[Federal Register Volume 61, Number 212 (Thursday, October 31, 1996)]
[Notices]
[Pages 56274-56322]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-27473]



[[Page 56273]]


_______________________________________________________________________

Part II





Environmental Protection Agency





_______________________________________________________________________



Reproductive Toxicity Risk Assessment Guidelines; Notice

  Federal Register / Vol. 61, No. 212 / Thursday, October 31, 1996 / 
Notices  

[[Page 56274]]



ENVIRONMENTAL PROTECTION AGENCY

[FRL-5630-6]


Guidelines for Reproductive Toxicity Risk Assessment

AGENCY: U.S. Environmental Protection Agency.

ACTION: Notice of availability of final Guidelines for Reproductive 
Toxicity Risk Assessment.

-----------------------------------------------------------------------

SUMMARY: The U.S. Environmental Protection Agency (EPA) is today 
publishing in final form a document entitled Guidelines for 
Reproductive Toxicity Risk Assessment (hereafter ``Guidelines''). These 
Guidelines were developed as part of an interoffice guidelines 
development program by a Technical Panel of the Risk Assessment Forum. 
They were proposed initially in 1988 as separate guidelines for the 
female and male reproductive systems. Subsequently, based upon the 
public comments and Science Advisory Board (SAB) recommendations, 
changes made included combining those two guidelines, integrating the 
hazard identification and dose-response sections, assuming as a default 
that an agent for which sufficient data were available on only one sex 
may also affect reproductive function in the other sex, expansion of 
the section on interpretation of female endpoints, and consideration of 
the benchmark dose approach for quantitative risk assessment. These 
Guidelines were made available again for public comment and SAB review 
in 1994. This notice describes the scientific basis for concern about 
exposure to agents that cause reproductive toxicity, outlines the 
general process for assessing potential risk to humans from exposure to 
environmental agents, and addresses Science Advisory Board and public 
comments on the 1994 Proposed Guidelines for Reproductive Toxicity Risk 
Assessment. Subsequent reviews have included the Agency's Risk 
Assessment Forum and interagency comment by members of subcommittees of 
the Committee on the Environment and Natural Resources of the Office of 
Science and Technology Policy. The EPA appreciates the efforts of all 
participants in the process and has tried to address their 
recommendations in these Guidelines.

EFFECTIVE DATE: The Guidelines will be effective October 31, 1996.

ADDRESSES: The Guidelines will be made available in the following ways:
    (1) The electronic version will be accessible on EPA's Office of 
Research and Development home page on the Internet at http://
www.epa.gov/ORD/WebPubs/repro/.
    (2) 3\1/2\-inch high-density computer diskettes in WordPerfect 5.1 
will be available from ORD Publications, Technology Transfer and 
Support Division, National Risk Management Research Laboratory, 
Cincinnati, OH; telephone: 513-569-7562; fax: 513-569-7566. Please 
provide the EPA No. (EPA/630/R-96/009a) when ordering.
    (3) This notice contains the full document. In addition, copies of 
the Guidelines will be available for inspection at EPA headquarters in 
the Air and Radiation Docket and Information Center and in EPA 
headquarters and regional libraries. The Guidelines also will be made 
available through the U.S. Government Depository Library program and 
for purchase from the National Technical Information Service (NTIS), 
Springfield, VA; telephone: 703-487-4650; fax: 703-321-8547. Please 
provide the NTIS PB No. (PB97-100093) when ordering.

FOR FURTHER INFORMATION CONTACT: Dr. Eric D. Clegg, National Center for 
Environmental Assessment--Washington Office (8623), U.S. Environmental 
Protection Agency, 401 M Street, S.W., Washington, DC 20460; telephone: 
202-260-8914; e-mail: [email protected].

SUPPLEMENTARY INFORMATION:

A. Application of the Guidelines

    The EPA is authorized by numerous statutes, including the Toxic 
Substances Control Act (TSCA), the Federal Insecticide, Fungicide, and 
Rodenticide Act (FIFRA), the Clean Air Act, the Safe Drinking Water 
Act, and the Clean Water Act, to regulate environmental agents that 
have the potential to adversely affect human health, including the 
reproductive system. These statutes are implemented through offices 
within the Agency. The Office of Pesticide Programs and the Office of 
Pollution Prevention and Toxics within the Agency have issued testing 
guidelines (U.S. EPA, 1982, 1985b, 1996a) that provide protocols 
designed to determine the potential of a test substance to produce 
reproductive (including developmental) toxicity in laboratory animals. 
Proposed revisions to these testing guidelines are in the final stages 
of completion (U.S. EPA, 1996a). The Organization for Economic 
Cooperation and Development (OECD) also has issued testing guidelines 
(which are under revision) for reproduction studies (OECD, 1993b).
    These Guidelines apply within the framework of policies provided by 
applicable EPA statutes and do not alter such policies. They do not 
imply that one kind of data or another is prerequisite for action 
concerning any agent. The Guidelines are not intended, nor can they be 
relied upon, to create any rights enforceable by any party in 
litigation with the United States. This document is not a regulation 
and is not intended to substitute for EPA regulations. These Guidelines 
set forth current scientific thinking and approaches for conducting 
reproductive toxicity risk assessments. EPA will revisit these 
Guidelines as experience and scientific consensus evolve.
    The procedures outlined here in the Guidelines provide guidance for 
interpreting, analyzing, and using the data from studies that follow 
the above testing guidelines (U.S. EPA 1982, 1985b, 1996a). In 
addition, the Guidelines provide information for interpretation of 
other studies and endpoints (e.g., evaluations of epidemiologic data, 
measures of sperm production, reproductive endocrine system function, 
sexual behavior, female reproductive cycle normality) that have not 
been required routinely, but may be required in the future or may be 
encountered in reviews of data on particular agents. The Guidelines 
will promote consistency in the Agency's assessment of toxic effects on 
the male and female reproductive systems, including outcomes of 
pregnancy and lactation, and inform others of approaches that the 
Agency will use in assessing those risks. More specific guidance on 
developmental effects is provided by the Guidelines for Developmental 
Toxicity Risk Assessment (U.S. EPA, 1991). Other health effects 
guidance is provided by the Guidelines for Carcinogen Risk Assessment 
(U.S. EPA, 1986a, 1996b), the Guidelines for Mutagenicity Risk 
Assessment (U.S. EPA, 1986c), and the Proposed Guidelines for 
Neurotoxicity Risk Assessment (U.S. EPA, 1995a). These Guidelines and 
the four cited above are complementary.
    The Agency has sponsored or participated in several conferences 
that addressed issues related to evaluations of reproductive toxicity 
data which provide some of the scientific bases for these risk 
assessment guidelines. Numerous publications from these and other 
efforts are available which provide background for these Guidelines 
(U.S. EPA, 1982, 1985b, 1995b; Galbraith et al., 1983; OECD, 1983; U.S. 
Congress, 1985, 1988; Kimmel, C.A. et al., 1986; Francis and Kimmel, 
1988; Burger et al., 1989; Sheehan et al., 1989; Seed et al., 1996). 
Also, numerous resources provide background information on the

[[Page 56275]]

physiology, biochemistry, and toxicology of the male and female 
reproductive systems (Lamb and Foster, 1988; Working, 1989; Russell et 
al., 1990; Atterwill and Flack, 1992; Scialli and Clegg, 1992; Chapin 
and Heindel, 1993; Heindel and Chapin, 1993; Paul, 1993; Manson and 
Kang, 1994; Zenick et al., 1994; Kimmel, G.L. et al., 1995; Witorsch, 
1995). A comprehensive text on reproductive biology also has been 
published (Knobil et al., 1994).

B. Environmental Agents and Reproductive Toxicity

    Disorders of reproduction and hazards to reproductive health have 
become prominent public health issues. A variety of factors are 
associated with reproductive system disorders, including nutrition, 
environment, socioeconomic status, lifestyle, and stress. Disorders of 
reproduction in humans include but are not limited to reduced 
fertility, impotence, menstrual disorders, spontaneous abortion, low 
birth weight and other developmental (including heritable) defects, 
premature reproductive senescence, and various genetic diseases 
affecting the reproductive system and offspring.
    The prevalence of infertility, which is defined clinically as the 
failure to conceive after one year of unprotected intercourse, is 
difficult to estimate. National surveys have been conducted to obtain 
demographic information about infertility in the United States (Mosher 
and Pratt, 1990). In their 1988 survey, an estimated 4.9 million women 
ages 15-44 (8.4%) had impaired fertility. The proportion of married 
couples that was infertile, from all causes, was 7.9%.
    Carlsen et al. (1992) have reported from a meta analysis that human 
sperm concentration has declined from 113 x 10\6\ per mL of semen prior 
to 1960 to 66 x 10\6\ per mL subsequently. When combined with a 
reported decline in semen volume from 3.4 mL to 2.75 mL, that suggests 
a decline in total number of sperm of approximately 50%. Increased 
incidence of human male hypospadias, cryptorchidism, and testicular 
cancer have also been reported over the last 50 years (Giwercman et 
al., 1993). Several other retrospective studies that examined semen 
characteristics from semen donors have obtained conflicting results 
(Auger et al., 1995; Bujan et al., 1996; Fisch et al., 1996; Ginsburg 
et al., 1994; Irvine et al., 1996; Paulsen et al., 1996; Van Waeleghem 
et al., 1996; Vierula et al., 1996). While concerns exist about the 
validity of some of those conclusions, the data indicating an increase 
in human testicular cancer, as well as possible occurrence of other 
plausibly related effects such as reduced sperm production, 
hypospadias, and cryptorchidism, suggest that an adverse effect may 
have occurred. However, there is no definitive evidence that such 
adverse human health effects have been caused by environmental 
chemicals.
    Endometriosis is a painful reproductive and immunologic disease in 
women that is characterized by aberrant location of uterine endometrial 
cells, often leading to infertility. It affects approximately five 
million women in the United States between 15 and 45 years of age. Very 
limited research has suggested a link between dioxin exposure and 
development of endometriosis in rhesus monkeys (Rier et al., 1993). 
Gerhard and Runnebaum (1992) reported an association in women between 
occurrence of endometriosis and elevated blood PCB levels, while a 
subsequent small clinical study found no significant correlations 
between disease severity in women and serum levels of halogenated 
aromatic hydrocarbons (Boyd et al., 1995).
    Even though not all infertile couples seek treatment, and 
infertility is not the only adverse reproductive effect, it is 
estimated that in 1986, Americans spent about $1 billion on medical 
care to treat infertility alone (U.S. Congress, 1988). With the 
increased use of assisted reproduction techniques in the last 10 years, 
that amount has increased substantially.
    Disorders of the male or female reproductive system may also be 
manifested as adverse outcomes of pregnancy. For example, it has been 
estimated that approximately 50% of human conceptuses fail to reach 
term (Hertig, 1967; Kline et al., 1989). Methods that detect pregnancy 
as early as eight days after conception have shown that 32%-34% of 
postimplantation pregnancies end in embryonic or fetal loss (Wilcox et 
al., 1988; Zinaman et al., 1996). Approximately 3% of newborn children 
have one or more significant congenital malformations at birth, and by 
the end of the first post-natal year, about 3% more are recognized to 
have serious developmental defects (Shepard, 1986). Of these, it is 
estimated that 20% are of known genetic transmission, 10% are 
attributable to known environmental factors, and the remaining 70% 
result from unknown causes (Wilson, 1977). Also, approximately 7.4% of 
children have low birth weight (i.e., below 2.5 kg) (Selevan, 1981).
    A variety of developmental alterations may be detected after either 
pre- or postnatal exposure. Several of these are discussed in the 
Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA, 1991), 
and developmental neurotoxicity is discussed in the Proposed Guidelines 
for Neurotoxicity Risk Assessment (U.S. EPA, 1996a). Relative to 
developmental reproductive alterations, chemical or physical agents can 
affect the female and male reproductive systems at any time in the life 
cycle, including susceptible periods in development. The reproductive 
system begins to form early in gestation, but structural and functional 
maturation is not completed until puberty. Exposure to toxicants early 
in development can lead to alterations that may affect reproductive 
function or performance well after the time of initial exposure. 
Examples include the actions of estrogens, anti-androgens or dioxin in 
interfering with male sexual differentiation (Gill et al., 1979; Gray 
et al., 1994, 1995; Giusti et al., 1995; Gray and Ostby, 1995). Adverse 
effects such as reduced fertility in offspring may appear as delayed 
consequences of in utero exposure to toxicants. Effects of toxic agents 
on other parameters such as sexual behavior, reproductive cycle 
normality, or gonadal function can also alter fertility (Chapman, 1983; 
Dixon and Hall, 1984; Schrag and Dixon, 1985b; U.S. Congress, 1985). 
For example, developmental exposure to environmental compounds that 
possess steroidogenic (Mattison, 1985) or antisteroidogenic (Schardein, 
1993) activity affect the onset of puberty and reproductive function in 
adulthood.
    Numerous agents have been shown to cause reproductive toxicity in 
adult male and female laboratory animals and in humans (Mattison, 1985; 
Schrag and Dixon, 1985a, b; Waller et al., 1985; Lewis, 1991). In adult 
males and females, exposure to agents of abuse, e.g., cocaine, disrupts 
normal reproductive function in both test species and humans (Smith, 
C.G. and Gilbeau, 1985). Numerous chemicals disrupt the ovarian cycle, 
alter ovulation, and impair fertility in experimental animals and 
humans. These include agents with steroidogenic activity, certain 
pesticides, and some metals (Thomas, 1981; Mattison, 1985). In males, 
estrogenic compounds can be testicular toxicants in rodents and humans 
(Colborn et al., 1993; Toppari et al., 1995). Dibromochloropropane 
(DBCP) impairs spermatogenesis in both experimental animals and humans 
by another mechanism. These and other examples of toxicant-induced 
effects on reproductive function have been reviewed (Katz and 
Overstreet, 1981; Working, 1988).

[[Page 56276]]

    Altered reproductive health is often manifested as an adverse 
effect on the reproductive success or sexual behavior of the couple 
even though only one of the pair may be affected directly. Often, it is 
difficult to discern which partner has reduced reproductive capability. 
For example, exposure of the male to an agent that reduces the number 
of normal sperm may result in reduced fertility in the couple, but 
without further diagnostic testing, the affected partner may not be 
identified. Also, adverse effects on the reproductive systems of the 
two sexes may not be detected until a couple attempts to conceive a 
child.
    For successful reproduction, it is critical that the biologic 
integrity of the human reproductive system be maintained. For example, 
the events in the estrous or menstrual cycle are closely interrelated; 
changes in one event in the cycle can alter other events. Thus, a short 
or inadequate luteal phase of the menstrual cycle is associated with 
disorders in ovarian follicular steroidogenesis, gonadotropin 
secretion, and endometrial integrity (McNatty, 1979; Scommegna et al., 
1980; Smith, S.K. et al., 1984; Sakai and Hodgen, 1987). Toxicants may 
interfere with luteal function by altering hypothalamic or pituitary 
function and by affecting ovarian response (La Bella et al., 1973a, b).
    Fertility of the human male is particularly susceptible to agents 
that reduce the number or quality of sperm produced. Compared with many 
other species, human males produce fewer sperm relative to the number 
of sperm required for fertility (Amann, 1981; Working, 1988). As a 
result, many men are subfertile or infertile (Amann, 1981). The 
incidence of infertility in men is considered to increase at sperm 
concentrations below 20  x  10\6\ sperm per mL of ejaculate. As the 
concentration of sperm drops below that level, the probability of a 
pregnancy resulting from a single ejaculation declines. If the number 
of normal sperm per ejaculate is sufficiently low, fertilization is 
unlikely and an infertile condition exists. However, some men with low 
sperm concentrations are able to achieve conception and many subfertile 
men have concentrations greater than 20  x  10\6\ illustrating the 
importance of sperm quality. Toxic agents may further decrease 
production of sperm and increase risk of impaired fertility.

C. The Risk Assessment Process and Its Application To Reproductive 
Toxicity

    Risk assessment is the process by which scientific judgments are 
made concerning the potential for toxicity to occur in humans. In 1983, 
the National Research Council (NRC) defined risk assessment as 
comprising some or all of the following components: hazard 
identification, dose-response assessment, exposure assessment, and risk 
characterization (NRC, 1983). In its 1994 report, Science and Judgment 
in Risk Assessment, the NRC extended its view of the paradigm to 
include characterization of each component (NRC, 1994). In addition, it 
noted the importance of an interactive approach that deals with 
recurring conceptual issues that cut across all stages of risk 
assessment. These Guidelines adopt an interactive approach by 
organizing the process around the components of hazard 
characterization, the quantitative dose-response analysis, the exposure 
assessment, and the risk characterization where hazard characterization 
combines hazard identification with qualitative consideration of dose-
response relationships, route, timing, and duration of exposure. This 
is done because, in practice, hazard identification for reproductive 
toxicity and other noncancer health effects include an evaluation of 
dose-response relationships, route, timing, and duration of exposure in 
the studies used to identify the hazard. Determining a hazard often 
depends on whether a dose-response relationship is present (Kimmel, 
C.A. et al., 1990). This approach combines the information important in 
comparing the toxicity of a chemical to potential human exposure 
scenarios identified as part of the exposure assessment. Also, it 
minimizes the potential for labeling chemicals inappropriately as 
``reproductive toxicants'' on a purely qualitative basis.
    In hazard characterization, all available experimental animal and 
human data, including observed effects, associated doses, routes, 
timing, and duration of exposure, are examined to determine if an agent 
causes reproductive toxicity in that species and, if so, under what 
conditions. From the hazard characterization and criteria provided in 
these Guidelines, the health-related database can be characterized as 
sufficient or insufficient for use in risk assessment (Section III.G.). 
This approach does not preclude the evaluation and use of the data for 
other purposes when adequate quantitative information for setting 
reference doses (RfDs) and reference concentrations (RfCs) is not 
available.
    The next step, the quantitative dose-response analysis (Section 
IV), includes determining the no-observed-adverse-effect-level (NOAEL) 
and/or the lowest-observed-adverse-effect-level (LOAEL) for each study 
and type of effect. Because of the limitations associated with the use 
of the NOAEL, the Agency is beginning to use an additional approach, 
the benchmark dose approach (Crump, 1984; U.S. EPA. 1995b), for a more 
quantitative dose-response evaluation when allowed by the data. The 
benchmark dose approach takes into account the variability in the data 
and the slope of the dose-response curve, and thus, provides more 
complete use of the data for calculation of the RfD or RfC. If the data 
are considered sufficient for risk assessment, and if reproductive 
toxicity occurs at the lowest toxic dose level (i.e., the critical 
effect), an RfD or RfC, based on adverse reproductive effects, could be 
derived. This RfD or RfC is derived using the NOAEL or benchmark dose 
divided by uncertainty factors to account for interspecies differences 
in response, intraspecies variability and deficiencies in the database.
    Exposure assessment identifies and describes populations exposed or 
potentially exposed to an agent, and presents the type, magnitude, 
frequency, and duration of such exposures. Those procedures are 
considered separately in the Guidelines for Exposure Assessment (U.S. 
EPA, 1992). However, unique considerations for reproductive toxicity 
exposure assessments are detailed in Section V.
    A statement of the potential for human risk and the consequences of 
exposure can come only from integrating the hazard characterization and 
quantitative dose-response analysis with human exposure estimates in 
the risk characterization. As part of risk characterization, the 
strengths and weaknesses in each component of the risk assessment are 
summarized along with major assumptions, scientific judgments, and to 
the extent possible, qualitative descriptions and quantitative 
estimates of the uncertainties.
    In 1992, EPA issued a policy memorandum (Habicht, 1992) and 
guidance package on risk characterization to encourage more 
comprehensive risk characterizations, to promote greater consistency 
and comparability among risk characterizations, and to clarify the role 
of professional judgment in characterizing risk. In 1995, the Agency 
issued a new risk characterization policy and guidance (Browner, 1995) 
that refines and reaffirms the principles found in the 1992 policy and 
outlines a process within the Agency for implementation. Although 
specific program policies and procedures are still evolving, these 
Guidelines discuss attributes of the Agency's risk

[[Page 56277]]

characterization policy as it applies to reproductive toxicity.
    Risk assessment is just one component of the regulatory process. 
The other component, risk management, uses risk characterization along 
with directives of the enabling regulatory legislation and other 
factors to decide whether to control exposure to the suspected agent 
and the level of control. Risk management decisions also consider 
socioeconomic, technical, and political factors. Risk management is not 
discussed directly in these guidelines because the basis for 
decisionmaking goes beyond scientific considerations alone. However, 
the use of scientific information in this process is discussed. For 
example, the acceptability of the margin of exposure (MOE) is a risk 
management decision, but the scientific bases for generating this value 
are discussed here.

    Dated: October 15, 1996.
Carol M. Browner,
Administrator.

Contents

List of Tables
Part A. Guidelines for Reproductive Toxicity Risk Assessment
I. Overview
II. Definitions and Terminology
III. Hazard Characterization for Reproductive Toxicants
    III.A. Laboratory Testing Protocols
    III.A.1. Introduction
    III.A.2. Duration of Dosing
    III.A.3. Length of Mating Period
    III.A.4. Number of Females Mated to Each Male
    III.A.5. Single- and Multigeneration Reproduction Tests
    III.A.6. Alternative Reproductive Tests
    III.A.7. Additional Test Protocols That May Provide Reproductive 
Data
    III.B. Endpoints for Evaluating Male and Female Reproductive 
Toxicity In Test Species
    III.B.1. Introduction
    III.B.2. Couple-Mediated Endpoints
    III.B.2.a. Fertility and Pregnancy Outcomes
    III.B.2.b. Sexual Behavior
    III.B.3. Male-Specific Endpoints
    III.B.3.a. Introduction
    III.B.3.b. Body Weight and Organ Weights
    III.B.3.c. Histopathologic Evaluations
    III.B.3.d. Sperm Evaluations
    III.B.3.e. Paternally Mediated Effects on Offspring
    III.B.4. Female-Specific Endpoints
    III.B.4.a. Introduction
    III.B.4.b. Body Weight, Organ Weight, Organ Morphology, and 
Histology
    III.B.4.b.1. Body weight
    III.B.4.b.2. Ovary
    III.B.4.b.3. Uterus
    III.B.4.b.4. Oviducts
    III.B.4.b.5. Vagina and external genitalia
    III.B.4.b.6. Pituitary
    III.B.4.c. Oocyte Production
    III.B.4.c.1. Folliculogenesis
    III.B.4.c.2. Ovulation
    III.B.4.c.3. Corpus luteum
    III.B.4.d. Alterations in the Female Reproductive Cycle
    III.B.4.e. Mammary Gland and Lactation
    III.B.4.f. Reproductive Senescence
    III.B.5. Developmental and Pubertal Alterations
    III.B.6. Endocrine Evaluations
    III.B.7. In Vitro Tests of Reproductive Function
    III.C. Human Studies
    III.C.1. Epidemiologic Studies
    III.C.1.a. Selection of Outcomes for Study
    III.C.1.b. Reproductive History Studies
    III.C.1.c. Community Studies and Surveillance Programs
    III.C.1.d. Identification of Important Exposures for 
Reproductive Effects
    III.C.1.e. General Design Considerations
    III.C.2. Examination of Clusters, Case Reports, or Series
    III.D. Pharmacokinetic Considerations
    III.E. Comparisons of Molecular Structure
    III.F. Evaluation of Dose-Response Relationships
    III.G. Characterization of the Health-Related Database
IV. QUANTITATIVE DOSE-RESPONSE ANALYSIS
V. EXPOSURE ASSESSMENT
VI. RISK CHARACTERIZATION
    VI.A. Overview
    VI.B. Integration of Hazard Characterization, Quantitative Dose-
Response, and Exposure Assessments
    VI.C. Descriptors of Reproductive Risk
    VI.C.1. Distribution of Individual Exposures
    VI.C.2. Population Exposure
    VI.C.3. Margin of Exposure
    VI.C.4. Distribution of Exposure and Risk for Different 
Subgroups
    VI.C.5. Situation-Specific Information
    VI.C.6. Evaluation of the Uncertainty in the Risk Descriptors
    VI.D. Summary and Research Needs
VII. REFERENCES
PART B. RESPONSE TO SCIENCE ADVISORY BOARD AND PUBLIC COMMENTS
I. INTRODUCTION
II. RESPONSE TO SCIENCE ADVISORY BOARD COMMENTS
III. RESPONSE TO PUBLIC COMMENTS

List of Tables

1. Default Assumptions in Reproductive Toxicity Risk Assessment
2. Couple-Mediated Endpoints of Reproductive Toxicity
3. Selected Indices That May Be Calculated From Endpoints of 
Reproductive Toxicity in Test Species
4. Male-Specific Endpoints of Reproductive Toxicity
5. Female-Specific Endpoints of Reproductive Toxicity
6. Categorization of the Health-Related Database
7. Guide for Developing Chemical-Specific Risk Characterizations for 
Reproductive Effects

PART A. GUIDELINES FOR REPRODUCTIVE TOXICITY RISK ASSESSMENT

I. Overview

    These Guidelines describe the procedures that the EPA follows in 
using existing data to evaluate the potential toxicity of environmental 
agents to the human male and female reproductive systems and to 
developing offspring. These Guidelines focus on reproductive system 
function as it relates to sexual behavior, fertility, pregnancy 
outcomes, and lactating ability, and the processes that can affect 
those functions directly. Included are effects on gametogenesis and 
gamete maturation and function, the reproductive organs, and the 
components of the endocrine system that directly support those 
functions. These Guidelines concentrate on the integrity of the male 
and female reproductive systems as required to ensure successful 
procreation. They also emphasize the importance of maintaining the 
integrity of the reproductive system for overall physical and 
psychologic health. The Guidelines for Developmental Toxicity Risk 
Assessment (U.S. EPA, 1991) focus specifically on effects of agents on 
development and should be used as a companion to these Guidelines.
    In evaluating reproductive effects, it is important to consider the 
presence, and where possible, the contribution of other manifestations 
of toxicity such as mutagenicity or carcinogenicity as well as other 
forms of general systemic toxicity. The reproductive process is such 
that these areas overlap, and all should be considered in reproductive 
risk assessments. Although the endpoints discussed in these Guidelines 
can detect impairment to components of the reproductive process, they 
may not discriminate effectively between nonmutagenic (e.g., cytotoxic) 
and mutagenic mechanisms. Examples of endpoints affected by either type 
of mechanism are sperm head morphology and preimplantation loss. If the 
effects seen may result from mutagenic events, then there is the 
potential for transmissible genetic damage. In such cases, the 
Guidelines for Mutagenicity Risk Assessment (U.S. EPA, 1986c) should be 
consulted in conjunction with these Guidelines. The Guidelines for 
Carcinogen Risk Assessment (U.S. EPA, 1986a, 1996b) should be consulted 
if reproductive system or developmentally induced cancer is detected.
    For assessment of risk to the human reproductive systems, the most 
appropriate data are those derived from human studies having adequate 
study

[[Page 56278]]

design and power. In the absence of adequate human data, our 
understanding of the mechanisms controlling reproduction supports the 
use of data from experimental animal studies to estimate the risk of 
reproductive effects in humans. However, some information needed for 
extrapolation of data from experimental animal studies to humans is not 
generally available. Therefore, to bridge these gaps in information, a 
number of default assumptions are made. These default assumptions, 
which are summarized in Table 1, should not preclude inquiry into the 
relevance of the data to potential human risk and should be invoked 
only after examination of the available information indicates that 
necessity. These assumptions provide the inferential basis for the 
approaches to risk assessment in these Guidelines. Each assumption 
should be evaluated along with other relevant information in making a 
final judgment as to human risk for each agent, and that information 
summarized in the risk characterization.

 Table 1.--Default Assumptions in Reproductive Toxicity Risk Assessment 
------------------------------------------------------------------------
                                                                        
-------------------------------------------------------------------------
1. An agent that produces an adverse reproductive effect in experimental
 animals is assumed to pose a potential threat to humans.               
2. Effects of xenobiotics on male and female reproductive processes are 
 assumed generally to be similar unless demonstrated otherwise. For     
 developmental outcomes, the specific effects in humans are not         
 necessarily the same as those seen in the experimental species.        
3. In the absence of information to determine the most appropriate      
 experimental species, data from the most sensitive species should be   
 used.                                                                  
4. In the absence of information to the contrary, an agent that affects 
 reproductive function in one sex is assumed to adversely affect        
 reproductive function in the other sex.                                
5. A nonlinear dose-response curve is assumed for reproductive toxicity.
------------------------------------------------------------------------

    An agent that produces an adverse reproductive effect in 
experimental animal studies is assumed to pose a potential reproductive 
threat to humans. This assumption is based on comparisons of data for 
agents that are known to cause human reproductive toxicity (Thomas, 
1981; Nisbet and Karch, 1983; Kimmel, C.A. et al., 1984, 1990; Hemminki 
and Vineis, 1985; Meistrich, 1986; Working, 1988). In general, the 
experimental animal data indicated adverse reproductive effects that 
are also seen in humans.
    Because similar mechanisms can be identified in the male and female 
of many mammalian species, effects of xenobiotics on male and female 
reproductive processes are assumed generally to be similar across 
species unless demonstrated otherwise. However, for developmental 
outcomes, it is assumed that the specific outcomes seen in experimental 
animal studies are not necessarily the same as those produced in 
humans. This latter assumption is made because of the possibility of 
species-specific differences in timing of exposure relative to critical 
periods of development, pharmacokinetics (including metabolism), 
developmental patterns, placentation, or modes of action. However, 
adverse developmental outcomes in laboratory mammalian studies are 
presumed to predict a hazard for adverse developmental outcome in 
humans.
    When sufficient data are available (e.g., pharmacokinetic) to allow 
a decision, the most appropriate species should be used to estimate 
human risk. In the absence of such data, it is assumed that the most 
sensitive species is most appropriate because, for the majority of 
agents known to cause human reproductive toxicity, humans appear to be 
as or more sensitive than the most sensitive animal species tested 
(Nisbet and Karch, 1983; Kimmel, C.A. et al., 1984, 1990; Hemminki and 
Vineis, 1985; Meistrich, 1986; Working, 1988), based on data from 
studies that determined dose on a body weight or air concentration 
basis.
    In the absence of specific information to the contrary, it is 
assumed that a chemical that affects reproductive function in one sex 
may also adversely affect reproductive function in the other sex. This 
assumption for reproductive risk assessment is based on three 
considerations: (1) For most agents, the nature of the testing and the 
data available are limited, reducing confidence that the potential for 
toxicity to both sexes and their offspring has been examined equally; 
(2) Exposures of either males or females have resulted in developmental 
toxicity; and (3) Many of the mechanisms controlling important aspects 
of reproductive system function are similar in females and males, and 
therefore could be susceptible to the same agents. Information that 
would negate this assumption would demonstrate that either a 
mechanistic difference existed between the sexes that would preclude 
toxic action on the other sex or, on the basis of sufficient testing, 
an agent did not produce an adverse reproductive effect when 
administered to the other sex. Mechanistic differences could include 
functions that do not exist in the other sex (e.g., lactation), 
differences in endocrine control of affected organ development or 
function, or pharmacokinetic and metabolic differences between sexes.
    In a quantitative dose-response analysis, mode of action, 
pharmacokinetic, and pharmacodynamic information should be used to 
predict the shape of the dose-response curve when sufficient 
information of that nature is available. When that information is 
insufficient, it has generally been assumed that there is a nonlinear 
dose-response for reproductive toxicity. This is based on known 
homeostatic, compensatory, or adaptive mechanisms that must be overcome 
before a toxic endpoint is manifested and on the rationale that cells 
and organs of the reproductive system and the developing organism are 
known to have some capacity for repair of damage. However, in a 
population, background levels of toxic agents and preexisting 
conditions may increase the sensitivity of some individuals in the 
population. Thus, exposure to a toxic agent may result in an increased 
risk of adverse effects for some, but not necessarily all, individuals 
within the population. Although a threshold may exist for endpoints of 
reproductive toxicity, it usually is not feasible to distinguish 
empirically between a true threshold and a nonlinear low-dose 
relationship. The shift to the term nonlinear does not change the RfD/
RfC methodology for reproductive system health effects, including the 
use of uncertainty factors.

II. Definitions and Terminology

    For the purposes of these Guidelines, the following definitions 
will be used: Reproductive toxicity--The occurrence of biologically 
adverse effects on the reproductive systems of females or males that 
may result from exposure to environmental agents. The toxicity may be 
expressed as alterations to the female or male reproductive organs, the 
related endocrine system, or pregnancy outcomes. The manifestation of 
such toxicity may include, but not be limited to, adverse effects on 
onset of puberty, gamete production and transport, reproductive cycle 
normality, sexual behavior, fertility, gestation, parturition, 
lactation, developmental toxicity, premature reproductive senescence, 
or modifications in other functions that are dependent on the integrity 
of the reproductive systems.
    Fertility--The capacity to conceive or induce conception.

[[Page 56279]]

    Fecundity--The ability to produce offspring within a given period 
of time. For litter-bearing species, the ability to produce large 
litters is also a component of fecundity.
    Fertile--A level of fertility that is within or exceeds the normal 
range for that species.
    Infertile--Lacking fertility for a specified period. The infertile 
condition may be temporary; permanent infertility is termed sterility.
    Subfertile--A level of fertility that is below the normal range for 
that species but not infertile.
    Developmental toxicity--The occurrence of adverse effects on the 
developing organism that may result from exposure prior to conception 
(either parent), during prenatal development, or postnatally to the 
time of sexual maturation. Adverse developmental effects may be 
detected at any point in the lifespan of the organism. The major 
manifestations of developmental toxicity include (1) death of the 
developing organism, (2) structural abnormality, (3) altered growth, 
and (4) functional deficiency (U.S. EPA, 1991).

III. Hazard Characterization for Reproductive Toxicants

    Identification and characterization of reproductive hazards can be 
based on data from either human or experimental animal studies. Such 
data can result from routine or accidental environmental or 
occupational exposures or, for experimental animals, controlled 
experimental exposures. A hazard characterization should evaluate all 
of the information available and should:
     Identify the strengths and limitations of the database, 
including all available epidemiologic and experimental animal studies 
as well as pharmacokinetic and mechanistic information.
     Identify and describe key toxicological studies.
     Describe the type(s) of effects.
     Describe the nature of the effects (irreversible, 
reversible, transient, progressive, delayed, residual, or latent 
effects).
     Describe how much is known about how (through what 
biological mechanism) the agent produces adverse effects.
     Discuss the other health endpoints of concern.
     Discuss any nonpositive data in humans or experimental 
animals.
     Discuss the dose-response data (epidemiologic or 
experimental animal) available for further dose-response analysis.
     Discuss the route, level, timing, and duration of exposure 
in studies as compared to expected human exposures.
     Summarize the hazard characterization, including:

--Major assumptions used,
--Confidence in the conclusions,
--Alternative conclusions also supported by the data,
--Major uncertainties identified, and
--Significant data gaps.

    Conduct of a hazard characterization requires knowledge of the 
protocols in which data were produced and the endpoints that were 
evaluated. Sections III.A. and III.B. present the traditional testing 
protocols for rodents and endpoints used to evaluate male and female 
reproductive toxicity along with evaluation of their strengths and 
limitations. Because many endpoints are common to multiple protocols, 
endpoints are considered separately from the discussion of the overall 
protocol structures. These are followed by presentation of many of the 
specific characteristics of human studies (Section III.C.) and limited 
discussions of pharmacokinetic and structure-activity factors (Sections 
III.D. and III.E.).

