U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
PB-273 195
Survey and Evaluation of Techniques Used
in Testing Chemical Substances for
Teratogenic Effects
Tracer Jitco, Inc, Rockville, Md
Prepared for
Environmental Protection Agency, Washington, D C
Oct 77
-------
EPA-560/5-77-007
SURVEY AND EVALUATION OF TECHNIQUES
USED IN TESTING CHEMICAL SUBSTANCES
FOR TERATOGENIC EFFECTS
October 1977
^
FINAL REPORT
.Contract 68-01-2204, Task 29
Office of Toxic Substances
Environmental Protection Agency
Washington, D.C. 20460
-------
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGEN.CY. ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE, IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE.
-------
EPA-560/5-77-007
Survey and Evaluation of Techniques
Used in Testing Chemical Substances
for Teratogenic Effects
Final Report
October 1977
Contract 68-01-2204, Task 29
Office of Toxic Substances
Environmental Protection Agency
Washington, B.C. 20460
Lois Jacob
Project Officer
/ *-
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-560/5-77-007
4, TITLE AND SUBTITLE
(Survey and Evaluation of Techniques Used in Tesi
Chemical Substances for Teratogenic Effects
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME ANO ADDRESS
Tracer Jitco, Inc.
1776 East Jefferson Street
Rockville, Maryland 20852
12. SPONSORING AGENCY NAME ANO ADDRESS
Office of Toxic Substances
U.S. Environmental Protection Agency
Washington, D.C. 20460
IS. SUPPLEMENTARY NOTES
3. Rij^lP BNT^ ACCESSION" «Q.£"
i* 0 £. / J i 7_J
6. REPORT DATE
ting October 1977
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
H. CO'NTRACT/GFtANT NO.
68-01-2204, Task #29 '
13. TYPE OP REPORT ANO PERIOD COVERED
Final
14. SPONSORING AGENCY COOS
15. ABSTRACT
This survey summarizes and evaluates the methods currently used or potentially
useful for testing chemicals for teratogenic effects. ;
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS b.lDENTU
Teratology
Testing methods
Teratogenic chemicals
18. DISTRIBUTION STATEMENT , 19. SECUF
, 20. SECUF
= IERS/OPEN ENDED TERMS C. COSATI Fieid/GlOUp
\ I TY CLIASS (Thi3 Report) \
?ITY CLASS (This page) 22. PRICE I M f
EPA Form 2220-1 (9-73)
-------
TABLE OF CONTENTS
Page
I. Introduction - Objectives and Scope 1
A. Objectives 1
B. Scope 2
C. Difficulties of Teratogenicity Testing 3
II. Problems of Design 4
A. Extrapolation from Animals to Man 4
B. Designing a Prospective Animal Testing
for Teratology 6
C. Selection of Control Parameters 8
D. Knowledge of the Chemical to be Tested 10
E. Expected Types of Malformations 11
III. Currently Available Methodologies and Their Evaluation 12
A. In Vivo Methods 12
1. Species and Strain of Test Animal 12
2. Administration of Chemical 19
3. The Test Environment 36
4. Observations 38
5. Interpretation of Results 45
B. In Vitro Systems 50
IV. Currently Used Methodologies 53
V. Surveillance '' 57
VI. Economics 61
VII. List of Known Teratogens 64
VIII. Recommendations 65
A. Screening Chemicals Already in the Environment
for Teratogenic Effects 65
B. Prospective Screening of Chemicals Not Yet in the
Environment for Potential Teratogenicity 69
IX. Possible Future Methodologies 74
A. Postnatal Evaluations 74
B. Possible Shortcuts in Current Procedures 74
C. Additional Species of Test Animals 75
D. Improved Monitoring of Human Populations 76
X. Concluding Remarks 77
-------
LIST OF ILLUSTRATIONS
Figure # Page #
1 " 43 Cleared, Alizarin - stained monkey fetus removed by
hysterotomy from an untreated rhesus female on day
100 of pregnancy.
2 44 Radiograph of a monkey fetus removed by hysterotomy
from an untreated rhesus female on day 100 of
pregnancy.
ii
-------
LIST OF TABLES
Table # Page # ' Title
1 5 Gestation Stages
2 Hi Some Animals Used or Available for Teratogenicity
Testing
3 15 Animals Recommended or Required by Various Gentries for
Drug Teratogenicity Testing
4 15 A New Concept on Teratogenicity Testing Based on
Multilevel Tests in Different Types of Animals
5 17 Comparative Thalidomide Teratogenicity
6 17-18 Thalidomide Teratogensis in Primates
7 18 Percent Malformations Produced in Four Inbred Strains
of Mice by Galactoflavin
8 19 Range of Spontaneous Malformations in Animals
9 20 Teratogenicity Toxicity Relationships for Some Drugs in
the Rat .,
10 27 Episodes on the Reproductive Cycle of Mammals
11 28 Gestation Times for Some Laboratory Animals and Man
12 29 Critical Periods of Organogensis in Animals
13 30 Some Chemical Agents Known to Influence Rates of
Metabolic Degradation of Themselves and/or Other
Repeated Dosages.
1U 31 Ways in Which Repeated Treatment Prior .to the Peak
Susceptible Period of the Embryo May Produce Misleading
Results
15 32 Numbers of Pregnant Rats Treated at Consecutive 3-day
Periods
16 32 Types of Tests for Developmental Abnormality Depending
on Duration of Treatment
17 42 .Weights and Measures on 15 Control Monkey Fetuses
Removed by Hysterotomy on Gestation Days 99-101 From
Untreated Rhesus Females
18 52 Some In Vitro Systems that have been Employed in
Teratology
19 62 Typical Costs for Some Carcinogenesis Bioassay
Functions Determined by Tracor Jitcb, Inc. for the
Cancer Institute in 1976.
20 63 Laboratory Procurement and Maintenance Costs Determined
by the American Association of Medical Colleges for
Year Ending June 30, 1973-
iii
-------
I. INTRODUCTION - OBJECTIVES AND SCOPE.
Adverse effects on human reproduction during early gestation are fre-
quent. According to the March of Dimes National Foundation, of the three
million births in the United States each year, about 200,000 (6.7t) of all
newborns have birth defects and more than 560,000 (18.7$) of all pregnancies
terminate iri*"spontaneous abortions, stillbirths or miscarriages caused by
maldevelopment (101).
Concern over the potentially harmful effects of drugs taken during preg-
nancy is relatively recent, dating back to the "thalidomide tragedy" which
occurred in the 1960s following introduction of this sedative-hypnotic com-
pound, into the market in 1956. Its use during early pregnancy was shown to
result in phocomelia. Since then other compounds such as androgens, pro-
gestrogens, aminopterin, methotrexate, and antithyroid drugs have been posi-
tively identified as human teratogens. Compounds regarded as "suspect tera-
togens" include anticonvulsants, neurotropic, anoexogenics, oral hypoglycemics
and alkalylating agents, while those regarded as "possible teratogens" include
aspirin, antibiotics, quinine, barbiturates, etc. (128).
Compounding the problem of women of child-bearing age exposed to drugs
during pregnancy is the problem of occupational exposure to chemicals in
various laboratories and industrial work places. It is estimated that. 2,000
or more new chemicals are produced each year, and at least 200 of these are
released into the environment in amounts that could cause developmental
disturbances (108). For this reason the testing of chemicals used in agri-
culture, food processing and production and industry is now considered manda-
tory by most regulatory agencies. The safety assessments of these chemicals
are based on experimental data obtained for several species tested by
employing various teratologic approaches.
A. Objectives
This survey is intended to summarize and evaluate the methods currently
used or potentially useful for testing chemicals for teratogenic effects. The
scope of the survey is limited as described hereunder.
-------
B. Scope
A teratogen, as conventionally understood, is an agent that causes struc-
tural defects or anomalies of prenatal development that are present, but not
necessarily detectable, at birth. Such defects are often called congenital
malformations. This concept may be considered narrow because it neglects
other facets of embryotoxicity such as prenatal death, overall growth retar-
dation, and functional deficits. Usually such deficits, when caused by a
teratogen, become detectable long after birth and only then may be traced to a
prenatal origin.
The environmentally-induced developmental defects covered in this report
include intrauterine death, fetal growth retardation, structural defects, and
early-appearing functional deficits. A limitation imposed on this report is
that the defects shall have originated after implantation, which excludes
defects that are hereditary, whether genie or chromosomal. A limitation on
the coverage of functional deficits is that both behavioral studies and obser-
vations later than the immediate postnatal period are specifically excluded.
Many environmental agents have been found to cause developmental defects
in experimental animals, including many that are outside the scope of this
report (e.g., radiation, noise, temperature extremes). The report considers
only methods used for testing whether chemical substances introduced into the
environment cause developmental defects when females have been exposed to them
during pregnancy. Defects caused by chemical deficiencies are included only
if they result from positive introduction of other chemical(s) into the
environment, and the test methodology involves a test for the deficiency.
Studies to evaluate food chemicals and therapeutic drugs for terato-
genicity were not analyzed in depth unless there was some unique aria* appro-
priate aspect of the experimental design.
-------
This survey is evaluative and based( on a literature search and by informa-
tion from consultants it covers all methods that have been used or have been
recommended for use during the past 50 years. It also covers some methods
which the evaluation suggested might be considered or reconsidered in the
future. It does not cover every method that has ever been used for terato-
genicity testing.
The methods surveyed are those particularly directed to (a) testing
chemicals for teratogenicity and (b) surveillance of human populations for
teratologic manifestations related to environmental chemicals. The chemical
may enter the proximate or ambient environment by whatever means; the report
is .not concerned with how or in what amounts it enters the environment.
However, a central question is, How may pregnant human females be exposed? and
tests for teratogenicity are generally designed according to the likely
route(s) of human exposures. Test findings are considered only to the extent
that they are criteria for the evaluation of methods.
C. Difficulties of Teratogenicity Testing
The ultimate question that a test method is designed to answer is, Could
this chemical harm unborn children? Direct experimentation being impossible,
the question becomes, How can an answer to this question be inferred? or, What
is the probability of adverse developmental effects? It is clear that no
available prospective methodology achieves the full objectives. At best,
available methods estimate degrees of probability that a chemical may be tera-
togenic for human beings. Even retrospective surveillance methods often reach
equivocal conclusions. This report therefore addresses questions such as What
animal testing methods are currently used, what available methods come nearest
to providing a valid answer, what are the limits of confidence of the best
available methods, and what other methods may be devised?
-------
II. PROBLEMS OF DESIGN
A. Extrapolation from Animals to Man
The thorniest problem undoubtedly is that of extrapolability to man. No
single factor - whether related to duration and level of exposure, response,
timing, or environment - can be transferred directly with acceptable confi-
dence. Retrospective surveillance of human populations is the only direct
approach that is at present fully acceptable and feasible. Even then, sur-
veillance is often incomplete and unrewarding in practice. The thalidomide
disaster demonstrated that nonspecific retrospective surveillance was inade-
quate. Prospective animal tests must therefore be designed to best infer and
assess any potential teratogenicity of agents for man.
In the absence of knowledge of the many possible factors (metabolic,
, *
embryological, and other) that may underlie susceptibility of human embryos to
a given chemical, species and strains of test animal cannot be chosen
rationally. These choices must normally be arbitrary. Also, the variables of
the test design must be approached empirically. For example, the nature of
the test environment, amounts of exposure to the chemical, timing of
exposures, responses to be looked for, and frequency of such responses in the
test animal population must to some extent be decided for each suspect
chemical individually. Final, conclusive evidence of teratogenicity for man
must be sought in the offspring of women exposed during pregnancy after a
chemical has entered their environment.
There are many reasons for this uncertainty. An important one is that the
time course of intrauterine development varies among species without regard to
lifespan or gestation period. For instance, the preimplantation period
(during which cells are undifferentiated, so there is no basis for terato-
genesis) varies from M.5 days in hamsters to 6.5 days in man to 10 days in
sheep while the total gestation period varies from 16 days in hamsters to 278
days in man (69). Organogenesis occupies the first 6 weeks of human develop-
ment, but overlapping and telescoping of the steps of development vary among
species, so that comparisons are usually difficult (Table 1).
-------
Table 1 (98)
STAND.
HORIZ.
RAT
CHICK
MAN
7-8
III
IV
« 6 t 6
GESTATION STAGES
42.5 tf
4 6.5. *| &5 [t 1Z5 »
12
VIII
8.5
1.5
19
4- 2 >>
4 1.5 *
4 8 *
16
XI
10.5
3
27
4 1 »
4 0.25 >
4 2 *
18
XII
11.5
3.25
29
4 1^-»
41. 75-*
4 7*
25
XIV
12.5
5
36
49.5-*
416*
4 231-*
-ft t + f Jt
FERTILIZATION * PRIM. STREAK TAIL BUD
IMPLANTATION NEURULA END OF EMB. PHASE
35-
38
XXIII
f
22
21
267
4
B
1
R
T
H
Abbreviations: Horizons shown in Roman numerals, represent a series of 23 stages in the
development of the human embryo (139); Standard, shown by Arabic numerals
represent a series of 36 stages in the development of the human embryo (169),
-------
The chronology of structural developments in the central nervous system
(CNS), especially at the histological level, varies greatly among species, and
the final steps (postnatal) probably occur later in man than in most test
animals. The CNS is a prime target for teratogenesis. But there is little
comparative knowledge"about development, as was stated in 1967 by the WHO
Scientific Group (173).
Other differences between man and many test animals that may be important
in teratology include: placental differences; single implantation in humans
versus multiple implantations in many test animals (relevant to resorption of
embryos and fetuses, and to abortions); and differences in endocrinology,
metabolism, pharmacology, pharmacokinetics, and nutrition. Exposed human
populations are usually larger and more randomly bred.than are groups of test
animals, and so they are genetically more heterogeneous.
Finally when a compound has been demonstrated to be teratogenic in
animals, it must be determined whether the mechanism is specific to the animal
under investigation.or whether it is also relevant to man. Information about
the mechanisms involved in maternal and fetal toxicity and teratogenesis could
allow some extrapolations to be made from animal models to man (IIU).
B. Designing & Prospective Animal Test for Teratogenicity
i
Experimental conditions and test designs contain many variables that can
influence the interpretation of results. These must be recognized and then
carefully standardized or regulated.
Experimental variables include: all aspects of husbandry such as quality
of feed, bedding, drinking water; temperature, humidity, barometric pressure,
amount and periodicity of light and noise; cage size, material, type of racks;
group size unless animals are caged individually; state of health and standard
of laboratory care; species and strain of test animals; parity and time of
mating of test females; exposures of test females to the chemical (dose range,
dose timing, route of exposure, duration of treatment) decided largely by the
nature and purpose of the test chemical; and treatment of concurrent control
, t
animals.
-------
An important factor that must be considered in the design of an experi-
mental model is the pharmacological activity between the maternal and fetal
compartments (U8, 8*0. In general, factors that alter the level of free drug
or its metabolite in the maternal compartment are capable of altering the
level of free drug in the fetus as well. The degree of transfer of drugs from
the placenta is directly proportional to the free drug concentration in the
maternal plasma. Since the fetal system has no other means of drug excretion
except through redistribution to the maternal compartment, any metabolism of
the drug within the fetus or the placenta would result in containing that drug
within the fetus and lead to drug-induced teratogenesis (55).
The design of -concurrent control treatments should never be approached in
a routine manner (see Section II, 3). The valid observation of results can
demand skilled insight at the design stage. What to look for and when to look
for it should be decided in detail according to the priorities of each
particular test. For example, a choice of methods or of timing may be
available for recovery of offspring from treated females, for examinations of
fetuses, and for postnatal examinations of offspring not killed at birth. On
the other hand, all teratogenicity tests require meticulous record-keeping to
be kept throughout, necropsy of any animal dying spontaneously, and reporting
of all dead and resorbed conceptuses.
0
When observations are complete, the ways in which the data should be
analyzed and interpreted will have been decided as part of the test design.
The methods for determining statistical significance and the uses to be made
of significant data should be chosen according to the possible sources of
error, as foreseen, and the anticipated limits of confidence.
Therefore, only well-qualified experts should undertake the design of
teratogenicity tests. Such experts will know the extent to which they can
control each of the experimental variables in the actual circumstances. They
will make informed choices when certainty is impossible and will be aware of
controversies.
-------
For example, the approach to dosage with potential teratogena has been
much debated. A program of single doses to test specific sensitivity of every
tissue at the time of its rapid proliferation has been recognized as ideal but
far too costly (115). However, multiple doses have been stated to produce, on
the one hand, cumulative effects (115), or on the other, fewer malformations
owing to development of metabolic tolerance (166). To escape this difficulty!
dose levels weighted for known tissue sensitivities ("equivalent" doses) have
been advocated (115). Although the "threshold" concept of dose-responsiveness
is not accepted universally, it has been recommended for planning purposes
(166). The concept of a "teratogenic ratio" (toxicity of a chemical to the
test female versus its toxicity to the fetus) has been suggested as a possible
predictor of teratogenicity for man, but the usefulness of this ratio is
disputed (122). Thus, the well-informed investigator has a wide choice of
approaches to dosage available at the design stage.
Studies intended to discern relationships between embryonic exposure and
teratogenic response should incorporate the following: (1) ability to
distinguish between the parent compound and its metabolite, (2) equality of '
dosages, (3) similarity of treatment schedule, CO use of comparable develop-
mental stages and (5) determination of embryonic levels before and after
attainment of peak levels. Other essentials to be considered include the
refinement of techniques to measure levels of teratogens to minimize errors
from embryonic loss or leaching and t.ie use of large sampling units to allow
for variances within and among experimental litters.
C. Selection of Control Parameters
That controls should be appropriate goes without saying. However, the
difficulty lies in defining what is appropriate, since one can never ascertain
all of the relevant variables.
Concurrent controls should, first, be contemporaneous with the test
procedures. They should duplicate the test conditions in every conceivable
way (except one - the test chemical): namely, administration of the test
vehicle, all aspects of animal husbandry, health, selection of the test group
animals.
8 '
-------
In general, however, background knowledge of so-called spontaneous malfor-
mations and the range of yariant3 (skeletal and visceral) characteristic of
the selected strain of test animal may be of great value. An increase in an
infrequent type of defect may not occur in the small groups used in any one
test, or if it does occur, it may not be correctly interpretable in that
context alone.
On the other hand, information from "positive controls" (responses of test
animals to known teratogens) is regarded by many investigators as either
irrelevant to the substance being screened, or simply redundant. If animals
are available for this purpose, it may be more meaningful to add them to the
treatment groups.
Many pitfalls in the design of controls have been discovered. Cases tend
to be individual but some examples of the sort of problem will perhaps
illustrate why it is necessary to think carefully in advance.
In a study of a pesticide, intermittent exposure was designed. Later, the
aerosol was found to have been adsorbed onto room surfaces from which it had
evaporated slowly and continuously, thus eliminating the intended no-treatment
periods (149).
