3
15
106
-
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.
033
Folder
The folder containing the original item.
0620
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/.
Creator
An entity primarily responsible for making the resource
Dwyer, James H.
Date
A point or period of time associated with an event in the lifecycle of the resource
August 1 1982
Title
A name given to the resource
Summary of Proceedings of a Conference on Herbicide Exposure and Reproductive Epidemiology in Viet Nam held April 10, 1982 at the Department of Medical Genetics, Mount Sinai School of Medicine
Subject
The topic of the resource
congenital birth defects
health effects
human exposure
Southeast Asia
ao_seriesIII
-
https://www.nal.usda.gov/exhibits/speccoll/files/original/b16814fdac2f29877f45793f8be6fcb0.pdf
c1e1fcfaf6d2bfd917a866e309f073f4
PDF Text
Text
Item ID Number
°0722
Author
Hatch, Maureen
Corporate Author
American Public Health Association, Washington, D.C.
ROPOrt/ArtidB TltlO Spontaneous Abortion and Exposure During Pregnancy
to the Herbicide 2,4,5-T
Journal/Book Title
Year
Month/Day
February
Color
^
Number of Images
66
DeSCrlptOU NOtBS
Alvin L
Friday, March 02, 2001
- Young filed this item under the category
"Human Exposure to Phenoxy Herbicides and TCDD";
report no: EPA 560/6-81-006, contract no: 68-01-3859
Page 722 of 865
�f
»• » —
V\A-VcK', M Spontaneous Abortion and Exposure during
Pregnancy to the Herbicide 2,4,5-T
American Public Health Association
Washington, DC
Prepared for
Environmental Protection Agency
Washington, DC
Feb 81
U.S.
of Commerce
Technical Information Service
PB81-165789
._
�EPA 560/6-81-006
EPIDEMIOLOGY STUDIES
SPONTANEOUS ABORTION AND EXPOSURE
DURING PREGNANCY TO THE HERBICIDE 2,4,5-T
FEBRUARY 1981
FINAL REPORT
U.S. E N V I R O N M E N T A L PROTECTION AGENCY
OFFICE OF PESTICIDES AND
TOXIC SUBSTANCES
WASHINGTON, D.C.
«««' 01:
NATIONAL TECHNICAL
INFORMATION SERVICE
us D!nii«(m or comma
\
trtMCIKlO. v» 11M
�TECHNIC,'.1 .REPORT DATA
(fletae read Intinittiunt o. t.ic it vent tie tore compleiint.1
3. RECIPIENT'S ACCESSION NO.
a.
i.BtPonrNo.
EPA 56Q/6-Q1-fir>6
ffjff
4. TITLE AND SUBTITLE
1A«?7» 9
.
5. REPORT OATE
February 1981
Spontaneous Abortion and Exposure During Pregnancy to the Herbicide 2,4,5-T
6. PERFORMING ORGANIZATION CODE
• !•
• • . • . * "
8. PERFORMING ORGANIZATION REPORT N(
7. AUTMOR(S)
Maureen Hatch * and Jennie Kline *
•
*
10. PROGRAM ELEMENT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
American Public Health Association
101515th Street, NW
*<
Washington. DC 20005
'
. . .
'
. '
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Toxic Substances
''.'''
1VV
'
•
•
Washington, DC 20460
"
'
' ll.CONTriACT/GRANT NO.
68-01-3859
.
.• •
US. Environmental Protection Agency
dflt M Orcot
,
,,-.
. ..
•"*'*.'..
•'
.- ;.:
,'...'
'
13. TYPE OF REPORT AND PERIOD COVERS'
Final Report
14. SPONSORING uGENCY CODE
'
'
.'
IS. SUPPLEMENTARY NOTES
•
•
.
•
•• . . . ' .
• • • • . -
•
•.
'
•
.
.
•
Columbia University
16. ABSTRACT
The purpose of this study is to identify locations in the United States where community exposure to ihe
phenoxy herbicide 2,4,5-T-can be examined in relation to the risk of spontaneous'abortion, and other
adverse pregnancy outcomes. Prior ecologic studies evaluating the reproductive consequences of exposure
to 2,4^-T have produced equivocal results. la tliis report, \ve consider what research strategies and what
populations may be used to generate evidence that can be more readily interpreted.
*• Issues relating to the nature of the exposure (e.g.,'seasonal use, unknown.dose level and fate in the
environment) and to ascertainment c'f spontaneous .abortions- (e.g., memory bias, variability in medical
service use) are discussed. Two candidate study areas are investigated. Research approaches suitable to each
isite -are set out and assessed for .their potential to provide a powerful, valid test of the relationslvip of
2;4,5-T use to spontaneous abortion.
.
• '
' '
' * " " ' . "
' •
17.
1.
•
.
KEY WOODS AND DOCUMENT ANALYSIS
DESCRIPTORS
2,4,5-T
.
Community Exposure
. Spontaneous Abortion
Reproductive Outcomes
.
Study Design
Statistical Power
/
b.lOENTIFIERS/OPEN E N D E D TEHMS
- •
.
%
.'
—
•
COSATI Field/Croup
Epidemiologic Study •
• Community Study
Spontaneous Abortion
Reproductive Outcomes
Environmental Research Strategies
Phenoxy Herbicides
•
IS. S E C U R I T Y CLASS (IhitHtporll
Unlimited
previous COITION i»
21. NO. <Jf- PAGES
30. S E C U R I T Y CLASS (Thupat*/
18. D I S T R I B U T I O N i T A T f c M i N T
F*tM 2220.1 (R«¥. 4*77)
c.
62
22. PRICE
�EPA 560/6-81-006
FEBRUARY 1981
SPONTANEOUS ABORTION AND EXPOSURE
DURING PREGNANCY TO THE HERBICIDE 2,4,5-T
by
Maureen Hatch'
Jennie Kline 1
1.
Columbia University
Contract No. 68-01-3859
Project Officer
Jane E, Ke'llcr
Health and Environmental Review D i v i s i o n
Office of Pesticides and Toxic Substances
Washington, D.C. 20460
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
�-2-
DISCLAIMER
This project has b<»en funded with Federal funds from the
Environmental Protection Agency under contract number
68"01-3859 The content of this publication does not
necessarily reflect the views or policies of the U.S.
Environm' ot.al Pcotectir-u Agency, nor does mention of trade
names, commercial products, or organizations imply
endorseneut by the U.S. Government.
�-3-
ABSTRACT
The purpose of this study was to identify locations in the United
States where community exposure to the phenoxy herbicide 2,4,5 Trichlorophenoxyacetic acid (2,4,5T) c-mld be examined in relation to the
risk of spontaneous abortion and other adverse pregnancy outcomes. Prior
ecologic studies evaluating the reproductive consequences of exposure to
2,4,5T have produced equivocal results.
In this report, we consider
what research strategies and what populations may be used to generate
evidence that can be more readily interpreted.
Issues relating to the nature of the exposure (e.g., seasonal use,
unknown dose level and fate in the environment) and to ascertainment of
spontaneous abortions (e.g., memory bias, variability in medical service
use) are discussed.
Two candidate study areas are investigated.
Research approaches suitable to each site are set out and assessed for
their potential to provide a powerful, valid test of the relationship of
«
2,4,5T use to spontaneous abortion.
�-4-
ACKNOVLEDGEMENTS
We would like to thank:
...Wendy Chavkin for her help with the field study;
/
...Zena Stein, Mervyn Susser, and the faculty of the Sergievsky
Center for thoughtful discussion of the issues;
...The physicians, hospital and health department personnel
at the sites investigated, and representatives of federal,
state and private agricultural and timber concerns for their
cooperation and participation;
...The American Public Health Association for its assistance
and support;
...Carol Giles and Sarah Starkman for help in preparing
the manuscript.
�-5-
Table of Contents
I.
Background
It.
Objectives
III.
Evaluatioa of Uses of 2,4,5T: Potential for Human Exposure
IV.
Selection of Sites for Review
V.
DataCathering Procedures
VI.
Site Visit 1: Pacific Northwest
VII.
Site Visit 2: Arkansas
VIII.
i
IX.
Summary and Final Recommendations
.
Appendix
'
�-6-
I.
Background
!
In February 1979, the Environmental Protection Agency issued the
results of an investigation into the relation of aerial applications of
the herbicide 2,4,5-Trichlorophenoxyacetic acid (2,4,5-T) it. a forested
area of Oregon and spontaneous abortion in populations residing nearby
(U.S. Environmental Protection Agency, 1979). In the report, it was
concluded that exposure to 2,4,5-T in connection with forest management
increased the risk of spontaneous abortions in the population exposed.
This conclusion was based on:
(1) A comparison of spontaneous abortion ratios (abortions:
livebirths) in the study area compared with ratios in
two nonforested areas of Oregon.
(2) A comparison of the frequency of abortions occurring
in months relating to herbicide use with that occurring
in months unassociated with herbicide use.
\
(3) A correlation of monthly amounts of herbicide applied
and monthly spontaneous aoortion ratios.
The report received considerable attention, partly because human
evidence on this question is scant and partly because the results appeared to concur with evidence from animal studies showing an increased
rate of fetal loss in association with exposure to 2,4,5-T and to its
contaminant 2,3,7,8-Tetrachlorodiben5p-p-dloxin (TCDD). Nevertheless,
the conclusions may have been unwarranted.
Unfortunately, the Oregon study had been conducted under considerable time pressure in order to serve as a guide for regulatory action
in the next spraying season; subsequent evaluation of the data and the
analysis have pointed up several serious limitations.
The following aspects are of primary importance in evaluating this
study:
1.
Herbicide application data had been obtained for only a
portion (25%) of the study area and the assumption was
made that the pattern of use (in terms of amounts and
timing) would be "similar" in the rest of the study area.
2.
Ascertainment of spontaneous abortions was confined to
abortions noted in hospital records. Since in this region,
most care is sought through private physicians, the use of
hospital records will seriously underestimate the
number of abortions.
�-7-
3.
The ecological nature of the study, that is, the
absence of data to confirm that couples had indeed
been at their usual residence during pregnancies, would
prohibit, in any case, strong conclusions regarding the
relationships of 2,4,5T exposure to abortion. At
best, the results could only be taken to suggest a
relationship which would need to be confirmed in a study
carried out at the individual level.
Not only were there significant limits to what could be learned
from this study about the risk to human reproduction in connection with
the use of 2,4,5T in forest management, but clearly nothing could be
inferred about possible risks associated with other types of uses of
2,4,51, such as applications to rangelands, rice, or orchards.
Accordingly, this feasibility study was undertaken to determine
whether and how the relation of 2,4,5-T to human reproduction could be
examined at the individual level of analysis.
II. Objectives of the Study
The objectives of the feasibility study were:
1.
To identify areas with: (a) a potentially high level of
exposure to 2,4,5T; (b) available documentation of herbicide
use; -md (c) an exposed population of sufficient size to
provide a sample of pregnancies large enough for studying
the relation of 2,4,5T to the risk of spontaneous abortion.
2.
To outline preliminary study designs for conducting research
on this relationship in the location(s) so identified.
It was understood that the work would entail both a review of the
original study area (Alsea, Oregon) as a site for an expanded investigation, as well as assessment of other locations which might serve as
alternative or concurrent sites for testing the hypothesis that exposure
to 2,4,5T under typical conditions of use is associated with an increased risk of spontaneous abortion and other adverse reproductive
events. Thus, the first task was to assess the potential for exposure
associated with different types of use of this phenoxy herbicide in
different areas of the country.
III. Evaluation of_Uses of2,4,5T;Potential for Human Exposure
The herbicide 2,4,5T and related phenoxies are defoliants used in
brush and weed control. The principal applications are on: forests;
crops, such as rice, wheat and fruits; brush along roadsides and power
line right-of-way; private pasturelands and gardens; and rangelands. Of
the various types of use rice and timber cultivation seem to hold the
greatest promise for studying the possible effects of 2,4,5-T on the
reproductive health of exposed populations. The limitations associated
with the other uses are summarized below.
�-8-
A.
Wheat Crops '
The phenoxy chiefly used is 2,4-dichlorophenoxyacetic acid (2,4-D)
which is not contaminated with dioxin. The application rate, though
similar to the rate associated with other agricultural uses, is half
that associated with forest use (0.5-1.5 pounds (Ibs) per acre versus
2-3 Ibs per acre); since wheat growing areas are sparsely settled and
wheat fanning is highly mechanized, the size of the population likely to
be exposed is very small; it is not certain that records pertaining to
herbicide applications on wheat are routinely kept.
B. Fruit Orchards
2,4,5-T has not been used often on fruit, although 2-2,4,5tricblorophenoxy propionic acid (2,4,5-TP) (Silvex) is used to some
extent; since the herbicide is applied using air blast sprayers mounted
on trucks, it may be less likely to drift onto nearby settlements than
when applied aerially; farm workers are typically migrants and likely to
be difficult to trace.
C. Roadside and Right-of-Ways
This use of herbicides is extensive, and the application rate is
high (4-10 Ibs per acre); however, depending upon the target vegetation,
compounds other than 2,4,5-T may be used. Applications are tractormanaged, and considered to have a limited potential for drift. Roadside
applications are carried out by county, state, corporate and federal
crews, and the existence and availability of records vary greatly depending on the site.
D. Raageland
The average application rate on range is from 0.5-2 Ibs per acre,
and in. some states is reputed to be as high as 10 Ibs per acre. By
definition, rangeland is very sparsely settled; thus it is unlikely that
large numbers of people are exposed.
E. Pa.sturel.and and Home and Garden Uses
It has been argued that applications of this kind may carry the
highest exposure. There are, however, no records of such applications
and it seems unlikely that accurate data regarding the timing and use of
herbicides in the past could be obtained from personal interview. The
uses which seem feasible to explore and likely to yield relevant c!ata
are the aerial applications of 2,4,5-T in connection with timber production and rice cultivation. Brief descriptions of the conditions and
function of the herbicide use in each of these settings, and a summary
table comparing the potential for human exposure associated with each
use follow.
�-9-
F. Conditions of Use on Timber
When forests are "managed" with the aim of commercial timber production, herbicide use is common. The purpose is to control the growth
of broadleaf plants, or hardwood, which compete with the conifers for
light and moisture. The need for herbicides increases as a function of
precipitation since rainfall encourages the growth of these competing
species. In a managed forest, the acreage is subdivide^ into cutting
units, each of which is treated with herbicide? perhaps 2-3 times in the
life cycle of the species: once for site preparation (The cutting unit
is cleared and the site sprayed to rid it of all growth prior to planting); again 35 years later for "conifer release" (the site is sprayed to
cut back the faster-growing young alder); and perhaps again the following year if "release" was not totally successful. The cutting unit —
which may range in size from 10-«1,000 acres — is not treated again
until the stand has attained maturity and been harvested and the land is
ready for replanting. Both federal and corporate owners report that
they aim to treat 2-4% of their total acreage in any one year. The time
of treatment depends upon weather conditions at the cutting unit but it
typically takes place in the late spring. The average application rate
is 2-4 Ibs of active ingredient per acre, and the herbicide is sprayed
on by helicopter.
G. Conditions of Use on Rice
In rice cultivation, 2,4,5-T is used to control weeds, particularly
indigo, whose presence in the harvested crop reduces its commercial
value by affecting the appearance of the product. 2,4,5-T is typically
applied annually, at an average rate of 1-1.5 Ibs per acre. (2,4,5-T
amine is used rather than the 2,4,5-T ester used in brush control.)
Planting takes place in late April and the rice is grown under
irrigation. Herbicide is applied before the rice head is formed. The
herbicide is sprayed from a fixed-wing aircraft, usually 3-10 feet from
the ground. At the proper time in internodal growth, the paddies are
drained, 2,4,5-T laid down, and then the areas are flooded once again
until harvesting in September and October.
There are short-grain, medium-grain and long-grain varieties of
rice; each nodes at a different rate. Thus, treatment time is a function of rice variety. Owners typically plant a mixture of rice varieties and hence, may treat their acreage at different times during the
season. The average size of a treatment area is 120-150 acres. Most
applications occur in late June or early July, but the range extends
from mid-May through mid-August.
For the sake of the soil, rice is alternated with another crop,
usually soybean, once or twice in a three-year period. Soybeans are
broadleaf plants and hence 2,4,5-T is not used on this crop. Because
these sensitive crops are frequently contiguous to rice fields, appli-
�-lo-
cators do attempt to minimize drift.
»
It is important to note that other pesticides are used in combination with 2,4,5-T in caring for rice crops; thus, populations living
near rice fields may be exposed to several chemicals.
H. Comparison of Rice and Timber Uses in Respect to Factors Affecting
Potential for Hunan Exposure
1
• Mode of Application
In the forest, the applications are made by helicopter at a considerable distance from the ground; in the rice paddies the applications are
made by low-flying fixed-wing aircraft. One might expect that there
would be more drift in connection with the timber applications, but that
smaller amounts of herbicide would reaching the ground.
The mountainous terrain of forests, which makes control of the
spray nozzles difficult, may provide a greater potential for exposure as
a result of drift than the flat rice paddies.
3. Population Density
•
Population is likely to be smaller and less evenly distributed in
mountainous areas than in flat agricultural areas (Census and topographical map data show this to be true.)
4. Sunlight
Oioxin breaks down readily only in the presence of sunlight; there
is less exposure to sunlight under a forest canopy than in an agricultural setting.
5.
Intensity of Exposure
t
The application rate on rice is half that of timber 1 Ib per acre
on average compared with 2 Ibs per acre.
6. Frequency of Treatment
Rice paddies are treated annually while individual cutting units
within a forest are treated only 2-3 times over several decades.
7. Proportion of the Population Potentially Exposed
Given that rice is typically alternated with another crop, we
estimate that from 50-60% of the population residing in rice farming
areas is likely to be exposed in any year. In the forests only 2-4% of
the total acreage is treated in any year, and thus, only a small portion
of the population is likely to be exposed each year.
�-118. Availability of Records
~
ii
Regulation of 2,4,5-T and other phenoxy herbicides has largly been
directed at the protection of susceptible non-target crops. For this
reason, agricultural uses have traditionally been much more closely
regulated than forest uses, and public records of such applications are
fairly complete. There has been far less public regulation of herbicide
applications in timbered areas, where damage from drift onto sensitive
crops is less likely. Maintenance of application records has been left
largely up to the individual owners or commercial applicators, and hence
the quality and availability of documented herbicide use in this sector
varies.
I.
SUMMARY
The uses of 2,4,5-T in rice and timber cultivation represent distinctly different types of exposure to this chemical. The use of the
herbicide on rice may produce a chronic, low-level exposure. Since farm
populations terd to be stable, many individuals living in rice-growing
areas may have had lifetime exposure to accumulating concentrations of
2,4,5-T as well as other pesticides with unknown reproduction risks to
human. Of special interest is the fact that there is a continuing use
of 2,4,5-T presenting the possibility of either a cross-sectional or
prospective study. Although the level of exposure in agricultural areas
is probably lower than in timbered areas, the population at risk is
larger.
In contrast, herbicide use on forests may produce sporadic and
more intense exposures, which occur perhaps 2-4 times in the life of any
individual. Although conditions in forests may lead to larger doses of
exposure than in the rice area, the population at risk is smaller. In
contrast to the residents around rice fields, herbicide-use on timber
has been suspended and its effects can only be studied retrospectively.
IV. Selection of Sites for Review
Queries to individuals knowledgeable about herbicide use on timber
and rice identified the following areas of heavy use:
1)
Timber^ The Pacific Northwest, specifically the coastal
corridor and especially Oregon; The Great Lakes; the
South Central and Gulf Coast regions.
2)
Rice; Mississippi, Louisiana, Texas, Arkansas, and
California.
In the timbeelands, even though-application practices are similar
from place to place, climate and topography vary so that 2,4,5-T use is
heaviest in the Pacific Northwest, with the South ranking next and the
Great Lakes region following.
�-12-
Of the heavy rice-growing states, Arkansas proved to be the heaviest user of 2,4,5-T. Only MCPA is used in California and Texas because
of the problem of drift onto non-target vegetation. 2,4,5-T is used in
Louisiana and Mississippi, but less acreage is planted in rice in these
states than in Arkansas. In Arkansas 2,4,5-T has also been used in a
large expanse of timberlands for more than 20 years.
Thus the Pacific Northwest, both Oregon and Washington and Arkansas
were selected for site visits.
V. Data-Gathering Procedures
A. Basic Approach
,
In Table 1, we outline the types of data we attempted to obtain for
each site visited.
�-13-
TABLE 1
2,4,5-T FEASIBILITY STUDY:
Da^a Gathering Objectives
EXPOSURE
EXPOSED POPULATION
REPRODUCTIVE OUTCOME
UNEXPOSED
Areas of high use of
Size and distribution of
population in sxoosed
2,4,5-T: proportion of use
for forest, crops, roadside, area.
range, right-of-ways.
Estimate of medically
unattended reproductive
events.
Patterns of ownership by
areas and type of use.
Size and fertility of
population of reproductive age.
Where and when in pregnancy
medical attention is sought;
what information is recorded
regarding the event.
Samples of existing data on
2,4,5-T applications, by
type of owner: dates:
locations; amounts;
composition.
Estimates of the number
of pregnancies at risk of
exposure per year.
What information is recorded
regarding the patient (i.e.,
data on potentially confounding variables). SEE APPENDED
INTERVIEW FORMS FOR FURTHER
DETAIL.
Accessibility of exposure
records.
Information relating to
demographic and control
variables:
race
mother's education
urban/rural residence
% illegitimate births
% births to women 19
% births to women 35
Accessibility of records.
Concomitant exposures:
— fungicides
-- insecticides
— other chemicals
Medical care patterns.
Fate in the environment:
possibility of continuing
exposure.
Deferred, except for decision
about whether to draw the
unexposed sample from the
same geographic areas as the '
exposed sample.
�-14-
The organisation of the table reflects the basic approach. In each
area we attempted first to find where there had been heavy use of
2,4,5-T, or 2,4,5-T and 2,4,-D in combination; to identify the major
users; and to evaluate the type and accessibility of records of herbicide applications.
Having determined where herbicide use had been heaviest, we then
gathered information about the population potentially exposed. Our
principal objective here was to establish whether Che residents in or
near areas of herbicide applications provided sufficient numbers of
reproductive events to permit examination of the research question. We
also had to estimate the proportion of reproductive events for which
there might be no official records. In addition to these quantitative
data, we also attempted to construct a social and health profile of the
local population, particularly those of reproductive age. Finally, we
investigated medical servic. use, medical practices, and the nature and
availability of medical records on reproductive outcomes. The questionnaires we developed for physicians and hospitals, inquiring about their
obstetric and recordkeepinp practices, are apended. These data will be
discussed below in connection with the site visits.
