xvEPA
United States
Environmental Protection
Agency
Office of Pesticide Programs
Washington, DC 20460
EPA-540/9-85-018
June 1985
Hazard Evaluation Division
Standard Evaluation Procedure
Teratology Studies
-------
EPA-540/9-85-018
June 1985
HAZARD EVALUATION DIVISION
STANDARD EVALUATION PROCEDURE
TERATOLOGY STUDIES
Prepared by
Laurence D. Chitlik, D.A.B.T.
Quang Q. Bui, Ph.D.
Gary J. Burin, M.P.H.
Stephen C. Dapson, Ph.D.
Standard Evaluation Procedures Project Manager
Stephen L. Johnson
Hazard Evaluation Division
Office of Pesticide Programs
United States Environmental Protection Agency
Office of Pesticide Programs
Washington, D.C. 20460
-------
STANDARD EVALUATION PROCEDURE
PREAMBLE
This Standard Evaluation Procedure (SEP) is one of a set of
guidance documents which explain the procedures used to evaluate
environmental and human health effects data submitted to the
Office of Pesticide Programs. The SEPs are designed to ensure
comprehensive and consistent treatment of major scientific topics
in these reviews and to provide interpretive policy guidance
where appropriate. The Standard Evaluation Procedures will be
used in conjunction with the appropriate Pesticide Assessment
Guidelines and other Agency Guidelines. While the documents were
developed to explain specifically the principles of scientific
evaluation within the Office of Pesticide Programs, they may also
be used by other offices in the Agency in the evaluation of
studies and scientific data. The Standard Evaluation Procedures
will also serve as valuable internal reference documents and will
inform the public and regulated community of important consider-
ations in the evaluation of test data for determining chemical
hazards. I believe the SEPs will improve both the quality of
science within EPA and, in conjunction with the Pesticide Assess-
ment Guidelines, will lead to more effective use of both public
and private resources.
tfohn W. Melone, Director
Hazard Evaluation Division
-------
TABLE OF CONTENTS
Page
PREFACE
I. INTRODUCTION
A. Definition 1
B. When Required 3
C. Usefulness of the Teratology Study in
a Regulatory Setting 3
II. DATA ACCEPTABILITY
A. Range Finding Study 4
B. Primary Study 5
III. EVALUATION AND INTERPRETATION
A. Evaluation of Study Conduct
1. Choice of Test Compound 6
2. Study Protocol 6
3. Choice of vehicle 6
4. Choice of Species 7
5. Route of Exposure 7
6. Reporting of Maternal and Fetal Findings ... 7
7. Concurrence of Dosing of Test Groups 7
8. Final Reports 8
B. Study Interpretation
Manifestations of Maternal Toxicity
1. Suitability of Test Animals 8
2. Dose-Levels 9
3. Mortality and Clinical Observations 10
4. Maternal Body Weight 10
5. Organ Weights and Enzyme Markers 12
6. Conception Rate 13
7. Corpora Lutea 13
8. Preimplantation Loss 14
Manifestations of Developmental Toxicity
1. Death 15
2. Altered Growth 16
3. Anomalies 17
4. Review End Points to be Considered 20
5. Functional Defects 23
IV. DEVELOPMENTAL TOXICITY RISK ASSESSMENT 24
V. BIBLIOGRAPHY 28
-------
PREFACE
The purpose of this standard Evaluation Procedure (SEP) for
teratology studies is not to provide a complete discussion of the
field of teratology but to act as a guide to the Regulatory
Toxicologist for practical application of our current knowledge
of this area. This SEP attempts to standardize, within the Office
of Pesticide Programs, some of the common terminology employed in
the evaluation of teratology studies and to provide useful infor-
mation for risk assessment. For example, the term developmental
toxicity is introduced in this SEP to more accurately describe
the variety of end points that are of concern in a teratology
study. It should also be noted that this document is consistent
with the Agency's "Interim Guidelines for the Health Assessment
of Suspect Developmental Toxicants."
The study of developmental toxicity of agents is rapidly
developing and changing. Refinements in methodologies and
techniques are almost a daily event and many issues and definitions
basic to this area are still under debate. Further, no universal
dictionary of terata is available and definitions of malformations
and variations vary dramatically from laboratory to laboratory.
Even new areas such as "behavioral teratology" are being rapidly
re-defined as more inclusive fields such as "functional develop-
mental toxicology."
The SEP points out some of the fundamental controversies
in this area in order to caution the toxicologist. For example,
the study reviewer should understand that findings in a study may
either be transient or permanent. Where possible, methods of
dealing with these types of issues are offered in the SEP.
The toxicologist is also informed as to why the classical studies,
as required by the Guidelines, sometimes fall short of resolving
such issues and basic guidance is provided to the reviewer. Further-
more, it should be stressed that the toxicologist must maintain an
understanding of the developmental toxicity issues faced by this
program and, when necessary, follow up on these issues by requiring
the most appropriate ancillary studies, e.g., post-natal studies.
Efforts have been made in this document to organize the
subject into different sections: data acceptability, data evaluation
and interpretation, and risk assessment.
The Data Acceptability section deals with specific questions
regarding the adequacy of the experimental design, and the utility
and the completeness of the data reported.
-------
The two main end points of a classical teratology study,
maternal toxicity and developmental toxicity (embryo/fetotoxicity)
which includes teratogenicity, are discussed in the section on
Data Evaluation and Interpretation along with the utility of
developmental hazard indices. Factors influencing the outcome of
toxicological manifestations such as diseases and methods of
impregnation are addressed. Experimental animal data and attempts
to differentiate malformations and variations are also discussed
in this section. Effects of environmental agents on post-natal
development are covered in the section on Functional Defects.
Another area discussed is risk assessment as it relates to
developmental toxicants. The Office of Pesticide Programs has used
the "Margin of Safety Approach" for a number of years. This approach
considers the NOEL for developmental toxicity with the estimated
exposure as a ratio. The only difference between this approach
and the Safety Factor Approach is the a. priori consideration of
exposure data. Clearly, however, the area of developmental
toxicity risk assessment needs considerable work and changes are
to be anticipated in the near future.
This SEP should be considered as only a beginning and will
require periodic updates.
-------
TERATOLOGY STUDIES
I. INTRODUCTION
A. Definition
There are a variety of definitions for teratology or terato-
genicity which may be found in the literature. One definition may
be found under § 83-3 of the 1982 Subdivision F Guidelines:
"Teratogenicity is the property of a chemical that causes permanent
structural or functional abnormalities during the period of embry-
onic development." Teratology has also been defined as the study
of abnormal development and congenital malformations. No matter
what the definition, it should be understood that malformations
are only one manifestation of developmental toxicity. A wide
variety of end points are of interest in teratology studies. In
addition to malformations and other structural alterations,
embryo-fetal death, and other signs of developmental toxicity, are
important indicators of an effect on the conceptus that should be
assessed.
The period during which exposure takes place in teratology
studies is usually the period of major organogenesis, but in some
studies it may be extended from implantation of the conceptus up
to parturition (Subdivision F Guidelines). The manifestations of
developmental toxicity may be observed at any time during the
life span of the test animal.
There is a lack of uniformity in the use of key terms in the
field of developmental toxicology. For our purposes, the following
definitions will be used:
Developmental Toxicology; The study of adverse effects on the
developing organism which may result from exposure prior to
conception (of either or both parents), during prenatal development
(as in a classical Guidelines teratology study), or postnatally to
the time of sexual maturation. Adverse developmental effects may
be detected at any point in the life span of the organism. The
major manifestations of developmental toxicity include: 1) death
of the developing organism, 2) structural abnormality (malformations
and variations), 3) altered growth, and 4) functional deficiency.
Embryotoxicity and fetotoxicity; The subset of developmental
toxicity referring to adverse effects on the developing conceptus
prior to parturition. The distinguishing feature between the
two terms is the stage of development during which the injury
occurred. Effects which occurred during the embryonic period
are referred to as embryotoxicity and events during the fetal
period are referred to as fetotoxicity. Oftentimes, embryotoxicity
and fetotoxicity are difficult to distinguish and in those cases
-------
the term embryo/fetotoxicity is appropriate. This subset of
developmental toxicity includes: 1) in utero death, 2) structural
abnormalities (malformations and variations), and 3) altered growth.
Teratogenicity; As defined in the 1982 Subdivision F Pesticide
Assessment Guidelines § 83-3, "Teratogenicity is the property of
a chemical that causes permanent structural or functional abnor-
malities during the period of embryonic development". In the
Interim Guidelines for the Health Assessment of Suspect Develop-
mental Toxicants, the term teratogenicity is only used to describe
a "permanent structural change which may adversely affect survival,
development or function." in reality, both definitions are consis-
tent since all functional abnormalities must have a structural
basis, even if only at the molecular level.
Altered Growth; Alterations (other than malformations) in the
soft tissues, the skeletal system or body weight or size of the
developing offspring. Altered growth may be induced at any stage
of development and may be permanent or reversible.