Intermittent exposure can also be rendered continuous by an affinity of
the substance for the subject's own tissues. "Paraquat," a herbicide with
delayed lethality, was found to be retained by lung-surface tissues, while
systemic uptake depended on how rapidly the tissues could elaborate ATP; thus
after intermittent exposure, the actual uptake was continuous (6).
Another example of brief or intermittent exposure producing a continuous
challenge and resulting in delayed fetotoxicity involves the azo dyes, such as
trypan blue. These dyes become concentrated in macrophages and are released
slowly; after brief exposure of a pregnant test female, the dye persists in
her plasma, so that the conceptuses are exposed for several days after the
brief maternal exposure. Strontium-90 is another chemical that is released
slowly and continuously after brief exposures.
-------
D. Knowledge of the Chemical to be Tested-
It is important to know in advance how the chemical is metabolized, as
Cahen has pointed out (13). Often there are differences between species or
strains of animals in rates and pathways of metabolism and elimination, which
might affect the exposure of a conceptus to teratogenic effects of a chemical
or its metabolites. Usually the manufacturer or principal user of a chemical
is the major or only source of such knowledge. Pertinent information
includes, at the least, purity and stability of the chemical, its cumulative
toxicity, physiologic (tissue) storage capacity, rate of transport through
maternal blood and storage and metabolism by the developing fetus.
Until a few years ago, placental membranes were regarded as impermeable to
drugs and chemicals. It has now been shown that most drugs readily traverse
the placenta, with permeability modulated by lipid solubility, molecular
weights, age of placenta, etc. (28).
Measurements of plasma half-life, protein binding metabolism and tissue
distribution in adult animals alone cannot provide information on the
differences in susceptibility among offspring of a litter. A combined
analysis of these parameters, along with placental transfer kinetics and
embryonic localization of the test chemical and its metabolites will enable
one to relate both the dose administered to the mother and the dose reaching
the fetus to the observed teratological effect.
It is also necessary to consider effects that originate from metabolic
interactions within the maternal-fetal system. These effects may be due to
microsomal activity of the liver, placenta and/or fetus, which could be
inhibited or stimulated. For example, thalidomide has been regarded as tera-
togenic mainly during early-stage organogenesis. However, free glutamic acid
has been found to produce fetal abnormalities during late-stage CNS develop-
ment (150), and free glutamic acid Is a major metabolite of thalidomide.
Although this metabolite does not seem to be responsible for the early-stage
effects of thalidomide, the question of whether thalidomide-derived glutamic
acid could cause late-stage CNS defects remains unanswered.
10
-------
E. Expected Types of Malformations ,
t '
The investigator who looks only for particular types of malformation may
be misled. To some extent, the time of exposure can influence the type of
malformation when a chemical is more toxic to the conceptus than to the mother
(13). More importantly, the possible effects of a previously untested
chemical cannot always be entirely foreseen. Therefore, when a chemical is to
be screened for potential teratogenicity, the observations should be designed
to detect any type of developmental anomaly.
11
-------
Ill. CURRENTLY AVAILABLE METHODOLOGIES A*ND THEIR EVALUATION
A. In Vivo Methods
1. Species and Strain of Test Animals
Animals used for teratogenicity studies, according to a 20-year literature
search, are listed in Table 2. Species recommended or required by various
countries for testing teratogenicity of drugs are listed in Table 3. .Rats,
mice, and rabbits have been used for the most part.
Wilson (163, 165) has recommended a multilevel approach, using a rodent
(rat, mouse, hamster, or rabbit) at the first level, a carnivore (dog, cat, or
ferret) or ungulate (pig or sheep) at the second and third levels, and a non-
human primate (monkey or baboon) at the fourth level^, as shown in Table 4.
Properly applied, this approach takes into account such factors as numbers of
individuals expected to be exposed, levels of exposure, and cost-risk-benefits.
The multilevel approach acknowledges that there is no ideal animal species
for testing for potential teratogenicity in man. At present, there is little
experience recorded using this approach.
The ideal species would resemble man as to absorption, metabolism, .and
elimination of chemical substances, placental transfer of those substances and
their metabolites, and course of development of the conceptus (165). Because
these and other attributes are either known or believed to influence fetal
responses to environmental factors, much research has concentrated on degrees
of resemblance for one or another of these attributes. Other research has
looked for similar responses to particular substances in man and various
animal species.
Neither of these directions of search has been outstandingly productive.
Wilson points out (165) that similarity of response between any species and
man varies from one substance to another. Kalter notes that reported
similarities of attributes tend to be .incomplete and >poorly
12
-------
substantiated, that the reasons for the .differences must be better understood
before a test species can be chosen on a fully rational basis, and that mean-
while it might be more helpful to compare various species with one another
than with man.
For example, placental similarity is recognized as desirable (163). On
the other hand, many major structural malformations arise before a true
placenta is established in any species, even man (115). Kalter notes that
much teratogenicity occurs before placental differences become important.
Thus, Tuchmann-Duplessis (152) urged the use of small rodents for screening,
and use of primates only when specially indicated. One such indication might
be a suspicion of late-stage teratogenicity for man.
Responses to particular substances vary greatly among species. This is
shown by Kalter for thalidomide in Table 5, and he also emphasizes the genetic
variations among strains of a species which influence the teratogenic
response. Table 6 shows the responses of various primates to thalidomide, and
Table 7 shows the responses of four strains of mouse to galactoflavin.
The reaction to a specific teratogen may also be the result of differences
in the rate of metabolism and the qualitative differences in metabolic path-
ways among species. For example, the role of the fetal liver in the rat,
guinea pig, rabbit and swine, with reference to drug metabolism is negligible
or absent, whereas in humans the fetal hepatic site is very active. Although
the plasma half-life of thalidomide is the same in the rabbit (susceptible
species) and the rat (nonsusceptible species), the drug is more readily
absorbed and its metabolites more slowly excreted in the rabbit than in the
rat (128). Even when two species metabolize a drug at the same rate, the
metabolic products may be different and cause different teratogenic responses
in the two species (e.g, imipramine) (153).
A number of different animal species have been used in an attempt to
determine the most satisfactory model for teratological research for predic-
ting the hazard to man. However, none have been totally adequate because of
high spontaneous malformation rate (mouse) (Table 8), low
13
-------
Table
Some Animals Used or Available
Animals
for Teratogenicity Testing
Referencess
Invertebrates;
Echinoderms:
Sea urchins
Sand dollars
Insects:
Fruit fly
Vertebrates;
Fish:
Amphibians:
Frogs
Toads
Salamanders
Reptiles:
Turtles
Birds (embryonated eggs):
Chicken
Quail
Pheasant
Mammals:
Marsupials:
Opossum
Edentates:
Armadillo
Rodents:
Mouse
Rat
Guinea pig
Hamster
Lagomorphs:
Rabbit
Hare
Carnivores:
Mink
Ferret
Cat
Dog
Ungulates:
Pig (regular, miniature)
Goat
Sheep
Cattle
Nonhuman primates:
Prosimian:
Bushbaby
Simian:
Monkeys (macaques, marmoset)
Apes (baboon)
139
75, 76, 169
.15, *5
11, 67,
50, 169, 17*
91, 116, 147, 17*
*3, 61
96
5, 13, 16, 3*. 115, 161, 169
22
135
69, 169
5, 13, 33, 69, 115,
169
5, 11, 38, 69, 122, 165, 173
5, 18, 69, 122, 153, 163,
165, 173
5, 13, 31, 60, 69, 169
5, 13, 19, 31, 69, 163, 165
5, 11, 69, 122, 163, 165, 173
165
5, 27
5, 99, 139, 163, 165,
5, 69, 81, 115, 153, 163, 173
5, 69, 115, 153, 163, 165, 173
5, 69, 115, 121, 163, 165, 173
69
5, 69, 110, 115, 163
69
5, 110, 153, 173
5, 69, 153, 163, 165, 169, 173
5, 58, 69, 163, 165
14
-------
Country
Table 3
Animals Recommended or Required by Various Countries
for Drug Teratogenicity Testing
Animal
USA (FDA)
Canada "
Great Britain
France
Japan
Sweden
Reference
At least 2 species. Mouse, rat, 78, 153
rabbit, acceptable.
At least 2 species (one a non- 5
rodent). Mouse, rat, hamster,
rabbit.
Rat and rabbit. 122, 153
Three species. Rat, mouse and . 153
rabbit.
Two species. Rat, mouse, rabbit used. 115
Two species. 34
Switzerland
Rat, mouse, rabbit.
115
Table H
A New Concept in Teratogenicity Testing Based on
Multilevel Tests in Different Types of Animals (165)
Order
of test8
First
level
Second
level
Third
level
Fourth
level
Purpose
Find embryotoxic dose range
Confirm or adjust above
Only if second level results
are equivocal
Only if use in human
pregnancy needed or likely
No. of
Suitable Pregnant
species Animals
Rat, mouse, hamster
or rabbit
A carnivore or an
ungulate
Alternate to that
used in second level
Macaque monkey or
baboon
130-150
40-60
40-60
40-50
would terminate at second or third levels in most instances.
Source: Environment and Birth Defects by J.G. Wilson, reprinted with
permission of the Academic Press. Year of first publication, 1973-
15
-------
sensitivity to teratogens (rat), absence of- pure strains (rabbit) high cost,
and limited availability and breeding capabilities (dog, monkey). Although
the monkey was an excellent predictive teratologieal model for thalidpmide the
use of this animal for testing other drugs (e.g aspirin, aminopterin) gave
disappointing results and the response differed greatly from that of man.
According to a WHO Scientific Group (173), all substances known to be
teratogenic for man have shown some teratogenicity in the rat, mouse, or
rabbit; yet "there is no absolute assurance" that negative tests in these
species will predict absence of teratogenicity for man, or that substances
that are teratogenic at high doses in these species will be teratcgenic for
man at lower doses.
Therefore neither current knowledge of physiology nor teratologieal
response patterns give secure grounds for preferring any one species of mammal
for predictive .screening for potential teratogenicity for man. Nor are there
grounds for believing that two or more species would be more predictive than
two or more sufficiently divergent strains of one species. Under these
circumstances other considerations become important, such as aspects of
husbandry, reproduction rates to generate "significant" numbers of offspring,
and cost-risk-benefit factors.
16
-------
Species
Tabl<* 5
Comparative Thalidomide Teratogenicity (69)
(mg/kg/day)
Smallest Dose Largest Dose
Producing Producing
Defects No Defects
Man
Baboon
Monkey, Cynomolgus
Rabbit
Mouse
Rat
Armadillo
Dog
Hamster
Cat
Source : Reprinted
0.5-1.0
5
10
30
31
50
100
100
350
from Teratology of the Central
7
50
nooo
nooo
200
8000
500
Nervous System by H.
Kalter, reprinted with permission of the University of Chicago Press. Year
of first publication, 1968.
Table 6
Thalidomide Teratogenesis in Primates (128)
Species
Man
Cynomologous
Monkey
Baboon
Rhesus Monkey
Teratogenic
dose mg/kg
oral route
0.5-1
10
5
12-19
Gestation
days treated
20-36
22-32
18-M4
2M-26; 27
or 30
Defects
Limbs (80$); ear (20$)
Limbs (67$);
teratomas (33$)
Limbs and tail (U0$)
Limbs (100$)
17
-------
Table 6 (Gont'd)
Thalidomide Teratogenesis in Primates (128)
Species
Bushbaby
Japanese Monkey
Stump-tailed
Monkey
Marmoset
Bonnet Monkey
Tera to genie
dose mg/kg
oral route
20
20
5-10
15
5-30
Gestation
days treated
16-30
21-26
21-30
25-35
21-29 or
11-11
Defects
Limbs (100?); tail
(17$); CNS (17$)
Limbs (100$); tail
Limbs, ear and jaw
(100$)
(20$)
Limbs (15$); Visceral (52$)
Source: Reprinted from Drugs as Teratogens, by J.L. Schardein, CRC Press, 1976.
Table 7
Percent Malformations Produced in Four Inbred Strains
of Mice by Galactoflavin (69)
Strain
A/J
DBA/1
129
C57BL/6
Source:
Cleft
Palate
3
11
8
13
Reprinted
Skeletal
Defects
38
92
81
0
from Teratology
Brain
Defects
55
11
9
5
of the Central
Atresia of
Esophagus
78
83
30
2
Nervous System by H
reprinted with permission of the University of Chicago Press. Year of first
publication, 1968.
18
-------
Table 8'
Range of Spontaneous Malformations in Animals
Species %_ Incidence Reported
Mouse 0.4-18.6
Rat 0.02-0.85
Rabbit 0.74-6.3
Dog (Beagle) 0.17-1.9
Monkey 0.11-16.8
Cat 1.24
Sheep 0.2-0.3
Pig 0.6-9.8
Source: Data abstracted from reference 128 (Tables 2-10).
2. Administration of Chemical
Dosage levels, routes of administration, durations of treatment, vehicles
and other control aspects are all considered together in designing a test.
For instance, in screening chemicals for potential teratogenicity, there is
debate about whether to give well-timed large doses to uncover any propensity
to cause malformations under test conditions, or to give chronic treatment
with small doses that approximate the usual conditions of human exposure (39,
165). Both sorts of information may be required, and since environmental
"accidents" can cause occasional acute exposures, most initial screening
seeks to uncover any potential teratogenicity. In this subsection of the
report, the parameters of dosage are examined one by one.
a. Dosage level. Four major overlapping effects of dose level have been
described under-test conditions: lethal to the mother, lethal to the
conceptus, teratogenic, and no-effect ((165), and see Figure 1). The terato-
genic range is looked for empirically (69), either from above (13), or from
below.
19
-------
A commonly preferred reference point is'the maternal LD_Q (39» 165),
normally established by acute toxicity tests before a chemical is screened
for potential teratogenicity. However, since in most instances there is
great disparity between the maternal toxic dose and the teratogenic dose
(Table 9) (18), the embryo usually has a higher susceptibility than the
adult. This is due to a particular vulnerability of certain embryonic cells
which are absent in the adult (153). Drugs with the greatest difference
between the two values present the greatest hazard.
Table 9
Teratogenicity Toxicity Relationships for Some Drugs in the Rat (18)
Approximate Minimal
Drug
Cyclophosphamide
Uracil mustard
Chlorambucil
Nitrogen mustard
Triethylene melamine
Triethylene thiophoaphoramide
Busulfan
6-Mercaptopurine riboside
6-Mercaptopurine-3N-oxide
6-hydroxylamino purine
5-Fluoro- 2 ' -deoxyuridine
5-Bromo-2 ' -deoxyuridine
Cytosine arabinoside
Aminonicot inami de
Hadacidin
Procarbazine
LD5Q (1P) Teratogenic Ratio
Dose (IP)
40
1,25
24
2.0
1.25
8.0
60
2000
200
800
1500
1500
3000
15
5000
400
7
0.3
8
0.5
0.3
3.0
. 18-
15
50
400
0.15
100
20
5
3500
25
0.17
0.24
0.33
0.25
0.24
0.37
0.3
0.007
0.25
0.5
0.0007
0.06
0.006
0.33
0.7
0.06
a) Data abstracted from Reference 18.
b) i.p., intraperitoneal
20
-------
When the LD__ data refer to another strain of animal or to different
test conditions, either the appropriate LD_0 should be ascertained, or the
figures used should be properly defined. The dosage plan is expressed as
fractions of the LD_Q; there is usually a threshold for lethality, and the
range LD. through LD-- is relatively narrow, so that (for example) the
LD._ will hardly ever be one-half of the LD Therefore, percentages of
the LD other than LD^- are not used. However, the MTD (median toxic dose)
and ,ED_- data are used by some workers as reference points (173).
Dose measurements for teratogenicity testing are usually based on body
weight (mg/kg) (173). Little discussion was found on whether this is really
the best basis for evaluating exposures to environmental chemicals that are
not ingested. For example, specific skin surface area (cm /kg body weight)
might conceivably be a better basis of measurement in view of the large
differences in this ratio between most test animals and man (11, 120). But
data have not been expressed on this basis in most studies reviewed (112).
The LD5Q is obtained as an acute toxicity measurement (see above).
Usually a test of chronic LD_Q would give different data (lower or higher)
but the test would contain more variables, and the amount of time that would
be required to produce the chronic effects would vary from chemical to
chemical. Chronic toxicity^ measurement might be more appropriate for some
environmental chemicals, but not for others (71, 167).
In any case, the basis of measurement is only a starting point, and the
dose level actually found to reveal the embryotoxic threshold Is used for
setting of safe tolerance levels for the pregnant animal (71, 167).
To screen a chemical for potential teratogenicity, Wilson (165) recommends
starting with three dose levels: 0.5, 0.25, and 0.125 of the maternal
LD5Q. If these do not reveal the embryotoxic range, other fractions (binary
system, usually smaller) are added. Three levels of dosage, determined
empirically or by some arbitrary formula, are recommended or required in
Canada (5), Great Britain (122), France (153), and by the WHO (173). Two
levels are required in Sweden (31). The United States FDA requires two levels
21
-------
for single-generation tests for teratogenic'ity of drugs, and three for multi-
generation studies (35, 78). Multigeneration studies are not relevant to the
present survey and may obscure the assessment of post-implantation terato-
genicity.
b. Route of administration. An accepted principle is that the test
substance should be administered to the animals by the route that most closely
resembles that of the anticipated human exposure (5, 71, 165, 167). There may
be exceptions when differences of physiology affect distribution of the test
substance.
Thus test substances are given by mouth, by inhalation, by intranasal
instillation, by skin application, etc. (5). Occasionally, to facilitate
exposure of the test animal, the principle is ignored, and the substance is
injected parenterally. In some cases the route of Administration has
influenced the degree of teratogenic response (16).
In all cases the route will affect some aspect of bioavailability: for
instance, its rate, the available proportion of the dose, the available
chemical form of the test substance, the relative systemic exposures of the
body tissues including the conceptus, or the durations of such exposures. In
practice, the principle of resemblance is often breached to make sure that the
test animals receive measured doses. Another breach of the principle is
administration by a single route when human exposures can be anticipated by
more than one route. Every such addition of certainty to the test situation
detracts from its interpretability in terms of anticipated human exposures and
of their potential teratogenic effects.
These difficulties involve each of the usual routes of administration of
test substances:
(1) Oral doses are given preferably by stomach-tube (gavage). Other
methods, such as in capsules, in drinking-water, or mixed with diet.
either introduce uncertainties into the records or, by altering
appetite, may affect actual'intakes (5, 69,' 165, 169, 173). Then,
22
-------
too, animals are commonly fasted before the dose (21); this itself
may be embryotoxic or may alter a teratogenic effect of a test
substance.