VI-
Site Visit 1: Pacific Northwest
In the Pacific Northwest our aim was to reevaluate the original
Alsea area as a site for further study and to assess the research potential of other locations in Oregon and Washington.
A. Exposure Data
1. Patterns of 2,4,5-T Use in Northwest Timberlands; Use of 2,4,5-T in
the Northwest occurs almost entirely west of the Cascade Mountain Ridge,
in the Coastal Range. This is chiefly a reflection of precipitation
patterns. As rainfall increases, so does the growth of the broadleaf
trees like the alder, which compete with conifers for sunlight and
moisture. Average annual rainfall is 12 inches east of the Cascades, 35
inches just west of the ridge, and 80 inches along the coast. The U.S.
Forest Service estimates that 80% or more of all phenoxy use in Oregon
and Washington occurs from Gray's Harbor County in Washington south
along the coast. North of Gray's Harbor there is more hemlock than Douglas
fit, and hemlock has an easier time penetrating broadleaf trees.
2. gpecies Produced: Douglas fir.
3. Application Rate: An average of 2 Ibs of active ingredient per
acre.
�-15-
4. Size of Treatment Unit; The average cutting unit in the Oregon
nationa* forests is about 40 acres. This may mean that depending upon
where one lives, one may be exposed to treatment at multiple sites.
5- Topography; The steep hills and narrow valleys characteristic of
the Coastal Range are thought to promote herbicide exposure through a
funneling effect. There is also a network of steams running through the
mountains which despite buffer strips becomes contaminated with herbicide.
6. Population Density; Since the rugged terrain restricts accessibility, population tends to cluster in the narrow valley bottoms.
However, particularly in the Siuslaw, the slopes are punctuated with
valleys and there is more distribution of population through the forest
than might occur if the drainage were very long and there were fewer
valleys.
7. Land Ownership; Ownership ot the timberlands along the coast and
east to the Cascades is of three types:
(1) Federal (owned either by
the U.S. Department of Agriculture Forest Service or the Department of
Interior Bureau of Land Management); (2) State; and (3) corporate.
Following is a Forest Service map of the area, showing county
boundaries and locations of national forests. In counties like Coos,
Clatsop and Tillamook, and parts of Lincoln, the forests are owned by
corporations or the state. Gray's Harbor, Pacific, and Lewis counties
in Washington, considered to contain extremely productive timberlands,
are largely the domain of corporate holders. In the strip just to the
east, between the coast and the Cascades, much of the timbered areas
belongs to the Bureau of Land Management and to smaller private holders.
Both Federal agencies consider themselves minor users of phenoxies
compared to the state and corporate owners because of the Federal governraent's more cautious policy regarding the use of chemicals.
8. Availability of Records, by Type of Owner
a
* Federal: Federal agencies expressed an interest and a willingness
to cooperate. Samples of the forms used to document herbicide applications were obtained. These records are said to be kept on file for a
minimum of five years.
b- Oregon State: While state officials view themselves as closely
allied to the private sector, they are willing to cooperate in "responsible" research. Records are said to be accessible and are maintained
for a minimum of three years. Sample forms were obtained.
c. Washington State: This agency reported that, with the exception of
the last few years, neither locations, amounts nor formulations of
herbicide applications are accurately reported in their records.
d. Corporate Owners: The Washington State Forest Association represents
�-16-
all of the commercial users. In order to gain access to the data on
herbicide applications of the corporations represented by the Association it will be necessary to obtain approval of the proposed research
from the Association.
9. Summary
A. Pacific Northwest as a Site for Study
In regard to exposure, the following points oust be addressed in considering the Pacific Northwest as a site for study:
a. the low frequency of exposure;
b. seasonality of applications; and
c. documentation of herbicide use.
a. Exposure Frequency; The potential for human exposure associated
with forest use of 2,4,5-T may be high, but because the portion of
forested land treated in any year is small, it will be difficult to
develop the sample size of exposed pregnancies sufficient to give a
study reasonable statistical power.
b. Seasonality; The use of 2,4,5-T on forests occurs in the spring,
and sometimes in the summer. Therefore, exposure as defined by application practices is seasonal. It is not clear whether the frequency of
spontaneous abortion varies with season, since this question has been
examined oaly on occasion and never in rural areas of the United States.
McDonald (1971) in Canada found a slight, but nonsignificant, increase
in the loss of pregnancies conceived in March-June which she related to
winter-spring infectious. Kelson et al. (1971), of the Center for
Disease Control, reported no season::! pattern in a three-year series of
aborted fetuses. A study of seasonality in relation to reported fetal
loss in New York City, which included spontaneous abortions less than 20
weeks gestation, observed a trough for August-November conceptions and a
peak among January-April conceptions (N.Y.C. Dept. of Health, 1971).
These variations were slight and were confined to early losses and to
the Puerto Rican and non-white. In our hospital-based study in New York
City, we h-ave not detected a seasonal trend in the frequency of spontaneous abortion. Since these results were all obtained in urban areas,
it is possible that a seasonal trend in abortion exists in rural areas.
Thus it is prudent to control for season (month) of conception in any
design to investigate the relation of herbicide exposure to abortion.
Cl
Do cumentat ion; The availability and accessibility of data on herb
icide applications present a potential problem, especially in the state
of Washington.
�-17-
B. Population at Rx-jk of Exposure
Vital statistics, population data and discussions with state and
county health officials formed the basis for understanding the general
health and habits of the population which might be exposed to spraying,
as well as their fertility rate and patterns of medical service use.
The demographic and spray data were used to estimate the number of
pregnancies per year at risk of exposure.
To supplement visual impressions, U.S. Geological Survey dwelling
maps were obtained to determine the distribution of population throughout the forested areas.
Anecdotal evidence was gathered regarding the attitudes of local
residents toward herbicide use and toward research on its relationship
to health effects.
The following issues bear on the suitability of the Pacific Northwest coastal region for research on the reproductive effects.of herbicide use:
1. the lifestyle of residents;
2. possible reporting bias; and
3. potential sample size.
*• Ei fgstyle: One criticism of EPA's Alsea study was that the lifestyle of the population living in the forested areas of the Pacific
Northwest particularly that of recent and younger residents, might have
operated to confound the analysis. It was claimed that this alternative
lifestyle involved drugs, diet and habits of hygiene that themselves
were risk factors for spontaneous abortion. In another vein, local
informants have noted that among this group there is a trend to alternative health care that might make it difficult to document pregnancy
outcome.
On the first point, it was not possible to develop hard information. Impressionistic evidence, however, suggests that under the
rubric "alternative" fall a variety of lifestyles which may, or may not,
incorporate drug use or extreme diets which, in turn, may or may not
increase the risk of spontaneous abortion. To our knowledge, organic or
macrobiotic diets have not been examined in relation to pregnancy outcome. In our study in New York City, heroin use, but not occasional
marijuana use, was associated with spontaneous abortion.
If one considers the frequency of illegitimate births or month of
firat prenatal visit as indirect measures of lifestyle, then statistics
for counties along the coastal corridor suggest there is considerable
variation. In one county for instance, 67% of women have s«Jc a pre-
�-18-
natal visit by the second trimester of pregnancy and only 6% of births
are illegitimate. The comparable figures for another county located are
37% and 17%.
Whether and how an alternative lifestyle affects the risk of spontaneous abortions may be a matter for speculation; however, it is unlikely that lifestyle varies with exposure. If a sample is drawn solely
from the forest area, it is reasonable to expect that the unknown effects
of "lifestyle" will be similar among exposed and unexposed residents.
2. Reporting Bias; In Oregon particularly, but also in Washington and
Northern California, herbicide use in forest management is a major local
issue pitting those with economic ties to the timber industry against
environmentalists. The clash of attitudes gets considerable coverage in
the local media, and groups have formed to promote the opposing
positions' e.g., Women for Timber and Citizens Against Active Sprays.
Although we were not able to examine the question of bias systematically, clearly attitudes have been formed regarding the health effects
of herbicides, giving grounds for concern about the potential for biased
reporting. If overreporting were the only possibility, one could confront the problem by requiring validation of all pregnancies reported.
Local health officials currently believe, however, that health problems
may be underreported because local residents fear the loss of their
economic base. The risk of this bias, which is far more difficult to
assess since one cannot check records on pregnancies that are not re-•
ported, argues against collecting reproductive information by interview.
3. Potential Sample Size; In designing a study, we would require
a sample of sufficient size to have at least an 80% chance of detecting
an increase in risk of predetermined size. All other things being
equal, as sample size decreases, so does the statistical power to detect
an effect, if it is indeed present. The sample size for a study of
2,4,5-T exposure and reproductive outcome in the Pacific Northwest is
likely to be small for several reasons:
a. The forested areas are sparsely settled, and there is little point
to including more populous urban centers in the sample since they are
not at risk of exposure and hence cannot be informative as to the presence or absence of effects.
b. Only a proportion of the population is susceptible to the effect
under study; that is, only the population of reproductive age.
c. Since .only a small percentage of total forest area is treated with
herbicides in any one year, only a correspondingly small proportion of
the population can be exposed in any year. Estimates of sample size
have been developed below as part of the preliminary study design.
�•19-
C. Reproductive Outcome Data
Physician records iu two sample counties (counties A and B) in
Oregon were reviewed in order to assess their utility either as a source
for ascertaining pregnancies or validating reports of pregnancy obtained
through interview. The attempt was made to interview virtually all
physicians who were, or had been, treating obstetric patients in the
sample counties, rather than visiting a portion of the physicians in a
larger number of counties. This approach seemed particularly useful
since the.strong possibility of reporting bias in the population suggested that a review of medical records, rather than personal interviews, was likely to comprise the data source.
1. Data-Collection; Appendix 1 contains the interview schedules which
were administered to physicians and to hospital personnel. In each
office we asked to review a random sample of charts to note the scope
and legibility of information included and to determine whether shortcuts existed to locating pregnancies within the chart. The central
questions we sought to answer were:
a.
What proportion of local obstetric care does tiie office deliver?
This information helped to weight the data from each office when
we attempted to describe the characteristics of the entire
county.
b.
What number of deliveries and of spontaneous abortions are treated
by this facility each year?
This information was useful in estimating the potential sample
size.
c.
How soon after their last menstrual period are women usually seen
for diagnosis of pregnancy?
This information helped us gauge how many early spontaneous
abortions might be missed in a medical record review.
d.
How often, if ever, are spontaneous abortions documented in the
medical records?
In particular, we sought to determine whether a telephone report
of a miscarriage, in the absence of an office visit, would be
recorded.
e.
What information is routinely obtained regarding the characteristics
of the couple who have experienced a spontaneous abortion?
Information on this question provides knowledge of whether
records could supply the necessary data on potentially confounding variables.
f.
What proportion of women experiencing spontaneous abortions are
hospitalized?
�-20-
g.
Lastly, would the physician agree to a record search were a new
study to be undertaken?
Following are background information and a summary of findings from the
medical records survey.
County A
County B
Long, narrow coastal county.
Population = 31,000
Smaller but more populous than
county A. Population = 68,000
Economy: depends on tourism,
tio.ber, fishing.
Economy: Somewhat higher
socioeconomic status than county
A; economy is university and
industry related.
Medical facilities: Three small
hospitals of 40-50 beds each.
Obstetric patients are treated
by general practitioners; almost
all obstetric admissions to one
hospital are welfare patients.
Medical facilities: One large
hospital of 160 beds. About 15
physicians have seen obstetric
patients at some time over the last
10 years. Included here are nine
obstetricians and six general
practitioners.
In 1978, there were 471 births
(15.1/1000); 270 (60%) of these
took place within the county. Of
the 60%, 9% occurred out-of-hospital.
In other words, some 50% of births
were not attended by doctors from
County A. It is believed that most
of the obstetric cases going out of
county A deliver in an adjacent county.
Those delivering in-county tend to be
less affluent; the ilJegitmacy rate
among these women is 27%.
In 1978, there were 804 births
(11.7/1000) to county residents,
yet almost 1200 births occurred
within the county, presumably
because the obstetricians provide
care to women from other counties.
About 3% of births occur out-ofhospital; illegitimacy rate is 6%.
�-21-
2. Conclusion: From our tally of the questionnaire data, we draw the
following tentative conclusions:
a.
The great majority of physician records would be open to us if we
were to undertake a study.
b.
At least 90%, probably close to 100%, of livebirths could be
ascertained or validated from these records since even those
delivering out-of-hospital usually have at least one prenatal
visit.
c.
The major loss of information on reproductive events, should
information be confined to these records, would be:
1.
Very early, recognized but unreported spontaneous abortions,
believed by physicians to be a very small proportion of all
spontaneous abortions.
2.
Early spontaneous abortions, reported but not preceded by a
prenatal visit: It appears from the data that as many as 50%
of spontaneous abortions prior to 12 weeks gestation are
either not recorded in the medical record or are recorded, but
with little data.
d.
It appears that data on potentially confounding variables (i.e.,
maternal characteristics) will be reasonably complete in these
records from about 1970 on for term pregnancies, but incomplete
for pregnancies ending in spontaneous abortion.
e.
A survey of all pregnancies would be possible from these charts,
with the limitations mentioned above, but would require a review
of the entire chart for every woman in order to identify all
spontaneous abortions. Alhough many uf the records have a
separate obstetric page, it may not contain information on
spontaneous abortions, but only data on pregnancies going to
term.
f.
Physician records should extend data on spontaneous abortion
considerably beyond that obtained through hospital records.
D.
Development of a Study Design in the Northwest Forests
The development of a research design to examine the relation of an
environmental factor like 2,4,5-T to reproductive outcome is contingent
both on the hypotheses to be tested and the special characteristics of
the population to be studied. We consider below: (1) tne previous
literature regarding the relation of 2,4,5-T to reproductive outcome and
the hypotheses generated by this work; and (2) special characteristics
relating to exposure and data collection in the Pacific Northwest.
Based on these considerations, we .then set out the conclusions we have
�-22-
rcached regarding study design for this location, and discuss the
strengths and limitations of this approach.
1. Review of Literature and Hypotheses
a. Route of Exposure to the Parent; Exposure via ingestion is the only
route which has been examined in laboratory studies of 2,4,5-T and/or
TCDD. Exposure via inhalation and skin absorption — the routes of
exposure one would associate with herbicide applications —has not been
studied in animals, so it is not known whether response varies with the
route of exposure.
Whether and to what extent 2,4,5-T remains in the environment following
spray applications is difficult to gauge. There may be continued exposure via ingestion of contaminated food or water; however, this is
difficult to determine since residue data are sporadic, conflicting and
subject to reporting bias. In the absence of data on the presence of
2,4,5-T in the water and food, it will be necessary to define exposure
as residential proximity to aerial application sites and to infer the
route and timing of exposure from spray data.
b. Route and Timing of Exposure to the Conceptus: We reviewed laboratory and epidemiologic evidence relating either TCDD or 2,4,5-T to
reproduction in order to determine whether the route to the conceptus is
likely to be through the female parent, the male parent, or both.
A paternal influence has been suggested by the allegations of
Vietnam veterans exposed to Agent Orange, and by unpublished data which
compared the occurrence of birth defects among offspring born to exposed
Vietnamese fathers and unexposed mothers with that among offspring of
couples where both parents were unexposed to herbicides (Tung et al,
1979). Although the data are very suggestive, a systematic investigation has not been carried out, and firm conclusions can not be drawn.
As yet, there is no published report in which exposure of male animals
to 2,4,5-T has been examined in relation to reproductive outcome. A
"three-generation rat study does not permit evaluation of a paternal
effect separate from maternal effects (Murray et al, 1978).
However, the question of male influence has been examined in relation to TCDD. Unlike the study of 2,4,5-T, a three-generation study of
TCDD included a cross-mating experiment, where the effects of both
maternal and paternal exposure were assessed in rats (Smith et al,
1979). Following exposure of female rats, nonsignificant trend to
fever implantations and a significant increase in resorptions were
observed; exposure of the male was not associated with infertility or
pre- or post implantation loss. Negative results have also been reported for exposure of the male on the dominant lethal test in the mouse
and in the rat; however, this test is insensitive to all but strong
mutageus.
�-23-
At this time, it is not possible to rule out the possibility of an
effect operating through the father as well as through the mother. This
suggests that exposure prior to conception must be considered as well
as exposure during pregnancy. When a paternal influence is
hypothesized, the relevant period for exposure in studies of reproductive effects is
sometimes defined as the approximately three mouths it takes for sperm
to fully regenerate.
In regard to the female, the animal studies establish the presence
of effects in association with exposure during pregnancy but do not allow
one to determine whether preconception maternal exposure of the female,
separate from exposure during gestation, has adverse reproductive
effects.
There is no information upon which to judge at what time during pregnancy
this agent might operate. However, since spontaneous abortion, the
major outcome of interest, is by definition fetal death occurring before
28 weeks gestation, we have considered exposures only up to week 24 of
gestation following the last menstrual period.
c. Sumnu'ry; In the absence of strong evidence that a reproductive
effect of 2,4,5-T is confined to one or the other parent, or a
particular time period, we propose to examine whether exposure *~o
2,4,5-T (defined by proximity to aerial applicatons) occurring to either
parent from six months prior to conception until 24 weeks after the last
menstrual period (LMP) is associated with spontaneous abortion. As we
discuss below, our definition of exposure prohibits our distinguishing
between maternal and paternal preconception exposure in the proposed
Study. We aim, however, to evaluate the effects of preconception «:nd
post conception exposure separately.
2. Conditions Affecting Study Design Decisions: Based on the nature of
the 2,4,5-T use under consideration and of the candidate study area,
certain factors emerge which influence the design of the study. These
factors have been described in the preceding text.
Following is a table which sets out the factors or problems to be
dealt with, and the corresponding decision about study design which was
made to cope wiJi each condition This table is followed by:
a. a statement of the hypothesis to be tested in such a study;
b. a discussion of the study design options;
c. estimates of the sample size potentially available for study and
a description of the analysis;
d. a discussion of tlie statistical power of the analysis; and
e. Conclusion.
�-24-
Table- 2
Conditions Affecting Design Decisions for 2,4,5-T Study in
Pacific Northwest Forests
Condition
1. Suspended use of herbicide2. Possibility of reporting bias
3. Different period at risk of
exposure for spontaneous abortions and livebirtbs
A.
Design Decision
T-Retrospective study
Indirect measure of exposure
Record review rather than interview
as source of data
. Control length of gestation to
' equalize period at risk
Case-control rather than cohort
design
Possible confounding effects of,,
lifestyle
5. Low prevalence of exposure-
Restrict analysis to forest
-} population
6. lack of data on potentially,
confounding variables
7, Seasonality of exposure (andpossibly of outcome)
Control for time of conception
Extend study area down coastal
I corridor
"8. Small sample residing in forest——>> Maximize time period of study
2^ Increase size of comparison
population
�-25-
a. Hypothesis to be Tested: The hypothesis which could be tested in
the Pacific Northwest timberlands is whether acute exposure to 2,4,5-T,
either prior to conception or during pregnancy, is associated with
spontaneous abortion. Although it is theoretically possible for the
same agent to have an effect at both periods, it is more likely that an
effect will be specific to one or the other time in i. elation to conception and hence it is desirable to examine the questions separately.
In addition to spontaneous abortion, other outcomes of pregnancy
such as birthweight and birth defects could be studied in relation to
2,4,5-T exposure.
b. Discussion of Study Design Options
1. Retrospective Data Collection: The use of 2,4,5-T in forest management has been suspended since after the 1978 spraying season. Therefore, any study of reproductive effects in timberlands must be retrospective, attempting to relate history of reproductive events to history
of exposure. E <en if reliable and noninvasive biochemical measures of
2,4,5-T existed, the retrospective design obliges use of an indirect
measure of exposure. The available definition of exposure — residential proximity to aerial application sites — serves as an indicator
that the potential for exposure existed: clearly absorbed dose could
vary widely in a population so defined as "exposed." The possibility of
misclassifying exposed and unexposed individuals when using this operational measure of exposure should be borne in mind.
2- Source of Data on Pregnancies; The major determinant of study
design has been the threat of reporting bias in forest communities.
Given its importance, it is regrettable that we have no firm evidence
that the threat is real. The fact that the relation of herbicide use to
reproductive health has already been studied in the Northwest, however,
argues strongly for use of a convincingly bias-free source of data in
subsequent research. Thus, we have selected physician records rather
than interviews as the method of data collection.
The decision to use physician records to collect data pertaining to
all reported reproductive events has strengths as well as limitations.
With this data source, neither biased responses nor a high nonresponse
rate can affect study results. Using medical records as the data source
means, however, that some recognized spontaneous abortions may be missed; that data on potentially confounding variables may not be present
consistently enough to permit analysis; and that confirmation of residence and additional exposure information which might be obtained by
interview must be forsaken. Since the residential information contained
in the chart may not accurately reflect residence during spraying, when
couples may have been away from home, the proposed analysis using place
of residence to evaluate exposure status is ecologic. The potential for
error is greater for full-term pregnancies, where more absences from the
�-26-
home may occur in the nine gestation months, than for spontaneous abortions, which are of shorter gestation.
3' Case-Control Design| The difference in length of gestation for
spontaneous abortions and for term pregnancies has other more serious
implications for study design. Ordinarily, as in this instance, when
exposu re to the factor under study is less common than the outcome of
interest (the prevalence of exposure in the population is estimated to
be about 2-4%, while the usual rate of spontaneous abortion is about 15%
of recognized pregnancies), a cohort analysis provides the most powerful
test of the study hypothesis. Special problems arise, however, in
applying this approach to the study of spontaneous abortion in relation
to an exposure occurring during pregnancy. Si. v ^ >.he period at risk of
exposure increases with increasing length of pregnancy, th.?re is a
built-in bias to define term pregnancies as "exposed", and thus, associate exposure with the satisfactory outcome.
A case-control analysis of the data, where cases are pregnancies
ending in spontaneous abortion (fetal death at 28 weeks gestation or
earlier) and controls are pregnancies ending after 23 weeks gestation,
can meet the problem outlined above by considering exposure in the
control pregnancy only up to the date of abortion of the referent case.