III.A. Laboratory Testing Protocols

III.A.1. Introduction
    Testing protocols describe the procedures to be used to provide 
data for risk assessments. The quality and usefulness of those data are 
dependent on the design and conduct of the tests, including endpoint 
selection and resolving power. A single protocol is unlikely to provide 
all of the information that would be optimal for conducting a 
comprehensive risk assessment. For example, the test design to study 
reversibility of adverse effects or mechanism of toxic action may be 
different from that needed to determine time of onset of an effect or 
for calculation of a safe level for repeated exposure over a long term. 
Ideally, results from several different types of tests should be 
available when performing a risk assessment. Typically, only limited 
data are available. Under those conditions, the limited data should be 
used to the extent possible to assess risk.
    Integral parts of the hazard characterization and quantitative 
dose-response processes are the evaluation of the protocols from which 
data are available and the quality of the resulting data. In this 
section, design factors that are of particular importance in 
reproductive toxicity testing are discussed. Then, standardized 
protocols that may provide useful data for reproductive risk 
assessments are described.
III.A.2. Duration of Dosing
    To evaluate adequately the potential effects of an agent on the 
reproductive systems, a prolonged treatment period is needed. For 
example, damage to spermatogonial stem cells will not appear in samples 
from the cauda epididymis or in ejaculates for 8 to 14 weeks, depending 
on the test species. With some chemical agents that bioaccumulate, the 
full impact on a given cell type could be further delayed, as could the 
impact on functional endpoints such as fertility. In such situations, 
adequacy of the dosing duration is a critical factor in the risk 
assessment.
    Conversely, adaptation may occur that allows tolerance to levels of 
a chemical that initially caused an effect that could be considered 
adverse. An example is interference with ovulation by chlordimeform 
(Goldman et al., 1991); an effect for which a compensatory mechanism is 
available. Thus, with continued dosing, the compensatory mechanism can 
be activated so that the initial adverse effect is masked.
    In these situations, knowledge of the relevant pharmacokinetic and 
pharmacodynamic data can facilitate selection of dose levels and 
treatment duration (see also section on Exposure Assessment). Equally 
important is proper timing of examination of treated animals relative 
to initiation and termination of exposure to the agent.
III.A.3. Length of Mating Period
    Traditionally, pairs of rats or mice are allowed to cohabit for 
periods ranging from several days to 3 weeks. Given a 4- or 5-day 
estrous cycle, each female that is cycling normally should be in estrus 
four or five times during a 21-day mating period. Therefore, 
information on the interval or the number of cycles needed to achieve 
pregnancy may provide evidence of reduced fertility that is not 
available from fertility data. Additionally, during each period of 
behavioral estrus, the male has the opportunity to copulate a number of 
times, resulting in delivery of many more sperm than are required for 
fertilization. When an unlimited number of matings is allowed in 
fertility testing, a large impact to sperm production is necessary 
before an adverse effect on fertility can be detected.

[[Page 56280]]

III.A.4. Number of Females Mated to Each Male
    The EPA test guidelines prepared pursuant to FIFRA and TSCA specify 
the use of 20 males and enough females to produce at least 20 
pregnancies for each dose group in each generation in the 
multigeneration reproduction test (U.S. EPA, 1982, 1985b, 1996a). 
However, in some tests that were not designed to conform to EPA test 
guidelines (OECD, 1983), 20 pregnancies may have been achieved by 
mating two females with each male and using fewer than 20 males per 
treatment group. In such cases, the statistical treatment of the data 
should be examined carefully. With multiple females mated to each male, 
the degree of independence of the observations for each female may not 
be known. In that situation, when the cause of the adverse effect 
cannot be assigned with confidence to only one sex, dependence should 
be assumed and the male used as the experimental unit in statistical 
analyses. Using fewer males as the experimental unit reduces ability to 
detect an effect.
III.A.5. Single- and Multigeneration Reproduction Tests
    Reproductive toxicity studies in laboratory animals generally 
involve continuous exposure to a test substance for one or more 
generations. The objective is to detect effects on the integrated 
reproductive process as well as to study effects on the individual 
reproductive organs. Test guidelines for the conduct of single- and 
multigeneration reproduction protocols have been published by the 
Agency pursuant to FIFRA and TSCA and by OECD (U.S. EPA, 1982, 1985b, 
1996a; Galbraith et al., 1983; OECD, 1983).
    The single-generation reproduction test evaluates effects of 
subchronic exposure of peripubertal and adult animals. In the 
multigeneration reproduction protocol, F1 and F2 offspring 
are exposed continuously in utero from conception until birth and 
during the preweaning period. This allows detection of effects that 
occur from exposures throughout development, including the peripubertal 
and young adult phases. Because the parental and subsequent filial 
generations have different exposure histories, reproductive effects 
seen in any particular generation are not necessarily comparable with 
those of another generation. Also, successive litters from the same 
parents cannot be considered as replicates because of factors such as 
continuing exposure of the parents, increased parental age, sexual 
experience, and parity of the females.
    In a single- or multigeneration reproduction test, rats are used 
most often. In a typical reproduction test, dosing is initiated at 5 to 
8 weeks of age and continued for 8 to 10 weeks prior to mating to allow 
effects on gametogenesis to be expressed and increase the likelihood of 
detecting histologic lesions. Three dose levels plus one or more 
control groups are usually included. Enough males and females are mated 
to ensure 20 pregnancies per dose group for each generation. Animals 
producing the first generation of offspring should be considered the 
parental (P) generation, and all subsequent generations should be 
designated filial generations (e.g., F1, F2). Only the P 
generation is mated in a single-generation test, while both the P and 
F1 generations are mated in a two-generation reproduction test.
    In the P generation, both females and males are treated prior to 
and during mating, with treatment usually beginning around puberty. 
Cohabitation can be allowed for up to 3 weeks (U.S. EPA, 1982, 1985b), 
during which the females are monitored for evidence of mating. Females 
continue to be exposed during gestation and lactation.
    In the two-generation reproduction test, randomly selected F1 
male and female offspring continue to be exposed after weaning (day 21) 
and through the mating period. Treatment of mated F1 females is 
continued throughout gestation and lactation. More than one litter may 
be produced from either P or F1 animals. Depending on the route of 
exposure of lactating females, it is important to consider that 
offspring may be exposed to a chemical by ingestion of maternal feed or 
water (diet or drinking water studies), by licking of exposed fur 
(inhalation study), by contact with treated skin (dermal study), or by 
coprophagia, as well as via the milk.
    In single- and multigeneration reproduction tests, reproductive 
endpoints evaluated in P and F generations usually include visual 
examination of the reproductive organs. Weights and histopathology of 
the testes, epididymides, and accessory sex glands may be available 
from males, and histopathology of the vagina, uterus, cervix, ovaries, 
and mammary glands from females. Uterine and ovarian weights also are 
often available. Male and female mating and fertility indices (Section 
III.B.2.a.) are usually presented. In addition, litters (and often 
individual pups) are weighed at birth and examined for number of live 
and dead offspring, gender, gross abnormalities, and growth and 
survival to weaning. Maturation and behavioral testing may also be 
performed on the pups.
    If effects on fertility or pregnancy outcome are the only adverse 
effects observed in a study using one of these protocols, the 
contributions of male- and female-specific effects often cannot be 
distinguished. If testicular histopathology or sperm evaluations have 
been included, it may be possible to characterize a male-specific 
effect. Similarly, ovarian and reproductive tract histology or changes 
in estrous cycle normality may be indicative of female-specific 
effects. However, identification of effects in one sex does not exclude 
the possibility that both sexes may have been affected adversely. Data 
from matings of treated males with untreated females and vice versa 
(crossover matings) are necessary to separate sex-specific effects.
    An EPA workshop has considered the relative merits of one- versus 
two-generation reproductive effects studies (Francis and Kimmel, 1988). 
The participants concluded that a one-generation study is insufficient 
to identify all potential reproductive toxicants, because it would 
exclude detection of effects caused by prenatal and postnatal exposures 
(including the prepubertal period) as well as effects on germ cells 
that could be transmitted to and expressed in the next generation. For 
example, adverse transgenerational effects on reproductive system 
development by agents that disrupt endocrine control of sexual 
differentiation would be missed. A one-generation test might also miss 
adverse effects with delayed or latent onset because of the shorter 
duration of exposure for the P generation. These limitations are shared 
with the shorter-term ``screening'' protocols described below. Because 
of these limitations, a comprehensive reproductive risk assessment 
should include results from a two-generation test or its equivalent. A 
further recommendation from the workshop was to include sperm analyses 
and estrous cycle normality as endpoints in reproductive effects 
studies. These endpoints have been included in the proposed revisions 
to the EPA test guideline (U.S. EPA, 1996a).
    In studies where parental and offspring generations are evaluated, 
there are additional risk assessment issues regarding the relationships 
of reproductive outcomes across generations. Increasing vulnerability 
of subsequent generations is often, but not always, observed. 
Qualitative predictions of increased risk of the filial generations 
could be strengthened by

[[Page 56281]]

knowledge of the reproductive effects in the adult, the likelihood of 
bioaccumulation of the agent, and the potential for increased 
sensitivity resulting from exposure during critical periods of 
development (Gray, 1991).
    Occasionally, the severity of effects may be static or decreased 
with succeeding generations. When a decrease occurs, one explanation 
may be that the animals in the F1 and F2 generations 
represent ``survivors'' who are (or become) more resistant to the agent 
than the average of the P generation. If such selection exists, then 
subsequent filial generations may show a reduced toxic response. Thus, 
significant adverse effects in any generation may be cause for concern 
regardless of results in other generations unless inconsistencies in 
the data indicate otherwise.
III.A.6. Alternative Reproductive Tests
    A number of alternative test designs have appeared in the 
literature (Lamb, 1985; Lamb and Chapin, 1985; Gray et al., 1988, 1989, 
1990; Morrissey et al., 1989). Although not necessarily viewed as 
replacements for the standard two-generation reproduction tests, data 
from these protocols may be used on a case-by-case basis depending on 
what is known about the test agent in question. When mutually agreed on 
by the testing organization and the Agency, such alternative protocols 
may offer an expanded array of endpoints and increased flexibility 
(Francis and Kimmel, 1988).
    A continuous breeding protocol, Fertility (or Reproductive) 
Assessment by Continuous Breeding (FACB or RACB), has been developed by 
the National Toxicology Program (NTP) (Lamb and Chapin, 1985; Morrissey 
et al., 1989; Gulati et al., 1991). As originally described, this 
protocol (FACB) was a one-generation test. However, in the current 
design (RACB), dosing is extended into the F1 generation to make 
it compatible with the EPA workshop recommendations for a two-
generation design (Francis and Kimmel, 1988). The RACB protocol is 
being used with both mice and rats. A distinctive feature of this 
protocol is the continuous cohabitation of male-female pairs (in the P 
generation) for 14 weeks. Up to five litters can be produced with the 
pups removed soon after birth. This protocol provides information on 
changes in the spacing, number, and size of litters over the 14-week 
dosing interval. Treatment (three dose levels plus controls) is 
initiated in postpubertal males and females (11 weeks of age) seven 
days before cohabitation and continues throughout the test. Offspring 
that are removed from the dam soon after birth are counted and examined 
for viability, litter and/or pup weight, sex, and external 
abnormalities and then discarded. The last litter may remain with the 
dam until weaning to study the effects of in utero as well as perinatal 
and postnatal exposures. If effects on fertility are observed in the P 
or F generations, additional reproductive evaluations may be conducted, 
including fertility studies and crossover matings to define the 
affected gender and site of toxicity.
    The sequential production of litters from the same adults allows 
observation of the timing of onset of an adverse effect on fertility. 
In addition, it improves the ability to detect subfertility due to the 
potential to produce larger numbers of pregnancies and litters than in 
a standard single- or multigeneration reproduction study. With 
continuous treatment, a cumulative effect could increase the incidence 
or extent of expression with subsequent litters. However, unless 
offspring were allowed to grow and reproduce (as they are routinely in 
the more recent version of the RACB protocol) (Gulati et al., 1991), 
little or no information will be available on postnatal development or 
reproductive capability of a second generation.
    Sperm measures (including sperm number, morphology, and motility) 
and vaginal smear cytology to detect changes in estrous cyclicity have 
been added to the RACB protocol at the end of the test period and their 
utility has been examined using model compounds in the mouse (Morrissey 
et al., 1989).
    Another test method combines the use of multiple endpoints in both 
sexes of rats with initiation of treatment at weaning (Gray et al., 
1988). Thus, morphologic and physiologic changes associated with 
puberty are included as endpoints. Both P sexes are treated (at least 
three dose levels plus controls) continuously through breeding, 
pregnancy, and lactation. The F1 generation is mated in a 
continuous breeding protocol. Vaginal smears are recorded daily 
throughout the test period to evaluate estrous cycle normality and 
confirm breeding and pregnancy (or pseudopregnancy). Pregnancy outcome 
is monitored in both the P and F1 generations at all doses, and 
terminal studies on both generations include comprehensive assessment 
of sperm measures (number, morphology, motility) as well as organ 
weights, histopathology, and the serum and tissue levels of appropriate 
reproductive hormones. As with the RACB, crossover mating studies may 
be conducted to identify the affected sex as warranted. This protocol 
combines the advantages of a continuous breeding design with 
acquisition of sex-specific multiple endpoint data at all doses. In 
addition, identification of pubertal effects makes this protocol 
particularly useful for detecting compounds with hormone-mediated 
actions such as environmental estrogens or antiandrogens.
III.A.7. Additional Test Protocols That May Provide Reproductive Data
    Several shorter-term reproductive toxicity screening tests have 
been developed. Among those are the Reproductive/Developmental Toxicity 
Screening Test, which is part of the OECD's Screening Information Data 
Set protocol (Scala et al., 1992; Tanaka et al., 1992; OECD, 1993a), a 
tripartite protocol developed by the International Conference on 
Harmonization (International Conference on Harmonization of Technical 
Requirements of Pharmaceuticals for Human Use, 1994; Manson, 1994), and 
the NTP's Short-Term Reproductive and Developmental Toxicity Screen 
(Harris, M.W. et al., 1992). These protocols have been developed for 
setting priorities for further testing and should not be considered 
sufficient by themselves to establish regulatory exposure levels. Their 
limited exposure periods do not allow assessment of certain aspects of 
the reproductive process, such as developmentally induced effects on 
the reproductive systems of offspring, that are covered by the 
multigeneration reproduction protocols.
    The male dominant lethal test was designed to detect mutagenic 
effects in the male spermatogenic process that are lethal to the 
offspring. A female dominant lethal protocol has also been used to 
detect equivalent effects on oogenesis (Generoso and Piegorsch, 1993).
    A review of the male dominant lethal test has been published as 
part of the EPA's Gene-Tox Program (Green et al., 1985). Dominant 
lethal protocols may use acute dosing (1 to 5 days) followed by serial 
matings with one or two females per male per week for the duration of 
the spermatogenic process. An alternative protocol may use subchronic 
dosing for the duration of the spermatogenic process followed by 
mating. Dose levels used with the acute protocol are usually higher 
than those used with the subchronic protocol. Females are monitored for 
evidence of mating, killed at approximately midgestation, and examined 
for incidence of pre- and postimplantation loss (see Section III.B.2. 
for discussions of these endpoints).

[[Page 56282]]

    Pre- or postimplantation loss in the dominant lethal test is often 
considered evidence that the agent has induced mutagenic damage to the 
male germ cell (U.S. EPA, 1986c). A genotoxic basis for a substantial 
portion of postimplantation loss is accepted widely. However, methods 
used to assess preimplantation loss do not distinguish between 
contributions of mutagenic events that cause embryo death and 
nonmutagenic factors that result in failure of fertilization or early 
embryo mortality (e.g., inadequate number of normal sperm, failure in 
sperm transport or ovum penetration). Similar effects (fertilization 
failure, early embryo death) could also be produced indirectly by 
effects that delay the timing of fertilization relative to time of 
ovulation. Such distinctions are important because cytotoxic effects on 
gametogenic cells do not imply the potential for transmittable genetic 
damage that is associated with mutagenic events. The interpretation of 
an increase in preimplantation loss may require additional data on the 
agent's mutagenic and gametotoxic potential if genotoxicity is to be 
factored into the risk assessment. Regardless, significant effects may 
be observed in a dominant lethal test that are considered reproductive 
in nature.
    An acute exposure protocol, combined with serial mating, may allow 
identification of the spermatogenic cell types that are affected by 
treatment. However, acute dosing may not produce adverse effects at 
levels as low as with subchronic dosing because of factors such as 
bioaccumulation. Conversely, if tolerance to an agent is developed with 
longer exposure, an effect may be observed after acute dosing that is 
not detected after longer-term dosing.
    Subchronic toxicity tests may have been conducted before a detailed 
reproduction study is initiated. In the subchronic toxicity test with 
rats, exposure usually begins at 6-8 weeks of age and is continued for 
90 days (U.S. EPA, 1982, 1985b). Initiation of exposure at 8 weeks of 
age (compared with 6) and exposure for approximately 90 days allows the 
animals to reach a more mature stage of sexual development and assures 
an adequate length of dosing for observation of effects on the 
reproductive organs with most agents. The route of administration is 
often oral or by gavage but may be dermal or by inhalation. Animals are 
monitored for clinical signs throughout the test and are necropsied at 
the end of dosing.
    The endpoints that are usually evaluated for the male reproductive 
system include visual examination of the reproductive organs, plus 
weights and histopathology for the testes, epididymides, and accessory 
sex glands. For the females, endpoints may include visual examination 
of the reproductive organs, uterine and ovarian weights, and 
histopathology of the vagina, uterus, cervix, ovaries, and mammary 
glands.
    This test may be useful to identify an agent as a potential 
reproductive hazard, but usually does not provide information about the 
integrated function of the reproductive systems (sexual behavior, 
fertility, and pregnancy outcomes), nor does it include effects of the 
agent on immature animals.
    Chronic toxicity tests provide an opportunity to evaluate toxic 
effects of long-term exposures. Oral, inhalation, or dermal exposure is 
initiated soon after weaning and is usually continued for 12 to 24 
months. Because of the extended treatment period, data from interim 
sacrifices may be available to provide useful information regarding the 
onset and sequence of toxicity. In males, the reproductive organs are 
examined visually, testes are weighed, and histopathologic examination 
is done on the testes and accessory sex glands. In females, the 
reproductive organs are examined visually, uterine and ovarian weights 
may be obtained, and histopathologic evaluation of the reproductive 
organs is done. The incidence of pathologic conditions is often 
increased in the reproductive tracts of aged control animals. 
Therefore, findings should be interpreted carefully.

III.B. Endpoints for Evaluating Male and Female Reproductive Toxicity 
in Test Species

III.B.1. Introduction
    The following discussion emphasizes endpoints that measure 
characteristics that are necessary for successful sexual performance 
and procreation. Other areas that are related less directly to 
reproduction are beyond the scope of these Guidelines. For example, 
secondary adverse health effects that may result from toxicity to the 
reproductive organs (e.g., osteoporosis or altered immune function), 
although important, are not included.
    In these Guidelines, the endpoints of reproductive toxicity are 
separated into three categories: couple-mediated, female-specific, and 
male-specific. Couple-mediated endpoints are those in which both sexes 
can have a contributing role if both partners are exposed. Thus, 
exposure of either sex or both sexes may result in an effect on that 
endpoint.
    The discussions of endpoints and the factors influencing results 
that are presented in this section are directed to evaluation and 
interpretation of results with test species. Many of those endpoints 
require invasive techniques that preclude routine use with humans. 
However, in some instances, related endpoints that can be used with 
humans are identified. Information that is specific for evaluation of 
effects on humans is presented in Section III.C.
    Although statistical analyses are important in determining the 
effects of a particular agent, the biological significance of data is 
most important. It is important to be aware that when many endpoints 
are investigated, statistically significant differences may occur by 
chance. On the other hand, apparent trends with dose may be 
biologically relevant even though pair-wise comparisons do not indicate 
a statistically significant effect. In each section, endpoints are 
identified in which significant changes may be considered adverse. 
However, concordance of results and known biology should be considered 
in interpreting all results. Results should be evaluated on a case-by-
case basis with all of the evidence considered. Scientific judgment 
should be used extensively. All effects that may be considered as 
adverse are appropriate for use in establishing a NOAEL, LOAEL, or 
benchmark dose.
III.B.2. Couple-Mediated Endpoints
    Data on fertility potential and associated reproductive outcomes 
provide the most comprehensive and direct insight into reproductive 
capability. As noted previously, most protocols only specify 
cohabitation of exposed males with exposed females. This complicates 
the resolution of gender-specific influences. Conclusions may need to 
be restricted to noting that the ``couple'' is at reproductive risk 
when one or both parents are potentially exposed.
    III.B.2.a. Fertility and Pregnancy Outcomes. Breeding studies with 
test species are a major source of data on reproductive toxicants. 
Evaluations of fertility and pregnancy outcomes provide measures of the 
functional consequences of reproductive injury. Measures of fertility 
and pregnancy outcome that are often obtained from multigeneration 
reproduction studies are presented in Table 2. Many endpoints that are 
pertinent for developmental toxicity are also listed and discussed in 
the Agency's Guidelines for Developmental Toxicity Risk Assessment 
(U.S. EPA, 1991). Also included in Table 2 are measures that

[[Page 56283]]

may be obtained from other types of studies (e.g., single-generation 
reproduction studies, developmental toxicity studies, dominant lethal 
studies) in which offspring are not retained to evaluate subsequent 
reproductive performance.

      Table 2.--Couple-Mediated End-points of Reproductive Toxicity     
------------------------------------------------------------------------
                                                                        
-------------------------------------------------------------------------
Multigeneration studies:                                                
  Mating rate, time to mating (time to pregnancy*)                      
  Pregnancy rate*                                                       
  Delivery rate*                                                        
  Gestation length*                                                     
  Litter size (total and live)                                          
  Number of live and dead offspring (Fetal death rate*)                 
  Offspring gender* (sex ratio)                                         
  Birth weight*                                                         
  Postnatal weights*                                                    
  Offspring survival*                                                   
  External malformations and variations*                                
  Offspring reproduction*                                               
Other reproductive endpoints:                                           
  Ovulation rate                                                        
  Fertilization rate                                                    
  Preimplantation loss                                                  
  Implantation number                                                   
  Postimplantation loss*                                                
  Internal malformations and variations*                                
  Postnatal structural and functional development*                      
------------------------------------------------------------------------
*Endpoints that can be obtained with humans.                            

    Some of the endpoints identified above are used to calculate ratios 
or indices (NRCl, 1977; Collins, 1978; Schwetz et al., 1980; U.S. EPA, 
1982, 1985b; Dixon and Hall, 1984; Lamb et al., 1985; Thomas, 1991). 
While the presentation of such indices is not discouraged, the 
measurements used to calculate those indices should also be available 
for evaluation. Definitions of some of these indices in published 
literature vary substantially. Also, the calculation of an index may be 
influenced by the test design. Therefore, it is important that the 
methods used to calculate indices be specified. Some commonly reported 
indices are in Table 3.

[[Page 56284]]

Table 3.--Selected Indices That May Be Calculated From Endpoints of 
Reproductive Toxicity in Test Species

Mating Index
[GRAPHIC] [TIFF OMITTED] TN31OC96.000

    Note: Mating is used to indicate that evidence of copulation 
(observation or other evidence of ejaculation such as vaginal plug 
or sperm in vaginal smear) was obtained.

Fertility Index
[GRAPHIC] [TIFF OMITTED] TN31OC96.001

    Note: Because both sexes are often exposed to an agent, 
distinction between sexes often is not possible. If responsibility 
for an effect can be clearly assigned to one sex (as when treated 
animals are mated with controls), then a female or male fertility 
index could be useful.

Gestation (Pregnancy) Index
[GRAPHIC] [TIFF OMITTED] TN31OC96.002

Live Birth Index
[GRAPHIC] [TIFF OMITTED] TN31OC96.003

Sex Ratio
[GRAPHIC] [TIFF OMITTED] TN31OC96.004

4-Day Survival Index (Viability Index)
[GRAPHIC] [TIFF OMITTED] TN31OC96.005

    Note: This definition assumes that no standardization of litter 
size is done until after the day 4 determination is completed.

Lactation Index (Weaning Index)
[GRAPHIC] [TIFF OMITTED] TN31OC96.006

    Note: If litters were standardized to equalize numbers of 
offspring per litter, number of offspring after standardization 
should be used instead of number born alive. When no standardization 
is done, measure is called weaning index. When standardization is 
done, measure is called lactation index.

Preweaning Index
[GRAPHIC] [TIFF OMITTED] TN31OC96.007

    Note: If litters were standardized to equalize numbers of 
offspring per litter, then number of offspring remaining after 
standardization should be used instead of number born.

                                                                        
------------------------------------------------------------------------

    Mating rate may be reported for the mated pairs, males only or 
females only. Evidence of mating may be direct observation of 
copulation, observation of copulatory plugs, or observation of sperm in 
the vaginal fluid (vaginal lavage). The mating rate may be influenced 
by the number of estrous cycles allowed or required for pregnancy to 
occur. Therefore, mating rate and fertility data from the first estrous 
cycle after initiation of cohabitation should be more discriminating 
than measurements involving multiple cycles. Evidence of mating does 
not necessarily mean successful impregnation.
    A useful indicator of impaired reproductive function may be the 
length of time required for each pair to mate after the start of 
cohabitation (time to mating). An increased interval between initiation 
of cohabitation and evidence of mating suggests abnormal estrous 
cyclicity in the female or impaired sexual behavior in one or both 
partners.
    The time to mating for normal pairs (rat or mouse) could vary by 3 
or 4 days depending on the stage of the estrous cycle at the start of 
cohabitation. If the

[[Page 56285]]

stage of the estrous cycle at the time of cohabitation is known, the 
component of the variance due to variation in stage at cohabitation can 
be removed in the data analysis.
    Data on fertilization rate, the proportion of available ova that 
were fertilized, are seldom available because the measurement requires 
necropsy very early in gestation. Pregnancy rate is the proportion of 
mated pairs that have produced at least one pregnancy within a fixed 
period where pregnancy is determined by the earliest available evidence 
that fertilization has occurred. Generally, a more meaningful measure 
of fertility results when the mating opportunity was limited to one 
mating couple and to one estrous cycle (see Sections III.A.3. and 
III.A.4.).
    The timing and integrity of gamete and zygote transport are 
important to fertilization and embryo survival and are quite 
susceptible to chemical perturbation. Disruption of the processes that 
contribute to a reduction in fertilization rate and increased early 
embryo loss are usually identified simply as preimplantation loss. 
Additional studies using direct assessments of fertilized ova and early 
embryos would be necessary to identify the cause of increased 
preimplantation loss (Cummings and Perreault, 1990). Preimplantation 
loss (described below) occurs in untreated as well as treated rodents 
and contributes to the normal variation in litter size.
    After mating, uterine and oviductal contractions are critical in 
the transport of spermatozoa from the vagina. In rodents, sufficient 
stimulation during mating is necessary for initiation of those 
contractions. Thus, impaired mating behavior may affect sperm transport 
and fertilization rate. Exposure of the female to estrogenic compounds 
can alter gamete transport. In women, low doses of exogenous estrogens 
may accelerate ovum transport to a detrimental extent, whereas high 
doses of estrogens or progestins delay transport and increase the 
incidence of ectopic pregnancies.
    Mammalian ova are surrounded by investments that the sperm must 
penetrate before fusing with ova. Chemicals may block fertilization by 
preventing this passage. Other agents may impair fusion of the sperm 
with the oolemma, transformations of the sperm or ovum chromatin into 
the male and female pronuclei, fusion of the pronuclei, or the 
subsequent cleavage divisions. Carbendazim, an inhibitor of microtubule 
synthesis, is an example of a chemical that can interfere with oocyte 
maturation and normal zygote formation after sperm-egg fusion by 
affecting meiosis (Perreault et al., 1992; Zuelke and Perreault, 1995). 
The early zygote is also susceptible to detrimental effects of mutagens 
such as ethylene oxide (Generoso et al., 1987).
    Fertility assessments in test animals have limited sensitivity as 
measures of reproductive injury. Therefore, results demonstrating no 
treatment-related effect on fertility may be given less weight than 
other endpoints that are more sensitive. Unlike humans, normal males of 
most test species produce sperm in numbers that greatly exceed the 
minimum requirements for fertility, particularly as evaluated in 
protocols that allow multiple matings (Amann, 1981; Working, 1988). In 
some strains of rats and mice, production of normal sperm can be 
reduced by up to 90% or more without compromising fertility (Aafjes et 
al., 1980; Meistrich, 1982; Robaire et al., 1984; Working, 1988). 
However, less severe reductions can cause reduced fertility in human 
males who appear to function closer to the threshold for the number of 
normal sperm needed to ensure full reproductive competence (see 
Supplementary Information). This difference between test species and 
humans means that negative results with test species in a study that 
was limited to endpoints that examined only fertility and pregnancy 
outcomes would provide insufficient information to conclude that the 
test agent poses no reproductive hazard in humans. It is unclear 
whether a similar consideration is applicable for females for some 
mechanisms of toxicity.
    The limited sensitivity of fertility measures in rodents also 
suggests that a NOAEL, LOAEL, or benchmark dose (see Section IV) based 
on fertility may not reflect completely the extent of the toxic effect. 
In such instances, data from additional reproductive endpoints might 
indicate that an adverse effect could occur at a lower dose level. In 
the absence of such data, the margin of exposure or uncertainty factor 
applied to the NOAEL, LOAEL, or benchmark dose may need to be adjusted 
to reflect the additional uncertainty (see Section IV).
    Both the blastocyst and the uterus must be ready for implantation, 
and their synchronous development is critical (Cummings and Perreault, 
1990). The preparation of the uterine endometrium for implantation is 
under the control of sequential estrogen and progesterone stimulation. 
Treatments that alter the internal hormonal environment or inhibit 
protein synthesis, mitosis, or cell differentiation can block 
implantation and cause embryo death.
    Gestation length can be determined in test animals from data on day 
of mating (observation of vaginal plug or sperm-positive vaginal 
lavage) and day of parturition. Significant shortening of gestation can 
lead to adverse outcomes of pregnancy such as decreased birth weight 
and offspring survival. Significantly longer gestation may be caused by 
failure of the normal mechanism for parturition and may result in death 
or impairment of offspring if dystocia (difficulty in parturition) 
occurs. Dystocia constitutes a maternal health threat for humans as 
well as test species. Lengthened gestation may result in higher birth 
weight; an effect that could mask a slower growth rate in utero because 
of exposure to a toxic agent. Comparison of offspring weights based on 
conceptional age may allow insight, although this comparison is 
complicated by generally faster growth rates postnatally than in utero.
    Litter size is the number of offspring delivered and is measured at 
or soon after birth. Unless this observation is made soon after 
parturition, the number of offspring observed may be less than the 
actual number delivered because of cannibalism by the dam. Litter size 
is affected by the number of ova available for fertilization (ovulation 
rate), fertilization rate, implantation rate, and the proportion of the 
implanted embryos that survives to parturition. Litter size may include 
dead as well as live offspring, therefore data on the numbers of live 
and dead offspring should be available also.
    When pregnant animals are examined by necropsy in mid- to late 
gestation, pregnancy status, including pre- and postimplantation losses 
can be determined. Postimplantation loss can be determined also by 
examining uteri from postparturient females. Preimplantation loss is 
the (number of corpora lutea minus number of implantation sites)/number 
of corpora lutea. Postimplantation loss, determined following delivery 
of a litter, is the (total number of implantation sites minus number of 
full-term pups)/number of implantation sites.
    Offspring gender in mammals is determined by the male through 
fertilization of an ovum by a Y- or an X-chromosome-bearing sperm. 
Therefore, selective impairment in the production, transport, or 
fertilizing ability of either of these sperm types can produce an 
alteration in the sex ratio. An agent may also induce selective loss of 
male or female fetuses. Further, alteration of the external sexual 
characteristics of offspring by agents that disrupt sexual development 
may produce apparent

[[Page 56286]]

effects on sex ratios. Although not examined routinely, these factors 
provide the most likely explanations for alterations in the sex ratio.
    Birth weight should be measured on the day of parturition. Often 
data from individual pups as well as the entire litter (litter weight) 
are provided. Birth weights are influenced by intrauterine growth 
rates, litter size, and gestation length. Growth rate in utero is 
influenced by the normality of the fetus, the maternal environment, and 
gender, with females tending to be smaller than males (Tyl, 1987). 
Individual pups in large litters tend to be smaller than pups in 
smaller litters. Thus, reduced birth weights that can be attributed to 
large litter size should not be considered an adverse effect unless the 
increased litter size is treatment related and the subsequent ability 
of the offspring to survive or develop is compromised. Multivariate 
analyses may be used to adjust pup weights for litter size (e.g., 
analysis of covariance, multiple regression). When litter weights only 
are reported, the increased numbers of offspring and the lower weights 
of the individuals tend to offset each other. When prenatal or 
postnatal growth is impaired by an acute exposure, compensatory growth 
after cessation of dosing could obscure the earlier effect.
    Postnatal weights are dependent on birth weight, sex, and normality 
of the individual, as well as the litter size, lactational ability of 
the dam, and suckling ability of the offspring. With large litters, 
small or weak offspring may not compete successfully for milk and show 
impaired growth. Because it is not possible usually to determine 
whether the effect was due solely to the increased litter size, growth 
retardation or decreased survival rate should be considered adverse in 
the absence of information to the contrary. Also, offspring weights may 
appear normal in very small litters and should be considered carefully 
in relation to controls.
    Offspring survival is dependent on the same factors as postnatal 
weight, although more severe effects are necessary usually to affect 
survival. All weight and survival endpoints can be affected by toxicity 
of an agent, either by direct effects on the offspring or indirectly 
through effects on the ability of the dam to support the offspring.
    Measures of malformations and variations, as well as postnatal 
structural and functional development, are presented in the Guidelines 
for Developmental Toxicity Risk Assessment and the Proposed Guidelines 
for Neurotoxicity Risk Assessment (U.S. EPA, 1991, 1995a). These 
documents should be consulted for additional information on those 
parameters.

Adverse Effects

    Table 2 lists couple-mediated endpoints that may be measured in 
reproduction studies. Table 3 presents examples of indices that may be 
calculated from couple-mediated reproductive toxicity data. Significant 
detrimental effects on any of those endpoints or on indices derived 
from those data should be considered adverse. Whether effects are on 
the female reproductive system or directly on the embryo or fetus is 
often not distinguishable, but the distinction may not be important 
because all of these effects should be cause for concern.
    III.B.2.b. Sexual Behavior. Sexual behavior reflects complex 
neural, endocrine, and reproductive organ interactions and is therefore 
susceptible to disruption by a variety of toxic agents and pathologic 
conditions. Interference with sexual behavior in either sex by 
environmental agents represents a potentially significant human 
reproductive problem. Most human information comes from studies on 
effects of drugs on sexual behavior or from clinical reports in which 
the detection of exposure-effect associations is unlikely. Data on 
sexual behavior are usually not available from studies of human 
populations that were exposed occupationally or environmentally to 
potentially toxic agents, nor are such data obtained routinely in 
studies of environmental agents with test species.
    In the absence of human data, the perturbation of sexual behavior 
in test species suggests the potential for similar effects on humans. 
Consistent with this position are data showing that central nervous 
system effects can disrupt sexual behavior in both test species and 
humans (Rubin and Henson, 1979; Waller et al., 1985). Although the 
functional components of sexual performance can be quantified in most 
test species, no direct evaluation of this behavior is done in most 
breeding studies. Rather, copulatory plugs or sperm-positive vaginal 
lavages are taken as evidence of sexual receptivity and successful 
mating. However, these markers do not demonstrate whether male 
performance resulted in adequate sexual stimulation of the female. 
Failure of the male to provide adequate stimulation to the female may 
impair sperm transport in the genital tract of female rats, thereby 
reducing the probability of successful impregnation (Adler and Toner, 
1986). Such a ``mating'' failure would be reflected in the calculated 
fertility index as reduced fertility and could be attributed 
erroneously to an effect on the spermatogenic process in the male or on 
fertility of the female.
    In the rat, a direct measure of female sexual receptivity is the 
occurrence of lordosis. Sexual receptivity of the female rat is 
normally cyclic, with receptivity commencing during the late evening of 
vaginal proestrus. Agents that interfere with normal estrous cyclicity 
also could cause absence of or abnormal sexual behavior that can be 
reflected in reduced numbers of females with vaginal plugs or vaginal 
sperm, alterations in lordosis behavior, and increased time to mating 
after start of cohabitation. In the male, measures include latency 
periods to first mount, mount with intromission, and first ejaculation, 
number of mounts with intromission to ejaculation, and the 
postejaculatory interval (Beach, 1979).
    Direct evaluation of sexual behavior is not warranted for all 
agents being tested for reproductive toxicity. Some likely candidates 
may be agents reported to exert central or peripheral neurotoxicity. 
Chemicals possessing or suspected to possess androgenic or estrogenic 
properties (or antagonistic properties) also merit consideration as 
potentially causing adverse effects on sexual behavior concomitant with 
effects on the reproductive organs.