(2) Air pollutants and anesthetics have been tested by inhalation (17,
88). Direct "environmental exposure" of rats on the property of a
synthetic rubber factory has been reported (97). But amounts inhaled
have been hard to measure reliably. When air concentrations of a
test substance are increased, the animals may respond by breathing
less deeply or less often. So, reliable dose-response curves are
hard to obtain. Intranasal instillation does not avoid this diffi-
culty.
(3) Skin applications must be protected from molestation by the test
animals if results are to be valid (165). Molestation transfers some
of the dose to the oral route. But a small test animal has many
times the specific skin surface area of a woman. Clipped or shaved
skin may respond differently than bare human skin, and the animal's
thermoregulation may be disturbed. The smaller the prepared area,
the more concentrated the dose must be. These considerations limit
the interpretability of test doses given to animals by this route.
(1) Ocular instillation is used to test eye medications (115). Some;
air-borne environmental chemicals may be suitable for testing by this
route, but lacrimation would introduce uncertainties.
(5) The parenteral routes are mainly subcutaneous, intramuscular, and
intravenous. Intraperitoneal injection is sometimes used as an
alternative to gavage, but microorganisms, pH, enzymes and other
conditions of the gut lumen are bypassed. Intra-amniotic,
intrauterine and yolk sac injections, and placing of drug treated
millipore filter on the amnion or the placenta are also current
research procedures (26, 128); these bypass the metabolic systems of
the mother and the placenta, and could give misleading information if
used to evaluate environmental hazards. Also, the procedures in
themselves can be teratogenic (80).
23
-------
It may be noted that none of these routes adequately reproduce the condi-
tions of uptake by man of an environmental chemical, when a reliable terato-
genic dose-response curve has to be obtained for test animals. The question
raised by Kalter, whether it is logical to make the attempt, seems to be
justified; this survey does not answer it. It may be that any route that
produces equivalent post-hepatic plasma levels of the test substance or an
appropriate metabolite in the female animal could satisfy criteria for
validity of the test. But evidence is lacking, and meanwhile, the principle
of approximation is accepted.
c. Duration of treatment. Human exposure to environmental chemicals
varies from continuous at low levels to intermittent at high levels.
Continuous exposure of a human female will usually have started long
before her pregnancy. But, for example, a newly married female may move into
an exposure area and become pregnant immediately. Continuous exposure may
produce either cumulative effects or, if metabolic tolerance develops, reduced
effects.
Cumulative effects tend to prolong exposure of the conceptus beyond the
period of maternal exposure. Metabolic tolerance, on the other hand, tends to
reduce the period of exposure of the conceptus. An investigator may be able
to predict, which sort of effect is more likely from preliminary test data on
metabolism of the substance, especially in pregnant females.
Intermittent high-level exposure of a pregnant female may, depending on
how she metabolizes the substance, deliver a short, sharp challenge to the
conceptus, or a more prolonged challenge at a lower level if, for instance,
the substance is rapidly stored in body fat.
If human exposure is expected to be low-level, continuous, and everywhere,
without episodes of higher exposure, a teratogenicity screening test may not
need to include short-term exposures of test animals. On the other hand, at
least one group of test animals should probably always be screened for cumula-
tive effects by continuously administering the substance throughout pregnancy.
24
-------
In general, however, the uncertainties suggest that durations of treatment
with a substance should rarely be based on expected human exposures to the
environment, but rather should be directed to revealing thresholds of sensi-
tivity at its greatest in the test animals.
Most protocols and guidelines for teratogenicity tests (69, 163, 165),
particularly those concerned with regulation, specify that the substance shall
be administered to the test female throughout the period of major organo-
genesis (Tables 10, 11 and 12). This time period, the "critical period of
organogenesis" differs among the species; it is in part dependent on the
duration of gestation and is the period during which the embryo is highly
susceptible to teratogenic insult. During the predifferentiation. period in
early gestation and following differentiation, the conceptus is generally
resistant to production of congenital malformations although embryonic death
and/or abortion may occur. Sometimes teratogenic treatment reveals more than
one period of maximum effect (Table 12).
Administration of drugs that dissolve and are absorbed, metabolized and
excreted rapidly can be confined to the "critical periods." But drugs that
are absorbed or metabolized more slowly will have effects that extend beyond
this period, and result in apparent absence of teratogenic effect if tested
only during the critical periods. This factor represents one of the most
common pitfalls in teratogenic testing and accounts for the observation of
teratogenic activity outside the critical period. Drugs acting in this manner
(actinomycin D, cyclophosphamide, striptonigrin) generally require activation
in situ for a period of time before they can be effective (18).
Wilson (163. 165) points out that continuous treatment during organo-
genesis may produce maternal adaptations that decrease the effect on the
conceptus (Tables 13 and 1U). He suggests the use of several treatment
spans: one group to be treated continuously to reveal any cumulative effects,
other groups to be treated for 3-** days at a time to avoid adaptations of
maternal enzyme systems (Table 15). Table 16 shows an extension of this idea
that is comprehensive and beyond the scope of this report.
25
-------
To avoid attribution of any pre-implantation defects to the teat
substance,- the start of pregnancy is observed and recorded. For example, in
rats and mice a vaginal plug or sperm in the vaginal smear can be seen after
mating. The morning this evidence is first seen marks the first day of preg-
nancy. The sensitive period of major organogenesis is then worked out, using
tables.
Other periods of sensitivity to teratogenesis have received less attention
in the past, but may prove more important in the future. One of these is
late-stage teratology of the CNS (69). Recently Langman and his associates
have been developing information on the timing and morphology of sensitivity
in specific areas of the CS (87, 123, 136). However, protocols for the
routine screening of environmental chemicals for late-stage CNS teratology
have not yet been fully developed.
d. Dose vehicle. The vehicle is more important than is often supposed.
Arguments that its properties do not matter so long as control and treatment
groups all receive like amounts have been shown to be untenable (5).
The Canadian guidelines for teratogenicity testing (5) discuss the vehicle
at length. The major requirements can be summarized as follows:
o No alteration of the chemical properties of the test compound.
o No serious influence on absorption, distribution, metabolism or
retention of the test compound.
o No increase, decrease, or other alteration of toxic properties of
the test compound.
o No local or systemic properties of its own.
o No effect on consumption or utilization of foods or drinking-water.
o No interactions with other conditions of the proposed test.
o If possible, the vehicle should be comparable with that to which
humans are expected to be exposed in practice.
If possible, the vehicle should dissolve the test substance.
26
-------
Table 10
Episodes in'the Reproductive Cycle of Mammals (163)
1. Preservation of the germinal line
2. Gametogenesis
3. Release and transport of gametes
4. Fertilization
5. Cleavage and blastocyst stages
6. Implantation
7. Metabolic changes in pregnant dam
8. Embroyonic period - organogenesis
9. Fetal period - histogenesis, growth and beginning
functional maturation
10. Placenta, maternal-conceptal relations
11. Birth and postnatal adjustments
12. Lactation and maternal care
13. Postnatal growth, functional maturation
Source: Reprinted from "Critique of Current Methods for Teratogenicity
Testing and Suggestions for Their Improvement" by J.G. Wilson, in Methods for
Detection of Environmental Agents That Produce Congenital Defects, T.H.
Shepard, J.R. Miller and M. Marols, Eds., by permission of Excerpta Medica -
North Holland Publishing Company. Year of first publication, 1975.
27
J
-------
Table 11 '
Gestation Times for Some Laboratory Animals and Man (128)
Species Mean duration of gestation days
Chick
Rat
Hamster (golden)
Mouse
Rabbit
Ferret
Cat
Dog
Guinea pig
Pig
Sheep
Monkey ( rhesus )
Monkey (baboon)
Armadillo i
Human
21
21
16
19
31
H3
63
..63
68
114
150
168
175
225
278
28
-------
Table 12
Critical Periods of Organogenesis in Animals (128)
Species
Critical period days
Chick
Rat
Hamster (golden)
Mouse
Rabbit
Ferret
Cat
1-3"
9-17
4-14
7-16
8-21
8-28
5-58; 5-15
most favorable
Dog
Guinea pig
Pig
Sheep
Monkey (rhesus) f.
Monkey (baboon)
Armadillo
Human
1-48; 8-20
estimated
11-20
12-34
14-36
20-65
most
22-47
1-30
20-55
; 22-30
susceptible
Period of embryological organogenesis or period of known
susceptibility to teratogens.
Taken from reference 128.
29
-------
Table 13
Some Chemical Agents Known to Influence Rates
of Metabolic Degradation of Themselves and/or
Other Repeated Dosage* (165)
Increase
metabolic degradation
Decrease
metabolic degradation
Barbiturates
Thyroxine
Some insecticides (DDT, chlordane, aldrin,
dieldrin, heptachlor)
Some tranquilizers and antipsychotics
(meprobamate, Librium, chlorpromazine)
Some antihistamines (chlorcyclizine,
diphenhydramine)
Several hypcglycemic agents
3,^-Eenzypyrene
3-Methylchclanthrene
Steroid hormones
SKF 525 -A
Chlorthione
Iproniazid
Metopirone
Actinomycin D
Puromycin
Triparanol
Chloramphenicol
Any that competitively., inhibit
catabolic enzymes
From various sources.
Source: Reprinted from Environment and Birth Defects by J.G. Wilson, by permission of
the Academic Press. Year of first publication, 1973.
30
-------
.Table 1U
Ways ,in Which Repeated Treatment Prior to the
Peak Susceptible Period of the Embryo May Produce
Misleading Results (165.)
Time of treatment
Primary effect
Secondary effect capable of
altering test results
1. Before implantation
2. Early organogenesis
3. Before peak
susceptibility
U. Before peak
susceptibility
5. Before peak
susceptibility
6. Before peak
susceptibility
7. Before peak
susceptibility
Interference with
implantation
Early embryonic death
Induction of cataboliz-
ing enzymes
Inhibition of cataboliz-
ing enzymes
Liver pathology or re-
duced function
Kidney pathology or
reduced function
Saturation of protein-
binding sites
No issue
No issue
Reduced blood level during
susceptible period
Increased blood level during
susceptible period
Increased blood level during
susceptible period
Increased blood level during
susceptible period
Increased blood level during
susceptible period
Source: Reprinted from Environment and Birth Defects by J.G. Wilson,
by permission of Academic Press. Year of first publication, 1973-
31"
-------
Table 15
Numbers of Pregnant Rats Treated At
Consecutive 3-day Periods (165)
Dose
level
xa
X/2
X/1
9-11
10
10
' 10
Gestation Days
12-14
10
10
10
15-17
10
10
10
X - highest tolerated dose over 10-day period, or
half of adult LD_Q, or other appropriate effect-level.
Source: Reprinted from Environment and Birth Defects by J.G. Wilson,
by permission of the Academic Press. Year of first publication, 1973-
Table 16
Types of Test for Developmental Abnormality
Depending on Duration of Treatment (163)
1. Throughout reproductive cycle - whole generation(s)..
2. Throughout pregnancy - conception to term.
3. Throughout organogenesis, primitive node to palatal closure.
4. Short-term (3-4 d) sequences during organogenesis.
5. Aimed at specific events - mutagenesis, postnatal, etc.
6. Combinations of above, e.g., 3, 4, and 5.
Source: Reprinted from "Critique of Current Methods for Teatogenicity
Testing
and Suggestions for Their Improvement" by J.G. Wilson, in Methods
for Detection of Environmental Agents That Produce Congenital Defects,
T.H. Shepard, J.R. Miller and M. Marois, eds., by permission of
Excerpta Medica - North Holland Publishing Company. Year of first
publication, 1975.
1
32
-------
Instances are cited (5) of seemingly unimportant properties of vehicles that
made test data "difficult if not impossible" to interpret. Dimethyl sulfoxide
enhanced penetration of some test substances through the skin and altered
their distributions in the body. Even the provision of distilled drinking-
water instead of deionized or tap water affected litter size, weight gain, and
water-consumption of pregnant rats.
Information may be lacking on the extent to which a vehicle satisfied
these requirements in the quantities to be used during a teratogenicity test.
In cases of doubt, a vehicle should be evaluated separately against a group of
completely untreated controls under the conditions of the proposed test (5).
Others have noted that a vehicle should maintain the physical state,
concentration, and other requirements of exposure to the test compound, and if
water is used, it should have no contaminants such as chlorine or phenols (71,
167).
e. Controls. The purpose of all controls in teratogenicity testing is to
help the investigator answer the question, Were these malformations that I
have observed caused by the test substance, or not? Often this question is
difficult or impossible to answer, however good the controls.
*
One reason for the difficulty is the range of normal variants and
"spontaneous" defects found in man and most test animals, from which induced
defects need to be distinguished. As Wilson points out (165), "clearcut
embryotoxicity.,. is readily recognized but clear-cut cases are by no means
the rule" because most highly embryotoxic agents will have been eliminated
before the stage of testing for teratogenicity.
A more important reason for good controls, according to Kalter, is the
possibility that unknown variables in the test animals or their environment
may modify the teratogenicity of the test substance, making it more or less
teratogenic, or making the animals more or less sensitive, or both.
33
-------
Therefore, the most useful control is the investigator's background
knowledge of the strain of animal employed - the "historical" control group
that comprises the accumulated data on all defects observed in the past in
untreated animals of that particular strain (69, 71, 167). If, for example,
anophthalmia had been observed at the rate of 1 in 200 offspring, and no cases
of spina bifida had been seen, in the historical controls, and then if one
case of each were to be found in a treated group, those two cases would be
interpreted differently (165).
The quantity of the historical control group is an important factor in
selecting the species and strain of test animal. Usually, a species or strain
with adequate historical controls will be preferred over one that seems
developmentally more suitable but lacks historical controls. This, for
example, is one reason why mice rather than guinea pigs have been used for
late-stage CNS teratogenicity tests.
Equally necessary are concurrent "negative" controls, groups of animals
that are similar to the test groups and are treated similarly but without the
test substance. Concurrent negative controls have three functions:
o They are compared to groups treated with the
test substance.
, o They are compared to the historical control in
order to detect the presence of unexpected
variables.
o They contribute to the accumulated historical
control for future tests.
All concurrent control groups are matched meticulously with test groups for:
o Number of animals in group.
o Characteristics of the animals (age, size or
weight, history, health)
o Test environment (all aspects of husbandry,
amount and times of handling, and identity
of handler).
34
-------
The purpose is to reduce the known variables to one: the differences that
need to be detected and measured.
The treatment of a test group will in part determine that of the con-
current negative control group, and may itself be partly determined by control
requirements.
An important instance is attention to the test vehicle (5, 71, 165, 173).
If a concurrent negative control group is to be compared with a group
receiving the test substance, the controls will receive the vehicle comparably
with the treated group as to amounts, times, route(s), and all other
particulars. If omission of the test substance alters the quantity or
character (e.g., volume or caloric density) of what is administered, the
change is not rectified by automatically increasing the amount of vehicle, but
alternatives are considered (e.g., a known inert placebo).
If activity of the vehicle is in question, the same principle applies, and
an untreated control group is compared to the vehicle treated group which now
acts as a test group. The concurrently untreated group is then compared with
the historical control. In this format there are essentially three tiers of
comparisons:
o Substance-plus-vehicle versus vehicle alone.
o Vehicle versus no treatment.
o No treatment versus historical control.
This set of comparisons can be expanded or condensed according to need, so
long as the need is understood to be for maximum validity.
As for numbers of animals per group, the practice is to use 20 mice or
rats, or 10 rabbits, at each dose level when the test substance is given
throughout organogenesis. These numbers are halved for short treatments given
at intervals during organogenesis or gestation (see Table 14). Recommenda-
tions for numbers of more expensive species vary (165).
35
-------
In the past, groups of animals designated as concurrent positive controls
used to be given doses of some compound that for them was a known teratogen.
This was considered to indicate how sensitive the particular strain of animals
was to teratogenic influences in general. The degree of sensitivity to a
known teratogen is no sure guide to the degree of sensitivity to a different
test substance.
3. The- Test Environment
Many aspects of the test environment have been discovered to affect the
development of the conceptus and its sensitivity to teratogens. They will be
considered under the headings (112) of (a) diet, (b) caging, (c) climate, (d)
stresses. When test conditions vary among laboratories, or from time to time
in a laboratory, test results become hard to interpret. At present, more
constancy and uniformity is needed. This implies high standards of animal
, >
care and laboratory records (5).
a* Dietary factors that can cause defects or modify effects of teratogens
include:
o Inappropriate feeding times, e.g., fasting (5, 69, 165).
o Composition of drinking-water (71, 167).
o Balance of the major nutrients (5, 118, 165).
o Composition of chows can vary among sources and among batches
from any source, affecting palatability and intake; thus
composition is often a trade secret. Knowledge is needed,
in order to maintain constancy.
o Decomposition of diets, especially chows, with age or spoilage.
o Contamination with endogenous alkaloids (69, 77), with
incorporated antibiotics (18, 71), with pesticides, dirt, and
excreta, or with microorganisms.
o Serious deficiencies or excesses of vitamins (5, 46, 69, 159,
165).
o Mineral deficiencies, e.g., Zn, Mg, Mn, I, Cu, (16, 63, 71,
159, 165) or excesses, e.g., I, Hg, Pb (1, 16),.
36
-------
Reviews are available of nutrients known to affect development of the concep-
tus (62), of methods of evaluating animal feeds (1, 30), and of the nutri-
tional requirements of certain test animals (3, ^6). It is very important to
control the quality of the diet. For example a 30J reduction in food intake
in mice produces effects on the teratological outcome of the results, while a
50/6 reduction is required in the rat to produce similar results (9).
o
b. Caging is mainly a matter of quality (5, 173); a poor surface can
harbor microorganisms or yield variable amounts of unidentified minerals.
Bedding is sometimes eaten or chewed, and if treated, it may contain
substances that can affect test results, e.g., ethylene oxide, organochlorine
compounds.
c. Climate includes many potentially teratogenic factors (112):
o Extremes of temperature, whether of heat or cold, can produce
malformations or potentiate other teratogenic influences (5, 25, 32,
69, 107, 112, 165). Drafts are to be avoided; the temperature should
remain even throughout an animal room.
o Relative humidity should be constant or recorded, extremes or sudden
change can stress test animals or affect the behavior of pathogens.
o Extremes of high or low atmospheric pressure or oxygen tension can be
teratogenic (5, 69).
o Light should be sufficient, of appropriate spectra quality, and evenly
distributed. Light cycles must be regular for reproducible results
'(165), and they can affect determinations of gestational age (69). If
cycles cannot be controlled, there may be seasonal differences of
reproduction (173).
o Some test animal species have shown seasonal sensitivities to
teratogens (69, 165).
o Any possible source of harmful radiations should, of course, be
eliminated (69, 165).
Sudden or excessive noise should be avoided or recorded (5, 117).
