SO-51 �51 109 A ASK * ItKOUOH ) fOH EACH US c. A. Thinking •tout IK *»!«• Iroa Injury, Kblck I!•*• or Do roll Mill r«rlod «>el« "»r« k»«« ICOOIIIOHI b«l>««« (DAIt Of l*SI IMIfmifNI •II*C««4I • IK* IDAIC Of IASI (on llr>) aotlc* (COMH1IOHII 2t-lt/ i LJ_J a xv Ho. PLEASE GO OH TO NEXT JACK J> 30 I Ir. 2 JO/ �-112- 92. DECK (Co.Ilnu.d) <Contlwi«dl H. fro. IDAIE Of 1*51 INTERVIEW md IK». during >k»t mtln •nd y.w(tl IMS Ik. (CONDI TION) noif l_LJ LU "-'^ 10 OH yo» » . . . doctor lor IOONDIIIOHII • t M« tk. dl«gnotl>t Mwl l> Ik. *tm» t*4 .ddr.il of tk* doctor •to *•** tk. dlogooiU or tk. wdlc.l laelllty .fcw. tk. 2 '. SKIP DurlM| .k.t Kntk wd >rar M> tk.II or Mdlc.l iKllltr you l.st con»«lt.d •bout ICONOIIIOHIt LL_l LU TO •hoi !• tk. nww >nd •ddr.n ol tk. docto „-«»/ y*»r did you l.tt consult •NAME FROM lit LU LLI 4,-,o/ •to. O.SJ l_U LU ricllltr w Durl«4 .k.t vontk and Str*.t JtMrxi cm -ww «•••«* ***-•« c»» -V.7.' *r. SI �DBUK SI 53. SECTION 81 (Maiden the prescribed medicines you 'told me about) are you currently taking any (other) prescribed medicines? Tea (ASK ' ft) S1 OECK 51 OFC:4«2BSec-fl HFALTH HABITS / The next net of questions refers to smoking habits* No...(SKIP TO SECTION 6 ) . . . . 2 I. A. For what conditions are the medicines? 1) 2) 3) What other condition*? ...__, . Have you ever smoked at least as many as 5 packs of cigarettes, that ls f IOO cigarettes, durlnq your entire life? . I Tea «/ No . (SKIP TO p.22) 5 J/ 2 _. 2. Do you now smoke cigarettes? 1 Yes No (SKIP TO p. 11) 54/ 2 CURRENT CIGARETTE SMOKE* SECTION 3. o^ o On average, how many cigarettes do you smoke a day? IHTEMVIENEMi ir « MiSHERS BT GIVINC NUHBEft OF PACKS OF CIGARETTES, RECORD VERBATIM. THEN MULTIPLY THE NtMBKIt OF PACKS BY 20 AND EHTEK THE M«BER OF CIGARETTES SMOKED. ENTER NUMBER OF CIGARETTES PER DAY I (IF HOT EVERY DATi) 11 | | | | PER MONTH | 55-5fi/ 57-SB/ OR (IF NOT EVERT DAYi ) 11 I j PER YEAR 59- 60/ �DECKS si-52 4. For how Many years have you been Moklnq (NUMBER IN p.3) (clqaretlea per day/per Month/per year)? Less than 2 yeara O1 2-5 yeara 7. 61-62/ Hhat type of ciqarettea are they? Are they . . . (READ EACH PAIR TOGETHER) CODE OHE A. Filter Lip or 1 O2 Non-filter tip?. CODE ONE 6-tO yeara*................02 II-1S year* O4 21-25 B. 03 16-2O year* Reqular size '. 1 7S/ OS yeara Kinq size o r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 100 H i l l i M t e r ? 8. 26-3O yeara e .•,..,,.,..... .06 JI-3S year* OS 36-40 year* O9 Nor* than 4O year* ..IO 3 Now I an qoinq to ehow you a dlaqran of d i f f e r e n t aize ciqarettea. Meaae look at the picture of the (KINO OF CIGARETTE IN p. 7A AND O.7B). Now, considering your atyle of snokinq--for exawple, how lonq you usually leave the ciqarette In an ashtray or just hold it in your hand — tell He the nmber which indicates now Much of the ciqarette you actually sanke. Section 1 ................... 1 76/ Section 2 5. 74/ Mhat brand of ciqarettea do you usually anok*T (IP HOUR THAN OHC BRNO OR NO REGULAR BRAND MENTIONED ASd Milch on* do you 'moke the Boat?) ................... 2 Section 3 ................... 3 Section 4 ................... 4 orrjce use ENTER BRAND 9e Mo reqular brand. .996 Durinq the period when you were Mokinit the nost heavily on a reqular basis* about how Many ciqarettes did you usually saoke in a day? ENTER NUMBER 1 6. For how lonq now have, yon been aanklna this particular brand? ENTER OATSi LJLJ (IF NOT EVERT DAT.) | I 70-71/ OR TEARS) 72-73/ (IF NOT EVERT OATl > 68-69/ OR MONTHSI 77-7H/ 79-80/ OR S6-67/ OR KEEKS t I PER DAY | | PER MONTH | | A. I | PEGIN DECK 52 ( | PER TF.AP 10-1 I/ When was that? FROM I_L_L_LJ Ho 12-1S/ Tr TO LJ_J_LJ 16-19/ �1O. when you saoke cigarettes, how deeply do you usually Inhale? sayi Would you As deeply Into the chest as possible. 13. Haw lonq hirl you been waking (NUMBER IN O. 12) (ciqarclles per day/per week/per Month)? 20/ Less than 2 years. As far back as the throat .Ol 2-5 years Only partly Into the chest .02 6-10 y e a r s . . . . . . . . .03 Hell back Into the mouth, or 4 11-15 years .04 Just puff and don't really draw It In at all 5 16-20 years OS DON'T UKM • 21-25 years O6 26-30 years O7 31-35 years OB 16-40 years O9 SKIP TO Q.22 FORMER CIGARETTE SMOKER SECTION II. How long haa It been since you smoked cigarettes fairly regularly (RECORD NUMBER) More than 4O y e a r s . . . . . . . . . I O ENTER DAYSt | |J _ OR WEEKS t || | 2J-24/ OR HONTHSi T || | 2S-26/ OR YEARSi || 2I-22/ j NEVER SMOKED REGULARLY. 12. >4. 27-2S/ (SKIP TO Q.22). 29/ ENTER NUMBER OP CIGARETTES PER DATl | l(_|_| PER MONTH ) | JO-31/ 32-3J/ OR (IF NOT EVERY DATl) f| | ,| PER YEAR I Yes IS. (SKIP TO P. 16) 38/ 2 Haw lonq did you stay off cigarettes at that tine? FNTCR DATS. | | 39-40/ | | | 4I-42/ OR HONTHSi I | | 43-44/ OR YEARSi 16. f OR WEEKS I IP R ANSWERS BT GIVING NUMBER OT PACKS OP CIGARETTES, RECORD VERBATIM. THEN MULTIPLY THE NUMBER OP PACKS BT 20 AND ENTER THE NUHBER OP CIGARETTES SMOKED. (IP NOT EVERT OAT.l You Benlioned that you have not sucked regularly for (TIME IN 0 . 1 1 ) . Did you ever stay off cigarettes for • longer period of time? No on the average, about how tuny cigarette* • day were you evoking at that time? INTERVIEWER! 36-37/ | | | 4S-46/ What brand of cigarettes did you usually smoke just before you stopped ssnklng cigarettes regularly? (IF MORE THAN ONE BRAND OR NO REGULAR BRAND MENTIONED, ASKi Which one did you smoke the most? OFFICE USE ENTER BRAND _ - 34-3S/ No regular brand ........................... 996 47-49/ �I > R « K 52 17. 2O. For how lonq did you smoke this particular brand? ENTER DAYSt I I I 50-St/ OR WEEKS 1 | | | 52-S3/ OR MONTHS I | | | S4-55/ OR YEARS 1 | | | During the period when you were smokint the most heavily on a regular basis, about how many cigarettes did you usually smoke in a day? 56- 57/ ENTER NUMBER I (IF NOT EVERY DAYi) | I 61-62/ | | PER DAY | 6J-64/ | | PER HONTH OR UF NOT EVERY DAYi) IB. A. Hhat type of cigarettes wore they? I | Filter tip or 1 Non-filter tip? SB/ 2 L_LJ_LJ Mo • CODE ONR B. Regular size 1 King site or... LJ_L_L_I Mo O UI 19. 7I-74/ Yr 3 21. CO 67-70/ Yr TO S9/ ..2 IOO Millimeter? | PER YEAR FROM CODE OMB A. | When was that? Now I an going to show you • diagram of different sire cigarettes. Please look at the picture of the (KIND or CIGARETTE IN p.ISA AMD O.IBB). How, considering yowr style of smoking—for example, how lonq you usually leave the cigarette In an ashtray or Just hold It In your hand—tell me the Busbar which indicates how such of the cigarette yon actually smoked. Mien you smoked cigarettes, how deeply did you usually inhale? you aayi Would As deeply Into the chest as possible Only partly Into the chest.. As far hack as the throat Hell back Into the mouth, or Section t I Section 2 . . 2 Just puff and don't really draw I t In at 3 DON'T KNOW Section 3 SO/ all. Section 4. CURRENT PIPE SMOKER SECTION 22. Doting your entire l i f e , have you smoked at least as many as SO plpefuls of tobacco? Yes Mo I (SKIP TO p.)S) 2 76/ �5} DECKS 52-53 2). 26. no you now seioke a I Yea (Continued) A. 77/ Mwn was that? FROM No (SKIP TO Q.26I 2 • I—LI—I—I No Yr 24. About how aany average elzed pipeful* of tobacco do you usually e»oke In a day? ENTER NUMBER OF PIPEFULS OF TOBACCO PER DAYi (IF HOT EVERY DAYlJ l| | | FBI MONTH 22-25/ TO LJ_U_I l__l__l 7B-79/ BB3IN DECK 53 tO-11/ Mo 27. OR Z6-29/ Yr Mien you exoke a pipe, how deeply do you usually inhale? Would you aayi As deeply into the cheat aa p o s a i b l e . . . . . . . . . . . . 1 3O/ 12-I3/ (IF NOT EVERY DAYl) Only partly into the cheat 25. Foe how many years have you been Mtokiit? (NUH0BR I* 0.34) lpipeful.1 per Aa far day/per wocvth/pec year)? .....2 back aa the throat 3 Hell back into the anuth, or 4 Juat puff and don't really draw I t in at a l l . . . . 5 DON'T KNOW cri (SKIP TO O.35) ...OJ 2^ R CURD FORMER PIPE SHOKEP SECTION ...OS 2B. Row long haa it NUMBER) ENTER DAYS i OR WEEKS i •tore than 40 year*. .»•.. been aince you emoked a pipe f a i r l y regularly? (RECORD U_l I_U 31-32/ S3-34/ OR MONTHSi 26. ENTER NUMBERI ( I F NOT EVERY D A Y i I I—l__l f | | rB * »¥ NOT EVERY OAYl) (SKIP TO 0.35). 39/ .1 I8-I9/ 29. OR (IF 37-3B/ NEVER SMOKED REGULARLY D | PER MONTH 35-36/ OR YEARS i During the period whan yon were aaoklng the mat heavily, about how •any pipafule of tobacco did you usually evoke in a day? 2O-2I/ On the average, about how «iany pipefula of aanklna at that tla«7 tobacco a day were you ENTER NUMBER OF PIPEFULS OF TOBACCO PER DAYi (IF NOT EVERY D A Y l J l| ( I F NOT EVER! O A Y l ) l| | | PER MONTH | | PER YEAR | | | 4O-41/ 42-43/ OR 44-4S/ �-12330. DECK SI DECT 53 14. For how Ions did you smoke (NUMBER IN p.29$ (plpetuls of tobacco per day/per week/per month)? , Ixtss than 2 years Ol 2-5 years 4S-47/ Mien you s«oked a pipe, how deeply did you usually Inhale? say. Mould you As deeply Into the chest as possible. O2 7I/ Only partly Into the chest 6-IO years.................03 HAHD CARD • T As far back as the throat 11-15 yesrs O4 16-2O yesrs OS Juat puff and don't really draw It In at all....S 21-25 years 06 DON'T KNOW 26-3O years O7 31-3S years O8 3»-4O years..... O9 4O or more years..... It. 12. IS. to 1 Yes (SKIP TO 0-311 | | | During your entire life have you smoked at least as »any as SO cigars? T 4B/ 3«. 2 > No....(SKIP TO O..S1) Do you now smoke clgsre? No ' 72/ 2 73/ TSS Haw long did you not smoke a pipe at that time? ENTER DATS t 8 CURRENT CIGAR SMOKER SECTION You mentioned that you have not smoked regularly for (TIME IN Q.3O). Old you ever not smoke a pipe for a longer period of time? Ho Hell back Into the muth, or (8K1F TO Q.42). 49-50/ OR WEEKS I .L_I_J SI-SV «« MONTHS. L_LJ S3-54/ OR YEARS I L_I_J SS-56/ 37. On average, about bow many cigars e day do you now smoke? ENTER NUMBER OF CIGARS PER DAIl ( I F NOT EVERT D»Tl » | | | »H_I *** MONTH 74-7S/ 76-77/ OR 3). During the period when you Mere smoking the most heavily on basis, about hew many plpefals of tobacco did you usually day? pl * °*T ENTER NUMBER I L_LJ (IT NOT EVERT DAYl ) •(_ |_ | PC* HONTH regular .• In a S7-58/ 59-60/ OR (IF HOT EVERT DAIl) K. t|_l_J PER TEM 61-62/ U_LU 61-66/ Mhen was that? Ho Tr TO I_L_U_I Ho Tr 67-70/ (IT NOT EVERT DMl> l| ( | PER TEAR 78-79/ �BEGIN OFCK 54 30. For how many years have you been anoking ( I in O.37) cigars per day/per month/per year)? Less than 2 years.. .Ol 2-5 yearn 11-15 years 16-2O years 21-25 Nhen you smoke cigars, how deeply do you u s u a l l y inhale? sav: As deeply into the chest as possible IO-II/ .02 6-IO years 4O. Hould you ............ 1 As far back as the throat ....................... .04 Hell back into the mouth, or .05 , ...................... .03 years Only partly into the chest 26/ Just puff and don't really draw it in at a l l . ...5 DON'T KNOW..... O6 26-JO years O9 Hare than 40 years 8 08 36-4O years ................................. 3 4 O7 31-35 years..- .................... 2 1O 41. What type of cigars do you usually smoke? Filter tip or... Non-filter tip?. 39. During the period when you war* smoking tike most heavily on a regular basis, about how nany cigars did you usually smoke In • day. am* NUMBER, (IF NOT EVERT DAI I ) | | ) pen Mr ij | | PKH MONTH A. 12-13/ Now I an going to show you a diaqran of different size cigars, please look at the picture of the (Kim or CIGAR IN 0 . 4 1 ) . Now considering your style of smoking—for example? how lonq you usually leave the cigar in an ashtray or just hold It In your hand—tell me the number which indicates how Much of the cigar you actually snoke. 14-IS/ Section I I Section 2 28/ 2 O« (I F HOT EVERT DM 11 'LJU *** TIUUI 16-17/ Section 1 A. 3 Section 4 Mien was that? 4 ntOM LJ_L_LJ SKIP TO Q.S1 1B-2I/ FORMER CIGAR SMOKER SECTION TO LjLLLJ Ho 22-2S/ 42. How long has it been since yousmoked cigars fairly regularly? Tr ENTER DATSi OR WEEKS. OR MONTHS OR TEARS L_I_J L_LJ L.LJ L_U NEVER SMOKER REGULARLY 29- JO/ 3I-32/ 33-J4/ 35- 36/ (SKIP TO p.5l)....t 37/ �47. 43. On the average, about how many ctqars a day were you smoking at that time? What type of ciqars did you usually smoke jusl tiefnrf you slopped smokinq ciqars reqularly? COOK ONE ENTER NUMBER OF CIGARS PER DAYi (IF HOT RVERY DAYi 1 l| | | 3B-39/ | F i l l e r lip or N o n - f i l l e r tip?. | | PER HOHTH 4O-4I/ | | PER YEAR 42-43/ OR (IF MOT EVERY D A Y i ) 44. l| 4fl. Pal how long did you smoke (HUMBFR PER DM IN O.43) cigars per oay7 Less than 2 years O1 2-5 years HAND I CARD R I * Section 1 16-20 years 21-2S 07 31-35 years OS 36-4O years O9 More than 4O years 4 During the period when you were smoking the most on a regular basis, about how many cigars did you usually smoke in a day? 1O ENTER NUMBER I ( I F NOT EVERY DAYi I a) a* 56/ O6 26-3O years 49. 3 Section 4..... OS yeara 2 Section 3 04 1 Section 2 44-45/ O3 II-IS years 5V Mow 1 am qoinq to show you a diagram of d i f f e r e n t nite ciqars. Please look at the picture of Ihe (KIND OP CIGAR IN O - 4 7 ) . Now, constderinq your a t y l e of aMokinq—for example, how long you usually leave the cigar in an aahlray or just hold it in your hand—tell ne the number which indicates how much of the cigar you actually moke. O2 6-10 year*.. 1 | | | PER DAY 11 S7-58/ | | PER MONTH 59-60/ | | PER YEAR 6I-62/ OR ( I F NOT EVERY DAYiJ A. 45. You Mentioned that you have not smoked regularly for (TIME IN Q.42). Did you ever stay off cigars for a longer period of time? 1 Yes l| When was that? FROM 46/ LJ_LJ_I Mo Mo {SKIP TO O.47J 63-6S/ Yr t TO 46. LJ_L_L_I How long did you stay oft cigars at that time? Mo ENTER DAYSI L_U 49-SO/ Oil WEEKS I OH MONTHSI OR YEARS I 47-4S/ L_U L_LJ SI-52/ S3-S4/ Yr 67-70/ �DECKS 54-5S -130- 5O. When you smoked a clqar, how deeply did you unually inhale? aayi Aa deeply into the cheat aa poaalble Hould you ............ 1 .............. 54. Approximately how many houra a week are you exposed to this smoke in your home? It/ 2 10 houra or less. Only partly Into the cheat ....... . 11 to 15 hours... to far back aa the throat ....................... 3 16 to 20 houra... Hell back Into the wroth, or .................... 4 21 to 25 hours... Just puff and don't really draw It in at A l l . . . . 5 26 or more houra. DON'T KNOH ................ . ..................... B 55. 51. INTERVIEWER > DORS R CURRENTLY HHVK It SPOUSE OR PMTMER? COOED IN SECTION 5, p'a 2.6c or Be 7 ) . YES .............. . (IS ANY "MO* .......... NO ..... (SKIP TO O.5JI ....... t for how many yeara have you bean expoaed to amoke in this way? (CHECK ONLY ONE) 72/ 1 to 2 S 52. Does your (epouae/partner) YES Cigarette* 01 BAND CARD HO. DON'T KHOM 1 2 • 16 to 21 to 73/ Clqars 1 2 8 1 2 a O7 Nor* 74/ Pipe OO 11 to M .a reqularly any of the following? O3 to 7S/ Don't BEGIN DECK 55 53. Approximately how Mich amoka la there In the air In your hone? A lot I A little 5*. 2 Hone (SKIP TO 0,56) 3 INTERVIEWER! DOES It WORK? (IS "YES* COOED AT SEC 2. JOB* CODED AT SEC 3. O.1F?) p.4 1 Yea Ho A. OR "CURRENT (SKIP TO O.62) IO/ 2 Approximately how such amok* i* there in the air In the transportation you take to and from work (For example, your car, the train, the bus, etc.)? A lot 1 A little 2 Hone....(SKIP TO p.59) 3 I1/ �-III- 57. . DECK Approximately how many hours a week are you exposed to this smoke? -132- 55 61. OfXT 55 for how nany years have you heen exposed to thia smoke at work? 10 hours or lesa Lesa than 1 year Ot 2 I to 4 y e a r s * • . .02 16 to 2O hours 3 S to IO years.. .03 21 to 25 hours 4 II to 15 years. .04 26 or store hours............5 16 to 2O years. .05 21 to 3O years..... .06 •tore than 3O years. 58. I 11 to IS hours .07 For how nany years have you been exposed to this smoke? Less than 1 year Ol 1 to 4 years.... O2 S to IO years HAND I CARD » I 1C to 20 years OS ...O6 O? Don't Know M Approximately how such eeoke Is there In the air where you work? A lot 1 A little None 60. (SKIP TO 0.621....J Approximately how many hours a week are you exposed to this smoker I tl to IS hours 2 1C to 20 hours Ihere are sope questions that are asked in survey research that are difficult to ask directly because Many people think they are too personal, while it is understandable that people feel this way, there Is a real ne.ed for the information for the population as a whole. He now have a way that Mkea it possible [or people to qive infcreation, without telling anyone about their own situation. Let aw show you how this worksi we will use the next question 1 have here as an example. HAND II CARD X. As you see. there are two questions on the card. One deals with the "real* question that the research is concerned with, the other Is completely unrelated. Both questions can be answered *yea* or *no.* n>e of the two questions is selected by chance and you anawer it. (I'll show you how thai works in a minute). I do not know which question you are answering. Mien all the questionnaires have been tallied, the researchers can tell how many people have smoked marijuana, but they have no way of knowing whether It was you or sny other person In particular who has smoked marijuana. please take the coin that you have been handed and flip it now. Don't tell me which side came up. If the coin shows heads, please answer only question t. If the coin shows tails, please answer only Question 2. I won't look to aee If the coin comes up heads or tailsi and you don't tell me which question you are answering. Just tell me if your answer is "yes" or "no". Te 1 IV 4 2C or cure hours........... ....S 16/ It Is very simple, as you will see. YOU will flip the coin, the question you will answer is selected by chance. In no way can a truthful answer prove harmful to you. There ia no identifying Information that can link you to your answers. 3 21 to 25 hours 17-1H/ HAND II COIN IS/ 2 10 hours or less HAND I CARD V I 62. HAND I CARD I 1 | O4 More than JO years 59. 13-14/ OJ it to II years 21 to 3O years ea 12/ Ho Don't Know. A. Now let's do that again, using the next question. HAND R CARD y. Flip the coin again. If the coin turns up heads, please anawer only question number 1. If the coin comes up tails, please anawar only question nwber 2. Don't tell me the question. la your answer "yea" or "no"? Tee No Don't Know. 20/ �S5 UBCK $5 63. 63. Have you ever been arrested tt)i * felony? Yes I Mo....(SKIP TO Q.64> (Continued) H. INTERVIEWER! HAS R EVER BEEN CONVICTED OF A THIRD FELONY? IN O.638 EQUAL TO 3 OR MORE?) 2 21 / Yes....(GO TO NEK OUEX) 1 No A. (IS I JB 2 Have you ever been convicted of a felony? 1 Ye* Bo (SKIP TO Q . 4 .6) 22/ 2 Next, I'd like SOM information about drinking alchollc beverages. B. How Many felonies have you been convicted of ? 64. ENTER HUMBERt C. | (_ | I3-24/ No 2S-2./ • (SKIP TO SECTION 9) JV 2 • 65. D. 1 Yes Hhat- Month and year were you convicted of this/your first) felony? l_L_LU Ho Yr Nave you had any alcoholic beveraqea. Including beer, wine, or liquor, since (DATE OF LAST INTERVIEW)? Since (DATE OF LAST INTERVZEN) have you had a drink of beer? 1 Yes On what charge were you convicted? No 29-JO/ 66. .(SKIP TO O.7II 40/ 2 Mov long has it been since your last drink of beer? ca t. INTERVIEHERi HAS ft BVER BEEH CONVICTED OT Ik SECOND FELONY? IN 0.66B EQUAL TO 2 OK NORC?) Yes Ho....(8Kir TO Q.64) r. I I 2 jl/ Mist mnth and year war* you convicted of this/your first) felony? U_l_U Ho G. (IS 32-J5/ Yr On what charge were you convicted? 36-S7/ 67. As you think back over the period of tlM between (DATE OF LAST INTERVIEW) and now, about how nany cans or hottlee -of beer would you drink on a typical day when you drank beer? ENTER NUMBER Of CANS OR BOTTLES I | | | 43-44/ �OFCS442BSOC-8 68. Main, thinking back over the period of tine between (DATE or LAST INTERVIEW) and now, about how regularly did you drink beer? PROBR IF HECESSARii it's BOBetUw* hard to reveaher. Jus I give me your beat quean. More often than once • day Every day 01 , 71. Since (DATE or LAST INTERIVEH) have you had a drink of wine? Yes 4S-46/ O2 72. 1 No. . . . ( S K I P TO p. 76) 2 How long has it been aince your laat drink of wine? 5 or 6 day* • week.........................•••...03 .02 8-14 day* ago .03 1S-3O day* ago .04 I Month aqo. . . . . . . . . . OS .01 1-7 day* ago O4 1 or 2 day* • week Today .05 2-3 Month* ago J or 4 day* • week.... .06 Lea* often than one* a week......................O4 IF CANNOT OBCIOEiDoa't know 69. 96 Haw large were the can* or bottle.* th*t you usually drank? 4-6 Month* ago....... 47/ 16 ox. (half quart) can* or bottle*. SI-S2/ .07 7-12 Months ago .OB More than 1 year ago. Standard I X ox. can* or bottle* 50/ .09 32 oc. (full quart) can* or bottle*. Le** than 12 ox. can* or bottlea.. More than 32 oc. can* or bottle* 70. 9 Don't drink can* or bottle* of beer 73. 4 A* you think back over the period of tiM between (DATE OF LAST INTERVIEW) and now, about how Many glasses of wine would you drink on a typical day when you drank wine? « ENTER NUMBER OP GLASSES: | During the laat 12 Month* that yon drank el nee (DATE OP LAST INTERVIEW), how often did yon have • or nor* can* of beer IB a *lngle day (3 quart* or •ore)? Every day or nearly every day O1 3-4 tlaws a week 74. 48-49/ Every day OS 3-6 tin** a year Once or twice • year.... Never 3 or 4 day* a w e e k . . . . . . . . . . . . . . . . . . . . 4 1 or 2 day* a week O6 .............O7 OB 2 S or 6 day* a w e e k . . . . . . . . . . . . . . . . . . . . 3 O4 ti«*s a year 53-S4/ Hore often than once a d a y . . . . . . . . . . . . 1 Once or twice a wa*k.............................O3 7-11 | Again, thinking back over the period of tin between (DATE OF LAST INTERVIEW) and now, about how regularly did you drink wine? PROBE If NECESSARY! It'* coMtlMea hard to reaenber. Just give Me your best one**. O2 1-3 tlM** a Month | 5 Leva often than once a week...........6 IP CANNOT DECIDE! Don't know 6 55 �OFCt442RSec-R 75. DECK 55 Durinq the last 12 months that you drank since (DATE or LAST INTERVIEW), how often did you have B or More qlasaea o( wine in a slnqle day (more than a f i f t h ) ? Every day or nearly every day Ol 3-4 times a week. DECK 55 79. 56-S7/ Again, thinking back over the period of time between (DATE OF LAST INTERVIEW) and now, about how reqularly did you drink hard liquor? PROBE IF NrCFSSARYi It'a sometimes hard to remember. Just qlve me your best queas. More often than once a day.. Once or twice a w e e k . . . . . . . . . . . . . . . . . 0 3 Every d y . . . . . . . . . a.......... 1-3 time* a month O4 S or 6 day* a week 7-11 OS J or 4 days a week O6 1 or 2 days a w e . . . . . ek..... once or twice, a year.................07 HAND CARD O2 Less often than once a week. time* a year CC 3-6 time* a year Never 76. 08 IF CANNOT DCCIDEl Since (DATE OF LAST INTERVIEH) have you had a drink containing liquor, such as whiskey, vodka, qln, brandy, etc.? Te* 1 No...(SHIP TO 0.81) 0O. 63/ Don't know. About how Many ounce* of hard liquor are there in the drinks that you usually drink? One ounce (one ahot) 2 58/ 1.5 ounce* (one jlqqer) 1 64/ 2 CO 2 ounces (2 shota) 3 3 ounce* (2 Jlqqer* or 3 shots) 4 4 ounce a (4 shots) 77. 5 How long ha* it been ainc* your la*t drink of hard liquor? Today Ol 1-7 day* aqo S9-6O/ .....O2 5 or aore ounces It or sure ]igqers)..6 8-14 day* aqo... 03 Don* t know. 1S-3O day* aqo I month aqo OS it O4 2-3 snath* aqo O6 4-6 month* aqo... 81. O7 Durinq Urn Isat 12 south* that you drank since (DATE or LAST IHTMIIVCH), how often did you have • or sore drinks of hard liquor in a • inqle day (a half pint or sore)? Hare than 1 year ago. 7«. 3-4 times a week 09 As you think back over the period of tiam between (DATE or LAST INTERVIEW) and now, about how many drinks of hard liquor would you drink on a typical day in which yon drank hard liquor? ENTER HIMMEK Of DRINKS I | ( | 61-62/ .02 Once or twice a week. OS .01 .03 1-3 tt«e» a Month 7-12 months aqo Every day or nearly every day. .04 7-M tines a year .05 3-6 times a y a . . . . . . . er....... .06 Once or twice a year .07 Never .OB 65-66/ �OFCi4428Sec-8 62. -MO- Have you had a drink of beer, wine or hard liquor Months? In the laat 12 8S. INTERVIEWER: ...... (SKIP 10 SECTION 9) ........... 56 HAS R WORKED THE PAST YEAR? 10/ YES...(ASK A THROUGH E). 67/ No BEGIN DPCK 2 NO...(SKIP TO 0.86) 81. About how often during the past 12 Month* did you drink enough to feel high — that la, happier or More carefree than usual, Maybe a little flushed or dl»y. but not drunk, for More than 24 hours In a row? Durinq the past yean No Yes S or More tinea ....... . .............. 01 4 tlana .............................. O2 HAND CARD 3 tiMss .............................. 03 EE 2 tlawa .............................. 04 Once ................................. OS 84. Have you atayed away froM work because of a hangover?..1 2 !!/ Have you qotten drunk when on the Job? 1 2 12/ drinking? < 2 IV D. Has drlnkinq led to your quitting a job? 1 2 147 E. Has drinking hurt your chances for promotion or .1 2 IS/ C. .07 Have you lost a Job, or nearly lost one, because of .06 Merer In My life. A. B. Never In the past year, but lMS before that 00 «8-69/ Now I would Ilk* to ask you scan questions about experiences that Many people have had with drinking. During the paat year. . . Tea U) A. raiaea or a better Job? Ho 2 7/ 0 •«. when you were growing up, do you think your father drank occasionally, drank frequently, bad a drinking probleM. or didn't he drink? B. Have you gotten Into a heated argument while drinking?..! 2 " / C. 2 T2/ 2 «/ 2 •>*/ 2 7*7 2 7«/ 2 tl/ Drank occasionally 1 2 ?/ • Drank frequently 2 Had a drinking probleM 3 Didn't drink 4 DON'T KHOH 8 D. Drank occasionally Drank frequently 1 Didn't drink 4 DON'T KNOW E, Her* you afraid you Might be an alcoholic or that 2 Had a drinking probleM Have you deliberately tried to cut down or quit 1 8 F. One* you * tar ted drinking, waa It difficult for you G. Have you awakened the next day not being able to reMenber thing* you had don* while drinking? H. Have you often taken a drink the ftret thing when I. Have your hand* abaken a lot the Morning after J. Have you eoMetlaoe gotten drunk when drinking by 1 2 K. •/ O When you were growing up, do you think your Mother drank occasionally, drank frequently, had a drinking probleM, or didn't ahe drink? 7»/ 2 •7. Have yon sonatinas kept on drinking after I7/ �OFC>442BSec-9 OFr>442BSrc-ll> SFCTIOM lOt SECTION 9i t. RECREATION, LEISURE, AND PHYSICAL ACTIVITIES Now we would like you to answer some questions about your leisure time activities. INTERVIEWER! HAND R SELF-ADMINISTERED FORM 1. Have any of the recreation, leisure, and/or physical actlvitlea you've participated In since (DATE OF LAST IHTrftVlEW) brought you in contact with any of the following substances . . . TE8 HO Listed below are a serlee of Leisure Time and Physical Actlvitlea. Related activities are grouped under general headings. I) Please read the list and circle number 1 for each activity you have performed In the laat 12 Months, and circle number 3 for those you have not. 2) Check which months you performed those activities. 3) Record the average number of days per month you performed those activities, and 4) Hecord how many hour* and Minutes you performed those activities, on am average day. 1 28/ -M2TOXIC SUBSTANCES FOR EACH SUBSTANCE COOED YES, ASK A THROUGH D. Since (DATE OF LAST INTERVIEW), In what Month and year did your recreation, leisure and/or physical activities first bring you In contact with (SUBSTANCE)? com MONTH Since (DATE OF LAST INTFJIVIEH), for how nany years did you continue to come In contact with (SUBSTANCE)? I I I years YEAR W-12/ 1. Have you participated three or MOT* tlmei 1 in (READ EACH ITEM),* Yee Scuba diving 1 Auto, boat, or Motorcycle racing t 2 industrial chemicals?.....! 1 IV IS/ ED ,..,. CD-CD MONTH YEAR 40-41/ 2O/ 2 2I/ 1 Plane racing or plane acrobatics, not Including flight training or any assignments for the Armed Forces 2 1 Hang gliding 2 22/ Skiing fast down a high mountain slop* 1 2 2 2S/ 1 2 2C/ 41-46/ I I 47-4B/ 24/ 2 I YEAR 42/ 23/ 1 Sailing long distance In email sailing craft . corn Insecticides or MONTH 1 Surf board riding INTERVIEHERl !/ • 1 Mountain climbing 2. 2 16- 19/ Skydiving W No degreaslng chemicals? 1 ' 49/ CD-CD MONTH years YEAR SO-Sl/ 54-S5/ HAS R ANSWERED "IBS* TO AMY QUESTIONS OM BELT-ADMINISTERED FORM 3 AMD TO 0.1 ABOVE? defoliants or I -! i MONTH MO years YEAR S7-60/ S6/ (SKIP TO 8BCTIOM I * . . . 2 D.... X-ray or nuclear . 61/ mm OD — 1 MONTH YEAR 61-63, 64-67/ 68-69 �I1FCKS 56-5 J OU'.ltiJHK&t - I I I (Continued} Since (DATE OF LAST IHTERVIFH), bow many days per year did you cone in contact with (SUBSTANCE)? On the days you came in contact with (SUBSTANCE) how often did you use protective clothing or qear or wash to remove (SMISTANCE)—all at the time, •one of the time, or never? Which of the following did you use? CODE ALL THAT APPLY. IV HED 75/ 77/ 78/ 70-72/ Self contained or auppli«d DATS SO/ BEGIN DECK M/ »3/ 57 is/ f«c^ vhleld All of UM tiM.{J.8K El«..1 3 IO-12/ 2O/ 2*a/ (ace •Hivld. .**•••*.. .**».] All of the tin. (ASK B>..*1 26/ 27/ 21-23/ -~J 29/ VJ1 Self contained or *upplled 3O/ 36/ J7/ 3V All of the ttM»(MK E)***1 39/ 32-J4/ Self contained or anpplted DAYS . 46/ All of the tine. (ASK B|..*1 ODD 41-45/ DAYS 47/ 40/ face ehield*.* ••.••»•••• ..3 «/ 5O/ 5I/ Seir contained or •applied 53/ se/ «7/ 59/ 6O/ 54-S6/ *2/ DATS Self contained or • applied air breathinq apperatne.. .6 NANO K SELF-AOrllNISTERFD FOKM 4 AND GENERAL PURPOSE ANSHER SHEET 16/ 17/ wachi.»q r«ciltti««... 5 Seir contained or aupplted 24/ INTER VIEWER I 63/ The Jenkins Activity Survey aaka questions about aspects of behavior that have been found helpful in Mdlcal diaqnosls. Each peraon la different, ao there are no "right* or 'wronq* answers. Por each question, choose the anawer that la true for you. on your answer sheet, fill 1» the circle below the letter of your answer. Use a black lead pencil, and Bake your Mrka heavy-and dark. Hark only one anawer for each question. If you change your Hind, eraae the old vark completely. �DECK 57 OFC:442HK«c-ll SECTION 1 1 1 2. INCOME Did you earn any income froei any Job during 1964? from retirement plans or pensions. Do not Include Income How I have acme questions about your incoM. Yes (ASK A) I 1. No (SKIP TO 0.3) 2 Please tell M which letter on this card best represents the total household income In 1984 before taxes or other deductions (or all people in your household, not including roomers. Itils Mount should include vages, net income llom business, interest, dividends, pensions, and any other money Income. Toll me the letter that COMS closest. A. In which of these groups did your earnings from jobs in 1984 f a l l — that la, before taxes or other deductions? Tell M the letter that come closest. (25,000 - f29,M9 OS »3O.OOO - (34,999 06 (35,000 - (39.999 07 (40,000 - (44,999 08 OS r. »JO,000 - (34,999 OC $3S,OOO - 139,999....... 07 H. (40,000 - 144,999 OB HAND CARD I. (45.000 - (49,999 09 rr J. (5O.OOO - $54,999 1O K. (55,000 - (59.999 11 L. $60,000 - $64,999 12 H. (65.000 - ««9,999 13 H. «70,000 - »74.999 14 " C. ea $25,000 - $29,999 04 B. .04 H. (20,000 - (24,999 $20,000 - $24,999. 03 D. .03 G. (15,000 - (19,999 $15,000 - (19,999. 02 C. 69-70/ .02 F. * 10.OOO - (14,9*9 $10,000 - $14.999. E. B. S6-67/ .01 D. Ol $5.0OO - (9,999... C. $5,OOO - (9,999 A. B. A. 68/ HAND (65,OOO - (69,999 13 $7O,OOO - (74,999 14 $75,000 - (79,999 15 P. JBO.OOO - (84,999 16 0. $85.000 - (89,999 17 R. S9O.OOO - (94,999 18 S. $95,OOO - $99,999... 19 T. $1OO,OOO or Bore 2O 17 $9O,OOO - (94,999 12 16 H. (60,000 - (64,999 IS $85,000 .- $89,999 11 O. Q. (55.OOO - (59,999 N. $80,000 - $84,999 IO H. P. O9 L. $75,OOO - |79,999 (4S.OOO - (49,999... (50,000 - (54.999 K. rr O. I. J. CAM) 18 8. $95,000 - (99.999 19 T. (tOO.OOO or sore 2O 3. INTERVIfMeRt 71-74/ AH RECORD TIME ENDED �OFCi4428REMKRK DECKS S7-5H INTERVIEWER REMARKS IHTERVIEHBRt OFCI4428RKHARK 6. Please complete these reurka as soon as you have finished the questionnaire. List the questions that confused, angered, or caused discovfort to the respondent or questions that you feel the respondent did not answer truthfully. EXPLAIN. NONE I. Lenqth of the interviawi (Section 1. p.1 throuqh section 12) 2O-21/ BBC IN DECK SB 2. C. Date ol the interviewi I_U L_LJ DM TKM 3. ?. ! Black 2 Other S2-34/ Describe Problem Race of Respondent! Nhlte 27-29/ JO-JI/ 7S-77/ 22-24X 2S-26/ L-LJLJ Himires List questions with skip errors, questions that were confusing to you, or questions that otherwise didn't work. EXPLAIN |6/ 1 NONE SECTION o. In general, -hat was the respondent's attitude toward the Interview? Friendly and interested I IT/ 39-41/ 42-43X 44-46/ C. 4. 37-38Y 47-487 49-5I/ Describe Problem S2/ Cooperative but not particularly interested ..........2 {•patient and restless.... Hostile 5. ] 4 0. 9. In general, was the respondent's understanding of the questions.... Good? 1 Fair? 2 IB/ Please record your interviewer ID li Please sign your nanei 5 3-SB/ �•one 4428 5/85 AIR FOBCB BKALTI BOBTXT roni Complete tale for* and enoloee m copy with eaoh oaaa Balled to Chicago. Int. 10 * Int. **m»i Date Halledi vounm It mot eaoloaad. a>plala ioa Bbvat Chlldraa** Record For* BuppleMBtary ChlldraK'e Record Fora Queetlonaalre Seir-AdBialetered Fora I ~~i Seir-ktelnlatered Fora * oo eeir-Aomlelexered For* J Medical Cooeeat Feral FOB OTFIC* on ont CIROU ' Date recelred la Chloafoi »1I required f«ra» vreeeati Tee �APPENDIX C* Physical Examination Methodology *0riginal forms were color coded; limited photocopy quality. �5) 0 'D 'D ® < 0 0 (D 0 '$ ® 0 0 ® 0;.l* 0 (D ® ••'*.' -I' VERSION 1.0 JRW:SCF 585 FAMILY HISTORY - PLEASE BLACKEN THE CIRCLE FOR ANY FAMILY MEMBER THAT HAS HAD ANY OF THE FOLLOWING: BLOOD RELATIVES -» NONE MOTHER FATHER SISTER BROTHER CHILD REVIEWER'S COMMENTS: O 0 0 DIABETES EPILEPSY CANCER HIGH BLOOD .PRESSURE O HEART DISEASE STROKE ALLERGY STOMACH TROUBLE O O O /— . f~\ NERVOUS TROUBLE w BLOOD DISEASE DEFORMITIES O O0 ARTHRITIS OTHER FAMILIAL DIS. '-» PLEASE UST HERE: Q © O O O O ;-> 6 c /""; O O O O O 0 O O O C'! O "••** 6 o '•-^' 0 .•'~v, 6 r-i '~^' o o O o o o o 0 o o o o o p) o o o '• 2 o • o •^^' o o o o o o c o o^ • o 0 0 o ; O • o -^ 0 Q •-. CURRENT FAMILY STATUS FATHER: LIVING-AGE CONDITION OF HEALTH? O EXCELLENT DEAD-AGE MOTHER: LIVING-AGE SISTERS: / LIVING NUMBER ARE YOU MARRIED? CHILDREN: QFAIR OPOOR CAUSE OF DEATH? CONDITION OF HEALTH? O EXCELLENT QGOOD OPOOR CAUSE OF DEATH? DEAD-AGE BROTHERS: / LIVING NUMBER^ D£AD QGOOD AGES CAUSES AGES CAUSES NO. OF YEARS WIFE'S I AGE HEALTH _ OF WIFE? O EXCELLENT IF WIFE IS DEAD, PLEASE GIVE AGE, YEAR, AND CAUSE OF DEATH: BOYSAGES GIRLS' AGES DO YOU HAVE ANY PHYSICAL OR NERVOUS COMPLAINTS? ALL HEALTHY? QGOOO ANY DEAD? ANY BIRTH DEFECTS? DO YOU HAVE ANY ALLERGIES OR SEVERE REACTIONS TO MEDICINES, FOODS, PLANTS. CHEMICALS. ETC.? OYES ONO EXPLAIN: PLEASE DESCRIBE: NCS Trans-Optic ® EP01-21182:321 C-l A8900 �PERSONAL HISTORY VES NO YES NO NO YES ® ~ ® HEPATITIS ACNE ® ® WORMS EXCESS HAtt GROWTH ® ® COUTIS ® ® ARTHRITIS < ® ® SCLERODERMA CIRCLE NEXT TO ANY OF ® ® RHEUMATIC FEVER OTHER SKIN TROUBLE ® ® CANCER OR TUMOR THESE CONDITIONS THAT YOU NOW HAVE OR HAVE HAD IN THE PAST, ® ® HEMORRHOIDS ® ® KIDNEY STONES . ® ® KIDNEY TROUBLE ®®VARieOSE VEINS ® ® PHLEBITIS BLACKEN THE © OTHERWISE BLACKEN ® . ; ® ® HERMA. (RUPTURE^ REVIEWER'S COMMENTS: ® ® ANEMIA ® ® CATARACTS @ ® PROSTATE TROUBLE ®®T!ONStLmS ® ® SINUSITIS ® ® GONORRHEA ®®HAY FEVER ®®FAWTING @® FITS OR CONVULSIONS ® ® MUMPS ®®MALARJA ®®GOUT ® ® DEPRESSION ® ® DIABETES ® ® NERVOUS BREAKDOWN ® ® MEASLES ® ® DYSENTERY 0® POUO © ® ASTHMA ® ® BRONCHITIS ® ® PARALYSIS ® ® PNEUMONIA ® ® MUSCLE PAIN ® ® MUSCLE WEAKNESS TUBERCULOSIS ® ® HEART TROUBLE ® ® STOMACH TROUBLE ® ® NUMBNESS ® ® LOSS OF SENSATION ® ® LOSS OF SEX DRIVE , ® ® RHEUMATOID ARTHRITIS ® ® SEVERE ARTHRITIS ® ® JAUNDICE g>®UVeR TROUBLE 6RYTHEMATOSIS LIST THE AVERAGE FOR EACH OF THE FOLLOWING DURING THE LAST 90 DAYS: FOR THE PAST 90 DAYS OR MORE: DID YOU TAKE REGULAR EXERCISE? PER DAY NUMBER CIGARETTES HOURS SLEEP PER NIGHT ALCOHOLIC DRINKS DAYS WORKED PER WEEK WHAT IS YOUR USUAL WEIGHT? LBS WHAT IS THE MOST YOU EVER WEIGHED? HOURS WORKED PER DAY LBS CUPS COFFEE DAYS VACATION PER YEAR AT WHAT AGE? HAVE YOU RECENTLY LOST OR GAINED WEIGHT? CHEWING TOBACCO IF SO, HOW MUCH? (+/-) SNUFF PAST HISTORY PLEASE LIST PREVIOUS OPERATIONS, INJURIES, AND SERIOUS ILLNESSES, INCLUDING THOSE INDICATED ABOVE YEAR LBS PLEASE BLACKEN THE CIRCLE IF YOU HAVE HAD REPEATED CASES OF ANY OF THE FOLLOWING IN THE PAST YEAR: YES NO DESCRIPTION OF OPERATION. INJURY, OR SERIOUS ILLNESS ® ® PNEUMONIA ® ® KIDNEY INFECTIONS ® ® SKIN BOILS DO NOT MARK IN THIS SPACE WHEN WAS YOUR LAST I PHYSICAL EXAM? I DID YOU TAKE ANY MEDICATIONS OR TREATMENT NOW OR OCCASIONALLY ANY ABNORMALITY FOUND? • YES O NO NO PLEASE DESCRIBE: NAME YOUR PERSONAL PHYSICIAN YES ® ® OTHER INFECTIONS SPECIFY STREET ADDRESS CITY, STATE, & ZIP C-2 ARE YOU UNDER ANY MEDICAL TREATMENT NOW? YES NO �NAME: DATE OF BIRTH: CASE NUMBER (0: 1 2 3 * 5 8 7 3 9 •: 1 2 3 #' 5 6 7 8 9 C '* 'i 2 3 4 5 6 7 3 9 0 'O, 1 f! 3 4 5 8 7 3 9 5 6: .'7 •'§. 9 f FEB APR WAV '.8' f- MH .9; MAR — ' I - 'f E 3 , i, MO TODAY'S DAY DATE YR BRQUP « fll (0. J . , i 301 20; 86, 90) REVIEW OF SYSTEMS FORM AFHS-2A 'Oi :T ''2; '••.-* < -i., .»', '•( '?; (8, '•) S•• 3') 4 . _V •—• • „. j ... -..- • v^- ''^ '-,V_ VERSION 1.0 JRW:SCF 585 INSTRUCTIONS IF YOU HAVE ANY OF THE FOLLOWING COMPLAINTS, BLACKEN THE CIRCLE IN, THE "YES" COLUMN, IF NOT, BLACKEN THE CIRCLE IN THE "NO" COLUMN. THE DOCTOR OR NURSE WILL ASK ABOUT THE DETAILS LATER. ANSWER ALL QUESTIONS. IF IN DOUBT, GUESS YES OR NO. QUESTIONNAIRE YES NO ••'£}•'$ ANY FOODS THAT TEND TO DISAGREE (WHICH ONES?) ••I*ITCH OR RASH (WHERE?) 0® SWELLING, LUMP, OR SORENESS ANYWHERE ON BODY (WHERE?) ® ® NUMBNESS OR TINGLING (WHERE?) 0 ® TWITCHING MUSCLES (WHERE?) 0 (NJ GET UP NIGHTS TO URINATE - HOW MANY TIMES A NIGHT? REVIEWER'S COMMENTS: YES NO ® ® 1. SEVERE HEADACHES OB HEAD PAINS • COMMENT YES NO 0 ® 31. WORRIED ABOUT YOUR HEART 0 (5) 2. ANY DISTURBANCE IN VISION 0 ® 32. BLOOD PRESSURE TOO HIGH ® ® 3. PAIN OR DISCOMFORT IN EYES 0 ® 33. BLOOD PRESSURE TOO LOW ® ® 4 . WEAR GLASSES (OR CONTACT LENSES?) 0 ® 34. PAINS IN HEART OR CHEST 0® 5. CONSTANT NOISE IN EARS 0 (N) 36. POUNDING OR SKIPPING OF HEART ® ® 8- HARD OF HEARING 0 ® 36. HEART STARTS RACING SUDDENLY 0 ® 7 ® ® 37. SHORTNESS OF BREATH OR WHEEZING - EAR ACHE WITH COLDS 0 (N,"' 8. EAR ACHE WITH PLANE FLIGHTS .Y, ..N,; 38. TROUBLE GETTING A DEEP BREATH 0 ® 0 ® 39. SWELLING ANKLES 0 9. CHRONIC RUNNING EARS ® 1°- CHRONIC STUFFY OR RUNNY NOSE ; 0 ® 11. NEED TO USE NOSE DROPS FREQUENTLY Y) 'S) 40. LEG CRAMPS IN BED OR SITTING STILL 0 ® 41. LEG CRAMPS WHILE WALKING 0 ® 12. BAD NOSE BLEEDS AT TIMES 0 ® 42. PAIN OR TROUBLE WITH SWALLOWING 0 ® 13. FREQUENT SEVERE COLDS OR SORE THROAT 0 ® 43. POOR APPETITE RECENTLY 0 ® 14. ANY KNOWN DENTAL PROBLEMS 0 (N) 44. POOR APPETITE ALWAYS 0 ® 16. SORENESS OR BLEEDING OF GUMS 0 ® 45. NAUSEA OR VOMITING 0 ® 16. MORE THAN A YEAR SINCE TEETH CHECKED 0 ® 46. VOMITING OF BLOOD 0 ® 17. SORE MOUTH OR TONGUE ® ® 47. BELCHING. BLOATING OR INDIGESTION 0 ® 18 - GOITER OR THYROID TROUBLE 0 ® 48. YELLOW SKIN OR EYES (JAUNDICE) ® ® 19. THYROID TEST TOO HIGH 0 ® 49. BURNING OR HUNGER PAINS IN STOMACH 0 ® 20. THYROID TEST TOO LOW 0 ® SO. USE ANTACIDS FOR STOMACH BURNING ® ® 21. FEELING OF LUMP IN THE THROAT 0 ® 51. SORENESS OR PAIN IN STOMACH. ABDOMEN 0 ® 22. NEED TO TAKE THYROID MEDICINE 0 (Sj) 23. HOARSENESS AT TIMES 0 ® 52. SUSPECT ULCERS OR STOMACH TROUBLE t 0 ® 53. CRAMPS IN STOMACH OR LOW DOWN 0 ® 24. RECENT OR CHRONIC COUGH 0 ® 54. LOOSE BOWELS OR DIARRHEA 0 ® 25. CHRONIC COUGHING UP OF SPUTUM ® ® 5B. BLACK OR TARRY STOOLS isowa MovEMtm) 0 'Mi 26. EVER COUGHING UP OF SPUTUM 0 ® 56. FRESH OR BRIGHT BLOOD WITH STOOLS 0 'S) 27. ACHE AU OVER 0 ® 57. MUCUS (SUME OR PHLEGM) IN STOOLS .Y' ..N; 28. HAVING CHILLS OR FEVER 0 ' N ) 53. CONSTIPATION 0 0 Cjjl 59. USE LAXATIVES FREQUENTLY *'> 29. SEVERE SOAKING NIGHT SWEATS Y"' N 30. LIVED WITH ANYONE HAVING T.B. ;?• .N, 60. USE ENEMAS FREQUENTLY NCS Trans-Optic ® EP01-21 161:321 C-3 A8900 �REVIEWER'S COMMENTS: YES NO YES NO ® ® 81. RECENT CHANGE IN BOWEL HABITS 0® 91. NAIL BITING 0 ® 62. RECTAL TROUBLE OR PAIN ® ® 9Z SLEEP WALKING © (N) 63. PAIN IN THE KIDNEY REGION ©® 91 BED WETTING AFTER AGE 12 ® ® 64. BLOOD OR PUS IN URINE 0® 94 ® ® 66. ALBUMN IN URINE ©® 95. IRREGULAR LIVING HABITS 0 ® 66. SUGAR IN URINE ©® 96. CANT GO TO SLEEP OR STAY ASLEEP 0 ® 67. SPELLS OF FREQUENT URINATION ® ® 97. NEARLY ALWAYS IN POOR HEALTH © ® 68. SEVERE BURNING OR PAIN ON URINATION ® ® 98. CONSIDERED TO BE A NERVOUS PERSON © ® 69. PAINS OVER BLADDER OR LOW DOWN ® ® 99. FROM SICKLY OR NERVOUS FAMILY © ® 70. TROUBLE STARTING URINE ® ® 100. TREMBLE AND SWEAT EASILY ® ® 71. URINARY STREAM HAS BECOME WEAK © ® 101. HAVE TROUBLE MAKING UP YOUR MIND - - CHRONICALLY TIRED OR OVERWORKED © ® 72. HARD TO EMPTY BLADDER COMPLETELY ® ® 10Z EASILY MIXED UP OR CONFUSED ©®71LOSE CONTROL Of PASSWGUR9« 0® © ® 74. PAINFUL OR SORE GENITALS (PRIVATES) 0® 104. FEEL SAD. LONELY OR DEPRESSED 103. CLUMSY OR HAVE FREQUENT ACCCENTS , "®© 1 * ON OFTEN 0 © ® 106. WISH I WERE DEAD 76. STIFFNESS OF MUSCLES OR JOINTS . SEVERE PAWS IN ARMS OR LEGS ® ® 107, WORRY COWTNUALLY 78. PAINFUL FEET © ® 108. UPSET BY LITTLE THINGS 0® 10ft A PERFECTIONIST ® ® 80. PAINS IN NECK 0® 110. SENSITIVE OR FEELINGS EASILY HURT ® ® St. EASYTOSUNBURN? ® 111. OFTEN MISUNDERSTOOD © ® 82. SUBJECT TO ACNE 112. OFTEN ACT ON SUDDEN IMPULSE ®® 83, SUBJECT TO SOLS Ofl INFECTION © ® 84. SUBJECT TO ATHLETE'S FOOT, SKIN FUNGUS (V)® 114. FREQUENTLY KEYED UP AND JITTERY ® ® 88. SUBJECT TO HIVES OR SKIM REACTIONS © ® 116. EASILY SCARED BY SUDDEN NOISE ® ® 86. EASY BLEEDING OR BRUISING ® ® 116. HAVE BAD DREAMS OR THOUGHTS © ® 87. MOLE OR SORE WHICH IS NOT HEALING ® ® 117. SUSPECT A SERIOUS DISEASE OR CANCER © ® 88. SEVERE DIZZINESS ® ® 118. HAVING TROUBLE GETTING ALONG WITH SOMEONE AT HOME OR AT WORK ® ® 88. GENERALIZED WEAKNESS ® ® 90. MUSCLE WEAKNESS EXPOSURE HISTORY HAVE YOU EVER BEEN EXPOSED TO ANY OF THE FOLLOWING SUBSTANCES OR TYPES OF RADIATION? EXPOSURE IS DEFINED AS SKIN OR RESPIRATORY CONTACT OF MORE THAN ONE DAY'S DURATION. FOR EACH "YES" RESPONSE, PLEASE COMPLETE ONE OF THE THREE BLOCKS ON FORM AFHS-2B. 0®, COMMENT REVIEWER'S COMMENTS: YES NO YES NO © ® CHLOROMETHYL ETHER 0® COAL TAR © ® CREOSOTE © ® ARSENIC ® ® ANTHRACENE 0®CHROMATES ® ® BENZENE ©©ASBESTOS 0® CUTTING OILS ® ® NAPHTHYLAMINE ® ® TRICHLOROETHYLENE ©<3) AMWODIPHENYL ® ® ULTRAVIOLET LIGHT (OTHER THAN SUN) ® ® MUSTARD GAS © ® X-RAYS (OTHER THAN ROUTINE) ® ® VINYL CHLORIDE © ® IONIZING RADIATION (OTHER THAN X-RAYS) REVIEWER'S NAME (PRINTED): C-4 �CASE NUMBER B NAME OF PARTICIPANT j PAGE OF FORM AFHS-2B EXPOSURE HISTORY DETAILS 0 FOR EACH "YES" EXPOSURE AT THE END OF FORM AFHS-2A, PLEASE FILL OUT ONE OF THE FOLLOWING BLOCKS. USE ADDITIONAL SHEETS IF NECESSARY. TYPE OF EXPOSURE (COAL TAR, ETC.) WAS EXPOSURE RECEIVED ON THE JOB? YES NO YES NO YES NO YES NO IF ON-THE-JOB EXPOSURE, JOB TITLE IF NOT ON-THE-JOB EXPOSURE, HOW EXPOSURE RECEIVED CHECK FREQUENCY OF EXPOSURE THAT BEST FITS YOUR EXPERIENCE DAILY WEEKLY MONTHLY YEARLY TYPE OF EXPOSURE (COAL TAR, ETC.) IN WHAT YEAR(S) WERE YOU EXPOSED? WAS EXPOSURE RECEfVED ON THE JOB? IF ON-THE-JOB EXPOSURE, JOB TITLE IF NOT ON-THE-JOB EXPOSURE, HOW EXPOSURE RECEIVED CHECK FREQUENCY OF EXPOSURE THAT BEST FITS YOUR EXPERIENCE DAILY WEEKLY MONTHLY YEARLY TYPE OF EXPOSURE (COAL TAR. ETC.) IN WHAT YEARIS) WERE YOU EXPOSED? WAS EXPOSURE RECEIVED ON THE JOB? IF ON-THE-JOB EXPOSURE, JOB TITLE IF NOT ON-THE-JOB EXPOSURE, HOW EXPOSURE RECEIVED CHECK FREQUENCY OF EXPOSURE THAT BEST FITS YOUR EXPERIENCE DAILY WEEKLY MONTHLY YEARLY TYPE OF EXPOSURE (COAL TAR, ETC.) IN WHAT YEARIS) WERE YOU EXPOSED? WAS EXPOSURE RECEIVED ON THE JOB? IF ON-THE-JOB EXPOSURE, JOB TITLE IF NOT ON-THE-JOB EXPOSURE, HOW EXPOSURE RECEIVED CHECK FREQUENCY OF EXPOSURE THAT BEST FITS YOUR EXPERIENCE DAILY WEEKLY MONTHLY YEARLY C-5 IN WHAT YEARIS) WERE YOU EXPOSED? �CASE NUMBER GROUP I NAME: DATE OF BIRTH: VERSION 1.0 JRW:SCF 585 GENERAL APPEARANCE • • • • • • ^ • • ^ • ^ • • • W B ^ ^ ^ M M ^ ^ M^ ^ ^ M ^ ^ ^ i ^ ^ ^ APPEARANCE VS STATED AGE HAIR DISTRIBUTION APPEARANCE OF ILLNESS OR DISTRESS WELL NOURISHED YOUNGER THAN YES NORMAL OBESE UNDER-NOURISHED OLDER THAN SAME AS NO ABNORMAL I v N COMMENT DESCRIBE ANY ABNORMAL HAIR DISTRIBUTION: NOTE: FILL IN VITAL SIGNS WITH "0"» IF REFUSED. I VITAL SIGNS HEIGHT CM ~ rtft r«N 00© ®®® CD CD® WEWHT (UNOMSSeO) KG ""(?! TIMniWTUM OHM. T '971 ''Oj (pi Q\ arrow BLOOD MIESSOBE SYSTOUC oV?i -^ 0® 0 «® 96) *.'f} (2) 7$® 'j) ftCs^ d) DIASTOtIC 'Oi • Oi : ^ X0 T) /2 , (2 t "11 '31 (Ji @'D® VPB* 2) -,3^ ?( PULSE RATE MINUTE •'S ' • f ''n'- NONOOMWAWT AflM, Hf-AIIT Lim '0' • ' O i •Ti rf) T: C*) (|; 'ii '.I^X '3. 1": vj) ' 2 ', :2i 4: '4~; •'•*) .:*• PER •'3^ • 3 , 0® 8®® '90® ••£ .?: (i) Js'i iia •'» :• T. 9 S 5 5 5 S 6 ei 8 '8 8 8 7 7 120 7 7 7 7 7 7 7 7 a a 9 9" 8 a a a a 8 9 9 9 N COMMENT 7 8'- Y' •«; 8 '»•: DESCRIBE ANY IRREGULARITIES: S. 5 "8; iob i'» I '5* -•. iib e~'i PULSE IS: .' REGULAR IRREGULAR 9; 9 '<£' T? 3 io> :«'.:.•«• i; 104 7 ' 7 a a" a9' a • ! ' »,' '4) '-t 9 FUNDUSCOPIC NORMAL YES NO YES EXTERNAL NO X' N. f LIGHT REFLEX v N HEMORRHAGES T- N! A-V NICKING v N) EXUDATES Y N ARTERIOLAR SPASM * N PAPILLEDEMA NO N; ARCUS SENILIS PRESENT N DISK PALLOR r; YES r: N. f CUPPING ABNORMAL REFUSED v Y N, ABNORMAL OCULAR PIGMENTATION COMMENT DESCRIBE VASCULAR LESIONS, HEMORRHAGES, EXUDATES, OR PAPILLEDEMA: NCS Trans-Optic ^ EP01-21157:321 C-6 A8900 �C NORMAL 0 ABNORMAL YES NO Y• N. Y N) Y a, RIGHT TYMPANIC MEMBRANE INTACT LEFT TYMPANIC MEMBRANE INTACT NASAL ULCERATIONS C REFUSED •"Y: :N) COMMENT DESCRIBE ABNORMALITY: C NORMAL ;Y- C THYROID GLAND -;MI PALPABLE r >!' NODULES LEFT •'." ".'. RIGHT ": : PAROTID GLAND ENLARGEMENT CAROTID PULSE ABSENT CAROTID BRUIT PRESENT Y N COMMENT ABNORMAL 'Y N .: REFUSED DESCRIBE ABNORMALITY: ENLARGED Y N TENDERNESS THORAX AND LUNGS O NORMAL O EXPMATOW ABNORMAL C CHEST CIRCUMFERENCE (CM) AT NIPPLE LEVEL REFUSED YES : Y• NO NI ASYMMETRICAL EXPANSION •Y) N. WHEEZES Y N 'D ' ,S> DULLNESS Y ; a"* ;'<»'> N) HYPERRESONANCE .Y; WSWIATION 0® Csi ft) ,7 7 RALES (NOTE LOCATION IN BOX BELOW) DESCRIBE ABNORMALITY: (Y,I : N) COMMENT HEART O NORMAL ,_ C ABNORMAL YES O . J HEART SOUNDS NORMAL DISPLACEMENT OF APICAL IMPULSE PRECORDIAL THRUST NO O ~ O ABNORMAL HEART SOUND(S) NO YES 'Si 'N: N; N' SI S2 S3 34 Y Y Y Y Y N COMMENT REFUSED MURMUR .. NO YES SYSTOLIC AORTIC Y N DIASTOLIC Y N PULMONIC Y N Y CHEST AREA APEX Y: N> N Y DESCRIBE ANY ENLARGEMENT, IRREGULARITY OF RATE, MURMUR, OR THRILL: C-7 N MITRAL Y, N' Y N �CASE NUMBER (o,, 'r '.*'• V •'".?• (o "V '2 ' 3 , 4. 5 (?: 'r f 3 4 ?,. 6 1 0,:<1; GROUP NAME: DATE OF BIRTH: 8 9 C" 8: 9 0 2 3; : 4) 5'. 8, 7 f: •>. I .2:.3j i-I) 'a.;.? • a .•.;(. F (a> '.!• '. MO TODAY'S DAY DATE YR FES MAR 30 201 101 38i 86! APR JUN M. Aue .9 JAM 'f 6 5) •98 87! 91' set i' «r» *i riov OCT 3' 94' LilJ 2 96) PHYSICAL EXAMINATION (CONTINUED) FORM AFHS-38 VERSION 1.0 JRW:SCF 585 CM YES NO O NORMAL O ABNORMAL C REFUSED ®<3> HEPATOMEGALY '£; '"N) SPLENOMEGALY Y N' TENDERNESS LIVER r; .r' TENDERNESS SPLEEN OTHER TENDERNESS Y N OTHER MASS WAIST MEASUREMENT 1!!^^^ (10) lj!0l -301 LIVER SPAN Y' N; COMMENT DESCRIBE ABNORMALITY: EXTREMITIES YES NO YES NO O NORMAL 0® PITTING EDEMA O ABNORMAL ®® CLUBBING OF NAILS ® ® VARICOSITIES ® ® LOSS OF HAIR ON TOES RIGHT ® ® LOSS OF HAIR ON TOES LEFT 'Y; •'$ NON-PITTING EDEMA •'»} ABSENCE (SPECIFY IN •NJ COMMENT BOX BELOW) DESCRIBE EDEMA. SIGNS OF VASCULAR INSUFFICIENCY, OR ABSENCE OF PART OR ALL OF EXTREMITY: O REFUSED PERIPHERAL PULSES RADIAL FEMORAL POPLITEAL DORSALIS PEDIS POSTERIOR TIBIAL COMMENTS: NORMAL .•~\ *—' DIMINISHED ABSENT REFUSED O O 0 "> 6 o 6 O O O O O 'o* 0 Q o o 6 ® (s) NCS Trans-Optic ® EP01-21156:321 C-8 A8900 �MUSCULOSKELETAL O NORMAL SPINE MUSCLES O ABNORMAL PRESENT ABSENT O O WEAKNESS O O TENDERNESS O O ATROPHY O O ABNORMAL CONSISTENCY PRESENT ABSENT O 0 SCOLIOSIS o o o o KYPHOSIS 0 PELVIC TILT 0 DECREASED RANGE OF MOTION 0 "0 O REFUSED STRAIGHT LEG RAISING O NORMAL O ABNORMAL TENDERNESS LEVEL © ® COMMENT o TENDERNESS O CERVICAL THORACIC O O LUMBAR SACRAL COMMENTS: GENITOURINARY/RECTAL/HERNIA YES NO VES NO ; ;O" NORMAL O; ABNORMAL. "®®KT INGUINAL HERNIA ®®LT INGUINAL HERNIA ®®VARICOCELE ® ® HEMORRHOIDS ® ® PROSTATE ENLARGEMENT ®®EPIDIDYMIS ®® RECTAL MASS ®®SCROTALMASS ®®®®®®®@®® O REFUSED 0® COMMENT TSSTE5 LEFT RIGHT NORM. O O ABS. ENGL. ATROPH. O O 0 o o o (DIAMETER-CM) COMMENTS: NON PALPABLE ENLARGED O NORMAL O ABNORMAL O REFUSED ' TENDER HARD CERVICAL O O O O OCCIPITAL SUPRACLAVICULAR AXILLARY EPTTROCHLEAR INGUINAL FEMORAL O O O O O O O O O O O O O O O O O O O O O O O O OTHER TESTS ORDERED ® YES ® NO DESCRIBE: ® ® COMMENT PRINTED NAME OF EXAMINING PHYSICIAN SIGNATURE C-9 oooooool o LYMPH NODES CONFLUENT O o o o o 0 o �• E EM E I ^ r B I • Z l S HJ ^ t B ™ I 1 2 3 * 5 8 7 3 0 1 2 3 4 5 8 » 8 9 C 0 ~ 0 9 * 1 * 3 4 5 6 7 8 9 0 0 1 . 4 3 4 9 8 7 8 9 NAME: DATE OF BIRTH: MO TODAY'S DAY DATE YR SROUP 1 JAN FEB MAR APR MAY JUN JUt AUG SEP OCT MOV 30 20 10 9 8 7 6 3 4 3 2 1 86 88 87 88 89. 90< 91 9Z 93) 94 95 98 <tt> G FORM AFHS-4 ^ DEC DERMATOLOGIC EXAMINATION AND BIOPSY VERSION 1.0 JRWtSCF 585 ^^^^^^^^^^^^^^^^^^^ BLACKEN CIRCLE IF LESION IS OBSERVED REFUSED EXAMINATION INDICATE TYPE AND LOCATION OF LESION ON ATTACHED ANATOMICAL CHART YES NO YES NO c I PALMAR KERATOSES COMEDONES ACTINIC KERATOSES ACNEFORM LESIONS . ACNEFORM SCARS PETECHIAE » DEPIGMENTATION C ECCHYMOSES INCLUSION CYSTS O CONJUNCT1VAL ABNORMALITY : CUTIS RHOMBOIDALIS ORAL MUCOSAL ABNORMALITY HYPERPIGMENTATION FINGER NAIL ABNORMALITY JAUNDICE TOE NAIL ABNORMALITY SPIDER ANGIOMATA O PALMAR ERYTHEMA 1" SUSPECTED MELANOMA DERMATOGRAPHIA SUSPECTED BASAL CELL CARCINOMA -.". SUSPECTED SQUAMOUS CELL CARCINOMA BIOPSY YES SKIN BIOPSY PERFORMED NO Y N BIOPSY LOCATION(S): YES CONSENT FORM OBTAINED Y NO N BIOPSY DIAGNOSIS: PHYSICAL FEATURES EYE COLOR BROWN HAZEL GREEN GREY BLUE HAIR COLOR SKIN COLOR ;T; •'*) -i: ;4.-..s; a" 7" KM) 10: .20! 30) '401 .BOI Wl fa S& 3$ «' i- •.? 'I! ® & •'•) 0rD ( PRINTED NAME OF EXAMINING PHYSICIAN SIGNATURE NCS Trans-Optic ! EP01-21160:321 C-10 A8900 �CASE NUMBER NAME: DATE OF BIRTH: MO TODAY'S DAY DATE YR BROUP I MAR APR WAV JUM -it* \Q '?) '?' (§) g) ^ §) |) l WOV '§) HE« >"C' 87J ($CT AU£ 101 sSl 36) FE8 20! JAM .961 .10 X- :|) • 4, .3, i;£) 1 ;£) X . ® 0 CD -X1 •'«) *'*><?} (i! ••£'0 ' i (f) O .(ji (D0 FORM AFHS-5 VERSION 1.0 JRW:SCF 585 HEAD AND NECK YES NO ® ® NORMAL TO PALPATION/INSPECTION 0 0 © NEUROLOGICAL EXAMINATION NECK RANGE OF MOTION O NORMAL O ABNORMAL O REFUSAL NO 0 ® ASYMMETRY ® DEPRESSION ® SCAR YES ® ® ® ® ,Y> 0 0 IMOTOR SYSTEM HANDEDNESS O LEFT I^^^^^^^^^^H ASSOCIATED MOVEMENT ARM SWING GAIT O NORMAL O ABNORIV AL O REFUSED O RIGHT ® 0 0 O ® BROAD BASED ® SMALL STEPPED (S)ATAXIC DECREASED LEFT DECREASED RIGHT DECREASED FORWARD DECREASED BACKWARD T O NORMAL O ABNORMAL LEFT O ABNORMAL RIGHT (SPECIFY) REFUSED MUSCLE STATUS BULK TONE: UPPER EXTREMITIES LOWER EXTREMITIES STRENGTH: DISTAL WRIST EXTENSORS ANKLE/TOE FLEXORS PROXIMAL DELTOIDS HIP FLEXORS O REFUSED ABNORMAL MOVEMENTS NORMAL ABNORMAL O O O O O O O O INCREASED LEFT RIGHT O O O O ONO OYES > 0®@0 TENDERNESS ONO OYES —> 0 ® ® ® TREMOR ONO OYES-I i LOWER UPPER LOWER UPPER LEFT LEFT RIGHT RIGHT O O O O O O OTHER —» O REFUSED EQUIUBRATORY (ROMBERG) FINGER - NOSE - FINGER HEEL -KNEE' -SHIN RAPIDLY ALTERNATING MOVEMENTS PRONATION/SUPINAT10N OF HANDS RAPID PATTING DEEP TENDON REFLEXES O O O ABNORMAL RIGHT BOTH O O O O O O O O O O O LEFT O O O O (0 - ABSENT, 1 - SLUGGISH, 2 - ACTIVE, 3 - VERY ACTIVE, 4 - THAN SIENT CLONUS, 3 - SUSTAINED CLONUS) EXTREMITY O O O O O O : COORDINATION NORMAL TICS, CHOREAS, FASCICULATIONS DECREASED LEFT RIGHT O O O O O O O O O O O O O O O O O O BICEPS TRICEPS PATELLAR ACHILLES BABINSKI LEFT ® © ® ® ® ® ® ® <I) ® ® © ® ® ® ® © ® ® ® O PRESENT O ABSENT O RESTING NORMAL O RESTING ABNORMAL O ESSENTIAL NORM. SKILLED ACTS O ESSENTIAL AB. O INTENTION NORM. SPEECH (ARTICULATION, APHASIA, AGNOSIA) O INTENTION AB. O NORMAL O GROSSLY ABNORMAL ® ® ® ® RIGHT (« Q) © © ® ® ® 0 (< ) 0 ® ® © ® £)PRESENT O ABSENT O DYSARTHRIA O APHASIA O AGNOSIA NCS Trans-Optic s> EP01-21169:321 c-ii A8900 �1 MENINGEAL IRRITATION AND SENSORY SYSTEM •^^^•^•^••^••^^^^^^^^^^^^^^^^^^^^^"^^^ ••" ' " — - ABNORMAL LEFT STRAIGHT LEG RAISING RIGHT O NORMAL 0 O UGHT TOUCH O PINPRICK VIBRATION AT ANKLE (128 HZ) O O O O POSITION (GREAT TOE) O o O O 0 0 0 BOTH 0 o . '"0 0 ' O O COMMENT NERVE STATUS (TENDERNESS, TUMORS, ETC.): It^^mm^^m CRANIAL NERVES • • • I B * H M M MM M ^ RIGHT LEFT O 0 o ABSENT 'scm "o o £'/' 0' ; 'b o am Io 0 f? B$I' ABNORMAL ^.asjc'wuttpR/ATfipw . EXUDATE f . -, Pa. -.;>0 "o •*i ,o o SENSORY NORMAL o o © puwts ?*<i" OROUAL • ., :..v : O o o o SS^jJBBSj? ®£K J& "d" :I '":-.0?'V. * >»vMTV ^ ' o" o - -O •' ''X3' -"~ ~ "0" (DBAW POSITION NYSTAGMUS VERTICAL ^-Ri«iP(K^;.FISSUlfeMpllli»a^^|^i^^? PALPEBRAL FISSURE ABNORMAL MOVEMENT DEVIATED i NYSTAGMUS HORIZONTAL o 0 FORM) • : ' PAUtfAfc'REFt^NORSiM^ PALATAL REFLEX ABNORMAL " i i i ON CONTINUATION PTOSIS PRESENT o • O'"; d" 0 • ' '^ •- ' 0 ^''TONGUE:PfKnlRUO£D»iMjaU^ ATROPHY OF TONGUE i '' • "„ . " . , • ' * , J• ' ' "" •.'" '"^"'": UGHT REACTION ABNORMAL '•-• 'Q. i NO YES PTOSIS ABNORMAL -.'-•0-' • : ^UOHrtReACTON NORMAL MENTAL STATUS (ALERT, CLEAR, COOPERATIVE, ETC.) EYEBALL POSmON NORMAL 'o SMILE ABNORMAL XII NYSTAGMUS ROTARY. (SROSS ABNORMALITIES? O NO SPECIFY : CORNEAL REFLEX ABNORMAL IX, XI PALATE AND UVULA 0 NYSTAGMUS ABSENT ROUND *,°r* ^^•'iSmjBB *'•&&;} ;;..;{*[• utHcR- 0 ; ... ay»£&'fiEn£$Mam^J;:i^*,'^ VII O O EYE MOVEMENT ABNORMAL SENSORY ABNORMAL >'^O»^' p::S*^;p9«p^!?0S|ftf|i^^^^^^^ ':'-': PUPIL SHAPE/POSITION O O OoNEQUAt D*f6RENC?(WMt;^®©®®^, EYE MOVEMENT NORMAL ; @@© •••":- III, IV, VI o o o : ; CLENCH JAW DEVIATED O NORMAL FIELDS ABNORMAL O Q 0 O O TQ • ' "O >Q-0 O 1 -f rw^tuMM^ ::^:- . ;::;5"^" ONORMAK:!-' •r-.'O"'•?.; O"" PI " r, fV' *£Qsl-, NORMAL HEMORRHAGE O NORMAL '??."""0!'/5: f?ifeDSiNORMAL (TO CONFRONTATION> !TOK O CLENCH JAW SYMMETRIC 0© FUNDU6 ;:.'' O o o PRESENT II ?!Cjtst RIGHT O SENSE OF SMELL o- V LEFT 1 , O YES 1 ( i 1 i i EYEBALL POSITION ABNORMAL i PRINTED NAME OF EXAMINING PHYSICIAN SIGNATURE I I I YES O NO O ADDITIONAL COMMENTS ENTERED ON FORM AFHS 7 C-12 �CASE NUMBER 0 i 2 3 4 5 8 7 3 9 0 1 2 3 4 3 8 7__ I 9 C 0 i 3 4 5 8 7 8 9 ^l 2 8 0 1 2 3 4 5 8 7 '8 9 0 1 2 3 4 5 6 7 8 £ 0 1 2 3 4 S 8' 7 3 f 0.'! VERSION 1.0 JRW SCF 385 TRAILS A TRAILS B 100 200 300 400 500 wo zoo 300 400 sw 10 20 30 40 50 60 70 80 90 to 20 30 40 50 60 70 30 90 f LOGICAL l '1 . IMMEDIATE 2 3 4 5 6 7 8 2 3 4 0 9 MEMORY 10 20 MEMORY 0 1 5 6 7 8 1 2 3 IMMEDIATE 5 3 7 3 9 10 20 F O R 9 4 0 1 2 3 4 5 6 7 8 9 6 7 3 9 DESIGNS 30' DELAY 30' DELAY 10 20 10 20 0 FIVE 1 3 2 3 4 5 9 'b 20 30 4a . 0 1 2 TOTAL 3 4 5 6 7 BO 10 SIX THREE 10 20 30 40 SEVEN 10 20 1 2 3 2 50) 4 5 6 7 3 9 0 1 2 3 4 5 6 7 8 5 7 0 9 4 «!! 10 20 30 40 0 1 30) 200 0 20. 10 20 30 40 TWO 6 80 10 20 30 40 7 90 FOUR 3 1 T P T WIN 9 °™ 0 1 2 3 4 5 6 7 3 9 S E C 0 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1S MEMORY 10 20 30 4O50 0 WIN SEC TOTAL 0 1 ° ' 2 3 2 3 . 2 3 4 5 8 DOMINANT 1 7 3 9 10 4 5 4 8 7 8 9 1 0 2 0 : 30 4 O 3 0 LOG 0 1 8 MIN NON DOMINANT SEASHORE RHYTHM * CORRECT SPEECH PERCEPTION 0 1 * & 30 * OF ERRORS 3 4 9 9 7 3 • 1 2 3 4 5 8 7 8 DOMINANT 9 10 » 30 40 50, 60 70 30 90 NQN DOMINANT 2 5 8 4 7 8 9 0 1 2 3 4 5 8 7 3 9 o RANK 9 ^ DOMINANT 2 3 4 PRINTED NAME OF EXAMINER 3 4 5 10 i 2 3 4 5 6 7 s 9 10 DOMINANT IS: 5 6 7 8 5~ 6 LEFT " RIGHT 9 ,0 20, 30 4« 50. 60. 70. 80 » 1 2 4 9 i 2 3 4 5 6 7 3 9 10 20 30 40 so so 70 80 90 1 3 3 10 11 12 13 14 1! TAPPING 10 20 30 40. so so ?o ao so 1 3 10 20 30 4» so 50 GRIP STRENGTH DOMINANT 2 2 7 7 I9 IMPAIRMENT INDEX .0 .1 1 .3 .4 .5 ii ,7 '.I' 9 1.0 SIGNATURE NCS Trans-Optic • EP01-21155:321 C-13 A3900 �CASE NUMBER B NAME OF PARTICIPANT LAST FIRST Ml CONT NUATION OF FORM AFHS-7 CONTINUATION FORM AFHS- • USE THIS FORM TO RECORD COMMENTS, OBSERVATIONS, OR PHYSICAL FINDINGS FOR WHICH THERE IS INADEQUATE SPACE ON THE OPTICAL MARK SENSE FORMS. PLEASE INDICATE THE ORGAN SYSTEM, IF APPROPRIATE, IN EACH COMMENT BLOCK. ORGAN SYSTEM: COMMENTS: ORGAN SYSTEM: COMMENTS: ORGAN SYSTEM: COMMENTS: ORGAN SYSTEM: COMMENTS: PRINTED NAME OF EXAMINING PHYSICIAN SIGNATURE C-14 DATE �CASE NUMBER 0 1 2 * 4 5 6 7 3 9 * 0 1 2 3 4 5 6 7 S * C 3 4* 0 1 2 GROUP * NAME: DATE OF BIRTH: 8 7 8 9 0 ,€} T; 2 3 4 5 3 7 " a ;£i ,'E. jHiX'T ?, * ? 3 7 'a sj 'p. ft (Tj 'I;' 5) 4) ''»'.- «... 7 8';• 9'; • 0 VERSION 1.0 JRW:SCF 585 INSTRUCTIONS FES 30 MAR 20 MO TODAY'S DAY DATE JUN JU(i AUQ 7 a S 90: APR MAY 3 10 91. 87 YR SEP OCT NOV DEC 3 4 93; 94i VIETNAM COMBAT INDEX FORM AFHS-8 INSTRUCTIONS ARE INCLUDED WITH EACH QUESTION. BELOW IS AN EXAMPLE OF THE CORRECT WAY TO ANSWER EACH QUESTION. EXAMPLE: DO YOU PLAN TO DO ANY OF THE FOLLOWING NEXT WEEK? (PLEASE BLACKEN EITHER "YES" OR "NO") YES NO • •«!' VISIT A RELATIVE ^••GO TO A MUSEUM TO A MOVIE (I WILL VISIT A RELATIVE AND GO TO A MOVIE NEXT WEEK) AIRCRAFT PLEASE INDICATE WHETHER YOU SERVED OR FLEW IN ANY OF THE FOLLOWING AIRCRAFT WHILE IN VIETNAM: (DO NOT INCLUDE TRANSPORTATION TO OR FROM VIETNAM) YES NO f ^_mm ®YES ®NO YES NO YES NO N)C-7 M. F-4 WERE YOU EVER A CREW MEMBER? ®C-54 ®C-118 ®C-123 ®C-130 ®F-105 ($8-52 ®B-66 ® ® HELICOPTER GUNSHIP 0® OTHER AIRCRAFT SPECIFY 3 EXPERIENCES SELOW IS A LIST OF DIFFERENT COMBAT ROLES AND FLYING EXPERIENCES THAT AIR FORCE PERSONNEL HAD DURING THE VIETNAM WAR. FOR EACH STATEMENT, PLEASE BLACKEN THE "YES" CIRCLE IF YOU HAD THAT EXPERIENCE DURING THE VIETNAM WAR OR THE "NO" CIRCLE IF YOU DID NOT. PLEASE BLACKEN EITHER "YES" OR "NO" FOR EACH EXPERIENCE. YES NO YES NO COMBAT PAY ®®FtEW IN AN AIRCRAFT THAT RECEIVED BATTLE DAMAGE '. :: / /: ® ® RECEIVED INCOMING ARTILLERY OR ROCKET FIRE AT HOME BASE OR CAMP ®®ENCOUNTERED MINES OR-BOOBY TRAPS :.;: • ® ® CRASH LANDED, BAILED OUT, OR SHOT DOWN f <3>® RECEIVED SNPER OR SAPPER FIRE IN. OR I ;"•'•;•'' AROUND BASE @ ® MOVED KILLED OR WOUNDED PERSONNEL ® ® KILLED VC OR NVA IN STRAFING OR BOMBING RUNS ®®WOUNDED: . ., ..-. "•-..:: •,": ;.;-: - ©® SERVED AS A FORWARD AIR CONTROLLER (FAC> ®®FLEW IN THE SAME AIRCRAFT WHEN FELLOW CREWMEMBER WAS WOUNDED OR KILLED ® ®FLEW IN THE SAME FORMATION OR ON 1• - - ; • • THE SAME SORTIE WHEN A FELLOW '•'•• CREWMEMBER WAS WOUNDED OR KILLED ®®HAD A CLOSE FRIEND KILLED IN ACTION VC OR IWA . ® ® CAPTURED BY THE ENEMY NCS Trent-Optic ® EP01-21134:321 C-15 AS900 �CASE NUMBER LAST FIRST NAME OF PARTICIPANT B Ml ATTACHMENT TO FORM FORM AFHS-9 ANATOMICAL CHART AFHS- RIGHT LEFT (OVER) C-16 �LEFT LEFT PRINTED NAME OF EXAMINING PHYSICIAN SIGNATURE C-17 DATE �CASE NUMBER B DATE ECG RECORDED NAME OF PARTICIPANT LAST FIRST Ml MO DAY YR FORM AFHS-10 ELECTROCARDIOGRAM REPORT ECG TECHNICIAN (INITIALS) t* 12- LEAD SCALAR ELECTROCARDIOGRAM NORMAL COMMENTS: ABNORMAL RBBB LBBB NON-SPECIFIC T-WAVE CHANGES TACHYCARDIA BRADYCARDIA ARRHYTHMIA PARTICIPANT COMPLIANCE WITH 4-HOUR ABSTINENCE: YES. No. RHYTHM STRIP INTERPRETATION OF ARRHYTHMIA: SIGNATURE PRINTED NAME OF CARDIOLOGIST C-18 DATE �CASE NUMBER B DATE OF X-RAY NAME OF PARTICIPANT MO DAY •V* FORM AFHS-11 PA CHEST X-RAY EXAMINATION X-RAY TECHNICIAN (INITIALS) RIGHT @ LEFT ANTERIOR PLEASE MARK THE LOCATION OF ANY SUSPECTED ABNORMALITY(IES) WITH AN ENCIRCLED NUMBER AND DESCRIBE BELOW. INTERPRETATION OF PA CHEST FILM NORMAL 1 ABNORMAL COMMENTS: SIGNATURE1 PRINTED NAME OF RADIOLOGIST C-19 DATE �CASE NUMBER B DATE OF ANTIGEN ADMINISTRATION LAST FIRST Ml NAME OF PARTICIPANT MO DAY YR TIME OF ANTIGEN ADMINISTRATION FORM AFHS-12 DELAYED SKIN TESTS © ADMINISTERED BY (INITIALS) RESULTS E - ERYTHEMA, MEASURED IN mm I - INDURATION, MEASURED IN mm ANTIGEN: 24-HOUR READING CANDIDA ALBICANS 1:1000 W/V MUMPS 2 CPU TRICHOPHYTON 1:1000 W/V STAPH-PHAGE-LYSATE 48-HOUR READING STAPH-6 TO 9 x 104 CPU PHAGE=0.5 TO 5 x 107 PFU TIME OF DAY THAT SKIN TESTS WERE READ SKIN TESTS READ BY (INITIALS) YES IS PARTICIPANT ON SYSTEMIC CORTICOSTEROIDS OR IMMUNOSUPPRESSANTS? NO DOSAGE: COMMENTS: PRINTED NAME OF REVIEWER SIGNATURE C-20 DATE �CASE NUMBER NAME OF PARTICIPANT B MO DAY DATE OF DIAGNOSIS YR FORM AFHS-16A OUTBRIEFIIMG CONDUCTED ICD9-CM CODE LAST YES DIAGNOSTIC SUMMARY (MEDICAL) NO CHECK ONE NEWLY PRE DIAGEXISTING NOSED DIAGNOSES BASED ON PHYSICAL EXAMS, ECG, DOPPLER. CHEST X-RAY, SKIN TESTS, AND LABORATORY STUDIES COMMENTS: PRINTED NAME OF DIAGNOSTICIAN SIGNATURE C-21 © �CASE NUMBER Ml NAME OF PARTICIPANT B MO DAY t-YR DATE OF DIAGNOSIS y OUTBRIEFINQ CONDUCTED ICD-9-CM CODE YES NO FORM AFH<! 1fiB FORM AFHS-16B CHECK ONE NEWLY DIAGNOSED PREEXISTING DIAGNOSTIC SUMMARY (PSYCHOMETRIC) DIAGNOSES BASED ON PSYCHOLOGICAL TESTING . COMMENTS: SIGNATURE PRINTED NAME OF DIAGNOSTICIAN r-?? m �APPENDIX D Study Selection and Participation Bias �TELEPHONE QUESTIONNAIRE 1. Compared to other people your age, would you say that your health is: Excellent Good Fair Poor 2. 1 2 3 4 Are you currently taking prescribed medicines for any illness or condition? Yes 1 (ANSWER A) No 2 (GO TO QUESTION 3) A. For what illness or condition are you taking prescribed medicines? 3. Within the last six months, did illness or injury keep you from going to work or from holding a job? Please do not count work around the house as a job. Yes 1 (ANSWER A and B) No 2 (GO TO QUESTION 4) A. How many days or weeks did you miss from work within the past six months? Days Missed OR Weeks Missed | | | | | | | | Not Able to Work at all 995 B. What illnesses or conditions caused you to miss work or prevented you from holding a job? D-l �Do you need the help of another person, or use of special equipment, to take care of personal needs such as eating, bathing, dressing, or getting up or around the house? [By "special equipment" we mean anything you use to help yourself, like a wheelchair, railing, walker, cane or other device, which is not usually used by most people.] Yes 1 (ANSWER A and B) No 2 (GO TO QUESTION 5) A. Is this help required because of a condition you've already mentioned, or because of some other health condition? Another condition 1 (ANSWER B) Condition 2 (GO TO already mentioned QUESTION 5) B. What illness or condition is that? 5. Now on to a different subject. Did you earn any income from any job in 1984? Yes 1 (ANSWER A) No 2 (DO NOT ANSWER A) A. What was your income in 1984? Less than $20,000 $20,000 to $40,000 More than $40,000 1 .. .2 3 NAME CASE ID I__|__|_J_|_J_| PHONEft( D-2 ) �TAHED-1. —„ ., . . -y as Efetemrined bv 1.154 Jfrriilarir <T^nB^T*! <WVQ — D B P £ V B C S K I N M M P I D B U N u S s G G P N U U C L V S E S EE E S E E M M E N R B B C —+ P U U —9 - G L L U M CC M E C G P L A T T I C Q N S S N C V V A A C H O 0 L A L K P P H 0 S C 0 O P P R R 0 A L L A T 4 T E S T T E D G T P TRTT.T .06 -.08 .01 .02 .02 .02 -.04 <.01 .02 .05 .02 .04 .01 -.10 -.01 -.01 .02 .08 -.01 -.06 -.07 .03 <.01 -.08 .14 EBP <.01 .03 .06 .02 -.04 .05 -.02 <.01 .03 <.01 .20 -.10 -.03 .03 .15 -.09 .23 .08 -.01 <.01 <.01 -.19 .12 .16 HBC .09 .07 -.04 -.09 .03 -.06 .02 .15 -.07 .12 -.05 .24 -.18 .04 -.07 .06 .15 .10 .10 .10 .05 .14 <.01 90N .06 -.01 .03 <.01 -.05 .05 .02 -.04 .06 -.04 .06 -.07 .01 -.06 <.01 .04 .10 .06 .04 -.02 .03 <.01 HMPID -.04 -.04 .02 -.04 .03 .04 .02 .12 -.03 -.01 -.IS .05 -.01 .10 .06 -.01 -.03 .03 -.01 .10 .03 BUN .22 .02 -.02 -.01 .01 <.01 .02 -.02 -.04 .07 -.03 -.01 .03 -.04 -.15 -.06 <0.1 -.09 .06 -.09 USG <.01 -.02 -.02 .06 .09 .U .03 -.08 <D.l -.05 -.01 -.07 <.01 -.12 .01 .02 -.11 -.02 -.07 PULSE <-01 .08 <.01 -.06 .11 -.10 .06 -.04 .05 -.06 .05 .07 -.08 -.02 .04 -.05 .08 -.01 NOTE -.06 .08 .12 -.06 .04 .05 .07 -.10 .30 .02 <.01 -.02 -.01 .02 .02 -.10 -.08 SEMEN -.02 -.10 .05 -.04 .04 <.10 -.02 .01 <.01 <.01 .03 -.08 <.01 <.01 .11 .02 BBC .07 .08 -.03 <.01 -.10 -.05 .10 .13 .07 -.11 .01 .20 .02 -.20 -.08 PULM .01 <.01 -.07 -.03 -.10 .14 <.01 -.04 -.04 .07 -.03 -.06 -.02 -.02 GLUC .07 .03 -.05 .06 -.10 .17 .12 -.05 -.02 .05 -.23 .B .09 BOG .01 .01 -.06 .03 -.02 -.07 .03 -.01 .03 .02 -.07 -.08 PLAT -.03 <.01 -.02 .07 .05 .09 -.05 .08 .02 .08 <.01 IQ -.06 -.07 -.04 -.08 .02 <.01 -.18 -.06 -.09 -.04 CMS -.13 .03 .04 .05 .01 -.03 .04 .12 .15 NCVA <.01 -.04 -.06 -.04 .07 .08 -.06 -.02 CHOL .03 <.01 -.04 .06 -.05 .18 .09 ALKPHOS -.03 -.01 .06 -.04 .13 .21 OOPRO .22 -.06 .08 .02 .05 ALA -.12 -.03 -.04 -.03 T4 .04 .07 .02 TEST -.13 <.01 SED .06 �TABLE D-2. Correlation Matrix of 20 Variables at First Follovup, as Determined by 1,293 Comparisons D B P TBILI DBP WBC SKIN MMPID ^ BUN USG PULSE RBC GLUC ECG PLAT CNS GGTP CHOL ALKPHOS COPRO T4 SED W B C .01 -.14 -.03 S K I N M M P I D B U N U S G P U L S E R B C G L U C E C G .01 .15 -.01 -.04 -.01 .06 .04 .05 01 .04 <.01 .05 <.01 01 .02 .07 .06 05 .14 .03 .03 .08 .09 -.11 <.01 02 <.01 -.01 .03 -.02 <.01 <.01 <.01 -.02 .03 .05 .03 .10 .09 05 02 -.01 <.01 .09 02 .01 .06 -.01 <.01 .05 .05 <.01 <.01 .10 P L A T -.08 <.01 .28 -.02 .07 -.06 -.03 -.01 .01 -.07 .01 C N S G G T P C H 0 L A L K P H 0 S .15 .05 .03 -.01 <.01 .01 <.01 <.01 .05 .03 .06 .17 -.03 <.01 -.01 -.04 .08 .03 .13 .02 -.05 .10 .06 -.05 .03 <.01 .03 .08 .01 .03 <.01 <.01 -.04 .14 .08 -.01 -.02 .22 .15 .07 .06 .08 .07 .08 .04 -.06 .05 .08 .16 .09 -.04 .06 .35 .03 C 0 P R 0 T 4 S E D .01 .03 .13 <.01 <.01 .15 -.05 .04 -.08 .10 <.01 <.01 <.01 .12 -.04 .04 .05 -.02 .13 <.01 .02 -.03 -.02 .01 .22 .01 .03 .04 .01 .03 .02 .08 <.01 .01 <.01 .18 <.01 .13 <.01 .03 .08 -.26 .16 .13 .12 .09 .17 .18 .27 .17 .05 T E S T <.01 !os .04 .01 -.08 -.17 -.05 -.05 .04 -.26 -.09 .04 .01 -.05 -.06 -.07 .04 .13 -.23 �APPENDIX E Statistical Methods �RUN: Initial screening models: Model A = all main effects Model B = Model A + all 2-way interactions Model C = Model B + all 3 way interactions involving exposure group (E) TEST: compare C with B for 3-way interactions: Is p-value > 0.05 for the global test of all 3-way interactions AND p value > 0.05 for each individual 3-way interaction? (i.e., can all 3-way interactions be dropped?) No Yes TEST: compare B with A for 2-way interactions: Is p-value > 0.05 for the global test of all 2-way interactions AND p value > 0.15 for each individual 2-way interaction? (i.e., can all 2-way interactions be dropped?) No Proceed stepwise eliminating individual effects with p-value > 0.05 (those with largest p-value first). "Initial Model" = Model A + all 2-way interactions with individual p-value <= 0.15 in Model B. I Yes •'Initial Model" = Model A. Proceed stepwise, eliminating individual effects with p-value > 0.05 (those with largest p-value first). Proceed stepwise eliminating individual effects with p-value > 0.05 (those with largest p-value first). Proceed stepwise, eliminating individual effects with p-value > 0.05 (those with largest p-value first). Yes Do all 2-way interactions drop out? I No Final model = "best model"; use for testing significance of E; use LSMEANS. "Initial Model" = Model B + all 3-way interactions with individual p-value <= 0.15 in Model D. Stratify and customize. Figure E-1. Modeling Strategy Yes Do all 3-way interactions drop out? No Stratify and customize. �Continuous Dependent Variables Use SAS® general linear models (GLM) procedure and follow the chart. Dichotomous Dependent Variable: Use BMDP®-LR (logistic regression) with MLR option, and follow the chart, If the number of covariates is huge, use ACE option for Models A, B, C. Polychotomous Dependent Variable: Use BMDP«-4F (log-linear model) adding "delta" =0.1 to each cell, and follow the chart. Use LAMB and COV options for parameter estimates when "best model" found. E-2 �APPENDIX F Exposure Index �TABLE F-l. Herbicide Orange Equivalent Gallons and Ranch Hand Manning by Month Mo./Yr. 10/61 11/61 12/61 01/62 02/62 03/62 04/62 05/62 06/62 07/62 08/62 09/62 10/62 11/62 12/62 01/63 02/63 03/63 04/63 05/63 06/63 07/63 08/63 09/63 10/63 11/63 12/63 01/64 02/64 03/64 04/64 05/64 06/64 07/64 08/64 09/64 10/64 11/64 12/64 01/65 02/65 03/65 Gallons Sprayed Pilot (Occ. 1) Navigator (Occ. 2) 0 0 0 191,426 324,216 191,426 0 0 0 0 0 334,126 334,126 0 90,879 0 0 0 0 0 174,024 259,150 0 0 339,588 377,172 942,630 121,454 363,758 755,312 56799 152,271 612,709 282,789 777,669 1,413,945 1,413,945 1,413,945 1,413,945 1,296,116 1,437,510 730,538 0 5 9 14 14 15 16 15 12 13 11 12 9 10 8 9 7 12 12 10 10 11 8 10 7 6 5 7 5 8 9 10 7 9 9 8 9 11 10 11 12 13 0 11 2 2 2 2 2 3 2 2 2 2 1 0 0 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 2 3 3 3 3 3 3 3 4 5 4 . F-l OfficerOther (Occ. 3) 0 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Enlisted Flyer (Occ. 4) EnlistedOther (Occ. 5) 0 6 7 7 7 7 6 6 5 5 5 5 5 5 4 5 5 5 5 5 4 8 8 9 9 10 6 7 7 5 6 5 5 6 5 4 4 4 6 6 6 6 0 14 20 23 23 20 14 13 7 4 5 6 6 5 5 4 4 6 6 7 7 6 4 4 6 6 6 5 4 4 2 2 2 3 3 2 2 1 1 1 1 1 �TABLE F-l. (continued) Herbicide Orange Equivalent Gallons and Ranch Hand Manning by Month Mo./Yr. 04/65 05/65 06/65 07/65 08/65 09/65 10/65 11/65 12/65 01/66 02/66 03/66 04/66 05/66 06/66 07/66 08/66 09/66 10/66 11/66 12/66 01/67 02/67 03/67 04/67 05/67 06/67 07/67 08/67 09/67 10/67 11/67 12/67 01/68 02/68 03/68 04/68 05/68 06/68 Gallons Sprayed Pilot (Occ. 1) Navigator (Occ. 2) 659,841 1,767,431 0 942,630 26,500 44,650 78,850 106,900 148,525 152,450 129,150 135,600 141,050 183,900 191,830 112,300 192,050 213,970 122,040 164,800 212,100 202,360 363,830 285,400 208,300 251,320 335,860 253,884 162,895 298,615 265,335 372,425 383,605 333,595 27,450 48,200 307,740 336,300 226,325 14 15 16 19 19 22 23 24 23 21 22 21 22 21 20 21 26 28 34 41 45 49 59 51 50 53 55 51 63 60 55 55 58 54 65 69 72 75 77 3 4 4 4 4 4 4 6 5 6 6 4 5 6 6 8 8 9 8 8 9 9 13 13 14 15 13 15 13 18 19 17 18 19 19 20 20 18 18 OfficerOther (Occ. 3) 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 3 4 5 5 5 4 4 4 3 3 4 5 5 6 6 6 6 5 6 6 6 Enlisted Flyer (Occ. 4) 6 6 7 7 7 6 6 10 11 10 10 10 10 9 10 9 11 12 16 18 28 28 28 28 33 34 36 37 32 33 36 33 34 33 35 34 36 32 37 EnlistedOther (Occ. 5) 2 2 4 3 3 3 6 12 12 16 26 32 37 38 41 45 46 62 85 104 123 123 116 114 108 101 105 163 160 161 149 145 129 127 141 160 161 160 164 �TABLE F-l. (continued) Herbicide Orange Equivalent Gallons and Ranch Hand Manning by Month Mo./Yr. 07/68 08/68 09/68 10/68 11/68 12/68 01/69 02/69 03/69 04/69 05/69 06/69 07/69 08/69 09/69 10/69 11/69 12/69 01/70 02/70 03/70 04/70 05/70 06/70 07/70 08/70 09/70 10/70 11/70 12/70 01/71 02/71 03/71 04/71 05/71 06/71 07/71 08/71 09/71 10/71 Gallons Sprayed 258,100 289,160 216,300 72,250 189,100 218,750 264,450 197,450 356,500 339,800 353,800 383,533 287,425 299,100 206,800 181,000 205,100 276,900 186,350 152,100 153,730 45,700 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pilot (Occ. 1) 84 91 89 89 101 94 98 91 90 94 93 88 91 85 83 83 90 76 66 58 59 54 51 47 44 40 40 34 30 25 23 23 23 23 23 28 29 29 29 29 Navigator (Occ. 2) 19 18 22 20 17 17 19 18 17 20 19 19 16 16 15 17 16 16 15 15 13 13 14 14 11 9 7 6 5 4 4 4 4 4 4 4 4 4 4 4 OfficerOther (Occ. 3) 7 9 8 8 7 8 7 5 5 6 6 7 6 6 6 6 6 5 5 5 5 5 5 3 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Source: Baseline Morbidity Report, 24 February 1984. Enlisted Flyer (Occ. 4) EnlistedOther (Occ. 5) 42 45 44 49 53 51 51 51 53 54 54 57 55 55 61 61 60 52 54 41 39 37 29 18 16 14 13 14 15 13 14 14 14 14 14 14 14 14 14 14 187 192 147 155 153 154 • 154 166 172 161 151 155 152 155 142 122 118 114 116 122 125 109 94 84 74 63 43 37 35 30 28 28 28 28 28 28 28 28 28 28 �APPENDIX G General Health �APPENDIX G: General Health Contents G-l Summary Statistics for General Health Covariates by Group G-l G-2 Unadjusted Analysis for Self-Perception of Health by Group (Original Comparisons Only) G-l G-3 Unadjusted Analysis for Appearance of Acute Illness or Distress by Group (Original Comparisons Only) G-2 G-4 Unadjusted Analysis for Appearance of Relative Age by Group (Original Comparisons Only) G-2 G-5 Adjusted Relative Risks of Appearance of Relative Age by Occupation (Original Comparisons Only) G-3 G-6 Unadjusted Analysis for Sedimentation Rate by Group (Original Comparisons Only) G-3 G-7 Unadjusted Analysis for Percent Body Fat by Group (Original Comparisons Only) G-4 G-i �TABLE G-l. Summary Statistics for General Health Covariates by Group Group Ranch Hand Comparison Covariate Covariate Category Percent Percent Race Black Nonblack 5.9 94.1 6.4 93.6 0.673 Occupation Officer Enlisted Flyer Enlisted Groundcrew 37.4 17.4 45.2 37.4 16.2 46.3 0.727 p-Value Mean±SE Age (at Baseline) Personality Score 43.94±0.24 — 43.85±0.22 0.789 -1.18±0.29 — Meant SE -1.56±0.25 0.332 — Covariate not categorized for these results. TABLE G-2. Unadjusted Analysis for Self-Perception of Health by Group (Original Comparisons Only) Self-Perception of Health Excellent Group Good Pair Poor Number Percent Number Percent Number Percent Number Percent Total Ranch Hand 490 48.2 434 42.7 74 7.3 18 1.8 1,016 Original Comparison 490 51.3 403 . 42.2 55 5.8 7 0.7 0.7 955 p=0.075 G-l �TABLE G-3. Unadjusted Analysis for Appearance of Acute Illness or Distress by Group (Original Comparisons Only) Acute Illness or Distress Yes Group No Number Percent Number 4 0.4 1,010 Ranch Hand Percent Total p-Value* 1,014 99.6 0.340 Original Comparison 6 0.6 949 955 99.4 *Fisher's exact test, 1-sided. TABLE G-4. Unadjusted Analysis for Appearance of Relative Age by Group (Original Comparisons Only) Appearance of Relative Age Younger Group Ranch Hand Same Older Number Percent Number Percent Number Percent Total 16 1.6 957 94.3 42 4.1 p-Value 1,015 0.454 Original Comparison 9 0.9 906 94.9 G-2 40 4.2 955 �TABLE G-5. Adjusted Relative Risks of Appearance of Relative Age by Occupation (Original Comparisons Only) Adjusted Relative Risk 95% C.I. p-Value Officer 3.27 (0.67,15.8) 0.142 Enlisted Flyer 0.47 (0.20,1-12) 0.089 Enlisted Groundcrew 1.17 (0.65,2.08) 0.603 Occupation TABLE G-6. Unadjusted Analysis for Sedimentation Rate by Group (Original Comparisons Only) Sedimentation Rate Normal <20 mm/hr Group Ranch Hand Abnormal >20 mm/hr Number Percent Number Percent Total 957 94.2 59 5.8 p-Value 1,016 0.059 Original Comparison 918 96.1 37 G-3 3.9 955 �TABLE G-7. Unadjusted Analysis for Percent Body Fat by Group (Original Comparisons Only) Percent Body Fat Lean/Normal <25% Group Ranch Hand Number Percent 831 81.8 Obese >25% Number Percent 185 18.2 Total p-Value 1,016 0.230 Original Comparison 759 79.6 195 G-4 20.4 954 �APPENDIX H Malignancy �APPENDIX H: Malignancy Contents Table H-l H-2 H-3 H-4 H-5 H-6 H-7 H-8 H-9 Unadjusted Analyses of Followup Participants with Verified and Suspected Neoplasms in the Baseline-Followup Interval by Group (Nonblacks and Blacks, Original Comparisons Only).. H-l Unadjusted Analyses of Nonblack Followup Participants with Verified and Suspected Malignant Skin Neoplasms in the Baseline-Followup Interval by Cell Type and Group (Original Comparisons Only) H-3 Association Between Baseline-Followup Interval Basal Cell Carcinoma Incidence and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants H-4 Association Between Baseline-Followup Interval Sun Exposure-Related Skin Malignancy Incidence and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants H-9 Summary Results of Main Effects Models for Selection of Covariates for Basal Cell Carcinomas in the Baseline-Followup Interval , H-14 Adjusted Analyses of Nonblack Followup Participants for Malignant Skin Neoplasm Incidence During Baseline-Followup Interval (Original Comparisons Only) H-15 Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Followup Interval (Nonblacks Only) H-16 Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Followup Interval (Original Comparisons Only) H-18 Followup Participants with Verified Malignant Systemic Neoplasms in Baseline-Followup Interval by Group (Original Comparisons Only) H-22 H-10 Unadjusted Analyses of Followup Participants with Verified and Suspected Malignant Systemic Neoplasms in the Baseline-Followup Interval by Group (Original Comparisons Only) H-23 H-ll Association Between Baseline-Followup Interval Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for the Combined Followup Ranch Hand and Comparison Groups H-24 H-i �APPENDIX H: Malignancy Contents (continued) Table H-12 Adjusted Analyses of Followup Participants for the Incidence of All Malignant Systemic Neoplasms During the Baseline-Followup Interval (Original Comparisons Only)...... H-28 H-13 Summary of Group-by-Occupation Interaction for All Malignant Systemic Neoplasms (Verified Plus Suspected) During the Baseline-Followup Interval H-29 H-14 Summary of Group-by-occupation Interaction for All Malignant Systemic Neoplasms (Verified Plus Suspected) During the Baseline-Followup Interval (Original Comparisons Only) , H-30 H-15 Unadjusted Analyses of Followup Participants with Lifetime Occurrence of Verified and Suspected Lifetime Neoplasms by Group (Nonblacks and Blacks, Original Comparisons Only) . H-31 H-16 Unadjusted Analyses of Nonblack Followup Participants with Lifetime Occurrence of Verified and Suspected Lifetime Malignant Skin Neoplasms by Cell Type and Group (Original Comparisons Only) H-33 H-17 Association Between Lifetime Incidence of Sun-Exposure Related Skin Malignancies and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants.... H-34 H-18 Summary of Group-by-Covariate Interactions for Lifetime Malignant Skin Neoplasm Incidence (Nonblacks Only) H-39 H-19 Adjusted Analyses of Nonblack Followup Participants for Lifetime Malignant Skin Neoplasm Incidence (Original Comparisons Only) H-40 H-20 Summary of Group-by-Covariate Interactions for Lifetime Malignant Skin Neoplasm Incidence (Original Comparisons Only*) H-21 Summary of Followup Participants with Lifetime Incidence of Verified Malignant Systemic Neoplasms by Group (Original Comparisons Only) H-22 Unadjusted Analyses of the Lifetime Incidence Rates of All Malignant Systemic Neoplasms Combined, by Group (Original Comparisons Only) H-23 Summary of Group-by-occupation Interactions for Lifetime Incidence of All Malignant Systemic Neoplasms Combined H-ii H-41 H-45 H-46 H-47 �APPENDIX H: Malignancy Contents (continued) Table H-24 Adjusted Analyses for Lifetime Incidence of All Malignant Systemic Neoplasms Combined (Original Comparisons Only) H-25 Summary of Group-by-Occupation Interactions for Lifetime Incidence of All Malignant Systemic Neoplasms Combined (Original Comparisons Only) H-49 H-50 H-26 Unadjusted Exposure Index Analyses for Followup Participants for Occurrence of Malignant Neoplasms in the Baseline-Followup Interval H-52 H-27 Unadjusted Exposure Index Analysis for Followup Participants for Lifetime Occurrence of Malignant Neoplasms ". H-55 H-iii �TABLE H-l. Unadjusted Analyses of Followup Participants with Verified and Suspected Neoplasms in the Baseline-Follovup Interval by Group (Nonblacks and Blacks, Original Comparisons Only) Group* Ranch Hand Site Skin Neoplasm Behavior and Status Original Comparison Number** Percent Number** Percent Total** p-Value*** 37 47 3.6 4.6 31 45 3.2 4.7 68 92 0.711 0.999 Benign Verified Verified and Suspected 76 78 7.5 7.7 65 68 6.8 7.1 141 146 0.600 0.667 Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected 1 1 0.1 0.1 1 1 0.1 0.1 2 2 0.999 0.999 Any Skin Neoplasm* Verified Verified and Suspected Systemic Malignant Verified Verified and Suspected 114 124 11.2 12.2 96 113 10.1 11.8 210 237 0.422 0.835 Malignant Verified Verified and Suspected 8 12 0.8 1.2 6 11 0.6 1.2 14 23 0.791 0.999 Benign Verified Verified and Suspected 42 48 4.1 4.7 41 49 4.3 5.1 83 97 0.911 Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected 6 6 0.6 0.6 6 9 0.6 0.9 12 15 0.999 0.442 Any Systemic Neoplasm1* Verified Verified and Suspected 55 65 5.4 6.4 51 66 5.3 6.9 106 131 0.999 0.652 H-l 0.679 �TABLE H-l. (continued) Unadjusted Analyses of Followup Participants with Verified and Suspected Neoplasms in the Baseline-Followup Interval by Group (Nonblacks and Blacks, Original Comparisons Only) Group* Ranch Hand Site All Neoplasms Neoplasm Behavior and Status Malignant, Benign, Uncertain Behavior. Unspecified Nature Verified Verified and Suspected Original Comparison Number** Percent Number** Percent 161 177 15.9 17.4 140 169 14.7 17.7 Total** p-Value*** 301 346 *Sample sizes: 1,016 Ranch Hands, 955 Original Comparisons. **Number of participants. ***Fisher's exact test. "Participant has one or more malignant, benign, or unspecified skin neoplasms. b Participant has one or more malignant, or benign skin or systemic neoplasms. c Participant has one or more malignant, benign, or unspecified systemic neoplasms. H-2 0.491 0.906 �TABLE B-2. Unadjusted Analyses of Nonblack Follovup Participants with Verified and Suspected Malignant Skin Neoplasms in the Baseline-Followup Interval by Cell Type and Group (Original Comparisons Only) 4 Group** Cell Type Status Statistic* Ranch Hand Original Comparison Est. Relative Risk (95Z C.I.) p-Value Verified Number/Z Total Neoplasms 29 42 3.0Z 22 31 2.5Z 1.24 (0.71,2 .18) 0.480 Verified & Suspected Number/Z Total Neoplasms 36 53 3.8Z 36 46 4.0Z 0.94 ( . 8 1 . 0 05, 5) 0.811 Verified Number/Z Total Neoplasms 4 6 0.4Z 3 3 0.3Z 1.25 ( . 8 5 .61) 02, 0.999 Verified & Suspected Number/Z Total Neoplasms 4 6 0.4% 3 3 0.3Z 1.25 (0.28,5 .61) 0.999 Verified Number/Z Total Neoplasms 1 2 0.1Z 3 3 0.3Z 00, 0.31 ( . 3 3 .01) 0.359 Verified & Suspected Number/Z Total Neoplasms 1 2 0.1Z 4 4 0.4Z 0.23 (0.03,2 .10) All Malignant Verified Skin Neoplasms Verified & Suspected Number/Z Total Neoplasms 37 56 3.9Z 31 41 3.5Z 1.13 (0.69,1 .83) 0. 711 Number/Z Total Neoplasms 47 70 4.9Z 45 58 5.0Z 0.98 ( . 4 1 .49) 0.999 06, Sun-Exposure Related Malignant Neoplasms' Verified Number/Z Total Neoplasms 32 47 3.4Z 25 33 2.8Z 0) 1.21 (0.71,2 . 6 Verified & Suspected Number/Z Total Neoplasms 39 58 4.1Z 39 50 4.4Z 0.94 (0.59,1 .47) 0.817 Basal Cell Carcinoma Squamous Cell Carcinoma Melanoma *Huraber and percent of participants; total number of malignant neoplasms of specified cell type. **Number of participants — 956 Ranch Hands, 897 Original Comparisons. *Basal cell carcinoma, melanoma, and malignant epithelial neoplasms NOS. 0. 204 0..504 �TABLE H-3. Association Between Baseline-Follovup Interval Basal Cell Carcinoma Incidence and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Age Born >1942 Born 1923-41 Born <1922 11 43 5 1.3 3.6 5.8 0.001 14 61 9 1.6 5.1 10.3 <0.001 Occupation Officer Enlisted Flyer Enlisted Groundcrew 31 11 17 3.7 3.1 1.8 0.047 47 14 23 5.5 3.9 2.4 0.003 Total Lifetime Smoking (Pack- Years) 0 >0-20 >20-40 >40 16 21 12 10 2.6 2.1 3.1 6.4 0.023 21 28 24 11 3.4 2.8 6.1 7.0 0.005 Total Lifetime Alcohol Consumption (Drink- Years) 0 >0-5 >5-30 >30-100 >100 4 20 18 13 4 2.8 2.8 2.8 2.7 3.9 0.980 5 27 24 24 4 3.6 3.8 3.7 5.0 3.9 0.800 Ethnic Background3 A B C D 49 8 1 0 3.1 1.9 1.6 0.0 0.333 72 9 1 0 4.6 2.1 1.6 0.0 0.046 Skin Color Dark Medium Pale Dark Peach Pale Peach 0 1 3 43 12 0.0 1.4 1.0 3.4 2.3 0.156 0 2 5 56 21 0.0 2.7 1.6 4.4 4.0 0.232 �TABLE H-3. (continued) Association Between Baseline-Followup Interval Basal Cell Carcinoma Incidence and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Eye Color Brown Hazel Green Grey Blue 18 17 2 2 20 2.8 3.7 1.7 2.2 2.4 0.590 24 18 5 4 33 3.7 4.0 4.2 4.3 3.9 0.998 Hair Color Black Dark Brown Light Brown Red Blond 11 25 18 2 3 2.5 2.4 3.2 12.5 2.8 0.152 16 37 24 3 4 3.6 3.6 4.3 18.8 3.7 000 .4 Residential History (Average Latitude) >37° <37° 21 38 1.9 3.7 008 .0 31 53 2.7 5.2 0.004 Skin Reaction to First 30 Min. of Sun Exposure Burns Usually Burns Burns Mildly Rarely Burns 7 24 17 11 2.8 5.6 2.1 1.6 0.001 11 32 22 19 4.5 7.5 2.7 2.8 <0.001 Skin Reaction to >2 Hrs. of Sun After First Exposure Burns Painfully Burns Becomes Red No Reaction 2 21 22 14 1.7 6.2 2.1 2.1 <0.001 4 23 35 22 3.3 6.8 3.4 3.3 0.027 Skin Reaction After Repeated Exposure to Sun Freckles, No Tan Tans Mildly Tans Moderately Tans Deep Brown 0 20 22 16 0 6,4 2.2 2.0 <0.001 1 25 33 24 2.2 8.0 3.2 3.1 0.001 PC I Cn �TABLE H-3. (continued) Association Betveen Baseline-Pollowup Interval Basal Cell Carcinoma Incidence and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Sun Reaction Index Tends to Burn. Mild Reaction Tends to Tan Exposures to Carcinogens Number* Percent Verified and Suspected p-Value Number* Percent p-Value 1.4 6.0 1.9 <0.001 4 32 47 2.8 7.1 3.0 <0.001 Asbestos Yes 10 No 49 2.2 2.9 0.519 17 67 3.7 3.9 0.999 Nonmedical x rays Yes 17 No 42 3.4 2.5 0.272 26 58 5.3 3.5 0.084 Industrial Chemicals Yes 28 No 31 2.5 3.0 0.511 40 44 3.6 4.2 0.437 Herbicides Yes 40 No 19 3.2 2.1 0.142 57 27 4.5 3.0 0.072 Insecticides Yes 40 No 19 3.1 2.2 0.281 55 29 4.2 3.4 0.365 Degreasing Chemicals Composite Carcinogen Exposure 2 27 29 Yes 35 No 24 2.8 2.7 0.894 47 37 3.7 4.1 0.735 Yes 15 No 42 3.1 2.5 0.523 18 64 3.7 3.9 0.999 �TABLE H-3. (continued) Association Between Baseline-Follovup Interval Basal Cell Carcinoma Incidence and the Covariates for Combined Pollovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value ~~ Yes No 0 59 0.0 2.7 0.999 0 84 0.0 3.9 0.999 Yes 2 No 56 4.9 2.6 0.301 3 80 7.3 3.8 0.206 Benzene PC iJ Anthracene Arsenic Exposure to Individual Carcinogens Yes No 4 55 5.4 2.6 0.140 5 79 6.8 3.8 0.207 Benzidine Yes 1 No 58 9.1 2.7 0.263 1 83 9.1 3.9 0.354 Chroma tes Yes No 3 54 3.6 2.6 0.484 4 78 4.8 3.8 0.558 Coal Tar 1 Yes No 58 1.5 2.8 0.999 2 82 2.9 3.9 0.999 Creosote Yes 6 No 53 3.8 2.6 0.441 7 77 4.4 318 0.669 Aminodiphenyl Yes No 0 58 0.0 2.7 0.999 0 83 0.0 3.9 0.999 Chloromethyl Ether 1 Yes No 58 4.8 2.7 0.442 1 83 4.8 3.9 0.566 Mustard Gas Yes No 0 59 0.0 2. 7 0.999 0 84 0.0 3.9 0.,999 * �TABLE H-3. (continued) Association Between Baseline-Follovup Interval Basal Cell Carcinoma Incidence and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Naphthylamine Yes No 3 55 5.8 2.6 0.161 4 79 7.7 3.7 0.136 Cutting Oils Yes 9 No 50 4.0 2.6 0.200 12 72 5.3 3.7 0.271 Trichloroethylene Yes 4 No 55 2.2 2.8 0.814 6 78 3.3 3.9 0.842 Ultraviolet Light Yes 2 No 57 4.6 2.7 0.339 2 82 4.6 3.9 0.688 Vinyl Chloride Exposure to Individual Carcinogens (continued) Yes 0 No 59 0.0 2.8 0.999 1 83 3.2 3.9 0.999 sa CO *Number of participants with basal cell carcinoma. a Ethnic Background: A B C D - English, Welsh, Scottish, Irish, Scandinavian, German, Polish, Russian, other Slavic, Jewish, French. Spanish, Italian, Greek. Mexican, American Indian, Asian. �TABLE H-4. Association Between Baseline-Foliovup Interval Sun Exposure-Related Skin Malignancy Incidence and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Age 14 46 5 1.6 3.8 5.8 0.004 19 64 9 2.2 5.4 10.3 <0.001 Occupation 33 I Born >1942 Born 1923-41 Born <L922 Officer Enlisted Flyer Enlisted Groundcrew 33 12 20 3.9 3.3 2.1 0.077 50 15 27 5.9 4.2 2.8 0.006 Total Lifetime Smoking (Pack- Years) 0 >0-20 >20-40 >40 19 22 13 11 3.1 2.2 3.3 7.0 0.012 24 31 25 12 3.9 3.1 6.4 7.6 0.007 Total Lifetime Alcohol Consumption (Drink-Years) 0 >0-5 >5-30 > 30-100 >100 4 22 18 15 5 2.8 3.1 2.8 3.1 4.8 0.858 5 29 26 26 5 3.6 4.0 4.0 5.4 4.8 0.736 Ethnic Background A B C D 54 8 1 0 3.4 1.9 1.6 0.0 0.213 78 10 1 0 4.9 2.4 1.6 0.0 0.036 Skin Color Dark Medium Pale Dark Peach Pale Peach 0 1 4 48 12 0.0 1.4 1.3 3.8 2.3 0.116 0 2 6 62 22 0.0 2.7 2.0 4.9 4.2 0.213 �TABLE H-4. (continued) Association Between Baseline-Follovup Interval Sun Exposure-Related Skin Malignancy Incidence and the Covariates for Combined Follovup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Eye Color Brown Hazel Green Grey Blue 20 19 2 2 22 3.1 4.2 1.7 2.2 2.6 0.461 27 20 5 4 36 4.2 4.4 4.2 4.3 4.2 0.999 Hair Color Black Dark Brown Light Brown Red Blond 12 27 20 2 4 2.7 2.6 3.6 12.5 3.7 0.171 17 40 26 3 6 3.9 3.9 4.6 18.8 5.6 0.051 Residential History (Average Latitude) >37° <37° 24 41 2.1 4.0 0.011 34 58 3.0 5.7 0.003 Skin Reaction to First 30 Min. of Sun Exposure Burns Usually Burns Burns Mildly Rarely Burns 8 26 20 11 3.2 6.1 2.5 1.6 <0.001 13 34 26 19 5.3 7.9 3.2 2.8 <0.001 Skin Reaction to >2 Hrs. of Sun After First Exposure Burns Painfully Burns Becomes Red No Reaction 4 22 24 15 3.3 6.5 2.3 2.3 0.001 6 25 38 23 5.0 7.4 3.6 3.5 0.016 Skin Reaction After Repeated Exposure to Sun Freckles, No Tan Tans Mildly Tans Moderately Tans Deep Brown 0 22 23 19 0 7.0 2.3 2.4 <0.001 1 27 35 28 2.2 8.6 3.4 3.6 <0.001 EC 4 �TABLE H-4. (continued) Association Between Baseline-Followup Interval Sun Exposure-Related Skin Malignancy Incidence and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Sun Reaction Index Tends to Burn Mild Reaction Tends to Tan Exposures to Carcinogens Number* Percent Verified and Suspected p-Value Number* Percent p-Value <0.001 6 34 51 4.1 7.5 3.3 <0.001 Asbestos Yes 10 No 55 2.2 3.2 0.283 17 75 3.7 4.4 0.602 Yes 18 No 47 3.6 2.8 0.367 27 65 5.5 3.9 0.129 Industrial Chemicals Yes No 33 32 2.9 3.1 0.900 45 47 4.0 4.5 0.594 Herbicides Yes No 43 22 3.4 2.4 0.203 61 31 4.8 3.4 0.130 Insecticides Yes 42 No 23 3.2 2.7 0.523 58 34 4.4 4.0 0.664 Degreasing Chemicals Composite Carcinogen Exposure 2.8 6.2 2.1 Nonmedical X Rays as 4 28 32 Yes No 37 28 2.9 3.1 0.899 50 42 4.0 4.6 0.451 Yes 15 No 48 3.1 2.9 0.879 18 72 3.7 4.4 0.608 �TABLE H-4. (continued) Association Between Baseline-Follovup Interval Sun Exposure-Related Skin Malignancy Incidence and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants Verified Covariate Exposure to Individual Carcinogens Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Anthracene Yes No 0 65 0.0 3.0 0.999 0 92 0.0 4.3 0.999 Arsenic Yes No 2 62 4.9 2.9 0.343 3 88 7.3 4.1 0.246 Benzene Yes No 4 61 5.4 2.9 0.280 5 87 6.8 4. 2 0.241 Benzidine 1 Yes No 64 9.1 3.0 0.285 1 91 9. 1 4.2 0.380 Chromates Yes No 3 60 3.6 2.9 0.734 4 86 4.8 4. 1 0.777 Coal Tar 1 Yes No 64 1.5 3.1 0.721 2 90 2.9 4.3 0.999 Creosote Yes 6 No 59 3.8 2.9 0.473 7 85 4.4 4.2 0.839 Aminodiphenyl Yes No 0 64 0.0 3.0 0.999 0 91 0.0 4.2 0.999 Chloromethyl Ether Yes No 1 64 4.8 3.0 0. ,475 1 91 4.8 4.3 0.,600 Mustard Gas Yes 0 No 65 0.0 3.0 0.,999 0 92 0.0 4.3 0.999 �•' TABLE H-4. (continued) Association Between Baseline-Followup Interval Sun Exposure-Related Skin Malignancy Incidence and the Covariates for Combined Followup Ranch Hand and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value 4 Exposure to Individual Carcinogens (continued) Naphthylamine Yes 3 No 61 5.8 2.9 0.197 4 87 7.7 4.1 0.277 Cutting Oils Yes No 9 56 4.0 0.406 12 80 5.3 4.1 0.384 Yes 4 No 61 2.2 0.653 6 86 3.3 4.4 0.701 Yes No 2 63 4.6 0.384 2 90 4.6 4.3 0.711 Yes No 0 65 0.0 0.999 1 91 3.2 4.3 0.999 Trichloroethylene so Ultraviolet Light Vinyl Chloride 2.9 3.1 3.0 3.1 *Number of participants with sun exposure-related skin malignancies. "Ethnic Background: A B C D - English, Welsh, Scottish, Irish. Scandinavian, German, Polish, Russian, other Slavic, Jewish, French. Spanish, Italian, Greek. Mexican, American Indian, Asian. �TABLE H-5. Summary Results of Main Effects Models for Selection of Covariates for Basal Cell Carcinomas in the Baseline-Followup Interval Basal Cell Carcinoma in Baseline-Follovup Interval p-Value Covariate Model 1* Category Model 2b Group Ranch Hand, Comparison 0.264 0.259 Age Born XL942, 1923-1941, <1922 0.001 0.002 Occupation Officer, Enlisted Flyer, Enlisted Groundcrew 0.683 0.609 Lifetime Smoking 0-20, >20-40, >40 Pack-years 0.091 0.096 Lifetime Alcohol 0-30, >30 0.743 0.819 Ethnic Background0 A, Not A 0.391 0.425 Skin Color Peach, Not Peach 0.084 0.055 Hair Color Blond or Red, Other 0.880 0.761 Reaction of Skin to 30 Minutes Sun Burns, Usually Burns, Other 0.195 Reaction of Skin to >2 Hours of Sun Burns Painfully, Burns, Other 0.007 Reaction of Skin to Repeated Sun Exposure Freckles With no Tan, Tans Easily, Other Sun-Reaction Index Residential History <0.001 Burns With Freckles, Intermediate Reaction, Tans Easily — <37° N latitude, >37° N latitude 0.006 *Model includes three reactions of skin to sun variables. Model includes sun-reaction index. °Ethnic background: A - English, Welsh, Scottish, Irish. H-14 _ <0.001 0.008 �TABLE H-6. Adjusted Analyses of Nonblack Followup Participants for Malignant Skin Neoplasm Incidence During Baseline-Followup Interval (Original Comparisons Only) Variable Status Adj. Relative Risk (95% C.I.) p-Value Basal Cell Carcinoma Verified 1.46 (0.82,2.61) 0.198 AGE (p<0.001) LAT (p=0.015) SUNREAC (p=0.001) GRP*SUNREAC (marginal: p=0.051) **** **** GRP*SUNREAC (p=0.007) AGE (p<0.001) LAT (p=0.006) **** **** GRP*SKIN (p=0.036) AGE (p<0.001) SUNREAC (p=0.003) LAT (p=0.015) **** **** GRP*SUNREAC (p=0.030) SKIN*SUNREAC (p=0.042) AGE (p<0.001) LAT (p=0.008) Verified & Suspected Verified Sun-Exposure Skin Neoplasms Verified & Suspected Covariate Remarks* *Abbreviations: LAT: average lifetime residential latitude SUNREAC: sun reaction index GRP: group OCC: occupation SKIN: skin color ****Group-by-covariate interactions—adjusted relative, confidence interval, and p-value are not presented. H-15 �TABLE H-7. Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Followup Interval (Nonblacks Only) Group Ranch Hand re Interaction Basal Cell Group-byCarcinoma Occupation (Verified Only) Percent Number Percent Adj . Relative Risk (95% C.I.) p- Value Stratification Statistic Officer Variable Number Comparison n Yes No 372 14 358 3.8 96.2 475 16 459 3.4 96.6 1.14 (0.54,2.40) 0.726 Enlisted Flyer n Yes No 166 9 157 5.4 94.6 193 2 191 1.0 99.0 6.50 (1.36,31.01) 0.019 Enlisted Groundcretf n Yes No 416 6 410 1.4 98.6 539 11 528 2.0 98.0 0.81 (0.29,2.23) 0.680 �TABLE H-7. (continued) Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Followup Interval (Nonblacks Only) Group Ranch Hand Variable Interaction Stratification Statistic Burns Easily Intermediate Basal Cell Group-byCarcinoma Sun Reaction Reaction (Verified Index Plus Suspected) Tans Easily a Number Percent Adj. Relative Number Percent Risk (95% C.I.) p- Value n Yes No 68 0 68 0 100 77 4 73 5.2 94.8 n Yes No 192 17 175 8.9 91.1 262 15 247 5.7 94.3 1.76 (0.84,3.69) 0.134 n Yes No 693 19 674 2.7 97.3 867 28 839 3.2 96.8 0.90 (0.50,1.64) 0.731 Adjusted relative risk not calculated because of a zero cell count. Fisher's exact test. Comparison 0.121b �TABLE H-3. Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Follovup Interval (Original Comparisons Only*) Group Ranch Hand Variable Stratification Statistic Burns Easily 93 I t-» 00 Interaction Basal Cell Group-byIntermediate Carcinoma Sun Reaction (Verified Index Only) Tans Easily Original Comparison Number Percent Adj. Relative Risk (95% C.I.) p-Value Number Percent n Yes No 68 0 68 0 100 56 2 54 3.6 96.4 0.202b n Yes No 192 15 177 7.8 92.2 186 6 180 3.2 96.8 2.81 (1.05,7.55) 0.040 n Yes No 694 14 680 2.0 98.0 653 13 640 2.0 98.0 1.13 (0.52,2.43) 0.758 �TABLE H-8. (continued) Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Follovup Interval (Original Comparisons Only*) Group Ranch Hand Variable CE l-> vO Interaction Basal Cell Intermediate Carcinoma Group-by(Verified Sun Reaction Plus Index Suspected) Tans Easily Adj. Relative Percent Risk (95% C.I.) p-Value Number Percent Number n Yes No 68 0 68 0 100 55 4 51 7.3 92.7 —a n Yes No 192 17 175 8.9 91.1 185 8 177 4.3 95.7 2.38 (0.98,5.76) 0.055 n Yes No 694 19 675 2.7 97.3 646 23. 623 3.6 96.4 0.86 (0.46,1.60) 0.627 Stratification Statistic Burns Easily Original Comparison 0.038b �TABLE H-8. (continued) Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Follovup Interval (Original Comparisons Only*) Group Ranch Hand Number Percent Original Comparison Number Q3 NS 0 Percent Adj. Relative Risk (95% C . I . ) p-Value Variable Interaction Stratification Statistic Sun Exposure Related Group-byNeoplasms Skin Color (Verified Only) Peach n Yes No 768 31 737 4.0 96.0 744 20 724 2.7 97.3 0.20 (0.02,1.80) 0.150 Not Peach n Yes No 185 1 184 0.5 99.5 150 4 146 2.7 97.3 1.70 (0.95,3.04) 0.073 �TABLE H-8. (continued) Summary of Group-by-Covariate Interactions for Malignant Skin Neoplasms in the Baseline-Pollovup Interval (Original Comparisons Only*) Group Ranch Hand Original Comparison EC Interaction Statistic n Yes No 68 2 66 2.9 97.1 56 4 52 7.1 92.9 0.47 (0.08,2.74) 0.401 n Yes No 192 .18 174 9.4 90.6 186 8 178 4.3 95.7 2.74 (1.14,6.63) 0.025 n Yes No 693 19 674 2.7 97.3 652 26 626 4.0 96.0 0.75 (0.41,1.37) 0.343 Sun Exposure Related Neoplasms Group-byIntermediate (Verified Sun Reaction Plus Index Suspected) Tans Easily Number Percent *Blacks excluded. "Adjusted relative risk not calculated because of a zero cell count. b Fisher's exact test. Number Adj . Relative Risk (95% C.I.) p-Value Stratification Burns Easily Variable Percent �TABLE H-9. Followup Participants With Verified Malignant Systemic Neoplasms in Baseline-Followup Interval by Group (Original Comparisons Only) Group Site Ranch Hand Original Comparison Total Oral Cavity and Pharynx 2a ' b 0 2 Thyroid Gland 0 1 1 Bronchus and Lung 1 0 1 Colon 0 3C ' d 3 Kidney and Bladder 2 1 3 Prostate 1 1 2 Testicles 1 0 1 _1 _0 J. 8 6 14 Connective and Other Soft Tissue Total "includes one Ranch Hand with separate malignancies of tongue and epiglottis and also malignant neoplasm of bone. Includes one Ranch Hand with separate malignant neoplasms of tongue and oropharynx and secondary malignant neoplasm of other site. c Incudes one comparison with secondary malignant neoplasms of liver and bone and bone marrow. Includes one comparison with secondary malignant neoplasm of liver. H-22 �TABLE B-10. Unadjusted Analyses of Followup Participants with Verified and Suspected Malignant Systemic Neoplasms in the Baseline-^ollowp Interval by Group (Original Comarisons Only) Group Statistic Original Ranch Hand Comparison Est. Relative Risk ( 5 C.I.) 9% p-Value Verified Number/% Total Neoplasms 8 08 6 06 .