Variations and Malformations: A variation is a divergence beyond
the usual range of structural constitution but which may not adversely
affect survival, development, or function. A malformation is a perma-
nent structural change which may adversely affect survival, develop-
ment, or function. Because development is considered as a continuum
of events, there is no universally accepted classification for many
structural alterations as either variations or malformations (See
Section III.B of this document for examples of both malformations
and variations).
Functional Developmental Toxicology; The study of the causes,
mechanisms, and manifestations of alterations or delays in functional
competence of the organism or organ system following exposure to an
agent during critical periods of development pre- and/or post-
natally.
Risk Assessment: The qualitative and quantitative determination
of potential risk to humans.
The guidance provided in this document is intended to be used
in conjunction with the Core Classification System for determining
the acceptability of a teratology study for regulatory purposes.
In a regulatory setting, teratology studies are required for
the purpose of attempting to determine the potential of an agent
to induce developmental toxicity. The test data must be approached
cautiously in attempting to extrapolate to humans due to fundamental
differences in conditions of exposure and biological differences
between the test species and humans. However, for obvious ethical
reasons, regulatory agencies have little choice but to use test
results from experimental animals treated under laboratory conditions
to predict the risk of human developmental toxicity.
-------
-3-
B. When Required
The Office of Pesticide Programs requires teratology testing
in two species to support the registration of a product intended
for food use (i.e., when tolerances or exemptions from tolerances
are considered) and for nonfood uses if significant exposure of
human females of child bearing age may reasonably be expected.
Temporary tolerances generally require testing in only one species.
The reviewer must be aware that the above requirements are
minimum data needs and that additional testing or modifications
of routine protocols (such as the inclusion of a post-natal phase
which may include functional assessment as well as information on
the reversibility of altered growth) will sometimes be necessary
for a more meaningful assessment of potential developmental toxicity
Structural similarity to developmental toxicants including known
teratogens may trigger testing when it otherwise would not be
required, but the reviewer should also be aware that small differ-
ences in molecular structure may lead to major differences in the
potential for developmental toxicity and that structure-activity
relationships in developmental toxicity testing are not well esta-
blished. Nevertheless, the reviewer should consult published
reference materials (e.g., Shepard, 1983; Schardein, 1985) and
the Toxicology Branch files for structurally related compounds
prior to determining data requirements for teratologic assessment
if there is any doubt as to the need for testing.
The reviewer should also be aware that it is sometimes
possible to waive the need for further teratogenicity testing if
the "limit test" requirements in Subdivision F are met or exceeded.
This section states that "if a dose level of 1000 mg/kg produces
no evidence of embryo toxicity or teratogenicity, studies at other
dose levels may not be considered. If a preliminary study at this
high dose level, with definite evidence of maternal toxicity, shows
no adverse effect on the embryos, studies at other dose levels may
not be necessary."
C. Usefulness of the Teratology Study in a Regulatory Setting
Information gathered from teratology studies may be the only
information that is available regarding the effects of in utero
exposure or it may be supplemented by the results of a reproduction
study. In either case, the teratology study yields important and
often unique information regarding the effects of a compound on
the conceptus. Furthermore, although it is the dam that is being
dosed, the reviewer of a teratology study must focus on both the
litter as well as the fetus. Maternal toxicity data is useful to
confirm that appropriate dose-levels have been used to allow the
greatest potential for eliciting an effect.
-------
-4-
As noted previously, caution should be exercised in extrapol-
ating animal developmental toxicology results to humans. However,
these studies are the best indicators available in the prediction
of potential adverse developmental effects in humans. Positive
findings in one or more animal species triggers a regulatory response
which consists of assessment of exposure (worker, dietary, or other)
and a comparison (ratio) of these exposure levels with the No Obser-
ved Effect Level (NOEL) in the most sensitive species (Margin of
Safety). The conduct of a developmental toxicity risk assessment
is described in detail in Section IV of this document.
II. DATA ACCEPTABILITY
A. Range Finding Study
The purpose of the range finding study is solely to establish
the proper selection of dose-levels for the primary teratology
study. The information required in the range finding study should
be adequate to assess maternal toxicity and should be consistent
with that required in Section 80-4 of the Subdivision F Guidelines.
Maternal weight gain during dosing and over the course of gestation
is of great importance and data should be available for each animal
on test. Gross necropsy results for each dam, detailed clinical
observations, uterine and ovarian findings (such as the incidence
of corpora lutea, implantations and resorptions), and food consump-
tion data (if available) are necessary to fully assess maternal
toxicity.
Developmental toxicity is not fully assessed in the range
finding study and information relating to embryo/fetoxicity is
usually limited to a rudimentary assessment of prenatal mortality
and external fetal findings. Also, range finding studies often
lack soft tissue and skeletal examinations of fetuses and utilize
small numbers of litters and fetuses. Therefore, they are of
limited utility for assessment of developmental toxicity, including
teratogenicity.
The acute oral toxicity study in the same strain of animal may
also provide useful information for the setting of dose levels.
However, it should be recognized that the acute studies are the
result of only a single dosing and that the sensitivity of a dam
during pregnancy may be different than that of a nonpregnant
animal. Other studies such as subchronic and metabolism studies
may be useful in judging the appropriateness of dose-levels but
rarely are adequate for this purpose-.
In reviewing the range finding teratology study, the reviewer
should pay particular attention to potential manifestations of
maternal toxicity. Maternal deaths and increased incidence of
abortions will often be observed at the higher dose-levels.
-------
-5-
Although these are forms of maternal toxicity, dose-levels associated
with a large number of abortions or maternal deaths are usually not
optimal for the high dose in the primary teratology study (see
Section IIIB). However, in some cases the margin between the dose-
level inducing lethality or abortion and that associated with other
forms of maternal toxicity such as decreased body weight gain (or
absolute body weight loss), tremors or other manifestations of com-
pound toxicity, may be small. in those cases, the deaths or abortions
may be seen at the lowest levels which also induce other forms of
maternal toxicity. Optimally, the spacing of the dose-levels should
be adequate to select a high dose-level for the primary study which
does not result in more than 10% maternal mortality.
A range finding study cannot be classified as more than Core
Supplementary Data since it is not adequate for the assessment of
developmental toxicity.
B. Primary Study
Information required to be in the primary study is clearly
stated in "Subdivision F: Hazard Identification: Humans and Domestic
Animals" Sections 80-4 and 83-3.
It is particularly important to be able to associate all
maternal and fetal findings with individual animals. For example,
it is necessary to be able to relate soft tissue, skeletal, and
external findings to any individual fetus and to determine the dam
from which the fetus was removed as well as the maternal toxicity
data specific for that dam.
A frequent shortcoming in the reporting of a teratology study
is the lack of detail regarding the visceral and skeletal examination
techniques and findings. If there is any question regarding the
adequacy of these examinations, the registrant should be requested
to furnish detailed protocols. In many cases the laboratory which
conducted the study can furnish a Standard Operating Procedure
which describes in detail the method by which technicians are to
conduct such examinations. It is also often necessary to request
the registrant to furnish photographs of malformations, detailed
definitions, and descriptions of findings and the grading of severity
of findings. The need for this information as well as historical
control data must be considered on a case-by-case basis.
Historical control data must be from the lab which has con-
ducted the test and should be reported on both a study-by-study
basis as well as collectively, with findings reported in terms of
the number of litters with a finding per number of litters examined
and the number of fetuses with a finding per number of fetuses
examined. The historical data base should preferably span two
years prior to and (when available) after the submitted study.
-------
-6-
All historical studies should be dated, and if a vehicle control
was used, that vehicle should be identified. Historical maternal
body weight gain data is also useful.
III. EVALUATION AND INTERPRETATION
A. Evaluation of Study Conduct
The Subdivision F Guidelines identify reporting requirements
for teratology tests. It is important for the reviewer to pay
particular attention to the following:
1. Choice of Test Compound
The test compound should be that intended for commercial use
which in most cases is the technical material. Testing of formu-
lations is usually not required. Both the purity of the test compound
and the identification of impurities are important information and
must be requested if these data cannot be ascertained from the study
report or the compound registration. The reviewer should be aware
that exposure to impurities in the test material may be an important
aspect of the potential for developmental toxicity of a compound
(e.g., dioxins).
2. Study Protocol
The study protocol should be compared to that presented in
Section 83-3 of the Subdivision F Guidelines as well as to the Core
Minimum standards. Deviations from basic protocols are sometimes
necessary due to the need for additional information. For example,
a post-natal phase may be necessary to distinguish the reversibility
of dilated renal pelvis (apparent hydronephrosis; Woo and Hoar, 1972)
and true hydronephrosis. Histopathology or dual skeletal staining
may sometimes be necessary to follow up on unusual findings. The
method of visceral and skeletal examination is left up to the dis-
cretion of the testing laboratory provided that a thorough exami-
nation and a sufficient number of animals are used. For rats,
mice, and hamsters, approximately equal numbers should be examined
for both skeletal and visceral findings (Wilson, 1965). It is
possible to examine all fetuses for both visceral and skeletal
findings, but this is not required for rats, mice, and hamsters.