Environmental odors can affect food intake and health of test animals.
d. Stresses of many kinds can cause or enhance teratogenicity (112):
37
-------
o Overcrowding (69, 173), or immobilization (1U), or excessive solitude.
o Odor of a strange animal (15).
o Physical trauma (165),
o Excessive, clumsy, or unexpected handling by laboratory staff (5, 78)..
o Pathogenic microorganisms and parasites (5, 1^).
o Medications (69, 165), and chemicals applied to the animal rooms (165).
o Psychological stresses (112) which often are hard to define but
include failures to follow routine.
The above list is not exhaustive. Usually all possible sources of stress
cannot be totally eliminated , but they can be recognized, recorded, and -
most important - unequal incidence among the animal groups can be avoided.
4. Observations
The four signs of embryopathy are death, growth retardation, malformation,
and functional deficit. Malformations may reflect cytotoxicity or brief,
local episodes of retarded growth; the observed categories are external,
skeletal, and visceral malformations.
To look for all of these signs, procedures must be timed carefully, so
timing will be considered before methods for observation. But the first
question, because it affects the approach to observation, is whether any con-
dition of pregnant females can be teratogenic of itself.
a. Maternal effects. Can a conceptus be malformed indirectly, as a
result of toxicity to a pregnant female, when no toxic substance or metabolite
reaches the conceptus? The answer seems to be qualifiedly negative (165).
According to Wilson, no examples of this have been seen in test animals,
although some observations in man have been interpreted as secondary malforma-
tions. Secondary fetal deaths have been reported, for example, after sero-
tonin was given to pregnant rats, but there were no malformations nor poten-
tiation of teratogenicity of other substances (51).
38
-------
If a pregnant female receives a diet deficient in certain respects, this
can produce a malformed conceptus, but such "teratogens" are not toxic
substances or metabolites. Nutritional antagonists have been hypothesized
but, according to Wilson, have not been demonstrated. Malformations
attributable to position effects in multigravid uteri are considered marginal
or trivial.
So, according to Kalter and Wilson, secondary embryopathy results either
in death of the conceptus or in retarded growth observed at birth and
reversed, usually, by proper feeding of the neonate. Teratological effects
would not be reversible. When a pregnant female fails to gain weight at the
expected rate, secondary embryopathy can be expected.
Another important factor is the age and parity of the animal. In poly-
tocous animals such as rats, rabbits, mice and hamsters the reproductive
capacity has been shown to decline with increasing age. In mice the highest
rate of spontaneous malformations occurs in the first litter and then
decreases rather significantly by the fourth litter. In rats and rabbits, on
the other hand, the size of the litter decreases with age, but the incidence
of spontaneous malformations is not affected by age and parity.
b. Timing. Pregnant ^females are watched throughout gestation for weight
gain and for signs of ill health.
As a rule, the females are killed a few hours before expected delivery to
avoid cannibalism of stillborn, moribund, or malformed offspring, and to count
fetal deaths and resorptions (5, 11, 35, 78, 153, 165).
When postnatal observations on neonates or older offspring are needed,
some females are allowed to litter normally (5, 35, 78, 153, 165). Cross-
fostering may be needed if the treatment has interfered, for example, with
lactation.
39
-------
One exception has been reported to the general rule of looking for birth
defects at or soon after birth. Neonatal CNS damage from absorption of free
acidic amino acids is phagocytized rapidly, within about 24 hours (12, 89,
114), and animals are killed 2-5 hours after treatment (113). Such damage can
occur transplacentally in late pregnancy (Olney, personal communication 197*0,
and examinations can be timed accordingly if this is suspected.
c. Preparations. The pregnant female is killed by any means that does
not injure the uterus or its contents. Unusual methods, e.g., perfusion-
fixation, are employed when specially needed.
The abdomen is opened; the uterus is inspected _in situ for total implanta-
tion sites including all states of resorption. Corpora lutea are counted and
compared with the total of implantation sites. Faulty implantation tends to
scar if it survives for 24 hours or more. By these or other suitable methods
(depending on the expected accuracy for the species)'/ preimplantation losses
are estimated.
The uterus is opened; living fetuses are distinguished from intact dead
ones by color, movement, etc., and macerated ones are counted as resporticns.
In the living fetuses, the usual vital signs are assessed, and the umbilical
cords are clamped or cauterized. After preliminary examination, these fetuses
are blotted, weighed, sexed, and prepared for detailed examination.
i
Fixation in Bouin's fluid is recommended by Kalter for animals to be
sectioned either by Wilson's razor-blade method or by microtome. Sectioned
tissues are stained appropriately. The head is often detached from the body
and sectioned coronally, while the body is sectioned transversely. If there
is reason to suspect selective teratology of a particular system, e.g.,
nervous system, special preparations are made at this stage for appropriate
histology.
d. External malformations. The preliminary examination is by eye and, if
necessary, low-power microscopy as described in detail by Wilson (161).
Information on what to look for in various test species has been given by
40
-------
Warkany (158), Kalter (69), and Wilson (164, 165, 169), and in man by Potts
(119). Gross structural features are inspected, and any deviations of propor-
tion are measured and recorded. This examination is conducted by a scientist
with in-depth, relevant knowledge of the species employed.
e. Skeletal malformations. One-third to one-half of the offspring are
fixed in 95$ alcohol for skeletal visualization by a technique using the
alizarin red S stain'(14, 23, 24, 38, 74, 129, 165). The method of Schnell
and Newberne is preferred (129» 165, and see Figure 2). But alizarin stains
only ossified bone, and rodent skeletons will still be largely cartilaginous.
Methylene blue may be used to stain rodent cartilage, and the Noback modifica-
tion of Van Wijhe's method is recommended (111). Large animals have been
examined by radiography (165, and see Figure 3) but good visualization may
pose problems. The examiner should distinguish major or harmful malformations
from normal skeletal variations (71, 167).
f. Soft-tissue malformations. Large animals, e.g., the 100-day monkey
fetus, are examined by standard autopsy procedures. Any local or general
growth retardation is revealed when organ measurements are compared with
normal values (see Table 17) (165). Small animal offspring not used for
skeletal examination are fixed in Bouin's fluid or other nonhardening fixative
*
that decalcifies the skeleton. They are dissected by standard methods to
reveal soft-tissue abnormalities. Wilson's freehand razor-blade technique
(169) is more often used, both for regulatory tests and for research (24, 35,
38, 78, 151, 173). Some examiners section by microtome, or di.ssect the major
viscera individually after fixation. A portion of liver is usually examined
separately. Slices of tissues are compared with reference sections or with,
e.g., Wilson's Atlas of Sections.
If there is reason to suspect that a substance may selectively damage a
particular system, such as the nervous system, the examiner chooses an appro-
priate technique that will reveal histological changes - for example, serial
sectioning. If special fixation is needed, or special times of observation,
for example a need to kill test animals by perfusion-fixation, a sufficient
additional group of animals is reserved for this.
41
-------
Table 17
Weights and Measures on 15 Control Monkey Fetuses Removed by Hysterotomy on
Gestation Days 99-101 From Untreated Rhesus Females (165).
Body Amniotic
Weight fluid
Fetus Sex (gro) (ml)
13d
16e
29d
41c
40e
47c
48c
60c
- 75b
39da
7 Ob
94a
98a
77c
58f
Mean
Standard
deviation
F
M
M
M
M
F
F
F
M
M
M
M
F
F
M
133
154
149
151
148
119
142
130
125
143
159
146
155
154
143
143
12
The mothers of this and
days 20-45 of gestations
Source:
Reprinted
from
70
135
150
125
130
95
84
105
85
85
71
93
86
84
47
96
28
the following 5
Environment and
Brain
20
22
23
21
15
25
21
18
23
22
21
21
.6
.5
.4
-
.1
.3
.4
.0
.0
.1
.2
.0
.1
fetuses were given
Birth Defects by J
Weights of major
Eyes Lungs Heart
1.10
1.02
0.96
0.96
1.16
0.92
0.96
0.98
0.90
1.06
1.23
1 . 08
0.93
1.11
1.05
1.03
0.10
a vechicle
. G. Wilson
2.4
2.6
2.6
3.5
2.8
3.0
2.7
3.1
2.5
2.7
3.2
3.0
3.5
2>9
2.2
2.8
0.4
, 15
, by
0.80
1.00
1.10
1.10
1.00
0.70
1.08
0.90
1.33
1.10
1.30
1.26
1.50
1.09
1.08
1.09
0.20
ml of 0.3%
permission
viscera (gm)
Liver
6.3
6.1
5.2
5.4
5.3
4.5
5.8
4.5
4.8
-
7.0
6.1
6.3
5.5
5.1
5.2
0.8
tragacanth daily
of the Academic
Kidneys
1.29
0.86
1.25
1.12
1.06
0.78
0.90
0.97
0.91
0.96
0.98
0.97
0.91
0.80
0.82
0.97
0.15
by stomach
Press.
Spleen
0.18
0.22
0.33
0.40
0.34
0.26
0.24
0.30
0.25
0.20
0.35
0.35
0.33
0.33
0.29
0.29
0.06
tube,
Year of first publication, 1973.
-------
Figure 1
Cleared, alizarin-stained monkey fetus removed by hysterotomy from an
untreated rhesus female on day 100 of pregnancy. These specimens allow
excellent visualization of the ossified skeleton but require 2 to 3 months
for preparation. The sternum was displaced forward during postmortem
examination of the thoracic viscera (165).
Source: Reproduced from Environment and Birth Defects by J.G. Wilson, by
permission of the Academic Press. Year of first publication, 1973.
43
-------
Figure 2
Radiograph of a monkey fetus removed by hysterotomy from an untreated
rhesus female on day 100 of pregnancy. Although the definition of
skeletal structures is not as clear as that in Figure 2, X-rays of this
type have proved adequate for evaluation of the skeleton, and they have
the advantage of being made quickly and inexpensively (165).
Source: Reproduced from Environment and Birth Defects by J.G. Wilson, by
permission of the Academic Press. Year of first publication, 1973-
-------
There is increasing, legitimate concern about possible postnatal func-
tional defects not revealed by these procedures or by test animals that are
born in a less developed state than the human. These defects are beyond the
scope of this survey but are referred to briefly in the concluding sections.
g. Microscope observations. Freehand razor-blade sections are usually
examined by dissecting microscope (29, 16M). Microtome sections of whole
animals or selected tissues are examined by ordinary light microscope (1U, 2M,
153, 157, 158, 173). Localized lesions and ultrastructures are examined by
transmisson or scanning electron microscope (7, 131).
h. Biochemical observations. Some blood dyscrasias, congenital errors of
metabolism, specific membrane defects, and other molecular abnormalities (171)
may under some circumstances arise by teratological processes. These methods
are no doubt useful for the study of teratogenesis; however, the extent to
which they are suitable for screening purposes is an open question. However,
as yet there is no sub-discipline of molecular teratology, and testing methods
are beyond the scope of this survey.
i. Other observations. Subtle damage to reproductive systems may be
detected by multi-generation tests, but in 1970 the FDA Advisory Committee
* *
warned of hazards in interpreting the results (12). Such tests are beyond the
scope of this survey.
5. Interpretation of Results
People vary in their development, and so do test animals. Variability is
easier to interpret if it is measured. Most laboratories draw constantly
from the same few colonies of test animals. Thus they can accumulate records
of anatomical variations, rates of intrauterine death, malformations, and
growth-retarded offspring that occur in untreated control groups. These
records tell them more than can any one concurrent control group. Some
problems of interpretation are discussed below.
45
-------
a. Growth retardation. Retarded growth in the uterus may be inferred if
a fetus or neonate weighs more than two standard deviations below the mean of
the matched controls. Unexplained cases occur in man (93) and undoubtedly in
other mammals too. Localized growth retardation, e.g., of the palatal shelves
that results in cleft palate, is a different phenomenon from the overall
growth retardation that is the subject of this subsection. However, the. two
phenomena can overlap or coexist. Some chemicals can produce both indepen-
dently, or one can lead to the other, or either can result in intrauterine
death.
Overall growth retardation usually involves delayed development of the
ossified parts of the skeleton. It is generally seen at doses that are toxic
to the maternal organism, to the fetus or both. Stunted fetuses are often
malformed and are entirely different from "runts" which are caused by
placental fusion and position of the fetus in the uterus, at least in mice.
To interpret a finding of growth retardation, two questions are often
asked: Is the effect permanent or reversible? Is it undesirable? Growth can
be permanently impaired by, for instance, prolonged exposure to low levels of
radiations (66), which are clearly undesirable. On the other hand, reversible
growth retardation can result indirectly from maternal toxicity or under-
feeding that interferes with the supply of some or all nutrients to the fetus.
i
The point at which relative 'smallness' becomes undesirable is not well
.defined. On the one hand, it has been known for more than 40 years that a
less-than-maximal growth rate and adult size can conduce to health and
longevity (e.g., 49). On the other hand, a recent 12-generation study of rats
fed a marginally protein-deficient diet produced overall growth retardation
with unevenly altered proportions of various viscera to body weight and with
behavioral changes; not all of these deviations could be reversed by transfer
of animals to a full diet, and some of them were considered undesirable. The
authors concluded that this study was extrapolable to man in principle, if not
in detail (140).
46
-------
Because these questions generally cannot be answered within the time
available for routine tera,togenicity tests, an animal weighing at least two
standard deviations below the mean of matched controls is usually assumed to
be undesirably growth-retarded.
b. Malformations. In human teratology a malformation can be defined as
an anatomical abnormality that either kills or requires surgical intervention
to maintain a normal existence, or is a cosmetic defect that reduces the
quality of life. When cadavers are dissected, circulatory system variants are
found in a continuum from near-infinity in the microcirculation to large,
infrequent deviations of the aorta. Most of these caused little or no dis-
comfort or restriction of function during life. The same applies to super-
numerary spleens. The concept of detriment to the individual raises questions
that are unanswered except according to particular circumstances. For
example, detriment has been defined as: "actual or potential cause of
disease, disability or psychologic impairment"; alternatively, as "actual or
potential hazard to survival, health, or usual activity."
Thus, detriments to people often do not correspond with detriments to test
animals. For example, cleft lip is a cosmetic detriment to man but a lethal
detriment to a rodent which thereby cannot suckle. On the other hand, poly-
dactyly and ectrodactyly ai*e classed as malformations in test animals because
they are malformations in people, while agenesis of tail is normal in people
but a detriment to some species of test animal. Skeletal variations of ribs
and thoracic and lumbar vertebrae are not considered abnormal in mice, rats,
rabbits, hamsters, guinea pigs, or rhesus monkeys even if frequencies increase
after treatment with a known teratogen; control treatments (inert injections)
also sometimes increase the frequency. Wilson suggests that such increases
may indicate that embryotoxicity is "being approached" or that irregular
ossification "may simply reflect growth retardation," so that body weight
records are needed for interpretation (165).
Therefore it is hard to define what is a malformation in a test animal,
especially in terms of what might be a malformation if it were found in
people. Defensible criteria, e.g., limits of magnitude or of statistical
rarity, do not seem to have emerged..
47
-------
One suggested criterion has been that a treatment shoudl increase the
frequency of known 'background1 (accumulated control) variations (2, 61, 86,
125, 126). However, some teratogens may not do this. When a variant or
malformation has been reported in accumulated controls, a significant increase
of the same in treated animals is cause for suspicion of teratogenicity, but
no more.
An unequivocal criterion of teratogenicity, regardless of control
findings, is a dose-related increase of malformations in treated animals,
c. Death in utero. The proportion of dead implants per female appears
to be a more appropriate measure of fetal death than the number of dead
implants per female because the analyses based on proportion take into consi-
deration the total number of implants per female (57).
, X
The relationship between deaths and malformations can be complex and
according to Wilson (165), is not always interpreted correctly. This is
particularly evident in cases where both manifestations tend to appear at
comparable dose levels during organogenesis and have parallel dose-response
until, at high doses, deaths replace malformations (e.g., cyclophosphami.de)
(18).
However, there are some drugs which produce embryolethality but not
malformations. This suggests that embryolethality and teratogenesis are two
separate phenomena produced by different modes of toxic action (e.g.,
methotrexate, 6-aminonicotinamide) (18). Obviously, selection of the proper
dose is of paramount importance. If the dose is too high, a 100$ lethality
will occur. If the dose is too low, no effect will be produced on the fetus
and the results will provide a false sense of security about the drug's safety.
Any increase of either deaths or malformations denotes toxicity. When
rhesus monkeys are given toxic doses, deaths always seem to occur more readily
than malformations. If this also applies to people, it "would in most
circumstances be regarded as preferable" to an increase of malformations (165).
48
-------
d. Positive controls. These are controls treated with a known
teratogen for comparison with the substance whose teratogenicity is unknown.
They ought not to be needed, for three reasons:
o The investigator should already know the general sensitivity of his
animals,
o Sensitivity to one agent is no guide to sensitivity to another; for
example, rats are sensitive to trypan blue and to late antagonist
but not to thalidomide or cortisone, and
o The range of sensitivities differs among strains of test animals,
including man.
Rather, the aim should be to find embryotoxicity in two or more
unrelated animal species and, if animals are available for a positive control
group, it would be more profitable to add them to a treatment group (165).
e. The "litter effect". Often malformations are concentrated in some
litters, sparing others (65, 170). This usually denotes maternal toxicity
(UO) but may also reflect the genetic similarity of littermates. The question
of how to interpret such clustering is much debated, but it is complex and
there are not enough facts to provide answers.
Many statisticians prefer to count the litter (pregnant female) rather
than the fetus as the basic experimental unit because they feel that ithe per
litter evaluation, which is based on the average reaction within the litter,
automatically includes the reaction of the individual fetus (5, 138, 162).
The number per litter tests has been recommended by Haseman and Hogen (57) for
analysis of teratologic incidence data.
Other investigators prefer to count individual fetuses, pooling the
total in a treatment group or comparing malformation rates per litter among
groups (10, 60, 165). This approach is misleading and may seriously
exaggerate the significance level (57).
Besides the selection of the litter as the experimental unit, there are
other factors involved in the appropriate statistical analysis of the
49
-------
teratologic data. These include the number and size of the test population,
distribution of the variables of interest, nature of comparison desired, the
magnitude and degree of litter effect, and fetal age and size at the time of
treatment (85, 90).
Questions have been raised about how much difference is involved in the
use of different methods, but they have not as yet been resolved (10, 138).
f. Sources of error. If results are not clear-cut, the first impulse
is to repeat the test. This will be of little use if the first test contained
errors that are not corrected. Such errors may be hard to recognize because
of unknown variables, the difficulty of controlling known variables, and
disappointing reproducibility (17). Possible causes of poor reproducibility
that require vigilance, from the design of a test to its interpretation,
include seasonal fluctuations of effects, spontaneous mutations, spontaneous
or infectious abortions, inaccurate timing of fertilization, and irregular
administration of test doses (165). . .
g. Statistical methods. For interpretation of teratologic results, the
following statistical methods are employed: two-way analysis of variance,
regression analysis and tests for best fit. Analyses based on individual
fetuses employ the chi square test, Fisher's exact test and the t-test. It is
believed that more rigorous statistical analysis is required to maximally
evaluate the teratologic results and risks.
