United States Science Advisory EPA-SAB-EHC-94-005
Environmental Board (1400F) March 1994
Protection Agency
AN SAB REPORT:
ASSESSMENT OF POTENTIAL
2,4-D CARCINOGENICITY
REVIEW OF THE
EPIDEMIOLOGICAL AND
OTHER DATA ON POTENTIAL
CARCINOGENICITY OF 2,4-D
BY THE SAB/SAP JOINT
COMMITTEE
-------
NOTICE
This report has been written as a part of the activities of the Science Advisory Board,
a public advisory group providing extramural scientific information and advice to the
Administrator and other officials of the Environmental Protection Agency. The Board is
structured to provide balanced, expert assessment of scientific matters relating to
problems facing the Agency. This report has not been reviewed for approval by the
Agency and, therefore, the contents of this report do not necessarily represent the views
and policies of the Environmental Protection Agency, nor of other agencies in the
Executive Branch of the Federal government, nor does mention of trade names or
commercial products constitute a recommendation for use.
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
ROSTER
SAB/SAP SPECIAL JOINT COMMITTEE
Chair
Dr. Genevieve Matanoski, Department of Epidemiology, School of Hygiene and Public Health, The
Johns Hopkins University, Baltimore, MD
Members and Consultants
Dr. Donald F. Austin, Center for Disease Prevention & Epidemiology, Oregon Health Division,
Portland, OR
Dr. Raymond Greenberg, Emory School of Public Health, Atlanta, GA
Dr. David Hoel, Department of Biostatistics, Epidemiology and Systems Science, Medical
University of South Carolina, Charleston, SC
Dr. Eugene McConnell, Raleigh, NC
Dr. Barbara McKnight, Department of Biostatistics, University of Washington,
Seattle, WA
Dr. Richard Monson, Department of Epidemiology, Harvard University School of Public Health,
Boston, MA
Dr. James Popp, Sterling Winthrop, Inc., Collegeville, PA
Dr. Walter Stewart, Department of Epidemiology, School of Hygiene and Public Health
The Johns Hopkins University, Baltimore, MD
Dr. Daniel Wartenberg, Environmental and Occupational Health Sciences Institute of New Jersey,
Piscataway, NJ
EPA STAFF
Mr. Samuel Rondberg, Designated Federal Officer Science Advisory Board (1400F), U.S. Environ-
mental Protection Agency, Washington, DC 20460
Mr. Bruce Jaeger, Designated Federal Officer, Scientific Advisory Panel (H7509C), U.S. EPA, 401
M St., SW, Washington DC 20460
Ms. Mary L. Winston, Support Secretary, Science Advisory Board (1400F), U.S. Environmental
Protection Agency, Washington, DC 20460
-------
ABSTRACT
In August 1980, the EPA required oncogenicity testing of 2,4-D
(2,4-dichlorophenoxyacetic acid. EPA reviewed the results of those studies completed
to date (some of which reported an association of phenoxy herbicides, including 2,4-D,
and non-Hodgkin's lymphoma (NHL)) and requested that a joint Committee of the
Science Advisory Board and the Scientific Advisory Panel review the epidemiologic
studies and other available relevant data. A joint Committee was formed, and met in
Arlington, Virginia on April 1-2, 1993 to review human/canine epidemiological studies
and animal toxicology studies re possible human carcinogenicity and mutagenicity.
Epidemiologic cohort studies have generally shown no increased risk of cancer,
albeit that all of the populations for which specific exposure to 2,4-D have been identified
were small, and the follow-up period usually short. Some case-control studies have
shown a risk of Non-Hodgkin's Lymphoma (NHL) in association with farming but many of
these studies did not control for exposure to other agents in addition to 2,4,D. The
Committee concluded that current studies cannot distinguish whether observed risks
reported are due to the use of 2,4,D. The single canine epidemiologic study suggested
that pet dogs may be at risk from exposure to 2,4,D or to areas treated by a lawn care
service. Although this study is supportive of a finding of carcinogenicity, there are ques-
tions about its applicability to human carcinogenicity because of poor information on
exposure and possible non-comparability between canine and human lymphomas.
Toxicology studies show that rats (but not other animal species tested) may develop
astrocytomas from exposure to 2,4,D, but this outcome has not been reported in the
human studies. An ongoing rat study at higher doses will clarify whether this finding is
treatment-related or not. Tests of 2,4-D have not shown any mutagenic changes under
experimental situations.
The Committee concludes that the data are not sufficient to find that there is a
cause and effect relationship between the exposure to 2,4,D and NHL. Because there is
some evidence that NHL occurs in excess in populations that are likely to have been
exposed to 2,4,D, there should be continued examination of the issue through further
studies. Other data gaps exist, and decision-making on 2,4-D would benefit from com-
pletion of rodent studies previously requested by EPA, particularly further animal
carcinogenicity studies that test 2,4-D jointly with other substances that might reflect the
human exposure situation; a replication of the dog epidemiology study; additional
case/control studies, with careful attention to exposures; additional human cohort studies
designed to assess both relative risk of NHL and the comparative risk of all mortality;
and additional follow up and analysis of worker cohorts involved in the production of 2,4-
D.
KEYWORDS: 2,4,D; 2,4-dichlorophenoxyacetic acid; carcinogenicity; astrocytoma; Non-
Hodgkin's Lymphoma (NHL); farming; canine epidemiology
-------
IV
-------
TABLE OF CONTENTS
1 EXECUTIVE SUMMARY 1
2 INTRODUCTION 4
2.1 Background 4
2.2 Charge 4
3. DETAILED FINDINGS 7
3.1 Epidemiologic Studies and Human Carcinogenicity 7
3.1.1 Epidemiologic Evidence 7
3.1.2 Types of Human Cancer Associated with 2,4-D Exposure 10
3.1.3 Specificity of Information Relating to 2,4-D 11
3.1.4 Strengths and Limitations of the Human Epidemiology Data-
base 12
3.2 Canine Epidemiologic Study 14
3.2.1 Exposure to 2,4-D and Cancer in Dogs 14
3.3 Animal Testing and Carcinogenicity 17
3.4 Mutagenicity of 2,4-D 18
3.5 Other Data and Data Gaps 19
3.6 Weight of Evidence for 2,4-D Carcinogenicity 21
4 CONCLUSIONS 23
5 REFERENCES 25
-------
1 EXECUTIVE SUMMARY
In August 1980, the EPA required oncogenicity testing in rats and mice of 2,4-D
(2,4-dichlorophenoxyacetic acid - one of the first herbicides to be registered in the
United States. To date, the registrant has not completed all of the required studies).
The EPA reviewed the results of those laboratory studies that were completed, and
extant epidemiologic studies reporting an association of phenoxy herbicides, including
2,4-D, and non-Hodgkin's lymphoma (NHL). Continuing reviews of the data prompted
the EPA to issue notice that the Agency had deferred decision on the final determination
of the Special Review to obtain additional information. Subsequently, the Agency
requested that a joint Committee of the Science Advisory Board and the Scientific
Advisory Panel review the epidemiologic studies and other available relevant data. Such
a joint Committee was formed, and met in Arlington, Virginia on April 1-2, 1993. The
meeting was structured around the Charge summarized below:
a. (1) Do the human epidemiologic studies provide evidence that 2,4-D is a
human carcinogen?