Rabbit fetuses, being larger, are more easily examined both viscerally
and skeletally and both examinations are in fact required for each
rabbit fetus (Staples and Wilson techniques or modified techniques).
In any case, the methods of visceral and skeletal examination should
be clearly stated in the study report.
3 . Choice of Vehicle
The vehicle should be appropriate for the delivery of the test
compound, should not interfere with absorption, and should not
-------
-7-
induce maternal or developmental toxicity- If there is question
as to the toxicity of the vehicle, the registrant should be
required to justify the choice of vehicle. A sham-control may
sometimes be required if the potential toxicity of the vehicle is
not completely understood.
4. Choice of Species
Generally, rats, mice, hamsters, or rabbits are used for
testing. Studies conducted using other species may be acceptable
but justification for the deviation should be requested from the
registrant.
5. Route of Exposure
Dosing by gavage is the preferred route of administration in
teratology studies because this route is considered to be closely
related to human dietary exposure. However, studies conducted by
other routes of administration, such as dermal and inhalation,
are also acceptable to the Office of Pesticide Programs on a
case-by-case basis if these routes are the primary situations for
human exposure. Dietary administration is not preferred but may
be considered acceptable if maternal toxicity is adequately
demonstrated.
6. Reporting of Maternal and Fetal Findings
Reporting requirements have been stated in the Subdivision F
Guidelines and will not be discussed in detail here. Clearly,
however, it must be possible to associate all reported maternal
and fetal findings with individual animals. It also must be
possible to identify which dam showed any given clinical signs on
any given day and to identify fetuses with each variation or
malformation.
All reported mean data should be carefully compared to
submitted individual data for possible inconsistency. The statistical
methods used should be referenced and/or described. Statistical
evaluations of all data including skeletal variation data should
be available in the study report. The appropriate application of
statistical methods should be verified.
7. Concurrence of Dosing of Test Groups
Dose groups should be run concurrently. The staggering of
induction of pregnancy within dose groups is acceptable but the
mean time of induction of pregnancy must not differ significantly
from one dose group to another.
-------
8. Final Reports
Study reports should be signed and dated and submitted as a
final report. This, of course, is not necessary for reports pub-
lished in the open literature. If a study report is not signed
and dated, it must be assumed that it is subject to change and does
not represent the final position of the investigator. it should be
noted in the review of such a study that it is a draft and these
studies should not be accepted as fully meeting regulatory require-
ments and classified as Core-Supplementary Data.
B. Study Interpretation
The following are the two major areas of a teratogenicity study
that require careful assessment:
Maternal toxicity
Developmental toxicity
MANIFESTATIONS OF MATERNAL TOXICITY
The Subdivision F Guidelines specify that maternal toxicity
must be observed in teratology studies. The actual wording reads:
Unless limited by the physical/chemical nature or
biological properties of the substance, the highest
dosage level should include some overt maternal
toxicity such as slight weight loss, but not more
than 10% maternal deaths.
Compared to subchronic or chronic toxicity studies, teratology
studies determine maternal toxicity based on a very limited number
of parameters. These routinely include body weight, food consumption,
clinical signs of toxicity, necropsy data, and reproductive data.
Interpretation of maternal toxicity is further hindered by the short
duration of exposure (during the period of major organogenesis) and
the relatively small number of animals on test at each dose-level
(especially in the case of rabbits).
1. Suitability of Test Animals
Among the mammals, rodents (rat and mouse) and rabbits have been
used most frequently as animal models (Wilson and Eraser, 1978).
Animals selected should be disease-free and of the same age and
parity in order to eliminate variability in the results.
Basic husbandry information for commonly used animals is given
as follows:
-------
— Q —
Rat Mouse Rabbit
Gestation (days) 21+/- 1 20+/- 1 31+/- 1
Mating system pair or pair or pair or
colonyt colonyt artificial*
Source; Hafez, E.S.E., Reproduction and Breeding
Techniques for Laboratory Animals
eds, Lea & Febiger, Philadelphia, 1970.
t 1 male to several females
* Discussed in "Reproductive Status Section"
Pasteurellosis and coccidiosis are common diseases encountered
in rabbits which may be manifested as: nasal discharge, diarrhea,
and congested lungs, as well as pitted kidneys and brain lesions
observed at necropsy (Benirschke et al., 1978). A high incidence
of such observations is suggestive of a questionable health status
of the test animals.
In some studies, especially with rabbits, there may be treat-
ment with antimicrobial agents prior to use in teratology studies.
Clearly, it is more desirable to use disease-free animals since the
impact of drug treatment and disease on development are clearly
unwanted variables. However, these studies may be considered accept-
able on a case-by-case basis, depending upon the adequacy of the
quarantine period prior to treatment and other factors.
Nulliparous animals should be selected for testing since confir-
mation of pregnancy in parous dams cannot be accurately determined.
Furthermore, if corpora lutea are counted for the determination of
preimplantation loss, nulliparous animals are preferred (WHO, Techni-
cal Report Series No. 364, 1967; See section on corpora lutea).
2. Dose-Levels
As stated in the Guidelines, at least three treatment levels
and a control group (usually vehicle control) should be used. If
there is any question as to toxicity of the vehicle, a sham or
untreated control is also required. Although not required by the
Guidelines, a concurrent positive control group may also be included.
Dosing should be performed at the same time each day and reported.
Dosing at levels that produce excessive maternal toxicity may
result in unreasonable numbers of deaths and abortions. Results
such as these may limit the utility of the study and lead to study
rejection. On the other hand, a study might also be rejected if
the highest dose used is too low, i.e., without significant toxicity,
Although the Subdivision F Guidelines state that "the lowest
dose level should not produce any evidence of toxicity," if maternal
toxicity is the only finding at this dose-level, the study may
still be considered acceptable.
-------
-10-
3. Mortality and Clinical Observations
Understanding of the clinical signs of toxicity characteristic
of the test compound should be gained from the range finding study
and other toxicity studies prior to evaluation of the primary
study.
Maternal death and/or abortion may be due to many factors,
such as the test compound, diseases, environmental factors, and
technical errors. Environmental factors, such as variations in
housing conditions (temperature, humidity, light-cycle, caging,
etc.), are known to influence the welfare of test animals. Also,
technical errors such as intubation error and mishandling of
animals in general can alter the outcome of test results and lead
to maternal death and/or excessive stress.
The cause of death should be clarified from the necropsy data
where possible. High incidences of congested lungs, reddening of
the tracheal lining, and fluid accumulation in the lungs are suggest-
ive of either technical error (e.g., gavage error) or disease. Conse-
quently, care must be taken in the assessment of maternal death as
an index of maternal toxicity since the two are often unrelated. The
reviewer should always consider other possible reasons for maternal
deaths (especially in rabbit studies). A dose-response relationship
should be evident for one to conclude that effects are compound
related to administration of the test compound.
4. Maternal Body Weight
Body weight and body weight gain data can be sensitive
indicators of toxicity and are often used as a basis for the
determination of the NOEL for maternal toxicity.
Test animals must be randomized for the body weight data to
be useful as potential indicators of toxicity. This allows all dose
groups to start with similar maternal weight means and variance.
Body weight measurements should be available at least on day
0 or 1 of gestation, at the start of dosing, on the final day of
dosing, and on the day of sacrifice (usually one day prior to
expected term).
Theoretically, the body weight gain (or percent change in
body weight) of the treated groups should be comparable to that
of the controls during the period prior to treatment. During the
treatment period, a body weight reduction may be due to toxicity
of the test compound and/or anorexia.
Anorexia is a common indicator of maternal toxicity and may
occur soon after the initial dosing or may require repeated
dosing before becoming evident. The effects of maternal anorexia
may be assessed by calculating the food efficiency index which is
-------
-11-
a measure of the efficacy of food utilization (grams of food con-
sumed per kilograms of body weight gained).
The following table presents a typical pattern of compound
related anorexia in a rabbit teratology study:
Assessment of a Typical Pattern of Food Intake During Gestation
Food Intake (g/dam/day)
Dose level Number of Pregnant Rabbits Gestation Days
mg/kg/day Available Per Interval* 0-5 6-18 19-27
0 14,13,10 106 76 53
250 13,13,10 115 56 79
500 15,14,4 102 23t 48
750 13,13,5 124 12t 26t
* Excluding spillers.
t Significantly different from control at P < 0.05.