Adverse Effects

    Effects on sexual behavior (within the limited definition of these 
Guidelines) should be considered as adverse reproductive effects. 
Included is evidence of impaired sexual receptivity and copulatory 
behavior. Impairment that is secondary to more generalized physical 
debilitation (e.g., impaired rear leg motor activity or general 
lethargy) should not be considered an adverse reproductive effect, 
although such conditions represent adverse systemic effects.
III.B.3. Male-Specific Endpoints
    III.B.3.a. Introduction. The following sections (III.B.3. and 
III.B.4.) describe various male-specific and female-specific endpoints 
of reproductive toxicity that can be obtained. Included are endpoints 
for which data are obtained routinely by the Agency and other endpoints 
for which data may be encountered in the review of chemicals. Guidance 
is presented for interpretation of results involving these endpoints 
and their use in risk assessment. Effects are identified that should be 
considered as adverse reproductive effects if significantly different 
from controls.
    The Agency may obtain data on the potential male reproductive 
toxicity of

[[Page 56287]]

an agent from many sources including, but not limited to, studies done 
according to Agency test guidelines. These may include acute, 
subchronic, and chronic testing and reproduction and fertility studies. 
Male-specific endpoints that may be encountered in such studies are 
identified in Table 4.

       Table 4.--Male-Specific Endpoints of Reproductive Toxicity       
------------------------------------------------------------------------
                                                                        
------------------------------------------------------------------------
Organ weights................  Testes, epididymides, seminal vesicles,  
                                prostate, pituitary.                    
Visual examination and         Testes, epididymides, seminal vesicles,  
 histopathology.                prostate, pituitary.                    
Sperm evaluation *...........  Sperm number (count) and quality         
                                (morphology, motility)                  
Sexual behavior *............  Mounts, intromissions, ejaculations.     
Hormone levels *.............  Luteinizing hormone, follicle stimulating
                                hormone, testosterone, estrogen,        
                                prolactin.                              
Developmental effects........  Testis descent*, preputial separation,   
                                sperm production*, ano-genital distance,
                                structure of external genitalia*.       
------------------------------------------------------------------------
* Reproductive endpoints that can be obtained or estimated relatively   
  noninvasively with humans.                                            

    III.B.3.b. Body Weight and Organ Weights. Monitoring body weight 
during treatment provides an index of the general health status of the 
animals, and such information may be important for the interpretation 
of reproductive effects (see also Section III.B.2.). Depression in body 
weight or reduction in weight gain may reflect a variety of responses, 
including rejection of chemical-containing food or water because of 
reduced palatability, treatment-induced anorexia, or systemic toxicity. 
Less than severe reductions in adult body weight induced by restricted 
nutrition have shown little effect on the male reproductive organs or 
on male reproductive function (Chapin et al., 1993a, b). When a 
meaningful, biologic relationship between a body weight decline and a 
significant effect on the male reproductive system is not apparent, it 
is not appropriate to dismiss significant alteration of the male 
reproductive system as secondary to the occurrence of nonreproductive 
toxicity. Unless additional data provide the needed clarification, 
alteration in a reproductive measure that would otherwise be considered 
adverse should still be considered as an adverse male reproductive 
effect in the presence of mild to moderate body weight changes. In the 
presence of severe body weight depression or other severe systemic 
debilitation, it should be noted that an adverse effect on a 
reproductive endpoint occurred, but the effect may have resulted from a 
more generalized toxic effect. Regardless, adverse effects would have 
been observed in that situation and a risk assessment should be pursued 
if sufficient data are available.
    The male reproductive organs for which weights may be useful for 
reproductive risk assessment include the testes, epididymides, 
pituitary gland, seminal vesicles (with coagulating glands), and 
prostate. Organ weight data may be presented as both absolute weights 
and as relative weights (i.e., organ weight to body weight ratios). 
Organ weight data may also be reported relative to brain weight since, 
subsequent to development, the weight of the brain usually remains 
quite stable (Stevens and Gallo, 1989). Evaluation of data on absolute 
organ weights is important, because a decrease in a reproductive organ 
weight may occur that was not necessarily related to a reduction in 
body weight gain. The organ weight-to-body weight ratio may show no 
significant difference if both body weight and organ weight change in 
the same direction, masking a potential organ weight effect.
    Normal testis weight varies only modestly within a given test 
species (Schwetz et al., 1980; Blazak et al., 1985). This relatively 
low interanimal variability suggests that absolute testis weight should 
be a precise indicator of gonadal injury. However, damage to the testes 
may be detected as a weight change only at doses higher than those 
required to produce significant effects in other measures of gonadal 
status (Berndtson, 1977; Foote et al., 1986; Ku et al., 1993). This 
contradiction may arise from several factors, including a delay before 
cell deaths are reflected in a weight decrease (due to preceding edema 
and inflammation, cellular infiltration) or Leydig cell hyperplasia. 
Blockage of the efferent ducts by cells sloughed from the germinal 
epithelium or the efferent ducts themselves can lead to an increase in 
testis weight due to fluid accumulation (Hess et al., 1991; Nakai et 
al., 1993), an effect that could offset the effect of depletion of the 
germinal epithelium on testis weight. Thus, while testis weight 
measurements may not reflect certain adverse testicular effects and do 
not indicate the nature of an effect, a significant increase or 
decrease is indicative of an adverse effect.
    Pituitary gland weight can provide valuable insight into the 
reproductive status of the animal. However, the pituitary contains cell 
types that are responsible for the regulation of a variety of 
physiologic functions including some that are separate from 
reproduction. Thus, changes in pituitary weight may not necessarily 
reflect reproductive impairment. If weight changes are observed, 
gonadotroph-specific histopathologic evaluations may be useful in 
identifying the affected cell types. This information may then be used 
to judge whether the observed effect on the pituitary is related to 
reproductive system function and therefore an adverse reproductive 
effect.
    Prostate and seminal vesicle weights are androgen-dependent and may 
reflect changes in the animal's endocrine status or testicular 
function. Separation of the seminal vesicles and coagulating gland 
(dorsal prostate) is difficult in rodents. However, the seminal vesicle 
and prostate can be separated and results may be reported for these 
glands separately or together, with or without their secretory fluids. 
Differential loss of secretory fluids prior to weighing could produce 
artifactual weights. Because the seminal vesicles and prostate may 
respond differently to an agent (endocrine dependency and developmental 
susceptibility differ), more information may be gained if the weights 
were examined separately.

Adverse Effects

    Significant changes in absolute or relative male reproductive organ 
weights may constitute an adverse reproductive effect. Such changes 
also may provide a basis for obtaining additional information on the 
reproductive toxicity of that agent. However, significant changes in 
other important endpoints that are related to reproductive function may 
not be reflected in organ weight data. Therefore, lack of an organ 
weight effect should not be used to negate significant changes in other 
endpoints that may be more sensitive.
    III.B.3.c. Histopathologic Evaluations. Histopathologic evaluations 
of test animal tissues have a prominent role in male reproductive risk 
assessment. Organs that are often evaluated include

[[Page 56288]]

the testes, epididymides, prostate, seminal vesicles (often including 
coagulating glands), and pituitary. Tissues from lower dose exposures 
are often not examined histologically if the high dose produced no 
difference from controls. Histologic evaluations can be especially 
useful by (1) providing a relatively sensitive indicator of damage; (2) 
providing information on toxicity from a variety of protocols; and (3) 
with short-term dosing, providing information on site (including target 
cells) and extent of toxicity; and (4) indicating the potential for 
recovery.
    The quality of the information presented from histologic analyses 
of spermatogenesis is improved by proper fixation and embedding of 
testicular tissue. With adequately prepared tissue (Chapin, 1988; 
Russell et al., 1990; Hess and Moore, 1993), a description of the 
nature and background level of lesions in control tissue, whether 
preparation-induced or otherwise, can facilitate interpreting the 
nature and extent of the lesions observed in tissues obtained from 
exposed animals. Many histopathologic evaluations of the testis only 
detect lesions if the germinal epithelium is severely depleted or 
degenerating, if multinucleated giant cells are obvious, or if sloughed 
cells are present in the tubule lumen. More subtle lesions, such as 
retained spermatids or missing germ cell types, that can significantly 
affect the number of sperm being released normally into the tubule 
lumen may not be detected when less adequate methods of tissue 
preparation are used. Also, familiarity with the detailed morphology of 
the testis and the kinetics of spermatogenesis of each test species can 
assist in the identification of less obvious lesions that may accompany 
lower dose exposures or lesions that result from short-term exposure 
(Russell et al., 1990). Several approaches for qualitative or 
quantitative assessment of testicular tissue are available that can 
assist in the identification of less obvious lesions that may accompany 
lower-dose exposures, including use of the technique of ``staging.'' A 
book is available (Russell et al., 1990) which provides extensive 
information on tissue preparation, examination, and interpretation of 
observations for normal and high resolution histology of the germinal 
epithelium of rats, mice, and dogs. Included is guidance for 
identification and quantification of the various cell types and 
associations for each stage of the spermatogenic cycle. Also, a 
decision-tree scheme for staging with the rat has been published (Hess, 
1990).
    The basic morphology of other male reproductive organs (e.g., 
epididymides, accessory sex glands, and pituitary) has been described 
as well as the histopathologic alterations that may accompany certain 
disease states (Fawcett, 1986; Jones et al., 1987; Haschek and 
Rousseaux, 1991). Compared with the testes, less is known about 
structural changes in these tissues that are associated with exposure 
to toxic agents. With the epididymides and accessory sex glands, 
histologic evaluation is usually limited to the height and possibly the 
integrity of the secretory epithelium. Evaluation should include 
information on the caput, corpus, and cauda segments of the epididymis. 
Presence of debris and sloughed cells in the epididymal lumen are 
valuable indicators of damage to the germinal epithelium or the 
excurrent ducts. The presence of lesions such as sperm granulomas, 
leucocyte infiltration (inflammation) or absence of clear cells in the 
cauda epididymal epithelium should be noted. Information from 
examinations of the pituitary should include evaluation of the 
morphology of the cell types that produce the gonadotropins and 
prolactin.
    The degree to which histopathologic effects are quantified is 
usually limited to classifying animals, within dose groups, as either 
affected or not affected by qualitative criteria. Little effort has 
been made to quantify the extent of injury, and procedures for such 
classifications are not applied uniformly (Linder et al., 1990). 
Evaluation procedures would be facilitated by adoption of more uniform 
approaches for quantifying the extent of histopathologic damage per 
individual. In the absence of standardized tissue preparation 
techniques and a standardized quantification system, the evaluation of 
histopathologic data would be facilitated by the presentation of the 
evaluation criteria and procedure by which the level of lesions in 
exposed individuals was judged to be in excess of controls.
    If properly obtained (i.e., proper preparation and analysis of 
tissue), data from histopathologic evaluations may provide a relatively 
sensitive tool that is useful for detection of low-dose effects. This 
approach may also provide insight into sites and mechanisms of action 
for the agent on that reproductive organ. When similar targets or 
mechanisms exist in humans, the basis for interspecies extrapolation is 
strengthened. Depending on the experimental design, information can 
also be obtained that may allow prediction of the eventual extent of 
injury and degree of recovery in that species and humans (Russell, 
1983).

Adverse Effects

    Significant and biologically meaningful histopathologic damage in 
excess of the level seen in control tissue of any of the male 
reproductive organs should be considered an adverse reproductive 
effect. Significant histopathologic damage in the pituitary should be 
considered as an adverse effect but should be shown to involve cells 
that control gonadotropin or prolactin production to be called a 
reproductive effect. Although thorough histopathologic evaluations that 
fail to reveal any treatment-related effects may be quite convincing, 
consideration should be given to the possible presence of other 
testicular or epididymal effects that are not detected histologically 
(e.g., genetic damage to the germ cell, decreased sperm motility), but 
may affect reproductive function.
    III.B.3.d. Sperm Evaluations. The parameters that are important for 
sperm evaluations are sperm number, sperm morphology, and sperm 
motility. Data on those parameters allow more adequate estimation of 
the number of ``normal'' sperm; a parameter that is likely to be more 
informative than sperm number alone. Although effects on sperm 
production can be reflected in other measures such as testicular 
spermatid count or cauda epididymal weight, no surrogate measures are 
adequate to reflect effects on sperm morphology or motility. Similar 
data can be obtained noninvasively from human ejaculates, enhancing the 
ability to confirm effects seen in test species or to detect effects in 
humans. Brief descriptions of these measures are provided below, 
followed by a discussion of the use of various sperm measures in male 
reproductive risk assessment.

Sperm Number

    Measures of sperm concentration (count) have been the most 
frequently reported semen variable in the literature on humans (Wyrobek 
et al., 1983a). Sperm number or sperm concentration from test species 
may be derived from ejaculated, epididymal, or testicular samples (Seed 
et al., 1996). Of the common test species, ejaculates can only be 
obtained readily from rabbits or dogs. Ejaculates can be recovered from 
the reproductive tracts of mated females of other species (Zenick et 
al., 1984). Measures of human sperm production are usually derived from 
ejaculates, but could also be obtained from spermatid counts or 
quantitative histology using testicular biopsy tissue samples. With

[[Page 56289]]

ejaculates, both sperm concentration (number of sperm/mL of ejaculate) 
and total sperm per ejaculate (sperm concentration x volume) should be 
evaluated.
    Ejaculated sperm number from any species is influenced by several 
variables, including the length of abstinence and the ability to obtain 
the entire ejaculate. Intra- and interindividual variation are often 
high, but are reduced somewhat if ejaculates were collected at regular 
intervals from the same male (Williams et al., 1990). Such a 
longitudinal study design has improved detection sensitivity and thus 
requires a smaller number of subjects (Wyrobek et al., 1984). In 
addition, if a pre-exposure baseline is obtained for each male (test 
animal or human studies when allowed by protocol), then changes during 
exposure or recovery can be better defined.
    Epididymal sperm evaluations with test species usually use sperm 
from only the cauda portion of the epididymis, but the samples for 
sperm motility and morphology may be derived also from the vas 
deferens. It has been customary to express the sperm count in relation 
to the weight of the cauda epididymis. However, because sperm 
contribute to epididymal weight, expression of the data as a ratio may 
actually mask declines in sperm number. The inclusion of data on 
absolute sperm counts can improve resolution. As is true for ejaculated 
sperm counts, epididymal sperm counts are influenced directly by level 
of sexual activity (Amann, 1981; Hurtt and Zenick, 1986).
    Sperm production data may be derived from counts of the distinctive 
elongated spermatid nuclei that remain after homogenization of testes 
in a detergent-containing medium (Amann, 1981; Meistrich, 1982; Cassidy 
et al., 1983; Blazak et al., 1993). The elongated spermatid counts are 
a measure of sperm production from the stem cells and their ensuing 
survival through spermatocytogenesis and spermiogenesis (Meistrich, 
1982; Meistrich and van Beek, 1993). If evaluation was conducted when 
the effect of a lesion would be reflected adequately in the spermatid 
count, then spermatid count may serve as a substitute for quantitative 
histologic analysis of sperm production (Russell et al., 1990). 
However, spermatid counts may be misleading when duration of exposure 
is shorter than the time required for a lesion to be fully expressed in 
the spermatid count. Also, spermatid counts reported from some 
laboratories have large coefficients of variation that may reduce the 
statistical power and thus the usefulness of that measure.
    The ability to detect a decrease in testicular sperm production may 
be enhanced if spermatid counts are available. However, spermatid 
enumerations only reflect the integrity of spermatogenic processes 
within the testes. Posttesticular effects or toxicity expressed as 
alterations in motility, morphology, viability, fragility, and other 
properties of sperm can be determined only from epididymal, vas 
deferens, or ejaculated samples.

Sperm Morphology

    Sperm morphology refers to structural aspects of sperm and can be 
evaluated in cauda epididymal, vas deferens, or ejaculated samples. A 
thorough morphologic evaluation identifies abnormalities in the sperm 
head and flagellum. Because of the suggested correlation between an 
agent's mutagenicity and its ability to induce abnormal sperm, sperm 
head morphology has been a frequently reported sperm variable in 
toxicologic studies on test species (Wyrobek et al., 1983b). The 
tendency has been to conclude that increased incidence of sperm head 
malformations reflects germ-cell mutagenicity. However, not every 
mutagen induces sperm head abnormalities, and other nonmutagenic 
chemicals may alter sperm head morphology. For example, microtubule 
poisons may cause increases in abnormal sperm head incidence, 
presumably by interfering with spermiogenesis, a microtubule-dependent 
process (Russell et al., 1981). Sperm morphology may be altered also 
due to degeneration subsequent to cell death. Thus, the link between 
sperm morphology and mutagenicity is not necessarily sensitive or 
specific.
    An increase in abnormal sperm morphology has been considered 
evidence that the agent has gained access to the germ cells (U.S. EPA, 
1986c). Exposure of males to toxic agents may lead to sperm 
abnormalities in their progeny (Wyrobek and Bruce, 1978; Hugenholtz and 
Bruce, 1983; Morrissey et al., 1988a, b). However, transmissible germ-
cell mutations might exist in the absence of any warning morphologic 
indicator such as abnormal sperm. The relationships between these 
morphologic alterations and other karyotypic changes remains uncertain 
(de Boer et al., 1976).
    The traditional approach to characterizing morphology in 
toxicologic testing has relied on subjective categorization of sperm 
head, midpiece, and tail defects in either stained preparations by 
bright field microscopy (Filler, 1993) or fixed, unstained preparations 
by phase contrast microscopy (Linder et al., 1992; Seed et al., 1996). 
Such an approach may be adequate for mice and rats with their 
distinctly angular head shapes. However, the observable heterogeneity 
of structure in human sperm and in nonrodent species makes it difficult 
for the morphologist to define clearly the limits of normality. More 
systematic, quantitative, and automated approaches have been offered 
that can be used with humans and test species (Katz et al., 1982; 
Wyrobek et al., 1984). Data that categorize the types of abnormalities 
observed and quantify the frequencies of their occurrences are 
preferred to estimation of overall proportion of abnormal sperm. 
Objective, quantitative approaches that are done properly should result 
in a higher level of confidence than more subjective measures.
    Sperm morphology profiles are relatively stable and characteristic 
in a normal individual (and a strain within a species) over time. Sperm 
morphology is one of the least variable sperm measures in normal 
individuals, which may enhance its use in the detection of 
spermatotoxic events (Zenick et al., 1994). However, the reproductive 
implications of the various types of abnormal sperm morphology need to 
be delineated more fully. The majority of studies in test species and 
humans have suggested that abnormally shaped sperm may not reach the 
oviduct or participate in fertilization (Nestor and Handel, 1984; Redi 
et al., 1984). The implication is that the greater the number of 
abnormal sperm in the ejaculate, the greater the probability of reduced 
fertility.

Sperm Motility

    The biochemical environments in the testes and epididymides are 
highly regulated to assure the proper development and maturation of the 
sperm and the acquisition of critical functional characteristics, i.e., 
progressive motility and the potential to fertilize. With chemical 
exposures, perturbation of this balance may occur, producing 
alterations in sperm properties such as motility. Chemicals (e.g., 
epichlorohydrin) have been identified that selectively affect sperm 
motility and also reduce fertility. Studies have examined rat sperm 
motility as a reproductive endpoint (Morrissey et al., 1988a, b; Toth 
et al., 1989b, 1991b), and sperm motility assessments are an integral 
part of some reproductive toxicity tests (Gray et al., 1988; Morrissey 
et al., 1989; U.S. EPA, 1996a).

[[Page 56290]]

    Motility estimates may be obtained on ejaculated, vas deferens, or 
cauda epididymal samples. Standardized methods are needed because 
motility is influenced by a number of experimental variables, including 
abstinence interval, method of sample collection and handling, elapsed 
time between sampling and observation, the temperature at which the 
sample is stored and analyzed, the extent of sperm dilution, the nature 
of the dilution medium, and the microscopic chamber employed for the 
observations (Slott et al., 1991; Toth et al., 1991a; Chapin et al., 
1992; Schrader et al., 1992; Weir and Rumberger, 1995; Seed et al., 
1996).
    Sperm motility can be evaluated in fresh samples under phase 
contrast microscopy, or sperm images can be recorded and stored in 
video or digital format and analyzed later, either manually or by 
computer-aided semen analysis (Linder et al., 1986; Boyers et al., 
1989; Toth et al., 1989a; Yeung et al., 1992; Slott and Perreault, 
1993). For manual assessments, the percentage of motile and 
progressively motile sperm can be estimated and a simple scale used to 
describe the vigor of the sperm motion.
    The recent application of video and/or digital technology to sperm 
analysis allows a more detailed evaluation of sperm motion including 
information about the individual sperm tracks. It also provides 
permanent storage of the sperm tracks which can be re-analyzed as 
necessary (manually or computer-assisted). With computer-assisted 
technology, information about sperm velocity (straight-line and 
curvilinear) as well as the amplitude and frequency of the track are 
obtained rapidly and efficiently on large numbers of sperm. Using this 
technology, chemically induced alterations in sperm motion have been 
detected (Toth et al., 1989a, 1992; Slott et al., 1990; Klinefelter et 
al., 1994a), and such changes have been related to the fertility of the 
exposed animals (Toth et al., 1991a; Oberlander et al., 1994; Slott et 
al., 1995). These preliminary studies indicate that significant 
reductions in sperm velocity are associated with infertility, even when 
the percentage of motile sperm is not affected. The ability to 
distinguish between the proportion of sperm showing any type of motion 
and those with progressive motility is important (Seed et al., 1996).
    Changes in endpoints that measure effects on spermatogenesis and 
sperm maturation have been related to fertility in several test 
species, but the ability to predict infertility from these data (in the 
absence of fertility data) is not reliable. This is in part due to the 
observation, in both test species and humans, that fertility is 
dependent not only on having adequate numbers of sperm, but also on the 
degree to which those sperm are normal. If sperm quality is high, then 
sperm number must be substantially reduced before fertility is 
affected. For example, in a rat model that employs artificial 
insemination of differing numbers of good quality sperm, sperm numbers 
can be reduced substantially before fertility is affected (Klinefelter 
et al., 1994b). In humans, the distribution of sperm counts for fertile 
and infertile men overlap, with the mean for fertile men being higher 
(Meistrich and Brown, 1983), but fertility is likely to be impaired 
when counts drop below 20 million/mL (WHO, 1992). Similarly, if sperm 
numbers are normal in rodents, a relatively large effect on sperm 
motility is required before fertility is affected. For example, rodent 
sperm velocity must be substantially reduced, in the presence of 
adequate numbers of sperm, before fertility is affected (Toth et al., 
1991a; Slott et al., 1995). These models also show that relatively 
modest changes in sperm numbers or quality may not cause infertility, 
but can nevertheless be predictive of infertility. On the other hand, 
fertility may be impaired by smaller decrements in both number and 
motility (or other qualitative characteristics).
    Thus, the process of reproductive risk assessment is facilitated by 
having information on a variety of sperm measures and reproductive 
organ histopathology in addition to fertility. Specific information 
about reproductive organ and gamete function can then be used to 
evaluate the occurrence and extent of injury, and the probable site of 
toxicity in the reproductive system. The more information that is 
available from supplementary endpoints, the more the risk assessment 
can be based on science rather than uncertainty.

Adverse Effects

    Human male fertility is generally lower than that of test species 
and may be more susceptible to damage from toxic agents (see 
Supplementary Information). Therefore, the conservative approach should 
be taken that, within the limits indicated in the sections on those 
parameters, statistically significant changes in measures of sperm 
count, morphology, or motility as well as number of normal sperm should 
be considered adverse effects.
    III.B.3.e. Paternally Mediated Effects on Offspring. The concept is 
well accepted that exposure of a female to toxic chemicals during 
gestation or lactation may produce death, structural abnormalities, 
growth alteration, or postnatal functional deficits in her offspring. 
Sufficient data now exist with a variety of agents to conclude that 
male-only exposure also can produce deleterious effects in offspring 
(Davis et al., 1992; Colie, 1993; Savitz et al., 1994; Qiu et al., 
1995). Paternally mediated effects include pre- and postimplantation 
loss, growth and behavioral deficits, and malformations. A large 
proportion of the chemicals reported to cause paternally mediated 
effects have genotoxic activity, and are considered to exert this 
effect via transmissible genetic alterations. Low doses of 
cyclophosphamide have resulted in induction of single strand DNA breaks 
during rat spermatogenesis which, due in part to absence of subsequent 
DNA repair capability, remain at fertilization (Qiu et al., 1995). The 
results of such damage have been observed in the F2 generation 
offspring (Hales et al., 1992). Other mechanisms of induction of 
paternally mediated effects are also possible. Xenobiotics present in 
seminal plasma or bound to the fertilizing sperm could be introduced 
into the female genital tract, or even the oocyte directly, and might 
also interfere with fertilization or early development. With humans, 
the possibility exists that a parent could transport the toxic agent 
from the work environment to the home (e.g., on work clothes), exposing 
other adults or children. Further work is needed to clarify the extent 
to which paternal exposures may be associated with adverse effects on 
offspring. Regardless, if an agent is identified in test species or in 
humans as causing a paternally mediated adverse effect on offspring, 
the effect should be considered an adverse reproductive effect.
III.B.4. Female-Specific Endpoints
    III.B.4.a. Introduction. The reproductive life cycle of the female 
may be divided into phases that include fetal, prepubertal, cycling 
adult, pregnant, lactating, and reproductively senescent. Detailed 
descriptions of all phases are available (Knobil et al., 1994). It is 
important to detect adverse effects occurring in any of these stages. 
Traditionally, the endpoints that have been used have emphasized 
ability to become pregnant, pregnancy outcome, and offspring survival 
and development. Although reproductive organ weights may be obtained 
and these organs examined histologically in test species, these 
measures do not necessarily detect abnormalities in dynamic processes 
such as estrous cyclicity or follicular atresia unless degradation is 
severe. Similarly, toxic effects on onset of

[[Page 56291]]

puberty have not been examined, nor have the long-term consequences of 
exposure on reproductive senescence. Thus, the amount of information 
obtained routinely to detect toxic effects on the female reproductive 
system has been limited.
    The consequences of impairment in the nonpregnant female 
reproductive system are equally important, and endpoints to detect 
adverse effects on the nonpregnant reproductive system, when available, 
can be useful in evaluating reproductive toxicity. Such measures may 
also provide additional interrelated endpoints and information on 
mechanism of action.
    Adverse alterations in the nonpregnant female reproductive system 
have been observed at dose levels below those that result in reduced 
fertility or produce other overt effects on pregnancy or pregnancy 
outcomes (Le Vier and Jankowiak, 1972; Barsotti et al., 1979; Sonawane 
and Yaffe, 1983; Cummings and Gray, 1987). In contrast to the male 
reproductive system, the status of the normal female system fluctuates 
in adults. Thus, in nonpregnant animals (including humans), the ovarian 
structures and other reproductive organs change throughout the estrous 
or menstrual cycle. Although not cyclic, normal changes also accompany 
the progression of pregnancy, lactation, and return to cyclicity during 
or after lactation. These normal fluctuations may affect the endpoints 
used for evaluation. Therefore, knowledge of the reproductive status of 
the female at necropsy, including the stage of the estrous cycle, can 
facilitate detection and interpretation of effects with endpoints such 
as uterine weight and histopathology of the ovary and uterus. Necropsy 
of all test animals at the same stage of the estrous cycle can reduce 
the variance of test results with such measures.
    A variety of measures to evaluate the integrity of the female 
reproductive system has been used in toxicity studies. With appropriate 
measures, a comprehensive evaluation of the reproductive process can be 
achieved, including identification of target organs and possible 
elucidation of the mechanisms involved in the agent's effect(s). Areas 
that may be examined in evaluations of the female reproductive system 
are listed in Table 5.

                          Table 5.--Female-Specific Endpoints of Reproductive Toxicity                          
----------------------------------------------------------------------------------------------------------------
                                                                                                                
----------------------------------------------------------------------------------------------------------------
Organ weights...............................................................  Ovary, uterus, vagina, pituitary. 
Visual examination and histopathology.......................................  Ovary, uterus, vagina, pituitary, 
                                                                               oviduct, mammary gland.          
Estrous (menstrual *) cycle normality.......................................  Vaginal smear cytology.           
Sexual behavior.............................................................  Lordosis, time to mating, vaginal 
                                                                               plugs, or sperm.                 
Hormone levels *............................................................  LH, FSH, estrogen, progesterone,  
                                                                               prolactin.                       
Lactation *.................................................................  Offspring growth, milk quantity   
                                                                               and quality.                     
Development.................................................................  Normality of external genitalia *,
                                                                               vaginal opening, vaginal smear   
                                                                               cytology, onset of estrous       
                                                                               behavior (menstruation *).       
Senescence..................................................................  Vaginal smear cytology, ovarian   
                                                                               histology (menopause *).         
----------------------------------------------------------------------------------------------------------------
* Endpoints that can be obtained relatively noninvasively with humans.                                          

    Reproductive function in the female is controlled through complex 
interactions involving the central nervous system (particularly the 
hypothalamus), pituitary, ovaries, the reproductive tract, and the 
secondary sexual organs. Other nongonadotrophic components of the 
endocrine system may also modulate reproductive system function. 
Because it is difficult to measure certain important aspects of female 
reproductive function (e.g., increased rate of follicular atresia, 
ovulation failure), assessment of the endocrine status may provide 
needed insight that is not otherwise available.
    To understand the significance of effects on the reproductive 
endpoints, it is critical that the relationships between the various 
reproductive hormones and the female reproductive organs be understood. 
Although certain effects may be identified routinely as adverse, all of 
the results should be considered in the context of the known biology.
    The format used below for presentation of the female reproductive 
endpoints is altered from that used for the male to allow examination 
of events that are linked and that fluctuate with the changing 
endocrine status. Particularly, the organ weight, gross morphology, and 
histology are combined for each organ. Endpoints and endocrine factors 
for the individual female reproductive organs are discussed, with 
emphasis on the nonpregnant animal. This is followed by examination of 
measures of cyclicity and their interpretation. Then, considerations 
relevant to prepubertal, pregnant, lactating, and aging females are 
presented.
III.B.4.b. Body Weight, Organ Weight, Organ Morphology, and Histology
    III.B.4.b.1. Body weight. Toxicologists are often concerned about 
how a change in body weight may affect reproductive function. In 
females, an important consideration is that body weight fluctuates 
normally with the physiologic state of the animal because estrogen and 
progesterone are known to influence food intake and energy expenditure 
to an important extent (Wang, 1923; Wade, 1972). Water retention and 
fat deposition rates are also affected (Galletti and Klopper, 1964; 
Hervey and Hervey, 1967). Food consumption is elevated during 
pregnancy, in part because of the elevated serum progesterone level. 
One of the most sensitive noninvasive indicators of a compound with 
estrogenic action in the female rat is a reduction in food intake and 
body weight. Also, growth retardation induced by effects on 
extragonadal hormones (e.g., thyroid or growth hormone) can cause a 
delay in pubertal development, and induce acyclicity and infertility. 
Because of these endocrine-related fluctuations, the weights of the 
reproductive organs are poorly correlated with body weight, except in 
extreme cases. Thus, actual organ weight data, rather than organ to 
body weight ratios, should be reported and evaluated for the female 
reproductive system.
    Chapin et al. (1993a, b) have studied the influence of food 
restriction on female Sprague-Dawley rats and Swiss CD-1 mice when body 
weights were 90%, 80%, or 70% of controls. Female rats were resistant 
to effects on reproductive function at 80% of control weight whereas 
mice showed adverse effects at 80% and a marginal effect at 90%. These 
results indicate that differences exist between species (and probably 
between strains) in the response of the female rodent reproductive 
system to reduced food intake or body weight reduction.
    III.B.4.b.2. Ovary. The ovary serves a number of functions that are 
critical to reproductive activity, including production and ovulation 
of oocytes.

[[Page 56292]]

Estrogen is produced by developing follicles and progesterone is 
produced by corpora lutea that are formed after ovulation.

Ovarian Weight

    Significant increases or decreases in ovarian weight compared with 
controls should be considered an indication of female reproductive 
toxicity. Although ovarian function shifts throughout the estrous 
cycle, ovarian weight in the normal rat does not show significant 
fluctuations. Still, oocyte and follicle depletion, persistent 
polycystic ovaries, inhibition of corpus luteum formation, luteal cyst 
development, reproductive aging, and altered hypothalamic-pituitary 
function may all be associated with changes in ovarian weight. 
Therefore, it is important that ovarian gross morphology and histology 
also be examined to allow correlation of alterations in those 
parameters with changes in ovarian weight. However, not all adverse 
histologic alterations in the ovary are concurrent with changes in 
ovarian weight. Therefore, a lack of effect on organ weights does not 
preclude the need for histologic evaluation.

Histopathology

    Histologic evaluation of the three major compartments of the ovary 
(i.e., follicular, luteal, and interstitial) plus the epithelial 
capsule and ovarian stroma may indicate ovarian toxicity. A number of 
pathologic conditions can be detected by ovarian histology (Kurman and 
Norris, 1978; Langley and Fox, 1987). Methods are available to quantify 
the number of follicles and their stages of maturation (Plowchalk et 
al., 1993). These techniques may be useful when a compound depletes the 
pool of primordial follicles or alters their subsequent development and 
recruitment during the events leading to ovulation.

Adverse Effects

    Significant changes in the ovaries in any of the following effects 
should be considered adverse:
     Increase or decrease in ovarian weight.
     Increased incidence of follicular atresia.
     Decreased number of primary follicles.
     Decreased number or lifespan of corpora lutea.
     Evidence of abnormal folliculogenesis or luteinization, 
including cystic follicles, luteinized follicles, and failure of 
ovulation.
     Evidence of altered puberty or premature reproductive 
senescence.
    III.B.4.b.3. Uterus.

Uterine Weight

    An alteration in the weight of the uterus may be considered an 
indication of female reproductive organ toxicity. Compounds that 
inhibit steroidogenesis and cyclicity can dramatically reduce the 
weight of the uterus so that it appears atrophic and small. However, 
uterine weight fluctuates three- to four-fold throughout the estrous 
cycle, peaking at proestrus when, in response to increased estrogen 
secretion, the uterus is fluid filled and distended. This increase in 
uterine weight has been used as a basis for comparing relative potency 
of estrogenic compounds in bioassays (Kupfer, 1987). As a result of the 
wide fluctuations in weight, uterine weights taken from cycling animals 
have a high variance, and large compound-related effects are required 
to demonstrate a significant effect unless interpreted relative to that 
animal's estrous cycle stage. A number of environmental compounds 
(e.g., pesticides such as methoxychlor and chlordecone, mycotoxins, 
polychlorinated biphenyls, alkylphenols, and phytoestrogens) possess 
varying degrees of estrogenic activity and have the potential to 
stimulate the female reproductive tract (Barlow and Sullivan, 1982; 
Bulger and Kupfer, 1985; Hughes, 1988).
    When pregnant or postpartum animals are examined, the numbers of 
implantation sites or implantation scars should be counted. This 
information, along with corpus luteum counts, can be used to calculate 
pre- and postimplantation losses.

Histopathology

    The histologic appearance of the normal uterus fluctuates with 
stage of the estrous cycle and pregnancy. The uterine endometrium is 
sensitive to influences of estrogens and progestogens (Warren et al., 
1967), and extended treatment with these compounds leads to hypertrophy 
and hyperplasia. Conversely, inhibition of ovarian activity and reduced 
steroid secretion results in endometrial hypoplasia and atrophy, as 
well as altered vaginal smear cytology. Effects induced during 
development may delay or prevent puberty, resulting in persistence of 
infantile genitalia.