% .% 12 9 1.26 ( . 3 3 6 ) 04,.3 071 .9 Verified & Suspected Number/% Total Neoplasms 12 1 2 11 1 2 1.03 ( . 5 2 3 ) .% .% 04,.4 23 14 099 .9 Status H-23 �TABLE H-ll. Association Between Baseline-Follovup Interval Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for the Combined Followup Ranch Hand and Comparison Groups Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Age l' 10 4 0.1 0.8 4.6 <0.001 4 16 4 0.4 1.3 4.6 0.001 Race en ro Born XL942 Born 1923-41 Born <1922 Nonblack Black 15 0 0.7 0.0 0.999 23 1 1.1 0.7 0.999 Occupation Officer Enlisted Flyer Enlisted Groundcrew 10 2 3 1.2 0.5 0.3- 0.056 13 4 7 1.5 1.0 0.7 0.193 Total Lifetime Smoking (Pack- Years) 0 >0-20 >20-40 >40 2 6 5 2 0.3 0.6 1.2 1.3 0.220 4 11 7 2 0.6 1.0 1.7 1.3 0.374 Total Lifetime Alcohol Consumption (Drink- Years) 0 >0-5 >5-30 >30-100 >100 0 2 3 7 1 0 0.3 0.4 1.4 0.9 0.082 1 5 5 9 2 0.7 0.7 0.7 1.8 1.9 0.233 Exposures to Carcinogens Asbestos Yes 3 No 12 0.6 0.7 0.999 5 19 1.0 1.1 0.999 Nonmedical X Rays Yes 5 No 10 0.9 0.6 0.364 10 14 1.9 0.8 0.049 �TABLE H-ll. (continued) Association Between Baseline-Follovup Interval Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for the Combined Followup Ranch Hand and Comparison Groups Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Industrial Chemicals Yes No 5 10 0.4 0.9 0.196 11 13 0.9 1.2 0.682 Herbicides Yes No 9 6 0.7 0.6 0.999 14 10 1.1 1.0 0.999 Insecticides Yes No 4 11 0.3 1.2 0.014 10 14 0.7 1.5 0.091 Degreasing Chemicals Exposures to Carcinogens (continued) Yes No 6 9 0.5 0.9 0.191 14 10 1.0 1.0 0.999 Yes No 4 11 0.8 0.6 0.757 5 18 1.0 1.0 0.999 Anthracene Yes No 0 15 0.0 0.7 0.999 0 24 0.0 1.0 0.999 Arsenic Yes No 0 15 0 0.7 0.999 2 22 4.8 1.0 0.069 Benzene Yes No 1 14 1.2 0.6 0.424 1 23 1.2 1.0 0.587 Yes No 1 14 7.1 0.6 0.088 1 22 7.1 1.0 0.131 EC N> Composite Carcinogen Exposure Exposure to Individual Carcinogens Benzidine „ �TABLE H-ll= (continued) Association Between Baseline-Follovup Interval Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for the Combined Follovup Ranch Hand and Comparison Groups Verified Covariate Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Chroma tes Yes No 2 13 2.3 0.6 0.110 2 22 2.3 1.0 0.232 Coal Tar Yes No 2 13 2.7 0.6 0.079 2 22 2.7 1.0 0.175 Creosote Exposure to Individual Carcinogens (continued) Yes No 1 14 0.6 0.7 0.999 3 21 1.8 1.0 0.240 Aminodiphenyl Yes No 0 15 0 0.7 0.999 1 23 16.7 1.0 0.061 Chloromethyl Ether Yes No I 14 4.4 0.6 0.140 2 22 8.7 1.0 0.023 Mustard Gas Yes 0 No 15 0 0.7 0.999 1 23 11.1 0.090 Naphthylamine Yes No 2 13 3.6 0.6 0.050 3 21 5.4 0.9 0.019 Cutting Oils Yes No 3 12 1.2 0.6 0.204 5 19 2.1 0.9 0.100 Trichloroethylene Yes No 3 12 1.5 0.6 0.135 4 20 2.0 1.0 0.149 CO i OS 1.0 �TABLE H-ll. (continued) Association Between Baseline-Follovup Interval Incidence of All Malignant Systemic Neoplasms Combined and the Covariates for the Combined Followup Ranch Hand and Comparison Groups Verified Covariate Exposure to Individual Carcinogens (continued) BC to Covariate Category Number* Percent Verified and Suspected p-Value Number* Percent p-Value Ultraviolet Light Yes No 1 14 2.0 0.6 0.286 I 23 2.0 1.0 0.417 Vinyl Chloride Yes No 0 15 0.0 0.7 0.999 1 23 3.0 1.0 0.294 *Number of participants with malignant systemic neoplasms . �TABLE H-12. Adjusted Analyses of Follovup Participants for the Incidence of All Malignant Systemic Neoplasms During the Baseline-Followup Interval (Original Comparisons Only) Variable Systemic Neoplasms (Verified) Systemic Neoplasms (Verified & Suspected) Adj. Relative Risk (95X C.I.) 1.47 (0.50,4.33) p-Value Covariate Remarks AGE (p<0.001) **** **** 0.481 GRP*OCC (p=0.028) AGE (p<0.001) RACE*PACKYR (p=0.032) ****Group-by-covariate interaction—adjusted relative risk, confidence interval, and p-value not presented. H-28 �TABLE H-13. Summary of Group-by-Occupation Interaction for All Malignant Systemic Neoplasms (Verified Plus Suspected) During the Baseline-Follovup Interval Group Ranch Hand Variable Percent Number 380 5 375 1. 3 98.7 484 8 476 1.7 98.3 N5 VD n Yes No 175 4 171 2. 3 97.7 209 0 209 0 100 n Yes No 457 3 454 0.7 99.3 598 4 594 0.7 99.3 Statistic n Yes No Enlisted Flyer Enlisted Groundcrev Systemic Cancer (Verified Group-byoccupation Plus Suspected) Stratification Officer . CO Interaction Number 'Adjusted relative risk not calculated because of a zero cell. 'Fisher's exact test. Comparison Percent Adj. Relative Risk (95% C.I.) p-Value 0.79 (0.25,2.44) 0.679 —a 0.042b 0.93 (0.20,4.26) 0.923 �TABLE H-14. Summary of Group-by-Occupation Interaction for All Malignant Systemic Neoplasms (Verified Plus Suspected) During the Baseline-Follovup Interval (Original Comparisons Only) Group • Ranch Hand Variable Number CO O Statistic n Yes No 380 5 375 1.3 98.7 350 7 343 2.0 98.0 Enlisted Flyers n Yes No 175 4 171 2.9 97.1 174 0 174 0.0 100.0 Enlisted Groundcrew All Systemic Cancer Group-by(Verified Occupation Plus Suspected) Stratification Officers ffi Interaction Original Comparison n Yes No 457 3 454 0.7 99.3 430 4 426 0.9 99.1 Percent 'Adjusted relative risk not calculated because of a zero cell count. 'Fisher's exact test. Number Percent Adj. Relative Risk (95% C.I.) p-Value 0.72 (0.22,2.30) 0.577 _ —a u. 0.123b 0.71 (0.16,3.28) 0.666 �TABLE H-15. Unadjusted Analyses of Followup Participants with Lifetime Occurrence of Verified and Suspected Lifetime Neoplasms by Group (Nonblacks and Blacks, Original Comparisons Only) Group Ranch Hand Site Skin Neoplasm Behavior and Status Comparison Number** Percent Number** Percent Total** p-Value*** 66 75 6.5 7.4 48 62 5.0 6.5 114 137 0.177 0.479 Benign Verified Verified and Suspected 84 86 8.3 8.5 67 70 7.0 7.3 151 156 0.310 0.360 Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected 1 1 0.1 0.1 1 1 0.1 0.1 2 2 0.999 0.999 Any Skin Neoplasm* Verified Verified and Suspected Systemic Malignant Verified Verified and Suspected 150 159 14.8 15.7 111 128 11.6 13.4 261 287 0.046 0.160 Malignant Verified Verified and Suspected 17 21 1.7 2.1 14 19 1.5 2.0 31 40 0.722 0.999 Benign Verified Verified and Suspected 51 57 5.0 ' 5.6 52 60 5.5 6.3 103 117 0.687 0.568 Uncertain Behavior and Unspecified Nature: Verified Verified and Suspected 15 15 1.5 1.5 11 14 1.2 1.5 26 29 0.560 0.999 Any Systemic Neoplasm13 Verified Verified and Suspected 81 91 8.0 9.0 71 86 7.4 9.0 152 177 0.673 0.999 H-31 �TABLE H-15. (continued) Unadjusted Analyses of Follovup Participants vith Lifetime Occurrence of Verified and Suspected Lifetime Neoplasms by Group (Nonblacks and Blacks, Original Comparisons Only) Group Ranch Hand Site Neoplasm Behavior and Status Comparison Number** Percent Number** Percent Total** p-Value*** All Malignant, Benign, Neoplasms Uncertain Behavior, Unspecified Nature Verified 216 Verified and Suspected 231 21.3 22.7 167 196 17.5 20.5 383 427 *Sample sizes: 1,016 Ranch Hands, 955 Original Comparisons. **Number of participants. ***Fisher's exact test. "Participant has one or more malignant, benign, or unspecified skin neoplasms. b Participant has one or more malignant, benign, or unspecified systemic neoplasms. Participant has one or more malignant or benign skin or systemic neoplasms. H-32 0.035 0.251 �TABLE H-16. Unadjusted Analyses of Nonblack Follovup Participants with Lifetime Occurrence of Verified and Suspected Lifetime Malignant Skin Neoplasms by Cell Type and Group (Original Comparisons Only) Group Cell Type Basal Cell Carcinoma Status Verified Verified & Suspected Squamous Cell Carcinoma Verified Verified & Suspected as i Melanoma Verified Verified & Suspected All Malignant Verified Skin Neoplasms Verified & Suspected Sun-Exposure Verified Related Malignant Neoplasms9 Verified & Suspected Statistic Ranch Hand Number /% Total Neoplasms 53 77 5.5% Number /% Total Neoplasms 59 88 6.2% Number/% Total Neoplasms 4 6 0.4% Number /% Total Neoplasms 4 6 0.4% Number /% Total Neoplasms 5 6 0.5% Number/% Total Neoplasms 5 6 0.5% Number/% Total Neoplasms 66 103 6.9% Number /% Total Neoplasms 75 117 7.9% Number /% Total Neoplasms 59 90 6.2%. Number /% Total Neoplasms 65 101 6.8% Original Comparison 34 Est. Relative Risk ( 5 C.I.) p-Value 9% 3.8% 1.49 (0.96,2 .32) 0.079 5.4% 1.16 ( . 9 1 .72) 07, 0.486 0.6% 0.75 ( . 0 2. 0 02, 8) 0.747 0.6% 0.75 8) (0.20,2 . 0 0.747 0.6% 0.94 (0.27,3 .25) 0.999 0.7% 0.78 (0.24,2 .57) 0.768 5.4% 1.31 ( . 9 1 .93) 08, 0.176 6.9% 1.15 (0.81,1 .63) 0.478 4.4% 1.45 (0.96,2 .19) 006 .9 5.9% 1.16 (0.80,1 .69) 0.448 52 48 68 5 6 5 6 5 6 6 7 48 73 62 90 39 58 53 75 *Number and percent of participants; total number of malignant neoplasms of specified cell type. **Number of participants — 956 Ranch Hands and 897 Original Comparisons. a Basal cell carcinoma, melanoma, and malignant epithelial neoplasms NOS. �TABIEH-17. Association Between Lifetime Incidence of Sun-Exposure Belated Skin Malignancies and the Covariates for Gcnbined Followup Ranch Band and Gcnparison Nonblack Participants Verified Covariate Covariate Category Number* Verified and Suspected Percent p-Value Number* Percent p-Value Age Bom >1942 Bom 1923-41 Born <L922 8 83 23 9.2 6.9 2.6 <0.001 28 99 12 3.2 8.3 13.8 <0.001 Occupation Officer fiilisted Flyer Enlisted Groundcrev 60 17 37 7.1 4.7 3.9 009 .0 75 20 44 8.8 5.6 46 . 001 .0 Total Lifetime Smoking (Pack-Years) 0 X)-20 >20-40 >40 35 40 24 15 5.7 40 . 6.1 96 . 0.021 40 49 34 16 6.5 4.9 8.7 10.2 002 .1 Total Lifetime Alcohol Consumption (Drink-Years) 0 X)-5 >5-30 >30-100 >100 8 39 29 24 11 5.7 5.4 4.4 5.0 10.6 0.140 9 45 36 35 11 6.4 6.3 5.5 7.3 10.6 0.341 Ethnic Background3 A B C D 95 16 1 0 60 . 3.8 1.6 00 . 005 .5 118 17 1 0 7.5 40 . 1.6 00 . 007 .0 Skin Color Dark Medium Pale Dark Peach Pale Peach 0 2 10 76 26 00 . 2.7 3.3 60 . 5.0 025 .8 0 2 12 90 35 00 . 2.7 3.9 7.1 6.7 010 .9 �TABLE B-17. (continued) Association Between Lifetime Incidence of SUV-Exposure Related Skin Malignancies and the Covariates for Combined BoUowup Ranch Bend and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Verified and Suspected Percent p-Value Number* Percent p-Value Eye. Color Brown Hazel Green Grey Blue 34 31 4 5 40 5.3 6.8 3.4 5.4 4.7 046 .6 40 32 7 7 53 6.2 7.0 5.9 7.5 6.2 090 .6 Hair Color Black Dark Brown Light Brown Red Blond 22 46 36 2 8 5.0 44 . 6.4 12.5 7.4 0.230 26 58 42 3 10 5.9 5.6 7.5 18.8 9.3 006 .9 Residential History (Average Latitude) >37° <37° 48 66 4.2 6.5 006 .2 56 83 4.9 8.1 003 .0 Skin Reaction to First 30 Min. of Sun Exposure Burns Usually Bums Bums Mildly Rarely Bums 23 39 34 17 9.3 9.1 4.2 2.5 O01 .0 28 47 38 25 11.3 11.0 4.7 3.7 O01 .0 Skin Reaction to >2 Hrs of Sun After First Exposure Bums Painfully Bums Becomes Red No Reaction 13 32 46 23 10.8 9.5 4.4 3.5 O01 .0 15 35 59 30 12.5 10.4 5.7 4.5 O01 .0 Skin Reaction After Repeated Exposure to Sun Freckles, No Tan Tans Mildly Tans Moderately Tans Deep Brown 5 33 40 35 11.1 10.5 3.9 4.5 <D.001 6 38 51 43 13.3 12.1 5.0 5.5 001 .0 �TABLE B-17. (continued) and the Covariates for Grinned FoUowup Ranch Eland and Comparison Nonblack Participants Verified Covariate Covariate Category Sun Reaction Index Tends to Burn Mild Reaction Tends to Tan Exposures to Carcinogens Number* Verified and Suspected Percent p-Value Number* Percent p-Value 9.7 9.3 3.7 O01 .0 16 48 74 11.0 10.6 4.7 <0.001 Asbestos Yes 19 No 95 4.2 5.6 0.288 26 113 5.7 6.6 0.520 Nonmedical X Rays Yes 31 No 83 6.3 5.0 0.252 39 100 7.9 6.0 0.143 Industrial Chemicals Yes 54 No 60 4.8 5.8 0.336 65 74 5.8 7.1 0.220 Herbicides Yes 74 No 40 5.9 4.4 014 .4 91 48 7.2 5.3 006 .7 Insecticides Yes 76 No 38 5.8 4.5 0.200 90 49 6.9 5.7 0.324 Degreasing Chemicals Composite Carcinogen Exposure 14 42 57 Yes 64 No 50 5.1 5.5 067 .9 77 62 6.1 6.9 0.534 Yes 21 No 91 4.3 5.5 0.355 24 113 4.9 6.8 0.141 �TABLE H-17. Continual) ad the Covariates for Conbined Followup Ranch Band and Comparison Nonblack Participants Verified Covariate Covariate Category Number* Verified and Suspected Percent p-Value Number* Percent p-Value Vs Anthracene Yes 0 No 114 00 . 5.3 099 .9 0 139 00 . 6.4 099 .9 Arsenic Exposure to Individual Carcinogens Yes 4 No 109 9.8 5.1 0.163 5 133 12.2 6.3 0.181 Benzene Yes 6 No 108 8.1 5.2 0.281 7 132 9.5 6.3 0.327 Benzidine Yes 1 No 113 9.1 5.3 049 .4 1 138 9.1 6.4 0.519 Chromates Yes 4 No 108 4.8 5.2 099 .9 5 132 6.0 6.4 099 .9 Coal Tar Yes 2 No 112 2.9 5.3 050 .8 3 136 4.4 6.5 079 .9 Creosote Yes 9 No 105 5.7 5.2 0.715 10 129 6.3 6.4 099 .9 Aminodiphenyl Yes 0 No 113 00 . 5.2 099 .9 0 138 00 . 6.4 099 .9 Chloromethyl Ether Yes 2 No 112 9.5 5.2 0.305 2 137 9.5 6.4 0.641 Mustard Gas Yes 0 No 114 00 . 5.3 099 .9 0 139 00 . 6.4 099 .9 �TAHE H-17. (continued) and tiie Govariates for Combined Followup Ranch Band and Comparison NonUack Participants Verified Covariate Category Verified and Suspected Number* 5.8 5.2 0.751 4 134 7.7 6.3 050 .7 Yes 12 No 102 5.3 5.3 099 .9 15 124 6.6 6.4 086 .8 Yes 5 No 109 2.7 5.5 010 .2 7 132 3.8 6.7 0.156 Ultraviolet Light Yes 5 No 109 11.4 5.1 008 .7 5 134 11.4 6.3 020 .0 Vinyl Chloride Exposure to Individual Carcinogens (continued) p-Value Trichloroethylene Covariate Yes 0 No 114 0.0 5.3 046 .0 1 138 3.2 6.5 0.718 Naphtnylamine , Cutting Oils £ Number* Percent Yes 3 No 110 *Number of participants with sun-exposure related skin malignancies. a • Ethnic Background• , Welsh, Scottish, Irish. B—Scandinavian, German, Polish, Russian, other Slavic, Jewish, French. C—Spanish, Italian, Greek. D—Mexican, American Indian, Asian. Percent p-Value �TABLE H-18. Summary of Group-by-Covariate Interactions for Lifetime Malignant Skin Neoplasm Incidence (Nonblacks Only) Group Ranch Hand Number Percent 03 I M3 Interaction Stratification Statistic Burns Easily Variable n Yes No 68 3 65 n Yes No 192 25 167 n Yes No 693 31 662 Basal Cell Group-byIntermediate Carcinoma Sun Reaction Reaction (Verified Index Plus Tans Easily Suspected) Original Comparison Number Percent Adj. Relative Risk ( 5 C.I.) p-Value 9% 11.7 88.3 0.29 (0.07,1.28) 0.102 95.6 77 9 68 13.0 87.0 262 21 241 8.0 1.97 (1.04,3.73) 0.038 92.0 867 36 831 95.8 4.4 4.5 95.5 4.2 1.15 (0.70,1.88) 0.591 �TABLE H-19. Adjusted Analyses of Nonblack Follovup Participants for Lifetime Malignant Skin Neoplasm Incidence (Original Comparisons Only) Adj. Relative Risk (95% C.I.) Variable Status Basal Cell Carcinoma Verified **** **** GRP*SUNREAC (p=0.010) SKIN*SUNREAC (p<0.001) OCC*AGE (p=0.010) OCC*SUNREAC (p=0.007) LAT (p=0.006) Verified & Suspected **** **** GRP*SUNREAC (p=0.003) AGE (p<0.001) LAT (p=0.002) SUNREAC*OCC (p=0.043) SKIN*SUNREAC (p<0.001) Verified **** **** GRP*SUNREAC (p=0.045) SKIN*LAT (p=0.028) SKIN*SUNREAC (p<0.001) OCC*AGE (p=0.023) Verified & Suspected **** **** GRP*SUNREAC (p=0.016) AGE (p<0,001) SUNREAC*LAT (p=0.017) SKIN*SUNREAC (p<0.001) Sun-Exposure Skin Neoplasms p-Value Covariate Remarks ****Group-by-covariate interactions—adjusted relative risk, confidence interval, and p-value not presented. H-40 �TABLE H-20. Summary of Group-by-Covariate Interactions for Lifetime Malignant Skin Neoplasm Incidence (Original Comparisons Only*) Group Ranch Hand Number Percent EC Interaction Stratification Statistic Burns Easily Variable n Yes No 68 3 65 n Yes No 192 23 169 n Yes No 693 27 666 Basal Cell Carcinoma Group-byIntermediate (Verified Sun Reaction Only) Index , Tans Easily Original Comparison Number Percent 95.6 56 6 50 10.7 89.3 12.0 88.0 186 8 178 95.7 652 19 633 97.1 Adj. Relative Risk ( 5 C . . p-Value 9% 1) 4.4 3.9 96.1 4.3 2.9 0.17 ( . 2 1.67) 0.128 00, 3.76 (1.60,8.86) 0.002 1.45 (0.79,2.67) 0.226 �TABLE H-20. (continued) Sunmary of Group-by-Covariate Interactions for Lifetime Malignant Skin Neoplasm Incidence (Original Comparisons Only*) Group Ranch Hand Number Percent Original Comparison Number Percent 83 to Interaction Adj. Relative Risk (95% C.I O p-Value Stratification Statistic Burns Easily Variable n Yes No 68 3 65 4.4 95.6 56 8 48 14.3 85.7 0.19 (0.03,1.15) 0.070 n Yes No 192 25 167 13.0 87.0 186 10 176 5.4 94.6 3.25 (1.47,7.18) 0.003 n Yes No 693 31 662 4.5 95.5 652 29 623 4.4 95.6 1.10 (0.65,1.86) 0.719 Basal Cell Intermediate Carcinoma Group-by(Verified Sun Reaction Plus Index Suspected) Tans Easily �TABLE H-20. (continued) Summary of Group-by-Covariate Interactions for Lifetime Halignant Skin Neoplasm Incidence (Original Comparisons Only*) Group Original Variable Comparison Number Percent Adj. Relative Risk ( 5 C.I . ) p-Value 9% Stratification Statistic Burns Easily at i u> Interaction Ranch Hand Number Percent n Yes No 68 7 61 10.3 89.7 56 8 48 14.3 85.7 1.20 ( . 4 4 22) 0.777 03,. n Yes No 192 26 166 13.5 86.5 186 10 176 5.4 3.98 (1.70,9.33) 0.002 94.6 n Yes No 693 32 661 652 34 618 94.8 Sun Intermediate Exposure Group-byRelated Sun Reaction Cancers Index (Verified Tans Easily Plus Suspected) 4.6 95.4 5.2 1.17 (0.31,4.41) 0.819 �TABLE H-2Q. (continued) Summary of Group-by-Covariate Interactions for Lifetime Malignant Skin Neoplasm Incidence (Original Comparisons Only*) Group Variable Original Comparison Number Percent Adj. Relative Risk (95% C.I.) p-Value w Stratification Statistic Burns Easily Sun Exposure Related Cancers (Verified Only) Interaction Ranch Hand Number Percent n Yes No 68 7 61 10.3 89.7 56 6 50 10.7 89.3 0.80 (0.25,2.53) 0.700 n Yes No 192 24 168 12.5 87.5 186 8 178 4.3 95.7 3.42 (1.56,7.51) 0.002 n Yes No 693 28 665 4.0 96.0 652 24 628 3.7 96.3 0.97 (0.59,1.61) 0.911 Group-byIntermediate Sun Reaction Index *Blacks excluded. Tans Easily �TABLE H-21. Summary of Followup Participants With Lifetime Incidence of Verified Malignant Systemic Neoplasms by Group (Original Comparisons Only) Group Ranch Hand Original Comparison Total Eye I 0 1 Oral Cavity and Pharynx -a , b 0 3 Larynx 0 1 1 Thyroid Gland 0 2 2 Bronchus and Lung 2 0 2 Colon 0 Kidney and Bladder 4 3 7 Prostate 2 2 4 Testicles 3 0 3 Connective and Other Soft Tissue 1 1 2 J.e _l f _2 17 14 31 Site Ill-Defined Sites Total 4c,d 4 a Includes one Ranch Hand with separate malignancies of tongue and epiglottis and also malignant neoplasm of bone. Includes one Ranch Hand with separate malignant neoplasms of tongue and oropharynx and secondary malignant neoplasm of other site. c lncudes one Comparison with secondary malignant neoplasms of liver and bone and bone marrow. d lncludes one Comparison with secondary malignant neoplasm of liver. 'Malignant neoplasm of thorax. Malignant neoplasm of face, head, or neck. H-45 �TABLE H-22. Unadjusted Analyses of the Lifetime Incidence Bates of Ml Malignant Systemic Neoplasms Combined, by Group (Original Comparisons Only) Statistic Group Original Ranch Hand Comparison Est. Relative Risk ( 5 C.I.) 9% p-Value Verified Number/% Total Neoplasms 17 1 7 14 1.5% .% 25 19 1.14 ( . 6 2 3 ) 05,.3 0.722 Verified & Suspected Number/% Total Neoplasms 21 2.1% 36 1 0 (.619) . 4 05,.5 Status IM6 19 2 0 .% 24 099 .9 �TABLE H-23. Summary of Group-by-Occupation Interactions for Lifetime Incidence of All Malignant Systenic Neoplasms Combined Group Variable Interaction Stratification Statistic Ranch Hand Number Percent Comparison Number Percent Officer Group-by- . Occupation a 380 8 372 2.1 97.9 484 12 472 2.5 97.5 Enlisted Flyer n Yes No 177 5 172 2.8 97.2 210 0 210 0 100 Enlisted Groundcrew Malignant Systemic Neoplasm (Verified Only) n Yes No n Yes No 459 4 455 0.9 99.1 599 5 594 0.8 99.2 Adjusted relative risk not calculated because of a zero cell. Fisher's exact test. Adj. Relative Risk (95% C.I.) p-Value 0.84 (0.34,2.08) 0.698 _ w. 0.019b 1.07 (0.28,4.03) 0.922 �TABLE H-23. (continued) Summary of Group-by-Occupatioh Interactions for Lifetime Incidence of All Malignant Systemic Neoplasms Combined Group Ranch Hand Number Percent ta i 00 Interaction Comparison Number Percent Malignant Systemic Group-byNeoplasm Occupation (Verified plus Suspected) a Stratification Statistic Officer Variable n Yes No 380 9 371 2.4 97.6 484 14 470 2.9 97.1 Enlisted Flyer n Yes No 175 7 168 4.0 96.0 209 0 209 0 100 Enlisted Groundcrew n Yes No 457 5 452 1.1 98.9 598 8 590 1.3 98.7 Adjusted relative risk not calculated because of a zero cell. Fisher's exact test. Adj. Relative Risk (95% C.I.) p-Value 0.80 (0.34,1.89) 0.616 0.004b 0.79 (0.26,2.47) 0.690 �TABLE H-24. Adjusted Analyses for Lifetime Incidence of All Malignant Systemic Neoplasms Combined (Original Comparisons Only) Variable Malignant Systemic Neoplasms (Verified) Malignant Systemic Neoplasms (Verified & Suspected) Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks **** **** GRP*OCC (p-0.034) AGE (p<0.001) **** **** GRP*OCC (p=0.003) AGE (p<0.001) RACE*PACKYR (p=0.033) ****Group-by-covariate interaction—adjusted relative risk, confidence interval, and p-value not presented. H-49 �TABLE H-25. Summary of Group-by-Occupation Interactions for Lifetime Incidence of All Malignant Systemic Neoplasms Combined (Original Comparisons Only) Group Variable Interaction Stratification Statistic Ranch Hand Number Percent Original Comparison Number Percent Officer EC Malignant Systemic Group-byNeo plasms Occupation (Verified Only) n Yes No 380 8 372 2.1 97.9 350 9 341 2.6 97.4 Enlisted Flyer n Yes No 177 5 172 2.8 97.2 174 0 174 0 100 Enlisted Groundcrev n Yes No 459 4 455 0.9 99.1 431 5 426 1.2 98.8 Adj. Relative Risk (95% C.I.) p-Value 0.90 (0.34,2.38) 0.834 0.061b 0.81 (0.21,3.07) 0.757 �TABLE H-25. (continued) Summary of Group-by-Occupation Interactions for Lifetime Incidence of All Malignant Systemic Neoplasms Combined (Original Comparisons Only) Group Ranch Hand Number Percent Original Comparison Number Percent CC Un a Statistic n Yes No 380 9 371 2.4 97.6 350 11 339 3.1 96.9 Enlisted Flyer n Yes No 175 7 168 4.0 96.0 174 0 174 0 100 Enlisted Groundcrew Malignant Systemic Group-byNeoplasms Occupation (Verified Plus Suspected) Stratification Officer Variable Interaction n Yes No 457 5 452 1.1 98.9 430 8 422 1.9 98.1 Adjusted relative risk not calculated because of zero cell. b Fisher's exact test. Adj. Relative Risk (95% C.I.) p-Value 0.81 (0.33,1.98) 0.638 0.015a 0.59 (0.19,1-85) 0.369 �TABLE H-26. unadjusted Exposure Index Analyses for FdUxwup Participants in the Bas^iTv>_Rn llnuin Tnfprual Exposure Index3 Low Variable Medium High Contrast Est. Relative Risk (95%C.I.) p-Value 2.5 97.5 Overall M vs. L H vs. L 018 .7 2.73 ( . 1 1 . 5 0.216 07,05) 1.03 ( . 1 5 2 ) 0 9 9 02,.3 .9 52 1 51 1.9 98.1 Overall M vs. L H vs. L 0.073 0.27 ( . 5 1 4 ) 0.145 00,.1 0 1 ( . 2 1 3 ) 0.113 .6 00,.5 1.3 98.7 129 1 128 08 . 99.2 Overall 0.627 01,.2 M vs. L 0.61 ( . 0 3 7 ) 0.674 00,.2 H vs. L 0.35 ( . 4 3 4 ) 0.623 127 12 115 9.4 90.6 121 4 121 3.3 96.7 Overall 002 .4 M vs. L 3.16 ( . 9 1 . 7 0 0 8 09,00) .6 H vs. L 1.03 ( . 5 4 2 ) 0 9 9 02,.3 .9 11.1 88.9 61 2 59 3.3 96.7 52 1 51 1.9 98.1 Overall 003 .7 00,.1 M vs. L 0.27 ( . 5 1 4 ) 0.145 H vs. L 0.16 ( . 2 1 3 ) 0.113 00,.5 2.9 97.1 149 2 147 1.3 98.7 129 1 128 08 . 99.2 Overall M vs. L H vs. L Nunber Percent Number Percent Number Percent Occupation Statistic Officer n Abnormal Normal • 125 3 122 2.4 97.6 127 8 . 119 6.3 93.7 121 3 118 n Abnormal Normal 54 6 48 11.1 8. 89 61 2 59 3.3 96.7 n Abnormal Normal 138 3 135 2.2 97.8 149 2 147 Officer n Abnormal Normal 125 4 121 3.3 96.8 Enlisted Flyer n Abnormal Normal 54 6 48 Enlisted Groundcrew n Abnormal Normal 138 4 134 Pagal Cell Enlisted Flyer Carcinoma (Verified Only)b Enlisted Groundcrew Basal Cell Carcinoma (Verified and Suspected) 0.372 0 4 ( . 8 2 5 ) 0.432 .6 00,.3 ,0.26 ( . 3 2 3 ) 0.372 00,.7 �TABLE B-26. (continued) Unadjusted Exposure Index Analyses for PoLLovup Participants for Occurrence of Malignant Neoplasms in the Baseline-Followp Interval Exposure Indexa Medium Low High Contrast Est. Relative Risk ( 5 C I ) p-Value 9% .. 2.5 97.5 Overall M vs. L H vs. L 006 .9 3.10 ( . 2 1 . 4 0 1 6 08,17) . 3 1.03 ( . 1 5 2 ) 0 9 9 02,.3 .9 52 1 51 1.9 98.1 Overall M vs. L H vs. L 0.073 0.27 ( . 5 1 4 ) 0.145 00,.1 0.16 ( . 2 1 3 ) 0.113 00,.5 u> 2.0 9. 80 129 2 127 1.6 98.4 Overall M vs. L H vs. L 0.929 0.93 ( . 8 4 6 ) 0.999 01,.6 0.71 ( . 2 4 3 ) 0 9 9 01,.1 .9 127. 13 114 10.2 8. 98 121 4 117 3.3 96.7 Overall M vs. L H vs. L 0.021 3.45 ( . 9 1 . 9 0 0 2 10,08) .4 1.03 ( . 5 4 2 ) 0 9 9 02,.3 .9 Sun-Exposure Belated Malignancies (Verified and Suspected) 11.1 88.9 61 2 59 3.3 96.7 52 1 51 1.9 98.1 Overall M vs. L H vs. L 0.073 0.27 ( . 5 1 4 ) 0.145 00,.1 0.16 ( . 2 1 3 ) 0.113 00,.5 2.9 97.1 149 3 146 2.0 9. 80 129 2 127 1.6 98.4 Overall M vs. L M vs. L 0.742 0.69 ( . 5 3 1 ) 0.714 01,.3 0.53 ( . 0 2 9 ) 0.685 01,.3 n Abnormal Normal 125 3 122 2.4 97.6 127 9 118 7.1 92.9 121 3 118 Bilisted Flyer n Abnormal Normal 54 6 43 11.1 88.9 61 2 59 3.3 96.7 138 3 135 2.2 97.8 149 3 146 Officer Ul Statistic n Bilisted Abnormal Groundcrev Normal Sun-Exposure Belated Malignancies (Verified Cnly)b Occupation Officer Variable Number Percent Number Percent Number Percent n Abnormal Normal 125 4 121 3.2 96.8 Enlisted Flyer n Abnormal Normal 54 6 48 n Bilisted Abnormal Groundcrev Normal 138 4 134 �TABLE H-26. (continued) Unadjusted Exposure Index Analyses for Followjp Participants for Occurrence of Malignant Neople in the Baseline-Follouup Interval Exposure Index8 Medium Low Variable Number Percent High Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9% . . Occupation Statistic Officer n Abnormal Normal 127 1 126 0.8 99.2 130 3 127 2.3 97.7 123 0 123 00 . 100 0. Overall 016 .8 M vs. L 2.98 ( . 1 2 . 0 0.622 03,90) H vs. L 0.34 ( . 1 8 4 ) 0 9 9 00,.6* . 9 n Abnormal Normal 55 1 54 1.8 98.2 65 1 64 1.5 98.5 57 0 57 00 . 100 0. Overall 0.612 M vs. L 0 8 ( . 5 1 . 1 0 9 9 .4 00,38) .9 H vs. L 0.32 ( . 1 7 9 ) 0 4 1 00,.2* . 9 Enlisted Groundcrev n Abnormal Normal 154 0 154 00 . 100 0. 163 1162 0.6 99.4 142 1 141 0.7 99.3 Overall 0.598 M vs. L 2.85 ( . 2 7 . 5 * 0 9 9 01,05) .9 H vs. L 3.28 ( . 3 8 . 6 * 0 4 0 01,10) . 8 Officer n Abnormal Normal 127 1 126 08 . 99.2 130 4 126 3.1 96.9 123 0 123 00 . 100 0. Overall H vs. L H vs. L 001 .3 40 (.43.9 030 .0 04,62) .7 0.34 ( . 1 8 4 ) 0 9 9 00,.6* . 9 Enlisted Flyer n Abnormal Normal 55 2 53 3.6 96.4 65 1 64 1.5 98.5 57 1 '6 5 1.8 98.2 Overall M vs. L H vs. L 0 4 (.446) . 1 00,.9 04 (.453) .7 00,.7 Enlisted Groundcrev n Abnormal Normal 154 0 154 00 . 100 0. 163 1 162 06 . 99.4 142 2 140 1.4 98.6 Overall M vs. L H vs. L 0.323 2.85 ( . 2 7 . 5 * 0 9 9 01,05) .9 5.50 ( . 6 1 5 5 ) 0.229 02,1.1* Enlisted Systemic Malignancies Flyer (Verified Only)c Systemic Malignancies (Verified and Suspected)0 a Nunber and percent b Nonblacks only. c of total participants. Blacks and nonblacks. *0.5 added to each cell to calculate relative risk and confidence interval, due to zero cell frequency. 078 .0 0.593 0.615 �TABLE B-27. unadjusted Exposure Index Analysis for FaUowp Participants for Lifef-imp Occurrence of Malignant Neoplasms Low Number Percent Exposure Index3 Medium High Number Percent Number Percent Contrast Est. Relative Risk ( 5 C.I.) p-Value 9% 7.4 92.6 Overall M vs. L H vs. L 061 .4 1.60 ( . 0 4 2 ) 0 4 4 06,.7 .6 1.36 ( . 9 3 7 ) 0.613 04,.6 52 1 51 1.9 98.1 Overall M vs. L H vs. L 000 .3 0.23 ( . 5 1 1 ) 0.081 00,.5 0.13 ( . 2 1 1 ) 0 0 0 00,.1 .6 4.0 9. 60 129 5 124 3.9 96.1 Overall M vs. L H vs. L 094 .8 1.12 ( . 3 3 7 ) 0 9 9 03,.4 .9 1.07 ( . 0 3 8 ) 0 9 9 03,.0 .9 127 14 113 11.0 89.0 121 10 111 8.3 91.7 Overall M vs. L H vs. L 049 .1 1.81 ( . 3 4 4 ) 0.265 07,.9 1.32 ( . 0 3 4 ) 0 6 0 05,.6 .3 13.0 8. 70 61 2 59 3.3 96.7 52 1 51 1.9 98.1 Overall 000 .3 M vs. L 0.23 ( . 5 1 1 ) 0 0 1 00,.5 .8 00,.1 H vs. L 0.13 ( . 2 1 1 ) 0 0 0 .6 4.3 95.7 149 6 143 40 . 96.0 129 5 124 3.9 96.1 Overall M vs. L H vs. L Occupation Statistic Officer n Abnormal Normal 125 7 118 5.6 94.4 127 11 116 8.7 91.3 121 9 112 n Abnormal Normal 54 7 47 13.0 87.0 61 2 59 3.3 96.7 Enlisted Groundcrew n Abnormal Normal 138 5 133 3.6 96.4 149 6 143 Officer Variable n Abnormal Normal 125 8 117 6.4 93.6 Enlisted Flyer n Abnormal Normal 54 7 47 Enlisted Groundcrew n Abnormal Normal 138 6 132 Basal Cell Enlisted Carcinoma Flyer (Verified Only)b Basal Cell Carcinoma (Verified and Suspected) 090 .8 0.92 ( . 9 2 9 ) 0 9 9 02,.3 .9 0.89 ( . 6 2 9 ) 0 9 9 02,.8 .9 �TABLE H-27. (ctntinued) Unadjusted Exposure Index Analysis for Followp Participants for Lifetime Occurrence of Malignant Neoplasms Exposure Index3 Contrast Est. Relative Risk (95%C.I.) p-Value 7.4 92,6 Overall M vs. L H vs. L 0.625 1.47 ( . 1 3 5 ) 0 5 4 06,.7 .0 1.04 ( . 0 2 7 ) 0 9 9 04,.0 .9 52 1 51 1.9 98.1 Overall M vs. L H vs. L 008 .5 0.35 ( . 9 1 4 ) 0.186 00,.2 0.13 ( . 2 1 1 ) 0 0 0 00,.1 .6 4.7 95.3 129 5 124 3.9 96.1 Overall M vs. L H vs. L 0.910 1.31 ( . 1 4 2 ) 0.772 04,.3 1.07 ( . 0 3 8 ) 0 9 9 03,.0 .9 127 16 111 12.6 87.4 121 10 111 8.3 91.7 Overall M vs. L H vs. L 0.382 1.66 ( . 2 3 8 ) 0.301 07,.1 1.04 ( . 2 2 5 ) 0 9 9 04,.9 .9 13.0 87.0 61 3 58 49 . 95.1 52 1 51 1.9 98.1 Overall M vs. L H vs. L 008 .5 0.35 ( . 9 1 4 ) 0.186 00,.2 0.13 ( . 2 1 1 ) 0 0 0 00,.1 .6 4.3 95.7 149 7 142 4.7 95.3 129 5 124 3.9 96.1 Overall M vs. L M vs. L 0.945 1.09 ( . 6 3 3 ) 0 9 9 03,.1 .9 0.89 ( . 6 2.98) 0 9 9 02, .9 Low Number Percent Medium High Number Percent Number Percent Sun-Exposure Related Malignancies (Verified and Suspected) n Abnormal Normal 125 9 116 7.2 92.8 127 13 114 10.2 89.8 121 9 112 Enlisted Flyer n Abnormal Normal 54 7 47 13.0 87.0 61 3 58 4.9 95.1 138 5 133 3.6 96.4 149 7 142 Officer 1° Statistic n Enlisted Abnormal Groundcrew Normal Sun-Exposure Belated Malignancies (Verified Only) Occupation Officer Variable n Abnormal Normal 125 10 115 80 . 92.0 Enlisted Flyer Abnormal Normal 54 7 47 n Enlisted Abnormal Groundcrew Normal 138 6 132 �TABIE H-27. (continued) Unadjusted Exposure Index Analysis for EoUowup Participants for T.iferimp Occurrence of Malignant Naoplc Low Number Percent Exposure Index3 Medium High Number Percent Number Percent Contrast Est. Relative Risk (95ZC.I.) p-Value 2.4 97.6 Overall M vs. L H vs. L 1.48 ( . 4 8 9 ) 02,-9 1.56 ( . 6 9 5 ) 02,.2 086 .7 099 .9 060 .8 57 1 56 1.8 98.2 Overall M vs. L H vs. L 08 (.261) .4 01,.8 0.47 ( . 4 5 3 ) 00,.7 0.825 099 .9 0.615 1.8 98.2 142 1 141 0.7 99.3 Overall M vs. L H vs. L 025 .0 6.74 ( . 5 1 1 5 ) 0 2 8 03,3.3* .4 3.28 ( . 3 8 . 6 * 0 4 0 01,10) . 8 130 4 126 3.1 96.9 123 3 120 2.4 97.6 Overall M vs. L H vs. L 0.730 1.98 ( . 6 1 . 3 0 6 4 03,10) .8 1.56 ( . 6 9 5 ) 0 6 0 02,.2 .8 5.5 94.5 65 2 63 3.1 96.9 57 2 55 3.5 96.5 Overall M vs. L H vs. L 0.55 ( . 9 3 4 ) 00,.2 0.63 ( . 0 3 9 ) 01,.3 0.0 100.0 163 3 160 1.8 98.2 142 2 140 1.4 98.6 Overall M vs. L H vs. L 0.261 6.74 ( . 5 1 1 5 ) 0.248 03,3.3* 5.50 ( . 6 1 5 5 ) 0 2 9 02,1.1* .2 Occupation Statistic Officer n Abnormal Normal 127 2 125 1.6 98.4 130 3 127 2.3 97.7 123 3 120 n Abnormal Normal 55 2 53 3.6 96.4 65 2 63 3.1 96.9 Enlisted Groundcrev n Abnormal Normal 154 0 154 00 . 100.0 163 3 160 Officer n Abnormal Normal 127 2 125 1.6 98.4 Enlisted Flyer n Abnormal Normal 55 3 52 Enlisted Groundcrew Variable n Abnormal Normal 154 0 154 Enlisted Systemic Malignancies Flyer (Verified Only)c Systemic Malignancies (Verified and Suspected)0 a Number and percent b Nonblacks only. c of total participants. Blacks and nonblacks. *0.5 added to each cell to calculate relative risk and confidence interval due to zero cell. 074 .8 060 .6 066 .7 �APPENDIX I Neurological Assessment �APPENDIX I: Neurological Assessment Contents Table 1-1 Unadjusted Exposure Index Analyses for Neurological Variables by Occupation 1-1 1-2 Interaction Summaries of Adjusted Exposure Index Analyses for Selected Neurological Variables 1-16 1-3 Exclusions and Missing Data for Neurological Assessment by Group (Original Comparisons Only) 1-19 1-4 Unadjusted Analyses for Verified Neurological Diseases by Group— 1982-1985 (Original Comparisons Only) 1-20 1-5 Unadjusted Analyses for Verified Neurological Diseases by Group— Baseline and First Followup Studies Combined (Original Comparisons Only) 1-21 1-6 Unadjusted Analyses for Cranial Nerve Function by Group (Original Comparisons Only) 1-22 1-7 Adjusted Analyses for Selected Variables of Cranial Nerve Function by Group (Original Comparisons Only) 1-25 1-8 Unadjusted Analyses for Peripheral Nerve Function by Group (Original Comparisons Only) 1-26 1-9 Adjusted Analyses for Selected Variables of Peripheral Nerve Function by Group (Original Comparisons Only) 1-27 1-10 Summary of Group-by-Diabetic Class Interaction for Pin Prick (Original Comparisons Only) 1-28 1-11 Unadjusted Analyses for CNS Coordination Variables by Group (Original Comparisons Only) 1-29 1-12 Adjusted Analyses for Selected Variables of CNS Coordination by Group (Original Comparisons Only) 1-30 1-13 Longitudinal Analyses of Romberg Sign and Babinski Reflex: A Contrast of Baseline and First Followup Examination Abnormalities (Original Comparisons Only) 1-31 I-i �TftBtEI-1. Variable Exposure Index High Medium Number Percent Number Percent Contrast Est. Relative 'Risk ( 5 C I ) phValue 9% .. 08 . . 9. 92 Overall M vs. L H vs. L 0.361 0 1 ( . 1 4 0 ) 0.243 . 9 00,.5* 00,.2 0.51 ( . 5 5 7 ) 0 9 9 .9 57 2 55 3.5 96.5 Overall M vs. L H vs. L 0.739 08 (.51.1 099 .4 00,38) .9 01,23) .9 1.96 ( . 7 2 . 0 0 9 9 06 . 99.4 142 1 141 0.7 99.3 Overall M vs. L H vs. L 095 .9 09 (.61.3 099 .4 00,52) .9 10 (.71.1 099 .9 00,75) .9 130 0 130 00 . 100 0. 123 1 122 08 . 99.2 Overall 0.593 M vs. L 0.32 ( . 1 8 0 ) 0 4 4 00,.1* . 9 H vs. L 1.03 ( . 6 1 . 0 0 9 9 00,67) .9 1.8 98.2 65 0 65 00 . 100 0. 57 0 57 00 . 100 0. Overall M vs. L H vs. L 0.328 00,.5* 0.28 ( . 1 6 9 ) 0.458 0.32 ( . 1 7 9 ) 0 4 1 00,.2* . 9 1.3 98.7 163 0 163 00 . 100 0. 142 1 141 0.7 99.3 Overall M vs. L H vs. L 0.356 0 1 ( . 1 3 9 ) 0.235 . 9 00,.2* 00,.1 .9 0.54 ( . 5 6 0 ) 0 9 9 Low Number Percent n Abnormal Normal 127 2 125 1.6 98.4 130 0 130 00 . 100 0. 123 1 122 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 1 64 1.5 98.5 Enlisted Groundcrew n Abnormal Normal 154 1 153 0.7 99.3 163 1 162 Officer Visual Fields Statistic Officer Smell Occupation n Abnormal Normal 127 1 126 08 . 99.2 Enlisted Flyer n Abnormal Normal 55 1 54 Enlisted Groundcrew n Abnormal Normal 154 2 152 �TfiBLE 1-1. (continued) Unadjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Occupation Statistic n Officer * Abnormal Normal n Light Reaction Enlisted Flyer ^ Enlisted Groundcrew Abnormal Normal n Abnormal Normal n Officer Abnormal Normal n Ocular. Movements Enlisted Flyer Normal Enlisted Groundcrew n Abnormal Normal Abnormal Low Number Percent Medium High Number Percent Number •Percent Contrast Est. Relative Risk (95% C.I.) p-Value 1.6 98.4 130 0 130 0.0 100.0 123 0 123 0.0 100.0 Overall M vs. L H vs. L 0.133 0.19 (0.01,4.01)* 0.241 0.20 (0.01,4.24)* 0.498 55 1 54 1.8 98.2 65 0 65 0.0 100.0 57 0 57 0.0 100.0 Overall M vs. L H vs. L 0.328 0.28 (0.01,6.95)* 0.458 0.32 (0.01,7.92)* 0.491 154 2 152 1.3 98.7 163 2 161 1.2 98.8 142 1 141 0.7 99.3 Overall M vs. L H vs. L 0.94 (0.13,6.79) 0.54 (0.05,6.01) 127 1 126 0.8 99.2 130 0 130 0.0 100.0 123 0 123 0.0 100.0 Overall M vs. L H vs. L 0.368 0.32 (0.01,8.01)* 0.494 0.34 (0.01,8.46)* 0.999 55 1 54 1.8 98.2 65 1 64 1.5 98.5 57 0 57 0.0 100.0 Overall M vs. L H vs. L 0.612 0.84 (0.05, 13.81) 0.999 0.32 (0.01,7.92)* 0.491 154 1 153 0.7 99.3 163 1 162 0.6 99.4 142 1 141 0.7 99.3 Overall M vs. L H vs. L 0.995 0.94 (0.06,15.23) 0.999 1.09 (0.07,17.51) 0.999 126 2 124 0.866 0.999 0.999 �_», Unadjusted Exposure Index Anal\*ses for 1fenmlncriral \fariahle*: fw (VnnvtHm ^ —x ^ Variable Occupation Statistic Officer Abnormal Normal Enlisted Flyer Abnormal Normal Enlisted Groundcrew Abnormal Normal n n Facial Sensation n n Officer Abnormal Normal n Comeal Reflex Enlisted Flyer Abnormal Normal n Enlisted Groundcrew Abnormal Normal Low Number Percent Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk (95%C.I.) p-Value 126 1 125 0.8 99.2 130 1 129 0.8 99.2 123 0 123 0.0 100.0 Overall M vs. L H vs. L 0.617 0.97 (0.06,15.66) 0.999 0.34 (0.01,8.40)* 0.999 55 I 54 1.8 98.2 65 0 65 0.0 100.0 57 0 57 0.0 100.0 Overall M vs. L H vs. L 0.328 0.28 (0.01,6.95)* 0.458 0.32 (0.01,7.92)* 0.491 154 1 153 0.7 99.3 163 0 163 0.0 100.0 141 0 141 0.0 100.0 Overall M vs. L H vs. L 0.372 0.31 (0.01,7.74)* 0.486 0.36 (0.02,8.95)* 0.999 125 0 125 0.0 100.0 130 0 130 0.0 100.0 123 0 123 0.0 100.0 Overall M vs. L H vs. L 55 0 55 0.0 100.0 65 0 65 0.0 100.0 57 0 57 0.0 100.0 Overall M vs. L H vs. L — — — — 152 0 152 0.0 100.0 161 0 161 0.0 100.0 142 0 142 0.0 100.0 Overall M vs. L H vs. L — — — — — — — — ___ �T6HE T-l. (continued) Unadjusted Exposure Index Analyses for Neurological ^ariahles by Occupation Exposure Index Variable Low Number Percent Medium High Number Percent Number Percent Statistic Officer n Abnormal Normal 127 1 126 08 . 99.2 130 0 130 00 . 100 0. 123 0 123 00 . 100 0. Overall 0.368 M vs. L 0.32 ( . 1 8 0 ) 0 4 4 00,.1* . 9 00,.6* . 9 H vs. L 0.34 ( . 1 8 4 ) 0 9 9 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 0 65 00 . 100 0. 57 0 57 00 . 100 0. Overall 0.328 00,.5* M vs. L 0.28 ( . 1 6 9 ) 0.458 00,.2* H vs. L 0.32 ( . 1 7 9 ) 0.491 Enlisted Groundcrew n Abnormal Normal 154 0 154 00 . 100 0. 163 0 163 00 . 100 0. 142 0 142 00 . 100 0. Overall M vs. L H vs. L Officer n Abnormal Normal 127 1 126 08 . 99.2 130 1 129 08 . 99.2 123 0 123 00 . 100 0. Overall 0.618 M vs. L 0 9 ( . 6 1 . 9 0 9 9 .8 00,57) .9 H vs. L 0.34 ( . 1 8 4 ) 0 9 9 00,.6* . 9 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 1 64 1.5 98.5 57 0 57 00 . 100 0. Overall 0.612 M vs. L 0 8 ( . 5 1 . 1 0 9 9 .4 00,38) .9 00,.2* H vs. L 0.32 ( . 1 7 9 ) 0.491 Enlisted Groundcrew n Abnormal Normal 154 0 154 00 . 100 0. 163 1 162 0.6 9. 94 142 2 140 1.4 9. 86 Overall 0.323 01,05) M vs. L 2.85 ( . 2 7 . 5 * 0.999 H vs. L 5.50 ( . 6 1 5 5 * 0.229 02,1.) * Jaw Clench Snile Contrast Est. Relative 9% . . Risk ( 5 C I ) p-Value Occupation — — — — �t • uhadjusted Exposure Index Anal]rses for 1feiirrtlrvHipal \fariaKl«ag by Occupation Variable Low Number Percent n Abnormal Normal 127 1 126 0.8 99.2 Enlisted " Flyer n Abnormal Normal 55 0 55 0.0 100.0 Enlisted Groundcrev n Abnormal Normal 154 0 154 Officer Balance Statistic Officer Palpebral Fissures Occupation n Abnormal Normal Enlisted Flyer Enlisted Groundcrev Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk (952C.I.) p-Value 1.5 98.5 122 1 121 0.8 99.2 Overall M vs. L H vs. L 0.801 1.97 (0.18, 21.99) 0.999 1.04 (0.06, 16.84) 0.999 65 1 64 1.5 98.5 57 0 57 0.0 100.0 Overall M vs. L H vs. L 0.0 100.0 163 1 162 0.6 99.4 142 1 141 0.7 99.3 Overall M vs. L H vs. L 0.420 2.58 (0.10,64.65)* 0.999 — — 0.598 2.85 (0.12, 70.55)* 0.999 3.28 (0.13,81.06)* 0.480 127 0 127 0.0 100.0 130 0 130 0.0 100.0 123 0 123 0.0 100.0 Overall M vs. L H vs. L n Abnormal Normal 55 0 55 0.0 100.0 64 0 64 0.0 100.0 57 0 57 0.0 100.0 Overall M vs. L H vs. L n Abnormal Normal 154 1 153 0.7 99.3 163 0 163 0.0 100.0 142 1 141 0.7 99.3 Overall M vs. L H vs. L 130 2 128 — — — — — — — — 0.574 0.31 (0.01,7.74)* 0.486 1.09 (0.07,,17.51)* 0.999 �unadjusted Exposure Index Analyses for 1 Variable Low Number Percent cal %ruibles by Occupation Exposure Index High Medium Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9% .. n Abnormal Normal 126 0 126 00 . 100.0 130 0 130 00 . 100 0. 123 0 123 00 . 100 0. Overall M vs. L H vs. L Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 0 65 00 . 100 0. 57 0 57 00 . 100 0. Overall 0.328 00,.5* . 5 M vs. L 0.28 ( . 1 6 9 ) 0 4 8 H vs. L 0.32 ( . 1 7 9 ) 0 4 1 00,.2* . 9 Enlisted Groundcrew n Abnormal Normal 153 0 153 00 . 100 0. 163 0 163 00 . 100 0. 142 0 142 00 . 100 0. Overall M vs. L H vs. L Officer Speech Statistic Officer Gag Reflex Occupation n Abnormal Normal 127 0 127 00 . 100 0. 130 0 130 00 . 100 0. 123 0 123 00 . 100 0. Overall M vs. L Hvs. L Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 0 65 00 . 100 0. 57 0 57 00 . 100 0. Overall M vs. L H vs. L Enlisted Groundcrew n Abnormal Normal 154 1 153 0.7 99.3 163 0 163 00 . 100 0. 142 1 141 07 . 9. 93 038 .2 0.28 ( . 1 6 9 ) 0 4 8 00,.5* . 5 0.32 ( . 1 7 9 ) 0 4 1 00,.2* . 9 Overall 0.574 M vs. L 0.31 ( . 1 7 7 ) 0 4 6 00,.4* . 8 00,75) .9 H vs. L 1.09 ( . 7 1 . 1 0 9 9 �T^HIE 1-1. (continued) Unadjusted Exposure Index Analyses for Neurological Variables by Occupation Variable Low Nunber Percent Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9% .. 00 . 100 0. Overall M vs. L H vs. L 0.368 0.32 ( . 1 8 0 ) 0 4 4 00,.1* . 9 0.34 ( . 1 8 4 ) 0 9 9 00,.6* . 9 57 0 57 00 . 100 0. 0.328 Overall M vs. L 0.28 ( . 1 6 9 ) 0.458 00,.5* H vs. L 0.32 ( . 1 7 9 ) 0 4 1 00,.2* . 9 06 . 9. 94 142 0 142 00 . 100 0. Overall 044 .0 M vs. L 2.83 ( . 2 7 . 9 * 0 9 9 01,00) . 9 H vs. L — 130 0 130 00 . 100 0. 123 0 123 00 . 100 0. Overall 0.365 Mvs. L 0.32 ( . 1 7 9 ) 0 4 2 00,.4* . 9 H vs. L 0.34 ( . 1 8 4 ) 0 9 9 00,.0* . 9 1.8 98.2 65 0 65 00 . 100 0. 57 0 57 00 . 100 0. Overall 0.328 M vs. L 0.28 ( . 1 6 9 ) 0 4 8 00,.5* . 5 H vs. L 0.32 ( . 1 7 9 ) 0 4 1 00,.2* . 9 00 . 100 0. 163 0 163 00 . 100 0. 142 0 142 00 . 100 0. Overall M vs. L H vs. L 127 1 126 08 . 99.2 130 0 130 00 . 100 0. 123 0 123 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 0 65 00 . 100 0. n Abnormal Normal 153 0 153 00 . 100 0. 163 1 162 Officer Palate and Uvula Movement n Abnormal Normal Enlisted Groundcrev M Statistic Officer Tongue Position Relative to Midline Occupation n Abnormal Normal 126 1 125 08 . 99.2 Misted Flyer n Abnormal Normal 55 1 54 Enlisted Groundcrev n Abnormal Normal 153 0 153 — — — — — �Unadjusted Exposure Index Analyses for Neurological \foriables by Occupation Variable Low Number Percent Exposure Index Medium High Number Percent Number Percent Statistic Officer n Abnormal Normal 127 9 118 7.1 92.9 130 9 121 6.9 93.1 123 10 IB 8.1 91.9 Overall 094 .2 M vs. L 0 9 ( . 7 2 5 ) 0 9 9 .9 .8 03,.4 04,.) . 1 H vs. L 1.16 ( . 6 2 % 0 8 4 Enlisted Flyer n Abnormal Normal 55 5 50 9.1 9. 09 65 6 59 9.2 9. 08 57 4 53 7.0 93.0 081 .9 Overall M vs. L 1.02 ( . 9 3 5 ) 0 9 9 .9 02,.3 H vs. L 0.76 ( . 9 2 9 ) 0 7 0 01,.7 .4 Enlisted Groundcrev n Abnormal Normal 154 9 145 5.8 94.2 163 6 157 3.7 96.3 142 3 139 2.1 97.9 020 .5 Overall 02,.7 M vs. L 0.62 ( . 1 1 7 ) 0.433 H vs. L 0.35 ( . 9 1 3 ) 0.142 00,.1 Officer Neck Range of Motion Contrast Est. Relative Risk ( 5 C I ) p-Value 92.. Occupation n Abnormal Normal 122 15 107 12.3 87.7 130 12 118 9.2 9. 08 123 13 109 10.7 89.3 074 .3 Overall M vs. L 0.73 ( . 3 1 6 ) 0.542 03,.2 .4 H vs. L 0 8 ( . 9 1 8 ) 0 8 1 .5 03,.7 Enlisted Flyer n Abnormal Normal 55 7 48 12.7 87.3 64 8 56 12.5 87.5 57 5 52 88 . 91.2 n Enlisted Abnormal Grounder®/ Normal 151 15 136 9.9 90.1 161 12 149 7.5 92.5 141 9 132 6.4 93.6 £ Cranial Nerve Function Index Overall M vs. L 0 9 .8 H vs. L 0 6 .6 , Overall M vs. L 0.73 H vs. L 0.62 0.754 (.329) 099 03,.0 .9 ( . 0 2 2 ) 0.554 02,.2 0.512 (.316) 056 03,.2 .4 ( . 6 1 4 ) 0.290 02,.6 �— —— — x ' • "r Unadjusted Exposure Index Analyses for Neurological \fariables by Occupation Variable Low Number Percent Exposure Index Medium Higfc Number Percent Number Percent n Abnormal Normal 122 8 114 6.6 93.4 130 3 127 2.3 97.7 122 3 119 2.5 97.5 Overall 0.136 M vs. L 0.34 ( . 9 1 3 ) 0.127 00,.0 H vs. L 0.36 ( . 9 1 3 ) 0.216 00,-9 Enlisted Flyer n Abnormal Normal 55 3 52 5.5 94.5 64 2 62 3.1 96.9 57 2 55 3.5 96.5 Overall M vs. L H vs. L Enlisted Groundcrew n Abnormal Normal 151 8 143 5.3 94.7 161 6 155 3.7 96.3 141. 7 134 5.0 95.0 Overall 075 .8 M vs. L 0 6 ( . 3 2 0 ) 0.590 .9 02,.4 H vs. L 0.93 ( . 3 2 6 ) 0 9 9 03,.5 .9 Officer Pin Prick Statistic Officer Cranial Nerve Function Index (Neck Range of Motion Excluded) Contrast Est. Relative Risk ( 5 C I ) p-Value 9% .. Occupation n Abnormal Normal 126 9 117 7.1 9. 29 127 6 121 4.7 95.3 122 7 115 5.7 9. 43 Overall 0.713 02,.7 M vs. L 0.65 ( . 2 1 8 ) 0 4 0 .4 H vs. L 0 7 ( . 9 2 2 ) 0 7 7 . 9 02,.0 .9 Enlisted Flyer n Abnormal Normal 55 5 50 9.1 9. 09 64 3 61 4.7 95.3 57 5 52 88 . 91.2 Overall 0.584 M vs. L 0 4 ( . 1 2 1 ) 0 4 9 .9 01,.6 .6 . 02,.2 H vs. L 0 % ( . 6 3 5 ) 0 9 9 .9 Enlisted Groundcrew n Abnormal Normal 152 13 139 8.6 91.4 162 8 154 49 . 95.1 138 3 135 2.2 97.8 0.052 Overall M vs. L 0 5 ( . 2 1 3 ) 0.259 .6 02,.8 H vs. L 0.24 ( . 7 0 8 ) 0 0 0 00,.5 .2 0.791 0 5 ( . 9 3 4 ) 0.661 .6 00,.7 0.63 ( . 0 3 9 ) 0 6 6 01,.3 .7 �TABLE 1-1. (cmtinued) Unadjusted Exposure Index Analyses for Neurological variables by Occupation Exposure Index Variable Low Number Percent Medium High Number Percent Number Percent Officer 126 9 117 7.1 92.9 127 4 123 3.2 96.8 3 119 2.5 97.5 0.141 Overall M vs. L 0.42 ( . 3 1 4 ) 0.167 01,.1 H vs. L 0.33 ( . 9 1 2 ) 0.137 00,-4 Bilisted ' Flyer n Abnormal Normal 55 3 52 5.5 9. 45 64 2 62 3.1 9. 69 57 4 53 7.0 9. 30 Overall 069 .1 M vs. L 0 5 ( . 9 3 4 ) 0 6 1 . 6 00,.7 .6 H vs. L 1.31 ( . 8 6 1 ) 0 9 9 02,.3 .9 n Abnormal Normal 152 5 147 3.3 96.7 162 5 157 3.1 96.9 138 3 135 2.2 97.8 Overall 084 .3 M vs. L 0 9 ( . 7 3 3 ) 0 9 9 .4 02,.0 .9 H vs. L 0 6 ( . 5 2 7 ) 0.725 .5 01,.9 Officer Muscle Status n Abnormal Normal Enlisted Groundcrev Light Touch Statistic n Abnormal Normal 127 5 122 3.9 96.1 130 2 128 1.5 98.5 123 5 118 4.1 95.9 Overall 048 .2 M vs. L 0 3 ( . 7 2 0 ) 0 2 8 . 8 00,.0 .7 H vs. L 1.03 ( . 9 3 6 ) 0 9 9 02,.6 .9 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 1 64 1.5 98.5 57 1 56 1.8 98.2 Overall M vs. L H vs. L 092 .9 0 8 (.5 1 . 1 0 9 9 . 4 00, 3 8 ) . 9 09 (.61.1 099 .6 00,58) .9 Enlisted Groundcrew n Abnormal Normal 154 3 151 2.0 9. 80 163 5 158 3.1 96.9 142 3 139 2.1 97.9 Overall M vs. L H vs. L 071 .8 1.59 ( . 7 6 7 ) 0.724 03,.8 1.09 ( . 2 5 4 ) 0 9 9 02,.7 .9 199 Contrast Est. Eelative Risk ( 5 C I ) p-Value 9% . . Occupation �Unadjusted Exposure Index Analyses for Neurological Variables by Occupation Variable Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9%.. 2.5 97.5 Overall M vs. L H vs. L 052 .4 0.33 ( . 3 3 1 ) 0.370 00,-7 1.03 ( . 1 5 2 ) 0 9 9 02,.2 .9 57 1 56 1.8 98.2 Overall 068 .2 Mvs. L 2.62 ( . 1 6 . 8 * 0 9 9 01,56) . 9 H vs. L 2.95 ( . 2 7 . 0 * 0 9 9 01,39) .9 1.2 98.8 138 0 138 00 . 100 0. 130 1 129 08 . 99.2 123 3 120 2.4 9. 76 Overall 0.156 M vs. L 2.95 ( . 2 7 . 8 * 0 9 9 01,31) .9 H vs. L 7.41 ( . 8 1 4 9 ) 0 1 8 03,4.0* . 1 1.8 98.2 65 1 64 1.5 98.5 57 1 56 1.8 98.2 092 .9 Overall M vs. L 0 8 ( . 5 1 . 1 0 9 9 .4 00,38) .9 H vs. L 0 % ( . 6 1 . 1 0 9 9 . 00,58) .9 2.0 9. 80 163 0 163 00 . 100.0 142 1 141 0.7 99.3 Overall M vs. L H vs. L Low Number Percent n Abnormal Normal 126 3 123 2.4 97.6 127 1 126 08 . 99.2 122 3 119 Enlisted Flyer n Abnormal Normal 55 0 55 00 . 100 0. 64 1 63 1.6 9. 84 Enlisted Groundcrew n Abnormal Normal 152 0 152 00 . 100 0. 162 2 160 Officer Patellar Reflex Statistic Officer Vibratory Sensation Occupation n Abnormal Normal 127 0 127 00 . 100 0. Enlisted Flyer n Abnormal Normal 55 1 54 Enlisted Groundcrew n Abnormal Normal 154 3 151 Overall Mvs. L H vs. L 016 .6 4.75 ( . 3 9 . 6 * 0 4 9 02,97) .9 0.170 0.13 ( . 1 2 5 ) 0 1 4 00,.8* . 1 0.36 ( . 4 3 4 ) 0 6 4 00,.7 .2 �TABIE 1-1. (continued) Ihadjusted Exposure Index Analyses for Neurological Variables by Occupation Variable Low Nunber Percent Exposure Index Medium High Number Percent Number Percent Est. Relative Risk ( 5 C I ) p-Value 9% . . n Abnormal Normal 124 8 116 6.5 93.5 129 6 123 4.7 95.3 123 8 115 6.5 93.5 074 .7 Overall M vs. L 0.71 ( . 4 2 1 ) 0 5 1 02,.0 .9 03,.8 H vs. L 1.01 ( . 7 2 7 ) 0 9 9 .9 Enlisted Flyer n Abnormal Normal 55 5 50 9.1 90.9 64 0 64 00 . 100 0. 57 4 53 70 . 93.0 Overall M vs. L H vs. L Enlisted Groundcrew n Abnormal Normal 153 10 143 6.5 93.5 163 9 154 5.5 9. 45 141 8 133 5.7 94.3 Overall 090 .2 M vs. L 0 8 ( . 3 2 1 ) 0 8 4 .4 03,.2 .1 H vs. L 0 8 ( . 3 2 2 ) 0 8 2 .6 03,.5 .1 Officer Biceps Reflex Statistic Officer • Achilles Reflex Occupation n Abnormal Normal 127 1 126 0.8 99.2 130 0 130 00 . 100 0. 123 2 121 1.6 98.4 Overall M vs. L H vs. L 034 .4 0.32 ( . 1 8 0 ) 0 4 4 00,.1* . 9 2.08 ( . 9 2 . 7 0 6 8 01,32) .1 Enlisted Flyer n Abnormal Normal 55 0 55 00 . 100.0 65 1 64 1.5 98.5 57 0 57 00 . 100 0. Overall M vs. L H vs. L 040 .2 2.58 ( . 0 6 . 5 0 9 9 01,46) .9 — — Enlisted Groundcrew n Abnormal Normal 154 1 153 0.7 99.3 163 3 160 1.8 98.2 142 1 141 0.7 99.3 Contrast 009 .5 0 0 (.0,.2* 0 0 9 . 7 00413) . 1 0.76 ( . 9 2 9 ) 0 7 0 01,.7 .4 Overall 0.515 03,78) M vs. L 2.87 ( . 0 2 . 8 0.623 00,75) .9 H vs. L 1.09 ( . 7 1 . 1 0 9 9 �TAKE 1-1. (continued) Ibadjusted Exposure Index Analyses for Neurological Variables by Occupation Exposure Index Variable Low Number Percent n Abnormal Normal 127 0 127 Enlisted Flyer n Abnormal Normal Enlisted Groundcrew Officer Tremor Statistic Officer Babinski Reflex Occupation High. Medium Nunber Percent Number Percent 00 . 100 0. 128 0 128 00 . 100 0. 122 0 122 55 0 55 00 . 100.0. 65 0 65 n Abnormal Normal 153 2 151 1.3 98.7 n Abnormal Normal 127 5 Enlisted Flyer Enlisted Groundcrew Contrast Est. Relative Risk ( 5 C I ) p-Value 9% . . 00 . 100 0. Overall M vs. L H vs. L 00 . 100 0. 57 0 57 00 . 100 0. Overall M vs. L H vs. L 162 0 162 00 . 100 0. 142 2 140 1.4 9. 86 Overall M vs. L H vs. L 199 3.9 96.1 130 1 129 08 . 99.2 123 3 120 2.4 97.6 Overall 0.248 M vs. L 0.19 ( . 2 1 6 ) 0.117 00,.4 H vs. L 0.61 ( . 4 2 6 ) 0.722 01,.1 n Abnormal Normal 55 1 54 1.8 98.2 65 3 62 46 . 95.4 57 3 54 5.3 94.7 Overall M vs. L H vs. L n Abnormal Normal 154 5 149 3.3 96.7 163 4 159 2.5 97.5 142 1 141 07 . 99.3 0.312 Overall M vs. L 0 7 ( . 0 2 8 ) 0 7 4 . 5 02,.5 .4 H vs. L 0.21 ( . 2 1 8 ) 0.216 00,.3 — — — — __ — — — — 0.329 0.19 ( . 1 3 9 ) 0.235 00,.2* 10 (.577) 099 .8 01,.6 .9 069 .0 2.61 ( . 6 2 . 6 0 6 4 02,58) . 2 3.00 ( . 0 2 . 6 0.618 03,97) �TABLE 1-1. (continued) \y Occupation Unadjusted Exposure Index Analyses for Neurological Variable Occupation Statistic Low Number Percent Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk (95XC.I.) p-Value . Officer n Abnormal Normal 127 0 127 00 . 100 0. 130 1 129 08 . 99.2 123 1 122 08 . 99.2 Overall 063 .0 01, 31) . 9 M vs. L 2.95 ( . 2 7 . 8 * 0 9 9 01, 73) .9 H vs. L 3.12 ( . 3 7 . 9 * 0 4 2 Coordination Enlisted • Flyer n Abnormal Normal 55 0 55 00 . 100.0 64 1 63 1.6 98.4 57 0 57 00 . 100.0 0.415 Overall M vs. L 2.62 ( . 1 6 . 8 * 0 9 9 01, 56) . 9 H vs. L — — Enlisted Groundcrew n Abnormal Normal 154 2 152 1.3 98.7 163 1 162 06 . 99.4 142 3 139 2.1 97.9 Overall M vs. L H vs. L Officer n Abnormal Normal 127 0 127 00 . 100.0 130 0 130 00 . 100.0 123 0 123 00 . 100.0 Overall M vs. L H vs. L — — — — Enlisted Flyer n Abnormal Normal 55 0 55 00 . 100.0 64 0 64 00 . 100.0 57 0 57 00 . 100.0 Overall M vs. L H vs. L — — — — Enlisted Groundcrew n Abnormal Normal 154 1 153 0.7 99.3 163 0 163 00 . 100.0 142 1 141 07 . 99.3 Overall M vs. L H vs. L Romberg Sign 0.516 0.47 ( . 4 5.23) 0.613 00, 1.64 ( . 7 9 % 0 6 4 02, . ) . 7 054 .7 0.31 ( . 1 7 7 ) 0 4 6 00,.4 .8 1.09 ( . 7 1 . 1 0 9 9 00,75) .9 �— — —- X ' " f Unadjusted Exposure Index Anal}ses for Neurological \feriahles by Occupation Variable Exposure Index Medium High Number Percent Mmber Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9%.. 1.6 98.4 Overall M vs. L H vs. L 051 .9 0.32 ( . 3 3 1 ) 0.366 00,.2 0.68 ( . 1 4 1 ) 0 9 9 01,.6 .9 57 0 57 00 . 100 0. Overall M vs. L H vs. L 0.370 08 (.261) 099 .4 01,.8 .9 00,.7* . 3 0.19 ( . 1 3 9 ) 0 2 9 1.2 9. 88 142 4 138 2.8 97.2 Overall M vs. L H vs. L 050 .8 04 (.625) 047 .7 00,.8 .3 1.09 ( . 7 4 4 ) 0 9 9 02,.3 .9 130 2 128 1.5 9. 85 123 4 119 3.3 9. 67 Overall M vs. L H vs. L 0.123 02 (.511) 008 .3 00,.2 .5 0 5 ( . 5 1 7 ) 0.377 . 0 01,.1 5.5 94.5 64 5 59 7.8 92.2 57 3 54 5.3 9. 47 Overall M vs. L H vs. L 080 .1 03,.5 1.47 ( . 4 6 4 ) 0.724 .9 0 % (.949) 099 . 01,.9 6.5 93.5 163 7 156 43 . 95.7 142 6 136 42 . 95.8 0.585 Overall . 5 02,.4 .5 M vs. L 0 6 ( . 4 1 7 ) 0 4 9 H vs. L 0 6 ( . 3 1 8 ) 0 4 8 . 4 02,.0 .4 Lew Number Percent n Abnormal Normal 127 3 124 2.4 97.6 130 1 129 08 . 99.2 123 2 121 Enlisted Flyer n Abnormal Normal 55 2 53 3.6 96.4 65 2 63 3.1 9. 69 Enlisted Groundcrew n Abnormal Normal 154 4 150 2.6 97.4 163 2 161 Officer n Abnormal Normal 127 8 119 6.3 93.7 Enlisted Flyer n Abnormal Normal 55 3 52 Enlisted Groundcrew CMS Sunnary Index Statistic Officer Gait Occupation n Abnormal Normal 154 10 144 *Estimated relative risk and confidence interval calculated after adding 0.5 to each cell. —No abnormals present in contrast; estimated relative risk, confidence interval, and p-value not calculated. �TABLE 1-2. Interaction Sumaries of Adjusted Exposure variable Interaction (Occupation) Stratification Diabetic . Cranial Nerve Function Index Light Touch . Exposure Index-byDiabetic Class (Enlisted Groundcrew) Exposure Index-byInsecticide Exposure (Enlisted Flyer) Exposure Index Low Adj. Relative Medium High 9Z . . Statistic Number Percent Number Percent Number Percent Contrast Risk ( 5 C I ) n Abnormal Normal n Impaired Abnormal Normal p-Value 7 0 7 00 . 100 0. 12 3 9 25.0 75.0 14 2 12 14.3 85.7 5.53 ( . 5 1 4 4 ) * 02,2.0* 30 (.37.1* .0 01,13)* 0.339* 0.263* 0.533* 11 0 11 00 . 100.0 14 2 12 14.3 85.7 18 0 18 Overall . 0 02,0.0* 0 0 M vs. L 4 6 ( . 0 1 6 3 ) * . — 1 0 0 H vs. L 0. 0.114* 047 .8* — Overall M vs. L H vs. L Normal* n Abnormal Normal 133 15 118 11.3 88.7 135 7 128 5.2 9. 48 109 7 102 64 . 93.6 Overall M vs. L 0 5 ( . 1 1 4 ) .5 02,.4 H vs. L 0 5 ( . 2 1 4 ) .6 02,.4 036 .3 0.225 028 .2 Yesb n Abnormal Normal 38 1 37 2.6 97.4 45 2 43 44 . 9. 56 40 4 36 10.0 9. 00 Overall M vs. L 0 4 ( . 2 8 8 ) . 2 00,.1 H vs. L 5 4 ( . 1 7 . 4 . 8 04,15) 006 .8 053 .7 014 .9 No n Abnormal Normal 17 2 15 11.8 88.2 19 0 19 00 . 100 0. 17 0 17 Overall 0 0 M vs. L 0.16 ( . 1 3 5 ) * . 00,.6* 1 0 0 H vs. L 0.18 ( . 1 3 9 ) * 0. 00,.8* 0.111* 026 .1* 045 .8* �TABLE 1-2. (continued) DatezacticD ^••ai-iog of AHjigt**j Exposure Variable Interaction (Occupation) Stratification Exposure Index High Low Medium Adj. Relative Statistic Number Percent Number Percent Number Percent Contrast Risk ( 5 C I ) 9% .. p-Value Born >1942 AchiUes Reflex Exposure Index-byAge (Enlisted Groundcrew) 1.1 98.9 129 2 127 1.6 93.4 82 0 82 00 . 100 0. Overall M vs. L 1.35 ( . 2 1 . 7 * 01,51) H vs. L 0.35 ( . 1 8 7 ) * 00,.0* 059 .3* 099 .9* 099 .9* Bom 1923-1941 n Abnormal Normal 63 1 62 1.6 9. 84 33 2 31 6.1 93.9 54 3 51 5.6 94.4 Overall M vs. L 4 0 ( . 5 4 . 4 * . 0 03,58) H vs. L 3.65 ( . 7 3 . 3 * 03,61) 046 .3* 021 .7* 034 .3* n Abnormal Normal 4 1 3 2. 50 75.0 1 1 0 100 0. 00 . 6 0 6 Bom £L942b Exposure Ihdex-byAge (Enlisted Groundcrew) 87 1 86 Born <L922 Muscle Status n Abnormal Normal n Abnormal Normal 86 2 84 2.3 97.7 129 6 123 4.7 95.3 81 1 80 1.2 98.8 Overall M vs. L 2.18 ( . 3 1 . 2 04,11) H vs. L 0.57 ( . 5 6 3 ) 00,.7 039 .0 036 .4 064 .4 n Bom 1923-1941" Abnormal Normal 63 5 58 7.9 92.1 33 2 31 6.1 93.9 54 6 48 11.1 88.9 Overall M vs. L 0 6 ( . 1 3 7 ) .4 01,.5 H vs. L 1.40 ( . 8 5 1 ) 03,.9 069 .4 064 .2 066 .1 n Abnormal Normal 4 3 1 75.0 2. 50 1 1 0 100.0 0;0 6 1 5 16.7 83.3 Overall M vs. L 1.29 ( . 3 5 . 1 * 00,35)* H vs. L 0 0 ( . 0 , . 1 * . 7 00315) 010 .0* 099 .9* 010 .9* Born <L922 Overall 0 0 M vs. L 7.00 ( . 7 2 1 3 ) * . 01,9.4* 1 0 0 H vs. L 0.18 ( . 1 5 6 ) * 0. 00,.8* 001 .5* 040 .0* 040 .0* �TABLE 1-2. (continued) Index AralV9PS I™* Selected tinimlniriral UaHahilM Variable Central Nervous System Index Interaction (Occupation) Exposure Index-byInsecticide Exposure (Enlisted Groundcrew) Stratification Exposure Index High Low Medium Adj. Relative Statistic Nuriber Percent Number Percent Nunber Percent Contrast Risk (95Z C.I.) p-Value Yes" n Abnormal Normal 102 4 98 3.9 96.1 113 7 106 6.2 93.8 95 3 92 3.2 96.8 Overall M vs. L H vs. L 1.68 (0.48,5.95) 0.87 (0.19,3.99) 0.560 0.419 0.853 No n Abnormal Normal 52 6 46 11.5 88.5 50 0 50 0.0 100.0 47 3 44 6.4 93.6 Overall M vs. L H vs. L 0.07 (0.004,1.29)** 0.38 (0.10,1.54)* 0.050* 0.027* 0.209* *Estimated relative risk or p-value, based on stratified tables. **Estimated relative risk and confidence interval calculated after adding 0.5 to each cell. —No abnormal participants present in contrast; adjusted relative risk, p-value, and confidence interval not presented. 'Results adjusted for age. b Results adjusted for diabetic class. Note: Small sample sizes may affect validity of overall p-value. �TABLE 1-3. Exclusions and Missing Data for Neurological Assessment by Group (Original Comparisons Only) Group Data Category Ranch Hand Original Comparison Total Lifetime Alcohol History (Drink-years); Missing Data 39 27 66 Peripheral Edema (Exclusion Category for Pin Prick, Light Touch, and Ankle Vibration) 13 12 25 Diabetic Class (Missing Data) 0 3 3 Positive Syphilis Serology (RPR and FTA); Exclusion Category 0 0 0 1-19 �TABLE 1-4. Unadjusted Analyses for Verified Neurological Diseases by Group*—1982-1985 (Original Comparisons Only) Group Abnormalities Original Comparison Ranch Hand Disease Category Number Percent Number Percent Total p-Value** Inflammatory Diseases 0 0.0 0 ' 0.0 0 Hereditary and Degenerative Diseases 2 0.2 0 0.0 2 0.500 Peripheral Disorders 18 1.8 20 2.1 38 0.626 Disorders of the Eye 5 0.5 5 0.5 10 0.999 Disorders of the Ear 6 0.6 6 0.6 12 0.999 Other Disorders 8 0.8 3 0.3 11 0.288 *Based on 1,016 Ranch Hands and 955 Original Comparisons; some participants may be classified in more than one category. **Fisher's exact test. 1-20 �TABLE 1-5. Unadjusted Analyses for Verified Neurological Diseases by Group*—Baseline and First Followup Studies Combined (Original Comparisons Only) Group Abnormalities Original Comparison Ranch Hand Disease Category Number Percent Number Percent Total p- Value**'* Inflammatory Diseases 3 0.3 2 0.2 5 0.999 Hereditary and Degenerative Diseases 2 0.2 3 0.3 5 0.678 Peripheral Disorders 23 2.3 27 2.8 50 0.475 Disorders of the Eye 16 1.6 16 1.7 32 0.861 Disorders of the Ear 24 2.4 26 2.7 50 0.668 Other Disorders 15 1.5 12 1.3 27 0.703 *Based on 1,016 Ranch Hands and 955 Original Comparisons; some participants may be classified in more than one category. **Fisher's exact test. *Group significance for patterns of disease: 1-21 p=0.702. �TABLE 1-6. Unadjusted Analyses for Cranial Nerve nncn.cn iay uroup vunginajL \ ^jflnjujri^ riR uniy /j ••—** **•*«/ Gn3UP Original Comparison Number Percent Est. Relative Risk (95% C.I.) Variable Cranial Nerve Statistic Ranch Hand Number Percent Smell I Olfactory n Abnormal Normal 1,016 10 1,006 1.0 99.0 955 9 946 0.9 99.1 1.05 (0.42,2.58) 0.999 n Abnormal Normal 1,016 6 1,010 0.6 99.4 955 4 951 0.4 99.6 1.41 (0.40,5.02) 0.754 n Abnormal Normal 1,015 8 1,007 0.8 99.2 954 7 947 0.7 99.3 1.08 (0.39,2.98) 0.999 Oculomotor IV Trochlear VI Abducens n Abnormal Normal 1,016 6 1,010 0.6 99.4 955 5 950 0.5 99.5 1.13 (0.34,3.71) 0.999 Facial Sensation V Trigeminal n Abnormal Normal 1,014 4 1,010 0.4 99.6 953 2 951 0,2 99,8 1.88 (0.34,10.31) 0.688 Jaw Clench V Trigeminal n Abnormal Normal 1,016 2 1,014 0.2 99.8 955 1 954 0.1 99.9 1.88 (0.17,20.79) 0.999 n Abnormal Normal 1,016 7 1,009 0.7 99.3 955 2 953 0,2 99.8 3.31 (0.69,15.95) 0.181 n Abnormal Normal 1,015 7 1,008 0.7 99.3 955 5 950 0.5 99.5 1.32 (0.42,4.17) 0.775 Visual Fields Light Reaction n Optic m Oculomotor p-Value m Ocular Movements Smile vn Facial Palpebral Fissures vn Facial 1-22 �TABLE 1-6. (ocntiiuad) UHU justea UK uyses tor Lianiai. Nerve Function by Group (Original Comparisons Only) Group Variable Balance Cranial Nerve vm Acoustic Gag Reflex IX Glossopharyngeal Speech X Vagus Original Comparison Nunber Percent Statistic Ranch Hand Number Percent n Abnormal Normal 1,015 2 1,013 0.2 99.8 955 1 954 0.1 99.9 1.88 ( . 7 2 . 1 0.999 01,08) n Abnormal Normal 1,014 1 1,013 0.1 99.9 955 1 954 0.1 99.9 0.94 ( . 6 1 . 8 00,50) n Abnormal Normal 1,016 3 1,013 0.3 99.7 954 0 954 0.0 100.0 6.59 ( . 4 1 7 8 ) 0.250 03,2.0* n Abnormal Normal 1,015 3 1,012 0.3 99.7 955 0 955 0.0 100.0 6.61 ( . 4 1 8 0 ) 0.250 03,2.6* Est. Relative Risk (95%C.I.) p-Value 0.999 Tongue Position Relative to Midline X Vagus Palate and Uvula Movement H Spinal Accessory n Abnormal Normal 1,014 2 1,012 0.2 99.8 955 1 954 0.1 99.9 1.89 ( . 7 2 . 3 0.999 01,08) Neck Range of Motion xn Hypoglossal n Abnormal Normal 1,016 61 955 6.0 94.0 955 64 891 6.7 93.3 0.89 ( . 2 1 2 ) 06,.8 0.579 n Abnormal Normal 1,003 . 96 907 9.6 90.4 941 89 852 9.5 90.5 1.01 ( . 5 1 3 ) 07,.7 0.939 Cranial Nerve Function Index 1-23 �TABLE 1-6. (continued) Unadjusted Analyses for Cranial Nerve junction ay uroup tunguEu umpansons uuy; Group Variable Cranial Nerve Cranial Nerve Function Index (Neck. Range of Motion Excluded) Statistic Ranch Hand Number Percent n Abnormal Norral 1,003 42 961 4.2 9. 58 Original Comparison Number Percent 941 29 912 3.1 9. 69 Est. Relative Risk ( 5 C.I.) p-Value 9* 1.34 ( . 5 2 2 ) 0.226 08,.3 ^Estimated relative risk and confidence interval calculated after adding 0.5 to each cell. 1-24 �TABLE 1-7. Adjusted Analyses for Selected Variables of Cranial Nerve Function by Group (Original Caparisons Only) Group Variable Statistic Ranch Hand Member Percent Original Comparison Nurber Percent Adj. Relative Risk ( 5 C.I.) p-Value 9% Covariate Remarks* n Neck Range of Abnormal Motion Normal 1,016 61 955 6.0 94.0 955 64 891 6.7 93.3 0.92 ( . 4 1 5 ) 0.647 06,.7 ACE(p<0.001) Cranial n Nerve Function Abnormal Index Normal 1,003 % 907 9.6 90.4 941 89 852 9.5 90.5 1.04 ( . 6 1 4 ) 0.812 07,.1 AGE(p<0.001) 964 38 926 3.9 96.1 911 28 883 3.1 96.9 1.27 ( . 7 2 1 ) 0.350 07,.2 DIAB*INS(p=0.032) OCC*OKm(p=0.016) OCC*DIAB(p=0.032) Cranial Nerve Function Index (Neck Range of Motion Excluded) n Abnormal Nornel Alternative Model—Includes Missing Drink-Year Participants* 'b n Abnormal Normal 1,003 42 961 4.2 95.8 938 28 910 3.0 97.0 1.41 ( . 6 2 3 ) 0.170 08,.2 DIAB*INS(i>O.022) OCC*DIAB(p=0.027) ^Abbreviations; DIAB: diabetic class INS: insecticide exposure OCC: occupation CROR: drink-years 'lifetime alcohol consumption (drink-years) not used as a covariate. b 66 missing drink-year participants: 4/39 Ranch Hands abnormal; 0/27 Original Comparisons abnormal. 1-25 �TABLE 1-8. Unadjusted Analyses for Peripheral Nerve Function by Group (Original Comparisons Only) Group Variable Statistic Ranch Hand Number Percent Original Comparison Number Percent Est,. Relative Risk (95% C.I.) p-Value n Pin Prick Abnormal Normal 1,003 59 944 5.9 94.1 943 62 881 6.6 93.4 0.89 (0.62,1 .28) 0.573 Light Touch n Abnormal Normal 1,003 38 965 3.8 96.2 943 33 910 3.5 96.5 1.09 (0.68,1 .75) 0.809 Muscle Status n Abnormal Normal 1,016 26 990 2.6 97.4 955 26 929 2.7 97.3 0.94 (0.54,1 .63) 0.888 n Vibratory Abnormal Sensation Normal 1,003 11 992 1.1 98.9 943 7 936 0.7 99.3 1.48 (0.57,3 .84) 0.482 Patellar Reflex n Abnormal Normal 1,016 11 1,005 1.1 98.9 954 14 940 1.5 98.5 0.74 (0.33,1 .63) 0.547 Achilles Reflex n Abnormal Normal 1,009 58 951 5.7 94.3 950 62 888 6.5 93.5 0.87 (0.60,1 .26) 0.510 Biceps Reflex n Abnormal Normal 1,016 9 1,007 0.9 99.1 955 10 945 1.0 99.0 0.85 (0.34,2 .09) 0.819 Babinski Reflex n Abnormal Normal 1,011 4 1,007 0.4 99.6 951 4 947 0.4 99.6 0.94 (0.24,3 .77) 0.999 1-26 �TABLE 1-9. Adjusted Analyses for Selected Variables of Peripheral Nerve Function by Group (Original Comparisons Only) Group Variable Statistic n Pin Prick Abnormal Nonnal Ranch Band Number Percent Original Comparison Number Percent 1,003 59 944 5.9 94.1 941 61 880 6.5 93.5 Adj. Relative Risk (95XC.I.) p-Value **** **** Covariate Remarks* GBP*DIAB(p=0.001) AGE(p<0.001) Light Touch Abnormal Normal 964 37 927 3.8 96.2 914 31 883 3.4 96.6 1.12 ( . 9 1 8 ) 0.652 06,.2 AGE(p=0.049) Muscle Status n Abnormal Normal 977 25 952 2.6 97.4 925 24 901 2.6 97.4 1.04 ( . 7 1 8 ) 0.893 05,.9 EHOR*AGE(D=O.Q37) DIAB*INS(p"0.028) Achilles Reflex n Abnormal Normal 1,009 58 951 5.7 94.3 947 61 886 6.4 93.6 0.92 ( . 3 1 3 ) 0.677 06,.5 AGE(p<0.001) DIAB(p<0.001) n *Additional Abbreviation: GBP: group ***><toup-by-<»variate interaction—adjusted relative risk, confidence interval, and p-value are not presented. 1-27 �TABLE 1-10. Summary of Group-by-Diabetic Class Interaction for Pin Prick (Original Comparisons Only) Group Variable Interaction Stratification Statistic Ranch Hand Number Percent Original Comparison Number Percent Adj. Relative Risk (95% C.I.) p-Value Abnormal Group-byPin Prick Diabetic Class n Abnormal Normal 76 13 63 17.1 82.9 74 7 67 9.5 90.5 2.03 (0.76,5.43) 0.161 Impaired n Abnormal Normal 105 1 104 1.0 99.0 135 14 121 10.4 89.6 0.08 ( . 1 0 6 ) 0.015 00,.1 Normal n Abnormal Normal 822 45 777 5.5 94.5 732 40 692 5.5 94.5 1.01 (0.65,1.57) 0.951 M 1 to 00 �TABLE 1-11. lhadjusted Analyses for CMS Coordination Variables by Group (Original Comparisons Only) Group Ranch Hand Number Percent Original Comparison Est. Relative Number Percent Risk ( 5 C.I.) p-Value 9% Variable Statistic Tremor n Abnormal Normal 106 ,1 26 990 2.6 97.4 955 14 941 1.5 1.77 ( . 2 3 4 ) 0.110 09,.0 9. 85 n Coordination Abnormal Normal 1,015 9 106 ,0 0.9 99.1 955 7 948 0.7 1.21 ( . 5 3 2 ) 0 8 4 04,.7 .0 99.3 Romberg Sign n Abnormal Normal 1,015 2 1,013 0.2 9. 98 955 1 954 0.1 1 8 ( . 7 2 . 1 0 9 9 .8 01,08) .9 9. 99 Gait n Abnormal Normal 106 ,1 20 996 2.0 9. 80 954 11 943 1.2 1.72 ( . 2 3 6 ) 0.153 08,.1 9. 88 n Abnormal Normal 1,015 48 967 4.7 95.3 954 29 925 3.0 1.58 ( . 9 2 5 ) 0.062 09,.3 9. 70 CNS Sunmary Index 1-29 �TABLE 1-12. Adjusted Analyses for Selected Variables of CMS Coordination by Group (Original Comparisons Only) Group Original Comparison Number Percent Variable Statistic Ranch Hand Number Percent Adj. Relative Risk ( 5 C.I.) p-Value 9% Covariate Remarks Tremor n Abnormal Normal 1,016 26 990 2.6 97.4 952 14 938 1.5 98.5 1.68 ( . 7 3 2 ) 0.112 08,.5 DIAB(p=0.002) Gait n Abnormal Normal 1,016 20 996 2.0 98.0 951 10 941 1.1 98.9 1.85 ( . 6 3 9 ) 0.105 08,.9 DIAB(p=0.035) n Abnormal Normal 1,015 48 967 '4.7 95.3 951 28 923 2.9 97.1 1.58 ( . 8 2 5 ) 0.055 09,.5 DIAB(p=0.009) M I CNS Summary Index �TABLE 1-13. Longitudinal Analysis of Romberg Sign and Babinski Reflex: A Contrast of Baseline and Follovup Examination Abnormalities (Original Comparisons Only) 1985 1982 Followup Baseline Exam Odds p-Value Exam Abnormal Normal Ratio (OR)* (ORt>u vs ORO C n^) Rn Variable Group 2 0 188 777 0 0 1 180 690 0.006 Ranch Hand Abnormal Normal 1 3 7 953 0.43 Original Abnormal Comparison Normal Babinski Reflex Abnormal Normal Abnormal Original Comparison Normal Romberg Sign 0 4 1 863 Ranch Hand 0.31 0.07 4.00 Number Normal Baseline, Abnormal Followup *OHH<! Rat In; - .... . Number Abnormal Baseline, Normal Follovup 1-31 �APPENDIX J Psychological Assessment �APPENDIX J: Psychological Assessment Contents Table Page J-l Summary of Kolmogorov-Smirnov Tests on MMPI for Officers J-l J-2 Summary of Kolmogorov-Smirnov Tests on MMPI for Enlisted Flyers J-2 J-3 Summary of Kolmogorov-Smirnov Tests on MMPI for Enlisted Groundcrev J-3 J-4 Summary of Group-by-Covariate Interactions for Psychological Variables J-4 J-5 Summary of Kolmogorov-Smirnov Tests on CMI J-7 J-6 Unadjusted Exposure Index Analyses for Psychological Variables by Occupation J-8 J-7 Interaction Summaries of Adjusted Exposure Index Analyses for Psychological Variables J-17 J-8 Unadjusted Analyses for Reported Psychological Illnesses by Group; Baseline and First Followup Studies Combined (Original Comparisons Only) J-27 J-9 Summary of Kolmogorov-Smirnov Tests on MMPI for Officers (Original Comparisons Only) J-28 J-10 Summary of Kolmogorov-Smirnov Tests on MMPI for EnlistedFlyers (Original Comparisons Only) J-29 J-ll Summary of Kolmogorov-Smirnov Tests on MMPI for Enlisted Groundcrew (Original Comparisons Only) J-30 J-12 Summary of Kolmogorov-Smirnov Tests on CMI (Original Comparisons Only) J-31 J-13 Unadjusted Analyses for MMPI by Group (Original Comparisons Only) J-32 J-14 Adjusted Analyses for MMPI by Group (Original Comparisons Only) J-34 J-15 Unadjusted Analyses for the Cornell Medical Index (CMI) by Group J-36 J-16 Adjusted Analyses for CMI Variables by Group (Original 'Comparisons Only) J-37 J-i �APPENDIX J: Psychological Assessment Contents Table J-17 Summary Results for the Halstead-Reitan Battery Impairment Index Unadjusted and Adjusted Analyses (Original Comparisons Only) J-38 J-18 Summary of Group-by-Covariate Interactions for Psychological Variables (Original Comparisons Only) J-39 J-ii �TABLE J-l. Summary of Kolmogorov-Smirnov Tests on MMPI for Officers Parameter* Group Mean Score Standard Deviation KolmogorovSmirnov p-Value Anxiety Ranch Hand Comparison 52.82 53.64 7.38 7.77 0.281 Consistency Ranch Hand Comparison 49.23 49.23 6.92 6.62 0.995 Defensiveness Ranch Hand Comparison 49.64 49.61 6.64 6.97 0.990 Denial Ranch Hand Comparison 59.47 59.81 7.31 7.46 0.877 Depression Ranch Hand Comparison 55.11 55.27 10.17 10.28 0.999 Hypochondria Ranch Hand Comparison 55.60 55.02 8.85 8.76 0.970 Hysteria Ranch Hand Comparison 60.45 59.97 7.00 7.47 0.627 Mania/Hypomania Ranch Hand Comparison 54.36 53.39 8.94 9.12 0.092 Masculinity/ Femininity Ranch Hand Comparison 58.83 58.10 8.64 9.07 0.366 Paranoia Ranch Hand Comparison 53.71 53.91 7.34 7.34 0.975 Psychopathic/ Deviate Ranch Hand Comparison 56.82 57.08 8.87 10.02 0.837 Schizophrenia Ranch Hand Comparison 54.80 55.14 8.48 9.02 0.942 Social Introversion Ranch Hand Comparison 46.56 47.19 7.43 7.77 0.594 Validity Ranch Hand Comparison 0.93 0.89 3.11 3.20 0.776 *n=380 Ranch Hands; n=480 Comparisons (except for validity, where n=483 Comparisons). J-l �TABLE J-2. Summary of Kolmogorov-Smirnov Tests on MMPI for Enlisted Flyers Group Mean Score Standard Deviation KolmogorovSmirnov p-Value Anxiety Ranch Hand Comparison 54.78 55.09 9.73 10.09 0.678 Consistency Ranch Hand Comparison 52.24 51.83 10.30 8.22 0.995 Defensiveness Ranch Hand Comparison 53.05 52.00 7.96 8.31 0.270 Denial Ranch Hand Comparison 55.95 56.95 8.73 8.50 0.807 Depression Ranch Hand Comparison 58.21 58.32 11.79 11.86 0.744 Hypochondria Ranch Hand Comparison 58.48 59.25 12.16 12.85 0.496 Hysteria Ranch Hand Comparison 60.00 61.32 8.94 9.47 0.366 Mania/Hypomania Ranch Hand Comparison 55.05 54.99 10.00 9.87 0.496 Masculinity/ Femininity Ranch Hand Comparison 55.79 56.57 7.85 8.46 0.776 Paranoia Ranch Hand Comparison 52.77 51.92 8.43 7.83 0.695 Psychopathic/ Deviate Ranch Hand Comparison 57.48 58.95 11.40 10.79 0.088 Schizophrenia Ranch Hand Comparison 56.56 57.06 13.03 11.49 0.178 Social Introversion Ranch Hand Comparison 50.35 49.48 8.89 8.14 0.678 Validity Ranch Hand Comparison 3.93 6.31 42.48 54.96 0.864 Parameter* *n=176 Ranch Hands; n=208 Comparisons (except for validity, where n=177 Ranch Hands and n=210 Comparisons). J-2 �TABLE J-3. Summary of Kolmogorov-Srairnov Tests on HMPI for Enlisted Groundcrew Group Mean Score Standard Deviation KolmogorovSmirnov p-Value Anxiety Ranch Hand Comparison 55.86 55.61 10.96 9.58 0.993 Consistency Ranch Hand Comparison 53.05 52.03 10.64 8.43 0.877 Defensiveness Ranch Hand Comparison 52.42 52.60 8.27 8.35 0.970 Denial Ranch Hand Comparison 55.53 56.52 8.57 8.88 0.118 Depression Ranch Hand Comparison 58.79 57.66 12.98 10.78 0.594 Hypochondria Ranch Hand Comparison 58.44 57.13 13.22 11.12 0.142 Hysteria Ranch Hand Comparison 60.36 59.78 9.80 8.99 0.577 Mania/Hypomania Ranch Hand Comparison 55.26 56.33 9.78 10.17 0.450 Masculinity/ Femininity Ranch Hand Comparison 56.18 57.18 8.24 8.82 0.292 Paranoia Ranch Hand Comparison 53.12 53.73 9.05 8.47 0.393 Psychopathic/ Deviate Ranch Hand Comparison 59.01 59.26 11.29 10.66 0.728 Schizophrenia Ranch Hand Comparison 58.20 57.36 14.22 11.12 0.837 Social Introversion Ranch Hand Comparison 51.86 50.43 9.82 8.71 0.107 Validity Ranch Hand Comparison 1.97 0.89 26.48 4.80 0.999 Parameter* *n=458 Ranch Hands; n=598 Comparisons (except for validity, where n=459 Ranch Hands and n=i599 Comparisons). J-3 �TABLE J-4. SuBBary of Group-by-Covariate Interactions for Psychological Variables Group Variable Interaction Anxiety Group-byEducation Statistic High School n Number/* Abnormal Normal 564 n Number/* Abnormal Normal 448 n Number /Z Abnormal Normal 537 •n Number/* Abnormal Normal 437 n Number /Z Abnormal Normal 192 n Number/* Abnormal Normal 368 n Number/* Abnormal Normal 414 College Consistency Group-byEducation High School College Depression Group-byCombat Index Ranch Hand Stratification Lov Medium High 57 507 16 432 30 507 6 431 28 164 37 331 43 371 Comparison Adj. Relative Risk (95Z C.I.) p-Value 685 10. 1Z 89. 9Z 49 636 7.2Z 92. 8Z 1.39 (0.92,2.08) 0.114 5.0Z 95. OZ 0.68 (0.36,1.28) 0.233 2.9Z 97. 5Z 1.81 (1.00,3.28) 0.051 2.6Z 97. 4Z 0.46 (0.17,1.19) 0.110 8.2Z 91.8* 1.73 (1.03,2.91) 0.039 9.1Z 90. 9Z 1.09 (0.67,1.75) 0.657 13.9* 86.1* 0.74 (0.48,1.16) 0.159 600 3.6Z 96. 4Z 30 570 661 5.6Z 94. 4Z 19 642 585 1.4Z 98.6* 15 570 490 14.6* 85.4* 40 450 417 10. 1Z 89. 9Z 38 379 339 10.4* 89.6* 47 292, �TABLE J-4. (continued) Summary of Group-by-Covariate Interactions for Psychological Variables Group Variable Paranoia Interaction Group-by Age Statistic Born >1942 n Number/Z Abnormal Normal 410 n Number/Z Abnormal Normal 602 n Number/Z Abnormal Normal 537 n Number/Z Abnormal Normal 439 n Number/Z Abnormal Normal 198 n Number/Z Abnormal Normal 384 n Number/* Abnormal Normal 430 Born <1942 Schizophrenia Group-byEducation High School «-. I College Social Introversion Group-byCombat Index Ranch Hand Stratification Low Hedium High 15 395 16 586 72 465 18 421 11 187 6 378 9 421 Adj. Relative Comparison Risk (95Z C.I.) p-Value 3.5Z 96. 5Z 0.88 (0.43,1.77) 0.712 1.2Z 98.8* 2.63 (1.11,6.20) 0.027 9.5Z 90. 5Z 1.51 (1.05,2.16) 0.033 6.3Z 93. 7Z 0.63 (0.35,1.12) 0.119 1.2Z 98.8* 4.86 (1.77,13.36) 0.002 0.9* 99. 1Z 1.59 (0.44,5.68) 0.478 2.6* 97.4* 0.82 (0.32,2.09) 0.677 545 3.7Z 96.3* 19 526 740 2.7Z 97. 3Z 9 731 665 13.4* 8.* 66 63 602 585 4.1Z 95. 9Z 37 548 509 5.6Z 94. 4Z 6 503 427 1.6* 98.4* 4 423 349 2.1Z 97. 9Z 9 340 �TABLE J-4. (continued) Summary of Group-by-Covariate Interactions for Psychological Variables Group Variable Interaction Validity Group-by- Stratification Nonblack Statistic Ranch Hand n 208 746 Risk ( 5 C.I.) 9% p-Value 1,206 954 Number/% >0 0 Race Adj. Relative Comparison 235 971 1.20 (0.97,1.49) 0.095 42.2% 57.8% 0.46 ( . 2 0 9 ) 02,.6 0.038 31.43 95% C.I.* ( 0 2 ,45.65) (25 . 6 3 . 6 3.5 4,87) Black 19.5% 80.5% (-1 Total CMI I Group-byEducation High School — <0.001 574 n 433 Number/% Adj. Mean* 23.72 24.22 95% C.I.* (19.10,29.41) (19 .57,29.91) — 0.657 n 15 45 83 60 Number/% >0 0 21.8% 78.2% n Number/% 25.0% 75.0% 35 48 655 529 Adj. Mean* 37.18 College M-R Subscore Group-byEducation High School College n Number/% 0 (Low) l-lO(Medium) >10 (High) n Number/% 675 555 246 267 42 44.3% 48.1% 7.6% 0.030 352 283 40 52.1% 41.9% 5.9% Medium vs. Low 1.37 (1.07,1.75) High vs. Low 1.33 (0.82,2.15) 0.014 63.2% 33.9% 2.9% Overall: Medium vs. Low 0.91 (0.70,1.18) High vs. Low 0 8 (0.35,1.84) .0 590 443 0 (Low) 292 l-lO(Medium) 141 >10 (High) 10 Overall: 65.9% 31.8% 2.3% 373 200 17 transformed from log(X+l) scale, where X was the number of questions answered "yes." —No relative risk given for Total CMI, which was analyzed as a continuous variable. 0.254 0.699 0.465 0.597 �TABLE J-5. Summary of Kolmogorov-Smirnov Tests on CHI Group Mean Score Ranch Hand Comparison 402 543 15.88 13.07 15.21 12.40 <0.001 Born 1923-1941 Ranch Hand Comparison 562 674 15.49 14.64 12.99 12.52 0.281 Born <1922 Ranch Hand Comparison 36 51 18.56 16.57 » 15.94 9.27 0.970 941 Ranch Hand Comparison 1,187 15.63 13.94 13.83 12.00 0.003 Ranch Hand Comparison 59 81 17.81 15.68 16.91 17.01 0.481 Ranch Hand Comparison 556 677 18.88 15.77 15.80 13.50 <0.001 Ranch Hand Comparison 444 591 11.85 12.08 10.18 10.63 0.999 Current Alcohol Use Drinker 851 Ranch Hand Comparison 1,092 15.16 13.72 13.12 12.00 0.024 Stratification Age Born XL942 Race Nonblack Black Education High School College Standard Deviation KolmogorovSmirnov p-Value Sample Size Ranch Hand Comparison 148 176 18.84 16.06 17.71 14.38 0.149 Ranch Hand Comparison 377 476 11.23 10.88 9.01 9.09 0.807 Enlisted Flyer Ranch Hand Comparison 174 204 17.72 15.63 14.84 13.53 0.220 Enlisted Groundcrev Ranch Hand Comparison 449 588 18.80 16.06 16.06 13.69 0.007 Nondrinker Occupation Officer J-7 �TABLE J-6. Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Exposure Index Lov Medium n Number/Z Abnormal Normal 127 n Number/Z Abnormal Normal 54 65 Enlisted Flyer 8 14.82 46 85 .2Z 4 6.2Z 61 93. 8Z n Number/Z Abnormal Normal 154 163 Enlisted Groundcrev 19 12 .3Z 135 87 .7Z Officer Anxiety Contrast p-Value Statistic n Number/Z Abnormal Normal n Number/Z Abnormal Normal 54 65 57 Overall 0.476 Enlisted Flyer 4 7.4Z 50 92 .6Z 2 3.1Z 63 96.92 2 3.5Z 55 96. 5Z M vs. L H vs. L 0.40 (0.07,2. 26) 0.409 0.46 ( . 8 2 59) 0.430 00,. n Number/Z Abnormal Normal 154 163 Overall 0.813 Enlisted Groundcrev 9 5.82 145 94 .2Z 9 5.5Z 6 4.3Z 154 94. 5Z 135 95. 7Z H vs. L H vs. L 0.94 (0.36,2.44) 0.999 0.72 (0.25,2.07) 0.603 130 High Est. Relative Risk (95Z C.I .) Occupation Officer Variable Overall 0.876 M vs. L H vs. L 1.48 (0.24,8.99) 0.999 1.56 (0.26,9.52) 0 6 0 .8 57 Overall 0.211 4 7.0Z 53 93. OZ H vs. L H vs. L 0.38 (0.11,1.33) 0.137 0.43 (0.12,1.54) 0.230 Overall 0.560 18 11. OZ 12 8.5Z 145 89. OZ 129 91. 5Z M vs. L B vs. L 0.88 (0.44,1. 75) 0.730 0.66 (0.31,1.42) 0.343 127 130 Overall 0.380 2 1. Z 6 125 98 .4Z 0 O.OZ M vs. L 2 1.5Z 128 98.52 123 100. OZ H vs. L 0.98 (0.14,7.04) 0.999 0.20 (0.01,4. 28)* 0.498 123 2 1 .62 3 2.3Z 3 2.4Z 125 98 • 4Z 127 97. 7Z 120 97. 6Z 4 I 00 Consistency 141 123 141 �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Low Exposure Index Medium High Est. Relative Risk (95% C.I.) Occupation Statistic n Number/% Abnormal Normal 127 Officer n Number/% Abnormal Normal 54 n Number /% Abnormal Normal 154 Enlisted Groundcrew n Number/% Abnormal Normal 127 Officer Variable n Number/% Abnormal Normal 54 65 Overall 0.366 Enlisted Flyer 2 3.7% 52 96.3% 2 3.1% 0 0.0% M vs. L 63 96.9% 57 100.0% H vs. L 0.83 ( .11,6.06) 0.999 0 0.18 ( .01,3.89)* 0.234 0 n Number/% Abnormal Normal 154 Enlisted Groundcrew Defensiveness Enlisted Flyer Denial 130 123 Contrast p-Value Overall 0.603 0 0.0% 1 0.8% 1 0.8% M vs. L 127 100.0% 129 99.2% 122 99.2% H vs. L 2.95 (0.12,73.18)* 0.999 3.12 (0.13,77.39)* 0.492 65 57 Overall 0.991 1 1.9% 1 1.5% 1 1.8% M vs. L 53 98.1% 64 98.5% 56 98.2% H vs. L 0 0.83 ( .05,13.56) 0.999 0.95 ( .06,15.52) 0.999 0 Overall 0.697 163 141 5 3.1% 6 3.9% 7 5.0% M vs. L 148 96.1% 158 96.9% 134 95.0% H vs. L 130 123 Overall 1 0.8% M vs. L 6 4.7% 2 1.5% 121 95.3% 128 98.5% 122 99.2% H vs. L 57 141 0.78 ( .23,2.61) 0 0 1.29 ( .42,3.93) 0.765 0.779 0.094 0.32 ( .06,1.59) 0 0.17 ( .02,1.39) 0 0.169 0.120 Overall 0.326 2 1.3% 0 0.0% 2 1.4% M vs. L 152 98.7% 163 100.0% 139 98.6% H vs. L 0.19 ( .01,3.92)* 0.235 0 0 1.09 ( . 15,7.87) 0.999 163 �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Low Exposure Index Medium High c_ ! Statistic n Number/% Abnormal Normal 127 130 Officer 8 6.3% 119 93.7% 6 4.6% 10 8. 1% 124 95.4% 113 91.9% n Number/% Abnormal Normal 54 65 Enlisted Flyer 11 20.4% 43 79.6% 7 10.8% 58 89.2% n Number/% Abnormal Normal 154 Enlisted Groundcrev Officer Depression Occupation Hypochondria Enlisted Flyer Enlisted Groundcrev 123 Est. Relative Contrast Overall Risk (95% C.I •) p-Value 0.517 M vs. L H vs. L 0.72 (0.24,2.14) 0.593 1.32 (0.50,3.45) 0.631 57 Overall 0.223 6 10.5% 51 89.5% M vs. L H vs. L 0.47 (0.17,1.32) 0.199 0.46 (0.16,1.35) 0.191 Overall 0.695 25 16.2% 21 12.9% 20 14.2% 129 83.8% 142 87.1% 121 85.8% M vs. L H vs. L 0.76 (0.41,1.43) 0.428 0.85 (0.45,1.62) 0.746 n Number/% Abnormal Normal 127 Overall 0.530 M vs. L H vs. L 0.57 (0.20,1.61) 0.312 0.92 (0.36,2. 36) 0.999 n Number/% Abnormal Normal 54 65 57 Overall 0.059 11 20.4% 43 79.6% 4 6.2% 61 93.8% 10 17.5% 47 82.5% H vs. L H vs. L 0.26 ( . 8 0 86) 0.026 00,. 0.83 (0.32,2.15) 0.810 n Number/% Abnormal Normal 154 163 25 16.2% 129 83.8% 18 11.0% 26 18.4% 145 89.0% 115 81.6% 163 130 141 123 9 7.3% 10 7.9% 6 4.6% 117 92.1% 124 95.4% 114 92.7% 141 Overall M vs. L H vs. L 0.176 0.64 (0.33,1.23) 0.192 1.17 (0.64,2. 13) 0.646 �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Low Exposure Index Medium High 130 7 5.5% 120 94.5% 9 6.9% 14 11.4% 121 93.1% 109 88.6% n Number/% Abnormal Normal 54 65 57 Overall Enlisted Flyer 11 20.4% 43 79.6% 6 9.2% 59 90.8% 7 12.3% 50 87.7% M vs. L H vs. L 0.40 (0.14,1.16) 0.115 0.55 (0.20,1.54) 0.307 n Number/% Abnormal Normal 154 163 Overall 0.065 Enlisted Groundcrew 28 18.2% 126 81.8% 16 9.8% 25 17.7% 147 90.2% 116 82.3% M vs. L H vs. L 0.49 (0.25,0.95) 0.035 0.97 (0.54,1.76) 0.999 n Number/% Abnormal Normal 127 130 Overall 0.585 Officer Mania/ Hypomania n Number/% Abnormal Normal 127 Officer Hysteria Statistic 8 6.3% 119 93.7% 8 6.5% 5 3.9% 125 96.1% 115 93.5% M vs. L H vs. L 0.60 (0.19,1.87) 0.407 1.04 (0.38,2.85) 0.999 n Number /% Abnormal Normal 54 65 57 Overall 0.623 Enlisted Flyer 3 5.6% 51 94.4% 6 9.2% 59 90.8% 6 10.5% 51 89.5% M vs. L H vs. L 1.73 (0.41,7.27) 0.509 2.00 (0.47,8.44) 0.491 n Number/% Abnormal Normal 154 163 Overall 0.568 Enlisted Groundcrew 11 7.1% 143 92.9% 6 4.3% 10 6.1% 153 93.9% 135 95.7% M vs. L H vs. L 0.85 (0.35,2.06) 0.823 0.58 (0.21,1.61) 0.326 123 141 123 141 Contrast Est. Relative Risk (95% C.I . ) Occupation p-Value Overall 0.200 M vs. L H vs. L 1.28 (0.46,3.53) 0.797 2.20 (0.86,5.66) 0.113 0.198 �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Occupation Statistic Officer n Number/% Abnormal Normal n Masculinity/ Femininity Enlisted Flyer Low Exposure Index Medium High 127 130 7 5.5% 120 94.5% 115 88.5% Est. Relative Contrast Risk (95% C • I.) p-Value 0.142 123 15 11.5% Overall 15 12.2% 108 87.8% M vs. L B vs. L 08, 2.24 ( . 8 5.68) 0.118 2.38 ( . 4 6.06) 0.075 09, Overall 0.478 M vs. L H vs. L 3.48 (0.38,32.06) 0.375 1.93 (0.17,21.89) 0.999 54 65 57 1 1.9% 53 98.1% 4 6.2% 61 93.8% 55 96.5% Number/% Abnormal Normal 2 3.5% 154 163 141 Overall 0.360 Normal 10 6.5% 144 93.5% 5 3.1% 158 96.9% 7 5.0% 134 95.0% M vs. L H vs. L 0.46 (0.15,1.37) 0.189 0.75 (0.28,2.03) 0.625 n Number/% 127 130 123 Overall 0.194 Abnormal 1 0.8% 126 99.2% 1 0.8% 129 99.2% 4 3.3% 119 96.7% M vs. L H vs. L 0.98 ( . 6 15.79) 0.999 00, 4.24 (0.47,38.44) 0.208 Overall 0.447 M vs. L H vs. L 0.27 (0.03,2.63) 0.328 0.94 (0.18,4.90) 0.999 141 Overall 0.735 7 5.0% 134 95.0% M vs. L H vs. L 1.14 (0.34,3.81) 0.999 1.56 ( . 8 5.02) 0.560 04, n Enlisted Groundcrev Officer Number/% Abnormal Normal Paranoia Enlisted Flyer n Number/% Abnormal Normal n Enlisted Abnormal 65 3 5.6% 51 94.4% 1.5% 3 5.3% 64 98.5% 54 94.7% 1 154 163 5 3.3% 149 96.7% 157 96.3% Number/% Groundcrev 57 54 Normal 6 3.7% �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Low Exposure Index Medium High Est. Relative Occupation Statistic n Number/% Abnormal Normal 127 130 Officer 9 7.1% 118 92.9% 7 5.4% 11 8.9% 123 9 . % 112 91.1% 46 n Number/% Abnormal Normal 54 65 8 14.8% 46 85.2% 11 16.9% 6 10.5% 54 83.1% 51 89.5% n Number/% Abnormal Normal 154 163 Enlisted Groundcrev 25 16.2% 129 83.8% Officer n Number/% Abnormal Normal n Number/% Abnormal Normal 54 65 57 Overall 8 14.8% 46 85.2% 6 9.2% 59 9 . % 08 5 8.8% 52 91.2% M vs. L H vs. L 0.59 (0.19,1 . 0 0 4 0 8) .0 0.55 (0.17,1 .81) 0.385 n Number/% Abnormal Normal 154 163 141 Overall 0.600 18 11.7% 136 88.3% 25 15.3% 138 84.7% 21 14.9% 120 85.1% M vs. L H vs. L 1.37 (0.71,2 .62) 0.413 1.32 (0.67,2 . 0 0.492 6) Psychopathic/ Enlisted Deviate Flyer - Schizophrenia Enlisted Flyer Enlisted Groundcrev 123 Contrast Risk (95% C.I.) p-Value Overall 0.545 M vs. L H vs. L 0.75 (0.27,2 . 7 0.614 0) 05, .4 1.29 ( . 1 3 .23) 0 6 6 Overall 0.593 M vs. L H vs. L 1.17 (0.43,3 .16) 0 8 6 .0 0.68 (0.22,2 .10) 0.574 Overall 0.220 18 11.0% 25 17.7% 145 8 . % 116 82.3% 90 M vs. L H vs. L 0.64 (0.33,1 .23) 0.192 1.11 (0.61,2 . 4 0.758 0) 127 130 Overall 0.546 5 3.9% 122 96.1% 2 1.6% 4 3.1% 126 96.9% 121 98.4% M vs. L H vs. L 0.78 (0.20,2 • 95) 0.747 0.40 ( . 8 2 .12) 0.447 00, 57 141 123 0.519 �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Exposure Index Medium High p-Value Statistic n Number/Z Abnormal Normal 127 n Number/Z Abnormal Normal 54 65 57 Overall 0.696 2 3.7Z 52 96. 3Z 1 1.5Z 64 98.5% 1 1.8Z 56 98. 2% M vs. L H vs. L 0.41 ( . 4 4 .61) 0.590 00, 0.46 ( . 4 5 .27) 0.611 00, n Number/Z Abnormal Normal 154 Overall 0.635 Enlisted Groundcrew 6 4.3Z 8 5.2Z 5 3.1Z 146 94. 8Z 158 96. 9Z 135 95. 7Z M vs. L H vs. L 0.58 (0.19,1 .81) 0.403 4) 0.81 (0.27,2 . 0 0.788 n Number/Z >0 0 127 Overall 0.052 Officer 32 25. 2Z 95 74. 8Z 34 26. 2Z 18 14. 6Z 96 73. 8Z 105 85.4% H vs. L H vs. L 1.05 ( . 0 1 .84) 0.887 06, 0.51 (0.27,0 .97) 0.041 n Number/Z Abnormal Normal 55 65 57 Overall 0.310 Enlisted Flyer 10 18. 2Z 45 81. 8Z 8 12. 3% 57 87. 7Z 13 22. 8Z 44 77.2% M vs. L H vs. L 0.63 (0.23,1 .73) 0.445 1.33 (0.53,3 .35) 0.642 n Number/Z >0 0 154 Overall 0.802 Enlisted Groundcrev M vs. L H vs. L 1.11 (0.66,1 .87) 0.790 1.20 (0.70,2 .05) 0.584 Enlisted Social Introversion Flyer c_ i _ Validity Low Contrast Est. Relative Risk (95% C.I.) Occupation Officer Variable Overall 0.385 0 O.OZ 1 0.8% M vs. L 2 1.5Z 126 99.2% 128 98. 5Z 123 100. OZ H vs. L 1.97 (0.18,21.99) 0.999 4) 0.34 (0.01,8 . 6 * 0.999 130 163 130 163 123 141 123 142 34 22. 1Z 39 23. 9% 36 25. 4Z 120 77. 9Z 124 76. 1Z 106 74. 6Z �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Variable Exposure Index Medium High Contrast Cn Statistic Officer Enlisted Flyer n 53 Mean** 13.69 95% C.I.** (10.63, 17.56) 65 13.13 (11.16, 15.42) 56 13.96 (11.51, 16.88) Overall M vs. L H vs. L n 153 Mean** 14.46 95% C.I.** (12.64, 16.52) 159 13.03 (11.57, 14.65) 137 14.92 (13.05, 17.05) Overall M vs. L M vs. L n Number/% >0 0 125 Officer 41 32.8% 84 67.2% 37 28.9% 91 71.1% n Number/% Abnormal Normal 53 22 41.5% 31 58.5% n Number/% >0 0 154 Est. Relative Risk (95% C.I.) 128 123 n 126 Overall 8.55 Mean** 10.04 M vs. L 7.99 95* C.I.** (6.84,9.31) (7.52,9.69) (8.99,11.20) H vs. L Enlisted Groundcrev Total Cornell Medical Index Occupation M-R Subscore Enlisted Flyer Enlisted Groundcrev Low 91 59.1% 63 40.9% 121 — — " p-Value 0.049 0.506 0.018 0.906 0.784 0.906 0.304 0.252 0.745 Overall 0.247 47 38.8% 74 61.2% M vs. L H vs. L 0.83 (0.49,1.42) 0.586 1.30 (0.77,2.19) 0.353 65 56 Overall 0.400 35 53.8% 30 46.2% 26 46.4% 30 53.6% H vs. L H vs. L 1.64 (0.79,3.42) 0.200 1.22 (0.57,2.61) 0.700 Overall 0.592 M vs. L H vs. L 0.86 (0.55,1.35) 0.567 0.79 (0.50,1.25) 0.346 128 157 87 55.4% 70 44.6% 139 74 53.2% 65 46.8% �TABLE J-6. (continued) Unadjusted Exposure Index Analyses for Psychological Variables by Occupation Est. Relative Exposure Index Medium c_ I Statistic n Number/% >0 0 113 57 50.4% 56 49.6% 56 50.0% 56 50.0% n Number/2 >0 0 50 Enlisted Flyer 33 66.0% 17 34.0% n Number/% >0 0 136 Enlisted Groundcrev A-H Area Subscore Occupation Officer Variable n Number/% Abnormal Normal 126 n Number/% Abnormal Normal Officer HRB Impairment Index Enlisted Flyer n Enlisted Groundcrev Number/% Abnormal ' Normal Low 112 High 114 Contrast Risk (95% C.I.) p-Value Overall 090 .8 56 49.1% 58 50.9% M vs. L H vs. L 0.98 ( . 8 1 .66) 0 9 9 05, .9 0.95 (0.56,1 . 0 0.895 6) 60 50 Overall 0.946 39 65.0% 21 35.0% 34 6 . % 80 16 32.0% H vs. L H vs. L 0.96 (0.43,2 .11) 04, 1.10 ( . 8 2 .52) 0.999 0.999 Overall 0.557 M vs. L H vs. L 0.82 ( . 0 1 •34) 0.455 05, 1.07 (0.63,1 . 0 0 8 4 8) .9 Overall 0.401 29 23.0% 97 77.0% 21 17.4% M vs. L 31 23.9% 99 76.1% 100 82.6% H vs. L 1.05 ( . 9 1 .87) 0.884 05, 0.70 (0.38,1 .32) 0.342 52 65 57 Overall 0.253 20 38.5% 32 61.5% 35 53.9% 30 46.1% 27 47.4% H vs. L 30 52.6% H vs. L 1.87 ( . 9 3 .92) 0.136 08, 1.44 (0.67,3 . 9 0.439 0) Overall 0.272 93 68.4% 43 31.6% 153 64 41.8% 89 58.2% 150 96 6 . % 40 54 36.0% 130 163 72 44.2% 91 55.8% 129 90 69.8% 39 30.2% 121 139 49 35.3% 90 64.7% M vs. L H vs. L 1.10 (0.71,1 .72) 0.76 (0.47,1 .22) ^Estimated relative risk and confidence interval calculated after adding 0.5 to each cell. **Converted from log (X+l) scale, where X was the number of questions answered "yes." —No relative risk given for Total Cornell Medical Index, which was analyzed as a continuous variable. 0.733 0.280 �TffilE J-7. Variable Interaction (Occupation) Anxiety Exposure Index-byAge (Officer) High Adj. Relative Contrast Risk (95% C I ) .. p-Value 138 n Number/% Abnormal Normal 16 14 5 31.3% 11 68.7% 2 12 n Number/% Abnormal Normal 43 22 11 2 4.7% 41 95.3% 0 00 .% 22 1 0 0 0.% 00,.2* 0 0.0% M vs. L 0.37 ( . 2 8 0 ) * 0.545* 0.% 00,61)* . 9 * 11 1 0 0 H vs. L 0.72 ( . 3 1 . 2 * 0 9 9 n Number/% Abnormal Normal 79 93 Born 192314() 91a 3 3.8% 76 96.2% 1 92 n Number/% Abnormal Normal 5 15 Born <=1922 . 1 20.0% 4 80.0% 1 14 Black Born >=1942 Denial Exposure Index Medium n Number/% Abnormal Normal Nonblack Exposure Index-byRace (Enlisted Groundcrew) Low Stratification Statistic 14 124 149 10.1% 89.9% 16 133 128 1.% 07 89.3% 12 9.4% M vs. L 1.07 ( . 0 2 2 ) 05,.7* 116 90.6% H vs. L 0.92 ( . 1 2 0 ) 04,.6* 13 14.3% 85.7% Overall Overall 1 0.9% M vs. L 0.27 ( . 3 2 6 ) 00,.1 108 99.1% H vs. L 0.25 ( . 3 2 4 ) 00,.9 3 6.7% 93.3% Overall 092 .3* 099 .9* 099 .9* 004 .7* 0 0.0% M vs. L 0.37 ( . 6 2 2 ) 00,.9* 0 3 9 .9* 0.% 13 1 0 0 H vs. L 0 0 ( . 0 , . 6 * 0 0 8 . 8 00415)* . 4 * 109 1.1% 9.% 89 Overall Overall 0.455* 0.337 0.255 0.238 0.558* .4* 0 0.0% M vs. L 0 2 ( . 1 5 6 ) . 9 00,.6* 0 4 7 3 1 0 0 H vs. L 0 4 ( . 1 1 . 8 * 0 9 9 0.% . 3 00,40)* . 9 * �— -- X —- — • • •f Interaction Summaries of Adjusted Exposure Index Analyses for Psychological Variables "Variable Interaction (Occupation) Stratification Statistic Nonblack Depression Exposure Index-byRace (Enlisted Groundcrew) Black High. School Exposure Hypochondria Index-byEducation (Officer) College Nonblack Exposure Index-byIfypochondria Race (Enlisted Groundcrew) Low Exposure Index Medium n Number/% Abnormal Normal 138 n Number/% Abnormal Normal 16 14 4 25.0% 12 75.0% 1 13 n Number/% Abnormal Normal 11 17 n Number/% Abnormal Normal 116 IB 6 5.2% 110 9 . % 48 3 110 n Number/% Abnormal Normal 134 146 n Number/% Abnormal Normal 14 21 117 149 15.2% 8.% 48 4 3.% 64 7 63.6% 17 117 12.7% 87.3% 20 129 3 14 15 131 High 128 13.4% 86.6% 086 .5* 077 .3* 099 .9* 002 .9* Overall 001 .1* .% . 8 00,.6* 0 3 1 0 0 0 M vs. L 0 3 ( . 7 2 1 ) .8* 24 1 0 0 H vs. L 0.03 ( . 0 , . 1 * 0 0 6 0.% 00207)* . 0 * 99 2.7% 97.3% Overall p-Value 0 0 0 M vs. L 0.23 ( . 2 2 3 ) .% 0 0 , . 7 * 0.336* 0.% 00,.1* 13 1 0 0 H vs. L 0.10 ( . 1 2 1 ) * 0 1 7 .0* 24 17.7% 82.3% Overall 04,.7* 20 15.6% M vs. L 0.86 ( . 5 1 6 ) 44 108 8 . % H vs. L 1.03 ( . 3 2 0 ) 05,.1* 13 7.1% 92.9% Adj. Relative Contrast Risk (95%C.I.) Overall 9 9.1% M vs. L 0 5 ( . 2 2 0 ) . 0 01,.5* 90 90.9% H vs. L 1.83 ( . 3 5 3 ) 06,.4* 010 .2* 049 .9* 023 .9* 119 10.3% 8.% 97 .2 03,.4 25 21.0% M vs. L 0 8 ( . 9 1 7 ) 94 79.0% H vs. L 1.67 ( . 3 3 3 ) 08,.3 066 .0 018 .4 o Black 6 42.9% 8 57.1% 14 3 11 12 21.4% 78.6% 1 8.3% M vs. L 0.31 ( . 6 1 7 ) 00,.0 11 91.7% H vs. L 0 1 ( . 1 1 0 ) . 0 00,.2 0.179 002 .5 �TABLE J-7. (continued) U — —-^ - TntorarHm &BBari«>>3 nf A/HnstpH Rnwam> Tnrlear AnalLvaes Variable Interaction (Occupation) Stratification Statistic High School Hysteria Exposure Index-byEducation (Officer) College<b) Nonblack <r Hysteria Exposure Ihdex-byRace (Enlisted Groundcrev) Black 0 Mania/ Hypomania Exposure Index-byDrink-years (Officer) 0-50 >50 Low Exposure Index High Medium n Number/% Abnormal Normal 11 n Number/% Abnormal Normal 116 n Number/% Abnormal Normal 138 n Number/% Abnormal Normal 16 n Number/% Abnormal Normal 5 n Number/% Abnormal Normal 105 n Number/% Abnormal Normal 15 18 0 0.0% 15 100.0% 0 18 2 9 5 11 1 23 115 5 11 17 18.2% 81.8% 11.8% 88.2% 4.3% 95.7% 7 106 149 16.7% 83.3% 13 136 14 31.2% 68.8% 3 11 21.4% 78.6% 25 19.5% H vs. L 103 80.5% H vs. L 0.029* 0.48 (0.23,0.99)* 1.21 (0.65,2.27)* 0.050* 0.633* 0.094* 0 0.0% M vs. L 13 100.0% H vs. L 0.60 (0.11,3.15)* 0.689* 0.08 (0.004,1.56)** 0.048* Overall — — — — 0.207* 0.58 (0.18,1.84)* 0.27 (0.06,1.30)* 0.402* 0.108* Overall 2 2.2% M vs. L 90 97.8% H vs. L 22 0.0% 100.0% 0.795 0.024 Overall 92 4.6% 95.4% 1.17 (0.35,3.92) 3.49 (1.17,10.32) Overall 0 0.0% M vs. L 6 100.0% H vs. L 109 0.025 Overall 14 14.1% M vs. L 85 85.9% H vs. L 6 — — 5 104 0.60 (0.07,5.03)* 0.999* 0.08 (0.003,1.77)** 0.092* 13 0 7.6% 92.4% 0 0.0% M vs. L 24 100.0% H vs. L 128 8.7% 91.3% p-Value 0.129* 99 6.2% 93.8% Adj. Relative Contrast Risk (95ZC.I.) Overall 24 113 0 0.0% 5 100.0% 8 97 2 15 — for Psvchnlnoricail feriahles Overall 0.016* 5 22.7% M vs. L 17 77.3% H vs. L — — 9.74 (0.50,190.8)** 0.067* �TABLE J-7. Variable Interaction (Occupation) Stratification Statistic High School Exposure Masculinity/ Index-byFemininity Education (Officer) College Born >=1942 Paranoia Exposure Index-byAge (Officer) Born <L942 Born >1942 Paranoia Exposure Ihdex-byAge (Enlisted Flyer) Born <1942 (continued) Low Exposure Index Medium High Adj. Relative .. Contrast Risk (95% C I ) p-Value n Number/% Abnormal Normal 11 16 1 9.1% 10 9 . % 09 4 12 n Number/% Abnormal Normal 114 n Number/% Abnormal Normal 43 22 0 00 .% 43 1 0 0 0.% 1 21 n Number/% Abnormal Normal n Number/% Abnormal Normal 11 18 0 00 .% 11 1 0 0 0.% 1 17 n Number/% Abnormal Normal 43 47 48 3 7.0% 40 93.0% 0 00 .% 47 1 0 0 0.% 3 6.3% M vs. L 0.12 ( . 1 2 4 ) * 00,.3* 45 93.7% H vs. L 0 8 ( . 7 4 6 ) .9 01,.6* 6 5.3% 108 94.7% 24 25.0% 75.0% 0.150 047 .6 % 110 11 99 1 4.2% M vs. L 5 % ( . 3 6 . 9 . 05,75) 00,.5 23 95.8% H vs. L 0.33 ( . 2 6 5 ) 1.% 00 9.% 00 13 13.5% M vs. L 3.03 ( . 1 9 0 ) 10,.8 83 86.5% H vs. L 4 4 ( . 9 1 . 4 .8 14,34) 11 Overall 008 .4 008 .0 028 .8* 4.5% 95.5% 84 108 02,5.0* 0 0.0% M vs. L 6.07 ( . 4 1 5 3 ) * 0.338* 11 1 0 0 H vs. L 0.% — — 112 Overall 016 .0* 1 1.2% 83 9 . % 88 0 00 .% 108 1 0 0 0.% 4 3.6% M vs. L 0 2 ( . 1 6 3 ) * 0 4 7 . 6 00,.8* .3* 108 96.4% H vs. L 3.07 ( . 4 2 . 2 * 0.394* 03,80) 9 5.6% 9.% 44 Overall 055 .6* 0 0.0% M vs. L 1.97 ( . 7 5 . 1 * 0 9 9 00,27)* . 9 * — 0.% 9 1 0 0 H vs. L — Overall 0.195* 0.105* 099 .9* �———_ V V- X" * — • • « - f Interaction Sunmarles of Adjusted Exposure Index Analyses for Psychological Variables Variable Interaction (Occupation) Stratification Statistic Nonblack Paranoia Exposure Index-byRace (Enlisted Groundcrew) Black High School Psychopathic/ Deviate Exposure Jndex-byEducation (Enlisted Groundcrew) College Low Exposure Index Medium High Adj. Relative 9%.. Contrast Risk ( 5 C I ) Overall 038 .4* 7 5.5% M vs. L 121 94.5% H vs. L 1.56 ( . 7 6 6 ) 0 3 , . 6 * 0.724 .0* 2.60 ( . 6 1 . 0 * 0 2 4 06,03) Overall 042 .2* n Number/% Abnormal Normal 138 149 3 2.2% 135 97.8% 5 144 n Number/% Abnormal Normal 16 14 2 12.3* 14 87.5% 1 13 n . 110 Number/% Abnormal 16 14.5% Normal 94 85.5% 131 . 15 116 11.5% 8.% 85 n Number/% Abnormal Normal 38 3 26 3.4% 9.% 66 13 7.12 92.9% 29 8 21.1% 30 78.9% 128 .% 0 0 0 M vs. L 0 5 ( . 4 6 6 ) . 4 00,.7* 0.% 13 1 0 0 H vs. L 0.22 ( . 1 4 8 ) * 00,.9* 108 10.3% 8.% 97 099 .9* 048 .3* Overall 03,.4 24 22.2% M vs. L 0.75 ( . 5 1 6 ) 84 77.8% H vs. L 1.50 ( . 2 3 0 ) 07,.9 23 p-Value 045 .7 0.278 Overall 1 4.3% M vs. L 0.52 ( . 2 2 2 ) 01,.3 22 95.7% H vs. L 0.12 ( . 1 1 1 ) 00,.1 037 .7 002 .6 �TABLE J-7. Variable Interaction (Occupation) Stratification Statistic Low (continued) Exposure Index Medium High Adj. Relative Contrast Risk ( 5 C I ) p-Value 9% .. () c () c _ 0.228 M vs. L H vs. L () c () c 094 .6 008 .5 M vs. L H vs. L () c () c 007 .0 000 .8 0 Total Off Exposure Index-byDrink-years (Officer Enlisted) n Adj. Mean 95% C I .. 5 0 40 .1 (.7 80) — 17, .6 — M vs. L 6 M vs. L 6.87 (.7 35, (25) 1.3 0-50 n Adj. Mean 95%C.I. 104 6.95 (.0 51, 9.37) 107 6.98 9.39) (5.13; 92 84 .0 (.9 61, 11.29) 15 13.74 (.1 89, 2.2 09) 17 7.17 (.8 46, 1.5 07) 22 9.27 (.4 63, 13.36) >50 n Adj. Mean 95% C.I. �_• . - X-""-— ""t Interaction Sunmaries of Adjusted Exposure Index Analyses for Psychological Variables Variable Interaction (Occupation) Stratification Statistic Low Exposure Index Medium High Adj. Relative Contrast Risk ( 5 C I ) 9% . . p-Value High School 0 n Adj. Mean 95% C.I. 2 4 4.00 17.79 (.11.8 (.83.9 10,14) 86,54) 4 11.32 (.1 53, 2.9 30) M vs. L M vs. L () c () c 0.018 0.106 High School 0-50 'n Adj. Mean 95%C.I. 25 36 27 16.06 10.87 12.68 (12,28) (.71.3 (.2 1.32.1 92,72) 78, 1.7 49) M vs. L M vs. L () c () c 0.191 005 .4 Exposure Ihdex-byEducation High School >50 n Adj. Mean 95% C I .. 