Mean food intake for days 0-5 varied between 102 and 124 grams
per day prior to administration of test compound. After test com-
pound administration (days 6-18), a dose-related decrease in food
consumption was observed in treated animals. The apparent decrease
at the low dose-level was slight and within the normal range of
variability. The mid and high dose-levels, however, were clearly
less than the control values. After cessation of dosing (days 19-27),
increased food consumption was observed in all treated groups as com-
pared to the dosing period. This compensatory increase is common
after a depression of food consumption during dosing. The overall
pattern in food consumption indicates that compound-related effects
were observed at the mid and high dose-levels during the period of
dosing. The extent of the decrease clearly indicates maternal toxi-
city at these dose-levels. The low dose-level is comparable to the
control group and, despite an apparent slight decrease, may be
regarded as the No Observed Effect Level for maternal toxicity if
there are no other indications of toxicity in the study (Burin, 1982).
If the food efficiency index is similar between treated and con-
trol groups, then anorexia may not be the main factor. Food consump-
tion data are not presently required by the 1982 EPA Guidelines
except for a dosed-feeding study. However, these data are considered
useful for assessment of maternal effects regardless of route of
administration.
-------
-12-
Consequently, relevant interpretation of maternal toxicity may
be obtained from the body weight gain data determined at different
periods throughout gestation. Some of those periods are listed
below:
a. Body weight gain throughout gestation
b. Body weight gain prior to dosing
c. Body weight gain during dosing
d. Body weight gain after dosing to study termination
Calculation of the "corrected" mean maternal weight gain (as
measured by the difference in mean initial and terminal maternal
body weight less the gravid uterus weight) may also serve as an
index of maternal toxicity. This approach is considered more
reliable by many investigators to assess maternal weight gain/loss,
since it removes the uterine weight variability. An alternate
but less desirable estimate of maternal weight change during
gestation can be obtained by subtracting the litter weight from
the maternal weight gain.
Assessment of maternal weight gain data in rabbits is difficult
since erratic body weight gain is commonly found in this species.
Nevertheless, comparison to concurrent control or historical control
data may still provide the reviewer with valuable data for assessment,
Although reductions in maternal weight gain are generally dose-
related, there are rare instances in which the low dose is affected
while the high dose is not (Kavlock et al., 1984). This may be
true when high maternal mortality is observed at high dose levels,
eliminating sensitive animals, and consequently reducing the number
of animals for comparison. In this case, the lack of a dose-response
does not imply that there was no compound-related effect.
5. Organ Weights and Enzyme Markers
At the present time, neither organ weight data nor enzyme
markers and clinical chemistry data are required by the Guidelines
and as a result are often not available from studies for evaluation.
However, dose-related effects on absolute and relative organ weights
can be useful in the assessment of maternal toxicity. For example,
the liver often shows the earliest signs of toxicity and therefore
weight of this organ is occasionally reported and should receive
careful consideration. Enzyme markers may sometimes provide addi-
tional data concerning exposure but must also be interpreted care-
fully as to whether or not a change constitutes toxicity.
Plasma, red blood cell and brain cholinesterase inhibition
are rarely determined in a teratology study, but if these data are
available and demonstrate pronounced effects (such as > 75%
inhibition for red blood cell ChE), they may be used as a basis
for establishing maternal toxicity (even in the absence of other
clinical signs).
-------
-13-
6. Conception Rate
The conception rate (# pregnant/t mated) is an important index
for assessing the reproductive performance capability of the test
animals selected. A low conception rate may suggest maternal
health problems, poor animal husbandry, or may be due to dosing
prior to completion of implantation. Regardless of the mechanism
by which the conception rate is decreased, a reduction in available
litters will result in a less meaningful interpretation of embryo/
fetotoxicity (which includes assessment of teratogenic potential).
Both mating and artificial insemination have been used in
rabbit teratology studies. Artificial insemination presumably will
ensure a knowledge of the exact time of conception (Gibson et al.,
1966). However, a high percentage of preimplantation loss and
small litter size have been described with this method (Woo and
Hoar, 1982; Woo, 1984). A conception rate of 80% and greater in
rabbits is usually obtained by artificial insemination (Gibson et al.,
1966, Adams, 1961; Woo, 1984) if the number and concentration of
sperm used in the inseminating procedures are adequate (1 million
sperm; Walton, 1927; Gibson et al., 1966). Artificial insemination
techniques, semen dilution factors, and sperm motility are a few
of the important parameters that should be included in the report.
The time between luteinizing hormone and insemination is also criti-
cal since ova lose their fertilizability within 12 hours after
hormonal treatment (Adams and Chang, 1962).
A conception rate of 88% or greater is usually obtained in
rodents by pair or colony mating. Lower conception rates are
common when pregnant animals are shipped from a supplier.
Since dosing in teratology studies is not to be initiated
prior to the completion of implantation, as per the 1982 FIFRA
Guidelines, no significant differences in conception rates between
any test groups and controls should be observed. If significant
differences are observed, the utility of the study may come into
question and it may be rejected. Compound-related alterations in
conception rate should generally be elucidated from data collected
from multi-generation studies.
7. Corpora Lutea
The number of corpora lutea from both ovaries should be counted.
Corpora lutea are evidence of released ova and generally outnumber
implantation sites. Nulliparous animals are preferred in teratology
studies to allow adequate determination of corpora lutea. This is
to prevent potential counting error when distinguishing between
corpora lutea and corpora albicans (from earlier ovulations).
Theoretically, the number of corpora lutea among test groups
should be similar, since administration of test compound is to be
performed after implantation has been completed.
-------
-14-
8. preimplantation Loss
Preimplantation loss is determined as the difference between
the number of corpora lutea and implantation sites. An increase in
preimplantation loss suggests a time-dosing error (dosing prior to
the completion of implantation), which results in compound induced
and/or maternal stress related embryolethality.
High preimplantation loss may obscure embryo/fetotoxicity
and/or teratogenicity due to the reduced number of offspring left
available for examination. In any case, an effort should be made
to determine the cause of the high preimplantation loss.
MANIFESTATIONS OF DEVELOPMENTAL TOXICITY
The manifestations of developmental toxicity include death,
structural abnormalities, altered growth, and functional disorder.
Furthermore, there is an association between the manifestation
observed and the developmental stage in which it is induced (Wilson
and Fraser, 1977). Each type of manifestation is discussed in its
respective section.
As previously indicated, developmental toxicity may or may not
be related to maternal toxicity (Wilson and Fraser, 1977). For ex-
ample, developmental toxicity and maternal toxicity are often observed
at similar doses with alkylating agents, whereas thalidomide is devoid
of maternal toxicity at doses that produce developmental toxicity.
For organophosphate compounds, however, significant maternal toxicity
is sometimes demonstrated at doses that result in little or no develop-
mental toxicity. It should be obvious to the reviewer that agents
of the greatest concern are those that induce developmental toxicity
at levels significantly below those inducing maternal toxicity (also
see "Review End Points to be Considered").
Developmental toxicity can be dependent upon the dosing schedule.
Two dosing procedures can be utilized. The first limits dosing of
the dams to the period of major organogenesis which is days 6-15
for rat and mouse, 6-14 for hamster, and 6-18 for rabbits.
Alternatively, the dosing period may be extended to approximately 1
day before the expected day of delivery (Subdivision F Guidelines,
1982) .
Exposure during the period of major organogenesis is based
upon the principle that the embryo is most susceptible to induction
of anomalies during this period; hence, exposure beyond this period
is considered by some investigators to contribute little additional
informative data. However, many organs such as the central nervous
system, the heart, the lung, and the sexual organs, have not function-
ally or morphologically completed development by the end of major
organogenesis. Consequently, other investigators believe that
exposure of the conceptus from the day of implantation through
gestation may disclose defects that normally would not be detected
-------
-15-
under conventional teratology testing. It must be recognized that
treatment of dams from the end of major organogenesis to sacrifice
usually results in a higher incidence of fetotoxicity characterized
mainly by altered growth and possibly fetal lethality. The reviewer
should be aware that both dosing schedules are acceptable to the
Office of Pesticide Programs.
1. Death
Depending on the stage of development affected, intrauterine
death may be evident at sacrifice as resorptions and/or dead fetuses.
These deaths may be related to maternal toxicity/stress or embryo/
fetotoxicity including lethal malformations. These may be either
spontaneous or compound-related. The total number of intrauterine
deaths (resorptions and dead fetuses) is usually referred to as
postimplantation loss or fetal wastage. Under certain circumstances
where post-natal mortality data are available, these data should
also be considered as part of this assessment.
Postimplantation loss is crucial to consider in assessing the
embryo/fetotoxicity of an agent in these studies. In order to
properly assess intrauterine death, fetuses should be recovered
prior to expected delivery to prevent loss of prenatal material.
Such losses frequently occur in many species after birth since the
dams often cannibalize stillborn and/or abnormal pups. Furthermore,
evidence of resorptions can only be accounted for from the maternal
reproductive tract.
Before differentiation begins, the early embryo has great
regenerative capability (totipotent), but if the dosage is suffi-
ciently high, death of the conceptus may ensue. During early
stages of organogenesis, susceptible embyros may die and be resorbed.