Adverse Effects

    Effects on the uterus that may be considered adverse include 
significant dose-related alteration of weight, as well as gross 
anatomic or histologic abnormalities. In particular, any of the 
following effects should be considered as adverse.
     Infantile or malformed uterus or cervix.
     Decreased or increased uterine weight.
     Endometrial hyperplasia, hypoplasia, or aplasia.
     Decreased number of implantation sites.
    III.B.4.b.4. Oviducts.
    Typically, the oviducts are not weighed or examined histologically 
in tests for reproductive toxicity. However, information from visual 
and histologic examinations is of value in detecting morphologic 
anomalies. Descriptions of pathologic effects within the oviducts of 
animals other than humans are not common. Hypoplasia of otherwise well-
formed oviducts and loss of cilia result most commonly from a lack of 
estrogen stimulation, and for this reason, this condition may not be 
recognized until after puberty. Hyperplasia of the oviductal epithelium 
results from prolonged estrogenic stimulation. Anomalies induced during 
development have also been described, including agenesis, segmental 
aplasia, and hypoplasia.
    Anatomic anomalies in the oviduct occurring in excess of control 
incidence should be considered as adverse effects. Hypoplasia or 
hyperplasia of the oviductal epithelium may be considered as an adverse 
effect, particularly if that result is consistent with observations in 
the uterine histology.
    III.B.4.b.5. Vagina and external genitalia.

Vaginal Weight

    Vaginal weight changes should parallel those seen in the uterus 
during the estrous cycle, although the magnitude of the changes is 
smaller.

Histopathology

    In rodents, cytologic changes in the vaginal epithelium (vaginal 
smear) may be used to identify the different stages of the estrous 
cycle (see Section III.B.4.d.). The vaginal smear pattern may be useful 
to identify conditions that would delay or preclude fertility, or 
affect sexual behavior. Other histologic alterations that may be 
observed include aplasia, hypoplasia, and hyperplasia of the vaginal 
epithelial cell lining.

Developmental Effects

    Developmental abnormalities, either genetic or related to prenatal 
exposure to compounds that disrupt the endocrine balance, include 
agenesis, hypoplasia, and dysgenesis. Hypoplasia of the vagina may be 
concomitant with hyperplasia of the external genitalia and can be 
induced by gonadal or adrenal steroid exposure. In rodents,

[[Page 56293]]

malpositioning of the vaginal and urethral ducts is common in steroid-
treated females. Such developmentally induced lesions are irreversible.
    The sex ratio observed at birth may be affected by exposure of 
genotypic females in utero to agents that disrupt reproductive tract 
development. In cases of incomplete sex reversal because of such 
exposures, female rodents may appear more male-like and have an 
increased ano-genital distance (Gray and Ostby, 1995).
    At puberty, the opening of the vaginal orifice normally provides a 
simple and useful developmental marker. However, estrogenic or 
antiestrogenic chemicals can act directly on the vaginal epithelium and 
alter the age at which vaginal patency occurs without truly affecting 
puberty.

Adverse Effects

    Significant effects on the vagina that may be considered adverse 
include the following:
     Increases or decreases in weight
     Infantile or malformed vagina or vulva, including 
masculinized vulva or increased ano-genital distance
     Vaginal hypoplasia or aplasia
     Altered timing of vaginal opening
     Abnormal vaginal smear cytology pattern
    III.B.4.b.6. Pituitary.

Pituitary Weight

    Alterations in weight of the pituitary gland should be considered 
an adverse effect. The discussion on pituitary weight and histology for 
males (see Section III.B.3.b.) is pertinent also for females. Pituitary 
weight increases normally with age, as well as during pregnancy and 
lactation. Changes in pituitary weight can occur also as a consequence 
of chemical stimulation. Increased pituitary weight often precedes 
tumor formation, particularly in response to treatment with estrogenic 
compounds. Increased pituitary size associated with estrogen treatment 
may be accompanied by hyperprolactinemia and constant vaginal estrus. 
Decreased pituitary weight is less common but may result from decreased 
estrogenic stimulation (Cooper et al., 1989).

Histopathology

    In histologic evaluations with rats and mice, the relative size of 
cell types in the anterior pituitary (acidophils and basophils) has 
been reported to vary with the stages of the reproductive cycle and in 
pregnancy (Holmes and Ball, 1974). Therefore, the relationship of 
morphologic pattern to estrous or menstrual cycle stage or pregnancy 
status should be considered in interpreting histologic observations on 
the female pituitary.

Adverse Effects

    A significant increase or decrease in pituitary weight should be 
considered an adverse effect. Significant histopathologic damage in the 
pituitary should be considered an adverse effect, but should be shown 
to involve cells that control gonadotropin or prolactin production to 
be called a reproductive effect.
III.B.4.c. Oocyte Production
    III.B.4.c.1. Folliculogenesis. In normal females, all of the 
follicles (and the resident oocytes) are present at or soon after 
birth. The large majority of these follicles undergo atresia and are 
not ovulated. If the population of follicles is depleted, it cannot be 
replaced and the female will be rendered infertile. In humans, 
depletion of oocytes leads to premature menopause. Ovarian follicle 
biology and toxicology have been reviewed by Crisp (1992).
    In rodents, lead, mercury, cadmium, and polyaromatic hydrocarbons 
have all been implicated in the arrest of follicular growth at various 
stages of the life cycle (Mattison and Thomford, 1989). Susceptibility 
to oocyte toxicity varies considerably between species (Mattison and 
Thorgeirsson, 1978).
    Environmental agents that affect gonadotropin-mediated ovarian 
steroidogenesis or follicular maturation can prolong the follicular 
phase of the estrous or menstrual cycle and cause atresia of follicles 
that would otherwise ovulate. Estrogenic as well as antiestrogenic 
agents can produce this effect. Also, normal follicular maturation is 
essential for normal formation and function of the corpus luteum formed 
after ovulation (McNatty, 1979).
    III.B.4.c.2. Ovulation. Chemicals can delay or block ovulation by 
disrupting the ovulatory surge of luteinizing hormone (LH) or by 
interfering with the ability of the maturing follicle to respond to 
that gonadotropic signal. Examples for rats include compounds that 
interfere with normal central nervous system (CNS) norepinephrine 
receptor stimulation such as the pesticides chlordimeform and amitraz 
(Goldman et al., 1990, 1991) and compounds that interfere with 
norepinephrine synthesis such as the fungicide thiram (Stoker et al., 
1993). Compounds that increase central opioid receptor stimulation also 
decrease serum LH and inhibit ovulation in monkeys and rats (Pang et 
al., 1977; Smith, C.G., 1983). Delayed ovulation can alter oocyte 
viability and cause trisomy and polyploidy in the conceptus (Fugo and 
Butcher, 1966; Butcher and Fugo, 1967; Butcher et al., 1969, 1975; Na 
et al., 1985). Delayed ovulation induced by exposure to the pesticide 
chlordimeform has also been shown to alter fetal development and 
pregnancy outcome in rats (Cooper et al., 1994).
    III.B.4.c.3. Corpus luteum. The corpus luteum arises from the 
ruptured follicle and secretes progesterone, which has an important 
role in the estrous or menstrual cycle. Luteal progesterone is also 
required for the maintenance of early pregnancy in most mammalian 
species, including humans (Csapo and Pulkkinen, 1978). Therefore, 
establishment and maintenance of normal corpora lutea are essential to 
normal reproductive function. However, with the exception of 
histopathologic evaluations that may establish only their presence or 
absence, these structures are not evaluated in routine testing. 
Additional research is needed to determine the importance of 
incorporating endpoints that examine direct effects on luteal function 
in routine toxicologic testing.

Adverse Effects

    Increased rates of follicular atresia and oocyte toxicity leads to 
premature menopause in humans. Altered follicular development, 
ovulation failure, or altered corpus luteum formation and function can 
result in disruption of cyclicity and reduced fertility, and, in 
nonprimates, interference with normal sexual behavior. Therefore, 
significant increases in the rate of follicular atresia, evidence of 
oocyte toxicity, interference with ovulation, or altered corpus luteum 
formation or function should be considered adverse effects.
    III.B.4.d. Alterations in the Female Reproductive Cycle. The 
pattern of events in the estrous cycle may provide a useful indicator 
of the normality of reproductive neuroendocrine and ovarian function in 
the nonpregnant female. It also provides a means to interpret hormonal, 
histologic, and morphologic measurements relative to stage of the 
cycle, and can be useful to monitor the status of mated females. 
Estrous cycle normality can be monitored in the rat and mouse by 
observing the changes in the vaginal smear cytology (Long and Evans, 
1922; Cooper et al., 1993). To be most useful with cycling females, 
vaginal smear cytology should be examined daily for at least three 
normal estrous cycles prior to treatment, after onset of treatment, and 
before necropsy (Kimmel, G.A. et al., 1995). However, practical

[[Page 56294]]

limitations in testing may limit the examination to the period before 
mating or necropsy.
    Daily vaginal smear data from rodents can provide useful 
information on (1) cycle length, (2) occurrence or persistence of 
estrus, (3) duration or persistence of diestrus, (4) incidence of 
spontaneous pseudopregnancy, (5) distinguishing pregnancy from 
pseudopregnancy (based on the number of days the smear remains 
leukocytic), and (6) indications of fetal death and resorption by the 
presence of blood in the smear after day 12 of gestation. The technique 
also can detect onset of reproductive senescence in rodents (LeFevre 
and McClintock, 1988). It is useful further to detect the presence of 
sperm in the vagina as an indication of mating.
    In nonpregnant females, repetitive occurrence of the four stages of 
the estrous cycle at regular, normal intervals suggests that 
neuroendocrine control of the cycle and ovarian responses to that 
control are normal. Even normal, control animals can show irregular 
cycles. However, a significant alteration compared with controls in the 
interval between occurrence of estrus for a treatment group is cause 
for concern. Generally, the cycle will be lengthened or the animals 
will become acyclic. Lengthening of the cycle may be a result of 
increased duration of either estrus or diestrus. Knowing the affected 
phase can provide direction for further investigation.
    The persistence of regular vaginal cycles after treatment does not 
necessarily indicate that ovulation occurred, because luteal tissue may 
form in follicles that have not ruptured. This effect has been observed 
after treatment with anti-inflammatory agents (Walker et al., 1988). 
However, that effect should be reflected in reduced fertility. 
Conversely, subtle alterations of cyclicity can occur at doses below 
those that alter fertility (Gray et al., 1989).
    Irregular cycles may reflect impaired ovulation. Extended vaginal 
estrus usually indicates that the female cannot spontaneously achieve 
the ovulatory surge of LH (Huang and Meites, 1975). A number of 
compounds have been shown to alter the characteristics of the LH surge 
including anesthetics (Nembutal), neurotransmitter receptor binding 
agents (Drouva et al., 1982), and the pesticides chlordimeform and 
lindane (Cooper et al., 1989; Morris et al., 1990). Persistent or 
constant vaginal cornification (or vaginal estrus) may result from one 
or several effects. Typically, in the adult, if the vaginal epithelium 
becomes cornified and remains so in response to toxicant exposure, it 
is the result of the agent's estrogenic properties (i.e., DES or 
methoxychlor), or the ability of the agent to block ovulation. In the 
latter case, the follicle persists and endogenous estrogen levels bring 
about the persistent vaginal cornification. Histologically, the ovaries 
in persistent estrus will be atrophied following exposure to estrogenic 
substances. In contrast, the ovaries of females in which ovulation has 
been blocked because of altered gonadotropin secretion will contain 
several large follicles and no corpora lutea. Females in constant 
estrus may be sexually receptive regardless of the mechanism 
responsible for this altered ovarian condition. However, if ovulation 
has been blocked by the treatment, an LH surge may be induced by mating 
(Brown-Grant et al., 1973; Smith, E.R. and Davidson, 1974) and a 
pregnancy or pseudopregnancy may ensue. The fertility of such matings 
is reduced (Cooper et al., 1994). Significant delays in ovulation can 
result in increased embryonic abnormalities and pregnancy loss (Fugo 
and Butcher, 1966; Cooper et al., 1994).
    Persistent diestrus indicates temporary or permanent cessation of 
follicular development and ovulation, and thus at least temporary 
infertility. Prolonged vaginal diestrus, or anestrus, may be indicative 
of agents (e.g., polyaromatic hydrocarbons) that interfere with 
follicular development or deplete the pool of primordial follicles 
(Mattison and Nightingale, 1980) or agents such as atrazine that 
interrupt gonadotropin support of the ovary (Cooper et al., 1996). 
Pseudopregnancy is another altered endocrine state reflected by 
persistent diestrus. A pseudopregnant condition also has been shown to 
result in rats following single or multiple doses of atrazine (Cooper 
et al., 1996). The ovaries of anestrous females are atrophic, with few 
primary follicles and an unstimulated uterus (Huang and Meites, 1975). 
Serum estradiol and progesterone are abnormally low.

Adverse Effects

    Significant evidence that the estrous cycle (or menstrual cycle in 
primates) has been disrupted should be considered an adverse effect. 
Included should be evidence of abnormal cycle length or pattern, 
ovulation failure, or abnormal menstruation.
    III.B.4.e. Mammary Gland and Lactation. The mammary glands of 
normal adults change dramatically during the period around parturition 
because of the sequential effects of a number of gonadal and 
extragonadal hormones. Milk letdown is dependent on the suckling 
stimulus and the release of oxytocin from the posterior pituitary. 
Thus, mammary tissue is highly endocrine dependent for development and 
function (Wolff, 1993; Imagawa et al., 1994; Tucker, 1994).
    Mammary gland size, milk production and release, and histology can 
be affected adversely by toxic agents, and many exogenous chemicals and 
drugs are transferred into milk (American Academy of Pediatrics 
Committee on Drugs, 1994; Oskarsson et al., 1995; Sonawane, 1995). 
Reduced growth of young could be caused by reduced milk availability, 
palatability or quality, by ingestion of a toxic agent secreted into 
the milk, or by other factors unrelated to lactational ability (e.g., 
deficient suckling ability or deficient maternal behavior). Perinatal 
exposure to steroid hormones and other chemicals can alter mammary 
gland morphology and tumor potential in adulthood. Because of the 
tendency for mobilization of lipids from adipose tissue and secretion 
of those lipids into milk by lactating females, milk may contain 
lipophilic agents at concentrations equal to or higher than those 
present in the blood or organs of the dam. Thus, suckling offspring may 
be exposed to elevated levels of such agents.
    Techniques for measuring mammary tissue development, nucleic acid 
content, milk production and milk composition in rodents are discussed 
by Tucker (1994). During lactation, the mammary glands can be dissected 
and weighed only with difficulty. RNA content of the mammary glands may 
be measured as an index of lactational potential. More direct estimates 
of milk production may be obtained by measuring litter weights of milk-
deprived pups taken before and after nursing. Milk from the stomachs of 
pups treated similarly can also be weighed at necropsy. Cleared and 
stained whole mounts of the mammary gland can be prepared at necropsy 
for histologic examination. The DNA, RNA, and lipid content of the 
mammary gland and the composition of the milk have been measured 
following toxicant administration as indicators of toxicity to this 
target organ.
    Significant reductions in milk production or negative effects on 
milk quality, whether measured directly or reflected in impaired 
development of young, should be considered adverse reproductive 
effects.
    III.B.4.f. Reproductive Senescence. With advancing age, there is a 
loss of the regular ovarian cycles and associated normal cyclical 
changes in the uterine and vaginal epithelium that

[[Page 56295]]

are typical of the young-adult female rat (Cooper and Walker, 1979). 
Although the mechanisms responsible for this loss of cycling are not 
thoroughly understood, age-dependent changes occur within the 
hypothalamic-pituitary control of ovulation (Cooper et al., 1980; Finch 
et al., 1984). Cumulative exposure to estrogen secreted by the ovary 
may play a role, as treatment with estrogens during adulthood can 
accelerate the age-related loss of ovarian function (Brawer and Finch, 
1983). In contrast, the principal cause of the loss of ovarian cycling 
in humans appears to be the depletion of oocytes (Mattison, 1985).
    Prenatal or postnatal treatment of females with estrogens or 
estrogenic pesticides can also cause impaired ovulation and sterility 
(Gorski, 1979). These observations imply that alterations in ovarian 
function may not be noticeable immediately after treatment but may 
become evident at puberty or influence the age at which reproductive 
senescence occurs.

Adverse Effects

    Significant effects on measures showing a decrease in the age of 
onset of reproductive senescence in females should be considered 
adverse. Cessation of normal cycling, which is measured by vaginal 
smear cytology, ovarian histopathology, or an endocrine profile that is 
consistent with this interpretation, should be included as an adverse 
effect.
III.B.5. Developmental and Pubertal Alterations

Developmental Effects

    Alterations of reproductive differentiation and development, 
including those produced by endocrine system disruption, can result in 
infertility, functional and morphologic alterations of the reproductive 
system, and cancer (Steinberger and Lloyd, 1985; Gray, 1991). Prenatal 
and postnatal exposure to toxicants can produce changes that may not be 
predicted from effects seen in adults, and those effects are often 
irreversible. Adverse developmental outcomes in either sex can result 
from exposure to toxic agents in utero, through contact with exposed 
dams, or in milk. Dosing of dams during lactation also can result in 
developmental effects through impaired nursing capability of the dams.
    Effects observed in rodents following developmental exposure to 
agents can include alterations in the genitalia (including ano-genital 
distance), inhibited (female) or retained (male) nipple development, 
impaired sexual behavior, delay or acceleration of the onset of 
puberty, and reduced fertility (Gray et al., 1985, 1994, 1995; Gray and 
Ostby, 1995; Kelce et al., 1995). Effects may include altered sexual 
behavior or ability to produce gametes normally that are not observed 
until after puberty. Hepatic enzyme systems for steroid metabolism that 
are imprinted during development may be altered in males. Testis 
descent from the abdominal cavity into the scrotum may be delayed or 
may not occur. Generally, the type of effect seen may differ depending 
on the stage of development at which the exposure occurred.
    Many of these effects have been detected in human females and males 
exposed prenatally to diethylstilbestrol (DES), other estrogens, 
progestins, androgens, and anti-androgens (Giusti et al., 1995; 
Harrison et al., 1995). Accelerated reproductive aging and tumors of 
the reproductive tract have been observed in laboratory animal and 
human females after pre- or perinatal exposure to hormonally active 
agents. However, capability to alter sexual differentiation is not 
limited to agents with known direct hormonal activity. Other agents, 
for which the mode of action is not known (e.g., busulfan, nitrofen), 
or which affect the endocrine system indirectly (e.g., PCBs, dioxin), 
may act via different mechanisms during critical periods of development 
to alter sexual differentiation and reproductive system development.

Effects on Puberty

    In female rats and mice, the age at vaginal opening is the most 
commonly measured marker of puberty. This event results from an 
increase in the blood level of estradiol. The ages and weights of 
females at the first cornified (estrous) vaginal smear, the first 
diestrous smear, and the onset of vaginal cycles have also been used as 
endpoints for onset of puberty. In males, preputial separation or 
appearance of sperm in expressed urine or ejaculates can serve as 
markers of puberty. Body weight at puberty may provide a means to 
separate specific delays in puberty from those that are related to 
general delays in development. Agents may differentially affect the 
endpoints related to puberty onset, so it is useful to have information 
on more than one marker.
    Puberty can be accelerated or delayed by exogenous agents, and both 
types of effects may be adverse (Gray et al., 1989, 1995; Gray and 
Ostby, 1995; Kelce et al., 1995). For example, an acceleration of 
vaginal opening may be associated with a delay in the onset of 
cyclicity, infertility, and with accelerated reproductive aging 
(Gorski, 1979). Delays in pubertal development in rodents are usually 
related to delayed maturation or inhibition of function of the 
hypothalamic-pituitary axis. Adverse reproductive outcomes have been 
reported in rodents when puberty is altered by a week or more, but the 
biologic relevance of a change in these measures of a day or two is 
unknown (Gray, 1991).

Adverse Effects

    Effects induced or observed during the pre- or perinatal period 
should be judged using guidance from the Guidelines for Developmental 
Toxicity Risk Assessment (U.S. EPA, 1991) as well as from these 
Guidelines. Significant effects on ano-genital distance or age at 
puberty, either early or delayed, should be considered adverse as 
should malformations of the internal or external genitalia. Included as 
adverse effects for females should be effects on nipple development, 
age at vaginal opening, onset of cyclic vaginal smears, onset of estrus 
or menstruation, or onset of an endocrine or behavioral pattern 
consistent with estrous or menstrual cyclicity. Included as adverse 
effects for males should be delay or failure of testis descent, as well 
as delays in age at preputial separation or appearance of sperm in 
expressed urine or ejaculates.
III.B.6. Endocrine Evaluations
    Toxic agents can alter endocrine system function by affecting any 
part of the hypothalamic-pituitary-gonadal-reproductive tract axis. 
Effects may be induced in either sex by altering hormone synthesis, 
storage, release, transport, or clearance, as well as by altering 
hormone receptor recognition or postreceptor responses. The involvement 
of the endocrine system in female reproductive physiology and 
toxicology has been presented to a substantial degree as a necessary 
component in Section III.B.4. (Female-specific Endpoints). The 
information in that section should be considered together with the 
following material.
    The male reproductive system can be affected adversely by 
disruption of the normal endocrine balance. In adults, effects that 
result in interference with normal concentrations or action of LH and/
or follicle stimulating hormone (FSH) can decrease or abolish 
spermatogenesis, affect secondary sex organ (e.g., epididymis) and 
accessory sex gland (e.g., prostate, seminal vesicle) function, and 
impair sexual behavior (Sharpe, 1994). In mammals, a female 
reproductive tract develops unless androgen is produced and utilized 
normally by the fetus (Byskov and Hoyer, 1994; George and Wilson, 
1994).

[[Page 56296]]

Therefore, the consequences of disruption of the normal endocrine 
pattern during development of the male reproductive system pre- and 
postnatally are of particular concern. Differentiation and development 
of the male reproductive system are especially sensitive to substances 
that interfere with the production or action of androgens (testosterone 
and dihydrotestosterone). Sexual differentiation of the CNS can be 
affected also. Therefore, interference with normal production or 
response to androgens can result in a range of abnormal effects in 
genotypic males ranging from a pseudohermaphrodite condition to 
reduction in sperm production or altered sexual behavior. Chemicals 
with estrogenic or anti-androgenic activity have been identified that 
are capable, with sufficient exposure levels, of causing effects of 
these types in males (Gray et al., 1994; Harrison et al., 1995; Kelce 
et al., 1995). While sensitivity may differ, it is likely that 
mechanisms of action for these endocrine disrupting agents will be 
consistent across mammalian species. Chemicals with the ability to 
interact with the Ah receptor (e.g., dioxin or PCBs) may also disrupt 
reproductive system development or function (Brouwer et al., 1995; 
Safe, 1995). Several of the effects seen with exposure of male and 
female rats and hamsters differ from those caused by estrogens, 
indicating a different mechanism of action.
    The developing nervous system can be a target of chemicals. In 
rats, sexual differentiation of the CNS can be modified by hormonal 
treatments or exposure to environmental agents that mimic or interfere 
with the action of certain hormones. Prior to gender differentiation, 
the brain is inherently female or at least bipotential (Gorski, 1986). 
Thus, the functional and structural sex differences in the CNS are not 
due directly to sex differences in neuronal genomic expression, but 
rather are imprinted by the gonadal steroid environment during 
development.
    Chemicals with endocrine activity have been shown to masculinize 
the CNS of female rats. Examples include chlordecone (Gellert, 1978), 
DDT (Bulger and Kupfer, 1985), and methoxychlor (Gray et al., 1989). 
Exposure of newborn female rats to these agents during the critical 
period of sexual differentiation can alter the timing of puberty and 
perturb subsequent reproductive function, presumably by altering the 
development of the neural mechanisms that regulate gonadotropin 
secretion.
    In females, the situation is more complex than in males due to the 
female cycle, the fertilization process, gestation and lactation. All 
of the functions of the female reproductive system are under endocrine 
control, and therefore can be susceptible to disruption by effects on 
the reproductive endocrine system.
    As with males, disturbance of the normal endocrine patterns during 
development can result in abnormal development of the female 
reproductive tract at exposure levels that tend to be lower than those 
affecting adult females (Gellert, 1978; Brouwer et al., 1995). 
Consistent with the differentiation mechanism described above, exposure 
of genotypic females to androgens causes formation of 
pseudohermaphrodite reproductive tracts with varying degrees of 
severity as well as alteration of brain imprinting. However, exposure 
to estrogenic substances during development also results in adverse 
effects on anatomy and function including, in rats, malformations of 
the genitalia. Exposure of human females to diethylstilbestrol in utero 
has been shown to cause an increased incidence of vaginal clear cell 
adenoma (Giusti et al., 1995). Dioxin, presumably acting through the Ah 
receptor, also disrupts development of the female reproductive system 
(Gray and Ostby, 1995).
    Endpoints can be included in standardized toxicity testing that are 
capable of detecting, but are not specific for, effects of reproductive 
endocrine system disruption. For effects of exposure on adults, 
endpoints can be incorporated into the subchronic toxicity protocol or 
into reproductive toxicity protocols. For effects that are induced 
during development, protocols that include exposure throughout the 
development process and allow evaluation of the offspring 
postpubertally are needed. Data from specialized testing, including in 
vitro screening tests, may be useful to evaluate further the site, 
timing, and mechanism of action.
    Endpoints that can detect endocrine-related effects with adult-only 
exposure in standardized testing include evaluation of fertility, 
reproductive organ appearance, weights, and histopathology, oocyte 
number, cycle normality and mating behavior. Endpoints that can detect 
effects induced by endocrine system disruption during development 
include, in addition to those identified for adult-exposed animals, the 
reproductive developmental endpoints identified in Section III.B.5. 
Significant effects on any of these measures may be considered to be 
adverse if the results are consistent and biologically plausible.
    Levels of the reproductive hormones are not available routinely 
from toxicity testing. However, measurements of the reproductive 
hormones in males offer useful supplemental information in assessing 
potential reproductive toxicity for test species (Sever and Hessol, 
1984; Heywood and James, 1985; NRC, 1989). Such measurements have 
increased importance with humans where invasiveness of approaches must 
be limited. The reproductive hormones measured often are circulating 
levels of LH, FSH, and testosterone. Other useful measures that may be 
available include prolactin, inhibin, and androgen binding protein 
levels. In addition, challenge tests with exogenous agents (e.g., 
gonadotropin releasing hormone, LH, or human chorionic gonadotropin) 
may provide insight into the functional responsiveness of the pituitary 
or Leydig cells.
    Interpretation of endocrine effects is facilitated if information 
is available on a battery of hormones. However, in evaluating such 
data, it is important to consider that serum hormones such as FSH, LH, 
prolactin, and androgens exhibit cyclic variations within a 24-hour 
period (Fink, 1988). Thus, the time of sampling should be controlled 
rigorously to avoid excessive variability (Nett, 1989). Sequential 
sampling can allow detection of treatment-related changes in circadian 
and pulsatile rhythms.
    The pattern seen in levels of reproductive system hormones can 
provide useful information about the possible site and type of effect 
on reproductive system function. For example, if a compound acts at the 
level of the hypothalamus or pituitary, then serum LH and FSH may be 
decreased, leading to decreased testosterone levels. On the other hand, 
severe interference with Sertoli cell function or spermatogenesis would 
be expected to elevate serum FSH levels. An agent having antiandrogenic 
activity in adults might elevate serum LH and testosterone. Testis 
weight might be unaffected, while the weight and size of the accessory 
sex glands may be reduced. The endocrine profile presented by exposure 
to specific antiandrogens can differ markedly because of differences in 
tissue specificity and receptor kinetics, as well as age at which 
exposure occurred.

Adverse Effects

    In the absence of endocrine data, significant effects on 
reproductive system anatomy, sexual behavior, pituitary, uterine or 
accessory sex gland

[[Page 56297]]

weights or histopathology, female cycle normality, or Leydig cell 
histopathology may suggest disruption of the endocrine system. In those 
instances, additional testing for endocrine effects may be indicated. 
Significant alterations in circulating levels of estrogen, 
progesterone, testosterone, prolactin, LH, or FSH may be indicative of 
existing pituitary or gonadal injury. When significant alterations from 
control levels are observed in those hormones, the changes should be 
considered cause for concern because they are likely to affect, occur 
in concert with, or result from alterations in gametogenesis, gamete 
maturation, mating ability, or fertility. Such effects, if compatible 
with other available information, may be considered adverse and may be 
used to establish a NOAEL, LOAEL, or benchmark dose. Furthermore, 
endocrine data may facilitate identification of sites or mechanisms of 
toxicant action, especially when obtained after short-term exposures.
III.B.7. In Vitro Tests of Reproductive Function
    Numerous in vitro tests are available and under development to 
measure or detect chemically induced changes in various aspects of both 
male and female reproductive systems (Kimmel, G.L. et al., 1995). These 
include in vitro fertilization using isolated gametes, whole organ 
(e.g., testis, ovary) perfusion, culture of isolated cells from the 
reproductive organs (e.g., Leydig cells, Sertoli cells, granulosa 
cells, oviductal or epididymal epithelium), co-culture of several 
populations of isolated cells, ovaries, quarter testes, seminiferous 
tubule segments, various receptor binding assays on reproductive cells 
and transfected cell lines, and others.
    Tests of sperm properties and function that have been applied to 
reproductive toxicology include penetration of sperm through viscous 
medium (Yeung et al., 1992), in vitro capacitation and fertilization 
assays (Holloway et al., 1990a, b; Perreault and Jeffay, 1993; Slott et 
al., 1995), and evaluation of sperm nuclear integrity (Darney, 1991). 
In addition, evaluation of human sperm function may include sperm 
penetration of cervical mucus, ability of sperm to undergo an acrosome 
reaction, and ability to penetrate zona pellucida-free hamster oocytes 
or bind to human hemi-zona pellucidae (Franken et al., 1990; Liu and 
Baker, 1992).
    The diagnostic information obtained from such tests may help to 
identify potential effects on the reproductive systems. However, each 
test bypasses essential components of the intact animal system and 
therefore, by itself, is not capable of predicting exposure levels that 
would result in toxicity in intact animals. While it is desirable to 
replace whole animal testing to the extent possible with in vitro 
tests, the use of such tests currently is to screen for toxicity 
potential and to study mechanisms of action and metabolism (Perreault, 
1989; Holloway et al., 1990a, b).

III.C. Human Studies

    In principle, human data are scientifically preferable for risk 
assessment since test animal to human extrapolation is not required. At 
this time, reproductive data for humans are available for only a 
limited number of toxicants. Many of these are from occupational 
settings in which exposures tend to be higher than in environmental 
settings. As more data become available, expanding the number of agents 
and endpoints studied and improving exposure assessment, more risk 
assessments will include these data. The following describes the 
methods of generation and evaluation of human data and the relative 
weight the various types of human data should be given in risk 
assessments.
    ``Human studies'' include both epidemiologic studies and other 
reports of individual cases or clusters of events. Typical 
epidemiologic studies include (1) cohort studies in which groups are 
defined by exposure and health outcomes are examined; (2) case-referent 
studies in which groups are defined by health status and prior 
exposures are examined; (3) cross-sectional studies in which exposure 
and outcome are determined at the same time; and (4) ecologic studies 
in which exposure is presumed based typically on residence. Greatest 
weight should be given to carefully designed epidemiologic studies with 
more precise measures of exposure, because they can best evaluate 
exposure-response relationships. This assumes that human exposures 
occur in broad enough ranges for observable differences in response to 
occur. Epidemiologic studies in which exposure is presumed, based on 
occupational title or residence (e.g., some case-referent and all 
ecologic studies), may contribute data for hazard characterization, but 
are of limited use for quantitative risk determination because of the 
generally broad categorical groupings of exposure. Reports of 
individual cases or clusters of events may generate hypotheses of 
exposure-outcome associations, but require further confirmation with 
well-designed epidemiologic or laboratory studies. These reports of 
cases or clusters may support associations suggested by other human or 
test animal data, but cannot stand by themselves in risk assessments.
III.C.1. Epidemiologic Studies
    Good epidemiologic studies provide valuable data for assessment of 
human risk. As there are many different designs for epidemiologic 
studies, simple rules for their evaluation do not exist. Risk assessors 
should seek the assistance of professionals trained in epidemiology 
when conducting a detailed analysis. The following is an overview of 
key issues to consider in evaluation for risk assessment of 
reproductive effects.
    III.C.1.a. Selection of Outcomes for Study. As already discussed, a 
number of endpoints can be considered in the evaluation of adverse 
reproductive effects. However, some of the outcomes are not easily 
observed in humans, such as early embryonic loss, reproductive capacity 
of the offspring, and invasive evaluations of reproductive function 
(e.g., testicular biopsies). Currently, the most feasible endpoints for 
epidemiologic studies are (1) indirect measures of fertility/
infertility; (2) reproductive history studies of some pregnancy 
outcomes (e.g., embryonic/fetal loss, birth weight, sex ratio, 
congenital malformations, postnatal function, and neonatal growth and 
survival); (3) semen evaluations; (4) menstrual history; and (5) blood 
or urinary hormone measures. Factors requiring control in the design or 
analysis (such as effect modifiers and confounders, described below) 
may vary depending on the specific outcomes selected for study.
    The reproductive outcomes available for epidemiologic examination 
are limited by a number of factors, including the relative magnitude of 
the exposure, the size and demographic characteristics of the 
population, and the ability to observe the outcome in humans. Use of 
improved methods for identifying some outcomes, such as embryonic loss 
detected by more sensitive urinary hCG (human chorionic gonadotropin) 
assays, change the spectrum of outcomes available for study (Wilcox et 
al., 1985; Sweeney et al., 1988; Zinaman et al., 1996). Other, less 
accessible, endpoints may require invasive techniques to obtain samples 
(e.g., histopathology) or may have high intra- or interindividual 
variability (e.g., serum hormone levels, sperm count).
    Demographic characteristics of the population, such as marital 
status, age, education, socioeconomic status (SES), and prior 
reproductive history are associated with the probability of

[[Page 56298]]

whether couples will attempt to have children. Differences in birth 
control practices would also affect the number of outcomes available 
for study.
    In addition to the above-mentioned factors, reproductive endpoints 
may be envisioned as effects recognized at various points in a 
continuum starting before conception and continuing through death of 
the progeny. Many studies, however, are limited to evaluating endpoints 
at a particular time in this continuum. For example, in a study of 
defects observed at live birth, a malformed stillbirth would not be 
included, even though the etiology could be identical (Bloom, 1981). 
Also, a different spectrum of outcomes could result from differences in 
timing or in level of exposure (Selevan and Lemasters, 1987).

Human Reproductive Endpoints

    The following section discusses various human male and female 
reproductive endpoints. These outcomes may be an indicator of sub- or 
infertility. These are followed by a discussion of reproductive history 
studies.