B. In Vitro Systems
In vitro systems, including tissue cultures and avian eggs, have con-
tributed to basic research on cellular mechanisms of teratology, but have
virtually no role as short-cuts in screening chemicals for potential terato-
genicity for man. Authorities do no agree on the value of avian eggs as
coarse indicators of potential teratogenicity. In this test system untreated
eggs can exhibit teratogenic effects from careless handling. In addition,
embryos have no way of excreting the test compound or its metabolite. For
50
-------
example, malathion was found to have ter'atogenic effects (monster formation)
when evaluated by this test method, but there were no adverse effects when
malathion was tested on mammals or in humans exposed during its manufacture
(79).
However, cultures of embryos and cells may become useful in the future
for studying metabolites of environmental chemicals that arise in mammals.
Conceptually, the questions that in vitro systems could answer await develop-
ment; this requires parallel correlative studies with particular chemicals in
vivo and _in vitro, and few such studies have been reported. Table 18
summarizes a selection of reported in vitro systems.
51
-------
Table 18
Some In Vitro Systems That Have Been Employed in Teratology
System
References
Embryos, vertebrate
Mammals:
Rat
Mouse
Birds:
Chick explanted
embryos
Amphibians:
Frog
Toad
Newt
Salamander
Fish:
102, 106, 121,
5, 82, 102, 105
56, 83, 103, 156
20, 21
12, 91, 132
19, 20
133
Embryos, invertebrate
Echinoderms:
Sea Urchin
Sand Dollar
Insects:
Fruit Fly
Organ Cultures
General:
Mouse long bones
Mouse limb buds
Mouse tail vertebrae
Rat eyes
Chick embryo organs
Tissue Cultures
Chick embryo tissues
Microorganisms
Protozoa:
Flagellates
Algae:
Deamidiaceaa
52, 117
75
15
5,
127
109
175
8
.14
14
, 151
14
116
52
-------
IV. CURRENTLY USED METHODOLOGIES
No one routine is considered reliable enough for use without question. In
1966, the FDA and in 1967, the WHO developed principles for testing chemicals
for teratogenicity. The recommendations issued constitute only "guidelines"
to methods (155). The onus is on the manufacturer, who has prior and superior
knowledge of his product, to propose methods that the FDA will then approve or
reject. The guidelines have been developed for new drug applications and have
received comment on this basis (35,78). No comparable methodologies have been
developed specifically for environmental chemicals, and to extrapolate from
the FDA guidelines one should bear in mind that conditions of exposure and the
reasons for it may be very different.
The FDA guidelines (155) divide testing into three "segments": I. Study
of Fertility and General Reproductive Performance; II. Teratological Study;
III. Perinatal and Postnatal Study.
Segment I covers the entire span of reproduction in both sexes and is
considered a pilot series, to "serve as a guide for subsequent studies in
depth" (155). Kelsey (78), an FDA scientist, noted that this segment is often
run along with chronic toxicity tests for convenience, using the same test
animals. She described a typical protocol: females are dosed from two weeks
before mating until they are killed or their offspring are weaned, and males
are dosed for 68-80 days before mating. Half of the females are killed on day
13 and examined for number, distribution, and condition of embryos,
implantation sites, etc. The others litter naturally and are observed for
duration of pregnancy, litter size, stillbirths, and malformations of
newborn. According to results, reproduction may be tested in the offspring,
or the same females may be studied through a second litter. The guidelines
themselves (155) give no protocols, but Segment I clearly has a wider scope
than teratology.
Segment II is intended to detect both teratology and embryotoxicity
(155). The guidelines restrict doses to the period of organogenesis and also.
envisage second-litter and second-generation studies. They include some
precise instructions. At least two species should be used (mouse, rat,
53
-------
and rabbit are used most often). Males are untreated. Mice and rats are
dosed on days 5-15, rabbits on days 6-18. Cesarean section is performed 1-2
days before expected delivery.
Observed are: number and placement of fetuses, living and dead, early and
late resorptions, and corpora lutea. Each fetus is weighed and examined for
external defects. Internal defects are always looked for: skeletal defects
using alizarin stain, and visceral defects by dissection or the Wilson
method. Rat fetuses are randomized into subgroups for these searches. Rabbit
fetuses are examined for all types of defect, and half of them may be
incubated for 24 hours with emphasis on observing the 6-hour survival rate.
Kelsey commented (78) on the need to choose a species "known or believed
to metabolize the drug in a similar fashion as man", and to use two or more
dose levels including the maximal tolerated dose.
. *
Segment III specifies (155) that animals should be dosed for the last
third of pregnancy and afterwards until weaning. The studies therefore are
beyond the scope of this survey.
The guidelines also mention a number of specific points. Segment II
studies can be made more specific by dividing the dosage period into
sub-periods that add up to the prescribed period. Chick embryo assays should
be used "for ancillary" data only". At least 20 female rats or mice (some
laboratories use 50) or 10 rabbits are required per group, and control groups
"should be "relatively large". Each laboratory should report its background
data, i.e., accumulated negative controls.
Kelsey (78) pointed out that the FDA accepts studies conducted outside the
United States if the investigator and laboratory are known and approved by
them. She emphasizes the need to know more about environmental long-term,
low-dose exposures, for which improved methodologies are required.
In 1973 the Canadian Department of Health and Welfare issued the report of
a working party in the form of a review with recommendations (5). In general,
these are parallel to the FDA guidelines, but they cover more ground, and
there are some differences.
54
-------
The investigator is advised to choose two out of three basic treatment
>
schedules: on certain days of gestation, throughout gestation, or before and
during gestation. The peri'od of organogenesis is given as days 6-15 in mice
and rats, 6-18 in rabbits, H-1H in hamsters, 6-20 in guinea pigs, 7-35 in
pigs, and 9-^0 in monkeys. Two species are chosen: the rabbit, and the
mouse, rat or hamster. The significance of placentation differences is
treated as uncertain, and so are the effects of maternal age and parity,
regarded as more important. In vitro test systems are not recommended, but
existing data should be reported.
Much emphasis is placed on husbandry, in particular on ensuring.that test
animals are given an inert environment free from, for example, disinfectant
residues. It is thought important to copy the expected route(s) of human
exposure to the test substance, so far as possible. The dose vehicle .is
expected to comply with rigorous standards of inertness and to allow the dose
to be given in solution. If "drug" combination effects are anticipated, these
should be studied.
Observations generally follow the FDA guidelines with added emphasis on
delayed effects and a recommendation to observe some animals till they are
weaned (42). Background information on human malformations is required, and
problems of extrapolation of animal data to man are discussed, as are
* '
statistical problems of interpretation such as litter effects. The report
concludes with three model protocols given as typical examples rather than as
specific recommendations: (i) for mammalian species generally, (2) for
rabbits, and (3) for rats.
In Britain three dose-levels are reportedly used, and Robson commented
that the additional level permits logarithmic spacing and also better
separation of embryolethality from teratogenicity (122).
In summary, no standard methodology is required for regulatory tests of
teratogenicity, but some general principles are recommended fairly uniformly.
These include use of the rabbit plus one or more of several small rodent
species, emphasis on background knowledge of the species chosen, dosage with
at least two levels of the test substance, administration by a route similar
55
-------
to anticipated human exposure, large negative-control groups and meticulous
husbandry with special attention to a teratologically inert environment.
Protocols are expected to vary according to the compound to be tested, and the
responsibility is on the manufacturer to propose in detail according to his
superior knowledge of the product. Teratology is not the first aspect of
toxicity to be tested when a chemical is screened by the manufacturer.
56
-------
V. SURVEILLANCE
Surveillance is the approach to meaningful information on the incidence of
human malformations. It is the only usual source of data on frequencies of
particular malformations in large human populations, and it can reveal much
that is unexpected. As an indicator for human teratolcgic risks, surveillance
is considered equal or superior to animals tests.
Four sorts of information are provided, according to the FDA Advisory
Committee on Protocols for Safety Evaluations (42): (1) accumulated exposure
data; (2) tissue samples from human surgery or autopsies; (3) notification by
doctors; (4) epidemiological surveys especially in occupational groups. A
problem has been that such information is acquired very slowly (5^). Indeed,
surveillance information can be reliable, comprehensive, complete, or prompt,
but seldom has it been all of these at one time. So a choice of surveillance
methods can be determined by one's priorities.
Trends in surveillance methods were reviewed in 197^ by Flynt (36). He
reported that six priorities had emerged at a recent WHO conference: (1)
monitoring; (2) epidemiologic studies; (3) registries; CO detection of new
syndromes; (5) education of doctors and the public; (6) public relations.
Each of these priorities had its own requirements. For example:
(1) Monitoring required prompt information (within 1-2 months of birth)
from selected sources. A source could be unrepresentative so long as
it remained constant over time.
(2) Epidemiologic studies required large, mixed populations and complete
reports. Promptness was not required.
(3) Registries were used for genetic studies, planning of health
services, and follow-up casework. Only the casework required prompt
reports.
CO New syndromes were detected through accurate and complete reports
from selected sources with adequate facilities. Promptness was
secondary.
57
-------
(5) and (6) Education and public relations were ancillary but were
essential to any acceptance of a surveillance program by doctors and
the public.
The history of surveillance in the United States was undistinguished until
recently. In 19^0, for example, birth certificates in New York State were
printed on the back with a form on which to report certain malformations as
Yes/No. By 1963, a 95$ response rate was noted, but the information had not
been regularly analyzed, and "shortcomings and biases" were admitted (92).
After the thalidomide tragedy more attention was given to surveillance. By
1971*, four programs were in operation: at Atlanta, Georgia; Albany, New York;
Olympia, Washington; and Jacksonville, Florida.
The Metropolitan Atlanta Congenital Defects Program, started in 1967, is
run jointly by the Center for Disease Control (CDC), the State of Georgia, and
Emory University. In 1971, this program covered a population of 1.3 million
(77? white) and was served by 21 hospitals and M chromosome laboratories.
Malformed infants were registered mainly through visits to hospital nurseries,
delivery rooms, and record departments. Stillbirth and infant-death
certificates were supplied by the Georgia Health Department. Case-reports
were collected from doctors who consulted the program's professional staff.
For surveillance purposes, a malformation was defined as "any structural,
chromosomal or biochemical abnormality in an infant diagnosed before his first
birthday", but comments were that diagnosis before 7 days tended to be more
reliable than later diagnoses (37).
A monthly report analyzing data gathered is circulated to those who
contribute data. If clustering is seen, families are interviewed about
possible environmental exposures. Since 1971, data have been monitored by
computer, and since 1971, a CDC computer automatically prints out
statistically significant increases above 'background'. By 1971*, 23 defects
had shown such increases, but no environmental causes had been identified. In
1971, the program cost about $50,000 a year and made people aware of
malformation problems (36,37).
i
In 1975 Dr. J.W. Flynt told us informally that CDC is now processing data
on 65,000 births a year from various programs throughout the United States.
58
-------
Expected incidences of some malformations,will be calculated from 1970-1973
data, when assembled. The automatic printout of significant increases will be
used by CDC to offer help to state health departments in three different
ways: (1) reviewing individual hospital case records; (2) interviewing
families; (3) offering services of trained epidemiologists. In Dr. Flynt's
opinion, there were still too many unknowns to permit the valid design of
systems for automated data-analysis or the monitoring of special occupation
groups of women, and such activity might be interpreted as invasion of
privacy. However, highly organized societies outside the United States, such
as Japan and Poland, have in the past welcomed American support for studies in
special groups of their populations.
In Britain voluntary notification started in 196H, a midwife or doctor
observing effects "in any way he wishes" and reporting to the General Register
Office on a form listing 100 categories of birth defects (59). Information
readily acquired included: locations of birth and of mother's residence,
place of birth (home/hospital/other), birthdate, sex, living/stillborn,
mother's age, and mother's total of live and stillbirths. Less often acquired
were mother's name, history of multiple births, and infant's birth weight.
Currently, the UK Office of Population Censuses and Surveys receives
notifications and alerts local authorities whenever the frequency of a defect
rises by a significant ^% above the previous average. About one-half of the
local authorities have followed up such notifications (160). Comments have
been (59, 160, 161) that: efficiency of notification is less than complete,
especially when the defect was lethal postnatally; a rise of defects is hard
to distinguish from improved efficiency of notification; statistical
approaches need more study; and therefore, such data must be interpreted with
caution.
Informally we understand that British records have started to include the
occupations of pregnant women with a view to future monitoring for
teratogens. Surveillance of exposures to environmental chemicals does not
seem to have been considered practical and an informal search for association
between diseased potatoes and anencephaly (a recent topic in Britain) was
negative.
59
-------
In 1971 Miller at the National Cancer Institute, NIH, suggested (95) that
a death-certificate registry for malformations, like that for cancers, might
prove useful. The data were routinely available but seldom coded. The
quality of diagnoses would vary, but defects lethal in the first month of life
(or during infancy) could be used. Fetal death certificates would be coded
separately. These sources of information were, .in the author's opinion,
better than birth certificates. Other sources advocated by the author
included Health Maintenance Organizations and veterinary observations; the
latter had been useful in identification of mercury as the contaminant at
Minnamata, and more recently in a Kentucky epidemic of lethal skeletal defects
in pigs grazing on vegetables that had been sprayed.
The Federal Meat and Poultry Program of the Animal and Plant Health
Inspection Service of the USDA does, in fact, operate a National Monitoring
Program in order to ensure "a meat and poultry supply that is wholesome and
contains no violative levels of any drug or chemical residue" (100). Sample
animals for inspection are selected by computer, and^tissues analyzed include
fat, liver, kidney, and muscle. In 1973 over 110,000 analyses were performed
on 18,000 animals for.47 different chemicals, and over 800. special ...
surveillance programs were conducted after violations had been identified.
The chemicals included teratogenic pesticides, and continual re-evaluation of
the Program permitted rapid responses to new information.
In summary, the techniques under levelopment appear to offer improvement
in the surveillance of environmental teratogens. However, much better
information can be obtained about total exposure from all sources if
laboratory measurements to determine concentrations are combined with surveys
to assess total intake. Amount of food, water, medication consumed, duration
of work in different jobs or of residence in communities that differ in their
potential for exposure to environmental teratogens are some of the factors
that must be taken into account.
60 '
-------
VI. ECONOMICS
I
The costs of teratogenic studies, as now required, are enormous. One
estimate puts the minimum cost of a complete study at $10,400 for rats,
$10,000 for rabbits, $6,500 for mice, $8,600 for hamsters and $55,900 or more
for monkeys (4M). A recent determination of typical costs for bioassay
functions in general was made by Tracor Jitco, Inc., as part of another
project, and some relevant excerpts are shown in Table 19. It is emphasized
that these determinations are at best approximate, for conditions and test
requirements vary enormously. This Table refers to work performed on contract
by laboratories specializing in such work. Earlier estimates for work
performed at university laboratories specializing in research, shown only for
procurement and maintenance of test animals (Table 20) and in fact equally
vague and variable, indicate much higher expenses in a university setting than
in a contract laboratory. However, this comparison may not be true under all
conditions.
If a manufacturer has the primary duty of having teratogenicity tests
performed, it is assumed that in most cases the work will be done not
'in-house', but in some laboratory accustomed to doing such work. It is also
assumed that the manufacturer will conduct a risk-benefit analysis for his own
purposes, on the assumption that he will be held liable for the full costs of
damage. The stopper here is the question of insurability and premia.
61
-------
Table 19
Typical Costs4 for Some Careinogenesis Bioassay Functions
Determined by Tracer Jitco, Inc., for the National Cancer
Institute in 1976.
Function Costs
Animal care: $0.15 per day per animal for an
animal population of about 4,000
mice and 4,000 rats.
Necropsy: $8.50 per adult animal.
Trimming; $6.50 per animal, 30 organs.
Pathological diagnosis: $1.60 per tissue.
Histology: $2.50 per slide, 8 slides per
mouse 12 slides per rat or
hamster.
Total costs including labor, materials, overheads, and
fees in an average contract laboratory.
62
-------
Table 20
Laboratory Procurement and Maintenance Costs Determined by the American
Association of Medical Colleges for Year Ending June 30, 1973
Animal
Procurement, ready
for experiment, $
Husbandry, full
costs per day, $
Dogs
Cats
Mice
Rats
Guinea pigs
Hamsters
Rabbits
Chickens
Calves
Swine
Frogs
Rhesus monkeys
Squirrel monkeys
45.52-59
27.50
0.50-0.
1.65-2.
4.40
1.93
5.50
2.20
110.00
33.00
2.20
92.62
49.00
.65
61
20
Ungulates
Amphibians
Primates
1.602
0.9^8
0.034
0.170
0.170
0.034
0.446
0.258
2.022
0.105
1 281
1 <. w *
63
-------
VII. LISTS OF KNOWN TERATOGENS
One comprehensive listing of known teratogens appears to be recognized in
the field of teratology: Catalog of Teratogenic Agents by Thomas H. Shepard
(130). Published in 1973, this list is computerized with a view to regular
updating. Copies-have been delivered to the EPA, and the list is not repeated
here.
Shepard (130) commented that about 3% of all human newborns have a
congenital anomaly requiring medical attention (about 1$ being life-
threatening) and that over twice this amount are detected with increasing
age. About 15$ of the total have genetic origins, and less than 3% of the
remainder are caused by teratogens. The author listed over 600 teratogens for
experimental animals, and stated that only about 20 are known to cause defects
in man. He noted that there is a wide gap of knowledge between experimental
teratology and the role of external agents in human teratology.
64
-------
VIII. RECOMMENDATIONS
t
A. Screening Chemicals Already in the Environment for Teratogenic Effects
1. How chemicals already in the environment would be screened for
teratogenic effects is not clear, since there is no readily apparent basis for
selecting particular chemicals for screening for potential teratogenicity,
from the many thousands presently in the environment, except perhaps a basis
of the numbers of people exposed.
2. Nevertheless, it is recommended that a watch be kept on chemicals
already in the environment, for the following alterations:
a. Quantity increases that may overstep teratogenicity thresholds, and
b. Interactions, in the environment or in exposed individuals, between
chemicals already present and "other chemicals," with teratogenic results.
"Other chemicals" might be newly introduced and shown to be nonteratogenic by
themselves, or already present and the amount is increased, or introduced as
drugs or food chemicals rather than as industrial chemicals, or other perhaps
complex alterations of the present state of the environment. Such a watch
would be activated by results' of population surveillance, mentioned below.
Thus if a general or local increase of birth defects, total or specific, were
to be revealed by surveillance, inquiry about possible environmental causes
would not be limited to newly introduced chemicals but would also cover
alterations as outlined above.
3. It is recommended that special exposure groups, which include women,
be monitored by surveillance as outlined below for possible teratogenic
effects of chemicals in their existing special environments.