This type of resorption is usually referred to as an early resorption
which is generally complete or leaving only a very small amount of
macerated tissue at the site of implantation or an implantation
scar (metrial gland). As organogenesis progresses, toxic insults
to the embryo may still result in resorption. The latter is referred
to as a late resorption which is evidenced by the presence of both
fetal tissue and placental tissue at the implantation site. After
the period of major organogenesis, the fetus is increasingly resistant
to the development of major malformations but still may display
various types of fetotoxicity. Also, appreciable fetal lethality
may still occur resulting in dead fetuses. Thus, evaluation of
postimplantation loss (fetal wastage) takes into consideration both
early and late resorption, as well as dead fetuses.
Analysis of postimplantation loss data may provide the most
useful information regarding sensitive stages of development. This
parameter is considered as a comprehensive expression of developmental
toxicity since it can include all stages of development.
-------
-16-
When embryolethal doses are reached, lethality increases at
the expense of malformation. An increase in fetal wastage (post-
implantation loss) may mask the teratogenic potential of an agent
since affected malformed fetuses may be resorbed (Beck and Lloyd,
1963) and therefore unavailable for examination. Thus, in a tera-
tology study, both a dose-related increase in fetal wastage (post-
implantation loss) and malformations must be considered as indicative
of embryo/fetotoxicity.
Only dead fetuses that do not show any significant degree of
maceration should be included in the examination for malformations
because of the frequent distortions and artifacts encountered in
such specimens (Wilson, 1973).
One of the main problems in assessing fetal wastage is when
pregnancy is suggested in a dam (proof of mating and subsequent
weight gain) but no implantations and/or abortions are reported.
In such case, it may not be possible to conclude whether or not
embryo/fetotoxicity (including teratogenesis) has occurred or
whether there has been an effect on fertility.
2. Altered Growth
In addition to known genetic, endocrine, and nutritional
factors that influence growth, many agents which are capable of
producing death and/or malformations may also cause altered growth.
Two parameters usually employed to assess altered growth are fetal
weight and fetal size (crown-rump length). The latter measurement
is presently not required by the Guidelines.
It is generally assumed that fetal body weight is inversely
proportional to litter size, i.e., higher fetal weight is usually
expected from a smaller litter size and lower fetal weight is
associated with larger litter size. Decreased fetal weight in
conjunction with comparable litter size to controls is generally
regarded as an indication of an embryo/fetotoxic effect.
Offspring that are two or three standard deviations below the
mean control weight are classified as growth-retarded pups. If
significant decreases in weight and length are found in one pup in
comparison to the normal control range, this pup is then usually
referred to as a runt. Runts are often considered manifestations
of reversible embryo/fetotoxicity, but this is not always the case as
is noted below.
Altered growth can be a transient or a permanent effect. In
the case of a transient effect (reduced fetal weight, ossification
retardation, etc.), a normal increase in weight, size, and maturation
will be expected after birth. On the other hand, failure to recover
from growth retardation is readily accepted as a defect (permanent
stunting). The term runt is normally used to describe offspring
exceptionally reduced in both size and weight. However, the criteria
-------
-17-
used in making this distinction varies dramatically among labora-
tories (Palmer, 1977). Thus, the reversibility of growth retardation
(including runting) must be assessed by considering appropriate data
from conventional teratology, post-natal, and multi-generation
reproduction studies. It should be emphasized that reproduction
studies are required for all food-use pesticides and are useful for
determining the reversibility of some aspects of embryo/fetotoxicity.
The reviewer should be aware that altered growth, like malfor-
mations, may only affect certain fetuses within a litter. Therefore
if individual fetal data are available, altered growth should be
assessed on both an individual and litter basis rather than solely
on a mean litter basis.
Although effects on fetal sex ratio are quite rare, such data
should always be examined since chemical agents may preferentially
affect a particular sex (Scott et al. , 1972). Other aspects of
altered growth are discussed under the section on "anomalies."
3. Anomalies
The fetus is usually examined for external, soft tissue, and
skeletal anomalies. Most investigators attempt to classify anomalies
as either variations or malformations, although criteria for these
classifications vary between laboratories. Variations are generally
regarded as anomalies that may not adversely affect the fetuses and
have no fatal outcome. Malformations, on the other hand, are
anomalies that are considered to have a significant adverse effect
on the fetus with or without fatal consequence. Distinction between
variations and malformations is quite difficult in some cases.
Many skeletal anomalies are so common in laboratory animals
that they are regarded as variations (if they do not have an adverse
effect on the offspring). For example, changes in the numbers or
degree of ossification of ribs and sternebrae in rabbits are con-
sidered variations (Gibson et al., 1966; Palmer, 1968; Woo, 1984).
Although skeletal variations are not truly regarded as terata, they
may be indicative of maternal toxicity/stress (Kavlock et al.,
1984) and/or fetal toxic effects if a significant dose-related
increase of a particular variation is noted above the concurrent
controls. Historical control data should also be considered as part
of this assessment.
Statistical analysis may support a determination as to whether
results differ from those of the controls. In some aspects of
assessment, N (on a litter basis) may be small and restricts the
statistical sensitivity of the evaluation.
a. External Anomalies
All fetuses should be examined for external anomalies which
are detected by means of a dissecting microscope. Limbs, tail,
-------
axial skeleton, face, palate, eyes, and head are systematically
examined.
An external malformation finding should be corroborated with
soft tissue and/or skeletal examination data whenever possible.
Some examples of external malformations are listed below:
Club foot
Dome shaped head (Hydrocephaly)
Agnathia
Micrognathia
Spina bifida
Omphalocele
Gastroschisis
Phocomelia
Micromelia
Meningoencephalocele
Open eyes
Ectrodactyly
Syndactyly
Exencephaly
Cyclopia
As noted on page 2 under the definition of "malformations", no
comprehensive list of malformations is possible.
b. Soft Tissue Anomalies
The Guidelines indicate that for rodents, approximately equal
numbers of fetuses from each litter should be prepared and examined
for skeletal anomalies and for soft tissue anomalies using appro-
priate methods. For rabbits, "each fetus should be examined by care-
ful dissection for visceral anomalies and then examined for skeletal
anomalies." The reviewer should understand that despite the Guide-
lines' reference to "visceral", the intention was that soft tissue
examination data for both the head and torso are necessary. Devia-
tions from the recommended protocol must be explained by the regis-
trant.
The selected fetuses are examined by the method of Staples
(1977) or fixed and examined for visceral anomalies by free-hand
razor sectioning (Wilson, 1965). This sectioning technique has
been found to be sufficiently adequate for rodent and rabbit fetuses,
Examples of some visceral anomalies are listed below:
Hydronephrosis
Dilated ureter
Anophthalmia
Diaphragmatic hernia
Ectocardia
-------
-19-
Ectopic testis
Renal agenesis or ectopia
c. Skeletal Anomalies
The staining method most commonly used is the Dawson's technique
(1926) using Alizarin Red S.
Distinctions between variations and malformations in skeletal
development are sometimes problematic in animals most commonly used
in teratology studies, such as the mouse, rat, hamster, and rabbit.
As previously indicated, no universal distinction has been made
between variations and malformations. Some investigators (see
Palmer, 1968, 1972, 1977) divide skeletal anomalies into common
variants, minor malformations, and major malformations on the basis
of incidence and degree of deviation from "normal." Skeletal anoma-
lies may be influenced by the genetic make-up of the species selected
(Sawin et al., 1967), by maternal factors (Green, 1962), or compound
induced (Palmer, 1972 and 1977).
Variations may be divided into groups such as delayed (retarded)
ossification, extra ossification centers, or minor irregularities
in shape and size.
As previous discussion has illustrated, skeletal anomalies should
not necessarily be regarded as compound induced. However, the pre-
sence of a dose-related increase in the incidence of individual and/or
anatomically-related skeletal variations is indicative of develop-
mental toxicity (compound induced) and/or of a maternally related
effect. For example, the presence of supernumerary ribs may be
regarded as either embryo/fetotoxicity (Kimmel and Wilson, 1973) or
maternally related effects (Kavlock et al., 1985).
Other examples of skeletal anomalies that are usually consi-
dered as skeletal variations include but are not restricted to, the
follow ing:
Skull, delayed ossification
Vertebrae, extra or absent ossification centers
Sternebrae, extra, absent, or incompletely ossified
Metacarpals (metatarsals), incomplete ossification or
unossified
Ribs, extra, rudimentary, or incomplete ossified
Talus, incomplete ossification
The best rationale for interpretation of skeletal variations
is comparison to concurrent and historical control data collected
from various studies with the same species, strain of animals, and
vehicle. Since there may be genetic drift in the strain of animals
used, the reviewer must make a comparison with historical control
data limited to a specific timeframe (that period being approximately
2 years prior to and possibly a comparable period subsequent to the
-------
-20-
study). For each variation, both the number of pups and litters
affected are important in making a final conclusion. Also, the
pattern of variations should be considered. An increase in the
incidence of variations can be assumed to be treatment related if
the increase is statistically significant over the control range
and particularly if the increase is dose-dependent. Trends should
also be carefully considered.