14 11 M vs. L 10 16.32 18.27 M vs. L 19.64 (05,48) (23,08) (16, 1.92.7 1.83.5 1.8 28.27) () c () c 059 .0 060 .8 Exposure Index-byDrink-Years (Enlisted Flyer) College 0 n Adj. Mean 95% C.I. 1 0.84 (,.6 056) College 0-50 n Adj. Mean 95% C.I. College >50 n Adj. Mean 95% C.I. Total CMT 1 0 10.17 (.33.6 — 21,88) — 6 8 10 11.92 6.43 11.84 ( . 5 1 . 7 ( . 3 1 . 1 (.6 3 2 , 1 9 ) 7 3 , 8 8 ) 70 19.72) M vs. L M vs. L () c () c 009 .4 — M vs. L M vs. L () c () c 0.100 0.112 1 20.55 (.4 50, 7.4 58) M vs. L M vs. L () c () c — — 0 — — 0 — — �——— — * - x f Interaction Smmarles of Adjusted Exposure Index Analyses for Psychological Variables Variable Interaction (Occupation) Stratification Statistic 0 M-R Subscore Exposure Index-byDrink-years (Enlisted Flyer 05() -0a >50(a) n Number/% X) 0 Low Exposure Index Medium 3 4 1 4 5 .% 0 00 3 100 0.% n Number/% X) 0 31 n Number/% X) 0 14 12 19 8.% 00 20.0% 7 50.0% 7 50.0% 25 21 p-Value 000 .6* 3 7 . % M vs. L 50 1 25.0% H vs. L 21.00 ( . 4 , 9 . ) * 0 1 3 06 6 0 0 * . 4 * 16.33 ( . 8 , 5 . ) * 0.143* 04 5 5 6 * Overall 004 .5 50.0% 50.0% 10 28.6% M vs. L 1.98 ( . 7 5 0 ) 07, . 9 25 71.4% H vs. L 0.65 ( . 3 1.83) 02, 0.158 040 .1 12 54.3% 45.7% 10 5 5 Adj. Relative Contrast Risk ( 5 C. . 9% 1) Overall 35 46 38.7% 61.3% High 040 .4 Overall 9 75.0% M vs. L 1.00 ( . 0 5.07) 02, 3 25.0% H vs. L 2.67 ( . 9 1 . 6 04, 4 4 ) 1.000 0.255 �TABLE J-7. (continued) Variable Interaction (Occupation) Stratification Statistic Low Exposure Index Medium High Adj. Relative Contrast Risk ( 5 C.I.) 9% p-Value 2 4 4 Normal 0 00 . 2 100 0. 4 100 0. 0 00 . 4 1 0 0 M vs. L 4 . 0 ( . 7 3 4 ) * 0 0 7 0. 5 0 06,03* .6* 0 0 0 H vs. L 45.00 ( . 7 3 4 ) * 0 0 7 . 06,03* .6* n Number/% 22 33 22 Overall 005 .3* Abnormal 18 8 . 18 4 18.2 18 15 10 45.5 12 54.5 M vs. L 0 2 ( . 7 0 % * . 7 00,.) Y vs. L 0 1 ( . 5 0 7 ) . 9 00,.3* 007 .4* 007 .2* 14 9 11 Overall 0.583* 10 71.4 4 28.6 8 1 8. 89 11.1 9 81.8 2 18.2 M vs. L 3.20 ( . 0 3 . 9 * 0.611* 03,45) 02,23) .6* H vs. L 1.80 ( . 6 1 . 0 * 0 6 1 n 0 0 .— 0 — 6 1 0 00 . 1 100 0. 1 0 00 . 1 100 0. M vs. L H vs. L 10 7 Overall 1 16.7 5 83.3 6 4 5 71.4 2 28.6 M vs. L 7.50 ( . 2 9 . 5 * 0.145* 06,06) H vs. L 12.50 ( . 4 1 6 3 ) 0.103* 08,8.0* 0 0 1 Overall 0 — 0 — 0 0 1100 0. 0 00 . M vs. L H vs. L High School n 0 Drink-years Nanber/% Ahnprmal High School 0-50 Drinkyears Normal Exposure Index-byEducation ft A-fl Area Subscore Index-byDrink-years (Enlisted Flyer) High School >50 Drinkyears n Number/% Abnormal Normal Exposure College 0 Drink-years Abnormal Normal College 0-50 Drinkyears n Njrober/% Abnormal Normal College >50 Drinkyears n Number/% Abnormal Normal 54.5 45.5 6. 00 4. 00 — — Overall 007 .0* Overall — — — — — 0.115* — — — — — �TftBLE J-7. (continued) Variable Interaction (Occupation) Nonblack HRB Impairment Index Exposure Ihdex-byRace (Enlisted Groundcrew) Low Stratification Statistic Black Exposure Index Medium n Number/% Abnormal Normal 132 n Number/% Abnormal Normal 13 14 11 84.6% 2 15.4% 10 4 48 84 59 85 Adj. Relative 9% . . Contrast Risk ( 5 C I ) p-Value 115 144 36.4% 63.6% High 4.% 10 59.0% 10,.5 40 3 . % M vs. L 1.88 ( . 9 3 2 ) 48 75 6 . % H vs. L 0 7 ( . 3 1 4 ) 52 .7 04,-0 004 .2 040 .0 12 71.4% 28.6% 3 2 . % M vs. L 0 3 ( . 5 2 7 ) 50 .9 00,.2 9 7 . % H vs. L 0 0 ( . 1 0 4 ) 50 .6 00,.4 030 .4 006 .0 ^Unadjusted estimate of relative risk and confidence interval, or p-value, based on stratified tables. **unadjusted estimate of relative risk and confidence interval calculated after adding 0.5 to each cell. (a)Results have been adjusted for education. (b)Results have been adjusted for age. —Zero counts in cells do not allow for calculation of percent, relative risk, confidence interval, or p-value. (c)No relative risk given for total Off, which was analyzed continuously. Note: Snail sample sizes may affect validity of overall p-value. Note: Results without (*) or ( * are adjusted for all other main effects in model (age, race, education, drink-years, and combat index *) [for MMPI variables]), unless otherwise noted. Note: Participants judged to have PTSD are eliminated from analysis of Total .Off, Section M-R Subscore, Section A-H Area Subscore, and HRB Impairment Index. �TABLE J-8. Unadjusted Analyses for Reported Psychological Illnesses by Group: Baseline and First Followup Studies Combined* (Original Comparisons Only) Group Abnormalities Type of Illness Original Ranch Hand Comparison Number Percent Number Percent Total p-Value** 14 1.4 8 0.8 22 0.289 Alcohol Dependence 9 0.9 4 0.4 13 0.268 Anxiety 7 0.7 8 0.8 15 0.798 72 7.1 46 4.8 118 0.037 Psychoses Other Neuroses ^Analyses based on 1,016 Ranch Hands and 955 Original Comparisons; some participants had more than one illness. **Fisher's exact test. J-27 �TABLE J-9. Summary of Kolmogorov-Smirnov Tests on MMPI for Officers (Original Comparisons Only) Parameter* Mean Score Group Standard Deviation KolmogorovSmirnov p-Value Anxiety Ranch Hand Original Comparison 52.82 53.22 7.38 7.06 0.328 Consistency Ranch Hand Original Comparison 49.23 49.02 6.92 5.50 0.995 Defensiveness Ranch Hand Original Comparison 49.64 49.69 6.64 7.14 0.822 Denial Ranch Hand Original Comparison 59.47 59.46 7.31 7.40 0.901 Depression Ranch Hand Original Comparison 55.11 54.86 10.17 9.38 0.975 Hypochondria Ranch Hand Original Comparison 55.60 54.96 8.85 8.09 0.990 Hysteria Ranch Hand Original Comparison 60.45 59.79 7.00 6.81 0.711 Mania/Hypomania Ranch Hand Original Comparison 54.36 53.68 8.94 9.10 0.436 Masculinity/ Femininity Ranch Hand Original Comparison 58.83 58,14 8.64 8.94 0.577 Paranoia Ranch Hand Original Comparison 53.71 53.57 7.34 6.86 0.995 Psychopathic/ Deviate Ranch Hand Original Comparison 56.82 56.61 8.87 9.65 0.644 Schizophrenia Ranch Hand Original Comparison 54.80 54.66 8.48 7.89 0.661 Social Introversion Ranch Hand Original Comparison 46.56 47.13 7.43 7.60 0.627 Validity Ranch Hand Original Comparison 0.93 0.80 3.11 2.86 0.594 *n=380 Ranch Hands; n=350 Original Comparisons. J-28 �TABLE J-10. Summary of Kolmogorov-Smirnov Tests on MMPI for Enlisted Flyers (Original Comparisons Only) Group Mean Score Standard Deviation KolmogorovSmirnov p-Value Anxiety Ranch Hand Original Comparison 54.78 54.16 9.73 8.78 0.993 Consistency Ranch Hand Original Comparison 52.24 51.41 10.30 8.15 0.923 Defensiveness Ranch Hand Original Comparison 53.05 52.09 7.96 8.34 0.353 Denial Ranch Hand Original Comparison 55.95 56.81 8.73 8.44 0.760 Depression Ranch Hand Original Comparison 58.21 57.27 11.79 10.35 0.760 Hypochondria Ranch Hand Original Comparison 58-48 58.44 12.16 11.71 0.577 Hysteria Ranch Hand Original Comparison 60.00 60.70 8.94 8.77 0.465 Mania/Hypomania Ranch Hand Original Comparison 55.05 55.50 10.00 9.42 0.155 Masculinity/ Femininity Ranch Hand Original Comparison 55.79 56.22 7.85 8.39 0.877 Paranoia Ranch Hand Original Comparison 52.77 51.72 8.43 7.68 0.877 Psychopathic/ Deviate Ranch Hand Original Comparison 57.48 57.85 11.40 9.72 0.163 Schizophrenia Ranch Hand Original Comparison 56.56 56.27 13.03 10.59 0.496 Social Introversion Ranch Hand Original Comparison 50.35 49.00 8.89 7.94 0.194 Validity Ranch Hand Original Comparison 3.93 7.18 42.48 60.31 0.864 Parameter* *n=176 Ranch Hands; n=172 Original Comparisons (except for validity, where n=177 Ranch Hands and n^!74 Original Comparisons). J-29 �TABLE J-ll. Summary of Kolmogorov-Smirnov Tests on MMPI for Enlisted Groundcrew (Original Comparisons Only) Group Mean Score Standard Deviation KolmogorovSmirnov p-Value Anxiety Ranch Hand Original Comparison 55.86 55.72 10.96 9.34 0.877 Consistency Ranch Hand Original Comparison 53.05 51.78 10.64 8.19 0.728 Defensiveness Ranch Hand Original Comparison 52.42 52.69 8.27 8.51 0.923 Denial Ranch Hand Original Comparison 55.53 56.80 8.57 8.72 0.112 Depression Ranch Hand Original Comparison 58.79 57.85 12.98 10.63 0.661 Hypochondria Ranch Hand Original Comparison 58.44 57.50 13.22 11.30 0.577 Hysteria Ranch Hand Original Comparison 60.36 59.75 9.80 8.84 0.644 Mania/Hypomania Ranch Hand Original Comparison 55.26 55.69 9.78 9.72 0.822 Masculinity/ Femininity Ranch Hand Original Comparison 56.18 57.24 8.24 8.49 0.304 Paranoia Ranch Hand Original Comparison 53.12 53.73 9.05 8.48 0.340 Psychopathic/ Deviate Ranch Hand Original Comparison 59.01 59.34 11.29 10.70 0.644 Schizophrenia Ranch Hand Original Comparison 58.20 57.58 14.22 10.64 0.328 Social Introversion Ranch Hand Original Comparison 51.86 50.99 9.82 8.63 0.353 Validity Ranch Hand Original Comparison 1.97 0.97 26.48 5.41 0.999 Parameter* *n=458 Ranch Hands; n=430 Original Comparisons (except for validity, where n=459 Ranch Hands and n=431 Original Comparisons). J-30 �TABLE J-12. Summary of Rolmogorov-Smirnov Tests on CHI (Original Comparisons Only) KolmogorovSmirnov Standard Deviation p-Value Sample Size Mean Score Ranch Hand Original Comparison 402 366 15.88 12.95 15.21 11.57 <0.001 Born 1923-1941 Ranch Hand Original Comparison 562 528 15.49 14.55 12.99 11.88 0.776 Ranch Hand Original Comparison 36 45 18.56 16.47 15.94 9.28 0.980 Ranch Hand Original Comparison 941 882 15.63 13.96 13.83 11.47 0.038 Ranch Hand Original Comparison 59 57 17.81 14.82 16.91 14.62 0.561 Ranch Hand Original Comparison 556 524 18.88 15.62 15.80 13.12 <0.001 Ranch Hand Original Comparison 444 415 11.85 11.99 10.18 9.16 0.577 Current Alcohol Use Drinker Ranch Hand Original Comparison 851 812 15.16 13.77 13.12 11.32 0.170 Ranch Hand Original Comparison 148 127 18.84 ' 15.61 17.71 13.69 0.162 Ranch Hand Original Comparison 377 345 11.23 10.99 9.01 7.53 0.970 Enlisted Flyer Ranch Hand Original Comparison 174 170 17.72 15.31 14,84 12.51 0.270 Enlisted Groundcrew 449 424 18.80 15.96 16.06 13.48 0.003 Stratification Age Born XL942 Born <1922 Race Nonblack Black Education High School College Nondr inker Occupation Officer Group Ranch Hand Original Comparison J-31 �TABLE J-13. Unadjusted Analyses for HMFI by Group (Original Comparisons Only) Group Variable Anxiety Consistency Defensiveness i Ranch Hand Statistic Number Percent Original Comparison Number Percent Est. Relative Risk ( 5 C.I.) 9% p-Value n Abnormal Normal 1,014 73 941 7.2 92.8 952 49 903 5.1 94.9 1.43 (0.98,2.08) 0.062 n Abnormal Normal 1,014 36 978 3.6 96.4 952 22 930 2.3 97.7 1.56 (0.91,2.67) 0.111 n Abnormal Normal 1,014 23 991 2.3 97.7 952 30 922 3.2 96.8 0.71 (0.41,1.24) 0.265 n Abnormal Normal 1,014 17 997 1.7 98.3 952 37 915 3.9 96.1 0.42 (0.24,0.75) 0.003 n Abnormal Normal 1,014 114 900 11.2 88.8 952 87 865 9.1 90.9 1.26 (0.94,1.69) 0.136 n Abnormal Normal 1,014 119 895 11.7 88.3 952 91 861 9.6 90.4 1.26 (0.94,1.68) 0.125 n Abnormal Normal 1,014 123 891 12.1 87.9 952 89 863 9.4 90.6 1.34 (1.00,1.79) 0.050 ro Denial Depression Hypochondria Hysteria �TABLE J-13. (continued) Unadjusted Analyses for MMFI by Group (Original Comparisons Only) Group Variable Original Ranch Hand Comparison Statistic Number Percent Number Percent Est. Relative Risk (95% C.I.) p-Value Mania/Hypomania n Abnormal Normal 1,014 63 951 6.2 93.8 952 56 896 5.9 94.1 1.06 (0.73,1.54) 0.777 Masculinity/ Femininity n Abnormal Normal 1,014 66 948 6.5 93.5 952 83 869 8.7 91.3 0.73 (0.52,1.02) 0.073 Paranoia n Abnormal Normal 1,014 31 983 3.1 96.9 952 17 935 1.8 98.2 1.73 (0.95,3.16) 0.079 Psychopathic/ Deviate n Abnormal Normal 1,014 120 894 11.8 88.2 952 104 848 10.9 89.1 1.09 (0.83,1.45) 0.570 Schizophrenia n Abnormal Normal 1,014 94 920 9.3 90.7 952 71 881 7.5 92.5 1.27 (0.92,1.75) 0.166 Social Introversion n Abnormal Normal 1,014 26 988 2.6 97.4 952 14 938 1.5 98.5 1.76 ( . 2 3 4 ) 09,.0 0.109 Validity n >0 0 1,016 224 792 22.0 78.0 955 198 757 20.7 79.3 1.08 (0.87,1.34) 0.510 CH I �TABLE J-14. Adjusted Analyses for MMPI by Group (Original Comparisons Only) Group Variable Ranch Hand Total Original Comparison Total 1,012 950 Consistency 976 926 Defensiveness 976 926 Denial 974 Depression Hypochondria Anxiety Hysteria Mania/Hypomania Adj. Relative Risk (95% C.I.) 1.39 (0.95,2.05) 0.092 Covariate Remarks* EDUC (p<0.001) AGE*CI (p=0.010) **** AGE (p=0.025) RACE*DRKYR (p=0.024) GRP*EDUC (p=0.034) 0.68 (0.38,1.19) 0.175 EDUC (p<0.001) AGE*DRKYR (p=0.043) 924 0.43 (0.23,0.80) 006 .0 AGE*DRKYR (p=0.026) EDUC*CI (p=0.040) 1,014 952 1.26 ( . 4 1.70) 09, 0.120 EDUC (p<0.001) 1,014 . **** p-Value 952 1.29 (0.96,1.72) 0.091 AGE (p=0.005) EDUC (p<0.001) 1,014 952 1.37 (1.02,1.83) 0.033 AGE (p=0.001) EDUC (p<0.001) 974 924 0.97 (0.66,1.43) 0.882 AGE (p=0.048) CI (p=0.010) DRKYR (p=0.016) �TABLE J-14. (continued) Adjusted Analyses for MMFI by Group (Original Comparisons Only) Group Variable Ranch Hand Total Original Comparison Total Adj. Relative Risk ( 5 C.I.) 9% p-Value Covariate Remarks* Masculinity/ Femininity 952 Paranoia (-1 I 1,014 1,012 950 Psychopathic/ Deviate 974 924 1.13 (0.84,1.52) 0.433 EDUC (p=0.010) AGE*CI (pXLOOl) RACE*DRKYR (p=0.001) Schizophrenia 976 926 1.27 (0.91,1.77) 0.153 EDUC (p<0.001) RACE*DRKYR (p=0.003) Social Introversion 1,014 952 1.72 (0.89,3.31) 0.099 Validity 1,014 953 0.71 (0.51,1.00) **** 0.047 **** Ul **** **** EDUC (p<0.001) RACE* AGE (p=0.006) AGE*CI (p=0.004) GRP*AGE (p=0.040) AGE (p=0.005) AGE*CI (p=0.034) GRP*RACE (p=0.016) GRP*CI (p=0.030) *Abbreviations : EDUC: education CI: combat index DRKYR: drink-years GRP: group ****Group-by-covariate interaction: presented. adjusted relative risk, confidence interval, and p-value are not �TABLE J-15. Unadjusted Analyses for the Cornell Medical Index (CHI) by Group (Original Comparisons Only) Group Variable Statistic Ranch Hand Total CMI n Mean* 95% C.I.* 1,000 11.74 (11.17,12.35) M-R Subscore Subscore C«i i n Number/% 0 (Low) 1-10 (Medium) (High) 998 538 408 52 Original Comparison 939 10.55 (10.02,11.11) 531 370 35 n Number/% 0 (Low) 1-3 (Medium) 4-8 (High) 1,148 914 360 449 105 39 .4% 49 .1% 11 .5% 375 394 82 0.004 44 .1% 46 .3% 9.6% 0.198 Medium vs. Low 1.09 (0.91,1.31) High vs. Low 1.47 (0.99,2.29) 0.371 Overall 56 .7% 39 .5% 3.8% u> A-H Area Subscore — p-Value Overall 936 53 .9% 40 .9% 5.2% Est. Relative Risk (95% C.I.) 0.106 Medium vs. Low 1.19 (0.97,1.45) High vs. Low 1.33 (0.97,1.84) 0.090 0.095 0.086 transformed from log (X+l) scale, where X was the number of questions answered "yes." —No relative risk given for Total CMI, which was analyzed as a continuous variable. �TABLE J-16. Adjusted Analyses for CMI Variables by Group (Original Comparisons Only) Variable Statistic Total CMI LO -vl n Adj. Mean 95% C.I. M-R Subscore A-H Area Subscore Ranch Hand Group Original Comparison 962 **** **** 998 935 n 914 850 ****Group-by-covariate not presented. p-Value Covariate Remarks* **** 913 **** **** n Adj. Relative Risk (95% C.I.) **** Overall Medium vs. Low: 1.20 ( . 9 1 4 ) 09,.7 High vs. Low: 1.32 (0.92,1.89) PTSD (p<0.001) RACE*DRKYR (p=0.027) AGE*EDUC (p=0.010) GRP*EDUC (p=0.002) **** AGE (p<0.001) PTSD (p<0.001) GRP*EDUC (p=0.006) 0.108 0.063 0.132 AGE (p<0.001) EDUC (p<0.001) PTSD (p<0.001) interaction — adjusted mean/relative risk, confidence interval, and p-value are —No relative risk given for total CMI, which was analyzed as a continuous variable. *Additional Abbreviations: PTSD: post-traumatic stress disorder �TABLE J-17. Summary Results for the Halstead-Reitan Battery Impairment Index Unadjusted and Adjusted Analyses (Original Comparisons Only) Group Analysis Unadjusted Analysis 00 n Abnormal Normal 1,006 348 658 Adjusted Analysis u> Ranch Hand Statistic Number Percent n 1,006 34.6 65.4 Original Comparison Number Percent 947 319 628 947 33.7 66.3 Est. Relative Risk (95% C.I). p-Value 1.04 (0.86,1.26) 0.703 1.10 (0.90,1.34) 0.359 Covariate Remarks AGE (p<0.001) RACE (p<0.001) EDUC (p<0.001) �TfiBlE J-18. for Psychological Variables (Original Comparisons Only) Group Variable Consistency Interaction Group-byEducation Statistic High School n Number/% Abnormal Normal 537 n Number/% Abnormal Normal 437 n Number/% Abnormal Normal 410 n Number/% Abnormal Normal 602 College V ^> Paranoia Group-byAge Ranch Hand Stratification Bom XL942 Born <1942 30 507 6 431 15 395 16 586 Original Comparison Adj. Relative Risk ( 5 C I ) 9% . . p-Value 2.5% 97.5% 2.31 ( . 8 4 5 ) 11,.0 0.014 2.6% 97.4% 0 6 (.217) . 1 02,.4 0.359 3.5% 96.5% 0 % (.421) . 04,.3 0.927 1.2% 98.8% 3 8 (.61.4 . 5 13,09) 0.011 516 5.6% 94.4% 13 503 585 1.4% 98.6% 15 570 545 3.7% 96.3% 19 526 740 2.7% 97.3% 9 731 �TABLE J-18. (continued) Summary of Group-by-Cbvariate Interactions (Original Comparisons Only) Group Variable Interaction Validity Group-byRace Ranch Hand Stratification Statistic Nonblack CX=Low n Nunfcer/% X) 0 178 n Number/% X) 0 367 n Number/% X) 0 409 n Number/% X) 0 20 n Number/% X) 0 18 n Number/% X) 0 22 Nonblack CkMediun Nonblack CI=Bigh Group-by Combat Index Black CT=low Black Q=Medium Black d=fligii 47 131 78 289 83 326 6 14 6 12 3 19 Original Conparison Adj. Relative Risk (95% C.I.) p-Value 20.7% 79.3% 1.49 (0.98,2.26) 0.059 14.6% 85.4% 1.60 (1.07,2.38) 0.021 23.1% 76.9% 0.83 (0.57,1.20) 0.319 37.5% 62.5% 0.55 (0.24,1.29) 0.170 41.7% 58.3% 0.59 (0.25,1.41) 0.239 50.0% 50.0% 0.31 (0.13,0.71) 0.006 353 26.4% 73.6% 73 280 295 21.3% 78.7% 43 252 247 20.3% 79.7% 57 190 24 30.0% 70.0% 9 15 12 33.3% 66.7% 5 7 22 13.6% 86.4% 11 11 �TABLE J-38. (ccntuued) for Psychological Variables (Original CoDfEorisons Only) Group Interaction Total Off Group-byEducation M-R Subscore Group-byEducation Stratification Statistic Ranch Hand Original Comparison High School n Adj. Mean* 95% C I * .. 529 35.50 (80,47) 2.94.9 511 29.65 (32,77) 2.73.0 — <0.001 College Variable n Adj. Mean* 95% C I * .. 433 22.70 (77,89) 1.62.5 402 23.49 (83,99) 1.72.7 — 048 .9 High School n 555 Number/% 0 (Low) 246 1-10 (Med) 267 >10 (High) 42 0.015 281 213 27 53.9% 40.9% 5.2% Overall Med vs. Low 1.42 ( . 1 1 8 ) 11,.2 High vs. Low 1.51 ( . 4 2 7 ) 08,.0 414 Overall Med vs. Low 60.4% 0.77 ( . 8 1 0 ) 05,.2 37.7% High vs. Low 1.9% 0.99 ( . 4 2 8 ) 03,.7 01 .% College 443 n Nunber/% 0 (Low) 292 1-10 (Med) 141 >10 (High) 10 521 44.3% 48.1% 7.6% 65.9% 31.8% 2.3% 250 156 8 ^Converted from log (X+l) scale, where X was the number of questions answered "yes". —No relative risk given for Total Off, which was analyzed as a continuous variable. Adj. Relative Risk ( 5 C I ) 9% .. p-Value 0.005 0.167 0.071 0.991 �APPENDIX K Gastrointestinal Assessment �APPENDIX K: Gastrointestinal Assessment Contents Table K-l Page Summary of Group-by-Covariate Interactions for Hepatic Function Variables and Porphyrin Determinations Analyzed Continuously K-l K-2 Summary of Group-by-Covariate Interactions for Hepatic Function Variables and Porphyrin Determinations Analyzed Categorically K-4 K-3 Unadjusted Categorical Exposure Index Analyses for Hepatic Function Variables by Occupation K-5 K-4 Unadjusted Continuous Exposure Index Analyses for Hepatic Function Variables and Tvo Porphyrin Determinations by Occupation K-8 K-5 Summary of Adjusted Exposure Index, Analyses Involving Interactions with a Continuous Covariate for Hepatic Function Variables K-13 K-6 Summary of Adjusted Exposure Index, Analyses Involving Interactions with a Categorical Covariate for Hepatic Function Variables K-15 K-7 Unadjusted Analyses for Baseline and Interval History of Liver Disease by Group (Verified by Medical Record Review) (Original Comparisons Only) K-16 K-8 Unadjusted Analyses of Enlarged Livers Diagnosed at Physical Examination by Group (Original Comparisons Only) K-17 K-9 Unadjusted Continuous and Categorical Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Group (Original Comparisons Only) K-18 K-10 Adjusted Continuous and Categorical Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Group (Original Comparisons Only) K-21 K-ll Summary of Group-by-Covariate Interactions for Hepatic Function Variables (Original Comparisons Only) K-27 K-12 Summary of Group-by-Covariate Interactions for Variables and Porphyrin Determinations Analyzed Categorically (Original Comparisons Only) K-29 K-i �APPENDIX K: Gastrointestinal Assessment (continued) Contents Table Page K-13 Unadjusted Analysis for Interval History of Skin Bruises, Skin Patches, or Skin Sensitivity by Group (Original Comparisons Only) K-30 K-14 Unadjusted Analyses for Porphyrin Abnormalities by Group and Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire (Original Comparisons Only) K-31 K-15 Unadjusted Analyses for Uroporphyrin Abnormalities by Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire and by Group (Original Comparisons Only) K-32 K-16 Longitudinal Results for SCOT, SGPT, and GGTP: A Contrast of Baseline and First Followup Examination Test Means (Original Comparisons Only) K-ii K-33 �TABLE K-l. v Summary of Group-by-Covariate Interactions for Hepatic Function Variables and Porphyrin Determinations Analyzed Continuously Group Ranch Hand Comparison n Mean Adj . Mean 95% C.I. 652 32.2 33.4 (32.4,34.4) 843 32.3 33.4 (32.5,34.4) 1-4 n Mean Adj . Mean 95% C.I. 275 35.3 36.5 (35.1,38.0) 335 33.4 34.4 (33.2,35.7) 0.010 >4 Mean Adj . Mean 95% C.I. 76 37.5 38.2 (35.7,40.9) 108 38.1 39.4 (37.2,41.8) 0.477 <1 Group-bycur rent Alcohol Use (Drinks/day) Statistic n SCOT Interaction Stratification <1 Variable n Mean Adj . Mean 95% C.I. 652 20.5 19.8 (18.6,21.2) 843 21.9 21.1 (19.8,22.5) 0.011 1-4 n Mean Adj . Mean 95% C.I. 275 23.4 22.2 (20.2,24.3) 335 22.6 21.5 (19.7,23.5) >4 n Mean Adj . Mean 95% C.I. 76 23.9 23.3 (19.7,27.5) 108 27.0 26.7 (22.8,31.3) • SGPT Group-bycur rent Alcohol Use (Drinks/day) p-Value 0.889 0.428 0.058 �TABLE R-l. (continued) Summary of Group-by-Covariate Interactions for Hepatic Function Variables and Porphyrin Determinations Analyzed Continuously Group Statistic Ranch Hand n Mean Ad j . Mean 95% C.I. 504 681 94.8 89.9 102.4 97.2 <0.001 (97.9,107.2) (93.1,101.5) Not Exposed to Industrial Chemicals n Mean Adj . Mean 95% C.I. 499 88.9 88.5 (84.9,92.1) 604 88.7 88.5 (85.0,92.1) Born XL942 Interaction Stratification Exposed to Industrial Chemicals Variable n Mean Adj . Mean 95% C.I. 408 122.2 124.8 (121.9,127.7) 547 121.2 0.425 123.6 (121.0,126.3) Born 1923-1941 n Mean Adj . Mean 95% C.I. 565 123.8 127.0 (124.5,129.7) 691 125.3 0.217 128.6 (126.1,131.1) Born <1922 n Mean Adj . Mean 95% C.I. 36 135.1 138.9 (130.8,147.5) 51 135.5 139.1 0.966 (132.2,146.5) Comparison p-Value Alkaline Group-byPhosphatase Industrial Chemicals 0.973 to LDH Group-by-Age �TABLE K.-1. (continued) Summary of Group-by-Covariate Interactions for Hepatic Function Variables and Porphyrin Determinations Analyzed Continuously Group Variable Ranch Hand Comparison n Mean Ad j . Mean 95% C.I. 406 107.1 95.6 (88.0,103.9) 546 110.2 99.7 0.321 (92.5,107.4) Born 1923-1941 n Mean Adj . Mean 95% C.I. 689 561 126.0 124.5 0.784 116.8 115.6 (108.6,125.5) (108.0,123.8) n Mean Adj. Mean 95% C.I. 51 36 139.0 105.6 130.7 . 98.3 0.039 (105.4,161.9) (81.9,118.0) BUN <L4 Uroporphyrins Statistic Born <1922 Group-by-Age Stratification Born >1942 Triglycerides Interaction n Mean Adj . Mean 95% C.I. 547 16.2 16.3 (15.4,17.4) 701 18.4 18.6 (17.6,19.7) <0.001 BUN >14 n Mean Adj . Mean 95% C.I. 453 17.8 17.9 (16.8,19.2) 582 17.4 17.6 (16.6,18.7) 0.686 p-Value Group-byBlood Urea Nitrogen (BUN) �TABLE K-2. Sunnary of Group-by-Cbvariate Interactions for Hepatic Function Variables Group Ranch Hand Variable Interaction Stratification Exposed to Industrial Chemicals Direct Bilirubin Statistic n High Normal Number 507 27 480 Percent 5.3 94.7 Comparison Number 683 20 663 Percent Adj. Relative Risk(95%C.I.) p-Value 2.9 97.1 1.89 ( . 5 3 4 ) 0 0 5 10,.2 .3 * Group-byIndustrial Chemicals 4 n High Normal 4 % 11 485 2.2 97.8 603 23 580 3.8 96.2 0.58 ( . 8 1 2 ) 0 1 3 02,.0 . 4 Officer n High Normal 376 34 342 90 . 91.0 484 26 458 5.4 94.6 1.77 ( . 4 3 0 ) 0 0 5 10,.1 . 3 Enlisted flyer n High Normal 177 15 162 8.5 91.5 209 14 195 6.7 93.3 1.27 ( . 9 2 7 ) 0 5 9 05,.1 .3 Enlisted Groundcrev n High Normal 456 19 437 4.2 95.8 5% 39 557 6.5 93.5 0.63 ( . 6 1 1 ) 0 1 4 03,.0 .0 Not Exposed to Industrial Chemicals Triglycerides Group-byOccupation �TABLE K-3. unadjusted Categorical Exposure Index Analyses for Hepatic Function Variables by Occupation Exposure Index Low Variable Medium High Contrast Est. Relative Risk ( 5 C I ) p-Value 9% . . 7.5 92.5 Overall M vs. L H vs. L 059 .1 1.57 ( . 5 3 7 ) 0.384 06,.7 1.05 ( . 0 2 7 ) 0 9 9 .9 04,.5 57 4 53 7.0 93.0 0.234 Overall M vs. L 5.49 ( . 4 4 . 9 0.123 06,70) H vs. L 4 0 ( . 4 3 . 7 0.364 .8 04,76) 8.1 91.9 142 17 125 12.0 8. 80 005 .6 Overall .4 M vs. L 1.86 ( . 2 4 7 ) 0 2 9 07,.9 H vs. L 2.86 ( . 5 7 1 ) 0.031 11,.1 130 20 110 15.4 8. 46 120 16 104 13.3 86.7 069 .6 Overall M vs. L 0 8 ( . 4 1 6 ) 0.736 .6 04,.7 H vs. L 0 7 ( . 6 1 4 ) 0 3 4 .8 . 3 03,.6 3.6 96.4 65 10 55 15.4 8. 46 57 6 51 10.5 89.5 0.105 Overall M vs. L 4.82 ( . 1 2 . 3 0.037 10,30) H vs. L 3.12 ( . 0 1 . 7 0.272 06,61) 11.0 8. 90 160 27 133 16.9 83.1 142 17 125 12.0 88.0 Overall M vs. L H vs. L 126 9 117 7.1 92.9 130 14 116 10.8 89.2 120 9 111 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 6 59 9.2 90.8 n Abnormal Normal 154 7 147 4.5 95.5 160 13 147 Officer SGPT n Abnormal Normal Enlisted Groundcrev E Statistic Officer SOOT Occupation Number Percent Number Percent Number Percent n Abnormal Normal 126 22 104 17.5 82.5 Enlisted Flyer n Abnormal Normal 55 2 53 Enlisted Groundcrew n Abnormal Normal 154 17 137 025 .6 1.64 ( . 5 3 1 ) 0.147 08,.4 1.10 ( . 4 2 2 ) 0 8 6 05,.4 .5 �TABIE K-3. (continued) unadjusted Categorical Exposure Index Analyses for Hepatic Function Variables by Occupation Exposure Index Low Medium n Abnormal Normal 126 11 115 8.7 91.3 130 9 121 6.9 93.1 120 9 111 7.5 92.5 089 .5 Overall M vs. L 0 7 ( . 1 1 9 ) 0 6 6 .8 03,.5 .4 H vs. L 0.73 ( . 6 1 4 ) 0.384 03,.6 Enlisted Flyer n Abnormal Normal 55 5 50 9.1 9. 09 65 7 58 1. 08 89.2 57 6 51 1. 05 89.5 Overall 0.950 M vs. L 1.21 ( . 6 4 0 ) 0 9 9 03,.4 .9 H vs. L 1.18 ( . 4 4 1 ) 0 9 9 03,.0 .9 n Abnormal Normal 154 16 138 1. 04 89.6 160 13 147 8.1 91.9 142 14 128 99 . 90.1 Overall 0.773 M vs. L 0.76 ( . 5 1 6 ) 0.561 03,.4 H vs. L 0 9 ( . 4 2 0 ) 0 9 9 . 4 04,.1 .9 Officer Alkaline Phosphatase Statistic Enlisted Groundcrew GGEP Number Percent Number Percent Est. Relative Risk ( 5 C I ) p-Value 9%.. Occupation Officer Variable Number Percent High n Abnormal Normal 126 4 122 3.2 96.8 130 7 123 5.4 94.6 120 3 117 2.5 97.5 Overall 048 .4 M vs. L 1.74 ( . 0 6 0 ) 0.540 05,.8 H vs. L 0 7 ( . 7 3 5 ) 0 9 9 .8 01,.7 .9 Enlisted Flyer n Abnormal Normal 55 1 54 1.8 98.2 65 5 60 7.7 92.3 57 5 52 88 . 91.2 0.258 Overall M vs. L 4.50 ( . 1 3 . 4 0.217 05,97) H vs. L 5.19 ( . 9 4 . ) 0 2 6 05,5% .0 Enlisted Groundcrew n Abnormal Normal 154 8 146 5.2 9. 48 160 10 150 6.3 93.8 142 13 129 9.2 9. 08 0.378 Overall M vs. L 1.22 ( . 7 3 1 ) 0.810 04,.7 H vs. L 1.84 ( . 4 4 5 ) 0.257 07,.8 Contrast �T&BLEK-3. (continued) 4 Unadjusted Categorical Exposure Index Analyses for Hepatic Ruction \fariables by Occupation Exposure Index Medium Low n Abnormal Normal 126 3 123 2.4 97.6 130 2 128 1.5 98.5 120 3 117 2.5 97.5 Overall 086 .4 M vs. L 0 6 ( . 1 3 9 ) 0 6 0 . 4 01,.0 .8 H vs. L 1.05 ( . 1 5 3 ) 0 9 9 02,.1 .9 Enlisted Flyer n Abnormal Normal 55 0 55 00 . 100.0 65 1 64 1.5 9. 85 57 2 55 3.5 96.5 Overall 0.353 M vs. L 21.58 ( . 0 6 . 5 0 9 9 01,46) .9 H vs. L 5.00 ( . 4 1 6 5 ) 0 4 02,0.4 .% n Abnormal Normal 154 7 147 4.5 95.5 160 3 157 1.9 98.1 142 6 136 4.2 95.8 Overall M vs. L H vs. L Officer Direct Bilirubin Statistic Enlisted Groundcrew Total Bilirubin Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9%.. Occupation Officer Variable Number Percent High n Abnormal Normal 126 2 124 1.6 98.4 130 5 125 3.8 96.2 120 6 114 5.0 95.0 Overall 0.327 04,30) .4 M vs. L 2.48 ( . 7 1 . 2 0 4 7 H vs. L 3.26 ( . 5 1 . 0 0.164 06,65) Enlisted Flyer n Abnormal Normal 55 2 53 3.6 9. 64 65 5 60 7.7 92.3 57 3 54 5.3 94.7 Overall 0.624 M vs. L 2.21 ( . 1 1 . 6 0 4 1 04,18) .5 H vs. L 1.47 ( . 4 9 1 ) 0 9 9 02,.7 .9 Enlisted Groundcrew n Abnormal Normal 154 4 150 2.6 97.4 160 6 154 3.8 96.3 142 5 137 3.5 96.5 Overall 0.834 M vs. L 1.46 ( . 0 5 2 ) 0 7 0 04,.8 .5 H vs. L 1.37 ( . 6 5 2 ) 0 7 2 .4 03,.0 034 .7 0 4 ( . 0 1 5 ) 0.211 .0 01,.8 0.93 ( . 0 2 8 ) 0 9 9 03,.3 .9 �TABLE K-4. Unadjusted Continuous Exposure Index Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Occupation Statistic Low Officer n Mean 95% C.I. 126 33.3 (31.5,35.3) Enlisted Flyer n Mean 95% C.I. Enlisted Groundcrew Variable High Contrast p-Value 130 34.5 (32.6,36.5) 120 33.8 (31.9,35.8) Overall M vs. L H vs. L 0.695 0.400 0.757 55 30.9 (28.5,33.4) 65 33.3 (31.0,35.8) 57 33.2 (30.7,35.8) Overall M vs. L H vs. L 0.289 0.151 0.192 n Mean 95% C.I. 154 32.9 (31.4,34.5) 160 33.2 (31.7,34.8) 142 34.4 (32.8,36.1) Overall M vs. L H vs. L 0.418 0.801 0.211 Officer SCOT Medium n Mean 95% C.I. 126 22.5 (20.5,24.6) 130 22.1 (20.2,24.2) 120 21.3 (19.4,23.39) Overall M vs. L H vs. L 0.712 0.804 0.421 Enlisted Flyer n Mean 95% C.I. 55 18.5 (16.4,20.8) 65 21.9 (19.7,24.5) 57 22.0 (19.6,24.7) Overall M vs. L H vs. L 0.061 0.037 0.041 Enlisted Groundcrew n Mean 95% C.I. 154 20.9 (19.3,22.6) 160 22.3 (20.7,24.1) 142 21.5 (19.9,23.4) Overall M vs. L H vs. L 0.495 0.237 0.595 00 SGPT �TABLE K-4. (continued) Unadjusted Continuous Exposure Index Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Occupation Statistic Low Officer n Mean 95% C.I. 126 31.4 (27.8,35.5) Enlisted Flyer n Mean 95% C.I. Enlisted Groundcrew GGTP Alkaline Phosphatase High Contrast p-Value 130 32.7 (29.0,36.9) 120 32.6 (28.7,36.9) Overall M vs. L H vs. L 0.885 0.653 0.688 55 29.7 (24.6,35.9) 65 34.9 (29.3,41.5) 57 37.1 (30.8,44.7) Overall M vs. L H vs. L 0.243 0.225 0.104 n Mean 95% C.I. 154 34.7 (31.1,38.7) 160 33.0 (29.6,36.7) 142 31.1 (27.8,34.9) Overall M vs. L H vs. L 0.412 0.527 0.183 Officer Variable Medium n Mean 95% C.I. 126 87.2 (83.5,91.1) 130 86.8 (83.2,90.6) 120 88.9 (85.0,92.9) Overall M vs. L H vs. L 0.735 0.893 0.546 Enlisted Flyer n Mean 95% C.I. 55 88.9 (83.1,95.0) 65 96.3 (90.5,102.4) 57 95.2 (89.1,101.6) Overall M vs. L H vs. L 0.189 0.086 0.153 Enlisted Groundcrew n Mean 95% C.I. 154 93.2 (89.7,96.9) 160 95.1 (91.5,98.8) 142 96.2 (92.4,100.2) Overall M vs. L H vs. L 0.535 0.484 0.270 �TABLE K-4. (continued) Unadjusted Continuous Exposure Index Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Occupation Statistic Low Officer n Mean 95% C.I. 126 0.76 (.208) 07,.0 Enlisted Flyer n Mean 95% C.I. Enlisted Groundcrew Total Bilirubin High . Contrast p-Value 130 0.73 (.907) 06,.7 120 0.77 (.308) 07,.1 Overall M vs. L H vs. L 0.382 0.324 0.722 55 0.66 (0.61,0.72) 65 0.73 (0.67,0.78) 57 0.76 (0.70,0.83) Overall M vs. L H vs. L 0.045 0.092 0.014 n Mean 95% C.I. 126 0.74 (.007) 07,.8 130 0.75 (.107) 07,.8 120 0.78 (.408) 07,.2 Overall M vs. L H vs. L 0.373 0.783 0.182 Officer Variable n Mean 95% C.I. 126 0.18 (0.17,0.21) 130 0.17 (0.16,0.19) 120 0.19 (0.17,0.21) Overall M vs. L H vs. L 0.606 0.484 0.784 Enlisted Flyer n Mean 95% C.I. 55 0.16 (.402) 01,.0 65 0.18 (0.15,0.21) 57 0.18 (0.15,0.21) Overall M vs. L H vs. L 0.684 0.455 0.440 Enlisted Groundcrew n Mean 95% C.I. 126 0.17 (0.15,0.19) 130 0.15 (0.17,0.20) 120 0.19 (0.16,0.19) Overall M vs. L H vs. L 0.541 0.277 0.728 i o Direct Bilirubin Medium �TABLE K~4. (continued) Unadjusted Continuous Exposure Index Analyses for Hepatic Function Variables and Two Porphyrin Determinations by Occupation Exposure Index Occupation Statistic Low Officer Variable n Mean 95% C.I. Enlisted Flyer High p-Value 126 130 120 125.1 121.8 124.2 (121.2,129.1) ( 2 . , 2 . ) (117.9,125.8) 104181 Overall M vs. L H vs. L 0.489 0.742 0.246 n Mean 95% C.I. 55 65 57 120.2 126.2 117.6 (114.4,126.4) (112.4,123.3) (120.2,132.5) Overall M vs. L H vs. L 0.118 0.533 0.175 n Mean 95% C.I. 154 160 142 123.1 127.9 122.3 (120.1,126.2) (119.4,125.3) (124.6,131.2) Overall M vs. L H vs. L 0.031 0.710 0.037 Officer Cholesterol Contrast Enlisted Groundcrew LDH Medium n Mean 95% C.I. 126 130 120 220.7 210.7 211.8 (213.3,228.3) (204.8,219.0) (203.5,218.2) Overall M vs. L H vs, L 0.122 0.093 0.063 Enlisted Flyer n Mean 95% C.I. 55 65 57 221.1 227.7 213.9 (210.2,232.6) (204.2,224.2) (216.6,239.3) Overall M vs. L H vs. L 0.203 0.351 0.418 Enlisted Groundcrew n Mean 95% C.I. 154 160 142 211.4 213.3 211.1 (205.4,217.6) (205.2,217.2) (207.0,219.8) Overall M vs. L H vs. L 0.873 0.984 0.677 �TABLE K-4. (continued) Unadjusted Continuous Exposure Index Analyses for Hepatic Function Variables and Tvo Porphyrin Determinations by Occupation Exposure Index Triglycerides Occupation Statistic Low Officer Variable n Mean 95% C.I. Enlisted Flyer Enlisted Groundcrew Medium High Contrast p-Value 126 130 120 114.2 117.9 125.1 (100.2,130.1) (103.8,134.1) (109.5,142.9) Overall M vs. L H vs. L 0.624 0.728 0.338 n Mean 95% C.I. 65 57 55 123.8 119.9 123.1 (105.2,145.7) (103.2,139.3) (104.9,144.5) Overall M vs. L H vs. L 0.953 0.775 0.811 n Mean 95% C.I. 142 154 160 118.7 113.8 117.9 (108.1,130.3) (103.9,124.7) (107.0,129.9) Overall M vs. L H vs. 'L 0.796 0.528 0.919 �TABLE K-5. Summary of Adjusted Exposure Index, Analyses Involving Interactions with a Continuous Covariate for Hepatic Function Variables p-Value Interaction Exposure Index Level Slope p-Value on Test of Slope Against Null Hypothesis of Zero Slope Variable Occupation Interaction Covariate SCOT Enlisted Flyer Current Alcohol Use <0.001 Low Medium High -0.012 0.068 0.029 0.266 004 .0 0.138 SCOT Enlisted Groundcrew Current Alcohol Use <0.001 Low Medium High 0.004 007 .4 0.062 0.628 <0.001 <0.00i SGPT Enlisted Groundcrew Current Alcohol Use 0.006 Low Medium High -0.026 0.027 0.071 0.097 0.189 0.013 GGTP Enlisted Flyer Current Alcohol Use 0.003 Low Medium High 0.022 0.183 0.154 0.455 <0.001 0.006 GGTP Enlisted Groundcrew Current Alcohol Use 009 .0 Low Medium High 0.035 0.091 0.170 0.129 <0.001 <0.001 Direct Bilirubin Enlisted Flyer Current Alcohol Use 0.017 Low Medium High 0.013 0.097 0.065 0.445 <0.001 0.090 Cholesterol Enlisted Groundcrew Current Alcohol Use 0.006 Low Medium High 0.003 -0.008 0.033 0.591 0.285 0.002 �TABLE K-5. (continued) Summary of Adjusted Exposure Index, Analyses Involving Interactions vith a Continuous Covariate for Hepatic Function Variables Exposure Index Level Slope Variable I Enlisted Groundcrew Current Alcohol Use 0.025 Low Medium High -0.005 0.016 0.003 0.322 0.020 0.728 Triglycerides *> Occupation Interaction Covariate LDH l-» p-Value Interaction p-Value on Test of Slope Against Null Hypothesis of Zero Slope Enlisted Flyer Age 0.012 Low Medium High 0.014 0.015 0.038 0.286 0.247 0.032 �TABLE K-6. Summary of Adjusted Exposure Index, Analyses Involving Interactions vith a Categorical Covariate for Hepatic Function Variables Variable Occupation Total Bilirubin Enlisted Groundcrew Interaction Covariate Stratification Exposure Index Level Adjusted Mean (n) l-» Ul Black Low Medium High 0.63 (14) 0.90 (13) 0.70 (13) Nonblack Race Low Medium High 0.75 (138) 0.73 (147) 0.80 (127) Interaction p- Value 0.007 �TABLE K-7. Unadjusted Analyses for Baseline and Interval History of Liver Disease by Group (Verified by Medical Record Review) (Original Comparisons Only) Group Comparison Ranch Hand Disease Statistic Number Percent Est. Relative Number Percent Risk (95% C.I.) p-Value Hepatitis (Viral and Alcoholic) n Yes No 1,016 37 979 3.6 96.4 955 34 3.6 921 96.4 1.02 (0.64,1.65) 0.999 Jaundice n Yes No 1,016 20 996 2.0 98.0 955 21 2.2 934 97.8 0.89 (0.48,1.66) 0.754 n Yes No 1,016 3 1,013 0.5 99.5 955 1 0.1 954 99.9 2.83 (0.29,27.21) 0.625 Enlarged Liver n Yes No 1,016 17 999 1.7 98.3 955 15 1.6 940 98.4 1.07 (0.53,2.15) 0.999 Miscellaneous Liver Disorders n Yes No 1,016 17 999 1.7 98.3 955 7 0.7 948 99.3 2.31 (0.95,5.58) 0.065 Cirrhosis K-16 �TABLE K-8. Unadjusted Analyses of Enlarged Livers Diagnosed at Physical Examination by Group* (Original Comparisons Only) Enlarged Liver Yes Group Number No Percent Number Percent Total p-Value 0.227 Ranch Hand 8 0.8 1,002 99.2 1,010 Original Comparison 3 0.3 950 99.7 953 *Excludes participants with positive HBgAg. K-17 �TAHEK-9. and Two Borphyrin Determinations by Group (Original Comparisons Only) Group Variable Statistic Ranch Hand Original Comparison Est. Relative Risk (95% C.I.) SOOT n Mean 95% C.I. Nunber/% Normal High 1,009 33.5 (32.9,34.1) 952 33.4 (32.8,34.0) — 929 92.3% 80 7.9% 872 91.6% 80 8.4% n Mean 95% C.I. »jmber/% Normal High 1,C09 21.6 (21.0,22.3) 952 ' 22.5 (21.8,23.2) 872 86.4% 137 13.6% 816 85.7% 136 14.3% n Mean 95% C.I. Nunber/% Normal High 1,009 32.8 (31.4,34.3) 952 32.7 (31.3,34.2) 919 91.1% 90 8.9% 862 90.6% 90 9.4% n Mean 95% C.I. Number/% Normal High 1,009 91.8 (90.4,93.3) 952 89.7 (88.3,91.2) 953 94.5% 56 5.5% 911 95.7% 41 4.3% SOT QGTP Alkaline Phosphatase 0.94 (0.68,1.30) — 0.94 (0.73,1.22) — 0.94 (0.69, 1.28) — 1.31 (0.87,1.97) p-Value 0.903 0.701 0.060 0.651 0.921 0.682 0.044 0.205 �TABIEK-9. (ccntinuad) and Two Borphyrin Determinations by Group (Original Comparisons Cbly) Group • Variable Statistic Ranch Hand Original Comparison Total Bilirubin n Mean 95% C I .. Number/% Normal High 1,009 0.74 (.307) 07,.6 952 07 .5 (.307) 07,.6 982 9 . % 73 27 2 7 .% 919 9 . % 65 33 3 5 .% n Mean 95% C I .. Number/% Normal High 1,009 02 .8 (.702) 02,.8 952 02 .8 (.702) 02,.8 971 9 . % 62 38 38 .% 920 9 . % 66 32 3 4 .% n Mean 95% C I .. Number/% Normal High 1,039 123.5 (2.,2.) 122148 952 124.3 (2.,2.) 128157 999 9 . % 90 10 1 0 .% 941 11 n Mean 93% C I .. Number/% Normal High 109 ,0 214.3 (1.,1.) 218268 952 216.7 (1.,1.) 240293 863 8 . % 55 146 1 . % 45 778 8 . % 17 174 IS.3% - Direct Bilirubin LDH Cholesterol . 98.8% 12 .% Est. Relative Risk ( 5 C I ) 9%.. — 0.77 ( . 6 1 2 ) 04, . 8 — 1.13 ( . 0 1 8 ) 07, . 2 — 08 (.620) .6 03,.2 — 0.76 ( . 0 0 % 06,.) p-Value 0.772 030 .1 075 .9 069 .2 049 .4 0.724 0.205 003 .2 �TABLE K-9. (continued) and Two Parphyrin Determinations by (koup (Original < Xflll|Xil. 'SOPS 'fll'yj Grcwp Variable Statistic Ranch Band Original Comparison Triglycerides n Mean 95%C.I. Nunber/% Normal High 1,009 118.5 (113.8,123.3) 952 118.1 (113.4,122.9) 941 93.3% 68 6.7% 891 93.6% 61 6.4% n Mean 95% C.I. 1,006 16.9 (16.2,17.7) 949 17.7 (16.9,18.4) — 0.176 n Mean 95% C.I. 1,008 119.1 (116.2,122.0) 950 115.0 (112.0,118.0) — 0.053 Uroporphyrin Coproporphyrin — No relative ride or confidence interval given for continuous analyses. Est. Relative Risk (95% C.I.) — 1.06 (0.74,1.51) p-Value 0.908 0.767 �TABLE K-10. and Two Borphyrin Determinations by Group (Original Comparisons Oily) Group Ranch Hand Original Comparison n Adj. Mean 95% C.I. 1,003 34.69 (33.66,35.75) 949 34.64 (33.61,35.70) 0.908 n Adj. Mean 95% C.I. 1,003 35.97 (34.82,37.17) 949 35.99 (34.83,37.18) 0.974 n 1,003 949 CC SGPT Statistic DD SOOT Analysis OC Variable n Adj. Mean 95% C.I. 1,003 21.52 (20.89,22.17) 949 22.53 (21.86,23.23) 0.035 ALC*DC (p=0.002) AGE*ALC (p=0.006) n Adj. Mean 95% C.I. 1,003 21.49 (20.21,22.86) 949 22.52 (21.18,23.%) 0.032 ALODC (p=O.043) AGE*ALC (p=0.008) n 1,003 949 CD CD DD Adj. Relative Risk (95%C.I.) 0.93 (0.66,1.29) 0.91 (0.70,1.18) p-Value 0.655 0.467 Covariate Remarks* ALC*DC (p<0.001) AGE*ALC (p=0.033) RACE (p=0.003) RACE (p=0.004) AIC (pO.OOl) 1C (p=0.032) DC (p=0.044) AGE*ALC (p<D.OOi) CCC*ALC (p<0.001) RACE (p=0.026) AGE (p=0.008) ALC (p^O.003) �TABLE K-10. (continued) and T\» Pdrphyrin Determinations by Group (Original Comparisons Only) Group Adj. Relative Risk (95%C.I.) Govariate Reniarks* Ranch Hand Original Comparison n Adj. Mean 95% C.I. 1,003 37.47 (34.93,40.20) 949 37.39 (34.84,40.12) — 0.941 ALC*DC (pO.OOl) AG£*DC (p=0.029) RACE*IC (p=0.005) n Adj. Mean 952C.I. 1,003 42.35 (38.17,46.98) 949 42.24 (38.08,46.86) — 0.932 AG£*ALC (p=0.006) RACE (pO.OOl) n 1,003 949 OC Alkaline Fhosphatase Statistic ED QGTP Analysis OC Variable n Adj. Mean 9SSC.I. 1,003 91.5 (89.1,93.9) 950 89.1 (86.8, 91.5) n Adj. Mean 95% C.I. 1,003 950 -I t i t , ffTCKff AAXA AA AA AAAA n 1,003 950 CD CD DD 0.94 (0.68,1.28) p-Value 0.680 — 0.020 _ **** 1.39 (0.91,2.12) 0.121 AGE*ALC (p=0.005) RACE (p=0.049) RACE*IC (p=0.002) WINE (pO.OOl) AGE (p<0.001) OCC (p<0.001) GEP*IC (p=0.006) AGE*IC (p=0.040) RACE*IC ()M).002) WINE (pO.OOl) OCC (pO.OOl) WINE*DC (p=0.010) AGE*IC (p=0.006) RACE*IC (p=O.C04) OCC*IC (p=0.011) GRP*IC (marginal) (p=0.052) �TfiBlEK-lO. (continued) and Two Barphyrin Determinations by Qxup (Original Comparisons Only) Group Total Bilirubin U) Direct Bilirubin Statistic Ranch Hand Original Comparison OC n Adj. Mean 95* C.I. 1,003 0.78 (0.75,0.81) 949 0.78 (0.75,0.82) 0.714 CCC*RACE (p4D.002) ALC*RACE (p=O.008) AGE*DC (p=0.043) n Adj. Mean 95* C.I. 1,003 0.84 (0.80,0.88) 949 0.84 (0.80,0.88) 0.774 OCC*RACE (p=0.003) OCOALC (p4).047) RACE*ALC (p=0.002) n 1,009 952 0.314 RACE (p<O.C01) OC - Analysis ED Variable n Adj. Mean 95% C.I. 1,003 (0.278) (0.271,0.285) 949 0.276 (0.269,0.283) 0.746 n Adj. Mean 95%C.I. 1,003 0.308 (0.291,0.325) 949 0.305 (0.288,0.323) 0.673 1,003 949 **** CD CD ED Adj. Relative Risk (95% C.I.) 0.77(0.46,1.29) p-Value Covariate Remarks* ALC (p<0.001) IC*DC (p=0.014) ALC*DC (p=0.012) ALC««ACE (p=0.030) ALC*CCC (p=0.007) GEP*IC (p=0.010) RACE (p=0.014) ALC (p=0.016) �TABLE K-10. (continued) and fao Borphyrin Determinations by Group (original Canparisons Only) Group Ranch Band Original Comparison n Adj. Mean 95% C.I. 1,009 952 "fCfCffX "X7CK7C TfifcxX "JfXffJf n Adj. Mean 95% C.I. 1,009 129.4 (126.5,B2.2) 952 130.0 (127.2,132.9) n 1,003 949 CC Cholesterol Statistic ED U3H Analysis OC Variable n Adj. Mean 95% C.I. 1,003 219.3 (214.1,224.6) 949 221.3 (216.0,226.6) — 0.301 n Adj. Mean 95% C.I. 1,003 223.7 (217.3,230.3) 949 226.0 (219.6,232.5) — 0.243 n 1,003 949 CD CD ED Adj. Relative Risk (95% C.I.) p-Value Covariate Renarks* GRP*A2E (p=0.038) IOBACE (jMD.029) IC*OGC (^0.049) — **** — 0.555 M3S, (pO.OOl) RACE (p=O.006) 0.706 ALC (p=0.046) 0.85 (0.36,2.01) 0.74 (0.58,0.94) 0.015 OOO«ACE (p=0.003) DC*RACE (p=0.006) AGE (jxD.COl) AI£ (pOD.OOl) 1C (j^O.016) OCC««ACE (p=0.029) ALC (p<0.001) AGE (ixO.OOl) RACE*ALC (p=0.009) AGE (p=0.020) OCC (p=0.037) �TfiBLEK-10. (continued) and Two Borphyrin Daterminatians by Group (Original Comparisons Only) Group 'feriable Analysis CC Statistic Ranch Hand Original Comparison n 1,003 949 95% C.I. n 1,003 112.7 (103.4,122.9) 949 112.3 (103.1,122.4) n 1,003 949 Uroporphyrin CC — 0.904 946 Adj. Mean ED **** AAAA 95% C.I. CD p-Value _ Adj. Mean Triglycerides Adj. Relative Risk (95% C.I.) n 1,000 Adj. Mean AAA A 95% C.I. "XTfTCX 1.06 (0.74,1.51) — AAA A' 0.770 **** Covariate Remarks* GBP*AGE (p=0.012) ALC (p<0.001) RACE (pO.COl) OCC (pO.OOl) OCC (p<0.001) RACE (p<0.001) AGE (p<0.001) ALC (p=0.007) AGE*ALC (p=0.023) RACE (p=0.044) GRP*BUN (p4).014) OCO«CC (p=0.006) ALC (p=0.048) �T&BLEK-10. (continued) Adjusted and Categorical Analyses for Hepatic Ruction Variables and Tuo Porphyrin Determinatians by Group (Original Confiarisons Ckily) . Group Variable Analysis Statistic Ranch Hand Original Comparison Coproporphyrin CC n Adj. Mean 95% C I .. 1,002 119.1 (1.,2.) 162119 947 115.2 (1.,1.) 124182 Adj. Relative Risk ( 5 C I ) 9%.. — p-Value 006 .6 *Abbreviations; GRP: group OCC: occupation ALC: current alcohol use VINE: vine consumption DC: exposure to degreasing chemicals 1C: exposure to industrial chemicals BUN: blood urea nitrogen — No relative risk or confidence interval given for continuous analyses. ***^Group-by-covariate interaction—adjusted mean/relative risk, confidence interval, and p-value not presented. Covariate Remarks* ALC ( < . 0 ) p001 BUI ( < . 0 ) p001 �TABLE K-U. LY) (Origii Group Variable Interaction Stratification Statistic Ranch Hand Original Comparison Alkaline Phosphatase Grcqp-by-Industrial Chemicals Exposed to Industrial Chemicals n Mean Adj. .Mean 95% C.I. 504 94.8 96.9 (93.5,100.6) 506 90.0 92.1 (88.8,95.5) Not Exposed to Industrial Chemicals n Mean Adj. Mean 95% C.I. 499 88.9 92.5 (89.3,95.8) 444 89.5 93.3 (90.0,96.7) Bom >1942 n Mean Adj. Mean 95% C.I. 408 •122.2 124.1 (121.1,127.1) 369 121.7 123.6 (120.6,126.7) n Mean Adj. Mean 95% C.I. 565 123.8 126.4 (123.7,129.1) 538 125.2 127.8 (125.1,130.5) n Mean Adj. Mean 95% C.I. 36 135.1 138.0 (129.9,146.5) 45 134.6 137.4 130.1,145.1) Lffi Group-by-rAge Born 1923-1941 Bom <1922 p-Value 0.001 0.608 0.756 0.295 0.913 �TABLE K-ll. x« • • -1 — . * » i v (QciginaJ- firmpangmg Only) Group Variable Interaction Stratification Statistic Ranch Band Original Comparison Triglycerides Group-by-rAge Born XL942 n Mean Adj. Mean 95% C.I. 406 107.14 95.55 (87.62,104.20) 368 110.24 98.49 (90.12,107.63) n Mean Adj. Mean 95%C.I. 561 125.98 116.13 (107.59,125.35) 536 124.23 115.14 106.64,124.31) n Mean Adj. Mean 95%C.I. 36 139.00 129.40 (104.04,160.94) 45 106.00 98.40 (80.80,119.83) n Mean Adj. Mean 95% C.I. 547 16.23 16.39 (15.41,17.44) 509 18.18 18.45 (17.33,19.65) n Mean Adj. Mean 952C.I. 453 17.79 18.02 (16.85,19.27) 547 17.15 17.45 (16.30,18.67) Bom 1923-1941 • Bom <L922 Uroporphyrin Group-by-BIN BIW <L4 BUM >14 p-Value 0.055 0.824 0.510 0.005 0.479 �TABLE K-12. Sunnary of Group-by-Covariate Interactions for Variables and (Original Comparisons Only) Group Ranch Hand Variable Interaction Stratification Exposed to Industrial Chemicals Direct Bilirubin Statistic Number Percent Comparison Number Percent Adj. Relative Risk ( 5 C I ) p-Value 9% .. n High Normal 507 27 480 5.3 9. 47 506 14 492 2.8 97.2 1.96 ( . 2 3 8 ) 0 0 5 10,.0 .4 n High Normal 4% 11 485 2.2 97.8 443 18 425 4.1 95.9 0.53 ( . 5 1 1 ) 0 1 7 02,.5 .0 Group-byIndustrial Chemicals Not Exposed to Industrial Chemicals �TABLE K-13. Unadjusted Analysis for Interval History of Skin Bruises, Skin Patches, or Skin Sensitivity by Group (Original Comparisons Only) Bruises, Patches, or Sensitivity Yes Group No Number Percent Number Percent Ranch Hand 265 26.2 746 73.8 1,011 Original Comparison 195 20.5 758 79.5 953 K-30 Total p-Value 0.003 �B/796I/disk 11/document 11 TABLE K-14. Unadjusted Analyses for Porphyrin Abnormalities by Group and Skin Patch, Bruise, or Sensitivity Reported at Followup Questionnaire (Original Comparisons Only) Abnormal Porphyrin Findings for a Participant Group Skin Patch, Bruise, or Reported Sensitivity 0 Number 1 Percent Number 2 Percent Number Percent Total p-Value* Both Groups Yes No 412 1,361 89.8 91.0 45 129 9.8 8.6 2 6 0.4 0.4 459 1496 0.754 Ranch Hand Yes No 239 670 90.5 90.3 24 70 9.1 9.4 1 2 0.4 0.3 264 742 0.950 Original Comparison Yes No 173 691 88.7 91.6 21 59 10.8 7.8 1 4 0.5 0.5 195 754 0.419 *Chi-square test, 2 d . f . �TABLE K-15. Unadjusted Analyses for Uroporphyrin Abnormalities by Skin Patch, Bruise, or Sensitivity Reported at Follovup Questionnaire and by Group (Original Comparisons Only) Group Ranch Hand Original Comparison Stratification Statistic Skin Patch, Bruise, or Sensitivity Reported Variable n Abnormal Normal 264 12 252 4.5 95.5 195 9 186 Skin Patch, Bruise, or Sensitivity Not Reported n Abnormal Normal 742 42 700 5.7 94.3 754 40 714 Number Percent Number Est* Relative Risk (95Z C.I.) p-Value 4.6 95.4 0.98 ( . 1 2 3 ) 04,.8 0.999 5.3 94.7 1.07 ( . 9 1 6 ) 06,.7 0.821 Percent Uroporphyrins �TABLE K-16. Longitudinal Results for SCOT, SGPT, and GGTP: A Contrast of Baseline and First Follovup Examination Test Means (Original Comparisons Only) Means Variable SCOT SGPT GGTP Group Total 1982 Baseline 1985 Followup Ranch Hand Original Comparison 971 32.91 33.73 872 32.93 33.69 971 20.08 21.82 872 20.51 22.36 971 39.26 33.16 872 38.76 32.56 Ranch HandOriginal Comparison Ranch Hand Original Comparison *Analyzed in log units. K-33 p-Value* (Equality of Difference) 0.90 0.86 0.78 �APPENDIX L Dermatological Evaluation �APPENDIX L: Dermatological Evaluation Contents Table Page L-l Unadjusted Exposure Index Analyses for Dermatological Variables by Occupation L-l L-2 Interaction Summaries of Adjusted Exposure Index Analyses of Dermatological Variables L-5 L-3 Unadjusted Analysis for Reported Historical Occurrence of Acne by Group (Original Comparisons Only) L-8 L-4 Unadjusted Analysis for Reported Historical Occurrence Relative to 1961 by Group (Original Comparisons Only) L-8 L-5 Unadjusted Analysis for Reported Historical Occurrence of Acne Relative to SEA Tour of Duty for Post-1961 Acne by Group (Original Comparisons Only) L-9 L-6 Adjusted Analysis for Duration of Acne (in Years) for Post-1961 Acne by Group L-9 L-7 Results Discussed Narratively in Questionnaire Data Section of Chapter (Original Comparisons Only) L-10 L-8 Unadjusted Analyses for Dermatological Variables by Group (Original Comparisons Only) L-ll L-9 Adjusted Analyses for Dermatological Variables by Group (Original Comparisons Only) L-13 L-10 Summary of Group-by-Presence of Pre-SEA Acne Interaction for the Dermatology Index (Original Comparisons Only) L-15 L-ll Longitudinal Analysis of the Dermatology Index: A Contrast of Baseline and First Follovup Examination Abnormalities (Original Comparisons Only) L-16 Figure L-l Location of Post-SEA and Pre- and Post-SEA Acne by Group (Original Comparisons Only) L-17 L-2 Location of Post-SEA Acne by Group (Original Comparisons Only) ... L-18 L-i �TABLE Unadjusted Exposure Index Analyses for Dennatological Variables by Occupation Variable Low Number Percent Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 92.. 20.3 79.7 Overall M vs. L H vs. L 0.198 09,.5 1.78 ( . 5 3 3 ) 0 0 4 .8 1.45 ( . 5 2 8 ) 0.320 07,.0 57 16 41 28.1 71.9 0.225 Overall M vs. L 0.58 ( . 7 1 2 ) O.lflO 02,.3 H vs. L 0 5 ( . 5 1 1 ) 0 1 5 .6 .4 02,.9 22.1 77.9 142 41 101 28.9 71.1 Overall M vs. L H vs. L 0.393 0 8 ( . 8 1 3 ) 0.433 .1 04,.5 1.16 ( . 9 1 9 ) 0 6 3 06,.3 .0 130 21 109 16.2 83.8 123 23 100 18.7 81.3 Overall M vs. L H vs. L 070 .2 05,.5 1.10 ( . 6 2 1 ) 0 8 4 .6 06,.4 1.31 ( . 7 2 5 ) 0 5 0 .0 14.5 85.5 65 9 56 13.8 86.2 57 8 49 1. 40 8. 60 Overall M vs. L H vs. L 094 .9 0 9 (.426) 099 . 4 03,.4 .9 0 % (.327) 099 . 03,.7 .9 20.1 79.9 163 36 127 22.1 77.9 142 33 109 23.2 76.8 Overall M vs. L H vs. L 086 .0 1.13 ( . 6 1 9 ) 0.682 06,.3 06,.9 1.20 ( . 9 2 0 ) 0.573 n Abnormal Normal 127 19 108 15.0 85.0 130 31 99 23.8 76.2 123 25 98 Enlisted Flyer n Abnormal Normal 55 23 32 4. 18 58.2 65 19 46 29.2 7. 08 Enlisted Groundcrew n Abnormal Normal 154 40 114 26.0 7. 40 163 36 127 Officer Acneiform Lesions Statistic Officer •Comedones Occupation n Abnormal Normal 127 19 108 15.0 85.0 Enlisted Flyer n Abnormal Normal 55 8 47 Enlisted Groundcrew n Abnormal Normal 154 31 123 �.___— _ _ unadjusted &Kposure Index Analvses for 1Iprmat-nlncriretl rf Variable Low Number Percent 'Obi-iaKloQ YKT (V-mmrim —. ^ -_^-_ _ _ Exposure Index Medium High Number Percent Number Percent Statistic Officer n Abnormal Normal 127 7 120 5.5 94.5 130 13 117 1. 00 9. 00 123 15 108 12.2 87.8 0.175 Overall 07,.4 M vs. L 1.91 ( . 3 4 9 ) 0.245 H vs. L 2.38 ( . 4 6 0 ) 0 0 5 09,.6 .7 Enlisted " Flyer n Abnormal Normal 55 12 43 21.8 78.2 65 12 53 18.5 81.5 57 7 50 12.3 87.7 041 .0 Overall M vs. L 0 8 ( . 3 1 9 ) 0.655 .1 03,.9 H vs. L 0.50 ( . 8 1 3 ) 0.214 01,.9 Enlisted Groundcrew n Abnormal Normal 154 22 132 14.3 85.7 163 29 134 17.8 82.2 142 33 109 23.2 76.8 Overall M vs. L H vs. L Officer • Acneiform Scars Contrast Est. Relative Risk ( K . . p-Value 9CI) Occupation n Abnormal Normal 127 14 113 11.0 8. 90 130 5 125 3.8 96.2 123 19 104 15.4 8. 46 Overall 008 .0 01,.3 M vs. L 0.32 ( . 1 0 9 ) 0.033 H vs. L 1 4 ( . 0 3 0 ) 0.352 . 8 07,.9 Enlisted Flyer Abnormal Normal 55 10 45 18.2 81.8 65 7 58 1. 08 89.2 57 8 49 1. 40 8. 60 Overall Hvs. L H vs. L 0.509 01,.4 0.54 ( . 9 1 5 ) 0.298 02,.3 0.74 ( . 7 2 0 ) 0.613 Enlisted Groundcrew n Abnormal Normal 154 15 139 9.7 90.3 163 7 156 4.3 95.7 142 17 125 12.0 8. 80 Overall M vs. L M vs. L 0.045 01,.5 0.42 ( . 7 1 0 ) 0 0 6 .7 1.26 ( . 0 2 6 ) 0.578 06,.3 0.135 07,.8 1.30 ( . 1 2 3 ) 0 4 6 .4 1.82 ( . 0 3 3 ) 0.053 10,.0 E Depigmentation �TfiBLB L-l. (continued) Unadjusted Exposure Index Analyses for Dermatological \fariables by Occupation Variable Lew Number Percent Exposure Index Medium High Number Percent Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9% . . 10.6 8. 94 Overall M vs. L H vs. L 0.221 1.92 ( . 8 4 1 ) 0.125 08,.9 1.25 ( . 4 2 9 ) 0.671 05,.0 57 8 49 1. 40 8. 60 Overall M vs. L H vs. L 087 .5 1.31 ( . 4 3 9 ) 0 7 3 04,.5 .8 1.33 ( . 3 4 1 ) 0 7 7 04,.3 .7 9.2 9. 08 142 16 126 11.3 8. 87 Overall 086 .3 M vs. L 0.87 ( . 2 1 8 ) 0 8 0 04,.4 .5 H vs. L 1.10 ( . 3 2 2 ) 0 8 3 05,.8 .5 130 23 107 17.7 82.3 123 20 103 16.3 83.7 Overall M vs. L H vs. L 0.922 0.97 ( . 1 1 8 ) 0 9 9 05,.4 .9 08 (.617) 0 7 9 .8 04,.0 .3 23.6 76.4 65 D 52 2. 00 8. 00 57 16 41 28.1 71.9 Overall M vs. L H vs. L 059 .7 .6 08 (.419) 0 6 2 .1 03,.3 05,.5 .6 1.26 ( . 4 2 9 ) 0 6 9 26.6 73.4 163 47 116 2. 88 71.2 142 32 110 22.5 77.5 Overall M vs. L H vs. L 0.452 1.12 ( . 8 1 8 ) 0 7 7 06,.3 .0 .2 08 (.713) 0 4 2 .0 04,.6 Statistic Officer n Abnormal Normal 127 11 116 8.7 91.3 130 20 110 15.4 8. 46 123 13 110 Enlisted ' Flyer n Abnormal Normal 55 6 49 10.9 89.1 65 9 56 13.8 86.2 Enlisted Groundcrev n Abnormal Normal 154 16 138 10.4 89.6 163 15 148 Officer • Inclusion Cysts Occupation n Abnormal Normal 127 23 104 18.1 81.9 n Abnormal Normal 55 13 42 n Abnormal Normal 154 41 113 Hyperpigmen- Enlisted tation Flyer Enlisted Groundcrev �TAHE L-l. (contiiued) Unadjusted Exposure Index Analyses for Dermatological Variables by Occupation Exposure Index Variable Low Number Percent Occupation Statistic Officer n Abnormal Normal 127 73 54 n Abnormal Normal Enlisted Groundcrew n Abnormal Normal Hsdli™ High Number Percent Number Percent Est. Relative Risk ( 5 C I ) p-Walue 92 . . 67.0 33.0 123 85 38 69.1 30.9 Overall 0.122 M vs. L 1.50 ( . 0 2 4 ) 0.125 09,.9 H vs. L 1.66 ( . 8 2 7 ) 0 0 7 09,.8 .6 65 43 22 66:2 33.8 57 33 24 57.9 42.1 Overall 001 .7 M vs. L 0.55 ( . 4 1 2 ) 0.160 02,.4 H vs. L 0.38 ( . 7 0 8 ) 0.027 01,.8 56.5 43.5 163 78 85 47.9 52.1 142 79 63 55.6 4. 44 Overall 0.237 M vs. L 0.71 ( . 5 1 1 ) 0.144 04,.0 H vs. L 0.97 ( . 1 1 5 ) 0 9 7 06,.3 .0 n 127 X) (Abnormal) 46 0 (Normal) 81 36.2 63.8 130 56 74 43.1 56.9 123 56 67 45.5 54.5 Overall 030 .0 Mvs. L 1.33 ( . 1 2 2 ) 0 3 8 08,.0 .0 H vs. L 1.47 ( . 9 2 4 ) 0.157 08,.4 Enlisted Flyer n 55 X) (Abnormal) 33 0 (Normal) 22 6. 00 4. 00 65 34 31 52.3 47.7 57 29 28 5. 09 49.1 Overall 059 .7 03,.1 Mvs. L 0.73 ( . 5 1 5 ) 0 4 2 .6 H vs. L 0 6 ( . 3 1 4 ) 0.349 .9 03,.6 Enlisted Groundcrew n 154 X) (Abnormal) 74 0 (Normal) 80 48.1 51.9 163 77 86 47.2 52.8 142 78 64 5. 49 45.1 Overall 0.348 M vs. L 0 9 ( . 2 1 5 ) 0.911 .7 06,.0 H vs. L 1.32 ( . 3 2 0 ) 0.247 08,.8 57.5 42.5 130 87 43 55 43 12 78.2 21.8 154 87 67 Officer Other Enlisted Abnormalities Flyer Dermatology Index Contrast �TABLE L-2. Index Analyses of Dermatological Variables Variable Interaction (Occupation) Stratification Acneiform Scars' Acneiform Scars Exposure Index-byPresence of pre-SEA Acne (Officer) Exposure Index-byPresence of pre-SEA Acne (Enlisted Groundcrew) p-Value n Abnormal Normal 99 0 99 00 . 100 0. 77 6 71 7.8 92.2 76 8 68 10.5 89.5 Overall M vs. L H vs. L n Abnormal Normal 25 5 20 2. 00 8. 00 49 7 42 14.3 85.7 44 6 38 13.6 8. 64 Overall H vs. L 0 7 ( . 0 2 6 ) . 3 02,.8 H vs. L 0 6 ( . 7 2 5 ) .7 01,.6 083 .3 060 .3 0.555 Black, No pre-SEA Acne n Abnormal Normal 1 1 0 100 0. 00 . 2 0 2 00 . 100 0. 1 0 1 00 . 100 0. Overall M vs. L 0 0 ( . 0 , . 9 * . 7 00154)* H vs. L 0.11 ( . 0 , 0 2 ) * 0011.7* 0.135* 0.333* 099 .9* Black, pre-SEA Acne n Abnormal Normal 1 1 0 100 0. 00 . 1 0 1 00 . 100 0. 1 1 0 100 0. 00 . Overall ' M vs. L 0.11 ( . 0 , 0 2 ) * 0011.7* H vs. L — 0.135* 099 .9* — No pre-SEA Acne n Abnormal Normal 101 16 85 15.8 84.2 105 15 90 14.3 85.7 99 16 83 16.2 83.8 M vs. L H vs. L 08 (.918) .6 03,.6 1.11 ( . 1 2 3 ) 05,.9 065 .9 079 .9 Pre-SEA Acne n Abnormal Normal 51 6 45 11.8 88.2 57 14 43 24.6 75.4 41 17 24 41.5 58.5 M vs. L H vs. L 2.24 ( . 9 6 3 ) 07,.9 5.38 ( . 5 1 . ) 14,9% 0.131 002 .1 Nonblack, No pre-SEA Acne Exposure Index-byRace Exposure Index Low Mediun Est. Relative Hitfi Statistic Number Percent Number Percent Number Percent Contrast Risk ( 5 C I ) 9* . . Nonblack, pre-Sea Acne 18.09 ( . 0 3 6 3 ) * 10,2.1* 2.9 (.044%* 46 14,3.)* 006 .0* 006 .0* 001 .0* �TABLE L-2. (continued) Index Analyses of variable Interaction (Occupation) Depigmentation Exposure Index-byPresence of pre-SEA Acne (Enlisted Groundcrev) Stratification lYmnt-n' «T Irwical \Variables — ~ Exposure Index Median Lev High Est. Relative Statistic Number Percent Number Percent Number Percent Contrast Risk (95ZC.I.) p-Value Dermatology Index • Exposure Index-byPresence of pre-SEA Acne (Officer) n Abnormal Normal 101 13 88 12.9 87.1 105 7 98 6.7 93.3 99 11 88 11.1 88.9 Overall M vs. L 0.48 ( . 9 1 2 ) 01,.7* H vs. L 0.85 ( . 6 1 9 ) 03,.9* 034 .1* 011 .6* 088 .2* Pre-SEA Acne n Abnormal Normal 51 2 49 3.9 96.1 57 0 57 00 . 100.0 41 6 35 14.6 85.4 Overall M vs. L H vs. L 006 .0* 021 .2* 0.133* Nonblack, No pre-SEA Acne Exposure Index-byRace No pre-SEA Acne n Abnormal Normal 99 32 67 32.3 67.7 77 40 37 51.9 48.1 76 35 41 46.1 53.9 Overall M vs. L 2.26 ( . 3 4 1 ) 12..8* H vs. L 1.78 ( . 6 3 3 ) 09,.1* 005 .2* 003 .1* 004 .8* Nonblack, pre-SEA Acne n Abnormal Normal 25 12 13 48.0 52.0 49 15 34 30.6 69.4 44 20 24 45.5 54.5 Overall M vs. L 0.48 ( . 8 1 2 ) 01,.9* H vs. L 0.90 ( . 4 2 4 ) 03,.1* 022 .2* 022 .0* 099 .9* Black, No pre-SEA Acne n Abnormal Normal 1 1 0 100.0 00 . 2 0 2 00 . 100.0 1 0 1 00 . 100.0 Overall M vs. L H vs. L 0.07 ( . 0 , . 9 * 00154)* 0.11 ( . 0 , 0 2 ) * 0011.7* 0.135* 033 .3* 099 .9* Black, pre-SEA Acne n Abnormal Normal 1 1 0 100.0 00 . 1 0 1 00 . 100.0 1 1 0 100.0 00 . Overall M vs. L 0.11 ( . 0 , 0 2 ) * 0011.7* — H vs. L 0.223* 099 .9* — 0.17 ( . 1 3 6 ) * 00, .7* 4.20 ( . 0 2 . 5 * 08,20) �TABLE L-2. (continued) Interaction Suamaries of Adjusted Exposure Index Analyses of Dermatological Variables FOOTNOTES, Table L-2 a Eesults presented for acneiform scars (officers) are based on stratification into the four categories shown. Unadjusted results are presented for all strata except nonblack Banch Hands with pre-SEA. acne. These results are adjusted for age. *Uhadjusted estimate of relative risk, and confidence interval, or p-value, based on stratified tables. **uhadjusted estimate of relative risk and confidence interval calculated after adding 0.5 to each cell. —-No normal participants present in contrast; estimated relative risk, confidence interval and p-value not presented. Note: Snail sample sizes may affect validity of overall p-value. ip --j Note: Results without (*) or ( * are adjusted for all other main effects in model (age, race, presence of pre-SEA acne), unless otherwise *) noted. �TABLE L-3. Unadjusted Analysis for Reported Historical Occurrence of Acne by Group (Original Comparisons Only) Acne Yes Group Number No Percent Number Percent Total Ranch Hand Original Comparison 412 40.6 604 59.4 1,016 349 36.5 606 63.5 955 Total 761 Summary Statistics Est. RR: 1.18 95% C.I. 5 (0.99,1.42) p-Value: 0.071 1,971 1,210 TABLE L-4. Unadjusted Analysis for Reported Historical Occurrence Relative to 1961 by Group* (Original Comparisons Only) Occurrence of Acne Post-1961 Group Number Percent Pre-1961 Number Percent Total Ranch Hand Original Comparison 239 58.3 171 41.7 410 192 55.2 156 44.8 348 Total 431 327 Summary Statistics Est. RR (for post1961 cases): 1.14 95* C.I.: (0.85, 1.52) p-Value: 0.418 758 *Three participants deleted due to missing data on time of occurrence. L-8 �TABLE L-5. Unadjusted Analysis for Reported Historical Occurrence of Acne Relative to SEA Tour of Duty for Post-1961 Acne by Group* (Original Comparisons Only) Post-1961 Acne Pre-SEA Group Ranch Hand Original Comparison Total Number Percent . Post-SEA Pre- and Post-SEA Total Number Percent Number Percent p-Value 0.155 58 25.4 80 35.1 90 39.5 228 48 106 26.2 81 161 44.2 54 144 29.5 183 411 *Twenty post-1961 participants with acne deleted due to missing data on time of occurrence. TABLE L-6. Adjusted Analysis for Duration of Acne (in Years) for Post-1961 Acne by Group* Group Ranch Hand Original Comparison Total Total Adjusted Mean** 219 8.18 (7.43,8.96) 175 394 7.15 Covariate Remarks (6.38,7.97) 952 C.I.** p-Value 0.070 Time Reference to SEA (p<0.001) *Seventeen participants deleted due to missing data on time of occurrence. **Converted from square root scale. L-9 �TABLE L-7. Results Discussed Narratively in Questionnaire Data Section of Chapter (Original Comparisons Only) Description p-Value • Effect due to SEA category in continuous analysis of duration of acne • Interaction between group and SEA category in continuous analysis of duration of acne <0.001 0.286 • Categorical analysis of duration of acne — p-value for group difference SEA Category pre-SEA post-SEA pre- and post-SEA 0.718 0.592 0.753 • Intersection of Venn diagram circles (temples, ears, and eyes) — p-value for group difference SEA Category post-SEA and pre- and post-SEA post-SEA L-10 0.133 0.627 �TABLE L-8. Unadjusted Analyses for Dermatological Variables by Group (Original Comparisons Only) Group Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk ( 5 C.I.) 9% p-Value 26.5 73.5 0.90 (0.74,1.11) 0.352 954 159 795 16.7 83.3 1.14 (0.90,1.43) 0.288 14.8 85.2 954 128 826 13.4 86.6 1.12 (0.87,1.44) 0.401 1,016 102 914 10.0 90.0 954 113 841 11.8 88.2 0.83 (0.63,1.10) 0.219 n Abnormal Normal 1,016 114 902 11.2 88.8 954 118 836 12.4 87.6 0.90 (0.68,1.18) Hyperpigmentation n Abnormal Normal 1,016 228 788 22.4 77.6 954 222 732 23.3 76.7 0.95 (0.77,1.18) Variable Comedones t-1 Statistic n Abnormal Normal 1,016 • 250 766 24.6 75.4 954 253 701 Acneiform Lesions n Abnormal Normal 1,016 188 828 18.5 81.5 Acneiform Scars n Abnormal Normal 1,016 150 866 Depigmentation n Abnormal Normal Inclusion Cysts • 0.442 0.668 �TABLE L-8. (continued) Unadjusted Analyses for Dermatological Variables by Group (Original Comparisons Only) Group Ranch Hand Number Percent Original Comparison Number Percent Variable Statistic Other Abnormalities n Abnormal Normal 1,016 608 408 59.8 40.2 954 565 389 59.2 40.8 Dermatology Index n 0 1 2 3 4 1,016 533 318 121 34 10 52.5 31.3 11.9 3.3 1.0 954 497 301 116 35 5 52.1 31.6 12.2 3.7 0.5 Est . Relative Risk ( 5 C.I.) 9% 1.03 (0.86,1 .23) Overall 1 vs. 0 0 . 9 9 2 vs. 0 0 .97 3 vs. 0 0 ,91 4 vs. 0 1 .87 ( .81,1. 20) 0 ( .73,1. 29) 0 (0.56,1. 48 ^ ( .63,5. 49) 0 p-Value 0.783 0.816 0.919 0.885 0.711 0.304 �TABLE L-9. Adjusted Analyses for Dermatological Variables by Group (Original Comparisons Only) Group Variable Ranch Hand Total Original Comparison Total Adj . Relative Risk (95% C.I.) Comedones 1,007 946 0.93 (0.75,1.14) 0.469 RACE (p<0.001) AGE (p<0.001) OCC*SEA ACNE (p=0.047) Acneiform Lesions 1,007 946 1.13 (0.89,1.43) 0.315 AGE (p<0.001) RACE (p=0.034) SEA ACNE (p=0.007) Acneiform Scars 1,007 946 1.12 (0.86,1.45) 0.417 AGE (p<0.001) RACE (p<0.001) SEA ACNE (p<0.001) Depigmentation 1,016 954 0.83 (0.63,1.10) 0.202 RACE (p=0.010) Inclusion Cysts 1,016 954 0.90 (0.68,1.18)* 0.442* — Hyperpigmentation 1,007 946 0.95 (0.76,1.17) 0.608 RACE (p<0.001) SEA ACNE (p=0.008) p- Value Covariate Remarks �TABLE L-9. (continued) Adjusted Analyses for Dermatological Variables by Group (Original Comparisons Only) Group Ranch Hand Total Original Comparison Total Other Abnormalities 1,016 954 Dermatology Index 1,007 946 Variable Adj. Relative Risk (95% C.I.) 1.07 (0.89,1.30) **** p-Value 0.459 **** Covariate Remarks AGE (p<0.001) RACE (p<0.001) OCC (p=0.013) GROUP*SEA ACNE (p=0.030) I l-» *- *No significant covariates in adjusted analysis; consequently, estimated relative risk, confidence interval, and p-value are from unadjusted analysis. ****Group-by-covariate interaction; adjusted relative risk, confidence interval, and p-value are not presented. Abbreviations; OCC: Occupation SEA ACNE: Presence of pre-SEA acne �TABLE HO. Summary of Group-by-Presence of. Pre-SEA Acne Interaction for Dermatology Index (Original Comparisons Only) Group Variable Interaction Stratification Statistic Ranch Hand Number Percent n No r K Dennatology Index Group-bypresenceof-pre-SEA. acne 684 0 1 2 3 4 360 234 69 16 5 0 1 2 3 4 167 82 51 18 5 Contrast Adj. Relative Risk(95%C.I.) p-Value 653 52.6 34.2 10.1 2.3 07 . 323 Yes Original Conparison Number Percent 370 202 61 19 1 56.7 3. 09 9.3 2.9 0.2 1 vs. 0 1.19 2 vs. 0 1.17 3 vs. 0 0.87 4 vs. 0 4.16 41.6 33.4 1B.8 5.1 1.0 1 vs. 0 2 vs. 0 3 vs. 0 4 vs. 0 (0.94,1.52) (0.80,1.70) (0.44,1.71) (0.61,28.48) 0.153 0.422 0.689 0.146 293 51.7 25.4 15.8 5.6 1.5 122 98 55 15 3 0.62 (0.42,0.90) 0.68 (0.44,1.07) 0.89 (0.43,1.83) 1.20 (0.30,4.88) 0.012 0.097 0.745 0.797 �TABLE L-ll. Longitudinal Analysis of the Dermatology Index: A Contrast of Baseline and First Followup Examination Abnormalities (Original Comparisons Only) Group 1982 Baseline Exam 1985 Followup Exam Abnormal Normal Odds Ratio (OR)* Ranch Hand Abnormal Normal 241 228 136 366 1.68 Original Comparison Abnormal Normal 222 202 109 339 1.85 p-Value (ORRH vs. ORQC) 0.53 Number Normal Baseline, Abnormal Followup *OrMa Rat in; Number Abnormal Baseline, Normal Followup L-16 �Ranch Hand 53 Other Sites n=80 Original Comparison p = 0.374 (8 Categories) 36 Other Sites n=48 Figure L-1. Location of Post-SEA and Pre- and Post-SEA Acne by Group (Original Comparisons Only) L-17 p = 0.310 (7 Categories; Other Sites Deleted) �Ranch Hand 103 Other Sites n=169 Original Comparison p - 0.492 (8 Categories) p = 0.449 (7 Categories; Other Sites Deleted) 85 Other Sites n=129 Figure L-2. Location of Post-SEA Acne by Group (Original Comparisons Only) L-18 �APPENDIX M Cardiovascular Assessment �APPENDIX M: CARDIOVASCULAR ASSESSMENT Contents Table M-l M-2 M-3 M-4 M-5 M-6 M-7 Page Association Between Reported Essential Hypertension and the Covariates in the Combined Ranch Hand and Comparison Groups M-l Association Between Reported Heart Disease and the Covariates in the Combined Ranch Hand and Comparison Groups M-2 Association Between Reported Myocardial Infarction and the Covariates in the Combined Ranch Hand and Comparison Groups M-3 Cardiovascular Morbidity-Mortality Analysis (Ranch Hands and Comparisons) M-4 Summary of Group-by-Covariate Interactions for Central Cardiac Function Variables (Diabetics Excluded) M-5 Summary of Group-by-Covariate Interactions for Peripheral Vascular Function Variables (Diabetics Excluded) M-6 Unadjusted Exposure Index Analyses for Reported and Verified Heart Disease by Occupation M-7 Unadjusted Exposure Index Analyses for Central Cardiac Function Variables by Occupation M-10 Unadjusted Exposure Index Analyses for Diastolic Blood Pressure, Funduscopic Abnormalities, Carotid Bruits, and Manual Pulse Readings by Occupation M-15 M-10 Unadjusted Exposure Index Analyses for Peripheral Vascular System Doppler Pulse Readings by Occupation . . . . M-21 M-8 M-9 M-ll Association Between Central and Peripheral Abnormalities and Verified Heart Disease in the Combined Ranch Hand and Comparison Groups M-25 M-12 Unadjusted Analyses for Reported and Verified Heart Disease by Group (Original Comparisons Only) M-26 M-13 Adjusted Analyses for Reported and Verified Heart Disease (Original Comparisons Only) M-27 M-i �APPENDIX M: CARDIOVASCULAR ASSESSMENT Contents (continued) Table Page M-14 Unadjusted Analyses for Central Cardiac Function by Group (Diabetics Excluded) (Original Comparisons Only) M-28 M-15 Adjusted Analyses for Central Cardiac Function (Diabetics Excluded) (Original Comparisons Only) M-30 M-16 Summary of Group-by-Covariate Interactions for Central Cardiac Function Variables (Diabetics Excluded) (Original Comparisons Only) M-32 M-17 Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) (Original Comparisons Only) . . . M-33 M-18 Adjusted Analyses for Peripheral Vascular Function (Diabetics Excluded) (Original Comparisons Only) M-36 M-19 Summary of Group-by-Covariate Interactions for Peripheral Vascular Function Variables (Diabetics Excluded) (Original Comparisons Only) M-38 M-20 Longitudinal Analysis of Pulse Index and Overall ECG: A Contrast of Baseline and First Followup Examination Abnormalities (Original Comparisons Only) M-39 M-ii �TABLE M-l. Association Between Reported Essential Hypertension and the Covariates in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Percent Abnormal p-Value Age Born >1942 Born <1942 934 1,214 20.4 29.6 <0.001 Race Black Nonblack 126 2,022 27.8 25.5 0.647 Occupation Officer Enlisted Flyer Enlisted Groundcrev 807 354 987 27.6 27.1 23.5 0.108 Current Smoking 0 >0 - 20 >20 1,262 . 463 422 28.8 22.0 20.4 <0.001 Pack-Years Smoking 0 >0 - 10 >10 512 760 869 28.7 23.6 25.7 0.120 Cholesterol <200 >200 - 230 >230 766 650 732 19.7 26.6 31.0 <0.001 HDL <40 >40 - 50 >50 719 754 675 27.5 25.3 24.0 0.309 Cholesterol-HDL Ratio <4.2 >4.2 - <5.5 >5.5 717 743 688 20.6 26.0 30.5 <0.001 Percent Body Fat <10 10 - 25 >25 10 1,758 379 0.0 21.0 47.8 <0.001 Personality Score <-5 -5-5 >5 829 731 580 26.8 26.7 22.9 0.202 Differential Cortisol <0.6 >0.6 - 4.0 >4.0 704 745 683 29.1 24.4 23.3 0.030 592 809 738 25.2 22.5 29.1 0.011 691 719 666 25.9 21.8 29.4 0.005 Current Alcohol 0 Use (Drinks/Day) >0 - 1 >1 Drink-Years Alcohol <1.25 >1.25 - 25 >25 M-l �TABLE M-2. Association Betveen Reported Heart Disease and the Covariates in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Percent Abnormal p-Value Age Born >1942 Born <1942 934 1,214 22.7 30.3 <0.001 Race Black Nonblack 126 2,022 31.0 26.8 0.354 Occupation Officer Enlisted Flyer Enlisted Groundcrev 807 354 987 30.6 24.9 24.8 0.014 Current Smoking 0 >0 - 20 >20 1,262 463 422 27.5 28.1 24.2 0.342 Pack- Years Smoking 0 >0 - 10 >10 512 760 869 28.1 25.9 27.3 0.667 Cholesterol <200 >200 - 230 >230 766 650 732 26.5 25.5 28.8 0.361 HDL <40 >40 - 50 >50 719 754 675 26.6 25.6 29.0 0.326 Cholesterol-HDL Ratio <4.2 >4.2 - <5.5 >5.5 717 743 688 29.6 23.0 28.6 0.009 Percent Body Fat <10 10 - 25 >25 10 1,75& 379 30.0 27.2 25.9 0.849 Personality Score <-5 -5-5 >5 829 731 580 27.5 24.9 29.1 0.214 Differential Cortisol <0.6 >0.6 - 4.0 >4.0 704 745 683 24.6 27.9 28.4 0.212 Current Alcohol 0 Use (Drinks/Day) >0 - 1 >1 592 809 738 27.4 28.6 25.1 0.297 Drink-Years Alcohol 691 719 666 26.8 25.7 28.4 0.537 <1.25 >1.25 - 25 >25 M-2 • �TABLE M-3. Association Between Reported Myocardial Infarction and the Covariates in the Combined Ranch Hand and Comparison Groups Covariate Covariate Category Total Percent Abnormal p-Value Age Born >1942 Born <1942 934 1,214 0.4 3.1 <0.001 Race Black Nonblack 126 2,022 0.8 2.0 0.523 Occupation Officer Enlisted Flyer Enlisted Groundcrew 807 354 987 1.5 2.3 2.2 0.477 Current Smoking 0 >0 - 20 >20 1,262 463 422 1.4 3.0 2.4 0.083 Pack-Years Smoking 0 >0 - 10 >10 512 760 869 0.2 1.8 3.1 0.001 Cholesterol <200 >200 - 230 >230 766 650 732 1.2 2.0 2.7 0.093 HDL <40 >40 - <50 >50 719 754 675 2.8 1.9 1.2 0.096 Cholesterol-HDL Ratio <4.2 >4.2 - <5.5 >5.5 717 743 688 1.0 1.5 3.5 0.002 Percent Body Fat <10 10 - 25 >25 10 1,758 379 0.0 1.9 2.1 0.867 Personality Score <-5 -5-5 >5 829 731 580 1.9 2.6 1.2 0.195 Differential Cortisol <0.6 >0.6 - 4.0 >4.0 704 745 683 2.0 1.9 2.0 0.973 Current Alcohol 0 Use (Drinks/Day) >0 - 1 >1 592 809 738 2.4 2.1 1.4 0.365 Drink-Years Alcohol 691 719 666 2.6 0.8 2.2 0.034 <1.25 >1.25 - 25 >25 M-3 �TABLE M-4. Cardiovascular Morbidity-Mortality Analysis (Ranch Hands and Comparisons*) Ranch Hands n-1,257 Event Endpoint Status Dead Living, Fully Compliant* Living, not Fully Compliant Event Yes Death (Any Cause) or Verified Heart Disease Comparisons* n.1,253 Estimated Percent Events Death (Any Cause) or Verified Myocardial Infarction Dead Living, Fully Compliant* Living, not Fully Compliant Estimated Percent Events Fatal or Nonfatal Verified Heart Disease Dead Living, Fully Compliant* Living, not Fully Compliant Estimated Percent Events Fatal or Nonfatal Verified Myocardial Infarction or Fatal Heart Disease *Comparison: Hands. Dead Living, Fully Compliant* Living, not Fully Compliant Estimated Percent Events Total No Total Yes 66 0 66 77 0 77 251 823 1,074 187 768 955 27.4° 89. 6C 117 27.4% 66 43. 3C No 177. 7C 221 24.5% 0 66 77 0 77 1,074 9 946 955 1.4° 115.6° 117 2.1° 218. 9C 221 6.4% 7.0% 13 1 ,061 5 61 66 3 74 77 251 823 1,074 187 768 955 27.4C 89.6° 117 22.5% 1 13 1.4C 43.3C 177.7C 221 18.6% 65 66 2 75 77 1,061 1,074 9 946 955 2.1° 218.9C 221 115.6° 117 1.2% 1.0% first Comparison of the randomly ordered set matched to the Ranch *Fully compliant at Baseline or third-year followup examination. Not fully compliant at Baseline and not fully compliant at third-year followup examination. c Estimated using event rate in living, fully compliants. M-4 �TABLE M-5. Summary of Group-by-Covariate Interactions for Central Cardiac Function Variables (Diabetics Excluded) Variable Interaction Stratification Adj. Relative Risk (95% C.I.) p-Value 4.56 (-1.91,11-03)* -16.01 (-27.52,-4.50)* 0.170 008 .0 Systolic Blood Pressure (Continuous Analysis) Group-by-Age (Blacks) Age (Baseline): 35 Age (Baseline): 53 ECG (Overall) Group-by-Pack-years Smoking Pack-years: 0 Pack-years: 30 0.70 (0.50,0.98) 1.25 (0.89,1.76) 0.038 0.197 ECG: Arrythmia Group-by-pack-years Smoking (21% Body Pat) Pack-years: 0 Pack-years: 30 0.58 (0.30,1.10) 1.59 (0.83,3.04) 0.093 0.162 Group-by-Percent Body Fat (7 Pack-years Smoking) 10% Body Fat 30% Body Fat 0.23 (0.07,0.78) 1.88 (0.66,5.34) 0.018 0.234 *Difference in group means (Ranch Hand-Comparison) and associated p-value given, rather than relative risk, for continuous analysis of dependent variables. �TABLE M-6. Sumnary of Group-by-Covariate Interactions for Peripheral Vascular Function Variables (Diabetics Excluded) Variable Interaction Stratification Adj. Relative Risk (95X C.I.) p-Value Popliteal Pulses (Manual) Group-by-Race Black Nbnblack 6.74 (0.72,63.40) 0.55 (0.28,1.12) 0.095 009 .9 Dorsalis Pedis Pulses (Manual) Group-by-Occupation Officer Enlisted Flyer Enlisted Groundcrew 0.70 ( . 5 1 0 ) 04,.9 1.70 ( . 6 3 3 ) 08,.4 1.32 ( . 4 2 0 ) 08,.6 0.116 0.124 0.222 Posterior Tibial Pulses (Manual) Group-by-Occupa t i on Officer Enlisted Flyer Enlisted Groundcrew 0.33 ( . 9 1 1 ) 00,.8 4.33 (1.14,16.50) 1.22 (0.58,2.57) 0.087 0.032 0.603 Leg Pulses (Manual) Group-by-occupation (21% Body Fat) Officer Enlisted Flyer Enlisted Groundcrew 0.62 ( . 1 0 9 ) 04,.4 1.55 (0.87,2.76) 1.24 (0.85,1.81) 0.026 0.136 0.271 Group- by-Percent Body Fat (Officer) Nonobese (<25X) Obese (>25X) 0.66 ( . 2 1 0 ) 04,.4 0.44 (0.17,1.12) 004 .8 0.072 Group- by-Percent Body Fat (Enlisted Flyer) Nonobese (<25%) Obese (>25X) 1.36 (0.73,2.55) 1.83 ( . 6 4 3 ) 07,.8 0.177 0.332 Group-by-Percent Body Fat (Enlisted Grounder ew) Nonobese (<25Z) Obese (>25Z) 1.20 (0.79,1.83) 0.91 (0.39,2.10) 0.818 0.390 Peripheral Pulses (Manual) Group-by-occupation Officer Enlisted Flyer Enlisted Groundcrev 0.64 ( . 2 0 9 ) 04,.6 1.45 (0.82,2.56) 1.16 ( . 0 1 6 ) 08,.9 000 .3 0.204 0.441 All Pulses (Manual) Group-by-Occupation Officer Enlisted Flyer Enlisted Groundcrev 0.64 (0.42,0.96) 1.45 (0.82,2.57) 1.14 (0.78,1.66) 0.030 0.204 0.490 3; �TABLE H-7. Unadjusted Exposure Index Analyses for Reported axl \ferified Heart niappgg by Occupation Variable Low Nunber Percent Exposure Index Medium Number Percent High Number Percent n Abnormal Normal 119 31 88 2. 60 74.0 122 33 89 27.0 73.0 109 32 77 29.4 70.6 Overall M vs. L H vs. L 1.05 ( . 9 1 8 ) 05,.6 1.18 ( . 6 2 1 ) 06,.1 080 .5 087 .5 0.575 Enlisted Flyer n Abnormal Normal 54 14 40 25.9 74.1 56 14 42 25.0 75.0 53 18 35 34.0 6. 60 Overall M vs. L H vs. L 0 9 (.024) . 5 04,.8 1.47 ( . 4 3 3 ) 06,.8 0.525 0.912 0.363 Enlisted Groundcrev n Abnormal Normal 148 33 115 22.3 77.7 151 42 109 27.8 72.2 130 30 100 23.1 76.9 Overall M vs. L H vs. L 1.34 ( . 9 2 2 ) 07,.7 1.05 ( . 0 1 8 ) 06,.3 040 .9 0.271 0.881 Officer Verified Essential Hypertension Statistic Officer Reported Essential Hypertension Contrast Est. Relative Risk ( 5 C I ) p-Value 9% . . Occupation n Abnormal Normal 119 23 % 19.3 80.7 122 22 100 18.0 82.0 109 27 82 2. 48 75.2 Overall M vs. L H vs. L 0.92 ( . 8 1 7 ) 04,.6 1.37 ( . 3 2 5 ) 07,.8 0.413 075 .9 0.322 Enlisted Flyer n Abnormal Normal 54 11 43 20.4 7. 96 56 8 48 14.3 85.7 53 12 41 22.6 77.4 Overall M vs. L 0.65 ( . 4 1 7 ) 02,.7 H vs. L 1.14 ( . 5 2 8 ) 04,.8 0.514 041 .0 0.772 Enlisted Groundcrev n Abnormal Normal 148 28 120 18.9 81.1 151 38 113 25.2 7. 48 130 26 104 2. 00 8. 00 Overall M vs. L H vs. L 0.375 0.194 088 .1 1.44 ( . 3 2 5 ) 08,.0 1.07 ( . 9 1 9 ) 05,.4 �TABLE H-7. (continued) Unadjusted Exposure Index Analyses for Reported and Verified Heart Disease by Occupation Exposure Index Low Number Percent Mgjjiin High. Number Percent Number Percent Contrast 109 • 28.4 31 78 71.6 Overall M vs. L H vs. L n Abnormal Normal 119 41 78 34.4 65.6 122 40 82 32.8 67.2 Enlisted Flyer n Abnormal Normal 54 15 39 27.8 72.2 56 19 37 33.9 66.1 53 11 42 n Abnormal Normal 148 42 106 28.4 71.6 151 25 126 16.6 83.4 Officer Verified Heart Disease Statistic Enlisted Groundcrew Reported Heart Disease Occupation Officer Variable n Abnormal Normal 119 37 82 31.1 68.9 122 35 87 Enlisted Flyer n Abnormal Normal 54 14 40 25.9 74.1 Enlisted Groundcrew n Abnormal Normal 148 37 'ill 25.0 75.0 Est. Relative Risk ( 5 C.I.) p-Value 9% 0.93 ( . 4 1 5 ) 05,.8 0.76 ( . 3 1 3 ) 04,.3 0.607 077 .8 0.332 2. 08 79.2 Overall M vs. L 1.34 ( . 9 3 0 ) 05,.1 H vs. L 0 6 ( . 8 1 6 ) .8 02,.6 0.306 044 .8 041 .0 130 41 89 31.5 68.5 Overall M vs. L 0 5 ( . 9 0 8 ) . 0 02,.7 H vs. L 1.16 ( . 0 1 9 ) 07,.5 008 .0 005 .1 059 .6 28.7 71.3 109 28 81 25.7 74.3 Overall M vs. L 0 8 ( . 1 1 5 ) .9 05,.5 H vs. L 0 7 ( . 3 1 3 ) .7 04,.7 065 .6 0.682 0.368 56 16 40 28.6 71.4 53 9 44 17.0 8. 30 Overall M vs. L 1.14 ( . 9 2 6 ) 04,.5 H vs. L 0.58 ( . 3 1 5 ) 02,.0 0.335 0.757 023 .6 151 16 135 10.6 89.4 130 32 98 24.6 75.4 Overall M vs. L 0.36 ( . 9 0 6 ) 01,.7 H vs. L 0 9 ( . 7 1 7 ) .8 05,.6 002 .0 001 .0 0.994 �TKBLE M-7. (continued) Unadjusted Exposure Index Analyses for Reported aid Verified Heart Disease by Occupation Occupation Statistic Officer Variable n Abnormal Normal Low Number Percent Exposure Index Median Number Percent Higi Number Percent Contrast Est. Relative Risk ( 5 C I ) p-Value 9% .. 08 . 99.2 122 4 118 3.3 96.7 109 1 106 09 . 99.1 Overall M vs. L 4 0 ( . 4 3 . ) . 0 04,63 H vs. L 1 0 ( . 7 1 . ) . 9 00,77 026 .5 0.219 092 .5 54 0 00 . 54 100.0 56 1 55 08 . 98.2 53 4 49 7.6 92.4 Overall M vs. L H vs. L 001 .6 — — 3.1 96.9 Overall M vs. L 0 6 ( . 1 3 9 ) . 5 01,.4 H vs. L 1 5 ( . 4 6 9 ) . 3 03,.8 119 1 118 n Abnormal Normal n Abnormal Normal 148 3 145 2.0 9. 80 151 2 149 1.3 98.7 130 4 126 Officer n Abnormal Normal 119 0 00 . 119 1 0 0 0. 122 4 118 3.3 96.7 109 0 00 . 109 100.0 Enlisted Flyer n Abnormal Normal 54 0 00 . 54 1 0 0 0. 56 1 55 1.8 98.2 53 1 52 1.9 98.1 Overall Mvs. L — H vs. L — Enlisted Groundcrev Verified Myocardial Infarction Enlisted Flyer Enlisted Groundcrev Reported Myocardial Infarction n Abnormal Normal 151 00 . 0 151 100.0 130 1 129 08 . 99.2 Overall M vs. L H vs. L 148 2 146 — Analysis not performed due to sparse cells. 1.4 98.6 — — Overall Mvs. L — H vs. L — — — 052 .9 068 .3 052 .8 — — - — — ___ — — �TABLE M-8. Unadjusted Exposure Index Analyses for Central Cardiac Function Variables ty Occupation Exposure Index Low Number Percent Medium Number Percent HisJi Number Percent n Abnormal Normal 119 8 111 6.7 93.3 122 8 114 6.6 93.4 109 7 102 64 . 93.6 Overall M vs. L 0.97 ( . 5 2 6 ) 03,.9 H vs. L 0.95 ( . 4 2 6 ) 03,.3 09 .% 090 .6 098 .2 Enlisted Flyer n Abnormal Normal 54 4 50 7.4 92.6 56 4 52 7.1 92.9 53 3 50 5.7 94.3 Overall M vs. L 0 % ( . 3 4 0 ) . 02,.5 H vs. L 0.75 ( . 6 3 5 ) 01,.3 097 .2 090 .6 0.719 n Abnormal Normal 148 5 143 3.4 96.6 151 13 138 86 . 91.4 130 8 122 6.2 93.8 Overall M vs. L H vs. L 2.69 ( . 4 7 7 ) 09,.6 1.88 ( . 0 5 8 ) 06,.8 016 .6 006 .6 020 .8 Officer Heart Sounds Statistic Enlisted Groundcrev Systolic Blood Pressure n Abnormal Normal 119 4 115 3.4 96.6 122 4 118 3.3 96.7 109 5 104 46 . 95.4 Overall H vs. L 0.97 ( . 4 3 9 ) 02,.9 H vs. L 1.38 ( . 6 5 2 ) 03,.9 084 .4a 098 .6 068 .3 Enlisted Flyer n Abnormal Normal 54 1 53 1.8 98.2 56 2 54 3.6 96.4 Enlisted Groundcrew n Abnormal Normal 148 6 142 40 . 96.0 150 2 148 1.3 98.7 53 0 00 . 53 1 0 0 0. 130 7 123 5.4 94.6 Contrast Est. Relative Risk ( 5 C I ) p-Value 9%.. Occupation Officer Variable — — Overall M vs. L 0.32 ( . 6 1 6 ) 00,.1 H vs. L 1.35 ( . 4 4 1 ) 04,.2 — — 0.167 018 .6 0.603 �TABLE H-8. (continued) Ubadjusted Exposure Index Analyses for Central Cardiac Ruction Variables by Occupation Exposure Index Medium Number Percent n Abnormal Normal 119 13 ; 106 10.9 89.1 122 21 101 17.2 82.8 109 14 95 12.8 87.2 Overall M vs. L H vs. L 1.70 ( . 1 3 5 ) 08,.7 1.20 ( . 4 2 6 ) 05,.9 037 .4 0.165 063 .5 Enlisted Flyer n Abnormal Normal 54 9 45 16.7 83.3 56 13 43 23.2 76.8 53 9 44 17.0 8. 30 Overall M vs. L H vs. L 1.51 ( . 9 3 9 ) 05,.0 1.02 ( . 7 2 8 ) 03,.2 064 .1 0.395 098 .6 n Abnormal Normal 148 17 131 11.5 88.5 151 10 141 6.6 93.4 130 15 115 11.5 88.5 Overall M vs. L 0.55 ( . 4 1 2 ) 02,.4 H vs. L 1.01 ( . 8 2 1 ) 04,.0 026 .6 0.147 092 .9 Officer Right Bundle Branch Block Statistic Enlisted Groundcrew EGG (Overall) Occupation Officer Variable Low Number Percent n Abnormal Normal 119 0 00 . 119 100.0 122 1 121 08 . 99.2 109 0 00 . 109 100.0 — 53 0 00 . 53 100.0 — Enlisted Flyer n Abnormal Normal 54 1 53 1.8 98.2 56 00 . 0 56 1 0 0 0. Enlisted Groundcrew n Abnormal Normal 148 1 147 0.7 99.3 151 0 00 . 151 100.0 High Number Percent 130 2 128 Contrast Est. Relative Risk ( 5 C I ) p-Value 92.. — — — 1.5 98.5 — �TABLE M-8. (continued) Uiadjusted Exposure Index Analyses for Central Cardiac Ruction Variables Variable Low Number Percent Exposure Index Median Number Percent High Number Percent Statistic Officer Left Bundle Branch Block Occupation n Abnormal Normal 119 00 . 0 0. 119 1 0 0 . 122 00 . 0 122 1 0 0 0. n Abnormal Normal 54 00 . 0 0. 54 1 0 0 56 00 . 0 0. 56 1 0 0 53 0 00 . 53 1 0 0 0. Est. Relative Risk ( 5 C I ) p-Value 9% . . 109 0 00 . 109 1 0 0 0. Enlisted Flyer Contrast — — 4 Enlisted Groundcrew 148 0 00 . 143 1 0 0 0. 151 0 00 . 151 1 0 0 0. 130 0 00 . 130 100=0 Officer Nonspecific Sr-THave Changes n Abnormal Normal n Abnormal Normal 119 8 111 6.7 93.3 122 15 107 12.3 87.7 109 11 96 10.1 8. 99 Overall M vs. L H vs. L Enlisted Flyer n Abnormal Normal 54 4 50 7.4 92.6 56 11 45 19.6 8. 04 53 7 46 13.2 8. 68 Overall M vs. L 3.06 ( . 1 1 . 8 09,02) H vs. L 1.90 ( . 2 6 9 ) 05,.2 0.171 0.072 0.327 Enlisted Groundcrew n Abnormal Normal 148 15 133 10.1 89.9 151 7 144 46 . 95.4 130 7 123 5.4 94.6 Overall M vs. L 0.43 ( . 7 1 0 ) 01,.9 H vs. L 0.50 ( . 0 1 2 ) 02,.8 016 .2 005 .7 0.150 1.94 ( . 9 4 7 ) 07,.7 1.56 ( . 0 4 0 ) 06,.3 030 .4 0.147 033 .6 �T&BLE M-8. (continued) Unadjusted Exposure Index .Analyses for Central Cardiac Function Variables by Peculation Exposure Index Variable law Number Percent Medium Number Percent High Number Percent Contrast Est. Relative Risk (952C.I.) p-Value 119 7 112 5.9 94.1 122 8 114 66 . 93.4 109 7 102 6.4 93.6 Overall M vs. L 1.12 ( . 9 3 2 ) 03,.0 H vs. L 1.10 ( . 7 3 2 ) 03,-4 095 .7 086 .2 085 .6 Enlisted Flyer n Abnormal Normal 54 1 53 1.8 98.2 56 5 51 8.9 91.1 53 2 51 3.8 96.2 Overall M vs. L H vs. L 0.205 0.139 0.555 n Abnormal Normal 148 9 139 6.1 93.9 151 2 . 1.3 149 98.7 130 4 126 3.1 96.9 Overall M vs. L 0.21 ( . 4 0 9 ) 00,.8 H vs. L 0.49 ( . 5 1 6 ) 01,.3 Officer Tachycardia n Abnormal Normal Enlisted Groundcrew G Statistic Officer Bradycardia Occupation n Abnormal Normal 119 0 00 . 119 100.0 122 0 00 . 122 100.0 109 0 00 . 109 100.0 — — — — Enlisted Flyer n Abnormal Normal 54 0 00 . 54 100.0 56 0 00 . 56 100.0 53 00 . 0 53 100.0 — — — — Enlisted Groundcrew n Abnormal Normal 148 0 00 . 148 100.0 151 00 . 0 151 100.0 130 00 . 0 130 100.0 5.20 ( . 9 4 . 4 05,60) 2.08 ( . 8 2 . 3 01,36) 008 .7 007 .4 026 .4 _ — — — — �TAHE H-8. (continiad) unadjusted Exposure Index Analyses for Central Cardiac Ruction Variables by Occupatic Low Number Percent Exposure Index Medium Number Percent High Nunber Percent Est. Relative Risk (95%C.I.) p-Value n Abnormal Normal 119 14 105 11.8 88.2 122 21 101 17.2 8. 28 109 12 97 11.0 8. 90 Overall M vs. L 1.56 ( . 5 3 2 ) 07,.3 H vs. L 0 9 ( . 1 2 1 ) .3 04,.0 0.311 0.234 087 .5 Enlisted Flyer n Abnormal Normal 54 9 45 16.7 83.3 56 5 51 89 . 91.1 53 4 49 7.6 92.4 M vs. L 0 4 ( . 3 1 5 ) .9 01,.7 01,.2 H vs. L 0.42 ( . 2 1 4 ) 0.265 0.230 0.159 n Abnormal Normal 148 13 135 8.8 91.2 151 9 142 60 . 9. 40 130 10 120 7.7 92.3 M vs. L 0 6 ( . 7 1 5 ) .6 02,.9 H vs. L 0 8 ( . 7 2 0 ) .7 03,.4 065 .4 0.352 0.741 Officer n Abnormal Normal 119 6 113 5.0 9. 50 122 3 119 2.5 97.5 109 1 108 09 . 99.1 Overall .7 01,.4 M vs. L 0 4 ( . 2 1 9 ) H vs. L 0.17 ( . 2 1 4 ) 00,.7 016 .6a 028 .9 0.110 Enlisted Flyer n Abnormal Normal 54 2 52 3.7 96.3 56 0 00 . 56 1 0 0 0. 53 1 52 1.9 98.1 Enlisted Groundcrew Arrhythmia Statistic Enlisted Groundcrew Other BOG Diagnoses Occupation Officer Variable n Abnormal Normal 143 7 141 4.7 95.3 130 7 123 5.4 94.6 a 151 4 147 Overall analysis; sparse cells, chi-square test may not be valid. — Analysis not performed due to sparse cells. 2.6 97.4 Contrast — — Overall M vs. L H vs. L 0.55 ( . 6 1 9 ) 01,.1 1.15 ( . 9 3 3 ) 03,.6 — — 042 .8 037 .4 052 .8 �TABLE H-9. 4 Unadjusted Exposure Index Analyses for Diastolic Blood Pressure, Fuoduscopic Abnormalities, Carotid Bruits, and Manual Pulse Readings by Occupation Exposure Index Lew Number Percent Medium Number Percent High Number Percent Est. Relative Risk (95%C.I.) p-Value n Abnormal Normal 119 8 111 6.7 93.3 122 7 115 5.7 94.3 109 10 99 9.2 9. 08 Overall .4 03,.1 M vs. L 0 8 ( . 0 2 4 ) H vs. L 1.40 ( . 3 3 6 ) 05,.9 0.585 079 .4 047 .9 Enlisted Flyer n Abnormal Normal 54 5 49 9.3 90.7 56 5 51 89 . 91.1 53 6 47 11.3 88.7 Overall M vs. L H vs. L 0 % (.635) . 02,.3 1.25 ( . 6 4 3 ) 03,.8 093 .0 0.952 0.726 n Abnormal Normal 148 11 137 7.4 92.6 151 17 134 11.3 88.7 130 10 120 7.7 92.3 Overall M vs. L H vs. L 1.58 ( . 1 3 5 ) 07,.0 1.04 ( . 3 2 5 ) 04,.3 044 .3 0.258 096 .3 Officer Funduscopic Examination Statistic Misted Groundcrew Diastolic Blood Pressure Occupation Officer Variable n Abnormal Normal 119 2 117 1.7 98.3 122 0 00 . 122 1 0 0 0. 1.8 98.2 — — 0.337a — — Enlisted Flyer n Abnormal Normal 54 1 53 1.8 98.2 56 0 00 . 56 1 0 0 0. 53 0 00 . 53 IflO.O — — Enlisted Groundcrew n Abnormal Normal 148 1 147 — _ — 0.7 99.3 150 Q 150 00 . 100 0. 109 2 107 130 1 129 Contrast Overall 08 . 99.2 �TABLE K-9. Continued) Unadjusted Exposure Index Analyses for Diastolic Blood Pressure, Funduscopic Abnormalities, Carotid Bruits, and Manual Pulse Readings by Occupation Variable Occupation Statistic Officer n Abnormal Normal Low Number Percent 120 3 117 1.7 96.3 Exposure Index Medium Number Percent 122 1 121 08 . 99.2 High Number Percent 109 0 00 . 109 1 0 0 0. Contrast Est. Relative Risk ( 5 C.I.) p-Value 9% Overall — — 038 .8a — — n Abnormal Normal 54 0 00 . 0. 54 1 0 0 56 0 00 . 56 1 0 0 0. 53 2 51 3.8 96.2 — — n Abnormal Normal 148 0 00 . 0. 148 1 0 0 150 0 00 . 150 100.0 130 2 128 — — _ 1.5 98.5 — — — — Officer Radial Pulses Enlisted Flyer Enlisted Groundcrev Carotid Bruits n Abnormal Normal 118 2 116 1.7 98.3 119 0 00 . 119 1 0 0 0. 108 2 106 — — 0.342a — — .. . ,,_ Overall 1.8 98.2 Enlisted Flyer n Abnormal Normal 54 00 . 0 0. 54 1 0 0 55 0 00 . 55 1 0 0 0. 53 0 00 . 53 1 0 0 0. Enlisted Groundcrev n Abnormal Normal 146 0 00 . 146 100.0 149 0 00 . 149 1 0 0 0. 127 00 . 0 127 1 0 0 0. — — — — _ — — — — �TAKE M-9. (continued) Unadjusted Exposure Index Analyses foe Dutstolic Blood Pressure, Funduscopic AbnoociLities, Carotid Bruits, and Manual Pulse Readings by Occupation Exposure Index Variable Low timber Percent n Abnormal Normal 118 1 117 Enlisted Flyer n Abnormal Normal Enlisted Groundcrev Officer . Popliteal Pulses (Manual) Statistic Officer Femoral Pulses (Manual) Occupation Medium Number Percent 0.8 99.2 119 1 118 54 1 53 1.8 98.2 55 2 53 n Abnormal Normal 146 ' 5 141 3.4 96.6 n Abnormal Normal 118 0 00 . 118 1 0 0 0. Bilisted Flyer n Abnormal Normal 54 2 52 Enlisted Groundcrev n Abnormal Normal 146 4 142 3.7 96.3 2.7 97.3 Higjh Number Percent Contrast Overall 08 . 99.2 108 3 105 2.8 97.2 3.6 96.4 53 4 49 7.6 92.4 149 0 00 . 149 1 0 0 0. 127 3 124 2.4 97.6 119 1 118 08 . 99.2 108 1 107 09 . 99.1 55 0 00 . 55 1 0 0 0. 53 3 50 149 2 147 1.3 98.7 127 3 124 Est. Relative Risk ( 5 C I )p-Value 9% .. — — 0.334 Overall M vs. L 2 0 ( . 8 2 . ) 0.575 . 0 01,27 H vs. L 4 3 ( . 7 3 . ) 0.197 . 3 04,98 Overall H vs. L — 6 ( . 6 2 9 ) 0 8 01,.1 . — — Overall 5.7 94.3 2.4 97.6 0.379a — — — — Overall M vs. L 0 4 ( . 9 2 6 ) . 8 00,.8 H vs. L 0 8 ( . 9 3 9 ) .6 01,.2 008 .8 063 .— 0 — — 0.224a — — 0.692a 047 .0 081 .4 �TABLE M-9. (continued) Unadjusted Exposure Index Analyses foe Diastolic Blood Pressure, Funduscopic Abnormalities, Carotid Bruits, and Manual Pulse Readings by Occupation Low Number Percent Exposure Index Medium Number Percent High Number Percent Est. Relative Risk (95%C.I.) p-Value Statistic n Abnormal Normal 118 15 103 12.7 87.3 119 13 106 10.9 89.1 108 8 100 7.4 92.6 Overall 03,.6 M vs. L 0.84 ( . 8 1 8 ) H vs. L 0.55 ( . 2 3 3 ) 02,.5 048 .1 067 .6 0.194 Enlisted Flyer n Abnormal Normal 54 6 48 11.1 88.9 55 9 46 16.4 83.6 53 8 45 15.1 8. 49 Overall M vs. L H vs. L 1.57 ( . 2 4 7 ) 05,.5 1.42 ( . 6 4 4 ) 04,.2 0.716 040 .3 0.542 Enlisted Groundcrew Dorsalis Pedis Pulses (Manual) Occupation Officer Variable n Abnormal Normal 146 17 129 11.6 88.4 149 12 137 80 . 9. 20 127 14 113 11.0 8. 90 Overall M vs. L 0.66 ( . 1 1 4 ) 03,.5 H vs. L 0.94 ( . 4 1 9 ) 04,.9 0.555 0.303 081 .4 n Abnormal Normal 118 1 117 08 . 99.2 119 2 117 1.7 98.3 108 0 00 . 108 1 0 0 0. Enlisted Flyer n Abnormal Normal 54 3 51 5.6 9. 44 55 3 52 Enlisted Groundcrew n Abnormal Normal 146 4 142 2.7 97.3 149 5 144 Officer Posterior Tibial Pulses (Manual) 5.4 94.6 53 4 49 3.4 96.6 127 5 122 Contrast — — — — 7.6 92.4 Overall M vs. L 0.98 ( . 9 5 0 ) 01,.9 H vs. L 1.39 ( . 0 6 5 ) 03,.2 0.879a 094 .8 064 .7 3.9 96.1 Overall M vs. L H vs. L 089 .5 0.757 0.582 1.23 ( . 2 4 6 ) 03,.8 1.45 ( . 8 5 5 ) 03,.4 �TABLE M-9. (continued) unadjusted Exposure Index Analyses for Diastolic Blood Pressure* Funduscopic Abnomalities, Carotid Bruits, and Manual Pulse BRad-jiggf" by Occupation Exposure Index Low Nunber Percent HpfHim High Nunber Percent Nunber Percent Est. Relative Risk ( 5 C I ) p-Value 9% . . n Abnormal Normal 118 15 103 12.7 87.3 119 14 105 11.8 88.2 108 12 96 11.1 8. 90 Overall M vs. L H vs. L 0 9 (.219) . 2 04,.9 08 (.819) .6 03,.3 0.932 086 .2 071 .1 Enlisted Flyer n Abnormal Normal 54 9 45 16.7 83.3 55 11 44 2. 00 8. 00 53 11 42 20.8 79.2 Overall M vs. L H vs. L 1.25 ( . 7 3 3 ) 04,.1 1.31 ( . 9 3 4 ) 04,.8 088 .4 063 .5 059 .8 n Abnormal Normal 146 23 123 15.8 84.2 149 18 131 12.1 87.9 127 18 109 14.2 85.8 Overall M vs. L H vs. L 0.73 ( . 8 1 4 ) 03,.3 08 (.517) .8 04,.2 069 .5 0.363 0.719 Officer Peripheral Pulses (Manual) Statistic Enlisted Groundcrev All Leg Pulses (Manual) Occupation Officer Variable n Abnormal Normal 118 16 102 13.6 86.4 119 14 105 11.8 88.2 108 13 95 12.0 8. 80 Overall M vs. L 0 8 ( . 9 1 8 ) .5 03,.3 H vs. L 0.87 ( . 0 1 9 ) 04,.1 094 .0 067 .9 0.734 Enlisted Flyer n Abnormal Normal 54 9 45 16.7 83.3 55 11 44 2. 00 8. 00 53 11 42 20.8 79.2 Overall M vs. L H vs. L 1.25 ( . 7 3 3 ) 04,.1 1.31 ( . 9 3 4 ) 04,.8 088 .4 063 .5 059 .8 Enlisted Groundcrev n Abnormal Normal 146 23 123 15.8 84.2 149 18 131 12.1 87.9 127 18 109 14.2 8. 58 Overall M vs. L H vs. L 0.73 ( . 8 1 4 ) 03,.3 08 (.517) .8 04,.2 069 .5 0.363 0.719 Contrast �TABLE H-9. (continued) Unadjusted Exposure Index Analyses for Diastolic Blood Pressure, Fuoduscopic Abnornalities, Carotid Bruits, aid Manual Pulse Readings by Occupation Variable Low timber Percent Exposure Index Medium Number Percent High Number Percent Statistic Officer n Abnormal Normal 118 16 102 13.6 8. 64 119 14 105 11.8 88.2 106 13 95 12.0 8. 80 M vs. L 0 8 ( . 9 1 8 ) .5 03,.3 B vs. L 0 8 ( . 0 1 9 ) .7 04,.1 094 .0 067 .9 0.734 Misted Flyer n Abnormal Normal 54 9 45 16.7 83.3 55 11 44 2. 00 8. 00 53 11 42 20.8 79.2 M vs. L H vs. L 1.25 ( . 7 3 3 ) 04,.1 1.31 ( . 9 3 4 ) 04,.8 088 .4 0.653 059 .8 Enlisted Groundcrev All Pulses (Manual) n Abnormal Normal 146 23 123 15.8 84.2 149 18 131 12.1 87.9 127 18 109 14.2 8. 58 M vs. L 0.73 ( . 8 1 4 ) 03,.3 H vs. L 0 8 ( . 5 1 7 ) .8 04,.2 069 .5 0.363 079 .1 a Overall analysis; sparse cells, chi-square test may not be valid. — Analysis not performed due to sparse cells. Contrast Est. Relative Risk ( 5 C.I.) p-Value 9% Occupation �TABLE H-10. Unadjusted Exposure Index Analyses for Peripheral Vascular System Popple** Pulse BparKr^gg by Occupation Occupation Statistic Officer Variable n Abnormal Normal Exposure Index Medium Number Percent High Number Percent 08 . 99.2 122 0 00 . 122 100.0 109 0 0.0 109 100.0 — — 54 0 00 . 54 100.0 56 0 00 . 56 100.0 53 0 0.0 53 100.0 — — — — Low Number Percent 119 I 118 Est. Relative Contrast Risk ( 5 C.I.) 9% p-Value — — n Abnormal Normal n Abnormal Normal 148 1 147 0.7 99.3 151 0 00 . 151 100.0 130 I 129 0.8 99.2 — — — — Officer Femoral Pulses Enlisted Flyer Enlisted Groundcrew Radial Pulses n Abnormal Normal 119 1 118 08 . 99.2 122 0 00 . 122 100.0 109 0 0.0 109 100.0 — — — — — — _ — — Enlisted Flyer n Abnormal Normal Enlisted Groundcrew n Abnormal Normal 54 0 00 . 54 100.0 148 1 147 0.7 99.3 56 1 55 1.8 98.2 53 1 52 151 0 00 . 151 100.0 130 2 128 .— .,_ 1.9 98.1 Overall 1.5 98.5 — — 034 .0a — — �TABLE MHO. (continued) Unadjusted Exposure Index Analyses for Peripheral Vascular Systen Doppler Pulse Readings by Occupation Occupation Statistic Officer Variable n Abnormal Normal Low Number Percent Exposure Index Medium Number Percent Higfc timber Percent 08 . 99.2 122 2 120 1.6 98.4 54 0 00 . 54 100.0 56 2 54 3.6 96.4 53 1 52 Contrast 1.9 98.1 119 1 118 109 00 . 0 109 100.0 Etet. Relative ) Risk (95% C.I. p-Value — — — — — — — — 0.123* — — n Abnormal Normal n Abnormal Normal 148 1 147 0.7 99.3 151 00 . 0 151 100.0 130 3 127 2.3 97.7 Officer Dorsalis Pedis Pulses Enlisted Flyer Enlisted Groundcrew Popliteal Pulses n Abnormal Normal 119 33 86 27.7 72.3 121 35 86 28.9 71.1 109 30 79 27.5 72.5 Overall M vs. L H vs. L 097 .6 1.06 ( . 0 1 8 ) 0 8 1 06,. 6 .4 .6 0.99 ( . 5 1 77) 0 9 8 05,. Enlisted Flyer n Abnormal Normal 54 11 43 20.4 79.6 56 13 43 23.2 76.8 52 13 39 25.0 75.0 Overall M vs. L H vs. L 088 .4 1.18 ( . 8 2 93) 0 7 9 04,. .1 1.30 ( . 2 3 25) 0 5 9 .6 05,. Enlisted Groundcrew n Abnormal Normal 147 32 115 21.8 78.2 151 32 119 21.2 78.8 129 29 100 22.5 77.5 Overall M vs. L H vs. L 097 .6 0.97 ( . 6 1 68) 0 9 4 .0 05,. 1.04 ( . 9 1 ,84) 0 8 9 .8 05,. . Overall — — �TABLE M-1D. (continued) Unadjusted Exposure Index Analyses for Vascular Systen Doppler Pulse BptytiT^gg by Qccufiaticn Posterior Tibial Pulses Occupation Statistic Officer Variable n Abnormal Normal Low Number Percent Exposure Index Medium Number Percent 1.7 98.3 121 5 116 54 0 00 . 54 100.0 56 3 53 119 2 117 High. Number Percent 4.1 95.9 109 2 107 5.4 9. 46 52 2 50 Contrast Est. Relative Risk ( 5 C I ) p-Value 9% . . Overall M vs. L H vs. L 040 .1 2.52 ( . 8 1 . ) 0 2 6 04,33 .7 1.09 ( . 5 7 9 ) 0 9 8 01,.0 .2 3.8 96.2 1.8 98.2 — — 0.248a — — 0.193* — — — — n Abnormal Normal Enlisted Groundcrew n Abnormal Normal 147 2 145 1.4 98.6 151 0 00 . 151 1 0 0 0. 130 3 127 2.3 97.7 Officer Leg Pulses Enlisted Flyer n Abnormal Normal 119 33 86 27.7 72.3 121 38 83 31.4 68.6 109 32 77 29.4 70.6 Overall M vs. L H vs. L 082 .2 1.19 ( . 8 2 0 ) 0.535 06,.8 1.08 ( . 1 1 9 ) 0 7 7 06,.3 .8 Enlisted Flyer n Abnormal Normal 54 11 43 20.4 79.6 56 15 41 26.8 73.2 52 14 38 26.9 73.1 Overall M vs. L H vs. L 066 .6 1.43 ( . 8 3 4 ) 0 4 0 05,.8 .3 1.44 ( . 8 3 5 ) 0 4 0 05,.5 .3 Enlisted Groundcrew n Abnormal Normal 147 33 114 22.4 77.8 151 32 119 21.2 78.8 129 29 100 22.5 77.5 Overall M vs. L H vs. L 095 .5 09 (.416) 075 .3 05,.1 .9 1.00 ( . 7 1 7 ) 1.000 05,.7 Overall Overall �TAHE tt-10. (cmtinued) ttiadjusted Exposure Index Analyses for Peripheral Vascular Systen Doppler Pulse Readings by Occupation Low Variable Number Percent Exposure Index Medium Number Percent High Number Percent Statistic Officer n Abnormal Normal 119 33 86 27.7 72.3 121 38 83 31.4 68.6 109 32 77 29.4 70.6 Overall M vs. L H vs. L 082 .2 1.19 ( . 8 2 0 ) 0.535 . 06,.8 1.08 ( . 1 1 9 ) 0 7 7 06,.3 .8 Enlisted Flyer n Abnormal Normal 54 11 43 20.4 79.6 56 15 41 26.8 73.2 52 14 38 26.9 73.1 Overall M vs. L H vs. L 066 .6 1.43 ( . 8 3 4 ) 0 4 0 05,.8 . 3 1.44 ( . 8 3 5 ) 0 4 0 05,.5 .3 Enlisted Groundcrew n Abnormal Normal 147 34 113 23.1 76.9 151 32 119 21.2 78.8 129 30 99 23.3 76.7 Overall M vs. L H vs. L 084 .9 08 (.215) 069 .9 05,.4 .8 1.01 ( . 