Male Endpoints--Semen Evaluations

    The use of semen analysis was discussed in Section III.B.3.d. Most 
epidemiologic studies of potential effects of agents on semen 
characteristics have been conducted in occupational groups and patients 
receiving drug therapy. Obtaining a high level of participation in the 
workforce has been difficult, because social and cultural attitudes 
concerning sex and reproduction may affect cooperation of the study 
groups. Increased participation may occur in men who are planning to 
have children or who are concerned about existing reproductive problems 
or possible ill effects of their exposures. Unless controlled, such 
biased participation may yield unrepresentative estimates of risk 
associated with exposure, resulting in data that are less useful for 
risk assessment. While some studies have response rates greater than 
70% (Ratcliffe et al., 1987; Welch et al., 1988), response rates are 
often less than 70% in such studies and may be even lower in the 
comparison group (Egnatz et al., 1980; Lipshultz et al., 1980; Milby 
and Whorton, 1980; Lantz et al., 1981; Meyer, 1981; Milby et al., 1981; 
Rosenberg et al., 1985; Ratcliffe et al., 1989). Some of the low 
response rates may be caused by inclusion of vasectomized men in the 
total population, although this could vary widely by population (Milby 
and Whorton, 1980). Participation in the comparison group may be biased 
toward those with preexisting reproductive problems. The response rate 
may be improved substantially with proper education and payment of 
subjects (Ratcliffe et al., 1986, 1987).
    Several factors may influence the semen evaluation, including the 
period of abstinence preceding collection of the sample, health status, 
and social habits (e.g., alcohol, recreational drugs, smoking). Data on 
these factors may be collected by interview, subject to the limitations 
described for pregnancy outcome studies.
    Reports of studies with semen analyses have rarely included an 
evaluation of endocrine status (hormone levels in blood or urine) of 
exposed males (Lantz et al., 1981; Ratcliffe et al., 1989). Conversely, 
studies that have examined endocrine status typically do not have data 
on semen quality (Mason, 1990; McGregor and Mason, 1991; Egeland et 
al., 1994).

Female Endpoints

    Reproductive effects may result from a variety of exposures. For 
example, environmental exposures may be toxic to the oocyte, producing 
a loss of primary oocytes that irreversibly affects the woman's 
fecundity. The exposures of importance may occur during the prenatal 
period, and beyond. Oocyte depletion is difficult to examine directly 
in women because of the invasiveness of the tests required; however, it 
can be studied indirectly through evaluation of the age at reproductive 
senescence (menopause) (Everson et al., 1986).
    Numerous diagnostic methods have been developed to evaluate female 
reproductive dysfunction. Although these methods have been used rarely 
for occupational or environmental toxicologic evaluations, they may be 
helpful in defining biologic parameters and the mechanisms related to 
female reproductive toxicity. If clinical observations are able to link 
exposures to the reproductive effect of concern, these data will aid 
the assessment of adverse female reproductive toxicity. The following 
clinical observations include endpoints that may be reported in case 
reports or epidemiologic research studies.
    Reproductive dysfunction also can be studied by the evaluation of 
irregularities of menstrual cycles. However, menstrual cyclicity is 
affected by many parameters such as age, nutritional status, stress, 
exercise level, certain drugs, and the use of contraceptive measures 
that alter endocrine feedback. Vaginal bleeding at menstruation is a 
reflection of withdrawal of steroidogenic support, particularly 
progesterone. Vaginal bleeding can occur at midcycle, in early 
miscarriage, after withdrawal of contraceptive steroids, or after an 
inadequate luteal phase. The length of the menstrual cycle, 
particularly the follicular phase (before ovulation), can vary between 
individuals and may make it difficult to determine significant effects 
on length in populations of women (Burch et al., 1967; Treloar et al., 
1967). Human vaginal cytology may provide information on the functional 
state of reproductive cycles. Cytologic evaluations, along with the 
evaluation of changes in cervical mucus viscosity, can be used to 
estimate the occurrence of ovulation and determine different stages of 
the reproductive cycle (Kesner et al., 1992). Menstrual dysfunction 
data have been used to examine adverse reproductive effects in women 
exposed to potentially toxic agents occupationally (Lemasters, 1992),
    Reports of prospective clinical evaluations of menstrual function 
(Kesner et al., 1992; Wright et al., 1992), have shown urinary 
endocrine measures to be practical and useful. The endocrine status of 
a woman can be evaluated by the measurement of hormones in blood and 
urine. Progesterone can also be measured in saliva. Because the female 
reproductive endocrine milieu changes in a cyclic pattern, single 
sample analysis does not provide adequate information for evaluating 
alterations in reproductive function. Still, a single sample for 
progesterone determination some 7 to 9 days after the estimated 
midcycle surge of gonadotropins in a regularly cycling woman may 
provide suggestive evidence for the presence of a functioning corpus 
luteum and prior follicular maturation and ovulation. Clinically 
abnormal levels of gonadotropins, steroids, or other biochemical 
parameters may be detected from a single sample. However, a much 
stronger design involves collection of multiple samples and their 
observation in conjunction with events in the menstrual cycle.
    The day of ovulation can be estimated by the biphasic shift in 
basal body temperature. Ovulation can also be detected by serial 
measurement of hormones in the blood or urine and analyses of estradiol 
and gonadotropin status at midcycle. After ovulation, luteal phase 
function can be assessed by analysis of progesterone secretion and by 
evaluation of endometrial histology. Tubal patency, which could be 
affected by abnormal development, endometriosis or infection, is an 
endpoint that can be observed in clinical evaluations of reproductive

[[Page 56299]]

function (Forsberg, 1981). These latter evaluations of endometrial 
histology and tubal patency are less likely to be present in 
epidemiologic studies or surveillance programs because of the 
invasiveness of the procedures.
III.C.1.b. Reproductive History Studies

Measures of Fertility

    Subfertility may be thought of as nonevents: a couple is unable to 
have children within a specific time frame. Therefore, the 
epidemiologic measurement of reduced fertility or fecundity is 
typically indirect and is accomplished by comparing birth rates or time 
intervals between births or pregnancies. These outcomes have been 
examined using several methods: the Standardized Birth Ratio (SBR; also 
referred to as the Standardized Fertility Ratio) and the length of time 
to pregnancy or birth. In these evaluations, the couple's joint ability 
to procreate is estimated. The SBR compares the number of births 
observed to those expected based on the person-years of observation 
preferably stratified by factors such as time period, age, race, 
marital status, parity, and (if possible) contraceptive use (Wong et 
al., 1979; Levine et al., 1980, 1981, 1983; Levine, 1983; Starr et al., 
1986). The SBR is analogous to the Standardized Mortality Ratio (SMR), 
a measure frequently used in studies of occupational cohorts and has 
similar limitations in interpretation (Gaffey, 1976; McMichael, 1976; 
Tsai and Wen, 1986). The SBR was found to be less sensitive in 
identifying an effect when compared to semen analyses (Welch et al., 
1991). These data can also be analyzed using Poisson regression.
    Analysis of the time between recognized pregnancies or live births 
is a more recent approach to indirect measurement of fertility (Dobbins 
et al., 1978; Baird and Wilcox, 1985; Baird et al., 1986; Weinberg and 
Gladen, 1986; Rowland et al., 1992). Because the time between births 
increases with increasing parity (Leridon, 1977), comparisons within 
birth order (parity) are more appropriate. A statistical method (Cox 
regression) can stratify by birth or pregnancy order to help control 
for nonindependence of these events in the same woman or couple.
    Fertility may also be affected by alterations in sexual behavior. 
However, data linking toxic exposures to these alterations in humans 
are limited and are not obtained easily in epidemiology studies (see 
Section III.C.1.d.).

Developmental Outcomes

    Developmental outcomes examined in human studies of parental 
exposures may include embryo or fetal loss, congenital malformations, 
birth weight effects, sex ratio at birth, and possibly postnatal 
effects (e.g., physical growth and development, organ or system 
function, and behavioral effects of exposure). Developmental effects 
are discussed in more detail in the Guidelines for Developmental 
Toxicity Risk Assessment (U.S. EPA, 1991). As mentioned above, 
epidemiologic studies that focus on only one type of developmental 
outcome or exposures to only one parent may miss a true effect of 
exposure.
    Evidence of a dose-response relationship is usually an important 
criterion in the assessment of exposure to a potentially toxic agent. 
However, traditional dose-response relationships may not always be 
observed for some endpoints (Wilson, 1973; Selevan and Lemasters, 
1987). For example, with increasing dose, a pregnancy might end in 
embryo or fetal loss, rather than a live birth with malformations. A 
shift in the patterns of outcomes could result from differences either 
in level of exposure or in timing (Wilson, 1973; Selevan and Lemasters, 
1987) (for a more detailed description, see Section III.C.1.d.). 
Therefore, a risk assessment should, when possible, attempt to look at 
the relationship of different reproductive endpoints and patterns of 
exposure.
    In addition to the above effects, exposure may produce genetic 
damage to germ cells. Outcomes resulting from germ-cell mutations could 
include reduced probability of fertilization and increased probability 
of embryo or fetal loss and postnatal developmental effects. Based on 
studies with test species, germ cells or early zygotes are critical 
targets of potentially toxic agents. Germ-cell mutagenicity could be 
expressed also as genetic diseases in future generations. 
Unfortunately, these studies are difficult to conduct in human 
populations because of the long time between exposure and outcome and 
the large study groups needed. For more information and guidance on the 
evaluation of these data, refer to the Guidelines for Mutagenicity Risk 
Assessment (U.S. EPA, 1986c).
    III.C.1.c. Community Studies and Surveillance Programs. 
Epidemiologic studies may be based on broad populations such as a 
community, a nationwide probability sample, or surveillance programs 
(such as birth defects registries). Some studies have examined the 
effects of environmental exposures such as potential toxic agents in 
outdoor air, food, water, and soil. These studies may assume certain 
exposures through these routes due to residence (ecologic studies). The 
link between environmental measurements and critical periods of 
exposure for a given reproductive effect may be difficult to make. 
Other studies may go into more detail, evaluating the above routes and 
also indoor air, house dust, and occupational exposures on an 
individual basis (Selevan, 1991). Such environmental studies, relating 
individual exposures to health outcomes should have less 
misclassification of exposure.
    Exposure definition in community studies has some limitations in 
the assessment of exposure-effect relationships. For example, in many 
community-based studies, it may not be possible to distinguish 
maternally mediated effects from paternally mediated effects since both 
parents spend time in the same home environment. In addition, the 
presumably lower exposure levels (compared with industrial settings) 
may require very large groups for the study. A number of case-referent 
studies have examined the relationship between broad classes of 
parental occupation in certain communities or countries and embryo/
fetal loss (Silverman et al., 1985; McDonald et al., 1989; Lindbohm et 
al., 1991), birth defects (Hemminki et al., 1980; Kwa and Fine, 1980; 
Papier, 1985), and childhood cancer (Fabia and Thuy, 1974; Hemminki et 
al., 1981; Peters et al., 1981; Gardner et al., 1990a, b). In these 
reports, jobs are classified typically into broad categories based on 
the probability of exposure to certain classes or levels of exposure. 
Such studies are most helpful in the identification of topics for 
additional study. However, because of the broad groupings of types or 
levels of exposure, these studies are not typically useful for risk 
assessment of any one particular agent.
    Surveillance programs may also exist in occupational settings. In 
this case, reproductive histories (including menstrual cycles) or semen 
evaluations could be followed to monitor reproductive effects of 
exposures. With adequate exposure information, these could yield very 
useful data for risk assessment. Reproductive histories tend to be 
easier and less costly to collect, whereas, a semen evaluation program 
would be rather costly. Success with such programs in the workplace 
will be determined by the confidence the worker has that reproductive 
data are kept confidential and will not affect employment status 
(Samuels, 1988; Lemasters and Selevan, 1993).
    III.C.1.d. Identification of Important Exposures for Reproductive 
Effects. For all examinations of the relationship between reproductive 
effects and

[[Page 56300]]

potentially toxic exposures, defining the exposure that produces the 
effect is crucial. Preconceptional exposures of either parent and in 
utero exposures have been associated with the more commonly examined 
outcomes (e.g., fetal loss, malformations, low birth weight, and 
measures of in- or subfertility). These exposures, plus postnatal 
exposure via breast milk, food, and the environment, may also be 
associated with postnatal developmental effects (e.g., changes in 
growth or in behavioral and cognitive function).
    A number of factors affect the intensity and duration of exposure. 
General environmental exposures are typically lower than those found in 
industrial or agricultural settings. However, this relationship may 
change as exposures are reduced in workplaces and as more is learned 
about environmental exposures (e.g., indoor air exposures, home 
pesticide usage). Larger populations are necessary to achieve 
sufficient power in settings with lower exposures which are likely to 
have lower measures of risk (Lemasters and Selevan, 1984). In addition, 
exposure to individuals may change as they move in and out of areas 
with differing levels and types of exposures, thus affecting the number 
of exposed and comparison events for study.
    Data on exposure from human studies are frequently qualitative, 
such as employment or residence histories. More quantitative data may 
be difficult to obtain because of the nature of certain study designs 
(e.g., retrospective studies) and limitations in estimates of historic 
exposures. Many reproductive effects result from exposures during 
certain critical times. The appropriate exposure classification depends 
on the outcomes studied, the biologic mechanism affected by exposure, 
and the biologic half-life of the agent. The half-life, in combination 
with the patterns of exposure (e.g., continuous or intermittent) 
affects the individual's body burden and consequently the actual dose 
during the critical period. The probability of misclassification of 
exposure status may affect the ability to recognize a true effect in a 
study (Selevan, 1981; Hogue, 1984; Lemasters and Selevan, 1984; Sever 
and Hessol, 1984; Kimmel, C.A. et al., 1986). As more prospective 
studies are done, better estimates of exposure should be developed.
    III.C.1.e. General Design Considerations. The factors that enhance 
a study and thus increase its usefulness for risk assessment have been 
noted in a number of publications (Selevan, 1980; Bloom, 1981; Hatch 
and Kline, 1981; Wilcox, 1983; Sever and Hessol, 1984; Axelson, 1985; 
Tilley et al., 1985; Kimmel, C.A. et al., 1986; Savitz and Harlow, 
1991). Some of the more prominent factors are discussed below.

The Power of the Study

    The power, or ability of a study to detect a true effect, is 
dependent on the size of the study group, the frequency of the outcome 
in the general population, and the level of excess risk to be 
identified. In a cohort study, common outcomes, such as recognized 
fetal loss, require hundreds of pregnancies to have a high probability 
of detecting a modest increase in risk (e.g., 133 pregnancies in both 
exposed and unexposed groups to detect a twofold increase; 
=0.05, power=80%), while less common outcomes, such as the 
total of all malformations recognized at birth, require thousands of 
pregnancies to have the same probability (e.g., more than 1,200 
pregnancies in both exposed and unexposed groups) (Bloom, 1981; 
Selevan, 1981, 1985; Sever and Hessol, 1984; Stein, Z. et al., 1985; 
Kimmel, C.A. et al., 1986). Semen evaluation may require fewer subjects 
depending on the sperm parameters evaluated, especially when each man 
is used as his own control (Wyrobek, 1982, 1984). In case-referent 
studies, study sizes are dependent upon the frequency of exposure 
within the source population. The confidence one has in the results of 
a study showing no effect is related directly to the power of the study 
to detect meaningful differences in the endpoints.
    Power may be enhanced by combining populations from several studies 
using a meta-analysis (Greenland, 1987). The combined analysis could 
increase confidence in the absence of risk for agents showing no 
effect. However, caution must be exercised in the combination of 
potentially dissimilar study groups.
    Results of a negative study should be carefully evaluated, 
examining the power of the study and the degree of concordance or 
discordance between that study and other studies (including careful 
examination of comparability in the details such as similarity of 
adverse endpoints and study design). The consistency among results of 
different studies could be evaluated by comparing statistical 
confidence intervals for the effects found in different studies. 
Studies with lower power will tend to yield wider confidence intervals. 
If the confidence intervals from a negative study and a positive study 
overlap, then there may be no conflict between the results of the two 
studies.

Potential Bias in Data Collection

    Bias may result from the way the study group is selected or 
information is collected (Rothman, 1986). Selection bias may occur when 
an individual's willingness to participate varies with certain 
characteristics relating to exposure or health status. In addition, 
selection bias may operate in the identification of subjects for study. 
For example, in studies of very early pregnancy loss, use of hospital 
records to identify the study group will under-ascertain events, 
because women are not always hospitalized for these outcomes. More 
weight would be given in a risk assessment to a study in which a more 
complete list of pregnancies is obtained by, for example, collecting 
biologic data (e.g., human chorionic gonadotropin [hCG] measurements) 
of pregnancy status from study members. The representativeness of these 
data may be affected by selection factors related to the willingness of 
different groups of women to continue participation over the total 
length of the study. Interview data result in more complete 
ascertainment than hospital records; however this strategy carries with 
it the potential for recall bias, discussed in further detail below. 
Other examples of different levels of ascertainment of events include: 
(1) use of hospital records to study congenital malformations since 
hospital records contain more complete data on malformations than do 
birth certificates (Mackeprang et al., 1972; Snell et al., 1992) and 
(2) use of sperm bank or fertility clinic data for semen studies. Semen 
data from either source are selected data because semen donors are 
typically of proven fertility, and men in fertility clinics are part of 
a subfertile couple who are actively trying to conceive. Thus, studies 
using the different record sources to identify reproductive outcomes 
need to be evaluated for ascertainment patterns prior to use in risk 
assessment.
    Studies of women who work outside the home present the potential 
for additional bias because some factors that influence employment 
status may also affect reproductive endpoints. For example, because of 
child-care responsibilities, women may terminate employment, as might 
women with a history of reproductive problems who wish to have children 
and are concerned about workplace exposures (Joffe, 1985; Lemasters and 
Pinney, 1989). Thus, retrospective studies of female exposure that do 
not include terminated women workers may be of

[[Page 56301]]

limited use in risk assessment because the level of risk for these 
outcomes is likely to be overestimated (Lemasters and Pinney, 1989).
    Information bias may result from misclassification of 
characteristics of individuals or events identified for study. Recall 
bias, one type of information bias, may occur when respondents with 
specific exposures or outcomes recall information differently than 
those without the exposures or outcomes. Interview bias may result when 
the interviewer knows a priori the category of exposure (for cohort 
studies) or outcome (for case-referent studies) in which the respondent 
belongs. Use of highly structured questionnaires and/or ``blinding'' of 
the interviewer reduces the likelihood of such bias. Studies with lower 
likelihood of such bias should carry more weight in a risk assessment.
    When data are collected by interview or questionnaire, the 
appropriate respondent depends on the type of data or study. For 
example, a comparison of husband-wife interviews on reproduction found 
the wives' responses to questions on pregnancy-related events to be 
more complete and valid than those of the husbands, and the 
individual's self-report of his/her occupational exposures and health 
characteristics more reliable than his/her mate's report (Selevan, 
1980; Selevan et al., 1982). Studies based on interview data from the 
appropriate respondents would carry more weight than those from proxy 
respondents.
    Data from any source may be prone to errors or bias. All types of 
bias are difficult to assess; however, validation with an independent 
data source (e.g., vital or hospital records), or use of biomarkers of 
exposure or outcome, where possible, may suggest the degree of bias 
present and increase confidence in the results of the study. Those 
studies with a low probability of biased data should carry more weight 
(Axelson, 1985; Stein, A. and Hatch, 1987; Weinberg et al., 1994).
    Differential misclassification (i.e., when certain subgroups are 
more likely to have misclassified data than others) may either raise or 
lower the risk estimate. Nondifferential misclassification will bias 
the results toward a finding of ``no effect'' (Rothman, 1986).

Collection of Data on Other Risk Factors, Effect Modifiers, and 
Confounders

    Risk factors for reproductive toxicity include such characteristics 
as age, smoking, alcohol or caffeine consumption, drug use, and past 
reproductive history. Groups of individuals may represent susceptible 
subpopulations based on genetic, acquired (e.g., behavioral), or 
developmental characteristics (e.g., greater effect of childhood 
exposures). Known and potential risk factors should be examined to 
identify those that may be confounders or effect modifiers. An effect 
modifier is a factor that produces different exposure-response 
relationships at different levels of that factor. For example, age 
would be an effect modifier if the risk associated with a given 
exposure changed with age (e.g., if older men had semen changes with 
exposure while younger ones did not). A confounder is a variable that 
is a risk factor for the outcome under study and is associated with the 
exposure under study, but is not a consequence of the exposure. A 
confounder may distort both the magnitude and direction of the measure 
of association between the exposure of interest and the outcome. For 
example, smoking might be a confounder in a study of the association of 
socioeconomic status and fertility because smoking may be associated 
with both.
    Both effect modifiers and confounders need to be controlled in the 
study design and/or analysis to improve the estimate of the effects of 
exposure (Kleinbaum et al., 1982). A more in-depth discussion may be 
found elsewhere (Epidemiology Workgroup for the Interagency Regulatory 
Liaison Group, 1981; Kleinbaum et al., 1982; Rothman, 1986). The 
statistical techniques used to control for these factors require 
careful consideration in their application and interpretation 
(Kleinbaum et al., 1982; Rothman, 1986). Studies that fail to account 
for these important factors should be given less weight in a risk 
assessment.

Statistical Factors

    As in studies of test animals, pregnancies experienced by the same 
woman are not fully independent events. For example, women who have had 
fetal loss are reported to be more likely to have subsequent losses 
(Leridon, 1977). In test animal studies, the litter can be used as the 
unit of measure to deal with nonindependence of response within the 
litter. In studies of humans, pregnancies are sequential, requiring 
analyses which consider nonindependence of events (Epidemiology 
Workgroup for the Interagency Regulatory Liaison Group, 1981; Kissling, 
1981; Selevan, 1981; Zeger and Liang, 1986). If more than one pregnancy 
per woman is included, as is often necessary with small study groups, 
the use of nonindependent observations overestimates the true size of 
the groups being compared, thus artificially increasing the probability 
of reaching statistical significance (Stiratelli et al., 1984). 
Analysis problems may occur when (1) prior adverse outcomes are due to 
the same exposures or (2) when prior adverse outcomes could result in 
changes in behaviors that could reduce exposures. Some approaches to 
deal with these issues have been suggested (Kissling, 1981; Stiratelli 
et al., 1984; Selevan, 1985; Zeger and Liang, 1986). These approaches 
include selecting one pregnancy per family (Selevan, 1985) or using 
generalized estimating equations (Zeger and Liang, 1986).
III.C.2. Examination of Clusters, Case Reports, or Series
    The identification of cases or clusters of adverse reproductive 
effects is generally limited to those identified by the individuals 
involved or clinically by their physicians. The likelihood of 
identification varies with the gender of the exposed person. 
Identification of subfecundity in either gender is difficult. This 
might be thought of as identification of a nonevent (e.g., lack of 
pregnancies or children), and thus is much harder to recognize than are 
some developmental effects, including malformations, resulting from in 
utero exposure.
    The identification of cases or clusters of adverse male 
reproductive outcomes may be limited because of cultural norms that may 
inhibit the reporting of impaired fecundity in men. Identification is 
also limited by the decreased likelihood of recognizing adverse 
developmental effects in their offspring as resulting from paternal 
exposure rather than maternal exposure. Thus far, only one agent 
causing human male reproductive toxicity, dibromochloropropane (DBCP), 
has been identified after observation of a cluster of infertility that 
resulted from male subfecundity. This cluster was identified because of 
an atypically high level of communication among the workers' wives 
(Whorton et al., 1977, 1979; Biava et al., 1978; Whorton and Milby, 
1980).
    Adverse effects identified in females through clusters and case 
reports have, thus far, been limited to adverse pregnancy outcomes such 
as fetal loss and congenital malformations. Identification of other 
effects, such as subfertility/subfecundity or menstrual cycle 
disorders, may be more difficult, as noted above.
    Case reports may have importance in the recognition of agents that 
cause reproductive toxicity. However, they are

[[Page 56302]]

probably of greatest use in suggesting topics for further 
investigation. Reports of clusters and case reports/series are best 
used in risk assessment in conjunction with strong laboratory data to 
suggest that effects observed in test animals also occur in humans.

III.D. Pharmacokinetic Considerations

    Extrapolation of toxicity data between species can be aided 
considerably by the availability of data on the pharmacokinetics of a 
particular agent in the species tested and, when available, in humans. 
Information on absorption, half-life, steady-state or peak plasma 
concentrations, placental metabolism and transfer, comparative 
metabolism, and concentrations of the parent compound and metabolites 
in target organs may be useful in predicting risk for reproductive 
toxicity. Information on the variability between humans and test 
species also may be useful in evaluating factors such as age-related 
differences in the balance between activation and deactivation of a 
toxic agent. These types of data may be helpful in defining the 
sequence of events leading to an adverse effect and the dose-response 
curve, developing a more accurate comparison of species sensitivity, 
including that of humans (Wilson et al., 1975, 1977), determining 
dosimetry at target sites, and comparing pharmacokinetic profiles for 
various dosing regimens or routes of exposure. EPA's Office of 
Prevention, Pesticides, and Toxic Substances has published protocols 
for metabolism studies that may be adapted to provide information 
useful in reproductive toxicity risk assessment for a suspect agent. 
Pharmacokinetic studies in reproductive toxicology are most useful if 
the data are obtained with animals that are at the same reproductive 
status and stage of life (e.g., pregnant, nonpregnant, embryo or fetus, 
neonate, prepubertal, adult) at which reproductive insults are expected 
to occur in humans.
    Specific guidance regarding both the development and application of 
pharmacokinetic data was agreed on by the participants of the Workshop 
on Dermal Developmental Toxicity Studies (Kimmel, C.A. and Francis, 
1990). This guidance is also applicable to nondermal reproductive 
toxicity studies. Participants of the Workshop concluded that 
absorption data are needed both when a dermal study does or does not 
show effects. The results of a dermal study showing no effects and 
without blood level data are potentially misleading and are inadequate 
for risk assessment, especially if interpreted as a ``negative'' study. 
In studies where adverse effects are detected, regardless of the route 
of exposure, pharmacokinetic data can be used to establish the internal 
dose in maternal and paternal animals for risk extrapolation purposes.
    The existence of a Sertoli cell barrier (formerly called the blood-
testis barrier) in the seminiferous tubules may influence the 
pharmacokinetics of an agent with potential to cause testicular 
toxicity by restricting access of compounds to the adluminal 
compartment of seminiferous tubules. The Sertoli cell barrier is formed 
by tight junctions between Sertoli cells and divides the seminiferous 
epithelium into basal and adluminal compartments (Russell et al., 
1990). The basal compartment contains the spermatogonia and primary 
spermatocytes to the preleptotene stage, whereas more advanced germ 
cells are located on the adluminal side. This selectively permeable 
barrier is most effective in limiting the access of large, hydrophilic 
molecules in the intertubular lymph to cells on the adluminal side. An 
analogous barrier in the ovary has not been found, although the zona 
pellucida and granulosa cells may modulate access of chemicals to 
oocytes (Crisp, 1992).
    The reproductive organs appear to have a wide range of metabolic 
capabilities directed at both steroid and xenobiotic metabolism. 
However, there are substantial differences between compartments within 
the organs in types and levels of enzyme activities (Mukhtar et al., 
1978). Recognition of these differences can be important in 
understanding the potential of agents to have specific toxic effects.
    Most pharmacokinetic studies have incompletely characterized the 
distribution of toxic agents and their subsequent metabolic fate within 
the reproductive organs. Generalizations based on hepatic metabolism 
are not necessarily adequate to predict the fate of the agent in the 
testis, ovary, placenta, or conceptus. For example, the metabolic 
profile for a given agent may differ in the male between the liver and 
the testis and in the female between the maternal liver, ovary, and 
placenta. Detailed interspecies comparisons of the metabolic 
capabilities of the testis, ovary, placenta, and conceptus also have 
not been conducted. For some xenobiotics, significant differences in 
metabolism have been identified between males and females (Harris, R.Z. 
et al., 1995). This is, in part, attributable to organizational effects 
of the gonadal steroids in the developing liver (Gustafsson et al., 
1980; Skett, 1988). Also, in adults, the sex steroids have been shown 
to affect the activity of a number of enzymes involved in the 
metabolism of administered compounds. Thus, the blood levels of a toxic 
agent, as well as the final concentration in the target tissue, may 
differ significantly between sexes. If data are to be used effectively 
in interspecies comparisons and extrapolations for these target 
systems, more attention should be directed to the pharmacokinetic 
properties of chemicals in the reproductive organs and in other organs 
that are affected by reproductive hormones.

III.E. Comparisons of Molecular Structure

    Comparisons of the chemical or physical properties of an agent with 
those of agents known to cause reproductive toxicity may provide some 
indication of a potential for reproductive toxicity. Such information 
may be helpful in setting priorities for testing of agents or for 
evaluation of potential toxicity when only minimal data are available. 
Structure-activity relationships (SAR) have not been well studied in 
reproductive toxicology, and have had limited success in predicting 
reproductive toxicity. The early literature has been reviewed and a set 
of classifications offered relating structure to reported male 
reproductive system activity (Bernstein, 1984). Data are available that 
suggest structure-activity relationships with limited utility in risk 
assessment for certain classes of chemicals (e.g., glycol ethers, some 
estrogens, androgens, other steroids, substituted phenols, retinoids, 
phthalate esters, short-chain halogenated hydrocarbon pesticides, 
alkyl-substituted polychlorinated dibenzofurans, PCBs, vinylcyclohexene 
and related olefins, halogenated propanes, metals, and azo dyes). 
McKinney and Waller (1994) have studied the qualitative SAR properties 
of PCBs with respect to their recognition by thyroxine, Ah and estrogen 
receptors. Although generally limited in scope and in need of 
validation, such relationships provide hypotheses that can be tested.
    In spite of the limited information available on SAR in 
reproductive toxicology, under certain circumstances (e.g., in the case 
of new chemicals), this procedure can be used to evaluate the potential 
for toxicity when little or no other data are available.

III.F. Evaluation of Dose-Response Relationships

    The description and evaluation of dose-response relationships is a 
critical component of the hazard characterization. Evidence for a dose-
response relationship is an important

[[Page 56303]]

criterion in establishing a toxic reproductive effect. It includes the 
evaluation of data from both human and laboratory animal studies. When 
possible, pharmacokinetic data should be used to determine the 
effective dose at the target organ(s). When adequate dose-response data 
are available in humans and with a sufficient range of exposure, dose-
response relationships in humans may be examined. Because quantitative 
data on human dose-response relationships are available infrequently, 
the dose-response evaluation is usually based on the assessment of data 
from tests performed in laboratory animals.
    The dose-response relationships for individual endpoints, as well 
as the combination of endpoints, must be examined in data 
interpretation. Dose-response evaluations should consider the effects 
that competing risks between different endpoints may have on outcomes 
observed at different exposure levels. For example, an agent may 
interfere with cell function in such a manner that, at a low dose 
level, an increase in abnormal sperm morphology is observed. At higher 
doses, cell death may occur, leading to a decrease in sperm counts and 
a possible decrease in proportion of abnormal sperm.
    When data on several species are available, the selection of the 
data for the dose-response evaluation is based ideally on the response 
of the species most relevant to humans (e.g., comparable physiologic, 
pharmacologic, pharmacokinetic, and pharmacodynamic processes), the 
adequacy of dosing, the appropriateness of the route of administration, 
and the endpoints selected. However, availability of information on 
many of those components is usually very limited. For dose-response 
assessment, no single laboratory animal species can be considered the 
best in all situations for predicting risk of reproductive toxicity to 
humans. However, in some cases, such as in the assessment of 
physiologic parameters related to menstrual disorders, higher nonhuman 
primates are considered generally similar to the human. In the absence 
of a clearly most relevant species, data from the most sensitive 
species (i.e., the species showing a toxic effect at the lowest 
administered dose) are used, because humans are assumed to be at least 
as sensitive generally as the most sensitive animal species tested 
(Nisbet and Karch, 1983; Kimmel, C.A. et al., 1984, 1990; Hemminki and 
Vineis, 1985; Meistrich, 1986; Working, 1988).
    The evaluation of dose-response relationships includes the 
identification of effective dose levels as well as doses that are 
associated with low or no increased incidence of adverse effects 
compared with controls. Much of the focus is on the identification of 
the critical effect(s) (i.e., the adverse effect occurring at the 
lowest dose level) and the LOAEL and NOAEL or benchmark dose associated 
with the effect(s) (see Section IV).
    Generally, in studies that do not evaluate reproductive toxicity, 
only adult male and nonpregnant females are examined. Therefore, the 
possibility that pregnant females may be more sensitive to the agent is 
not tested. In studies in which reproductive toxicity has been 
evaluated, the effective dose range should be identified for both 
reproductive and other forms of systemic toxicity, and should be 
compared with the corresponding values from other adult toxicity data 
to determine if the pregnant or lactating female may be more sensitive 
to an agent.
    In addition to identification of the range of doses that is 
effective in producing reproductive and other forms of systemic 
toxicity for a given agent, the route of exposure, timing and duration 
of exposure, species specificity of effects, and any pharmacokinetic or 
other considerations that might influence the comparison with human 
exposure scenarios should be identified and evaluated. This information 
should always accompany the characterization of the health-related 
database (discussed in the next section).
    Because the developing organism is changing rapidly and is 
vulnerable at a number of stages, an assumption is made with 
developmental effects that a single exposure at a critical time in 
development may produce an adverse effect (U.S. EPA, 1991). Therefore, 
with inhalation exposures, the daily dose is usually not adjusted to a 
24-hour equivalent duration with developmental toxicity unless 
appropriate pharmacokinetic data are available. However, for other 
reproductive effects, daily doses by the inhalation route may be 
adjusted for duration of exposure. The Agency is planning to review 
these stances to determine the most appropriate approach for the 
future.

III.G. Characterization of the Health-Related Database

    This section describes evaluation of the health-related database on 
a particular chemical and provides criteria for judging the potential 
for that chemical to produce reproductive toxicity under the exposure 
conditions inherent in the database. This determination provides the 
basis for judging whether the available data are sufficient to 
characterize a hazard and to conduct quantitative dose-response 
analyses. It also should provide a summary and evaluation of the 
existing data and identify data gaps for an agent that is judged to 
have insufficient information to proceed with a quantitative dose-
response analysis. Characterizing the available evidence in this way 
clarifies the strengths and uncertainties in a particular database. It 
does not address the level of concern, nor does it completely address 
determining relevance of available data for estimating human risk. 
Issues concerning relevance of mechanisms of action and types of 
effects observed should be included in the hazard characterization. 
Both level of concern and relevance are discussed further as part of 
the final characterization of risk, taking into account the information 
concerning potential human exposure. Data from all potentially relevant 
studies, whether indicative of potential hazard or not, should be 
included in the hazard characterization.
    A complex interrelationship exists among study design, statistical 
analysis, and biologic significance of the data. Thus, substantial 
scientific judgment, based on experience with reproductive toxicity 
data and with the principles of study design and statistical analysis, 
may be required to evaluate the database adequately. In some cases, a 
database may contain conflicting data. In these instances, the risk 
assessor must consider each study's strengths and weaknesses within the 
context of the overall database to characterize the evidence for 
assessing the potential hazard for reproductive toxicity. Scientific 
judgment is always necessary and, in many cases, interaction with 
scientists in specific disciplines (e.g., reproductive toxicology, 
epidemiology, genetic toxicology, statistics) is recommended.
    A scheme for judging the available evidence on the reproductive 
toxicity of a particular agent is presented below (Table 6). The scheme 
contains two broad categories, ``Sufficient'' and ``Insufficient,'' 
which are defined in Table 6. Data from all available studies, whether 
or not indicative of potential concern, are evaluated and used in the 
hazard characterization for reproductive toxicity. The primary 
considerations are the human data, if available, and the experimental 
animal data. The judgment of whether data are sufficient or 
insufficient should consider a variety of parameters that contribute to 
the overall quality of the data, such as the power of the studies 
(e.g., sample size and variation in the data), the number and types of 
endpoints examined,

[[Page 56304]]

replication of effects, relevance of route and timing of exposure for 
both human and experimental animal studies, and the appropriateness of 
the test species and dose selection in experimental animal studies. In 
addition, pharmacokinetic data and structure-activity considerations, 
data from other toxicity studies, as well as other factors that may 
affect the overall decision about the evidence, should be taken into 
account.