(2) Does the canine malignant lymphoma epidemiologic study demonstrate a
relationship between exposure to 2,4-D and cancer in dogs? Is this study
supportive or non-supportive of a finding of human carcinogenicity for
2,4-D?
b. Do data from laboratory animal testing of 2,4-D provide evidence of
carcinogenicity in animals? Do these studies provide evidence of human
carcinogenicity for 2,4-D?
c. What are the endpoints of concern, if any, from the mutagenicity database
for 2,4-D? Do any of these data elicit concern as to the carcinogenicity of
2,4-D in animals or humans?
d. (1) What other existing data or information should be considered in the
weight-of-evidence evaluation of the carcinogenicity of 2,4-D and how
should they be incorporated?
(2) Are there informational gaps in the existing carcinogenicity database?
-------
e. Considering all of the above items, what is the overall perception of the
panel as to the weight of evidence that 2,4-D is a human carcinogen?
f. If it is concluded that there is some evidence of carcinogenicity, what
methods might be used to evaluate the dose-response (potency) relation-
ships either qualitatively or quantitatively?
The extant epidemiological cohort studies on persons involved with the manufac-
ture or use of 2,4-D generally do not have sufficient power (because of short follow-up
time and small study populations) to determine if there is any association between
modest excess risk of STS or NHL and exposure to 2,4-D. Case-control studies have
shown an excess risk of NHL in association with the occupation of farming, but, because
of likely exposure to multiple chemical agents, it is not clear if this risk is indeed
associated with exposure to 2,4-D itself.
The single epidemiologic study of pet dogs suggested that these animals may be
at risk from exposure to 2,4,D or use of a lawn care service. This study has the same
constraints of selection bias of cases and controls and recall and misclassification
biases that could occur in human case-control studies. The applicability of the canine
epidemiology data to lymphoma in humans is questionable since the cancers may not be
similar in dogs and humans, and exposures to 2,4-D are not clearly established. To
substantiate these results, the study should be replicated with improvement in the
exposure measures and verification of the comparability of the cancers in dogs and
humans.
Requested animal test data at appropriate dose levels have not yet been
completed. Nonetheless, the studies completed so far show that rats (but not other
animal species tested) may develop astrocytomas from exposure to 2,4,D, an outcome
which has not been reported in the human studies. An ongoing rat study at higher doses
will clarify whether this finding is treatment-related or not. The chemical has not shown
mutagenic changes under experimental situations - it would have strengthened the
observations in humans if there were laboratory data, especially in toxicologic studies of
animals, to support any observations in humans.
The Committee concludes that, at this time, the data are not sufficient to conclude
that there is a cause and effect relationship between the exposure to 2,4,D and NHL.
-------
Because there is some evidence that NHL occurs in excess in populations that are likely
to have been exposed to 2,4,D, there should be continued examination of the issue
through further studies. The following research and data would be of most value:
a) completion of EPA requested rodent carcinogenicity studies and dog
toxicology studies
b) animal carcinogenicity studies that test 2,4-D with other substances that
might reflect the human exposure situation
c) a replication of the dog epidemiology study
d) additional case/control studies, with careful attention to exposures, particu-
larly multiple exposures
e) human studies, particularly cohort studies designed to assess both relative
risk of NHL and the comparative risk of all mortality (or all disease inci-
dence, if possible)
f) additional follow up and analysis of worker cohorts involved in the produc-
tion of 2,4-D
-------
2 INTRODUCTION
2.1 Background
2,4-D (2,4-dichlorophenoxyacetic acid) was one of the first herbicides to be regis-
tered in the United States. This herbicide and its derivatives are commonly used on
crops such as corn, wheat, sugar cane, rice, and on pasture land, to control broadleaf
weeds. It is also used for weed control in residential settings, in forest management, and
for growth control on certain crops.
In August 1980, EPA required oncogenicity testing in rats and mice of 2,4-D. The
EPA reviewed the results of these laboratory studies and requested additional rat and
mouse carcinogenicity studies which have not yet been completed. At that time, extant
epidemiologic studies reported an association of phenoxy herbicides, including 2,4-D,
and non-Hodgkin's lymphoma (NHL). This evidence prompted EPA to issue a prelimi-
nary notification on September 22,1986 announcing that there would be a Special
Review. Continuing reviews of the data in the subsequent time periods have prompted
the EPA to issue notice that the Agency had deferred the decision on the final determi-
nation of the Special Review for the purpose of obtaining additional information.
Subsequently, to evaluate the additional information it had gathered, the Agency
requested that a joint Committee of the Science Advisory Board and the Scientific
Advisory Panel review the epidemiologic studies and other available relevant data. Such
a joint Committee was formed, and met in Arlington, Virginia on April 1-2, 1993. The
meeting was structured around the Charge described below.
2.2 Charge
The Agency asked the Special Joint Committee of the Science Advisory Board
and the Scientific Advisory Panel to evaluate the weight of evidence of carcinogenicity of
2,4-D. Such an evaluation requires consideration of all evidence that supports and does
not support a conclusion of carcinogenicity, along with the strengths, weaknesses, and
uncertainties of the database. Although the Agency was interested in any insights the
Committee may wish to share related to the carcinogenicity of 2,4-D, there was particular
interest in the Committee's response to the following specific questions:
a. (1) Do the human epidemiologic studies provide evidence that 2,4-D is a
-------
human carcinogen? What types of human cancer, if any, are associated
with exposure to 2,4-D? How specific is the information to 2,4-D versus
other pesticidal and non-pesticidal sources? What are the strengths and
limitations of the human epidemiology database?
(2) Does the canine malignant lymphoma epidemiologic study demonstrate a
relationship between exposure to 2,4-D and cancer in dogs? Is this study
supportive or non-supportive of a finding of human carcinogenicity for
2,4-D? What are the strengths and limitations of the canine lymphoma
study?
b. Do data from laboratory animal testing of 2,4-D provide evidence of
carcinogenicity in animals? Do these studies provide evidence of human
carcinogenicity for 2,4-D?
c. What are the endpoints of concern, if any, from the mutagenicity database
for 2,4-D? Do any of these data elicit concern as to the carcinogenicity of
2,4-D in animals or humans?
d. (1) What other existing data or information should be considered in the
weight-of-evidence evaluation of the carcinogenicity of 2,4-D and how
should they be incorporated? Examples might include information on: expo-
sure, absorption, metabolism in humans and animals, contaminants and
inert materials in 2,4-D products.