It would be impossible in the small Oregon sample to equalize the period
at risk for births and abortions in a cohort analysis. A lifetable
analysis in days (or weeks) of gestation would be needed to ensure that
successful pregnancies were not at greater risk of exposure than pregnancies ending in abortion, because of their increased gestation. The
number of pregnancies in the Pacific Northwest forest area are too few
to contemplate this approach. Failure to control gestation in relation
to risk of exposure will not merely dilute an effect, if one exists, but
will actually bias a study against detecting an effect. Thus, the gain
in statistical power that would ordinarily result from organizing the
data into cohorts of exposed and unexposed pregnancies is offset by the
inherent bias that exists in studying spontaneous abortion in relation
to an exposure which occurs during pregnancy. This argues for a casecontrol analysis. Since age is associated with abortion and often with
habits, such as smoking, which also are associated with abortion, we
propose also that cases and controls be matched loosely for maternal age
( 30 years, 30+ years).
4-6. The Forest Residents: The decision to restrict analysis to
comparisons of women living in forest areas provides an indirect control
for the unknown effects of lifestyle on reproduction. If the alleged
higher risk of spontaneous abortion in timbered areas as compared with
other areas is truly a reflection of the population's lifestyle and not
of their exposure to herbicides, then comparisons made within forest
areas, where lifestyle and service use patterns should be similar, will
reveal this. If, on the other hand, herbicide exposure does have an
effect, that too should be detectable given that herbicide applications
within.the forests are highly variable with respect to place and year.
�-27-
The focus on forested areas also maximizes the prevalence of exposure by
excluding noninformative pregnancies — that is, those that are never at
risk of exposure. It also compensates somewhat for the lack of systematically collected data on potentially confounding variables such as
smoking, alcohol drinking and drug use because of the similarity in
lifestyle of the residents.
7. Seasonality: In order to control for the potential confounding of
the effects of season of conception with timing of spraying, cases and
controls should be matched for month of LMP so that the frequency of
exposure can be compared among pregnancies conceived at similar periods.
c. Estimates of Sample Size and Description of the Analysis
The limiting factor on sample size is the number of spontaneous
abortions occurring in the Oregon coastal forests over a five-year
period. To boost sample size, we have extended the study period to five
years, the longest period for which exposure data are reported to be
reliably available, and have considered the study area to be the entire
coastal corridor of Oregon forests.
The estimated number of cases was generated as follows:
1. The number of livebirths occurring in 1978 was calculated for the
population falling within the Oregon coastal forests: n=l400 livebirths.
2. The number of spontaneous abortions occurring each year was estimated assuming that livebirths represent 85% of recognized pregnancies:
n=241 spontaneous abortions.
3. Based on results of our medical records survey, it was assumed that
as few as 50% of ths abortions occurring would be found (with usable
data) in physician records. (This assumption may be somewhat strigent.)
n=120 spontaneous abortion found.
4. The number of ascertained abortions per year was multiplied by five
to get the total for the study period.
We estimate that 600 abortions will be ascertained through a review
of all physician records in the forest region.
Since spraying occurs only in the spring-summer, pregnancies occurring at different times of the year vary in the risk of exposure either
prior to conception or during gestation. We propose that the exposure
status of each conception, whether a case or control, c<<Lld be described
in one of five categories: exposed 4-6 months prior to LHP; exposed 1-3
months prior to LMP; exposed in the 1-3 months following LMP; exposed in
the 4-6 months following LMP; not exposed 1-6 months prior to LMP or 1-6
months following LMP, that is, not exposed in any of these four time
�-28-
periods. For pregnancies ending within 28 weeks of the LMP, these
categories of exposure status are mutually exclusive, provided spraying
does not occur twice in the same location in any one year. As mentioned
earlier, the gestation of controls considered in assigning exposure
status ends at the time of abortion for the referent case. A contingency table analysis would be used to compare the distributions of cases
and controls with respect to these five categories of exposure.
Given that approximately 4% of the forest is sprayed each year, we
estimate that 1% of the sample will fall into each of the four exposure
categories and the remaining 96% will be unexposed both in the six
months prior to conception and in the first six months of gestation in
any year.
d. Statistical Power
In the next table we have illustrated the size of the effect which
could be detected with 80% power ( =.05; two-tailed) if the effect of
exposure is confined to one three-month period prior to or during gestation and 600 cases are compared with 600 controls. (In the example, we
show the effect as occurring only with exposure during the first trimester of pregnancy.) We have also carried out the same calculation also
assuming two controls were selected for each case. The effect is shown
as the increase in exposure frequency among cases relative to their
controls.
As may be seen from the second column of the table, only large
effects (odds ratio = 4.1, or 2.7 in the case of two controls per case)
could be excluded with confidence if the distributions did not differ
significantly. On the other hand, if exposure had an effect during more
than one three-month period (e.g., throughout the first two trimesters
of pregnancy) an odds ratios of 2.7 could be detected with equal numbers
of cases and controls.
Perhaps it should be mentioned that three factors known to be associated
with spontaneous abortion each have odds ratios less than those set out
in the table. The odds associated with heavy smoking are 1.7; with
drinking daily, 2.5; .and with prior spontaneous abortions, 2.0. Thus,
the effects likely to be detected in the study described here, are
larger in size than those of previously identified risk factors. Also,
it should be mentioned that the sample described here may include multiple pregnancies to the same couple, raising the problem of noni
ndependence in the unit of analysis. If the sample were restricted to
one pregnancy per couple, the sample size, and consequently the power of
the test, would be further reduced.
�-29-
Table 3
ILLUSTRATION OF THE EFFECT SIZES WHICH COULD BE DETECTED IN THE
OREGON FOREST SAMPLE (WITH 80% POWER*) USING ONE CONTROL PER CASE
AND USING TWO CONTROLS PER CASE
Expected %
Distribution
of Sample
% Distribution
Which Could Be
Detected as
Significantly
Different From
Expected If
N cases =600 &
N controls =600
% Distribution
Which Could Be
Detected as
Significantly
Different From
Expected If
N cases =600 &
N controls =1200
Exposed 4-6 months
prio;. to LMP
1.0
1.0
1.0
Exposed 1-3 months
prior to LMP
1.0
1.0
1.0
Exposed 1-3 months
after LMP
1.0
4.0 (odds=4.1)
2.7 (odds=2.7)
Exposed 4-6 months
after LMP
1.0
1.0
1.0
96.0
93.0
Not exposed
*
=,05, two-tailed
94.0
�-30-
e. Conclusion
The statistical power of the analysis might be greater than the
above estimate if:
1. The prevalence of exposure is substantially higher than has been
estimated, perhaps because the topography of the West Coast forests
produces exposures in a larger proportion of population than is suggested by the percentage of treated land. For example, if the prevalence of exposure was as high as 8% rather than the 4% estimated, the
detectable odds ratio for an effect confined to one three-month period
would be 2.8.
•
2. Ascertainment of abortions in physicians' records is better than the
50% estimated; if ascertainment improved from 50% to 70%, the number of
cases would increase from 600 to 840. As an indication of how this
would improve power, the detectable odds ratio for an effect confined to
one three-month period is changed from 4.1 to 3.3.
. 3. Assuming the preceding estimates of exposure frequency and sample
size set out in (d) above are correct, an increase in sample size would
produce an improvement in the power to detect a more modest effect.
This could be accomplished by extending the study to include residents
in the Northern California forests. This last suggestion may be the
most pragmatic. Certainly based on the present estimates of the sample
available, the power of the study set out here is not adequate to search
for moderate effects of exposure to 2,4,5-T (odds ratios of 1.8 or 2.0).
In sum, we consider that the Oregon coastal forests cannot provide
a sample sufficient to test whether exposure to 2,4,5-T has a moderate
association with spontaneous abortion.
In addition to the problem of sample size one must worry about the
further reduction in power stemming from the imprecision of the operational measure of exposure.
Further, the use of medical records as the data source means that
the analysis is essentially ecologic and also requires the assumption
that use of medical care does not relate to risk of exposure.
�-31-
Refereaces Cited
McDonald, A.D. 1971. Seasonal distribution of abortions. Br J Pev
Soc Med. 25: 222-224.
Murray, F.J., Smith, F.A., Nitschke, K.D., Humiston, C.G., Kociba,
R.J., and B.A. Schwetz. 1979. Three-generation reproduction study
of rats given 2,3,7,8- Tetrachlorodibenzo-p-dioxin (TCDD) in the
diet. Tpxicol. Appl. Pharmacol. In press.
Nelson, T., Oakley, G.P. and T.H. Shepard. 1971. Collection of human
embryos and fetuses. In Monitoring, Birth Defects, and Environment;
The Problem of Surveillance, Hook, E.B., Janerich, D.T. and I.H.
Porter, eds., pp. 45-64.
New York City Department of Health. 1971. The Influence of Season of
Conception on Obstetric Casualties. New York: Health Services
Administration.
Smith, F.A., Schwetz, B.A., Murray, F.J., Crawford, A.A., John, J.A.,
Kociba, R.J. and C.G. Humiston. 1973. Three generation reproduction
study of rats ingesting 2,4,5-Trichlorophenoxyacetic acid in the
diet. Unpublished manuscript.
Tung, Ton that, Lang, Ton due, and Do due Van. 1979. The question of
mutagenic effects on the second generation after exposure to herbicides. Unpublished manuscript.
U.S. Environmental Protection Agency, Environmental Studies Program. 1979.
Report of Assessment of a Field Investigation of Six-Year Spontaneous
Abortion Rates in Three Oregon Areas in Relation to Forest 2,4,5-T
Practices.
�-32-
VII.
Site Visit 2: Arkansas
Arkansas was of special interest as a potential study site because
it presented an opportunity to examine the use of 2,4,5-T on both timber
and rice in contiguous areas and thus, in relation to a homogeneous
population. The map which follows shows the areas within the state
where rice and timber are cultivated commercially. Prior to the site
visit, two rice counties (counties C and D) and two timber counties
(counties E and F) were selected for investigation. State and local
officials considered these counties to be typical of the rice and timber
cultivating areas. Our findings are summarized below.
TIMBER
ARKANSAS
�•33-
A. Exposure Data
1.
Timber
The similarities and differences which were noted between the
treated forests in Arkansas and in the Pacific Northwest are descibed
below.
a
- Species produced; In Arkansas, pine is produced, rather than
the Douglas fir of the Northwest. Pine natures more rapidly than fir.
**• Application Rate: Despite the fact that the rate is the same:
Arkansas has a broader spectrum of target species to deal with that is,
2,4,5-T is applied at an average of 2-3 J.bs per acre.
c. Size cf Treatment Unit: Cutting units in Arkansas range from
10-1000 acres, with an average of 200-500 acres. The average cutting
unit in the Oregon national forests is on the order of 40 acres. This
may suggest that the Arkansas population is at risk of exposure from
fewer treatment sites than the Oregon population.
d. Topography; The commercial timber area in Arkansas consists of
gently rolling hills rather than the steep slopes and narrow valleys of
Oregon. This would seem to reduce somewhat the potential for drift, A
network of streams like the one which runs through the Oregon forests
does not exist. In Arkansas forests, the water supply comes from wells
sunk over 1000 feet down. Hence, contamination of water supply is less
likely.
e. Population Density: In comparison to the Pacific Northwest,
the less rugged terrain of the Arkansas Plains produces a higher population density, as sample data indicate, and a more even distribution of
population throughout the forested area.
f. Extent of Annual Treatments; In both Oregon and Arkansas,
owners of timberland claim to treat 2-4% of their acreage per year.
8- Land Ownership: Commercial timberlands in Arkansas are entirely privately owned, mostly by to a few large corporations. There
are two national forests in the northwestern half of the state but they
are hardwood, not pine producers, and thus do not use much 2,4,5-T.
In summary, although conditions in the Oregon forests may lead to a
higher dose level of exposure, the Arkansas forests seem to have a
larger population at risk of exposure.
h. Exposure records; The state of Arkansas regulates the use of
2,4,5-T, 2,4-D, and other hormonal herbicides, and requires that records
of every application be filed with the State Plant Board. Private
applicators, however, are exempted from this regulation. Most corporate
timber operations have their own helicopters and hence qualify as
"private" rather than "public" applicators. Thus, Plant Board regarding
applications of 2,4,5-T on timber are substantially incomplete.
�-34-
The corporate owners, who are banded together into a consortium
called the Arkansas Forestry Association, keep their own records. If
the study design met their approval and confidentiality of data was
guarded, it would probably be possible to arrange to abstract their
records. The proposed concomitant assessment of effects of use in the
adjacent rice area was viewed as a strength of the study design by tne
local timber representative with whom we spoke.
2. Rice
a. Application and Acreage: The agricultural section of Arkansas
is essentially the southeast portion of the state, particularly the
region known as the Delta, located between the Arkansas and Mississippi
Rivers. The principal crops are rice, soybean and cotton. The trend
over time has been to increase the total acreage cropped, with a shift
away from cotton to rice and soybean.
Currently the total acreage harvested in rice is about 1.2
million, with the two sample counties having the largest number of rice
acres: in 1978, 114,000 acres of rice were planted in county C and in
county D, 89,000 acres were planted in rice.
The use of 2,4,5-T in connection with rice cultivation has been
described on pages 9-10 ol this report. There it was noted that rice is
typically alternated with soybean to restore the soil. Like cotton,
soybeans are sensitive to 2,4,5-T; however, a number of other pesticides
are recommended for use on cotton and soybean. In fact, pesticides
other than 2,4,5-T are also applied to rice fields. Thus, the agricultural population is at risk of exposure to many chemicals. The table
which follows lists all pesticides that may be used on rice, cotton or
soybean. It is not possible to state with certainty that these other
agents are without risk to human reproduction.
Table 4
Pesticides Recommended for Use on Rice, Soybean and Cotton
Recommended
Pesticides
Rice
Soybean
Cotton
Ordram
Modown
Propanil
Ronstar & Pronpanil
Propanil 3 & 3
Propanil & Ordram 3
Ordram 10 G
2,4,5-T
Basalin
Tolban
Treflan
Cobex
Prowl
Lasso
Lorox
Sencor-Lexone
Basalin
Cobex
Prowl
Tolban
Treflan
Tolban & Cotoran
Treflan & Cotoran
Bladex 80 W
�-35-
HERBICIDES
FUNGICIDES
2,4,5-T & Propanil
Silvex
Basagran
Basagraa & Propanil
Propanil
Arastan 70%
Busan 30-A
Captan
Difolatan
Dithane M-45
Terra-Coat L-205
Vitavax
Benlate 50%
Metalkamate
Carbofuran
Malathion
Methyl Parathion
Carbaryl
Toxaphene
INSECTICIDES
Treflan or Tolban & Sencor-Lexone
Prowl & Sencor-Lexone
Lasso & Senco or Lexone
Treflan & Vernam 7£
Lasso & Lore;:
Lasso & Sencor Lexone
Surflan & Lorex
Surflan & Seucor-Lexone
Arasaa
Captan
Terra-Coat L-205
Vitavax -200
Terraclor Super X 10-2.5D
Benlate
Mertect 340-F
Acephate
Carbaryl
Methomyl
Methyl Parathion
Toxaphene
Bladex 4L
Cotoran or Lanex 80 W
Karmex or Dynex 80 W
Zorial 80 W
Basan 30
Busan 30 & Demosaa
Deroosan-Thiram
Dexon Oaconil (72%)
PCNB & Dexon •
Terra-Coat L-21
Terra-Coat L-205
Vitavax & Captan
Vitavax & Thiram
Demosan D
Demosan G
Terraclor Super XG
Terraclor Super XD
Terraclor Super XE
PCNB & Captan 30D
PCND & Manab D
PCNB-& Thiraro D
Acephate
Aldicarb
Azinphosmethyl
Carbaryl
Chlorpyrifos
Chlordineform
Propargite
Dicrotophos
Dimethosate
Disulfoton
Endrin
EPN
Fenralerate
Malathion
Methidiathion
Methomyl
Methyl Parathion
Methyl Parathion FM
Methamidophos
Monocrotophos
Permethrin
Phorate
Sulprofos
Toxaphene
Trichlorfon
Compiled from data supplied by the Arkansas Cooperative Extension Service
�-36-
Because susceptible crops are often contiguous to rice paddies,
applicators attempt to control drift. Nevertheless, since many dwelling
units are in or adjacent the fields and crops extend right to the edge of
the towns and along both sides of the roads, there is clearly potential
for human exposure associated with spray applications, even for individuals
living on the edge of town.
Since this use of 2,4,5-T was not part of the EPA suspension process,
2,4,5~T will again be applied in Arkansas rice fields this coining summer.
However, only half as much 2,4,5-T will be available for use as in the
past because of declining stocks; manufacturers have ceased producing
2,4,5-T amine until after the current court action (involving a possible
total ban on use of 2,4,5-T) has been settled.
b. Records of Exposure: As mentioned above, Arkansas requires that
commercial applicators file both a notice of intent to apply the herbicide
and a record of each application, containing information on the client;
time and location of treatment (including distance and direction from
nearest town); wind and air conditions; amount and composition of chemical
used; crop treated; equipment used; and distance from susceptible crops.
These records are kept on file at the State Plant Board for a minimum of
three years and are available for public scrutiny. The Plant Board considers
its records to be quite complete with respect to agricultural use of 2,4,5-T.
3. Summa ry
The following factors are noteworthy in considering Arkansas as a
site for studying exposure of 2,4,5-T in connection with applications on
timber and rice:
Timber: Arkansas provides a larger population at risk of exposure to
2,4,5-T through timber use than does Oregon because commercial timber production
extends over a larger area and the topography is more conducive to settlement,
Rice; The major issues to be considered in studying the effect of 2,4,5-T
in agricultural settings are: (1) Exposure to 2,4,5-T is accompanied by exposure
to many other chemicals, none of which have been systematically examined in
relation to reproductive outcome; this is similar to the problem of multiple
exposures that is frequently encountered in studying occupational cohorts.
Thus, isolation of the effects of 2,4,5-T from effects of other exposures may be
difficult. (2) Populations living near rice fields may be exposed to 2,4,5-T
once or twice every three years. Therefore, a study in this setting offers the
opportunity (and difficulty) to distinguish the effects of lifetime low-dose
exposure from those of sporadic higher-dose exposure.
A great advantage of studying the rice setting is that there is continued
use; thus, cross-sectional or prospective research designs could be considered.
Population At Risk Of Exposure
Two main factors emerge as important in considering the Arkansar
population as a potential study group. First, it appears that sides have
not formed nor have attitudes hardened on the issue of whether pesticides have
adverse health effects. Certainly there is not the local furor in Arkansas
�-37-
that exists on the West Coast and, presumably, not the same threat of
reporting bias, in this largely unresearched population.
Second, the population at risk of exposure to 2,4,5-T in Arkansas
is, as on the West Coast forests, distinctive. In Oregon, the purported
lifestyle of the forest residents is thought to relate to reproductive
risk. In Arkansas the location of herbicide applications coincides with
the location of a high risk obstetric population.
Among the poorest states in the nation, Arkansas has an excess of
adverse pregnancy outcomes. The state's perinatal mortality rate of
20.8/1000 is similar to that of New York City. Its maternal mortality
rate is double the national rate (22/100,000 livebirths v. 11/100,000).
Its rate of teenage pregnancy may be the highest in the United States.
Within the state, indicators such as infant mortality, teenage
pregnancy, and low birthweight show that the population with the worst
reproductive outcome clusters in the southern part of the state; if one
draws a diagonal line from Polk County in the southwest through Little
Rock at the center of the state and up to Mississippi County ou the
eastern border, one finds that counties falling below the diagonal rank
poorly on these indicators. This area also represents the site of
2,4,5-T use within the state. Thus, a high-risk population comprised of
women with diabetes, hypertension, cardiac problems and other obstetric
risk factors, is also the population at risk of exposure to 2,4,5-T.
Because the type of exposure in rice and in timber areas is different, we attempted to discover whether the two populations differed.
Discussions with sociologists and anthropologists at the University of
Arkansas, and with others knowledgeable about the state, suggest that
they are fairly similar. Socioeconomically, the Delta and Plains areas
(i.e., rice and timber areas, respectively, are similarly stratified,
with timber companies and plantations functioning in analogous ways to
produce communities where wealth and poverty coexist. Racially, there
may be slightly fewer blacks in the Plains than in the Delta.
C-
Reproductive Outcome Data
The medical care systems for rich and poor are separate. In addition, prenatal care and care for delivery are also frequently separated, with women receiving prenatal care locally but traveling to population centers to deliver.
1. Public Patients; The separation of place of prenatal care from
place of delivery occurs largely with poor patients, who typically
deliver at the one indigent care center in the state, located in Little
Rock. Prenatal care for the poor, including pregnancy testing, is
offered through local Health Department clinics. Two and a half years
ago, the state set up a fairly extensive prenatal care system throughout
the rural areas, using a cadre of nurse practitioners specially trained
in maternity care. Although poorer patients typically come in rather
late (after the first trimester), the Health Department follows them
closely and keeps a log of their progress on file. Patients reporting a
spontaneous abortion are referred to Little Rock or to general practitioners for care.
�-38-
Until recently, there was a widespread network of granny midwives
operating in the rural areas, but the state has tried to eliminate this
type of care.
2- Private Patients:
Private patients receive their obstetric care either from OB/GYN
specialists located in the urban areas, or from family practitioners.
Specialist obstetric care is at a premium in the state, and private
patients often wait 2 to 3 months to see an obstetrician. (In fact, the
Health Department has begun to offer interim services to such patients.)
Some of the physicians, we have interviewed expressed reluctance about
having their records utilized in research, because of their concern
regarding the confidentially of the patient/chart or because such research might disrupt the already busy office. The Health Deprtment
expressed willingness to share their records with us, and several hospitals offered use of their premises for our research.
3. S mama ry: The following factors emerge as relevant to a study
of reproductive outcome in Arkansas:
(a) Difficulty in finding obstetric data on local women;
(b) Incomplete data on reproductive events in medical records;
and
(c) Prerequisites for access to medical records.
, a Tracing obstetric events of^ local women; Because of the organi()
zation ob obstetric services, women from the same area may seek care
from a number of different sources at a considerable distance from home.
Conversely, the same practitioner may deliver care to women from many
different areas. Thus, the use of records to ascertain the reproductive
events of a certain local population would be difficult and inefficient.
(b) Incomplete data on reproductive events: There may be a considerable underestimate in medical records of reporductive events, especially
of early fetal loss, and especially those occurring to the poor. Poor
patients come for care quite late in pregnancy, later on average than
when most spontaneous abortions occur. Also, because prenatal services
for the poor are a fairly recent development, some women may continue to
rely on midwives or self-care. Opinion varies as to the extent of loss
these factors produce.
(c) Prerequisites for access to medical records: For those reproductive outcomes which are reported in the medical records, it may be
necessary to obtain prior consent from the patient to review the medical
chart.