         Table 6.--Categorization of the Health-Related Database        
------------------------------------------------------------------------
                                                                        
-------------------------------------------------------------------------
                           Sufficient Evidence                          
                                                                        
  The Sufficient Evidence category includes data that collectively      
 provide enough information to judge whether or not a reproductive      
 hazard exists within the context of effect as well as dose, duration,  
 timing, and route of exposure. This category may include both human and
 experimental animal evidence.                                          
                                                                        
                        Sufficient Human Evidence                       
                                                                        
  This category includes agents for which there is convincing evidence  
 from epidemiologic studies (e.g., case control and cohort) to judge    
 whether exposure is causally related to reproductive toxicity. A case  
 series in conjunction with other supporting evidence also may be judged
 as Sufficient Evidence. An evaluation of epidemiologic and clinical    
 case studies should discuss whether the observed effects can be        
 considered biologically plausible in relation to chemical exposure.    
                                                                        
       Sufficient Experimental Animal Evidence/Limited Human Data       
                                                                        
  This category includes agents for which there is sufficient evidence  
 from experimental animal studies and/or limited human data to judge if 
 a potential reproductive hazard exists. Generally, agents that have    
 been tested according to EPA's two-generation reproductive effects test
 guidelines (but not limited to such designs) would be included in this 
 category. The minimum evidence necessary to determine if a potential   
 hazard exists would be data demonstrating an adverse reproductive      
 effect in a single appropriate, well-executed study in a single test   
 species. The minimum evidence needed to determine that a potential     
 hazard does not exist would include data on an adequate array of       
 endpoints from more than one study with two species that showed no     
 adverse reproductive effects at doses that were minimally toxic in     
 terms of inducing an adverse effect. Information on pharmacokinetics,  
 mechanisms, or known properties of the chemical class may also         
 strengthen the evidence.                                               
                                                                        
                          Insufficient Evidence                         
                                                                        
  This category includes agents for which there is less than the minimum
 sufficient evidence necessary for assessing the potential for          
 reproductive toxicity. Included are situations such as when no data are
 available on reproductive toxicity; as well as for data bases from     
 studies on test animals or humans that have a limited study design or  
 conduct (e.g., small numbers of test animals or human subjects,        
 inappropriate dose selection or exposure information, other            
 uncontrolled factors); data from studies that examined only a limited  
 number of endpoints and reported no adverse reproductive effects; or   
 data bases that were limited to information on structure-activity      
 relationships, short-term or in vitro tests, pharmacokinetic data, or  
 metabolic precursors.                                                  
------------------------------------------------------------------------

    In general, the characterization is based on criteria defined by 
these Guidelines as the minimum evidence necessary to characterize a 
hazard and conduct dose-response analyses. Establishing the minimum 
human evidence to proceed with quantitative analyses based on the human 
data is often difficult because there may be considerable variations in 
study designs and study group selection. The body of human data should 
contain convincing evidence as described in the ``Sufficient Human 
Evidence'' category. Because the human data necessary to judge whether 
or not a causal relationship exists are generally limited, few agents 
can be classified in this category. Agents that have been tested in 
laboratory animals according to EPA's two-generation reproductive 
effects test guidelines (U.S. EPA, 1982, 1985b, 1996a), but not limited 
to such designs (e.g., a continuous breeding study with two 
generations), generally would be included in the ``Sufficient 
Experimental Animal Evidence/Limited Human Data'' category. There are 
occasions in which more limited data regarding the potential 
reproductive toxicity of an agent (e.g., a one-generation reproductive 
effects study, a standard subchronic or chronic toxicity study in which 
the reproductive organs were well examined) are available. If 
reproductive toxicity is observed in these limited studies, the data 
may be used to the extent possible to reach a decision regarding hazard 
to the reproductive system, including determination of an RfD or RfC. 
In cases in which such limited data are available, it would be 
appropriate to adjust the uncertainty factor to reflect the attendant 
increased uncertainty regarding the use of these data until more 
definitive data are developed. Identification of the increased 
uncertainty and justification for the adjustment of the uncertainty 
factor should be stated clearly.
    Because it is more difficult, both biologically and statistically, 
to support a finding of no apparent hazard, more data are generally 
required to support this conclusion than a finding for a potential 
hazard. For example, to judge whether a hazard for reproductive 
toxicity could exist for a given agent, the minimum evidence could be 
data from a single appropriate, well-executed study in a single test 
species that demonstrates an adverse reproductive effect, or suggestive 
evidence from adequately conducted clinical or epidemiologic studies. 
As in all situations, it is important that the results be biologically 
plausible and consistent. On the other hand, to judge whether an agent 
is unlikely to pose a hazard for reproductive toxicity, the minimum 
evidence would include data on an array of endpoints and from studies 
with more than one species that showed no reproductive effects at doses 
that were otherwise minimally toxic to the adult animal. In addition, 
there may be human data from appropriate studies that are supportive of 
no apparent hazard. In the event that a substantial database exists for 
a given chemical, but no single study meets current test guidelines, 
the risk assessor should use scientific judgment to determine whether 
the composite database may be viewed as meeting the ``Sufficient'' 
criteria.
    Some important considerations in determining the confidence in the 
health database are as follows:
     Data of equivalent quality from human exposures are given 
more weight than data from exposures of test species.
     Although a single study of high quality could be 
sufficient to achieve a relatively high level of confidence, 
replication increases the confidence that may be placed in such 
results.
     Data are available from one or more in vivo studies of 
acceptable quality with humans or other mammalian species that are 
believed to be predictive of human responses.
     Data exhibit a dose-response relationship.
     Results are statistically significant and biologically 
plausible.
     When multiple studies are available, the results are 
consistent.
     Sufficient information is available to reconcile 
discordant data.
     Route, level, duration, and frequency of exposure are 
appropriate.
     An adequate array of endpoints has been examined.
     The power and statistical treatment of the studies are 
appropriate.
    Any statistically significant deviation from baseline levels for an 
in vivo effect warrants closer examination. To determine whether such a 
deviation constitutes an adverse effect requires an understanding of 
its role within a complex system and the determination of whether a 
``true effect'' has beenobserved. Application of the above criteria, 
combined with guidance presented in Section III.B. can facilitate such 
determinations.
    The greatest confidence for identification of a reproductive hazard 
should be placed on significant adverse effects on sexual behavior, 
fertility or development, or other endpoints that are directly related 
to reproductive function such as menstrual (estrous) cycle normality, 
sperm evaluations, reproductive histopathology, reproductive organ 
weights, and reproductive endocrinology. Agents producing adverse 
effects on these endpoints can be assigned to the ``Sufficient 
Evidence'' category if study quality is adequate.

[[Page 56305]]

    Less confidence should be placed in results when only measures such 
as in vitro tests, data from nonmammals, or structure-activity 
relationships are available, but positive results may trigger follow-up 
studies that extend the preliminary data and determine the extent to 
which function might be affected. Results from these types of studies 
alone, whether or not they demonstrate an effect, may be suggestive, 
but should be assigned to the ``Insufficient Evidence'' category.
    The absence of effects in test species on the endpoints that are 
evaluated routinely (i.e., fertility, histopathology, prenatal 
development, and organ weights) may constitute sufficient evidence to 
place a low priority on the potential reproductive toxicity of a 
chemical. However, in such cases, careful consideration should be given 
to the sensitivity of these endpoints and to the quality of the data on 
these endpoints. Consideration also should be given to the possibility 
of adverse effects that may not be reflected in these routine measures 
(e.g., germ-cell mutation, alterations in estrous cyclicity or sperm 
measures such as motility or morphology, functional effects from 
developmental exposures).
    Judging that the health database indicates a potential reproductive 
hazard does not mean that the agent will be a hazard at every exposure 
level (because of the assumption of a nonlinear dose-response) or in 
every situation (e.g., the type and degree of hazard may vary 
significantly depending on route and timing of exposure). In the final 
risk characterization, the summary of the hazard characterization 
should always be presented with information on the quantitative dose-
response analysis and, if available, with the human exposure estimates.

IV. Quantitative Dose-Response Analysis

    In quantitative dose-response assessment, a nonlinear dose-response 
is assumed for noncancer health effects unless mode of action or 
pharmacodynamic information indicate otherwise. If sufficient data are 
available, a biologically based approach should be used on a chemical-
specific basis to predict the shape of the dose-response curve below 
the observable range. It is plausible that certain biologic processes 
(e.g., Sertoli cell barrier selectivity, metabolic and repair 
capabilities of the germ cells) may impede the attainment or 
maintenance of concentrations of the agent at the target site following 
exposure to low-dose levels that would be associated with adverse 
effects. The assumption of a nonlinear dose-response suggests that the 
application of adequate uncertainty factors to a NOAEL, LOAEL, or 
benchmark dose will result in an exposure level for all humans that is 
not attended with significant risk above background. With a linear 
dose-response, it is assumed that some risk exists at any level of 
exposure, with risk decreasing as exposure decreases.
    The NOAEL is the highest dose at which there is no significant 
increase in the frequency of an adverse effect in any manifestation of 
reproductive toxicity compared with the appropriate control group in a 
database having sufficient evidence for use in a risk assessment. The 
LOAEL is the lowest dose at which there is a significant increase in 
the frequency of adverse reproductive effects compared with the 
appropriate control group in a database having sufficient evidence. A 
significant increase may be based on statistical significance or on a 
biologically significant trend. Evidence for biological significance 
may be strengthened by mode of action or other biochemical evidence at 
lower exposure levels that supports the causation of such an effect. 
The existence of a NOAEL in an experimental animal study does not show 
the shape of the dose-response below the observable range; it only 
defines the highest level of exposure under the conditions of the study 
that is not associated with a significant increase in an adverse 
effect. Alternatively, mathematical modeling of the dose-response 
relationship may be performed in the experimental range. This approach 
can be used to determine a benchmark dose, which may be used in place 
of the NOAEL as a point of departure for calculating an RfD, RfC, MOE, 
or other exposure estimates.
    Several limitations in the use of the NOAEL have been described 
(Kimmel, C.A. and Gaylor, 1988; U.S. EPA, 1995b): (1) Use of the NOAEL 
focuses only on the dose that is the NOAEL and does not incorporate 
information on the slope of the dose-response curve or the variability 
in the data; (2) Because data variability is not taken into account 
(i.e., confidence limits are not used), the NOAEL will likely be higher 
with decreasing sample size or poor study conduct, either of which are 
usually associated with increasing variability in the data; (3) The 
NOAEL is limited to one of the experimental doses; (4) The number and 
spacing of doses in a study can influence the dose that is chosen for 
the NOAEL; and (5) Because the NOAEL is defined as a dose that does not 
produce an observable change in adverse responses from control levels 
and is dependent on the power of the study, theoretically the risk 
associated with it may fall anywhere between zero and an incidence just 
below that detectable from control levels (usually in the range of 7% 
to 10% for quantal data). The 95% upper confidence limit on 
developmental toxicity risk at the NOAEL has been estimated for several 
data sets to be 2% to 6% (Crump, 1984; Gaylor, 1989); similar 
evaluations have not been conducted on data for other reproductive 
effects. Because of the limitations associated with the use of the 
NOAEL, the Agency is beginning to use the benchmark dose approach for 
quantitative dose-response evaluation when sufficient data are 
available.
    Calculation and use of the benchmark dose are described in the EPA 
document The Use of the Benchmark Dose Approach in Health Risk 
Assessment (U.S. EPA, 1995b). The Agency is currently developing 
guidance for application of the benchmark dose, including a decision 
process to use for the various steps in the analysis (U.S. EPA, 1996c). 
The benchmark dose is based on a model-derived estimate of a particular 
incidence level, such as a 5% or 10% incidence. The BMD/C for a given 
endpoint serves as a consistent point of departure for low-dose 
extrapolation. In some cases, mode of action data may be sufficient to 
estimate a BMD/C at levels below the observable range for the health 
effect of concern. A benchmark response (BMR) of 5% is usually the 
lowest level of risk that can be estimated adequately for binomial 
endpoints from standard developmental toxicity studies (Allen et al., 
1994a, b). For fetal weight, a continuous endpoint, a 5% change from 
the control mean was near the limit of detection for standard prenatal 
toxicity studies (Kavlock et al., 1995). The modeling approaches that 
have been proposed for developmental toxicity (U.S. EPA, 1995b) are, 
for the most part, curve-fitting models that have biological 
plausibility, but do not incorporate mode of action. Similar approaches 
can be applied to other reproductive toxicity data to derive dose-
response curves for data in the observed dose range, but may or may not 
accurately predict risk at low levels of exposure. Further guidance on 
the use of the BMD/C is being developed by the Agency (U.S. EPA, 
1996c).
    The RfD or RfC for reproductive toxicity is an estimate of a daily 
exposure to the human population that is assumed to be without 
appreciable risk of deleterious reproductive effects over a lifetime of 
exposure. The RfD or RfC is derived by applying uncertainty factors to 
the NOAEL, or the LOAEL if a NOAEL is not available, or to the

[[Page 56306]]

benchmark dose. Because of the short duration of most studies of 
developmental toxicity, a unique value (RfDDT or RfCDT) is 
determined for adverse developmental effects. For adverse reproductive 
effects on endpoints other than those of developmental toxicity, no 
special designator is attached. Data on reproductive toxicity 
(including developmental toxicity) are considered along with other data 
on a particular chemical in deriving an RfD or RfC.
    The effect used for determining the NOAEL, LOAEL, or benchmark dose 
in deriving the RfD or RfC is the most sensitive adverse reproductive 
endpoint (i.e., the critical effect) from the most appropriate or, in 
the absence of such information, the most sensitive mammalian species 
(see Sections II and III.B.1.). Uncertainty factors for reproductive 
and other forms of systemic toxicity applied to the NOAEL or benchmark 
dose generally include factors of 3 or 10 each for interspecies 
variation and for intraspecies variation. Additional factors may be 
applied to account for other uncertainties that may exist in the 
database. In circumstances where only a LOAEL is available, the use of 
an additional uncertainty factor of up to 10 may be required, depending 
on the sensitivity of the endpoints evaluated, adequacy of dose levels 
tested, or general confidence in the LOAEL.
    Other areas of uncertainty may be identified and modifying factors 
used depending on the characterization of the database (e.g., if the 
only data available are from a one-generation reproductive effects 
study; see Section III.G.), data on pharmacokinetics, or other 
considerations that may alter the level of confidence in the data (U.S. 
EPA, 1987). The total size of the uncertainty factor will vary from 
agent to agent and requires scientific judgment, taking into account 
interspecies differences, variability within species, the slope of the 
dose-response curve, the types of reproductive effects observed, the 
background incidence of the effects, the route of administration, and 
pharmacokinetic data.
    The NOAEL, LOAEL, or the benchmark dose is divided by the total 
uncertainty factor selected for the critical effect in the most 
appropriate or most sensitive mammalian species to determine the RfD or 
RfC. If the NOAEL, LOAEL, or benchmark dose for other forms of systemic 
toxicity is lower than that for reproductive toxicity, this should be 
noted in the risk characterization, and this value should be compared 
with data from other studies in which adult animals are exposed. Thus, 
reproductive toxicity data should be discussed in the context of other 
toxicity data.
    It has generally been assumed that there is a nonlinear dose-
response for reproductive toxicity. This is based on known homeostatic, 
compensatory, or adaptive mechanisms that must be overcome before a 
toxic endpoint is manifested and on the rationale that cells and organs 
of the reproductive system and the developing organism are known to 
have some capacity for repair of damage. However, in a population, 
background levels of toxic agents and preexisting conditions may 
increase the sensitivity of some individuals in the population. Thus, 
exposure to a toxic agent may result in an increased risk of adverse 
effects for some, but not necessarily all, individuals within the 
population.
    Efforts are underway to develop models that are more biologically 
based. These models should provide a more accurate estimation of low-
dose risk to humans. The development of biologically based dose-
response models in reproductive toxicology has been impeded by a number 
of factors, including limited understanding of the biologic mechanisms 
underlying reproductive toxicity, intra- and interspecies differences 
in the types of reproductive events, lack of appropriate 
pharmacokinetic data, and inadequate information on the influence of 
other types of systemic toxicity on the dose-response curve. Current 
research on modes of action in reproductive toxicology is promising and 
may provide data that are useful for appropriate modeling in the near 
future.

Utilization of Information in Risk Characterization

    The hazard characterization and quantitative dose-response 
evaluations are incorporated into the final characterization of risk 
along with information on estimates of human exposure. The analysis 
depends on and should describe scientific judgments as to the accuracy 
and sufficiency of the health-related data in experimental animals and 
humans (if available), the biologic relevance of significant effects, 
and other considerations important in the interpretation and 
application of data to humans. Scientific judgment is always necessary, 
and in many cases, interaction with scientists in specific disciplines 
(e.g., reproductive toxicology, epidemiology, statistics) is 
recommended.

V. Exposure Assessment

    To obtain a quantitative estimate of risk for the human population, 
an estimate of human exposure is required. The Guidelines for Exposure 
Assessment (U.S. EPA, 1992) have been published separately and will not 
be discussed in detail here. Rather, issues important to reproductive 
toxicity risk assessment are addressed. In general, the exposure 
assessment describes the magnitude, duration, schedule, and route of 
human exposure. Ideally, existing body burden as well as internal 
circulating and target organ exposure information for the agent of 
concern and other synergistic or antagonistic agents should be 
described. It should include information on the purpose, scope, level 
of detail and approach used, including estimates of exposure and dose 
by pathway and route for populations, subpopulations, and individuals 
in a manner that is appropriate for the intended risk characterization. 
It also should provide an evaluation of the overall level of confidence 
in the estimate(s) of exposure and dose and the conclusions drawn. This 
information is usually developed from monitoring data, from estimates 
based on modeling of environmental exposures, and from application of 
paradigms to exposure data bases. Often quantitative estimates of 
exposures may not be available (e.g., workplace or environmental 
measurements). In such instances, employment or residential histories 
also may be used in characterizing exposure in a qualitative sense. The 
potential use of biomarkers as indicators of exposure is an area of 
active interest.
    Studies of occupational populations may provide valuable 
information on the potential environmental health risks for certain 
agents. Exposures among environmentally exposed human populations tend 
to be lower (but of longer duration) than those in studies of 
occupationally exposed populations and therefore may require more 
observations to assure sufficient statistical power. Also, 
reconstruction of exposures is more difficult in an environmental study 
than in those done in workplace settings where industrial hygiene 
monitoring may provide more detailed exposure data.
    The nature of the exposure may be defined at a particular point in 
time or may reflect cumulative exposure. Each approach makes an 
assumption about the underlying relationship between exposure and 
outcome. For example, a cumulative exposure measure assumes that total 
exposure is important, with a greater probability of effect with 
greater total exposure or body burden. A dichotomous exposure measure 
(ever exposed versus never exposed) assumes an irreversible effect of 
exposure. Models that define exposure only at a

[[Page 56307]]

specific time may assume that only the present exposure is important 
(Selevan and Lemasters, 1987). The appropriate exposure model depends 
on the biologic processes affected and the nature of the chemical under 
study. Thus, a cumulative or dichotomous exposure model may be 
appropriate if injury occurs in cells that cannot be replaced or 
repaired (e.g., oocytes); on the other hand, a concurrent exposure 
model may be appropriate for cells that are being generated continually 
(e.g., spermatids).
    There are a number of unique considerations regarding the exposure 
assessment for reproductive toxicity. Exposure at different stages of 
male and female development can result in different outcomes. Such age-
dependent variation has been well documented in both experimental 
animal and human studies. Prenatal and neonatal treatment can 
irreversibly alter reproductive function and other aspects of 
development in a manner or to an extent that may not be predicted from 
adult-only exposure. Moreover, chemicals that alter sexual 
differentiation in rodents during these periods may have similar 
effects in humans, because the mechanisms underlying these 
developmental processes appear to be similar in all mammalian species 
(Gray, 1991).
    The susceptibility of elderly males and females to chemical insult 
has not been well studied. Although procreative competence may not be a 
major health concern with elderly individuals, other biologic functions 
maintained by the gonads (e.g., hormone production) are of significance 
(Walker, 1986). An exposure assessment should characterize the 
likelihood of exposure of these different subgroups (embryo or fetus, 
neonate, juvenile, young adult, older adult) and the risk assessment 
should factor in the susceptibility of different age groups to the 
extent possible.
    The relationship between time or duration of exposure and 
observation of male reproductive effects has particular significance 
for short-term exposures. Spermatogenesis is a temporally synchronized 
process. In humans, germ cells that were spermatozoa, spermatids, 
spermatocytes, or spermatogonia at the time of an acute exposure 
require 1 to 2, 3 to 5, 5 to 8, or 8 to 12 weeks, respectively, to 
appear in an ejaculate. That timing may vary somewhat depending on 
degree of sexual activity. It is possible that an endpoint may be 
examined too early or too late to detect an effect if only a particular 
cell type was affected during a relatively brief exposure to an agent. 
The absence of an effect when observations were made too late suggests 
either a reversible effect or no effect. However, an effect that is 
reversible at lower exposures might become irreversible with higher or 
longer exposures or exposure of a more susceptible individual. Thus, 
the failure to detect transient effects because of improper timing of 
observations may be important. If information is available on the type 
of effect expected from a class of agents, it may be possible to 
evaluate whether the timing of endpoint measurement relative to the 
timing of the short-term exposure is appropriate. Some information on 
the appropriateness of the protocol can be obtained if test animal data 
are available to identify the most sensitive cell type or the putative 
mechanism of action for a given agent.
    Compared with acute exposures, the link between exposure and 
outcome may be more apparent with relatively constant subchronic or 
longer exposures that are of sufficient duration to cover all phases of 
spermatogenesis (Russell et al., 1990). Assessments may be made at any 
time after this point as long as exposure remains constant. Time 
required for the agent or metabolite to attain steady-state levels 
should also be considered. Again, application of models of exposure 
(e.g., dichotomous, concurrent, or cumulative) depends on the suspected 
target and chemical mechanism of action.
    The reversibility of an adverse effect on the reproductive system 
can be affected by the degree and duration of exposure (Clegg, 1995). 
The degree of stem cell loss is inversely related to the degree of 
restoration of sperm production, because repopulation of the germinal 
epithelium is dependent on the stem cells (Meistrich, 1982; Foote and 
Berndtson, 1992). For agents that bioaccumulate, increasing duration of 
exposure may also increase the extent of damage to the stem cell 
population. Damage to other spermatogenic cell types reduces the number 
of sperm produced, but recovery should occur when the toxic agent is 
removed. Less is known about the effects of toxicity on the Sertoli 
cells. Temporary impairment of Sertoli cell function may produce long-
lasting effects on spermatogenesis. Destruction of Sertoli cells or 
interference with their proliferation before puberty are irreversible 
effects because replication ceases after puberty. Sertoli cells are 
essential for support of the spermatogenic process and loss of those 
cells results in a permanent reduction of spermatogenic capability 
(Foster, 1992).
    When recovery is possible, the duration of the recovery period is 
determined by the time for regeneration (for stem cells) and 
repopulation of the affected spermatogenic cell types and appearance of 
those cells as sperm in the ejaculate. The time required for these 
events to occur varies with the species, the pharmacokinetic properties 
of the agent, the extent to which the stem cell population has been 
destroyed, and the degree of sublethal toxicity inflicted on the stem 
cells or Sertoli cells. When the stem cell population has been 
partially destroyed, humans require more time than mice to reach the 
same degree of recovery (Meistrich and Samuels, 1985).
    Unique considerations in the assessment of female reproductive 
toxicity include the duration and period of exposure as related to the 
development or stage of reproductive life (e.g., prenatal, 
prepubescent, reproductive, or postmenopausal) or considerations of 
different physiologic states (e.g., nonpregnant, pregnant, lactating). 
For infertility, a cumulative exposure measure assumes destruction of 
increasing numbers of primary oocytes with greater lifetime exposure or 
increasing body burden. However, humans may be exposed to varying 
levels of an agent within the study period. Exposures during certain 
critical points in the reproductive process may affect the outcomes 
observed in humans (Lemasters and Selevan, 1984). In test species, 
perinatal exposure to androgens or estrogens such as zearalenone, 
methoxychlor, and DDT (Bulger and Kupfer, 1985; Gray et al., 1985) have 
been shown to advance puberty and masculinize females. Similar effects 
have been reported in humans (both sexes) exposed neonatally to 
synthetic estrogens or progestins (Steinberger and Lloyd, 1985; 
Schardein, 1993). Studies using test species also have shown that 
exposure to some environmental agents such as ionizing radiation 
(Dobson and Felton, 1983) and glycol ethers (Heindel et al., 1989) can 
deplete the pool of primordial follicles and thus significantly shorten 
the female's reproductive lifespan. Furthermore, exposure to compounds 
at different stages of the ovarian cycle can disrupt or delay 
follicular recruitment and development (Armstrong, 1986), ovulation 
(Everett and Sawyer, 1950; Terranova, 1980), and ovum transport 
(Cummings and Perreault, 1990). Compounds that delay ovulation can lead 
to significant alterations in egg viability (Peluso et al., 1979), 
fertilizability of the egg (Fugo and Butcher, 1966; Butcher and Fugo, 
1967; Butcher et al., 1975), and a reduction in litter size (Fugo and 
Butcher, 1966). After ovulation, single exposures to

[[Page 56308]]

microtubule poisons such as carbendazim may impair the completion of 
meiosis in the fertilized oocyte with adverse developmental 
consequences (Perreault et al., 1992; Zuelke and Perreault, 1995). 
Thus, knowledge of when acute exposures occur relative to the female's 
lifespan and reproductive cycle can provide insight into how an agent 
disrupts reproductive function.
    DES is a classic example of an agent causing different effects on 
the reproductive system in the developing organism compared with those 
in adults (McLachlan, 1980). DES, as well as other agents with 
estrogenic or anti-androgenic activity, interferes with the development 
of the Mullerian and Wolffian duct systems and thereby causes 
irreversible structural and functional damage to the developing 
reproductive system. In adults, the reproductive effects that are 
caused by the estrogenic activity of DES do not necessarily result in 
permanent damage.
    Unique considerations for developmental effects are duration and 
period of exposure as related to stage of development (i.e., critical 
periods) and the possibility that even a single exposure may be 
sufficient to produce adverse developmental effects. Repeated exposure 
is not a necessary prerequisite for developmental toxicity to be 
manifested, although it should be considered in cases where there is 
evidence of cumulative exposure or where the half-life of the agent is 
long enough to produce an increasing body burden over time. For these 
reasons, it is assumed that, in most cases, a single exposure at the 
critical time in development is sufficient to produce an adverse 
developmental effect. Therefore, the human exposure estimates used to 
calculate the MOE for an adverse developmental effect or to compare to 
the RfD or RfC are usually based on a single daily dose that is not 
adjusted for duration or pattern (e.g., continuous or intermittent) of 
exposure. For example, it would be inappropriate to use time-weighted 
averages or adjustment of exposure over a different time frame than 
that actually encountered (such as the adjustment of a 6-hour 
inhalation exposure to account for a 24-hour exposure scenario) unless 
pharmacokinetic data were available to indicate an accumulation with 
continuous exposure. In the case of intermittent exposures, examination 
of the peak exposures as well as the average exposure over the time of 
exposure would be important.
    It should be recognized that, based on the definitions used in 
these Guidelines, almost any segment of the human population may be at 
risk for a reproductive effect. Although the reproductive effects of 
exposures may be manifested while the exposure is occurring (e.g., 
menstrual disorder, decreased sperm count, spontaneous abortion) some 
effects may not be detectable until later in life (e.g., endocrine 
disruption of reproductive tract development, premature reproductive 
senescence due to oocyte depletion), long after exposure has ceased.

VI. Risk Characterization

VI.A. Overview

    A risk characterization is an essential part of any Agency report 
on risk whether the report is a preliminary one prepared to support 
allocation of resources toward further study, a site-specific 
assessment, or a comprehensive one prepared to support regulatory 
decisions. A risk characterization should be prepared in a manner that 
is clear, reasonable, and consistent with other risk characterizations 
of similar scope prepared across programs in the Agency. It should 
identify and discuss all the major issues associated with determining 
the nature and extent of the risk and provide commentary on any 
constraints limiting more complete exposition. The key aspects of risk 
characterization are: (1) bridging risk assessment and risk management, 
(2) discussing confidence and uncertainties, and (3) presenting several 
types of risk information. In this final step of a risk assessment, the 
risk characterization involves integration of toxicity information from 
the hazard characterization and quantitative dose-response analysis 
with the human exposure estimates and provides an evaluation of the 
overall quality of the assessment, describes risk in terms of the 
nature and extent of harm, and communicates results of the risk 
assessment to a risk manager. A risk manager can then use the risk 
assessment, along with other risk management elements, to make public 
health decisions. The information should also assist others outside the 
Agency in understanding the scientific basis for regulatory decisions.
    Risk characterization is intended to summarize key aspects of the 
following components of the risk assessment:
     The nature, reliability, and consistency of the data used.
     The reasons for selection of the key study(ies) and the 
critical effect(s) and their relevance to human outcomes.
     The qualitative and quantitative descriptors of the 
results of the risk assessment.
     The limitations of the available data, the assumptions 
used to bridge knowledge gaps in working with those data, and 
implications of using alternative assumptions.
     The strengths and weaknesses of the risk assessment and 
the level of scientific confidence in the assessment.
     The areas of uncertainty, additional data/research needs 
to improve confidence in the risk assessment, and the potential impacts 
of the new research.
    The risk characterization should be limited to the most significant 
and relevant data, conclusions, and uncertainties. When special 
circumstances exist that preclude full assessment, those circumstances 
should be explained and the related limitations identified.
    The following sections describe these aspects of the risk 
characterization in more detail, but do not attempt to provide a full 
discussion of risk characterization. Rather, these Guidelines point out 
issues that are important to risk characterization for reproductive 
toxicity. Comprehensive general guidance for risk characterization is 
provided by Habicht (1992) and Browner (1995).

VI.B. Integration of Hazard Characterization, Quantitative Dose-
Response, and Exposure Assessments

    In developing each component of the risk assessment, risk assessors 
must make judgments concerning human relevance of the toxicity data, 
including the appropriateness of the various test animal models for 
which data are available, and the route, timing, and duration of 
exposure relative to the expected human exposure. These judgments 
should be summarized at each stage of the risk assessment process. When 
data are not available to make such judgments, as is often the case, 
the background information and assumptions discussed in the Overview 
(Section I) provide default positions. The default positions used and 
the rationale behind the use of each default position should be clearly 
stated. In integrating the parts of the assessment, risk assessors must 
determine if some of these judgments have implications for other 
portions of the assessment, and whether the various components of the 
assessment are compatible.
    The description of the relevant data should convey the major 
strengths and weaknesses of the assessment that arise from availability 
and quality of data and the current limits of understanding of the 
mechanisms of toxicity. Confidence in the results of a risk assessment 
is a function of confidence in the results of

[[Page 56309]]

these analyses. Each section (hazard characterization, quantitative 
dose-response analysis, and exposure assessment) should have its own 
summary, and these summaries should be integrated into the overall risk 
characterization. Interpretation of data should be explained, and risk 
managers should be given a clear picture of consensus or lack of 
consensus that exists about significant aspects of the assessment. When 
more than one interpretation is supported by the data, the alternative 
plausible approaches should be presented along with the strengths, 
weaknesses, and impacts of those options. If one interpretation or 
option has been selected over another, the rationale should be given; 
if not, then both should be presented as plausible alternatives.
    The risk characterization should not only examine the judgments, 
but also should explain the constraints of available data and the state 
of knowledge about the phenomena studied in making them, including:
     The qualitative conclusions about the likelihood that the 
chemical may pose a specific hazard to human health, the nature of the 
observed effects, under what conditions (route, dose levels, time, and 
duration) of exposure these effects occur, and whether the health-
related data are sufficient and relevant to use in a risk assessment.
     A discussion of the dose-response patterns for the 
critical effect(s) and their relationships to the occurrence of other 
toxicity data, such as the shapes and slopes of the dose-response 
curves for the various other endpoints; the rationale behind the 
determination of the NOAEL, LOAEL, and/or benchmark dose; and the 
assumptions underlying the estimation of the RfD, RfC, or other 
exposure estimate.
     Descriptions of the estimates of the range of human 
exposure (e.g., central tendency, high end), the route, duration, and 
pattern of the exposure, relevant pharmacokinetics, and the size and 
characteristics of the various populations that might be exposed.
     The risk characterization of an agent being assessed for 
reproductive toxicity should be based on data from the most appropriate 
species or, if such information is not available, on the most sensitive 
species tested. It also should be based on the most sensitive indicator 
of an adverse reproductive effect, whether in the male, the female 
(nonpregnant or pregnant), or the developing organism, and should be 
considered in relation to other forms of toxicity. The relevance of 
this indicator to human reproductive outcomes should be described. The 
rationale for those decisions should be presented.
    If data to be used in a risk characterization are from a route of 
exposure other than the expected human exposure, then pharmacokinetic 
data should be used, if available, to extrapolate across routes of 
exposure. If such data are not available, the Agency makes certain 
assumptions concerning the amount of absorption likely or the 
applicability of the data from one route to another (U.S. EPA, 1985a, 
1986b). Discussion of some of these issues may be found in the 
Proceedings of the Workshop on Acceptability and Interpretation of 
Dermal Developmental Toxicity Studies (Kimmel, C.A. and Francis, 1990) 
and Principles of Route-to-Route Extrapolation for Risk Assessment 
(Gerrity et al., 1990). The risk characterization should identify the 
methods used to extrapolate across exposure routes and discuss the 
strengths and limitations of the approach.
    The level of confidence in the hazard characterization and 
quantitative dose-response evaluation should be stated to the extent 
possible, including placement of the agent into the appropriate 
category regarding the sufficiency of the health-related data (see 
Section III.G.). A comprehensive risk assessment ideally includes 
information on a variety of endpoints that provide insight into the 
full spectrum of potential reproductive responses. A profile that 
integrates both human and test species data and incorporates both 
sensitive endpoints (e.g., properly performed and fully evaluated 
histopathology) and functional correlates (e.g., fertility) allows more 
confidence in a risk assessment for a given agent.
    Descriptions of the nature of potential human exposures are 
important for prediction of specific outcomes and the likelihood of 
persistence or reversibility of the effect in different exposure 
situations with different subpopulations (U.S. EPA, 1992; Clegg, 1995).
    In the risk assessment process, risk is estimated as a function of 
exposure, with the risk of adverse effects increasing as exposure 
increases. Information on the levels of exposure experienced by 
different members of the population is key to understanding the range 
of risks that may occur. Where possible, several descriptors of 
exposure such as the nature and range of populations and their various 
exposure conditions, central tendencies, and high-end exposure 
estimates should be presented. Differences among individuals in 
absorption rates, metabolism, or other factors mean that individuals or 
subpopulations with the same level and pattern of exposure may have 
differing susceptibility. For example, the consequences of exposure can 
differ markedly between developing individuals, young adults and aged 
adults, including whether the effects are permanent or transient. Other 
considerations relative to human exposures might include pregnancy or 
lactation, potential for exposures to other agents, concurrent disease, 
nutritional status, lifestyle, ethnic background and genetic 
polymorphism, and the possible consequences. Knowledge of the molecular 
events leading to induction of adverse effects may be of use in 
determining the range of susceptibility in sensitive populations.
    An outline to serve as a guide and formatting aid for developing 
reproductive risk characterizations for chemical-specific risk 
assessments can be found in Table 7. A common format will assist risk 
managers in evaluating and using reproductive risk characterization. 
The outline has two parts. The first part tracks the reproductive risk 
assessment to bring forward its major conclusions. The second part 
pulls the information together to characterize the reproductive risk.

 Table 7.--Guide for Developing Chemical-Specific Risk Characterizations
                        for Reproductive Effects                        
------------------------------------------------------------------------
                                                                        
-------------------------------------------------------------------------
                                Part One                                
                                                                        
         Summarizing Major Conclusions in Risk Characterization         
                                                                        
                       I. Hazard Characterization                       
                                                                        
A. What is (are) the key toxicological study (or studies) that provides 
 the basis for health concerns for reproductive effects?                
   How good is the key study?                                   
   Are the data from laboratory or field studies? In a single or
   multiple species?                                                    
   What adverse reproductive endpoints were observed, and what  
   is the basis for the critical effect?                                
   Describe other studies that support this finding.            