(2) Are there informational gaps in the existing carcinogenicity database, and if
so, can the committee suggest specific studies or other data which should
be developed and/or submitted to the Agency?
e. Considering all of the above items, what is the overall perception of the
panel as to the weight of evidence that 2,4-D is a human carcinogen?
f. If it is concluded that there is some evidence of carcinogenicity, what
methods might be used to evaluate the dose-response (potency) relation-
ships either qualitatively or quantitatively? What data might be needed to
enhance this capability?
-------
For the purposes of this review, the Committee was asked to consider the use of
the following descriptive terms to express its conclusions about the weight-of-evidence of
carcinogenicity of 2,4-D as requested in Question (5) above:
proven human carcinogen
highly probable that 2,4-D is a human carcinogen
probable that 2,4-D is a human carcinogen
somewhat probable that 2,4-D is a human carcinogen
improbable that 2,4-D is a human carcinogen
highly improbable that 2,4-D is a human carcinogen
inadequate data for conclusions about human carcinogenicity
These terms do not refer to any existing classification system for carcinogenicity
used by governmental or international organizations. The precise way in which the
Committee will consider the evidence for and against the human carcinogenicity of 2,4-D
and the relative weights given individual elements is left to the committee. We do
request that the committee develop a rationale for their conclusions that will accompany
the descriptive term it applies to the weight of evidence of carcinogenicity.
-------
3. DETAILED FINDINGS
3.1 Epidemiologic Studies and Human Carcinogenicity
3.1.1 Epidemiologic Evidence
The epidemiologic data reviewed by the Committee included case-control studies
(primarily non-Hodgkin lymphoma cases), cohort studies of manufacturers of, and
industrial users, of pesticides, and several reviews of the information available from
epidemiologic studies. Recent studies received a more thorough review than those in
the past. Following the discussion of the studies and the reviews, the Committee
deliberated on all of the information to reach its conclusions.
The available epidemiologic studies have different characteristics. Some include
occupational groups which might have had exposure to phenoxyherbicides in applica-
tions as farmers, lawn specialists and forest workers. Other studies include workers who
have been involved in production of the chemicals in the workplace. Many of these
occupational studies have attempted to identify the specific chemical exposures of the
workers. In addition, studies have included case-control studies of NHL (non-Hodgkin's
lymphoma) to try to identify the specific risk factors for cases with this condition. One
problem which all studies have faced is the need to reconstruct exposures in the distant
past. In addition, most of the populations under study have had exposures to many
chemicals and contaminants in addition to the 2,4-D.
Numerous studies over the past several years have examined the association
between farming, agricultural work or exposures to pesticides and the risk of NHL
(Johnson, 1990). Most of the studies were non-specific in regard to the exact exposure
that might have occurred in the population. However, in general, almost all of these
studies have shown a positive association between this general occupational grouping
and NHL. One particular problem underlies all of these studies, however: farming is
associated with many exposures that depend on activities common to all farming, as well
as those exposures which may depend on the specific type of farm (e.g., livestock versus
agriculture), and on the type of crop and the geographic area in which that crop is
raised. Farmers commonly are one of the large groups of heavy users of herbicides
identified in case-control studies. If a study wishes to attribute a risk of NHL to
herbicides, then it must prove that the risk is not due to farming in general but to the
-------
specific chemical of interest. Most studies have not done this and we are to must
resolve whether any elevated risk is due to 2,4-D, to farming, or to other exposures.
Several case-control studies have examined specific histories of exposures to
pesticides and herbicides in the study subjects. Hardell etal. (1981) in Sweden found
significant associations between soft tissue sarcomas (STS), Hodgkin's disease (HD)
and NHL and phenoxyherbicide use. Vineis etal. (1986) suggested that STS was
associated with exposure to phenoxyherbicides in women in a rice-growing region of
Northern Italy. In a recent update of that study which compared the rates of STS, HD
and NHL in regions with high and low soil levels for 2,4-D and 2,4,5-TP, the incidence of
NHL (but not of STS) was significantly higher in the regions with the high levels. Other
studies from New Zealand and the U.S. found no excesses of STS. Therefore the review
has focused primarily on NHL.
The Hoar et al. (1986) study of 170 Kansas males with NHL reported that farming
(OR 1.4; 95 % Cl 0.9, 2. 1) and phenoxyherbicide use (OR 2.2; 95 % Cl 1.2, 4.1) were
associated with a risk of NHL when non-farmers were used as a comparison. This
observation, in part, prompted the 1986 announcement of a Special Review. The odds
ratios for phenoxyherbicide use in general, and 2,4-D use in particular, increased with
increasing duration of use and frequency of use per year. Almost all subjects reporting
phenoxyherbicide use had used 2,4-D. Most subjects had exposure to other chemicals
in addition to 2,4-D. A population-based study in Washington state (Woods et al., 1987)
reported that among 576 male NHL cases and 694 male controls, there was an associa-
tion between the disease and farming (OR 1.33, 95% Cl 1.03, 1.7). No statistically
significant association was seen between NHL and phenoxyherbicides even at "high"
exposure levels (OR 1.24; 95% Cl 0.8, 1.9). Forest sprayers had a high odds ratio (OR
4.80; 95% Cl 1.2, 19.4) but there were only 7 cases in this subgroup and all of them
were exposed to a combination of 2,4-D and 2,4,5-T One hundred-eighty- three male
NHL cases in New Zealand (Pearce et al., 1987) showed no association between the
disease and exposure to phenoxyherbicides (OR 1.0; 90% Cl 0.7,1.5). However,
reportedly 2,4,5-T is usually used as a herbicide in New Zealand rather than 2,4-D
(Zahm and Blair, 1992). Another group in Sweden (Persson et al., 1989) has investi-
gated exposures of 106 NHL cases identified from a local oncology registry and found a
strong negative association with farming (OR 0.3; 90% Cl 0.1, 0.7) but an odds ratio of
4.9 (90% Cl 1.3, 18) with exposure to phenoxyherbicides. The Committee notes that this
study included a different mix of cases in the category of NHL since the group includes
8
-------
multiple myeloma and chronic lymphocytic leukemia but excludes the usual NHL
classification (ICD category 202). This study does not provide complete information for
evaluation. Since the number of exposed cases is small, these data have been consid-
ered separately from the other studies.
Hoar-Zahm etal., (1990) conducted a population based, case-control study of
NHL in Eastern Nebraska. Among 201 white male cases and 725 white male controls,
they found no increase in NHL in subjects with a history of ever having worked or lived
on a farm (OR 0.9; 95% Cl 0.6, 1.4). However, when cases who had exposure to mixing
or applying 2,4-D were compared to non-farm workers (who, in this case, never worked
or lived on a farm), the odds ratio increased to 1.5 (95% Cl 0.9,2.5) and there was an
increase in the odds ratio associated with the number of days per year of application.
However, in this study, only three cases were in the subset of those cases with 21 days
or more of exposure. In addition, the authors report that organophosphate exposures
demonstrate similar increases in odds ratios with increasing days of annual exposure.