D. Development^o£ a Study Design in Arkansas: In Arkansas, as in
Oregon, the research design for studying the relationship of 2,4j5-T to
reproductive outcome must evolve from both the hypotheses to be tested
and the characteristics of the study population. We discuss each in
turn, concluding with a description of the approach suggested by the
nature of the site; an estimate of the size of the sample available for
study; and a summary and recommendations.
1. Hypothesis to Be Tested: The use of Arkansas rather than Oregon as
the site for study allows the test of a more comprehensive set of hypo-
�-39-
theses. Three types of exposure could be investigated. In the timbered
area, the population is at risk of a relatively high-dose exposure
occurring very occasionally over a lifetime. Thus this population permits one to examine the effects of a single intense exposure to 2,4,5-T.
The residents of the rice areas are at risk of a low-dose exposure which
occurs repeatedly over a lifetime. In this population, one could examine
two types of effects: the effects of acute, low-dose exposure; ?nd the
effects of chronic and possibly accumulating exposure. Thus, Arkansas
provides a setting in which to study the relationship of pregnancy
outcome to 2,4,5-T exposure which occurs either long before conception
or immediately prior to coucept-tion, as well as during pregnancy. It is
therefore possible to consider whether exposures in childhood or the
very early reproductive years have an impact on the outcome of later,
unexposed pregnancies.
2. Factors Affecting Study Design Decisions: Characteristics of the
types of exposure, the data and the study population have been alluded
to in the preceding text but are set out systematically in the following
table. Next to each condition mentioned, we specify a research strategy
which may answer the problem posed by the condition. A discussion of
the strengths, limitations and unresolved issues associated with the
various study design options follow the table.
�-40-
Table 5
ARKANSAS;
Conditions
Conditions
A continuing, as well as
suspended, herbicide use
available for study.
2. Incomplete and inaccessible
medical records. • ••
Design Decision
Cross-sectional, prospective, or
retrospective approaches could be
taken.
Interviews with residents rather
than records as the primary source
of data.
3. No evidence of potential bias,,
in reporting by residents.
4. Population at risk has high-,
risk obstetric profile.
Restrict analysis to comparisons
within the same population.
5. Possible cumulative ef£ects
from chronic agricultural
exposure.
Compare With a population never
exposed to 2,4,5-T.
6. Problem of multiple
exposures of farm population
Compare to population with similar
exposures but not to 2,4,5-T, or
establish effect first, disentangle
later.
7. Low prevalence of exposure
in timbered areas.
~
8. Seasonality of exposure.9. Different periods at risk
of exposure for spontaneous"
abortions and livebirths.
_ Select samples depending on treatment status during the study years
of interest.
Control for time of conception.
Control length of gestation to
'equalize period at risk.
�-41-
*• Discussions of Study Design Options
1. Possible Cross-sectional and/or Prospective Data Collection:
Be-cause the use of 2,4,5-T in agriculture is continuing, albeit on a
limited basis, it may be possible to collect at least some of the data
on exposure and pregnancy outcome either cross-sectionally or prospectively. The use of systematic drift and residue data would permit
develop-ment of a better index of individual exposure. With respect to
spontaneous abortion, it would be advantageous to arrange to examine the
abortuses since such an examination would add specifity to the description of the outcome (e.g. chromosomally normal conceptus or abnormal
conccptus with specific anomalies, or abortus with morphologic abnormalities). Concurrent measures of both exposure and outcome would certainly
strengthen inferences regarding an association between the use of
2,4,5-T and poor pregnancy outcome.
There is at least .one facility in a rice growing county where a
hospital-based study of spontaneous abortions might be set up. The
number of women treated for spontaneous abortion in this hospital is not
large (N 150 annually), and thus a cross-sectional study limited to this
setting would be inadequate to test the hypothesis that exposure has a
modest effect on the frequency and nature of spontaneous abortion.
If, however, the effect of 2,4,5-T is to induce chromosomal or
morphologic anomalies which are ordinarily rare, then a comparison of
the chromosomal and morphologic data from this hospital with the data we
have obtained in our New York City study should point up this effect.
Thus, depending on the costs involved in carrying out such a study, it
may be a valuable adjunct to a larger investigation based on historical
data about previous pregnancy outcomes.
Given that 2,4,5-T is no longer used in the timber area, a retrospective analysis essentially identical to that described for Oregon is
proposed. We consider it likely that a similar analytic approach would
be taken in the rice areas. A major difference, however, between the
Oregon and Arkansas studies is in the source of the data on previous
obstetric events: record review in Oregon and personal interview in
Arkansas.
2-3. Source of Data on Pregnancies; For the reasons cited in Table
5 aud described above, the preferred approach to identifying pregnancies
appears to be through personal interviews with residents rather than a
search of medical records. Ordinarily, interview data generate a more
complete account of reproductive events than other methods, although it
may not *•* possible to confirm all information. An attempt could be
made, however, to validate the reproductive outcomes for which care was
sought at a medical facility or from midwives and to evaluate the reliability of reports through interviews with husbands, friends.or relatives.
Invalidated and validated outcomes could also be analyzed separately,
and the results of these analyses compared.
In addition to a more complete obstetric history, interviews with
residents provide an opportunity to develop supplementary data on other
sources of exposure and on potentially confounding variables.
�-42-
It is not clear at this time how the sample to be interviewed would
best be selected.and approached. A further feasibility study is needed
to determine: (1) whether sample selection should be based in health
care institutions, other institutions (e.g. church groups) or in entire
cotnm-unities.; (2) whether residents will willingly discuss their reproductive history; (3) whether the racial or regional characteristic of
the inter-viewer will influence the quality of the interview data (opini
ons of local sociologists and anthropologists vary on this question);
(4) whether the respondents can provide valid reliable and detailed data
on previous preg-nancies.
4. Obstetric Risk Factors; The fact that the use of 2,4,5-T within Arkansas coincides with the location of a high risk obstetric population is similar to th« problem faced ia Oregon in regard to the lifes
tyle of the forest residents, and can be dealt with similarly. That is,
the possibility of confounding by such risk factors may be limited by
restricting data collection to homogeneous populations, within which
some members are exposed and others are not exposed to 2,4,5-T. Potentially confounding variables may be controlled statistically in the
analysis.
5- I-ifstime Risk: In order to test whether there are effects
assoc-iated with chronic, but not current, exposure to 2,4,5-T, the
comparison population must necessarily be one that has never been exposed
to 2,4,5-T. It seems likely that the timbered area of the state, where
exposure occurs so infrequently, should generate a suitable control
group for the residents of the rice-growing areas who have been exposed
in the past but not in relation to their current pregnancy. A comparison
between these two groups, who are said to be similar in other relevant
respects, should provide a test of "ever"/ "never" exposure.
A nonpositive finding from such a comparison woul^ provide
convincing evidence that chronic exposure to 2,4,5-T is without long
-term effects. A positivs association, however, would not implicate
2,4,5-T alone since individuals chronically exposed to this herbicide
are also chronically ex-posed to a wide range of other pesticides. In
the event of a positive assoc-iation, data on a population exposed to
similar agents, excepting 2,4,5-T, would be needed to determine whether
exposure to 2,4,5-T was the relevant factor. Perhaps the population in
the Texas rice fields, where MCPA is used in lieu of 2,4,5-T, might
provide suitable controls. In view of the difficulty and expense involved
in carrying out the proposed study, it seems most practical to attempt
at the start to establish that an effect exists in association with
agricultural herbicide use, and defer the problems of targeting the
specific agent responsible.
6. Concomitant Exposures: The multiple chemical exposures
sustained by Arkansas' agricultural population pose less of a problem in
evaluating the effects of acute exposure in the agricultural setting.
Within any one year, 50-60% of the population is exposed to 2,4,5-T.
Exposed and unexposed individuals are similar in terms of their lifetime
exposure to 2,4,5-T and other chemicals. Within any one year, however,
�-43-
tbe other chemical ex-posures of individuals exposed to 2,4,5-T may
differ from those of in-dividuals not exposed to 2,4,5-T. It is apparent
that there is a marked overlap in insecticides and fungicides used in
rice and soybean fanning (see Table 4, pp 34. Thus, any difference in
exposure to other chemicals between residents living near rice (exposed
to 2,4,5-T) and soybean (not exposed to 2,4,5-T) crops may be limited to
one or a few specific herbicides. If the application times for these
different herbicides vary, it may well be possible to disentangle the
effects of 2,4,5-T from those of other exposures encountered during the
same season. Insofar as the findings oo acute exposure are similar in
the rice and timber areas, the inference that an association in the rice
area is linked to 2,4,5-T, rather than to other chemicals, will be
strengthened.
7. Sampling Procedures: Since timber use of 2,4,5-T covers a great
expanse of Arkansas, it would be most efficient to select the study
samples depending on treatment status during the years for which spray
data are available. By consulting records of past herbicide applications, it should be possible to target areas known to have been
treated during the relevant years, thus making the population selected
for study as informative as possible.
8-9. Seasona 1 ity/Length^ of Period^ at^ Ri s k: The seasonality of
2,4,5-T use and the problem of fivebirths and spontaneous abortions
being at risk of exposure for different durations of gestation are the
same problems as were raised and discussed in connection with the Oregon
study, and can be hand-led similarly here; that is, by matching pregnancies ending in abortion and term pregnancies for time of conception
and by curtailing the period used to define exposure in term
pregnancies.
b. Estimates of Sample Size
The number of cases of spontaneous abortions and of term births
(controls) occurring per year in both the 18 counties where timber is
grown commercially and the 13 counties which have substantial acreage
planted in rice have been estimated. As vas described in the section on
sample size determinations for Oregon, the expected number of spontaneous abortions was derived by considering the number of livebirths,
from published vital stat-istics, to represent 85% of pregnancies. (The
expected rate of abortion is estimated at 15% of recognized conceptions.) We consider that at least 80% of women will consent to be
interviewed and thus 80% of abortions will be identified. The numbers
set out below represent the estimated numbers of spontaneous abortions
and terra births for one year. The length of the study poriod has not
been determined; however, the State Plant Board keeps records for a
minimum of three years, and so an upper limit of three years of pregnancies is available.
Although the two exposed areas generate similar numbers of abortions per year, one might need study periods of differing lengths in the
timber and the rice areas. Given that the prevalence of exposure is so
�-44-
diffcrent in the two populations, the power provided by samples of the
same size will vary. Thus the low frequency of exposure in forested
areas requires that a larger sample of pregnancies be studied to test
for a modest effect.
It should also be noted, however, that the forested areas of
Arkansas provide more cases in one year than could be garnered in Oregon
over five years (N = 742 per year in Arkansas, v. N =600 per 5 years in
Oregon.)
Sample Sizes Available
18 Timber Counties: 5,255 births and an estimated 742 spontaneous
abortions per year.
13 Rice Counties: 5,450 births and an estimated 769 spontaneous
abortions per year.
In summary, it appears that Arkansas may provide a suitable site in
which to examine the relation of 2,4,5-T exposure to spontaneous abortions.
VIII.
iiuncviry and Recommendations
We have reviewed several issues bearing on the decision to examine
the relation of 2,4,5-T exposure to spontaneous abortion. Two sites
have been considered: the Pacific Northwest where 2,4,5-T is applied to
timber, aud Arkansas where 2,4,5-T is applied both to timber and rice
iiel'lrf. The primary strategy proposed to study this relation is similar
for both sites. We propose that the frequency of exposure to 2,4,5-T be
compared among women experiencing spontaneous abortions (cases) and
women delivering after 28 weeks gestation (controls). Controls would be
matched to cases for month of last menstrual period and maternal age.
Exposure status ftr controls would be assigned only for that period
which corresponds to the gestation of the matched case at the time of
spontaneous abortion.
Apart from the similarity in the analytic approaches which would be
taken to study the relation of 2,4,5-T exposure to spontaneous abortion
in either setting (the Pacific Northwest or Arkansas), there are differences in the range of hypotheses which may be tested, the source of data
and the sample sizes available, which bear on the choice of a study
setting. Each of these is discussed below.
1. Range of Hypotheses
The hypotheses which may be tested differ depending on whether the
herbicide use is for timber or rice production. In the timber areas in
the Pacific Nortlwest and in Arkansas, previous exposures to 2,4,5-T
were infrequent, brief and possibly of fairly high dose. Therefore, in
either of these areas, the test would be of the relation of an acute
exposure, either prior to or during pregnancy, to spontaneous abortion.
Although it is possible that contamination of food and water supplies
may follow spraying and provide a possible chronic exposure, this route
�-45-
of continued exposure has not been broadly demonstrated. Indeed, the
observation which«drew attention to the possibility that 2,4,5-T exposure is related to abortion suggests that an effect, if present, coincides with the time of spraying. We consider, therefore, that the
hypothesis best tested in a timber area is that acute, rather than
chronic, exposure is associated with spontaneous abortion.
In the rice areas in Arkansas it is possible to examine the relation of both acute and chronic 2,4,5-T exposure to spontaneous abortion.
In each year, approximately 50-60% of the acreige is planted with rice,
and it is a fair inference that approximately the same proportion of the
population may be exposed.
In the agricultural setting numerous pesticides are used, making it
difficult to isolate the effects of 2,4,5-T. Nevertheless, because rice
is alternated with broadleaf plants, there will be a segment of the
population in any year that is not exposed to 2,4,5-T. Thus enabling a
comparison between reproductive experience in a group currently exposed
and a group currently unexposed to 2,4,5-T will be possible. This
approach is similar to the one frequently employed in occupational
studies: the exposed workers are often compared to a group of workers
from the same industry who are exposed to chemicals other than the agent
under study.
A study which examines the effects of both timber and agricultural
use has the advantage that the effects of acute exposure associated with
these two uses may be compared. If the findings from the rice area are
similiar to those from the timber area, where the population is not
exposed to multiple pesticides, the inference that 2,4,5-T, rather than
Some other pesticide, is associated with abortion will be strengthened.
Since the fields sprayed with 2,4,5-T may also be treated with other
herbicides not used in areas which are not sprayed with 2,4,5-T, the
finding of an association of 2,4,5-T exposure with abortion in the rice
areas but not in the timber area will be difficult to interpret. It may
be possible to distinguish the effects of 2,4,5-T from other herbicides
(on which data will also be collected) if the effect of 2,4,5-T is
limited to a .short time interval, either during gestation or immediately
prior to conception.
In Arkansas it is also possible to examine whether long-term residence in a rice-growing community is associated with spontaneous abortion by comparing the .rates of spontaneous abortion in the rice counties
with that in the timber counties. The failure to detect an association
will suggest that chronic exposure to 2,4,5-T is not associated with
spontaneous abortion. \ possible association between residence in the
rice area and spontaneous abortion would not specifically implicate
2,4,5-T because of the concomitant exposures to other pesticides. As
mentioned earlier, in the event of a positive association, it would be
necessary to collect data from another rice-growing area in which
2,4,5-T is not used in order to determine whether the observed effect is
owed to exposure to this pesticide alone.
�-46-
2. Sources of Data
The data source differs depending cm the location (Pacific Northwest or Arkansas) of the study. In the Pacific Northwest, physician
records would serve as the data source for information on pregnancy
outcomes. This data source is unlikely to be affected by the current
debate regarding the merits of pesticide use. This debate would probably affect the quality of interview data. A strength of —:c"*rd-based
data is that it will be possible to distinguish spontaneous abortions
where pregnancy was confirmed either by a pregnancy test or pathologic
examination of the abortus, from those where it was not. There are,
however, several limitations to using this data source: (1) for seme
spontaneous abortions which were known to the physician at the time of
abortion there will not be sufficient data in the record to permit use
in the analysis; (2) pregnancies where medical attention was not sought
will not be included in these records; (3) data on potentially confounding variables (e.g. smoking) will not be available. It is not
possible to estimate the way in which these limitations may affect the
results; this depends in part on whether there is an association between
exposure and the use of medical facilities for care of a spontaneous
abortion. Such an association would arise, for example, if exposure
leads to early spontaneous abortions and a low proportion of women
experiencing early abortions (e.g. 40%) attend physicians. In this
example, the association of exposure with abortion may be missed.
In Arkansas, data on previous obstetric events would be collected
by interview with women. This method of data collection ensures more
complete information on all three types of variables: pregnancy outcomes, including exposure through occupation and home and garden use of
pesticides; and potentially confounding variables. A limitation in
using this method of data collection in this setting is that it may not
be possible to validate all reports of pregnancy outcomes.
Frequently the limitations of record-based data are offset by the
ease with which they are collected. This is not the case, however, in
the Pacific Northwest, where it will be necessary to abstract the obstetric records of all physicians practicing ia the area.
While the proposed interview study in Arkansas will also entail an
extensive effort, the data obtained in this study will be far more
comprehensive than those obtained in the record review study. In regard
to the exposure variable, in the Arkansas study it will be possible to
gather information on residence both prior to and during pregnancy (i.e.
whether vacations were taken) and whether there may have been other
routes of exposure. Furthermore, in the rice areas, data on the drift
of spray and residue in foods could be collected during the spraying
season.
3.
Sajn£leJ5ize
The population available for study ia the Pacific Northwest is
small. In order to develop a sample of reproductive events of sufficient size to insure that the study will have adequate power to detect
�-47-
a moderate effect (odds ratio =2), both the Oregon and Northern California regions must be studied. Conducting a study over a large a-ea
(and across state lines) may have implications for both cost and efficiency, since exposure data would need to be obtained from several
sources. Furthermore, exposure and reproductive .data extending over a
five-year period (1973-78) would be needed; and the quality of these
data may vary with study year.
In Arkansas a considerably larger population is at risk of exposure. For example, in a single year, the forest area of this state can
generate more cases of spontaneous abortions than are expected in the
Oregon forest over a five-year period. The larger the sample, the
greater the power of the study to detect modest effects.
The large population available for study in Arkansas also has the
advantage that only one pregnancy for each couple would need to be used
in the central analysis, although data on all pregnancies to each couple
will be collected. In the power calculations presented earlier for the
Pacific Northwest, we have permitted more than one pregnancy per couple
to enter the analysis. We have not yet resolved whether this approach
is consistent with the assumption that observations must be independent
in order for a contingency table analysis of the type proposed here to
be appropriate. If, on further consideration of this issue, we determine that the assumption of independence requires that only one pregnancy per couple enter the analysis, then the power of the Pacific
Northwest study is even lower than that set out previously. A similar
problem does not arise for the Arkansas study, where, if needed, there
are sufficient numbers of couples to provide the needed numbers of
single reproductive events.
In sum, we consider that the Arkansas timber and rice areas are
likely to permit a more thorough examination of the relation of 2,4,5-T
exposure to spontaneous abortion than the Pacific Northwest timber
areas. Arkansas offers strong advantages both in the range of hypotheses to be tested, the quality of the reproductive and exposure data,
and the size of effects which may be detected with adequate statistical
power. We have not yet evaluated the feasibility of collecting data by
interview in this region, and so a firm conclusion regarding the potential of this site for the proposed study awaits this assessment.
4. Summary
A study carried out in the Pacific Northwest coastal region
would provide an ecologic test of the hypothesis that 2,4,5-T exposure
is associated with spontaneous abortion. Given that the data would be
confined to obstetric events recorded in physician records (possibly
about 50% of all recognized abortions) and that proximity of any individual to spray application site can not be confirmed, the results will be
limited in their implications for a casual relation between 2,4,5-T
exposure and spontaneous abortion. If both the obstetric and spray data
were already computerized and available for linkage, it would be efficient and expedient to carry out the proposed ecologic analysis. However, given that these data must be abstracted in order to carry out
this analysis, the costs of such data collection must be weighed against
�-48-
the limited hypotheses and inferences which attend an ecologic study in
this small population.
In contrast, Arkansas provides a setting in which the relation
between 2,4,5-T and spontaneous abortion might be examined on an individual level. The effects of both chronic and sporadic exposure may be
examined in this state where two major types of herbicide use are found
in adjacent areas.
�-49-
Appendix A
Reproductive Outcome Data
A. PHYSICIAN INTERVIEWS
Name
Address_
Phone
Time and date of interview
For physicians practicing in a group, please note:
number of physicians in the practice
whether procedures and record are uniform_
Period over which the office has done obstetrics:
1. Ask for an estimate of the following data by month or by year,
whichever the interviewee finds easier: (For group practices,
get totals for the group, not just for the physician interviewed.)
total deliveries:
total spontaneous abortions:
seen in office
telephone report only__
total spontaneous abortions hospitalized by this office:_
at what hospital(s)
2. What proportion of obstetric cases in (name the area of_interest)
does this office handle?
�-50-
3. Characteristics of obstetric population:
(a) patient residence
(b) race
(c) education
(d) age
(e) other demographic characteristics
4.
Ask for estimate of the proportion of women in (nam>i area of interest)
with (a) livebirths and (b) spontaneous abortions who do not attend a
hospital, physician, or other medical facility for care. (We need to
gauge what proportion of reproductive events could not be either
ascertained or validated in physician or hospital records.)
4a. What is the basis for this estimate?
�•51-
How soon after their last menstrual period do patients usually
come in fo"r confirmation of pregnancy? (NOrE: IF physician gives a
broad range, ask what percent comes in at the early end and what
proportion towards the end of the range. The purpose of the question
is to find out how many first trimester abortions may not be medically
attended.)
5a. What proportion are confirmed by pregnancy test?
by physician exam?
What proportion
5b. Where pregnancy tests have been used, ascertain:
type of pregnancy test
__
where results are analyzed
where results are recorded
6.
What proportion of patients experiencing spontaneous abortion have
had their pregnancy confirmed before the abortion?
6a. Of that proportion, how many by pregnancy tests? how many by PE?
�-52-
7. What is the office protocol with respect to a spontaneous abortion when:
(a) patient gives self-report by phone, pregnancy unconfirmed prior
to phone call.
(b) patient phones in self-report, pregnancy has been confirmed
prior to phone call.
Is the phone contact recorded iu the chart? Is date of abortion
recorded? Is LIP recorded? What other data are recorded?
Is a patient visit always or sometimes required following report
of spontaneous abortion?
What proportion of patients-reporting spontaneous abortion are
hospitalized? and for what reasons?
Are pathologic examinations done on the products of conception
to confirm pregnancy? When and where?
Pregnancy unconfirmed:
Pregnancy confirmed
�-53-
8.
Hand the interviewee the sheet with the list of variables (checklist
of variables) and have him/her fill out.
9.
How many total charts does this office have?
How are they filed and stored?
10. In this office, is there any way to pull only those charts of women
who have had pregnancies — e.g., an ongoing list of OB patients;
list of those pregnancy tested; list of path exams on POC; etc.