Skeletal anomalies that are generally recognized as malform-
ations include but are not restricted to:
Scoliosis
Kyphosis
Hemivertebrae
Syndactyly
Cranial vault, dome-shaped
Brachydactyly
Micromelia
As noted previously, the distinction between variations and
malformations is sometimes quite difficult. In such cases, the
reviewer should consult with a Toxicology Branch expert in this area,
This distinction can be quite important since only an agent that
causes a demonstrable increase in the incidence of malformations
above the spontaneous rate (primarily concurrent but also historical
control data) can be suspected of having a teratogenic effect.
Finally, it is critical for the reviewer to recognize two impor-
tant facts (palmer, 1977):
"1. That almost without exception every type of malformation
ever recorded can occur sporadically in any species, and
2. That almost every type of malformation can arise from more
than one cause."
In consideration of the above two points by Palmer, caution
should be exercised by the reviewer in assessing teratogenic poten-
tial. For example, an individual occurrence of an extremely rare
malformation in the absence of a dose-response curve may be spon-
taneous in origin. Conversely, if a larger population (N) were
sampled in a repeat study, the possibility also exists that a
treatment-related effect may be demonstrated.
4. Review End Points to be Considered
After following the assessment procedures presented in this
document, the reviewer must make the following determinations:
Maternal Toxicity
As stated previously in this SEP, maternal toxicity must be
demonstrated at the high dose-level and a No Observed Effect Level
-------
-21-
(NOEL) and Lowest Observed Effect Level (LOEL) should be determined
by the reviewer. If the reviewer cannot determine a NOEL for
maternal toxicity even at the lowest dose, as long as a NOEL for
developmental toxicity (embryo/fetotoxicity) is available from this
study, a repeat study is not necessary.
0 Developmental Toxicity
As defined previously, developmental toxicity includes death
of the developing organism, structural abnormalities (malformations
and variations), altered growth, and functional deficiency. At
present, the latter end point is not assessed in a conventional
Guidelines teratology study but useful data is sometimes available
from reproductive studies and post-natal studies (if required).
These additional studies (and especially post-natal studies) also
offer useful information for assessing the reversibility of altered
growth as well as providing data on other aspects of developmental
toxicity. It must also be understood that in a classical teratology
study, developmental toxicity is restricted to adverse effects
manifested in the developing organism prior to parturition. There-
fore, NOELs for adverse effects which were observed in classical
teratology studies (studies which encompass the embryonic and fetal
periods of the organism) are more accurately referred to as "develop-
mental toxicity (embryo/fetotoxicity) NOELs" while adverse effects
which are observed in post-natal studies should be referred to as
"developmental toxicity (post-natal) NOELs."
a. Embryo/Fetotoxicity
Assessment of developmental toxicity (embryo/fetotoxicity)
must include determinations as to whether there are dose-related
increases in postimplantation loss [early resorptions (embryo-
lethality), plus late resorptions, plus death] as well as variations,
fetal body weight changes and fetal crown-rump length changes (when
available). As stated previously in the definitions section of
this SEP (page 2), the reviewer should also be aware that develop-
mental toxicity (embryo/fetotoxicity) includes malformations.
Therefore, the NOEL and LOEL determined by the reviewer should be
based on all of the above end points and a separate NOEL for tera-
togenicity (as has been customary in the past) will not be necessary.
Malformation and fetal wastage data will now be presented in text
form as described below. If no developmental toxicity (embryo/
fetotoxicity) NOEL can be determined, a repeat study may be required.
b. Teratogenic Potential
The reviewer must determine whether a study is positive or
negative for teratogenicity but this will not necessitate a separate
teratogenic NOEL and LOEL. As discussed earlier, since teratogeni-
city is only a subset of developmental toxicity, distinction between
NOELs for subsets of developmental toxicity may not be relevant.
-------
-22-
The concept of developmental toxicity is not a new one.
Wilson (1973) indicated that there are four signs of developmental
toxicity: 1) death and resorption, 2) birth of either live or dead
malformed offspring, 3) developmental delays, and 4) decrement of
expected post-natal function. All four end points are of toxico-
logical concern if produced at a particular exposure level and in a
dose-related manner (Johnson and Christian, 1984). For example, in
utero death or resorptions are as significant as production of live
but malformed offspring. Therefore, the reviewer should indicate:
A developmental toxicity (embryo/fetotoxicity) NOEL and and
LOEL for all teratology studies following § 83-3 of the Guidelines.
- If a dose-related increase in the incidence of malformations
was noted and/or fetal wastage (postimplantation loss) was observed,
the reviewer, in addition to the developmental toxicity (embryo/
fetotoxicity) NOEL and LOEL, should accurately describe in the
conclusions/recommendations section of the review the types of
effects and the dose-levels at which these adverse effects occurred.
0 Developmental Toxicity Index
A number of investigators have suggested that a useful way to
assess the relative developmental hazard is through calculation of
a developmental toxicity index.
Johnson (1980; 1981) describes agents as "coeffective terato-
gens" if they affect the embryos only at doses near the adult toxic
range and as "non-coeffective" if effects are observed in the off-
spring at dose-levels below the adult toxic range. Using the above
terminology, it is the "non-coeffective" developmental toxins which
are usually, but not always, of greatest concern to the Agency.
Quantitative estimates of developmental hazards have been
proposed by several investigators. The "Relative Teratogenic
Index" was introduced by Fabro et al. (1982, 1985). These invest-
igators proposed a quantitative estimate of teratogenic potency
using the "Relative Teratogenic Index" which is a calculated ratio
of the maternal adult toxic dose (LDQI) and the tDg5 (minimum tera-
togenic dose based upon dose-response analysis of teratogenicity
fitted to a probit model). Johnson and Gabel (1982) suggested the
use of the A/D ratio which is the minimal effective concentration
toxic to the adult (A) and to the developing organism (D). However,
it should be noted that although all proposed estimates of relative
developmental hazard are of merit, none of them should automatically
be applied and used by the reviewer without proper consideration of
the limitations of the data. It should be recognized that the utility
of all of these index systems is limited by (1) the dose selection
(both spacing of the dose-levels and the few levels actually tested)
which is used to define maternal and developmental toxicity, (2) the
limited number of parameters used for assessing maternal and develop-
mental toxicity, (3) species selection and number of animals tested
-------
-23-
per group, (4) route of administration, etc.. It must also be under-
stood that these calculations are indices of hazard but do not repre-
sent measures of risk since they do not consider levels of human
exposure. In conclusion, it is important for the reviewer to be
aware of the many limitations of the A/D ratio and other index
systems.
The A/D ratio (Johnson and Gabel, 1982) takes into consideration
the maternal LOEL and developmental toxicity (embryo/fetotoxicity)
LOEL. A ratio of less than 1 indicates that developmental toxicity
occurs at doses higher than those producing maternal toxicity. A
ratio of 1 indicates that developmental effects are found at doses
which also produced maternal toxicity. However, a ratio of more
than 1 reveals that developmental toxicity occurs at doses lower
than those producing maternal toxicity. Using this approach, the
higher A/D ratio would suggest a higher developmental hazard for
that species, route, etc..
0 Acceptability of the Study for Regulatory Purposes
The guidance provided in this document is intended to be used
in conjunction with the Core Classification System for determining
the acceptability of a teratology study for regulatory purposes.
5. Functional Defects - Assessment of Post-Natal Studies
Developmental toxicity includes any developmental changes
induced at any stage of gestation and detected not only at birth
but also at any time post-natally. The protocol, as listed in the
Subdivision F Guidelines, is restricted to an examination at cesarean
section. Consequently, functional changes in the offspring (and
even viability) cannot be fully assessed from the Guidelines protocol.
Furthermore, there are data which demonstrate that some teratogenic
manifestations of developmental toxicity which appear after birth
are not detected by the classical Guidelines study (Gray et al.,
1982; Chernoff and Kavlock, 1982; Bui et al. , 1983).
If assessment of the classical study data raises any question
relative to the permanent status of an abnormality or raises other
concerns, the reviewer should consider requesting post-natal studies.
Other supportable rationales may also be used as a basis for requesting
these studies. Examples of dose-related findings that may trigger
such requests include (but are not restricted to): dilated renal
pelvis and/or dilated ureters, runt ing, and abnormalities of the
central nervous system.
A reviewer should consider data from the reproduction study in
assessment of possible post-natal effects; however, one must recognize
that (1) the route of exposure is usually via the diet rather than
gavage, (2) the dose-levels are usually lower and extended over longer
periods of time, and (3) these studies do not include the level of
fetal examination employed in classical teratology studies. Thus,
-------
-24-
although these data should be considered, they can rarely provide
an adequate assessment of all potential post-natal effects.
As stated previously, the NOEL and LOEL for the adverse effects
which are observed in a post-natal study should be referred to as
"developmental toxicity (post-natal)" NOELs and LOELs.