When the data are corrected for exposure to this other chemical, the odds ratio for ever
handling 2,4-D drops to 1.1. As the authors themselves point out, "Because of the small
numbers of subjects and the high proportion of subjects with multiple exposures, it Is not
possible in this study to entirely disentangle these relationships. There may be some
residual confounding." The recent population-based case-control study in Iowa and
Minnesota of 622 white men with NHL and 1245 white controls (Cantor etal., 1992)
reported an odds ratio associated with the history of being a farmer of 1.2 (95% Cl 1.0,
1.5), with adjustment for other variables such as age, state, vital status, smoking, etc.
The odds ratio for use of phenoxyacetic acids compared to non-farmers was 1.2 (95% Cl
0.9, 1.6). The specific use of 2,4-D showed similar odds ratios in various analyses.
These investigators provided additional information from a subset of re-interviews of
subjects with a history of use of pesticides with exposure days per year evaluated.
These data indicated odds ratios of 0.6 and 0.4 for exposure periods under 10 days of
use and an odds ratio of 1.1 (95% Cl 0.5, 2.4) for those exposed 10+ days per year.
Another recent study from Australia included 52 men in a combined group of HD and
NHL who were compared to cancer and population controls (Smith et al., 1992). There
were no significantly elevated odds ratios but, for the subset who had used
phenoxyherbicides or chlorophenols for more than 30 days total exposure, the odds ratio
was 2.7 (95% Cl 0.7, 9.6). This study did not separate 2,4-D exposure from other expo-
sures, nor HD from NHL .
-------
Johnson (1990) has reviewed the findings from cohort studies of manufacturing
facilities and suggested that few cohorts had sufficient numbers of deaths to enable the
observation of the risk of STS, HD or NHL Thus, all cohorts needed additional follow-up.
Bond's early study (Bond etal., 1988) did show a significant excess of all lymphopoietic
cancers in workers involved in 2,4-D production (0/E, 5/1.6, SMR=312) but there were
only 68 total deaths in the cohort. The subsequent update did not analyze the data
separately for this work area. A study of 4,459 Danish workers (Lynge, 1985) employed
in herbicide manufacture identifies 940 workers who produced the herbicide but showed
no risk of NHL. Wiklund etal. (1986) reported no risk of NHL in Swedish agriculture and
forestry workers but did identify an excess of HD in silviculture and non-traditional
agricultural occupations. In a second study (Wiklund et al., 1989), the incidence rates of
cancers in Swedish pesticide applicators were examined. No significant excess risks
were identified although the authors noted excesses of testicular cancer (SIR = 1.55),
other endocrine cancers (SIR = 1.33) and nervous system cancers (SIR = 1.27). A
Canadian study of 70,000 farmers identified from various farm registries (Wigle, 1990)
reported no overall excess of NHL, However, there was an excess of NHL mortality in
the subgroups which reported $900 or more in fuel costs annually and, for smaller farms
under 1000 acres, the risk of NHL increased with the number of acres sprayed. The
authors report that the chlorophenoxy compound in general use in the area was 2,4-D
but the exposure was not tied directly to cases of the disease. In the large International
Agency for Research on Cancer (IARC, 1990) international study of 18,910 production
workers or sprayers, no excess of NHL was observed but the authors noted
non-significant excesses of STS, cancers of the testicles, thyroid, other endocrine glands
and nose and nasal cavity based on small numbers of deaths.
3.1.2 Types of Human Cancer Associated with 2,4-D Exposure
In the early 1980s, several case-control studies from Sweden suggested a
positive association between exposure to chlorophenoxy herbicides, including 2,4-D,
and cancer in humans. Specifically, associations were suggested with STS and with
NHL. In the late 1980s, two case-control studies in Kansas and Nebraska suggested an
association between exposure to 2,4-D and NHL. No positive association was seen with
STS. These two forms of cancer in humans are the only ones for which a positive
association with 2,4-D has been suggested in any study. There are other case-control
studies of NHL which have failed to reveal an association. Most studies have only
demonstrated an association in subgroups in the populations.
10
-------
In general, follow-up studies of persons engaged in the manufacture and
application of 2,4-D have two limitations. Study populations are small or follow-up time
is (on average) relatively short. As such, these studies generally have insufficient power
at this time to determine if exposure to 2,4-D is associated with a modest excess risk of
STSorNHL.
3.1.3 Specificity of Information Relating to 2,4-D
The epidemiologic studies often do not provide information on exposures specific
to the chemical 2,4-D. Most of the studies relate the risk to the general category of
phenoxyherbicides-a group which might include 2,4,5-T and substances contaminated
with TCDD. If no association is shown and 2,4-D use was limited in either amount or
time, these studies might have missed an association. If there is an apparent risk and
the information on specific exposures is missing, then chemicals other than 2,4-D may
account for the apparent risk.
Several case-control studies and cohort studies are specific with respect to
exposure to 2,4-D only. In the case-control studies in Nebraska and Iowa/Minnesota,
questions were asked specifically about exposure to 2,4-D. However, because the
information on exposure was obtained only by questionnaire, some degree of uncertainty
is attached to the exposure assessment. In addition, the persons have identified several
exposures to chemicals which are often correlated with exposure to 2,4-D. In these
situations it is often impossible to be sure that confounding is removed. Few studies
have examined the risk by duration of exposure in order to establish a cumulated dose
response. In the recent studies the investigators have identified a dose response
relationship, using the number of days per year the subject was exposed as a variable.
However, the measure does not take into account the number of years in which the
subject was exposed at that dose. This measure does appear to have an exposure
response relationship but, when years of exposure are considered there is no response.
This "dose response" is not similar to that usually seen in epidemiology. The exposure
in one year, regardless of when it occurred during a lifetime, creates a higher risk of
disease than multiple exposures over many years. This finding would require further
biologic explanation if it is real. The finding may occur because of the individuals who
fell in the "high" category of yearly use. Selective analysis of small sub-groups in a large
study may lead to erroneous results because the investigators are simply describing the
characteristics of the outliers in the data set
11
-------
3.1.4 Strengths and Limitations of the Human Epidemiology Database
The strength of human epidemiology is that the data relate directly to humans. No
extrapolation from non-human species is needed.
Evaluation of epidemiologic data is difficult because the information is obtained
from non-experimental, observational studies. Biases may arise in the collection of the
data. The possible effect of confounding factors which are related to both the exposure
of interest and the risk of disease may make it difficult to interpret the results. The ability
of epidemiologic studies to provide convincing evidence of causation under such
circumstances is limited. Causation is suspected if several studies are consistent in their
findings; if the association between the agent and the risk of disease is strong (the odds
ratio is high); and if the association adheres to biologic theory. Support from animal data
will help to make the case for causation, particularly by establishing biologic plausibility
and the existence of potential mechanisms. However, failure to detect an association,
particularly if studies have limited statistical power due to small sample size or short
follow up time, are not sufficient to disprove an association. In the case of 2,4-D, the
studies are not consistent and the associations found are weak. The data can not
determine whether the weak association is due to the agent being a weak carcinogen,to
an exposure level which may be low, to uncorrected sources of confounding or bias, or
to random variation in response from study to study which might be experienced if the
relative risk was unity. In general, unless they are extremely large, epidemiologic
studies are not able to provide convincing evidence of health effects for a weak causal
association.