11. Within the chart, is there any quick way to retrieve all data on
pregnancies?
�-54-
12. Would be/she agree to a record search if a study were to be undertaken?
/ 13. Ask whether doctor would be willing to provide us access to names
of patients for the purpose of linking across facilities. (We
would
of course respect confidentiality of patients.)
14. Is amniocentesis available to patients?
reasons?
If so, where, and for what
�-55-
15. FOR AIL INTERVIEWEES: Ask to review 2-5 randomly selected charts.
(In case of a group practice, ask for two charts pe* physician.)
(a) check especially for information on smoking, alcohol, residence,
occupation.
(b) note whether chart:
i. is typed or handwritten; if latter, is it legible?
ii. does or does not have a facesheet?
iii.
is organized from current to past, or vice versa?
iv. provides a continuous history or has separate chart
for each calendar year?
(c) in case of GP, is reproductive data difficult to cull from record?
Comments:
�j
:
*
.
-56-
Appendix B
I
Checklist of Variables
i
Interviewee
Address
Phone__i
|
.
For the following variables, please estimate what proportion of your
records contain this information:
a) for a live or still birth
b) for a spontaneous abortion
DATE OF LMP
DATE OF ABORTION OF BIRTH
OTHER ESTIMATES OF GESTATION
(e.g. iotrauterine size)
;_
MOTHER'S DATE OF BIRTH
FATHER'S DATE OF BIRTH
'
._
MOTHER'S PREVIOUS REPRODUCTIVE HISTORY
,_
RACE OF MOTHER
RACE OF FATHER
•
EDUCATION OF MOTHER
EDUCATION OF FATHER
.'
;
•
OCCUPATION OF FATHER
OCCUPTATION OF MOTHER
PREGNANCY TEST RESULTS
PATHOLOGICAL REPORT ON
PRODUCT OF ABORTION
<p» •**
ZIP CODE OR OTHER INDICATOR
OF PATIENT. ADDRESS
�•57-
MATERNAL SMOKING
MATERNAL ALCOHOL CONSUMPTION
MATERNAL IRRADIATION EXPOSURES
MATERNAL VIRAL INFECTION
SEX OF LIVEBIRTH
BIRTHWEIGHT OF LIVEBIRTK
CONGENITAL MALFORMATION IN
LIVEBIRTH
CONGENITAL MALFORMATION
IN STILLBIRTH
�-58-
Appendix C
Reproductive Outcome Data
B. HOSPITAL INTERVIEWS
Name
Address
Phone
Time and date of interview_
Name of Hospital
Number of beds total_
Number of OB beds
Number of physicians admitting OB/Gyn patients_
1. Ask for an estimate of the following data by month or year, whichever
the interviewee finds easier:
total deliveries:
total spontaneous abortions:
Inpatients
Outpatients_
Clinics_
ER
total induced abortions:
2. Does this hospital have a 'Birth Registrar' - someone who keeps summary
data on births, stillbirths, induced and spontaneous abortions?
�-59-
3. Are fetal death certificates filled out in hospital?
what gestation?
4. Please characterize your OB population in terms of:
(a) residence
(b) race
(c) education
(d) age
(e) other demographic characteristics
Beginning at
�-60-
Please estimate the proportion of women in the area with
(a) livebirths and (b) spontaneous abortions who do not attend
a hospital, go to a phyiscian, or other facility for care.
(We need to gauge what proportion of 'reproductivity events
could not: be either ascertained or validated in physician
or hospital records.)
5a. What is the basis for this estimate?
6.
Are most patients with livebirths or spontaneous abortions admitted
through a private physician, or do some patients have no physician?
7.
a. How are charts filed - by name, unit number, admitting or
discharge diagnosis? It is possible to access charts by
diagnosis?
b. Are the records computerized?
computerized as well?
If so, are outpatient records
c. What data are entered onto the computer?
�-61-
8.
How would one identify every spontaneous abortion that occurred
at your hospital? IF RECORDS ARE COMPUTERIZED: Have you ever
checked whether your computer records pick up every case from the
admissions book, path, log, OB clinic log, etc.?
9.
Hand the interviewee the sheet with the list of variables (checklist of
variables) and have him/her fill out. In the interest of clarity, ask
them to distinguish those data that are coded from those that appear in
the chart only.
�•62-
10. Ask for a printout, to take home, for several years' worth of both
livebirtns and spontaneous abortions, by zip jde, maternal age,
date of IMP, and date of event.
11. If we were to do a study could we have access to their records?
12. Could we have access to patient names for the purpose of linking
across facilities?
13. Do they offer amniocentesis?
If so, for what reason?
��
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.
035
Folder
The folder containing the original item.
0722
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/.
Creator
An entity primarily responsible for making the resource
Hatch, Maureen
Jennie Kline
Description
An account of the resource
<strong>Corporate Author: </strong>American Public Health Association, Washington, D.C.
Date
A point or period of time associated with an event in the lifecycle of the resource
1981-02-01
Title
A name given to the resource
Spontaneous Abortion and Exposure During Pregnancy to the Herbicide 2,4,5-T
Subject
The topic of the resource
human exposure
Oregon
congenital birth defects
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.
035
Folder
The folder containing the original item.
0742
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/.
Creator
An entity primarily responsible for making the resource
Hay, Alastair
Source
A related resource from which the described resource is derived
Nature
Date
A point or period of time associated with an event in the lifecycle of the resource
November 18 1976
Title
A name given to the resource
Seveso: The Problems Deepen
Subject
The topic of the resource
Italy
biodegradation
soil decontamination
congenital birth defects
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.
045
Folder
The folder containing the original item.
1114
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/.
Creator
An entity primarily responsible for making the resource
Tuchmann-Duplessis, H.
Source
A related resource from which the described resource is derived
Bulletin de l'Academie Nationale de Medecine
Date
A point or period of time associated with an event in the lifecycle of the resource
October 19 1982
Title
A name given to the resource
Seveso, predictions et realite
Subject
The topic of the resource
herbicide toxicology
congenital birth defects
Italy
industrial exposure
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.
028
Folder
The folder containing the original item.
0431
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
The Lancet
Date
A point or period of time associated with an event in the lifecycle of the resource
October 3 1981
Title
A name given to the resource
Seveso After Five Years
Subject
The topic of the resource
chloracne
congenital birth defects
civilian impact
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.
070
Folder
The folder containing the original item.
1890
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/.
Source
A related resource from which the described resource is derived
Worldwide Report, Environmental Quality
Date
A point or period of time associated with an event in the lifecycle of the resource
1981
Title
A name given to the resource
Secret Imports of Agent Orange Suspected in Birth Defects Increase
Subject
The topic of the resource
Australia
congenital birth defects
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.
040
Folder
The folder containing the original item.
0926
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/.
Creator
An entity primarily responsible for making the resource
Nelson, C. J.
J. F. Holson
H. G. Green
D. W. Gaylor
Source
A related resource from which the described resource is derived
Teratology
Date
A point or period of time associated with an event in the lifecycle of the resource
1979
Title
A name given to the resource
Retrospective Study of the Relationship between Agricultural Use of 2, 4, 5 -T and Cleft Palate Occurrence in Arkansas
Subject
The topic of the resource
congenital birth defects
dioxin
herbicide toxicology
agricultural exposure
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.
045
Folder
The folder containing the original item.
1150
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/.
Creator
An entity primarily responsible for making the resource
Williams, Margie
Source
A related resource from which the described resource is derived
The Wheat Grower
Date
A point or period of time associated with an event in the lifecycle of the resource
1983-09-01
Title
A name given to the resource
Reproduction Not Affected By Pesticide Exposure NAWG Study Shows
Subject
The topic of the resource
agricultural exposure
congenital birth defects
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.
043
Folder
The folder containing the original item.
1062
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/.
Creator
An entity primarily responsible for making the resource
Smith, Allan H.
D.P. Matheson
D.O. Fisher
C.J. Chapman
Source
A related resource from which the described resource is derived
The New Zealand Medical Journal
Date
A point or period of time associated with an event in the lifecycle of the resource
March 25 1981
Title
A name given to the resource
Reprint with letter: Preliminary Report of Reproductive Outcomes Among Pesticide Applicators Using 2, 4, 5- T
Subject
The topic of the resource
agricultural exposure
congenital birth defects
health effects
human studies
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.
040
Folder
The folder containing the original item.
0909
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/.
Description
An account of the resource
<strong>Corporate Author: </strong>University of Milan; Chair of Obstetric and Gynecological Pathology
Date
A point or period of time associated with an event in the lifecycle of the resource
1982
Title
A name given to the resource
Report: Ultrastructure Findings on Placental Tissue from Pregnant Women Living in the TCDD-Polluted Zone (Seveso) Subjected to Voluntary Termination of Pregnancy between March 1, 1980 and June 30, 1982. Final Report
Subject
The topic of the resource
Italy
congenital birth defects
industrial exposure
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.
0411
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
New Zealand Medical Journal
Date
A point or period of time associated with an event in the lifecycle of the resource
April 23 1980
Title
A name given to the resource
Report to the Minister of Health of an investigation into allegations of an association between human congenital defects and 2,4,5-T spraying in and around Te Kuiti
Subject
The topic of the resource
New Zealand
congenital birth defects
herbicide poisoning
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.
028
Folder
The folder containing the original item.
0447
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/.
Creator
An entity primarily responsible for making the resource
Aldred, J. E.
Description
An account of the resource
<strong>Corporate Author: </strong>Consultative Council on Congenital Abnormalities in the Yarram District
Date
A point or period of time associated with an event in the lifecycle of the resource
September 26 1978
Title
A name given to the resource
Report of the Consultative Council on Congenital Abnormalities in the Yarram District
Subject
The topic of the resource
AVHS
congenital birth defects
ao_seriesIII
-
https://www.nal.usda.gov/exhibits/speccoll/files/original/83e9967701deb14815fe16c14594eaf9.pdf
f2c521d5b52571c48677e07e9a93d151
PDF Text
Text
Item ID Number
00421
Author
Anonymous
Epidemiologic Studies Program, Human Effects Monitor!
RODOrt/ArtiClO TltlO ReP°rt of Assessment of a Field Investigation of SixYear Spontaneous Abortion Rates in Three Oregon
Areas in Relation to Forest 2,4,5-T Spray Practices
Journal/Book Title
Year
1979
Month/Day
February 28
Color
Number of Images
OeSCriptOU Notes
104
Alvin L Youn
9 filed tnis itern under the category
"Human Exposure to Phenoxy Herbicides and TCDD"
Tuesday, February 06, 2001
Page 422 of 526
�RESORT OF ASSESSMEOT OF A FIELD
OF SIX-YEAR SPOnaMEOUS ABORTION RATES IN THREE OREGON AREAS
IN RELATION TO FOREST 2,4,5-T SPRAY PRACTICES
Prepared by the Epidemiologic Studies Program^
Human Effects Monitoring Branch ,
Benefits and Field Studies Division,
\JiOA
February 28, 1979
�TABLE OF CONTENTS
Page
List of Tables
ii
List of Figures
iv
Chapter
I Introduction
II Purpose and Scope
1
6
III Methods
8
IV Results
37
V Summary and Conclusions
77
Appendix
A Questionnaire: Oregon Miscarriage Investigation
80
B Labor Room Statistics, Jackson Memorial Hospital
99
�LIST OF TABLES
Table
1.
2.
3.
Page
Selected Agricultural Statistics for Counties
Including the Study, Control and/or Urban Areas
18
Zip Codes Corresponding to the Study,
Urban, and Control Areas
20
Total Acres Treated and Pounds 2,4,5-T Used
in Alsea Basin (1972-1977)
4.
22
Applications of 2,4,5-T by Day and Month in the
Alsea Basin (1972-1977)
5.
6.
24
Annual Admissions and Births in Hospitals in Study
Urban, and Control Areas
,
30
Number of Hospitalized Spontaneous Abortion Cases for
• Study, Urban, and Control Areas
7.
8.
9.
10.
33
Number of Hospitalized Spontaneous Abortion Cases for
Study, Urban, and Control Areas, Accummulated
by Month..
39
Hospitalized Spontaneous Abortion Cases by Study,
Urban, Control Area, and Age Group of Patient
40
Number and Percent of Spontaneous Abortions According
to Length of Gestation in the Study, Urban and
Control Areas
41
Total Number of Births by Months for the
Study, Urban, and Control Areas
11.
42.
Number of Births Corresponding to the Month of
Spontaneous Abortions in the Study, Urban and
Control Areas
43
11
�LIST OF TABLES
Table
12.
Page
Monthly Spontaneous Abortion Index for the
Study , Urban, and Control Areas ............................. 44
13 . Abortion Index by Period, Area, and Month...................47
14.
Analysis of Variance of Abortion Index by
Period, Area, and Month ..................................... 51
15.
Parameters of the Model "Index « A + B Sin (2J t-D)"
for the various areas.......................T...............57
16.
Study Area Index Corrected for Phase Shifts Study
minus Control — 1 month lag................................58
17.
Study Area Index Corrected for Phase Shift, Study
minus Urban --- 2 month lag....................... ..........59
18.
Study Area Index Corrected for Phase Shifty Study
minus (Urban plus Control)-'—2 month lag. . .................60
.
19.
Number of Births per Month and Unweighted Abortion
Index for the Study, Urban, and Control Areas ............... 64
20.
Cross - Correlation of Corrected Study Area
Index with Spray Data ....................................... 68
21.
Pounds of 2,4,5-T Applied in the Alsea Basin and
Abortion Index for the Study Area by Month and
Period.
22.
Cross-Correlation of Study Area Abortion Index
with Spray Data ............................................. 70
23.
Study and Urban Area Physician Interviews: Estimated
Numbers of Spontaneous Abortions Treated Annually and
Percentages Hospitalized................................... . 74
iii
�LIST OF FIGURES
Figure Number
Page
1.
Oregon
2.
Study Area in Lane, Lincoln and Benton
Counties of Oregon .......................................... 10
3.
Land Cover in Lincoln, Benton, and Lane (Partial)
Counties, Oregon .............................................
11
4 . Control Area in Malheur County, Oregon ...................... 14
5 . Land Cover in Malheur County, Oregon........................15
6 . Urban Area in Benton County, Oregon.........................17
7.
Acres Sprayed with 2,4,5-T in Alsea Basin
Accumulated by Respective Month, 1972-1977 .......... ........ 25
8.
Pounds of 2,4,5-T Sprayed in Alsea Basin
Accumulated by Respective Month, 1972-1977.............. ---- 26
9 . Hospitals in Study and Urban Areas..........................28
10 . Hospitals in Control Area................................'. . . 29
11.
Plot of Monthly Spontaneous Abortion Index .
for the Study, Urban, and Control Areas ..................... 45
12.
Abortion Index for the Study Area by
Month and Period.......................................... . 48
.
13.
Abortion Index for the Control Area by
Month and Period ............................................
49
14.
Abortion Index for the Urban Area by
Month and Period ............................................. 50
15.
Spontaneous Abortion Index, 1972-1977 For the Study Area
and the Combined Control and Urban Areas ..................... 55
iv
�LIST OF FIGURES
Figure Number
16.
Page
Study and Control Areas: a) Combined Urban and Control Index;
b) Study Area Index; c) Urban and Control Index Shifted Back
Two Months
56
11.
Predicted and Observed Points for the Study Area
61
18.
Predicted and Observed Points for the Urban & Control Area
67
19.
a) Study Area Abortion Index Corrected for Phase Shifted
Urban Index b) Pounds of 2,4,5-T Sprayed by Month
65
Cross-Correlation of Study Area Abortion Index
with Spray Pattern in Total Pounds 2,4,5-T applied
by Month, 1972-1977
66
20.
�Chapter I
INTRODUCTION
One of the widely used chlorophenoxy herbicides is the chemical
2,4,5-trichlorophenoxy acetic acid, commonly referred to as 2,4,5-T.
A selective herbicide, 2,4,5-T is especially useful for brush control
on rangeland, along right-of-ways, and in conifer forest habitats.
The parent acid is formulated in a variety of emulsifiable esters
and specific amine salts registered for use in both Canada and the
United States.
Many studies have been conducted to determine the fate of TCDD in the
environment. Since TCDD is reportedly not mobile in soil, it is not
considered to be a risk in ground water. TCDD does not accumulate
in vegetation (1).
In comparison with other pesticide products, the phenoxy herbicides are
relatively non-toxic to mammals, and 2,4,5-T is classified as moderately
toxic with an acute oral LD_
(rat) at 500 mg (acid basis) Ag
(2).
However, low levels of TCDD have been shown to have oncogenic effects
in rodents and related effects in primates. Fetotoxicity and teratogenicity have been demonstrated at low levels, with the appearance of
cleft palate and kidney anomalies in rats, mice, and hamsters.
toxic effects in avian species have also been reported (1).
Embryo-
�In July, 1978, Staff of the EPA Office of Pesticide Programs (OPP) and
Colorado State University epidemiologists met in Oregon with local and
State health officials to begin an investigation into a group of women
living in the vicinity of Alsea who claimed they had experienced
miscarriages because of herbicide spraying in the course of forest
management. The investigation was precipatated in late June, 1978,
when EPA received a letter signed by eight women living in this area
who had experienced 10 miscarriages since 1973.
The women claimed
to be surrounded by forest land which has been sprayed for years with
herbicides known to contain dioxin. They also charted their dates of
miscarriage and related dates the forest areas were sprayed. Each of
the women was under a physician's care at the time of miscarriage, and
neither the women nor their doctors could ascertain the reasons for
abortion. The investigators met with four of the women to discuss the
circumstances of their reported exposure and any possible relationship
to their subsequent miscarriages.
Following the visit by EPA and Colorado State University staff, an
extensive health questionnaire (Appendix A) was designed jointly by
OPP's Epidemiologic Studies Programs (ESP's) in Miami, Iowa, South
Carolina and Colorado, in which detailed questions on self and family
pregnancy and medical histories, environmental aspects, diet, occupational and household pesticide usage were included. During the first
week of August, 1978, an epidemiologist from Colorado's ESP administered the questionnaire to nine women who had experienced 13
confirmed miscarriages from May, 1973 to March, 1978.
�During late August and early September, the health questionnaire and
related data (spray application, vital statistics, etc.) were evaluated
independently by 10 experts in the fields of obstetrics, gynecology,
epidemiology, biostatistics, reproductive endocrinology, and perinatal
medicine. The consensus of the reviewers was that:
1.
the spontaneous abortions did appear to follow a seasonal
pattern (two of the reviewers noted the seasonal relationship
but drew no conclusions);
2.
Good Samaritan Hospital records for Benton County women
for the years 1975-77 showed numerically higher rates of
spontaneous abortions per live births during Jan,, Feb., Mar.,
and Oct., Nov., and Dec., of each year than during April
through Sept. (these records were based on stillbirths of
terms
greater than 20 weeks). Conversely, 10 of the 13 miscar-
riages reported by the Alsea participants occurred during the
months of.April through September (1973-1978);
3.
there was a high numerical incidence of March to June
miscarriages (nine of 13) among the Alsea participants.
However,
there was concern that the reports might comprise a biased
sample (albeit unintended) of all miscarriages that occurred
within the area and years under investigation;
�4.
a causal relationship between forest herbicide spraying
and reproductive wastage had not been demonstrated from the data
presented. Opinions ranged from ". . .no evidence of a
causal relationship . . ."to "I cannot support or refute a
cause and effect relationship ..." Reviewers either stated
or inferred that there was no real evidence of an epidemic
based on the data presented.
An analysis of the data by staff of OPP's Human Effects.Monitoring
Branch (HEMB) identified:
1)
the possibility of a relationship between time of
spraying and conception and subsequent abortions among
the Alsea women; and
2)
the fact that, while State and county records of spontaneous
abortions are given for terms of 20 weeks or greater, 12 of
the 13 miscarriages experienced by the women in Alsea were
for terms of less than 20 weeks.
Based upon the comments of the reviewers and HEMB staff, OPP undertook
to develop data on spontaneous abortions of 20 weeks duration or less
in the Study area and in a comparable control population.
o
�The current study of 6-year spontaneous abortion rates in three Oregon
areas was initiated by the Human Effects Monitoring Branch, Office of
Pesticide Programs, in October of 1978. The study was accomplished
under contract to the Epidemiologic Studies Program (ESP) projects in
Colorado, Florida and Idaho. Scientists from the Colorado Project,
under the direction of Dr. Eldon Savage, organized and conducted the
field investigations, developed the data including the spontaneous
abortion index, and prepared much of the report.
Idaho project staff
assisted in collection of hospital data in Malheur County. Statistical
analysis and tie interpretative narrative were developed by Drs. Robert
Duncan and Thomas Keefe of the Florida and Colorado projects, respectively.
The following scientist were instrumental in the successful development
and conduct of the study:
Dr. Eldon P. Savage, Director, Colorado ESP, Colorado State
University
Colorado ESP Staff: Drs. Thomas Keefe, Robert Zimmerman and
Richard Hayes; Messrs. William Wheeler, Lawrence Mounce and Jerry .
Rench; Ms. Lois Cox and Ms. Barbara Stevens.
Idaho ESP Staff (Dr. Charles Brokopp, Director): Ms. Jill Wyatt
and Ms. Pamela Smith.
Dr. Robert C. Duncan, Director, Florida ESP, School of Medicine,
University of Miami.
HEMB Staff: Drs. Jack Griffith and Charles Miller, Mr. Robert
Heath and Ms. Mary Frankenberry.
This report was organized and edited by:
Jack Griffith, Ph.D.
Robert Heath, M.S.
Mary Frankenberry, B.A.
�Chapter II
PURPOSE AND SCOPE
The purpose of this study was to assess the rates of spontaneous abortion occurring in a forested region of Oregon's Coastal Range, centered'
about the Alsea basin, where 2,4,5-T has been commonly used in forest
management/ and to compare those rates with rates occurring in a comparable control area.
Specific objectives of the study were to test the following hypotheses:
a.
whether or not differences in spontaneous abortion rates exist
between the study and control populations;
b.
whether or not seasonal variations in rates exist within the
study and control populations;
c.
whether or not such variations, 'if they exist, can be associated with time and concentration of spray applications in
the study area.
The data are limited to in-patient records of women hospitalized for
spontaneous abortions of less than 20 weeks term, based upon;
a.
the observation that 12 of the 13 miscarriages reported by the
original nine participants were for terms of less than 20
weeks duration;
�b.
the assumption that a possible chemical effect would be most
likely during the first trimester of pregnancy and could
therefore be masked in abortion data spanning the entire nine
months of term.