8 1 7 ) 6 9 4 05,.6 .8 Officer n Abnormal Normal 119 33 86 27.7 72.3 121 38 83 31.4 68.6 109 32 77 29.4 70.6 Overall M vs. L H vs. L 082 .2 1.19 ( . 8 2 0 ) 0 5 5 06,.8 . 3 1.08 ( . 1 1 9 ) 0 7 7 06,.3 . 8 Enlisted Flyer n Abnormal Normal 54 11 43 20.4 79.6 56 15 41 26.8 73.2 52 14 38 26.9 73.1 Overall M vs. L H vs. L 066 .6 1.43 ( . 8 3 4 ) 0 4 0 05,.8 . 3 1.44 ( . 8 3 5 ) 0 4 0 05,.5 . 3 Enlisted Groundcrew Peripheral Pulses n Abnormal Normal 147 34 113 23.1 76.9 151 32 119 21.2 78.8 129 30 99 23.3 76.7 Overall M vs. L H vs. L 084 .9 08 (.215) 069 .9 05,.4 .8 1.01 ( . 3 1 7 ) 0 9 4 03,.6 .8 _ All Pulses a Overall analysis; sparse cells, chi-square test may not be valid. — Analysis not performed due to sparse cells. Contrast Est. Relative Risk (95%C.I.) p-Value Occupation �TABLE H-ll. Association Between Central and Peripheral Abnormalities and Verified Heart Disease in the Combined Ranch Hand and Comparison Groups Variable Stratification Total .Hypertension Essential, Percent Percent p-Value Abnormal Abnormal Heart Disease* Percent p-Value Abnormal Myocardial Infarction p-Value Systolic Pressure 2,002 145 17.5 61.4 <0.001 21.7 24.1 0.556 1.0 2.1 0.386 Diastolic Pressure s i NJ <140 XL40 <90 >90 1,966 180 14.8 81.7 <0.001 22.4 15.6 0.043 1.1 0.6 0.790 ECG Normal Abnormal 1,858 290 18.8 30.7 <0.001 17.4 50.3 <0.00i 0.3 5.9 <0.001 Heart Sounds Normal Abnormal 2,084 63 20.1 31.8 0.036 21.4 34.9 0.017 0.9 6.4 <0.001 Funduscopic Examination Normal Abnormal 2,134 13 20.2 53.8 0.008 21.6 53.8 0.014 1.0 0.0 0.999 Carotid Bruits Normal Abnormal 2,132 14 20.3 42.9 0.080 21.8 28.6 0.775 1.0 0.0 0.999 Peripheral Pulses (Manual) Normal Abnormal 1,806 314 20.0 21.7 0.547 21.5 22.6 0.725 0.9 1.6 0.391 Peripheral Pulses (Doppler) Normal Abnormal 1,611 529 20.6 19.8 0.754 20.4 26.1 0.007 0.9 1.5 0.306 Ul *Excluding hypertension. �TABLE M-12. Unadjusted Analyses for Reported'and Verified Heart Disease by Group (Original Comparisons Only) Group Variable Statistic Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk (95% C.I.) p-Value Reported Essential Hypertension 942 247 695 26.2 73.8 888 231 657 26.0 74.0 1.01 (0.82,1.24) 0.920 Verified Essential Hypertension a: to n Yes No n Yes No 942 195 747 20.7 79.3 888 184 704 20.7 79.3 1.00 (0.80,1.25) 0.992 Reported Heart Disease (Excluding Hy pe r t ens ion) n Yes No 942 265 677 28.1 71.9 888 234 654 26.4 73.6 1.09 (0.89,1.34) 0.390 Verified Heart Disease (Excluding Hypertension) n Yes No 942 224 718 23.8 76.2 888 180 708 20.3 79.7 1.23 (0.98,1.53) 0.070 Reported Myocardial Infarction n Yes No 942 20 922 2.1 97.9 888 18 870 2.0 98.0 1.05 (0.55,1.99) 0.889 Verified Hyocardial Infarction n Yes No 942 9 933 1.0 99.0 888 11 877 1.2 98.8 0.77 (0.32,1.87) 0.562 �TABLE M-13. Adjusted Analyses for Reported and Verified Heart Disease (Original Comparisons Only) Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks* Reported Essential Hypertension 1.12 (0.90,1.40) 0.312 AGE (p<0.001), CSMOK (p<0.001), CHOL (p<0.001), %BFAT (p<0.001), ALC (p=0.003) Verified Essential Hypertension 1.10 (0.86,1.39) 0.447 CSMOK (p=0.015), CHOL (p<0.001), %BFAT (p<0.001), PS (p=0.024), ALC (p=0.028) Reported Heart 1.14 (0.92,1.40) Disease 0.230 AGE (p<0.001), RACE (p=0.046) Verified Heart 1.28 (1.02,1.60) Disease 0.032 AGE (p<0.001) Reported Myocardial Infarction 1.16 (0.58,2.32) 0.682 AGE (p<0.001), OCC (p=0.005), CHOL (p»0.040) Verified Myocardial Infarction 0.90 (0.36,2.24) 0.810 AGE (p<0.001), CHOL (p=0.021) Variable Abbreviations; CSMOK: Current smoking CHOL: Cholesterol 2BFAT: Percent body fat ALC: Current alcohol use (drinks/day) PS: Personality score OCC: Occupation M-27 �TABLE H-14. Unadjusted Analyses for Central Cardiac Function by Group (Diabetics Excluded) (Original Comparisons Only) Group Variable Statistic Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk (95% C.I.) p-Value Systolic Blood Pressure n Abnormal Normal 942 60 882 6.4 93.6 888 67 821 7.6 92.4 0.83 (0.58,1.20) 0.322 Heart Sounds n Abnormal Normal 941 31 910 3.3 96.7 888 27 861 3.0 97.0 1.09 (0.64,1.84) 0.757 n Abnormal Normal 942 121 821 12.8 87.2 888 125 763 14.1 85.9 0.90 (0.69,1.18) 0.441 n Abnormal Normal 942 5 937 0.5 99.5 888 6 882 0.7 99.3 0.79 (0.24,2.58) 0.689 n Abnormal Normal 942 0 942 0.0 100.0 888 0 888 0.0 100.0 — — n Abnormal Normal 942 85 857 9.0 91.0 888 78 810 8.8 91.2 1.03 (0.75,1.42) 0.857 i CO ECG (Overall) ECG: RBBB ECG: LBBB . ECG: Nonspecific T-Have Changes �TABLE M-14. (continued) Unadjusted Analyses for Central Cardiac Function by Group (Diabetics Excluded) (Original Comparisons Only) Group Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk (95% C.I.) p- Value 1.12 (0.72,1.75) 0.610 Variable Statistic ECG: Bradycardia n Abnormal Normal 942 45 897 4.8 95.2 888 38 850 4.3 95.7 n Abnormal Normal 942 0 942 0.0 100.0 888 0 888 0.0 100.0 n Abnormal Normal 942 31 911 3.3 96.7 888 32 856 3.6 96.4 0.91 (0.55,1.50) 0.711 n Abnormal Normal 942 97 845 10.3 89.7 888 92 796 10.4 89.6 0.99 (0.74,1.34) 0.968 ECG: Tachycardia KJ SO ECG: Arrhythmia ECG: Other Diagnoses * —No relative risk, given, since no abnormals are present. — — �TABLE M-15. Adjusted Analyses for Central Cardiac Function (Diabetics Excluded)* (Original Comparisons Only) Adj. Relative Risk (95% C.I.) p-Value Systolic Blood 0.95 (0.65,1.38) Pressure (Discrete) 0.772 Variable Covariate Remarks** AGE (p<0.001) CHOL (p=0.001) %BFAT (p<0.001) PS (p=0.001) Systolic Blood -0.152 (-1.401,1.097)* 0.811" Pressure (Continuous) AGE*OCC (p=0.018) CSMOK (p<0.001) CHOL (p<0.001) %BFAT (p<0.001) PS (p=0.001) ALC (p<0.001) Heart Sounds AGE (p<0.001) RACE (p=0.002) OCC (p»0.048) CHOL/HDL (p<0.001) ECG (Overall) ECG: 1.26 (0.73,2.17) **** 0.407 **** AGE (p<0.001) RACE (p=0.023) %BFAT (p<0.001) GRP*PACKYR (p=0.034) RBBB 0.85 (0.26,2.81) 0.787 AGE (p»0.015) ECG: Nonspecific ST-T-Wave Changes 1.20 (0.85,1.69) 0.294 AGE (p<0.001) RACE (p=0.033) CHOL (p-0.009) JSBFAT (p<0.001) ECG: Bradycardia 1.16 (0.74,1.82) 0.509 OCC (p=0.019) CHOL/HDL (p<0.001) M-30 �TABLE M-15. (continued) Adjusted Analyses for Central Cardiac Function (Diabetics Excluded)* (Original Comparisons Only) Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks** EGG: Arrhythmia **** **** AGE (p<0.001) OCC (p<0.001) GRP*PACKYR (p=0.034) GRP*%BFAT (p=0.016) ECG: Other Diagnoses **** **** AGE (p<0.001) GRP*OCC (p=0.036) Variable *Some adjusted analyses did not explore effects of all covariates due to sparse number of abnormalities (see text). **Additional Abbreviations; CHOL/HDL: cholesterol to HDL ratio GRP: group PACKYR: pack-years of smoking ****Group-by-covariate interaction—relative risk/difference in group means, 95% confidence interval, and p-value not presented. "Difference in group means (Ranch Hand-Original Comparison) and associated p-value given, rather than relative risk, for continuous analysis of dependent variables. M-31 �TABLE M-16. Summary of Group-by-Covariate Interactions for Central Cardiac Function Variables (Diabetics Excluded) (Original Comparisons Only) Variable Interaction Stratification Adj. Relative Risk (95% C.I.) p-Value ECG (Overall) 0.76 (0.53,1.10) 1.24 (0.87,1.78) 0.142 0.234 Group-by-Pack- Years Smoking (21% Body Fat) Pack- Years: 0 Pack- Years: 30 0.59 (0.30,1.18) 1.55 (0.72,3.33) 0.134 0.263 Group-by-Percen t Body Fat (7 Pack- Years Smoking) to Pack- Years: 0 Pack- Years: 30 ECG: Arrhythmia 3C Group-by-Pack- Years Smoking 10% Body Fat 30% Body Fat 0.17 (0.05,0.62) 2.46 (0.76,7.99) 0.007 0.136 Officer Enlisted Flyer Enlisted Groundcrew 1.73 (1.06,2.85) 0.75 (0.38,1.47) 0.73 (0.44,1.22) 0.030 0.401 0.234 ECG: Other Diagnoses Group-by-Occupation �TABLE M-17. Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) (Original Comparisons Only) Group Variable Statistic Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk (95% C.I.) p- Value Diastolic Blood Pressure 942 79 863 8.4 91.6 887 76 811 8.6 91.4 0.98 (0.70,1.36) 0.889 Funduscopic Examination 3. i 0 0 u> n Abnormal Normal n Abnormal Normal 941 7 934 0.7 99.3 888 6 882 0.7 99.3 1.10 (0.37,3.29) 0.865 Carotid Bruits n Abnormal Normal 941 7 934 0.7 99.3 887 6 881 0.7 99.3 1.10 (0.37,3.28) 0.865 Radial Pulses (Manual) n Abnormal Normal 929 4 925 0.4 99.6 876 7 869 0.8 99.2 0.54 ( . 6 1 8 ) 01,.4 0.322 Radial Pulses (Doppler) n Abnormal Normal 942 3 939 0.3 99.7 886 4 882 0.4 99.6 0.70 (0.16,3.16) 0.646 Femoral Pulses (Manual) n Abnormal Normal 929 20 909 2.2 97.8 876 21 855 2.4 97.6 0.90 (0.48,1.66) 0.726 �TABLE M-17. (continued) Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) (Original Comparisons Only) Group Variable Statistic Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk (95% C.I.) p-Value Femoral Pulses (Doppler) n Abnormal Normal 942 6 936 0.6 99.4 887 3 884 0.3 99.7 1.89 (0.47,7.57) 0.368 Popliteal Pulses (Manual) n Abnormal Normal 929 16 913 1.7 98.3 876 21 855 2.4 97.6 0.71 (0.37,1.38) 0.312 Popliteal Pulses (Doppler) n Abnormal Normal 942 10 932 1.1 98.9 887 6 881 0.7 99.3 1.58 (0.57,4.35) 0.379 n Dorsalis Pedis Pulses Abnormal (Manual) Normal 929 102 827 11.0 89.0 876 95 781 10.8 89.2 1.01 (0.71,1.45) 0.928 n Dorsalis Pedis Pulses Abnormal (Doppler) Normal 938 228 710 24.3 75.7 885 203 682 22.9 77.1 1.08 (0.87,1.34) 0.490 Posterior n Tibial Pulses Abnormal (Manual) Normal 929 27 902 2.9 97.1 876 24 852 2.7 97.3 1.06 (0.61,1.85) 0.834 Posterior n Tibial Pulses Abnormal (Doppler) Normal 939 19 920 2.0 98.0 885 18 867 2.0 98.0 0.99 (0.52,1.91) 0.984 �TABLE M-17. (continued) Unadjusted Analyses for Peripheral Vascular Function by Group (Diabetics Excluded) (Original Comparisons Only) Group Variable Statistic Ranch Hand Number Percent Original Comparison Number Percent Est. Relative Risk (95* C.I.) p-Value Leg Pulses (Manual) n Abnormal Normal 929 131 798 14.1 85.9 876 134 742 15.3 84.7 0.91 (0.70,1.18) 0.472 Leg Pulses (Doppler) n Abnormal Normal 938 237 701 25.3 74.7 885 212 673 23.9 76.1 1.07 (0.87,1.33) 0.516 Peripheral Pulses (Manual) n Abnormal Normal 929 133 796 14.3 85.7 876 138 738 15.8 84.2 0.89 (0.69,1.16) 0.395 Peripheral Pulses (Doppler) n Abnormal Normal 938 239 699 25.5 74.5 885 214 671 24.2 75.8 1.07 (0.87,1.33) 0.522 All Pulses (Manual) n Abnormal Normal 929 133 796 14.3 85.7 876 138 738 15.8 84.2 0.89 (0.69,1.16) 0.395 All Pulses (Doppler) n Abnormal Normal 938 239 699 25.5 74.5 884 214 670 24.2 75.8 1.07 (0.86,1.32) 0.529 Ul �TABLE H-1B. Adjusted Analyses for Peripheral Vascular Ruction (Diabetics Excluded)* (Original Comparisons Only) Variable Statistical/Clinical Analysis Adj. Relative Risk (95% C.I.) p-Value Covariate Remarks** Discrete 1.04 ( . 4 1.46) 07, 0.818 CHOL ( < . 0 ) p001 %BFAT (p«O.C01) Continuous 0.250 (-0.530, 1.030)a 0.537* AGE (p=0.016) CSMOK ( < . 0 ) p001 CHOL ( < . 0 ) p001 %BFAT (p<0.001) ALC ( = . 0 ) p004 Rmduscopic Examination 1.31 (0.42,4.09) 0.646 AGE ( < . 0 ) p001 RACE ( < . 0 ) p001 CHOL/HDL ( = . 1 ) p000 ALC (p=0.032) Carotid Bruits 0.95 ( . 0 2 9 ) 03,.7 098 .2 AGE (p=0.018) DRK2R ( < . 0 ) p001 Diastolic Blood Pressure Radial Pulses Manual Doppler 0.57 ( . 7 1 % 01,.) 07 (.631) .0 01,.6b 0.373 066 .4b AGE (p=0.047) Femoral Pulses Manual 1.20 ( . 0 2 3 ) 06,.6 0.710 AGE (p<0.001) OCC (p=0.050) CHOL/HDL (p=0.043) %BFAT (p<0.001) DIFOCRT (p=O.C05) Doppler 1.87 (0.45,7.74) 0.384 AGE (p=0.001) CSMCK (p=0.001) CHOL/HDL (p=0.032) Popliteal Pulses AGE (p=O.C02) RACE (p*Q.C02) PACm (p=0.006) CHOL/HDL (p=Q.021) GRP*OCC (p=0.048) rCrfrfft Doppler Dorsalis Pedis Pulses 1.66 (0.59,4.72) 0.337 AGE (p<D.C01) RACE (p=0.015) CSMOK (p<0.001) Manual 1.06 (0.78,1.44) 0.704 AGE (p=O.C04) Doppler 1.07 ( . 6 1 3 ) 08,.3 0.569 AGE ( = . 3 ) p006 RACE ( 4 . 0 ) p)07 %BFAT ( = . 1 ) p000 M-36 �TABLE M-18. (continued) Adjusted Analyses for Peripheral Vascular Function (Diabetics Excluded)* (Original Comparisons Chly) Variable Posterior Tibial Pulses Statistical/Clinical Analysis Manual Doppler Leg Pulses **** **** AGE (p=0.003) RACE (p<D.001) PACKXR ( = . 1 ) p001' GRP*OCC (p=0.026) 0.849 AGE (p<0.001) RACE (p<0.001) CSMCK (p=0.010) CHOL/HDL (p=0.002) **** AGE (p=O.C01) GRP*OCC (p=0.038) GRP**BFAT 1.07 (0.54,2.13) 1.07 ( . 6 1 3 ) 08,.3 1.06 ( . 6 1 3 ) 08,.2 0.575 **** Manual Doppler 0.516 **** Manual Doppler All Pulses p-Value Manual Doppler Peripheral Pulses Adj. Relative Risk (95% C.I.) 1.06 ( . 6 1 3 ) 08,.2 0.582 Covariate Remarks** AGE (p=0.016) RACE (p=0.024) 2BFAT (p=0.034) AGE (pO.OOl) GRP*%BFAT (p=0.036) AGE (p=£.028) RACE (p=0.042) 2BFAT AGE (p<0.001) GRP*%BFAT (p=0.036) AGE (p=0.027) RACE (p=0.042) %BFAT (p=0.038) *Some adjusted analyses did not explore effects of all covariates due to sparse number of abnormalities (see text). **Additional Abbreviations; DRKBR: Drink-years of alcohol DIFCORT: Differential cortisol "Difference in group means (Ranch Hand - Original Comparison) and associated p-value given, rather than relative risk, continuous analysis of dependent variables. Unadjusted for any covariates—same as for unadjusted results. —No covariates significant. ****Group-by-covariate interaction—relative risk, 95% confidence interval, and p-value not presented. M-37 �TABLE H-19. Summary of Group-by-Covariate Interactions for Peripheral Vascular Function Variables (Diabetics Excluded) (Original Comparisons Only) Variable Interaction Stratification Adj. Relative Risk (95% C.I.) p-Value Popliteal Pulses (Manual) Officer Enlisted Flyer Enlisted Groundcrev 0.21 (0.02,2.54) 4.68 (0.53,41.1) 0.99 (0.38,2.55) 0.219 0.165 0.976 Posterior Tibial Pulses (Manual) Group-by-Occupation Officer Enlisted Flyer Enlisted Groundcrew 0.31 (0.05,1.82) 3.68 (0.96,14.1) 1.27 (0.56,2.87) 0.197 0.057 0.575 Leg Pulses (Manual) 3: i Group-by-Occupation Group-by-occupation (21% Body Fat) Officer Enlisted Flyer Enlisted Groundcrew 0.65 (0.41,1.01) 1.52 (0.83,2.77) 1.23 (0.81,1.87) 0.054 0.174 0.327 Peripheral Pulses (Manual) Group-by-Percent Body Fat 10% Body Fat 30% Body Fat 1.87 (0.94,3.72) 0.58 (0.34,0.98) 0.073 0.042 All Pulses (Manual) Group-by-percent Body Fat 10% Body Fat 30% Body Fat 1.87 (0.94,3.72) 0.58 (0.34,0.98) 0.073 0.042 CO �TABLE M-20. Longitudinal Analysis of Pulse Index and Overall EGG: A Contrast of Baseline and First Follovup Examination Abnormalities (Original Comparisons Only) Variable Group 1982 Baseline Exam 1985 Followup Exam Odds* Ratio p-Value (OR) (ORRH vs ORQC) Abnormal Normal Ranch Hand Abnormal Normal 50 104 72 743 1.44 Original Comparison Abnormal Normal 35 115 51 668 2.25 Ranch Hand Abnormal Normal 86 43 192 650 0.22 Original Comparison All Pulses (Manual) Abnormal Normal 83 44 147 598 0.30 0.05 ECG • (Overall) 0.23 Number Normal Baseline, Abnormal Follovup *OHH<3 Rat-In? . Number Abnormal Baseline, Normal Followup M-39 �APPENDIX N Hematological Evaluation �APPENDIX N: Hematological Evaluation Contents Table N-l Summary Statistics for Hematological Covariates by Group • N-l N-2 Summary of Group-by-Covariate Interactions for Hematological Variables N-2 N-3 Interaction Summaries for Adjusted Continuous Exposure Index Analyses for Hematological Variables N-4 N-4 Unadjusted Categorical Analyses for Hematological Variables by Group (Original Comparisons Only) N-6 N-5 Unadjusted Continuous Analyses for Hematological Variables (Contrast of Group Means) (Original Comparisons Only) N-7 N-6 Adjusted Categorical Analyses for Hematological Variables (Abnormal Versus Normal, Adjusted for Age, Race, Occupation, and Smoking) (Original Comparisons Only) N-8 N-7 Adjusted Continuous Analyses for Hematological Variables (Ranch Hand-Original Comparison Group Differences) (Original Comparisons Only) N-9 N-8 Summary of Group-by-Covariate Interactions for Hematological Variables (Original Comparisons Only) N-10 N-9 Longitudinal Analyses for MCV, MCH, and PLT: A Contrast of Baseline and First Followup Examination Test Means (Original Comparisons Only) N-15 N-i �TABLE N-l. Summary Statistics for Hematological Covariates by Group Group Covariate Covariate Category Ranch Hand Comparison Percent Percent Race Black Nonblack 5.9 94 .1 6.4 93.6 0.679 Occupation Officer Enlisted Flyer Enlisted Groundcrew 37 .4 17 .5 45 .2 37.5 16.2 46.3 0.724 Mean±SE Mean±SE p-Value Age (At Baseline) — 43.94±0.24 43.86±0.22 0.804 Current Smoking* — 10.75±0.48 9.13±0.40 Smoking History (Pack-Years) — 13.24±0.52 12.98±0.47 0.708 "Equivalent cigarettes per day. —Covariate not categorized for these results. N-l 0.010 �TABLE N-2. Summary of Group-by-Covariate Interactions for Heoatological Variables RBC Occupation Smoking History Ranch Hand-Comparison Group Difference ± SB p-Value - . 0 + 0.030 009 -0.014 ± 0.035 0.775 0.686 -0.066 ± 0.053 -0.001 ± 0 0 6 .4 0.212 0.975 Enlisted Groundcrew Nonsmoker 30 Pack-Years Variable Interaction Covariates -0.010 + 0.029 -.9 ± 006 003 .3 0.736 000 .1 Nonblack Nonsmoker 35 53 -0.010 + 0.020 0 0 2 ± 0.021 .0 069 .0 0.942 0.115 + 0.031 -0.017 ± 0.019 <0.001 0.369 RH>C -0.550 + 0.184 -0.143 ± 0.186 0.003 0.446 ORH Enlisted Flyer 35 53 -0.341 ± 0.172 0.067 ± 0.134 0.050 0.618 ORH Enlisted Groundcrew 35 53 -0.121 ± 0.074 0.286 ± 0.149 0.104 0.057 Stratification Direction of Result* Officer Nonsmoker 30 Pack- Years Enlisted Flyer Nonsmoker 30 Pack- Years z: i to VBC* Race Smoking History ORH Occupation Age (At Baseline) 30 Pack- Years 35 53 Black Officer 35 53 RH>C �TABLE N-2. (continued) Summary of Group-by-Covariate Interactions for Hematological Variables Variable PLTa Interaction Covariates Race Stratification Ranch Hand-Comparison Group Difference ± SE Direction p-Value of Result* 0.003 ± 0.012 0.789 0.103 ± 0.042 0.014 Black Nonsmoker 30 Pack-Years, Currently 1 Pack/Day Smoking (Smoking History and Current Smoking) Nonblack Nonsmoker 30 Pack-Years, Currently 1 Pack/Day -0.050 ± 0.050 0.317 0.258 ± 0.096 0.007 Variable mean in Ranch Hand group greater than variable mean in Comparison group. ORE: Variable mean in Comparison group greater than variable mean in Ranch Hand group. a Units are on log scale. RH>C RH>C �TABLE N-3. Interaction Summaries for Adjusted Continuous Exposure Index Analyses for Hematological Variables . Interaction Variable (Occupation) Stratification Contrast Exposure Index Level Difference ± SE p-Value Direction of Result* HGB HGB HCT 0.014 + 0.295 -0.180 ± 0.372 0.962 0.630 M vs. L H vs. L -0.047 + 0.402 0.516 ± 0.502 0.908 0.304 M vs. L H vs. L -0.119 + 0.128 0.261 ± 0.151 0.356 0.086 H > L M vs. L H vs. L 0.033 + 0.102 -0.172 ± 0.102 0.748 0.093 L > H Nonblack M vs. L H vs. L -0.007 + 0.122 0.060 ± 0.123 0.955 0.626 Black RBC M vs. L H vs. L Black RBC Nonblack M vs. L H vs. L 0.089 + 0.874 2.227 ± 0.960 0.919 0.021 Exposure Index35 by-Age (At Baseline) (Enlisted Flyer) 53 M vs. L H vs. L -0.854 + 0.389 0.626 ± 0.457 0.029 0.172 M vs. L H vs. L 0.425 + 0.308 -0.339 ± 0.308 0.169 0.272 Exposure Indexby-Race (Officer) Nonblack M vs. L H vs. L -0.093 + 0.359 0.206 ± 0.362 0.795 0.569 Black M vs. L -0.849 + 2.569 5.611 ± 2.823 0.741 0'4 .08 Exposure Indexby-Race (Officer) Exposure Index35 by-Age (At Baseline) (Enlisted Flyer) 53 Exposure Indexby-Race (Officer) H vs. L H > L L > M H > L �TABLE N-3. (continued) Interaction Summaries for Adjusted Continuous Exposure Index Analyses for Hematological Variables Variable Stratification Contrast Exposure Index Level Difference ± SE p- Value Exposure Index 35 by-Age (At Baseline) (Enlisted Flyer) 53 PLT* PUT* M vs. L H vs. L -2.332 ± 1.141 1.921 ± 1.341 0.042 0.154 M vs. L H vs. L 1.074 + 0.904 -0.887 ± 0 9 4 .0 Exposure Indexby -Current Smoking (Packs per Day) (Enlisted Flyer) 0 M vs. L H vs. L 0.097 ± 0.054 -0.005 ± 0.053 0.075 0.932 1 M vs. L H vs. L -0.015 ± 0.042 -0.062 ± 0.045 0.710 0.169 M vs. L H vs. L -0.128 ± 0 0 8 .6 -0.118 ± 0.078 0.064 0.132 Nonblack M vs. L H vs. L -0.015 ± 0.025 -0.004 ± 0.026 0.557 0.885 Black M vs. L H vs. L 0.156 ± 0.079 -0.036 ± 0.081 0.049 0.659 Exposure Indexby-Race (Enlisted Groundcrew) Variable Variable Variable Variable Units are in log *H L L M > > > > L: H: M: L: Direction of Result* 0.236 0.328 2 HCT z i Interaction (Occupation) mean in mean in mean in mean in scale. L >M M > L L >H M > L high exposure index category greater than variable mean in low exposure index category. low exposure index category greater than variable mean in high exposure index category. low exposure index category greater than variable mean in medium exposure index category medium exposure index category greater than variable mean in low exposure index category �TABLE N-4. Unadjusted Categorical Analyses for Hanatological Variables by Group (Original Comparisons Only) Abnormally Low Variable Group RBC WBC HGB HOT MCV MCH MCHC PIT Normal Number Percent Number Percent Abnormally High Number Percent Total p-Valuea Ranch Hand 30 Original Comparison 36 3.0 976 96.2 8 0.8 1,014 3.8 910 95.4 8 0.8 954 Ranch Hand 45 Original Comparison 47 4.4 906 89.4 62 6.1 1,013 4.9 857 89.8 50 5.2 954 3.8 752 74.2 . 223 22.0 1,014 3.6 717 75.2 203 21.3 954 Ranch Hand 11 Original Comparison 13 1.1 1,001 98.7 2 0.2 1,014 1.4 938 98.3 3 0.3 954 Ranch Hand 10 Original Comparison 11 1.0 857 84.5 147 14.5 1,014 1.2 809 84.8 134 14.0 954 Ranch Hand Original Comparison 7 0.7 943 93.0 64 6.3 1,014 7 0.7 890 93.3 57 6.0 954 1 0.1 1,013 99.9 0 0.0 1,014 0 0.0 954 100 0. 0 0.0 954 5 0.5 987 97.4 21 2.1 1,013 2 0.2 935 98.0 17 1.8 954 Ranch Hand 39 Original Comparison 34 Ranch Hand Original Comparison Ranch Hand Original Comparison 0.598 0.631 0.867 0.722 0.904 0.947 0.760 a Chi-square test, 2 d.f., except for HCT and PLT, which were obtained from continuity adjusted chi-square tests on 1 d.f. (Abnormally high category pooled with normal, and abnormally low category pooled with normal for HCT and PLT, respectively.) —Only one abnormal MCHC value; p-value not given. N-6 �TABLE N-5. Unadjusted Continuous Analyses for Hematological Variables (Contrast of Group Means) (Original Comparisons Only) Group Mean±SE Variable Ranch Hand Original Comparison Difference ±SE RBC 4.964±0.012 4.974±0.012 WBC1 7.003 6.844 t-Statistic p-Value -0.010±0.017 -0.61 0.540 1.78 0.076 0.50 0.619 HGB 15.624±0.033 15.600±0.034 — 0.024±0.047 HCT 45.904±0.097 45.874±0.097 0.030±0.137 0.22 0.827 MCV 92.596±0.150 92.355+0.157 0.241±0.217 1.11 0.268 MCH 31.544±0.055 31.435±0.058 0.108±0.080 1.35 0.177 MCHC 34.040±0.021 34.009±0.613 0.030±0.029 1.06 0.291 1.39 0.165 PLT1 265.2 261.7 — 1 Means transformed from log scale. —Difference and standard errors (SE) not presented, since variables were analyzed on logarithmic scale. N-7 �TABLE N-6. Adjusted Categorical Analyses for Hematological Variables (Abnormal Versus Normal, Adjusted for Age, Race, Occupation, and Smoking) (Original Comparisons Only) Abnormally Low vs. Normal Abnormally High vs. Normal Variable Adj. Relative Risk (95% C.I.) p-Value Adj. Relative Risk (95% C.I.) p-Value RBC 0.81 (0.51,1.30) 0.380 0.95 (0.41,2.17) 0.898 WBC 0.95 (0.63,1.43) 0.730 1.16 (0.80,1.70) 0.431 HGB 1.10 (0.70,1.72) 0.692 1.05 (0.85,1.30) 0.662 HCT 0.84 (0.41,1.71)* 0.628a MCV Nonblack Black 1.92 (0.64,5.76) 0.32 (0.09,1.11) 0.242 0.072 1.09 (0.84,1.42) 0.41 (0.12,1.34) 0.503 0.139 MCH 0.96 (0.40,2.30) 0.920 1.06 (0.74,1.52) 0.734 PLT Sparse Data 1.13 (0.63,2.05)b 0.682b "Abnormally low versus normal/abnormally high. 'Abnormally high versus normal/abnormally low. N-8 Sparse Data �TABLE N-7. Adjusted Continuous Analyses for Hematological Variables (Ranch Hand-Original Comparison Group Differences) (Original Comparisons Only) Variable Ranch Hand-Original Comparison Group Difference ±SE p-Value Covariate Remarks* AGE (p<0.001) CSMOK (p=0.001) OCC*PACKYR (p=0.043) RBC -0.018 ± 0.017 0.278 WBC **** **** GRP*RACE*OCC (p<0.001) GRP*AGE*RACE (p=0.013) GRP*AGE*PACKYR (p=0.013) GRP*RACE*PACKYR (p=0.022) PACKYR*CSMOK (p<0.001) HGB **** **** GRP*RACE*AGE (p=0.030) GRP*RACE*OCC (p=0.020) OCC*PACKYR (p=0.023) CSMOK (p<0.001) HCT **** **** GRP*RACE*AGE (p=0.026) GRP*RACE*OCC (p=0.011) AGE*OCC (p=0.009) CSMOK (p<0.001) MCV **** **** GRP*AGE*PACKYR (p=0.041) GRP*AGE*CSMOK (p=0.012) RACE*OCC (p=0.021) AGE*RACE (p<0.001) MCH **** **** GRP*AGE*CSMOK (p=0.026) AGE*RACE (p=0.001) OCC (p=0.001) MCHC 0.030 ± 0.028 0.296 PLT **** **** RACE (p<0.001) GRP*RACE*PACKYR (p=0.011) GRP*AGE (p=0.040) OCC (p=0.005) *CSMOK: current level of smoking (cigarettes per day) PACKYR: smoking history (pack years) OCC: occupation GRP: group ****Group-by-covariate interaction—group difference, standard error (SE), and p-value not presented. N-9 �TABLE N-8. Summary of Group-by-Covariate Interactions for Hematological Variables (Original Comparisons Only) Ranch Hand-Original Comparison Group Difference ±SE p-Value -0.014 + 0.021 0.014 ± 0.022 0.511 0.508 0.107 + 0.033 -0.011 ± 0.019 0.001 0.551 Black -0.053 ± 0.058 (No significant interactions) 0.364 Variable Interaction Covariates Stratification WBCa Nonblack Nonsmoker 35 53 Race Smoking History Age (At Baseline) 30 Pack-Years 35 53 I l-» o Direction of Result* RH>OC �TABLE N-8. (continued) Summary of Group-by-Covariate . Interactions for Hematological Variables (Original Comparisons Only) HGB Interaction Covariates Direction of Result* Race Occupation Ranch Hand-Original Comparison Group Difference +SE p-Value 0.151 ± 0.113 -0.035 ± 0.082 0.181 0.667 Enlisted Flyer 35 53 Variable 0.087 ± 0.132 -0.100 ± 0.121 0.509 0.410 Enlisted Groundcrew 35 53 0.026 + 0 0 0 .8 -0.161 ± 0.107 0.744 0.134 -1.034 + 0.627 0.209 ± 0.629 0.099 0.740 OORH Enlisted Flyer 35 53 -1.398 + 0.613 -0.155 ± 0.481 0.023 0.747 OORH Enlisted Groundcrew 35 53 -0.014 + 0.283 1.230 ± 0.526 0.962 0.020 RH>OC Stratification Nonblack Officer 35 53 Age (At Baseline) z Black Officer 35 . 53 �TABLE N-8. (continued) Summary of Group-by-Covariate Interactions for Hematological Variables (Original Comparisons Only) HCT Interaction Covariates Race Occupation Ranch Hand-Original Comparison Group Difference ±SE p-Value 0.506 + 0.325 -0.174 ± 0.237 0.120 0.462 Enlisted Flyer 35 53 Variable 0.390 + 0.377 -0.291 ± 0.347 0.301 0.403 Enlisted Groundcrew 35 53 0.063 + 0.230 -0.617 ± 0.307 0.783 0.045 -2.662 + 1.801 0.825 ± 1.805 0.140 0.648 Enlisted Flyer 35 53 -4.365 + 1.753 -0.878 ± 1.378 0.013 0.524 OORH Enlisted Groundcrew 35 53 -0.050 + 0.806 3.436 ± 1.497 0.950 0.022 RH>OC Stratification Nonblack Officer 35 53 Direction of Result* Age (At Baseline) as »-» to Black Officer 35 53 OORH �TABLE N-8. (continued) Summary of Group-by-Covariate Interactions for Hematological Variables (Original Comparisons Only) Variable Interaction Covariates MCV MCH p-Value 0.692 + 0.418 -0.136 ± 0.417 0.098 0.743 0.057 ± 0.635 -0.582 ± 0.412 0.928 0.158 Nonsmoker 35 53 0.126 + 0.141 0.158 ± 0.132 0.373 0.232 1 Pack/Day 35 53 0.254 + 0.144 -0.275 ± 0.149 0.079 0.066 RH>OC OORH 2 Pack/ Day 35 53 0.381 + 0.273 -0.707 ± 0.281 0.162 0.012 OORH Smoking Nonsmoker (Smoking History 35 and Current Smoking) 53 Age (At Baseline) Current Smoking Age (At Baseline) Direction of Result* Ranch Hand-Original Comparison Group Difference ±SE Stratification 30 Pack- Years, Currently 1 Pack/Day 35 53 RH>OC �TABLE N-8. (continued) Summary of Group-by-Covariate Interactions for Hematological Variables (Original Comparisons Only) Variable PLTa Interaction Covariates Race Smoking History Stratification Nonblack Nonsmoker 35 53 Ranch Hand-Original Comparison Group Difference ±SE p-Value Direction of Result* 0.030 + 0.016 -0.017 ± 0.018 0.061 0.354 RH>OC 0.050 + 0.020 0.003 ± 0.016 0.012 0.858 RH>OC -0.082 + 0.051 -0.129 ± 0.054 0.111 0.016 OORH 0.261 + 0.103 0.214 ± 0.102 0.011 0.036 RH>OC RH>OC Age (At Baseline) 3Q Pack- Years 35 53 I (-* *- Black Nonsmoker 35 53 30 Pack-Years 35 53 *RH>OC: Variable mean in Ranch Hand group greater than variable mean in Original Comparison group. OORH: Variable mean in Original Comparison group greater than variable mean in Ranch Hand group. a Units are on log scale. �TABLE N-9. Longitudinal Analyses for MCV, MCH, and PLT: A Contrast of Baseline and First Followup Examination Test Means (Original Comparisons Only) Means Variable Group Total Baseline Difference Followup (Followup-Baseline) Error* 88.89 92.60 +3.71 872 88.56 92.35 +3.79 Ranch Hand 971 30.81 31.55 +0.74 872 30.63 31.44 +0.81 Ranch Hand 971 276.90 271.50 -5.4 Original Comparison PLT 971 Original Comparison MCH Ranch Hand Original Comparison MCV 872 272.10 267.30 -4.8 p-Value (Equality of Difference) *Error - HSubj * Time/Group mean squares. N-15 2.808 0.65 0.873 0.21 35.50 0.80 �APPENDIX 0 Renal Assessment �TABLE 0-1. Unadjusted Exposure Index Analyses for Renal Variables by Occupation Occupation Statistic Officer Variable n Number /% Abnormal Normal Low 127 10 7.9% - 117 92.1% Exposure Index Medium High Contrast Est. Relative Risk (95% C.I.) p- Value 130 123 Overall 0.226 11 8.5% 119 91.5% 17 13.8% 106 86.2% M vs. L H vs. L 1.08 (0.44,2.64) 0.999 1.88 (0.82,4.28) 0.155 65 57 Overall 0.165 7 12.7% 48 87.3% 8 12.3% 57 87.7% 2 3.5% 55 96.5% M vs. L H vs. -L 0.96 (0.33 ,2.85) 0.999 0.25 ( . 5 ,1.26) 0.091 00 n Number /% Abnormal Normal 153 163 141 Overall 0.214 18 11.8% 135 88.2% 11 6.7% 152 93.3% 10 7.1% 131 92.9% M vs. L H vs. L 0.54 (0.25 ,1.19) 0.172 0.57 (0.26 ,1.29) 0.233 n Number/% Abnormal Normal 127 130 123 Overall 0.696 Officer Urinary Protein n Number/% Abnormal Normal 55 Enlisted Flyer Enlisted Groundcrev Kidney Disease 4 3.1% 123 96.9% 2 1.5% 3 2.4% 128 98.5% 120 97.6% M vs. L H vs. L 0.48 ( . 9 ,2.67) 0.443 00 0.77 (0.17 ,3.51) 0.999 n Number/% Abnormal Normal 55 65 57 Overall 0.708 Enlisted Flyer 2 3.6% 53 96.4% I 1.5% 64 98.5% 1 1.8% 56 98.2% M vs. L H vs. L 0.41 ( . 4 ,4.69) 0.593 00 0.47 ( . 4 ,5.37) 0.615 00 n Number/% Abnormal Normal 154 163 Overall 0.726 Enlisted Groundcrew 8 5.2% 146 94.8% 9 6.3% 7 4.3% 156 95.7% 133 93.7% M vs. L H vs. L 0.82 (0.29 ,2.32) 0.794 1.24 (0.46 ,3.29) 0.804 142 �TABLE 0-1. (continued) Unadjusted Exposure Index Analyses for Renal Variables by Occupation Variable Occupation Statistic Low Exposure Index Medium High Est. Relative Contrast Risk (95% C.I.) p-Value n Number/% Abnormal Normal 127 130 123 Overall 0.334 17 13.4% 110 86.6% 15 11.5% 115 88.5% 22 17.9% 101 82.1% H vs. L H vs. L 0.84 ( .40,1.77) 0.708 0 1.41 (0.71,2.81) 0.385 n Number/% Abnormal Normal 55 65 57 Overall 0.325 Enlisted Flyer 12 21.8% 43 78.2% 14 21.5% 51 78.5% 7 12.3% 50 87.7% M vs. L H vs. L 0.98 ( .41,2.35) 0.999 0 0.50 ( .18,1.39) 0.214 0 n Number /% Abnormal Normal 154 163 141 Overall 0.144 Enlisted Groundcrev 25 16.2% 129 83.8% 34 20,9% 129 79.1% 36 25.5% 105 74.5% H vs. L H vs. L 1.36 (0.77,2.41) 0.315 1.77 (1.00,3.13) 0.061 n Number/% Abnormal Normal 127 130 123 Overall 0.558 11 8.7% 116 91.3% 7 5.4% 123 94.6% 10 8.1% 113 91.9% M vs. L H vs. L 0.60 ( .23,1.60) 0.337 0 0.93 (0.38,2.28) 0.999 n Number /% Abnormal Normal 55 65 57 Overall 0.763 Enlisted Flyer 7 12.7% 48 87.3% 8 12.3% 57 87.7% 5 8.8% 52 91.2% M vs. L H vs. L 0.96 (0.33,2.85) 0.999 0.66 ( .20,2.22) 0.554 0 n Number /% Abnormal Normal 154 163 142 Overall 0.447 Enlisted Groundcrev 20 13.0% 134 87.0% 15 148 19 13.4% 123 86.6% M vs. L H vs. L 0.68 (0. 33,1.38) 0.290 1.04 (0. 53,2.03) 0.999 Officer Urinary Occult Blood o I Officer Urinary White Blood Cell Count 9.2% 90.8% �TABLE 0-1. (continued) Unadjusted Exposure Index Analyses for Renal Variables by Occupation Exposure Index Statistic Low Medium High Contrast n Mean* 95* C.I.* 127 14.62 (14.05, 15.19) 130 14.82 (14.20, 15.46) 123 14.74 (14.14, 15.35) Overall M vs. L H vs. L 0.891 0.633 0.775 Enlisted Flyer n Mean* 95* C.I.* 55 13.84 (12.79, 14.93) 65 13.80 (12.87, 14.77) Overall 57 13.93 M vs. L (13.15, H vs. L 14.72) 0.983 0.963 0.895 Enlisted Groundcrew n Mean* 95* C.I.* 154 14.08 (13.50, 14.67) 163 13.72 (13.15, 14.30) 142 14.00 (13.48, 14.52) Overall M vs. L H vs. L 0.639 0.389 0.841 Officer n Mean 95% C.I. 127 1.0144 (1.0132, 1.0156) 130 1.0149 (1.0137, 1.0161) Overall 123 1.0150 M vs. L (1.0137, H vs. L 1.0162) 0.786 0.554 0.541 Enlisted Flyer n Mean 95* C.I. 55 1.0159 (1.0139, 1.0178) 65 1.0155 (1.0138, 1.0173) 57 Overall 1.0141 M vs. L (1.0123, H vs. L 1.0158) 0.358 0.803 0.178 Enlisted Groundcrew Blood Urea Nitrogen Occupation Officer Variable n Mean • 95* C.I. 154 1.0167 (1.0156, 1.0178) 163 1.0168 (1.0158, 1.0179) 142 Overall 1.0166 M vs. L (1.0154, H vs. L 1.0179) 0.967 0.885 0.907 o Urine Specific Gravity ^Transformed from square root scale. —No relative risk given for variables analyzed continuously. Risk (95* C.I.) p-Value �TABLE 0-2. Interaction Summaries of Adjusted Exposure Index Analyses for Renal Variables Variable Interaction (Occupation) Stratification Statistic Diabetic Urinary Protein Exposure Index-byDiabetic Impaired flagg (Officer) Normal Nonblack Urinary Protein Exposure Index-byRace (Enlisted Groundcreu) Black Bom XL942 Urinary Occult Blood Exposure Index-byAge (Enlisted Groundcreu) Bom <1942 Lou Exposure Index Medium High Adj. Relative 9% Contrast Risk ( 5 C.I.) p-Value n Nunber/% Abnormal Normal 8 8 14 Overall 000 .1* 3 37.5% 5 62.5% 0 00 .% 8 100 0.% 0 0.0% M vs. L 0.% 14 1 0 0 H vs. L 0.09 ( . 0 , . 6 * 0 2 0 00421)* . 0 * 0.05 ( . 0 , . 2 * 0 0 6 00212)* .3* n Number/% Abnormal Normal 11 n Nunfcer/% Abnormal Normal 106 108 0 0.0% 0.% 108 1 0 0 2 106 n Number/% Abnormal Normal 138 149 4 2.9% 134 97.1% 6 143 n Number/% Abnormal Normal 16 14 4 25.0% 12 75,0% 1 13 n Number/* Abnormal Normal 87 n Number /% Abnormal Normal 67 1 10 11 76 14 53 14 9.1% 90.9% • 12.6% 87.4% 16 Overall 0.247* 0 00 .% 14 1 0 0 0.% 0 0.0% M vs. L 16 1 0 0 H vs. L 0.% 0.24 ( . 1 6 5 ) * 0 4 0 00,.3* .4* 0.21 ( . 0 , . 1 * 0 4 7 00857)* . 0 * Overall 0.190* 3 3.2% M vs. L 90 96.8% H vs. L 5.09 ( . 4 1 7 3 ) * 0 4 8 02,0.6* .9* 8.39 ( . 3 1 4 6 ) * 0 0 7 04,6.1* .9* Overall 0.259* 93 1.9% 98.1% 129 4.0% 96.0% 13 7.1% 92.9% 17.8% 82.2% 11 23 Overall 0.751* 0.158* 002 .9* 0 0.0% M vs. L 0.23 ( . 2 2 3 ) 0 0 , . 7 * 0.336* 0.% 13 1 0 0 H vs. L 0.10 ( . 0 , . 1 * 0.107* 00521)* 20 24.7% H vs. L 61 75.3% H vs. L 1.56 ( . 1 3 4 ) 07,.1 2.32 ( . 3 7 3 ) 07,.2 0.252 046 .6 1.73 ( . 8 4 4 ) 06,.2 1.36 ( . 9 3 1 ) 05,.2 0.265 0.152 60 34 20.9% 79.1% 1.41 ( . 9 5 0 ) 03,.9* 2.51 ( . 5 8 3 ) 07,.7* 81 129 23 106 9 7.0% M vs. L 120 93.0% H vs. L 32.4% 67.6% 16 26.7% M vs. L 44 73.3% H vs. L �TABLE 0-2. (continued) Interaction Summaries of Adjusted Exposure Index Analyses for Renal Variables Variable Interaction (Occupation) Stratification Statistic Low Exposure Index Medium High Adj. Relative Contrast Risk ( 5 C.I.) 9% p-Value Age £ 1 0 Overall 0 0 — 1 100 0.% 0 0.0% 0 0 — M vs. L H vs. L n Hunter/* Abnormal Nonnal 1 1 0 Overall 0.157* Bom >1942 Impaired Exposure Index-by- n Number/% Abnormal Normal 0 Born >1942 Diabetic 0 0.0% 1 100 0.% 1 100 0.% 0 0.0% 0 — 0 — M vs. L H vs. L 9.00 ( . 0 8 1 7 ) * 0 9 9 01,3.8* . 9 * n Number/% Abnormal Normal 10 16 9 Overall 0.065* Bom >1942 Normal 0 0.0% 16 1 0 0 0.% 1 11.1% M vs. L 8 88.9% B vs. L 0.07 ( . 0 , . 2 * 0 0 6 00314)* . 4 * 0.29 ( . 2 3 4 ) * 0.582* 00,.8* n Number/% Abnormal Normal 1 8 Bom <1942 Diabetic 0 0.0% 1 100 0.% 1 7 n Number/% Abnormal Normal 8 6 Bom <1942 Impaired 0 0.0% 8 100 0.% 2 4 n Number/% Abnormal Nonnal 35 33 Bom <1942 Nonnal 4 11.4% 31 88.6% 4 29 Urinary Unite Blood Cell Count ( ) a Exposure Index-byDiabetic Class (Enlisted Flyer) 3 7 30.0% 70.0% 5 12.5% 87.5% Overall 068 .6* 0 0.0% H vs. L 5 1 0 0 H vs. L 0.% 0.60 ( . 2 2 . 7 * 0 9 9 00,30)* . 9 * — — 6 33.3% 66.7% Overall 0.222* 1 16.7% 5 83.3% H vs. L H vs. L 9.44 ( . 7 2 2 1 ) * 0.165* 03,4.8* 4.64 ( . 6 1 5 5 ) * 0 4 9 01,3.7* .2* 37 12.1% 87.9% Overall 3 8.1% M vs. L 34 91.9% H vs. L 085 .0 0.79 ( . 6 3 9 ) 01,.0 0.58 ( . 1 3 0 ) 01,.3 0.776 055 .1 �TABLE 0 2 (continued) -. Interaction Summaries of Adjusted Exposure Index Analyses for Benal Variables Variable Interaction (Occupation) Low Stratification Statistic Exposure Index Medium High Adj. Relative Contrast Risk ( 5 C.I.> ] 9Z p-Value Nonblack Blood Urea Nitrogen Exposure Index-byRace (Officer) n . 125 127 121 M vs. L Adj. Mean(c) 1 . 5 46 14.71 14.55 H vs. L 95XC.I.(c) ( 3 9 , 5 4 ) ( 3 9 , 5 4 ) ( 3 8 , . 7 1.11.1 1.61.9 1.412) (b) (b) 087 .7 0.814 Black n Adj. Mean 95XC.I. (b) (b) 0.035 0.494 2 2 3 M vs. L 16.93 19.51 H vs. L 10.65 (22,23) (.21.9 (44,53) 1.42.8 76,41) 1.52.3 ^unadjusted estimate of relative risk and confidence interval, or p-value, based on stratified tables. **Uhadjusted estimate of relative risk and confidence interval calculated after adding 0.5 to each cell. —Zero counts in cells do not allow for calculation of percent, relative risk, confidence interval, or p-value. (a)Results presented for urinary white blood cell count are based on stratification into the six categories shown. Unadjusted results are presented for all strata except Ranch Bands in normal diabetic class, born before 1 4 . These results have been adjusted for race. 92 (b)No relative risk given for variables analyzed continuously. (c)Transforned from square root scale. Note: Small sample sizes may affect validity of overall p-value. Note: Results without (*) or ( * are adjusted for all other main effects in model (age, race, and diabetic class), unless otherwise noted. *) �TABLE 0-3. Unadjusted Analyses for Renal Variables by Group (Original Comparisons Only) Group Original Comparison Number Percent Variable Kidney Disease n Yes No 1,014 94 920 9.3 90.7 954 94 859 Urinary Protein o •-J Statistic Ranch Hand Number Percent n Abnormal Normal 1,016 37 979 3.6 96.4 Urinary Occult Blood n Abnormal Normal 1,015 182 833 Urinary White Blood Cell Count n Abnormal Normal 1,016 102 914 Blood Urea Nitrogen n Mean* 95% C.I.* 1,016 14 .21 (13.99,14.43) 955 14.37 (14.15,14.59) 0.318 Urine Specific Gravity n Mean 95% C.I. 1,016 1.0157 (1.0153,1.0162) 954 1.0154 (1.0150,1.0159) 0.365 : Est. Relative Risk ( 5 C.I.) 9% p-Value 9.9 90.1 0.93 (0.69,1.26) 0.701 954 28 926 2.9 97.1 1.25 (0.76,2.06) 0.449 17.9 82.1 954 155 799 16.2 83.8 1.13 (0.89,1.43) 0.338 10.0 90.0 954 83 871 8.7 91.3 1.17 (0.86,1.59) 0.316 *Converted from square root scale. —No relative risk given for variables analyzed continuously. �TABLE 0-4. Adjusted Analyses for Renal Variables by Group (Original Comparisons Only) Variable Stratification Statistic Group Ranch Original Hand Comparison Adj. Relative Risk (95% C.I.) Covariate p-Value Remarks* Kidney Disease 1,014 950 0.95 (0.70,1.28) 0.726 AGE (p=0.001) Urinary Protein 0 00 n n 1,016 942 **** **** OCC (p=0.007) AGE*DIAB (p=0.037) GRP*DIAB (p=0.003) Urinary Occult Blood GRP*OCC*RACE (p=0.020) Nonblack n 955 887 1.14 (0.89,1.46) 0.299 AGE (p=0.003) OCC (p<0.001) Black Enlisted n 53 51 1.43 (0.54,3.76) 0.465 GRP*OCC (p=0.051) AGE*DIAB (p=0.027) �TABLE 0-4. (continued) Adjusted Analyses for Renal Variables by Group (Original Comparisons Only) Group Variable Stratification Ranch Hand Statistic Original Comparison Adj. Relative Risk (95* C.I.) p-Value GRP*AGE*RACE (p=0.048) GRP*OCC*RACE (p=0.022) Urinary White Blood Cell Count Nonblack VO n 956 n Black Enlisted o i Covariate Remarks* * 893 **** **** OCC (p=0.010) GRP*DIAB (p=0.054) GRP*AGE-(p=0.053) 53 51 **** **** GRP*OCC (p=0.015) DIAB (p-0.017) GRP*AGE (p=0.010) **** AGE (p<0.001) OCC (p=0.025) RACE*DIAB (p=0.049) GRP*RACE (p=0.077) Blood Urea Nitrogen n Adj . Mean 95* C.I. 1,016 **** **** 952 **** **** • Urine Specific Gravity n Adj. Mean 952 C.I. 1,016 **** **** 951 **** **** **** DIAB (p=0,011) OCC (p<0.001) GRP*RACE (p=0.037) ^Abbreviations: OCC: occupation GRP: group DIAB: diabetic class ****Group-by~covariate interaction—adjusted relative risk/group means, p-value, and confidence interval are not presented. —No relative risk given for variables analyzed continuously. �TABLE 0-5. Summary of Group-by-Covariate Interactions for Renal Variables (Original Comparisons Only) Group Variable Interaction Stratification Statistic Diabetic Urinary Protein o Group-byDiabetic Class Impaired Normal Officer Urinary Occult Blood Group-byOccupation (Black) Enlisted Flyer Enlisted Groundcrev n Number/% Abnormal Normal Ranch Hand n Number/% Abnormal Normal n Number/% Abnormal Normal 7 25 807 0 7 n Number/% Abnormal Normal n Number/% Abnormal Normal 4.7% 95.3% 3.0% 97.0% 14.4% 85.5% 0.57 (0.20,1.60) 0.286 11 125 8.1% 91.9% 0.58 (0.19,1.75) 0.332 6 733 0.8% 99.2% 3.82 (1.56,9.37) 0.004 7 0.0% 100.0% 43 13 30 11 65 739 10 3 7 p-Valn*e 136 832 n Number/% Abnormal Normal 9.0% 91.0% 106 5 101 Adj. Relative Risk (95% C.I.) 76 78 7 •71 Original Comparison 3 4 42.9% 57.1% — 8.3% 91.7% — 12 30.0% 70.0% 1 11 39 30.2% 69.8% 9 30 23 U% 76.9% — — �TABLE 0-5. (continued) Summary of Group-by-Covariate Interactions for Renal Variables (Original Comparisons Only) Group Variable Interaction Adj. Relative Risk (95% C.I.) p-Value 16.7% 83.3% 0.90 (0.38,2.10) 0.799 33 4 29 12.1% 87.9% 4.12 (2.06,8.24) <0.001 11.0% 89.0% 298 13 285 4.4% 95.6% 2.08 (1.16,3.71) 0.014 57 6 51 10.5% 89.5% 57 12 45 21.1% 78.9% 0.38 (0.14,1.04) 0.059 n Abnormal Normal 80 14 66 17.5% 82.5% 89 6 83 6.7% 93.3% 1.77 (0.75,4.17) 0.191 n Abnormal Normal 439 31 408 7.1% 92.9% 404 37 367 9.2% 90.8% 0.89 (0.55,1.44) 0.639 Original Comparison Ranch Hand Stratification Statistic Born >1942 Diabetic n Abnormal Normal 14 I 13 7.1% 92.9% 12 2 10 Born <1942 Impaired n Abnormal Normal 22 2 20 9.1% 90.9% Born >1942; Normal n Abnormal Normal 344 38 306 Born <1942 Diabetic n Abnormal Normal Born <1942 Impaired Born <1942 Normal . Nonblack: Group-byAge o i Urinary White Blood Cell Count Group-byDiabetic Class 4 �TABLE 0-5. (continued) Summary of Group-by-Covariate Interactions for Renal Variables (Original Comparisons Only) Group Variable Group-byOccupation (Black) o i Stratification Statistic Ranch Band Officer Urinary Uhite Blood Cell Count Interaction n Number/% Abnormal Normal 7 Enlisted Flyer Enlisted Groundcrew Born >1942 Urinary Uhite Blood Cell Count Group-byAge (Black Enlisted) Born <1942 0 7 n Number/% Abnormal Normal n Number/% Abnormal Normal n Number/% Abnormal Normal 29 n Number/% Abnormal Normal 0.0% 100.0% 43 24 7 36 4 25 6 18 Adj. Relative Risk (95% C.I.) p-Value 7 10 3 7 Original Comparison 2 5 28.6% 71.4% 12 30.0% 70.0% 0 12 0.0% 100 0.% 39 16.3% 83.7% 7 32 17.9% 82.1% 26 13.8% 86.2% 5 21 19.2% 80.8% 0.65 (0.15,2.78) 0.565 8.0% 92.0% 4.99 (0.79,31.56) 0.088 25 25.0% 75.0% 2 23 �TABLE 0-5. (continued) Summary of Group-by-Covariate Interactions for Renal Variables (Original Comparisons Only) Group Variable Stratification Statistic Nonblack Blood Urea Nitrogen Interaction n Adj. Mean* 95* C.I.* Group-byRace Black o i Nonblack Urine Specific Gravity Group-byRace Black •Ranch Hand 956 14.16 (13.86,14.47) Original Comparison 894 14.22 (13.92,14.53) Adj. Relative Risk (95% C.I.) p-Value 0.726 n Adj. Mean* 95Z C.I.* 60 12.43 (11.41,13.51) 58 13.63 (12.66,14.64) 0.057 n Adj. Mean 95X C.I. 956 1.0163 (1.0157,1.0169) 893 1.0158 (1.0151,1.0164) 0.104 n Adj. Mean 95Z C.I. 60 1.0154 (1.0135,1.0173) 58 1.0177 (1.0158,1.0196) 0.081 —No relative risk given for variables analyzed continuously. Also, no relative risk or p-value given for interactions involving occupation for urinary occult blood and urinary white blood cell count; tables presented only for illustrative purposes. ^Transformed from square root scale. �APPENDIX P Endocrine Assessment �APPENDIX P: Endocrine Assessment Contents Table Page P-l Summary of Group-by-Covariate Interactions for Endocrinological Variables P-l P-2 Unadjusted Exposure Index Analyses for Endocrinological Variables by Occupation P-2 P-3 Interaction Summaries of Adjusted Exposure Index Analyses for Endocrinological Variables P-6 P-4 Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group (Original Comparisons Only)... P-8 P-5 Adjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group (Original Comparisons Only)... P-10 P-6 Summary of Group-by-Covariate Interactions for Endocrinological Variables (Original Comparisons Only) P-12 �TABLE P-l. Summary of Group-by-Covariate Interactions for Endocrinological Variables Group Comparison Differential Group-byCortisol Race-byAge n Mean Ad j . Mean 95% C.I. 4 6 522.6 1044.8 654.4 1042.8 0.012 (484.7,849.5) (803.5,1313.2) 10-25% n Mean Ad j . Mean 95% C.I. 815 627.8 603.3 (583.4,623.5) 1026 605.7 582.4 0.023 (563.9,601.3) n Mean Ad j . Mean 95% C.I. 179 470.5 463.0 (435.0,491.9) 257 470.3 456.7 0.706 (432.9,481.2) Nonblack Born >1942 Testosterone Group-by% Body Fat Ranch Hand >25% Interaction Statistic <10% Variable n Mean Ad j . Mean 95% C.I. 374 2.07 1.58 (0.64,2.51) 500 2.44 1.97 (1.06,2.88) 0.156 Nonblack Born <1942 n Mean Ad j . Mean 95% C.I. 572 2.53 2.08 (1.18,2.97) 697 2.46 1.99 (.028) 11,.8 0.708 Black Born >1942 n Mean Adj. Mean 95% C.I. 32 0.17 -0.46 (-2.11,1.18) 47 2.78 2.33 (0.89,3.76) Black Born <1942 n Mean Adj . Mean 95% C.I. 26 3.24 2.94 (1.22,4.67) 36 2.30 1.89 (0.32,3.46) Stratification p-Value 0.003 0.312 �TABLE P-2. Unadjusted Exposure Index Analyses for Endocrinological Variables by Occupation Low Exposure Index High Medium Contrast n Mean 95% C.I. 126 124 122 28.01 28.35 27.96 (27.66,28.36) (28.00,28.71) (27.61 ,28.32) Overall M vs. L H vs. L 0.250 0.177 0.857 Enlisted Flyer n Mean 95% C.I. 55 65 55 27.59 27.31 27.83 (26.82,27.81) (27.14,28.06) (27.32 ,28.34) Overall M vs. L H vs. L 0.362 0.411 0.155 n Mean 95% C.I. 153 142 161 27.68 27.65 27.53 (27.37,28.00) (27.34,27.95) (27.21 ,27.85) Overall M vs. L H vs. L 0.775 0.865 0.493 Officer •ja Statistic Enlisted Groundcrev T3 % Uptake Occupation Officer Variable n Mean 95% C.I. 126 124 122 1.167 1.278 1.181 (1.089,1.286) (1.076,1.271) (1.173 ,1.397) Overall M vs. L H vs. L 0.283 0.843 0.203 Enlisted Flyer n Mean 95% C.I. 55 65 55 1.153 1.033 1.204 (0.924,1.162) (1.035,1.292) (1.068 ,1.367) Overall M vs. L H vs. L 0.181 0.178 0.075 Enlisted Groundcrev n Mean 95% C.I. 153 161 1.117 1.124 (1.043,1.199) (1.051,1.204) Overall M vs. L B vs. L 0.738 0.899 0.465 NJ TSH 142 1.160 (1.079 ,1.250) p-Value �TABLE P-2. (continued) Unadjusted Exposure Index Analyses for Endocrinological Variables by Occupation Variable Low Exposure Index Medium High n Mean 95% C.I. 125 118 128 599.4 547.4 558.0 (563.8,636.1) (513.8,582.1) (522.7,594.4) Overall M vs. L H vs. L 0.099 0.041 0.112 Enlisted Flyer n Mean 95% C.I. 55 63 588.7 637.4 (531.6,648.6) (581.8,695.6) 57 584.9 (529.1,643.6) Overall M vs. L H vs. L 0.363 0.244 0.929 Enlisted Groundcrew n Mean 95% C.I. 162 139 153 641.4 603.4 609.3 (574.7,644.9) (606.9,676.9) (567.3,640.6) Overall M vs. L H vs. L 0.276 0.205 0.820 Officer *j3 Statistic Officer Testosterone Occupation n • Mean 95% C.I. 124 130 11.97 11.43 (11.33,12.65) (10.83,12.07) 121 12.28 (11.61,12.99) Overall M vs. L H vs. L 0.186 0.239 0.531 Enlisted Flyer n Mean 95% C.I. 57 55 65 11.97 11.08 11.13 (11.09,12.91) (10.33,11.88) (10.33,11.99) Overall M vs. L H vs. L 0.276 0.001 0.092 Enlisted Groundcrew n Mean 95% C.I. 142 154 161 11.50 11.69 11.33 (11.09,12.32) (10.76, 11. 92> (10.89,12.15) Overall M vs. L H vs. L 0.699 0.397 0.678 UJ Initial Cortisol Contrast p-Value �TABLE P-2. (continued) Unadjusted Exposure Index Analyses for Endocrinological Variables by Occupation Low Exposure Index Medium High Contrast Differential Cortisol n Mean 95% C.I. 124 9.27 (8.72,9.86) 130 9.16 (8.63,9.73) 121 9.69 (9.10,10.32) Overall M vs. L H vs. L 0.416 0.789 0.324 Enlisted Flyer n Mean 95% C.I. 55 8.43 (7.67,9.26) 65 9.48 (8.69,10.34) 57 9.06 (8.25,9.94) Overall M vs. L H vs. L 0.200 0.074 0.286 n' Mean 95% C.I. 154 161 9.64 9.29 (9.11,10.19) (8.79,9.81) 142 9.17 (8.65,9.72) Overall M vs. L H vs. L 0.453 0.358 0.229 Officer *a i Statistic Enlisted Groundcrev 2-Hour Cortisol Occupation Officer Variable n Mean 95% C.I. . 124 2.69 (2.02,3.36) 130 2.31 (1.66,2.96) 121 2.59 (1-91,3.26) Overall M vs. L H vs. L 0.709 0.424 0.829 Enlisted Flyer n Mean 95% C.I. 55 3.43 (2.50,4.36) 65 1.20 (0.34,2.06) 57 2.30 (1.39,3.22) Overall M vs. L H vs. L 0.003 <0.001 0.092 Enlisted Groundcrev n Mean 95% C.I. 154 1.96 (1.26,2.66) 161 2.12 (1.44,2.81) 142 2.37 (1.64,3.09) Overall M vs. L H vs. L 0.726 0.740 0.425 p-Value �TABLE P-2. (continued) Unadjusted Exposure Index Analyses for Endocrinological Variables by Occupation Variable Low Exposure Index Medium High •jo Ul Statistic Officer 2-Hour Postprandial Glucose Occupation n Mean 95% C.I. 121 113 124 109.0 107.5 107.6 (104.3,113.8) (103.0,112.3) (102.9,112.6) Overall M vs. L H vs. L 0.892 0.666 0.694 Enlisted Flyer n Mean 95% C.I. 54 62 56 100.9 118.0 110.9 (92.1,110.6) (108.3,128.5) (101.4,121.4) Overall M vs. L H vs. L 0.051 0.015 0.149 Enlisted Groundcrew n Mean 95% C.I. 150 156 140 105.5 106.7 109.1 (100.1,111.1) (101.4,112.4) (103.3,115.2) Overall M vs. L H vs. L 0.674 0.750 0.380 Contrast p-Value �TAHUEP-3. Variable Interaction Stratifi(Occupation) cation <LO% Testosterone Exposure Index-by. % Body Fat (Enlisted Groundcrew) io_25% >25% 10-25% 2-8ax Cbrtisol Exposure Index-by%Body Fat (Officer) >25% • Nonblack 2-Hour Cortisol Exposure Index-byRace (Enlisted Groundcrev) Black Low Statistic Exposure Index Medium High Contrast p-Value n Adj. Mean 95% C.I. 1 2 3 72.7 645 8. M vs. L 888 2. 57818.) 454 958 ( . , 7 . ) ( 3 . , 1 2 5 ( 6 . , 4 . ) H vs. L 03230 O01 .0 O01 .0 n Adj. Mean 95% C.I. 125 128 105 600 3. 646.2 647 4. M vs. L ( 8 . , 7 . ) ( 0 . , 9 . ) ( 9 . , 9 . ) H vs. L 561655 616625 582 629 0.527 055 .8 n Adj. Mean 95% C.I. 27 31 29 501.8 467.5 M vs. L 530 0. ( 3 . , 7 . ) ( 3 . , 7 . ) ( 0 . , 3 . ) H vs. L 439547 476530 407 5 9 6 090 .8 0.471 n Adj. Mean 95% C.I. 102 110 8.31 85 .0 (.095) (.197) 72,.8 74,.6 99 9.05 M vs. L ( . 6 1 . 3 H vs. L 78, 0 4 ) 069 .2 009 .7 n Adj. Mean 95% C.I. 22 20 10.03 7.46 (.71.8 (.291) 82,21) 61,.0 20 M vs. L 8.26 ( . 7 1 . 7 H vs. L 67, 00) 006 .0 008 .6 n 129 138 146 Adj. Mean 9.30 9.32 M vs. L 9.33 95% C.I. ( . 0 1 . 2 ( . 8 1 . 1 ( . 3 1 . 3 H vs. L 83,04) 83,04) 83, 0 4 ) 091 .2 0.961 11 14 8.41 M vs. L 9.36 ( 0 5 , 5 6 ) ( . 1 1 . 2 ( . 8 1 . 8 H vs. L 1.91.7 76,15) 74, 04) 003 .1 002 .0 n 16 Adj. Mean 12.89 95% C.I. �• — —- x t Interaction granaries of Adjusted Exposure Index Analyses for Endocrinological Variables Variable Interaction Stratifi(Occupation) cation Initial Cortisol Exposure Index-by%Body Fat (Officer) Statistic Low Exposure Index Medium Higfr Contrast p-Value n Adj. Mean 95% C.I. 102 110 99 11.51 11.19 12.25 M vs. L ( 0 1 , 3 0 ) ( . 9 1 . 6 , ( 0 8 , 3 9 ) H vs. L 1.31.8 98,26) 1.01.1 0.514 0.154 n Adj. Mean 95% C.I. 22 20 20 12.28 9.59 M vs. L 10.26 ( 0 3 , 4 6 ) ( . 3 1 . 5 ( . 9 1 . 6 M vs. L 1.31.0 80,14) 85,22) 000 .1 000 .6 Nonblack TypeA n Adj. Mean 952: C.I. 49 63 1.94 0% . (07,.6 (.834) -.426) 03,.9 53 2.00 M vs. L ( . 2 3 6 ) H vs. L 03,.8 0.237 0.223 Nonblack TypeB 76 n 89 83 1.36 M vs. L Adj.. Mean 2.29 1.64 95% C.I. ( . 0 3 7 ) ( 0 1 , . 4 ( . 4 3 1 ) H vs. L 08,.8 -.228) 01,.4 0.158 0.336 Black TypeA 5 n 8 4 Adj.. Mean -5.49 3.27 3.10 M vs. L 95% C.I. ( 8 7 , 2 2 ) ( 0 7 , . 5 ( 1 3 , . 9 H vs. L -.0-.8 -.172) -.074) <D.001 001 .0 Black TypeB n 8 9 7 M vs. L Adj,. Mean 1.16 2.99 0.48 95% C.I. ( 2 0 , . 7 ( 0 0 , . 5 ( 2 8 , . 6 H vs. L -.443) -.760) -.037) 0.384 0.762 10-25% Differential Cortisol Exposure Ihdex-bySace-byPersonality Type (Enlisted Groundcrew) >25K �TABLE P-4. Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group (Original Comparisons Only) Group Variable Statistic Ranch Hand Original Comparison T3 % Uptake n Mean • 95% C.I. Number/% Low Normal High 1,003 27.79 (27.67,27.91) 936 27.68 (27.55,27.80) 7 969 27 13 898 25 n Mean 95% C.I. Number/% Normal High 1,003 1.16 (1.13,1.19) n Mean 95% C.I. Number/% Low Normal High 952 1,000 597.31 570.90 (583.97,610.80) (557.69,584.27) TSH 00 Testosterone Initial Cortisol n • Mean 95% C.I. Number/% Low Normal High 996 7 38 949 13 0.7% 96.6% 2.7% 99.3% 0.7% 3.8% 94.9% 1.3% 1,009 11.62 (11.39,11.85) 52 950 7 5.2% 94.2% 0.7% 1.4% 95.9% 2.7% 936 1.11 (1.08,1 -14) 931 5 37 904 11 99.5% 0.5% 3.9% 95.0% 1.2% 949 11.67 (11.44,11.91) 44 898 7 4.6% 94.6% 0.7% Contrast Est. Relative Risk (95% C.I.) p-Value 0.203 Overall Low vs. Normal 0.50 (0.20,1.26) High vs. Normal 1.00 (0.58,1.74) — 1.31 (0.41,4.14) — Overall Low vs. Normal 0.98 (0.62,1.55) High vs. Normal 1.13 ( . 0 2.53) 05, — Overall Low vs. Normal 1.12 (0.74,1.69) High vs. Normal 0.95 (0.33,2.71) 0.322 0.177 0.999 0.040 0.775 006 .0 0.955 0.999 0.839 0.729 0.864 0.603 0.999 �TABLE P-4. (continued) Unadjusted Continuous and Categorical Analyses for Laboratory Endocrinological Variables by Group (Original Comparisons Only) Group Variable Ranch Hand Original Comparison 2-Hour Cortisol n Mean 95% C.I. Number/% Low Normal High 1,009 9.30 (9.10,9.51) 949 9.32 (9.11,9.52) 0 0.0% 1,005 99.6% 4 0.4% 0 946 3 Differential Cortisol n Mean 95% C.I. 1,009 2.30 (2.05,2.55) 949 2.38 (2.12,2.64) 2-Hpur Postprandial Glucose T3 ** Statistic n Mean 95% C.I. Number/% Normal Impaired Diabetic 976 909 107.93 110.03 (105.90,110.01) (108.03,112.07) n Number/% Yes No 1,016 • Diabetes (Composite Indicator) 836 106 34 85.7% 10.9% 3.5% 74 7.3% 942 92.7% 748 137 24 Contrast p-Value 0.924 0.0% 99.7% 0.3% 82.3% 15.1% 2.6% Est. Relative Risk (95% C.I.) 1.26 (0.28,5.62) 0.999 0.692 Overall Impaired vs. Normal Diabetic vs. Normal 0.154 0.69 (0.53,0.91) 1.27 (0.75,2.16) 0.017 0.009 0.423 1.04 (0.74,1.47) 0.861 955 67 888 7.0% 93.0% — No relative risk or confidence interval given for continuous analysis. �TABLE P-5. T fOriariral flnBrarianrv: (Wlv^ •* —' u — ___ s __ _.—-j ^ Group Variable Statistic Ranch Hand Original Comparison T3 % Uptake n 998 933 TSH n Adj. Mean 95%C.I. 1,003 1.16 (1.13,1.19) 936 1.11 (1.08,1.14) Contrast Adj. Relative Risk (95% C.I.) Overall — Low vs. Normal 0.52 (0.45,2.36) High vs. Normal 1.03 (0.59,1.78) — High vs. Normal 1.31 (0.41,4.14) n Adj. Mean 95%C.I. 1,000 n Testosterone 1,000 OOC (p=0.025) PERSTYPE (p=0.035) 0.020 AGE (pO.OOl) 0.775 RACE (p=0.015) AGE*BFAT (p=0.039) GRP*BFAT (p=0.022) AAA'A — 951 Initial Cortisol n Adj. Mean 95%C.I. 1,004 11.35 (10.49,12.27) 945 11.38 (10.52,12.32) Differential Cortisol 1,004 945 AAA'A AAAA AAAA Covariate Remarks 0.332 0.153 0.916 951 Overall Low vs. Normal 1.01 90.63,1.61) High vs. Normal 1.02 (0.42,2.52) n Adj. Mean 95% C.I. p-Value — 0.998 0.980 0.958 AGE (pO.OOl) %BFAT (pO.OOl) 0.820 AGE (pO.OOl) %BFAT (pO.OOl) PERSTYPE (p=O.028) OOC*RACE (p=0.026) **** GRP*AGE*RACE (p=0.01< PERSTYPE (p=0.002) XBFM- fn=0.006^ �TAKE P-5. (coatimad) (Original Cnmarisons Only) Group Variable Statistic Ranch Band Original Comparison 2-flour Postprandial Glucose n Adj. Mean 95% C I .. 976 114.8 (0.,2.) 172129 906 160 1. (0.,2.) 163142 Adj. Relative Risk ( 5 C I ) 9% . . p-Value Covariate Remarks 1,016 954 040 .3 %BFAT (pO.OOl) OCC ( = . 0 ) p002 AGE*RACE ( = . 0 ) p002 Overall Impaired vs. Normal 0 7 ( . 4 0 9 ) . 1 05,.3 Diabetic vs. Normal 1 2 ( . 1 2 3 ) . 9 07,.8 - Diabetes (Composite Indicator) Contrast 003 .2 0.014 045 .0 AGE (p<D.C01) RACE ( = . 2 ) p007 %BFAT ( < . 0 ) p001 1.13 ( . 0 1 6 ) 08,.1 044 .8 AGE ( < . 0 ) p001 %BFAT (pO.OOl) RACE ( = . 1 ) p005 interaction—adjusted mean/relative risk, confidence interval, and p-value are not presented. —No relative risk or confidence interval given for continuous analyses. GRP: Group OCC: Occupation FERSHFE: Personality type (A or B) %BFAT: Percent body fat �TABLE P-6. Variable Summary of Group-by-Covariate Interactions for Endocrinological Variables (Original Comparisons Only) Group Original Comparison Stratification Statistic Ranch Hand p-Value Interaction <10% 3 1173.8 1074.7 0.014 (794.4,1397.2) 10-25% n Mean Ad j . Mean 95% C.I. 815 627.8 602.3 (581.2,623.7) 757 595.9 577.7 0.013 (556.7,599.1) n Mean Ad j . Mean 95% C.I. 179 470.5 461.5 (432.9,491.1) 191 470.7 459.1 0.891 (431.8,487.2) Nonblack Born >1942 Differential Group-by Cortisol Race-by Age 6 522.6 640.1 (472.2,833.5) >25% Testosterone Group-by % Body Fat n Mean Ad j . Mean 95% C.I. n Mean Ad j . Mean 95% C.I. 374 2.07 1.36 (.823) 03,.4 342 2.34 1.65 (.726) 06,.4 0.331 n Mean Ad j . Mean 95% C.I. 572 2.53 1.85 (.227) 09,.9 545 2.36 1.66 (.126) 07,.0 0.419 n Mean Adj. Mean 95% C.I. 32 0.17 -0.68 (-2.35,0.99) 27 3.05 2.41 (0.64,4.17) 004 .0 n Mean Ad j . Mean 95% C.I. 26 3.24 2.72 (.844) 09,.7 31 2.22 0.274 1.54 (-0.14,3.23) Nonblack Born <1942 Black Born >1942 Black Born <1942 �APPENDIX Q Immunological Evaluation �APPENDIX Q: Immunological Evaluation Contents Table Page Q-l Summary of Group-by-Covariate Interactions for Immunological Variables " Q-l Q-2 Unadusted Exposure Index Analyses for Cell Surface Markers by Occupation Q-4 Q-3 Unadjusted Exposure Index Analyses for Functional Stimulation Tests by Occupation Q-8 Q-4 Interaction Summaries of Adjusted Exposure Index Analyses for Immunological Variables Q-10 Q-5 Unadjusted Analyses for Cell Surface Markers by Group (Original Comparisons Only) Q-ll Q-6 Adjusted Analyses for Cell Surface Markers by Group (Original Comparisons Only) Q-12 Q-7 Summary of Group-by-Covariate Interactions for Cell Surface Markers (Original Comparisons Only) Q-14 Q-8 Unadjusted Analyses for Functional Stimulation Tests by Group (Original Comparisons Only) Q-15 Q-9 Adjusted Analyses for Functional Stimulation Tests by Group (Original Comparisons Only) Q-16 Q-10 Summary of Group-by-Covariate Interactions for Functional Stimulation Tests (Original Comparisons Only) Q-17 �•DfflLEQ-1. Sunmary of Gcoup-by-Covariate Interactions for Innuiological Variables Group Comparison p-Value Variable Interaction Stratification Statistic Ranch Hand Total TCells Group-by-Race Nonblack n Adj. Mean 95% C.I. 424 532 0.619 1,616 1,599 (1,564, 1,669) (1,554, 1,646) Black n Adj. Mean 95% C.I. 18 35 0.039 1,566 1,888 (1,340, 1,810) (1,705, 2,080) 0 n Adj. Mean 95% C.I. 128 149 153.8 189.9 0.004 (135.6, 173.3) ( 7 . , 210.0) 109 X)-20 n Adj. Mean 95% C.I. 265 200 199.5 0.998 199.6 ( 8 . , 216.9) (185.3, 214.3) 129 >20-40 n Adj. Mean 95% C.I. 74 94 200.4 172.8 0.109 ( 7 . , 228.7) (151.5, 195.4) 140 >40 n Adj. Mean 95% C.I. 32 29 219.9 214.4 0.857 ( 7 . , 266.3) (172.5, 260.9) 180 0 n Adj. Mean 95% C.I. 35 30 44.19 0.060 32.33 (34.72, 56.23) ( 4 8 , 42.03) 2.8 XX-2 n Adj. Mean 95% C I .. 100 40.43 ( 4 3 , 47.59) 3.5 BCells Group-by(Nonblacks lifetime Smoking Only) (Pack-years) Monocytes Group-byOfficer Occupation Group-byCurrent Alcohol Use Officer (Drinks/day) Officer >2-4 n Adj. Mean 95% C.I. Officer >4 Enlisted 0 Flyer 131 38.38 0.569 (30, 4.4 3.0 46) 31 37 41.53 44.35 0.689 ( 2 0 , 53.91) (35.02, 56.17) 3.0 n Adj. Mean 95% C.I. 13 37.65 (25.58, 55.43) n Adj. Mean 95% C.I. 21 27 35.65 (26.32, 48.28) 45.10 0.232 (34.36, 59.20) Q-l 17 3.2 31 0.608 (25.18, 43.56) �TABLE Q-l. (continued) Group Comparison Ranch Hand Misted X)-2 Flyer n Adj. Mean 95% C.I. 35 50 34.95 0.159 43.22 3.0 (27.53, 44.36) ( 5 0 , 53.37) 8 9 0.097 32.71 56.23 (20.35, 52.56) (36.38, 86.89) n Adj. Mean 95% C.I. 11 5 36.17 0.729 41.10 2.9 ( 2 2 , 75.80) ( 3 9 , 54.54) 2.8 Enlisted 0 Groundcrew n Adj. Mean 95% C.I. % 76 0.421 45.04 48.98 ( 7 8 , 53.54) ( 1 7 , 57.50) 3.9 4.3 Enlisted >0-2 Groundcrew n Adj. Mean 95% C.I. 120 81 089 .3 42.42 41.59 ( 5 8 , 50.24) ( 6 1 , 47.91) 3.1 3.0 Enlisted >2-4 Groundcrew n Adj. Mean 95% C.I. 17 19 37.37 0.397 45.15 2.8 (33.05, 61.67) ( 7 0 , 51.57) Enlisted >4 Groundcrew HLA-CR Cells Statistic Enlisted >4 Flyer Monocytes (Cbnt.) Interaction Stratification Enlisted >2-4 n Flyer Adj. Mean 95% C.I. Variable n Adj. Mean 95% C.I. 16 23 35.32 0.003 68.86 2.2 ( 9 0 , 96.68) ( 6 4 , 47.23) 4.4 n Adj. Mean 95% C.I. 142 158 0.650 593.1 581.0 (555.1, 632.2) ( 4 . , 618.9) 544 n Adj. Mean 95% C.I. 223 306 0.232 559.5 582.8 (529.5, 590.3) ( 5 . , 609.2) 570 >2-4 n Adj. Mean 95% C.I. 65 58 0.750 553.4 565.9 ( 9 . , 611.9) (512.7, 621.7) 478 >4 n Adj. Mean 95% C.I. 51 36 0.052 563.7 472.7 ( 9 . , 640.6) (417.7, 531.0) 417 Group-by0 Current Alcohol Use (Drinks/day) >0-2 0-2 p-Value �TfiBUSQ-l. (continued) Smmary of Gcoup-hy-Covariate Interactions for Inmnological Variables Variable Interaction MLC Net Response Statistic Group Ranch H a n d C o m p a r i s o n n Adj. Mean 95% C.I. 129 156 68,921 77,232 0.053 (56,625, 82,424) (64,572, 91,025) X)-20 n Adj. Mean 95% C.I. 201 277 67,976 74,333 0.057 ( 6 9 0 79,930) (62,947, 86,665) 5,9, >2CMO n Adj. Mean 95% C.I. 71 88 76,511 67,758 018 .2 ( 3 0 7 91,216) (55,676, 81,025) 6,9, >40 n Adj. Mean 95% C.I. 29 29 71,116 63,991 0.444 ( 4 7 4 89,576) ( 8 8 7 81,189) 5,8, 4,3, Stratification Group-byLifetime Smoking (Pack-years) Q-3 p-Value �TABLE Q-2. Unadjusted Exposure Index Analyses for Cell Surface Markers by Occupation Exposure Index Medium High Contrast p-Value 63 1,550 (1429,1676) 56 1,578 (1449,1713) Overall M vs. L H vs. L 0.885 0.626 0.865 24 1,677 (1477,1889) 25 1,659 (1464,1866) 28 1,622 (1439,1815) Overall M vs. L H vs. L 0.923 0.901 0.697 n Mean 95% C.I. 69 1,759 (1623,1901) 78 1,555 (1434,1680) 59 1,586 (1446,1732) Overall M vs. L H vs. L 0.076 0.032 0.091 Officer n Mean 95% C.I. 62 55 63 852.7 838.3 839.6 (755.8,925. 1) (757.7,925.8) (764.5,945.8) Overall M vs. L H vs. L 0.970 0.983 0.821 Enlisted Flyer n Mean 95% C.I. 27 24 26 861.7 917.5 888.3 (789.3,1055 ,3)(767. 1,1018. 5)(744.5,987.5) Overall M vs. L H vs. L 0.832 0.755 0.545 Enlisted Groundcrev n Mean 95% C.I. 59 69 76 874.6 865.4 948.9 (867.6,1033 .9X791.3,942.7) (790.4,963.1) Overall M vs. L H vs. L 0.296 0.146 0.227 Statistic Lov n Mean 95% C.I. 62 1,594 (1471,1723) Enlisted Flyer n Mean 95% C.I. Enlisted Groundcrev Total T Cells Occupation Officer Variable o Helper T Cells �TABLE Q-2. (continued) Unadjusted Exposure Index Analyses for Cell Surface Markers by Occupation Occupation Statistic Low Officer n Mean 95% C.I. Enlisted Flyer Exposure Index Medium High p-Value 62 64 56 522.6 461.4 497.3 (467.9,583.6) (413.8,514.4) (442.7,558.7) Overall M vs. L H vs. L 0.287 0.117 0.546 n Mean 95% C.I. 28 24 26 557.7 533.9 528.1 (443.1,629.3) (451.1,631.9) (474.1,656.0) Overall M vs. L H vs. L 0.891 0.930 0.656 n Mean 95% C.I. 69 77 59 575.5 502.3 505.5 (518.9,638.3) (455.4,554.0) (451.9,565.3) Overall M vs. L H vs. L 0.123 0.063 0.097 Officer Suppressor T Cells Contrast Enlisted Groundcrew Variable n Mean 95% C.I. 55 62 64 167.4 154.1 176.8 (143.5,193.2) (131.5,178.4) (150.7,205.0) Overall M vs. L H vs. L 0.453 0.445 0.618 Enlisted Flyer n Mean 95% C.I. 24 25 26 233.2 228.6 183.3 (185.4,286.6) (182.1,280.3) (142.7,228.9) Overall M vs. L H vs. L 0.262 0.897 0.144 Enlisted Groundcrew n Mean 95% C.I. 69 73 59 202.4 196.7 199.2 (173.3,233.8) (168.8,226.8) (168.1,233.0) Overall M vs. L H vs. L 0.966 0.792 0.889 o Ul B Cells �TABLE Q-2. (continued) Unadjusted Exposure Index Analyses for Cell Surface Markers by Occupation Occupation Statistic Low Officer n Mean 95% C.I. Enlisted Flyer Exposure Index Medium p-Value 62 64 55 47.39 43.32 46.85 (35.48 ,52.89) (38.49,57.02) (38.34,58.58) Overall M vs. L H vs. L 0.799 0.584 0.546 n Mean 95% C.I. 24 26 27 37.52 41.76 36.91 (27.27 ,49.97) (28.05,50.20) (31.38,55.56) Overall M vs. L H vs. L 0.816 0.940 0.563 n Mean 95% C.I. 69 76 59 48.18 46.97 47.03 (39.65 ,55.64) (40.02,55.27) (40.12,57.87) Overall M vs. L H vs. L 0.975 0.992 0.842 Officer Monocytes Contrast Enlisted Groundcrew Variable n Mean 95% C.I. 62 63 56 564.9 527.5 544.1 (463.9 ,595.1) (480.0,612.2) (495.8,638.7) Overall M vs. L H vs. L 0.750 0.727 0.449 Enlisted Flyer n Mean 95% C.I. 24 26 28 568.3 570.3 547.1 (475.4 ,673.9) (457.7,644.5) (480.3,663.7) Overall M vs. L H vs. L 0.932 0.739 0.977 Enlisted Groundcrew n Mean 95% C.I. 69 58 76 580.7 578.0 587.3 (521.7 ,637.1) (533.2,644.0) (519.4,645.6) Overall M vs. L H vs. L 0.973 0.820 0.949 o I HLA-DR Cells High �TABLE Q-2. (continued) Unadjusted Exposure Index Analyses for Cell Surface Markers by Occupation o I Statistic Low n Mean 95% C.I. Enlisted Flyer Enlisted Groundcrew T4 /T ?8 Ratio Occupation Officer Variable Exposure Index Medium High Contrast p-Value 62 55 63 1.732 2.002 1.758 (1.500,1.964) (1.773,2.232) (1.512,2.004) Overall M vs. L H vs. L 0.207 0.106 0.881 n Mean 95% C.I. 27 24 26 1.625 1.793 1.781 (1.478,2.108) (1.478,2.083) (1.328,1.922) Overall M vs. L H vs. L 0.692 0.955 0.449 n Mean 95% C.I. 59 69 76 1.780 1.751 1.780 (1.580,1.922) (1.617,1.943) (1.595,1.965) Overall M vs. L H vs. L 0.965 0.812 0.822 �TABLE Q-3. Unadjusted Exposure Index Analyses for Functional Stimulation Tests by Occupation Low Exposure Index Medium High Contrast p-Value 64 1,428 (1249,1632) 56 1,809 (1568,2087) Overall M vs. L H vs. L 0.047 0.071 0.557 24 1,668 (1315,2114) 26 1,973 (1571,2478) 28 1,642 (1318,2045) Overall M vs. L H vs. L 0.466 0.320 0.926 n Mean 95% C.I. 69 1,814 (1583,2079) 78 1,923 (1691,2186) 58 1,843 (1588,2139) Overall M vs. L H vs. L 0.818 0.541 0.877 Officer Variable n Mean 95% C.I. 61 213,383 (192,608, 234,158) 64 205,057 (184,775, 225,339) 55 209,661 (187,782 231,539) Overall M vs. L H vs. L 0.853 0.575 0.809 Enlisted Flyer n Mean 95% C.I. 24 245,179 (212,269, 278,089) 26 209,656 (178,038, 241,275) 28 219,358 (188,890, 249,827) Overall M vs. L H vs. L 0.297 0.131 0.263 Enlisted Groundcrew n Mean 95% C.I. 69 227,062 (208,219, 245,905) 78 221,793 (204,070, 239,515) 58 208,537 (187,984, 229,089) Overall M vs. L H vs. L 0.414 0.690 0.194 Statistic Officer n Mean 95% C.I. 61 1,705 (1487,1955) Enlisted Flyer n Mean 95% C.I. Enlisted Groundcrev Unstimulated Response (PHA) Occupation 0 00 PHA Net Response �TABLE Q-3. (continued) Unadjusted Exposure Index Analyses for Functional Stimulation Tests by Occupation Exposure Index Medium High Contrast p-Value 64 86,250 (71,074, 102,894) 56 96,107 (78,992, 114,899) Overall M vs. L H vs. L 0.633 0.947 0.388 24 92,847 (71,008, 117,610) 26 101,256 (79,227, 125,985) 28 55,480 (40,058, 73,410) Overall M vs. L H vs. L 0.004 0.619 0.011 n Mean 95% C.I. 69 77,859 (64,872, 92,030) 78 92,188 (78,815, 106,610) 58 87,540 (72,540, 103,950) Overall M vs. L H vs. L 0.345 0.151 0.360 Officer Variable n Mean 95% C.I. 58 74,175 (59,443, 90,536) 63 78,069 (63,516, 94,123) 54 76,486 (61,000, 93,722) Overall M vs. L H vs. L 0.940 0.727 0.841 Enlisted Flyer n Mean 95% C.I. 23 71,959 (50,336, 97,436) 26 74,217 (53,426, 98,416) 26 62,391 (43,471, 84,721) Overall M vs. L H vs. L 0.724 0.892 0.550 Enlisted Groundcrew n Mean 95% C.I. 68 69,460 (56,591, 83,647) 77 85,901 (72,345, 100,620) 57 81,091 (65,912, 97,842) Overall M vs. L H vs. L 0.248 0.102 0.275 Low MLC Net Response n Mean 95% C.I. 62 85,479 (70,144, 102,328) Enlisted Flyer n Mean 95% C.I. Enlisted Groundcrew o Statistic Officer "Pokeweed Net Response Occupation �TABLE Q-4. Interaction Summaries of Adjusted Exposure Index Analyses for Immunological Variables Variable Interaction (Occupation) Stratificaton Statistic Low Exposure Index Medium High Contrast p-Value __ 4 3,273 0 — 5 1,557 M vs. L H vs. L 0.031 n Adj. Mean 22 118 2, 22 1,975 16 2,770 M vs. L H vs. L 0.655 0.125 n Adj. Mean 24 2,752 22 2,069 15 2,788 M vs. L H vs. L 0.071 0.941 n Adj. Mean 10 2,357 17 2,120 16 1,892 M vs. L H vs. L 0.628 0.319 >100 n Adj. Mean 1 983 1, 2 1,555 3 6,700 M vs. L H vs. L 0.725 0.049 Born >1942 n 18 Adj. Mean 261, 397 9 153,534 7 299 ,002 M vs. L H vs. L 0.002 0.302 Born n 40 1923-1941 Adj. Mean 234, 921 43 236,262 46 223 ,373 M vs. L H vs. L 0.943 0.534 Born <1922 11 229,757 1 191 ,111 M vs. L H vs. L 0.095 0.632 0 Exposure Index-byLi fe time Alcohol Use >0-5 (Drink/years) (Officer) >5-30 o i PHA Net Response n Adj. Mean >30-100 Unstimulated Response (PHA) Exposure Index-byAge (Officer) n Adj. Mean 3 139,307 �TABLE Q-5. Unadjusted Analyses for Cell Surface Markers by Group (Original Comparisons Only) Group Variable Statistic Ranch Hand Original Comparison p-Value Total T Cells n Mean 95% C.I. 464 1,606 (1551,1662) 424 1,592 (1537,1648) 0.711 Helper T Cells n Mean 95% C.I. 461 868.1 (833.3,903.6) 423 861.0 (826.4,896.3) 0.765 Suppressor T Cells n Mean 95% C.I. 465 521.9 (498.5,546.5) 425 529.7 (505.7,554.8) 0.630 B Cells n Mean 95% C.I. 457 185.2 (173.8,197.0) 418 188.1 (176.5,200.1) 0.708 n Mean 95% C.I. 462 46.19 (43.05,49.56) 424 45.47 (42.39,48.79) 0.744 n Mean 95% C.I. 462 571.6 (547.6,596.2) 424 569.9 (546.1,594.3) 0.917 n Mean 95% C.I. 461 1.597 (1.528,1.669) 421 1.558 (1.491,1.629) 0.410 Monocytes HLA-DR Cells T4/T8 Ratio Q-ll �TABLE Q-6. Adjusted Analyses for Cell Surface Markers by Group (Original Goqparisons Only) Variable Statistic Total TCells n Adj. Mean 95% C.I. Helper Cells Group Original Ranch Hand Comparison 442 p-Value 414 **** n 439 413 Adj. Mean 829 6. 867.5 95% C.I. (2.,8 6 6 (3., 9 1 1 898 9 . ) 846 0 . ) Suppressor TCells n 465 Adj. Mean 95% C.I. BCells n Adj. Mean 95% C.I. Monocytes n 440 Adj. Mean 95% C.I. HLA-CR Cells n Adj. Mean 95% C.I. 085 .3 425 455 416 155 8. 111 9. 041 .5 (174.2, 197.1) (179.6, 202.9) Covariate Remarks* BATCH ( < . 0 ) p001 GRP*RACE (p=0.028) CRKXR*PACKYR ( = . 0 ) p004 CSMCK (p<0.001) ALC (p=0.005) AGE (p=0.049) BATCH (p=0.014) DAY (BATCH) (p=0.007) AGE (p<0.001) CSMCK (p<0.001) ALC*OCC (p=0.025) (=.0) p001 BATCH (p<0.001) GRP*RACE (p=0.010) OCC (p=0.021) A3E (p=O.C06) CSMCK (p<0.001) BATCH (p<0.001) OOC (p=0.041) AGE (p=0.005) ALC (p=0.002) CSMCK (p<0.001) 414 BATCH (p<0.001) MY (BATCH) (p<0.001) GRP*AGE (p=0.040) OCC (p=0.010) EKCZR (p=0.008) CSMCK (p<0.001) PACKYR (p=0.020) 461 423 570.5 568.2 0.887 ( 4 . , 594.5) ( 4 . , 592.0) 569 550 BATCH (p<0.001) DAY (BATCH) (p4).014) OCC (p=0.017) CSMCK (p<0.001) AGE*PACKXR (p=0.007) CH12 �TAHE Q-6. (continued) Adjusted Analyses for Cell Surface Markets by Group (Original Gon|Brisans Only) Group Variable Statistic Ranch Hand T /T Ratio n 461 Adj. Mean **** 95% C.I. **** Original Comparison 421 **** **** p-Value **** Covariate Remarks* MICH (p<0.001) OOC (p=0.002) GRP*CSMCK (p=0.016) ^Abbreviations; BA3XH: batch-to-batch variation among examination groups DAY (MTCH): blood-draw day variation ALC: current alcohol use CSMGK: current smoking OOC: occupation GRP: group EHOR: lifetime alcohol use (drink-years) PACKER: lifetime smoking (pack-years) ****Group-by-covariate interaction - adjusted mean, confidence interval, and p-value not presented. Q-13 �TAHEQ-7. Sunnary of Group-4y-0ovarlate Interactions for Cell Surface Markers (Original Comparisons Only) Group Original Comparison p-Value Variable Interaction Stratification Statistic Banch Hand Total TCells Group-by-Race Nonblack n Adj. Mean 95% C.I. 424 391 0.324 1,601.7 1,567.3 (1,554.6, 1,649.6) (1,518.6, 1,161.6) Black n Adj. Mean 95% C.I. 18 23 1,558.3 1,888.6 0.043 (1,338.6, 1,794.8) (1,673.1, 2,117.1) Nonblack n Adj. Mean 95% C I .. 444 530.0 ( 0 . , 555.6) 555 402 530.8 ( 0 . , 556.6) 562 0.960 n Adj. Mean 95% C.I. 21 497.5 ( 1 . , 602.2) 410 23 708.8 ( 8 . , 852.2) 596 0.008 n Adj. Mean 95% C.I. 174 41.39 (68, 4.4 3.8 64) 162 48.02 ( 2 8 , 53.76) 4.9 0.048 Bom 1923-1941 n Adj. Mean 95% C.I. 247 48.20 ( 4 0 , 52.76) 4.4 232 42.82 ( 9 0 , 46.99) 3.2 0.058 Born <L922 n Adj. Mean .95% C.I. 19 47.88 ( 5 0 , 65.40) 3.6 20 46.90 (34.85, 63.13) 0.924 n Adj. Mean 95% C.I. 277 1.44 (1.36, 1 5 ) .2 273 1.48 (.0 1 5 ) 14, . 6 0.454 n Adj. Mean 95% C.I. 78 1.70 (.4 1 8 ) 15, . 8 69 1.73 (1.56, 1 9 ) .2 0.811 n Adj. Mean • 95% C.I. 90 1.84 (.8 2 0 ) 16, . 1 71 1.51 (1.36, 1 6 ) .7 0.004 n Adj. Mean 95% C.I. 16 1.70 (1.38, 2 1 ) .1 8 1.75 (1.30, 2 3 ) .5 0.895 Suppressor Group-by-Pace TCells Black Monocytes Group-by-Age T4/T 8 Ratio Group-byCurrent Smoking (Cigarettes/ day) Born >1942 0 X)-20 >2CWO >40 0-14 �TABLE Q-8. Unadjusted Analyses for Functional Stimulation Tests by Group (Original Comparisons Only) Group Ranch Hands Original Comparison p-Value 464 1,669 (1,578, 1,755) 426 1,640 (1,559, 1,724) 0.608 n Mean 95% C.I. 463 212,481 425 206,796 0.168 Pokeweed Net Response n Mean 95% C.I. 465 85,559 (81,373, 89,849) 426 83,724 (79,579, 87,974) 0.522 MLC n Net Response Mean 95% C.I. 452 79,451 (75,485, 83,519) 415 82,387 (78,363, 86,512) 0.285 Variable Statistic Unstimulated n Response (PHA) Mean 95% C.I. PHA Net Response (206,424, 218,539) (200,733, 212,858) Q-15 �TABLE Q-9. Adjusted Analyses for Functional Stimulation Tests by Group (Original Comparisons Only) Group Ranch Hand Original Comparison p-Value Unstimulated n Response Adj. Mean (PHA) 95% C.I. 464 1,813 (1,645, 1,997) 426 1,783 (1,620, 1,962) 0.613 PHA Net n 461 423 Response Adj. Mean **** Variable Statistic **** 95% C.I. Pokeweed n 465 426 Net Response Adj. Mean 92,684 91,738 95% C I . . (86,712, 98,855) (85,660, 98,023) MLCNet Response n 431 405 Adj. Mean 79,412 82,916 95% C I . . (73,680, 85,359) ( 6 8 8 89,214) 7,4, Covariate Remarks BATCH (p<0.001) DAY (BATCH) (p<0. 0 ) 01 RACE (p<0.001) AGE (p<0.001) BATCH (p<O.C01) DAY (BATCH) (p<0.001) GRP*OCC (p=0.017) RACE (p=0.005) AGE (p4).001) AUXSMOK (p*0.037) BATCH (p<0.001) 0.746 DAY (BATCH) (p<0.001) OCC (p=0.046) CSMOK (p<0.001) BATCH (p<0.001) 0 1 7 DAY (BATCH) (p<0.001) .9 AUC (p<0.001) EBK2R (p=0.001) CSMCK (p<0.001) ****Group-by-covariate interaction - adjusted mean, confidence interval, and p-value not presented. Q-16 �TABLE Q-10. Sunnary of Group-by-Covarlate Interactions for Functional Stimulation Tests (Original Comparisons Only) Group Original Comparison Variable Interaction Stratification Statistic Ranch Hand PHA Net Pesponse Group-byOccupation Officer n Adj. Mean 95% C.I. 180 163 217,397 210,363 0.280 (201,412, 233,382) (193,843, 226,883) Enlisted Flyer n Adj. Mean 95% C.I. 78 77 230,597 202,077 0.003 (212,032, 249,162) (183,749, 220,405) Enlisted Groundcrev n Adj. Mean 95% C.I. 203 183 208,347 212,473 0.504 (193,386, 223,308) (197,550, 227,397) 0 n Adj. Mean 95% C.I. 129 109 78,900 87,110 0.107 (69,737, 88,627) (77,309, 9 , % 74) >0-20 n Adj. Mean 95% C.I. 201 206 76,689 83,468 009 .7 (69,735, 83,973) (76,169, 91,101) n Adj. Mean 95% C.I. 71 67 86,904 75,456 0.086 (77,241, 97,135) (65,854, 85,711) n Adj. Mean 95% C.I. 29 23 80,525 75,984 0.679 (66,260, 96,180) ( 0 8 6 92,850) 6,0, MLC Net Response Group-byLifetime Smoking (Pack-years) >40 Q-17 p-Value �APPENDIX R Pulmonary Disease �APPENDIX R: Pulmonary Disease Contents Table R-l R-2 R-3 R-4 R-5 R-6 R-7 R-8 R-9 R-10 Page Unadjusted Analyses of Respiratory Variables by Group (Original Comparisons Only). . . R-l Adjusted Analyses of Respiratory Variables by Group (Original Comparisons Only) R-3 Summary of Group-by-Covariate Interactions for Respiratory Variables (Original Comparisons Only) R-5 Cross Tabulation of Bronchitis- (Abnormal, Total) byExposure Index Category-by-Age Category-by-Pack-Years for Ranch Hand Officers R-9 Cross Tabulation of Pneumonia- (Abnormal, Total) byExposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Officers R-10 Cross Tabulation by Hyperresonance- (Abnormal, Total) byExposure Index Category-by-Pack-Year Category for Ranch Hand Officers R-ll Cross Tabulation of Thorax- and Lung- (Abnormal, Total) by-Exposure Index Category for Ranch Hand Officers R-12 Cross Tabulation of X Ray- (Abnormal, Total) by-Exposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Officers R-13 Cross Tabulation of Bronchitis- (Abnormal, Total) by-Exposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Enlisted Flyers R-14 Cross Tabulation of Pleurisy- (Abnormal, Total) byExposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Enlisted Flyers R-15 R-ll Cross Tabulation of Pneumonia- (Abnormal, Total) byExposure Index Category-by-Age Category for Ranch Hand Enlisted Flyers R-16 R-12 Cross Tabulation of Hyperresonance- (Abnormal, Total) byExposure Index Category-by-Pack-Year Category for Ranch Hand Enlisted Flyers R-17 �APPENDIX R: Pulmonary Disease Contents (continued) Table R-13 Cross Tabulation of Thorax and Lung Abnormal!ties(Abnormal, Total) by-Exposure Index Category for Ranch Hand Enlisted Flyers R-18 R-14 Cross Tabulation of Asthma- (Abnormal, Total) byExposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Enlisted Groundcrev R-19 R-15 Cross Tabulation of Pleurisy- (Abnormal, Total) byExposure Index Category-by-Age Category for Ranch Hand Enlisted Groundcrew. R-20 R-16 Cross Tabulation of Wheezes- (Abnormal, Total) byExposure Index Category-by-Age Category for Ranch Hand Enlisted Groundcrew R-21 R-17 Cross Tabulation of Wheezes- (Abnormal, Total) byExposure Index Category-by-Pack-Year Category for Ranch Hand Enlisted Groundcrew R-22 R-18 Cross Tabulation of Bronchitis- (Abnormal, Total) byExposure Index Category for Ranch Hand Enlisted Groundcrew R-23 �TABLE R-l. Unadjusted Analyses of Respiratory Variables by Group (Original Comparisons Only) Group Ranch Hand Original Comparison Number Percent Est. Relative Risk (95% C.I.) p-Value Variable Statistic Number Percent Asthma n Abnormal Normal 1,016 44 972 4.3 95.7 954 42 912 4.4 95.6 0.98 (0.64,1.51) 0.92 Bronchitis n Abnormal Normal 1,015 129 886 12.7 87.3 954 131 823 13.7 86.3 0.91 (0.71,1.19) 0.50 Pleurisy n Abnormal Normal 1,016 47 969 4.6 95.4 953 45 908 4.7 95.3 0.98 (0.65,1.49) 0.92 Pneumonia n Abnormal Normal 1,016 195 821 19.2 80.8 953 191 762 20.0 80.0 0.95 (0.76,1.18) 0.64 Tuberculosis n Abnormal Normal 1,015 7 1,008 0.7 99.3 954 5 949 0.5 99.5 1.32 (0.43,3.97) 0.64 Thorax & Lungs n Abnormal Normal 1,015 61 954 6.0 94.0 955 45 910 4.7 95.3 1.29 (0.87,1.92) 0.20 �TABLE R-l. (contined) Unadjusted Analyses of Respiratory Variables by Group (Original Comparisons Only) Group Ranch Hand Original Comparison Number Percent n Asymmetrical Exp. Abnormal Normal 1,015 2 1,013 0.2 99.8 955 2 953 0.2 99.8 0.94 (0.16,5.49) 0.95 Hyperresonance n Abnormal Normal 1,015 30 985 3.0 97.0 955 28 927 2.9 97.1 1.01 (0.60,1.69) 0.98 Dullness n Abnormal Normal . 1,015 2 1,013 0.2 99.8 955 0 955 0.0 100.0 Wheezes n Abnormal Normal 1,015 24 991 2.4 97.6 955 16 939 1.7 98.3 1.42 (0.76,2.67) 0.28 Rales n Abnormal Normal 1,015 6 1,009 0.6 99.4 955 4 951 0.4 99.6 1.41 (0.42,4.70) 0.59 X ray n Abnormal Normal 1,012 102 910 10.1 89.9 951 113 838 11.9 88.1 0.83 (0.63,1.10) 0.20 Statistic Percent Est. Relative Risk (95% C.I.) p-Value Number Variable 0.17 �TABLE R-2. Adjusted Analyses of Respiratory Variables by Group (Original Comparisons Only) Group Variable Ranch Hand Original Comparison Adj. Relative Risk ( 5 C.I.) 9% p-Value Covariate Remarks* **** PACKYR (p=0.02) GRP*PACKYR (p=0.03) 0.44 GRP*PACKYR (Borderline: p=0.09) **** GRP*PACKYR (p=0.0006) 0.97 (0.77,1.22) 0.74 AGE (p=0.002) 953 1.32 (0.43,3.98) 0.60 GRP*PACKYR (Borderline: p=0.06) 1,011 954 1.33 (0.88,2.01) 0.14 AGE (p<0.0001) PACKYR (p<0.0001) Asymmetrical Exp. 1,011 954 0.94 (0.16,5.50) 0.97 AGE*PACKYR (Borderline: p=0.08) Hyperresonance 1,011 954 1.00 (0.58,1.74) 0.87 AGE (p<0.0001) PACKYR (p<0.0001) GRP*PACKYR (Borderline: p=0.09) As thma 1,012 953 Bronchitis 1,011 953 Pleurisy 1,012 952 Pneumonia 1,012 952 Tuberculosis 1,011 Thorax and Lungs **** 0.91 (0.70,1.18) **** �TABLE R-2. (continued) Adjusted Analyses of Respiratory Variables by Group (Original Comparisons Only) Group Variable Ranch Hand Original Comparison Adj. Relative Risk (95% C.I.) Dullness 1,011 954 ¥heezes 1,011 954 — 1.43 (0.75,2.76) 0.25 PACKYR (p<0.0006) Rales 1,011 954 1.42 (0.39,5.18) 0.59 AGE*PACKYR (p=0.04) GRP*AGE (p=0.09) X Ray 1,008 950 0.86 (0.64,1.16) 0.31 AGE (p<0.0001) PACKYR (p=0.02) GRP*PACKYR (Borderline: p=0.07) 50 **Abbreviations PACKYR: lifetime smoking history (pack-years) GRP: group p-Value Covariate Remarks* None �TABIER-3. Suomary of Group-by-Covariate Interactions for Respiratory Variables x~—o— —~ ~ i isons *—--j f ~ Group Variable Interaction Asthma Group-byPack-Year Stratification 0 Statistic Ranch Hand Number Percent Original Comparison Number Percent Adj. Relative Risk ( 5 C I ) p-Value 9%.. >10 £ Bronchitis Group-byPack-Year 0 >0-10 >10 291 11 280 3.78 96.22 268 3 265 1.12 98.88 4.78 ( . 9 1 . 6 0 0 13,55) .4 n Abnormal Normal 284 16 268 5.63 94.37 280 12 268 4.29 95.71 1.33 ( . 3 2 8 ) 0 4 06,.3 .6 Abnormal Normal XKLO n Abnormal Normal 437 17 420 3.89 96.11 405 27 378 6.67 93.33 0.57 ( . 1 1 0 ) 0 0 03,.5 . 7 n Abnormal Normal 291 38 253 1.6 30 8.4 69 268 25 243 9.33 90.67 1 4 (.624) 0 1 . 6 08,.8 .6 n Abnormal Normal 284 38 246 13.38 86.62 280 40 240 1.9 42 85.71 0.93 ( . 8 1 4 ) 0 7 05,.9 . 5 n Abnormal Normal 436 52 384 11.93 8.7 80 405 66 339 16.30 8.0 37 0.70 ( . 7 1 0 ) 0 0 04,.3 . 7 n �TABIE R-3. (continued) Sunnary of Group-by-Covariate Interactions for Respiratory Variables Comparisons Only) Group Variable Interaction Pleurisy Group-byPack-Year Stratification 0 XKLO >10 Tuberculosis Group-byPack-Year 0 >0-10 >10 Statistic Ranch Hand Number Percent Original Comparison Number Percent Adj. Relative Risk ( 5 C I ) p-Value 9% .. n Abnormal Normal 291 8 283 2.75 97.25 267 5 262 1.87 98.13 1.48 ( . 0 4 3 ) 0 4 05,.7 . 9 n Abnormal Normal 284 18 266 6.34 9.6 36 280 4 276 1.43 98.57 4.67 ( . 1 1 . 2 O 0 1 16,30) . 0 n Abnormal Normal 437 21 416 48 .1 95.19 405 36 369 88 .9 91.11 0.52 ( . 0 0 9 ) 0 0 03,.0 . 2 n Abnormal Normal 290 I 289 0.34 99.66 268 3 265 1.12 9.8 88 0.31 ( . 5 2 3 ) 0 2 00,.4 . 8 n Abnormal Normal 284 4 280 1.41 98.59 280 0 00 .0 280 1 0 0 0.0 n Abnormal Normal 437 2 435 04 .6 99.54 405 2 403 04 .9 99.51 — 00 .5 0.93 ( . 6 5 4 ) 0 9 01,.2 . 2 �TABLE R-3. (continued) (Original Comparisons Oily) Group Variable Interaction Byperresonnance Group-byPack-Year Stratification Adj. Relative Risk ( 5 C I ) p-Value 9%.. 291 2 289 0.69 99.31 268 00 .0 0 268 1 0 0 0.0 n Abnormal Normal 283 8 275 2.83 97.17 281 3 278 1.07 98.93 2.70 ( . 5 9 2 ) 0 1 07,.7 . 3 n Abnormal Normal 437 20 417 4.58 95.42 405 25 380 6.17 93.83 0.73 ( . 0 1 3 ) 0 3 04,-3 .0 n Abnormal Normal 384 1 383 0.26 9.4 97 341 0 00 .0 341 1 0 0 0.0 Born 1922-1942 n Abnormal Normal 600 5 595 0.83 99.17 577 4 573 >10 Group-byAge Original Comparison Number Percent n Abnormal Normal 0 XKLO Bales Statistic Ranch Hand Number Percent Bom >1942 Born <L922 n Abnormal Normal 27 0 00 .0 27 1 0 0 0.0 0.69 99.31 36 0 00 .0 36 1 0 0 0.0 — — 01 .7 03 .5 1.20 ( . 4 4 2 ) 0 7 03, . 1 .8 — — �TfiBLE R-3. (continued) Sunnary of Gxxip-by-Gowariate Interactions for Respiratory \feriables (Original Gonjjariscns Only) Group Variable Interaction XRay Group-byPack-Year Stratification 0 >0-10 >10 Statistic Ranch Band Hunter Percent Original Comparison Number Percent Adj. Relative Risk ( 5 C I ) p-Valxe 9% .. n Abnormal Normal 290 15 275 51 .7 9.3 48 266 29 237 1.0 09 8.0 91 n Abnormal NornBl 282 28 254 9.93 90.07 281 26 255 9.25 90.75 1.08 (0.62, 1.89) 0.79 n Abnornal Normal 436 59 377 13.53 86.47 403 58 345 14.39 85.61 0.93 (0.63, 1.37) 0.72 0 4 (.408) 0 0 . 5 02,.5 . 1 �OFFICERS The statistically significant bronchitis-by-exposure-by-age-by-pack-year shown in Table 20-5 is described here in Table R-4. interaction (p=0.009) TABLE R-4. Cross Tabulation of Bronchitis- (Abnormal, Total) by-Exposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Officers Exposure Index Pack-Years Age Low Abnormal Total Percent Medium Abnormal Total Percent High Abnormal Total Percent 0 50 VO Born >1942 Born 1922-1942 Born <1922 4 3 0 21 32 3 19 9 0 0 7 0 10 36 3 0 19 0 1 8 0 3 44 0 33 10 >0-10 Born >1942 Born 1922-1942 Born <1922 1 2 0 14 15 0 7 13 2 5 0 7* 19 3 29 26 0 0 1 0 0 26 1 4 0 Born >1942 Born 1922-1942 Born <1922 0 8 0 3 39 0 0 21 2 3 1 4 43 5 50 7 20 0 4 0 4 44 1 0 9 0 >10 �The statistically significant pneumonia-by-exposure-by-age-by-pack-year Table 20-5 is described here in Table R-5. interaction (p=0.040) shown in TABLE R-5. Cross Tabulation of Pneumonia- (Abnormal, Total) by-Exposure Index Category-byAge Category-by-Pack-Year Category for Ranch Hand Officers Exposure Index Low Pack-Years Age Abnormal Total Medium Percent Abnormal Total High Percent Abnormal Total Percent 0 Born >1942 Born 1922-1942 Born <1922 6 6 0 21 22 3 29 19 0 0 6 1 10 ' 36 3 0 17 33 1 10 0 3 44 0 33 23 >0-10 Born. >1942 Born 1922-1942 Born <1922 1 4 0 14 15 0 7 27 1 4 0 7 19 3 14 21 0 0 9 1 0 26 1 35 100 Born >1942 Born 1922-1942 Born <1922 0 9 0 3 39 0 0 23 1 6 1 4 43 5 25 14 20 0 8 0 4 44 1 0 18 0 >10 �The statistically significant hyperresonance-by-exposure-by-pack-year (p=0.07) seen in Table 20-5 is tabulated in Table R-6. interaction TABLE R-6. Cross Tabulation of Hyperresonance- (Abnormal, Total) by-Exposure Index Categoryby-Pack-Year Category for Ranch Hand Officers Exposure Index Low Pack-Years Abnormal Total Percent Medium Abnormal Total Percent High Abnormal Total Percent 0 0 56 0 0 49 0 1 47 2 >0-10 0 29 0 1 29 3 1 27 4 >10 5 42 12 1 52 2 0 49 0 �The statistically significant thorax- and lung-by-exposure category interaction (p=0.05) shown in Table 20-5 is tabulated in Table R-7. TABLE R-7. Cross Tabulation of Thorax and Lung- (Abnormal, Total) by- Exposure Index Category for Ranch Hand Officers Exposure Index Low Abnormal Total 9 l-» NS 127 Percent 7 Medium Abnormal Total Percent 3 130 2 High Abnormal Total Percent 5 123 4 �The statistically borderline significant x ray-by-exposure-category-byage-by-pack-year interaction (p=0.08) seen in Table 20-5 is tabulated in Table R-8. TABLE R-8. Cross Tabulation of X Ray- (Abnormal, Total) by-Exposure Index Categoryby-Age Category-by-Paek-Year Category for Ranch Hand Officers Exposure Index Pack-Years Age Low Abnormal Total Percent Medium Abnormal Total Percent Abnormal High Total Percent 0 50 h-» 00 Born >1942 Born 1922-1942 Born <1922 1 3 0 21 32 3 5 9 0 0 1 0 10 36 3 0 3 0 0 2 0 3 44 0 0 5 >0-10 Born >1942 Born 1922-1942 Born <1922 0 3 0 14 15 0 0 20 0 3 1 6 19 3 0 16 33 0 2 0 0 26 1 8 0 Born >1942 Born 1922-1942 Born <1922 0 7 0 3 39 0 0 18 0 7 1 4 43 5 0 16 20 0 2 1 4 44 1 0 5 100 >10 �ENLISTED FLYERS The statistically significant bronchitis-by-exposure-by-age-by-pack-year seen in Table 20-6 is tabulated in Table R-9. interaction (p=0.005) TABLE R-9. Cross Tabulation of Bronchitis- (Abnormal, Total) by-Exposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Enlisted Flyers Exposure Index Pack-Years Age Low Abnormal Total Percent Medium Abnormal Total Percent High Abnormal Total Percent 0 50 Born >1942 Born <1942 2 0 3 6 67 0 0 0 3 6 0 0 0 1 3 11 0 9 >0-10 Born >1942 Born <1942 1 1 4 13 25 8 1 1 6 9 17 11 2 0 2 7 100 0 >10 Born >1942 Born <1942 0 3 2 26 0 12 0 6 7 32 0 19 0 3 2 31 0 10 �The statistically borderline significant pleurisy-by-exposure-by-age-by-pack-year interaction (p=0.08) shown in Table 20-6 is tabulated in Table R-10. TABLE R-10. Cross Tabulation of Pleurisy- (Abnormal, Total) by-Exposure Index Category-byAge Category-by-Pack-Year Category for Ranch Hand Enlisted Flyers Exposure Index Pack-Years Age Low Abnormal Total Percent Medium Abnormal Total Percent High Abnormal Total Percent 0 Born >1942 Born <1942 0 0 3 6 0 0 0 0 3 6 0 0 0 0 3 11 0 0 >0-10 Born >1942 Born <1942 1 1 4 13 25 8 0 1 6 9 0 11 1 0 2 7 50 0 >10 Born >1942 Born <1942 1 1 2 26 50 4 0 1 7 32 0 3 0 2 2 32 0 6 �The statistically borderline significant pneumonia-by-exposure-by-age interaction (p=0.08) shown in Table 20-6 is tabulated in Table R-ll. TABLE R-ll. Cross Tabulation of Pneumonia- (Abnormal, Total) -by-Exposure Index Category-by-Age Category for Ranch Hand Enlisted Flyers Exposure Index Age Born >1942 Born <1942 JO . Low Abnormal Total Percent 4 10 9 45 44 22 Medium Abnormal Total Percent 2 10 16 47 13 21 High Abnormal Total 0 13 7 50 Percent 0 26 �The statistically borderline significant hyperresonance-by-exposure-by-pack-year category interaction (p=0.08) shown in Table 20-6 is tabulated in Table R-12. TABLE R-12. Cross Tabulation of Hyperresonance- (Abnormal, Total) by-Exposure Index Category-by-Pack-Year Category for Ranch Hand Enlisted Flyers Exposure Index Pack-Years Abnormal 0 >0-10 >10 I l-» -J 0 0 3 Low Total 9 17 28 Percent 0 0 11 Medium Abnormal Total Percent 0 2 4 9 15 39 0 13 10 High Abnormal Total 0 0 0 14 9 34 Percent 0 0 0 �The statistically significant (p=0.04) thorax and lung-by-exposure interaction shown in Table 20-6 is tabulated in Table R-13. TABLE R-13. Cross Tabulation of Thorax and Lung Abnormalities- (Abnormal, Total) by-Exposure Index Category for Ranch Hand Enlisted Flyers Exposure Index Low Abnormal Total 7 00 54 Percent 13 Medium Abnormal Total Percent 10 63 16 High Abnormal Total 2 57 Percent 4 �ENLISTED GROUND The statistically significant asthma-by-exposure-by-age-by-pack-year interaction (p=0.02) shown in Table 20-7 is tabulated in Table R-14. TABLE R-14. Cross Tabulation of Asthma- (Abnormal, Total) by-Exposure Index Category-by-Age Category-by-Pack-Year Category for Ranch Hand Enlisted Groundcrev Exposure Index Pack-Years 50 Age Low Abnormal Total Percent Medium Abnormal Total Percent High Abnormal Total Percent 0 Born >1942 Born 1922-1942 Born <1922 1 0 0 24 18 0 4 0 1 0 0 35 6 0 3 0 1 0 0 15 9 0 7 0 >0-10 Born >1942 Born 1922-1942 Born <1922 0 1 0 37 16 0 0 16 2 1 0 47 7 0 4 14 4 0 0 36 13 1 11 0 0 >10 Born >1942 Born 1922-1942 Born <1922 0 4 0 22 33 3 0 12 0 4 0 0 46 21 1 9 0 0 0 2 1 25 37 5 0 5 20 �The statistically significant pleurisy-by-exposure-by-age interaction (p=0.03) shown in Table 20-7 is tabulated in Table R-15. TABLE R-15. Cross Tabulation of Pleurisy-(Abnormal, Total) by-Exposure Index Category-byAge Category for Ranch Hand Enlisted Groundcrev Exposure Index Age Low Abnormal Total Born >1942 Born 1922-1942 Born <1922 to o 5 7 0 83 67 3 Percent 6 10 0 Medium Abnormal Total Percent 5 0 0 128 34 1 4 0 0 High Abnormal Total 0 5 0 76 59 6 Percent 0 8 0 �The statistically significant wheeze-by-exposure-by-age interaction (p=0.009) shown in Table 20-7 is tabulated in Table R-16. TABLE R-16. Cross Tabulation of Wheezes- (Abnormal, Total) by-Exposure Index Category-byAge Category for Ranch Hand Enlisted Groundcrev Exposure Index Age Low Abnormal Total Born >1942 Born 1922-1942 Born <1922 2 1 0 83 67 3 Percent 2 1 0 Abnormal 1 3 0 Medium Total Percent Abnormal High Total Percent 1 9 0 3 2 0 76 59 6 4 3 0 127 34 1 �The statistically significant wheeze-by-exposure-by-pack-year interaction (p=0.02)' shown in Table 20-7 is tabulated in Table R-17. TABLE R-17. Cross Tabulation of Wheezes- (Abnormal, Total) by-Exposure Index Category-byPack-Year Category for Ranch Hand Enlisted Groundcrew Exposure Index Low Pack-Years Abnormal Total Percent Medium Abnormal Total Percent High Abnormal Total Percent 0 42 0 3 41 0 0 24 0 >0-10 to to 0 2 53 4 3 53 6 1 50 2 >10 1 58 2 1 68 1 4 67 6 �The statistically borderline significant bronchitis-by-exposure category interaction (p=0.08) shown in Table 20-7 is tabulated in Table R-18. TABLE R-18. Cross Tabulation of Bronchitis- (Abnormal, Total) by-Exposure Index Category for Ranch Hand Enlisted Groundcrev Exposure Index Lov AbnormalTotalPercent 25 153 16 Medium Abnormal Total Percent 19 163 12 ' High Abnormal Total 11 141 Percent 8 �APPENDIX S Glossary of Abbreviations �APPENDIX S: Glossary of Abbreviations Contents Table Page S-l Glossary of Technical and Medical Abbreviations S-l S-2 Glossary of Covariate and Statistical Abbreviations S-5 S-i �TABLE S-l GLOSSARY OF TECHNICAL AND MEDICAL ABBREVIATIONS CHECO - replacement Comparisons interviewed by Air Force staff contemporary historical evaluation and combat operations CHOL cholesterol CMI -A- Cornell Medical Index ACA® - Automated Chemical Analyser CNS central nervous system ACTH - adrenocorticotropic hormone COPRO coproporphyrins AFHS - Air Force Health Study CPK creatine phosphokinase ALA - delta-aminolevulinic acid cpm counts per minute ALKPHOS - alkaline phosphatase CUSUM cumulative sum CV coefficient of variation -BBA - Bachelor of Arts BMDP-4F® - log-linear program DBF diastolic blood pressure BMDP-LR® - logistic regression program DNA deoxyribonucl'eic acid BS - Bachelor of Science BSL-2 - National Institutes of Health Biosafety Level 2 ECG electrocardiograph - blood urea nitrogen EEC elec t roencephalogram -D- -E- erythema BUN -C- -F- CAD - coronary artery disease FC fully compliant CAP - College of American Pathology FEF 25-75 forced midexpiratory flow rate CATI - computer-aided telephone interview FEV forced expiratory volume CBC - complete blood count CPU - colony forming unit CHD - coronary heart disease forced expiratory volume in one second FIR CUSUM S-l - fast initial response cumulative sum �TABLE S-l (continued) GLOSSARY OF TECHNICAL AND MEDICAL ABBREVIATIONS FSH follicle stimulating hormone FTA fluorescent treponemal antibody FTI free thyroxine index FVC forced vital capacity ICD-9-CMD - International Classification of Disease 9th Revision, Clinical Modification IQ -K- K-S -GGED General Equivalency Diploma GGTP - Kolmogorov-Smirnov -L- gamma-glutamyl transpeptidase GLUC - intelligence quotient glucose LBBB - left bundle branch block LDH - lactic dehydrogenase LH - leuteinizing hormone -HHB8Ag hepatitis B surface antigen HCT hematocrit HD Hodgkin's Disease HDL high density lipoprotein HGB hemoglobin HLA histocompatibility antigens HPF HRB -M- MCH - mean corpuscular hemoglobin MCHC - mean corpuscular hemoglobin concentration MCV - mean corpuscular volume MI - myocardial infarction ML - malignant lymphoma high-power field MLC - mixed lymphocyte culture Halstead-Reitan battery MMPI - Minnesota Multiphasic Personality Inventory MMPID - MMPI depression scale -II induration IARC International Agency for Research on Cancer -N- ICD International Classification of Disease S-2 NIOSH - National Institute for Occupational Safety and Health �TABLE S-l (continued) GLOSSARY OP TECHNICAL AND MEDICAL ABBREVIATIONS NC - noncorapliant PULM - FEVj/FVC ratio NCI - National Cancer Institute PULSE - pulse index NCVA - nerve conduction velocity above the ankle PWM - pokeweed mitogen NCVE - nerve conduction velocity above the elbow NHL - non-Hodgkin's lymphoma NKC - natural killer cell NORC - National Opinion Research Center NOS - not otherwise specified -QQA - quality assurance QC - quality control QRC - Quality Review Committee -RR RBBB -0- - replacement Comparisons - right bundle branch block 0 - original Comparisons RBC - red blood cell OHR - Optical Mark Recognition RIA - radioimmunoassay RPR - rapid plasma reagin RVN - Republic of Vietnam -PPBM - peripheral blood mononuclear PC - partially compliant PCB - polychlorinated biphenyl PCT - porphyria cutanea tarda PHA - phytohemagglutinin PLAT - platelet count PLT - platelet count PMR - proportionate mortality ratio POMS -' profile of mood states PTSD - post-traumatic stress disorder -SS SAIC - Science Applications International Corporation SAS® - Statistical Analysis System SAS®-GLM - Statistical Analysis System general linear model SCRF - Scripps Clinic and Research Foundation SEA S-3 - shifted original Comparisons - Southeast Asia �TABLE S-l (continued) GLOSSARY OF TECHNICAL AND MEDICAL ABBREVIATIONS SED sedimentation rate SEMEN semen count WAIS - Wechsler Adult Intelligence Scale SCOT serum glutamic-oxaloacetic transaminase WBC - white blood cell SGPT serum glutamic-pyruvic transaminase SIRL Scripps Immunology Reference Laboratory SKIN skin index STS -W- soft tissue sarcoma -T- triiodothyronine 4 total thyroxine TBILI total bilirubin TCDD 2,3,7,8-tetrachlorodibenzo-p-dioxin TEST testosterone TLC total lymphocyte count TSH thyroid stimulating hormone -U- USG urine specific gravity -V- VA Veterans Administration S-4 �TABLE S-2. Glossary of Covariate and Statistical Abbreviations d.f. Adj. RR - differential cortisol DRKYR - current alcohol use (drinks/day) - diabetic class DIFCORT ALC - degrees of freedom DIAB -A- - lifetime alcohol history (drink-years) - adjusted relative risk -B- BATCH - batch-to-batch variation among examination groups BUN - blood urea nitrogen -E- EDUC - education Est. RR - estimated relative risk -C-G- C.I. - confidence interval CC - continuous dependent variable, adjusted by continuous covariates GRP CD - group -H- - continuous dependent variable, adjusted by discrete covariates H - high exposure index level HDL - high density lipoprotein cholesterol CHOL - cholesterol CHOL/HDL - cholesterol to HDL ratio CI - combat index 1C - exposure to industrial chemicals CSMOK - current smoking (packs/day) INS - exposure to insecticides -I- -D- -L- DAY(BATCH) - blood-draw day variation vtihin batch LOG - discrete dependent variable, adjusted by discrete covariates - average lifetime residential latitude - logarithmic - exposure to degreasing chemicals DD - low exposure index level LAT DC L S-5 �TABLE S-2. (continued) Glossary of Covariate and Statistical -M- M medium exposure index level -N- NS not significant -0- OCC occupation OR odds ratio -P- PACKYR lifetime smoking history (pack-years) 2BFAT percent body fat PERSTYPE personality type PS personality score PTSD post-traumatic stress disorder -R- RR relative risk -S- SE standard error SEA ACNE presence of pre-SEA acne SKIN skin color SORT square root SUNREAC sun-reaction index -tfWINE wine consumption S-6 Abbreviations � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 059 Folder The folder containing the original item. 1586 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Lathrop, George D. William H. Wolfe Joel E. Michalek Judson C. Miner Michael R. Peterson Stella G. Machado Theodore G. Karrison William D. Grubbs Wanda F. Thomas Date A point or period of time associated with an event in the lifecycle of the resource 1987-10-01 Title A name given to the resource Air Force