4. If, as a result of inquiries in paragraphs 2 and 3 immediately above,
suspicion should fall on any chemical already in the general or special
environment being surveyed, then that chemical should be screened for
potential teratogenicity using criteria as outlined in Subsection B below, and
taking into account the circumstances revealed by the surveillance
procedures. It is envisaged that the responsibility for the work of screening
would be the producer's.
65
-------
5. Any surveillance of human populations for teratological
manifestations can achieve its objectives only if the design is adequate.
Following are some recommendations for the design of general surveillance
projects and for special surveillance of particular population segments.
a. In practice, areas for surveillance should be selected with geo-
graphical or other limits, for example, areas served by selected medical
centers already possessing the requisite skills and facilities.
b. All information should be fed into one national collecting center.
In Britain all information is fed to the Office of Census and Surveys. In the
United States, the Center for Disease Control at Atlanta, Georgia, might be
the proper recipient, or FDA or EPA.'
c. Within the selected area, reporting should be mandatory.
d. An objective of organized, comprehensive reporting will be to develop
baseline criteria, i.e., true background levels of malformations and func-
tional deficits, with the seasonal and other regular fluctations. Then even
small irregular 'blips' can be rapidly discerned. Inter-regional comparisons
of baseline data will be important.
e. The demography of each reporting area should* be recorded and kept
up-to-date. All reports of terata should include the mother's age and
occupation during pregnancy, place of work, and home address.
f. The environments in the reporting area should be charted (composition
of air, water, radiations, and other sources of potential teratogenic
exposures) and the map kept up-to-date.
g. In handling and processing the above types of data, the expected
timelags relevant to teratogenesis, both before and after exposure of the
mother, should not be omitted from the program.
h. The reporting area should be mapped for drug usage: (i) local
prescribing patterns by doctors, and (iiO local consumption of over-the-
counter drugs by the public, in each case with special reference to pregnant
women. 66
-------
i. A standard system of reporting'birth defects and functional deficits
is required. Problems in (the development, of such a system may include:
excessive dependence on the knowledge and diagnostic ability of local
physicians, excessive need for special procedures, inadequacies in any
itemized list of defects at the present time, considerations of privacy for
families, irregular performance of procedures by reporting doctors, lack of
indicators of many functional deficits and some malformations at time of birth
or reporting. Therefore, a special study is recommended to evaluate these and
analogous problems and to develop a provisional standardized reporting
procedure.
j. Miscarriages, abortions, stillbirths, and deaths in the perinatal
period should be included in the mandatory reporting procedures, together with
any concomitant malformations. Problems of performance may include:
excessive need for autopsy and pathological skills and facilities, resistance
by some families in the reporting area, uneven willingness of local doctors to
cooperate. The above-mentioned special study recommendation should cover this
subsection as well.
k. Notifications should include known pregnancies of employed women, to
be reported by personnel departments of the employers.
*
1. An informative adjunct to human surveillance would be surveillance,
at the same time, of hereditary and nonhereditary malformations and functional
deficits in the farm and domestic animals in the reporting area. Problems may
include: concealments by farmers for commercial reasons or to' safeguard
pedigree statuses, lack of a mandatory birth-and-death registration system for
animals, incomplete involvement of veterinary practitioners in problems of
animal health and disease.
m. At the same time note will be taken of current requirements of state
governments for notification of birth defects, the extent to which they are
fulfilled, the methods used and the problems met. Any state requirements for
notification of birth defects in farm animals will also be noted. Where a
reporting area includes portions of two or more state jurisdictions, the two
sets of requirements may differ. Whether or not this is the case, it is
presumed that state governments will be consulted during any surveillance
.planning and invited to collaborate.
67
-------
6. Women employed under industrial conditions may be exposed to special
teratological risks. Special surveillance is recommended, as follows.
a. Populations should be identified consisting of sufficient numbers of
women in one or more areas served by a medical center with the requisite
skills and facilities. These may, but need not, be the same as those identi-
fied for general surveillance but rather should be selected according to
location of types of potential exposure and concentrations of women poten-
tially at risk. .
b. Within a population, all women should be subjects, not only preg-
nant women. Such surveillance is practiced currently in the Soviet Union, and
information should be sought through the proper channels.
c. Where the basis of a population is employment in a particular
industry, personnel departments of that industry should supply all informa-
tion required for surveillance purposes, and medical officers already involved
in health services to that industry should be involved in the surveillance,
ancillary to officers employed or retained as consultants by the regulatory
agency.
d. The major routes of exposure for women in these populations are
expected to be the skin and the lungs. In addition, food and drink prepared
i
or sold to the population in the areas of potential exposure should also be
considered a potential source of exposure, and therefore ingestion may be a
potential route.
e. Other recommendations of detail, as mentioned above (Item 5), apply
also to special surveillance.
7. All designs for both general and special exposure surveillance
procedures should include measures for follow-up in case of positive indi-
cations.
68
-------
B. Prospective Screening of Chemicals Not Yet in the Environment for
i
Potential Teratogenicity
Although doubts will always remain where no evidence of teratogenicity is
found, screening for potential teratogenicity is both possible and practical
at the present time, and is recommended.
1. Because of the doubts, and because human beings are the best
indicators of human teratology, both general and special exposure surveil-
lance procedures are recommended whenever a chemical is permitted into the
environment for the first time. Operators of general surveillance will be
alerted to watch for appropriate reports and statistical effects at expected
times. Special exposure populations will be identified and surveillance
prearranged at the time permission is given, and preferably as a condition of
that permission.
2. Laboratory screening of "new" chemicals for potential teratogenicity
is recommended as follows.
a. Detailed methods must be selected in each case by the manufacturer,
in view of his prior and superior knowledge of his product.
^
b. It is assumed that standard toxicity data will have been determined
(acute, subacute, chronic) and that studies on the effects on.reproduction of
low-level, long-term exposures for several generations will at least have been
considered.
c. The teratogenicity study will involve only one generation.
d. It will involve two or more species of rodent or lagomorph, e.g.,
mouse, rat, hamster, guinea pig, rabbit. If the fetal CNS is a suspected
target, a species, such as the guinea pig, having a protracted fetal period
comparable to that in humans should be considered. In addition to different
species, the comparability of the genetic structure of the test species
population to that of the human population must be weighed.
69
-------
e. In each species selected for testing, one or more identified strains
should be used. If results are equivocal, further tests should be performed
in additional strains. Debate should be generated on whether inbred or hybrid
strains give better information. For example, inbred strains may be better
for studies of mechanism, where homogeneity of response is important. In many
other studies, outbred strains may give a better representation of the
variability to be expected in a heterogeneous population.
f. Each treatment group (test, concurrent controls) should include 20 or
more pregnant females, mated in the same laboratory where the assay is to be
done, i.e., not transported in the pregnant state. The control group
should resemble the test group as closely as possible (e.g., by bodyweight
averaging) and should be kept under the same conditions (e.g., temperature,
light, humidity, cage-positions, etc.).
g. The LD_Q of the test substance should be determined with similar
animals, e.g., LD _ values determined for aged males are irrelevant. If the
LD_Q is impractical, MTD or some other well defined toxic level should be..
established as a basal criterion.
h. Three dose levels should be employed to begin with. These should be
logical fractions of the LD e.g., 1/2, 1/U, 1/8; or successive 1/2-
dilutions; or log-intervals, or other suitable division, depending on the
substance and its characteristics known to the manufacturer. The LD _ value
should be of pregnant animals, not of the fetus, for the embryotoxic zone may
be narrow. If the first three dose-levels used do not result in (i) death of
most fetuses at one level, and (ii) survival of most fetuses at another level,
then further levels should be administered to other groups until these results
are achieved.
i. The typical or expected route of human exposure should be copied as
closely as possible. Drinking water, for instance, can be copied exactly, or
at least test animals should be dosed by gavage. Aerosols should be given in
suitable chambers, although the LD_Q may be hard to establish this way. The
overriding aim is to reproduce in test animals, by whatever route, the tissue
, t
levels and persistence of the test chemical and its metabolites that would
occur in typical humans under typical exposure conditions. This ideal may be
unattainable for many reasons.
70
-------
j. Dose-equivalents for specific skin surface area (body surface/body-
weight) should be adjusted for interspecies comparisons, even if this is not
i
the expected exposure route.
k. The dose schedule should approximate the expected patterns of human
exposures (e.g., continuous, intermittent) and conditions of exposure (e.g.,
sunlight, temperature, humidity), taking into account diurnal and other cycles
in the test animal compared with man.
1. The purity of the substance as marketed or released into the
environment would be an acceptable criterion, but this should be constant
throughout the test series even though it may vary in practice, e.g., between
batches. Purity as synthesized would also be acceptable. Alternatively, so
would purity in the expected environmental context. The ideal purity and form
of the substance is the expected form to which people are likely to be exposed.
m. The test vehicle for the substance to be tested should be inert and
should sustain the concentrations and the appropriate physical state (e.g.,
particulate, non-colloidal) of the substance throughout the experiment.
Concurrent controls should be given the vehicle without the substance. The
nature of the vehicle will be the manufacturer's responsibility.
! +'
n. Pregnant test females should be killed one day or so before expected
delivery, and the method should not traumatize the fetus or uterus. If there
is particular reason to modify this in order to obtain some meaningful data,
some animals may be killed at additional times or by additional or modified
methods, e.g., by perfusion-fixation.
o. The abdomen is opened, and the uterus is inspected in situ for total
implantation sites including all stages of resorption. The corpora lutea are
counted and the total compared with the total observed implantation sites.
Preimplantation or early-resorption losses should be estimated either by the
foregoing comparison or by other suitable methods. Faulty implantation in
mice, rats, and hamsters usually leaves a scar if the embryo survives for at
least 2M hours. The fact that accuracy of these estimates is not perfect, and
also that it varies by species, does not abolish their usefulness.
71
-------
p. The uterus is opened, and living fetuses are distinguished from dead,
intact fetuses by color, movement and other signs. Macerated fetuses are
counted as part resorptions. Early and late resorptions need not be counted
separately.
q. The umbilical cord of each living fetus is clamped or cauterized, and
the fetus is examined grossly (under magnification if need be) for deviations
of proportions, configuration or symmetry, and other gross structural features.
r. Fetuses are blotted and weighed individually.
s. Fetuses are observed for the usual vital signs.
t. The sex of each fetus is recorded.
u. One-half to one-third of fetuses are prepared for skeletal visuali-
zation in 95$ ethanol as preparation for clearing and staining. The rest are
fixed in Bouin's fluid or other nonhardening fixatives that can decalcify the
skeleton, as preparation for sectioning or dissection, by standard methods, to
reveal any soft-tissue abnormalities. Use of the same fetuses for both
skeletal and soft tissue examinations has been reported; the fetuses were
fixed in 70% alcohol, dissected, and finally were stained with alizarin (1^3).
v. If there is reason to suspect that subtle damage of particular
systems, e.g., the nervous system, may occur selectively, appropriate
procedures should be devised to reveal histological changes. This may need
modification of the above procedures or special fixation procedures, e.g.,
perfusion-fixation. If so, adequate samples should be fixed specially.
w. There is growing concern about possible postnatal functional defects
that are not revealed by size, growth, survival, gross anomalies, and other
measurements described above. Deviations from normal maturational milestones
should be observed, e.g., eyelid opening, hair growth, incisor eruption,
freeing of pinna, posture, locomotion. In addition, neuromuscular competence
should be tested by appropriate procedures, e.g., swimming, climbing, walking
a rod. Because young adults may compensate for such defects, which emerge
72
-------
when senescent animals are given functional tests, a proportion of test
animals may require lifetime observation.. This area of concern is considered
legitimate, but it evokes problems. Functions that develop in utero in people
may develop postnatally in rats. Until suitable animal models or accumula-
tions of background information become available to simulate human time in
utero, the best compromise model would involve use of rats because most is
known about them. Manufacturers should be prepared to respond to demands for
functional tests.
x. The uncertainties in teratogenicity testing require that risk-and-
benefit analyses be undertaken. The manufacturer should be responsible for
primary data, which can then be checked independently.
73
-------
IX. POSSIBLE FUTURE METHODOLOGIES
The purpose of this section is to indicate directions of technical
development that appear, from data and opinions received during preparation of
this report, to be both desirable and currently feasible. Four such items
have been identified: postnatal evaluations, shortcuts in current methods of
assessing potential teratogenicity, additional species of test animals, and
improved monitoring of human populations.
A. Postnatal Evaluations
Much should be known about postnatal physiological and behavioral
development of a species before it can be used profitably to assess terato-
genic effects that were not readily detected by existing methods at or before
birth. Much is known about one species, the rat, because it is used to
, *
research the basic processes of postnatal development (137). But its course
of postnatal development cannot be compared to that of humans, and so the rat
is judged to be less than ideal for the purposes outlined in this survey.
Nevertheless, techniques for postnatal evaluation are becoming adequate for
screening purposes in teratology, and it would seem desirable and feasible to
work up the background knowledge in some more suitable species.
The need for postnatal teratologic evaluation arises mainly because some
structures, particularly in the CNS, do not mature fully in human fetuses, in
which they may still be liable to certain teratogenic influences even in late
pregnancy. As the same structures develop postnatally in many laboratory
animals, such animals cannot be used meaningfully for prenatal screening for
this teratogenicity. This tends to rule out rats and mice for testing
potential late-stage teratogenicity of environmental substances and suggests
that the guinea pig (or other suitable species with a protracted fetal period)
might be a useful test animal. It seems both desirable and feasible to work
up enough background knowledge to qualify the guinea pig for this purpose.
B. Possible Shortcuts in Current Procedures
i
Shortcuts are desirable and need developing when long lists of chemicals
are to be screened for potential teratogenicity in a short time. One
74
-------
approach would be to consider use of submammalian animals: invertebrates,
sub-avian vertebrates, and birds. For example, the metabolism of a chemical
by liver microsomes might 'be studied in vitro, and then submammalian embryos
would be treated with it. However, to use this shortcut requires knowledge of
basic mechanisms involved, such as the molecular mechanisms by which the
chemical damages the tissues that are tested. Such knowledge is often
incomplete.
It is not enough to know that two chemicals are structurally related for
catalog purposes, for it may be the conformation of 'active' parts of
molecules that matters. These may (a) consist of only a minute portion of the
molecule, and (b) be similar in substances that would not be classified
together by their entire structures. Similar considerations apply to
binding-sites of test tissues and organisms. Therefore, testing by chemical
analogy demands specific foreknowledge, and there are possible pitfalls. For
example, the intact molecule of thalidomide is teratogenic to man during early
organogenesis, and its major component, glutamic acid, may be cytotoxic to
parts of the CMS in some late-stage animal fetuses. But there is no evidence
that glutamic acid is teratogenic during early organogenesis, or thalidomide
itself during late pregnancy. Therefore one suspects that these
manifestations are chemically unrelated. In summary, research would seem
worthwhile and technically feasible in order to identify tissue-chemical
*
interactions that occur both in man and in cell-free systems or simple animal
forms lacking maternal-conceptus placental units.
C. Additional Species of Test Animals
Many species that in prospect seem economically feasible have seldom or
never been used. The ferret is an example. The principle of species
diversification has worked in other fields such as food production, where the
gain is protection against epidemics. Suppose, for example, that a test
strain were to be standardized and then were to be found sensitive to a mutant
pathogen? The background knowledge would become useless. Therefore, "tiered"
or "multilevel" systems, in which the design of a test involves several
species, would seem worth looking into, even though the procedures might
become more elaborate. Occasions for use of this sort of procedure would be
determined in advance by risk-benefit analysis.
75
-------
D. Improved Monitoring of Human Populations
Improvements are needed in (a) monitoring, (b) reporting and analyzing
(c) dissemination of knowledge, and (d) follow-up action. Better surveillance
is needed of the general population and of special groups, mainly women in
industry. Interagency coordination seems capable of being improved: agencies
within the same government require more coordination, and all sources of
information need to be made available nationally and internationally. Some of
this may require legislation.
One approach to monitoring would be to select a few potentially high-risk
areas or industries in which notification would be compulsory. Data would be
collected on all possible anomalies of development, whether these were
observed at birth or later, and compiled to facilitate the finding of any
connection with any aspect of the environment. The discovery process would
involve value judgements. All women in the area would be monitored, and their
residences and places of work reported. All doctors, including both
obstetricians and other physicians responsible for postnatal health, would be
required to participate. Anything that conceivably could be defined as
developmental would be noted, and an improved format would be devised. A
record linkage system would permit lifetime follow-up of children born during
the monitoring. The provisions are of course idealized; to approach them
would require research and development of administrative procedures, as well
as convincing the public that their privacy is not being invaded.
To facilitate the exchange of information, the March of Dimes National
Foundation established the International Clearing House of Birth Defects
Monitoring System in 1975. The countries participating in this project are
Canada, France, Finland, Hungary, Israel, Norway, Sweden, England and Wales,
and the United States. The value of this exchange is, however, limited
because of the lack of uniformity in collecting and reporting of data,
defining the type of birth defects, etc., and the limited participation in
this program of the international community as a whole (91).
76
-------
X. CONCLUDING REMARKS
If a chemical is found to be teratogenic, should its use ever continue?
Under what circumstances? At what levels? These questions of public policy
are beyond the scope of this survey, but the answers will largely depend on
methods available for testing and in turn will influence the methodologies to
be sought.
This survey has revealed that no one routine of testing will reliably
indicate whether a chemical is potentially teratogenic for man. The available
methods are empirical, and much more fundamental knowledge is needed before
teratology testing becomes more rational. For example, extrapolation of data
from test animals to man might be closer if extrapolation were possible from
one animal species to another. The latter will require extensive comparative
teratologic and pharmacologic studies of various classes of chemicals in
various species of test animals. A complementary requirement is much improved
knowledge of human pharmacology and pharmacologic polymorphisms.
Thus the relationship of screening to research is important, for research
is needed to form a basis for organized development of screening methods in
teratology.
Even the definition of teratology has caused problems. If teratology
results from harmful interference with development of the organism, then to
define it by reference to the moment of birth can be misleading. Development
is rarely, if ever, complete at birth in any mammalian species, and humans are
born at a different stage of development from most test animals. Better
criteria of when development is complete might be: (a) the last mitotic cell
division in neurons, or (b) the completion of myelination in myelinated
neurons. (Sexual maturity is not considered relevant.) This difficulty of
definition will be minimized if teratology is considered in proper
relationship to other areas of toxicology.
77
-------
Only by standardizing teratogenic test procedures, by using the best
available methodology and well-trained personnel can the existing
confusion in comparing teratogenic data, emanating from various
laboratories using different animal species, be resolved.
Exposure of the public to teratogenic chemicals appears to be rare,
but known occasions have included major disasters. With the continued
development and release of potent chemicals into the environment,
continued research into screening methods appears to be essential.