�Chapter III
METHODS
The methods used in this study consisted of defining the study area,
developing/confirming a 2»4,5-T use history in that area, researching
U.S. Postal ZIP code boundaries, selecting a control area, defining an
urban area, abstracting spontaneous abortion data from the hospitals
in the three areas, interviewing area physicians, retrieving county
birth data from Oregon computer tapes, and collecting various descriptive data ancillary to the study.
Description of the Study Area
The Study Area comprises approximately 1600 square miles of Oregon's
forested Coastal Range (Figures 1 and 2). It was selected so as to be
centered around the "Al'sea basin", an area of approximately 400 square
miles. The Study area includes the western half of Benton County,
northwestern Lane County and all but the northern and northwestern
reaches of Lincoln County. It is bounded on the west by approximately
70 miles of the Pacific Coast entending from.Lincoln City southward to
Florence, and extends inland for distances ranging from 10 to 35 miles.
(Exact boundaries conform to U.S. Postal ZIP boundaries.)
The Study
area includes all but the northern and southern reaches of the Siuslaw
National Forest. Interspaced throughout one numerous commercially owned
and Bureau of Land Management forested acreages (Figure 3). Mountain
elevations of approximately 1,000 feet are not uncommon? peak elevation
is slightly more than 4,000 feet.
�Figure 1.
O
IO
OREGON
20
30
SO
60
�Figure 2.
Study
Ar
ea in Lane, Lincoln and Benton Counties of Oregon
ci^^±^rt^±^±i^^^
STUDY AREA
_3_0
SCALE
in
�3. Land Cover in Lincoln. Bento, and Lane (p.rtl.1, Co.nt.es O.e.on
FOREST
CROP &/or RANGE
�Drainage is primarily westward; principal rivers include the Siletz,
Alsea, Yaquina and the Siuslaw. Eastern fringes of the area drain
eastward into the Willamette Valley. Maximum runoff is reached generally during the winter months as the result of storms off the Pacific
occurring usually as rain. There is little snow accumulation.
The Study area is predominantly rural. The four hospitals in the area
are located in the four largest towns: Newport (pop. 5,200); Lincoln
City (4,200); Toledo (2,800); and Florence (2,250). Each of these hospitals was utilized in the study. With the exception of Philomath
(1,700), all other towns/villages have populations of less than 1,000.
Alsea has a population of 400 (1970 census).
All of the nine women who participated in the first phase of the investigation resided, at the time of pregnancy, in rural residences located
within 12 miles of Alsea. All but one of the women resided in the Alsea
River watershed; the ninth resided southwest of Philomath in the
Corvallis watershed.
Description of the Control Area
After careful review and consultation with staff of the Oregon State
Health Department, the Control area was selected in Malheur County,
Oregon. Selection was based on the following criteria:
1.
The area had little or no use of 2,4,5-T.
•?-
The area is primarily rural, as is the Study area.
�3.
Physician practices and hospital facilities were expected to
be similar to those in the Study area.
4.
The area bore topographic similarities to the study area,
being of similar elevation and, although not mountainous, having
rugged terrain (escarpments, rolling hills, arroyos, canyons).
The Control area comprises four contiguous postal ZIP code zones in the
northeastern part of Mal,heur County (See Figure 4). The area covers
approximately 1,000 square miles and is bounded on the east and northeast by the Snake River, which there forms the Oregon-Idaho boundary.
Several creeks drain the area eastward into the Snake River. Approximately 90 percent of the area is classified as rangeland, sagebrush
being the dominant vegetation. Cropland accounts for a small but
important percentage of the area along stream and river courses.
Twenty-one percent of the land in the county is in private ownership,
75 percent is Federal and the remainder (5%) is State, county, or
local governments. (See Figure 5)
The two hospitals in the Control area are located in the two largest
towns: Ontario (pop. 8,200) and Nyssa (pop. 2,900). Both hospitals
were utilized in the survey. The area also includes the town of Vale
(1,850) and the villages of Harper and Cairo (pop. less than 250).
13
�Figure 4.
Control Area in Malheur County Oregon
/
_1^1_Jj»1^1_f«1i
•*1_jiVjjVllj's_|1
COf4TROL AREAL ...._, ,._30Mil«»
SCALE
|
�Figure 5.
Land Cover in Malh'eur County Oregon
FOREST
CROP &/or RANGE
15
�Personnel from the Malheur County Cooperative Extension Agent's office
had no specific figures to report on pesticide use in the county;
however, it was stated that there is very/ very little use of 2,4,5-T
in the county, particularly in the cultivated areas. Personnel from
the area office of the BUM, which has the responsibility for management
of public land within Malheur County, reported that 2,4,5-T has not
been applied on BLM lands in the county since 1972 by either BLM or
ranchers with grazing permits. Additionally, BLM personnel stated that
no pesticides of any kind had been applied to BLM grazing land since
1968. A sagebrush control program used 2,4-D but not 2,4,5-T in Malheur
County.
•
Description of the Urban Area
The Urban area is comprised of the two connecting Postal ZIP zones that
encompass the cities of Corvallis and Albany, Oregon. Both sites are
located in the agricultural non-forested Willamette Valley. The
Corvallis ZIP zone is contiguous with the east-central boundary of the
Study area and the Albany zone connects the Corvallis zone at its northeast corner (see Figure 6). The populations of Corvallis and Albany are
over 37,000 and 21,000 (1970 census).
16
�Urban Area in Benton County, Oregon
Figure 6,
URBAN AREA
30 U l l « »
SCALE
17
�TABLE 1
Selected Agricultural Statistics for Counties Including
the Study, Control and/or Urban Areas
Oregon, 1972 - 1977
County and Area
Lincoln
Study
Malheur
Control
Lane
Study
Benton
Study &
Urban
1,357
258
1,840
575
Land in Farms, Acres
1,360,195
47,390
270,587
129,034
Approximate Land Area, Acres
6,309,760
631,104
2,913,280
427,520
21.6
7.5
9.3
30.2
5
38
158
58
1,278
208,726
553,042
457,813
935
59
900
301
23,040,163
495,703
11,780,769
5,317,112
991
214
1,152
423
20,063,573
1,028,121
9,319,416
2,667,955
All Farms, Number
.Percent in Farms
Forest Production, Number
of Farms
Forest Production, Dollars
Crops, Number of Farms
Crops, Collars
Livestock, Number of Farms
Livestock, Dollars
U.S. Bureau of the Census, Census of Agriculture, 1969 \fol. 1 area
Report Part 41. Oregon Section 2. County Data. U.S. Government
Printing Office, Washington, D.C. 1972.
18
�Women in both cities use Good Samaritan Hospital in Oorvallis for
gynecological and/or obstetrical care, as do various women throughout
the Study area. All spontaneous abortion records for terms of less than 20
weeks were obtained from the Hospital, first, to derive additional
data for the Study area.- and, second, to permit comparison of seasonal
spontaneous abortion patterns and frequencies in an unsprayed urban
area adjacent, to the Study area.
The abortion data for the Urban area are considered to be of limited
utility in this study because of an apparent tendency for firsttrimester abortions to be frequently handled in urban clinical
facilities of a type that do not exist in the Rural Study and Control
areas. The .data are useful, however, in providing a measure of
monthly and seasonal trends on patterns in abortion frequencies.
Research of zit* Code Boundaries
To facilitate an identification of the boundaries of the Study area,
ZIP code maps were developed with the cooperation of personnel in
local post offices.
The boundaries of the Study area, which coincide
with the zip code delivery routes, remained unchanged during the study
time frame of 1972-1977 (See Table 2).
19
�TABLE 2
ZIP Codes Corresponding to the Study, Urban,
and Control Areas
Oregon, 1972 - 1977
Area
Zip Codes
Urban
97321, 97330
Control
97906, 97913, 97914, 97918
Study
97324,
97357,
97369,
97388,
97439,
20
97326,
97365,
97370,
97390,
97453,
97341,
97366,
97376,
97391,
97480,
97343
97367
97380
97394
97498
�Spray Data
In the study of the Alsea are.5? wonen, spray data on the use of 2,^,5-7
were collected and plotter- for the immediate area referred to as the
"Alsea Basin". This ir.forrv-tior. was supplied '•// the following major
organizations that used the chemical: USFS-Ciuslaw National Forest;
USDI-Bureau of Land Management, Alsea Resources Area; Willamette Incustries, Inc., Pilonath, Oregon; and Starker Forest, Philomath, Oregon.
The supplier" data consisted of the elate(s) of application, rate of
application, formulation, :: ur.be r of acres treated, arV th.e location of
the treated land.
The locations of the sprayed areas were plotted to quarter-section on
township maps. In this manner the perimeter of the "Alsea Basin" could
then be defined as one covering approximately 40C square miles or
255,000 acres. During the o-year period fron 1S7P-1977, a total of 71"!
acres was treated with 9916 pounds of 2,4,^-T. Tlie poundage and acreage
varied fron year to year (TsMe 3). Tf^.e areas treated during this
period represent approximately 3% of the total acreage within the "Also?.
Basin".
*
As mentioned previously, the predominant feature of the study area is
the forested Coastal Range. Since the Coastal Range extends from north
ern California into Washington, it was considered necessary to ect.-.hlio',
that: 1) topography and vegetation are similar throughout Oregon; and 2)
forest management practices in the Alsea basin are representative of the
21
�Table 3
Total Acres Treated and Pounds 2,4,5-T Used in Alsea Basin
Oregon, 1972-1977
Year
Acres Treated
Treatment
Total 2,4,5-T (Ibs.)
1972
88
0.5 Ib/A
98
219
63
1.0 Ib/A
2.0 Ibs/A
44
98
438
189
769
3.0 Ibs/A
468
1973
444
25
469
1 Ib/A
2 Ibs/A
444
50
494
1974
207
80
287
1 Ib/A
207
240
447
1975
223
239
90
552
3 Ibs/A
1 Ib/A
2 Ibs/A
3 Ibs/A
223
478
270
971
1976
1256
2 Ibs/A
2512
1977
444
90
534
2 Ibs/A
3 Ibs/A
888
270
Total
3,566
1158
6,351
�entire forested area. Personal communications with representatives of
the U.S. Forest Service and commercial tree farm operators substantiated that the ecological characteristics of the Coastal Range were
consistent in the Oregon Coastal Range and that the chemical, 2,4,5-T,
is used as a camion forest management tool in this region.
The herbicide 2,4,5-T is applied almost exclusively by helicopter at an
average rate of two pounds per acre for control of undesirable vegetation such as red alder, vine iraple, salmonberry, and Thimbleberry.
Certain weather factors such as wind and precipitation dictate time of
application, but in general the compound is used in the spring (March,
April, or May) with a second application made, if needed, in middle to
late summer (July and/or August). These seasonal usage patterns are
shown in Table 4 and Figures 7 and 8.
To avoid contamination of water sources prior to 1978, the general application policy was to avoid spraying near homes and provide for a single
swath of 30 to 60 feet on each side of any major stream. In September,
1978, the Oregon Forest Practices Act provided extended guidelines.
These guidelines required that no spraying was to be made within 500
feet of an inhabited residence nor within 200 feet on either side of a
Class A stream. Class A streams are defined as major streams with fish
and/or ones that are used for domestic water supplies.
Thus, the data from the Alsea basin illustrate a pattern which may be
considered as representative of the Study area. In general, greater
amounts of the chemical were applied during the spring than
�Table 4
Applications of 2,4,5-T by Day and Month in the Alsea Basin
Oregon, 1972-1977
•Year
Dates
Total Days
Acres treatment
Amount Applied (Ibs
1972
March 17, 20, 31
April 4
3
1
1
4
9
148
296
121
48
151
468
no
144
219
769
5
1
6
444
25
469
444
50
494
3
1
1
1
6"
180
27
48
32
287
180
27
144
96
447
July 31
August 1, 19, 23, ?
1973
May 5,6,10,13,14
August 15
T974
April 26, 27, 29
May 4
July 29
August 2
1975
April 9, 16
May 25, 27, 29
July 31
August 20
2
3
1
1_
7
239
• 202
16
95
552
478
282
16
195
971
1976
April 5-10
May 6
5
1_
2512
35
6
1256
35
1291
2547
4
534
1158
. 1977
March 12-15, 19
�Figure 7.
Acres Sprayed with 2,4,5-T in ATsea
Basin Accumulated by Respective Month, 1972 through 1977
3
U
x
n
»S
2-4
25
�Figure 8.
.
.
Pounds* of 2,4,5-T Sprayed in AT sea
Basin Accumulated by Respective Month, 1972 through 1977.
1 x
3
3
1
-
' J
' F ' M ' A ' M
T
71 J ' A ' S
MONTH
*Active ingredient
» 0 ' N ' D'
�during the summer treatments. Additionally, the spray program is not
a month-long operation.
Usually it spans only a few days' time. The
duration of 'the spraying depends on the number of acres to be treated
and on the weather conditions.
Selection of Hospitals and Abstraction of Spontaneous Abortion Data
Spontaneous abortion data were abstracted from all five hospitals
located in the Study and Urban areas (See Figure 9) and from the two
hospitals located in the Control area (See Figure 10). These hospitals
were viewed as the primary source of health care delivery for the
respective areas. The seven Oregon hospitals contacted are located
in Corvallis, Lincoln City, Newport, Toledo, Florence, Nyssa, and
Ontario. Those hospitals and the number of admissions and live births
that each recorded in 1975 are found in Table 5. All field work was
completed during November and December, 1978.
Most of the hospitals were reluctant to participate in a record search
with non-hospital personnel because of patient confidentiality. Three
hospitals pjrmitted a complete record search by field epidemiologists,
while at the other four hospitals the data were abstracted by hospital
personnel.
Patients' names and addresses, excepting ZIP codes, were not
recorded.
For each of the spontaneous abortions the following information was
provided or abstracted:
H-ICDA diagnosis code, age of patient, date of
27
�Hospitals in Study and Urban Areas
Figure 9.
Lincoln/
City
INCOLN
COUNTY
Toledo
Newport/
x/Corvallis
BENTON
COUNTY /
LANE
COUNTY
Florencejx
28
�Figure '10.
/
Hospitals in Control Area
i
Ontario
Nyssa
/
MALHEUR
COUNTY
�TABLE 5
Annual Admissions and Births in Hospitals in
Study, Urban, and Control Areas
Oregon 1972.- 1977
Hospital
location
Hospital
Name
Urban Area
Study Area
Admissions
1975
Births
1975
Good Samaritan
Corvallis
7589
986
Western Lane
Florence
802
87
North Lincoln
Lincoln City
1367
77
Pacific Communities
Newport
1159
48
New Lincoln
Control Area
City
Ibledo
875
103
Malheur Memorial
Nyssa
710
110
Holy Rosary
Ontario
4097
474
3U
�spontaneous abortion, gestation period, and ZIP code of patient's residence. Only records of spontaneous abortions that occurred from 1972
through 1977 were sought. This time period coincided with the temporal
miscarriage pattern of the nine original Alsea women.
From 1972 to 1977, two ICDA texts were used by the hospitals. In 1973
H-ICDA was published and spontaneous abortions were listed by the codes
643.0, 643.1, 643.2, and 643.9. The 643.9 code includes any spontaneous
abortions not listed as induced or spontaneous.
Prior to 1973, hospitals
used the 8th revision of ICDA. Spontaneous abortions were coded as
643.0, 643.1, 643.2, and 643.9. The code number 644- was also used for
1972, since this code approximated ft-ICDA code number 643.9.
Physician Interviews
A list of private physicians who practice at each of the five hospitals
in the Study and the Urban areas was compiled from information provided by the hospitals and by local health officials. From the list,
30% of the physicians were randomly selected for interview. Ebwever,
since physicans tended to practice in groups, doctors not randomly
selected were also interviewed whenever available. In the Study area
19 of 27 (70%) of the physicians—all general practitioners—were
contacted. In Corvallis all eight of the obstetricians/gynecologists
were interviewed as well as 20% (five) of the general practitioners and
10% (two) of the internists.
�When possible, interviews were carried out in the doctor's office.
Each physician was asked to estimate the number of spontaneous
abortion cases of terms less than 20 weeks that he or she had treated
per year during the 1972 to 1977 time period. Each was also asked
to estimate the percentage of those cases which had been hospitalized.
Because of similarities in medical facilities and rural population
distribution in the Control and the Study areas, physician practices
were assumed to be similar in the two areas. Therefore, it was
decided not to conduct physician interviews in the Control area.
Data Preparation
The data on spontaneous abortions of less than 20 weeks from the seven
study hospitals were edited, coded, keypunched, and computer edited.
The number of monthly hospitalized spontaneous abortions during 19721977 were tabulated for each of the three areas. These data were also
tabulated and cross-tabulated according to the several variables of
interest. For example, the data on hospitalized spontaneous abortions
were cross-tabulated according to area (Study, Urban, and Control) and
gestation period (less than 4 weeks, 5-8 weeks, 9-13 weeks, 14-17
weeks, and 17-20 weeks).
Birth certificate data for 1972 through 1977 were obtained in the form
32
�of computer tapes from the Vital Statistics Section of the Oregon
State Health Department. These computer tapes were used to obtain the
number of births per month in the Study, Urban, and Control areas for
1972-1977.
The Spontaneous.Abortion Index; In order to make comparisons among
areas and among the months of the year, especially within areas, the
data on spontaneous abortions need to take into account the number of
births in the three areas during each of the twelve months. The index
described below is basically the ratio of the number of hospitalized
spontaneous abortions to the number of births corresponding to the
spontaneous abortions, based on the residence ZIP Code of the women
contributing to each event. Thus, the ratio is not a true rate but
rather an index of the hospitalized spontaneous abortion experience of
the women residing in the three areas.
In order to describe the spontaneous abortion index, the following
notation is required:
Y. • = the number of hospitalized spontaneous abortions in the
area (i * 1,2,3. . ) in the j
.
month (j » 1,2,...,
12) during 1972-1977;
X- = the number of births in the i— area and in the j
month during 1972-1977.
�An index that has been used in similar studies is simply the ratio
Y
/5f
ij ij. The inherent problem with such an index is that the
.numerator involves women of one conception period whereas the
denominator involves women of a different non-overlapping conception
period.
The spontaneous abortion index developed here is Z-• = Yji/C.:-:
where C^• is a five-month moving-average of the X^.'s which have
been appropriately shifted.
In particular, C^=
where the weights (W , . . W ) represent the proportions of
.,
abortions of varying gestation.
In the above definition of C.^/ the
second subscript on the number of births is taken modulo 12; for
example, j+3+k = 13 refers to January births. The rationale for the
use of GJJ rather than X^ • is that a woman who has a spontaneous
abortion in month j could have delivered between month j+4 and month
j+8, depending on her length of gestation.
In averaging the number of
births in the five months between month j+4 and month j+8, one could
use a simple average (i.e., W^ = W2 = . . = W5 = .20). Instead,
.
a weighted average of the number of births was applied with estimated
weights given by the proportions of abortions for the five gestation
periods.
Index values were calculated, by month and area, for the aggregated
six-year data and for the two three-year periods 1972-1974 and 1975-
1977.
34
�Statisticial Procedures
Statistical evaluation of the data is based on the following analyses;
1. Analysis of variance of the abortion index by area and
month, deriving a residual error term by dividing the data
into two 3-year periods.
2. Frequency table analyses, by chi-square, to test monthly
variation in the cumulated number of spontaneous abortions
analyzed as simple frequencies and as expected monthly
frequencies of spontaneous abortions calculated from the
frequencies of corresponding live births.
3. Tests of cyclic trends including:
(a) A power-spectrum analysis of the abortion index over
months to test for cyclic trends in the monthly data.
(b) Adjustment of the abortion index data to account for
phase differences between the Study area and the Control and Urban areas.
(c) Fit of a sine-wave model to the abortion index data.
(d) Cyclic analysis of the raw abortion data (unadjusted
for births).
35
�(e) Examination of birth data from Miami, Florida (Jackson
Memorial Hospital, 1976-77) for cyclic behavior.
(f) A test for cyclic trends after recalculating the
abortion index by raw numbers of births as the
denominator rather than the five-month moving average.
4. Cross-correlation analyses between Study area abortion
indices and spray patterns.
(a) Cross-correlation analyses, parametric and non-parmetric,
between the Study area abortion index and the monthly
spray pattern of 2,4,5-T in pounds sprayed per month.
(b) Cross-correlation analysis with the monthly spray pattern after adjusting the abortion index data for differences in cyclic trends.
(c) Cross-correlation analyses of Study area abortion index
and spray patterns for the periods 1972-74 and 1975-77.
5. Evaluation of physician interview data.
3G
�Chapter IV
RESULTS
The annual numbers of hospitalized spontaneous abortion cases in the
three areas appear in Table 6. There was a total of 477 cases: 188
in the Study area, 109 in the Control area and 180 in the Urban.
Table 7 depicts these cases accumulated by month of the year for the
1972 to 1977 time period. The same cases are also represented according to the age of the patient in Table 8. Table 9 presents, for each
area, the cases by weeks-of-term categories for those cases (456) for
which term was known.
As mentioned previously, the number of births in these three areas
needs to be taken into account in any comparisons made among the
areas with respect to the number of spontaneous abortions.
Table 10
presents the number of births per calendar month for the Study, Urban
and Control areas; the five-month moving average of the number of
births corresponding to the month of spontaneous abortion is given in
Table 11 for each area, where the weights were obtained from the
tabulation of spontaneous abortions according to length of gestation
(Table 9).
Table 12 presents the monthly spontaneous abortion index for each of
the three areas in the study; the monthly spontaneous abortion index
for these areas is also displayed graphically in Figure 11. An
obvious feature of this graph is the elevated index of the Study area
for the month of June.
37
�TABLE 6
Number of Hospitalized Spontaneous Abortion Cases for Study,
Urban, and Control Areas
Oregon, 1972 - 1977
Study Area
Urban Area
1972
37
44
9
1973
34
40
19
1974
23
38
27
1975
31
19
27
1976
33
20
8
1977
30
19
19
188
180
109
Year
Tbtal
38
Control Area
�TABLE 7
Number of Hospitalized Spontaneous Abortion Cases for Study,
Urban, and Control Areas, Accumulated by Month
Oregon, 1972 - 1977
Month
Study Area
Urban Area
January
10
25
12
February
17
17
4
March
18
15
7
April
11
16
14
May
16
17
9
24
15
6
July
20
5
7
August
17
11
10
17 '
11
June
September
.