Health Study: An Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides, Volume II, First Followup Examination Results, January 1985-September 1987 Subject The topic of the resource Air Force Health Study dioxin morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/f71efe00ef0e23a97920020525184974.pdf eec375a4d3e2e149cf799c9da6235ae0 PDF Text Text Item ID Number 01533 Wolfe > William H. Corporate Author RepOrt/ArtfClB Title Air Force Health Study: Results of the 1985 Morbidity Study Journal/Book Title Year 1987 Month/Day September 25 Color n Number of Images 22 Descrlpton Notes sAM-FM-34-1 Wednesday, May 23, 2001 Page 1584 of 1608 �RESULTS OF TH Y STUDY BRIEFER: COLONEL WILLIAM H. WOLFE SAM-FM-34-l 25 SEP 1987 �AIR FORCE HEALTH STUDY PHASE 111 RESULTS MAY 1985 - MARCH 1986 EXAMINATION OF 2309 PARTICIPANTS • 93% OF ALL PHASE I PARTICIPANTS - 1016. RANCH HANDS - 1293 COMPARISONS CONTRACTORS: • SAIC, MCLEAN, VA • SCRIPPS CLINIC, LA JOLLA, CA • NORC, CHICAGO, IL SAM-FM-34-2 25 SEP 1987 �CHARACTERISTICS OF THE PARTICIPANTS RANCH HANDS COMPARISONS 46.9 YRS 46.8 YRS CURRENT SMOKERS* 40.1% 35.0% ALCOHOL USE SINCE 1982 88.7% 88.7% COLLEGE EDUCATED 48.6% 49.3% RETIRED MILITARY 54.4% MEAN AGE • , 52.8% s I * STATISTICALLY SIGNIFICANT SAM-FM-34-3 25 SE? 1987 �PHYSICAL EXAMINATION FORMAT • GENERAL EXAMINATION • NEUROLOGICAL EXAMINATION • DERMATOLOGICAL EXAMINATION • DOPPLER EVALUATION OF PULSES • CHEST X-RAY • PSYCHOLOGICAL STUDIES • 33 CLINICAL LABORATORY TESTS • IMMUNOLOGICAL STUDIES SAM-FM-34--S 25 SE? 108 �QUALITY ASSURANCE PROGRAM PREDETERMINED STATISTICAL PLAN STRICT LABORATORY QUALITY CONTROL MARK-SENSE FORMS USED FOR DATA COLLECTION • EXAMINERS UNAWARE OF PARTICIPANT STATUS ON-SITE MONITOR PRESENT THROUGHOUT EXAMINATION PERIOD SAM-FM-34-5 25 SEP 1987 �GENERAL HEALTH • OVERALL GROUP SIMILARITY • NO DIFFERENCES IN APPEARANCE OF ILLNESS OR AGE • MORE RANCH HANDS PERCEIVED HEALTH AS FAIR OR POOR (9.1% VERSUS 7.3%) • PRIMARILY SEEN IN ENLISTED PARTICIPANTS • IMPROVED PERCEPTION OF HEALTH FROM THAT IN 1982 • RANCH HANDS HAD LOWER MEAN % BODY FAT • SEDIMENTATION RATE RESULTS MIXED • MEANS EQUAL (5.05 MM/HR VERSUS 4.93 MM/HR) • MORE "ABNORMALS" IN RANCH HANDS (5.8% VERSUS 3.6%) SAM-FM-34-6 25 SEP 1987 �MALIGNANCY • NO GROUP DIFFERENCES OVER INTERVAL PERIOD FOR SKIN OR SYSTEMIC MALIGNANCY • LIFETIME HISTORY OF BASAL CELL SKIN CANCER • CONTINUES TO BE ELEVATED IN RANCH HANDS • RELATIVE RISK = 1.56 • • SYSTEMIC CANCER EQUIVALENT IN BOTH GROUPS • 1 SOFT TISSUE SARCOMA IN EACH GROUP • 1 LYMPHOMA IN EACH GROUP • FINDINGS NOT DISTURBING BUT ARE CAUSE FOR CONTINUED SURVEILLANCE SAM-FM-3-5-7 25 SEP 1987 �NEUROLOGY FEWER NUMBER OF ABNORMALITIES THAN AT BASELINE AGE, ALCOHOL, DIABETES HAD SIGNIFICANT IMPACT TESTS OF CRANIAL, PERIPHERAL AND CENTRAL NERVOUS SYSTEMS ARE EQUIVALENT BABINSKI REFLEX DIFFERENCES AT BASELINE NO LONGER SEEN SAM-FM-34-8 25 SEP 1987 �PSYCHOLOGY HISTORY OF PSYCHOLOGICAL ILLNESS EQUIVALENT MMPI RESULTS: • INCREASED DENIAL AND MASCULINITY/FEMININITY SCALE ABNORMALITIES IN COMPARISONS • MARGINAL INCREASES IN HYSTERIA AND INTROVERSION SCALE ABNORMALITIES IN RANCH HANDS MORE CORNELL MEDICAL INDEX ABNORMALITIES IN RANCH HANDS (COMPLAINTS OF MEDICAL PROBLEMS) ; EQUIVALENT RESULTS ON HALSTEAD-REITAN PERFORMANCE TESTS SIGNIFICANT EFFECTS OF EDUCATION, AGE AND ALCOHOL SAM-FM-34-9 25 SEP 1987 �GASTROINTESTINAL/HEPATIC • EQUIVALENT HISTORY OF LIVER DISEASE AND ULCERS • 11 LABORATORY ASSAYS OF HEPATIC FUNCTION • INCREASED SGPT AND UROPORPHYRINS IN COMPARISONS • INCREASED ALKALINE PHOSPHATASE IN RANCH HANDS • • BORDERLINE ELEVATION OF COPROPORPHYRIN IN RANCH HANDS • NO EVIDENCE OF PORPHYRIA CUTANEA TARDA • ABNORMALITIES EQUALLY DIVIDED IN THE TWO GROUPS < SAM-FM-34-10 25 SEP 1987 �DERMATOLOGY NO CHLORACNE DIAGNOSED EQUIVALENT HISTORY OF SKIN DISEASE NO GROUP DIFFERENCES IN DIAGNOSED SKIN DISORDERS HEMATOLOGY EQUIVALENCE IN 8 LABORATORY TESTS / AGE, RACE, OCCUPATION SMOKING HAD SIGNIFICANT EFFECT SAM-FM-34-11 25 SEP 1987 �CARDIOVASCULAR NO DIFFERENCES IN EGG OR BLOOD PRESSURE PULSE DIFFERENCES SEEN AT BASELINE NO LONGER SEEN INCREASED HISTORY OF HEART DISEASE IN RANCH HANDS (24% VS 20%) NOT ACCOMPANIED BY OTHER ABNORMALITY CHOLESTEROL AND TRIGLYCERIDE LEVELS EQUIVALENT SAM-FM-34-12 25 SEP 1987 �RENAL • NO DIFFERENCES IN HISTORY OF RENAL DISEASE CHANGE NOTED FROM BASELINE • EQUIVALENCE IN 5 LABORATORY TESTS PULMONARY NO DIFFERENCES SEEN EXCEPT IN INCREASED RALES IN RANCH HANDS NO PATTERN SUGGESTING DISEASE DIFFERENTIAL SAM-FM-34-13 25 SE? 1987 �ENDOCRINOLOGY NO DIFFERENCE IN HISTORY OF DISEASE TSH AND TESTOSTERONE INCREASED IN RANCH HANDS (O.7% VS 0.5% AND 1.3% VS 1.1%) INCREASE IN IMPAIRED GLUCOSE TOLERANCE IN COMPARISONS (10.9% VS 14.3%) SAM-FM-34-14 25 Sir 1987 �IMMUNOLOGY 6 CELL MARKER AND 3 FUNCTIONAL TESTS PERFORMED ON 47% OF PARTICIPANTS GROUP EQUIVALENCE ON ALL TESTS SKIN TEST DATA NOT ANALYZED DUE TO EXCESSIVE INTER-READER VARIATION QUALITY CONTROL ENHANCED FOR PHASE III SAM-FM-34-15 25 SEP 1987 �LONGITUDINAL ANALYSES 19 VARIABLES STUDIED EACH PARTICIPANT ACTED AS HIS OWN "CONTROL" GROUP CHANGES OVER TIME ASSESSED 5 SIGNIFICANT CHANGES • SEDIMENTATION RATE ,• BABINSKI REFLEX "N • DEPRESSION (MMPI) I GROUP EQUIVALENCE • PLATELET COUNT f AT FOLLOWUP PULSES (METHOD CHANGE) J NO DETRIMENTS TO RANCH HANDS SAM-FM-34-:6 SEP 19B"? �EXPOSURE INDEX ANALYSES SPORADIC DIFFERENCES SEEN BUT NO CONSISTENT PATTERN EMERGES IN ANY CLINICAL AREA SAM-FM-34-17 25 SEP 1987 �CURRENT EXPOSURE INDEX ESTIMATE BASED ON: • GALLONS OF'HERBICIDE SPRAYED PER MONTH (1962-1971) • LEVELS OF DIOXIN CONTAMINATION • NUMBER OF MEN ASSIGNED EACH MONTH A DIFFERENT INDEX WAS MADE FOR EACH OCCUPATION GROUP NO METHOD TO RELATE ONE TO THE OTHER BETTER ESTIMATE OF EXPOSURE NEEDED SAM-FM-34-19 25 SEP 1987 �t '. „ USAF/CDC COOPERATIVEEFFORTS PILOT STUDY OF 200 GROUNDCREW PERSONNEL DIOXIN HALF-LIFE STUDY OF 100 PAIRS OF SERA (1982/1987) LARGE SCALE TESTING OF AFHS PARTICIPANTS SAM-FM-34-20 25 SEP 1987 �AFHS/CDC SERUM DIOXIN ASSAY STUDIES FIRST 76 ASSAY RESULTS • 23 COMPARISONS - MEAN: 5.3 PPT - RANGE: 3 - 9 PPT • • 53 RANCH HANDS - MEAN: 55.3 PPT - RANGE: 3 - 3 1 4 PPT SAM-FM-34-21 25 SEP 1987 �AFHS SERUM DIOXIN TESTING DRAWING 450cc OF BLOOD AT SCRIPPS CLINIC FROM ALL PARTICIPANTS PERFORM DIOXIN ASSAYS AT CDC SAM-FM-34-22 �CONCLUSIONS SUBTLE BUT CONSISTENT NARROWING OF DIFFERENCES SEEN AT BASELINE REASSURING EVIDENCE THAT CURRENT HEALTH OF RANCH HANDS IS NOT DIFFERENT THAN COMPARISONS NO EVIDENCE TO IMPLICATE LINK BETWEEN HERBICIDE AND ADVERSE HEALTH CONTINUED SURVEILLANCE STILL REQUIRED SAM-FM-34-18 25 StP 1 98 / � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 059 Folder The folder containing the original item. 1583 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Wolfe, William H. Date A point or period of time associated with an event in the lifecycle of the resource September 25 1987 Title A name given to the resource Air Force Health Study: Results of the 1985 Morbidity Study Subject The topic of the resource Air Force Health Study study protocol morbidity trends mortality trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/beed241e785b94fc9bd023e5a06109dc.pdf 3523666342d3d5bb26133cf78991dd4d PDF Text Text ItomlDNunber °1867 Author Anderson, Ron J. Corporate Author Texas Veterans Agent Orange Assistance Program, Tex Report/Article Title Annual Report Journal/Book Title Year 1985 Month/Day Au ust Color n Number of Images 91 9 Descrlpton Notes Wednesday, July 11, 2001 Page 1868 of 1870 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH AUSTIN, TEXAS A N N U A L August Ron J. Anderson, M.D. Chairman Texas Board of Health R E P O R T 1985 Robert Bernstein, M.D., F.A.C.P. Commissioner of Health Texas Department of Health �1036-1986 Texas Department of Health Robert Bernstein, M.D., F.A.C.P. Commissioner Robert A. Maclean, M.D. Deputy Commissioner Professional Services Hermas L. Miller Deputy Commissioner Management and Administration 1100 West 49th Street Austin, Texas 78756-3199 (512)458-7111 August 23, 1985 Members of the Board Ron J. Anderson, M.D., Chairman Laurance N. Mickey, M.D., F.A.A.P., ViceChairman Bob D. Glaze, D.C., Secretary Johnnie M. Benson, F.A.C.N.H.A. Sister Bernard Marie Borgmeyer, R.N., F.A.CH.A. Frank Bryant, Jr., M.D., F.A.A.F.P. Joaquin C. Cigarroa, Jr., M.O. Barry D. Cunningham, D.O.S. Ben M. Durr, M.H.A. Dennis K. Mclntosh, O.V.M. Robert D. Moreton, M.D., F.A.C.R. Joe N. Pyle, P.E. Arthur L. Raines, M.D. Isadore Roosth Barbara T. Slover, R.Ph. Max M. Stettner, D.O. Edward H. Zunker, O.D. The Honorable Mark W. White Governor of Texas State Capitol Austin, Texas 78711 Dear Governor White: Enclosed is the Annual Report of the T e x a s V e t e r a n s Agent Orange A s s i s t a n c e Program. Pursuant to Section 3 of Article U447w, VTCS, this report is being distributed to the Legislature, V e t e r a n s Administration, T e x a s V e t e r a n s Affairs Commission, veterans' organizations, and interested individuals. It reflects the work of the Texas Department of Health Agent Orange Program to date. The report contains research findings on the effects of exposure to chemical defoliants or herbicides or other causative agents, including Agent Orange, and statistical information compiled from reports submitted by physicans, hospitals, and veterans. I hope you find this report both useful and informative. Sincerely, Commissioner of Health Enclosure �18J6-I986 Texas Department of Health Robert Bernstein, M.D., F.A.C.P. Commissioner Robert A. Maclean, M.D. Deputy Commissioner Professional Services 1100 West 49th Street Austin, Texas 78756-3199 (512)458-7111 Members of the Board Ron \. Anderson, M.D., Chairman Laurance N. Nickey, M.D., F.A.A.P., Vice-Chairman Bob D. Glaze, D.C., Secretary Johnnie M. Benson, F.A.C.N.H.A. Sister Bernard Marie Borgmeyer, R.N., F.A.C.H.A. Hermas L. Miller Deputy Commissioner Management and Administration Frank Bryant, Jr., M.D., F.A.A.F.P. Joaquin C. Cigarroa, Jr., M.D. Barry D. Cunningham, O.D.S. Ben M. Durr, M.H.A. Dennis K. Mclntosh, D.V.M. Robert D. Moreton, M.O., F.A.C.R. Joe N. Pyle, P.E. Arthur L. Raines, M.O. Isadore Roosth PREFACE Barbara T. Slover, R.Ph. Max M. Stettner, D.O. Edward H. Zunker, O.D. From 1962 to 1971 during the Vietnam conflict, 152,000 Texans serving in the m i l i t a r y forces were exposed to v a r y i n g amounts of herbicides used to kill or d e f o l i a t e plants. Since t h a t t i m e v e t e r a n s have a t t r i b u t e d a n u m b e r of illnesses to Agent Orange which c o n t a i n e d a c o n t a m i n a t i n g chemical (TCCD); known to be highly toxic to animals, yet not well understood in its effects on humans. The T e x a s V e t e r a n s A g e n t O r a n g e A s s i s t a n c e P r o g r a m set into m o t i o n a cooperative program between the Texas Department of Health and the U n i v e r s i t y of Texas System to assist these veterans in establishing claims through pilot clinical studies designed to establish the cause and effect relationship of exposed v e t e r a n s and s u b s e q u e n t h e a l t h p r o b l e m s . The U n i v e r s i t y of Texas Agent Orange Project has selected 248 for study from whom 927 specimens have been a n a l y s e d in the v a r i o u s protocols. To date no f i n a l r e s u l t s of the s t u d i e s have been released since i n t e r p r e t a t i o n of the i n d i v i d u a l s t u d y reports requires correlation with controls and the results of the project as a whole. Activities at the federal level have increased, as evidenced by the activities of the Agent Orange Epidemiological Study at the Centers for Disease Control, the VA Chloracne Task force and completion of the morbidity study phase of the Air Force R a n c h H a n d Study, and the r e l e a s e of the B i r t h D e f e c t S t u d y in Atlanta by the Centers for Disease Control. Robert Bernstein Commissioner of Health �TABLE OF CONTENTS TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH . ANNUAL REPORT August 1985 Status Report and Data Sheet, July 31, 1985 Summary of Pilot Studies Protocols (FY 84-85 studies) Report on "Development and Preliminary Results of Pilot Clinical Studies, March 26, 1984 (FY 82, 83, 84 studies) Analysis of Major Demographic Statistics, May 3, 1985 (FY 82, 83, 84 studies) Herbicide Status Report by Department of the Army Veterans Health Survey—Report on Agent Orange studies conducted by the Centers for Disease Control, July 1985 �Texas Department of Health 1836-1986 Robert Bernstein, M.D., F.A.C.P. Commissioner 1100 West 49th Street Austin, Texas 78756 (512) 458-Trtr 7251 Robert A. Mac Lean, M.D. Deputy Commissioner Professional Services Hermas L. Miller Deputy Commissioner Management and Administration DATE: TO: FROM: RE: August 5, 1985 INTERESTED INDIVIDUALS AND ORGANIZATIONS HARRIET FRANSON, Program Manager Agent Orange Program TEXAS AGENT ORANGE PROGRAM STATUS REPORT FOR PERIOD ENDING JULY 31, 1985 Enclosed is the Texas Agent Orange Program Status Report for period ending July 31, 1985. This report is cumulative and reflects program activities since the inception of the program on September 1, 1981. As you may be aware, the Texas Legislature this year did not approve continued funding for the Agent Orange Program. Therefore, funding will expire on August 31, 1985 with resulting curtailment of program activities. Program data analyses are anticipated and study results published. It is our understanding that a nationwide search is underway for individuals to assist in the distribution plan for the $180 million settlement fund approved by the New York State Court in the class action suit against the seven chemical companies. Appointments will be made to an Advisory Group for the payment program, an Executive Director, and a Board of Directors for the Foundation. Names with resumes can be forwarded to Kenneth R. Feinberg, Suite 1150, 1575 Eye Street, N.W., Washington, D.C. 20005. The resumes should reflect relevant background and experience (see attached request). �^•-KAYE, SCHOLER, FIERMAN, HAYS & HANDLER 1575 EYE STREET, N.W. NEW YORK OFFICE 4«« PARK AVENUE NEW YORK. N.Y. IOOJI (til) 4 O 7 - 8 O O O FLORIDA OFFICE 125 WORTH AVCNUC PALM BEACH. FLA. J34* WASHINGTON, D.C. 2 O O O 5 CA1LE ADDRESSES KAYEMACLCR WASHINGTON KAYEMACLER NCW YORK (2O2) 783-I2OO TELEX HUMIERS WASHINGTON a»'*S8 NEW YORK DOMESTIC 128911 NEW YORK INTX 234860 HOMO KONG 62818 •AY NX (3O5) 8 J J - 5 I B I MONO KONO OFFICt CDINCUROH TOweft «OYH FLOOR IB OUCCN'S ROAO CENTRl HOMO KONO For Release: June 7, 1985 PRESS RELEASE Kenneth R. Feinberg, the Special Master reappointed by the Court in the Agent Orange litigation to help develop a distribution plan for the Settlement Fund, announced today that a nationwide search for individuals to assist in implementing the distribution plan approved by the Court would be undertaken as the first step in the distribution of the $180 million Fund. The settlement is between the seven defendant chemical companies and the plaintiff class, which consists of those veterans who served in or near Vietnam from 1961 to 1972, who were exposed to Agent Orange and have injuries allegedly related to that exposure. The class also includes spouses and children of the veterans. Over 240,000 claims have been filed with the Court by class members seeking to participate in the settlement distribution. The Court order requires a $150 million cash compensation program for veterans exposed to Agent Orange who are longterm totally disabled or to the families of those who have died, and a $45 million foundation to provide grants for services to the class, including those children of the class members suffering from birth defects. The Court ordered that members of the class play a significant role in the governance of all aspects of the distribution plan. To that end, the Court ordered the Special Master to take appropriate steps leading to the recommendation of names to the Court for appointment to an Advisory Group for the payment program and a Board of Directors of the foundation. Both the Advisory Group and the Board of Directors are to be comprised primarily, but not exclusively, of class members. �KAYE,SCHOLER. FIERMAN, HAYS & HANDLER - 2 - June 7, 1985 The Special Master announced today that he is soliciting names from all interested groups or individuals who would be willing to serve either as an Advisory Group member or on the Board of Directors. The names will then be submitted to the Court for final review and appointment. All Advisors and Board members will serve without compensation, other than reimbursement of reasonable travel and other per diem expenses. The Advisory Group and the Board will be as representative of the class as possible, cutting across economic, social, racial, gender, geographic and occupational lines. Persons with management, investment, budget and foundation experience would be particularly desirable as members. Resumes showing relevant background and experience should be included with any suggestions of persons for consideration. In another aspect of the outreach effort, Mr. Peinberg announced that a nationwide search would begin for an experienced professional to serve as Executive Director of the foundation. The Executive Director would administer the day-to-day operations of the foundation and would be compensated from the Fund. The Court will initially appoint the Executive Director, who will then serve at the pleasure of the Board of Directors. All inquiries or suggestions for the Advisory Group, the Board of Directors or the Executive Director should be directed, in writing with supporting data, to Kenneth R. Peinberg, Suite 1150, 1575 Eye Street, N.W., Washington, DC 20005. The final outreach effort announced by the Special Master*is the solicitation of insurance companies or other parties interested in bidding on contracts to implement the $150 million payment program. Contracts .for claims processing, investment consulting, claims adjudication, and auditing will be finalized by the Court within the next few months. Contractors interested in receiving bid specifications or in obtaining information concerning the payment program should contact, in writing, Lawrence B. Novey, consultant to the Special Master, who can be reached at the same address as Mr. Feinberg. �IfcAAbUtPAKIMtlN I Uh HtALI H AUSTIN THROUGH: CHIEF, BUREAU OF EPIDEMIOLOGY THROUGH: ASSOCIATE COMMISSIONER FOR PREVENTABLE DISEASES FROM TEXAS INTER-OFFICE THROUGH: DEPUTY COMMISSIONER FOR PROFESSIONAL SERVICES GEORGE R. ANDERSON, M.D. OCCUPATIONAL MEDICINE AND TOXICOLOGY/ AGENT ORANGE PROGRAM Robert Bernstein, M.D., F.A.C.P. Commissioner of Health TO Page 1 SUBJECT TEXAS VETERANS AGENT_ORANGE ASSISTANCE PROGRAM STATUS REPORT FOR 2 MONTH PERIOD ENDING JULY 31, 1985 REFERRALS No. of veterans referred into the program this reporting period (No.-of deceased veterans—1: TOTAL 22) TOTAL TO DATE (6/1/85-7/31/85) 1,962 29 Military and medical records have been requested for all referred veterans: Medical records reviewed to date: (Include VA and civilian records— 105 reviewed this reporting period) 2,073 Military records reviewed to date: (include combat history, DD211, and/or medical—205 reviewed this reporting period) 1,795 No. of veterans referred into program and not in compliance with residency requirements—ineligible 18 CONTACTS Direct contacts from veterans this reporting period 1,352 39 By phone—31 (total to date: 981) By letter—6 (total to date: 326) By visit—2 (total to date: 51) Contact from News Media: Channel 7 TV (Austin) Bryan Eagle (Bryan) 114 Boston Globe Contact from or with other states/countries: Massachusetts (1) Orgeon (1) Washington (1) 287 West Virginia (1) Wisconsin (3) StCNB> DATi _ -CONTINUEDAugust 2, 1985 FORM NO. AG-2-A �TEXAS DEPARTMENI Oh HEALIH AUSTIN TEXAS INTER-OFFICE FROM George R. Anderson, M.D. TO Robert Bernstein, M.D., F.A.C.P. Pa 8e 2 TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM STATUS REPORT FOR 2 MONTH PERIOD ENDING JULY 31, 1983 Continuing contact with Legislative offices (State Representatives Larry Don Shaw 4 Jerry Yost) Office of the Governor, Office of the Attorney General, State Auditor, Congressman J.J. Pickle, Texas Department of Corrections, Texas Veterans Affairs Commission, Texas Land Commission, U n i v e r s i t y of Texas System, V e t e r a n ' s A d m i n i s t r a t i o n , Vet Centers, Military Personnel Records Center, County Veteran Services Officers, Local Health D e p a r t m e n t s / C l i n i c s , Other State Agent Orange offices, counseling services/physicians/hospitals, veterans' organizations, Dow Chemical Company, law firms, and students 9 f o l l o w u p letters were sent this reporting period to veterans who p r e v i o u s l y inquired about the program but not yet participating. (TOTAL TO DATE: 684) Made/mailed 112 followup phone calls/letters to check on military/medical records requested but not yet received. (TOTAL TO DATE: 1,851) One feedback letter was sent this reporting period to v e t e r a n in our program to apprise him of the status of his case ( m i l i t a r y / m e d i c a l records received, pending, etc.). (TOTAL TO DATE: 540) 10 veterans in the program requested or were placed on inactive status this r e p o r t i n g period, p r i m a r i l y due to i n d i v i d u a l s m o v i n g w i t h no forwarding addressess available (TOTAL TO DATE: 127) Inactive veterans resuming participation in the program. (TOTAL TO DATE: 6) In response to our m a i l i n g to Texas v e t e r a n s on the VA Agent Orange Registry received 0 completed questionnaire (TOTAL TO DATE: 1,511) of which 0 asked to be registered with the Texas Agent Orange Program (TOTAL TO DATE: 1,217). 5 veterans requested and were sent copies of case file records, in preparation for filing a claim: (TOTAL TO DATE: 27) SIGNED DATE -CONTINUEDAugust 2, 1985 FORM NO. AG-2-A �UtrAKIMtIN I Oh HtALI M AUSTIN TEXAS INTER-OFFICE HIOM George R. Anderson, JJ.JK TO Robert Bernstein, M.D., F.A.C.P. SUBJECT TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM STATUS REPORT FOR 2 MONTH PERIOD ENDING JULY"31,1985 Page 3 PROTOCOL STUDIES The second phase of the clinical studies has begun, with the following studies conducted this fiscal year: Cytogenetics at UTS CANCER CENTER, Houston, by Dr. Hsu Bleomycin Test at UTS, CANCER CENTER, Houston, by Dr. Hsu Immune Profile at UT HEALTH SCIENCE CENTER, Houston, by Dr. Kerman Uroporphyrins at UT HEALTH SCIENCE CENTER, Houston, by Dr. Kerman Aryl Hydrocarbon Hydroxylase Induction at UT MEDICAL BRANCH, Galveston, by Dr. Ward The protocols were published in summary and complete format. Questionnaires received from selected veterans and proposed controls continue to be reviewed to establish proper matching of veterans with controls. 17 volunteer control questionnaires for the Agent Orange clinical studies were received. (TOTAL TO DATE: 148, of which 1 is TDH employee. Contacts made with selected veterans and controls to make appointments for the collection of specimens. Contacts made with the clinics/laboratories where specimens collected/delivered. are to be 53 appointments arranged for the collection of specimens (TOTAL TO DATE: 304). Total specimens collected for the 2nd/3rd collection of specimens for the Sperm Study (TOTAL TO DATE: 232). No reminder letters were sent re. collection of specimens (TOTAL TO DATE: 31). Collection of specimens for the Sperm Study is now completed. SIGNED DATE -CONTINUEDAugust 2. 1985 FORM NO. AG-2-A �ItAAS LJtrAKIMtlNI UMItALIH AUSTIN TEXAS INTER-OFFICE George R. Anderson. M.D. reOM SUBJECT TO Robert Bernstein, M.D., F.A.C.P. Pa 8e 4 TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM STATUS REPORT FOR 2 MONTH PERIOD ENDING~TULY 31, 1985 111 letters were mailed to veterans concerning their participation in the clinical studies (TOTAL TO DATE: 159) and 55 to proposed controls (TOTAL TO DATE: 226). Number of specimens collected and shipped to UTS: (6/1/85-7/31/85) TOTAL TO DATE " CYTOGENETICS STUDY 23 238 IMMUNE SUPPRESSION STUDY UROPORPHYRIN AHH (Enzymes) SPERM STUDY SPECIMEN N O . 2 SPECIMEN N O . 3 23 23 13 0 237 72 58 126 FAT TISSUE SPECIMEN 0 0 9 9 0 9 5 2 Two v e t e r a n / c o n t r o l s requested and were given results of i n d i v i d u a l study specimens analyses. (TOTAL TO DATE: 156) SELECTION PROCESS FOR REFERRAL TO THE UTS SYSTEM Review of cases is an ongoing process for eventual referral to the Agent Orange Selection Committee—700 were reviewed this period for referral to the committee. To d a t e the Selection C o m m i t t e e has r e v i e w e d 1,103 cases (117 being r e v i e w e d more than once), of w h i c h 2U8 have been selected for the clinical studies (of which 126 are for inclusion in the second study phase). 95 v e t e r a n s have also been selected as possible low-risk controls. BROCHURES/POSTERS To date a p p r o x i m a t e l y 35,500 brochures and 7,726 posters have been mailed. In addition to individual requests, brochures and posters have been p r o v i d e d to v e t e r a n s ' o r g a n i z a t i o n s , c o u n t y service o f f i c e r s , clinics/hospitals, and other states. SIGNED DATE -CONTINUEDAugust 2. 1985 FORM NO A<-,-•>. A �UtrAK I /VltIN I Ur MtALI M AUSTIN TEXAS INTER-OFFICE FROM George R. Anderson^J1.JK ____ TO Robert Bernstein, M.D., F.A.C.P. SUBJtCT _ J^MO^l^liy^AGjra^ ____ STATUS REPORT FOR 2 MONTH PERIOD ENDING JULY 31, Page 5 MAINTAINING STATISTICAL INFORMATION I n f o r m a t i o n is compiled each m o n t h from case files concerning the following medical conditions reported and substantiated by medical records. This information is provided to the Agent Orange Selection Committee and becomes part of our data information. Such information will be compiled for other medical conditions as the need arises. Cancer in Veterans Under Age 36 Cancer in Veterans Over Age 36 Tingling/Numbness in Extremities Post Traumatic Stress Disorder (PTSD) Current Rashes Children with Leg Deformities Miscarriages/Stillbirths Schizophrenia Diagnoses continue to be coded w i t h I n t e r n a t i o n a l Code for c o m p u t e r entry. AGENT ORANGE ADVISORY COMMITEE No meetings were held during this reporting period. SPECIAL ACTIVITIES Continue review of available literature for research on Agent Orange and related topics. Continue to purchase publications for reference library. E x t r a c t i o n of statistical data from case files concerning specific military data and medical conditions, etc. Utilize word processor for the storage/retrieval of data and for multiple reproduction of originally-typed letters when form letters are not warranted. In-house t r a i n i n g on use of computer equipment for R i c h a r d Smith and Harriet Franson (d-Base and Software Users Group) continues. SIGNED DATE -CONTINUEDAugust 2. 1985 FORM NO. AG-2-A �i CA/\a i>cr/m i IVICIN i isrncs\Lin AUSTIN TEXAS INTER-OFFICE FROM George R. Anderson. M.D. _________ TO Robert Bernstein, M.D. , F.A.C.P. SUBJECT TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM STATUS REPORT FOR 2 MONTH PERIOD ENDING JULY 31, 1985 _ Pa 8e _____ 6 Received "Wisconsin Vietnam Veteran Mortality Study" for review Ongoing communication with other state Agent Orange Commissions/Programs as their representative on the VA Advisory Committee on Health-Related Effects of Herbicides. One summer employee is assisting with coding and data entry for the Epidemiological Study. A senior citizen volunteer is also assisting the program on a limited basis. MAJOR ACCOMPLISHMENTS 1. Number of veterans in the program has increased to 1,962 — an increase of 1,565 since the beginning of FY 84. 2. To date 1,243 cases have been reviewed by the Subject Selection Committee, of which over 248 have been selected for referral to the U n i v e r s i t y of Texas clinical studies. A total of 906 blood/sperm specimens have been collected and shipped to the U n i v e r s i t y of Texas System laboratories and one fat tissue shipped for analysis in the V.A./E.P.A. Study of Dioxin Levels in Human Adipose Tissue. MEETINGS ATTENDED None attended or scheduled. Attachment — Data Sheet cc: Agent Orange Selection Committee Agent Orange Advisory Committee Veterans' Organizations and other • interested individuals SIGNED DATE August 2, 1985 FORM NO. AG-2-A �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH DATA SHEET (as of July 31, 1985) NOTE: Variation in totals is due to receipt of only questionnaires to date and/or lack of medical/military information. \ In some instances, the initial complaint and those listed under "Other Medical Problems" were supplied by the veteran rather than a physician. BY (When entering AGE program) 29 30 2 7 32 33 34 35 36 37 38 39 40 II 74 117 134 169 120 97 72 56 BY SEX MALE BY SERVICE Army FEMALE 4 791 Air Force 1958 118 Marines 193 BY RACE WHITE BLACK 244 HISPANIC NaVy 733 277 41 31 42 43 44 45 46 47 48 49 50 36 20 30 25 17 17 16 20 20 51 52 23 20 NO. OF DECEASED VETERANS REPORTED INTO THE PROGRAM 53 54 55 56 57 58 5 9 60 61 20 15 9 11 7 9 8 3 1 ' OTHER 9 NO. REPORTED INTO THE PROGRAM AND DETERMINED NOT TO BE ELIGIBLE ' ' ' 62 5 1 1 2 3 1 1 1 PROGRAM AND RESIDING IN ANOTHER STATE 18 NO. REPORTED INTO THE 63 64 65 66 68 70 72 22 I 26 5? �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) BYjCOUNTY Anderson 5 Andrews 1 Angelina 1 Aransas 3 Bandera 3 Bastrop 14 Bee 3 Bell 50 Bexar 101 Bosque 3 Bowie 9 Brazoria 12 Brazos 8 Brooks 1 Brown 1 Burleson 3 Burnet 5 Calhoun 2 Callahan 2 Cameron 13 Cass 5 Castro 2 Chambers 1 Cherokee 5 Collin 7 Collingworth 2 Comal 3 Concho 1 Cooke 4 Coryell Crockett Dallas Deaf Smith Denton Dewitt Dimmit Donley Duval Ector Ellis El Paso Falls Far-.nin Fayette Fisher Fort Bend Gaines Galveston 10 1 104 1 6 1 1 2 1 6 3 109 3 2 2 1 7 2 21 Goliad Grayson Gregg Guadalupe Hale Hamilton Hardeman Hardin Harris Harrison Haskell Hays Henderson Hidalgo Hill Hockley Hood Hopkins Howard Hunt Hutchinson Jackson Jasper Jefferson Jim Wells Johnson Karnes Kaufman Kendall Kerr Kimble Kleberg Lamar Lamb Lampasas Lavaca Leon Liberty Llano Lubbock Lynn Marion Matagorda Maverick McCullock McLennan Medina Midland 2 11 3 2 1 1 1 2 118 3 1 2 2 26 2 1 2 2 4 6 1 1 1 15 5 5 1 7 2 5 1 3 2 1 3 3 1 6 2 12 1 2 2 1 1 9 2 6 Milam Montague Montgomery Moore Morris Nacogdoches Navarro Newton Nueces Ochiltree Orange Palo Pinto Parker Parmer Potter Randall Reeves Richmond Robertson Rusk San Jacinto San Patricio Shelby Smith Starr Tarrant Taylor Tom Green Travis Tyler Upshur Uvalde Val Verde Van Zandt Victoria Walker Ward Webb Wharton Wichita Wilbarger Willacy Williamson Wilson Winkler Wise Wood Young Zapata 2 1 8 1 1 1 2 'I 53 1 9 1 3 1 10 5 1 1 1 3 2 12 1 6 1 71 3 5 82 1 6 3 6 4 7 38 4 6 1 11 1 1 14 2 1 5 2 1 1 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) DUTY PERFORMED DUTY PERFORMED Accounting Specialist Administrative Specialist Administrative & Supply ADP Officer Aerial Photo Interpretator Airborne Infantry Aircraft Technician Air Crew Air Frame Repair Specialist Air Mobile Air policeman Air Operation Supervisor/Spec. Air Traffic Control Ammunition Ammo Cargo Handler Armor Unit Artillery Base Maintenance Battalion Clerk Boatswain Mate Calibration Team Technician Career Counselor Cargo Handler Carpenter Chaplain Chemical Operations Combat Cook Combat Engineer Combat Military Police Combat News Correspondent Communications Specialist Construction Controller Convoy Escort Corpsman Counterinsurgency Specialist Courier Coxswain Crane Operator Crew Chief Deck Force Demolition Expert Dining Facility Manager Engineering Equipment Repair Explosives Finance Firefighter Food Service Forward Air Control Freight Handler Grave Registration Ground Crew Guard Gunfire Spotter 1 10 5 1 2 3 42 71 1 2 1 1 2 6 2 12 62 1 2 5 1 2 6 3 3 11 20 22 1 1 30 10 1 4 7 1 1 1 1 5 2 3 2 15 4 2 3 2 Gunner's Mate Harbor Defense Infantry Inspector General Intelligence Interrogator Investigator, Narcotic Journalist Lineman Machinist Maintenance Mechanic Medic Medical Advisor Medical Clerk Medical Corps Mess Steward Meteorologist Military Advisor Military Police Musician NCO Neuropsy Specialist Nurse Operator, Heavy Equipment Paratrooper Personnel Officer Petroleum Storage Supply Photographer Pilot Platoon Leader Plumber Polelinetnan Powerlineman Printer Guard Psychological Operations Radar Operator Radio Operator Radio Repair (field) Recon. Infantry Recon. Forward Observer River Rat 3 6 2 1 4 6 1 1 7 1 456 1 18 2 1 1 4 1 20 32 24 4 1 2 1 1 2 17 1 1 1 3 11 3 4 4 2 16 6 3 1 2 1 3 4 9 3 5 11 3 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS.DEPARTMENT OF HEALTH, AUSTIN, TEXAS DUTY PERFORMED Sea Bees Security Guard Sergeant/Clerk Ship Crew Ship Engine Man Signal Corps Small Missile Repairman Social Worker Physical Specialist Special Forces Advisory Group Supply Sergeant Supply Specialist Support Battalion Switchboard Operator Tank Crewman Telephone Repair Translator/Interpreter Transportation Truck Driver Tunnel Rat Warehouseman Watercraft Operator Weapons Mechanic Wireman 3 5 2 2 1 9 1 1 6 21 36 11 1 9 2 1 23 33 1 4 1 2 5 NO OF VETERANS REPORTING MISCARRIAGES/STILLBIRTHS 329 NO OF VETERANS REPORTING CHILDREN WITH BIRTH DEFECTS ' AND/OR MEDICAL PROBLEMS PRESENT SINCE BIRTH 307 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) CURRENT OCCUPATION CURRENT ^OCCUPATION Accountant Active Duty . Administrative Aircraft Air Conditioning/Refrig. Contractor Apartment Manager Applied Research Lab Army Depot Attorney Automotive Banking Barber Bellman Biomedical Engineering Technician Border Patrol Building Inspector Cable Company Carpenter Carpet Installer Cement Company Chemical Company Child Care Chrome Plater City Employee Civil Service Clergy Clerical Computers Construction Consultant Cook Correctional Institution Counselor Cowboy Custodian Disability Examiner Disabled, medically unemployed Draftsman Editor, publication Education Specialist Electrical Supply Electrician Electrician, Naval Aviation 11 3 34 13 4 2 1 4 3 1 2 3 1 2 1 1 4 18 1 1 6 2 1 8 8 3 8 6 28 5 5 1 2 2 11 1 96 3 1 1 1 14 1 Electric Technician Electronics Technician Employment Interviewer Engineering Environmentalist Equipment Operator Executive Executive, Oil Field Fence Builder Firebrick Company Fire Dept. Fisherman Floor Finisher Food Service Funeral Home Furniture Restoration Gas pipeline operator Glazier Grocer Hair Stylist Helicopter Technician Highway Dept. Inmate Inspector Quality Control Insurance Claims/Agent Investigator, State Ironworker IRS Laborer Laundry Lawman Legal Assistant Library Lineman Lumber Mill Worker 3 14 1 6 1 15 3 1 1 1 9 2 1 7 2 1 1 2 1 1 1 1 46 2 8 5 4 1 14 2 27 1 1 4 2 Machinist 18 Maintenance Management Analyst Meat Packer Mechanic Medical Assistant Medical Lab Tech. Military Base Millwright Mobile Court Owner Musician Newspaper Carrier 19 1 1 52 1 1 1 3 1 1 1 Nurse 2 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31 • 1985) CURRENT OCCUPATION CURRENT OCCUPATION <;ti short Oil Oilfield Operating Room Technician Optometry Oxygen Plant Painter Parks Service Pest Control Pharmacy Tech. Photographer Physical Therapist Physician Pipefitter Planner Estimator Plant Operator Plumber Porter Post Office Printer Private Investigator Probation Officer Production Psychologist Purchasing Railroad Ranching/Farming Real Estate Recruiter (service) Refinery/Boilermaker Rehabilitation Center Repair electrical equipment Retired Sales Sanitarian Sawmill Operation Seaman Security Self-Employed Service Station Shipping Clerk Shrimper Silver Smith Slaughterhouse Speech Therapist State Employee Steel Company Stocker 3 12 1 1 1 4 1 1 3 2 1 3 5 1 4 12 1 53 5 2 3 1 1 1 13 4 3 1 5 5 1 24 48 1 1 1 24 11 2 5 1 1 1 1 1 7 1 Store Manager Student Supervisor, Computer Supervisor, Production Supply Clerk Teacher Telephone Company Tool & Dye Tree Surgeon Typewriter Repair Unemployed Upholsterer Utility Company Vehicle Driver Warehouseman Welder Woodworker Writer '"> 14 4 9 1 18 7 1 1 1 110 2 6 63 13 26 4 1 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) INITIAL COMPLAINT (as reported by the veteran) Abdominal Pain Acne Allergies Anxiety Apnea (shortness of breath) Arthritis Asphyxia Asthenia (weakness) Atrophy Back Pain Birth Defect, child Blackouts Blood Disorders Body Aches Cancer Chest Pain Chloracne Confusion Constipation Cysts Depression Diabetes Diarrhea Dizziness Dysphasia (speech impairment) Dyspnea (labored breathing) Edema Emotional Problems Epilepsy Fatigue Fever, recurring Gastritis, chronic Gastrointestinal disorders Hair LOSS Headaches Heart murmur Hematoma Hepatitis, recurrent Hepatomegalia Hives Hyperlipidemia Hypertension Infected prostate Infections 3 17 6 12 10 8 1 9 1 6 37 2 10 4 62 15 10 1 2 9 17 1 9 13 1 4 11 52 1 10 5 1 84 7 107 1 1 2 1 2 1 24 2 11 Itching Joint pain Kidney Lesions Lethargy Liver damage Liver pain Loss of appetite Low potassium level Low resistance to disease Lumps on body Lumps in scrotum Lung disease Memory impairment Miscarriage Multiple sclerosis Muscle spasms Nausea Nerve problems Neuralgia (nerve pain) Numbness Paralysis, extremities Personality change Pneumothorax Pruritus, intense (itching) Rages Rash Rectal bleeding Renal failure, chronic Respiratory problems Reversed sperm travel Seizures Sensorial impairments Sexual problems Skin infection Skin blistering/peeling Skin pigmentation, loss of Sleeplessness Sores/boils Sore throat, chronic Sperm count—low Stroke Sunlight allergy Tendenitis 17 28 10 10 3 17 4 2 2 3 21 1 4 8 7 3 7 2 100 1 123 4 2 1 1 2 524 5 1 5 1 5 1 35 13 36 10 55 18 3 4 1 2 1 Tingling in extremities Tumors, skin Vomiting blood Ulcer Urination frequency Visual disturbance Weight loss/gain Withdrawal regression 42 5 5 8 5 11 5 1 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) PIAG_NOSIS T(JDx)—as reported by physician Aberrant innervation of third cranial nerve (child) Acalculous choleystitis Acre Acre keloidalis ruchae Acrocyanosis Acrokeratosis verrucofomis Actonic keratoses Adenoma, villous Agent orange symptomatology Alcoholism Allergic rhinitis Alopecia areata Amebic liver abscess Amputation, fingers, congential (chlla) Anemia Ankylosis spondylitis Anxiety, chronic Anxiety neurosis Aortic insufficiency Apnea (cessation of breath) Arteriosclerosis, advanced (carotids & femorals) Arthralgias (joint pain) Arthritis Arthritis, cervical Arthritis, degenerative Arthritis, gouty Arthritis, poly, seronegative Arthritis, post-traumatic Arthritis, rheumatoid Arthritic changes of joints Aspermia Asthma Atherosclerotic occlusive peripheral vascular disease Atrophy of kidney Azoospermia Baker's cyst Barlow's syndrome Bell's palsy Bilateral acanthosis Bilateral internal tibial torsion (child) Bilateral rnetatarus adductus (child) Bilateral mandibular tori 1 1 10 1 1 1 1 1 2 15 2 3 1 1 4 2 17 11 1 5 2 12 16 5 12 5 1 3 2 1 1 1 1 1 3 1 .1 1 1 2 1 1 Bilaterial calycealcalculi Bilateral supernumerary fingers, non-boney (child) Bipolar disorder Bowen's disease Brain syndrome, chronic Bronchitis, chronic Bullous emphysema Buerger's disease Bursitis Calcaneovarus deformity of feet (child) Cancer (Neoplasra)-total 74 Adenocarcinoma of esophagus Adenocarcinoma of rectum stage Dukes C w/ 5/15 lymph nodes positive for Ca Adenocarcinoma of rectum, Duke C Adenocarcinoma of rectum w/seeding of pararectal fat Adenocarcinoma of sigmoid colon Adenocarcinoma of sigmoid colon with metastasis Basal cell Bladder, low grade Bronchogenic carcinoma, squamous cell Carcincoma of esophagus Carcinoma of rectum w/metastasis to liver Differentiated lymphocytic lymphoma, nodular type, Stage 4 w/widespread metastasis Diffuse bilateral adenocarcinoma w/metastases multiple areas bone/brain Embryonal cell carcinoma, Stage 1 w/teratoma, left testicle Epidermoid carcinoma of esophagus Esophagus Fibrohistiocytoma (leg) Giant cell tumor of bone Glioblastoma multiforme, brain Hodgkin's disease Larynx Malignant melanoma Metastatic carcinoma in hilar lymph node Metastatic embryonal cell carcinoma (testis) 1 1 1 1 1 8 1 2 3 1 1 1 1 1 1 4 8 1 1 1 1 1 1 1 1 2 1 1 1 2 2 3 1 1 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) DIAGNOSIS (Dx)—as^reggrted by physician Metastatic malignant melanoma, bowel and subcutaneous tissue/ renal cell carcinoma, left kidney 1 Metastatic melanoma lesion, right parieto-occipital 1 Metastatic squamous carcinoma to scalene node 1 Mixoid liposarcoma w/chest & spine metastasis 1 Myeloma, multiple 1 Nasopharyngeal carcinoma 1 Neoplasm, malignant (transitional cell carcinoma) kidney 1 Oat cell Ca of lung w/liver & bone metastasis 1 Papillary (transitional cell carcinoma, Grade I) of bladder 1 Pituitary adenoma (brain) chromophobe type w/hypopituitarism 1 Plasmacytoma ilium, recurrent 1 Renal cell carcinoma 3 Right apiccal, large cell Ca w/ resultant Homer's syndrome 1 Right breast 1 Sarcoma, left leg 1 Semiroma, right testis 4 Squamous cell carcinoma of ear (epidermoid carcinoma) 1 Squamous cell of lung 4 Squamous cell carcinoma of anus, keratinizing 1 Squamous cell carcinoma of larynx, keratinizing, Grade I, invasive 1 Squamous cell carcinoma of pinna of ear 1 Testicular 3 Transitional cell carcinoma of bladder, Grade III 2 Undifferentiated liver carcinoma 1 Capillary hemangioma 1 Cardiomegaly 1 Carpal tunnel syndrome 3 Cataplexy 1 Cephalhematoma of foot postional deformity (child) 1 Cerebellar atrophy 2 Cerebellar tumor (child) 1 Cerebello insufficiency Cerebral convulsive disorder Cerebral palsy (child) Cervical adenopathy Charcot-Mar ie-Tooth Chest pain syndrome Chloracne Cholecystitis, chronic Chondromalacia, patella Chronic infection and subcutaneous papular eruption Chronic sclerosing glomerulonephritis Cirrhosis of liver Cleft palate (child) Club-Foot (child) Coagulopathy Colitis Collagen disorder Colon, mass in Colon, spastic Congential absence of tibia (child) Congenital athyriotic hypothyroidism (child) Congential dislocation of hip (child) Congenital heart disease, pulmonary valve artresia (child) Congenital polyneuropathy (child) Condyloma accuminata Congestive heart failure Constipation Convergence insufficiency by Hx Conversion reaction (numbness) Coronary Artery Disease S/P/ Coronary atherosclerosis Costochondritis Crohn's disease Crouzon1s disease (child) Cyst, sebaceous Cyst, vocal cord Degenerated nucleus pulposus Degenerative changes in joint Demorphic erythrocytosis Demyelination of peripheral nerves Depression atypical Depression w/anxiety Depression, endogenous 1 1 3 2 1 1 5 2 2 1 4 2 2 1 4 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 3 1 1 6 1 1 5 1 1 2 24 2 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) PI^JL^l^-J^l"!"3.^^reported by physician Depression syndrome, chronic Depressive disorder Depressive neurosis Dermatitis, l)f:rm-'i+,Ltis , '.i'-r'!.•>'-.: \. -,, .><• •r,;.-M MS, atopic ohroric V'r.taot <:r ythematous Dermatitis, perineal Dermatitis, photosensitivity Dermatitis, pruritis Dermatitis, scaly . _ Dermatitis, seborrheic .-., Dermatofibromas Dermatophytosis, recurrent Diabetes Diabetes mellitus Diastematomyelia (child) Dumping syndrome Duodenitis Dysethesias, diffuse Dyshidrosis (disorder of sweat glands) Dysmethic disorder Dysphasia Dyspnea Ecchymosis of legs Ecthyma Eczema Eczema, atopic Eczema, seborrheic Eczematous lesions Emphysema Encephalopathy End stage renal disease Eosinophilia Ependymoma, cerebellar (child) Epididymitis Epilepsy, idiopathic Epilepsy (child) Epistaxis, recurrent (child) Erythema multiforma Erythematous macular Erythematous papular Erythematous, resolving Esophagitis Esophoria (child) Exfoliative erythroderma . Extrarenal Wilms tumor (child) 3 3 4 5 14 6 3 1 2 3 1 11 6 3 4 17 1 1 1 1 4 6 1 3 1 2 12 4 2 5 4 2 1 1 1 7 1 3 1 2 4 1 2 1 1 1 1 10 Fatigue Fatty Metamorphosis Feet turned inward (child) Fibroepithelial papilloma Fibromas Fibromyalgia Fibromyositis syndrome Fibrosis Folliculitis Forefoot Adductus (child) Fundoplasty Furunculos.is (boils) Gastritis, chronic Gastroenteritis Gastroesophageal. reflux Globus heptericus Glomerulonephritis Granulatoma, -fit. scrotum Granuloma Granuloma Annulase Granulomatous colitis Granulomatous pulmonary disease Granulomatous skin lesions Granulomatous ulcer Guillain Barre Syndrome Gynecomastia, breast Headaches Headaches, cluster Headaches, vascular Hematoma Hematuria Hemoptysis, chronic Hemorrhoids Hemorrhaphies, bilateral inguinal Hepatitis, infectious Hepatocellular degeneration, focal Hepatocellular dysfunction Hepatomegaly Hernia Hernia, hiatal Hernia, inguinal Hernia, inguinal indirect (child) Herpes simplex Herpes zoster Hidrosadenitis, chronic, supprative (inflammation of sweat glands) 1 1 1 1 1 l 1 1 15 1 1 3 12 2 2 1 4 1 1 1 1 1 1 1 3 3 16 1 12 1 5 2 18 2 4 1 1 2 2 8 7 1 1 2 2 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) DIAGNOSIS (Dx)—as reported by physician 1 Hilar adenopathy w/calcification Hyaline membrane disease (child) 2 1 Hydrocele 2 Hydrocephalus (child) 1 Hygroma, cystic (child) 1 Hyperanxiety 1 Hyperbilirubinemia 1 Hypercalciuria 1 Hypercholesterum Hyperlipidemia 3 Hyper pigmental scaly plaques 3 1 Hyper pigmentation 2 Hypersomnia Hypertension 59 2 Hyperthesia, extremities 1 Hyperlipemia 1 Hyperlipoproteinemia (Type IV) Hyperthyroidism (child 1) Hypertryglycerdemia (Type IV) Hyperuricemia 1 Hypochondriasis Hypoglycemia 5 1 Hypopigmented areas (face) 1 Hypoplastic breast (child) 1 Hypospadias (child) Hypotension 3 1 Hypotonia (child) 1 Hysteronic personality Infection, persistent, soft tissue (child) Infundible pulmonary 1 stenosis (child) 1 Iritis, chronic Joint disease, degenerative 2 (of back) 1 Joint pain, peripheral 1 Keratitis 1 Keratoderma 1 Left spastic hemiparesis Leukemia, acute, lymphocytic 2 (child) 2 Leukemia, myelogenous, chronic Leukocytosis w/atypical lymphocytos 1 Lichen planus 3 Lichen simplex chronicus 3 Lipoma 13 1 Lipoma, cyst in lumbar area 2 Lipoma, spermatic cord II Liver pain Lumber sprain Lung disease, severe, chronic, obstructive Lupus, discoid Lupus erythematosis Lymphodenitis, chronic Lymphoid hyperplasia Lymphopranuloma inguinale Macular melasna Maculo-erythematous (rash) Mastoiditis, chronic sclerosing Meniere's disease Meningitis, cryptococcus Meningomyelocele (child) Mental Retardation (child) Metatarsus adductus (w/medial tibial torsion) (child) Microtia of ear (child) Microcephaly (child) Missing pectoralis (left) major muscle (child) Multiple sclerosis Musculoskeletal condition Myelomeningocele, lumbrosral (child) Myocardial infarction, acute Myofacial pain Narcolepsy Nephrolithiasis Neuralgias Neuralgia w/headache and recurrent fever Neurasthenia Neuritis Neurodermatitis Neuroma Neurosis, depressive Numbness in extremities Numbness ulnar aspect upper extremities Oligohydramnias (child) Oligospermia Onchomycosis Organic brain syndrome Osgood-Schlatter's disease (knee) Osler-Weber-Rondu disease Osteoarthritis, cervical 1 1 1 1 3 1 1 1 1 2 1 1 2 1 1 2 2 1 1 1 1 1 2 1 1 1 2 1 1 1 8 1 3 14 1 1 2 5 2 1 1 2 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) DIAGNQ3IS-(Dx)—as reported _by_ physician Osteoarthritis, degenerative Osteoarthrosis, degenerative Osteomylitis Pancreatitis, hyperlipidemia, chronic Papilloma Papular squamous rash Paralysis of vocal chords Paranoid schizophrenia disorder Paranoid state Parapsoriasis Parasthesias of extremities Paroxysmal atrial tachycardia Patent ductus arteriosus (child) Patent ductus arterosis of formen ovale cordus (child) Peptic esophagitis Peroneal palsy Peripheral neuropathy Peripheral ulnar palsy Personality disorder Peyronie's disease Photosensitivity Pityriasis alba Pityriasis rubra pilaris Pityriasis versicolor Plantar, hyperkeratosis, mild Pleural scarring Pneumothorax Pneumothorax, spontaneous Polyarthralgia Polyneuropathies Polyps, nasal/vocal cords Porphyria Porphyria cutanea tarda Posterior cervical pain Post traumatic stress syndrome Premature ejaculation Proctitis, inflammatory Prostatitis, chronic Proteinuria Pruritis Pruritus/onychomycosis of extremities Pseudofolliculitis Psoriasiform lichen simplex chronicus Psoriasis Psychosis, major 4 1 1 1 2 2 1 3 1 1 3 1 2 1 1 1 5 1 11 2 1 1 2 1 1 1 4 5 2 1 4 1 3 1 71 1 2 16 1 13 1 3 1 13 1 Psychotic depressive reaction (child 1) Pulmonary atresia Pulmonary disease, chronic obstructive Pulmonary emboli, massive Pulmonary embolism, ASC VC CVI Pulmonary nodule Pustules, recurrent Pyelonephritis Pylorospasm Radicular neuropathy Rash Rash, maculopapular Raynaud's phenomenon Reiter 1 s disease Renal glycasuria (no diabetes) Rhinitis Sarcoidosis Schizoid disorder Schizophrenia Schizophrenia, chronic, undifferentiated type Schizophrenia, paranoid Schizophrenia, schizo-affective type Schizo-type disorder Sciatica Scleroderma Scoliosis (child) Sebaceous cyst abscess Seborrhea Seizure disorder Soto's Syndrome (child) (cerebral gigantism) Spermatocelectomy Spermatoceles Sperm count, low Spina bifida (child) Spondylolisthesis Spondylosis Stenosis of larynx Sterility Stress syndrome Supple pes planus (child) Syncopy Syndactyly index w/absence and congenital absence of middle phalanx of 4 fingers (child) 3 1 2 1 1 1 2 1 1 1 7 1 2 1 1 3 4 5 19 11 31 3 4 1 1 1 4 12 4 1 1 3 3 4 5 3 1 5 1 1 1 1 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) DIAGNOSIS (Dx)—as ^reported by physician Tardive dyskinesia Telangiectasia Tendinitis Tenosynovitis (De Quervain's Disease) Testicular mass Testis, atrophic (child) Thrombocytopenia Tibial torsion of leg (child) Tietze's syndrome Tinea corpis Tinea cruris Tinea cruris pedis w/ onychomycosis Tinea pedis Tinea versicolor Tonsillitis, acute, chronic Transurethral resection Trichophytosis Tricuspid atresia, atrial septal defect, ventricular septal defect (child) Triglycerides, high Tropical fungus Truncal dystonia Ulcer, duodenal Ulcer, peptic Uroporphyria Urticaria, giant, recurrent Varicocele Xerosis of skin (dryness) 1 1 1 1 1 2 1 5 1 12 26 4 12 21 1 1 3 1 3 1 1 15 12 1 1 1 3 /3 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) CANCER IN VETS AGE 36 AND UNDER (substantiated by medical records) Case // A§e When Dx #13 #15 30 30 #40 #54 #98 (#106 31 31 35 33 (#106 33 #107 29 #121 * 31 #151 #180 * #242 #315 33 36 36 30 (#316 (#316 #469 * 30 33 36 #474 #523 * #551 34 34 25 #603 #621 #631 #676 #1072 #1213 (#1259 (#1259 * * * * * * 23 25 32 36 35 28 36 38 (#1672 (#1672 #1684 * #1732 #1751 * #1900 35 42 32 30 36 35 Type of Cancer/ICD No. Metastatic malignant melanoma 172.9 M8720/6 Squamous cell Ca w/adenocarcinomatous components, rt. lung 162.9 M8070/3 Basal cell Ca 173.9 M8090/3 Ca of esophagus 150.9 M8010/3 Polypoid carcinoma of sigmoid colon 153.9 M8050/3 Renal cell carcinoma, left kidney (died at age 33) 189.0 M8312/3 Metastatic malignant melanoma, bowel & subcutaneous tissue (died at age 33) 172.9 M8720/6 Multiple melanomas, Stage I (died at age 31) 172.9 M8720/3 Diffuse bilateral adenocarcinoma w/metastasis multiple areas bone/brain (died at age 31) 170.9 M8140/6 Ca of esophagus (died at age 34) 150.9 M8010/3 Ca of larynx 161.7 M8010/3 Giant Cell tumor of bone 170.9 M9250/3 Metastatic embryonal cell carcinoma (testis) with pulmonary involvement (died at age 30) 186.9 M9070/6 Basal cell carcinoma 173.9 M8090/3 Seminoma of right testis 186.9 M9061/3 Squamous cell carcinoma of anus, keratinizing 154.3. M8071/3 Seminoma, testis 186.9 M9061/3 Malignant melanoma 172.9 M8720/3 Pituitary adenoma (brain) chromophobe type w/ hypopitutarism 237.0 M8270/0 Basal cell epitheliomas, nose 173.3 M8090/3 Seminoma, right testicle 186.9 M9061/3 Sertoli cell carcinoma, testis 186.9 M8640/3 Seminona, right testis 186.9 M9061/3 Nasopharyngeal carcinoma 147.9 M8010/3 Renal cell carcinoma 189.0 M8312/3 Sarcoma, left leg 170.7 M8800/3 Osteosarcoma left leg w/metastasis right lung 170.7 M9180/6 Malignant melanoma (Level III) w/cerebral metastasis 172.9 M8720/6 Renal cell Ca. 189.0 M8312/3 Plasmacytoma right ilium, recurrent 203.8 M9731/3 Undifferentiated liver carcinoma 155.2 M8020/3 Myxoid liposarcoma w/spine & chest metastasis 171.9 M8852/6 Veterans who have sought treatment at a Veterans Administration Medical facility.for their malignancies. �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH DATA SHEET (as of July 31, 1985) CANCER IN VETS OVER AGE 36 (substantiated by medical records) Case Age When Dx #16 #24 * 51 47 #30 #70 » #76 #146 * #214 53 44 40 41 48 #239 48 #249 53 #295 52 #304 * 41 #360 * 58 #535 #567 * #622 * 39 46 50 #638 #806 * 45 68 #829 46 #890 * #894 * #1041 * 47 46 60 #1053 52 #1089 (#1093 (#1093 #1178 #1182 #1421 50 50 56 60 44 52 #1477 * * * * 41 Renal cell Ca left kidney 189.0 M8312/3 IGA, Multiple myeloma-lumbar spine (died at age 50) 170.2 M9730/3 Ca of lungs, squamous cell 162.9 M8070/3 Transitional cell Ca, Grade I 188.9 M8120/3 Ca breast w/metastasis to axilla 175.0 M8010/6 Basal cell Ca of nose 173.3 M8090/3 Squamous cell Ca anterior fascialpillar 146.2 M8070/3 Neoplasm, malignant (transitional cell carcinoma) right kidney 189.0 M8120/3 Glioblastoma multiforme, brain (died at age 53) 191.9 M9440/3 Squamous cell carcinoma of larynx, keratinizing, Grade I, invasive 161.9 M8070/3 Adenocarcinoma of lung w/brain metastases 162.9 M8140/6 Adenocarcinoma of sigmoid colon metastasized to liver 153.9 M8140/6 Basal cell Ca on scalp 173.9 M8090/3 Testicular cancer 186.9 M8010/3 Bronchogenic carcinoma, squamous cell, poorly differentiated; metastatic carcinoma in hilar lymph node 162.9 M8010/6 Fibrohistiocytoma, leg 170.7 M8831/3 Carcinoma of rectum w'/metastasis to liver 154.1 M8010/6 Metastatic squamous carcinoma to scalene node 195.0 M8070/6 Epidermoid carcinoma left lung 162.9 M8070/3 Epidermoid carcinoma of esophagus 150.9 M8010/3 Adenocarcinoma of sigmoid colon and metastatic to 1 of 3 lymph nodes (Duke C) w/metastasis to liver 153.9 M8140/6 Right apical (lung) large cell Ca w/ resultant Horner's syndrome 162.9 M8012/3 Chronic myelogenous leukemia 205.1 M9863/3 Basal cell epitheliomas 173.9 M8090/3 Bronchogenic carcinoma, left lung 162.9 M8010/3 Adenocarcinoma of rectum, Duke C 154.1 M8140/3 Ca of bladder, low grade 188.9 M8010/3 Adenocarcinoma of sigmoid colon w/metastasis to liver and lymph nodes 153.9 M8140/6 Basal cell carcinoma, nose 173.3 M8090/3 •Veterans who have sought treatment at a Veterans Administration medical facility for their malignancies. �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH DATA SHEET (as of July 31, 1985) CANCER IN VETS OVER AGE 36 (substantiatedI by medical records) Case # Age When Dx #1496 48 #1578 * 50 #1582 48 #1622 #1741 37 44 # 1820 55 #1830 37 #1833 39 #1881 65 Type of Cancer/ICD ^Nq,, Adenocarcinoma of rectum stage D u k e s C w/ 5/15 lymphnodes positive for Ca 154.1 M8140/6 Adencarcinoma of rectum w/seeding of pararectal fat 154.1 M8140/6 Bladder - papillary, transitional cell, grade I 188.9 M8130/3 Adenocarcinoma of esophagus 150.9 M8140/3 Squamous cell Ca of pinna right ear (epidermoid carcinoma) 173.2 M8070/3 Transitional cell Ca og bladder, Stage II 188.9 M8120/3 Differentiated lymphocytic lymphoma nodular type, Stage 4 w/widespread metastasis 202.8 M9620/6 Oat cell Ca of lung w/liver & bone metastasis 162.9 M8042/6 Basal cell Ca 173.9 M8090/3 "Veterans who have sought treatment at a Veterans Administration medical facility for their malignancies. �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) QUESTIONABLE RANGERS (substantiated by medical records) #45 #169 24 29 Granuloma Rt. scrotum Liporna cyst in lumbar area CHILDREN WITH LEG DEFORMITIES (substantiated by medical records) Club-foot, secondary to spina bifida Feet turned inward Congenital absence of rt. tibia Cephalhematoma rt. foot positional deformity at birth Metatarsus adductus w/medial tibial torsion Congenital dislocation of hip Club-Foot/Forefoot Adductus Supple pes pianus Bilateral internal tibial torsion Bilateral metatarsus adductus Metatarsus adductus, right foot Club-foot Short leg Tibial torsion, both legs Calcaneovarus deformity of feet Bilateral internal tibial torsion Tibial torsion of left leg #25 #40 //68 #152 #188 #207 #296 #330 #571 #582 #583 #770 #872 #978 #1603 #1622 #1754 CURRENT JRASHES ^(substantiated .by jrcedical records) No. of cases 62 II �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) TINGLING/NUMBNESS IN EXTREMITIES (substantiated by tnedioal records) Case # Year Dx #14 #22 #37 #41 #85 #119 #145 #192 #195 #206 #211 #212 #224 #229 #267 #306 #338 #341 #389 #415 #422 #423 #450 #500 #872 #1315 #1417 #1758 #1808 #1858 1981 1968 1981 1977 1980 1980/1981 1981 1978 1981 1981 1982 1982 1980 1964 1980 1982 1979 1978 1981 1982 1972 1981 1982 1981 1983 1984 1981 1983 1975 1983 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) MISCARRIAGE/STILLBIRTH (substantiated by medical records) Case // #36 #38 #41 #44 #63 #67 #82 #83 #85 #86 #97 #99 #142 #156 #173 #179 #181 #183 #209 #241 #310 #323 #328 #331 #386 #494 #568 #571 #591 #608 #647 #653 #688 #699 #722 #773 #774 #822 #841 #854 #879 #926 #930 #948 #953 #954 Case # Year Dx 1970 1972 (2) 1973 (2) ? 1980 (2) Between 1974-1981 (2) 1971, 1975, 1978 1974, 1977 1974 1979, 1980 1980 1974 1976, 1978 1976 1972 1970 1971 1978, 1980, 1981 1975 1979 1971,1979 1979 1977 1975 1977, 1980 1976 1983 1976 1982, 1983 1975, 1976 1971 1973 1980 1981, 1982 9 1974 1978, 1973, 1972 1982 1982 1981 1970 1977 1974 1978 1979 1974, 1978 #978 #997 #1019 #1095 #1275 #1278 #1283 #1395 #1421 #1655 #1677 #1754 #1780 #1809 #1821 #1835 Year Dx 1971 1981 1978, 1982 1975, 1976 1975 1976, 1977 1974 1972 1971, 1978 1979 1977 1983 1983 1980 & ? 1971 1968 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) SCHIZOPHRENIA (siftstantiated by medical records) Case # #16 #3^ #43 #47 #123 1125 #128 #139 #144 #164 #190 #229 #232 #238 #245 #248 #253 #256 #294 #301 #361 #371 #378 #381 #397 #401 #405 #431 #449 #455 #463 #552 #564 #565 #571 #634 #635 #809 #838 #841 #872 #953 #967 #101? #1080 #1084 #1107 Year Dx Case # #1183 #1198 #1200 #1291 #1303 #1323 #1336 #1519 #1581 #1625 #1631 #1615 #1669 #1678 #1708 #1718 #1718 #1776 #1871 1971 1976 1980 1981 1978 1974 1981 ? 1979 1977 1983 1982 1979 1982 1982 1968 1976 1975 1970 1971 1969 1970 1976 1976 1972 1971 1973 1971 1981 1971 1972 1972 1970 1969 1971 1977 1970 1967 1982 1981 1980 1981 1982 1976 1980 1982 1981 3.0 Year Dx 1971 1982 1982 1980 1973 1983 1981 1982 1968 1968 1977 1982 1969 ? 1976 1983 1977 1978 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) POST TRAUMATIC STRESS DISORDER (Substantiated by medical records) Case // Year Dx Case # #10 #32 . #50 #60 #78 #104 #128 1141 #173 #177 #223 #229 #270 #278 #298 #310 #361 #362 #364 #365 #366 #367 #378 #388 #430 #446 #449 #456 #459 #489 #600 #603 #623 #663 #697 #775 ? 1981 1981 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1982 1981 7 ? 1982 1983 7 7 1982 1982 1983 1982 1981 1982 1983 1981 1980 #1308 #1325 #1327 #1475 #783 #784 #838 #842 #849 #872 #875 #920 #967 #971 1983 1983 1983 1980 1981 1982 1981 1981 1982 1982 1982 1981 1985 1982 1982 #1516 #1541 #1604 #1613 #1776 #1936 Year _Dx 1981 1981 1981 1981 1982 1982 1981 1981 1981 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) OTHER MEDICAL CONDITIONS (as reported by veterans and physicians) Acne 19 Acne, cystic 2 Alcoholism 8 Allergic bronchitis 3 Allergies 19 Alopecia areata (spotty baldness) 1 Anemia, iron deficiency 5 Anorexia 3 Anxiety 23 Apnea (cessation of breath) 16 Arteriosclerosis 3 Arthralgia 1 Arthritis 34 Asthenia (weakness) 35 Asthma 10 Ataxia (lack of muscle coord) 7 Back pain 29 Birth defects/medical problems—child 13 Blackouts 15 Blisters 10 Blood disorders 214 Body cramps/aches 15 Boils 18 Bones, decaying 2 Burning sensation in back/extremities 8 Calcium loss .2 Cancer 44 Chest pains 43 Chloracne 23 Colitis 4 Comedones 1 Constipation 8 Cyst 22 Cyst, retention 2 Delayed healing by first intention 2 Dementia, in remission 1 Depression 86 Dennatitis 6 Diabetes 22 Diarrhea 28 Disruption of circadian rhythms 1 Dizziness 78 Drug addiction 1 Dumping syndrome 2 Dysmorphic erythrocytosis (unusual shape of blood cells) 1 Dyspnea (labored breathing) 11 Dysesthesias, diffuse 1 Ear fungus 1 Ecxema 2 Edema 16 Emotional problems 502 Endocrine problems 1 Epilepsy 1 Fatigue 60 Fat tissue lumps 11 Fever, recurrent 9 Fungus 10 Gall bladder w/mass 1 Gastrointestinal disorders 650 Glands, swollen 4 Gout 7 Growths, skin 32 Hallucinations 3 Hair loss 27 HBsag-Carrier 1 Headaches 566 Hearing problems 29 Heart attack 7 Heart problems 20 Hematuria 2 Hemorrhoids 4 Hepatitis 6 Hepatomegalia 1 Herpes 3 Hirsutism (abnormal body hair growth) 1 Hives 2 Hyperlipidemia 1 Hypertension 80 Hypertension, essential 1 Hypoglycemia 1 Hysteria 1 �TEXAS VETERANS AGENT ORANGE ASSISTANCE PROGRAM TEXAS DEPARTMENT OF HEALTH, AUSTIN, TEXAS DATA SHEET (as of July 31, 1985) OTHER_MED^AL^CpNDITIONS> (as^reportedI by veterans and[. Infections, chronic 20 Immuno suppression 3 Irritability 18 Itching 34 Joint pain 83 Kidney problems 24 Liver problems 161 Loss of appetite 11 Loss of concentration 10 Loss of smell 3 LOSS of taste 2 Low blood sugar 3 Lung problems 30 Melanomas 5 Memory loss 69 Meningitis, cryptococal 1 Moles 6 Muscle problems 37 Myocardial infarction, acute inferolateral. 1 Nails fall out 11 Nausea 24 Nerve problems 575 Neuritis, traumatic 1 Nose bleeds 1 Numbness 323 Pancreatitis, chronic 3 Paresthesias of distal arms/legs 3 Peripheral neuropathy 3 Peripheral tumescense (swollen extremities) 3 Personality change 11 Pleurisy, chronic 2 Polyps 1 Porphyria Post traumatic stress disorder Prostatitis Psoriasis Pulmonary embolisms, multiple 3 6 22 1 1 Pulmonary fibrosis 1 Rash 385 Rectal bleeding, history of 29 Renal cysts 1 Reproductive problems 90 Restricted blood flow 7 Rhinitis (nose inflammation) 2 Respiratory problems 25 Seizures 7 Sensitivity to change in heat/cold 6 Sex, pain during 1 Sexual dysfunction 268 Shingles 1 Sinus problems 24 Skin, dryness 23 Skin hyper/hypopigmentation 24 Sleep disturbance 338 Sores 12 Speech problems 1 Sterility 6 Sweating, excessive 5 Tachycardia 2 TB, subclinical 3 Teeth, loss of 3 Throat, sore, chronic 8 Thyroid problems 1 Tingling of extremities 163 Tinitus (ringing in ears) 22 Ulcer 37 Upper respiratory infections (URI) 3 Urinary infections 21 Vascular insufficiency 6 Venous thrombosis, deep 1 Vision, blurred 22 Vision, decreased 27 Vision, sensitivity to light 9 Vomiting 19 Warts 6 Weight loss/gain 43 �SUMMARY of PILOT STUDY PROTOCOLS for the TEXAS VETERANS AGENT ORANGE PROGRAM September 1984 Protocols for four pilot studies have been developed by faculty of the University of Texas for use in the Texas Veterans Agent Orange Program administered by the Texas Department of Health. They are: 1. Cytogenetic/Bleomycin Testing 2. Aryl,Hydrocarbon Hydroxylase (AHH) Assay 3. Immune Evalution of Veterans Exposed to Agent Orange 4. Uroporphyrin Testing These protocols are described in the attached documents. These studies have special subject selection requirements. They also have limitations and pitfalls that should be recognized by everyone interested in the outcome. SUBJECT SELECTION. Interpretation of data gathered by the proposed pilot studies will depend heavily on the criteria and care used in selecting study subjects. Three categories of age-matched study subjects are required: 1) Vietnam veterans (at high risk), 2) Vietnam veterans (at low risk), and 3) unexposed controls. Vietnam Veterans (at high risk). The establishment of a reliable exposure index is critical and very problematical. It may well be impossible to estimate degree, duration, and route of exposure to Agent Orange in most Vietnam veterans. However, it should be possible to identify some Vietnam veterans whose contact with Agent Orange, as reflected in personal histories and verified in military records submitted to the Texas Department of Health, was substantial and prolonged. From among this group of individuals, participants in the pilot studies should be selected on the basis of a detailed personal interview and medical history that would exclude subjects whose present or previous occupation, habits, or lifestyle might introduce obvious confounding factors into the study. -1- �Vietnam VeteransL (at 1j3W_r1_s_k). Veterans whose service records and personal histories indicate an improbable contact with Agent Orange or other herbicides. Obvious confounding factors would exclude these veterans the same as the "high risk". It would be safe to state that in assigning high or low risk that increasing probability of a difference prevails as the size of the total group increases and the number of veterans determined to be in the nonselected medium group increases. Unexposed Controjs. These individuals, selected to exclude confounding factors, will serve as the normal control group for the pilot studies. These individuals are matched for usual factors and may be civilians or non-Vietnam veterans. LIMITATIONS AND PITFALLS. The limitations of the laboratory-based pilot studies should be clearly understood. Some of these limitations are: 1. Uncertainty regarding degree, duration, and route of exposure to Agent Orange in individual study subjects is a major limitation of these studies. Efforts to cope with this uncertainty are discussed under the topic of Subject Selection. 2. Since exposure of veterans to Agent Orange or other herbicides used in Vietnam occurred more than a decade ago, any evidence or consequences of that exposure may have diminished to such an extent that it is no longer detectable. 3. The tests to be performed in these pilot studies will not detect effects that are specifically attributable to Agent Orange or any other herbicide. Chromosome damage, enzyme abnormalities, and suppressed immune responsiveness can result from any number of causes, some well-known and others yet unrecognized. Because of this, it will not be possible to conclude that disorders detected in any given individual are due to Agent Orange. However, this is not to say that populations of matched veterans whose detailed military, occupational, medical, and personal histories suggest that they differ as groups only in their exposure to Agent Orange. 4. Individuals whose test results are positive cannot be offered therapeutic manipulation or corrective intervention. There is no known way of reversing chromosome damage. -2- �5. Negative results from these tests would not be definitive. In other words, absence of overt chromosome damage in the study population would not mean that other, less easily recognized effects were absent. COSTS. The University of Texas has made every effort to minimize the costs that cannot be absorbed for tests done on veterans in these studies. For each matched set of exposed and unexposed individuals, the Texas Department of Health will cover the unabsorbable costs of $380 for cytogenetic testing, $800 for aryl hydrocarbon hydroxylase assay, arid $800 for immune evaluation and uroporphyrin testing. REPORTS. These studies will proceed at different rates. Upon completion, the investigators responsible for each study will present their data and results to the Texas Department of Health and, in addition, will prepare and'submit their findings for publication in the scientific literature. Interim progress reports will be provided according to a schedule to be decided by mutual agreement of the Texas Department of Health and the individual investigators. -3- �CYTOGENETIC/BLEOMYCIN TESTING. Peripheral blood samples are set up with the standard blood culture medium to stimulate lymphocytes to grow. Standard cytogenetic harvest method (Colcemid block for 1 hr., hypotonic solution treatment for 20 min., fix and air-dried) is used to prepare 48-hr and 72-hr culture samples. The slides are stained with Giemsa. Thus, each blood culture will have two harvest samples. Whenever possible, 100 metaphases are analyzed from each harvest sample to record chromatid-type and chromosome-type aberrations, and the aberrations are finally converted into breaks per cell for comparison. ARYL HYDROCARBON HYDROXYLASE (AHHLASSAY. The purpose of this study is to evaluate veterans for evidence of abnormalities in the levels, activity, or regulation of the cytochrome P-450 microsomal mono-oxygenase enzymes." Lymphocytes will be cultured from blood samples obtained from veterans and matched controls. The cultured lymphocytes will be assayed for levels of the enzyme aryl hydrocarbon hydroxylase both with and without a challenge by 3-methyl cholanthrene to induce the enzyme. Induction of the cytochrome P-450 associated enzymes is the most basic biological effect of TCDD, the toxic contaminant of Agent Orange. IMMUNE EVALUATION/UROPORPHYRIN TESTING. This study will examine various measured immune parameters of 1) Vietnam veterans from Texas who were at high risk for exposure to Agent Orange, 2) Vietnam veterans from Texas who were at low risk for exposure to Agent Orange and 3) matched veteran-controls who were not exposed to Agent Orange. In addition, urine specimens will be collected from these same groups and tests will be performed to measure levels of urinary porphyrins. �AGENT ORANGE ADVISORY COMMITTEE to the TEXAS DEPARTMENT OF HEALTH Development and Preliminary Results of Pilot Clinical Studies Report of the Chairman Guy R. Newell, M.D. Professor of Epidemiology and Chairman, Department of Cancer Prevention The University of Texas System Cancer Center Monday, March 26, 1984' �ACKNOWLEDGEMENTS The principal investigators, their colleagues and I wish to first acknowledge William B. Neaves, Ph.D. of The University of Texas Health Science Center at Dallas who was the first chairman of The University of Texas System Agent Orange Program Committee. It was under his skillful, thoughtful, and statesman approach that the initial pilot studies were reviewed, selected and initiated. We are all indebted to Ms. Harriet Franson, the Program Manager for the Texas Department of Health, who sees to our requests and to those of the veterans we are trying to assist with both speed and compassion. I want to thank Paul K. Mills, M.S., M.P.H. of the Department of Cancer Prevention, UTSCC, who prepared the section describing the criteria/selection methodology and analysed the criteria used in review of the first 255 Vietnam veterans. Without the unending wealth of first-hand knowledge of Vietnam and its environs provided by George R. Anderson, M.D., Director of the Texas Veterans Agent Orange Assistance Program, the task of estimating gross exposure of veterans would have been impossible. Finally, the support given to this program by Robert Bernstein, M.D., F.A.C.P., Commissioner of the Texas Department of Health has been unyielding. �Background During the Vietnam War, U.S. military personnel sprayed large quantities of a herbicide called "Agent Orange" over the Vietnamese countryside. The herbicide, named because of its shipment in orange-striped barrels, consisted of approximately equal portions of the n-butyl esters of 2,4-dichlorophenoxyacetic acid (2,4-D) and 2,4,5-trichlorophenoxyacetic acid (2,4,5-T). These shipments of Agent Orange were contaminated during the manufacturing process with traces of a highly toxic chemical, the dioxin dibenzo-p-dioxin (TCOD). 2,3,7,8-tetrachloro- Its concentration varied from batch to batch but averaged about 2 pom of 2,4,5-T. TCOO is known to be an exceptionally toxic chemical. Concerns have been frequently raised by Vietnam veterans that Agent Orange exposure may result in Infertility, genetic damage, birth defects in offspring, and cancer. No studies to date have confirmed these suspicions. In 1981, the Texas Legislature established a program to assist veterans who may have been exposed to certain chemical defoliants or herbicides, including Agent Orange. An important goal of the Texas Veterans Agent Orange Assistance Program was to determine 1f veterans have suffered physical damage as a result of substantial exposure to Agent Orange. The bill establishing this program called for a cooperative effort between the Texas Department of Health and The University of Texas System to conduct studies that would address the health effects of exposure to Agent Orange. Pi 1ot Studies of VIetnam Veterans Faculty of The University of Texas developed protocols for pilot studies of selected veterans in the Texas Veterans Agent Orange Program. �Three pilot studies were selected for implementation. These were cytogenetic testing, sperm evaluation, and analysis of the immune response in putatively exposed veterans and suitable control subjects. Cytogenetic testing to be conducted at The University of Texas System Cancer Center, will determine if Vietnam veterans presumed to have been exposed to Agent Orange during their military service have more genetic damage as measured by chromosomal abnormalities in cultured lymphocytes than does a suitable comparison group of veterans presumed not to have been exposed to Agent Orange. Sperm evaluation, to be conducted at The University of Texas Medical Branch at Galveston, will determine whether an association can be detected between current production of abnormal sperm and prior exposure to Agent Orange. The percentage of morphologically abnormal sperm and the incidence of nondysjunction of the Y chromosome will be assessed in this study. Analysis of the immune response, to be conducted at The University of Texas Health Science Center at Houston, will compare the immunocompetency of Vietnam veterans thought to have been exposed to Agent Orange with that of age-matched controls having no history of exposure to Agent Orange. In addition, a birth defects study was to be Initiated by the Division of Clinical Genetics of The University of Texas Health Science Center at Dallas. A summary of this study by Jan M. Friedman, M.D., Ph.D. is attached. Study Limitations Every attempt was made to explain the inherent limitations of these studies to all concerned with their outcome. These include, briefly: • Inability to establish a reliable index of exposure to Agent Orange for any individual Vietnam veteran. No exposure Index was available from the Department of Defense, Veterans Administration, or other official source. �• Control subjects could be selected on gross variables such as obvious lack of previous contact with Agent Orange. Ability to match on other variables was limited. « Because exposure occurred over a decade ago, damage or adverse consequences of such exposure may have diminished to an extent that they are no 1onger detectable. • The tests performed in the pilot studies are not specific for measuring effects of Agent Orange or any other specific agent. • Chromosome damage, sperm abnormalities, and altered immune responsiveness can result from any number of causes; therefore, it will not be possible to conclude that any abnormal findings in the group or in any individual are due to Agent Orange. « Individuals whose test results are positive cannot be offered therapeutic manipulations or corrective intervention in that there is no known way of reversing chromosomal damage or sperm abnormalities. 9 Negative results of the pilot tests would not prove the absence of other, less easily or impossible to measure, effects. Although these limitations are substantial, the laboratory-based pilot studies represent a positive step toward resolution of the Agent Orange dilemma. These studies have become part of a diverse and rapidly expanding national effort to answer pressing questions about the health effects of herbicides used in Vietnam. In addition, results of these initial pilot studies could suggest avenues for future scientific investigations of this national concern. �Agent Orange Subject SelectionCommittee Criteria/Selection Methodology The Agent Orange Subject Selection Committee was established to review evidence (military records, medical records, arid other supporting documents) which would indicate if a given veteran was Indeed exposed to Herbicide Orange in Vietnam, and if so how much exposure occurred. Seven criteria were used to evaluate a given veteran's category of exposure. Depending on the combination of exposure variables veterans were classified into one of six exposure categories. These categories included: highly exposed, medium to highly exposed, medium exposed, low to medium exposed, low exposed and disqualified. Those veterans deemed to be In the highly exposed category were then included in the Pilot Phase of the clinical studies. These studies included cytogenetic testing, immune competency, and sperm mobility and mOtility assessment. The criteria which the committee considered when reviewing the military records, medical records, questionnaire, and other supporting documents included the following. 1. Exposure to herbicides. The committee noted the amount (in gallons) of Herbicide Orange, White, and Blue sprayed in the area where the veteran was assigned during the time period he was assigned to that area. This criterion included estimated rates of exposure and exposures other than "Ranch Hand" exposures. �2. Repo rt ed symptoms A. At the time of exposure: since the chloracne rash is path- o gnomon ic of exposure to dioxin, the committee considered the appearance of a rash at exposure in evaluating the individual's exposure status. B. After time of exposure: reports of chloracne after initial exposure were also considered by the committee for evaluation of exposure. 3 * Cu r rent ned ica o b S * T"6 occurrence of current disease which could possibly be related to herbicide exposure was viewed by the committee as an important criterion for evaluating exposure status. *• Current or past^jpccufij^t0^"^L^ffUJ-Al, Jixpgsjjre. Since exposure to non-herbicide related chemicals could occur on the job outside the military, the committee regarded such occupational exposure as a potential confounding factor in the evaluation of exposure status. Such exposure could disqualify a veteran from participating in the Pilot Phase. s * Hi scarriages or sti 1 1 bi rths. The potential genotoxic effects of phenoxy herbicides, including Herbicide Orange were noted by the committee. Hence, the occurrence of miscarriage or stillbirth among the off- spring of the veterans was considered when evaluating the exposure status of a veteran. te 1n (5) "above, the phenoxy herbicides are potential teratogens in addition to being mutagens and carcinogens. Hence, the committee noted the occurrence of birth .defects in evaluating exposure status. �7 ' Dates and types of service dut^. The heaviest spraying of Herbi- cide Orange in Vietnam occurred between 1967 and 1969. operations ceased in early 1971. All spraying Hence, the committee closely evaluated the service dates in Vietnam in establishing the exposure status. duty type in Vietnam was considered. Also, Clerks, truck drivers, repairmen, and personnel assigned to base camps were not considered to be at high risk of exposure in comparison to infantrymen in the field where potential exposure was much higher. The following table {Table 1) demonstrates the relative importance of each of the selection criteria used by the committee in arriving at a judgment of exposure status. The percentages reflect the importance the committee placed on each criteria in placing veterans in a given exposure category. �8 Table 1. Percentage Summary of Criteria Considered in Exposure Classification of 255* Vietnam Veterans Exposure Status High Med./High Medium Low/Med. Low Criteria Exposure (Gallons) 96.4 100.0 94.7 Symptoms at Exposure 42.3 46.1 13.1 16.6 0.0 Symptoms After Exposure 42.3 61.5 18.4 16.6 0.0 Current Medical Problems 70.5 84.6 31.5 50.0 0.0 Occup. /Chemical Exposure 4.7 0.0 2.6 33.3 5.9 Miscarriages 18.8 15.3 10.5 50.0 1.4 Birth Defects 11.7 15.3 10.5 83.3 0.0 Dates and Type of Service 94.1 100.0 68.4 16.6 5.9 38 6 Total No. Veterans 85 13 100.0 52.2 57 *46 veterans were disqualified from the pilot phase of the study for various reasons. Veterans who had previously received chemotherapy were disqualified since such treatment would affect cytogenetic and immune parameters. In addition, veterans with occupational exposure to chemicals which could affect laboratory testing of sperm, cytogenetic or immunological parameters were removed from further consideration. �As of February 29, 1984, the Selection Committee reviewed 320 cases of which 99 were selected for the clinical studies (fifty cases were reviewedmore than once after more Information had been obtained). The goal set for the pilot studies was 50 veterans selected for having received the highest possible exposure to Agent Orange based on all available information. Thus, the study group was intentionally skewed toward exposure and was not intended to be "representative" of veterans who claimed exposure. Nor within the study group was there a gradient from high to low exposure. All veterans in the study group were selected for high exposure. A doseresponse effect was, therefore, not built in to the pilot study design. The controls, by contrast, were intentionally selected because of no possible exposure to Agent Orange in Vietnam. Matching for associated factors such as occupation or for other sources of exposure to dioxin was attempted, but was recognized to be imprecise. The intentional study design to Include maximum possible exposure among cases (Vietnam veterans) contrasted to least likely exposure among comparison subjects (matched controls) was selected because there was virtually no literature describing similar studies in humans. Since these pilot studies represented a "first," it was thought most desirable to design the study for maximum likelihood of detecting a biologic abnormality among the veterans, if one existed and could be measured by the available methods used. Collection of samples of specimens from both veterans and controls was arranged by staff of the TON and shipped to the individual investigators. Samples were coded so that the tests were performed in all three laboratories without knowledge of whether the sample was from a veteran or a control (specimens were "blinded"). After all specimens were analysed by the �10 laboratories the code was sent to each Investigator on the same day so that appropriate analyses could be performed. Preliminary Results of the Pilot Studies A summary of findings of the three pilot studies are presented. All three studies were performed on specimens from the same Vietnam veterans and controls. The total numbers in each group may vary from study to study and from specimen to specimen. These do not represent errors, rather they indicate variability among the techniques used for the studies. The investigator(s) along with their title and affiliation are given for each study. They can provide more technical details if requested. Cytogenetic Testing T. C. Hsu, Ph.D., Principal Investigator Professor of Cell Biology Sen Pathak, Ph.D., Collaborator K. L. Satya-Prakash, Ph.D., Collaborator The University of Texas System Cancer Center M. D. Anderson Hospital and Tumor Institute Each blood sample was set up for short-term culture with standard blood culture medium. Cell chromosomes were examined at 48 and 72 hours after initiation of cultures. This technique 1s standard and has been published by Dr. Hsu and his colleagues. Each cell speciman was critically examined for chromosome changes. These include: 1. Chromatid breaks, isochromatid breaks and exchanges. �11 2. Chromosomes showing acentric fragments, dicentrics, rings, and marker chromosomes indicating translocations. The percentage of cell specimens with chromosome breaks and chromatid breaks were recorded. The frequency of chromosome changes was calculated as breaks per cell (b/c). In previous studies of large numbers of patients, families, and population subjects the b/c ratio was found to be the most useful expression of genetic damage. The results of this pilot study of cytogenetics on veterans exposed to Agent Orange and matched controls are summarized below: Table 2. Cy toge n e t i c data_ojii j^eteranisL and_cp_nt_ro]_s_ Vietnam Veterans % cells with chromosome breaks 0.78 breaks/cell (b/c) 0.03 Matched Controls 0.62 - 0.02 �12 Table 3. Cases with Chromosome-type Aberrations and Breaks/Cell 'Cytogenetic Change Vietnam Veteran No. (*) Matched Control No. (*) 0.0 - 0.9 17 (S6.7) 22 (73.4) 1.0 - 1.9 9 (30.0) 4 (13.3) 2.0 - 2.9 2 { 6.7) 3 (10.0) 3.0 - 3.9 1 ( 3.S) 0 ( 0.0) 4.0 ~ 4.9 1 ( 3.3) 1 ( 3.3) 5.0 and over 0 ( 0.0) 0 ( 0.0) Metaphases with Chromosome -type abberrations * 30 100.0 30 100.0 0.00 - 0.02 16 (61.S) 20 (66.7) 0.03 » 0.07 7 (26.9) 10 (33.3) 0.08 - 0.12 2 ( 7.7) 0 ( 0.0) 0.13 and over 1 ( 3.9) 0 ( 0.0) Breaks/cell * 26 100.0 30 100.0 * Chi square not significantly different between veterans and controls. �13 It should be pointed out that the lack of positive results does not necessarily indicate the lack of genomic toxicity in persons soon after the Agent Orange exposure. Genetic effects induced by Agent Orange, if any, might have been sufficiently diluted by years of lymphocytic proliferation. In other words, we do not have a complete chronological study following persons, before, soon after, and long after exposure to a genotoxic agent. However, the present data, collected some 15 years after the exposure, appear negative. Sperm Tests Jonathan 6. Ward, Jr., Ph.D., Principal Investigator Marvin S. Legator, Ph.D., Collaborator Division of Environmental Toxicology, The University of Texas Medical Branch at Galveston Up to 3 semen specimens were obtained from each study subject at 2 and 3 month intervals. Upon receipt of the samples, a sperm count was determined, morphology (appearance) was classified by shape and size using standard, published methods. Reference slides were randomly included to serve as an internal control for scoring consistency. At least 500 sperm were examined per sample and the percentage of morphologically abnormal sperm was recorded. The percentage of fluorescent bodies (F-bodies) was recorded as well. �14 The results are shown in the table below: Table 4. Mean Values (± Standard Deviation) of Sperm Test Results for Veterans and Controls Sperm Characteristics Sperm Count (X 106) Vietnam Veterans (Mean ± SO) (No. Subjects/ Samples) Matched Controls (Mean ± SD) (No. Subjects/ Samples) 103.7 ± 76.0 116.3 ± 79.3 32 (76) 32 (64) P = 0.43* % Morphologically Abnormal 50.6 ± 14.8 48.7 ± 12.6 31 (73) 31 (61) 47.7 ± 2.1 47.8 ± 2.5 P = 0.78 % One F-body P - 0.96 % Two F-body P = 0.82 30 (70) 0.7 ± 0.2 30 (70) 30 (58) 0.7 ± 0.3 30 (58) *Kolmogorov-Smirnov 2 sample test used for significance of difference of mean values �15 Interpretation and Conclusion: The results of the sperm tests are reported for 32 pairs of veterans and non-veteran controls. No statistically significant differences were observed between the two groups for sperm count, abnormal morphology and 2 F-body frequency. The preliminary conclusion is that none of the three tests employed demonstrated any effect among individuals with prior military service in Vietnam where exposure to herbicide was probable. However, based on the numbers tested, large differences in sperm count could escape detection, while small differences in morphology and F-body frequency could exist, which would not have been detected. Imntjinologic Studies The immune system is charged with the defense of the body against both internal as well as external antigenic challenges. The cells which make up this system are several different types of lymphocytes - T and B cells, macrophages, and a poorly characterized cell referred to as null cell. T-lymphocytes (derived from the thymus gland, hence also called T-cells) play a central role in the overall regulation of immune responses, including both antibody synthesis and the development of cell-mediated immunity. �16 Several measures of T-cells and their functions were determined from blood lymphocytes of Vietnam veterans and matched controls. A brief descrip- tion of these is given below: Table 5. Test Performed Explanation of Test % Total T-RFC All T-cells in the peripheral blood leukocytes (PBL) as measured by sheep red blood cell rosette formation (RFC). % Pan-T cells All T-cells in PBL measured by monoclonal antibody (OKT 3). %Active T-RFC Subpopulation of T-cells which function as immune surveillance cells. % Helper/Inducer T cells "Helper T cells" required for antibody formation, measured by monoclonal antibody OKT 4. % Suppressor/Cytotoxic T cells "Suppressor T cells" Suppress antibody response after initiated, measured by monoclonal antibody OKT 8. Helper/Suppressor Ratio Ratio of T-helper to T-suppressor cells. % HNK Human natural killer cells measured by Leu 7. % OKT 9 T cell actlvational antigen measured by OKT 9. % OKT 10 T cell activational antigen measured by OKT 10. PMLC (S.I.) Panel of mixed lymphocyte culture, measures ability to respond to 3-5 peripheral blood leukocytes. �17 Table 5. (Continued) Test Performed Explanation of Test PHA (S.I.) Response to a mltogen stimulant, phytohemagglutinin S.I. = Stimulation Index Spont. Blasto. Spontaneous blastogenesis, measure of metabolic activity of round cells in peripheral blood leukocytes. The numbers of individuals tested, the mean values for the groups and the standard deviation are given in the table below: Table 6. Mean Values (± Standard Deviation) of Immune Tests Results for Veterans and Controls Immune Test Total T-RFC Pan T cells Active T-RFC Helper T cell (Inducer) Suppressor T cell (Cytotoxic) Helper/Suppressor Ratio HNK OKT-9 OKT-10 PMLC (S.I.) PHA {S.I.) Spont. Blasto. VTetnam Veterans (n*66) 38 61 20 39 24 1.8 11 3 ± 5 ± 29 ± 114 t 17.367 ± 15 13 Matched Controls (n=50) 44 ± 64 ± 15 14 ± 39 ± 11 23 ± 10 1.8 ± 08 . 6 12 ± 3± 5 11 6± 22 ± 21 90 98 ± 9.787 19,943 ± *Stati st1cal ly~TTTfer5nt~at'"P 1ess~than ~O57 19* 13 11* 10 8 0.7 7 1 7 13 78 10,136 �18 Interpretation and Conclusion; Of the 12 measures of the Immune system examined in this pilot study, the Active T-RFC was higher among the Vietnam veterans (20 t 15) than among the matched controls (14 t 11), (P less than 0.05). This test measures the % of Active T~cells which is the subpopulation of T-lymphocytes that function as immune surveillance cells. These cells are a subpopulation of the total T-RFC cells, which is reflected in a decrease of the % total T-RFC among Vietnam veterans (38 t 15) compared to matched controls (44 ± 19), (P less than 0.05). �i 19 .Summary: Because of concerns of Vietnam veterans that exposure to Agent Orange and its contaminants may have caused adverse health effects, The University of Texas System working closely with the Texas Department of Health, initiated three pilot research projects. These were (1) a study of the cellular characteristics of lymphocytes in the peripheral blood (cytogenetics), (2) a study of the number and physical appearance of sperm, and (3) several measures of the immune system. Vietnam veterans were purposely chosen who had the greatest likelihood of heavy exposure and were compared with age matched individuals with maximum likelihood of no exposure. The pilot phase called for 50 veterans and SO matched controls. Specimens were coded so that their Identities were blinded to the investigators when the laboratory tests were performed. The several limitations of these studies were made known from the beginning to concerned and interested Individuals. Preliminary r-esults of the three pilot studies are: Cytogenetlc Testing. No differences were found between the % of cells with chromosome breaks or the number of breaks per cell between Vietnam veterans and matched controls. Sperm Tests. Mo differences were found between the number of sperm, appearance of sperm, or percent of fluorescent bodies of sperm between Vietnam veterans and matched controls. Immunologic Studies. Of 12 tests performed to measure the immune system, the % Active T-RFC (which measures Immune surveillance cells) was higher among Vietnam veterans than among the matched controls (P less than 0.05). The % Total T-RFC was lower among veterans than among controls {P less than 0.05) �SUMMARY OF U.T. AGENT ORANGE BIRTH DEFECTS STUDY 26 MARCH, 1984 Data for the period 1 February, 1982 - 1 February, 1984 Center New Patients Seen 17 8 5 UTHSC Dallas UTHSC Houston UTHSC San Antonio UT Medical Branch TOTAL Disease Type Paternal Agent Orange Exposure* _2 33 = 0.6% Frequency in General Patient Population (Based on Partial Data) Frequency in Children of Agent Orange-Exposed Fathers* Possibly due to Agent Orange exposure in father (sporadic dominant or chromosomal anomaly) 15% 18% Not due to Agent Orange exposure in father (inherited dominant or chromosomal anomaly, autosomal recessive, or X-linked recessive) 18% 3% (Differences are marginally statistically significant) �Estimates of Frequency of Agent Orange Exposure* in Fathers of Children With Birth Defects of Certain Etiological Classes Class Conditions possibly due to Agent Orange exposure in father 0.75% (Difference is marginally statistically significant) Conditions not. due to Agent Orange exposure in father 0.1% CONCLUSION; Trend observed is consistent with fathers' exposure to Agent Orange causing birth defects in offspring, but numbers are very small; most fathers in the "exposed" group do not actually claim exposure; and the trend has become less clear as more data are collected. RECOMMEND; Continue data collection phase for 1 more year and re-evaluate at that time. J.M. Friedman, M.D., Ph.D. Associate Professor of Obstetrics and Gynecology and of Pediatrics Head, Division of Clinical Genetics *Defined as military service in Southeast Asia between 1969 and 1971. Most fathers were unaware of direct exposure to Agent Orange. �en c en «3> en cs> en SS.SiMlffiMiniKiTOWf^ijjlCIIIfiffiSffiWSSiaKffi- D en en cs:» KmiSqji!'Bii!i;i«fii9im KH* CJt ""^ O CO ni f ifii 1 5 Cu 11 %'"l:l; :~3l fJl ,,i. LJ "11 WIMIHtl u.» ni en rri —I LJ «.lH»m.ll " " I " " en CO CD 1 L™.* CO O nmitu, ui fn «jMc:.-wJ ANALYSIS OF MAJOR DEMOGRAPHIC STATISTICS FY 82,83,84 AGENT ORANGE CLINICAL STUDIES MAY 3, 1985 REPORT TO THE TEXAS AGENT ORANGE ADVISORY COMMITTEE BY GUY R. NEWELL, JR., M.D., CHAIRMAN �Demographic Characteristics of Vietnam Veterans and Controls Characteristic Vietnam Veterams (n-84) Controls (n=65) P Value (*) Race 0.44 Black Hispanic White Other 11.9 15.5 71.4 1.2 4.6 13.8 80.0 1.5 0.67 Language Spoken English Spanish Other 85.7 13.1 1.2 86.2 13.8 0.0 0.58 Religion Protestant Catholic Other 65.5 26.2 8.4 64.6 33.8 1.5 0.66 Marital Status Married Single Separated Divorced 66.7 13.1 2.4 16.7 76.9 9.2 1.5 12.3 0.0001 Education Through High School Through College Post College 35.7 56.0 8.3 13.8 53.8 32.3 0.0001 Income/Year <$5,000 5-<10,000 10-<20,000 20-OO.OOO 30-<40,000 40-<50,000 >50,000 13.2 17.1 21.1 23.7 17.1 3.9 3.9 0.0 3.1 10.8 27.7 27.7 18.5 12.3 �Current Consumption of Tobacco Type of Tobacco Vietnam Veterans (n=84) Controls (n=65) P-Value 52.5 35.5 0.04 20.0 20.0 17.5 22.5 20.0 28.6 19.0 23.8 9.5 19.0 1.8 98.2 3.7 96.3 0.97 8.1 91.9 5.5 94.5 0.85 8.9 91.9 3.7 96.3 0.46 Cigarettes (Pk/Yrs) Yes <9 9-<20 20-<26 26-<41 41+ Cigars Yes No I Pipe Yes No Chewing Yes No �Current Consumption of Beverages Vietnam Veterans (n-84) Controls (n=65) (X) Beverage (X) 0.24 Decaffeinated Coffee None 1-4 cups/day 5+ 76.9 18.5 4.6 84.9 15.1 0.0 Regular Coffee None 1-4 cups/day 5+ P Value 0.65 32.1 52.6 15.4 29.7 59.4 10.9 Tea 0.83 None 1-4 cups/day 36.8 63.2 33.3 66.7 0.27 Cola, Regular None 1-4 5+ 31.9 63.9 4.2 30.5 69.5 0.0 0.59 Cola, Dietetic None 1-4 85.5 14.5 80.0 20.0 0.92 Beer None 1-4 5+ 49.3 46.6 4.1 52.5 44.1 3.4 0.69 Wine None 1-4 89.1 10.9 92.9 7.1 Liquor None 1-4 0.21 89.4 10.6 79.7 20.3 �History of Drug Use Type of Drug Vietnam Veterans (n=84) (« Controls (n=65) m Medication, Regularly Prescribed Yes No 0.004 51.9 48.1 27.9 72.1 0.22 Medication, Past '60 Days Yes No 58.2 41.8 45.5 54.5 1.00 Recreational Drug Yes No P Value 8.5 91.5 7.7 92.3 �History of Exposure to Chemicals Vietnam Veterans (n-84) Exposure Controls (n=65) Routinely Exposed P value 0.85 Yes 30.5 27.7 No 69.5 72.3 Symptoms Related to Chemical Exposure 0.06 Yes 8.8 1.5 No 91.2 98.5 1.00 Solvent Yes 2.7 3.1 No 97.3 96.9 0.64 Paints Yes 10.1 14.1 No 89.9 85.9 1.00 Pesticides Yes 5.3 4.7 No 94.7 95.3 �History of Exposure to Specific Chemicals Chemical Vietnam Veteran Control P Value <« 1.00 Solvent Yes No 2.7 97.3 3.1 96.9 0.64 Paints Yes No 10.1 89.9 14.1 85.9 1.00 Pesticides Yes No 5.3 94.7 4.7. 