Thus, in assessing the weight-of-the-evidence of the epidemiologic studies which
relate to 2,4-D exposure, one has to make a subjective judgment as to the weights given
to the various studies and their conflicting conclusions. There are different results from
many of the studies; some have indeed shown a relationship between 2,4-D and NHL.
However, there are inconsistencies in the results (even of the positive studies) which
raise doubts as to whether the relationship is causal.
However, most epidemiologists acknowledge that epidemiologic studies are
insensitive and therefore unable to provide unambiguous evidence when exposure to
12
-------
some agent is a weak cause of cancer. Thus, if a toxicant increases the risk among
exposed persons by some small percentage, say ten percent, epidemiologic studies are
unlikely to provide a clear demonstration of that excess risk. This is because of the non-
experimental nature of data arising from epidemiologic studies and the small size of the
effect. Misclassification of exposure levels and/or of the presence of disease may distort
epidemiologic associations. Other unmeasured factors, including sampling variability,
may cause apparent associations in epidemiologic data. On the other hand, lack of an
observed association may be due to small study populations, insufficient follow up time,
or a low exposure to a carcinogenic agent. Ultimately any decision as to the meaning of
a body of epidemiologic data, or indeed any scientific data, is a matter of qualitative
judgment rather than of quantitative inference.
Over the past six years, a number of scientific panels, convened under the
auspices of various groups, have evaluated the human carcinogenicity of chlorophenoxy
herbicides in general and of 2,4-D in particular. Their findings are summarized below:
a) 1987 - International Agency for Research on Cancer (IARC)'. In Supple-
ment 7 of the IARC Monographs on the Evaluation of Carcinogenic Risks to
Humans, the IARC Working Group judged that there was limited evidence
that chlorophenoxy herbicides were carcinogenic to humans. There was
no clear delineation of data related to 2,4-D as opposed to other
chlorophenoxy herbicides, some of which contain dioxin (TCDD).
b) 1988 - Council on Scientific Affairs, American Medical Association: As a
general statement relating to all pesticides, the Council concludes: "At this
time, the results of these and other studies of agricultural workers (as well
as of Vietnam veterans who were allegedly exposed during military service
and pesticide production workers) offer only conjectural evidence at best
that pesticides may be carcinogenic." In a table, 2,4-D is listed as having
inadequate evidence relating to human carcinogenicity.
c) 1991 - Harvard School of Public Health Panel, Michel A. Ibrahim, Chair: (A
panel convened by the Center for Risk Analysis at the Harvard School of
Public Health under sponsorship of the National Association of Wheat
Growers Foundation. The panel concluded: "Although a cause-effect
relationship is far from being established, the epidemiological evidence for
13
-------
an association between exposure to 2,4-D and non-Hodgkin's lymphoma is
suggestive and requires further investigation. There is little evidence of an
association between use of 2,4-D and soft-tissue sarcoma or Hodgkin's
disease, and no evidence of an association between 2,4-D use and any
other form of cancer."
d) 1992 -1. C. Munro et al.: (A review relating to the safety of 2,4-D by a
panel of toxicologists and epidemiologists, supported by the Industry Task
Force II on 2,4-D Research Data. The authors conclude: "The case-control
epidemiological studies that have been the source of the cancer risk
hypothesis are inconclusive. Problems in assessing exposure based on
patient's memories make these studies difficult to interpret. Cohort studies
of exposed workers do not generally support the specific hypothesis that
2,4-D causes cancer. Taken together, the epidemiological studies provide,
at best, only weak evidence of an association between 2,4-D and the risk
of cancer."
Thus, the general conclusion reached by others as well as by the members
of this Panel is that there is, at most, weak evidence for an association between
exposure to 2,4-D and cancer in humans.
3.2 Canine Epidemiologic Study
3.2.1 Exposure to 2,4-D and Cancer in Dogs
To summarize the results of the Hayes et al., (1991) canine lymphoma epidemic-
logic study and its relevance to the evaluation of carcinogenicity of 2,4-D, it is useful first
to discuss the design of the study in generic terms and then to look at the specific details
of the Hayes et al. (1991) study.
The Hayes et al. (1991) study is one of a limited number of case-control studies
utilizing animal pathology records to identify subjects for investigation of environmental
risks (e.g. Glickman and Domanski 1986; Glickman etal., 1989; and Reif etal., 1992).
The study investigated the association between exposure to lawn chemicals and
14
-------
the incidence of lymphoma (as a model for human non-Hodgkin's lymphoma) in dogs
hospitalized at three veterinary hospitals. Two types of controls were selected from the
registry: dogs diagnosed with other malignancies and dogs in the veterinary hospital for
other reasons. Subjects were matched on age, year of hospital visit and hospital.
Exposure and confounder data were solicited by mail questionnaire. If no response was
received after a second mailing, a telephone interview was attempted. Nearly 20% of
the owners of the study subjects could not be located. Of the 1,436 owners who were
located, all but four completed either the questionnaire or interview. Information was
collected about the breed of dog, age of dog, lawn care practices and pesticide use of
the owner, and the dog's access to the pesticide treated yards. A statistically significant
association was found between case-control status and owner's use of 2,4-D and/or use
of a commercial lawn care service. When exposures were separated, the individual
odds ratios were similarly elevated although not statistically significant (an issue of
inadequate statistical power). The excess risk for both owner use of 2,4-D and use of a
commercial lawn care service was three times higher than either use alone. Significant
trends were found for the frequency of owner use (number of applications per year)
although not for the duration of owner use (number of years of application). No
significant trends were found for use of commercial lawn care service.
The investigators considered both confounding and bias as alternative explana-
tions for their findings. Adjustment for confounders did not affect the results. The
investigators were concerned that the pet owners' knowledge of their dogs disease
status might be associated with reported use of 2,4-D. The proportion of case owners
who reported no exposure to 2,4-D or who did not allow the dog access to the yard did
not differ by whether they know what disease the dog had. In addition, demographic and
socio-economic comparisons of the case and control groups did not reveal any striking
differences. Finally, because these dogs were seen in a veterinary hospital, upon death,
many underwent autopsy, limiting the opportunity for outcome misclassification.
At the Committee's meeting, Dr. Hayes clarified a number of issues not discussed
explicitly in the published study. Specific data (Hayes 1993) were presented to show
that there were not substantial differences in study results between tumor and non-tumor
controls, between telephone and mail respondents, and among the three participating
veterinary hospitals. Further, risks for exposure to other pesticides (e.g., malathion and
carbaryl) were not elevated. To evaluate the degree to which the reported exposure
data were representative of "real-world" experience, they were compared to a similar
15
-------
lawn care use survey conducted by EPA and found to be similar.