9
Control Area
October
15
19
7
November
16
7
8
December
15
16
14
188
180
109
Tbtal
39
�TABLE 8
Hospitalized Spontaneous Abortion Cases by Study, Urban,
Control Area, and Age Group of Patient
Oregon, 1972 - 1977
Age
Group
Study
Area
-Percent
Urban
Area
10-14
3
1.6
0
0
0
0
15-19
38
20.2
12
6.7
14
12.8
20-24
69
36.7
68
37.8
38
34.9
25-29
44
23.4
62
34.4
24
22.0
30-34
18
9.6
26
14.4
21
19.3
35-39
9
4.8
7
3.9
9
8.3
40-44
6
3.2
5
2.8
3
2.8
45-49
_!_
0.5
0
Total
188
100.0
180
'
Percent
0__
10.0
Control
Area
Percent
0
0_
109
100.1
�TABLE 9
Number and Percent of Spontaneous Abortions According to
length of Gestation in the Study, Urban, and Control Areas
Oregon, 1972 - 1977
Gestation Period
1. 18 - 20 weeks
Study
Area
Urban
Area
8
12
(4.6%)
2. ' 14 -17 weeks
(6.7%)
24
15
'(13.8%)
3. 10 - 13 weeks
68
(39.1%)
4. 5 - 9 weeks
(46.1%)
60
(33.3%)
12
(6.9%)
Obtal
83
62
(35.6%)
5, 4 weeks or less
(8.3%)
10
Control
Area
7
(6.8%)
14
(13.7%)
48
(47.1%)
28
(27.5%)
5
Cverall
27
(5.9%)
53
(11.6%)
199
(43.6%)
150
(32.9%)
27
(5.6%)
(4.9%)
(5.9%)
180
102
456
174
41
�TABLE 10
total Number of Births by Months for the Study,
Urban/ and Control Areas
Oregon 1972 - 1977
Study Area
Urban Area
Control Area
January
194
338
124
February
188
353
124
March
202
34.4
126
April
189
335
131
May
201
357
155
June
228
378
145
July
203
315
156
August
212
351
131
September
204
342
154
October
170
.343
136
November
172
337
155
December
181
327
129
2344
4120
1666
Six-Year Total
�TABLE 11
Number of Births Corresponding to the Month of Spontaneous Abortions
in the Study, Urban, and Control Areas
Oregon, 1972 - 1977
Month
Study Area
Urban Area
Control Area
January
208.8
338.3
146.3
February
206.8
345.0
142.6
March
191.8
341.5
April
177.8
340.4
143.6
May
177.9
334.8
142.3
June
184.0
334.3
130.5
July
189.7
342.0
126.5
August
193.0
345.7
125.4
Septeinber
195.6
342.5
129.0
October
196.7
346.9
138.7
November
208.7
358.1
147.3
December
213.4
350.7
148.1
Total
2344
4120
1666
Based on five-month moving average.
43
'
145.6
�TABLE 12
Monthly Spontaneous Abortion Index for the Study, Urban, and
Control Areas
Oregon, 1972 - 1977
Month
Study Area
Urban Area
Control Area
Average
January
48.1
73.9
82.0
68.0
February
82.2
49.3
28.1
53.2
March
93.8
43.9
48.1
61.9
April
61.9
47.0
97.5
68.8
May
89.9
50.8
63.2
68.0
June
130.4
44.9
46.0
73.8
July
105.4
14.6
55.3
58.4
August
88.1
31.8
79.8
66.6
September
46.0
49.6
85.3
60.3
October
76.2
54.8
50.5
60.5
November
76.7
19.6
54.3
50.2
December
70.3
45.6
94.5
70.1
Average
80.8
43.8
65.4
63.3
IV
The spontaneous abortion index is defined as the ratio of the number of
hospitalized spontaneous abortions to the corresponding number of live
births based on a five-month moving average, and is expressed as
abortions/1,000 births.
�Figure 11. Plot of Monthly Spontaneous Abortion Index
for the Study, Urban, and Control areas
Oregon, 1972-1977
•—.STUDY A R E A
M I M I M I M URBAN AREA
•r^^^pror-,CONTROL A R E A
i
I
r
i
j | -j
—
|j
p— p
J
F
M
A
M
A
O
J
J
S
N
�Statistical Analyses
In order to arrange the data in such a way that proper error terms
for hypothesis tests could be calculated and to see whether seasonal
spontaneous abortion patterns were consistent over time, -the data
were gathered into two 3-year periods as shown in Table 13. The
plots of these data are shown in Figures 12, 13, and 14.
From the graphical representations it is seen that the time
(seasonal) patterns within each area are remarkably similar for the
two periods.
The analysis of variance appropriate for the data in Table 13 is
given in Table 14. Although the three-year time periods could have
been viewed as a blocking factor, it was decided to test the various
interaction terms for possible significance. Clearly, the only
significant variation is among the three areas.
The mean values for the three areas are: Study area = 80.8; Urban
area = 43.8; and Control area = 65.4 (see Table 12). By the New
Duncan's Multiple Range Test, all three means are significantly
different from each other (p<.05, two tailed).
From Table 13 and Figure 14 it is seen that there is a decrease in the
overall Abortion Index for the Urban area during Period II. Analysis
showed that this decrease is significant (p<.0002). The mean
AC,
�TABLE 13
Abortion Index "by Period, Area, and Month
Oregon, 1972-1977
PERIOD I
PERIOD II
(1972 - 3 - 4)
(1975 - 6 - 7)
Study
Area
Urban.
Area
Control
Area
January
48
95
68
48
53
96
February
77
58
28
87
41
28
March
94
59
14
94
29
82
April
45
76
84
79
18
111
May
112
54
84
67
48
42
June
98
72
46
163
18
46
July
95
29
63
116
0
47
August
76
52
96
104
12
64
September
61
76
140
31
23
31
October
71
63
29
81
46
72
November
105
22
68
48
17
41
December
84
57
95
56
34
95
Average
80.5
59.4
67.9
81.2
28.3
62.9
Study
Area
Urban
Area
Hospitalized spontaneous abortions/1,000 live births adjusted for
gestational age distribution of observed abortions. Ihe computed
monthly births.for the six-year interval were divided equally between
the two periods,,
47
Control
Area
�Figure 12 Abortion Index for the Study Area By Month and Period
180
160
140
120
X
O)
"O
c
O
100
O
80
60
40
20
0
kt-Hfi* tS&y-***-- fni-m m^t
F
•lK&*f>W*r.i* U t a < | H f / 4 M I & £ * * ' »
4rmWt*-«BWM-h*MCM
M
M
J
J
Month
A
0
N
D
�Figure 13 Abortion Index for the Control Area By Month and Period
140
120
X
O)
100
TO
c
4J
80
O
60
40
20
• M «•*» •»*»|f*r-*«"'-I*^--**i-«V—W(vl.J.»»»
J
F
M
•••viftL.,,i fMIVI* ***•»*« if* I'#lr**of»*Jtr*v
M
J
Month
Irtfc-rfwi^uwf^Htn^rtm-ijh*.1,*
J
A
l.|>M<»t*««1M.Rf
0
N
,.2,
�Figure 14 Abortion Index for the Urban Area By Month and Period
X
01
-a
N
Month
D
�TABLE 14
Analysis of Variance of'Abortion Index by Period, Area, and Month
(See liable 13}
Source of
Variation
Degrees of
Freedom
Sum of
Squares
Mean
Square
F-Ratio
Periods
1
2,485.13
2,485.13
4.10
Areas
2
16,475.53
8,237.76
13.58*
Months
1.1
3,377.82
307.07
<1
Periods x Areas-
2
3,498.08
1,749.04
2.88
Periods x Months
11
7,74.9.38
704.49
1.16
Areas x Months
22
25,156.81
1,143.49
1.89
Non-additivity
1
0.34
0.34
21
12,738.58
606.60
Residual
p<.0002
51
<1
�abortion index of the Urban area for Period I is 59.4. When this is
oonpared to the six-year mean for the Study area (80.8) and that for the
Control area (65.4) using the New Duncan's Multiple Range Test it is
found that both the Urban area and the Control area differ from the
Study area (p<.05, two tailed), but not from each other. This seems
a more conservative analysis since the reason for the precipitous
decrease of the abortion index during Period II is presently unknown.
Figure 12 is noteworthy for two reasons. First, there is striking
similarity in the patterns, especially the "shoulders" on the curves
following the peaks for each period. Second, the peak for the second
period corresponding to a greatly increased pesticide usage is
higher than that for the first period.
Frequency Table Analyses
Uniform Hypothesis^ Since the adjusted numbers of births used for the
denominators in calculating indices appear sensibly flat (see Table 11)
monthly variations in the cumulated number of spontaneous abortions
were analyzed as simple frequencies (see Table 7).
Under the hypothesis that monthly accumulated spontaneous abortions
were uniformly distributed throughout the year, chi^square analyses
were performed for each area. The urban area showed a significant
variation among months due to a high number of spontaneous abortions in
December and low numbers in July and November. The other two areas did
52
�not show significant variation among months/ but in the Study area the
month of June had a significant contribution to the overall chi-square
value. Further, in each area, various months showed large
contributions to the overall chi-square.
Expected Frequencies Based on Live Births; When the expected monthly
frequencies of spontaneous abortions were calculated from the
frequencies of corresponding live births from Table 11, the chi-square
analyses were virtually the same as when based on the actual number
of spontaneous abortions.
Cyclic Trends
Spectral Analysis! The Analyses of Variance and the chi-square
analyses, together with a close study of Figure 11, strongly suggested
the possibility of cyclic trends in the data.
Power-spectrum analysis of the abortion index over months showed that
for each area there was only one cycle of significance and it had a
period of about four months in each area.
Adjustment For Cyclic Trends: The importance of indentifying cyclic
trends is two-fold. First, their presence might have some biological
meaning relevant to this study. Second, a month-to-month comparison
between the Study area and the Control areas should take into account
phase differences which might be related to fertility patterns; or some
other as-yet-unknown phenomenona.
�The need for this type of adjustment is shown in Figure 15. Since the
Study area and the combined Urban and Control areas seem to be almost
exactly 180° out of phase, any month-by-month differences could be
falsely amplified. Clearly, if the Control data are shifted back in
time as shown in Figure 16, monthly differences could possibly reflect
excess abortions in the Study area if they exist.
Ihe Sine Wave Model; Since the power-spectrum analysis identified
only one frequency for each set of data, it was decided to fit the
simplest cyclic model:
Index = A -f B Sin (,££ t - D) where t is in months. Table 15 shows
quite clearly that this model adequately represents the data in the
Urban and Control areas. Because of the peak in the Study area data
around June and July, the model does not fit the Study area as well.
Hie phase difference between the Study and the Urban and Control
areas is on the order of two months.
Tables 16, 17, and 18 show the Study area data adjusted for the Control
and the Urban area and for both areas combined. Figures 17 and 18 show
the sine wave models fitted to the abortion index data for the Study
area and the Control and Urban Areas.
54
�FIGURE~15f SPONTANEOUS ABORTION INDEX, 1972 - 77 FOR THE STUDY AREA AND THE
*" '
' COMBINED CONTROL AND URBAN AREAS.
(1404- STUDY AREA
ao-
- URBAN AND CONTROL
| no,
fioo'<
ui
C,
80,
llo
M
1
i
I
J
F
M
L__
A
L
M
1
J
Month
J
A
S
O
N
D
�FIGURE 16:
STUDY AND CONTROL AREAS: (a) COMBINED URBAN AND CONTROL INDEX,
<b) STUDY AREA INDEX, (c) URBAN AND CONTROL INDEX SHIFTED BACK
TWO MONTHS
(i) CONTROL AND URBAN AREA
70
g
so
40 tI
30 -
I
I
I
N
M
M
140
(b) STUDY AREA
120
*hoo
60
40
M
M
N
J
(e) URBAN AND CONTROL AREA
WITH 2-MONTH PHASE SHIFT
80
70
X
ui
O
~
\
»
60
40
90
M
M
J
J
O
N
J
-F
�TABLE 15
Parameters of the Model "Index = A + B Sin (? t - D)"
Ftor the Various Areas
Control
Urban +
Control
Study
Urban
A (Mean)
83.1
43.4
65.5
56.0
B (Amplitude)
15.6
15.2
24.7
15.6
T (Period in Months)
3.6
4.1
4.1
4.0
D (Phase Shift in Months)
1.1
3.1
2.2
2.8
Adequacy of the Model
(Correlation)
0.49
0.78*
^* 0.05<p<.06
p<.001
57
.92**
.84**
�TABLE 16
Study Area Index Corrected for Phase Shifts
Study minus Control —- 1 month lag
Study
Control
Difference
48.1 (J)
28.1 (F)
20.0
82.2 (F)
48.1 (M)
34.1
93.8 (M)
97.5 (A)
-3.7
61.9 (A)
63.2 (M)
-1.3
89.9 (M)
46.0 (J)
43.9
130.4 (J)
55.3 (J)
75.1
105.4 (J)
79.8 (A)
25.6
88.1 (A)
85.3 (S)
2.8
46.0 (S)
50.5 (0)
-4.5
76.2 (0)
54.3 (N)
21.9
76.7 (N)
94.5 (D)
-17.8
70.3 (D)
82.0 (J)
-11.7
�TABLE 17
Study Area Index Corrected for Phase Shift
Study minus Urban — 2 month lag
Study
Urban
Difference
48.1 (J)
43.9 (M)
4.2
82.2 (F)
47.0 (A)
35.2
93.8 (M)
50.8 (M)
43.0
61.9 (A)
44.9 (J)
17.0
89.9 (M)
14.6 (J)
75.3
130.4 (J)
31.8 (A)
98.6
105.4 (J)
49.6 (S)
55.8
88.1 (A)
54.8 (0)
33.3
46.0 (S)
19.6 (N)
26.4
76.2 (0)
45.6 (D)
30.6
76.7 (N)
73.9 (J)
2.8
70.3 (D)
49.3 (F)
21.0
59
�TABLE 18
Study Area Index Corrected for Phase Shift
Study minus (Urban plus Control) — 2 month lag
Difference
Study
Urban + Control
48.1 (J)
46.0 (M)
2.1
82.2 (F)
72.3 (A)
9.9
93.8 (M)
57.0 (M)
36.8
61.9 (A)
45.5 (J)
16.4
89.9 (M)
35.0 (J)
54.9
130.4 (J)
55.8 (A)
74.6
105.4 (J)
67.5 (S)
37.9
88.1 (A)
52.7 (0)
35.4
46.0 (S)
37.0 (N)
9.0
76.2 (0)
70.1 (D)
6.1
76.7 (N)
78.0 (J)
-1.3
70.3 (D)
58.3 (F)
12.0
60
�STUDY AREA
130 H
120
110
Sine Wave Model
100
90
80
70
60
60
M
M
J
Month
N
�FIGURE 18 PREDICTED AND OBSERVED POINTS ihOR THE URBAN + CONTr^ Ji. AREAS
80
-
Urban + Control Areas
76
70
-
65
60
er»
55
50
45
40
35
I
I
M
I
I
M
N
�Cyclic Analysis of the Numbers of Abortions; The raw numbers of
abortions showed the same time behavior as did the indices. In order
to find some rationale for cyclic variation in spontaneous abortions,
a physician board-certified in obstetrics and gynecology was
consulted. He was aware that births might show some high frequency
cyclic behavior superimposed on a yearly cycle. The data for 1976-77
for Jackson Memorial Hospital in Miami, Florida are shown in
Appendix B. A plot of these data suggests cycles of two-months
duration. Plots of the Oregon birth data in this report suggest
cycles of two-to-three-months duration. Clearly, these patterns need
investigation since they might be viewed as "noise" in the data and
thus obscure important relationships.
Simple Number of Births as the Denominator for the Abortion Index; In
order to see if any unanticipated algebraic relationships involved in
computing the denominator for the index could be causing a problem,
the raw numbers of births were used to compute abortion indices. The
results of the analyses did not change from the above. Table 19 shows
the distribution of births by month and the computed indices.
Cross-Correlation Between Study Area Abortion Indices and Spray
Pattern
Cross-Correlation Analysis; A cross-correlation was computed between
the abortion index for the Study area and the monthly pattern of
spraying 2,4,5-T in pounds sprayed by month.
63
This
�TABLE 19
Number of Births per Month and Unweighted Abortion Index
for the Study, Urban, and Control Areas
Oregon 1972 - 1977
Month
Study Area
Urban .Area
January
194
51.5
338
74.0
124
96.8
February
188
90.5
353
48.1
124
32.3
March
202
89.1
344
43.6
126
55.6
April
189
58.2
335
47.8
131
100.9
May
201
79.6
357
47.6
155
58.1
June
228
105.3
378
39.7
145
41.4
July
203
98.5
315
15.9
156
44.9
August
212
80.2
351
31.3
131
76.3
September
204
44.1
342
49.7
154
71.4
October
170
88.2
343
55.4
136
51.5
November
172
93.0
337
20.8
155
51.6
December
181
82.9
327
48.9
129
108.5
2344
80.1
4120
43.6
1666
65.8
Total/Average
64
Control Area
�FIGURE 19 (a) STUDY AREA ABORTION INDEX CORRfcCfED FOR PHASE-SHIFTED
URBAN INDEX; (b) POUNDS OF 2,45-T SPRAYED BY MONTH
(a)
100 —i
CORRECTED
INDEX
•
/
50
\
•
M
•p • •• te
\
m
us
0 —
4
2.4,5 - T
(POUNDS)
X 1,000
-I
3 .
2
1
0 .J
I
J
F
M
A
M
i
J
i
S
i
O
I
N
D
�FIGURE 20 CROSS-CORRELATION OF STUDY AREA ABORTION INDEX WITH
SPRAY PATTERN IN TOTAL POUNDS OF 2,4,5-T APPLIED BY
MONTH. 1972-77.
.8
.7.
0.72 **
0.61
.6
.5
.4-
z
o
H
.3
UJ
CC
CC
O
u
o
cc
u
.1
o.io
o.io
0
-.1 -
-0.09
-.2
-.3
-.5-
i f I iI
4 3 2 1 0
Month* Shift in Spray Data
�analysis showed that the abortion index for the Study area was
significantly correlated (p<.01) with the spray pattern after a lag
of two (r=.70) to three (r=.76) months (see Figure 20).
To avoid problems with a non-normal distribution of the spray data,
a Spearman rank correlation corrected for ties was computed. The twomonth lag was not quite significant (r=.46), but the three-month lag
was (r=0.58, p<.05).
Cross-Correlation Between the Adjusted Study Area Index and the Spray
Pattern
Removal of the Cyclic Trend: The differences shown in Tables 16, 17,
and 18 now appear to be free of cyclic trends. This is shown in
Figure 19. Thus, the adjusted indices over time can be related to the
spray pattern over tijne in a straight-forward way. This is a more conservative approach than to use the unadjusted differences between the
Study area and the Control or Urban areas when correlating with the
spray pattern,,
Cross-correlation of Spray Pattern with Adjusted Abortion Index; From
Figure 19 it appears that the peak in the corrected abortion index
for the Study area follows the peak in the spray pattern by two to
three months. This is borne out in Table 20 which shows significant
correlations (p<.01) after two months.
67
�TABLE 20
Cross - Correlation of Corrected Study Area Index with
Spray Data
lag (months)
1
2
3
-.07
0.46
0.88*
0.42
-0.04
Study vs. Control
-0.22
0.35
0.84*
0.33
-0,23
Study vs. Urban +
Control
0.10
0.48
0.81*
0.42
0.11
Study vs. Urban
JK.01
68
�TABLE 21
Pounds of 2,4,5-T Applied in the Alsea Basin and Abortion Index
For the Study Area by Month and Period
Oregon, 1972-1977
PERIOD I
PERIOD II
(1972 - 3 - 4)
Abortion
Index
(1975 - 6-7)
Pounds
2,4,5-T
Abortion
Index
Pounds
2,4,5-T
January
48
0
48
0
February
77
0
87
0
March
94
296
94
1158
April
45
290
79
2990
May
112
471
67
317
June
98
0
163
0
July
95
288
116
16
August
76
365
104
195
September
61
0
31
0
October
71
0
81
0
November
105
0
48
0
December
84
0
56
0
69
�TABLE 22
Cross-Correlation of Study Area Abortion Index with Spray Data
Lag (months)
0
1
2
3
4
Period I
0.19
-0.06
0.27
0.31
Period II
0.32
-0.24
0.31
0.66
Spearman .Rank Correlation corrected for ties
0.01 <p<0.05
70
**
0.00
0.00
�Abortion Index/Spray Pattern Cross-<torrelation Analyses for Periods
I and II
Table 21 shows the pounds of 2,4,5-T used by month and the abortion
index by month for Periods I and II (1972-74 and 1975-77). Crosscorrelation analyses were conpleted between the abortion indices for
each period and the respective monthly patterns of spraying 2,4,5-T in
pounds sprayed by month in the "Msea Basin." Since the spray data
are clearly not approximately normally distributed, a Spearman Rank
Correlation corrected for ties was done to find the cross-correlation
between spray patterns and the abortion index in the Study Area.
The results are shown in Table 22. These correlations show the same
pattern as those for the six-year aggregate of data (see Table 20),
although the correlation does not reach significance for the Period I
data.
Physician Interviews
Table 23 liscs the physicians' estimates of the number of spontaneous
abortion cases of terms less than 20 weeks treated per year (1972-77)
and the percentage that were hospitalized.
Of the 19 doctors contacted at the four hospitals in the Study area,
four refused to provide estimates and four indicated they did not
handle such cases. The remaining 11 physicians provided positive
71
�.responses on number of cases handled and/or percentage hospitalized.
Assuming that the four who refused to respond do handle such cases,
79% (15 of 19) of the physicians contacted have treated spontaneous
abortions during the study period.
From the physician interviews, it is estimated that approximately
70% of the spontaneous abortion cases are hospitalized among those
treated by physicians practicing at one of the Study area's four
hospitals. Ihe percentage is calculated by weighting the individual
hospitalized percentage estimates by the respective number of cases
treated by each responding physician.
It is concluded, therefore, that the abortion data presented for the
Study area are, indeed, representative of that area. It is further
assumed that because of the aforementioned medical/population
similarities between the Study and the Control areas, a similar
representative percentage of spontaneous abortions are hospitalized in
the Control area.
The percentage of spontaneous abortion cases hospitalized in the Urban
area is estimated to be approximately 30%. (Ihe estimate is adjusted
for general practitioners not selected for interview). Ihe percentage
is clearly less than that for the Study area, apparently due to those
cases treated at the clinic and medical center facilities not
available in the Study area. The data are considered to be reflective
72
�of seasonal spontaneous abortion patterns; however, Urban area index
values are undoubtedly biased downward relative to those for the Study
and Control areas.