In the design of acceptable protocols for post-natal studies,
the reviewer should consult with Toxicology Branch experts. Proto-
cols will vary widely and studies may in some cases be terminated
at weaning, at sexual maturation, or even later.
IV. DEVELOPMENTAL TOXICITY RISK ASSESSMENT
Developmental toxicity risk assessment includes consideration
of all relevant data such as pharmacokinetic, metabolism, structure-
activity relationships (although not well studied) and other studies.
Many methods have been proposed for the extrapolation of develop-
mental toxicity risk from test animals to humans and this area con-
tinues to be an active one for research and scientific discussion
(Hogan, 1982). The "Margin of Safety" and the closely related
"Safety Factor" approaches are widely used and accepted. The Office
of Pesticide Programs has used the Margin of Safety approach to
evaluate potential developmental toxicity risk (e.g., Chitlik, 1980).
It has recently been cited in the Agency's "Interim Guidelines for
the Health Assessment of Suspect Developmental Toxicants" as a
recommended approach that the Agency use for developmental toxicity
risk assessment. The primary difference between the two approaches
is the a_ priori consideration of exposure data in the Margin of
Safety approach. Thus, the Margin of Safety approach is a direct
comparison (ratio) between the appropriate No Observed Effect Level
(NOEL) and the estimated human exposure, while the Safety Factor
approach divides the NOEL by a somewhat arbitrary numerical value
(the Safety Factor) to reach a "safe" level of exposure.
If a concern for the effects of developmental toxicity as
indicated by this SEP is suggested by one or more scientifically
sound studies, the reviewer should determine the No Observed Effect
Level in the most sensitive species tested. The most sensitive
species is generally used for developmental toxicity risk assessment
purposes due to the great difficulty in determining the most relevant
species from which to extrapolate to humans (USEPA, 1984a).
Ideally, epidemiological studies of birth defects after exposure
to pesticides would be the most relevant type of study upon which to
estimate risk. These studies, if available, are usually inadequate
for quantitative risk assessment due to many inherent limitations
which are outside the scope of this document.
The reviewer should next examine the types and levels of
actual or potential human exposure, keeping in mind that develop-
-------
-25-
mental toxicity may result from a single exposure. Developmental
toxicity risk assessment routinely includes dietary and worker expo-
sure, as well as other forms of exposure such as drinking water or
home use. Particular attention should be given to worker exposure
as it is frequently a much greater acute exposure than that received
via the dietary route. Worker exposure estimates are the responsi-
bility of the Exposure Assessment Branch (EAB). If a worker exposure
estimate is not available, the reviewer should consult with the
Section Head for guidance on requesting an exposure assessment to
be conducted by EAB. The EAB estimates of exposure must be on a
daily basis and should be quantified for each route of exposure.
It is the responsibility of Toxicology Branch to determine the
rate of dermal absorption. In the proper assessment of potential
risk from dermal or other routes of exposure, one should also com-
pare pharmacokinetic data such as peak plasma concentrations and/or
area under the curve of the test material and/or metabolites when
dosing is via different routes. This may prove important in under-
standing differences in developmental toxicity potential to workers
(primarily dermal and inhalation versus dietary exposure hazards;
Chitlik and Bui, 1985). Metabolism data should also be considered
since this may vary somewhat with route of exposure. In the absence
of dermal absorption data, a 100% rate of absorption should be
assumed.
Dietary exposure must also be assessed on a daily basis. The
standard "food factor" approach (Schmitt, 1978) does not provide
information regarding expected daily exposure but is an annualized
average. When completed, the new Tolerance Assessment System
(USEPA, 1984b) will not have this limitation and will be the primary
source for reviewers to obtain this information. Another useful
source for estimating single serving size is the U.S. Department of
Agriculture's "Family Food Buying Guide" (USDA, 1977). Other refer-
ences or rationales may also be used in the determination of single
serving size. Regardless of the source used, the reviewer must
estimate the maximum amount '(kg) of a raw agricultural commodity
that is likely to be consumed on a given day by an individual.
The USDA guide provides the reviewer with the "size of market
unit" as well as the "number of servings or measures per market
unit." The reviewer can then calculate from this information a
number of conceivable single serving sizes and should use, for risk
assessment purposes, the largest value. A total dietary exposure
can then be estimated by multiplying the residue level (expressed
in ppm) by the maximum likely serving size per day (in kg) and divi-
ding this by the weight of the individual exposed. The Toxicology
Branch has historically used 60 kg as the body weight of females of
child bearing age for developmental toxicity assessment. As per
the new Tolerance Assessment System (TAS), females of child bearing
age are estimated to weigh 54.8 kg. The appropriate residue level
is that which is recommended by the Residue Chemistry Branch and
may either be at the tolerance level or at the actual exposure
-------
-26-
level (if such data are made available).
The reviewer must utilize judgment in selecting the relevant
dietary exposure estimates. The reviewer should consider individual
raw agricultural commodities as well as realistic combinations of
food items which may be consumed in a single day. In all cases,
the reviewer should attempt to be conservative but not unrealistic
in the exposure assumptions made in the risk assessment.
Drinking water risks may also be of concern and are assessed
in a manner similar to dietary risk. Determination of whether or
not a pesticide has a potential for groundwater or surface water
contamination is the responsibility of the Exposure Assessment
Branch. If actual contamination exists, the most relevant contam-
ination levels must be selected in consultation with the EAB. The
National Academy of Sciences has recommended that, for risk assess-
ment purposes, it be assumed that the average adult consumes two
liters of water per day (Office of Drinking Water/EPA and NAS,
1977). The estimated daily exposure (mg/kg/day) to a pesticide in
drinking water is therefore determined by multiplying the appropriate
estimate of the residue level (mg/liter) by two liters and dividing
that amount by body weight. A more detailed description of the
hazard evaluation of pesticides in drinking water can be found in
the 1983 document "Assessment of Groundwater Contamination by Pesti-
cides" (USEPA, 1983) and in the Federal Register of June 12, 1984.
The exposure estimates generated for the worker and the consumer
of food or water containing the chemical of concern can then be com-
pared as a ratio to the No Observed Effect Level for the relevant
effect to yield an estimate of the Margin of Safety (MOS). When
possible, pharmacokinetic and metabolic considerations should be
taken into account in extrapolation from one route of administration
to another.
This determination is useful to the risk managers who must bal-
ance the health risks against societal costs and benefits in the
determination of appropriate regulatory action. The Margin of Safety
approach is a means of relating the potency of an agent to induce
an effect with an estimate of human exposure.
In a modified form, the MOS is useful even when a NOEL cannot
be assessed from available data. In that case a "what if" scenario
can illustrate possible Margins of Safety associated with speculated
(or hypothetical) NOELs (Chitlik, 1983). The following table illu-
strates the effect of a reduction in a potential NOEL from 0.1 to
0.05 to 0.01 mg/kg/day upon the calculated Margin of Safety:
-------
-27-
Food Item
Broccoli
Brussel-
Sprouts
Cabbage
Celery
Cauliflower
Kohlrabi
Onions
(Dry bulbs)
Sugar beets
(sugar)
Serving
Size (kg)a
.103
.081
.140
.170
.140
.182
.202
.202
Tolerance
(ppm)
0.10
0.75
0.75
0.10
0.30
0.75
0.75
0.05
Exposure
(mg/kq bw/
day) "
0.00018
,00109
,00188
,00030
,00075
0.00244
0.00271
0.00018
MOS assuming
developmental tox,
NOEL of :
0.1, 0.05, 0.01
(mg/kg/day)
543 271
92
34
328
133
41
46
17
164
67
20
37 18
554 277
54
9
3
33
13
4
4
55
(a) Size of individual food services were obtained from Family Food
Buying, L. Fultons C. Davis and E. Matthews, USDA No.37, 08/1977.
(b) female body weight of 55.9 kg utilized based upon TAS draft
(Note: now revised to 54.8 kg; USEPA, 1984b)
Calculations of Margin of Safety must be qualified by description
of the strength of the evidence and biological uncertainties. The
National Research Council, in a widely cited report "Risk Assessment
in the Federal Government: Managing the Process" has recommended that
risk assessments "... set forth in detail the nature and quality of
the relevant scientific evidence concerning the substance in question
and should cover all relevant components of risk assessment" (NRC,
1983). For developmental risk assessment, the uncertainties are
great and have been discussed in detail in several Agency documents,
e.g., USEPA, 1980, and USEPA, 1984a. The routine assumptions made
in the extrapolation of animal data to humans need not be reiterated
for each risk assessment. The reviewer should concentrate on the
discussion of the quality of the studies supporting the concern for
developmental toxicity risk, the strength of the NOEL and the uncer-
tainties in the exposure assessment (such as the lack of information
regarding dermal absorption or uncertainties relative to residue
levels in the diet).
-------
-28-
V. BIBLIOGRAPHY
Adams, C.E., 1961. Artificial insemination. In: The semen of
animals and artificial insemination, ed. J.P. Maude. England.