[[Page 56310]]

                                                                        
   Discuss any valid studies which conflict with this finding.  
B. Besides the reproductive effect observed in the key study, are there 
 other health endpoints of concern? What are the significant data gaps? 
C. Discuss available epidemiological or clinical data. For              
 epidemiological studies:                                               
   What types of data were used (e.g., human ecologic, case-    
   control or cohort studies, or case reports or series)?               
   Describe the degree to which exposures were described.       
   Describe the degree to which confounding factors were        
   considered.                                                          
   Describe the degree to which other causal factors were       
   excluded.                                                            
D. How much is known about how (through what biological mechanism) the  
 chemical produces adverse reproductive effects?                        
   Discuss relevant studies of mechanisms of action or          
   metabolism.                                                          
   Does this information aid in the interpretation of the       
   toxicity data?                                                       
   What are the implications for potential adverse reproductive 
   effects?                                                             
E. Comment on any nonpositive data in animals or people, and whether    
 these data were considered in the hazard characterization.             
F. If adverse health effects have been observed in wildlife species,    
 characterize such effects by discussing the relevant issues as in A    
 through E above.                                                       
G. Summarize the hazard characterization and discuss the significance of
 each of the following:                                                 
   Confidence in conclusions                                    
   Alternative conclusions that are also supported by the data  
   Significant data gaps                                        
   Highlights of major assumptions                              
                                                                        
                  II. Characterization of Dose-Response                 
                                                                        
A. What data were used to develop the dose-response curve? Would the    
 result have been significantly different if based on a different data  
 set?                                                                   
   If laboratory animal data were used:                         
    Which species were used?                                            
    Most sensitive, average of all species, or other?                   
    Were any studies excluded? Why?                                     
   If epidemiological data were used:                           
    Which studies were used?                                            
    Only positive studies, all studies, or some other combination?      
    Were any studies excluded? Why?                                     
    Was a meta-analysis performed to combine the epidemiological        
     studies?                                                           
    What approach was used?                                             
    Were studies excluded? Why?                                         
B. Was a model used to develop the dose-response curve and, if so, which
 one? What rationale supports this choice? Is chemical-specific         
 information available to support this approach?                        
   How was the RfD/RfC (or the acceptable range) calculated?    
   What assumptions and uncertainty factors were used?          
   What is the confidence in the estimates?                     
C. Discuss the route, level, and duration of exposure observed, as      
 compared to expected human exposures.                                  
   Are the available data from the same route of exposure as the
   expected human exposures? If not, are pharmacokinetic data available 
   to extrapolate across route of exposure?                             
   How far does one need to extrapolate from the observed data  
   to environmental exposures? One to two orders of magnitude? Multiple 
   orders of magnitude? What is the impact of such an extrapolation?    
D. If adverse health effects have been observed in wildlife species,    
 characterize dose-response information using the process outlined in A 
 through C above.                                                       
                                                                        
                    III. Characterization of Exposure                   
                                                                        
A. What are the most significant sources of environmental exposure?     
  Are there data on sources of exposure from different media?           
  What is the relative contribution of different sources of exposure?   
  What are the most significant environmental pathways for exposure?    
B. Describe the populations that were assessed, including the general   
 population, highly exposed groups, and highly susceptible groups.      
C. Describe the basis for the exposure assessment, including any        
 monitoring, modeling, or other analyses of exposure distributions such 
 as Monte Carlo or krieging.                                            
D. What are the key descriptors of exposure?                            
  Describe the (range of) exposures to: ``average'' individuals, ``high-
   end'' individuals, general population, high exposure group(s),       
   children, susceptible populations, males, females (nonpregnant,      
   pregnant, lactating).                                                
  How was the central tendency estimate developed?                      
  What factors and/or methods were used in developing this estimate?    
  How was the high-end estimate developed?                              
  Is there information on highly exposed subgroups?                     
  Who are they?                                                         
  What are their levels of exposure?                                    
  How are they accounted for in the assessment?                         
E. Is there reason to be concerned about cumulative or multiple         
 exposures because of biological, ethnic, racial, or socioeconomic      
 reasons?                                                               
F. If adverse reproductive effects have been observed in wildlife       
 species, characterize wildlife exposure by discussing the relevant     
 issues as in A through E above.                                        
G. Summarize exposure conclusions and discuss the following:            
   Results of different approaches, i.e., modeling, monitoring, 
   probability distributions;                                           
   Limitations of each, and the range of most reasonable values;
   Confidence in the results obtained, and the limitations to   
   the results                                                          
                                                                        

[[Page 56311]]

                                                                        
                                Part Two                                
                                                                        
                    Risk Conclusions and Comparisons                    
                                                                        
                          IV. Risk Conclusions                          
                                                                        
A. What is the overall picture of risk, based on the hazard,            
 quantitative dose-response, and exposure characterizations?            
B. What are the major conclusions and strengths of the assessment in    
 each of the three main analyses (i.e., hazard characterization,        
 quantitative dose-response, and exposure assessment)?                  
C. What are the major limitations and uncertainties in the three main   
 analyses?                                                              
D. What are the science policy options in each of the three major       
 analyses?                                                              
  What are the alternative approaches evaluated?                        
  What are the reasons for the choices made?                            
                                                                        
                             V. Risk Context                            
                                                                        
A. What are the qualitative characteristics of the reproductive hazard  
 (e.g., voluntary vs. involuntary, technological vs. natural, etc.)?    
 Comment on findings, if any, from studies of risk perception that      
 relate to this hazard or similar hazards.                              
B. What are the alternatives to this reproductive hazard? How do the    
 risks compare?                                                         
C. How does this reproductive risk compare to other risks?              
  How does this risk compare to other risks in this regulatory program, 
   or other similar risks that the EPA has made decisions about?        
  Where appropriate, can this risk be compared with past Agency         
   decisions, decisions by other federal or state agencies, or common   
   risks with which people may be familiar?                             
  Describe the limitations of making these comparisons.                 
D. Comment on significant community concerns which influence public     
 perception of risk.                                                    
                                                                        
                      VI. Existing Risk Information                     
                                                                        
Comment on other reproductive risk assessments that have been done on   
 this chemical by EPA, other federal agencies, or other organizations.  
 Are there significantly different conclusions that merit discussion?   
                                                                        
                         VII. Other Information                         
                                                                        
Is there other information that would be useful to the risk manager or  
 the public in this situation that has not been described above?        
------------------------------------------------------------------------

VI.C. Descriptors of Reproductive Risk

    Descriptors of reproductive risk convey information and answer 
questions about risk, with each descriptor providing different 
information and insights. There are a number of ways to describe risk. 
Details on how to use these descriptors can be obtained from the 
guidance on risk characterization (Browner, 1995) from which some of 
the information below has been extracted.
    In most cases, the state of the science is not yet adequate to 
define distributions of factors such as population susceptibility. The 
guidance principles below discuss a variety of risk descriptors that 
primarily reflect differences in estimated exposure. If a full 
description of the range of susceptibility in the population cannot be 
presented, an effort should be made to identify subgroups that, for 
various reasons, may be particularly susceptible.
VI.C.1. Distribution of Individual Exposures
    Risk managers are interested generally in answers to questions such 
as: (1) Who are the people at the highest risk and why? (2) What is the 
average risk or distribution of risks for individuals in the population 
of interest? and (3) What are they doing, where do they live, etc., 
that might be putting them at this higher risk?
    Exposure and reproductive risk descriptors for individuals are 
intended to provide answers to these questions. To describe the range 
of risks, both high-end and central tendency descriptors are used to 
convey the distribution in risk levels experienced by different 
individuals in the population. For the Agency's purposes, high-end risk 
descriptors are plausible estimates of the individual risk for those 
persons at the upper end of the risk distribution. Given limitations in 
current understanding of variability in individuals' sensitivity to 
agents that cause reproductive toxicity, high-end descriptors will 
usually address high-end exposure or dose. Conceptually, high-end 
exposure means exposure above approximately the 90th percentile of the 
population distribution, but not higher than the individual in the 
population who has the highest exposure. Central tendency descriptors 
generally reflect central estimates of exposure or dose. The descriptor 
addressing central tendency may be based on either the arithmetic mean 
exposure (average estimate) or the median exposure (median estimate), 
either of which should be clearly labeled. The selection of which 
descriptor(s) to present in the risk characterization will depend on 
the available data and the goals of the assessment.
VI.C.2. Population Exposure
    Population risk refers to assessment of the extent of harm for the 
population as a whole. In theory, it can be calculated by summing the 
individual risks for all individuals within the subject population. 
That task requires more information than is usually available. 
Questions addressed by descriptors of population risk for reproductive 
effects would include: What portion of the population is within a 
specified range of some reference level, e.g., exceeds the RfD (a 
dose), the RfC (a concentration), or other health concern level?
    For reproductive effects, risk assessment techniques have not been 
developed generally to the point of knowing how to add risk 
probabilities, although Hattis and Silver (1994) have proposed 
approaches for certain case-specific situations. Therefore, the 
following descriptor is usually appropriate: An estimate of the 
percentage of the population, or the number of persons, above a 
specified level of risk or within a specified range of some reference 
level (e.g., exceeds the RfD, RfC, LOAEL, or other specific level of 
interest). The RfD or RfC is assumed to be a level below which no 
significant risk occurs. Therefore, information from the exposure 
assessment on the populations below the RfD or RfC (``not likely to be 
at risk'') and above the RfD or RfC (``may be at risk'') may be useful 
information for risk managers. Estimating the number of persons 
potentially removed from the ``may be at risk'' category after a 
contemplated action is taken may be particularly

[[Page 56312]]

useful to a risk manager considering possible actions to ameliorate 
risk for a population. This descriptor must be obtained through 
measuring or simulating the population distribution.
VI.C.3. Margin of Exposure
    In the risk characterization, dose-response information and the 
human exposure estimates may be combined either by comparing the RfD or 
RfC and the human exposure estimate or by calculating the margin of 
exposure (MOE). The MOE is the ratio of the NOAEL or benchmark dose 
from the most appropriate or sensitive species to the estimated human 
exposure level from all potential sources (U.S. EPA, 1985a). If a NOAEL 
is not available, a LOAEL may be used in the calculation of the MOE, 
but consideration for the acceptability would be different than when a 
NOAEL is used. Considerations for the acceptability of the MOE are 
similar to those for the selection of uncertainty factors applied to 
the NOAEL, LOAEL, or the benchmark dose for the derivation of an RfD. 
The MOE is presented along with the characterization of the database, 
including the strengths and weaknesses of the toxicity and exposure 
data, the number of species affected, and the information on dose-
response, route, timing, and duration. The RfD or RfC comparison with 
the human exposure estimate and the calculation of the MOE are 
conceptually similar, but may be used in different regulatory 
situations.
    The choice of approach is dependent on several factors, including 
the statute involved, the situation being addressed, the database used, 
and the needs of the decisionmaker. The RfD, RfC, or MOE are considered 
along with other risk assessment and risk management issues in making 
risk management decisions, but the scientific issues that should be 
taken into account in establishing them have been addressed here.
VI.C.4. Distribution of Exposure and Risk for Different Subgroups
    A risk manager might also ask questions about the distribution of 
the risk burden among various segments of the subject population such 
as the following: How do exposure and reproductive risk impact various 
subgroups? and What is the population risk of a particular subgroup? 
Questions about the distribution of exposure and reproductive risk 
among such population segments require additional risk descriptors.

Highly Exposed

    The purpose of this measure is to describe the upper end of the 
exposure distribution, allowing risk managers to evaluate whether 
certain individuals are at disproportionately high or unacceptably high 
risk. The objective is to look at the upper end of the exposure 
distribution to derive a realistic estimate of relatively highly 
exposed individual(s). The ``high end'' of the risk distribution has 
been defined (Habicht, 1992; Browner, 1995) as above the 90th 
percentile of the actual (either measured or estimated) distribution. 
Whenever possible, it is important to express the number or proportion 
of individuals who comprise the selected highly exposed group and, if 
data are available, discuss the potential for exposure at still higher 
levels.
    Highly exposed subgroups can be identified and, where possible, 
characterized, and the magnitude of risk quantified. This descriptor is 
useful when there is (or is expected to be) a subgroup experiencing 
significantly different exposures or doses from those of the larger 
population. These subpopulations may be identified by age, sex, 
lifestyle, economic factors, or other demographic variables. For 
example, toddlers who play in contaminated soil and consumers of large 
amounts of fish represent subpopulations that may have greater 
exposures to certain agents.
    If population data are absent, it will often be possible to 
describe a scenario representing high-end exposures using upper 
percentile or judgment-based values for exposure variables. In these 
instances, caution should be taken not to overestimate the high-end 
values if a ``reasonable'' exposure estimate is to be achieved.

Highly Susceptible

    Highly susceptible subgroups also can be identified and, if 
possible, characterized, and the magnitude of risk quantified. This 
descriptor is useful when the sensitivity or susceptibility to the 
effect for specific subgroups is (or is expected to be) significantly 
different from that of the larger population. Therefore, the purpose of 
this measure is to quantify exposure of identified sensitive or 
susceptible populations to the agent of concern. Sensitive or 
susceptible individuals are those within the exposed population at 
increased risk of expressing the adverse effect. Examples might be 
pregnant or lactating women, women with reduced oocyte numbers, men 
with ``borderline'' sperm counts, or infants. To calculate risk for 
these subgroups, it will be necessary sometimes to use a different 
dose-response relationship; e.g., upon exposure to a chemical, pregnant 
or lactating women, elderly people, children of varying ages, and 
people with certain illnesses may each be more sensitive than the 
population as a whole.
    In general, not enough is understood about the mechanisms of 
toxicity to identify sensitive subgroups for most agents, although 
factors such as age, nutrition, personal habits (e.g., smoking, 
consumption of alcohol, and abuse of drugs), existing disease (e.g., 
diabetes or sexually transmitted diseases), or genetic polymorphisms 
may predispose some individuals to be more sensitive to the 
reproductive effects of various agents.
    It is important to consider, however, that the Agency's current 
methods for developing reference doses and reference concentrations 
(RfDs and RfCs) are designed to protect sensitive populations. If data 
on sensitive human populations are available (and there is confidence 
in the quality of the data), then the RfD is based on the dose level at 
which no adverse effects are observed in the sensitive population. If 
no such data are available (for example, if the RfD is developed using 
data from humans of average or unknown sensitivity), then an additional 
3- to 10-fold factor may be used to account for variability between the 
average human response and the response of more sensitive individuals 
(see Section IV).
    Generally, selection of the population segments to consider for 
high susceptibility is a matter of either a prior interest in the 
subgroup (e.g., environmental justice considerations), in which case 
the risk assessor and risk manager can jointly agree on which subgroups 
to highlight, or a matter of discovery of a sensitive or highly exposed 
subgroup during the assessment process. In either case, once 
identified, the subgroup can be treated as a population in itself and 
characterized in the same way as the larger population using the 
descriptors for population and individual risk.
VI.C.5. Situation-Specific Information
    Presenting situation-specific scenarios for important exposure 
situations and subpopulations in the form of ``what if?'' questions may 
be particularly useful to give perspective to risk managers on possible 
future events. The question being asked in these cases is, for any 
given exposure level, what would be the resulting number or proportion 
of individuals who may be exposed to levels above that value?
    ``What if * * *?'' questions, such as those that follow, can be 
used to examine candidate risk management options:

[[Page 56313]]

     What are the reproductive risks if a pesticide applicator 
applies this pesticide without using protective equipment?
     What are the reproductive risks if this site becomes 
residential in the future?
     What are the reproductive risks if we set the standard at 
100 ppb?
    Answering such ``what if?'' questions involves a calculation of 
risk based on specific combinations of factors postulated within the 
assessment. The answers to these ``what if?'' questions do not, by 
themselves, give information about how likely the combination of values 
might be in the actual population or about how many (if any) persons 
might be subjected to the potential future reproductive risk. However, 
information on the likelihood of the postulated scenario would be 
desirable to include in the assessment.
    When addressing projected changes for a population (either expected 
future developments or consideration of different regulatory options), 
it usually is appropriate to calculate and consider all the 
reproductive risk descriptors discussed above. When central tendency or 
high-end estimates are developed for a scenario, these descriptors 
should reflect reasonable expectations about future activities. For 
example, in site-specific risk assessments, future scenarios should be 
evaluated when they are supported by realistic forecasts of future land 
use, and the reproductive risk descriptors should be developed within 
that context.
VI.C.6. Evaluation of the Uncertainty in the Risk Descriptors
    Reproductive risk descriptors are intended to address variability 
of risk within the population and the overall adverse impact on the 
population. In particular, differences between high-end and central 
tendency estimates reflect variability in the population but not the 
scientific uncertainty inherent in the risk estimates. As discussed 
above there will be uncertainty in all estimates of reproductive risk. 
These uncertainties can include measurement uncertainties, modeling 
uncertainties, and assumptions to fill data gaps. Risk assessors should 
address the impact of each of these factors on the confidence in the 
estimated reproductive risk values.
    Both qualitative and quantitative evaluations of uncertainty 
provide useful information to users of the assessment. The techniques 
of quantitative uncertainty analysis are evolving rapidly and both the 
SAB (Loehr and Matanoski, 1993) and the NRC (1994) have urged the 
Agency to incorporate these techniques into its risk analyses. However, 
it should be noted that a probabilistic assessment that uses only the 
assessor's best estimates for distributions of population variables 
addresses variability, but not uncertainty. Uncertainties in the 
estimated risk distribution need to be evaluated separately. An 
approach has been proposed for estimating distribution of uncertainty 
in noncancer risk assessments (Baird et al., 1996).

VI.D. Summary and Research Needs

    These Guidelines summarize the procedures that the EPA will follow 
in evaluating the potential for agents to cause reproductive toxicity. 
They discuss the assumptions that must be made in risk assessment for 
reproductive toxicity because of gaps in our knowledge about underlying 
biologic processes and how these compare across species. Research to 
improve the interpretation of data and interspecies extrapolation is 
needed. This research includes studies that: (1) more completely 
characterize and define female and male reproductive endpoints, (2) 
more completely characterize the types of developmental toxicity 
possible, (3) evaluate the interrelationships among endpoints, (4) 
examine quantitative extrapolation between endpoints (e.g., sperm 
count) and function (e.g., fertility), (5) provide a better 
understanding of the relationships between reproductive toxicity and 
other forms of toxicity, (6) explore pharmacokinetic disposition of the 
target, and (7) examine mechanistic phenomena related to 
pharmacokinetic disposition. These types of studies, along with further 
evaluation of a nonlinear dose-response for susceptible populations, 
should provide methods to more precisely assess risk.

VII. References

    Aafjes, J.H., Vels, J.M., Schenck, E. (1980) Fertility of rats 
with artificial oligozoospermia. J. Reprod. Fertil. 58:345-351.
    Adler, N.T., Toner, J.P. (1986) The effect of copulatory 
behavior on sperm transport and fertility in rats. In: Komisaruk, 
B.R., Siegel, H.I., Chang, M.F., Feder, H.H. Reproduction: 
Behavioral and Neuroendocrine Perspective. New York Academy of 
Science, New York. pp. 21-32.
    Allen, B.C., Kavlock, R.J., Kimmel, C.A., Faustman, E.M. (1994a) 
Dose-response assessment for developmental toxicity: II. Comparison 
of generic benchmark dose estimates with NOAELs. Fundam. Appl. 
Toxicol. 23:487-495.
    Allen, B.C., Kavlock, R.J., Kimmel, C.A., Faustman, E.M. (1994b) 
Dose-response assessment for developmental toxicity: III. 
Statistical models. Fundam. Appl. Toxicol. 23:496-509.
    Amann, R.P. (1981) A critical review of methods for evaluation 
of spermatogenesis from seminal characteristics. J. Androl. 2:37-58.
    American Academy of Pediatrics Committee on Drugs. (1994) The 
transfer of drugs and other chemicals into human milk. Pediatrics 
93:137-150.
    Armstrong, D.L. (1986) Environmental stress and ovarian 
function. Biol. Reprod. 34:29-39.
    Atterwill, C.K., Flack, J.D. (1992) Endocrine Toxicology. 
Cambridge University Press, Cambridge.
    Auger, J., Kunstman, J.M., Czyglik, F., Jouannet, P. (1995) 
Decline in semen quality among fertile men in Paris during the past 
20 years. N. Engl. J. Med. 332:281-285.
    Axelson, O. (1985) Epidemiologic methods in the study of 
spontaneous abortions: sources of data, methods, and sources of 
error. In: Hemminki, K., Sorsa, M., Vaino, H. Occupational Hazards 
and Reproduction. Hemisphere, Washington. pp. 231-236.
    Baird, D.D., Wilcox, A.J. (1985) Cigarette smoking associated 
with delayed conception. JAMA 253:2979-2983.
    Baird, D.D., Wilcox, A.J., Weinberg, C.R. (1986) Using time to 
pregnancy to study environmental exposures. Am. J. Epidemiol. 
124:470-480.
    Baird, S.J.S., Cohen, J.T., Graham, J.D., Shlyakhter, A.I., 
Evans, J.S. (1996) Noncancer risk assessment: a probabilistic 
alternative to current practice. Human Ecol. Risk Assess. 2:79-102.
    Barlow, S.M., Sullivan, F.M. (1982) Reproductive Hazards of 
Industrial Chemicals. Academic Press, London.
    Barsotti, D.A., Abrahamson, L.J., Allen, J.R. (1979) Hormonal 
alterations in female rhesus monkeys fed a diet containing 2,3,7,8-
TCDD. Bull. Environ. Contam. Toxicol. 21:463-469.
    Beach, F.A. (1979) Animal models for human sexuality. In: Ciba 
Foundation Symposium No. 62, Sex, Hormones and Behavior. Elsevier-
North Holland, London. pp. 113-143.
    Berndtson, W.E. (1977) Methods for quantifying mammalian 
spermatogenesis: a review. J. Anim. Sci. 44:818-833.
    Bernstein, M.E. (1984) Agents affecting the male reproductive 
system: effects of structure on activity. Drug Metab. Rev. 15:941-
996.
    Biava, C.G., Smuckler, E.A., Whorton, D. (1978) The testicular 
morphology of individuals exposed to dibromochloropropane. Exp. Mol. 
Pathol. 29:448-458.
    Blazak, W.F., Ernst, T.L., Stewart, B.E. (1985) Potential 
indicators of reproductive toxicity, testicular sperm production and 
epididymal sperm number, transit time and motility in Fischer 344 
rats. Fundam. Appl. Toxicol. 5:1097-1103.
    Blazak, W.F., Treinen, K.A., Juniewicz, P.E. (1993) Application 
of testicular sperm head counts in the assessment of male 
reproductive toxicity. In: Chapin, R.E. and Heindel, J.J. Methods in 
Toxicology: Male Reproductive Toxicology. Academic Press, San Diego. 
pp. 86-94.
    Bloom, A.D. (1981) Guidelines for reproductive studies in 
exposed human populations. Guideline for studies of human 
populations exposed to mutagenic and

[[Page 56314]]

reproductive hazards. Report of Panel II. March of Dimes Birth 
Defects Foundation, White Plains, NY, pp. 37-110.
    Boyd, J.A., Clark, G.C., Walmer, D.K., Patterson, D.G., Needham, 
L.L., Lucier, G.W. (1995) Endometriosis and the environment: 
biomarkers of toxin exposure. Abstract from Endometriosis 2000 
Workshop, May 15-17.
    Boyers, S.P., Davis, R.O., Katz, D.F. (1989) Automated semen 
analysis. Curr. Probl. Obstet. Gynecol. Fertil. 12:173-200.
    Brawer, J.R., Finch, C.E. (1983) Normal and experimentally 
altered aging processes in the rodent hypothalamus and pituitary. 
In: Walker, R.F., Cooper, R.L. Experimental and Clinical 
Interventions in Aging. Marcel Dekker, New York. pp. 45-65.
    Brouwer, A., Ahlborg, U.G., Vandenberg, M., Birnbaum, L.S., 
Boersma, E.R., Bosveld, B., Denison, M.S., Gray, L.E., Hagmar, L., 
Holene, E., Huisman, M., Jacobson, S.W., Jacobson, J.L., 
Koopmanesseboom, C., Koppe, J.G., Kulig, B.M., Morse, D.C., Muckle, 
G., Peterson, R.E., Sauer, P.J.J., Seegal, R.F., Smitsvanprooije, 
A.E., Touwen, B.C.L., Weisglaskuperus, N., Winneke, G. (1995) 
Functional aspects of developmental toxicity of polyhalogenated 
aromatic hydrocarbons in experimental animals and human infants. 
Eur. J. Pharmacol. 293:1-40.
    Browner, C.M. (1995) EPA risk characterization program. U.S. EPA 
Memorandum, March 21, 1995. Available from the EPA Air docket.
    Brown-Grant, K., Davidson, J.M., Grieg, F. (1973) Induced 
ovulation in albino rats exposed to constant light. J. Endocrinol. 
57:7-22.
    Bujan, L., Mansat, A., Pontonnier, F., Mieusset, R. (1996) Time 
series analysis of sperm concentration in fertile men in Toulouse, 
France between 1977 and 1992. Br. Med. J. 312: 471-472.
    Bulger, W.H., Kupfer, D. (1985) Estrogenic activity of 
pesticides and other xenobiotics on the uterus and male reproductive 
tract. In: Thomas, J.A., Korach, K.S., McLachlan, J.A. Endocrine 
Toxicology. Raven Press, New York. pp. 1-33.
    Burch, T.K., Macisco, J.J., Parker, M.P. (1967) Some 
methodologic problems in the analysis of menstrual data. Int. J. 
Fertil. 12:67-76.
    Burger, E.J., Tardiff, R.G., Scialli, A.R., Zenick, H. (1989) 
Sperm Measures and Reproductive Success. Alan R. Liss, New York.
    Butcher, R.L., Fugo, N.W. (1967) Overripeness and the mammalian 
ova. II. Delayed ovulation and chromosome anomalies. Fertil. Steril. 
18:297-302.
    Butcher, R.L., Blue, J.D., Fugo, N.W. (1969) Overripeness and 
the mammalian ova. III. Fetal development at midgestation and at 
term. Fertil. Steril. 20:223-231.
    Butcher, R.L., Collins, W.E., Fugo, N.W. (1975) Altered 
secretion of gonadotropins and steroids resulting from delayed 
ovulation in the rat. Endocrinology 96:576-586.
    Byskov, A.G., Hoyer, P.E. (1994) Embryology of mammalian gonads 
and ducts. In: Knobil, E., Neill, J.D. The Physiology of 
Reproduction. Raven Press, New York. pp. 487-540.
    Carlsen, E., Giwercman, A., Keiding, N., Skakkebaek, N.E. (1992) 
Evidence for decreasing quality of semen during past 50 years. Br. 
Med. J. 305:609-613.
    Cassidy, S.L., Dix, K.M., Jenkins, T. (1983) Evaluation of a 
testicular sperm head counting technique using rats exposed to 
dimethoxyethyl phthalate (DMEP), glycerol alpha-monochlorohydrin 
(GMCH), epichlorohydrin (ECH), formaldehyde (FA), or methyl 
methanesulphonate (MMS). Arch. Toxicol. 53:71-78.
    Chapin, R.E. (1988) Morphologic evaluation of seminiferous 
epithelium of the testis. In: Lamb, J.C., Foster, P.M.D. Physiology 
and Toxicology of Male Reproduction. Academic Press, New York. pp. 
155-177.
    Chapin, R.E., Heindel, J.J. (1993) Methods in Toxicology: Male 
Reproductive Toxicology. Academic Press, San Diego.
    Chapin, R.E., Filler, R.S., Gulati, D., Heindel, J.J., Katz, 
D.F., Mebus, C.A., Obasaju, F., Perreault, S.D., Russell, S.R., 
Schrader, S., Slott, V., Sokol, R.Z., Toth, G. (1992) Methods for 
assessing rat sperm motility. Reprod. Toxicol. 6:267-273.
    Chapin, R.E., Gulati, D.K., Barnes, L.H., Teague, J.L. (1993a) 
The effects of feed restriction on reproductive function in Sprague-
Dawley rats. Fundam. Appl. Toxicol. 20:23-29.
    Chapin, R.E., Gulati, D.K., Fail, P.A., Hope, E., Russell, S.R., 
Heindel, J.J., George, J.D., Grizzle, T.B., Teague, J.L. (1993b) The 
effects of feed restriction on reproductive function in Swiss CD-1 
mice. Fundam. Appl. Toxicol. 20:15-22.
    Chapman, R.M. (1983) Gonadal injury resulting from chemotherapy. 
In: Mattison, D.R. Reproductive Toxicology. Alan R. Liss, New York. 
pp. 149-161.
    Clegg, E.D. (1995) Reversibility of effects: overview and 
reproductive systems. Inhal. Toxicol. 7:881-889.
    Colborn, T., vom Saal, F.S., Soto, A.M. (1993) Developmental 
effects of endocrine-disrupting chemicals in wildlife and humans. 
Environ. Health Perspect. 101:378-384.
    Colie, C.F. (1993) Male mediated teratogenesis. Reprod. Toxicol. 
7:3-9.
    Collins, T.F.X. (1978) Multigeneration reproduction studies. In: 
Wilson, J.G., Fraser, F.C. Handbook of Teratology. Plenum Press, New 
York, pp. 191-214.
    Cooper, R.L., Walker, R.F. (1979) Potential therapeutic 
consequences of age-dependent changes in brain physiology. Interdis. 
Topics Gerontol. 15:54-76.
    Cooper, R.L., Conn, P.M., Walker, R.F. (1980) Characterization 
of the LH surge in middle-aged female rats. Biol. Reprod. 23:611-
615.
    Cooper, R.L., Chadwick, R.W., Rehnberg, G.L., Goldman, J.M., 
Booth, K.C., Hein, J.F., McElroy, W.K. (1989) Effect of lindane on 
hormonal control of reproductive function in the female rat. 
Toxicol. Appl. Pharmacol. 99:384-394.
    Cooper, R.L., Goldman, J.M., Vandenbergh, J.G. (1993) Monitoring 
of the estrous cycle in the laboratory rodent by vaginal lavage. In: 
Heindel, J.J., Chapin, R.E. Methods in Toxicology: Female 
Reproductive Toxicology. Academic Press, San Diego. pp. 45-56.
    Cooper, R.L., Barrett, M.A., Goldman, J.M., Rehnberg, G.R., 
McElroy, W.K., Stoker, T.E. (1994) Pregnancy alterations following 
xenobiotic-induced delays in ovulation in the female rat. Fundam. 
Appl. Toxicol. 22:474-480.
    Cooper, R.L., Stoker, T.E., Goldman, J.M., Parrish, M.B., Tyrey, 
L. (1996) Effect of atrazine on ovarian function in the rat. Reprod. 
Toxicol. 10: in press.
    Crisp, T.M. (1992) Organization of the ovarian follicle and 
events in its biology: oogenesis, ovulation or atresia. Mutat. Res. 
296:89-106.
    Crump, K.S. (1984) A new method for determining allowable daily 
intakes. Fundam. Appl. Toxicol. 4:854-871.
    Csapo, A.I., Pulkkinen, M. (1978) Indispensability of the human 
corpus luteum in the maintenance of early pregnancy: lutectomy 
evidence. Obstet. Gynecol. Surv. 33:69.
    Cummings, A.M., Gray, L.E. (1987) Methoxychlor affects the 
decidual cell response of the uterus but not other progestational 
parameters in female rats. Toxicol. Appl. Pharmacol. 90:330-336.
    Cummings, A.M., Perreault, S.D. (1990) Methoxychlor accelerates 
embryo transport through the rat reproductive tract. Toxicol. Appl. 
Pharmacol. 102:110-116.
    Darney, S.P. (1991) In vitro assessment of gamete integrity. In: 
Goldberg, A.M. In Vitro Toxicology: Mechanisms and New Technology. 
Mary Ann Liebert, Inc., New York. pp. 63-75.
    Davis, D.L., Friedler, G., Mattison, D., Morris, R. (1992) Male-
mediated teratogenesis and other reproductive effects: biologic and 
epidemiologic findings and a plea for clinical research. Reprod. 
Toxicol. 6:289-292.
    de Boer, P., van der Hoeven, F.A., Chardon, J.A.P. (1976) The 
production, morphology, karyotypes and transport of spermatozoa from 
tertiary trisomic mice and the consequences for egg fertilization. 
J. Reprod. Fertil. 48:249-256.
    Dixon, R.L., Hall, J.L. (1984) Reproductive toxicology. In: 
Hayes, A.W. Principles and Methods of Toxicology. Raven Press, New 
York. pp. 107-140.
    Dobbins, J.G., Eifler, C.W., Buffler, P.A. (1978) The use of 
parity survivorship analysis in the study of reproductive outcomes. 
Presented at the Society for Epidemiologic Research Conference, 
Seattle, WA: June, 1978.
    Dobson, R.L., Felton, J.S. (1983) Female germ cell loss from 
radiation and chemical exposure. Am. J. Ind. Med. 4:175-190.
    Drouva, S.V., Laplante, E., Kordon, C. (1982) Alpha 1-adrenergic 
receptor involvement in the LH surge in ovariectomized estrogen-
primed rats. Eur. J. Pharmacol. 81:341-344.
    Egeland, G.M., Sweeney, M.H., Fingerhut, M.A., Wille, K.K., 
Schnorr, T.M., Halperin, W.E. (1994) Total serum testosterone and 
gonadotropins in workers exposed to dioxin. Am. J. Epidemiol. 
139:272-281.
    Egnatz, D.G., Ott, M.G., Townsend, J.C., Olson, R.D., Johns, 
D.B. (1980) DBCP and testicular effects in chemical workers; an 
epidemiological survey in Midland. J. Occup. Med. 22:727-732.
    Epidemiology Workgroup for the Interagency Regulatory Liaison 
Group (1981)

[[Page 56315]]