Finally, the limitations that exist in predicting human results on
the basis of teralogical data obtained from animals led teratologists to
oppose legislation that would apply the "Delaney regulation" to
teratogens (148). Extension of a Delaney-type clause to include
teratogens (such as presently exists for tumor-inducing agents) would
have decreed that any agent found to cause malfornati-ons at any dose in
any experimental animal must be categorized as hazardous and be
prohibited from human use. As stated by Karnofsky (73), "the purpose of
evaluating drugs for teratogenicity is not to eliminate from use the
drugs which show teratogenic potential but rather to assess the hazard
their use presents to the human fetus."
78
-------
REFERENCES
1. Amin-Zaki, L., S. Elhassani, M.A..Majeed, T.W. Clarkson, R.A. Doherty,
and M. Greenwood. 1974. Intra-uterine me thy liner cury poisoning in
Iraq. Pediatrics 54(5): 587-595.
2. Anderson, D.H. 1949. Effects of diet during pregnancy upon the
incidence of congenital hereditary diaphragmatic hernia in the rat.
Am. J. Pathol. 25: 163-185.
3. Anon. 1971. Nutrient Requirements of Poultry, 6th Edition. National
Academy of Sciences, National Research Council, Washington, D.C.
4. Anon. 1972. Atlas of Nutritional Data on U.S. and Canadian Feeds.
National Academy of Sciences, National Research Council, Washington,
D.C.
5. Anon. 1973. The Testing of Chemicals for Carcinogenieity,
Mutagenicity, Teratogenicity. Health Protection Branch, Ministry of
National Health and Welfare, Canada, pp. vi and 185.
6. Anon. 1971. How Paraquat gets into the lung. New Scientist
(December 12, 197*0: 797.
7. Armstrong, P.B. 1971. Scanning electron microscope technique for
study of internal microanatomy of embryos. Microscope 19: 281.
8. Armstrong, R.C., and J.J. Elias. 1968. Development of embryonic rat
eyes in organ culture. I. Effect of glutamine on normal and abnormal
^development in a chemically defined medium. J. Embryol. Exp. Morphel.
19(3): 397-405.
9. Asling, C.W. 1968'.' Nutrition and teratogenesis. In: Methods for
Teratological Studies in Experimental Animals and Man. H. Nishimura,
J.R. Miller, and H. Yasuda, eds. Medical Examination Publishing Co.,
Flushing, N.Y. pp. 76-91.
10. Becker, B.A. 1971. The statistics of teratology. Teratology 9(3):
261-262.
11. Briggs, J.C., and J.G. Wilson. 1959. Comparison of the teratogenic
effects of trypan blue and low temperature in the medakafish (Oryzios
latipes). Quart. J. Fla. Acad. Sci. 22: 5*-68.
12. Burde, R.M., B. Schainker, and J. Kayes. 1971. Acute effect of oral
and subcutaneous administration of monosodium glutamate on the arcuate
nucleus of the hypothalamus in mice and rats. Nature 233: 58-60.
13. Cahen, R.L. 1964. Evaluation of the teratogenicity of drugs. Clin.
Pharmacol. Therap. 5: 480-514.
14. Cahen, R.L. 1966. Experimental and clinical chemoteratogenesis. Adv.
Pharmacol. 4: 264-349.
79
-------
15
16.
17.
18.
19.
20.
21.
22.
23.
25.
26.
27.
28.
29.
Clayton, F.C., and R.T. Francoeur. 1971. Some teratogenic effects of
vinblastine on the external morphology of Drosophilia melanogaster.
Oncology 25: 188-192.
Clegg, D.J. 1971. Teratology. Ann. Rev. Pharmacol. 11: 409-424.
Collaborative Study by 16 Laboratories. 1966. The evaluation of drugs
for foetal toxicity and teratogenicity in the rat. Proc. Eur. Spc.
Study Drug Tox. 7: 216-228.
Chaube, S., and M.L. Murphy. 1969. The teratogenic effects of recent
drugs active in cancer chemotherapy. In: Advances in Teratology.
D.H.M. Hoollam, ed. Logos Press Ltd., London. Vol 3« p. 205.
.Collins, T.F.X. , and C.H. Williams. 1971. Teratogenic studies with
2,4,5-T and 2,4-D in the hamster. Bull. Environ. Contain. Toxicol.
6(6): 559-567.
Cooke, A.S. 1972. The effects of DDT, Dieldrin, and 2,4-D on
amphibian spawn and tadpoles. Environ. Pollut. 3(1): 51-68.
Cooke, A.S. 1973. Effect of DDT, when used as a mosquito larvacide,
on tadpoles of the frog Rana temporaria. .Environ. Pollut. 5: 259-273.
Critchfield, C. , and J.C. Daniel, Jr. 1965. Teratogenic effects of
trypan blue on Coturnix quail when injected into the mother. Growth
29(3): 301-309.
Dawson, A.B. 1926. A note on the staining of the skeleton of cleared
specimens with alizarin red S. Stain Technol. 1: 123-124.
Delahunt, C.S. 1970. Detection of teratogenic actions. In: Methods
in Toxicology. G.E. Paget, ed. Blackwell Scientific Publications,
Oxford, England, pp. 132-157.
Domac-Tesar, B., and Z. Frank. 1963. Effects of local application
hypothermia during embryonic development of the white rat (Abstract) .
Acta Anat. 54: 353.
Dostal, M. , and R. Jelinek. 1971. Induction of cleft palate in rats
with intra-amniotic corticoids. Nature 230:
Duby, R.T., and H.F. Travis. 1971. Influence of dienestrol diacetate
on reproductive performance of female mink (Mustela vision) . Am. J.
Vet. Res. 32(10): 1599-1602.
Evans, M.A., M.W. Stevens, B. Mantilla-Plata, and R.D. Harbison.
1975. Drugs of abuse: teratogenic and mutagenic considerations.
Addictive Diseases. 2: 45-61.
Faherty, J.F. 1972. Surface staining of 1 mm (Wilson) slices of
fetuses for internal visceral examination. Stain Technol. 47: 53-58.
80
-------
30. Feldmann, B.M., L.M. Dudman, and M.S. Redfearn. 1971. Canine, feline
and lab primate dry feeds: guidelines to evaluation. Lab. Anim. Sci.
21(6): 862-861.
31. Ferm, V.H. 1965. The rapid detection of teratogenic activity. Lab.
Invest. 14: 1500-1505.
32. Perm, V.H. 1967. Potentiation of the teratogenic effect of vitamin A
with exposure to low environmental temperature. Life Scl. 6: 493-497.
33- Perm, V.H., and A.R. Beaudoin. 1965. Studies on the effect of trypan
blue in the pregnant armadillo, Dasypus novemcinctus. Anat. Rec.
15 KM): 571-578.
34. Ferngren, H., and U. Forsberg. 1971. The correlation of adverse
effects in man with observations in animals. Evaluation in animals of
teratogenic effects of drugs submitted to the Swedish Drug Control
1963-1968. Proc. Eur. Soc. Study Drug Tox. 12: 347-351.
35. Fitzhugh, O.G. 1968. Reproduction tests. In: Modern Trends in
Toxicology. E. Boyland, and R. Groulding, eds. Butterworth, London.
vol. 1. pp. 75-85.
36. Flynt, J.W., Jr. 1974. Trends in surveillance of congenital
malformations. In: Congenital Defects. D.T. Janerich, R.G. Skalko,
and I.H. Porter, eds. Academic Press, New York. pp. 119-127.
37. Flynt, J.W., Jr., A.J. Ebbin, G.P. Oakley, Jr., A Falek, and C.W.
Heath. 1971. Metropolitan Atlanta congenital defects program. In:
Monitoring, Birth Defects and Environment. E.B. Hook, D.T. Janerich,
and I.H. Porter, eds. Academic Press, New York. pp. 155-158.
38. Forsberg, U., and C.AV Grant. 1971. The correlation of adverse
effects in man with observations in animals. Current practices of
pharmaceutical companies for detecting and reporting foetal anomalies
obtained in animal tests. Proc. Eur. Soc. Study Drug Tox. 12:
352-356.
39. Fraser, F.C. 1961. Experimental teratogenesis in relation to
congenital malformations in man. In: Second International Conference
on Congenital Malformations. Lippincott, Philadelphia, pp. 277-287.
40.. Fraser, F.C., and V.A. McKusick, eds. 1970. Congenital Malformations.
Excerpta Medica, North Holland Publishing Co., Amsterdam.
41. Freireich, E.F., E. Frei, III, and M. Karon. 1962. Methylglyoxal
bis-(guanylhydrazone): a new agent active against acute myelocytic
leukemia. Cancer Chemother. Rep. 16: 183-186.
42. Friedman, L., C.M. Kunin, N. Nelson, J.L. Whittenberger, and J.G.
Wilson. 1970. Food and Drug Administration Advisory Committee on
Protocols for Safety Evaluations: Panel on Reproduction report on
reproduction studies in the safety evaluation of food additives and
pesticide residues. Toxicol. Appl. Pharmaeol. 16: 264-196.
81
-------
U3. Gebhardt, D.O.E., and J. Faber. 1966. The influence of aminopterin on
limb regeneration in Ambystoma mexicanun. J. Embryo!. Exp. Morphol.
16(1): 1M3-158.
14. Gehring, P.G., V.K. Howe, and S.B. McCollister. 1973. Toxicology:
Cost/time. Food Cosmet. Toxieol. 11: 1097-1110.
%5. Gilbert, E.F., B.C. Pitot, H.J. Bruyere, Jr., and A.L. Cheung. 1973.
Teratogenic effects of 5-bromodeoxyuridine on the external morphology
of Drosophila melanogaster. Teratology 7: 205-208.
46. Giroud, A. 1970. Nutrition of the Embryo. Thomas, Springfield, 111.
47. Gofmekler, V.A. 1969. Methods for studying the effects of chemical
substances on embryos. Hyg. Sanit. 3M: 198-202.
U8. Goldstein. A., L. Aronow, and S.M. Kalman. 1969. Principles of Drug
Action. Harper and Row,New York. pp. 180.
M9. Goodhart, R.S. 1961. Criteria of an adequate, diet. In: Modern
Nutrition in Health and Disease. M.G. Wohl, and R.S. Goodhart, eds.
Lea & Febiger, Philadelphia, pp. 617-632.
50. Greenhouse, G., and M. Hamburgh. 1968. Analysis of trypan blue
induced teratogenesis in Rang pipiens embroys. Teratology 1: 61-7M.
51. The effect of serotonin on the offspring and the fetal-maternal
relationship in rats. Arzneim.-Forsch. 25(U): 626-632.
52. Hagstrom, B.E., and S. Lcnning. 1973. The sea urchin egg as a testing
object in toxicology. Acta Pharm. Toxieol. 32(Suppl. 1): 1-M9.
53. Hamburgh, M., L.A. Mendoza, M. Rader, A. Lang, H. Silverstein, and K.
Hoffman. 197M. Malformations induced in offspring of crowded and
parabiotically-stressed mice. Teratology 10(1): 31-37.
5M. Hanley, T., V. Udall, and M. Weatherall. 1970. An industrial view of
current practice in predicting drug toxicity. Br. Med. Bull. 26:
203-207.
55. Harbison, R.D., and B.A. Becker. 1971. Effects of phenobarbital and
SKF 525A pretreatment on phenylhydantoin disposition in pregnant mice.
Toxieol. Appl. Pharmacol. 20: 571-581.
56. Hart, N.H., and M. Greene. 1971. LSD: teratogenic action in chick
blastoderms. Proe. Soc. Exp. Biol. Med. 137(2): 371-373-
57. Haseman, J.K., and M.D. Hogan. 1975. Selection of experimental unit
in teratology studies. Teratology 12: 165-171.
58. Hendrickx, A.G., L.R. Axelrod, and L.D. Clayborn. 1966. Thalidomide
syndrome in baboons. Nature 210: 958-959.
59. Hill, G.B., C.C. Spicer, and J'.A.C. Weatherall. 168. The computer
surveillance of congenital malformations. Br. Med. Bull. 21: 215-218.
82
-------
60. Horar, R.M., and A.J. Salem. 1961. Time of teratogenic action of
trypan blue in guinea pigs. Anati Rec. 111: 173-181.
61. Humphrey, R.H. 1967. Genetic andexperimental studies on a lethal
trait ("short toes") in the Mexican axolotl (Ambystoma mexieanum). J.
Exp. Zool. 161: 281-296.
62. Hurley, L.S. 1965. Nutritional Factors in Teratogenesis. Supplement
to the Teratology Workshop Manual, Pharmaceutical Manufacturers Assn.,
Washington, D.C. pp. 111-120.
63. Hurley, L.S., and H. Swenerton. 1966. Congenital malformations
resulting from zinc deficiency in rats. Proe. Soc. Exp. Biol. Med.
123: 692-696.
61. Ingalls, T.H., and F.J. Curley, 1957. Principles governing the genesis
of congenital malformations induced in mice by anoxia. New Engl. J.
Med. 257: 1121-1127.
65. Jensh, R.P., R.L. Brent, and M. Barr. 1970. The litter effect as a
variable in teratologic studies of the albino rat. Am. J. Anat. 128:
185-192.
66. Jones, D.C. 1969- Persistent and late effects of whole-body
irradiation in juvenile male rats. In: Radiation Biology of^ the Fetal
and Juvenile Mammal. M.R. Sikov and D.D. Mahlum, eds. U.S. Atomic
Energy Commission, Washington, D.C., pp. 139-117.
67. Jones, R.W., and M.N. Huffman. 1957. Fish embryos as bioassay
material in testing chemicals for effect on cell division and
differentiation. Trans. Am. Microscop. Soc. 76: 177-183.
68. Kalter, H. 1977. The structure and uses of genetically homogenous
lines of animals. Irf: Handbook of Teratology. J. C. Wilson and F.C.
Fraser, eds. Plenum Press, New York.
69. Kalter, H. 1968. Teratology of the Central Nervous System.
University of Chicago Press, Chicago, 111. :
70. Kalter, H. 1975. Choice of the number of sampling units in
teratology. Teratology 9(3): 257-258.
71. Kalter, H. 1975. Personal communication.
72. Kalter, H., and J. Warkany. 1959. Experimental production of
congenital malformations by metabolic procedures. Physiol. Rev. 39:
69-115.
73. Karnofsky, D.A. 1965. Drugs as teratogens in man. Ann. Rev.
Pharmacol. 5: 117-172.
71. Karnofsky, D.A. 1965. The chick embryo in drug screening: survey of
teratological effects observed in the 1-day chick embryo. In:
Teratology: Principles and Techniques. J.G. Wilson, and J. Warkany,
eds. University of Chicago Press, pp. 191-213.
83
-------
75. Karnofsky, D.A., and E.B. Simmel. 1963. Effects of growth-inhibiting
chemicals on the sand-dollar embryo (Echinarachinus parma). Prog. Exp.
Tumor Res. 3:2 254-295.
76. Karnofsky, D.A., and R.S. Basch. 1960. Effects of
5-fluorodeoxyuridine and related halogenated pyrimidines on the
sand-dollar embryo. J. Biophys. Biochem. Cytol. 7(1): 61-71.
77. Keeler, R.F. 1973- Lupin alkaloids from teratogenic and
nonteratogenic lupins. I. Correlation of crooked calf disease
incidence with alkaloid distribution determined by gas chromotography.
Teratology 7: 23-30.
78. Kelsey, P.O. 197U. Present guidelines for teratogenic studies in
experimental animals. In: Congenital Defects. D.T. Janerich, R.G.
Skalko, and I.H. Porter, eds. Academic Press, New York. pp. 195-201.
79. Kenaga, E.'E. 1976. Use of biological tests for evaluation of
pesticides. Pure Appl. Chem. 42: 285-299.
80. Kennedy, L.A., and T.V.N. Persaud. 1976. Experimental amniocentesis
and teratogenesis, Anat. Anz. Bd. 140: 267-276.
81. Khera, K.S. 1973. Teratogenic effects of me'thylmercury in the cat: a
note on the use of this species as a model for teratogenicity studies.
Teratology. 8(3): 293-303.
82. Kirkpatrick, J.F. 1973- Radiation induced abnormalities in early in
vitro mouse embryos. Anat. Rec. 176: 397-404.
83. Klein, M.tf. 1965. Teratological studies with explanted chick
embryos. In: Teratology; Principles and Techniques. J.G. Wilson,
and J. Warkany, eds. University of Chicago Press, pp. 131-1U1.
84. Krowke, R., and D. Neubert. 1977. Embryonic intermediary metabolism .
under normal and pathological conditions. In: Handbook of
Teratology. J.G. Wilson and F.C. Fraser, eds. Plenum Press, N.Y.
Vol. 2. pp. 117-151.
85. Kupper, L.L., and J.K. Baseman. 1977. On the use of a correlated
binomial model for the analysis of certain toxicological experiments.
"Tentatively Accepted" for publication in Biometrics.
86. Landauer, W. 1957. Phenocopies and genotype, with special reference
to sporadically-occurring developmental variants. Am. Natur. 91:
79-90.
87. Langman, J., W. Webster, and P. Rodier. 1975. Morphological and
behavioral abnormalities caused by insults to the CNS in the prenatal
period. In: Teratology: Trends and Applications. C.L. Berry, and
D.E. Poswillo, eds. Springer-Verlag, Berlin, Heidelberg, New York.
pp. 183-199.
84
-------
88. Lehman, J., D. Arnold, W. Passet, W. Horace, H. Guarde, E. Stockinger,
and J.W. Zapp. 1964. Principles 'and Procedures for Evaluating the
Toxicity of Household Substances. National Academy of Sciences,
Washington, D.C. .
89. Lemkey-Johnson, N., and W.A. Reynolds. 1972. Incidence and extent of
brain lesions in mice followoing ingestion of monosodium glutamate.
Anat. Rec. 172: 354.
90. Lin, P.O., and J.K. Baseman, 1976. A modified Jonckheere test against
ordered alternatives when ties are present at a single extreme value.
Biom. Z. Bd. 8: 623-631.
91. Lustig, E.S. 1971. Teratogenic effects induced in tail of Bufo
arenarum tadpoles following treatment wth carcinogens. Experientia
27: 555-556.
92. Milham, S. 1963. Malformation surveillance in New York State. New
York State Med. J. 63: 2823-2824.
93. Miller, B.C., and K. Hassanein. 1971. Diagnosis of impaired fetal
growth in newborn infants. Pediatrics 48: 511-522.
94. Miller, J.R. 1976. Birth defect monitoring system: An overview.
Teratology 14: 235.
95. Miller, R.W. 1971. Monitoring congenital malformations. In:
Monitoring, Birth Defects and Environment. E.B. Hook, D.T. Janerich,
and I.H. Porter, eds. Academic Press, New York. pp. 97-111.
96. Mitchell, J.T., and C.L. Yntema. 1973. Teratogenic effect of
malathion and captan in the embryo of the common snapping turtle
Chelydra serpentina. Anat. Ree. 175(2): 390.