95.3 �Miscellaneous History Vietnam Veterans (n=84) Controls (n=65) (« History w 0.76 Family History of Cancer Yes No Unknown 41.7 53.6 4.8 47.7 47.7 4.6 0.16 History of Cold or Flu Yes No 48.2 51.8 35.4 64.6 1.00 History of Vasectomy Yes No 20.2 79.8 20.0 80.0 0.001 X-Rays for Diagnosis Yes No P value 36.6 63.4 9.2 90.8 �Present or Past Employment by Occupation Occupational Titles Vietnam Veterans No. («) Professional, Technical, and Managerial 47 (04 2.) 107 (57 5.) Clerical and Sales 27 (.) H7 31 (61 1.) Services 33 (43 1.) 20 (04 1.) Agricultural, Fishery, Forestry 6 ( 26 .) 4 ( 2.1) Processing 6 ( 26 .) 3 ( 1-6) 42 (83 1.) 7 ( 36 .) 6 ( 26 .) 2 ( i.o) 7 ( 36 .) 11 ( 57 .) Machine Trades Benchwork Structural kbrk 41 (78 1.) Miscellaneous 22 ( 96 .) Total No. of Individuals No. per Individual X 2 = 80.22, P = 1.00 10 Control No. () % 230 192 83 65 2.8 3.0 �Present or Past Employment by Industry Standard Industrial Classification Vietnam Veterans No. () % Control No. () % Agriculture, Forestry and Fishing 5 ( 2.2) 4 ( 2.1) Mining 3 ( 1-3) 2 ( 10 .) Construction 20 ( 86 .) 5 ( 26 .) Manufacturing 28 (12.1) 18 ( 9-4) Transportation Communications Electric, Gas and Sanitary Services 25 (08 1.) 9 ( 47 .) 7 ( 30 .) 5 ( 26 .) Retail Trade 30 (29 1.) 22 (11.5) Finance, Insurance and Real Estate 10 ( 4-3) 5 ( 26 .) Services 36 (15.5) 60 (13 3.) Public Administration 64 (76 2.) 61 (18 3.) 4 ( 1-7) 1 ( 05 .) Wholesale Trade Not Classified 232 Total No. of Individuals No. per Individual X2 - 26.58, P = 1.00 192 83 65 2.8 3.0 �History of Major Health Problem Vietnam Veteran (n-77) History Control (n=64) Yes 32 ( 1 6 ) 4.% 13 ( 0 3 ) 2.% No 45 ( 8 4 ) 5.% 51 ( 9 7 ) 7.% = 6.3, P = 0.01 Twice as many Vietnam veterans gave a history of a major health problem as did controls. �Frequency of "Major" Health Problems « Veterans Controls No. Health Problem No. () * HBP 35.7 2.4 Hepatitis 15 1 0 0 2 1 2 0 0 0 0 0 1 1 1 3 1 1 1 1 1 1 1 1 1 3 1 1 1 Total 42 17 No. persons 32 13 Heart Irregularity LBP High triglycerides Heart disease, NOS Rh art Ulcerative colitis Low blood sugar Chronic bronchitis Hypersensitive insects High blood sugar Gout Ulcers Hidradenitis Feels sickly Anxiety Tbc Chloracne Liver cirrhosis Blood disorder Headaches Insomnia Chronic proctitis Spinal fracture Combat injury Diabetes Chronic discoid Edema Problem/person 18/42 — — — — — 2.4 2.4 2.4 7.1 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 7.1 2.4 2.4 2.4 1.3 Heart problems — —8 4 . 2.4 4.8 () * 7 1 1 1 1 1 0 1 1 1 1 1 0 0 0 41.2 5.9 5.9 5.9 5.9 5.9 -— 5.9 5.9 5.9 5.9 5.9 ___ 0 0 — — — __ — __ — 0 0 0 0 0 0 0 0 0 0 0 0 — — — — — __ --— 1.3 (42.9%) 11/17 (64.7%) �History of W>rk and Chemical Exposure Vietnam Veterans Controls (n=82) Exposure (n=65) Yes 25 No 57 (69.5%) X 2 = 0.14, P = 0.85 (30.5%) 18 (27.7%) 47 (72.3%) �History of Kbrk and Chemical Exposures Types of Veterans Controls Exposure No. () % No. 15 50.0 6 19.4 Radiation 2 6.7 4 12.9 Heat 4 13.3 1 3.2 Embalming fluids 0 1 3.2 Solvents 1 3.3 3 9.7 Fumes 3 10.0 1 3.2 Leaded gasoline 0 1 3.2 Paint thinner 0 1 3.2 Tylene 0 1 3.2 Phenol 0 2 6.5 Alcohols 0 1 3.2 Acids 0 1 3.2 Ether 0 1 3.2 Mold spray 0 1 3.2 Insecticides 2 1 3.2 Epoxy 0 1 3.2 Monomers 0 1 3.2 Miscellaneous 3 3 9.7 Noise/sound 6.7 10.0 Total exposures 30 31 No. exposed 25 18 Exposures/person 1.2 1.7 () % �DEPARTMENT OF THE ARMY OFFICE OF THE ADJUTANT GENERAL A R M Y AGENT ORANGE TASK FORCE ROOM 21O. 173O K STREET N.W. WASHINGTON. DC 2OOO6 R E P L Y TO A T T E N T I O N OF HERBICIDE STATUS REPORT The name Herbicide Orange comes from the identifying orange stripe painted on the drums containing a particular herbicide which contained equal proportions of the commercially-available herbicides 2,4-D and 2,4,5-T. These herbicides have been used extensively and in large quantities in agriculture and forest management in the United States (US) as well as worldwide for more than three decades. Only the 2,M,5,-T has been implicated as causing any potential health problems due to the presence of toxic contaminant - dioxin (2,3,7,8,-tetrachloro-dibenzo-paradioxin (TCDD)) - which is formed in low concentrations (parts per million) in the manufacturing process of the herbicide. At the request of the President of the Republic of Vietnam (RVN), the use of herbicides in Vietnam was approved by the President of the United States to primarily deny cover to the enemy and, secondarily, to deny food crops to the enemy. This was done only after testing in Florida, Hawaii, and South East Asia during 1961-1962, and limited operational use during 1962-1965. At that time, the herbicides used had the desired effects of improving visibility in dense jungles and were then believed to be harmless to humans. From 1965 to 1970, extensive aerial spraying was carried out over approximately 10 percent of the land mass of RVN, dispersing 11,300,000 gallons of Herbicide Orange in over 6,000 separate missions conducted by the U.S. Air Force under the code name "Ranch Hand". The missions were often carried out in remote or enemy-controlled areas as a result of the military need to improve observation of enemy activity and to reduce the potential for ambush. Each mission was carefully approved by identical staffing procedures within the US and RVN chains of command. The missions were flown under strict meteorological and operational conditions designed to minimize the drift of herbicide. Additionally, US and RVN commanders were advised to keep their troops out of the target areas at the time of spraying so that Vietcong grouncjfire might be returned by the fighter aircraft protecting the spraying missions. Nonetheless, spraying did occur over US troop positions. These missions are now recorded on computer tape (HERBS tape). In a typical spraying of dense jungle, tests have shown that only 6 percent of the herbicide reached the ground. At normal rates of application, this equals U millionths of a pound per acre of 'the contaminant 2,3,7,8-TCDD. Repeated testing reveals that 2,3,7,8-TCDD is rapidly detoxified by exposure to daylight in a matter of days, with a �half-life of approximately 6 hours. However, pure dioxin which has penetrated below the surface of the soil will persist for years, though it, too, will slowly detoxify. Dioxin is very insoluble in water and has a low vapor pressure. From 1965 on there are detailed computerized records of the dates, locations, types and amounts of herbicide used in fixed-wing "Ranch Hand" spray missions. The enclosed copies of maps, which were drawn from the records of spraying missions, show the locations of all "Ranch Hand" defoliation and crop destruction missions from 1965 to 1971. Herbicides were used, additionally, to clear the perimeter areas around US and RVN bases and along routes of communications to deny the enemy concealment capability and were applied with hand sprayers, and from tank trucks, riverine boats, and helicopters. While there are records of over 3.000 of these smaller scale applications, a complete compilation and computerization has not yet been accomplished, as documenting the instances and locations of firebase perimeter spraying is a painstaking, time consuming process. The DOD, however, considers this as another possible source of exposure and we are, therefore, continuing to search the records to determine the locations, dates, and magnitude of this type of perimeter herbicide spraying. The RVN armed forces are known to have used aerially "dispersed herbicides; however, no records exist of this usage. Finally, a small amount of herbicide was applied during 196? - 1969 in the Demilitarized zone (DMZ) in Korea. This was applied by hand spray apparatus and from trucks operated by Korean Army personnel. No US troops are known to have been involved or exposed in Korea. A study by Monsanto Chemical Company, of an accident which occurred at their Nitro, West Virginia facility in 19^9 has not shown an excess of deaths, cancers or heart disease among the 122 male workers who were conclusively proven to have been exposed to dioxin, in this incident when compared to the general US population. A similar study by Dow Chemical Company of 61 males exposed during a 1964 accident failed to establish a cause and effect relationship. However, because of the small population size in each of these studies, there is an acknowledged limited capacity for detection of normally infrequently occurring abnormalities or effects. Reflecting worldwide interest in the subject, studies of other similar accidents, including the one at Seveso, Italy, in 1976, are being conducted. Recent studies from Europe on forestry, agriculture and railroad workers suggest that two kinds of cancer, lymphoma and soft tissue sarcoma, may result from chronic, high exposure to dioxin. In animal studies, dioxin has been shown to be capable of acting as a promoter of cancer, 'fetal death and congenital defects but, to date, these effects have not been confirmed in humans. The reproductive effects have so far been observed only in pregnant rats and mice from large doses of dioxin, but not in rabbits, sheep or monkeys. There are marked species differences in sensitivity to dioxin1s effects. A recently completed study of male mice exposed to dioxin did not show any increase in fetal deaths or fetal abnormalities in the mated �females thus reducing concern about male-transmitted congenital abnormalities. An extensive study of the use and effects of herbicides in Vietnam was conducted by the National Academy of Sciences (NAS) and was reported to Congress in 1971. That study did not identify any specific health problems. Present interest in Herbicide Orange use in Vietnam centers on a wide range of exposures, from very low to high, actual and potential, htach of the present difficulty with the herbicide issue stems from the lack of concrete information about exposure and its consequences, expecially at low dose levels. There are no known, proven effects on health or reproduction from exposure to low levels of 2,4,5-T or dioxin. Nor do the health complaints voiced by those who believe they may have been exposed to Herbicide Orange fall into any discernible pattern. There is no significant marker or unusual condition such as chloracne, the rare skin condition which is a uniform sign of large, acute exposures to dioxin, to serve as a specific clue that low level exposure may have occurred. For example, with exposure to polyvinyl chloride or asbestos the remarkably consistent high incidence of otherwise very rare cancers substantially hastened an association of exposure to these substances and subsequent ill health. However, such a causal relationship has not been the case with dioxin. Thus, to date, there is no scientifically proven evidence that exposure to dioxin in very low doses leads to ill health or genetic defects. However, the matter is not being allowed to rest on that conclusion. There are many studies presently being carried on, both in and outside the Government, which are designed to investigate many of the unknown aspects of herbicide exposure. The lack of definitive information has heightened public and private concern about the possible human effects of exposure to dioxin. Within the DOD, the Air Force is conducting a study of the 1,200 men from "Ranch Hand" who performed the fixed-wing spraying of herbicides in Vietnam. The Ranch Hand study has been projected over a 20-year period and will be studying the long term health of the members of the "Ranch Hand" crews. The conclusions for the initial phase of this study, which was released in July 1983, were not Indicative of a cause and effect relationship. The conduct of an epidemiology study, originally to be by the Veterans Administration, has been assumed by the Centers for Disease Control (CDC) in Atlanta, and will examine the health of ground troops who were likely exposed to herbicide, as well as those who were likely not exposed to herbicides. Additionally, there will be considered the broader question of health effects of service in Vietnam in general, as it is possible that troops in Vietnam may have been exposed to other potentially toxic substances and exotic diseases. In addition to this large scale study (30,000 soldiers), the Centers for Disease Control is conducting a study to examine the possibility of increased incidence of congenital abnormalities among the offspring of Vietnam veterans. These studies will take several years to complete; however, �they offer the best possible hope of definitive answers to questions which at present have no answers. Critical to these studies, and to concerned individuals, will be information about whether a given individual was actually exposed to Herbicide Orange. In 1980, the Department of Defense initiated an intensive search of Army and Marine Corps unit operational records, morning reports/unit diaries, Combat After Action Reports, and other related troop movement records to determine if it would be possible to correlate locations of battalion and company size units with the Ranch Hand spray missions. We have found it is possible to identify certain selected companies as having been within close proximity of fixed-wing herbicide spray missions. The legislation of PL 96-151 mandated the Veterans Administration to conduct a study of possible health effects related to Agent Orange exposure. Following subsequent Congressional hearings, it was determined, since the majority of personnel who served in Vietnam were Array affiliated, that the Array would play the foremost role in providing the Department of Defense related data to support the VA's and related studies. Consequently, on 21 May 1980, The Adjutant General of the Array established the Array Agent Orange Task Force, drawing on the expertise of staff members already experienced in research methods and intensely familiar with the organization of the Vietnam War records collection. The Army Agent Orange Task Force, originally three full-time and two part-time members, now has a complement of 29 personnel and includes representation from the Air Force, Navy, and Marine Corps, comprising a joint services staff effort to support the veterans. The role of the Task Force involves in-depth research into the Vietnam War records of all branches of the services to locate units, identify those in relation to known herbicide spray missions, identify personnel within units, record incidents of herbicide sprays found in the records and previously undocumented, and to provide support to state and federal agencies conducting Agent Orange related studies. The records searches have demonstrated that there are significant differences in the quality, completeness and accuracy of the data contained in the records of the many units involved. It was never envisioned that these records, compiled and organized under combat conditions, would ever have to serve as th'e basis for scientific studies in determining exposure probabilities. Hence, some of the information needed is simply not available. During 1981, while DOD personnel were researching troop movement* records, another possible source of exposure to herbicides was uncovered — aircraft mission incidents. Records found to date indicate that over the years during which Ranch Hand missions were carried out, there were 155 incidents. These incidents were necessitated for a variety of reasons - engine failure, bad weather, radio malfunction, �navigational errors/problems and, in some instances, battle damage to aircraft. A mission incident did not necessarily mean that the pilot "dumped" the herbicide; however, the herbicide could be rapidly jettisoned through an emergency dump valve in less than a minute, to lighten the aircraft. To date, we have documented that emergency releases of herbicides took place 126 times, 58 of which definitely involved Herbicide Orange. The majority of these releases occurred at high altitudes, over the sea, or in remote areas in the vicinity of enemy held targets. A few, nonetheless, did occur near our bases. Those individuals who have unresolved health concerns from possible exposure to herbicides while serving in Vietnam may contact their nearest Veterans Administration hospital or regional office. Those persons still serving on active duty in the military services should contact their service medical facility. We remain dedicated to seeking answers to questions relative to Herbicide Orange and other dioxin-contaminated substances. �AGENT ORANGE STUDIES IN PROGRESS Compiled by the Veterans Administration STUDY AGENCY DESCRIPTION PROJECTED COMPLETION DATE Vietnam Veteran Mortality Study Veterans Administration To compare mortality To be Determined patterns and specific causes of death between those veterans who served in Vietnam and those veterans without Vietnam service. *Vietnam Veteran Identical Twin Study Veterans Administration To compare mental and physical health status of identical twin veterans, one who served in Vietnam and one who did not. 1986 Survey of Patient Veterans Treatment File Administration To identify morbidity patterns among Vietnam veterans from VA inpatient files. Retrospective Study of Dioxins and Furans in Adipose Tissue Veterans Administration To devise a method for 1985. determining levels of dioxins and furans in adipose tissue of Vietnamera veterans from samples in EPA's Survey of Human Adipose Tissue, to identify Vietnam veterans among the tissue samples and to analyze samples. Case-Control Study of SoftTissue Sarcoma Veterans Administration To determine whether Viet- 1985 nam service, Agent Orange exposure and other factors increase the risk of softtissue sarcoma. Department of Health A Human Services, Centers for Disease Control To evaluate possible longterm health effects of Agent Orange exposure on ground troops in Vietnam and to assess possible health effects of Vietnam service; 30,000 veterans expected to participate. *Epidemiological Study of Ground Troops Exposed to Agent Orange Initial 1983 1987 �Birth Defects and Military Service in Vietnam Department of Health & Human Services, Centers for Disease Control To determine possible Early association between Viet- 198U nam service and subsequent fathering of congenitally malformed children; based on Birth Defects Registry in Atlanta area which includes families of approx. 5,^00 case babies and 3iOOO '• control babies. Soft-Tissue Sarcoma . Investigation National Institute for Occupational Safety & Health To study tissues from Indefinite seven cases of soft-tissue sarcoma in U.S. (H who had been exposed to dioxin and 3 who may have been) in order to identify patterns of cancer that may be unique among those exposed to dioxin. Investigation of Leukemia in Madison County, KY National Institute for Occupational Safety & Health To determine possible association between cases of leukemia and exposure to wood ammunition boxes treated with hexadioxins. Dioxin Registry National Institute for Occupational Safety & Health To analyze causes of death 1985 among workers at 12 production sites where dioxincontaining products were manufactured. Internationa] Registry of Persons Exposed to Phenoxy Acid Herbicides & Contaminants National Institute of Environmental Health Sciences, with International Agency for Research on Cancer To establish an interIndefinite national registry of workers in some 20 plants where phenoxy acid herbicides were manufactured; mortality study planned when enough workers have been added to registry. Case-Control National. Cancer Study of Lymphoma Institute and Soft-Tissue Sarcoma Fall 1983 To compare herbicide 198U exposure among cases of soft-tissue sarcoma and lymphoma with controls of the same age, sex and Kansas county of residence. �Air Force Health Study Department of Defense To compare mortality and morbidity of Air Force personnel involved in Agent Orange spraying in Vietnam with a group of Air Force personnel who were not exposed •to the herbicide. Agent Orange Registry of Vietnam Veterans Biopsy Tissue Armed Forces Institute of Pathology To determine disease Indefinite patterns in biopsy tissue from Vietnam veterans; 1,200 specimens thus far show no unusual patterns,; especially of cancer. Preliminary Mortality 1983 Complete 1999 * Indicates those studies which are being supported through records research and review by the Array Agent Orange Task Force. . �Science Panel of the White House Agent Orange Working Group Represented by the following agencies: Department of State Department of Defense Department of Health and Human Services Department of Agriculture Department of Labor Environmental Protection Agency Office of Management and Budget Office of Science and Technology Veterans Administration Office of Technology Assessment Council on Policy Development of the White House �' VETERANS HEALTH SURVEY' Page l CDC continues to get inquiries regarding the status of its Agent Orange studies. Following is an update, which includes: Background Description of the CDC Research Project Agent Orange and Vietnam Experience Studies Selected Cancers Study Investigation Results h********4HH BACKGROUND Between August 1965 and February 1971 approximately 11.3 million gallons of the herbicide 'Agent Orange' (so named because of the orange markings on the drums in which it was shipped) were sprayed over much of South Vietnam in military operations designed to deprive the enemy of cover and food. A chemical contaminant, 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin, more often called TCDD, or simply dioxin, was created during manufacture of and contained in the Agent Orange which was sprayed. Dioxin has been shown to be a highly toxic substance. In January 1978 the Veterans' Administration (VA) received the first of what was to become many claims from veterans who felt that their current health problems had resulted from their being exposed to Agent Orange while serving in Vietnam. In January 1979 the U.S. Congress enacted legislation (Public Law 96-151) directing the VA to design and conduct an epidemiologic study to determine if exposure to Agent Orange had caused long-term adverse health effects in Vietnam veterans. In November 1981 the scope of the study was expanded (by Public Law 97-72) to include other factors in the 'Vietnam experience,' including medications and environmental hazards or conditions. In January 1983 the responsibility for designing and conducting the investigation was transferred from the VA to the Centers for Disease Control (CDC). In May 1983 CDC scientists completed detailed guidelines (protocols) for the Agent Orange and Vietnam Experience studies, recommending that a third investigation be conducted at the same time to determine the risk of Vietnam veterans developing selected types of cancers. Public 'Notice of Research Project Initiation' was published in the Federal Register on March 13, 1984. DESCRIPTION OF THE CDC RESEARCH PROJECT The study includes three separate but related components: 1) 2) 3) Agent Orange Study. (Study of the* long-term health effects of exposure to herbicides in Vietnam. ) Vietnam Experience Study. (Study of the long-term health effects of military service in Vietnam.) Selected Cancers Study. (Study to determine the risks of specific cancers among Vietnam veterans. ) July 1985 �Page 2 DESCRIPTIONi"AGENT"ORANGE'AND"VIETNAM'EXPERIENCE'STUDIES « Although both of these historical, or ''retrospective,1 studies are in some respects similar, each has a separate purpose. The Agent Orange study is designed to find out if troops who were exposed to the herbicide during service in Vietnam have suffered long-term adverse health effects as a result of that exposure. The Vietnam Experience study is designed to demonstrate whether or not there is any difference in the health of veterans of the Vietnam era who served in Vietnam compared to the health of veterans who served in other countries during the same period of time. The studies require the cooperation of a large number of Vietnam era veterans willing to be interviewed about their health status and experiences before, during, and after those years. TO ENSURE STATISTICAL ACCURACY, NO VOLUNTEERS CAN BE ACCEPTED AS PARTICIPANTS IN THE STUDIES. Participants are selected following scientific guidelines established by the research protocols. With the help of the Department of Defense and other agencies, CDC will identify a minimum of 30, 000 qualified veterans to participate in the studies: 6,000 in each of five separately defined groups or 'cohorts.1 The five cohorts are to be made up of veterans who: 1) Served during 1967-68 in a specified area of Vietnam, and were likely to have been exposed to Agent Orange. 2) Served during 1967-68 in the same area of Vietnam as cohort 1, and were less likely to have been exposed to Agent Orange. 3) Served during 1967-68 in another area of Vietnam than cohorts 1 and 2, and were not likely to have been exposed to Agent Orange. 4) Served in Vietnam during 1966-71. areas. 5) Served during 1966-71 in countries other than Vietnam. Randomly selected from all Data for the Agent Orange investigation will be gathered from cohorts 1, 2, and 3. Cohorts 4 and 5 will provide data for the Vietnam Experience study. PARTICIPATION IN THE CDC STUDY IS ENTIRELY VOLUNTARY. AGREEING OR DECLINING TO PARTICIPATE IN THE STUDY WILL HAVE NO EFFECT UPON BENEFITS A VETERAN MAY BE RECEIVING OR TO WHICH HE MAY BE ENTITLED IN THE FUTURE. �Page 3 fill information given by each veteran will be held in complete confidence. The names of the participants will never be associated with their answers in the statistical summaries studied by scientists. Names and other identifying information, such as addresses and social security numbers or service numbers, will be kept in a separate file that no one will have access to but the U.S. Public Health Service and the private research firms working on this study. No other researchers or government agencies, including the Veterans Administration and the Department of Defense, will be able to learn if a veteran participated or what his answers were. This promise of confidentiality is guaranteed by Federal laws—42 U.S. Code 242(b), (k), and (m). Unless the veterans gives written permission to CDC to release personal information, no one, including the veteran1 s family, will ever be ible to get the personal information provided by the veteran. The interview takes about 45 minutes and is conducted by telephone by CDC's contractor, Research Triangle.Institute (RTI), Inc. Veterans who are selected to be called by RTI receive a letter from CDC telling them to expect the call. From those being interviewed, approximately 2000 veterans from each cohort will have been preselected for the medical examination component of the study. The RTI interviewers have no control over which veterans will be asked to take the medical exams. Only veterans who have already been interviewed by RTI will be selected to be asked to take the medical exams which will take 3 days to complete. Several weeks after being interviewed, each veteran selected will receive a letter explaining the examinations and a telephone call from Lovelace Medical Center asking when he can come to Albuquerque. Veterans can select dates convenient to themselves. The 10,000 medical examinations are being conducted at non-hospital clinical facilities specially constructed for this project by another CDC contractor, the Lovelace Medical Foundation, in Albuquerque, NeW Mexico. All examinations are being done at the same place to ensure that standard testing procedures are used. The examination includes about 60 physical, psychological, and laboratory tests. Blood and urine samples are required, but no tests are included that most persons would find painful. Participants can refuse to take any test or to answer any question. Veterans who complete all the tests receive a $300 stipend. Veterans* expenses for travel to and from Albuquerque, food and lodging, etc., will be paid by!the government. Veterans will stay in private rooms at a first-class downtown hotel and have their evenings free. Each room will accommodate up to four persons without1 cost to the veteran. (The government cannot pay for family members travel or food.) Physicians and other health providers working on the CDC studies will not provide any treatment for individuals. If a veteran1s medical examination indicates the possible existence of a problem of any sort, the veteran will be advised immediately and encouraged to seek treatment from the VA, private, or other sources of medical services. �Veteran interviews for the CDC study began in September 1984, and M i l l continue until about October 1987. The first medical examinations were conducted in March 1985. fill examinations are expected to be completed by about January 1988. RTI, Lovelace, and other non-government research firms have been contracted to collect the data for these studies. These firms are monitored closely by CDC officials, fill analysis and interpretation of data is done by CDC. (((((((«((((((((((((((((«((((((((((((((((((((((((((((((((((((((( DESCRIPTIONS SELECTED CfiNCERS STUDY There is some scientific evidence that exposure to herbicides may increase the risk of several serious, but relatively rare, cancers in workers in industries which manufacture or use similar products. Because these cancers are so infrequently seen, the 30,000 veterans in the other study cohorts do not offer a large enough sample population upon which to base this investigation. Instead, two other groups w i l l be studied in a * case-control* investigation. Because of the design of this study, veterans and non-veterans will be included in both the case and control groups. The tumors selected for the study aret lymphoma, soft-tissue sarcoma, nasal and nasopharangeal cancer, and primary liver cancer. Other types of tumors may be added to the study later. The first (case) group in the Selected Cancers Study w i l l be made up of male patients who have actually had these tumors, and who could have been in the military during the Vietnam conflict. The second (control) group will include men of the same age and from the same current geographic area as the case cohort, but without the tumors. Using information from interviews and military records, CDC w i l l determine which men in both groups are veterans, which veterans served during the Vietnam era, and which veterans may have been exposed to figent Orange. Comparison of data collected from both groups may indicate significant differences in their risk of these cancers which could be associated with military service, service in Vietnam, and exposure to figent Orange. INVESTIGATION RESULTS » The exact rate of progress of epidemiological studies of this size cannot be forecast. Collection and analysis of the large amounts of data needed for scientifically valid findings takes time; particularly when so many thousands of veterans must be identified, located, interviewed, and examined. CDC w i l l report on each component of the study when it has been completed. Final reports on the figent Orange and Vietnam Experience components are expected by September 30, 1988. The final report on the Selected Cancers Study component is expected by September 30, 1989. CDC hopes that these studies w i l l provide answers to many of the important questions being asked about figent Orange and other factors related to service in Vietnam. But, as in every epidemiologic investigation — no matter how carefully designed and professionally conducted--the possibility exists that definitive answers to some questions may never be found. � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 069 Folder The folder containing the original item. 1867 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Anderson, Ron J. Robert Bernstein Description An account of the resource <strong>Corporate Author: </strong>Texas Veterans Agent Orange Assistance Program, Texas Department of Health, Austin, Texas Date A point or period of time associated with an event in the lifecycle of the resource 1985-08-01 Title A name given to the resource Annual Report Subject The topic of the resource state-funded Vietnam veterans study cancer risk assessment PSTD morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/2f079846877c4349c6f09e2beb20fe29.pdf 2f14b6260585640f7557acb72da2bea2 PDF Text Text Item D Number 0193 Author rott M J ° Corporate Author ROpOrt/ArtlOlO TltlB Australian Veterans Health Studies: Morbidity Study: Protocol for a Morbidity Survey of Vietnam Veterans and Controls Journal/Book Title Year 1982 Month/Day December 13 Color Number of Images 12 DeSCriptOU NotBS Cover sheet says, "A protocol was prepared for this study, but it was never carried out. The protocol is attached, and is a public document." ° Thursday July 26, 2001 Pago 1930 of 1957 �Australian Veterans Health Studies Morbidity Study A protocol was prepared for this study, but it was never carried out. The protocol is attached, and is a public document. �PROTOCOL FOR A MORBIDITY SURVEY OF VIETNAM VETERANS AND CONTROLS Written by: Dr M.J. Fett MB, BS (Hons), B Med Sc (Hons), MFH Consultant: Dr J.D. Mathews MD, BS, FRACP, PhD Director of Studies: Professor R.J. Walsh MB, BS, FRACP, FRCPA, FAA Date: 13 December 1982 �Ac k i iow ] c no front s The following people have provided specialist advice and information in their respective fields: Piof<:.;or G. Andrews - Psychiatry Dr S. Henderson - Psychiatry Profersor R. Kalucy - Psychiatry Professor J. McLeod - Neurology Dr C. Smith - Associate Professor C. Tennant - Psychiatry Dr K. Kalsh - Neuropsychology Professor Khitrod - Hepatology Psychiatry Many AVHS staff members have made contributions to this protocol. Among them are Drs A. Long, G. Nairn, E. Harding, B. O'Toole, Mr C. Fung, Mr I. Adams, Mr N. Kendrick and Ms J. Busby. This protocol was typed by Mrs J. Charles, Mrs S. Foster and Miss A. Micallef. �TABLE OF CONTENTS Page J. SUMMARY 1.1 1 Rationale for this Study of Morbidity Descriptive Hypotheses to be Tested 1 2 1.4 Summary of Study Design 3 1.5 1.6 2. Background 1.2 1.3 Cancer Morbidity to be a Separate Study Relationship to Other Studies 3 4 PACKGROUND TO PROBLEM 5 2.1 Stimulus to Study 5 2.2 Possible Effects of Phenoxy Herbicides and 2.3 2.4 3. Related Subr-tances Corlist Syndrome and Other Psychosocial Effects of War Service Effects Related to Specific Concomitants of War Service 5 10 13 BACKGROUND TO STUDY DESIGN 15 3.1 3.2 3.3 15 15 3.4 Rationale for Study of Former National Servicemen Enlistment and Training of National Servicemen Factors Influencing Selection of National Servicemen for Vietnam Service Comparability of Veteran and Non-veteran National Servicemen 3.5 3.6 3.7 3.8 Duration of National Service in the Army National Service Intakes Sources of National Service Cohort Data Base Verifying the Completeness of National Service 3.9 Determining the Accuracy of the National Service Cohort Rationale for Morbidity Study Subjects Being a Subset of Mortality Study Subjects Rationale for Subjects being N.S.W. Enlistees Only Subject Selection Obtaining Subject Addresses Areas to be Investigated Implications of Pilot Study Results Rationale- for Using MEDICHECK Health Screening Centre Rationale for Method of Measuring Morbidity Physician Assessment Rationale for Number of Subjects Index of Herbicide Exposure Index of Combat Exposure Army Dossier Data Held on Each Subject Cohort 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.27 3.18 3.19 3.20 3.21 3.22 16 19 20 21 21 23 24 24 25 27 28 29 31 32 34 42 44 47 49 50 �Page 4. STUDY PROTOCOL 54 4.1 4.2 4.3 55 56 59 4.4 4.5 4.6 4.7 4.8 4.9 4.10 5. Subject Selection Obtaining Current Address Methods of Obtaining Compliance for Medical Examination Outline of Design of Medical Examination Design of Examination Procedure for Subjects Information from Wife/Female Partner Follow-up of Non-Complying Subjects Follow-up of Non-Complying Wives/Female Partners Pilot Testing of Medical Examination Verifying Data in Medical Records 61 63 70 73 76 77 78 79 5.1 5.2 5.3 6. ADDITIONAL INFORMATION FROM ARMY SOURCES 79 79 80 CARD Dossier Central Medical Record Psychology Record Cards ' DATA ASSESSMENT AND ANALYSIS 81 6.1 6.2 6.3 6.4 6.5 81 82 91 93 97 Data Acquisition and Verification Adequacy and Utility of Morbidity Study Data Hypotheses to be Tested Outcome Measures to be Used Principles of Statistical Analysis References APPENDIX 1: APPENDIX 2: 105 SPECIFIC HYPOTHESES AND POWER CALCULATIONS PULHEEMS ARMY HEALTH RATING �1. PUKVARY 1.1 Background^ At the request of the Commonwealth Department of Veterans' Affairs, a study group was established in January 1980 to investigate the suggestion that herbicide exposure in Vietnam was responsible for health problems reported by Vietnam veterans and their families. The investigations, initially known as "Australian Veterans Herbicide Studies", are administered by the Commonwealth Institute of Health. In October 1981, the terms of reference were widened to include all disabilities related to Vietnam service, and not just those which might be due to herbicide expor.ure. In May 1982, the study group was renamed the "Australian Veterans Health Studies" (AVHS). 1.2 Rationale for This Study of Morbidity Comments from Vietnam veteran groups have led to suggestions that there is an increased illness rate among veterans, particularly in the following areas: psychological health, behaviour, liver, gastrointestinal system, skin and neurological system. There have also been claims of an increased incidence of cancer amongst veterans, and of marital and reproductive problems affecting veterans and their wives and offspring. The rationale for this study is to use information on current health and past medical treatment of veterans and their wives to test some of these claims. To achieve this it is proposed to follow-up and examine medically 5,000 former national servicemen who enlisted in NSW (3,000 veterans, 2,000 controls) and obtain reproductive histories from their wives/female partners. The results will be assessed to see whether the disabilities for veterans are more frequent than for national servicemen who did not go to Vietnam (controls). �In the event that, disabilities are more frequently found in veterans than in controls, additional data will be analysed to decide whether the excess disabilities are best explained in terms of the physical or social sequelae of combat stress, and/or in terms of herbicide exposure and/or in terms of individual differences which antedated the Vietnam experience (See Sec 3.4) . The rationale for restricting this study to national servicemen is explained in detail in Sec 3.1. 1.3 E^c.rJp.VLY? Hypotheses to be Tested The null hypothesis is that there is no difference in the frequency of disabilities between Vietnam veterans and controls (national servicemen who did not go to Vietnam) . This null hypotheses will be tested against each of the following alternative hypotheses: (i) That social and behavioural disabilities (unemployment, separation, divorce, motor accidents, alcohol abuse) are more frequent in veterans than in controls. (ii) That anxiety, depression and other psychiatric disabilities are more frequent in veterans than in controls. (iii) That disorders of the nervous system (including neuropsychological disorders) are more frequent in veterans than controls. (iv) That liver disorders are more frequent in veterans than controls. (v) That gastro-intestinal disorders are more frequent in veterans than controls. (vi) That skin disorders are more frequent in veterans than controls. (vii) That infertility, miscarriage or death or disability of children have been more frequent in the families of veterans than in the families of controls. �1.4 FuTu-nary of _Study _p_esiqn Veterans and controls will be former national servicemen who enlisted in those intakes from which veterans were chosen (from June 1965 to February 1971), who served at lear.t 13 weeks, and who were discharged alive or survived for 2 years after enlistment. , From the group of former national servicemen who enlisted in N.S.W., 3,000 veterans and 2,000 controls will attend a central examination site in Sydney. The medical examination will consist of an in-depth medical questionnaire, neuropsychialric testing, biochemical testing of blood and urine, and physical examination and health assessment by a doctor. Following this examination, certain subjects will undergo more detailed psychiatric and neuropsychological assessment (see Figure 1, chapter 4). Any subject with illness requiring treatment or urgent investigation will be referred back to his local doctor. Any subject with a suspected disability which requires further investigation for research purposes will be referred to an appropriate specialist, either immediately (option A) or after interim data analysis (option B). The wives of veteran and control subjects will be interviewed, to seek information about the health of any children of the subject, and about the outcome of all pregancies. 1.5 Cancer Morbidity Not Addressed by This Study This morbidity survey is not ideally suited to examining the incidence of cancer (see section 3.19) . �1.6 Kplat KMiship to Other Studies Separate studies have investigated the relationship of Vietnam service to birth defects in offspring (Case-Control Study of Congenital Anomalies and Vietnam Service), and are investigating mortality (Retrospective Mortality Study of Vietnam Veterans and Controls Revised Protocol). �2- BACKGROUND^ TO PROBLEM 2.1 St_im_u 1 u r, Jto_St udy In 1979 Vietnam veteran groups reported that there was an excess of veteran morbidity due to gastrointestinal, neurologic, skin and psychiatric disability and cancer, above that expected in a group of previously healthy young men. This supposed excess was attributed, by some veterans, to exposure to Agent Orange herbicide during Vietnam war service. Health effects other than morbidity are addressed in other documents (Case-Control Study of Congenital Anomalies and Vietnam Service Report, Mortality Study protocol). While the claimed effects of Agent Orange have yet to be substantiated, it has become apparent that several other environmental exposures, both in Vietnam and back in Australia, could be causally related to any increase in disability f in veterans. Among these exposures are non-phenoxy herbicides (e.g. cacodylic acid), insecticides, infectious tropical diseases and malaria prophylaxis, experiences of social dislocation and warfare, alcohol, tobacco and other drug consumption, and the experiences of homecoming, readjustment and re-establishing a satisfactory life style (Boman, 1982). 2.2 Possible Effects of Phenoxy Herbicides and Related Substances 2.2.1 General Literature Review The 'Review of Literature on Herbicides, Including Phenoxy Herbicides and Associated Dioxins1, (U.S. Veterans Administration 1981) summarises the known and suspected health effects of herbicides, and the gaps in current knowledge. study of Vietnam veterans are: Points of relevance to a morbidity �TCDD (Dioxin) o TCDD has been an important contaminant of 2,4,5-Tr and it has not always been possible to distinguish between the effects of the two substances. o Chloracne is the most consistently reported health effect of TCDD exposure in huir.ans. o Neurasthenia, a series of subjective complaints including irritability, fatigue and insomnia, has been reported after many industrial accidents and exposures, o Other neurological disorders (as peripheral neuritis) and hepatic disorders (as hepatomegaly) have been reported after several of the exposure incidents. o Porphyria cutanea tarda and gastrointestinal problems have not been commonly reported and seem to be associated with long-term exposure. o TCDD is a limited cumulative toxicant; cumulative effects of doses administered within a month of each other have been observed in animals, but not for doses administered beyond about one month, o The subacute effects of TCDD are porphyria and depletion of blood cells; these effects are not observed after acute doses, o In animal studies TCDD appears to act secondarily or indirectly in enhancing the carcinogenicity of other components (usually unidentified). 2,4-D, 2,4,5-T o Both 2,4-D and 2,4,5-T are cleared rapidly from the blood after they are absorbed, with half-times for plasma clearance in humans of 12-23 hours. �o Neither 2,4-D nor 2,4,5-T has been shown to accumulate in animal fat. o 2,4-D and 2,4,5-T are not cumulative toxicants. o In animals the cause of death from lethal doses of 2,4-D or 2,4,5-T is unknown; both compounds produce several non-specific effects, such as mild weight loss. o 2,4-D produces neurotoxicity in humans and animals, and 2,4,5-T produces neurotoxicity in animals. o Animal studies have not produced conclusive evidence that 2,4-D, 2,4,5-T, cacodylic acid or picloram are carcinogenic. o The effects of acute exposure to 2,4,5-T in humans are unknown. o There is no positive information on the carcinogenic potential in humans of diquat, diuron, dalapon, bromacil, picloram, and tandex and on 2,4-D, 2,4,5-T, or TCDD, (except in the case of concomitant exposure to trichlorophenol or other herbicides). IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Kan: Volume 15 gives the following effects of TCDD, 2,4-D and 2,4,5-T as having been described in humans: TCPD chloracne porphyria cutanea tarda hyperpigmentation and hirsutism liver damage raised serum hepatic enzyme levels disorders of fat metabolism disorders of carbohydrate metabolism cardiovascular disorders �urinary tract disorders respiratory disorders pancreatic disorders polyneuropathies lower extremity weakness scnr.orial impairment (sight, hearing, smell, taste) neurasthenic or depressive syndromes 2,4-D hyperthermia and tachycardia nausea, vomiting, diarrhoea anorexia and gastralgia increased salivation sweet taste in the mouth abdominal pain fatigue, malaise headache sensation of drunkenness peripheral neuropathy paralysis somnolence muscular weakness, twitching skeletal muscle damage severe leg pains joint swelling increased blood cholesterol abnormal serum protein �chloracne liver disorders neurological changes behavioural changes fat metabolism disorders signs of porphyria cutanea tarda gastrointestinal symptoms headache hypomania neurasthenic syndrome The effects listed have generally been observed in subjects after recent chemical exposure, and the pattern of effects varies between reports. 2.2.2 Possibility of Carcinogenic Risk in Man Soft-tissue sarcoma and lymphoma are suspect as outcomes of exposure to phenoxy herbicides or associated substances (Hardell and Sanstrom, 1979). The work of Hardell has received some support from reports of at least four cases of soft-tissue sarcoma which have occured in workers involved in the manufacture of phenoxy-herbicides (Hardell and Ericksson, 1981; Honchat and Halperin, 1981). Several other studies of the effects of herbicides on cancer mortality have failed to demonstrate an effect (Axelson and Sundell, 1974; Riihimaki et al, 1978) and no excess mortality has been demonstrated following the 1976 Seveso dioxin disaster (Regianni, 1980), although the latent period is short. �2.3 ?f?:Tll''?t._?ynF'?c?TTic!!:. ^nj-LPth.?1 Psychospcijl Effects of War Service There are no published epideroiologic studies of the effects of Vietnam service on the psychological adjustment of Australian soldiers, nor on their subsequent readjustment to civilian life; however, literature indicates three main areas of psychopathology in U.S. veterans: post traumatic stress disorder; depression; alcohol and substance abuse disorder (review by Boman, 1982; symptoms after DSM III). 2.3.1 ?9.st..~?.r_fiu!!!ati5: Stress Disorder Symptoms: Re-experiencing of the trauma; numbing of responses to or reduced involvment with the external world; hyperalertness, sleep disturbance, guilt about surviving; memory impairment or trouble concentrating. 2.3.2 Depression of various types Symptoms: Anorexia, weight loss or gain, increased appetite; sleep disturbance, psychomotor agitation or retardation; feelings of worthlessness; loss of interest, loss of energy, fatigue; complaints or evidence of decreased ability to think; recurrent thought of death or suicide; suicide attempt. 2.3.3 Alcohol and Substance Use Disorder Symptoms: Alcohol or drug consumption causing impairment in social or occupational function, with or without tolerance or withdrawal. 10 �2.3.4 Othc- r_ Psy chospci a 1_ Disorders Three other classes of disorder seem possible in view of media reports of the health of Australian Vietnam veterans. These are: anxiety disorders; adjustment disorder with anxious mood; somatization disorder (symptoms again taken from DSM III). Anxiety Disorders Symptoms: apprehension, fear or terror; dyspnoea, palpitations, chest pain or discomfort; choking or smothering sensation; dizziness, vertigo or unsteady feelings; feelings of unreality, paraesthesias, hot and cold flushes; sweating, faintness, trembling or shaking; fear of dying, going crazy or doing something uncontrollable. Adjustment Disorder with Anxious Mood Symptoms: as for anxiety disorder, with nervousness, worry or jitteriness predominating. Somatization Disorder Symptoms: Sickly: Believes that he or she has been sickly for a good part of his or her life. Conversion or pseudoneurological symptoms: Difficulty swallowing, loss of voice, deafness, double vision, blurred 11 �visionf blindness, fainting or loss of consciousness, memory loss, seizures or convulsions, trouble walking, paralysis or muscle weakness, urinary retention or difficulty urinating. Gastrointestinal symptoms: Abdominal pain, nausea, vomiting spells, bloating (gassy), intolerance (e.g. gets sick) of a variety of foods, diarrhoea. Psychosexual symptoms: For the major part of the individual's life after opportunities for sexual activity: Sexual indifference, lack of pleasure during intercourse, pain during intercourse. Pain: Pain in back, joints, extremities, genital area (other than during intercourse); pain on urination; other pain (other than headaches). Cardiopulmonary symptoms: Shortness of breath, palpitations, chest pain, dizziness. Other psychiatric abnormalities may occur, but the number of affected individuals will probably be very small. 12 �2.4 Ii(f_e.£tE_Rc3at_ed _to_Sppci_f ic Con corn Habits of War Service 2.4.1 Alcohol Use Because of combat stress and the social changes consequent on Vietnam service, it is likely that the consumption of alcohol by Vietnam veterans will prove to be higher than that in non-veterans. This would lead us to expect that there could be a greater incidence of alcohol related disability in veterans than in controls, as has been reported in other army and veteran populations (Mathews 1976). It is also necessary to consider the possibility that some of the personal qualities which are associated with being a "good soldier" (e.g. vigour, initiative, aggression) may show an intrinsic (genetic) correlation with the propensity to drink more alcohol (Mathews, 1981) . if this were the case, then any association of alcohol use with veteran status could be partly consequent on the processes of veteran selection. It will be possible to test this hypothesis in the morbidity study, as data on alcohol consumption will be collected from subjects. 2.4.2 Cigarette Smoking and Other Drugs Similar considerations apply to cigarette consumption. A particular attempt will also be made to identify any morbidity which could be attributable to illicit drug usage. 2.4.3 Infectious Diseases It is likely that Vietnam veterans have been at greater risk of several infectious diseases (e.g. malaria, hepatitis, melioidosis, strongyloidiasis), although there is no specific evidence to suggest that this has contributed to any disability in the post-Vietnam period. 13 �Tuberculosis might also be more prevalent in veterans, because of the greater exposure in Vietnam, and the greater susceptibility associated with a history of social disintegration and alcohol abuse. However, because of the efficacy of treatment, it is unlikely that there will be any subjects with tuberculosis. Because of the social dislocation of the war experience, venereal disease may also be more prevalent in veteran than in control cohorts. 14 �3 - 3• * J^CKJSROUND TO ST_IJDY_ DESIGN Rationale for Study of Former Kational Servicemen In studying the effect of Vietnam service on subsequent risk of disability, a group of subjects who served in Vietnam (veterans) is required, along with a comparison group of subjects who did not serve in Vietnam (controls). The veterans and controls should have been as similar as possible at the point in time at which the veterans departed for Vietnam. Specifically, the distribution of age, health status and socio-economic status (SES) should be similar between veterans and controls. Former national servicemen are considered to form a group most closely satisfying these requirements and consequently the study will be restricted to this group. At present it is considered that it would be too difficult to * define an appropriate control group for those Vietnam veterans who were in the regular army. If further investigation were to be performed, it may prove feasible to identify acceptable groups of veterans and controls among the regulars, but even so it would still be highly desirable to perform the study with national servicemen for two reasons: firstly, national service veteran and control groups would, in all probability, still be more comparable than groups of regulars, and secondly, within regulars the health effects of Vietnam service could be obscurred by the deleterious health effects of peace-time army service (see above). 3.2 Enjy.?j-!n±nt_A"d Training of National Servicemen Nineteen year old Australian males were required to register for national service (National Service Act 1957-1971), and those with birth dates selected by ballot were required to present for medical examination and interview by the Department of Labour and National Service. 15 Those accepted were required �to enlist in the army within the next few months, unless deferrment (educational, medical reasons) or exemption (clergyman, conscientious objector) was obtained. Then followed an enlistment medical examination. Some men volunteered for national service independent of the ballot. They had to reach the same medical standard as ballotees. These men can only be identified by examination of individual, dossiers. After enlistment followed 10 to 12 weeks of recruit training, then allocation to corps and 12 weeks Initial Employment (corps) training (except Infantry). The member was then posted to his service unit. Generally, between 6 months and 1 year after enlistment the veterans were sent to Vietnam for a 1 year tour of duty . » 3.3 Factors Influencing Selection of National Servicemen for Vietnam Service Selection for Vietnam service was based on whether an individual had been allocated to a unit which was subsequently selected for Vietnam service. Very few if any national servicemen were prevented from serving in Vietnam for medical conditions and the interval between enlistment and posting to Vietnam (less than 1 year) was insufficient for life-threatening conditions to develop in a significant number of national servicemen which would result in exclusion from Vietnam service. Many factors influenced the selection of a national serviceman for service in Vietnam. Those currently known to the Study Team are as follows: 16 �a) Indi_y ijdua 1^ Soldi^e£_E__I_nf luence Completing form NS24, giving details of education, special skills and t r a i n i ng. Completion of a "dream sheet1 which recorded desires to serve in particular corps. It is believed that in a considerable proportion of cases these preferences were recorded on the psychology record cards, which are available. It is thought to have been common knowledge which units were going to Vietnam in the next couple of years, and therefore a national serviceman could steer himself toward a corps which was more or less likely to be sent to Vietnam. Additionally, the field corps (Artillery, Armoured, Infantry, Engineers, Signals) were known to be the most dangerous. Volunteering for a reinforcement unit. Expressing a desire to serve in Vietnam. Performing in a sufficiently unsatisfactory manner so as to be regarded as unsuitable for Vietnam service. A fear of combat, or conscientious objection to the Vietnam war or combat. Lack of physical fitness becoming apparent during basic training but not necessitating discharge; lack of psychological fitness for combat service, as evident on enlistment testing or as determined by his commanding officer. 17 �An accident resulting in injury prior to posting to Vietnam, but not necessitating immediate discharge (due to its minimal severity or the need for prolonged medical treatment). Family circumstance, such as illness, social problems etc. resulting in the national serviceman being granted leave without pay or being posted to a base near home, precluding Vietnam service. b) Army influence Requirement lists. These listed the manpower requirements of corps and units. The allocation boards attempted to match 'dream sheets' to 'requirement lists', probably with varying success. The 49/51 rule. The Army maintained Vietnam service unit strength at 51% regulars or above, although it appears from 1968 onwards the lack of available regulars meant that the ratio sat on 49% national servicemen and 51% regulars. A national serviceman's superiors regarding him as unsuitable for Vietnam service. It is probable that these factors, while influencing Vietnam service, would also influence subsequent morbidity and mortality. The number of national servicemen who sought or avoided Vietnam service is not known. 18 �As this problem is one of confounding, its role may be evaluated, at least in part, at the data analysis stage once the requisite data has been obtained from the army. It appears possible to identify individuals transferring into or out of Vietnam bound units by searching through individual personnel dossiers. The proportion of national servicemen who became veterans is similar across all States of enlistment (see Table 3.1) although anecdotal evidence indicates that some units were comprised mainly of enlistees from certain States. OLD NSW vie SA- WA TAS NT Total subjects enlisted in State 14% 32% 30% 10% 10% 4% 0% % of subjects who are veterans 43% 42% 39% 40% 42% 44% Table 3.1 Origin of subjects by State of enlistment, and % veterans of subjects enlisted in each State. 3.4 Comparability of Veteran and Non-Veteran National Servicemen As indicated in the previous section, there is ample evidence that the decision to send a national serviceman to Vietnam was not random. In the absence of random allocation, it is likely that even before the Vietnam experience, those who eventually went (veterans) would have differed, in several important respects, from those who did not go to Vietnam (controls). 19 �This view is supported by evidence that veterans differed from controls in educational level and in scores on the SDI psychological scale at induction. In as much as these pre-Vietnam differences between veteran and control national servicemen are measurable, they can be treated as potential confounding factors. At the stage of analysis of results it will be possible to see whether factors such as education and psychological type are related to outcome, and if so, to make statistical adjustments to minimise the effects of the confounding. However, it is important to emphasise that because of (undetectable) errors in the measurement of these confounding factors, such statistical adjustments will always be incomplete. Furthermore, no statistical treatment could ever allow for the effects of confounding factors which are unmeasured (and possibly unsuspected). 3.5 Duration of National Service in the Army Discharge occurred after 2 years service (reduced to 18 months in 1971) unless discharge occurred early for extraordinary reasons (medically unfit, exceptional hardship, change of Government) or late (retention for medical treatment, voluntary prolongation of service). An analysis of 'discharge reason1 by 'duration of service1 for those intakes with veterans revealed that 53% of discharges as 'medically unfit1 occurred in the first 3 months, 67% by 6 months and by 12 months 80% of all such discharges had occurred (see Table 3.2). For 'expiration of term1, 0% occurred in the first 12 months, 29% in 13-24 months, 65% in the 25th month, and 5% after 25 months. The 5% of discharges after 25 months service may be due to errors in enlistment and discharge dates or retention of servicemen in 20 �the Army for medical treatment or voluntary prolongation of service. The reasons for those delayed discharges will be explored by manual searching of the CARD dossiers. Reason for Discharge % Discharged in Time Interval of Duration of Service 0-3m 7-12 m 13-24 m 25m + 4m- 6m Total No. 100% 0% 0% 14% 19% 4% 10% 0% 0% 13% 22% 13% 22% 29% 0% 15% 19% 23% 23% 70% 100% 5% 6% 52% 42% 40829 Other 0% 0% 53% 34% 6% 3% Total 5% 2% 3% 27% 63% 49881 Expiration of term Exceptional hardship Medically unfit Unsuitable, non-discipl. Unsuitable, disciplinary 63 3661 874 316 4138 Table 3.2 Reason for Discharge by Duration of Service 3.6 National Service Intakes Department of Defence Army Manning Reviews divide national service enlistments into 4 intakes each year. Each intake lasted up to 2 weeks, and they occurred in January, April, July and September. A minority of enlistments occurred in the other months, and for the purposes of this study an intake consists of all national servicemen who enlisted in 1 of the above months or enlisted in the month on either side of that month. 3.7 Sources of the National Service Cohort Data Base The following information is available from the army for current and former servicemen: Service number Surname Given names 21 �Date of birth Date of enlistment Marital status at enlistment Educational status at enlistment Occupation at enlistment Religion Nun-iber of dependents Place of birth Army health classification at enlistment Postings Dates of postings Date of discharge Discharge reason Army health classification at discharge » Medical record and psychology record data Other To date, two army sources have been used to provide data about national servicemen in the Vietnam period. These are the Central Army Records Office (CARD) and the Melbourne Regional Computer Centre (MRCC), both located in Melbourne. MRCC An MRCC-supplied computer file contains most of the above data for most study subjects. However, it has several major short-comings. Firstly, it contains initials, not given names (which are required for obtaining subject addresses) and does not contain postings data, medical or psychology data, or other miscellaneous service history data. Secondly, preliminary examination indicates that the data it contains are less accurate than those in the CARD 22 �dossier. As discussed below the first match to obtain addresses will be computerised: matching study subjects' MRCC-file derived names (surname, 2 initials) and date of birth with the Australian Electoral Register. It is proposed to overcome the problems of MRCC file inaccuracy by returning all unmatched subjects names and d.o.b. to CARO for manual verification and addition of 2 given names (CARO contains dossier records for all current and former servicemen). Where corrections have been made, the subjects will be matched with the Australian Electoral Register again, and the residue not matching will be sought in other registers. CARO In addition to the above limitations, the MRCC file does not indicate veteran/control status, and neither dates nor names of Vietnam postings. These are contained on a computer file held by AVHS compiled from data manually extracted from individual CARO dossiers. While this file does have 2 given names for all veterans, it only gives data for 17% of controls. There is evidence of misclassification of veteran/control status, probably less than 1%. 3• & Verifying the Completeness of the National Service Cohort The completeness of the cohort has two aspects: firstly, whether we know of the existence of all of the national servicemen, and secondly, whether we have complete details on those known. This work has already been performed as part of the Mortality Study. All national servicemen have been identified, and the data listed in 3.7 above is present in at least 98% of cases. 23 �3• 9 P?i.cJ.?r?r1 Ln.9 _yi£.^Jr-SmiacV °* the Nationa 1 JSe r_vi£e_ Cohort Generally, data on the MRCC file will be used for the purposes of the Morbidity Study, as this is the only file with complete coverage of national servicemen. The CARD computer file has much greater coverage of veterans than controls and is therefore a biased source. The MRCC - CARO match was used to determine the error rates in the MRCC data in the following way: If an inconsistency emerged in the MRCC - CARO match the origin of this inconsistency was determined. This involved checking that the computerised CARO data were correctly entered, and, if so, that they were correctly transcribed from the original CARO dossiers. The remaining discrepancies were corrected by referral to army personnel dossiers. 3.10 Rationale for Morbidity Study Subjects Being a Subset of Mortality Study Subjects Mortality Study subjects comprise all former national servicemen who saw service in Vietnam, and all other national servicemen from those intakes which also included veterans who stayed in the army at least 13 weeks. The 13 week minimum army service criterion is proposed for several reasons: Initial recruit training lasted from 10 to 12 weeks and the majority (53%) of discharges described as "medically unfit" occurred during this time. Recruit training therefore acted as a further screening procedure to identify and discharge these persons not suitable for army service. By the end of recruit training those remaining in the army would be considered suitable "material" for selection for Vietnam service. A minimum service duration of 13 weeks for subject selection therefore ensures that subjects, both veterans and controls, had an adequate health standard at the time of enlistment. 24 �This rationale for defining mortality study subjects applies equally to morbidity study subjects. Since all 48,600 national servicemen fulfilling the above criteria will be selected for the mortality study and only 5,000 need be studied for the morbidity study, the morbidity study subjects can be a subset of mortality study subjects. Vietnam service generally commenced 9 to 12 months after enlistment. Therefore servicemen discharged between 10 weeks and 9 months after enlistment were not eligible to become veterans. Nevertheless, they will be included, because their exclusion would prevent investigation of the relationship between morbidity and the factors associated with early discharge. If morbidity is found to be related to a history of early discharge in the control sample this will provide evidence of the magnitude of the effects on morbidity that can arise from differences which are not related to the Vietnam experience. If necessary, these 'short service1 controls could be excluded from the analysis at a later stage, to allow comparisons of morbidity to be made in veterans and controls with similar duration of army service. 3.11 Rationale for Subjects Being N.S.W. Enlistees Only To enhance the logistic feasibility of the morbidity study it is proposed to select subjects who enlisted in N.S.W. N.S.W is chosen as it is the most populous State and the study team is physically located in Sydney, within 200 kilometers of approximately 85% of the State's residents. N.S.W. enlistees who have moved interstate will be sought and encouraged to participate. N.S.W enlistees who have emigrated from Australia will be deemed to be unavailable until they return to Australia. 25 �The assumption underlying this state of enlistment approach is that a large majority of N.S.W. enlistees are still living in the State. This assumption has been examined by matching former national servicemens1 names and dates of birth with the Australian Electoral Register of July 1981 (see Table 3.3) and tabulating the State of residence on the register. Approximately 6,000 veterans and 9,000 controls enlisted in N.S.W, many more than required for the morbidity survey (see sec. 3.19) . Veteran 3515 6155 2728 4971 % of Names Matching 78% 81% % of Names Not Matching (b) Number of Names Sought Number of Names Matching (a) Control 22% 19% State Distribution of Names Matching with Australian Electoral Register. Table 3.3 State of Residence of N.S.W. Enlistees. 1981 Electoral Register State Veteran Control NSW 88% 91% VIC 2% 2% OLD 6% 5% SA, NT 2% 1% WA 2% 1% TAS 0.4% 0.4% 12% Not in NSW 26 9% �The results in Table 3.3 are likely to be the best possible expectation of reality ar> the 19-23% not matching on the Australian Electoral Register may be more evenly spread across Australia or overseas. In addition, errors in the Electoral Register State of residence due to delays in entering change of address data are likely to over-represent N.S.W. since N.S.W. enlistees started out there. Nevertheless, the result is encouraging, and also suggests that confining the medical examination facilities to N.S.W. may be the most economical way of executing the study. N.S.W. enlistees are comparable to enlistees from other states in terms of distribution across corps, proportion of veterans overall, and proportion of veterans within each year of enlistment. » An alternative approach is to select current N.S.W. residents irrespective of State of enlistment. There would be the potential for major bias, in that mobility after Vietnam service could be strongly correlated with morbidity, and, in terms current State of residence, differ among the States (e.g. mobile people heading to Qld and NT). 3.12 Subject Selection To be eligible for subject selection, a former national serviceman must fulfil the following requirements: (i) Served in the Army at least 13 weeks, to ensure uniform enlistment health status (see section 3.10); (ii) Be enlisted in an intake from which Vietnam veterans were subsequently drawn. (iii) Intakes after February 1971 contain no veterans; Be enlisted in N.S.W. (see section 3.11). 27 �The pool of former national servicemen fulfilling these three criteria will be stratified by veteran/control status and enlistment intake. With regard to the total number of veterans required (see section 3.19), random selection of a constant proportion of veterans within each intake will be made. For every intake from which veterans have been selected, 2 controls will be randomly selected for every 3 veterans. This will ensure that both the structure of ages and chronologic years of army and Vietnam service are similar for veterans and controls. 3.13 Obtaining Subject Addresses Inviting subjects to a medical examination requires knowledge of their current address. * Initially, address will be obtained from the Australian Electoral Register, (probably 80%) and drivers licence registers (a further 10-15%). Follow-up data from the Mortality Study would be useful in this regard. Additional negotiation will be required to obtain addresses from licence registers. However, in the Pilot Study it was found that only 87-90% of subjects could be located using a known Electoral Register address. It is proposed therefore, that when a morbidity study subject is unable to be located through an address obtained via the mortality study, additional address searching will be carried out. Since all subjects will be sought in the Australian Electoral Register and computerized licence registers, there will be no additional search requirement for these sources. Additional searching will involve manual drivers licence registers (Vic, Qld) the latest Electoral Register microfiche, commercial 28 �credit bureaux, telephone books and Telecom customer files, and Social Security records, if available. It is probable that 10% of morbidity subjects will have to be sought in this way. This second stage searching, specific to the morbidity study requirement for current addresses, will be carried out concurrently with the medical examination field work, as the failure to contact a subject will only become evident at this stage. If available, the current Australian Electoral Register tape will be used as the initial source of the addresses of subjects. Addresses, once obtained, will be held on computer file to facilitate the control and monitoring of subject contact and participation. 3.14 Areas to be Investigated Veterans have expressed concern about the health of themselves and their offspring in several areas. However, confining investigation to these areas alone would ignore many areas of disease which might have arisen from service in Vietnam, and which might, at a later stage, become significant to the veterans. 3.14.1 Veterans Areas of Concern Birth defects - see Case-Control Study of Congenital Anomalies; Death from a variety of causes - see Retrospective Mortality Study; Cancer - to be investigated in part via the Mortality Study; Abnormalities of behaviour (eg outbursts of rage); Substance abuse (eg excessive use of alcohol, tobacco, illicit drugs, prescription drugs); Relationship difficulties (eg divorce, social disabilities); Psychiatric disorders (eg depression, anxiety); Reproductive disorders (eg miscarriages, infertility); 29 �Liver disorders; Gastrointestinal disorders; Neurological disorders of both the C.N.S. and P.N.S.; Skin abnormalities. 3.14.2 Morbidity Related to Concomitants -of War Service. - ._ Corcbat injuries, disability and crippling Psychiatric disorders (see sec. 2.3): Post-traumatic stress disorder Substance abuse: alcohol - gastrointestinal ulceration, liver disease, hypertension, degeneration of nervous system and heart, psychoses tobacco - chronic obstructive pulmonary disease, respiratory tract infection, cardiovascular disease, cancer miscellaneous drug dependencies Depression and anxiety disorders: wide variety of concomitant symptoms Somatization disorders: wide variety of concomitant symptoms affecting nervous system, gastrointestinal system, cardiopulmonary system, psychosexual functioning Infectious diseases Only four diseases endemic to Vietnam could (arguably) still affect veterans: Melioidosis Strongyloidiasis Syphilis Tuberculosis 30 �All other infections will have either spontaneously resolved or become sufficiently florid to necessitate treatment and cure. Although vivax malaria can persist in a latent phase for long period of time it is not considered that this could be a significant cause of morbidity. 3.14.3 Morbidity Implicated in Herbicide Literature Occupational studies of the long term effect of human exposure to chlorinated phenols have revealed cases of soft tissue sarcoma, and the Swedish case-control studies, suggest that phenoxy herbicides could cause soft tissue sarcoma and lymphoma (Hardell and Sandstrom, 1979, see also section 2.2). 3.15 Implications of Pilot Study Results The Pilot Study has provided information about the value of many aspects of methodology which have been considered in the planning of this morbidity protocol. Information from the Pilot Study will also be used in the planning of detailed procedures if this present protocol is approved in principle. Briefly, the pilot study results show that: o telephone and face to face interviews were more expensive ($100 and $95 respectively) than self-administered questionnaires ($60) posted to the home o response rates were higher for telephone and face to face interviews than for self-administered questionnaires o veterans responded to the request for an interview more frequently than controls o self reports on army unit, subunit, corps and veteran status are likely to be useful 31 �o self reports on posting dates and operations in Vietnam are unlikely to be useful o sell reports on exposure to chemicals in Vietnam are unlikely to be useful o self reports on exposure to chemicals at work and home are more likely to be useful - - o cigarette smoking was reliably reported o the use of alcohol was reliably reported, but with the instrument used drinking frequently was less reliably reported o false positive and false negative rates for conditions reported in the medical history were high, especially for conditions occurring more than one year prior to interview o separation of the medical history and physical examination meant that the examining physicians were hindered by a lack of contextual clues in making judgments about morbidity in the subjects studied o morbidity was detected in both veteran and control subjects; target conditions based on subjective responses were found at highest frequency o the psychology tests used were insensitive in detecting psychopathology o a proportion of the reports from wives on miscarriages, birth defects and handicaps in children could not be confirmed. The implications of these findings have been considered in more detail elsewhere in this protocol. 3.16 Rationale for Using MEDICHECK jte a1th Screening Centre Comprehensive medical evaluation of several thousand men requires complex logistical arrangements. The MEDICHECK Centre in Sydney has extensive experience in processing large numbers of people through a medical evaluation procedure. As the majority of subjects will be living in or near Sydney, �compliance will he maximized by having the examination site in Sydney. MF.DICHECK is convenient, as it is located in the centre of Sydney, close to public transport. Detailed evaluation of the comparative economics of MEDICHECK versus AVHS established centres has not as yet been performed. However, initial indications are that MEDICHECK compares favourably with Pilot Study costs. Apart from economics, there are other advantages of using MEDICHECK: 1. The use of MEDICHECK facilities would significantly reduce the time required to commence the medical examination of subjects and would therefore bring forward the date of reporting of the morbidity survey. 2. The experience and skill of currently employed staff ensures maximum •• - efficiency in subject processing, even at the commencement of examinations. 3. MEDICHECK uses a computer-guided VDU-type questionnaire which is acceptable to clients, does not permit invalid or inconsistent answers, and does not require staff to administer. (It is constrained by permitting only yes/no answers and will therefore be supplemented by a pencil and paper instrument when more complex responses are required). 4. Facilities and staff are in place for blood pressure measurement, electrocardiogram, chest x-ray, pulmonary function, audiogram, anthropometric and biochemical evaluation. The Centre operates it own biochemistry and microbiology laboratories, and participates in a standards programme supervised by the College of Pathologists of Australasia. 33 �/ 5. The entry of all data onto computer files is the routine method of data handling, and will increase the efficiency of the morbidity study. While the use of MEDICHECK is indicated for the above reasons, certain additions and modifications to the usual programme will be made to ensure quality control of critical items {eg blood presure measurement) -and-to ensure that additional items of data of particular interest to the morbidity study are collected. These are discussed below. 3.17 Ra t_ionale for Method of Measuring Morbidity Additions to the usual MEDICHECK programme are required to meet the requirements of the morbidity study. The data to be collected by the usual MEDICHECK evaluation and those to be collected by additional sections of the examination are listed below, along with their rationale (see section 3.14). 3.17.1 VDU Questionnaire. Item Rationale Marital status Veterans, war service Job satisfaction block War service Financial status/problems Veterans, war service Sleep/worry Veterans, war service Depression and state of mind Veterans, war service, TCDD, 2,4-D, 2,4,5-T Drinking habits War service Smoking habits War service Exercise War service Tablets War service Bereavements and family history Confoundin- variable Coronary symptoms War service, TCDD 34 �Leg pain symptoms War service, TCDD Hoart beat/hypertension War service, TCDD, 2,4-D Breathlessness, numbness, varicose War service, TCDD, 2,4-D, 2,4,5-T veins Ankle oedema symptoms War service, TCDD Lung problems and diseases War service, TCDD Abdominal diseases and history Veterans, Vietnam service, TCDD, 2,4-D, 2,4,5-T Veterans, war service, TCDD, Sexual problems 2,4,5-T VD and U/G history, infections, War service, TCDD operations Joint & muscle pains/arthritis War service, 2,4-D Neurological symptoms Veterans, war service, TCDD, 2,4-D, 2,4,5-T Skin disease and allergies Veterans, war service Vision and eye problems TCDD, 2,4,5-T Ear, nose and throat problems 2,4-D Tropical diseases Veterans, Vietnam service Infections and miscellaneous Veterans diseases 3.17.2 Components of the AVHS Questionnaire Self Report of Current Conditions and Symptoms At reception, all subjects will be asked to list all current medical conditions and symptoms, to indicate a grading of severity (from 1-trivial to 5-incapacitating) and to give a duration for each complaint. This data is being sought in addition to questionnaire data for several reasons: 35 �to identify intercurrent and trivial illness which may affect pathology tests (e.g. white cell count) to permit an unprompted description of the subject's problems, which may reveal unsuspected or unconventional symptomatology to improve the accuracy of clinical judgements made by the doctor about the subject's health. Pencil and Paper Questionnaire Item Rationale (see section 3.14) Occupation, employment Veterans Education Veterans Social and behavioural functioning Veterans, war service Marital and offspring history Veterans, war service Wife/partner identification Veterans, war service Social desirability questions Confounding variable Reasons for Vietnam/non-Vietnam Confounding variable service Combat experience Explanatory variable Combat injuries Explanatory variable Alcohol and tobacco consumption War service - diary of past week Other drug consumption, including tea and coffee War service Medical and hospital consultations Confirmation of reports - reasons, when, duration, name and address. Confounding variable Herbicide exposure in Australia 36 �3.17.3 Neuropriychjatric Screening Test Rationale Interview: Veterans, war service present state examination post-traumatic stress disorder interpersonal relationships psychological well-being AVHS schedule of life events: Eysenck Personality Inventory Army Self Description Inventory Symbol-Digit Modalities Test Supra-Span Digit Learning Test Trail Making Test Nelson Adult Reading Test Army Speed and Accuracy Test. 3.17.4 Pathology and Other Tests Rationale Test Electrocardiogram War servicer TCDD Chest x-ray Vietnam service: tuberculosis Spirometry War Service Hearing War service Anthropometry War service 37 �Rationale Test Blood: Glucose, lipids TCDD, 2,4-D, 2,4,5-T Uric acid Alcohol related, 2,4-D Liver enzymes Veterans, Vietnam service, TCDD 2,4-D Haematology Alcohol, Vietnam service Hepatitis B serology Vietnam service Strongyloides, melioidosis, Vietnam service and syphilis serology Drug screen War service Urine 3.17.5 TCDD, 2,4,5-T Physical Examination Item Rationale Skin Veterans, Vietnam service TCDD Hepatosplenomegaly War service, TCDD, 2,4-D Neurological screen Veterans, war service, TCDD, 2,4-D, 2,4,5-T Thoracic auscultation War service, TCDD, 2f4-D Legs - vascular, reflexes, sensation Veterans, war service, TCDD, 2,4-D Blood pressure War service, TCDD, 2,4-D Auditory canals and tympanic Confounding variable in membranes hearing testing In addition, the examining doctor will have the questionnaire responses for each subject, and will elicit additional historical details and additional examination findings as indicated by the 38 �history. He will then record all physical signs found and make clinical judgements as to the presence or absence of particular conditions (see section 3.14), and record other diagnoses suggested by the data. Later, when test results are available he will have the opportunity to modify or add to his clinical judgements. 3.17.6 Hicrarchica1 Struetun; One component of the on-site medical examination will be conducted on a sample of subjects only. This is the phase 2 neuropsychiatric evaluation. This area is of particular importance to veterans and relates directly to the putative effects of both war service and herbicide exposure. It is therefore desirable that detailed evaluation be carried out in this area. As this will require 1 hour of additional testing per subject, it is not feasible to evaluate all subjects without reducing the total number of subjects studied. A random sample of 10% of all subjects plus subjects with high scores on the neuropsychiatric screen will undergo additional testing. This allows for accurate diagnosis of those with suspicious scores on the screening test, and the evaluation of the random sample of non-high scorers will enable inferences to be made about the prevalence of psychopathology in the total sample. 3.17.7 Referral Policy 3.17.7.1 Referrals Indicated for Medical Reasons Any subject with illness requiring treatment or urgent investigation will be referred back to his local doctor, with a brief note from the MEDICHECK physician explaining the problem and asking for follow-up information. Subsequent investigation and treatment would be arranged by the local doctor. 39 �With the prior approval of each subject, a summary of the complete evaluation will be sent to his local doctor. 3.]7.7.2 Pp_tigns__fgr Referrals Indicated for Research Reasons For some subjects, although there may be no indication for treatment or urgent investigation, there may be symptoms and/or signs which cannot be explained without specialist consultation or referral. There are two alternative methods for obtaining the opinions of nodical specialists in regard to subjects who are thought by the examining doctor to be suffering from a medical condition that the doctor is not able to diagnose accurately. The first alternative (option A) is to confine specialist examination to the 3 areas of particular concern to veterans and of significance in relation to herbicide exposure. gastroenterology and neurology. These are dermatology, For this option, if the examining doctor is of the opinion that the subject has a condition in any of the 3 areas of interest which the doctor is not able to diagnose accurately the subject will be referred to co-operative specialists for full clinical evaluation at AVHS expense. The results of the clinical evaluations of all subjects referred would then be available at the time of data analysis. The second alternative (option B) is to perform the standard medical examination on all subjects prior to referral of any subjects, and at the stage of analysis, if the data suggested that veterans were suffering from particular forms of disability more frequently than controls, selected veterans and controls would be referred to the 40 �appropriate specialists. This appioach has the advantage that there is no prejudgement of areas requiring specialist assistance, thereby throwing the net wider to catch unexpected areas of morbidity; specialist evaluation (and consequent expenditure) is confined to those areas in particular need of investigation; the number of subjects to be investigated is much more readily controlled-than vould be the case with a comparatively open-ended referral system; referral is confined to those subjects that, on full consideration of all data, have inexplicable disabilities. Difficulties are that in the interval between initial examination and subsequent referral some subjects will have changed addresses, and therefore all subjects will be required to notify the Study of all changes of address; subject motivation may have waned, resulting in lower compliance rates. Option B will also delay the completion of the study. In both option A and option B specialist reports would be made available to the Study. 3.17.8 Questionnaire jor Wife/ Female Partner Considerable anxiety has been experience by veterans in relation to decreased fertility and abnormal reproductive outcomes. The separate Case-Control Study of Congenital Anomalies has addressed part of this area, but not infertility, death or disability of offspring, or miscarriages. These will be investigated using a structured telephone interview with the current and previous partners. 41 Attention will be confined to �women who have cohabited with the subject for at least 12 months (or who became pregnant while cohabiting for a lesser period). If contact cannot be made by telephone, a home visit will be made to all wives who live in a capital city, otherwise a postal questionnaire will be sent, to be completed at home. There is no requirement for the partner or children to be present at the examination site, as all abnormal pregnancy outcomes will be verified through medical records. For any treatment received in relation to these outcomes, the year of treatment and name and address of the doctor or hospital will be obtained, to allow verification of the data supplied. For all children biologically fathered by the subject the following data will be sought (from the most recent child, back in time): Sex and birthdates *» Difficulties with any pregnancies Disabilities or death of any of the children. Additional data will also be obtained: Miscarriages leading to curettage in hospital Mothers date of birth. 3.18 Physician Assessment Based on a retrospective assessment of Pilot Study data, two AVHS physicians made judgments about the presence and absence of a number of target conditions, taking into account the symptoms, physical signs and results of special tests. This work has also been incorporated in the Pilot Study Report. The following recommendations can be made about the need for physician assessments in the proposed morbidity study: 42 �(i) An assessment by a physician has high face validity, provided that it is based on a contextual analysis of symptoms, signs and results of special tests. Therefore, in the proposed study each subject should have a brief interview (15 minutes) with a physician, working in the MEDICHECK environment, who will: o assess the results of the MEDICHECK and pencil and paper questionnaires o ask additional direct questions o carry out a physical examination o record judgments about: the quality of the history the presence of designated signs the presence of designated conditions - o other morbidity re-assess his judgments when the results of special tests have been made available. (ii) No case can be made for separating the assessment of the history from the assessment of signs and special tests. (iii) To minimise subjective bias in physician judgments, objective indices should be sought, wherever possible, to support the judgments based on the assessments of symptoms and signs. • For example, to precisely document the prevalence of peripheral neuritis in veterans and controls, a "council of perfection" would be to recommend that clinical examination and nerve conduction studies by • a trained neurologist be carried out on (a) all subjects with symptoms and signs suggestive of peripheral neuritis and (b) a random selection (e.g. 1 in 20) of subjects. 43 �However/ such an option is probably precluded by considerations of cost and acceptability to subjects. The lessei option, i.e. of referral on the basis of suspected signs alone, would be less informative because of the subjective nature of neurological signs when elicited by a physician who is not a practised neurologist. (iv) Ideally, the physicians who carry out the 15 minute interviews and assessments at KEDICHECK should have post-graduate training as physicians (i.e. FRACP qualifications or MRCP) or general practitioners (FRAOGP), and they should be specially selected for the purpose of this study. (v) The exact protocols and proformas for the recording of physicians judgments will be finalised after consultation between MEDICHECK physicians and AVHS physicians. 3.19 Rationale For Number of Subjects The number of subjects to be examined is influenced by several constraints: o The need to complete examinations in 12-18 months to allow the submission of a report in an appropriate time frame, o The total number of subjects available that fulfil the criteria outlined above, o The need for the study to have sufficient power to detect moderate relative risks for conditions of interest possibly associated with Vietnam service. o The need to have sufficient numbers of veterans to test the hypothesis that variables such as combat exposure, herbicide exposure and corps, are predictive of morbidity -within the Vietnam cohort. 44 �Approximately 120 subjects could be examined per week, for a total of 5,760 in a 48 week period. Taking into account delays introduced by not being able to locate subjects at known addresses, and initial failure of some subjects to keep examination appointments, a more realistic total in one year is probably 5,000 subjects. As approximately 6,000 veterans and 9,000 controls enlisted in N.S.W., total numbers available are adequate. For 5,000 subjects, maximum study power is achieved when the numbers of veterans and controls are equal (2,500 of each). Increasing the number of veterans studied to 3,000 (to increase power in relation to explanatory variables within the veteran group) and reducing the number of controls to 2,000 reduces the power of veteran/control comparisons by 4%, a minimal loss. The effect on study power of further increasing the number of veterans is shown in Table 3.4. Table 3.4 Effect of Veteran/Control Ratio on Power. Veterans Controls 2,500 2,500 0% 3,000 2,000 4% 3,500 1,500 16% 4,000 1,000 36% Loss of efficiency Therefore, it is proposed to examine 3,000 veterans and 2,000 controls. 45 �Foi selected conditions of particular interest, the minimum relative risk that would be detected (80% power, P(1) = 0.05) is shown in Table 3.5. The expected prevalence rates are mostly derived from complaints of controls in the Pilot Study. A more detailed table is given in Appendix 1. Table 3.5 Minimum Relative Risks Detectable as Statistically Significant with 3,000 Veterans and 2,000 Controls Condition Estimated prevalence Anxiety Sleep difficulties Depression Temper outbursts Hypertension Numbness and tingling Dizziness Loss of strength Severe acne Burning /itching of skin Persistent rash Minimum Relative Risk 10% 16% 10% 20% 1.2 1.1-1.2 1.2 1.1-1.2 5% 5% 7% 5% 1.3-1.4 1.3-1.4 1.2-1.3 1.3-1.4 3% 13% 13% 1.3-1.4 1.1-1.2 1.1-1.2 Up to 100 deaths are anticipated to have occurred since discharge (18 months to 2 years post enlistment) in this cohort of 7,000. These will be detected via the mortality study. A minimum detectable relative risk of 1.4 for the common conditions of interest indicates that 5,000 is a satisfactory number of subjects. Rare * conditions such as melioidosis, strongyloidiasis and syphilis will, if significant causes of morbidity among veterans, have much larger relative risks than 1.4. Lack of adequate general population prevalence rates makes power estimation impossible. Rare manifestations of common exposures such as 46 �alcohol (eg Korsakow's psychosis) are not of relevance here, as there are much more common manifest ions that will allow evaluation of veteran/control differences arising from any differences in patterns of alcohol use. Any infrequent but distinctive effect of Vietnam service or herbicide exposure might be provisionally identified on clinical grounds even if its frequency was not statistically increased in veterans (e.g. 6 cases in veterans, none in controls). For instance, there would be only 2 or 3 cases of a condition with the incidence of multiple sclerosis in this group of 5,000 men. This study will not specifically investigate cancer incidence for the following reasons: i. The study would have very low power to detect veteran/control *• differences due to the infrequency of cancer in this young age group, ii. The mortality associated with cancer would reduce the number of subjects giving a past history of cancer, iii. Subjects with cancer may not be sufficiently well or motivated to present for medical examination, iv. The long latent period for developing cancer. 