The strengths of the study are:
a) a very high response rate of those owners located;
b) no difference found between the responses of the two control groups;
c) no difference was found between the mode of response (phone or mail);
d) lawn care use data corresponded closely to a similar EPA survey;
e) a high rate of autopsy (diagnostic confirmation).
There are a number of limitations to the study relating to possible selection bias
and exposure misclassification:
a) there was a locational bias in the study, with fewer cases residing in rural
areas or within 2 miles of an industrial facility;
b) the questionnaire did not list specific pesticides but relied on owner recall
to provide the name of the products used;
c) reported pesticides were reclassified by the investigators into those
containing 2,4-D and those not containing 2,4-D;
d) questions regarding the frequency of use were about use of all lawn chemi-
cals as a whole rather than specific to each chemical identified by the
owner;
e) dogs roam and thus may experience exposures to 2,4-D not reported by
their owners;
f) there was no evaluation of possible viral etiology.
Another issue raised addressed the propriety of using dogs as a good model for
human carcinogenicity. It was pointed out that dogs are more acutely sensitive to 2,4-D,
and this is attributed to a different clearance mechanism which saturates at concen-
trations 10-100 times lower than for humans. However, data reported regarding
exposures likely experienced by these animals indicated that they were receiving
exposures at least 10 times lower than this clearance saturation level. Additionally, it
was noted the prevalence of malignant lymphoma was substantially higher in this data
16
-------
set than in other registries. However, this is due in part to the type of registry (veterinary
hospital, as opposed to population-based).
Overall, in spite of its limitations, this study was viewed by the Committee
as supporting an association between 2,4-D use and professional lawn care, and
malignant lymphoma in dogs. The study would be strengthened by providing
more precise exposure data and by demonstrating that canine lymphoma is an
appropriate model for human NHL.
A number of suggestions were made regarding improving the interpretability and
validity of this study. It was recommended strongly that the study be replicated on an
independent data set. When conducted, more attention should be placed on the details
of the exposure assessment and on the referral pattern by which the dogs ended up in
the registry.
Some additional suggestions were made for more detailed evaluation of data from
the Hayes etal. (1991) study and any replications. To evaluate possible referral bias,
the residence locations of cases and controls should be compared and evaluated with
respect to human demographics, socio-economic status, and case-control status.
3.3 Animal Testing and Carcinogenicity
The documentation submitted by the EPA and several published studies were
reviewed by the Committee. In 1987, the EPA Health Effects Division Carcinogenicity
Peer Review Panel (HED PRC) found that data on 2,4-D provided limited evidence for
Carcinogenicity in male rats and classified the compound as a Group "C" or "Possible
Human Carcinogen." The FIFRA Scientific Advisory Panel (SAP) concluded that this
evidence was "equivocal" and that"... 2,4-D should be classified as a Group "D," or "Not
Classifiable as to Human Carcinogenicity." The Agency agreed with the SAP that "...the
evidence for Carcinogenicity was not strong, and categorized 2,4-D as a Group D
Carcinogen pending receipt of the repeat rodent Carcinogenicity studies and additional
forthcoming epidemiological data." However, the animal testing database is still
incomplete six years later.
A subsequent review by a panel of 14 scientists was convened by the Harvard
School of Public Health in 1989 (Ibrahim etal., 1991). They essentially agreed with the
17
-------
FIFRA SAP review of the animal data and stated "Considered together, these two animal
studies do not provide impressive evidence that exposure to 2,4-D causes cancer in
animals. Based on the results from the rat study, the workshop participants concluded
that there was weak evidence supporting an excess of brain cancer occurrence in the
male Fischer 344 rats receiving the highest dose."
The Committee agrees that the above conclusions were warranted at the time
they were made, and since there is no additional animal data, we find no reason to
change the Harvard Panel's conclusions at this time. The Committee noted that
additional rodent bioassays are underway (approximately 12 months into a 2-year study)
in response to the criticism that the studies were not conducted at the Maximum
Tolerated Dose (MTD). The results of these studies should resolve the question as to
whether the finding of a marginal increase in brain neoplasms (astrocytomas) in male
rats in the original study was related to 2,4-D or not. If the neoplasms are truly treatment
related they should be reproduced in the ongoing study, particularly since it includes
higher doses than the original study. The Committee also noted that preliminary data on
the current rat subject studies suggest that the original studies may have been closer to
the MTD than originally thought, as evidenced by the 17 percent decrease in body
weight gain for male rats, and the 25 percent decrease for females
observed at 45 Mg/Kg dosage; this level is near the maximum dose of 75 Mg/Kg used in
the original study. This gives added weight to the original findings.
In summary the Committee felt that, at most, there was only equivocal
evidence of carcinogenic activity in animals and more studies are needed to
determine the compound's potential carcinogenicity.
3.4 Mutagenicity of 2,4-D
The industry supplied mutagenicity data (as summarized by EPA) demonstrating
that 2,4-D in various forms was uniformly negative in the Ames assay, mouse micronu-
cleus assay, and unscheduled DMA synthesis assays. The published data supplied by
the EPA for review did not provide any information to contradict the industry results.
A number of cytogenetic studies have been published with several reporting
negative results while others reported positive cytogenetic endpoints. Due to the
18
-------
variation in results, this data base deserves special evaluation. In 2 publications
(Linnainmaa, 1983; Mustonen etal. (1986), studies evaluating sister chromatid ex-
change or chromosome aberration were negative when blood samples were evaluated
from forest workers involved in 2,4-D spraying where exposure occurred under field
conditions. In each case either air concentrations and/or urine measurements were
made to determine exposure. Several studies evaluated cytogenetic effects in cultured
human lymphocytes with both negative (Mustonen etal., 1989) and positive results
(Mustonen etal., 1986; El Zoka and McKenzie; Turkula and Jalal, 1985; Korte and
Jalal,1982). The lack of a dose response decreases the confidence in the reported
positive results.
The form and purity of the 2,4-D is frequently not reported in the publications that
describe positive results in human peripheral lymphocytes exposed in culture. Evalua-
tion of rat peripheral lymphocytes exposed to 2,4-D in culture have produced negative
results for cytogenetic changes (Linnainmaa reference identified as chapter 23, source
unknown; Linnainmaa, 1984; Moustonen, 1989). Evaluation of cytogenetic effects in
CHO cells have been negative (Linnainmaa reference identified as chapter 23, source
unknown, Linnainmaa, 1984). Cytogenetic studies in bone marrow have been positive in
rats (Adhikari and Grover, 1988) and negative in mice (Schop etal., 1990). However,
the source and the purity of the test material is not identified in the positive study in rats
(Adihikari and Grover, 1988). Bone marrow cytogenetic studies in the Chinese hamster
have yielded negative results (Linnainmaa, Carcinogenesis , 1984). A recent study
described a positive result in the mouse hair follicle aberration assay (Schop etal.,
1990). Since this assay has not been widely utilized, the predictive value of a positive
result is uncertain.