73
�TABLE 23
Study and Urban Area Physician Interviews:
Estimated Numbers
of Spontaneous Abortions Treated Annually and Percentages
Hospitalized, 1972 - 1977
Annual Mean Number
of Spontaneous
Abortion Cases,
Percent
1972 - 1977
Hospitalized
Type of
Patients
A
GP
*
B
GP
0
GP
0
D2
GP
10.0
E3
Pacific Communities
Newporte, OR
(9 Physicians)
Contacted
Physician
C1
Hospital
GP
0
75-80
—
95
—
Western Lane Hospital
Florence, OR
(6 Physicians)
F
G
GP
H
1.5
GP
GP
I
20-30
Uhknown
6
GP, Surgery
25
1.5
95
**
New Lincoln Hospital
Toledo, OR
(4 Physicians)
GP
15
75
K
GP
7.1
50
L
GP
*
M
GP
0
N
GP
0.3
74
—
100
�TABLE 23 (Continued)
Annual Mean Number
of Spontaneous
Abortion Cases,
Percent
1972 - 1977
Hospitalized
North Lincoln Hospital
Lincoln City, OR
(8 Physicians)
Type of
Patients
O
GP
5.0
80
P
GP
5.0
80
Q
Hospital
Contacted
Physician
GP
*
GP
*
GP
*
R
S
—
—
—
Good Samaritan Hospital
Corvallis, OR
T
GP
5.0
80
U5
GP
1
50
V6
GP
3.9
30
w
GP
*
X
—
Internist
0
Y
OB/GYN
50.0
—
10
z
OB/GYN
50.0
10
AA
OB/GYN
50.0
10
BB
OB/GYN
50.0
10
CC
Internist
0
DD
GP
11
75
—
0
�TABLE 23 (Continued)
Hospital
Contacted
Physician
Type of
Patients
Annual Mean Number
of Spontaneous
Percent
Abortion Cases,
1972 - 1977
Hospitalized
EE
OB/GYN
50.0
FF
OB/GYN
0
GG
OB/GYN
50.0
10
KH
Good Samaritan Hospital
Gorvallis, OR
(continued)
OB/GYN
50.0
10
Refused to release data.
^•Physician C has practiced in the community since January 1978.
2
Physician D has practiced in the community since 1975.
^Physician E has practiced in the community since 1977.
Physician J has practiced in the community since 1976.
^Physician U has not taken any obstetric cases in past three years.
Physician V has practiced for 14 months.
10
�Chapter V
SUMMARY AND CONCLUSIONS
the objectives of the study were to test the following hypotheses:
(a) whether or not differences in spontaneous abortion rates
between the study and control population;
(b) whether or not seasonal variations in rates exist within
the study and control populations;
(c) whether or not such variations, if they exist, can be
associated with time and concentration of spray applications
in the Study area.
The statistical einalyses of the spontaneous abortion and spray data
presented herein to test these hypotheses have demonstrated that:
1.
The 1972-77 abortion rate index for the Study area is
significantly higher than those for either the Control or
the Urban area.
2.
There is a statistically significant seasonal cycle in the
abortion index in each of the areas with a period of about
4 months. In particular/ there is an outstanding peak in
�3.
There is a statistically significant cross-correlation
j
between the Study area spontaneous abortion index and
spray patterns in terms of pounds applied by months in the
Alsea basin, 1972-77, after a lag time of 2 or 3 months.
The results based on the six-year aggregate of abortion and spray data
are confirmed and enhanced by the analysis of the two 3-year
aggregates. The cyclic time patterns are consistent over time
periods; the Study area shows an elevated abortion index over the
Control and the Urban areas; and there is a peak in the abortion index
which correlates positively with the spray pattern in the Alsea basin
after a lag time of 2 to 3 months.
For all its complexity, however, this analysis is a correlational
analysis, and correlation does not necessarily mean causation.
78
S^
�REFERENCES
1.
Rebuttable Presumption Against Registration and Continued
Registration of Pesticide Products Containing 2,4,5-T.
Environmental Protection Agency, Office of Pesticide Programs,
Washington, D.C., Federal Register [6560-01] Vol. 43, No. 78 Friday, April 21, 1978.
2.
Farm Chemicals Handbook - 1978. R. L. Meister, Jr. (ed.),
Meister Publ. Co., Willoughby, Ohio, pp. 250-251, 1978.
79
�80
APPENDIX A
QUESTIONNAIRE: Oregon Miscarriage Investigation
NAME
DATE
ADDRESS
Prior addresses:
COUNTY
STATE
CITY
COUNTY
ccuNTr
STATE
to
19
From 19__
to
19
STATE
CITY
to 19
From 19
^YY
From 19
Family Physician:
NAME
ADDRESS
]
Date of Last Complete Physicial Examination
PHONE
�I'RKCNANCY DATA
1'lease list all pregnancies — If more than 5, ntUl pages. If the answer to any
question Is "yes" please specify time in weeks of gestation and describe fully.
PKECNANCIES
1st
2nd
3rd
4th
5th
Month_uml_year of conception
Residence during pregnancy:
(City, County, State)
Indicate live births
(give date)
(I) Full Term
___
I'remalire
S t i l l births/Spontaneous
°"
Induced abortions
([•ive date)
Weeks of gestation
., 02.)
Any hi rili defects, mental
-deficiency, or other
eongun 11 a 1 J mpa1rment ?
(Answer No, Yes_ (describe)
«ir unknown)
oo
�PREGNANCY DATA
CONTINUED
PREGNANCIES
1st
2nd
3rd
Ath
5th
For HvebJrths, did you
experience spotting during
pregnancy?
For inlHcnrrlageH and
jjtl 1 ll»i rl liJi, was there a
long period of illness
find spotting, or was
miscarriage sudden?
Weight change during
proRiiuncy (Ib).
Hid your doctor tell
you that your weight
gain WUH excesKive?
Old your doctor indicate
that you had high blood
pressure during
pregnancy?
Did your doctor Indicate
that you had kidney
In fuel lun during
pregnancy? (Weeks of
I'rsl at Ion)
a
00
to
�PREGNANCY DATA
CONTINUED
PREGNANCIES
1st
2nd
3rd
Ath
5th
Did lie prescribe any
iHtiiitr.u tions for weight
control, high blood
pressure, or kidney
Infections?
If yes
pleaae specify
mod i en t Ions.
Did you have any
respiratory Infections
during pregnancy?
i^mH.kji J*_L ^ejitat Ion)
Did you use any of the
following medicat Ion
prior to or during
pregnancy? If yes,
Indicate how often.
1. Orn.l contraceptives
2. Hormone H
3. Antlnauseants
Ant Ih istamines
A.
5. AiULL'LljL
0.
7.
8.
0.
10.
A n I I b l o t Irs (Specify)
SJ ji£ p I ng^ ,P_n_1_s_
V] _tanij n s
Traiu|u( L i z u r s
(specify)
Other medications
(spec i f y)
to
�PREHNANCY 1>ATA
CONTINUED
PRECNANCIES
1st
2nd
3rd
4th
5th
Hit! you use any of the
following prior to or
during pregnance? (Specify
kind, amounts, frequency
and when)
J. Tobacco
2.
Alcohol
j.
Marijuana
4.
"Drill's" (coccaine,
I.SD, etc.)
Did you suffer any physical
Jujuries during pregnancy?
(If yea, .show kind, weeks
of pregnancy, was physicians
<:;\fi". required.)
Were you X-rayed during
pregnancy? (If yes,
'-lLLy- LYIM1 and how often)
oo
.u
�PREGNANCY DATA
CONTINUED
PREGNANCIES
1st
2nd
3rd
4th
5th
What household chemicals
(paints, garment cleaners,
solvents, pesticides) were
yon exposed to prior to or
during pregnancy? (Specify
J2nil how often)
MUM your pregnancy confirmed
by a physician if an early
ml jU'nrr I age?
Was your ml.scarrlage
by a physician?
oo
in
�86
PREGNANCY OUTCOME OF SISTERS
For each of your sisters, please list pregnancy outcomes.
Total number of pregnancies
City of
Residence
1.
2.
3.
4.
5.
6.
Present
Age
Full
term
Premature
Spontan.
Induced
Abortions Abortions
�87
MEDICAL HISTORY
A. SUBJECT'S:
Have you had any of the following diseases or disorders in you
lifetime?
NO
1. Measles:
Regular
German
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Infectious heptatitis
Pneumonia, other
respiratory infections
Allergies (asthma, hay
fever, drug reactions)
Blood disorders, anemia
Heart disease
High blood pressure
Kidney disease; cystic
nephritis, etc.
Diabetes
Thyroid disorders
Stomach disorders, ulcer
Small bowel disorders,
ileitis
Large bowel disorders,
colitis
Nervous system disease,
central, peripheral
Epilepsy
Muscle system disorders,
paralysis
Skeletal (bone)
disorders
Skin disease
V.D.
Congenital defects
Mental illness
Other (Please List)
YES
If yes,
when/how long?
�88
MEDICAL HISTORY (CONTINUED)
B. Subject's hospitallzations, excluding pregnancies:
APPROXIMATE
DATE
PLACE
LENGTH Or
STAY
REASON
C. MENSTRUAL INFORMATION:
1. Are your menstrual periods regular?
No,
„ Yes.
2. Would you characterize your menstrual period as (check one):
a. Probably light
b. Probably average
c. Probably heavy
D. SUBJECT'S FAMILY
!• Did subject*s_father, or father's brother(s) or sister(s) suffer
from any congenital defect? ,.L
No,
Yes,
Not sure.
If yes, please list:
Father
His brother(s)
His sister(s)
Bid subject's mother, or mother's brother(s) or sister(s) suffer
from any congenital defect?
No,
Yes,
Not sure.
If yes, please list:
Mother
Her brother(s)
Her sister(s)
•
�89
MEDICAL HISTORY (CONTINUED-)
- Mother's pregnancies:
Total number
How many were full term
How many miscarriages
,
?
?
If known, during which week(s) of pregnancy
4. Did any of subject's brothers suffer from any congenital defect?
No, ___ Yes,
Not sure.
If yes, list type of defect:
5. Major illnesses or diseases of brothers:
6. Did any of subject's sisters suffer from any congenital defect?
No,
Yes,
No sure.
If yes, list type of defect: __
7. Major illnesses or diseases of sisters:
�ENVIRONMENTAL ASPECTS
PREGNANCIES
1st
How m:iny miles la (was) your
residence or place of work
from the nearest area sprayed
wil.li herbicides during each
pregnancy?
t.o«ihi you ever smell the
chemical at the time Jt
was applied to the Forest?
(If yes, specify during
wh^ch pregnancies.)
What IB (was) the source
of your water supply?
^VI» t
you ever aware of a.
iu the taste of
your water during or
soon after herbicide.
•MM'! ll'ilt: l()_u''
Were your garden &
flowers, trees, shrubs
ever damaged when
herhicide \j.\u applied
lo I lit- forest?
2nd
3rd
4th
5th
�F.NV1 RONMENTAI. ASPECTS
CONTINUED
PREGNANCIES
, _ . . . . _ . _ _ - . _ , . . _..., r ~
2nd
1st
3rd
4th
— —
••• ~
" ••
5th
hiring
_ No,
rise
etc.)
ever been tested for pesticide residues?
hen
s?
1
By whom
(list)
lere they may he obtained
No,
Yes,
Not sure,
vo
�92
FOOD SUPPLY AND DIET
A. What percentage of your meat and poultry products are raised
locally?
%,
B. What percentage of your fruits and vegetables are raised
locally?
%.
C. Does your milk come from local cows?
Some(%).
No,
D. Do you have a home garden?
Yes.
No,
Yes,
E. Do you eat game taken from local forest areas that have been
treated with herbicides?
No,
Yes. Kind?
Approximate number of meals per year
.
�93
HOUSEHOLD PESTICIDE USAGE
1. Has a commercial applicator treated your residence for pests within
the past five (5) years? _ Yes, _ No, _ Unknown.
2. Has your residence ever been treated for termites? __ No, _
Unknown.
rr
Yes,
3. Within the past five (5) years has your household used no-pest strips?
_ Yes, _ No, _ Unknown.
4. Within the past five (5) years has your household used any of the
following pesticides on pets?
Insecticide Collar _ Yes, _ No, _ NA
Insecticide Shampoo _ Yes, _ No, _ NA
Insecticide Powder _ Yes, _ No, _ NA
Other
.
__
;
__
5. Within the past five (5) years have you used moth balls, crystals,
flakes, or aerosols? _
Yes, _
No, _
Unknown.
6. Within the past five (5) years have pesticides been used in the:
House _ Yes,
No.
Garden _ Yes ,_ No, _ NA.
Yard _ Yes, __ No, _ NA..
7. Are any pesticides stored on the premises? _
,_ n _ Unknown.
Yes, __ No,,
3. Within the past five (5) years have you used any disinfectants?
_ Yes, _ No, _ Unknown.
�94
02H.G1S 10K
S3EIO
1KZJ3SVE
KD'dC;Z >1DY£
C3HS
33VHVO
1
i
I
IP** 03 S ',
app-psad STU3 st i
•TPB.IC aq3 uo aj»M*1 i
UZH10
G2E ^ZM01£
g
K3CKVD 31ffY13D3^i
3SBC sq3 UT: pssn (•OMB i uFq3 Z 3 U ^ C) CHYA
NWYT
sfq3 SBU sastn
30VEVO
sad aq3 uo ejeq^i
asnoH
isasfmaad aq3 uo pajoss X^suaaano
BIB appjassd STUS 10 saouno XUBTU «.OH
T pasn p-jouasnoq jnoX 30 aaqnam XUB ao
aABq apTOT3S9d STU3 io Eaouno Xusc WOK
i-thot- SJB s-[osoaav) icuoz pT^os anos
ao pTnbT~[ UT appT^sed s-pq i si ainbn
p-jo SO.A 8T>- ailED
•
P'[o SAA sT<-iinoY "asnio
31VK3£ - OV3H GlOHHSnOH
3TVK - OV3H QlOHHSnOH
stq3 jpasn ou_fi
TV13H3KKOO
t:
o
C? W
«H »-4
^J C
.
*- o
c
•uwouVj 57 aaqnnu
3"j3sad pasn-uou jfB go saireu
©q3 pue saBaX aAfg 3sed aq3
pioqasnoq aqs Xq pasn sap^o
-T3sad TJB 30 saniBU aac 3Bu/i
PESTTCtDE NAME
w
e.
_
—
1
—
—i
_
i
—
—
1
—
— — —
— —
—
—
—
.
—
i
ii
—
i
—
i
1
—
—
—•
—
—
—
i
__
*••
••*
_
—
—
—
—
—
—
—
—
—
—
—
—
-
*
—
•W
.
MOT
—
*•"
_
—
-
—
—
E
-
— —
—
****
—
•*•
—
—
—
-
—
— —
—
—
—
••"
—
MM
—
— —
—
—
—
— — —
—
—
_
—
—
—
—
V ^~s
H O
L/*' C^
1
i
i
i
1
—
-
K
s
;
!
—
KOOS nniin
«3K1C-K3E01I2
htws ^sQ!wr:-:ISEDIIX
|a3yc)1S IOK
j,aat:TE3uo: T EUlSTJO S3f
OK
;T tpaaoss app sad siq3 sj
S3i
CIHOIS IOX
i,A*avL pu'E spoi aapun
OK
paao3s apppsad STqs si
S31
3KOX
•
'HiKIO
i SlSa 5 K3EC1IED 01 SilKIl-iJO YZHY
SGNTn HSVfi
^apTOT3sac r
XSVK
S"j:q3 2n-isn uaqw ua^B3 T
S3AOTD
5 SUO-pn^Oaid JO SdA3 3^Uf.
S100E
^sqsubc 21 3SEC aq3 uf pasn
• sj^a/
i _
i __
^™*
—
— —
—
—
—
— —
—
— —
—
™*
—
—
�95
OCCUPATIONAL HISTORY OF SUBJECT
Please list any employment:
Employer
(location)
Describe
Type of Business
Job Title
Beginning and ending
Dates
Do you launder your husband 1 s work clothes with family wash?
Yes
No
OCCUPATIONAL HISTORY OF HUSBAND
Employer
(location)
1.
2.
3.
4.
5.
6.
Describe
Type of Business
Job Title
Beginning and ending
Dates
�96
THESE QUESTIONS REFER TO HUSBAND'S OCCUPATION:
Was This Job Was Protective
Considered
Equipment, or
Dangerous
Clothing Available
(1)
Did He Inhale
Did Chemical Solvents
Chemical Solvents
Oils,, Dusts, etc. Get
Dust or Other Fumes? on His Skin/Clothes?
Yes
No
If .yes,
why?
Yes
No
If yes, what
was it?
Yes
No
If yes, list if
known:
(2)
Yes
Yes
Yes.
No
No
If yes, "what
No
If yes, list if
known:
•If yes,
why?
was it?
Yes
No
If yes~n 1st
if known:
Yes
No
If yesTl ist
if known:
(3)
Yes
Yes
.
No
Yes
No
If yes, what
No
If yes" list if
known:
Yes
No
If yes, 1 1st
if known:
was it?
Yes
No
If yes"7 list if
known:
Yes
No
If yes, 1 ist
if known:
Yes
No
If yes, what
was it?
Yes
No
If yes, list if
known:
Yes
No
If yes, 1 ist
if known:
If yes,
why?
was it?
(4)
Yes
Yes
No
If yes,
why?
No
If yes, what
(5)
Yes
'
No
If yes,
why?
�97
THESE QUESTIONS REFER TO HUSBAND'S OCCUPATION CONTINUED:
Was This Job Was Protective
Considered
Equipment or
Dangerous
Clothing Available
(6)
Yes
No
If yes,
why?
Yes
No
If yes, what
was it?
Did He Inhale
Did Chemical Solvents
Chemical Solvents
Oils, Dusts, etc. Get
Dust or Other Fumes? on His Skin/Clothes?
Yes
Yes
No ~
~
If yes, list
if known:
No " ~
"
If yes, list if
known:
1. Did your husband serve in Vietnam?
service.
Yes,
No,
Months of
2. If ves, was he directly involved with spraying of Agent Orange ___
Yes.
No,
3. Was he otherwise exposed to Agent Orange during his military activities?
None;
Some;
Often.
4. Has your husband ever had a sperm examination?
If yes, were abnormalities suspected?
5. Has your husband ever had V.D.?
No knowledge.
6. Has your husband ever used marijuana?
No know!edge.
No,
No,
No,
Yes.
Yes (specify),
No,
Yes,
7. Has your husband ever used "drugs" such as coccaine, LOS, etc?
No,
Yes,
No knowledge.
Yes.
�EDUCATIONAL BACKGROUND
Did you graduate from high school?
No,
Yes, Class of 19_
If not, what was the highest grade completed?
Did you graduate from college?
degree's
and year's
If not, did you attend college?
No,
.
No, __i
.
Yes,
Yes,
Did your husband graduate from high school?
Class of 19 .
Years.
No,
Yes,
If not, what was the highest grade he completed?
.
Did your husband graduate from college?
No,
List degree1 s
and year' s
Yes,
.
If not, di'd your husband attend college?
No,
Yes,
Years,
�LA » 0 R
ROOK
S T A T I S T I C S
Jtdutaa Kescrial Sbspitzl
APPENDIX B
1977-1S7S
CC
liLl
76
77*76
WOV. DSC.
5S4
352
570
461
563
491
(•40
Z*
503
I
-.
MSL
I m I i*
t* nft
*
(471
t
I
IS73
I
f?c
4.74.
I
I
!
eat
I
I
1531 1 5 2 4
47
76-'
7-7 S
SS?T.
wVt
I go
.<tc
7i
•13
65
il jLiveSirt'-s
76-77
1
l 29
Staff
is -1
3C
•7
i= • ; iS5£
1515 it s s 1 52=
1570
=
1
i ...2_2
U45
L22
L 1?_
i 13
I
;~"
I 4=2
i 514C
ii:«
I
;
Private
I
I
1423
21
I
1425 |4S5 I 4 4 C 134?
1472
3
1273 1 3 7 =
1257 US!
I1C5 1125
2
I 94
I 67
I1C1 I 1 C 2
I
I 12
.14
K=J U42
I
1529
90
I 95
e
I
I £2
jiCC
i £5
I 9£
I 72 I 65
I rC I 95
95
i 95
I £5 I 52
I
t
5 I 10
S i
Staff
:ulti Pr*cr.ar.cias
7£-77
4
9
|
! 12
I 6:, I 9C
I
I
"
66
t 51
3
I
! I
I 12
i -i
1.6.
I
I
I 71 I 63
j 43
i
4
I
I
I
| 42
2
S
I
'
I
/o- 77-??
I IT !
*3C4
I 24
U3S
/ o— /•
Stiff
77-7S
I 19
1270 1422 121"
24
1545
Staff
i3=2 I 39C
I 13
HC2 I 77
I5C7 | 464
I
I
i'4i£ i 4 7 5
I
1470
Staff
I
t
| 54 j 73
*fi
J
L
i
(
j 1"
I
!
62
1123
�
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.
0421
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/.
Description
An account of the resource
<strong>Corporate Author: </strong>Epidemiologic Studies Program, Human Effects Monitoring Branch, Benefits and Field Studies Division, Environmental Protection Agency (EPA), Washington, D.C.
Date
A point or period of time associated with an event in the lifecycle of the resource
February 28 1979
Title
A name given to the resource
Report of Assessment of a Field Investigation of Six-Year Spontaneous Abortion Rates in Three Oregon Areas in Relation to Forest 2,4,5-T Spray Practices
Subject
The topic of the resource
dioxin
congenital birth defects
civilian impact
herbicide application
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.
072
Folder
The folder containing the original item.
1907
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
Jessop
Description
An account of the resource
<strong>Corporate Author: </strong>Commonwealth of Australia, Senate, Standing Committee on Science and the Environment
Date
A point or period of time associated with an event in the lifecycle of the resource
March 31 1982
Title
A name given to the resource
Reference: Pesticides Inquiry (Official Hansard Report), Sydney, Wednesday, 31 March 1982.
Subject
The topic of the resource
Australian Parliamentary hearings
veteran health and hygiene
herbicide toxicology
congenital birth defects
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.
040
Folder
The folder containing the original item.
0904
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/.
Description
An account of the resource
<strong>Corporate Author: </strong>International Scientific Advisory Committee for Seveso
Title
A name given to the resource
Recommendations for the 4th Meeting of the International Scientific Advisory Committee for Seveso, January 10-14, 1981
Subject
The topic of the resource
Italy
health studies
congenital birth defects
industrial exposure
ao_seriesIII