3.20 Index of Herbicide Exposure 3.20.1 Objective Determination of Herbicide Exposure Determining individual herbidide exposure in Vietnam from spraying mission and troop movement data is considered in other reports (Adams et al., 1981). 3.20.2 Subjective Reports of Herbicide Exposure The relationship between subjects reporting of herbicide exposure and morbidity will be difficult to interprete for several reasons: 47 �The events occured 12 to 17 years ago and therefore recollection is likely to be inaccurate. At the time the spraying was taking place the soldiers did not know what chemicals were being sprayed. It may be anticipated that as a result of the publicity surrounding the issue and the desire for compensation those subjects with morbidity will be more likely to recall being sprayed independent of whether they were actually sprayed. It may be possible to analyse this latter problem, one of recall bias, by determining the relationship between current morbidity, reported exposure to herbicide and the likelihood of herbicide exposure determined by the HOPPS programme. In spite of this potential bias it is important to collect and assess the utility of subjective reports » of herbicide exposure. 3.20.3 Quantity of Herbicide Sprayed Each Year The volume of herbicide sprayed in the RANCH HAND programme in Phuoc Tuy in each year of the Vietnam War is known (Table 3.6). Table 3.6 Annual Herbicide Usage in Phuoc Tuy Province by Agent According to HERBS Tapes 1965-1971 Amount of RANCH HAND herbicide sprayed in Phuoc Tuy (in thousands of litres) Year Agent Orange 1965 1966 1967 1968 1969+ Total White 90 200 490 240 140 570 "~ • * 1020 Blue ^^ ^ - 50 • 710 48 50 Total 90 200 680 810 " 1780 �Thus the chronologic year of service in Vietnam is useful as an explanatory variable, although herbicide usage and level of combat are probably highly correlated. 3.20.4 Exposure to Herbicide in kustralia If herbicide exposure does have a measurable effect on morbidity, the relationship between herbicide exposure in Vietnam and current morbidity is potentially confounded by herbicide exposure in Australia. Therefore data will be sought from subjects on occupational exposure to herbicides. 3.21 Index of Combat Exposure American literature suggests increased prevalence of war-related disorders amongst those troops closest to the 'front-line1 of combat (Penk et al., 1981). In analysing relationships within the veteran group it is therefore important to assess the degree of combat or danger to which veterans were subjected. This will be done in several ways: 1. Chronologic Year of Vietnam Service. The risk of death or wounding as a function of year of Vietnam service will be determined and used as an explanatory variable. 2. Corps Subject's Corps will be used as an explanatory variable, as some Army Corps (Engineers, Infantry, Armoured, Artillery, Signals) engaged in contact with the enemy, and other did not. 3. Subjects Injured in Combat The Central Medical Record of each subject will be reviewed 49 �and the presence of any combat injuries recorded, along with their nature, cause and duration of hospital treatment (see section 3.22.2). 4. Casualty Rate of Units (Combat Index) With "Casualty Reports' (the completeness of which is not yet known) and Vietnam Unit postings data, the incidence of combat wounds and deaths will be compiled for all Units, if feasible. With data on each subject's Vietnam postings from his CARD dossier (see section 3.22.1) the risk of combat injury and death will be calculated for each subject. 5. Subjective Combat Exposure Figley (1980) has developed a questionnaire which quantifies combat exposure, which he has validated on a small sample. Only minor modification is required to make it applicable to Australian Vietnam veterans. Only veterans will be asked to complete this instrument, as it is not applicable to controls. 3.22 Army Dossier Data Held on Each Subject The personal (CARD) dossier, Central Medical Record and psychology record cards together contain a wealth of information about almost every aspect of a serviceman's period of service. This material is potentially available, is not subject to recall bias, and constitutes a valuable source of baseline data for each subject, enabling comparisons of veterans and controls at the time of enlistment and during the period of service. are outlined below. 50 More specific uses of these data �For all 5,000 subjects it is proposed to extract certain data from these records for UKO in the analysis and interpretation of findings from the medical examinations and to compare the enlistment and Army service characteristics of subjects who present for examination with those who do not. The extraction of these data therefore need not precede the medical examinations, but will be concurrent. 3.22.1 CARP Dossier Data to be extracted: Verification of veteran/control status Verification of Vietnam postings for veterans. Disciplinary proceedings Promotions Volunteer/bal lotee The CARD dossiers of all current and former members of the army are held in the Central Army Records Office (CARD) in Melbourne, except for those of serving Officers (Military Secretary's Office, Canberra). These data will be used to derive each subject's risk of combat injury or death (Sec. 3.21), and to see if ability in the army or disciplinary problems are related to current morbidity and social and psychological functioning. 3.22.2 Central Medical Record; Data to be extracted: 1. From enlistment medical examination: Weight, height Significant past medical history PULHEEMS rating (Army fitness ratings, see Appendix 2). 51 �2. Hospital admissions: Type of disease Duration of admission Residual disability 3. From discharge medical examination: PULHEEMS Theso data will be used to determine to what extent current morbidity existed at the time of army service, or has arisen subsequently. 3.22.3 Psychology Record Card Three types of data will be extracted from the psychology record card: i. Enlistment psychology test results. The comparison of these test results between veterans and • controls, and with scores obtained when retesting subjects 10 to 15 years later, will permit valuable comparisons of veteran and control subjects both prior to and after army service. The relationship between test scores and subsequent morbidity is also of interest, ii. Corps preferences expressed by recruits. The preferences for corps reflects the desire of a member for combat (e.g. prefers Armoured, Infantry, etc) or non-combat (e.g. Ordnance, Catering, Band, Transport) roles, and are therefore potentially valuable explanatory variables in analysing the relationship between desire for Vietnam service and subsequent morbidity. The completeness of these data is adequate, although not 100%. 52 �iii. Referral for Psychological Opinion. The referral of a soldier for psychological evaluation may bear significantly on current morbidity and therefore such data will be obtained. Data from all three types of records will be transferred directly onto a data entry sheet for ease of computer entry and analysis. 53 �STUDY PROTOCOL 4. The outline of the design of the Morbidity Study is shown in Figure 1. Figure 1. Overall Plan of the Morbidity Study Development and testing of examination procedures (Section 4.9) Subject selection (Section 4.1) I I Subject location and invitation (Figures 2 and 3) V Search Army records (Section 5) Subject examination (Figure 4) reception questionnaires testing physical examination detailed testing of a sample Record data Specialist r e f e r r a l Option A Section 4.5.11 Partner/wife location (Figure 5) Verification of medical data (Section 4.10) I J \ Data entry (Section 6.1) Specialist r e f e r r a l Option B Data analysis (Section 6.2-6.5) Report findings 54 Partner interview (Section 4.6) �4.1 Subjcct Selection 4.1.1 Definition of Veteran Status A Vietnam veteran is a national serviceman who served at least 13 weeks who went to Vietnam within two years of enlistment during the period of the war, irrespective of the duration of Vietnam service, and who was discharged alive or survived for two years after enlistment. 4.1.2 Definition of Control Status A 'control' subject is a national serviceman who served at least 13 weeks, who did not go to Vietnam but was drafted in an intake from which national servicemen were sent to Vietnam (June 1965 to February 1971) , and who was discharged alive or survived for two years after enlistment. 4.1.3 Process of Subject Selection Former Vietnam war period national servicemen only; N.S.W enlistees only; Date of original enlistment February 1971 or before; Served at least 13 weeks; Discharged alive or survived for 2 years post enlistment, whichever occured first; Randomly select a uniform proportion of veterans from all intakes to yield 3,000 veterans; Randomly select 2 controls for every 3 veterans from each intake (to yield 2,000 controls); Randomly allocate veterans and controls to 10 blocks, each with 300 veterans and 200 controls; �Blocks of subjects are invited to participate sequentially until 5,000 subjects have been examined. For the subjects within each block selected, full measures will be taken to maximize the participation rate (see section 4.3). 4.2 Obtaining Current Address Current addresses will be sought only for the 5,000 subjects selected. 1. Current serving member of the army. Address from CARO and serving unit. If not a current serving member, go to stage 2. 2. Subjects names matched with computerised registers: Current Electoral Register, and N.S.W drivers licence register, and S.A. drivers licence register, and N.T. drivers licence register, and Tas drivers licence register, and (possibly) W.A. drivers licence register. If a name matches and an address(es) is found, it is stored on the study subject file. If no match occurs go to stage 3. 3. Subject whose names are unmatched after stage 2: Names matched with manual registers: QLD drivers licence register, and Vic drivers licence register, and Immigration and Ethnic Affairs 'arrivals and departures' microfiche. If a name matches and an address is found it is stored on the study subject file. Note; Matching in stages 1,2, and 3 has been performed in the Mortality Study. 56 �4. Subjects whose names are unmatched after 3: Names matched with: Commercial credit bureaux and Criminal Records Bureau (if available), and Department of social Security files (if available). If a name matches and an address is found it is stored on the study subject file. 5. If an address if found, yet subsequent tracing fails to locate the subject, the name will be matched with those registers in stages 2,3 and 4 not yet used for that name. This stage can only occur after an attempt at initial contact has been made. 6. Once an address has been found, telephone books and Telecom will be consulted to obtain a home telephone number if required. » See Figure 2 for plan for obtaining subject addresses. 57 �Fiaurc 2 Name_ & D.O.B. "t-'~ f ' • Not Currently Serving in Army -. ^ Currently Serving in Army^^ I Address exact match Computer Search of Computerised Matching Sources: \ ^ A D D R E S S Elect. Roll/Mot. Req. I no match 1 f Manual Search of r>mnmi4-ni- { r.~.A -r*A computer i sea ana Manual Motor Registries F I L E I noTO£i4"f'V'i Ilw Ilia Ui^ii SEARCH " Immigration & ^^ Ethnic Affairs j n ; arrived: address available not listed Army Record: Parents/N.O.K. ^ parent(s) ^ ^ N.O.K. Address ^ 1 UrM-4- al 'i t-\f List 1 SEARCH alive Reaistrar(s) General ^ name change ^ Initiate Search ™" using New Name »o name change 1 Other: V.A. n SEE FURTHER k PROCEDURES PROTOCOL Cr\r* i A 1 Qof^ DOC laX DCt, . Corr. Serv. Def. Forces H. Corp 58 �4.3 Methods of Obtaining Compliance for Medical Examination 4.3.1 Contact Procedures For each address obtained, a sequence of contacts will be made until an appointment is made for the subject to attend for examination (see Figure 3). A record will be kept of all attempts made to contact each subject. For subjects living in the country, telephone calls will be substituted for personal visits. Stage 1. Sequence until contact made: 1st mailing - introductory letter followed by an appointment card. 2nd mailing - letter plus appointment card. Telephoning (if possible) Personal visit to enlist participation. If no contact has been made with the subject go to stage 2A. If contact has been made with the subject but he has not presented for medical examination, go to stage 2B. Stage 2A Seek new address, firstly by visiting or telephoning the most recent address found on register searching, and then if no contact is made, seek another address on the registers (see section 4.2, stage 5). Stage 2B Contact subject again by telephone and make another appointment. If 2nd appointment not kept, or no telephone contact possible, make personal visit to arrange appointment if the subject is a capital city resident. If the subject still fails to attend, see section 59 4.7. �Figure 3 1 H3*? Subject Contact Procedures KNOWN ADDRESS Mail 1 1 Contact Participation Contact Participation Contact Participation no contact 1 no contact 1 I 1 Home Visit no contact Seek New Address 60 �4.3.2 F_i ni a nc_i_a 1. ..Compe n ca t_i on Full economy class return fares for travel between home and the examination site by taxi, country train or aeroplane will be paid to all subjects. Full compensation for lost earnings will be provided to the subject or his employer upon receipt of appropriate documentary evidence. 4.3.3 A_dd_i_ti_ona_l Means of Maximizing Compliance A covering letter signed by the Minister for Veterans' Affairs will be sent to all subjects, mentioning that the investigation is supported by the leaders of the major political parties, the Returned Servicemen's League and the Vietnam Veterans Association of Australia, if such endorsement can be obtained. Advertisements and feature articles and programmes in the public media would be • valuable. 4.4. Outline of Design of Medical Examination All subjects presenting for examination will undergo:- V.D.U. questionnaire Pencil and paper questionnaire Neuropsychology and psychiatric screening tests Blood and urine testing Electrocardiogram, lung function test, chest x-ray, hearing and vision test, height and weight measurement Physical examination and evaluation by a doctor See Figure 4 for plan for medical examination. 61 �Figure 4 Plan For Medical Examination Reception identification list of current complaints Blood and urine specimen collection NEDICHECK Questionnaire (administered using Visual Display Unit) Pencil and paper questionnaire I Neuropsychiatric screening Additional tests: ECG, CXR, spirometry, ht and wt, audiogram, visual acuity 1 t I Checking of incomplete or inconsistent responses Physician examination and assessment Neuropsychiatric evaluation Discharge 62 �High scorers on the neuropsychiatric screen plus a 10% random sample will undergo detailed neuropsychiatric testing. Under option A subjects with suspected undiagnosed gastroenterologic, dermatologic or neurologic conditions will be referred to appropriate co-operative specialists for full clinical evaluation. Under option B, any referrals for specialist opinion for research purposes will depend upon the results of initial analyses of the morbidity profiles of veterans and controls (see 4.5.11). The current and previous wives/female partners will be interviewed by telephone, and failing that will be interviewed at home or sent a pencil and paper questionnaire, to be returned by mail. Subjects who are located but do not present for examination will be asked to complete as much of the questionnaire (all pencil and paper) as possible, and give a blood and urine sample. Subjects will be referred to their local doctors for management, where indicated on clinical grounds. 4.5 Design of Examination Procedure for Subjects MEDICHECK 4.5.1 component Recept ion Name taken and identified on subject list. yes Given written request to answer all questions as honestly and completely as possible. no Given form to describe (unprompted) all current complaints, with severity and duration. 63 no �MED1CHECK component 4.5.2 Blood and Urine Samples taken yes Blood tests: Biochemistry Glucose, lipids, electrolytes, urea, creatinine yes Liver enzymes, serum protein yes Drug screen* - benzodiazepines, salicylate no Alcohol* no Microbiology VDRL Screen* no Melioidosis and strongyloides serology* no Hepatitis B serology*: no 1. Core antibody - indicator of past infection. 2. Surface antigen if core antibody (+) - indicator of current infection. 3. Be antigen if surface antigen (+) - indicator of infectiousness. 4. Surface antibody if surface antigen (-) - indicator of resolved infection. Contractual arrangements with additional laboratories will be required for these tests. 64 �MEDICHECK component Harmatology Hb, MCV, MCHC yes WCC, ESR yes Platelet count ' " yes Urine tests Culture yes 'Dipstix1 chemical testing - Hb, protein, glucose, bilirubin, pH yes Additional serum for storage 4.5.3 no MEDICHECK Questionnaire yes Questions in Yes or No format, will cover the following areas: Marital status Job satisfaction block Financial status/problems Sleep/worry Depression and state of mind Drinking habits Smoking habits Exercise Tablets Bereavements and family history Coronary symptoms Leg pain symptoms Heart beat/hypertension Breathlessness, numbness, varicose veins 65 �MEDICHECK component Ankle oedema symptoms Lung problems and diseases Abdominal diseases and history Sexual problems VD end U/G history, infections, operations Joint & muscle pains/arthritis Neurological symptoms Skin disease and allergies Vision and eye problems Ear, nose and throat problems Tropical diseases Infections and miscellaneous diseases » The questionnaire will take up to 40 minutes to complete. 4.5.4 AVHS Questionnaire no Pencil and Paper Questionnaire: Occupation, employment Education Marital and offspring history Wife/partner identification Reasons for Vietnam/non-Vietnam service Combat experience Combat injuries Alcohol and tobacco consumption - diary Chemical exposure in Vietnam Exposure to noxious substances 66 �MEDICHECK component Other drug consumption including tea and coffee Medical and hospital consultations - reasons, when, duration, name and address. Medical record release Herbicide exposure in Australia 4.5.5 Keuro-Psychiatric Screen Interview by clinical psychologist, assessing psychiatric no pathology using the Present State Examination (PSE) , assessing post traumatic stress disorder, interpersonal relationships, and psychological well-being. Tests of psychological functioning, given to all subjects: AVHS schedule of life events Eysenck Personality Inventory Army Self Description Inventory Symbol-Digit Modalities Test Supra-Span Digit Learning Test Trail Making Test Nelson Adult Reading Test Army Speed and Accuracy Test. 4.5.6 Additional Tests Electrocardiogram yes Chest x-ray yes Spirometry yes Audiogram yes Visual acuity yes Height and weight yes 67 �While these tests are being performed, the AVHS questionnaire and ncuropsychiatric screen will be checked for completeness and consistency of answers, and the neuropsychiatric tests will be scored, to determine whether the subject needs further psychiatric or neuropsychological testing as a 'high scorer1. If incomplete or inconsistent answers are detected, the interviewer will ask the subject to correct his answers at the end of the 'Additional Tests' section. 4.5.7 Doctors Examination and Evaluation The examining doctor will have the results of the MEDICHECK questionnaire and the AVHS questionnaire, and will seek further information from the subject to guide his/her clinical formulation of the subject's morbidity. Physical examination will be directed toward arriving at a conclusion about problems suggested by the questionnaires and the history. In addition, examination will be directed toward detecting abnormalities in the following areas: Skin Hepatosplenomegaly Neurological screening examination Thoracic auscultation Legs ~ vascular, reflexes, sensation Blood pressure Auditory canals and tympanic membranes (to facilitate interpretation of audiogram) 68 �Blood pressure measurement and ear examination could be performed by technicians. The doctor will then record all physical signs found and his clinical impression in the form of differential diagnoses, with confidence ratings. Three days later, when the results of pathology tests are available, he will have an opportunity to reassess and modify or add to his list of possible conditions. 4.5.8 Detailed ^euro-psychiatric Assessment. Subjects who are "high scorers' on the neuropsychiatric screen plus a 10% random sample will undergo detailed neuropsychiatric assessment consisting of both pencil and paper tests: Hostility Questionnaire Spielberger Anxiety Scale Depression Questionnaire. 4.5.9 Specialist Referral Two options have been identified with respect to specialist referral: Option A; If, in the opinion of the examining doctor, a neurological or gastroenterological condition is suspected, he will be referred to specialists who have agreed to co-operate with AVHS. The specialist will carry out a full evaluation as warranted by the subject's symptoms and signs. If, in the opinion of the examining doctor, a subject has an unusual skin condition, or if any doubt exists about the diagnosis of a skin 69 �condition, the subject will be referred to a dermatologist for consultation. Option^ B; At the stage of data analysis the frequencies of disabilities will be compared for veterans and controls. If, after adjusting for the relevant confounding variables (e.g. alcohol consumption in the case of liver disease), an apparent excess in veterans is evident, the affected veterans and controls will be identified on the subject file. Within this group, samples of veterans and controls with no obvious clinical explanation for their examination findings will be referred to the appropriate specialist for full clinical evaluation. The arrangement between AVHS and specialists will provide for a full report to be sent to AVHS with regard to every subject referred. 4.5.10 Local Doctor Referral If, in the opinion of the examining doctor a subject has a condition requiring treatment he will be advised to consult his local doctor immediately. With the prior approval of each subject, his medical report (including specialist reports, if applicable) will be sent to his local doctor. 4.6 Information From Wife/Female Partner 4.6.1 Definition A partner will be any woman who has cohabited with the subject for at least 12 months, or who became pregnant while cohabiting for a lesser period. Thus each subject could have more than one partner. 70 �4.6.2 Ver_if_ication of Address All subjects will be asked to provide the full names, dates of birthf telephone numbers and current addresses of their partners, if known. For partners where the address is unknown, it will be sought on the Electoral Register microfiche, and then telephone number sought in telephone directories (see Figure 5). 71 " �4.6.3 Figure 5 J Plan for Partner Follow-up KNOWN ADDRESS Telephone Participation Contact J no contact 1 J Participation Mail no contact Home Visit.. Seek Participation Contact 1 no contact 1 Seek New Address 72 �4.6.4 Questionnaire A structured telephone interview will seek the following information, for the most recent pregnancy and then for any other pregnancies going back in time: Mothers date of birth. Outcome of pregnancy Difficulties with the pregnancy. (If yes, year and name of treating doctor or hospital). Miscarriages leading to curettage in hospital. (If yes, year and name of treating doctor). For each child biologically fathered by the subject: sex and birthdate Congenital anomalies or death (If yes, year and name of treating doctor) . A medical record release will be sent to participating partners for signature and return by mail. 4.7 Follow-Up of Non-Complying Subjects 4.7.1 Additional Strategies for Obtaining Compliance. If a subject fails to present for examination after 3 appointments have been made, or, if at an earlier stage he expresses a desire not to present for examination, he will be contacted at home and a modified health evaluation performed. This contact will be by telephone if the subject lives in the country or personal visit (initially) if the subject lives in Sydney, or the capital city of another State. 73 �ESllow-up of Non-Complying Subjects Figure 6 Appointmont made Appointment kept End of follow-up ^.Appointment kept End of follow-up Appointment not kept i J 2nd and 3rd appointment made Appointment not kept State capital city resident Able to be contacted by telephone? Country . resident \ No Yes >me visit to .e appointment X \ Telephone interview Appointment kept i Appointment not kept Pencil and paper questionnaire 1 \ Returned? End of follow-up T 1' / 7 / Yes, / / End of End of follow-up follow-up No "< 2nd questionnaire sent I Home visit for interview and sample collection Returned? No End of follow-up Reminder letter 4 End of follow-up �4.7.2 He a 1J h Evaluation Modified for Home Administration. List of current complaints VDU questionnaire in pencil and paper format AVHS questionnaire Neuropsychiatric screen Reasons for reluctance to co-operate If nurse available: Blood samples Urine sample Blood pressure measurement 4.7.3 Hierarchy of Data to be Sought Data will be elicited in a particular order so that if the subject becomes reluctant to continue at any point the most critical data will have been gathered. A suggested sequence is given below but this will be modified in the light of experience with initial home interviews. Order of precedence: Blood pressure measurement List of current complaints Combat exposure (veterans only) Reasons for Vietnam/non-Vietnam service Sleep/worry Depression and state of mind Drinking habits Smoking habits Tablets Marital status 75 �Job satisfaction Medical and hospital consultations Medical record release Neuropsychiatric screening Social adaptation Family coherence Behavioural Education Occupation, employment Neurological symptoms Abdominal diseases and history Skin diseases and allergies The remainder of the MEDICHECK questionnaire The remainder of the AVHS questionnaire Blood samples Urine sample 4.8 Follow-up of Non-Complying Wives/Female Partners 4.8.1 Additional Strategies for Obtaining Compliance Partners not able to be contacted by telephone will be visited at home if residents of a State capital city, and failing that, will be sent a pencil and paper questionnaire. Partners failing to return a questionnaire will be sent another, and then a reminder letter. 4.8.2 Suggested Hierarchy of Data to be Sought Order of precedence: Sex and birthdate of all children Death of children 76 �Miscarriages leading to curettage in hospital Difficulties with pregnancies Mother's date of birth For each occurrence: when name and address of institution or treating doctor Medical record release 4.8.3 Data Not Able to be Obtained All data not obtained due to poor compliance will be coded to indicate this. 4.9 Pilot Testing of Medical Examination The components of the routine MEDICHECK evaluation that are being retained will not require specific pilot testing. The AVHS questionnaires for subjects and wives/partners will be tested on a small sample of men and women (not study subjects) prior to the commencement of the medical examinations to detect and overcome any ambiguities or problems which might arise during questionnaire administration. The sources of men and women for pilot testing have not yet been determined. The physician assessment will be pilot tested by MEDICHECK doctors prior to commencement of AVHS medical examinations, and the comments provided by the doctors used to modify both the physical examination protocol and the documents used to record clinical findings and judgments. Following this, it is intended to operate the examination centre for a 1 or 2 week period initially, to determine how long the various components of the examination take, how many subjects will keep a first appointment and to 77 �discover any logistic or organizational problems. This examination period will be followed by a 1 or 2 week break, while solutions are found to the problems revealed by the initial run. Following this, the examination centre will commence full scale operation. 4.10 Verifying Data in Medical Records All medical data volunteered by subjects or partners that is accompanied by the name of a treating doctor or hospital and signed release form will be verified. A letter giving the reported complaint and time of occurence will be sent to the relevant doctor or hospital with a request for substantiation of the condition and for the provision of more accurate or clarifying information. The verified data will not be used to replace data volunteered by the subject* but will be retained for complementary analyses. 78 �5. ADDITIONAL DATA FRQM_ARMY ^SOURCES 5.1 Caro Dossier Verification of: Dates of birth, enlistment, discharge Veteran/control status Corps Civil education Vietnam postings data Obt a in: Australian posting data Special courses completed Disciplinary offences Promotions Volunteer/ballotee 5.2 Central Medical Record Verify: Enlistment and discharge PULHEEMS (Army health rating APPENDIX 2) Obtain: From enlistment medical examination: Weight, height Significant past medical history PULHEEMS Hospital admissions: site (e.g. Vietnam, Australia) 79 �type of pathology (trauma - combat/non combat, sexually transmitted disease, other infection, drug induced, psychiatric, other stress-related,) duration of admission From discharge medical examination: PULHF.EMS 5.3 P_sychology Record Cards Obtain: All test results Corps preferences - 1st, 2nd, 3rd. Referrals - reason (application for course or promotion, disciplinary problem, psychological problem). - date. 80 �6 • 6.1 DATA ASSESSMENT AND ANALYSIS HaJr.£L Acqui si t ion and Vertical ion Baseline information from the MRCC tape has been checked for logical inconsistencies, and where possible, for inconsistences with the CARD tape. As error emerged, they have been checked against original dossiers in CARO. Follow-up information, including last known address will be stored on computer disc file and verified against the original source documents. This information will be used as a basis for a master file which will be used to record the appointment and compliance history of each subject. A parallel file will be developed for the corresponding information on wives and female partners. Combat exposure information will be derived from manual searching of army records, and the reliability of coding, punching and verification will be established during pilot studies and spot checks of the final data set. Verification of veteran/control status is essential. For some individuals on the MRCC tape there were logical inconsistencies - eg "veterans" with periods of service which were far too short for them to have served in Vietnam. This suggested that there are errors on the tape, either in relation to veteran status and/or in relation to duration of army service. All inconsistent records have been checked manually and manual spot checks will be made of other "self-consistent" computer records to assess the accuracy of the remaining information on the MRCC tape. 81 �Furthermore, if at the time of interview there is a conflict between the reports of the subject ("I went to Vietnam") and the computer record ("he is a control"), the details will be returned to CARD for clarification. Final data will be coded according to the correct (CARO) classification. Interview information will be obtained from computerised MEDICHECK records, from AVHS pencil and paper questionnaire, from psychiatric questionnaires, neuropsychology assessments and physician assessments. The reliability of coding, punching and verification will be established during pilot studies and spot checks of the final data set. 6.2 Adequacy and Utility of Morbidity Study Data 6.2.1 Comparability of Follow-up of Veterans and Controls As shown in Table 3.3, there is suggestive evidence that the proportion of NSW veterans found on the Electoral Register (78%) is less than the proportion of controls (81%); furthermore, of those found on the register, the proportion of veterans presently living outside NSW, (12%) was greater than the proportion of contols (9%). These differences may reflect differences between veterans and controls in relation to socio-economic factors, employment and social mobility. It should be noted that the men who will be most difficult to trace (namely those men who are single, divorced, separated, itinerant, unemployed, or alcoholic) are most likely to be at greatest risk of disability. Thus in order to minimise any bias between veteran and control families arising from incomplete follow-up, it will be 82 �important to reduce the number of untraced men in both groups to an absolute minimum. Some idea of the magnitude of bias arising from incomplete follow-up could be obtained by looking to see how the morbidity of each veteran and control subject varies with the amount of follow-up required to trace him (i.e. was his address found in a primary, secondary or other source) . 6.2.2. Complicance Rates for Veteran and Control Subjects Even if the follow-up is adequate for both veteran and control subjects (e.g. 95% or more followed to their most recent address), the results of the morbidity assessments could still be biased if there were a difference between veteran and control rates of compliance with the interview. The most likely bias would be for veteran subjects with a disability to be more compliant than control subjects with a disability. After maximising the compliance rate, the magnitude of any residual bias could be assessed by comparing the morbidity profiles of veteran and control subjects according to whether they attended the first, second or third appointment. These data can also be compared with the (incomplete) data obtained from non-compilers at a home interview. If the morbidity patterns do vary according to level of compliance (appointment) then this can be adjusted for (at least in part) during the analysis of the results. 83 �6.2.3 Subjectivity of Self Reported Information A major source of potential "bias" is that self reports from veterans are likely to be influenced by an expectation of disability which is greater than that of non-veterans. This tendency, whether it is conscious or sub-conscious, could be so general that it could lead to quite marked differences between veterans and controls in the prevalence of symptoms related to many different disabilities. The potential for such bias will lead to major problems in the interpretation of all subjective information collected in the course of this morbidity study. Similar problems of interpretation have been encountered in assessing the symptoms of "effort syndrome" and "combat syndrome" in servicemen from previous wars, and in assessing symptoms associated with "compensation neurosis" after injury at work or in road accidents. There are several approaches to this problem of interpretation. It may be appropriate to accept the reality of the symptoms, as such, and to explain them as being consequent on the psychogenic stimulus (of the war). This interpretation is, of course, more plausible if the symptoms can be identified as components of a depressive syndrome, an anxiety reaction or if they can plausibly be identified as somatic equivalents of psychogenic origin. In some circumstances the physical findings may support a functional diagnosis (eg tachycardia, sweating and hyperventilation if otherwise unexplained), and in other circumstances the functional origin of symptoms is supported by their 84 �pjpomorphic or protean nature, and by their failure to fit an organic syndrome. However, the functional or psychogenic origin of symptoms should only be accepted after steps have been taken to exclude an organic or biologic basis for the symptoms. An organic basis can be suggested by the pattern of symptoms: thus if a veteran complains of chronic productive cough, shortness of breath on exertion and give a history of heavy smoking, we would be justified in suspecting the presence of chronic obstructive airways disease. This could be confirmed by physical examination or by objective testing. The presence of dissimulation or malingering might be suspected if there were a constellation of plausible symptoms, together with an absence of supportive objective signs. In some subjects it may be possible to suspect dissimulation if there is a high score on the social desirability scale administered as part of the psychological assessment. However, in general it will be necessary to assume that all symptoms are real, to analyse the contexts in which the symptoms are found, and to look to the epidemiological data, the physical examination and objective tests to provide clues about the physical disabilities which may underlie the symptoms. In the context of this study, the analysis of subjective information is made easier by the fact that we are not necessarily required to make judgments about the physical bases of symptoms in individual veterans; it will suffice to show that in veterans as a group, such 85 �and such a symptom complex is associated with objective signs of disease significantly more often than in similar groups of control subjects. For those symptom complexes which are not supported by objective measures, the symptoms might be provisionally identified as being of functional or psychogenic origin. However, if these symptoms are found to be more frequent in veterans, and specifically if they are correlated with measures of combat exposure or herbicide exposure, then they can plausibly be regarded as real effects of war service. Certainly, if any functional syndrome is associated (see 6.5) with a measurable outcome which is more frequently observed in veterans (e.g. more frequent divorce) then it should probably be counted as one of the real hazards of army service. 6.2.4 Objective Measures Several objective measures are available. For example, lung function testing will provide an objective test for chronic obstructive airways disease, biochemical tests on plasma will provide objective evidence of abnormalities of liver function and help to detect heavy drinkers. Objective tests are also available for the detection of past hepatitis B infection and for the detection of past syphilitic infection. Neuropsychological tests for the detection of brain damage and neurophysiological test of nerve function can also be regarded as objective. The importance of these objective tests is not that they will provide a definitive medical diagnosis in their own right, but rather that they can provide independent support for disabilities which might be 86 �suspected from the pattern of symptoms and from signs reported by the examining physician. As argued in the previous section, although this objective support might not be evident in every case with symptoms, it should be sufficient to show whether, as a group, veterans symptoms are significantly associated with objective measures more often than in similar groups of control subjects. Because of the importance of objective measures, and because of the need to make inferences about the whole population of veterans and controls, we have argued that objective measures should be included either: (i) for all veterans and controls, or (ii) for randomly selected veterans and controls, plus those who have a clinical indication or those who fail a screening test. 6.2.5 Face Validity of Medical Diagnosis Made by a Physician Medical diagnoses made by a physician must be accorded face validity, because the practice of diagnosis is defined in terms of the judgment of the physician. This is not to say that a judgment of a physician is necessarily reliable and objective, but as it is based on a contextual analysis of symptoms, physical signs and the results of special tests (gestalt), it provides an assessment of the meaning of the data which can be obtained in no other way. Nevertheless, because a physician's judgments are subject to error, it is important to consider the possibility that there may be a systematic bias in diagnostic accuracy between veteran and control subjects. Such a bias could invalidate any conclusions which were 87 �based on physician judgments. To minimise this bias in the pilot study, a decision was made to "blind" the physician carrying out the physical examination to the veteran status of each subject and to the results of the medical history (and vice versa). As a result, the judgments made by the physicians were out of context, and they were less helpful than might otherwise have been expected. It would be unwise to separate the assessments of symptoms and physical signs in the proposed morbidity study, although it would be desirable to try to maintain "blinding" of the physician to the veteran status of each subject during the examination. However, regardless of the precautions taken, it is unreasonable to suppose that the physician will always remain ignorant of the veteran status of each subject; accordingly it will be impossible to always exclude physician bias as an explanation for (minor) differencies in the frequency of certain diagnoses in veteran and control groups. This conclusion is not as gloomy as it sounds, in that it will be possible to test some of the medical diagnoses made by the physician against objective data which are free from bias. For example, suppose that on the basis of physicians' dignoses, the frequency of alcoholic liver disease appears to be higher in veteran than in control subjects. This difference could be real, or it could be a result of a systematic bias in the physicians assessments. However, if the objective tests of liver function show more abnormalities in veterans than controls, this would suggest that the difference in diagnostic frequency reflected a real difference in disease frequency, and not just a diagnostic artefact. 88 �On the basis of this example, it might be argued that it would be wise to discard the physicians' judgments and to rely on objective tests alone. Such a policy would be misguided, for several reasons: (i) Objective tests are not available for all organ systems. (ii) Although objective tests can identify the organ system involved (eg liver), additional information is usually neded before an aetiological diagnosis can be made. (iii) Physician judgments are based on contextual clues, and on an "intuitive" synthesis of the available information. It is not possible to automate this synthetic function of the physician, if only because of the difficulty of capturing and codifying all the observations upon which his judgments are based. Any attempt to use the physician merely as an "observer" would be misguided, because it is impossible to separate "observation" from "theory" (contextual analysis and selective aquisition of data to test provisional diagnoses) in the course of medical diagnosis. (iv) The face validity of physicians judgments is widely accepted, both in the medical and in the lay mind. Thus a study which ignored the opinions of physicians could lack credibility in the eyes of the community. 6.2.6 Data From Wives and Female Partners Data from wives and female partners will be obtained via telephone, face-to-face interview or written questionnaire to assess pregnancy outcomes, birth defects and children's health. The data will suffer from subjective bias and selective recall, and even if there is no 89 �real difference, these results could suggest that there is a greater frequency of disability in the families of veterans than in the families of control subjects. Several strategies can be used to assess the validity of these subjective responses. The first is to verify the reported medical condition or event (eg stillbirth, birth defect, curettage for miscarriage) with the medical attendant or hospital authorities. This procedure is adequate as far as it goes, but it suffers from the defect that it is not possible to verify an event which has been forgotten or not reported in the first place. Thus, even using an outcome criterion such as hospital admission for miscarriage, it will not be possible to exclude the possiblity of selective bias in recall between veteran and control wives. The potential for biased will be even greater for those (early) miscarriages which did not result in a hospital admission. Hospital admissions occuring after 1978 are also likely to be subject to bias because of the publicity, from 1979 onwards, surrounding the alleged effects of herbicides. There is less likely to be bias for hospital admissions occuring before 1979. For those outcome conditions which leave a more or less permanent trace (eg surviving children with birth defects or spasticity) the validity of the wife's responses could be assessed, in part, by arranging for a follow-up medical examination of the children affected. This could undoubtedly confirm the diagnoses in the (most severly) affected children, but it would not exclude the possibility 90 �of bias in the initial reporting. For example, it is plausible that there would be less incentive for the wife of a control subject to report the presence of a disability in one of her children, and such a child, if unreported, would be missed from the follow-up study. 6.3 Hypotheses to be Tested 6.3.3 Descript ivc Hypotheses The null hypothesis is that there are no differences in the frequency of disabilities between Vietnam veterans and controls (national servicemen who did not go to Vietnam). This null hypotheses will be tested against each of the following alternative hypotheses: (i) That social and behavioural disabilities (unemployment, separation, divorce, motor accidents, alcohol abuse) are more frequent in veterans than in controls, (ii) That anxiety, depression and other psychiatric disabilities are more frequent in veterans than in controls, (iii) That disorders of the nervous system (including neuropsychological disorders) are more frequent in veterans than controls, (iv) That liver disorders are more frequent in veterans than controls, (v) That gastro-intestinal disorders are more frequent in veterans than controls, (vi) That skin disorders are more frequent in veterans than controls, (vii) That infertility, miscarriage or childhood disability or death have been more frequent in the families of veterans than in the families of controls. 91 �6.3.2 Actiological Hypotheses In the event that one (or more) of these disabilities is more frequently observed in veterans (or their wives and offspring), it will be necessary to explore the causal basis of the difference(s) observed. The following hypotheses need to be considered: (i) That for self-reported symptoms or disabilities, an apparent excess in veterans (or their wives) might be caused by bias between the subjective responses of veteran and control subjects. (ii) That an excess of some disabilities in veterans might be caused by non-comparability of the original groups of veterans and controls (eg in age, education, socio-economic status, and predisposition to subsequent disability), (iii) That an excess of some disabilities in veterans night be caused by the physical and psychosocial sequelea of war service and combat stress, (iv) That an excess of alcohol abuse in veteran, itself attributable to war service, might contribute to any observed excess of social, behavioural and physical disabilities, (v) That an excess of some disabilities might be caused by herbicide exposure in Vietnam, (vi) Than an excess of other disabilities (e.g. tuberculosis, strongyloides, VD) might be caused by other aspects of Vietnam service. 92 �6.4 Outcgmc_Measures_ to be Used 6.4.1 Keed for Simplicity As the protocol calls for the collection of a large amount of information, it is essential to specificy, a priori, a simple set of outcome measures which can be easily used to test the principal hypotheses of interest. Such a scheme is outlined in Table 6. It can be seen that most emphasis is placed on those outcome measures which are valid and unambiguous, potentially relevant and reliably measured. Consequently, at the primary stage of analysis most attention will be paid to objective measures, to physician assessments and to subjective self-reports using psychiatric scales which have been well validated. 93 �Table 6 Major Outcome Measures and Covariates To Be Used In Testing Principal Alternative Hypothesis ALTERNATIVE HYPOTHESES Veterans will show an increased frequency of: 1. Social disability MAJOR OUTCOME MEASURES MOST IMPORTANT CONFOUNDING FACTORS AND COVARIATES (in additions to age, Vietnam exposures to combat, herbicides, etc.) Employment/unemployment Level (status) of employment Pre-enlistment education level, Pre-enlistment psychological assessment Ever married/single Divorced, separated/Presently married Frequency of marital disputes Number of children Religion Substance use and abuse Alcohol use (GGT, urate, MCV) Cigarette use Teas and coffee use Other drug use Marital stutus S.E.S. Behavioural disability Uncontrollable rages Motor accidents Fighting at hotel/football etc Sexual problems Marital disability 2. Psychiatric disability SDI score O mfc• b • Marital Status S «E • S • Alcohol consumption 5 • £• • o • Depression (Hamilton scale) Anxiety (Spielberger) Present state examination Prevalence of psychoactive drug use Scores and components 3. Neur©psychologic disability Symbol digit substitution Scores S.E.S Trail making tests and Alcohol consumption Supra-span digit learning components Nelson adult reading test 4. Neurological disability Physician assessment of - peripheral neuritis & nerve conditions studies - other neurological disability Nerve conduction deafness Alcohol Occupation history (Patency of external ear) 5. Liver disease Liver enzymes Physician assessment of liver disease Alcohol consumption Hepatitis B 94 �6. Cardiovascular disability Blood pressure ECG abnormalities - individual items and components Physician assessment - angina Cigarettes Alcohol - myocardial infarction - stroke - palpitations Plasma cholesterol 7. Infectious disease Hepatitis B virus serology Tuberculosis - CXR report Melioidosis antibody titre Strongyloides antibody titre Venereal disease - VDRL 8. Gastrointestinal disease Physicians assessment irritable bowel syndrome - diarrhoea - ulcerative colitis 9. Skin disorders Physicians assessment 10. Other symptoms A. Symptom complexes, specified jj priori, which are potentially relevant to particular outcomes (above) B. component or factor analyses of symptoms to define their latent structure. These components can then be used as outcome measures to look for differences between veterans and controls. Alcohol consumption Veteran wives partners have an increased frequency of; 1. Infertility Complaints of inability to conceive No. of pregnancies No. of live-born children 2. Miscarriage Verified miscarriage resulting in hospitalization prior to 1979. 3. Birth defects spasticity Verified birth defect and/or spasticity in child born prior to 1979 4. Stillbirth Verified still birth 5. Health disability in surviving children Verified hospitalisation of child prior to 1979 95 �6.4.2 Need for Date Reduction Some of the outcome measures are simple and unambiguous (eg marital status). Others, such as the psychiatric scales contain numerous items which measure several different components, are relevant to psychiatric disability. The dimensions of these more complex outcome measures can be reduced by calculating, for each individual studied, a score on each of the known components. These component scores (e.g. for depression, anxiety and somatic symptoms) can then be used as measures to test for any differences in outcome between veteran and control subjects. In other situations (e.g. with the neuropsychology tests) it may be more appropriate to use the data obtained in the study (from pooled cases and controls to avoid bias) to define the components of interest. Component scores in the reduced number of dimensions can then be used to test for any differences between veterans and controls. 6.4.3 Approach to the Analysis of Subjective Self Reports Interpretation of subjective information presents many problems, and in sec. 6.2.3 some guidelines are given which should be helpful. In particular, it will often be wise to discount any disabilities for which the symptoms are not supported by objective data or physician assessments. In some circumstances (Table 6) it will be possible to define, a_ priori, those self-reported symptoms which are deemed to be relevant to particular outcomes (e.g. the questions defining coronary heart 96 �disease). Scores on these symptom patterns can then be used as outcome measures which should be complementary to those based on objective measures or physician assessments. Another approach is to use the self-reported information itself to explore the factors or components giving rise to the observed variation. Thus it should be appropriate to subject the entire response matrix to factor and/or component analysis. If "expectation of disability" or "response bias" is an important cause of variation in response pattern (i.e. if many people tend to say yes to many symptoms) then this would be reflected in the identification of a "disability factor" which loads for most of the questions. The residual factors or components would then help to identify patterns of symptoms which are not simply due to "general disability" or "response bias" but which are likely to be more useful as potential outcome measures. 6.5 Principles of Statistical Analysis 6.5.1 The Problem The major objective of the analysis is to examine the relationship of the several outcome measures to veteran status, to measures of exposure in Vietnam, and to those explanatory covariates which may be confounded with Vietnam service or with exposure while in Vietnam. The interpretation of morbidity study results would be moderately straghtforward if the allocation of national servicemen to Vietnam service had been completely at random (rather than being haphazardly selective, as was the case), if there had been no mortality while in Vietnam, and if there were no selective compliance with the proposed 97 �interview schedules. If this were the case, then any observed differences between the veteran and control subjects could be interpreted, at least in a very general sense, as being caused by the Vietnam experience. On this view, even the most subjective of self reported symptoms (see Sec. 6.2.3) could be attributed to the Vietnam experience acting through functional or psychogenic processes or through a (conscious or subconscious) desire for compensation. Unfortunately, because of non-random allocation to Vietnam, selective mortality and the likelihood of selective compliance, the veteran and control subjects studied will differ for reasons which may not be logically consequent on Vietnam service; therefore, much of the statistical analysis will be directed towards examining the effects of the variables which are confounded with Vietnam service. This will allow assessment of the effects of Vietnam service on outcome to be made which have been adjusted for the effects of confounding variables. 6.5.2 Basic Approaches to Analysis Because of the non-experimental design, it will be most appropriate to fit regression models to the data, using the outcome measures as the dependent variable. For those situations where the outcome variable is qualitative and binary, it would be appropriate to use a logistic regression model (Cox, 1970; Breslow and Day, 1980) , and for those situations where the outcome variables is quantitiative, the basic approach will be that of standard multiple regression (Draper and Smith, 1966). 98 �The advantage of the regression approach is that it provides a flexible method for dealing with confounding and for estimating the main effects (and interactions) of explanatory variables. This is achieved at the cost of making assumptions about linear effects and about the distribution of residuals. As required, these assumptions can be tested or relaxed at a later stage of the analysis. In some situations it may be appropriate to fit mixed models for the analysis of covariance to allow for the effects of factors which are related to outcome but which have random rather than fixed effects (Sokal and Rolf, 1972). In other situations, with qualitative outcome data and qualitative explanatory variables, it would be appropriate to use log-linear models for the analysis of multidimensional contingency tables (Bishop et al, 1975; GLIM manual - Baker and Nelder, 1978). For those outcome measures which are measured on random samples (because of the hierarchical design) the analyses will be modified accordingly. Special procedures will be developed to use data from the "random" samples and the "extreme value" samples to make efficient estimates about the distribution of the outcome measures over the entire population studied and to find the most efficient procedures for testing fcr differencies in outcome related to veteran status. 6.5.3 Confounding variables can be regarded as (nuisance) variables which have (potential) effects on outcome, and which are •accidentally" correlated with the main factor of interest (veteran status) and logically independent or antecedent in the causal chain. 99 �For example, age is likely to be confounded with veteran status in the present study. Thus because morbidity will be higher in older men, there could, due to confounding, be an artefactual association between morbidity and veteran status. - The appropriate analysis is to first fit a model which includes only the confounding variable(s) (age) and then to fit the factor of interest (veteran status). The test for improvement in fit then provides a measure of the significance of the factor after allowing for the effect of the confounding variable. In practice, it will be necessary to allow for the effects of a number of confounding variables, although care is needed to ensure that variables which are secondary to veteran status are not treated as if they were confounding variables. Effects of interactions can also be estimated. For example, consider the hypothesis that any effect of Vietnam service on subsequent morbidity was greater in men who where older at the time that they went to Vietnam. This hypothesis can be tested by first fitting the main effect of age at time of service (as a confounding variable) and then fitting the main effect of veteran status; the third term (age x veteran status interaction) will provide a test of the required hypothesis. At least in principle, it is also possible to allow for non linear (quadratic) effects of covariates. 100 �6.5.4 Strategy of Model Fitting After the descriptive stage has been completed, it will be important to first fit a model vhich includes all necessary confounding terms (age, educational attainment, religion etc) without regard to veteran status, exposure indices etc. It also seems plausible to allow some degree of overfitting for these confounding variables. Next it will be appropriate to fit (sequentially) the effects related to Vietnam service (veteran status, time in Vietnam, combat exposure) and to retain any significant effects in the model. Thirdly it will be appropriate to look for interactions of these service related variables with the (confounding) explanatory variables (eg age x Vietnam service). 6.5.5 Incompleteness of Adjustments for Confounding Factors If there is an apparent effect of Vietnam service on morbidity which is partly removed when confounding factors are fitted beforehand, it will be necessary to seriously consider the possiblity that the Vietnam effect might have been removed completely if it had been possible to measure the confounding factor(s) more precisely (R. Peto, 1973). For example, if a pre-Vietnam measure (eg psychological assessment at induction) were found to be predictive of outcome and also to be confounded with veteran status, then it might be found to "explain away" a considerable proportion of any effect of Vietnam service. As there is always considerable error associated with the measurement of such psychological scales, it can be argued that a more precisely measured scale might have "explained away* a greater proposition of the observed effect. This qualitative argument could 101 �be made somewhat more precise if data were available on scale reliability and validity. 6.5.6 Problems Arising from Non-orthogonality and Confounding It is important to remember that there may be particular problems in elucidating the significance for outcome of factors-vhich-are confounded with Vietnam service. For example, national servicemen were selected for Vietnam service (either by the army or by themselves), so that veterans will differ from controls for a number of factors, only some of which will have been measured. Consider a factor which is measurable and has been measured (e.g. psychological scale at army induction); furthermore, suppose that Vietnam service is selected partly on the basis of this factor ("they » make good soldiers"). Given this state of affairs, how are we to interpret the relationships between Vietnam service, the "confounding" factor, and outcome? The problems arises, in part, because the measured factor is, in one sense, a cause of Vietnam service, and hence some of the outcome which might be attributable to Vietnam service would, in any convential analysis, be partly attributed to the measured factor because it is a logically prior "confounding" factor and should be fitted first. Thus if we simply adopt the policy of fitting only main effects and of fitting the confounding effect first, we could increase the risk of a Type II error (i.e. of missing a real effect of Vietnam service). To minimise the risk of such Type II errors it will be important to always examine the interaction terms between Vietnam service and each of the "confounding" factors which have a significant main effect on 102 �outcome. For example, we would always be interested to know whether, after adjusting for main effects of Vietnam service and (say) the confounding (psychological) factor, there is a significant 2 way interaction effect on outcome. If there is such a significant interaction, then we are justified in concluding that both factors have real (causal) effects, even though the main effect of Vietnamservice may appear to be non-significant. Thus by fitting such interaction terms it is possible to reduce the Type II error rate with respect to detection of effects related to Vietnam service. It might be argued that an alternative strategy would be to fit the effect for Vietnam service before that for the confounding factor, or at the least to allow the effects to compete with each other at the same stage of model-fitting. Although these issues are complex, it is generally agreed that if the aim of the analysis is to make inferences about attributable (causal) risk, there is usually no justification for fitting first that factor which is logically (and/or causally) secondary. Thus in the context of the present example, we would not usually be justified in fitting the effect for Vietnam service before fitting an effect for a factor (e.g. psychological scale at induction) which is logically prior to Vietnam service. In other words, although it is plausible to postulate that a psychological factor could influence the probability of Vietnam service, it is much less plausible to postulate that Vietnam service could influence psychological measures measured at army induction (i.e. well before the process of selection for Vietnam service began) . 103 �6.5.7 Approach to the Testing qf_Aeti_ological Hypotheses Thus in developing statistical procedures for testing aetiological hypotheses, it is important firstly to impose a causal ordering on the explanatory variables, and secondly to see whether, after allowing for main effects of confounding variable and Vietnam service, the interaction terms also have significant effects on outcome. If the interaction terms are significant, this provides further evidence for rejecting the null hypothesis in relation to Vietnam service. 104 �References Adams, I.M., Hehir, P.J., Byth, K., Charlesworth, N., Mears, A. Report on the Feasibility of Establishing an Index of Exposure to Herbicides for Vietnam Veterans. (Confidential document) Australian Veterans Herbicide Studies, November 1981. American Psychiatric Association. Diagnositic and Statistical Manual of Mental Disorder, ed. 3., Washington, D.C, 1980. Army Manning Review 1973-74, Department of Defence, Canberra. Axelson, O. and Sundell, L. Herbicide exposure, mortality and tumour incidence. An epidemiological investigation on Swedish railroad workers. Work Environment Health, 11, 21-28 (1974). Baker, R.J. and Nelder J.A. The GLIM system release 3. Generalised linear interactive modelling. Numerical Algorithms Group, Oxford, (1978). Bishop, Y.M.M., Fienberg S.E., Holland, P.W. Discrete multivariate analysis: Theory and practice. MIT Press, Cambridge, (1975). Boman, B. Review: the Vietnam veteran ten years on. Aust. N.2. J. Psychiat. 1982; 16:107-127. Breslow, N.E. and Day, N.E. Statistical methods in cancer research. I. The analysis of case-control studies. IARC Scientific Publications No. 33 (Lyon) , (1980). Case-Control Study of Congenital Anomalies and Vietnam Service. (Confidential document) Australian Veterans Herbicide Studies, April, 1982. Cox, D.R. Analysis of binary data. Methuen, (1970) . Draper, N.R. and Smith, H. Applied regression analysis. Wiley, N.Y., (1966). Fett, M.J. Retrospective Mortality Study of Vietnam Veterans and Controls. Revised Protocol. (Confidential document). Australian Veterans Health Studies, May, 1982. Figley, C.R. and Stretch, R.H. Vietnam Veterans Questionnaire. manuscript. U.S. Veterans' Administration, (1980). Unpublished Hardell, L. and Eriksson, M. Soft-tissue sarcomas, phenoxy herbicides and chlorinated phenols. Lancet, j_i, 250, August 1, (1981) . Hardell, L. and Sandstrom, A. Case-control study: Soft-tissue sarcomas and exposure to phenoxyacetic acids or chlorophenols. Br. J. Cancer, 39, 711-717, (1979). Honchat, P.A. and Halperin, W.E. 2,4,5-T, trichlorophenol and soft tissue sarcoma. Lancet, \_, 268-9, January 31, ( 9 1 . 18) International Agency for Research on Cancer. IARC monographs on the evaluation of the carcinogenic risk of chemicals to man: Some fumigants, the herbicides 2,4-D and 2,4,5-T, chlorinated dibenzodioxins and miscellaneous industrial chemicals. 15, 41-299, (1977). 105 �Ma thews, J.D. Alcohol use as a possible explanation for socio-economic and occupational differentials in mortality from hypertension and coronary heart disease in England and Wales. Aust. N.Z. J. Med., j>, 393-397, (1976). Ma thews, J.D. Genetics and alcohol: Implications for human disease. N.Z.J. Med. n, 109-114, (1981). Aust. National Service Act 1951-1971, The Commonwealth of Australia. Penk, K.E., Robinowitz, W.R., Roberts, E.T., Patterson, M.P. , Dolan, M.P. and Atkins, H.G. Adjustment differences among male substance abusers varying in degree of combat experience in Vietnam. J. Consult. Clin. Psychol. , 4:9, 426-437, (1981). Peto, R. 1973, (Personal communication). Reggiani, G. Acute human exposure to TCDD in Seveso, Italy. J. Toxicol. Environ. Health, Jj, 27-43, (1980). Riihimaki, V., Asp, S., Seppalainen, A.M. and Heinberg, S., (1978) Mortality study of persons exposed to dioxin after an accident which occurred in the BASF on 13th November 1953. Working paper for the Workshop-on Long-term Hazards of Polychlorinated Dibenzodioxins and Polychlorinated Dibenzofurans. International Agency for Research on Cancer, Lyons, January 10-12. Cited in Kimbrough, (1980). Scientific Advisory Committee Report. (Confidential document) Australian Veterans Herbicide Studies, February, 1981. Sokal and Rolf. Biometry. Freeman, San Francisco, 1972. The PULHEEMS System of Medical Classification. Medical Pamphlet. Department of Defence (Army Office), Canberra, 1978. U.S. Veterans' Administration. Review of literature on herbicides, including phenoxy herbicides and associated dioxins. Volume I, Washington D.C., 1981. 106 �APPENDIX 1 SPECIFIC HYPOTHKSES AND POWER CALCULATIONS The null hypothesis is that there are no differences in the frequency of disabilities between Vietnam veterans and controls (national servicemen who did not go to Vietnam). These null hypotheses will be tested against the alternative hypotheses that each outcome listed in Table 1 is more frequent in veterans than controls. In interpreting Table 1 the following points should be borne in mind: (1) Smaller numbers of study subjects (sample sizes) than those given have the consequence of increasing the minimum size of an effect (the relative risk) that can be detected, that is, of reducing the sensitivity of the study to detect veteran/control differences. (2) In the process of adjusting for pre-Vietnam and other differences in the veteran/control group (for example, differences in marital status at enlistment), the study becomes less sensitive than shown in Table' 1. The data in Table 1 are therefore 'best possible case1 data, where no adjustment for veteran/control differences is required. (3) Many of the putative effects of Vietnam service appear to be related to particular aspects of Vietnam service (e.g. combat). Since it is unlikely that all veterans would be exposed to the factor of importance (e.g. combat), the number of truly 'exposed1 veteran subjects may be considerably below the total number of veteran �subjects. This further reduces the power of the study to detect veteran/control differences below those shown in Table 1. This consideration is the rationale for selecting veterans to controls in a 3 to 2 ratio, thereby permitting more powerful comparisons between veterans at different levels of exposure to possible causal variables for a given sample size, and at the same time reducing only slightly the sensitivity of veteran-control comparisons. It is not currently possible to perform power calculations within the veteran group, as the distribution of potentially causal variables among veterans is not yet known. �• I Outcome Source Estimated prevalence (limits) 2,500 subjects Rel. risk Min VN-caused detectable cases detectable/1000 Cancer (i) 1% 2.23 Husculoskeletal Arthritis/rheumatism Muscular aches & pains (h) (a) 3% 10-30% 1.65 1.16-1.32 Infectious Disease Syphilis (VDRM+) ) Melioidosis Strongyloidiasis (a) , ( j ) (k) , (1) (P) 0.5-1% 1-3% 1% in vets 5 , U U U siJD^eci-t. 12 Min VN-caused cases detectable/1000 Rel. r i s k detectable Min VN-causod cases detectable/1000 1.83 8 Rel. risk detectable 5 1.63 4 1.45 19-28 1.11-1.23 2.23-2.90 1.65-2.23 2.23 6-8 8-12 1.83-2.24 1.45-1.83 8 8 14-20 1.34 1.09-1.18 6 11-16 1.63-1.94 1.34-1.63 1.83 4-5 5-8 5 1.63 3-4 4-6 4 Medical Treatment Hospitalised in last 12 months Medication taken in last 2 days (h) 14% 1.26 22 1.19 16 1.14 12 (h) 37% 1.14 30 1.10 21 1.08 17 Reproductive Relative infertility (a) 5% 1.48 14 1.33 10 1.26 8 Social and Behavioural Never married Divorced Currently unemployed Recent accident (h) (h) (n) (b) 13% 2.5% 4% 6% 1.28 1.74 1.55 1.43 22 11 13 16 1.19 1.50 1.38 1.30 15 8 9 11 1.15 1.39 1.28 1.23 12 6 7 8 3-16% 1-19% 1.24-1.65 1.22-2.23 12-23 8-25 1.17-1.45 1.15-1.83 8-16 5-17 1.13-1.34 1.12-1.63 6-13 4-14 0.5% 1-2% 3-10% 2.90 1.83-2.23 1.32-1.65 6 8-10 12-19 2.24 4 5-7 8-14 1.94 1.43-1.63 1.18-1.34 3 4-5 6-11 Psychiatric (o) Depression Anxiety (0) Severe personality disorder (a) (a) ,(p) Psychosis j Alcoholism (100gms+/d) (g) Smoking: see Respiratory * 1.53-1.83 1.23-1.45 Table 1. Prevalence of Outcomes of Importance with Minimum Relative Risks Detectable as Statistically Significant and the Minimum Number of Vietnam-Caused Cases per 1,000 subjects that would be Detected with Power of 80% at PJJ_J 0.05 for Sample Sizes of 2,500, 5,000 and 18,000 subjects. �Outcome Kervous system Clinical peripheral neuropathy Neu r opsy cholog ica 1 abnormality Hearing abnormality Frequent headaches Skin Severe acne Fungal infections Liver Cirrhosis Hepatitis B(+) aerology Abnormal liver function tests Source Estimated prevalence (limits) 2,500 subjects Min VN-caused Rel. risk detectable cases detectable/1000 5,000 subjects Rel. risk Min VN-caused detectable cases detectable/1000 Rel. r i s k detectable 1.63-1.94 ( 2 ) , (a) 0.5-1% 2.23-2.90 6-8 1.83-2.24 4-5 (3) (3) (4),(b) 5% 5% 3-4% 1.48 1.48 1.33 1.33 1.55-1.65 14 14 12-13 10 10 8-9 (a) (a) 3% 4% 1.65 1.55 12 13 1.38-1.45 1.45 1.38 8 9 3-4 1.26 1.26 8 8 6-7 1.28-1.34 1.34 1.28 6 7 1.85 2-4 3 10 1.26 8 1.26 1.18 1.63 1.56-1.63 10 14 5 5-7 1.43-1.63 8 11 4 4-5 14 1..33 10 1.26 8-10 1.56-1.83 5-7 1.43-1.63 (a) (c) 0.1-1% 0.6% 2.70 3-8 6 (3) 5% 1.48 (a) (a) (b) (d) 5% 10% 1% 1-2% 1.48 1.32 2.33 (3) 1.83-4.3 2.11 14 1.33 1.33 1.23 1.83 1.83-2.23 14 19 8 8-10 5% 1.48 (e) 1-2% 1.83-2.23 (b) (f) 0.1% 0.6% 6.3 2.7 Respiratory Smoking currently Asthma Abnormal pulmonary (g) 40-46% (h) 2-4% 1.11-1.13 1.55-1.83 function tests (3) 5% 1.48 Cardiovascular Hypertension Receiving medication for hypertension Symptoms of ischemic heart disease ECG abnormality Min VN-caus«?d cases detectable/1000 % 2.23-6.3 2-5 4 Gastrointestinal Persistent vomiting Persistent diarrhoea Abdominal pain Peptic ulcer 8,000 subjects V. 3 ' 6 30-31 10-13 14 4.3 2.11 1.08-1.09 1.38-1.56 1.33 2 4 22 7-9 10 1.63-3.5 3.5 1.85 1.06-1.07 1.28-1.43 1.26 8 4-5 2 3 17 5-7 �Notes (1) The power calculations have been performed using the following parameters: o veteran to control ratio of 3 to 2 o the 'relative risk detectable1 is the minimum relative risk that would be statistically significant at the level ?„* 0.05, with power of 80% o (2) the excess cases in veterans are derived from the relative risk detectable and the veteran to control ratio of 3 to 2. Pilot Study prevalence data have been used only where other data have not been obtainable, since the validity of Pilot Study data is limited. The prevalence rates cited are for veteran and control subjects combined. (3) For this measure an abnormal (positive) result is defined as that result above which only 5% of the normal population resides. * (4) All data are for the appropriate age and sex group. Sources of Prevalence Data (a) Pilot Study, Australian Veterans Health Studies. (b) Australian Health Survey 1977-78 Recent Illness. ABS. (c) (d) Nelson (1975) Health Care Surveys Gosford/Wyong/Illawara 1975. ABS, HC NSW. (e) Australian Health Survey 1977-78. ABS. (f) Personal Communication: Dr T. Ireland, Director of Research, Medicheck Centre. (g) Alcohol and Tobacco Consumption Patterns. (h) (i) Social Indicators. No. 3. 1980. ABS. Cancer in NSW Incidence and Mortality 1977. NSW Cancer Registry. (j) Personal Communication: A. Lee, Serologist, NSW Red Cross Blood Transfusion Service, (k) Clayton et al (1973) . (1) Kishimoto et al (1971). (m) (n) Gilbert et al (1968). Labour Statistics 1980 Australia. (o) (p) Reynolds and Rizzo (1979). Goldberg (1972). ABS. February 1977. ABS. �References Australian Bureau of Statistics. Alcohol and Tobacco Consumption Patterns. February 1977. ABS Catalgoue No. 4312.0, Canberra, 1978. Australian Bureau of Statistics. Australian Health "Survey 1977-1978. ABS Catalogue No. 4311.0, Canberra, 1979. Australian Bureau of Statistics. Labour Statistics 1980 Australia. ABS Catalogue No. 6101.0, Canberra, 1980. Australian Bureau of Statistics. Social Indicators. No. 3. 1980. ABS Catalogue No. 4101.0, Canberra, 1980. Australian Bureau of Statistics. Australian Health Survey 1977-1978. Recent Illness. ABS Catalogue No. 4318.0, Canberra, 1981. Australian Bureau of Statistics and Health Commission of New South Wales. Health Care Surveys in Gosford-Wyong and Illawarra areas of NSW. 1975. AB^S Catalogue No. 4305.1, Sydney, Clayton, A.J., Lisella, R.S. and Martin, D.G. 1976. Melioidosis: A Serological Survey in Military Personnel. Milit. Med. 3J3 : 24-26, (1973). Gilbert, D.N., Moore, W.L.f Hedberg, C.L. and Sanford, J.P. Potential Medical Problems in Personnel Returning from Vietnam. 68 t 662-678, (1968). Review. Ann. Int. Med. �Goldberg, D.P. The Detection of Psychiatric Illness by Questionnaire. Institute of Psychiatry Maudsley Monographs No. 21. Oxford University Press, London, 1972. Kishimoto, R.A., Brown, G.L., Blair, E.B. and Wenkheimer, D. Melioidosis: Serologic Studies on US Army Personnel Returning from South East Asia. Milit. Med. 3j> : 694-698, (1971). Nelson, M. immunology and Epidemiology of the Hepatitis B (Australia) Antigen. PhD thesis. Unilversity of New South Wales, 1975. New South Kales Central Cancer Registry. Cancer in New South Wales Incidence and Mortality 1977. Health Commission of New South Wales, Sydney, 1981. Reynolds, I. and Rizzo, C. Psychosocial Problems of Sydney Adults. Commission of NSW and Medicheck Referral Centre. 1979. Health �APPENDIX 2 PULHEEMS ARMY HEALTH RATINGS The following tables outline the basis of the PULHEEMS ratings, and are extracts from "The PULHEEMS System of Medical Classification1, 1978. NOTES ON THE PULHEEMS QUALITIES 15. a. P - Physical Capacity. This indicates general physical development, potential capacity to acquire a high level of physical stamina, capacity for hard work. b. U - Upper Limbs. This indicates the functional use of hands, arms, shoulder girdle and upper spine. Htyere there is a degree of incapacity which would limit general physical capacity the U assessment will also affect the P assessment. c. L - Lower Limbs. This indicates the functional efficiency of feet, legs, thighs, pelvis, lumbar spine, ankle, knee and hip joints. As with the U assessment the L assessment may also affect the P assessment. d. H - Hearing Acuity. This indicates hearing acuity only. Diseases of ear are to be assessed under the P quality. e. EE - Eyesight. This indicates visual acuity only. Diseases of the eyes are to be assessed under the P quality. f. M - Mental Capacity. This is difficult to assess on the basis of a single medical examination. Some guidance is given by: (1) (2) School record and post-school occupational record; (3) g. Impression given on interview with regard to alertness and intelligence; Selection test results. S - Emotional Stability This is also difficult to assess on the basis of a single examination. There are no tests of temperament or personality available to estimate emotional stability. Reliance must be placed on careful history taking, including family background and employment record, and physical examination. �DEGREES OF PULHEEMS QUALITIES 16. There are nine degrees of qualities but not all of these are in use. The following table shows the degrees used under each quality. U 17. H M The broad correlation between degress of P, U, L, M, S, and functional capacity, combatant capacity, climatic restriction, is as follows: Degree Function Capacity Combatant Capacity Above Average Full Climatic Restriction None Non tropital Average None Non tropical Below Average None Non tropical Markedly Deminished Restricted Serve in Australia Note: The restriction "Service in Australia" applies only to P 7 and not to U7 or L7, Under P, U, L, E, M, S « medically unfit for. any form of service. � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 073 Folder The folder containing the original item. 1930 Series The series number of the original item. Series III Subseries IV Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Fett, M. J. J. D. Mathews R. J. Walsh Date A point or period of time associated with an event in the lifecycle of the resource December 13 1982 Title A name given to the resource Australian Veterans Health Studies: Morbidity Study: Protocol for a Morbidity Survey of Vietnam Veterans and Controls Subject The topic of the resource AVHS study protocol morbidity trends ao_seriesIII Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 075 Folder The folder containing the original item. 1941 Series The series number of the original item. Series III Subseries IV Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Walsh, R. J. Description An account of the resource <strong>Corporate Author: </strong>Commonwealth Institute of Health, University of Sydney, Australia Date A point or period of time associated with an event in the lifecycle of the resource 1983-07-01 Title A name given to the resource Australian Veterans Health Studies: Pilot Study Report, Volume I: A Report into the feasibility of an epidemiological investigation of morbidity in Vietnam veterans Subject The topic of the resource AVHS morbidity trends questionnaire ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/252560d2f96615aedef84b0d2d86d088.pdf f39caa3a4fda966f6b7e7d82bf31410b PDF Text Text Item ID Number 01858 Author McKinley, Thomas W. Corporate Author Georgia Department of Human Resources ROpOTt/ArtlGto HUB Geor 9'a Agent Orange Survey of Vietnam Veterans: Summary Journal/Book Title Year 1983 Month/Day Ju| Color Number of Images v || : 11 Descrlpton Notes Wednesday, July 11, 2001 Page 1859 of 1870 �v.v$? --•••£•*'* Georgia Survey of Vietnam Veterans -"'&:.: • IT: &* ;•«?» •SSfei 4^ •S" s^si^s stA^ �SUMMARY Published by GEORGIA DEPARTMENT James G. Ledbetter, PhD, Commissioner 47 Trinity Avenu® , S.W. Atlanta, Georgia 30334 JULY 1983 Prepared by Thomas W. McKintey, MPH Epidemiologist th& Direction of R. Keith Sikes, DVM, MPH Director, Office of Epidemiology James W. Alley, MD, MPH Director, Division of Public Health �flCKNQWLEDGEMENTS The Office of Epidemiology expresses grateful appreciation for the suggestions and guidance provided by the following persons who comprised an ad hoc fldvisory Committee for the Georgia ftgent Orange Study: Committee* Members Thomas W. McKinley, MPH (Chairman) Office of Epidemiology, DHR R. Keith Sikes, DVM, MPH Office of Epidemiology, DHR Douglas Huber, MD John Brady Nam Vets of Georgia •*» ^a. Dept. of Vet. Services D. S. Wilkerson Ga. D^nt. of Vet. Services Julian ft. Jarman, MD Decatur Vft Hor-p:'tal James R. Bishop Decatur VO Hosoitai Observers John D. Humphreys Division of Pub. Health,DHR Don Barrish Office of Community Inte*—Gov. Relations, and DHR �GEORGIA SURVEY The 198£ Georgia General ftssernbly passed House B i l l 1£00* entitled "Reports of Veterans Exposed to Agent Orange." ft sum of $67,525 was appropriated for the Department of Human Resources to conduct a questionnaire survey of Vietnam veterans exposed to flgent Orange during the Vietnam conflict. Recording to Veterans ndrr,inistrat ion (VO) estimates, approximately 58,008 Georgians Served in Vietnam. ft list of Vietnam veterans was not available from the Georgia Department of Veterans Services to use as a basis for the survey. Therefore, it was necess-ary to use registers of veterans who took the ficent Orange physical examination being offered by Vft hospitals and membership lists from organisations such as Nam Vets of Georgia. In addition, veterans were reached oy publicity campaigns and by placing posters, brochures. and Questionnaires in Georgia Department of Veterans Services Offices and other locations freauented by veterans throughout the state. firrancsments were also made with Ti.el.ine, the state telephone information and referral system, to allow Vietnam veterans to call toll free from anywhere in the state and request a questionnaire. ftpproximately £6, 030' questionnaires were distributed; 9.6% by direct mailing and 90.4% by placement in locations freauented by veterans. Participation was limited to Vietnam veterans residing in the state at the time of the survey. General objectives were to: 1. Obtain completed questionnaires by Marcn 31, 1383, from the largest possible number of veterans in Georgia who: (a) (b) currently reside in Georgia, (c) had known or presumed exposure to flgent Orange, and (d) £. served in Vietnam, period 196E-1974, Laos, or Cambodia during the have seen a physician for a health problem believed to be related to Agent Orange Exposure.' Verify medical histories given by veterans by querying physicians and/or hospitals identified on veteran questionnaires. *Sponsored by Representatives Eleanor L. Richardson, Joe T. Wood, Forest Hayes, Jr., Joe Frank Harris, and Paul S. Branch, Jr. �3. Analyse and summarize data from veterans, physicians, and hospitals. 4. Report findings to General flssembly. the 1384 session of the Georgia Results fls of June 30, 1983, quest ionnaires were received form 1905 veterans. These quest ionnaTres form the basis for a regTsTfry of~ Vietnam veterans in Georgia whose illnesses are allegedly due to flgent Orange exposure or who have health concerns about flgent Orange exposure. Of the total questionnaires received, 1£6S _ (67.6%) were e 1 i g i big for i >^gjLusjLon__i n the survey based on the abov"e~cr i t erTa". Questionnaires were received from 1£4 . o ^ f Georgia's 159« counties (Figure 1). " flpproxirnately 97% of the survey group were' males; 65"/« were white and 30"/. black. ftge ranged from £3-77 years; mean 39.4 years. Major findings of the survey are contained in the following statemants. Interpretation of these findings must take into consideration the fact that 1) the survey targeted veterans who had one or more health conditions which they believe to be related to flgent Orange exposure, £) a. subst ant_ial proport ion of Jnejalth conditions r^^£Il^JigL__by veterans^were not confirmed by 'thjnTr bhysi'cians^ and may have been reported on the basTsoTsUTf^ diagnosis, and 3) information regarding exposure to flgent Orange is totally dependent uoon recall of sometimes uncertain- events which occurred 10-15 years ago. 1. 1£88 Vietnam veterans in the State of Georgia reported having one or more health conditions which .they believe to be related to exposure to flgent Orange. Health conditions r_e_p_ortgd__ by more than half the veterans include s k in cond i t i ons (other than acne), emotional/adjustment problems, nervous system problems, and sleeplessness. £. Only 52?t of 'survey participants had taken Orange physical examination offered by Vfl. the flgent 3. fl substantial proportion of veterans (£9% during their first tour of duty) reported being sprayed with flgent Orange by aircraft. 4. Veterans reported 205 cases of acne with onset after service in Vietnam. Physicians confirmed £9 cases in 119 of these reports (£4. 4"/i), but there was no indication that the cases were chloracne (a specific type of acne caused by exposure to dioxin and other chlorinated biphenyls). .Vfl has acknowledged only two or three cases of chloracne in Georgia veterans. �5. Veterans who participated in Operation Ranch Hand (code name for the group who sprayed flgent Orange) reported a significantly higher prevalence of cancer, liver problems, respiratory problems, sexual dysfunction, and chronic pain than other veterans. 6. Veterans who remembered developing sorna type of illness within 43 hours of exposure to flgent Orange, reported a significantly higher prevalence of 12 of 30 medical conditions. 7. Veterans reported 99 cases of cancer, but physicians completing questionnaires on 47 of these confirmed only 13 (£1.3%). Theoretically, all Georgia Vietnam veterans (est. 58,000) could have participated in the survey if they have a health problem, including cancer, which they believe to be related to flgent Orange exposure. There are at least two ways to analyze the cancer data: (a) The first method of analysis involves a comparison of observed to expected cases. Using cancer surveillance data and assuming that the total population of Georgia Vietnam veterans has the same race, sex, and ace distribution as the survey group, Jbh_e expected number, ._._..Q.f_ cases i n . JLJlg_t ^!^gL'-!^QJ[li-ia ^ i'ietTvam^veFe'r an population _ - - . ^ 37 7. If the actual number of "cases ir\~t h e sorVey groTTp"" is 10^ this would only be three percent of the expected. If the actual number is 21, this would be six percent of expected. If the actual total is 99, this would be £3* of expected. (b) 0 second method of analysis consists of comparing the observed prevalence rate of living cancer cases in the survey group to the expected prevalence rate estimated for all Georgia Vietnam veterans. The expected prevalence rate of living cancer cases in the total population of Georgia Vietnam veterans was derived using cancer surveillance data arid the assumptions indicated in (a) above. If the actual number of cancer cases in the survey group is only 10, this would give a prevalence rate of 77S per 130,000 which is not significantly different from the expected prevalence rate of 613 per 100,000. If the actual number of cases is £1, the observed prevalence rate would be sigificantly higher than expected (p<.01; Chi-square test). However, these data must be interpreted with caution since the survey design tended to inflate the number of cases of illness in the survey group.. JJia—Sj^yjey^jdesigji,_ ijo_jHa.c_tJL__doGs not al low for a determination of whjstjier cancer rjat e s are higher i n~vTetnarn vc?tt?rans exposed to _ ~ unex posed popUIatjjgru This and similar determinations" "musTE await completion of the large population based study being conducted by the Centers for Disease Control. �8. Negative pregnancy outcomes reported by veterans were less than 6.5^ of the number expected for any negative pregnancy outcome among families of all 58,000 Georgia Vietnam veterans. Pregnancy outcomes were not confirmed by physician questionnaires or other means. 9, The rate of cancer, other than leukemia, for progeny of Vietnam veterans was not significantly different between those children born before and those born after the father's Vietnam service. Veterans reported two cases of leukemia in children born after Vietnam service, but meaningful comparisons were not possible since physician confirmation of these cases was not• obtained. 0 MORE DETAILED REPORT OF THE STUDY IS WftlLftBLE ON REQUEST �Figure STATE : .r>?st->:*zvfiv/ ^^^S iift," ^1 i *:'?i'5?!:''''"'>:''"1''';>'''"ViiJii;!!i>; CAL**OW* ! - «*<«.«» -' i /» v '-j '- ,•' /•,'--""/>" ' " vr«g^s;.sss.s«8;r' »«" ^^^^Ix ""*"' /; �RECOMMENDRTIONS This report completes the charge to conduct an flgent Orange survey which was given to the Department of Human Resources by the 1982 Georgia General assembly. The following recommendations are made as a result of that survey: 1. £. Consideration should be given to setting up an flgent Orange clearinghouse or phone center which would receive inquiries and complaints from veterans, dependents and others, and would transmit to interested persons information with resoect to flgent Orange or dioxin-related matters. •*• Veterans who have not taken.the Vfl flgent Orange physical examination should be encouraged to take the examination at the earliest time. 3. The list of veterans who indicated they participated in Operation Ranch Hand should be checked against military study records to determine whether all these veterans are enrol led in the Ranch Hand Study. 4. The Vfl should be asked to evaluate or re-evaluate, as the case may be, veterans whase physicians confirmed a diagnosis of acne after age 18 to determine whether they may have ch lor acne.. 5. figent Orange questionnaires, computer tapes' containing data on health conditions, and other pertinent files and records should be transferred to the Georgia Department of Veterans Service for safe keeping and possible use when results arc completed on the CDC epidemiologic study. S. Odditional studies regarding the question of Ogent Orange exposure and health of Vietnam veterans in Georgia should await the results of the CDC epideniiolog ical cohort study. �SUKMflRY OF HEflLTH EFFECTS OF DIOXIN EXPOSURE figent Orange consisted of an approximately equal mixture of two common herbicides, 2,4-D (£, 4-dichlorophenoxy acetic acid) and 2,4,5-T (2,4,S-trichlorophenoxy acetic acid). The latter herbicide contained a small amount (average 2 parts per million) of a chemical contaminant known as TCDD (2,3,7,8tetrachlorodibenzo-para-dioxin), also commonly referred to as "dioxin." This contaminant, which is formed if the reaction temperature becomes too high during synthesis of 2,4,5-T, has been called the "most toxic man made substance known" because of its highly lethal effects on certain strains of guinea pigs. To date__t_here ^rg__np conclusive studies which causal 1 y link TCDjD_ or 'fl"ge_nt Q*"anJe_j!JiP-gj?ure with_j?xcessive mortality or long term health effects___i-Ki—huwajas...— Information on'TfeaTth effects comes almostentirely from animal studies, which are not directly predictive of effects in hurnans, and from human occupational exposures to herbicides and other chemicals contaminated with TCDD. What is known regarding health effects is briefly summarized in the following paragraphs. Persons exposed to high concentrations of TCDD by reason of occupation or industrial accident were commonly observed to develop a painful skin, condition called chloracne. This condition usually appeared within weeks to months following exposure and persisted for one to several years, depending on the severity of exposure. Other health effects have also been observed in severely exposed persons. For example, a condition known as porphyria cutanea tarda, which is characterized by large blisters of the skin and liver involvement, was reported among at least two groups of exposed workers. In addition, Swedish investigators have recently suggested that there may be a relationship between exposure to TCDD containing herbicides and a form of cancer known as soft tissue sarcoma. However, information to date is not sufficiently completed to establish a cause and effect relationship. Birth defects were reported among children born to south Vietnamese refugees who sought sanctuary in north Vietnam. Pi higher rate of birth defects was also reported among infants born to women whose husbands fought in south Vietnam compared to those born to women whose husbands stayed in north Vietnam. Results of these observations are in doubt, however, "due to methodological problems attendant with ascertainment of information in a war-torn area. Increased abortion rates were also reported among women living in the PUsea, Oregon area where 2,4,5-T had been used for forest management. fln EPPt study tended to confirm this report, but the EPft study was later found to have serious problems with incomplete ascertainment of data. �flnimal studies have shown that rabbits and monkeys develop chloracne when exposed to subacute doses of TCDD. Subacute exposure has also been shown to produce severe weight loss and porphyria (a disorder of hemoglobin metabolism) in certain animal species. Carcinogenicity testing of TCDD in rats and mice has yielded results that are difficult to interpret. Increases were observed in cancerous tumors but only at doses which produced other toxic effects. There was a general lack of both organ specificity and linear dose response usually observed with cancer causing agents. In one study a certain strain of mice fed combinations of TCDD and S,4,5-trichlorophenoxyethanol showed a significantly higher incidence of liver cancer than controls. These observations led investigators to hypothesize that'TCDD may be a tumor promoter rather than a primary carcinogen. However, —in actual trials in rats and mice, TCDD was not shown, to be a tumor promoter. In test systems which employed TCDD and a carcinogenic polyaromatic hydrocarbon, TCDD was observed to inhibit tumor formation by inducing the production of enzymes which converted the polyarornatic hydrocarbons into non-cancer causing metabolites. In other animal studies, certain strains of pregnant mice showed fetatoxicity and birth defects in their offspring after TCDD exposure; however, exposed male mice were not shown to produce deformed offspring. � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 069 Folder The folder containing the original item. 1858 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource McKinley, Thomas W. Description An account of the resource <strong>Corporate Author: </strong>Georgia Department of Human Resources Date A point or period of time associated with an event in the lifecycle of the resource 1983-07-01 Title A name given to the resource Georgia Agent Orange Survey of Vietnam Veterans: Summary Subject The topic of the resource state-funded Vietnam veterans study questionnaire morbidity trends dioxin ao_seriesIII Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 073 Folder The folder containing the original item. 1926 Series The series number of the original item. Series III Subseries IV Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Description An account of the resource <strong>Corporate Author: </strong>Parliament of the Commonwealth of Australia, Senate Standing Committee on Science and the Environment, Canberra, Australia Date A point or period of time associated with an event in the lifecycle of the resource 1982-11-01 Title A name given to the resource Pesticides and the Health of Australian Vietnam Veterans: First Report, November 1982 Subject The topic of the resource Australian Parliamentary hearings mortality trends morbidity trends veteran psychological health ao_seriesIII Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 027 Folder The folder containing the original item. 0407 Series The series number of the original item. Series III Subseries I Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Source A related resource from which the described resource is derived Chemical and Engineering News Date A point or period of time associated with an event in the lifecycle of the resource October 29 1979 Title A name given to the resource Study Shows Few Health Effects From Dioxin Subject The topic of the resource Monsanto Seveso industrial exposure morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/72404de6080fcd82751a4b3b21072f06.pdf 1575cc3e0c554ac58a4b7d11a7c31708 PDF Text Text Item ID Number oieoo Author Corporate Author Typescript: Draft, Ranch Hand Morbidity Report Journal/Book Title Year 000 ° Month/Day Color Number of Images n 7 Desoripton Notes Wednesday, May 23, 2001 Page 1601 of 1608 �DRAFT RANCH HAND MORBIDITY REPORT 1. The A1r Force study of the health status of Individuals exposed to Herbicide Orange and D1ox1n during the Ranch Hand Operation 1n Vietnam was released on 24 February 1984. Over 1,000 Ranch Hand members completed the questionnaires and physical examination. Their results were compared with a carefully selected group of Individuals who also served in Southeast Asia but were not exposed to Herbicide Orange or D1ox1n. Copies of the press release and executive summary of the report are attached. 