The conflicting cytogenetic results do not provide evidence for genotoxicity of
2,4-D. Studies with positive results have significant experimental deficiencies as noted
above, thus limiting the value of these studies for assessing genotoxicity. Therefore,
although there are serious data deficiencies, the currently available evidence suggests
that 2,4-D is non-genotoxic. The lack of genotoxicity may reduce the concern for
potential carcinogenicity of 2,4-D, but it is recognized that not all carcinogens are
necessarily genotoxic.
3.5 Other Data and Data Gaps
19
-------
Human case control studies provide conflicting information. The findings of the
extant cohort studies may be interpreted as either positive or negative. The one
epidemiologic animal study (dogs) is consistent with some of the human studies, but
several reservations apply. Animal or in vivo mutagenesis and carcinogenesis studies
do not support an interpretation of carcinogenicity for 2,4,-D, but data gaps exist. The
chemical had apparent carcinogenicity only in rats, where the outcome was astrocytoma,
a cancer not reported in the human data. In addition, there is no proposed, plausible
biological mechanism that can explain the discrepant findings.
To provide an adequate evaluation of 2,4-D, we need the following research and
data:
a) rat, mouse, and dog toxicology studies requested by EPA
b) animal carcinogenicity studies that test 2,4-D with other substances that
might reflect the human exposure situation.
c) a replication of the dog epidemiology study, preferably done using newly
diagnosed cases, and with careful attention to exposure of 2,4-D and other
pesticides and solvents. We also need to have some additional analyses
of the collected data to see if a referral bias exists for the cases (eg.
referral from an area with a different exposure prevalence), as well as
biological and toxicological data which will tie the animal data to the human
situation.
d) additional case/control studies, with careful attention to exposures, particu-
larly multiple exposures. The single largest drawback to the existing
case/control studies is our equivocal and indirect measure of exposure.
Incidence (morbidity) studies should take precedence over mortality
studies and proxies should be avoided, if possible.
e) human studies, particularly cohort studies designed to assess both relative
risk of NHL and the comparative risk of all mortality (or all disease inci-
dence, if possible). This means that attention must be paid to attaining
nearly 100% follow-up and a comparison group must be selected for the
entire cohort, that is a realistic standard for the cohort. General population
20
-------
mortality is not a suitable comparison standard . A comparison cohort must
be found, or one constructed, that is a realistic predictor of the prevalence
of disease in the exposed cohort.
f) additional follow up and analysis of the human cohort studies of workers
involved in the production of 2,4-D. In addition, it would be useful to ask
the manufacturer to provide data for workers involved in 2,4-D production
from the update to Bond's (1988) study, to provide greater specificity for
this agent than was reported.
3.6 Weight of Evidence for 2,4-D Carcinogenicity
Epidemiologic studies of 2,4-D have included both case-control studies of NHL in
geographic areas where the numbers of farmers might be high and cohort studies of
manufacturers of the chemical as well as applicators and farming populations. The
case-control studies have focused on the association of the general class,
phenoxyherbicides, with NHL. Many studies did not specify the specific chemical so the
exposure to 2,4-D had to be inferred from the usual use in the area. In general the
studies indicated an approximately 20% to 30% increased odds ratio associated with
farming. If phenoxy acids were responsible for the observed excess of NHL in farmers,
the odds ratio for the association between the chemicals and NHL should be much
higher than the 1.2 to 1.3 seen for farmers because the etiologic exposure is now more
specific. This is not the case. In the NCI studies of NHL in Kansas and Nebraska (but
not in Iowa) the odds ratio increased with the number of days per year of exposure to
suggest a dose-response relationship. However, in these studies as well as some
others there was no sign of increasing risk with number of years of use. So, unlike many
other carcinogens, there is no indication of a cumulative dose effect on risk - only an
increasing risk with heavy exposure at some time during the lifetime. Thus the lack of an
increase in the risk ratio when we move from a non-specific exposure (as with farming to
a more specific exposure (as with 2,4-D) and the absence of a positive dose response
where cumulative exposure by years is used is not consistent a causal relationship
between the chemical and NHL.
The cohort studies in general have not suggested an increased risk of NHL for
individuals exposed to 2,4-D. However, many of these studies had a small exposed
population and did not have sufficient follow-up to be expected to show a risk even if it
21
-------
did exist. Therefore the negative results are relatively uninformative as to whether there
is an effect from this chemical. As with most epidemiologic studies the retrospective
assessment of exposure is suspect in all these studies. However, for most studies, the
subjects may well have had many chemical exposures which were not taken into account
in the analysis. Exposure to some of the chemicals are highly correlated, making
individual assessment of the exposure to a single chemical difficult or impossible. Thus,
while the epidemiologic studies suggest it is possible that 2,4-D may be carcinogenic in
humans, the evidence is not strong enough to support a causal relationship to the
specific phenoxyherbicide or any other farm exposure. However, there is suggestive
evidence of an association between exposure to 2,4-D and NHL in some of the studies,
and this observation requires further investigation. Future studies must try to establish
the exposure to 2,4-D and distinguish its effects from those of farming in general and
from other specific chemicals and pesticides used in the same environment.
When epidemiologic studies alone cannot establish a carcinogenic effect in
humans, it is often possible to combine that data with animal studies to establish with
high probability that the agent is carcinogenic. In the case of 2,4-D the chemical has
possibly produced astrocytomas, but only in rats. Since this cancer site is different from
that reported in man and, since the effect has not been seen in studies of 2,4-D in other
laboratory animals, these findings do not lend further support to the evidence from the
human studies . However, both rats and mice have shown changes in growth and
thyroxine levels from 2,4-D. Two cohort studies have reported non-significant increases
in thyroid, testicular and other endocrine cancers which deserves further study with an
increase in the follow-up of these groups. At this time, with the exception of the
suggestive association of lymphomas and exposure to 2,4-D in free-living dogs , the
animal data offer no support for the (conflicting) observations in humans, since the
carcinogenic effects are very weak, and are limited to a different cancer site and effects
which are demonstrable in only one species.
22
-------
4 CONCLUSIONS
The epidemiologic cohort studies which have tried to identify a hazard from
exposure to 2,4,D have generally shown no increased risk of cancer. However, all of the
populations which have identified specific exposure to the chemical have been small and
the follow-up period usually short. Thus, the number of deaths in the groups have been
too small to expect to demonstrate a risk of NHL. No other cancer site in humans has
been related to the specific exposure to 2,4,D. Sprayers in the international study had
an increase in STS, but the exposures were not specifically to 2,4,D.
Some case-control studies have shown a risk of NHL in association with the
occupation of farming but many did not indicate whether this relationship was due to a
specific exposure to 2,4,D. In the studies in which identification of specific exposures
was attempted, the risk did appear to be due to self-reported phenoxyherbicide exposure
in some studies but not in others. The risk did not seem to be much higher than the risk
from farming as a general work exposure and the risk increased primarily due to number
of days of use per year but not from duration of use. This is not what one would expect if
2,4,D were the agent causing the excess of NHL in farmers. Most of the association ap-
peared to be due to exposure at the highest number of days of use but the number in
this group was small. In the study where adjustment was made for other farm exposures,
the elevated risk associated with exposure to 2,4,D decreased to a non-significant OR of
1.1 after control for use of organophosphates. Thus the studies cannot distinguish
whether any observed risks reported in these studies are due to the use of 2,4,D or
some other aspect of farming as an occupation.