97. Mnatsakanyan, A.V., U.G. Pogosyna, K.Kh. Akopyan, V.A. Gofmekler, A.O.
Avoyan, and M.S. Andikyan. 1971. Certain aspects of the embryotoxic
effect of chloroprene in a field experiment. Hyg. Sanit. 36(7):
140-141.
98. Moni, I.W. 1965. Comparative development of rat, chick, and human
embryos. In: Supplement to Teratology Workshop Manual. PMA, Berkley.
pp. 154.
-99. Mould, G.P., S.H. Curry, and P. Beck. 1973. The ferret, a useful
model for teratogenic study. Naunyn-Schmiedeberg Arch. Pharmacol.
279(24): R-18.
100. Mussman, H.C. 1975. Drug and chemical residues in domestic animals.
Fed. Proc. 34(2): 197-201.
101. March of Dimes National Foundation. 1975. Facts. National
Foundation. New York.
85
-------
102. Neuweiler, W., and R.H.H. Richter. 1961. Etiology of gross
malformations. Schweiz. Med. Wochenachr. 91: 359-363.
103. New, D.A.T. 1955. A new technique for the cultivation of the chick
embryo in vitro. J. Embryol. Exp. Morphol. 3: 326-331.
101. New, D.A.T. 1966. Culture of Vertebrate Embryos. Academic Press, New
York, N.Y.
105. New, D.A.T., and K.F. Stein. 1963. Cultivation of mouse embryos in
vitro. Nature 199: 297-299.
106. New, D.A.T., and R.L. Brent. 19.72. Effect of yolk-sac antibody on rat
embryos grown in culture. J. Embryol. Exp. Morphol. 27: 5^3-553.
107. Nielsen, N.O. 1969. Teratogenic effects of hypothermia in chick
embryos. In: Teratology. A. Bertelli, and L. Donati, eds. Excerpta
Med. Found.., Amsterdam, pp. 102-107.
108. Nishimura, H., and T. Tanimura. 1976. Clinical Aspects of the
Teratogenicity of Drugs. Excerpta Medica, Amsterdam (American Elsevier
Publishing Co. Inc., New York). pp. 271-298.
109. Nishimura, H., and Y. Yasuda. 1969. Culture-of human and mouse limb
buds as a tool for testing the embryotoxicity of drugs. In: Methods
for Teratological Studies in Experimental Animals and Man. Medical
Examination Publishing Co., Flushing, N.Y. pp. 272-273.
110. Wishimura, H., J.R; Miller, and M. Yasuda, eds. 1969. Methods for
Teratological Studies in Experimental Animals and Man. Medical
Examination Publishing Co., Flushing, N.Y.
111. Noback, G.J. 1916-1917. The use of the Van Wijhe method for the
staining of the cartilaginous skeleton. Anat. Rec. 11: 292-29^.
112. Nomura, T. 1968. Management of animals for use in teratological
experiments. In: Methods for Teratological Studies_in Experimental
Animals and Man. Medical Examination Publishing Co." Flushing, N.Y.
pp. 3-15.
113- Olney, J.W. 1971. Glutamate-induced neuron necrosis in the infant
mouse hypothalamus. J. Neuropathol. Exp. Neurol. 30(1): 75-90.
114. Olney, J.W., and 0-L. Ho. 1970. Brain damage in infant mice following
oral intake of glutamate, asparatate, or cysteine. Nature 227:
609-611.
115. Palmer, A.K. 1969. The relationship between screening tests for drug
safety and other teratological investigations. In: Teratology. A.
Bertelli, and L. Donati, eds. Excerpta Medica Foundation, Amsterdam.
pp. 55-72.
86
-------
116. Paschke, M., and F. Schwanitz. 1968. On the problems of pretesting
(preliminary testing) drugs with the acid of lower and higher plants.
I. Effect of phthalylglutamic aci'd imide (thalidomide) and other
sedatives and hypnotics on rate of cell division, shape, and sexuality
of some Desnidiaceae'. Arzn.-Forsch'. 18(10): 1338-1311.
117. Peck, H.M. 1963. Preclinical evaluation of drugs for evidence of
teratogenic activity. J. Pharm. Sci. 52: 1115-1120.
118. Persaud, T.V.N. 1969. Fetal toxicity of leucine in the rat. West
Indian Med. J. 18(1): 31-39.
119. Potts, E.L. 1962. Pathology of the Fetus and the Infant. Year Book
Medical Publishers, Inc., Chicago, 111.
120. Rieselbach, R.E., E.E. Morse, D.P. Rail, E. Frei, III, and E.J.
Freireich. 1962. Treatment of meningeal leukemia with intrathecal
aminopterin. Cancer Chemother. Rep. 16: 191-196.
121. Robkin, M.A., T.H. Shepard, and T. Tanimura. 1972. A new in vitro
culture technique for rat embryos. Teratology 5: 367-376.
122. Robson, J.M. 1970. Testing drugs for teratogenicity and their effect
on fertility. Br. Med. Bull. 26: 212-216.
123. Rodier, P., W. Webster, and J. Langman. Morphological and behavioral
consequences of chemically-induced lesions of the CNS. In: Proe. 7th
Gatlinburg Conf. on Research and Theory in Mental Retardation. N.
Ellis, ed. Lawrence Erlbaum Assoc., Potomac, Md.
121. Rosenkrantz, J.G., W.W. Frost, and J.V. Krupp. Unpublished. Effects
of trypan blue on the Pitman-Moore miniature pig.
125. Runner, N.M. 1971. '-The chick embryo as a test system for teratogenic
drugs. In: Proc. 3rd Teratology Workshop. Boulder, Colorado.
I ;
126. Runner, N.M. 1959. Inheritance of susceptibility to congenital
deformity. Metabolic clues provided by experiments with teratologenic
agents. Pediatrics 23: 215-251.
127. Saxen, L., M. Karkinen-Joaskelainen, and I. Saxen. 1976. Organ
culture in teratology. Curr. Top. Pathol. 62: 123-113-
128. Schardein, J.L. 1976. Drugs as Teratogens. CRC Press, Cleveland,
Ohio. p. 22-25.
129. Schnell, V., and J.W. Newberne. 1970. Accelerated clearing and
staining of teratologic specimens by heat and light. Teratology 3'
315-317.
130. Shepard, T.H. 1973. Catalog of Teratogenic Agents. Johns Hopkins
University Press, Baltimore, Maryland.
87
-------
131. Shimizu, K., A. Mizutani, and M. Inoup. 1973. Electron microscopic
studies on the hypothalamic lesions in the mouse fetus caused by
monbsodium glutamate (Abstract). Teratology 8: 105.
132. Sit, K.H., an R. Kanagasuntheram. 1972. A structural analysis of
congenital limb deformities in experimental hyperthyroid tadpoles. J.
Eabryol. Exp. Morphol. 28: 223-234.
133. Sladecek, F., and A. Romanovsky. I960. The effect of cold and heat
shock on the eggs of axolotl, recorded to early larval stages. Folia
Biol. (Prague) 6: 42-49.
134. Smithberg, M. 1962. Teratogenic effects of tolbutamide on early
development of the fish, Oryzias latipes. Am. J. Anat. 111(2):
205-213.
135. Somers, J., E.T. Moran, Jr., and B.S. Reinhart. 197**. Effect of
external application of pesticides to the fertile egg on hatching
success and early chick performance. 2. Commercial-herbicide mixtures
of 2,4-D with picloram or 2,4,5-T using the pheasant. Bull. Environ.
Contam. Toxieol. 11(4): 339-342.
136. Spyker, J.M. 1975. Assessing the impact of low-level chemicals on
development: behavioral and latent effects. Fed. Proc. 34(9):
1835-1844.
137. Spyker, J.M. 1975. Behavioral teratology and toxicology. Environ.
Sci. Res. 5: 311-349.
138. Staples, R.E., and J.K. Haseman. 1974. Selection of appropriate
experimental units in teratology. Teratology. 9(3): 259-260.
139. Steffek, A.J., and A.C. Verrusio. 1972. Experimentally induced
oral-facial malformations in the ferret (Mustela putorius furo).
Teratology 5: 268.
'140. Stewart, R.J.C., R.F. Preece, and H.G. Sheppard. 1975. Twelve
generations of marginal protein deficiency. Br. J. Hutr. 33: 233-253.
141. Streeter, G.L. 1933. A human embryo of the presoaite period.
Carnegie Contributions. Embrypl. 9: 389.
142. Stula, E.F., and W.C. Kraus. 1977. Embryotoxicity in rats and rabbits
from cutaneous application of amide-type solvents and substituted
ureas. Toxieol. Appl. Pharmacol. 41: 35-55.
143. Sullivan, P.M., and P.R. McElhatton. 1975. Teratogenic activity of
the antiepileptic drugs phenobarbital, phenytoin and primidone in
mice. Toxieol. Appl. Pharmacol. 34: 271-282.
144. Sullivan, P.M. 1974. Animal tests to screen for human teratogens.
Pediatrics 53: 822-823-
88
-------
145. Szabo, K.T., and A.K. Palmer. 1974. Teratological evaluation of
ophthalmic preparations by the ocular, dermal, and intraperitoneal
routes of administration in the rabbit. Teratology 9(3): A38-A39.
146. Takagaki, S., and H. Kanamori. 1957. Developmental anomalies of
embryos of Bufo vulgaris induced by 2,3-dimercaptopropanol (BAL). Acta
Anat. Nippon. 32(3): 326.
1U?. Tencer", R. 1961. The effects of 5-fluorodeoxyuridine on amphibian
embryos. Exp. Cell Res. 23: 418-419.
148. Teratology Society. 1971. Teratogens and Delaney clause. Teratology
10: 1.
149. Tinker, J. 1974. Vapona: How much is too much? New Scientist
(December 12, 1974).
150. Tracer Jitco, Inc. 1974. GRAS Monograph Series; Glutamates.
National Technical Information Service, Springfield, Va. No.
PB.229856/AS.
151. Tuchmann-Duplessis, H. 1966. Methods for evaluating teratogenic
properties of new drugs. In: Methods in Drug Evaluation. P.
Mantegazza, and F. Piccini, eds. North Holland Publishing Co.,
Amsterdam, pp. 11-31.
152. Tuchmann-Duplessis, H. 1970. The problems of species difference and
statistics. Proc. Eur. Soc. Study of Drug Toxicity. 11: 33-49-
153. Tuchmann-Duplessis, H. 1972. Teratogenic drug screening. Present
procedures and requirements. Teratology 5: 271-285.
154. Turbow, M.M., and J.G. Chamberlain. 1968. Direct effects of
6-aminonicotinamide on1 the developing rat embryo in vitro and in vivo.
Teratology 1: 103-108.
155. U.S. Dept. of Health, Education and Welfare. 1966. Guidelines for
Reproduction Studies for Safety Evaluation of Drugs for Human Use.
Food and Drug Administration, Washington, D.C.
156. Wainwright, S.D., L.K. Wainwright, and I.H. Fraser. 1973. Hemoglobin
formation in the explanted chick blastodisc. A model system for
evaluation of embryotoxic agents as illustrated by effects of
vinblastine, 1,1,1-trichloro-2,2-bis-(p-chlorophenyl)-ethane (DDT), and
sodium diethyl barbiturate. Toxicol. Appl. Pharmaeol. 25: 123-129-
157. Warkany, J. 1969. Experimental production of mammalian limb
malformations. In: Limb Development and Deformity. Chester A.
Swinyard, ed. Thomas, Springfield, 111. pp. 140-160.
158. Warkany, J. 1971. Congenital Malformations. Year Book Medical
Publishers, Inc., Chicago, 111.
159. Warkany, J. 1973. Congenital malformations of the brain produced by
zinc deficiencies in rats. Am. J. Ment. Defic. 77: 645-653.
89
-------
160. Weatherall, J.A.C. 1969. An assessment of the efficiency of
notification of congenital malformations. Med. Off. 121(6): 65-68.
161. Weatherall, J.A.C. 1970. The detection of incidence of congenital
malformations in the community. Proc. Roy Soe. Med. 62(12):
1251-1252.
162. Weil, C.S. 1970. Selection of the valid number of sampling units and
a consideration of their combination in toxicological studies involving
reproduction, teratogenesis or carcinogenesis. Fd. Cosmet. Toxieol.
8: 177-182.
163. Wilson, J.G. 1975. Critique of current methods for teratogenic
testing in animals and suggestions for their improvement. In: Methods
for Detection of Environmental Agents That Produce Congenital Defects.
T.H. Shepard, J.R. Miller, and M. Marois, eds. North Holland
Publishing Co., Amsterdam, pp. 29-49.
164. Wilson, J.G. 1965- Methods for administering agents and detecting
malformations in experimental animals. In: Teratology; Principles
and Techniques. J.G. Wilson and J. Warkany, eds. University of
Chicago Press, Chicago, 111.
165. Wilson, J.G. 1973- Environment and Birth Defects. Academic Press,
New York.
166. Wilson, J.G. 1974. Factors determining the teratogenicity of drugs.
Ann. Rev. Pharmacol. 14: 205-217.
167. Wilson, J.G. 1975. Personal communication.
168. Wilson, J.G. 1975. Reproduction and teratogenesis: Current methods
and suggested improvements. J. Assoc. Off. Anal. Chem. 58: 657-667.
169. Wilson, J.G., and J. Warkany, eds. 1965. Teratology: Principles and
Techniques. University of Chicago Press,1 Chicago, 111.; p. 279.
170. Wilson, J.G., T.H. Maren, K. Takano, and A. Ellison. 1068.
Teratogenic action of carbonic anhydrase inhibitors in the rat.
Teratology 1(1): 51-60.
171. Winick, M. 1974. Cellular growth during normal and abnormal
development of the brain. In: Biochemistry of the Developing Brain,
W. Himwich, ed. Marcel Dekker Inc., New York. Vol 2. p. 199.
172. Witschi, E. 1956. Development of the Vertebrates. W.B. Saunders,
Philadelphia.
173. World Health Organization. 1967. Principles for the Testing of Drugs
for Teratogenicity. WHO Tech. Rep. Ser. No.
90
-------
175.
Yoshida, Y. , T. Ohashi, K. Takano,, N. Maekawa, N. Toyoda, R. Hujimoto,
K. Yamamoto, S. Iwata, H. Ito, and Y. Mikami. 1957. Visceral
malformations produped by vital dyes, Janus green, trypan blue and
methylene blue in amphibian embryos. Mie Med. J. 7(2): 273-285.
Yoshihara, H. 1966. The effect of 5-fluroruracil on development of
tail vertebrae of mice in organ culture. Anat. Ree. 15**:
91
-------
INDEX
Azo dyes
Chemicals
Cleft Lip
Cleft palate
Cortisone
Death Certificate, infant
Death, intrauterine
Defects, brain
Defects, central nervous system
Defects, definitions
Defects, digits
Defects, eye
Defects, functional
Defects, skeletal
Defects, soft-tissue
Development, postnatal
Development, structural
Dietary factors
Dosage
Dose vehicle
Drugs
Economics
Smbryotoxicity
Pages
9
1-3, 6, 9, 10, 19-36, 51, 53, 54,
65, 69, 75, 77, 78.
47
18, 46
49
58, 60
2, 46, 48, 54
18
6, 26, 34,^ 40, 69, 74, 75
2, 11, 40-41, 47-48, 58
47
34-
2, 37, 38, 41, 67, 72, 74
18, 34, 38, 41, 47, 54, 72
18, 34, 38, 41-45, 54, 72-73
6-7, 11, 27, 32, 41-45, 74, 77
4, 12, 16, 26, 27, 34, 41, 43, 44
2, 6, 10, 22-23, 26, 35-38, 46, 54, 60
65, 58
6-9, 19-26, 34-36, 47, 53-55, 70, 71
8-9, 19, 26-33, 42, 55, 71
3, 38, 49, 53, 55, 66
16., 58, 61-63, 67, 73, 75
8, 21-23., 33, 48-49, 53, 55
92
-------
INDEX (Cont'd)
Environmental factors
Examination", biochemical
Examination, visual
Examination, histological
Exposure
Extrapolability
Fetus
Galactoflavin
Gestation
Glutamic acid
Growth retardation
Guidelines
Herbicides
Litter effect
Malathion
Malformations
Metabolic tolerance
Metabolism
Necropsy
Organogenesis
Paraquat
Pesticides
Pages
1-6, 8-9, 12, 20, 22-25, 34-38, 50-51,
58, 64-66, 68, 70-71, 76-77
45, 75
7, 11, 38-45, 54, 62, 71-72
\
7,...11, 38, 41-45, 62, 72
2-6, 8-9, 19-20, 22-24, 55, 59-60,
64-66, 68-71, 70-76
3-6, 12, 19-20, 22-23, 47-48, 50-51,
64, 70, 72, 74-75, 77
6, 27-29, 38-46, 54-55, 69-72, 74 '
13, 18
4, 28, 39, 42, 43, 53-55
75
2, 38-39, 46-47
26, 33, 53-56
9
49-50
51
2, 7-11, 13, 18, 29, 37, 39, 40-41,
45, 47-48, 55, 58, 66-67
8, 24
4-6, 10, 12, 24, 26, 30, 50-51, 54, 75
7, 39-41, 57, 67, 71-72
4-6, 10, 25, 27, 31, 35, 48, 53, 55, 75
9
60
93
-------
INDEX (Cont'd)
Pages
Placenta 6, 13, 23, 27, 40, 75
Radiation 2, 46, 66
Reporting, centers for 58-60, 66-68, 76
Reporting, defects in farm and 60, 67
domestic animals
Reporting, monitoring defects 57-60, 66-68, 76
Risk-benefit analysis 12, 13, 4g, 58, 61-63, 73, 75-76
Resorptions 71
Route of exposure or administration 6-7, 22-24, 35, 55, 70
Screening, chemicals 11, 19, 21-22, 24-25, 56, 61, 65-73,
74, 77 '
Spina bifida 34
Stillbirth 2, 53, 58-60, 67
Strontium-90 9
Surveillance 3, 4, 57-60, 65-68, 69, 76
Teratogen ' 2, 9, 35, 37-38, 47, 49, 59, 64, 77
Teratogenic ratio 8
Test animals 4-6, 8-9, 12-16, 28, 33-36, 37-38, 52
53-60, 64, 69-72, 74-75, 77
Testing, controls 6, 8-9, 26-36, 45-49, 54, 69-71
Testing, errors 7, 50, 67
Testing, methods 1-9, 12-45, 49-56, 65, 69-72, 74-75, 77
Testing, multigeneration 22
Testing, multilevel 12, 15, 75
Testing, results 1-3, 8-10, 38-40, 45-51
94
-------
Testing, variables
Thalidomide
Trypan blue
INDEX (Cont'd)
Pages
4-6, 11, 13, 16, 21, 33-38, 50
4, 16, 17-18, 48, 75
9, 49
95
------- |