2. Important findings of Interest include the following: a. Conditions which have been attributed to exposure. No cases of soft tissue sarcoma (a form of cancer), porphyrla cutanea tarda (a Hver disorder), or chloracne (a skin condition) were found 1n any Ranch Hander, Various investigators have previously reported herbicide or dloxln exposure caused these specific conditions. b. Cancer There was no difference in occurrence of systemic cancer. Some types of systemic cancer were found only in comparison subjects. (See Attachment 1.) Skin cancer, of the type that can be completely cured by simple excision (basal cell carcinoma), was seen more frequently in Ranch Handers. However, 1t was not possible to determine sun exposure history 1n either group. Sun exposure 1s recognized as a leading cause of such cancers. Necessary data on sun exposure will be obtained at the scheduled follow-up examinations, c. Fertility/Infertility There were no differences 1n fertility/Infertility Indices, miscarriages, stillbirths, live birth rates, sperm amount and type, frequency /> �DRAFT of severe birth defects (life threatening), or moderate birth defects (require medical care for correction). There was an Increase 1n rate of reported lim- ited (minor) birth defects 1n the Ranch Hand group. As noted on attachment 2, these Include a number of Inconsequential skin conditions. If these frequently seen, clinically Insignificant anomalies are excluded from the calculations, no statistically significant difference exists 1n limited birth defects. Neonatal deaths (deaths within 28 days after delivery) were reported mpre frequently in Ranch HanderSi As noted 1n attachment 3, this difference may be due to marked underreporting of deaths 1n this category post SEA 1n the comparison subjects. Verification of the reported conditions by medical record and birth/death certificate review is underway. d. Liver No difference in frequency of liver disease was detected by laboratory i testing or physical examination. The majority of laboratory tests were slmi1a " In thi twf groupoi There were oome minor differences In means uf A rum laboratory tests but, as noted 1n Attachment 4, these values were still well within normal limits and the differences were slight. There was no difference 1n the past occurrence of hepatitis, jaundice, or cirrhosis. There were more verified past liver abnormalities in the unspecified or miscellaneous category. As noted, no difference 1n the two groups (comparison and Ranch Handers) was found at the examination and the significance of the past mlscellenous liver conditions 1s uncertain. e. Cardiovascular No differences were found 1n the frequency of abnormal electro- cardiogram or abnormal blood pressure. Pulse Intensity was decreased more �DRAFT frequently 1n two leg vessels 1n the Ranch Hand group. These determinations were made only by palpation, a relatively crude technique that can be affected by obesity, edema, and similar factors. Measurement by a more sophisticated technique allowing quantification (Doppler) will be accomplished at the next scheduled examination to determine whether this pulse difference actually exists or 1s due to artifact. f t Psychology No difference was found 1n the objective psychological tests, e.g., IQ. Some differences of uncertain clinical significance were noted 1n subjective measurements, particularly 1n the high-school educated portion of the examinees. 3. We believe the study measured the true health status of both the Ranch Hand and comparison Individuals to the extent medical science permits. The study repeatedly demonstrated the effects of classical risk factors (smoking, age, alc6hbl, etc.) to the same extent in both groups. Such demonstration of these effects reassures us that we would also have demonstrated a significant herbicide or contaminant exposure effect 1f it were present at this time. We consider the study reassuring 1n that no cases of conditions previously attributed to exposure were found, there was no increased Incidence of major clinical health problems in Ranch Handers, and both the Ranch Handers and comparisons appeared in good general health for their age. �MORBIDITY REPORT TYPE CANCER BY PERCENT RH (1045) ORIGINAL COMP (773) TOTAL COMP (1194) SKIN CANCER 3.3% 1.4% 2.1% SYSTEMIC CANCER 1.2% 1.0% 0.9% COLON 0.0% 0.3% 0.3% PANCREAS 0.0% 0.1% 0.1% GENITO-URINARY 0.6% 0.1% 0.3% OROPHARYNGEAL 0.4% 0.3% 0.2% �RANCH HAND REPORTED LIMITED SKIN BIRTH DEFECTS BIRTHMARKS CSi 1 • 5 SKIN DISCOLORATION 1 YELLOW COLOR, GONE IN ONE WEEK 1 SKIN TAGS 1 TWO NIPPLES ON BREAST 1 �NEONATAL DEATHS (UNVERIFIED SELF REPORTS; MEDICAL RECORDS, DEATH CERTIFICATES PENDING) RATE/1000 BEFORE RVN AFTER RVN RANCH HAND 13.4 16.8 COMPARISON 16.0 3. ATTACHMENT 3 �MORBIDITY REPORT BLOOD TESTS MEAN VALUES RH C LIVER SGOT ALK PHOS SGPT GGTP LDH CHOL 33.0 7.7 20.3 40.1 142.1 212.2 •STATISTICALLY SIGNIFICANT DIFFERENCE 33.1 7.5 20.5 39.3 141.7 216.6 NORMAL RANGE < 41 < 9.7 < 45 < 85 <200 <240 � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 060 Folder The folder containing the original item. 1600 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Typescript: Draft, Ranch Hand Morbidity Report Subject The topic of the resource Air Force Health Study morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/5605e5a6a27b16b9fca7221018cc7904.pdf d27c233dcd99c635eb516f8ae3f924c6 PDF Text Text Item ID Number oieo4 Author Corporate Author Report/Article Title Typescript: Project Ranch Hand II Journal/Book Title Year 000 ° Month/Day Color rj Number of linages 35 DOSCrlptOU NOtOS Appears to be printouts from a slide presentation. Wednesday, May 23, 2001 Page 1605 of 1608 �PROJECT RANCH HAND II A WHITE HOUSE DIRECTED EPIDEMIOLOGIC INVESTIGATION OF POSSIBLE ADVERSE HEALTH EFFECTS ATTRIBUTABLE TO HERBICIDE EXPOSURE AMONG OPERATION RANCH HAND PERSONNEL �STUDY DESIGN IDENTIFY EXPOSED POPULATION SELECT COMPARISON POPULATION DETERMINE BASELINE HEALTH STATUS OF THE TWO GROUPS • MORTALITY » DISEASE OR ABNORMALITY (MORBIDITY) COMPARE FINDINGS STATISTICALLY TO DELINEATE POSSIBLE HERBICIDE EFFECTS ACCOMPLISH FOLLOW UP STUDIES OF POPULATIONS �STUDY POPULATIONS ALL RANCH HAND MEMBERS WHO SERVED IN VIETNAM DURING 1962-1971 IDENTIFIED (1,200+) COMPARISON GROUP COMPOSED OF CARGO FLIGHT CREW MEMBERS AND SUPPORT PERSONNEL IN SOUTHEAST ASIA DURING SAME PERIOD BUT NOT OCCUPATIONALLY EXPOSED TO HERBICIDE (19,000+) EIGHT COMPARISON SUBJECTS MATCHED TO EACH RANCH HANDER BY: • JOB CATEGORY RANCH HAND _ DAPr RACE INDIVIDUAL • • AGE COMPARISON INDIVIDUALS , JJI 1:8 EXPOSURE INDEX DEVELOPED FOR RANCH HAND POPULATION �MORTALITY PORTION FIVE INDIVIDUALS RANDOMLY ORDERED IN EACH COMPARISON SET MORTALITY EXPERIENCE OF THE FIVE COMPARISON SUBJECTS COMPARED TO A RANCH HANDER RANCH HAND INDIVIDUAL COMPARISON INDIVIDUALS RANDOMLY ORDERED MORTALITY COMPARISONS 1-5 t �MORBIDITY PORTION FIRST INDIVIDUAL IN 5 MEMBER MORTALITY SET USED FDR COMPARISON STUDY • IF COMPARISON SUBJECT DECEASED OR NON-COMPLIANT, NEXT COMPLIANT SUBJECT SELECTED LIVING RANCH HAND INDIVIDUAL COMPARISON INDIVIDUALS RANDOMLY ORDERED MORTALITY COMPARISONS I 4- 1:1 - DEAD + UNWILLING (NONCOMPLIANT) I — I * I I I V A * I I J ** VOLUNTEER * * REPLACEMENT CANDIDATES �MATCHING PROCEDURES INITIAL MATCH COMPARISON INDIVIDUALS RANCH HAND INDIVIDUAL 1 :8 I I I I I I I COMPARISON INDIVIDUALS RANDOMLY ORDERED MORTALITY COMPARISON MATCH FOR MORTALITY PORTION RANCH HAND INDIVIDUAL 1 :5 lllliDOD COMPARISON INDIVIDUALS RANDOMLY ORDERED MORTALITY COMPARISONS MATCH FOR MORBIDITY PORTION LIVING RANCH HAND INDIVIDUAL + - * ** * ~ UNWILLING * * VOLUNTEER REPLACEMENT CANDIDATES �QUESTIONNAIRES DEVELOPED AND ADMINISTERED UNDER CONTRACT • INCORPORATED DATA FROM EXTENSIVE LITERATURE REVIEW AND VETERANS' COMPLAINTS AND CONCERNS QUESTIONNAIRES DEVELOPED FOR STUDY SUBJECTS, SPOUSES, AND NEXT OF KIN ADMINISTERED FACE-TO-FACE IN HOME BY OVER 80 ESPECIALLY SELECTED AND TRAINED PERSONNEL REQUIRED UP TO 3 HOURS TO COMPLETE PROVIDED OVER 300 ITEMS FROM EACH STUDY SUBJECT FOR ANALYSES OVER 5000 QUESTIONNAIRES ADMINISTERED �PHYSICAL EXAMINATION ACCOMPLISHED BY CONTRACT • CIVILIAN ORGANIZATION OF NATIONAL STATURE • EXAMINERS UNAWARE OF EXPOSURE HISTORY PARTICULAR EMPHASIS ON DERMATOLOGIC, NEUROPSYCHIATRIC, HEPATIC, IMMUNOLOGIC, REPRODUCTIVE, AND NEOPLASTIC ASPECTS • BASED ON EXTENSIVE SCIENTIFIC LITERATURE REVIEW EXAMINATION REQUIRED THREE DAYS PROVIDED OVER 1100 ITEMS FROM EACH STUDY SUBJECT FOR ANALYSES OVER 2200 EXAMINATIONS ACCOMPLISHED �MORBIDITY PORTION PARTICIPATION RANCH HAND 1206 POTENTIAL SUBJECTS COMPARISON ORIGINAL REPLACEMENTS QUESTIONNAIRE 1174 (97%) 956 (93%) 576 (90%) EXAMINATION 1045 (87%) 774 (76%) 450 (71%) �BASELINE MORTALITY PORTION RESULTS RELEASED JUNE 1983 EVALUATED 50 RANCH HAND AND 250 COMPARISON SUBJECTS WHO DIED FROM NONCOMBAT CAUSES SMALL NUMBER EMPHASIZES PRELIMINARY NATURE OF RESULTS RESULTS • MORTALITY EXPERIENCE NEARLY IDENTICAL IN RANCH HAND AND COMPARISON GROUPS • CAUSE SPECIFIC ANALYSES NOT STATISTICALLY DIFFERENT �MORBIDITY PORTION RESULTS LACK OF CLEAR CUT HALLMARKS REQUIRED COMPREHENSIVE EVALUATION OF NUMEROUS SYSTEMS • GENERAL PHYSICAL HEALTH • MALIGNANCY • FERTILITY/REPRODUCTIVE • NEUROLOGY • PSYCHOLOGY • HEPATIC • DERMATOLOGY • CARDIOVASCULAR • IMMUNOLOGY • HEMATOLOGY • PULMONARY • RENAL • ENDOCRINE • INDIVIDUAL HEALTH �GENERAL PHYSICAL HEALTH SIGNIFICANT GROUP DIFFERENCE FOUND IN SELF-PERCEPTION OF HEALTH WITH MORE RANCH HANDERS PERCEIVING THEMSELVES TO BE IN FAIR OR POOR HEALTH • NOT SUPPORTED BY EXPOSURE INDEX ANALYSES BORDERLINE SIGNIFICANT GROUP DIFFERENCE IN THE EXAMINER'S ASSESSMENT OF ILLNESS OR DISTRESS (RH=>C) • NOT SUPPORTED BY EXPOSURE INDEX ANALYSES NO SIGNIFICANT GROUP DIFFERENCES IN PERCENT BODY FAT OR HEMATOCRIT SIGNIFICANTLY MORE INDIVIDUALS WITH SEDIMENTATION RATE ABNORMALITIES IN THE YOUNGER COMPARISONS THAN IN THE RANCH HANDERS • NO OVERALL GROUP DIFFERENCES OBSERVED �MALIGNANCY NO SIGNIFICANT DIFFERENCE IN THE OCCURRENCE OF "SYSTEMIC" CANCER NO SOFT TISSUE SARCOMA FOUND RANCH HANDERS • ONE CASE FOUND IN COMPARISONS SIGNIFICANTLY MORE POST-SEA SKIN CANCER IN THE RANCH HAND GROUP THAN IN THE ORIGINAL COMPARISONS WHO COMPLETED PHYSICAL EXAMINATION • DIFFERENCE ONLY BORDERLINE WITH TOTAL COMPARISON GROUP • ANALYSES NOT FULLY ADJUSTED FOR SUN EXPOSURE (GEOGRAPHIC AREA OF RESIDENCE) • NON-MELANOMA CELL TYPE PREDOMINATES • FACE, HEAD AND NECK PREDOMINANT DISTRIBUTION �MALIGNANCY (CONTD) ANALYSES DEMONSTRATED ASSOCIATION BETWEEN SMOKING AND "SYSTEMIC" CANCER AND INDUSTRIAL CHEMICAL EXPOSURE AND SKIN CANCER IN BOTH GROUPS SLIGHTLY MORE GENITOURINARY CANCER (6 TO 3), MORE OROPHARYNGEAL CANCER (4 TO 2), AND LESS GASTROINTESTINAL CANCER (0 TO 5) IN RANCH HAND GROUP, BUT DIFFERENCE NOT SIGNIFICANT EXPOSURE INDEX ANALYSES REVEALED NO STATISTICALLY SIGNIFICANT OR SUGGESTIVE ASSOCIATIONS BETWEEN HERBICIDE EXPOSURE AND EITHER SKIN OR "SYSTEMIC" MALIGNANCY �VERIFIED SKIN AND SYSTEMIC CANCER • RANCH HAND ALL COMPARISONS NUMBER (%) SKIN CANCER ORIGINAL COMPARISONS NUMBER (%) NUMBER (%) 35 (3.3) 11 (1.4) 25 (2.1) SYSTEMIC SITE LIP, MOUTH, THROAT 4 2 2 DIGESTIVE SYSTEM 0 4 5 RESPIRATORY 3 1 2 GENITOURINARY BLOOD, LYMPH 6 2 3 0 0 1 OTHER 1 1 1 TOTAL 14 (1.3) 10 (1.3) 14 (1.2) �FERTILITY/REPRODUCTIVE OVER 7000 CONCEPTIONS EVALUATED NO SIGNIFICANT GROUP DIFFERENCES • SPERM COUNT, PERCENT ABNORMAL • FERTILITY/INFERTILITY • MISCARRIAGE, STILLBIRTH, LIVE BIRTH SIGNIFICANT GROUP DIFFERENCE (RH > C) • REPORTED BIRTH DEFECTS - NO SIGNIFICANT DIFFERENCE WHEN SKIN ANOMALIES (BIRTH MARKS, BLEMISHES, etc.) REMOVED • REPORTED PHYSICAL HANDICAPS • REPORTED NEONATAL DEATH MEDICAL RECORD VERIFICATION OF REPORTED EVENTS UNDER WAY NO CONSISTENT PATTERN WITH INCREASING EXPOSURE �REPORTED POST-SEA DEFECTS BY SEVERITY RANCH HANDERS (RH) VERSUS ORIGINALS (0) RH 0 SEVERE Y N 32 885 18 726 P 0.20 MODERATE Y N 22 895 20 724 P 0.71 LIMITED Y N 26 891 io P 0.04 734 RANCH HANDERS (RH) VERSUS ALL COMPARISONS (A) SEVERE Y N 32 885 34 1,275 P 0.22 MODERATE Y N 22 895 34 1,275 P 0.77 LIMITED Y N 26 891 18 1,291 P 0.01 �INDIVIDUAL HEALTH (CONT'D) NO SIGNIFICANT DIFFERENCE IN DISTRIBUTION OF ILL INDIVIDUALS GROUP DIFFERENCES ONLY FOR • SKIN CANCERS, NOT CORRECTED FOR SUN EXPOSURE • PULMONARY FUNCTION (MORE ABNORMALITIES IN COMPARISON) BOTH GROUPS IN GOOD HEALTH FOR AGE �NEUROLOGY NO SIGNIFICANT GROUP DIFFERENCES IN CRANIAL NERVE FUNCTION NO SIGNIFICANT GROUP DIFFERENCES IN TESTS OF PERIPHERAL NERVES EXCEPT FOR A BORDERLINE DIFFERENCE IN BABINSKI REFLEX • EFFECTS OF ALCOHOL USE AND ABNORMAL GLUCOSE METABOLISM DEMONSTRATED IN BOTH GROUPS NO GROUP DIFFERENCES IN CENTRAL FUNCTION (TREMOR, COORDINATION, ROMBERG, GAIT) NO SIGNIFICANT GROUP DIFFERENCES IN NERVE CONDUCTION VELOCITY NO DOSE RESPONSE EFFECT OBSERVED FOR PERIPHERAL NERVES, CRANIAL NERVES, CONDUCTION VELOCITY, OR CENTRAL FUNCTION �PSYCHOLOGY COMPREHENSIVE GROUP OF VALIDATED TESTS USED ANALYSES STRATIFIED BY EDUCATION • REFLECTED KNOWN SUBSTANTIAL EFFECT OF EDUCATION ON PSYCHOLOGICAL TESTING SUBJECTIVE MEASURES SHOWED SIGNIFICANT GROUP DIFFERENCES FOR HIGH SCHOOL EDUCATED PERSONNEL (QUESTIONNAIRE, CORNELL INDEX, MMPI) • NOT OBSERVED IN THE COLLEGE EDUCATED INDIVIDUALS NO DIFFERENCES IN IQ OR PERFORMANCE TESTING NO DOSE RESPONSE EFFECTS OBSERVED �HEPATIC RANCH HANGERS REPORTED MORE LIVER SYMPTOMS IN PAST • NO CLINICAL EVIDENCE OF SIGNIFICANT GROUP DIFFERENCES AT EXAMINATION OVERALL BIOCHEMICAL GROUP DIFFERENCES: GGPT (RH=>C), LDH (RH=>«, CHOLESTEROL (RH<C) NO OVERALL GROUP DIFFERENCES: SGOT, SGPT, ALK. PHOS. T. BILI, D. BILI, TRIGLYCERIDES, UROPORPHYRIN, CQPRQPQRPHYRIN, D-ALA NO PORPHYRIA CUTANEA TARDA DETECTED NO DOSE RESPONSE EFFECT OBSERVED �DERMATOLOGY QUESTIONNAIRE RESPONSES SHOWED NO GROUP DIFFERENCES FOR • OCCURRENCE OF PAST ACNE • OCCURRENCE OF PAST ACNE RELATIVE TO INDIVIDUAL'S SEA TOUR • SEVERITY OR DURATION OF PAST ACNE • ANATOMIC LOCATION OF PAST ACNE SUGGESTING CHLORACNE �DERMATOLOGY (CONT'D) PHYSICAL EXAMINATION SHOWED • NO CLINICAL DIAGNOSES OF CHLORACNE • NO POSITIVE BIOPSIES FOR CHLORACNE • NO GROUP DIFFERENCES IN THE PREVALENCE OF THE 5 MOST COMMON DERMATOLOGIC DIAGNOSES EXPOSURE INDEX ANALYSES WERE NEGATIVE �CARDIOVASCULAR PHYSICAL EXAMINATION SHOWED SIMILAR GROUP FINDINGS FOR • SYSTOLIC, DIASTOLIC BLOOD PRESSURES • ABNORMAL ECG'S • COMPARISON OF PAST ECG'S TO THE EXAM ECG • HEART SOUNDS • FUNDUSCOPIC ABNORMALITIES • CAROTID BRUITS NO PREMATURE HEART DISEASE IN THE RANCH HANDERS �CARDIOVASCULAR (CONT'D) SOME PERIPHERAL PULSES DECREASED IN RANCH HANDERS • SIGNIFICANCE UNCERTAIN ANALYSES DEMONSTRATED EFFECTS OF • AGE • SMOKING, PAST ANDPRESENT HERBICIDE EXPOSURE ANALYSES WERE ESSENTIALLY NEGATIVE FOR CARDIOVASCULAR ABNORMALITIES �IMMUNOLOGY NO GROUP DIFFERENCES SMOKING, ALCOHOL, AGE EFFECTS DEMONSTRATED IN BOTH GROUPS HEMATOLOGY NO DIFFERENCE IN CLINICALLY SIGNIFICANT BLOOD ABNORMALITIES DIFFERENCES IN CELL HEMOGLOBIN CONCENTRATION PRESENT • MOST VALUES STILL WITHIN NORMAL LIMITS �PULMONARY NO GROUP DIFFERENCES IN REPORTED DISEASE, FORCED VITAL CAPACITY, FORCED EXPIRATORY VOLUME (1 SECOND) OR FEV^FVC RATIO NO CONSISTENT ASSOCIATIONS BETWEEN PULMONARY FUNCTION AND EXPOSURE LEVELS EXPECTED EFFECTS OF SMOKING HISTORY SEEN IN BOTH GROUPS NO INDICATION THAT CURRENT PULMONARY FUNCTION WAS AFFECTED BY EXPOSURE TO HERBICIDES �RENAL NO SIGNIFICANT DIFFERENCES IN TESTS OF RENAL FUNCTION NO CLINICAL EVIDENCE OF HERBICIDE EFFECT EXPOSURE INDEX ANALYSES ESSENTIALLY NEGATIVE �ENDOCRINE SOME THYROID HORMONE GROUP DIFFERENCES • MOST VALUES STILL WITHIN NORMAL LIMITS TESTOSTERONE LEVELS INCREASED IN RANCH HANDERS • MOST VALUES STILL WITHIN NORMAL LIMITS BLOOD GLUCOSE LEVELS NOT SIGNIFICANTLY DIFFERENT �INDIVIDUAL HEALTH BOTH GROUPS SIMILAR IN • RISK TAKING ACTIVITIES • RELIGIOUS PREFERENCE • EDUCATION • PERSONAL AND FAMILY INCOME • MILITARY STATUS • REPORTED INJURIES AND POISONINGS • CURRENT AND PAST USE OF ALCOHOL • PAST HISTORY OF CIGARETTE, PIPE, CIGAR, USAGE MORE RANCH HANDERS CURRENTLY SMOKE THAN COMPARISONS �INTERPRETIVE COMPLEXITIES STATISTICAL ASSOCIATIONS DO NOT MEAN CAUSATION NO ANALOGOUS HUMAN STUDIES MANY STATISTICALLY SIGNIFICANT RANCH HAND-COMPARISON GROUP DIFFERENCES ARE NOT OF CLINICAL RELEVANCE EXPOSURE INDEX ANALYSES ARE UNREFINED POSITIVE AND NEGATIVE FINDINGS REQUIRE FOLLOW-UP �SUMMARY BASELINE MORBIDITY PORTION DID NDT DEMONSTRATE DEFINITIVE CLINICAL END POINTS CONCLUSIVELY ATTRIBUTABLE TO HERBICIDE EXPOSURE NO STS, PCT, CHLORACNE DIAGNOSED IN RANCH HANDERS DID FIND A NUMBER OF CLINICAL AND SUBCUNICAL DIFFERENCES • DEFINING SIGNIFICANCE OF SOME DEPENDENT ON ANALYSES OF DATA NOT YET COLLECTED (SUN EXPOSURE, BIRTH RECORDS) • MOST VALUES STILL WITHIN NORMAL RANGES SCHEDULED FOLLOW UP EXAMINATIONS WILL PROVIDE ESSENTIAL DATA NECESSARY TO DEFINE BOTH FALSE POSITIVES AND ANY FALSE NEGATIVES �PROJECT RANCH HAND II BASELINE MORTALITY AND MORBIDITY STUDIES ACCOMPLISHED RESULTS CONFIRM EFFECTIVENESS OF PROTOCOL DESIGN AND IMPLEMENTATION INITIAL FINDINGS UNDERSCORE NEED FOR FOLLOW UP OVER TIME EXTENSIVE DATA BASE ALREADY OBTAINED WILL ALLOW CONTINUING DATA ANALYSES �CONCLUSIONS STUDY MEASURED TRUE HEALTH STATUS TO MAXIMUM EXTENT POSSIBLE ALL SIGNIFICANT FINDINGS ARE BEING FOLLOWED UP INSUFFICIENT EVIDENCE TO SUPPORT HERBICIDE CAUSALITY AT THIS TIME FINDINGS TO DATE SHOULD BE REASSURING TO RANCH HANDERS • NO CHLORACNE MEANS LOW EXPOSURE VERSUS CHEMICAL WORKER POPULATIONS • NO MAJOR CLINICAL HEALTH PROBLEMS • OVERALL GOOD GENERAL HEALTH FOR AGE �STUDY COMPLEXITIES ENORMOUS DATA BASE (4 MILLION ITEMS) VALIDITY OF SELF REPORTING IN ABSENCE OF COMPLETE MEDICAL RECORD VERIFICATION EFFECTS OF MULTIPLE RISK FACTORS (AGE, SMOKING, etc.) POTENTIAL BIASES STUDY SCHEDULE � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 060 Folder The folder containing the original item. 1604 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Typescript: Project Ranch Hand II Subject The topic of the resource Air Force Health Study mortality trends morbidity trends ao_seriesIII https://www.nal.usda.gov/exhibits/speccoll/files/original/194b80b0af7aec76903fc7d330e9372f.pdf 338f12b2dd882e0f62b0909e9171fe7d PDF Text Text Item ID Number 01537 Author Albanese, Richard A. United States Air Force School of Aerospace Medicine, ROpOrt/ArtiClB TltlB Un ted ' States Air Force Personnel and Exposure to Herbicide Orange, Interim Report for Period March 1984-February 1988 Journal/Book Title Yoar 1988 Month/Day February Color a Number of Images 38 Descripton Notes Wednesday, May 23, 2001 Page 1588 of 1608 �USAFSAM-TR-88-3 UNITED STATES AIR FORCE PERSONNEL AND EXPOSURE TO HERBICIDE ORANGE Richard A. Albanese, M.D. February 1988 interim Report for Period March 1984 - February 1988 I Approved for public release; distribution Is unlimited. I USAF SCHOOL OF AEROSPACE MEDICINE Human Systems Division (AFSC) Brooks Air Force Base, TX 78235-5301 �NOTICES This interim report was submitted by personnel of the Radiation Analysis Branch, Radiation Sciences-.Division, USAF School of Aerospace Medicine, Human Systems Division, AFSC, Brooks Air Force Base, Texas, under job order SUPTXXRH. When Government drawings, specifications, or other data are used for any purpose other than in connection with a definitely Government-related procurement, the United States Government incurs no responsibility or any obligation whatsoever. The fact that the Government may have formulated or in any way supplied the said drawings, specifications, or other data, is not to be regarded by implication, or otherwise in any manner construed, as licensing the holder, or any other person or corporation; or as conveying any rights or permission to manufacture, use, or sell any patented invention that may in any way be related thereto. The Office of Public Affairs has reviewed this report, and it Is releasable to the National Technical Information Service, where it will be available to the general public, Including foreign nationals.. This report has been reviewed and is approved for publication. 777. xS RICHARD A. ALBANESE, M.D. ' Project Scientist .DAVIS, Colonel, USAF, MC 0HN C. MITCHELL, B.S. Chief, Radiation Sciences Division �IflCATIQN Qf THIS PAGE form Approved OMi NO. 0704-01 it REPORT DOCUMENTATION PAGE 1b. RESTRiaiVE MARKINGS la. REPORT SECURITY CLASSIFICATION Unclassified 2«. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/AVAILABILITY OF REPORT 2b. DECLASSIFICATION/DOWNGRADING SCHEDULE Approved for public release; distribution is unlimited. 4. PERFORMING ORGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION REPORT NUMBER(S) USAFSAM-TR-88-3 6*. NAME OF PERFORMING ORGANIZATION USAF School of Aerospace Medicine 66. OFFICE SYMBOL Of app/fcabfej 7«. NAME OF MONITORING ORGANIZATION USAFSAM/RZM 6c ADDRESS (Oty, Statt, and ZlfCodt) 7b. ADDRESS (Oty, Statt. tnd ZlPCotto) Human Systems Division (AFSC) Brooks Air Force Base, TX 78235-5301 Ba. NAME OF FUNDING/SPONSORING ORGANIZATION USAF School of Aerospace 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER 8b. OFFICE SYMBOL Of appflcab/t) USAFSAM/RZM Be ADDRESS (Oty, State, and HP Cot*) 10. SOURCE OF FUNDING NUMBERS Human Systems Division (AFSC) Brooks Air Force Base, TX 78235-5301 PROGRAM ELEMENT NO. PROJECT NO. 65306F WORK UNIT ACCESSION NO. TASK NO. SUPT XX RH 11. TITLE flnc/wde Security Oaoiffcattoft) United States Air Force Personnel and Exposure to Herbicide Orange 12. PERSONAL AUTHOR(S) Albanese, Richard A. 13a. TYPE OF REPORT 13b. TIME COVERED FROM 3/84 TO. Interim 2/88 14. DATE QF. T (Y«»r. Month, (toy) 15. PAGE COUNT 16. SUPPLEMENTARY NOTATION 7. FIELD 06 COSATI CODES GROUP SUB-GROUP 05 05 It. SUBJECT TERMS (Continue on reverie If neceoary and Mtntfty by block number) Epidemiologic Investigation Dioxin Phenoxy Herbicides Ranch Hand Herbicide Orange Air Force Health Study 9. ABSTRACT (Cont/nue on reverie if neceoary and Me/Wry by block number) The United States Air Force is conducting an epidemiological study, called the Air Force Health Study (AFHS), to determine whether or not military personnel associated with herbicide spraying during the Vietnam War have experienced any adverse health effects. This report reviews salient findings of the AFHS first morbidity report published in 1984, and presents new work by examining relationships between AFHS findings and the results of laboratory toxicological studies and other epideraiological studies addressing dioxin. Eleven clinical areas have been emphasized based on a toxicological profile developed from the literature and availability of data in the AFHS: weight loss, neoplasia, birth defects, neurological changes, psychological changes, hepatotoxicity, chloracne, cardiovascular changes, immunological deficits, endocrine changes, and mortality. In six of these eleven clinical areas, statistically significant group differences occurred, and in five of these six instances the group differences were in the direction of expected dioxin o. DISTRIBUTION/AVAILABILITY OF ABSTRACT 38 UNCLASSIFIED/UNLIMITED D SAME AS RPT. 2a. NAME OF RESPONSIBLE INDIVIDUAL Richard A. Albanese, M . D . DD Form 1473, JUN 86 2.1. ABSTRACT SECURITY CLASSIFICATION Unclassified Q OTIC USERS 22b. TELEPHONE f/nc/ude Area Code) 22c. OFFICE SYMBOL USAFSAM/RZM (512)536-3884 Prtviouj edition trt obsolttt. i SECURITY CLASSIFICATION OF THIS PAGE UNCLASSIFIED �UNCLASSIFIED SECURITY C L A S S i p l C A T i O N 0<= THIS PAGE 19. ABSTRACT (Continued) effects. Dioxin cannot be confidently identified as the causative agent of these findings at this time because of several reasons, including the absence of correlations with an \ exposure index and the incomplete clinical picture. However, dioxin is not exonerated as a causative agent because of the directionality of the observed group differences and the preliminary nature of the exposure index used in the AFHS first morbidity report. ijL UNCLASSIFIED SECURITY Ci-ASSlPICATICN C* *"iS c a G c �UNITED STATES AIR FORCE PERSONNEL AND EXPOSURE TO HERBICIDE ORANGE INTRODUCTION The United States Air Force is conducting an epidemiological study to determine whether or not military personnel associated with herbicide spraying during the Vietnam War have experienced any adverse health effects. The Air Force program responsible for the herbicide spraying was code named Operation Ranch Hand, and the personnel involved in the spray missions are termed Ranch Hands. The current epidemiological study is called the Air Force Health Study (AFHS). This report reviews salient findings of the AFHS first morbidity report.1 Building on prior reports, this article presents new work by examining relationships between AFHS findings and the results of laboratory toxicological studies and other epideraiological studies addressing dioxin. This report attempts to assess the extent to which these initial AFHS findings are compatible with toxic effects of 2,3,7,8-tetrachlorodibenzo para dioxin (TCDD) as known from animal experiments and, to a lesser extent, from observation in humans. In the preparation of this report, more than 100 dioxin-related published articles were studied, and a blomedical portrait of dioxin effects emerged. This portrait will change as research proceeds; nevertheless, the current literature is sufficiently mature to permit comparison with AFHS findings. METHODS The Protocol Investigators at the USAF School of Aerospace Medicine developed a comprehensive study protocol to govern the AFHS. The protocol underwent extensive peer review by military and civilian agencies, including: The University of Texas School of Public Health, an Air Force Scientific Advisory Board, the Armed Forces Epidemiological Board, the National Academy of Sciences, and the White House appointed Agent Orange Working Group. This last organization continues to act in an advisory role. Recommendations by these groups were incorporated into the protocol. �Study Design The study design is a matched cohort design in a nonconcurrent prospective setting. The study addresses mortality and morbidity and includes longterm follow-up activities. A detailed population ascertainment process identified 1,278 Ranch Hand personnel who served in Vietnam during the period 1962 through 1971, when the spraying operation was active. A comparison group was formed by identifying all individuals assigned to Air Force organizational units with a mission of flying cargo to, from, and in Vietnam during the same period. A computerized nearest neighbor selection process was used to match up to 10 comparison individuals to each Ranch Hand. This matching was done by job category, race, and age. The initial comparison group erroneously contained some individuals who did not in fact have any Southeast Asia experience. These individuals (18J) were removed from the comparison population after detailed hand record review, leaving an average of approximately eight comparison individuals matched to each Ranch Hand. The comparison individuals matched to each Ranch Hand were listed in random order within each set. The first five comparisons were included in the mortality analyses giving these studies a 1:5 design. For each living Ranch Hand, the first living member of his randomized comparison set was selected for participation in a morbidity study consisting of an in-home interview and a comprehensive physical examination. If the matched comparison subject declined to participate or subsequently withdrew from the morbidity study, that individual was replaced by the next living comparison subject from the randomized set who was willing to participate. Follow-up studies are an important aspect of the AFHS. .The follow-up studies consist of mortality and morbidity components. Each Ranch Hand and his set of comparisons will be the subject of mortality evaluations for the next 20 years. In addition, follow-up questionnaires and physical examinations are being offered to participants in 1985, 1987, 1992, 1997, and 2002 in order to bracket the latency periods associated with possible attributable disease. Inference Concerning Herbicide Causation In an experiment, members from a single population are randomly assigned to either a treatment (exposed) or control group. In the setting of such completely randomized designs, statistically significant differences between the treated and control group can often be reliably'ascribed to the effect of the exposure. The AFHS is the study of an unplanned environmental exposure and thus does not follow the above experimental design, the study has a nonrandomized design and is an observational Study. In such studies, while the comparison group is chosen to be similar to the exposed group with respect to r as many qualities as possible, except exposure status, in the absence of randomization, group differences or the lack thereof cannot be interpreted .. solely in terms of exposure. For example,.in the AFHS the exposed group was stationed in the Republic of Vietnam itself, while most of the comparison �group was quartered in surrounding areas such as Okinawa and Japan. Comparison aircrew members periodically flew into Vietnam while comparison ground support personnel, predominantly enlisted, remained in non-combat areas. Ranch Hand Vietnam tour length was approximately 1 year while comparison tour length was 3 years. These differences and others may or may not influence health and longevity. Thus, group differences or the lack thereof cannot unambiguously be ascribed to herbicide exposure. This emphasizes the importance of relating study results to other studies to see whether common patterns of effect emerge. Another approach to inference is the use of an exposure measure or index. If one knew exactly how much herbicide each Ranch Hand was exposed to, highly exposed individuals could be contrasted with less exposed individuals within the Ranch Hand group and an estimate of herbicide effect could be constructed. However, once again, since randomization was not employed in the dose assignment, estimates of herbicide effect must be viewed with great care due to the possibility of confounding factors. For example, it could happen that higher exposures occurred for a variety of reasons in the lower socio-economic strata (lower ranks) of the Ranch Hand cohort. Industrial hygiene data concerning herbicide exposure were not collected during the Vietnam era. In any case, however, the use of an exposure index provides another view of exposure effects which can augment interpretation of group differences. The exposure index used in this report relates to the TCDD-containing herbicides: Herbicide Orange, Herbicide Purple, Herbicide Pink, and Herbicide Green. Archived samples of Herbicide Purple had a mean TCDD concentration of approximately 33 ppm, and archived samples of Herbicide Orange had a mean concentration of 2 ppm. Herbicides Pink and Green contained twice the 2,K,5-T of Herbicide Purple and, therefore, have been estimated to contain TCDD at a concentration of approximately 66 ppm. Using mission records, it was possible to determine the amount of each herbicide sprayed each month in Vietnam as well as the number of Ranch Hands in each job category who were involved in spraying that month for the period 1962 through 1971. Tour data also allowed determination of the months each individual was involved in the Ranch Hand operation. Using these data, an exposure index was developed for each Ranch Hand. The exposure index is directly proportional to the number of gallons of herbicide sprayed in Vietnam during the individual's tour, where potential exposure to the higher TCDD.containing herbicides (Purple, Pink, Green) has been properly scaled according to dioxin concentration to place them on the same basis as Herbicide Orange. Also, the exposure index is inversely proportional to the number of airmen assigned to the specific subject's job category during his tour. From the description just given, it should be clear that the current 1 Ranch Hand exposure index is an estimate only, as it applies theater-wide spraying to a single individual, and, since it assumes that all individuals in a job category were equally exposed. Also, the degree to which this calculated index is associated with actual body burden of TCDD is unknown. In short, the absence of a positive association between the index and health outcomes cannot be taken as confirming a lack of herbicide effect, nor can the presence of an association be interpreted as an unambiguous herbicide effect without consideration of possible confounding factors. �Job category matching in the AFHS used five categories: (a) officerpilot, (b) officer-navigator, (c) officer-other, (d) enlisted-flying, and (e) enlisted-ground. Exposure-index analyses used three occupational categories: all officers were combined into one category called "officer" due to the fact that navigators and pilots, while having different jobs, were believed to be exposed in the same manner. For each exposure occupational group (officer, enlisted-flying, enlisted-ground), the calculated exposure index was trichotomized into three levels: low exposure, medium exposure, and high exposure. Since the mode of exposure was judged to be different in each occupational group, statistical analyses with the exposure index were occupational group specific. Questionnaire and Physical Examination The AFHS uses a broad medical history and physical examination. The medical history was collected by an extensive in-home questionnaire. ^ The purpose of the extensive questionnaire was to collect data that could be analyzed for the subjective presence of adverse health effects that might be related to herbicide exposure. Jn addition t'o the study participants, the questionnaire contractor was also required to interview the participants1 current and former wives, as well as the first-order next-of-kin of deceased individuals to obtain morbidity data as completely as possible. Physical examinations were performed at a single location by a contractor. All examiners evaluated the participants without knowledge of their exposure status. The number of examiners and the turnover of staff members were kept to a minimum to limit between-examiner variability. All laboratory tests were subjected to rigid standards of quality control, and laboratory and physical examination data were measured on a continuous scale whenever possible to improve statistical power in the analysis. A general summary of the major components of the examination is presented in Table 1, and the laboratory procedures conducted on each subject are listed in Table 2. �TABLE T. AFHS PHYSICAL EXAMINATION General Physical Examination Neurological Examination Dermatological Examination Electrocardiogram Pulmonary Function Study Chest X-ray Nerve Conduction Velocities Psychological Evaluation: Minnesota Multiphasic Personality Inventory (MMPI) Cornell Wechsler Memory Scale I Wechsler Adult Intelligence Scale (WAIS) Wide Range Achievement Test (WRAT) Halstead-Reitan Neuropsychological Battery �TABLE 2. LABORATORY PROCEDURES Chemistry Panel: Blood Urea Nitrogen (BUN) Creatinine Cholesterol High-Density Lipoprotein Triglyceride Total Bilirubin Direct Bilirubin Alkaline Phosphosphatase Glucose (Fasting and 2 hour) Cortisol (Fasting and 2 hour) Serum Glutamic-Oxaloacetic Transaminase (SCOT) Serum Glutamic-Pyruvic Transaminase (SGPT) Gamma Glutamyl Transpeptidase (GGTP) Lactic Acid Dehydrogenase (LDH) Creatine Phosphokinase (CPK) Blood Alcohol Hormone Assay: Luteinizing Hormone (LH) Follicle Stimulating Hormone (FSH) Testosterone Triiodothyronine (T3) Uptake Tetraiodothyronine (T4) Free Thyroxine Index (FTI) Hematology Panel: Erythrocyte Sedimentation Rate Prothrombin Time Serological Test for Syphilis (RPR) White Blood Cell Count (with 10,000 cell differential) Red Blood Cell Count Hemoglobin Hematocrit Red Cell Indices Platelet Count Urinalysis: 21-Hour Urine Volume Delta Amino Levulinic Acid Coproporphyrins Uroporphyrins Porphobilinogen Creatinine Semen Analysis: Volume Count Abnormal Forms Hepatitis B Testing: Surface Antigen Antibody to Surface Antigen . Core Antibody �RESULTS Morbidity The results of the analysis of the baseline morbidity data were released in February 198H. Of all Ranch Hand and comparison individuals who were selected for the questionnaire and physical examination phases of this study, 99.5/1 were contacted, eliminating the major concern of selection bias. One thousand one hundred and seventy-four (97J) of the Ranch Hand group and 956 (93/t) of the originally selected comparison group participated in the questionnaire portion of the Morbidity Study. An additional 576 comparison subjects were interviewed as substitute subjects, bringing the total number of comparison participants to 1,532. Substitute comparisons were selected to replace comparisons selected erroneously and to replace noncompliant comparisons. Two thousand seven hundred eight current and former wives of the study participants were also interviewed. One thousand and forty-five (87J) of the Ranch Hand group participated in the physical examination, and 773 (7650 of the originally selected comparison subjects participated in the examination process, giving a total of 2,269 participants. The analyses presented in the baseline morbidity report were largely performed using all available Ranch Hand data (1,045 participants) and data from originally selected comparison subjects (773 participants) yielding a total of 1,8>8 subjects. Data from the substitute comparison subjects were not used for inference. Due to logistic difficulties, the substitute comparisons were invited to participate in the physical examination later in the study period; therefore the substitute comparisons had a narrower time window within which to travel to the examination site. The substitute comparison subjects, entered early into the study to replace ineligible subjects, were found to be statistically comparable to the original comparisons when evaluated on clinical endpoints. Some principal investigators were concerned that the group substituting later for noncompliant comparisons may have self selected differently due to the reduced scheduling opportunity. Since opinions differed, a management decision was made to use only the originally/ invited comparisons for inference. The baseline morbidity report had 13 primary clinical chapters addressing: general health, neoplasia, reproduction, neurological status, psychological status, hepatic function, dermatological findings, cardiovascular conditions, immunological competence, hematopoietic status, and renal, pulnuv nary, and endocrine functions. More than 190 clinical variables were tabulated. In this report only a subset of these variables will be discussed. The variables selected for emphasis in this report were chosen because of the availability of corresponding or related evaluations in laboratory or other epidemiological studies related to dioxin effects. Thus, this report reflects on the degree to which AFHS findings are compatible with TCDD toxic effects as currently suggested by experiments with animals and observations in humans. Sample sizes may vary in adjusted analyses due to missing covariate or endpoint data. In all analyses, the phrase "statistically significant" refers to" a p value of less than or equal to 0.05. �The fixed sample sizes in this study impose limits on its ability to detect small relative risks for rare diseases. This ability is expressed in probabilistic terms using the statistical quantity called power, which is defined as the probability of detecting a group difference of interest. In the case of dichotomous response, such as presence or absence of disease, groups are generally compared with relative risk, defined as the ratio of the probability of disease in the exposed group to the probability of disease in . the comparison group. A relative risk of two, for example, would indicate a doubling of the disease rate in the exposed group relative to the comparisons. If the disease incidence in the control group is 1/1000, as is typical for some specific cancers, such as bladder cancer, the AFHS would require 22, 810 exposed and an equal number of comparisons to attain a power of 80J5 to detect a relative risk of two, assuming two-sided testing with a 5% significance level. In fact, the AFHS is unable to detect relative risks less than eight in diseases with a comparison incidence of less than or equal to 1/1000. 'There is a 0.351 chance of observing no cases at all of a rare disease of incidence 1/1000 in a group of 1,045 subjects. With even rarer diseases of incidence, 1/10,000, there is a 90$ probability of observing ho cases at all in a group of» 1 ,015 subjects. This study does have good power to detect small relative risks for diseases having an incidence of 5/100 or greater.. For example, the power for detecting a relative risk of 2, when the disease rate in the comparison population is 5/100, is 0.85, based on only 450 pairs in a matched pair analysis. In the case of continuously distributed response variables, such as blood pressure or cholesterol, this study has good power to detect small mean shifts. For example, the probability of detecting a mean shift of 5% in a matched pair analysis utilizing only 450 pairs is at least 0.90, assuming equal variances, two-sided testing and an 0.05 significance level. The power of the mortality component of this study is similarly constrained. The mortality study design consists of all 1 ,24? Ranch Hands and up to 5 matched controls per Ranch Hand. The mortality study has a power of 0.85 to detect a relative risk of two for causes of death, such as heart disease, having incidence 1/100 in the comparison population. The corresponding power is less than 0.25 for causes having an incidence of 1/1000 in the control population. This study, in summary, has good power for detecting relative risks on the order of two or three for common diseases and causes of death and has virtually no power for detecting relative risks'bf the same order of magnitude for rare diseases. It does have good power for detecting small mean shifts in continuously distributed response variables. Bearing these study power constraints in mind, the following ten areas of clinical morbidity are discussed. General Health Weight loss has been frequently reported as a consequence of subacute and chronic administration of TCDD to animals. McNulty^ noted marked weight loss in two male rhesus monkeys fed diets containing 2 or 20 ppb TCDD. Horses ingesting TCDD-contaminated waste oil sprayed on arenas in Missouri showed significant weight loss.5'6 Chapman and Schiller' report that decreased feed consumption did not account for weight loss in C57 mice given dioxin in their �o feed. Seefeld and colleagues0 conclude that TCDD affects the weight regulation set-point in rats. Weight loss has not been prominently commented on in studies of human exposure.. However, Oliver* indicates weight loss or loss of appetite in two of three reported cases. The toxicological literature mentioned here suggests that weight changes might be anticipated in a dioxin-exposed group. In the AFHS, body fat percent was calculated by a formula that uses height and weight as independent variables.10 No statistically significant differences in the distribution of estimated body fat were detected between the Ranch Hand and comparison group. The basic data are shown in Table 3. TABLE 3. DISTRIBUTION OF BODY FAT PERCENT Lean «10*) Number Percent Ranch Hand Comparison 13 7 (1) (1) Normal (10-25*) Number Percent 82U 607 (79) (79) Obese (>25*) Number Percent 207 157 (20) (20) Total 1,0*1 771 The sample sizes in Table 3 (1,014 and 771) are reduced due to missing data for three individuals. A chi-square statistical test using these data indicated no statistically significant difference between the distributions for the groups (p-0.89). Detailed analyses of percent body fat, adjusting for age, race, and occupational category, are described in the baseline report, and these analyses also indicated the absence of a group difference. Also, within the Ranch Hand group no relationship between body fat and the exposure index was found. Ne.oplasia The animal toxicology literature portrays dioxin as a carcinogen, a cocarcinogen, and as having anti-carcinogenic properties. Jackson 1 showed impairment in the functioning of the mitotic apparatus at 0.2 yg/1 in dividing endosperm cells of the African blood lily. Kociba et al12 fed male and female Sprague-Dawley rats on diets supplying 0,1, 0.01, or 0,001 ug of TCDD/kg/day for 2 years. Exposed male rats displayed more stratified squamous cell carcinomas' of the hard palate and tongue. However, fewer adenomas of the pancreas and pheochromocytomas of the adrenal were found. Kouri and colleagues" conclude that TCDD is a cocarcinogen. They propose that this effect is mediated through aryl hydrocarbon-hydroxylase induction. On the other hand, DiGiovanni and colleagues found that TCDD reduced cutaneous papilloma formation by various hydrocarbons, indicating an anti-carcinogenic effect. With respect to carcinogenesis in man, Coggon and Acheson •* reviewed the available �epidemiological studies .and concluded M. . . there is suggestive evidence of a biological association between phenoxy herbicides (or their contaminants) and soft-tissue sarcomas. The. evidence relating these products to the occurrence of lymphomas is weaker." Table 1 summarizes the cancer events that have occurred in the Ranch Hand and comparison groups since these individuals completed their Southeast Asia military tours. All shown cancer cases have been verified by personal medical or pathological records. One comparison individual has had both a skin and systemic cancer. In the table below he is shown as having only a systemic cancer for purposes of the statistical analysis. TABLE 1. CANCER VERIFIED BY INDIVIDUAL MEDICAL RECORDS OR PATHOLOGY REPORTS Group Ranch Hand Comparison No. with Skin Cancer (.%) 35 (3.3) 10 (1.3) No. with Systemic Cancer (%) 13 8 (1.2) (1.0) No. with No Cancer (?) Total 997 (95.1) 755 (97.7) 1,015 773 Of 1,015 Ranch Hands, 1.59$ have a skin or systemic cancer. Of the 773 comparison individuals, 2.33J have a skin or systemic cancer. Thus, the relative risk for any type of cancer is 1.97 and this relative risk has a probability of less than 0..01 of occurring by chance under the hypothesis of no difference. This statistical test for overall cancer rate difference was the hypothesis test formulated prior to examination of the cancer data set. After this statistical test was performed, detailed review of the cancer data file suggested that the file be partitioned into skin and systemic events based upon the observation that skin cancer comprised a large fraction of the cancer set. The relative risk for skin cancer is 2.59, and this risk has a p value of less than 0.01. The relative risk for systemic cancer is 1.20, and this risk has a probability of 0.67 of occurring by chance. In the first morbidity report, the authors felt the neoplastlc process was confined to skin. This inference cannot be affirmed because the separate skin and systemic hypothesis tests followed rather than preceded review of the cancer data file; thus the critical levels for these tests are unknown. Furthermore, important increments in relative risk for systemic cancer could be missed by chance mechanisms because.of the small sample sizes in the AFHS. Neither skin cancer nor systemic cancer rates were correlated with herbicide exposure level; however, these statistical analyses involved a very small number of cases in most of the nine occupation-exposure categories, thus decreasing the precision of rate determinations in these categories. 10 �Reproduction The literature suggests that dioxin has mutagenic and teratogenic capacity. Some bacterial tests have been positive for TCDD mutagenicity. The baby hamster kidney cell transformation assay was positive for TCDD mutagenicity. '" Chromosome 1 aberrations have been seen in bone marrow cells of male rats exposed to TCDD. ' Van Miller and Allen18 observed reduced spermatogenesis in rats experiencing chronic exposure to TCDD. Seller * observed reduced DNA synthesis in mouse testicle. Courtney and Moore20 observed kidney abnormalities in fetuses of female rats given 0.5 yg/kg/day of TCDD subcutaneously on days 6-15 of gestation. Cleft palate and renal abnormalities have been produced in the mouse after oral or subcutaneous administration of TCDD to females. Lamb and colleagues showed that exposure of male mice to toxic levels of TCDD with subsequent mating did not affect sperm, mating frequency, or quality of offspring. This sampling of the literature should be sufficient to indicate the possibility of reproductive changes in exposed human populations. Hanify and colleagues2^ have reported an association of aerial spraying of 2,4,5-T and an excess of talipes in New Zealand. Townsend et al., in a study of Dow chemical workers' wives, found no statistically significant differences in fetal wastage or birth defects. The Australian government's study of birth defects in Vietnam veterans showed no statistically significant differences in rates between veterans who served in Vietnam and those who did not.2-' A Centers for Disease Control study (CDC) also found Vietnam veterans to have the same overall risk for fathering abnormal offspring as nonveterans. In the CDC study some specific defects were associated with higher exposures, but interpretation of this finding was uncertain. Male exposure could theoretically lead to unfavorable reproductive outcomes by means of several mechanisms: (a) mutated DNA in sperm, (b) abnormal sperm or testicular function due to biochemical effects, (c) transmission of TCDD to the female by spermatic fluid, and (d) transmission of TCDD to the female by contact with clothes or other objects. 11 �Semen specimens from study participants without vasectomies or orchiectomies evidenced no statistically significant group differences in sperm count or percent abnormal sperm. The data are displayed in Table 5. Sample sizes reflect the number of compliant subjects not previously vasectomized. TABLE 5. DESCRIPTIVE STATISTICS OF SPERM VARIABLES BY GROUP Sperm Count (millions/ml) Ranch Hand (N-572) Comparison (N-421) 111.5 111.9 102.8 108.8 Percent Abnormal Sperm Ranch Hand (N-560) Comparison (N-409) 9.7 9.6 12 5.5 5.2 �Conceptions reported by study participants and their spouses were categorized as miscarriages, stillbirths, induced abortions, and live births. Numbers in each category are shown in Table 6 with indication of whether the conception occurred before or following the participant's Southeast Asia tour. TABLE 6. CONCEPTION OUTCOMES BY GROUP MEMBERSHIP AND TIME Yes Miscarriages Ranch Hand Pre-SEA (?) No Yes Post-SEA (t) No 239 (13.7) 1,505 156 (15.0) 883 172 (11.9) 1,276 101 (12.5) P Value 726 0.76 Comparison Stillbirths Ranch Hand 9 (0.5) 1,735 12 (1.2) Comparison 8 Induced Abortions 8 Ranch Hand . Comparison 7 Live Births Ranch Hand 1,187 (0.6) 1,110 8 (1.0) (0,5) 1,736 37 (3.6) (0.5) 1,111. 33 (1.0) 797 (85.3) 257 833 (80.2) 206 Comparison 1,258 (86.9) 190 682 (82.2) 118 1,027 1.00 822 1,002 0.89 0.91 . The data were analyzed using log-linear models2', adjusting for the factors of maternal age «35, £35), maternal smoking (yes/no), maternal alcohol use (yes/no), and paternal age (<35, £35). The four statistical tests all had p values greater than or equal to 0.76. Exposure analyses showed no consistent pattern with exposure level. 13 �Ranch Hand and comparison live births were further analyzed to determine the occurrence of learning disabilities, physical handicaps, infant death, neonatal death, and birth defects. Data on live birth outcomes by group membership and time are shown in Table 7. TABLE 7. LIVE BIRTH OUTCOMES BY GROUP MEMBERSHIP AND TIME Pre-SEA (*) No Yes Post-SEA (*) No. (3.8) 1,430 75 (9.0) 758 57 (1.5) 1,201 47 (6.9) 635 131 (9.0) 1,353 126 (15.1) 707 Comparison Infant Death Ranch Hand 103 (8.2) 1,155 77 (11.3) 605 7 (0.5) 1,480 3 (0.4) 830 Comparison Birth Defects Ranch Hand 2 (0.2) 1,256 1 (0.1) 681 (5.2) 1,409 76 (9.1) 757 80 (.) 64 1,178 44 (6.5) 638 20 (1.3) 1,467 14 (1.7) 819 (1.4) 1,241 Yes Learning Disability Ranch Hand 57 P Value 0.19 Comparison Physical Handicap Ranch Hand 0.45 0.81 Comparison Neonatal Death Ranch Hand Comparison 78 17 1 3 0.04 0.20 (0.4) 679 Analyses of these data, adjusting for maternal age, maternal smoking, maternal alcohol use, and paternal age, reveals a statistically significant increase in reported birth defects in the Ranch Hand group. Subsequent to this observation the birth defects were categorized as severe (life threatening or interfering with normal overall health of'socio-economic progress), moderate (not life threatening and, with health care, non-interfering with overall health or socio-economic progress), and limited (nbn-llfe threatening, non-interfering, and needing no care). These data are shown in Table 8. 14 �TABLE 8. SEVERITY OF REPORTED BIRTH DEFECTS BY GROUP MEMBERSHIP AND TIME Pre-SEA Moderate N Ranch Hand Comparison 51 (3.0) 50 (3.5) Limited N (*) <*) 32 (1.9) 27 (1.9) 7 (0.1) 10 (0.7) No reported Defects N Total (*) 1,633 (95) 1*348 (91) 1,723 1,435 Post-SEA N (*) Ranch Hand 32 (3 .5) Comparison 18 (2 .4) N 26 (2.8) 10 (1.3) 22 (2.4) 20 (2.7) N (*) 837 (91) 696 (94) 917 744 The above data set is larger than the prior data set since this set contains all reported live births while the prior set consisted of all live births on whom the covariates (maternal age, maternal smoking, maternal alcohol, and paternal age) were available. The larger data set was used since categorizing birth defects as severe, moderate, or limited reduced cell counts. Full covariate adjustment was not possible. The morbidity report can be consulted for details.' Once again, in this larger data set, an increase in reported defects is noted. Specifically, the Ranch Hand to comparison birth defect odds ratio is 0.85 for children born prior to Vietnam, while the post-Vietnam ratio is 1.39. A statistical analysis of the complete data set suggests that the birth defect severity pattern by group relationship changes with time (p-0.07). Visual inspection of these data suggests that this nearly significant change may be due to a relatively large number (26) of post-Vietnam Ranch Hand children reported as having limited birth defects; the Ranch Hand to comparison odds ratio in this category is 2.16. However, an excess of severe defects is also seen by visual inspection. A separate analysis on data for defective children only (with the reported non-defective children removed) suggests that the group by severity relationship does not change with time (p-0.15) and that the severity pattern does not change with group (p-0.78). The statistical analysis of the complete data set is more powerful than these last two analyses since the latter analyses use only the 305 reportedly defective children, which constitutes only 6.3$ of the total data set. In the AFHS first morbidity report, it was asserted that minor skin lesions accounted for the reported birth defects excess. That analysis was incomplete, and we are no longer confident in that inference. Also in the AFHS first morbidity report, it was properly suggested that differential reporting of birth defects could be responsible for the apparent excess. A 15 �preliminary analysis, of medical records of children reported abnormal, has indicated that overreporting of defects may not account for the excess; however, intensive work is in-;progress addressing both differential over- and underreporting. Exposure index analysis of the birth defect data yields inconsistent findings that are not interpretable as a herbicide effect. The finding of increased birth defects as reported by study participants, their wives, and partners is under further investigation by review of birth certificates and medical records of all 5,663 children to verify both positive and negative responses. Neurological Findings A variety of neurological symptoms have been described following industrial accidents involving TCDD including headaches, asthenia, sleep disturbances, irritability, and confusion. Peripheral polyneuropathy is a specific neurological condition that has also been linked to acute dioxin exposure, and is a condition that is amenable to direct clinical measurement.. Elovaara and colleagues2' found that acid proteinase activity was increased in the ' brains of Wistar rats after TCDD treatment. Acid proteinase is a lysosomal enzyme responsible for protein destruction in the mammalian brain, and may play a role in degenerative diseases and intoxications. In the AFHS, neurological examination of the twelve cranial nerves revealed no statistically significant group differences. Assessment of peripheral nerve status included sensitivity to touch, vibration, and test of the patellar, achilles, and biceps reflexes. Again, no statistically significant group differentials were observed. As shown in Table 9, the groups were not statistically different with respect to nerve conduction velocities. Of interest is the observation (data not shown) that the conduction velocities decreased as expected with increasing self-reported alcohol use (drink-years) and postprandial glucose levels (dichotomized as less than, or equal to or greater than 120 mg/dl). These effects appear to be consistent in both groups. All exposure index analyses were unremarkable. 16 �TABLE 9. NERVE CONDUCTION VELOCITY (M/SEC) Nerve Group (N) Ulnar (above the elbow) Ranch Hand (1,035) 55.9 Comparison (769) 56.2 Ranch Hand (1042) 60.5 Comparison (771) 60.7 Ranch Hand (1041) 48.2 Comparison (769) 48.1 Ulnar (below the elbow) Peroneal Unadjusted Mean P Value 0.30 0.39 0.74 Psychological Assessment 30 Working with rats, Creso et al.JU have noted that TCDD provokes irritability, aggressiveness, and restlessness. They found by in vitro studies that TCDD directlv stimulates the striatal and hypothalamic adenylate cyclase of rat. Oliver" reports that two of three TCDD-exposed individuals studied expressed the symptom of excessive fatigue, and one communicated loss of ability to concentrate. Bauer et al.31 studied nine workers with chloracne and noted fatigue and apathy alternating with anger and irritability. Rorshach tests showed a weakened emotional32reaction, slowed thought processes, and perseveration. Poland and Smith observed increased values on the MMPI mania scale in the group of workers with chloracne when compared to two groups with less severe acne. The AFHS disclosed several group differences in psychological testing. Indices developed from the questionnaire relating to fatigue, anger, mental erosion, anxiety, isolation, and depression all showed Ranch Hands to be statistically significantly less well than comparisons. The Cornell index results paralleled the questionnaire indices; however, no increase in Ranch Hand depression was seen. Education strongly related to results of all psychological testing with group differences tending to be most prominent in high-school-only educated individuals rather than college educated. For example, the MMPI among high-school-only educated individuals showed statistically significantly higher hypochondria, mania, and social introversion scores among Ranch Hands. The MMPI among college-educated participants showed only higher social introversion among Ranch Hands. In interpreting these data it must be remembered that there is a very high association between being college educated and having officer status. None of these data had been adjusted for the potentially confounding variable of combat stress. This area was pursued during the 1985 follow-up examination. Tests aiming at neuromuscular and intellectual functioning (WAIS and Halsted-Reitan) showed no group differences, and no consistent patterns emerged from any analyses using the exposure index. 17 �Thirty-six of 1,015 Ranch Handera ( . and 16 of 773 comparisons (2.1 Jt) 3W reported psychological illness (psychosis, alcohol dependence, anxiety, or other neurosis). This group difference is not statistically significant. Hepatic Examination Dioxin has been associated with the occurrence of hepatotoxicity. Proliferation of the smooth endoplasroic reticulum, distortion of liver architecture, and increase in liver weight relative to body weight have been seen in rats and mice. Changes in serum enzymes have been observed, but not with consistency. Porphyria has been observed after chronic dosing. A sampling of an extensive literature is given next. Weber et al." observed rats over a 32-week period following a single intraperitoneal dose (20 ug/kg) of TCDD. Centrolobular necrosis, mitochbndrlal lesions, smooth endoplasmic reticulum increase, and hepatic regeneration were seen. The abnormalities began to regress 16 weeks after exposure. Similar findings in rats have also been reported by King and Roesler* Gupta and colleagues^ orally administered single, daily, and weekly doses of TCDD to rats, guinea pigs, and mice. Severe liver lesions were only seen in the rat indicating species variation. Kociba et al. noted elevated liver enzyme levels in rats fed diets with TCDD for two years. Porphyria cutanea tarda has occurred in workers exposed to TCDD,'° and porphyria has been observed in rats. • Sweeney and colleagues^ have performed work that shows a synergism between the effects of TCDD and systemic iron. Iron deficiency was seen to prevent TCDD-induced porphyria in mice. In iron-deficient animals, TCDD was not able to decrease uridine decarboxylase levels. Iron deficiency protected mice against skin damage and the disruption of hepatic architecture seen with TCDD. Mixed function oxygenase activity was induced by TCDD in iron-deficient animals to a lesser degree than in non-deficient animals, but this difference was not significant. Sinclair and G-ranick^ had earlier seen that iron was needed for uroporphyrin formation induced by chlorinated hydrocarbons. Smith et al.39 noted an increase in hepatic iron content 3 weeks after a 75 pg/kg oral dose of TCDD to C57BL/10 and DBA/2 mice. Increased intestinal uptake of iron has been observed in mice and! rats after TCDD exposure. ° In the AFHS, nine biochemical determinations of liver function were made: SOOT, SGPT, GGTP, alkaline pnosphatase (Alk.Phos.), total bilirubin (T.Bill), direct bilirubin (D.Blli), lactic acid dehydrogenase (LDH), cholesterol (Choi), and triglycerides (Trig). In the analyses of these nine variables, statistical adjustments were made for four covariates: current alcohol ingestion (self-reported in drinks per day), self-reported days of exposure to nonherbicide industrial chemicals, self-reported days of exposure to degreasing chemicals, and presence or absence of antibody to hepatitis B surface antigen (anti-HBsAg). Table 10 provides unadjusted and adjusted means and percent abnormality by group for the nine hepatic-related variables. The standard age-adjusted criteria for abnormal laboratory values were used throughout. No obvious group differences are apparent in these data. However, the statistical modeling of the dependent variables with the covariables showed three group differences. The Ranch Hand SGOT-alcohol regression slope is 0.0178 base-10 logarithmic units per drink per day while the comparison slope is 0.0113. 18 �These slopes mean that among study participants who report one drink per day, Ranch Hand SCOT levels are 1.5% higher than comparison levels. Among study participants who had four drinks per day, Ranch Hands Have SCOT levels 6.3% higher than comparisons. TABLE 10. UNADJUSTED MEANS, ADJUSTED MEANS, AND PERCENT ABNORMALITY FOR NINE LIVER-RELATED VARIABLES Unadjusted Means Adjusted Means PCT Outside of Normal Range Variable Group SCOT RH* Com** 33.0 33.1 33.0 33.1 13.9 11.8 SGPT RH Com 20.3 20.5 20.3 20.5 7.8 8.6 GGPT RH Com i«0.2 39.3 10.1 39.3 10.8 10,3 Alk.Phos. RH Com 7.68 7.53 7.69 7.52 17.3 16.9 T. Bili RH Com 0.57 0.58 0.57 0.58 1.8 2.0 D. Bili RH Com 0.23 0.21 0.23 0.21 29.0 29.7 LDH RH Com 112. 1 111.7 112.1 111.7 1.7 2.1 RH Com 212.2 216.6 212.2 216.6 26.0 27.7 RH Com 121.8 124.3 121.9 121.1 31.7 36.1 • Choi Trig . *RH denotes Ranch Hand **Com denotes original fully compliant comparisons. LDH-alcohbl slopes were 0.0011 logarithmic units per drink per day in the Ranch Hand cohort and -0.0008 in the comparison group (p-0.011). The LDHdegreasing chemical slopes were 0.000005 and -0.0000008 logarithmic units per day degreasing chemical exposure in the Ranch Hand and comparison groups respectively (p-0.037). 19 �Twenty-four-hour urine collections were obtained for 620 Ranch Hands and 139 comparisons; uroporphyrins, coproporphyrins, and d-arainolevulinic acid were determined. Unadjusted group means are shown in Table 11. TABLE 11. UNADJUSTED GROUP MEANS FOR THREE COMPOUNDS RELATED TO PORPHYRIN METABOLISM Uroporphyrin RH Com 30.5 30.8 Coproporphyrin RH Com 31•2 30.8 d-aminolevulinic acid RH Com 2328.9 2383.2 No statistically significant differences are obvious in these data. Detailed statistical analyses were done, simultaneously adjusting for six covariates: current alcohol use, blood urinary nitrogen, creatinine clearance, days of exposure to industrial chemicals, days of exposure to degreasing chemicals, and presence/absence of antibody to hepatitis B surface antigen. A generalized linear model analysis was done for each of the three compounds with all six covariates examined simultaneously. The coproporphyrin-alcohol slope was +0.013 logarithmic units per drink per day in the Ranch Hand group and -0.008 logarithmic units per drink per day in the comparison group (p-O.OMS). No other group differences were statistically significant. The clinical relevance of these differences in slope is unclear. Sixteen of 1,027 Ranch Handers (1.56J) were diagnosed as having hepatomegaly at physical examination while six of 769 comparisons (0.78?) had that finding (p-0.138). Thirteen of 1,032 Ranch Handers had a verified medical history of liver disorder other than hepatitis, jaundice, or cirrhosis verified by medical record while two of 773 comparisons had the same (p-0.001*). Throughout the hepatic analyses no variable showed a meaningful relationship with the herbicide exposure index. In all the above analyses no adjustments for. iron metabolism were made. Dermatological Finding TCDD is known to cause chloracne. Chloracne is an acneiform lesion which tends to predominate in the areas of the face around the eyes, temples, and ears. Chloracne has been frequently seen in humans who have contacted TCDD in the context of industrial accidents. In one study the chloracne resolved within 1 year for the most part with no scarring. ' Chloracne has been noted 20 �by many investigators years after exposure and is generally recognized as a persistent effect of dioxin exposure. Bovey and Young^2 conclude that "the presence of active chloracne months to years after exposure does not necessarily mean continuing exposure." In the AFHS, no active chloracne was found in either the exposed or comparison group by examination or review of medical records. Also, as indicated in Table 12, there were no statistically significant group differences with respect to chloracne-related lesions. No statistically significant regression trends were noted with the exposure index. TABLE 12. PREVALENCE OF DERMATOLOGIC DIAGNOSES IN PERCENT Diagnoses Ranch Hand N - 1045 Comedones Acneiform lesions Acneiform scars Cysts Hyperpigmentation Other abnormalities Any abnormalities 21.7 18.3 11.2 11.6 8.3 12.6 45.0 Comparison N - 773 Relative RisK 0.60 0.66 0.57 0.46 0.35 0.03 0.97 20.7 17.5 10.4 10.5 7.1 16.3 44.9 P Value 1.05 1.05 1.08 1.10 1.17 .77 1.00 95% Conf Int (.87,1.26) (.85,1.29) (.82,1.43) (.84,1.46) (.84,1.65) (.81, .98) (.90,1.11) Cardiovascular System TCDD causes a rapid, dose-dependent elevation of lipofuscin in the hearts of female Fischer 344 rats. The authors of the research suggest that TCDD toxicity may be associated with radical-induced lipid peroxidation. 3 Kociba and colleagues12 saw an increased incidence of arteritis in rats. In 1958, Schmittle and colleagues^ reported hydr©pericardium in poultry following ingestion of feeds contaminated with industrial chemicals. In 1969 a dioxin was shown to be6 the hydropericardium-producing factor in poultry. 5 Jirasek and colleagues^ report that a 57-year-old male with chloracne developed unusually severe atherosclerosis and subsequently died. Moses and colleagues ' found that 17 of 116 workers with chloracne reported myocardial infarction (14.75E) while 7 of 85 (8.2?) without chloracne reported the same (p > 0.10). Zack and Suskind ° observed no excess in circulatory system deaths comparing events in Monsanto Company workers to standard US population rates. In the AFHS no statistically significant group differences were observed with respect to measurements of systolic and diastolic blood pressures. Also, the groups were not statistically significantly different with respect to numbers of abnormal.electrocardiograms. Abnormal funduscopic findings were not associated with group membership nor was the occurrence of carotid bruits. During the physical examination, 10 peripheral pulses were examined: the radial, femoral, popliteal, dorsalis pedis, and posterior tibial"pulses. One 21 �or more pulses in this set of pulses were found to be abnormal in 12.8* (106/ 829) of the non-black Ranch Hands, while 9.4J (56/596) were found abnormal in the comparison group (p-0..05). The group difference was not statistically significant in the data set on Black study participants, but this may only reflect smaller numbers. Peripheral pulse abnormalities tended to aggregate in older individuals (2 40 years) who smoked (> 10 pack years). The participants were allowed to smoke prior to the examination of the pulses, and more Ranch Hands smoked at the time of the examination than did comparisons (45.7J versus 40.5J, p-0.03). Data on the numbers of individuals in the Ranch Hand and the comparison group who had experienced some form of heart disease (ICD-9th edition, CM) or who had experienced a myocardial infarction are shown in Table 13. The numbers shown are supported by medical record verification of participants' selfreporting. These data do not suggest a difference in ischemic heart disease in the two groups. TABLE 13. HEART DISEASE AND HEART ATTACK IN THE AFHS Ranch Hand Yes No Verified Heart Disease Verified Heart Attack 147 7 898 1,038 Comparison Yes No 109 3 664 770 P Value 0.982 0.432 Immunological Effects Clark and colleagues ^ describe thymic atrophy as a consistent observation in all animal species following TCDD exposure. They also observed reduced delayed hypersensitivity reactions assessed by ear swelling following oxazalone sensitization in 6- to 8-week-old mice. Cytotoxic T cell lymphocyte generation was impaired by low doses (0.004 ug/kg) of TCDD. Following a variety of experiments, the authors suggest that TCDD may acutely decrease cytotoxic T lymphocyte generation by promoting the generation of suppressor T cells. In parallel results, Montovani and colleages-*0 observed decreased numbers of peritoneal macrophages and splenocytes in TCDD-treated 6- to 8-week-old mice, while cytotoxic capability per unit number of cells was not affected. Van Logten and colleagues51 conclude that the atrophy of the thymus observed following TCDD administration in rats is not mediated by the adrenal or pituitary glands. Clark and colleagues52 state that the immunotoxic effects of TCDD in C57B/6 and DBA/2 mice occur at dose levels below those needed to induce hepatic mixed function-oxidase enzymes. In the AFHS, immunological status was assessed in 592 participants by (1) the enumeration of T-lymphocytes, T-lymphocyte subsets, and B-lymphocytes using monoclonal surface marker analysis, and by (2) the assessment of lymphocyte ability to respond to selected antigen or mitogen stimuli.. 22 �The data were analyzed for statistically significant group distributional differences using the Kolmogorov-Smirnov two-sample test* The analysis of the immunological cell count data showed no statistically significant differences. These cell count data are shown in Table 14. TABLE 14. SELECTED PERCENTILES AND P VALUE FOR KOLMOGOROV-SMIRNOV TESTING OF NUMBERS OF SURFACE MARKER POSITIVE CELLS (THOUSANDS/MM3) Variable N Group 10* 50* 90% P Value (comparing distributions) 235 144 0.70 0.77 1.25 1.23 1.96 2.02 0.74 Com T3 RH Com 233 144 0.70 0.73 1.27 1.28' 1.96 2.13 0.39 Ti, RH Com 231 147 0.40 0.48 0.79 0.78 1.25 1.42 .0.81 RH Com 235 t47 0.30 0.28 0.57 0.60 0.99 1.17 0.34 B, 'RH Com 235 147 0.023 0.022 0.071 0.071 0.188 0.247 0.097 TLC* RH .Com 290 177 1.34 1.35 1.92 1.91 2.54 2.74 0.63 T RH 11 T8 - *Total lymphocyte count 23 �Statistical testing of the four stimulation and two control measurements assessing lymphocyte functional ability is shown in Table 15. TABLE 15. KOLMOGOROV-SMIRNOV TESTING OF T AND B CELL FUNCTION DATAs THYMIDINE INCORPORATION MEASURED AS COUNTS/MIN N Percentiles 50$ Variable Group Control #1 RH Com 279 168 140 138 374 D48 1,320 1,483 0.20 After ConA RH Com 279 168 17,741 13,596 54,190 58,394 91,724 99,104 0.38 After PHA RH Com 279 168 33,027 30,143 79,342 84,339 130,064 135,684 0.51 Control #2 RH Com 271 168 132 142 388 404 917 1,079 0.85 After PW RH Com 274 168 12,700 12,232 29,623 27,916 58,288 53,662 0.64 After TT RH Com 274 168 866 1,001 3,726 3,719 13,979 16,058 0.81 ConA PHA PW TT » - 10 % 90* P Value concanavallin A phytohemagglutin pokeweed mitogen tetenus toxoid. No statistically significant group differences are noted in these T and B cell function data. High laboratory variability and small sample sizes led to the decision to not use exposure index analyses with the immunological data. The control #1 and control #2 variables represent the unstimulated activity of the T cells. Endocrinological Effects Working with liver homogenates taken from adult male Wistar rats, Nienstedt and colleagues" sb°wecl tnat TCDD reduced the catabolism of testosterone. Hook and colleagues-* also reported reduced metabolism of testosterone. These reports would suggest that elevated testosterone levels could be observed in a recently dioxin-exposed population. 24 �Bastomsky55 observed serum T^ levels to be one-half of normal in TCDDtreated animals, but serum T, was elevated by 50%. Sephadex uptake of T, was statistically significantly decreased. Potter and colleagues56 observed serum levels of TH to fall to 46J of pair-fed controls in TCDD-dosed rats. However, no statistically significant change in serum Tg occurred. These authors also noted hypoglycemia after a single intraperitoneal dose of TCDO. Rozman and colleagues'7 emphasize the potentially important role of thyroid hormones in certain expressions of TCDD toxicity. Specifically, athyroid rats showed a markedly decreased mortality rate and less weight loss than nonthyroidectomized or thyroidectomized but euthyroid controls. In the endocrinological portion of the AFHS, five clinical variables were studied: T, uptake, serum T^,, free thyroxine index (FTI), 2-hr postprandial glucose, aha serum testosterone. One statistical analysis of these variables examined the number of participants below, in, or above the variables' normal ranges. This analysis is summarized in Table 16. Statistically significant difference is seen in these data for the Tg uptake comparison. The Ranch Hand group was also contrasted with the comparison group in terms of the five endocrinological variables using analysis of covariance, adjusting for age and percent body fat. These analyses are summarized in Table 17. Three group differences are noted in these analyses. In both the Ranch Hand and comparison groups, a decrease in To uptake is observed with advancing age, but the slope is -0.006851 per year In the comparison group and -0.0195X per year in the Ranch Hand group, and this group difference was statistically significant (p-0.026). Two-hour postprandial glucose levels increase with age in both the Ranch Hand and comparison groups, but the rate of increase is 1.53 mg/dl per year in the Ranch Hand group and 0.77 mg/dl per year in the comparison group (p-0.006). Lastly, Ranch Hands show a higher (but not statistically significant) testosterone level than do comparisons. Both increasing age and body fat were found to be associated with decreasing testosterone levels to the same extent in both groups. 25 �TABLE 16. UNADJUSTED PERCENTAGES FOR FIVE ENDOCRINOLOGICAL VARIABLES BY VARIABLE LEVEL AND GROUP Low Variable Level Normal High P Value For Group Difference Variable Group To Uptake (*) RH Com 1,032 767 5.72* 8.47* 93. 41* 91.26* 0.87* 0.26* 0.020 T RH Com 1,033 767 0.10$ 0.39* 99.13* 99.22* 0.77* 0.39* 0.250 FTI RH Com 1,033 767 0.00* 0.26* 99.71* 99:74* 0.29* 0.00* 0.085 2-hr Glucose (mg/dl) RH Com 1,010 770 NA NA 84.81* 82.73* 15.19* 17.27* 0.234 Testosterone RH Com (ng/dl) 1,034 769 1.93* 6.37* 91.58* 93.11* 0.48* 0.52* 0.414 4 (yg/dl) N , 26 �TABLE 17. RANCH HAND COMPARISON GROUP MEANS OF ENDOCRINE VARIABLES Variable Group T, RH Uptake (*) Com T i N P Value For Unadj. Unadj. Adjusted Mean Means Mean 1,037 30.11 . RH ., 1,038 8.16 Group-by-age interaction (p-0.026) 30.28 770 P Value For Remarks About Adjusted Adjusting Means Covariates * » 0.21 8.15 (yg/dl) t Com RH 770 8.39 1,038 2.51 770 2.51 0.31 8.39 0.38 2.51 FTI Com 2-hr Glucose (mg/dl) Testosterone (ng/dl) RH Com RH Com 1,015 102 0.13 101 773 . 2.51 0.07 1,039 772 * 0.37 652 651 631 * Group-by-age interaction (p-0.006) - 0.02 637 0.06 ^Signifies interaction present rendering group means noninformative. Mortality Administration of--.^fngle doses of dioxin to animals can result i;n lethality with marked species differences being observed. Repeated.daily doses can also lead to death. No studies with large numbers of animals (for instance, with 100 control and 100 exposed animals) are being conducted where the possibility of life shortening by very low doses of dioxin can be evaluated. The Australian government provides a very complex report addressing mortality among Vietnam veterans." Infantry exhibited a relative mortality rate of 0.96 with 95? confidence interval from 0.7 to 1.3. Engineers exhibited a relative mortality rate of 2.5 (confidence interval 1.1 to 1.0); armour and artillery, 1.06 (confidence interval 0.7 to 1.7); veterans with minor field presence, 1.5 (confidence interval 0.9 to 2.6); and non-field corps, 1.01 (confidence interval 0.7 to 1.5). Thus, only the engineers exhibited a statistically significant difference (p-0.001), and that difference involved increased Vietnam veteran mortality. 27 �In the AFHS, cumulative mortality as of 31 December 198M displays no statistically significant overall Ranch Hand-comparison differences. Summary counts of death and age standardized mortality ratios (SMR) by rank and occupation are given in Table 18. TABLE 18. SUMMARY COUNTS, SMRs AND P VALUES FOR DEATH BY RANK AND OCCUPATION Rank Officers Enlisted Ranch Hand At Risk Dead Rate Comparison At Risk Dead Rate 466 791 16 39 .OM9 2278 3893 646 611 2U 31 .037 .051 3163 3008 98 187 SMR P Value .OM8 .791 1.03 .37 .89 .051 .041 .726 1.23 .13 .33 Occupation Flying Ground 161 121 Further study of these mortality data discloses complex patterns with date of birth (which is related to date of service in Vietnam) and date of death. CONCLUSION In this report, eleven clinical areas have been emphasized based on a clinical toxicological profile developed from the literature concerning animal and human responses to dioxin and availability of data in the AFHS. Table 19 lists the general toxicological effects anticipated oh the basis of the literature and also summarizes Ranch Hand findings. The toxicological profile we have developed from the literature certainly has an element of subjectivity as articles were selected and interpreted from a very large literature. Similarly, each observation in the AFHS can be challenged, and some specific caveats have already been mentioned. 28 �TABLE 19. SUMMARY OF FINDINGS* lexicological Effect Suggested by Animal and Human Literature on Dioxin Observation In AFHS Weight Loss o Increased Neoplasia Increased Birth Defects Neurological Changes Psychological Changes Hepatotoxicity Chloracne Cardiovascular Changes Immunological Deficits Endocrine Changes Increased Mortality + + o + + o + o •*•/o "o" indicates no group differences observed "+" indicates group difference observed in expected direction "-" indicates group difference observed but in opposite direction At this time one cannot ascribe the observed group differences to an effect of dioxin. One cannot implicate dioxin for at least four reasons: Ca) the exposure index completely failed to demonstrate any association between increased exposure and increasing adverse outcome; (b) the full clinical profile for dioxin was not realized with the absence of chloracne being particularly noteworthy; (c) uninvestigated confounding variables remain for several of the target clinical endpoints, and resolution of these issues may alter the observed group differences; and (d) the effect of multiple statistical testing is not well defined. However, the AFHS does not exonerate dioxin as a causative agent of these group differences. This conclusion is supported by three reasons: (a) in six of eleven clinical variables, statistically significant group differences occurred, and in five of these six instances the group differences were in the direction of expected dioxin effects; (b) uninvestigated confounding variables remain for several of the targeted clinical endpoints and resolution of these issues could alter group differences; and (c) the currently available exposure index is only an indication of exposure with unknown precision. The overall probability of obtaining the AFHS results under the hypothesis of no group difference is not known. The summary in Table 19 cannot be. used for statistical inference at this time because the table summarizes the results of hundreds of potentially correlated tests of significance. Further clarification of the role of dioxin in human health must await the results of the follow-up phases of the AFHS and other ongoing epidemiologic studies of dioxin-exposed groups. 29 �ACKNOWLEDGMENT I thank Ms Edith E. Wood for her professional help in retrieving and organizing the extensive literature review summarized in this report. REFERENCES 1. Lathrop CD, Wolfe WH, Albanese RA, et al: An Epidemiological Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides: Baseline Morbidity Study Results. USAF School of Aerospace Medicine Report, Available from NTIS, Springfield, VA, 1984. 2. Lathrop GD,'Wolfe WH, Albanese RA, et al: Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides: Study Protocol. USAF School of Aerospace Medicine Technical Report 82-44, Available from NTIS, Springfield, VA, 1982. 3. Lathrop GD, Moynahan PM, Albanese RA, et al: Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides: Base1line Questionnaires. USAF School of Aerospace Medicine Technical Report 82-42, Available from NTIS, Springfield, VA, 1982. 4. McNulty WP: Toxicity of 2,3,7,8-Tetrachlorodibenzo-p-dioxin for Rhesus Monkeys: Brief Report. Bull Environ Contam Toxicol 1977;18,1:108-109. 5. Case AA, Coffman JR: Waste Oil: Toxic for Horses. Vet Clin North Am 1973;3:273-277. 6. Carter CD, Kimbrough RD, Liddle JA, et al: Tetrachlorodibenzodioxin: An Accidental Poisoning Episode in Horse Arenas. Science 19755188:738-740. 7. Chapman DE, Schiller CM: Dose-Related Effects of 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) in C57BL/6J and DBA/2J Mice. Toxicol Appl Pharmacol 1985;78,1 :147-157. 8. Seefeld MD, Peterson RE: Starvation-Like Syndrome and 2,3,7,8-Tetrachlorodibenzo-P-Dioxin: New Ideas on the Mode of Action at the Whole Animal Level, in Proc Conf Environ Toxicol 1982:237-249. 9. Oliver RM: Toxic effects of 2,3,7,8 tetrachlorodibenzo 1,1 dioxin in laboratory workers. Br J Ind Med 1975;32:49-53. 10. Hogdon JA, Marcinik EJ. Unpublished Data. Department of the Navy, Naval Health Research Center, San Diego, 1983. 11. Jackson WT: Regulation of Mitosis, III. Cytological Effects of 2,4,5Trichlorophenoxyacetic Acid and of Dioxin Contaminants in 2,4,5-T Formulations. J Cell Sci 1972;10:T5-25. 30 �12. Kociba RJ, Keyes DO, Beyer JE, et al: Results of a Two-Year Chronic Toxicity and Oncogenlcity Study of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin In Rats. Toxicol Appl Pharmaool 1978;46:279-303. f 13. Kouri RE, Rude TH, Joglekar R, et al: 2,3t7,8-Tetrachlorodibenzo-p-dioxin as Cocarcinogen Causing 3~Methylcholanthrene-initiated Subcutaneous Tumors in Mice Genetically "Nonresponsive" at Ah Locus. Cancer Res 1978;38: 2777-2783. 11. DiGiovanni J, Berry DL, Gleason GL, et al: Time-dependent Inhibition by 2,3,7,8-Tetrachlorodibehzo-p-dioxin of Skin Tumorigenesis with Polycyclic Hydrocarbons. Cancer Res 1980$10:1580-1587. 15. Coggon D, Acheson ED:.Do Phenoxy Herbicides Cause Cancer in Man? The Lancet 1982;1:1057-1059. 16. Hay A: The Chemical Scythe. Lessons of 2,1,5-T and Dioxin. Plenum Press, NY, 1982. 17. Green S, Moreland F, Sheu C: Cytogenetic Effect of 2,3t7,8-Tetrachlorodibenzo-p-dioxin on Rat Bone Marrow Cells. FDA By-Lines 1977;7,6:292-291. 18. Van Miller JP, Allen JR: Chronic Toxicity of 2,3.7,8-Tetrachlorodibenzo-pDioxin in Rats. Fed Proc Fed Am Soc Exp Biol 1977;36,1:396. 19. Seller JP^ Inhibition of Testicular DNA Synthesis by Chemical Mutagens and Carcinogens. ""Preliminary Results in the Validation of a Novel Short Term Test. Mutat Res 1977;16:305-310. 20. Courtney KD, Moore JA: Teratology Studies with 2,1,5-Trichlorophenoxyacetic Acid and 2,3,7,8^Tetrachlorodibenzo-P-dioxin. Toxicol Appl Pharmacol, 1971;20:396-103. 21. Smith FA, Schwetz BA, Nitschke KD: Teratogenicity of 2,3,7»8-Tetrachlorodibenzo-p-Dioxin in CF-1 Mice. Toxicol Appl Pharmacol 1976;38:517~523. *' 22. Lamb JC, Moore JA, Marks TA: Evaluation of 2,1-Dichlorophenoxyacetic Acid (2,1-D), 2,1,5-Trichlorophenoxyacetic Acid (2,1,5-T), and 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) Toxicity in C57BL/6 Mice: Reproduction and Fertility in Treated Male Mice and Evaluation of Congenital Malformations in Their Offspring. National Toxicology Program Report-80-11, 1980. 23. Hanify JA, Metcalf P, Nobbs CL, et al: Aerial Spraying of 2,1,5-T and Human Birth Malformations: An Epideraiological Investigation. Science 1981;212: 319-351. 21. Townsend JC, Bodner KM, Van Peenan PFD, et al: Survey of Reproductive Events of Wives of Employees Exposed to Chlorinated Dioxins. Am J Epidemiol 1982;115,5:695-713. 25. Donovan JW, Adena MA, Rose G, et al: Case-Control Study of Congenital Anomalies and Vietnam Service (Birth Defects Study). Australian Government Publishing Service, Canberra, 1983. 31 �26. Erickson JD, Mulinare J, McClain PW, et al: Vietnam Veterans' Risks for Fathering Babies with Birth Defects. JAMA 198*l;252,7:903-912. 27. Bishop YMM, Feinberg SE, Holland PW: Discrete Multivariate Analysis: Theory and Practice. The MIT Press, Cambridge MA, 1975. 28. Poc.chia.ri F, Silano V, Zampieri A: Human Health Effects From Accidental Release of Tetrachlorodibenzo-p-Dioxin (TCDD) at Seveso, Italy. Ann NY Acad Sci 1979;320:311-320. 29. Elovaara E, Savolainen H, Parkki MG, et al: Neurochemical Effects of 2,3»7,8-Tetrachlorodibenzo-p-Dioxin in Wistar and Gunn Rats. Res Commun Chem Pathol and Pharmacol 1977;l8,3:1«&7-1l9t». 30. Creso E, DeMarino V, Donatelli L, et al: Neuropsychopharmacological effects of TCDD. Boll Soc Intal Biol Sper 1978;5H, 17:1592-1596. 31. Bauer H, Schulz KH, Spiegelberg U: Berufliche Vergiftungen bei der Herstellung von Chlorphenol-Verbindungen [Occupational Intoxications in Manufacturing Chlorphenol Compounds], Archiv Gewerbepath 1961 ;18:538-5S5. 32. Poland AP, Smith D: A Health Survey of Workers in a 2.U-D and 2,1,5-T Plant. Arch Environ Health 1971 ,-22:316-327. 33. Weber G, Luzi P, Resi L, et al: Natural History of TCDD-Induced Liver Lesions in Rats as Observed1 by Transmission Electron Microscopy During a 32-Week Period after a Single Intraperitoneal Injection. J Toxicol Environ Health 1983;12:533-510. 3%. King ME, Roesler AR: Subacute Intubation Study on Rats with the Compound 2,3,7,8-Tetrachlorodioxin. IIT Res Inst Rep to EPA No IITRI-L6073-12. 35. Gupta BN, Vos JG, Moore JA, et al: Pathologic Effects of 2,3,7,8Tetrachlorodibenzo-p-dioxin in Laboratory Animals. Environ Health Perspect 1973;5:125-mo. 36. Bleiberg J, Wallen M, Brodkin R, et al: Industrially Acquired Porphyria. Arch Dermatol 196H;89:793-797. 37. Sweeney GD, Jones KG, Cole FM, et al: Iron Deficiency Prevents Liver Toxicity of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin. Science 1979;204:332-335. 38. Sinclair PR, Granick Ss Uroporphyrin Formation Induced by Chlorinated Hydrocarbons .(Lindane, Polychlorinated Biphenyls, Tetrachlorodibenzo-pDioxin). Requirements for Endogenous Iron, Protein Synthesis and DrugMetabolizing Activity. Biochem Biophys Res Commun 197*»;61,1:124-133. 39. Smith AG, Francis JE, Kay SJE, et al: Hepatic Toxicity and Uroporphyrinogen Decarboxylase Activity Following a Single Dose of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin to Mice. Biochem Pharmacol 1981;30,20:2825-2830. MO. Manis J, Kim G: Stimulation of iron absorption by polychlorinated aromatic hydrocarbons. Am J Physiol 1979;236:E763-E768. 32 �41. May G: Chloracne from the accidental production of tetrachlorodibenzodioxin. Br J Ind Med 1973;30:276-2S3. 42. Bovey RW, Young AL: The Science of 2,4,5-T and Associated Phenoxy Herbicides. John Wiley and Sons, NY, 1980. 13. Albro PW, Corbett JT, Harris M, et al: Effects of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin on Lipid Profiles in Tissue of the Fischer Rat. Chera Biol Interact 1978;23:315-330. 44. Schraittle SC, Edwards HM, Morris D: A Disorder of Chickens Probably Due to a Toxic Feed-Preliminary Report. J. Amer Vet Med Assoc 1958;132:216-219. 45. Cantrell JS, Webb NC, Mabis AJ: The Identification and Crystal Structure of a Hydropericardium-Producing Factor: 1,2,3,7,8,9-Hexachlorodibenzo-pdioxin. Acta Crystallogr 1969;B25:150-156. 16. Jirasek L, Kalensky J, Kubec K, et al: Chlorakne, Porphyria cutanea tarda und andere Intoxikationen durch Herbizide. [Chloracne, porphyria cutane tarda and other herbicide induced intoxications.] HaUtarzt 1976;27,7:328-333. 47. Moses M, Lilis R, Crow KD, et al: Health Status of Workers with Past Exposure to 2,3,7,8-Tetrachlorodibenzo-p-dioxin in the Manufacture of 2,4,5Trichlorophenoxyacetic Acid: Comparison of Findings With and Without Chloracne. Am J Ind Med 1984;5:161-182. • 48. Zack JA, Suskind RR: The Mortality Experience of Workers Exposed to Tetrachlorodibenzodioxin in a Trichlorophenol Process Accident. J Occup Med 1980;22,1:11-14. 49. Clark DA, Gauldie J, Szewczuk MR, et al: Enhanced Suppressor Cell Activity as a Mechanism of Immunosuppression by 2,3,7,8-Tetrachlorodibenzo-p-dioxin. Proc Soc Exp Biol Med 1981;168:290-299. 50. Montovani A, Vecchi A, Luini W, et al: Effect of 2,3,7,8-tetrachlorodibenzo-p-dioxin on macrophage and natural killer cell-mediated cytotoxicity in mice. Biomedicine .1.980(32,4:200-204. 51. Van Logten MJ, Gupta BN, McConnell EE, et al: Role of the Endocrine System in the Action of 2,3,7,8-Tetrachlorodibenzo-p-Dloxin (TCDD) on the Thymus. Toxicology 1980;15:135-144. 52. Clark DA, Sweeney G, Safe S, et al: Cellular and genetic basis for suppression of cytotoxic T cell generation by haloaromatic hydrocarbons. Immunopharmacology 1983.; 6,2:143-153. 53. Nienstedt W, Parkki M, Uotila P, et. alt Effect of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin on the Hepatic Metabolism of Testosterone in the Rat. Toxicology 1979;13:233-236. 54. Hook GER, Orton TC, Moore JA, et al: 2,3,7,8-Tetrachlorodibenzo-p-DioxinInduced Changes in the Hydroxylation of Biphenyl by Rat Liver Microsomes. • Biochem Pharmacol 1975;24:335-340. 33 �55. Bastomsky CH: Enhanced Thyroxine Metabolism and High Uptake Goiters in Rats After a Single Dose of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin. Endocrinology 1977; 101:292-296. 56. Potter CL, Sipes IG, Russel DH: .Hypothyroxinemia and Hypothermia in Rats in Response to 2,3,7,8-Tetrachlorodibenzo-p-dioxin Administration. Toxicol Appl Pharmacol 1983;69:89-95. 57. Rozman K, Rozman T, Greim H: Effect of Thyroidectomy and Thyroxine on 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) Induced Toxicity. Toxicol Appl Pharmacol 198H;72:372-376. 58. Fett MJ, Dunn M, Adena MA, et al: Australian Veterans Health Studies (The Mortality Report), Part I. Australian Government Publishing Service, Canberra, 198U. � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 059 Folder The folder containing the original item. 1587 Series The series number of the original item. Series III Subseries III Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Albanese, Richard A. Description An account of the resource <strong>Corporate Author: </strong>United States Air Force School of Aerospace Medicine, Human Systems Division (AFSC), Brooks AFB, Texas Date A point or period of time associated with an event in the lifecycle of the resource 1988-02-01 Title A name given to the resource United States Air Force Personnel and Exposure to Herbicide Orange, Interim Report for Period March 1984-February 1988 Subject The topic of the resource Air Force Health Study dioxin morbidity trends ao_seriesIII