The single epidemiologic study of pet dogs suggested that these animals have a
risk from owner-reported exposure to 2,4,D or use of a lawn care service. This study
has the same constraints of selection bias of cases and controls and recall and
misclassification biases that could occur in human case-control studies. The applicabili-
ty of the canine epidemiology data to lymphoma in humans is questionable since the
cancers may not be similar in dogs and humans, and exposures to 2,4-D are not clearly
established. To substantiate these results, the study should be replicated with improve-
ment in the exposure measures and verification of the comparability of the cancers in
dogs and humans.
The toxicologic data have shown that rats may develop astrocytomas from
23
-------
exposure to 2,4,D but other animal species have not shown this cancer, nor has it been
reported in the human studies. The animals clearly have changes in growth patterns
from heavy exposure to the chemical. An on-going rat study at higher doses will clarify
whether this finding is treatment-related or not.
The chemical has not shown mutagenic changes under experimental situations.
Although the Committee recognizes that it is not necessary for a carcinogen to also be a
mutagen, it would have strengthened the observations in humans if there were
laboratory data, especially in toxicologic studies of animals, to support any observations
in humans.
Therefore, our conclusion at this time is that while there is some evidence
that NHL may occur in excess in populations which are likely to be exposed to
2,4,D, the data are not sufficient to conclude that there is a cause and effect
relationship between the exposure to 2,4,D and NHL. The data are, however,
sufficient to require continued examination of the issue through further studies.
24
-------
5 REFERENCES
Adhikari, N. and Grover, I.S. 1988. Genotoxic effects of some systemic pesticides: in
vivo chromosomal aberrations in bone marrow cells in rats. Environ. Mol.
Mutagen. 12(2):235-242.
Bond, G.G., Witterstroem, N.H., Roush, G.L., McLarem, E.A., Lipps, I.E., and Cook, RR.
1988. Cause specific mortality among employees engaged in the manufacture,
formulation or packaging of 2,4-dichlorophenoxyacetic acid and related salts. Br.
J. Ind.Med. 45:98-105.
Carlo, G.L., Cole. P., Miller, A.B., Munro, I.C., Solomon, K.R. and Squire, R.A. 1992.
Review of a study reporting an association between 2,4-dichlorophenoxyacetic
acid and canine malignant lymphoma: Report of an expert panel. Reg Tox and
Pharm 16: 245-253.
El-Zoka, T.A. and McKenzie, W.H. 1980. Cytological effects of the herbicides 2,4-D and
2,4,5-T on human peripheral lymphocytes exposed in vitro. Am. J. Human
Genetics 32(4): 154A. (abstract)
EPA. Task Force Representing 2,4-D Pesticide Registrants Agrees to Exposure-
Reduction Label Changes and User Education Program As Alternative to EPA
Suspension for Overdue Data. Friday, October 9, 1992 Press Advisory
Glickman, L.T. and Domanski, L.M. 1986. An alternative to laboratory animal experimen-
tation for human health risk assessment: Epidemiological studies of pet animals.
ATLA 13: 267-285.
Glickman L.T., Schofer, F.S., McKee, L.J., Reif, J.S. and Goldschmidt, M.H.. 1989.
Epidemiologic study of insecticide exposures, obesity, and risk of bladder cancer
in household dogs. J Tox and Environ Health 28: 407-414.
Hayes, H.M., Tarone, R.E., Cantor, K.P, Jessen, C.R., McCurnin, D.M. and Richardson,
R.C. 1991. Case-control study of canine malignant lymphoma: Positive associa-
tion with dog owner's use of 2,4-dichlorophenoxyacetic acid herbicides. J Natl
Cancer Inst 83: 1226-1231.
Hayes, H.M. 1993. Presentation to EPA SAB/SAP Special Meeting to Evaluate the
Carcinogenicity of 2,4-D. April 1, 1993, Alexandria, VA.
IARC, 1990. Phenoxy acid herbicides and contaminants: description of the IARC
International Register of Workers. Am. J. Ind. Med. 18:39-45.
25
-------
Korte, C. and Jalal, S.M. 1982. 2,4-D induced clastogenicity and elevated rates of sister
chromatid exchange in cultured human lymphocytes. J. Heredity 73:994-226
Lilienfeld, D.E. and Gallo, M.A, 1989. 2,4-D, 2,5-T, and 2,3,7,8-TCDD: an overview.
Epidemiol. Rev. 11:28-58.
Linnainmaa, K. 1983. Nonmutagenicity of phenoxy acid herbicides 2,4-
dichlorophenoxyacetic acid and 4-methyl-2-chlorophenoxyacetic acid. In:
Choudhary et al. (eds.). Chlorinated Dioxins and Dibenzofurans in the Total
Environment, pp. 385. Bufferworth Publishers, Woburn, MA.
Linnainmaa, K. 1984. Induction of sister chromatid exchanges by peroxisome
proliferators 2,4-D, MCPA, and clofibrate in vivo and in vitro. Carcinogenesis
5(6):703707.
Mustonen, R., Elovaara, E., Zifting, A., Linnaiamaa, K., and Vainio, H. 1989. Effects of
commercial chlorophenate, 2,3,7,8-T4CDD, and pure phenoxyacetic acids on
hepatic peroxisome proliferation, xenobiotic metabolism and sister chromatid
exchange in the rat. Arch. Toxicol. 63:203-208.
Mustonen, R., Kangras, J., Vuojolahti, P., and Linnainmaa, K. 1986. Effect of
phenoxyacetic acids on the induction of chromosome aberrations in vitro and in
vivo. Mutagenesis 1:241 -255.
Persson, B., Dahlander, A-M., Frediickson, M., Brage, H.N., Ohlson, C-G., and Axelson,
0. 1989. Malignant lymphomas and occupation exposures. Br. J. Ind. Med.
45:516-520.
Reif, J.S., Dunn, K., Ogilvie, G.K. and Harris, C.K. 1992. Passive smoking and canine
lung cancer risk. Amer J Epidemiol 135: 234-239.
Schop, R.N., Hardy, M.H., and Goldberg, M.T. 1990. Comparison of the activity of
topically applied pesticides and the herbicide 2,4-D in two short-term in vivo
assays of genotoxicity in the mouse. Fundamental and Applied Toxicology
15:666-675.
Turkula, I.E. and Jalal, S.M. 1985 Increased rates of sister chromatid exchanges
induced by the herbicide 2,4-D. J. Heredity 76:213-214.
Wiklund, K., Dich, J., and Holm, L-E. 1989. Risk of soft tissue sarcoma, Hodgkin's
disease, and non-Hodgkin's lymphoma among Swedish licensed pesticide
applicators. Chemosphere 18(1-6):395-400.
26
------- |