Benirschke, K., Garner, P.M., and Jones, T.C., 1978. In: Pathology
of Laboratory Animals, eds. K. Benirschke, P.M. Garner, and T.C.
Jones, Springer Verlag.
Beck, F. and Lloyd, J.B., 1963. An investigation of the relationship
between foetal death and foetal malformation. J. Anat. 97, pp. 555-564
Bui, Q.Q., Sperling, F. and West, W.L., 1983. Developmental toxic
effect after subcutaneous injections of methadone in Charles River
CD-I mice. Drug and Chemical Toxicology, 6(1), pp. 41-70.
*Burin, G.J., 1982. Review of rabbit teratology studies of
Ethalfluralin. Memorandum to William Burnam, 11/24, USEPA.
Chang, M.C., 1958. Capacitation of rabbit spermatozoa in the uterus
with special referene to the reproductive phases of the female.
J. Endocrinol. 63, pp. 619-628.
Chernoff, N. and Kavlock, R.J., 1982. An in vivo teratology screen
utilizing pregnant mice. J. Toxicol. Environm. Health 10, pp. 541-550.
*Chitlik, L.D., 1980. TOK (Nitrofen) - Oncogenic, teratogenic, and
mutagenic evaluations. Memorandum to Marcia Williams, 05/21,
USEPA.
*Chitlik, L.D., 1983. Margin of Safety (MOS) assessment for Nitrofen
(TOK). Memorandum to M. Branagan, 06/15, USEPA.
*Chitlik, L.D. and Bui, Q.Q., 1985. Review of toxicology studies with
Karathane. Memorandum to Jay Ellenberger, 04/01, USEPA.
Dawson, A.B., 1926. A note on the staining of the skeleton of
cleared specimens with Alizarin Red S, Stain Technology 1, p. 123.
Fabro, S., 1985. On predicting environmentally-induced human
reproductive hazards: An overview and historical perspective.
Fund. Appl. Toxicol. 5, pp. 609-614.
Fabro, S., Shull, G. and Brown, N.A., 1982. The relative terato-
genic index and teratogenic potency: Proposed components of
developmental toxicity risk assessment. Terato. Carcino. and
Mutagenesis 2, pp. 61-76.
Federal Register, 1984. Volume 49, No. 114, pp. 24330-24335.
-------
-29-
Gibson, J.P., staples, R.E., and Newberne, J.W., 1966. Use of
rabbit in teratogenicity studies. Toxicol. Appl. Pharmacology
9, pp. 398-408.
Gray, L.E. Jr., Kavlock, R.J., Chernoff, N., et al., 1982. Prenatal
exposure to the herbicide 2,4-dichlorophenyl-p-nitrophenyl ester
destroys the rodent harderian gland. Science 215, pp. 293-294.
Green, E.L., 1962. Quantitative genetics of skeletal variations
in the mouse. II. Crosses between four inbred strains (C3H, DBA,
C57BL, BALB/c). Genetics 47: pp. 1085-1096.
Hafez, E.S.E., 1970. Reproduction and breeding techniques for
laboratory animals, eds. Lea and Febiger, Philadelphia'.
Hogan, M.D. and Hoel, D.G., 1982. Extrapolation to man. In:
Principles and methods of toxicology, ed., A.W. Hayes, Raven Press.
Johnson, E.M., 1980- A subvertebrate system for rapid determination
of potential teratogenic hazards. J. Environ. Pathol. Toxicol. 4,
pp. 153-156.
Johnson, E.M., 1981. Screening for teratogenic hazards: Nature of
the problems. Annu. Rev. Pharmacol. Toxicol. 21, 417-429.
Johnson, E.M. and Gabel, B.E.G., 1982. Application of the hydra
assay for rapid detection of developmental hazards. J. Amer. Col.
Toxicol. 1(3), pp. 57-71.
Johnson, E.M. and Christian, M.S., 1984. When is a teratology study
not an evaluation of teratogenicity? J. Amer. Col. Tox. 3, pp. 431-434
Kavlock, R.J., Chernoff, N., and Rogers, E.H., 1984. Supernumerary
ribs: Is there a relationship with alterations in maternal
health status. Teratogenesis Carcino. and Mutagen. 5(1).
Kavlock, R.J., Chernoff, N., and Rogers, E.H., 1985. The effect of
acute maternal toxicity on fetal development in the mouse. Terato-
genesis, Carcino. and Mutagen. 5, pp. 3-13.
Kimmel, C.A. and Wilson, J.G., 1973. Skeletal deviations in rats:
Malformations or variations? Teratology 8, pp. 309-316.
National Academy of Sciences, July 11, 1977. Drinking water and
health, Federal Register, 42, p. 132.
National Research Council, 1983. Risk assessment in the federal
government: Managing the process. National Academy Press, Washington
D.C., p. 159.
Palmer, A.K., 1968. Spontaneous malformations of the New Zealand
white rabbit: The background to safety evaluation studies. Lab.
Animals 2, pp. 195-206.
-------
-30-
Palmer, A.K., 1972. Sporadic malformations in laboratory animals
and their influence on drug testing. In: Drugs and Fetal Development
eds M.A. Klinberg, A. Abramovici, and J. Chemke. Plenum Press.
Palmer, A.K., 1977. Incidence of sporadic malformations, anomalies
and variations in random bred laboratory animals. In: Methods in
Prenatal Toxicology, eds D. Neubert, H.J. Merker, and I.E. Kwasigroch
Georg Thieme Pub., Stuggart.
Sawin, P.B. and Gow, M. 1967. Morphogenetic studies of the rabbit.
XXXVI. Effect of gene and genome interaction on homeotic variation
Anat. Rec. 157, pp. 425-435.
Schardein, J.L., 1985. In: Chemically induced birth defects.
ed. J.L. Schardein, Marcel Dekker Pub., New York
Schmitt, D., 1978. Update of food factor tables. Toxicology Branch,
HED, OPP, USEPA.
Scott, W.J., Butcher, R.L., Kindt, C.W., and Wilson, J.G., 1972
Greater sensitivity of female than male rat embryos to acetazolamide
teratogenicity. Teratology 6, pp. 239-240.
Shepard, T.H., 1983. Catalog of Teratogenic Agents. John Hopkins
University Press.
Staples, R.E., 1977. Detection of visceral alterations in mammalian
fetuses. Teratology 9, pp. A37 - A38.
U.S. Dept. of Agriculture, 1977. Family food buying guide.
U.S. Dept. of Agriculture, 1982. Foods commonly eaten by individuals;
Amount for day and per eating. Human Nutritional Information Service.
Home Econ. Res. Rep. 44.
USEPA, 1980. Assessment of risks to human reproduction and to
development of the human conceptus from exposure to environmental
substances. National Technical Information Service US. Dept. of
Commerce, No. DE 82-007897.
USEPA, 1982. Pesticide Assessment Guidelines, Subdivision F,
Hazard Evaluation: Human and Domestic Animals.
USEPA, 1983. Assessment of ground water contamination by pesticides
Hazard Evaluation Division, OPP, USEPA.
USEPA, 1984a., Proposed Guidelines for the health assessment of suspect
developmental toxicants and request for comments. Federal Register,
November 23.
USEPA, 1984b., Ethylene Dibromide (EDB) - Scientific support and
Decision. Document for grain and grain milling fumigation uses.
Office of Pesticide Programs, pp. 47-49.
-------
-31-
Walton, A., 1927. Relation between density of sperm suspension
and fertility as determined by artificial insemination of
rabbits. Proc. Roy. Soc. B 101, pp. 303-315.
WHO, 1967. Principles for the testing of drugs for teratogenicity.
WHO Technical Report Series No. 364, Geneva, Switzerland.
Wilson, J.G., 1965. Methods for administering agents and detecting
malformations in experimental animals. In: Teratology: Principles
and Techniques., eds J.G. Wilson and J. Warkany. Univ. of
Chicago Press.
Wilson, J.G. 1973. In: Environment and birth defects, ed., J.G.
Wilson, Academic Press.
Wilson J.G. and Eraser, F.C., 1977. in :Handbook of teratology.
I. General Principles and Etiology, eds. J.G. Wilson and F.C.
Fraser, Plenum press.
Woo, D.C. 1984. Control data of teratological studies in New
Zealand White rabbits (A joint study by MARTA), in Principles
in Teratology, Teratology Society.
Woo, D.C. and Hoar, R.M., 1972. "Apparent hydronephrosis" as a
normal aspect of renal development in late gestation of rats:
The effect of methyl salicylate. Teratology 6, pp. 191-196.
Woo, D.C. and Hoar, R.M., 1982. Reproductive performance and
spontaneous malformations in control New Zealand rabbits: A
joint study by MARTA. Teratology 25, p. 82A.
* These references are internal EPA documents which contain
confidential data. Removal of such data will eliminate the
instructional value of the document. Therefore, these documents
must not be released from HED without the proper authorization and
legal clearances.
------- |