Guidelines for documentation of epidemiologic studies. Am. J. 
Epidemiol. 114:609-613.
    Everett, J.W., Sawyer, C.H. (1950) A 24-hour periodicity in the 
``LH-release apparatus'' of female rats disclosed by barbiturate 
sedation. Endocrinology 47:198-218.
    Everson, R.B., Sandler, D.P., Wilcox, A.J., Schreinemachers, D., 
Shore, D.L., Weinberg, C. (1986) Effect of passive exposure to 
smoking on age at natural menopause. Br. Med. J. 293:792.
    Fabia, J., Thuy, T.D. (1974) Occupation of father at time of 
children dying of malignant disease. Br. J. Prev. Soc. Med. 28:98-
100.
    Fawcett, D.W. (1986) Bloom and Fawcett: A Textbook of Histology. 
W.B. Saunders, Philadelphia, PA.
    Filler, R. (1993) Methods for evaluation of rat epididymal sperm 
morphology. In: Chapin, R.E., Heindel, J.J. Methods in Toxicology: 
Male Reproductive Toxicology. Academic Press, San Diego. pp. 334-
343.
    Finch, C.E., Felicio, L.S., Mobbs, C.V. (1984) Ovarian and 
steroidal influences on neuroendocrine aging processes in female 
rodents. Endocrinol. Rev. 5:467-497.
    Fink, G. (1988) Gonadotropin secretion and its control. In: 
Knobil, E., Neill, J.D. The Physiology of Reproduction. Raven Press, 
New York. pp. 1349-1377.
    Fisch, H., Goluboff, E.T., Olson, J.H., Feldshuh, J., Broder, 
S.J., Barad, D.H. (1996) Semen analyses in 1,283 men from the United 
States over a 25-year period: no decline in fertility. Fertil. 
Steril. 65:1009-1014.
    Foote, R.H., Berndtson, W.E. (1992) The Germinal Cells. In: 
Scialli, A.R., Clegg, E.D. Reversibility in Testicular Toxicity 
Assessment. CRC Press, Boca Raton. pp. 1-55.
    Foote, R.H., Schermerhorn, E.C., Simkin, M.E. (1986) Measurement 
of semen quality, fertility, and reproductive hormones to assess 
dibromochloropropane (DBCP) effects in live rabbits. Fundam. Appl. 
Toxicol. 6:628-637.
    Forsberg, J.G. (1981) Permanent changes induced by DES at 
critical stages in human and model systems. Biol. Res. Pregnancy 
2:168-175.
    Foster, P.M.D. (1992) The Sertoli cell. In: Scialli, A.R., 
Clegg, E.D. Reversibility in Testicular Toxicity Assessment. CRC 
Press, Boca Raton. pp. 57-86.
    Francis, E.Z., Kimmel, G.L. (1988) Proceedings of the workshop 
on one-vs two-generation reproductive effects studies. J. Am. Coll. 
Toxicol. 7:911-925.
    Franken, D.R., Burkman, L.J., Coddington, C.C., Oehninger, S., 
Hodgen, G.D. (1990) Human hemizona attachment assay. In: Acosta, 
A.A., Swanson, R.J., Ackerman, S.B., Kruger, T.F., VanZyl, J.A., 
Menkveld, R. Human Spermatozoa in Assisted Reproduction. Williams 
and Wilkins, Baltimore. pp. 355-371.
    Fugo, N.W., Butcher, R.L. (1966) Overripeness and the mammalian 
ova. I. Overripeness and early embryonic development. Fertil. 
Steril. 17:804-814.
    Gaffey, W.R. (1976) A critique of the standard mortality ratio. 
J. Occup. Med. 18:157-160.
    Galbraith, W.M., Voytek, P., Ryon, M.S. (1983) Assessment of 
risks to human reproduction and development of the human conceptus 
from exposure to environmental substances. In: Christian, M.S., 
Galbraith, W.M., Voytek, P., Mehlman, M.A. Advances in Modern 
Environmental Toxicology. Princeton Scientific Publ., Princeton. pp. 
41-153.
    Galletti, F., Klopper, A. (1964) The effect of progesterone on 
the quantity and distribution of body fat in the female rat. Acta 
Endocrinol. 46:379-386.
    Gardner, M.J., Hall, A.J., Snee, M.P., Downes, S., Powell, C.A., 
Terrell, J.D. (1990a) Methods and basic data of case-control study 
of leukaemia and lymphoma among young people near Sellafield nuclear 
plant in West Cumbria. Br. Med. J. 300:429-434.
    Gardner, M.J., Snee, M.P., Hall, A.J., Powell, C.A., Downes, S., 
Terrell, J.D. (1990b) Results of case-control study of leukaemia and 
lymphoma among young people near Sellafield nuclear plant in West 
Cumbria. Br. Med. J. 300:423-429.
    Gaylor, D.W. (1989) Quantitative risk analysis for quantal 
reproductive and developmental effects. Environ. Health 79:243-246.
    Gellert, R.J. (1978) Kepone, mirex, dieldrin, and aldrin: 
estrogenic activity and the induction of persistent vaginal estrus 
and anovulation in rats following neonatal treatment. Environ. Res. 
16:131-138.
    Generoso, W.M., Piegorsch, W.W. (1993) Dominant lethal tests in 
male and female mice. In: Chapin, R.E., Heindel, J.J. Methods in 
Toxicology: Male Reproductive Toxicology. Academic Press, San Diego. 
pp. 124-139.
    Generoso, W.M., Rutledge, J.C., Cain, K.T., Hughes, L.A., 
Braden, P.W. (1987) Exposure of female mice to ethylene oxide within 
hours after mating leads to fetal malformation and death. Mutat. 
Res. 176:269-274.
    George, F.W., Wilson, J.D. (1994) Sex determination and 
differentiation. In: Knobil, E., Neill, J.D. The Physiology of 
Reproduction. Raven Press, New York. pp. 3-28.
    Gerhard, I., Runnebaum, B. (1992) Grenzen der hormonsubstitution 
bei Schadstoffbelastung und fertilitatsstorungen. Zentralbl. 
Gynakol. 114:593-602.
    Gerrity, T.R., Henry, C.J., Bronaugh, R., et al. (1990) Summary 
report of the workshops on principles of route-to-route 
extrapolation for risk assessment. In: Gerrity, T.R., Henry, C.J. 
Principles of Route-To-Route Extrapolation for Risk Assessment. 
Elsevier Science Publ. Co., New York. pp. 1-12.
    Gill, W.B., Schumacher, F.B., Bibbo, M., Straus, F.H., 
Schoenberg, H.W. (1979) Association of diethylstilbestrol exposure 
in utero with cryptorchidism, testicular hypoplasia and semen 
abnormalities. J. Urol. 122:36-39.
    Ginsburg, J., Okolo, S., Prelevic, G., Hardiman, P. (1994) 
Residence in London area and sperm density. Lancet 343:230.
    Giusti, R.M., Iwamoto, K., Hatch, E.E. (1995) Diethylstilbestrol 
revisited: a review of the long-term health effects. Ann. Intern. 
Med. 122:778-788.
    Giwercman, A., Carlsen, E., Keiding, N., Skakkebaek, N.E. (1993) 
Evidence for increasing incidence of abnormalities of the human 
testis: A review. Environ. Health Perspect. 101:65-71.
    Goldman, J.M., Cooper, R.L., Laws, S.C., Rehnberg, G.L., 
Edwards, T.L., McElroy, W.K., Hein, J.F. (1990) Chlordimeform-
induced alterations in endocrine regulation within the male rat 
reproductive system. Toxicol. Appl. Pharmacol. 104:25-35.
    Goldman, J.M., Cooper, R.L., Edwards, T.L., Rehnberg, G.L., 
McElroy, W.K., Hein, J.F. (1991) Suppression of the luteinizing 
hormone surge by chlordimeform in ovariectomized, steroid-primed 
female rats. Pharmacol. Toxicol. 68:131-136.
    Gorski, R.A. (1979) The neuroendocrinology of reproduction: an 
overview. Biol. Reprod. 20:111-127.
    Gorski, R.A. (1986) Sexual differentiation of the brain: a model 
for drug-induced alterations of the reproductive system. Environ. 
Health Perspect. 70:163-175.
    Gray, L.E. (1991) Delayed effects on reproduction following 
exposure to toxic chemicals during critical periods of development. 
In: Cooper, R.L., Goldman, J.M., Harbin, T.J. Aging and 
Environmental Toxicology: Biological and Behavioral Perspectives. 
Johns Hopkins University Press, Baltimore. pp. 183-210.
    Gray, L.E., Ostby, J.S. (1995) In utero 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) alters reproductive morphology 
and function in female rat offspring. Toxicol. Appl. Pharmacol. 
133:285-294.
    Gray, L.E., Ferrell, J.M., Ostby, J.S. (1985) Alteration of 
behavioral sex differentiation by exposure to estrogenic compounds 
during a critical neonatal period: effects of zearalenone, 
methoxychlor, and estradiol in hamster. Toxicol. Appl. Pharmacol. 
80:127-136.
    Gray, L.E., Ostby, J., Sigmon, R., Ferrell, J., Linder, R., 
Cooper, R., Goldman, J., Laskey, J. (1988) The development of a 
protocol to assess reproductive effects of toxicants in the rat. 
Reprod. Toxicol. 2:281-287.
    Gray, L.E., Ostby, J., Ferrell, J., Rehnberg, G., Linder, R., 
Cooper, R., Goldman, J., Slott, V., Laskey, J. (1989) A dose-
response analysis of methoxychlor-induced alterations of 
reproductive development and function in the rat. Fundam. Appl. 
Toxicol. 12:92-108.
    Gray, L.E., Ostby, J., Linder, R., Goldman, J., Rehnberg, G., 
Cooper, R. (1990) Carbendazim-induced alterations of reproductive 
development and function in the rat and hamster. Fundam. Appl. 
Toxicol. 15:281-297.
    Gray, L.E., Ostby, J.S., Kelce, W.R. (1994) Developmental 
effects of an environmental antiandrogen: the fungicide vinclozolin 
alters sex differentiation of the male rat. Toxicol. Appl. 
Pharmacol. 129:46-52.
    Gray, L.E., Kelce, W.R., Monosson, E., Ostby, J.S., Birnbaum, 
L.S. (1995) Exposure to TCDD during development permanently alters 
reproductive function in male Long Evans rats and hamsters: reduced 
ejaculated and epididymal sperm numbers and sex accessory gland 
weights in offspring with normal androgenic status. Toxicol. Appl. 
Pharmacol. 131:108-118.
    Green, S., Auletta, A., Fabricant, R., Kapp, M., Sheu, C., 
Springer, J., Whitfield, B. (1985) Current status of bioassays in 
genetic toxicology: the dominant lethal test. Mutat. Res. 154:49-67.

[[Page 56316]]

    Greenland, S. (1987) Quantitative methods in the review of 
epidemiologic literature. Epidemiol. Rev. 9:1-30.
    Gulati, D.K., Hope, E., Teague, J., Chapin, R.E. (1991) 
Reproductive toxicity assessment by continuous breeding in Sprague-
Dawley rats: a comparison of two study designs. Fundam. Appl. 
Toxicol. 17:270-279.
    Gustafsson, J.-A., Mode, A., Norstedt, G., Hokfelt, T., 
Sonnenschein, C., Eneroth, P., Skett, P. (1980) The hypothalamo-
pituitary-liver axis: a new hormonal system in control of hepatic 
steroid and drug metabolism. Biochem. Act. Hormones 14:47-89.
    Habicht, F.H. (1992) Guidance on risk characterization for risk 
managers and risk assessors. U.S. EPA, Memorandum to Assistant 
Administrators and Regional Administrators, February 26, 1992. 
Available from the EPA Air docket.
    Hales, B., Crosman, K., Robaire, B. (1992) Increased post-
implantation loss and malformations among the F2 progeny of male 
rats chronically treated with cyclophosphamide. Teratology 45:671-
678.
    Harris, M.W., Chapin, R.E., Lockhart, A.C., Jokinen, M.P., 
Allen, J.D., Haskins, E.A. (1992) Assessment of a short-term 
reproductive and developmental toxicity screen. Fundam. Appl. 
Toxicol. 19:186-196.
    Harris, R.Z., Benet, L.Z., Schwartz, J.B. (1995) Gender effects 
in pharmacokinetics and pharmacodynamics. Drugs 50:222-239.
    Harrison, P.T.C., Humfrey, C.D.N., Litchfield, M., Peakall, D., 
Shuker, L.K. (1995) IEH Assessment on Environmental Oestrogens: 
Consequences to Human Health and Wildlife. MRC Institute for 
Environment and Health. Leicester, UK.
    Haschek, W.M., Rousseaux, C.G. (1991) Handbook of Toxicologic 
Pathology. Academic Press, New York.
    Hatch, M., Kline, J. (1981) Spontaneous abortion and exposure to 
the herbicide 2,4,5-T: a pilot study. U.S. Environmental Protection 
Agency, Washington, D.C. EPA-560/6-81-006.
    Hattis, D., Silver, K. (1994) Human interindividual variability: 
a major source of uncertainty in assessing risks for noncancer 
health effects. Risk Analysis 14:421-431.
    Heindel, J.J., Chapin, R.E. (1993) Methods in Toxicology: Female 
Reproductive Toxicology. Academic Press, San Diego.
    Heindel, J.J., Thomford, P.J., Mattison, D.R. (1989) 
Histological assessment of ovarian follicle number in mice as a 
screen of ovarian toxicity. In: Hirshfield, A.N. Growth Factors and 
the Ovary. Plenum Press, New York. pp. 421-426.
    Hemminki, K., Vineis, P. (1985) Extrapolation of the evidence on 
teratogenicity of chemicals between humans and experimental animals: 
chemicals other than drugs. Teratogenesis Carcinog. Mutagen. 5:251-
318.
    Hemminki, K., Mutanen, P., Luoma, K., Saloniemi, I. (1980) 
Congenital malformations by the parental occupation in Finland. Int. 
Arch. Occup. Environ. Health 46:93-98.
    Hemminki, K., Saloniemi, I., Salonen, T. (1981) Childhood cancer 
and paternal occupation in Finland. J. Epidemiol. Community Health 
35:11-15.
    Hertig, A.T. (1967) The overall problem in man. In: Benirschke, 
K. Comparative Aspects of Reproductive Failure. Springer-Verlag, New 
York. pp. 11-41.
    Hervey, E., Hervey, G.R. (1967) The effects of progesterone on 
body weight and composition in the rat. J. Endocrinol. 37:361-384.
    Hess, R.A. (1990) Quantitative and qualitative characteristics 
of the stages and transitions in the cycle of the rat seminiferous 
epithelium: light microscopic observations of perfusion-fixed and 
plastic-embedded testes. Biol. Reprod. 43:525-542.
    Hess, R.A., Moore, B.J. (1993) Histological methods for 
evaluation of the testis. In: Chapin, R.E., Heindel, J.J. Methods in 
Toxicology: Male Reproductive Toxicology. Academic Press, San Diego. 
pp. 52-85.
    Hess, R.A., Moore, B.J., Forrer, J., Linder, R.E., Abuel-Atta, 
A.A. (1991) The fungicide Benomyl (methyl 1-(butylcarbamoyl)-2-
benzimidazolecarbamate) causes testicular dysfunction by inducing 
the sloughing of germ cells and occlusion of efferent ductules. 
Fundam. Appl. Toxicol. 17:733-745.
    Heywood, R., James, R.W. (1985) Current laboratory approaches 
for assessing male reproductive toxicity. In: Dixon, R.L. 
Reproductive Toxicology. Raven Press, New York. pp. 147-160.
    Hogue, C.J.R. (1984) Reducing misclassification errors through 
questionnaire design. In: Lockey, J.E., Lemasters, G.K., Keye, W.R. 
Reproduction: the new frontier in occupational and environmental 
health research. Alan R. Liss, Inc., New York. pp. 81-97.
    Holloway, A.J., Moore, H.D.M., Foster, P.M.D. (1990a) The use of 
in vitro fertilization to detect reductions in the fertility of male 
rats exposed to 1,3-dinitrobenzene. Fundam. Appl. Toxicol. 14:113-
122.
    Holloway, A.J., Moore, H.D.M., Foster, P.M.D. (1990b) The use of 
rat in vitro fertilization to detect reductions in the fertility of 
spermatozoa from males exposed to ethylene glycol monomethyl ether. 
Reprod. Toxicol. 4:21-27.
    Holmes, R.L., Ball, J.N. (1974) The Pituitary Gland: A 
Comparative Account. Cambridge University Press, Cambridge.
    Huang, H.H., Meites, J. (1975) Reproductive capacity of aging 
female rats. Neuroendocrinology 17:289-295.
    Hugenholtz, A.P., Bruce, W.R. (1983) Radiation induction of 
mutations affecting sperm morphology in mice. Mutat. Res. 107:177-
185.
    Hughes, C.L. (1988) Phytochemical mimicry of reproductive 
hormones and modulation of herbivore fertility by phytoestrogens. 
Environ Health Perspect. 78:171-175.
    Hurtt, M.E., Zenick, H. (1986) Decreasing epididymal sperm 
reserves enhances the detection of ethoxyethanol-induced 
spermatotoxicity. Fundam. Appl. Toxicol. 7:348-353.
    Imagawa, W., Yang, J., Guzman, R., Nandi, S. (1994) Control of 
mammary gland development. In: Knobil, E., O'Neill, J.D. The 
Physiology of Reproduction. Raven Press, New York. pp. 1033-1063.
    International Conference on Harmonization of Technical 
Requirements of Pharmaceuticals for Human Use (1994) ICH harmonized 
tripartite guideline, Detection of Toxicity to Reproduction for 
Medicinal Products. FR 59(1831):48746-48752.
    Irvine, S., Cawood, E., Richardson, D., MacDonald, E., Aitken, 
J. (1996) Evidence of deteriorating semen quality in the United 
Kingdom: birth cohort study in 577 men in Scotland over 11 years. 
Br. Med. J. 312:467-471.
    Joffe, M. (1985) Biases in research on reproduction and women's 
work. Int. J. Epidemiol. 14:118-123.
    Jones, T.C., Mohr, U., Hunt, R.D. (1987) Genital System. 
Springer-Verlag, New York.
    Katz, D.F., Overstreet, J.W. (1981) Sperm motility assessment by 
videomicrography. Fertil. Steril. 35:188-193.
    Katz, D.F., Diel, L., Overstreet, J.W. (1982) Differences in the 
movement of morphologically normal and abnormal human seminal 
spermatozoa. Biol. Reprod. 26:566-570.
    Kavlock, R.J., Allen, B.C., Kimmel, C.A., Faustman, E.M. (1995) 
Dose-response assessment for developmental toxicology: benchmark 
doses for fetal weight changes. Fundam. Appl. Toxicol. 26:211-222.
    Kelce, W.R., Stone, C.R., Laws, S.C., Gray, L.E., Kemppainen, 
J.A., Wilson, E.M. (1995) Persistent DDT metabolite p,p'-DDE is a 
potent androgen receptor antagonist. Nature 375:581-585.
    Kesner, J.S., Wright, D.M., Schrader, S.M., Chin, N.W., Krieg, 
E.F. (1992) Methods of monitoring menstrual function in field 
studies: Efficacy of methods. Reprod. Toxicol. 6:385-400.
    Kimmel, C.A., Francis, E.Z. (1990) Proceedings of the workshop 
on the acceptability and interpretation of dermal developmental 
toxicity studies. Fundam. Appl. Toxicol. 14:386-398.
    Kimmel, C.A., Gaylor, D.W. (1988) Issues in qualitative and 
quantitative risk analysis for developmental toxicology. Risk 
Analysis 8:15-20.
    Kimmel, C.A., Holson, J.F., Hogue, C.J., Carlo, G.L. (1984) 
Reliability of experimental studies for predicting hazards to human 
development. National Center for Toxicological Research, Jefferson, 
AR. NCTR Technical Report for Experiment No. 6015.
    Kimmel, C.A., Kimmel, G.L., Frankos, V. (1986) Interagency 
Regulatory Liaison Group workshop on reproductive toxicity risk 
assessment. Environ. Health 66:193-221.
    Kimmel, C.A., Rees, D.C., Francis, E.Z. (1990) Proceedings of 
the workshop on the qualitative and quantitative comparability of 
human and animal developmental neurotoxicity. Neurotoxicol. Teratol. 
12:173-292.
    Kimmel, G.L., Clegg, E.D., Crisp, T.M. (1995) Reproductive 
toxicity testing: a risk assessment perspective. In: Witorsch, R.J. 
Reproductive Toxicology. Raven Press, New York. pp. 75-98.
    Kissling, G. (1981) A generalized model for analysis of 
nonindependent observations. Dissertation. University of North 
Carolina.
    Kleinbaum, D.G., Kupper, L.L., Morgenstern, H. (1982) 
Epidemiologic Research: Principle and Quantitative Methods. Lifetime 
Learning Publications, London.
    Kline, J., Stein, Z., Susser, M. (1989) Conception to Birth: 
Epidemiology of

[[Page 56317]]

Prenatal Development. Oxford University Press, New York.
    Klinefelter, G.R., Laskey, J.W., Kelce, W.R., Ferrell, J., 
Roberts, N.L., Suarez, J.D., Slott, V. (1994a) 
Chloroethylmethanesulfonate-induced effects on the epididymis seem 
unrelated to altered Leydig cell function. Biol. Reprod. 51:82-91.
    Klinefelter, G.R., Laskey, J.W., Perreault, S.D., Ferrell, J., 
Jeffay, S., Suarez, J., Roberts, N. (1994b) The ethane 
dimethanesulfonate-induced decrease in the fertilizing ability of 
cauda epididymal sperm is independent of the testis. J. Androl. 
15:318-327.
    Knobil, E., Neill, J.D., Greenwald, G.S., Markert, C.L., Pfaff, 
D.W. (1994) The Physiology of Reproduction. Raven Press, New York.
    Ku, W.W., Chapin, R.E., Wine, R.N., Gladen, B.C. (1993) 
Testicular toxicity of boric acid (BA): relationship of dose to 
lesion development and recovery in the F344 rat. Reprod. Toxicol. 
7:305-319.
    Kupfer, D. (1987) Critical evaluation of methods for detection 
and assessment of estrogenic compounds in mammals: strengths and 
limitations for application to risk assessment. Reprod. Toxicol. 
2:147-153.
    Kurman, R., Norris, H.J. (1978) Germ cell tumors of the ovary. 
Pathol. Annu. 13:291.
    Kwa, S.L., Fine, L.J. (1980) The association between parental 
occupation and childhood malignancy. J. Occup. Med. 22:792-794.
    La Bella, F.S., Dular, R., Lemons, P., Vivian, S., Queen, M. 
(1973a) Prolactin secretion is specifically inhibited by nickel. 
Nature 245:330-332.
    La Bella, F.S., Dular, R., Vivian, S., Queen, G. (1973b) 
Pituitary hormone releasing activity of metal ions present in 
hypothalamic extracts. Biochem. Biophys. Res. Commun. 52:786-791.
    Lamb, J.C. (1985) Reproductive toxicity testing: evaluating and 
developing new testing systems. J. Am. Coll. Toxicol. 4:163-171.
    Lamb, J.C., Chapin, R.E. (1985) Experimental models of male 
reproductive toxicology. In: Thomas, J.A., Korach, K.S., McLachlan, 
J.A. Endocrine Toxicology. Raven Press, New York. pp. 85-115.
    Lamb, J.C., Foster, P.M.D. (1988) Physiology and Toxicology of 
Male Reproduction. Academic Press, New York.
    Lamb, J.C., Jameson, C.W., Choudhury, H., Gulati, D.K. (1985) 
Fertility assessment by continuous breeding: evaluation of 
diethylstilbestrol and a comparision of results from two 
laboratories. J. Am. Coll. Toxicol. 4:173-183.
    Langley, F.A., Fox, H. (1987) Ovarian tumors. Classification, 
histogenesis, etiology. In: Fox, H. Haines and Taylor's Obstetrical 
and Gynaecologic Pathology. Churchill Livingstone, Edinburgh. pp. 
542-555.
    Lantz, G.D., Cunningham, G.R., Huckins, C., Lipshultz, L.I. 
(1981) Recovery from severe oligospermia after exposure to 
dibromochloropropane. Fertil. Steril. 35:46-53.
    LeFevre, J., McClintock, M.K. (1988) Reproductive senescence in 
female rats: a longitudinal study of individual differences in 
estrous cycles and behavior. Biol. Reprod. 38:780-789.
    Lemasters, G.K. (1992) Occupational exposures and effects on 
male and female reproduction. In: Rom, W.N. Environmental and 
Occupational Medicine. Little, Brown, Boston, MA. pp. 147-170.
    Lemasters, G.K., Pinney, S.M. (1989) Employment status as a 
confounder when assessing occupational exposures and spontaneous 
abortion. J. Clin. Epidemiol. 42:975-981.
    Lemasters, G.K., Selevan, S.G. (1984) Use of exposure data in 
occupational reproductive studies. Scan. J. Work. Environ. Health 
10:1-6.
    Lemasters, G.K., Selevan, S.G. (1993) Toxic exposures and 
reproduction: a view of epidemiology and surveillance. In: Scialli, 
A.R., Zinaman, M.J. Reproductive Toxicology and Infertility. McGraw-
Hill, New York. pp. 307-321.
    Leridon, H. (1977) Human Fertility: The Basic Components. The 
University of Chicago Press, Chicago.
    Le Vier, R.R., Jankowiak, M.E. (1972) The hormonal and 
antifertility activity of 2,6-cis-diphenylhexamethylcyclotetra-
siloxane in the female rat. Biol. Reprod. 7:260-266.
    Levine, R.J. (1983) Methods for detecting occupational causes of 
male infertility: reproductive history versus semen analysis. Scand. 
J. Work Environ. Health 9:371-376.
    Levine, R.J., Symons, M.J., Balogh, S.A., Arndt, D.M., 
Kaswandik, N.R., Gentile, J.W. (1980) A method for monitoring the 
fertility of workers: I. Method and pilot studies. J. Occup. Med. 
22:781-791.
    Levine, R.J., Symons, M.J., Balogh, S.A., Milby, T.H., Whorton, 
M.D. (1981) A method for monitoring the fertility of workers: II. 
Validation of the method among workers exposed to 
dibromochloropropane. J. Occup. Med. 23:183-188.
    Levine, R.J., Blunden, P.B., DalCorso, R.D., Starr, T.B., Ross, 
C.E. (1983) Superiority of reproductive histories to sperm counts in 
detecting infertility at a dibromochloropropane manufacturing plant. 
J. Occup. Med. 25:591-597.
    Lewis, J.R. (1991) Reproductively Active Chemicals: A Reference 
Guide. Van Nostrand Reinhold, New York.
    Lindbohm, M.L., Hemminki, K., Bonhomme, M.G., Anttila, A., 
Rantala, K., Keikkila, P., Rosenberg, M.J. (1991) Effects of 
paternal occupational exposure on spontaneous abortions. Am. J. 
Public Health 81:1029-1033.
    Linder, R.E., Hess, R.A., Strader, L.F. (1986) Testicular 
toxicity and infertility in male rats treated with 1,3-
dinitrobenzene. J. Toxicol. Environ. Health 19:477-489.
    Linder, R.E., Strader, L.F., Barbee, R.R., Rehnberg, G.L., 
Perreault, S.D. (1990) Reproductive toxicity of a single dose of 
1,3-dinitrobenzene in two ages of young adult male rats. Fundam. 
Appl. Toxicol. 14:284-298.
    Linder, R.E., Strader, L.F., Slott, V.L., Suarez, J.D. (1992) 
Endpoints of spermatotoxicity in the rat after short duration 
exposures to fourteen reproductive toxicants. Reprod. Toxicol. 
6:491-505.
    Lipshultz, L.I., Ross, C.E., Whorton, D., Thomas, M., Smith, R., 
Joyner, R.E. (1980) Dibromochloropropane and its effect on 
testicular function in man. J. Urol. 124:464-468.
    Liu, D.Y., Baker, H.W.G. (1992) Tests of human sperm function 
and fertilization in vitro. Fertil. Steril. 58:465-483.
    Loehr, R.A., Matanoski, G.M. (1993) Letter to Carol M. Browner, 
EPA Administrator, re: quantitative uncertainty analysis for 
radiological assessments. U.S. EPA Science Advisory Board, July 23, 
1993 (EPA-SAB-RAC-COM-93-006).
    Long, J.A., Evans, H.M. (1922) The oestrous cycle in the rat and 
its associated phenomena. Mem. Univ. Calif. 6:1-111.
    Mackeprang, M., Hay, S., Lunde, A.S. (1972) Completeness and 
accuracy of reporting of malformations on birth certificates. HSMHA 
Health Reports 84:43-49.
    Manson, J.M. (1994) Testing of pharmaceutical agents for 
reproductive toxicity. In: Kimmel, C.A., Buelke-Sam, J. 
Developmental Toxicology. Raven Press, New York. p. 379.
    Manson, J.M., Kang, Y.J. (1994) Test methods for assessing 
female reproductive and developmental toxicology. In: Hayes, A.W. 
Principles and Methods of Toxicology. Raven Press, New York. pp. 
989-1037.
    Mason, H.J. (1990) Occupational cadmium exposure and testicular 
endocrine function. Hum. Exp. Toxicol. 9:91-94.
    Mattison, D.R. (1985) Clinical manifestations of ovarian 
toxicity. In: Dixon, R.L. Reproductive Toxicology. Raven Press, New 
York. pp. 109-130.
    Mattison, D.R., Nightingale, M.R. (1980) The biochemical and 
genetic characteristics of murine ovarian aryl hydrocarbon 
(benzo(a)pyrene) hydroxylase activity and its relationship to 
primary oocyte destruction by polycyclic aromatic hydrocarbons. 
Toxicol. Appl. Pharmacol. 56:399-408.
    Mattison, D.R., Thomford, P.J. (1989) The mechanisms of action 
of reproductive toxicants. Toxicol. Pathol. 17:364-376.
    Mattison, D.R., Thorgeirsson, S.S. (1978) Gonadal aryl 
hydrocarbon hydroxylase in rats and mice. Cancer Res. 38:1368-1373.
    McDonald, A.D., McDonald, J.C., Armstrong, B., Cherry, N.M., 
Nolin, A.D., Robert, D. (1989) Father's occupation and pregnancy 
outcome. Br. J. Ind. Med. 46:329-333.
    McGregor, A.J., Mason, H.J. (1991) Occupational mercury vapour 
exposure and testicular, pituitary and thyroid endocrine function. 
Hum. Exp. Toxicol. 10:199-203.
    McKinney, J.D., Waller, C.L. (1994) Polychlorinated biphenyls as 
hormonally active structural analogues. Environ. Health Perspect. 
102:290-297.
    McLachlan, J.A. (1980) Estrogens in the Environment. Elsevier 
North Holland, New York.
    McMichael, A.J. (1976) Standardized mortality ratios and the 
healthy worker effect: scratching beneath the surface. J. Occup. 
Med. 18:165-168.
    McNatty, K.P. (1979) Follicular determinants of corpus luteum 
function in the human ovary. Adv. Exp. Med. Biol. 112:465-481.
    Meistrich, M.L. (1982) Quantitative correlation between 
testicular stem cell survival, sperm production, and fertility in 
the mouse after treatment with different cytotoxic agents. J. 
Androl. 3:58-68.

[[Page 56318]]

    Meistrich, M.L. (1986) Critical components of testicular 
function and sensitivity to disruption. Biol. Reprod. 34:17-28.
    Meistrich, M.L., Brown, C.C. (1983) Estimation of the increased 
risk of human infertility from alterations in semen characteristics. 
Fertil. Steril. 40:220-230.
    Meistrich, M.L., Samuels, R.C. (1985) Reduction in sperm levels 
after testicular irradiation of the mouse: a comparison with man. 
Radiat. Res. 102:138-147.
    Meistrich, M.L., van Beek, M.E.A.B. (1993) Spermatogonial stem 
cells: assessing their survival and ability to produce 
differentiated cells. In: Chapin, R.E., Heindel, J.J. Methods in 
Toxicology: Male Reproductive Toxicology. Academic Press, San Diego. 
pp. 106-123.
    Meyer, C.R. (1981) Semen quality in workers exposed to carbon 
disulfide compared to a control group from the same plant. J. Occup. 
Med. 23:435-439.
    Milby, T.H., Whorton, D. (1980) Epidemiological assessment of 
occupationally related chemically induced sperm count suppression. 
J. Occup. Med. 22:77-82.
    Milby, T.H., Whorton, M.D., Stubbs, H.A., Ross, C.E., Joyner, 
R.E., Lipshultz, L.I. (1981) Testicular function among 
epichlorohydrin workers. Br. J. Ind. Med. 38:372-377.
    Morris, I.D., Bardin, C.W., Gunsalus, G., Ward, J.A. (1990) 
Prolonged suppression of spermatogenesis by oestrogen does not 
preserve the seminiferous epithelium in procarbazine-treated rats. 
Int. J. Androl. 13:180-189.
    Morrissey, R.E., Lamb, J.C., Schwetz, B.A., Teague, J.L., 
Morris, R.W. (1988a) Association of sperm, vaginal cytology, and 
reproductive organ weight data with results of continuous breeding 
reproduction studies in Swiss (CD-1) mice. Fundam. Appl. Toxicol. 
11:359-371.
    Morrissey, R.E., Schwetz, B.A., Lamb, J.C., Ross, M.D., Teague, 
J.L., Morris, R.W. (1988b) Evaluation of rodent sperm, vaginal 
cytology, and reproductive organ weight data from National 
Toxicology Program 13-week studies. Fundam. Appl. Toxicol. 11:343-
358.
    Morrissey, R.E., Lamb, J.C., Morris, R.W., Chapin, R.E., Gulati, 
D.K., Heindel, J.J. (1989) Results and evaluations of 48 continuous 
breeding reproduction studies conducted in mice. Fundam. Appl. 
Toxicol. 13:747-777.
    Mosher, W.D., Pratt, W.F. (1990) Fecundity and infertility in 
the United States, 1965-88. Report 192, National Center for Health 
Statistics, Hyattsville, MD.
    Mukhtar, H., Philpot, R.M., Lee, I.P., Bend, J.R. (1978) 
Developmental aspects of epoxide-metabolizing enzyme activities in 
adrenals, ovaries, and testes of the rat. In: Mahlum, D.D., Sikov, 
M.R., Hackett, P.L., Andrew, F.D. Developmental Toxicology of Energy 
Related Pollutants. Technical Information Center, U.S. Department of 
Energy, Springfield, VA. pp. 89-104.
    Na, J.Y., Garza, F., Terranova, P.F. (1985) Alterations in 
follicular fluid steroids and follicular hCG and FSH binding during 
atresia in hamsters. Proc. Soc. Exp. Biol. Med. 179:123-127.
    Nakai, M., Moore, B.J., Hess, R.A. (1993) Epithelial 
reorganization and irregular growth following carbendazim-induced 
injury of the efferent ductules of the rat testis. Anat. Rec. 
235:51-60.
    National Research Council (1977) Reproduction and teratogenicity 
tests. In: Principles and Procedures for Evaluating the Toxicity of 
Household Substances. National Academy Press, Washington, DC.
    National Research Council. (1983) Risk Assessment in the Federal 
Government: Managing the Process. National Academy Press, 
Washington, DC.
    National Research Council. (1989) Biologic Markers in 
Reproductive Toxicity. National Academy Press, Washington, DC.
    National Research Council. (1994) Science and Judgment in Risk 
Assessment. National Academy Press, Washington, DC.
    Nestor, A., Handel, M.A. (1984) The transport of morphologically 
abnormal sperm in the female reproductive tract of mice. Gamete Res. 
10:119-125.
    Nett, T.M. (1989) Hormonal evaluation of testicular function: 
species variation. J. Am. Coll. Toxicol. 8:539-549.
    Nisbet, I.C.T., Karch, N.J. (1983) Chemical hazards to human 
reproduction, Park Ridge, N.J., Noyes Data Corp.
    Oberlander, G., Yeung, C.H., Cooper, T.G. (1994) Induction of 
reversible infertility in male rats by oral ornidazole and its 
effects on sperm motility and epididymal secretions. J. Reprod. 
Fertil. 100:551-559.
    Organization for Economic Cooperation and Development (1983) 
First addendum to OECD guideline 415 for testing of chemicals, 
``One-Generation Rreproduction Toxicity''. OECD, Paris, pp. 1-8.
    Organization for Economic Cooperation and Development (1993a) 
Draft guidelines for testing chemicals: combined repeated dose 
toxicity study with the reproduction/developmental toxicity 
screening test. #422. OECD, Paris.
    Organization for Economic Cooperation and Development (1993b) 
First amendment to OECD guidelines 416, ``Two Generation 
Reproduction Toxicity''. OECD, Paris, pp. 1-8.
    Oskarsson, A., Hallen, I.P., Sundberg, J. (1995) Exposure to 
toxic elements via breast milk. Analyst 120:765-770.
    Pang, C.N., Zimmerman, E., Sawyer, C.H. (1977) Morphine 
inhibition of preovulatory surges of plasma luteinizing hormone and 
follicle stimulating hormone in the rat. Endocrinology 101:1726-
1732.
    Papier, C.M. (1985) Parental occupation and congenital 
malformations in a series of 35,000 births in Israel. Prog. Clin. 
Biol. Res. 163:291-294.
    Paul, M. (1993) Occupational and Environmental Reproductive 
Hazards. Williams and Wilkins, Baltimore.
    Paulsen, C.A., Berman, N.G., Wang, C. (1996) Data from men in 
greater Seattle area reveals no downward trend in semen quality: 
further evidence that deterioration of semen quality is not 
geographically uniform. Fertil. Steril. 65:1015-1020.
    Peluso, J.J., Bolender, D.L., Perri, A. (1979) Temporal changes 
associated with the degeneration of the rat oocyte. Biol. Reprod. 
20:423-430.
    Perreault, S.D. (1989) Impaired gamete function: implications 
for reproductive toxicology. In: Working, P.K. Toxicology of the 
Male and Female Reproductive Systems. Hemisphere, New York. pp. 217-
229.
    Perreault, S.D., Jeffay, S.C. (1993) Strategies and methods for 
the functional evaluation of oocytes and zygotes. In: Heindel, J.J., 
Chapin, R.E. Methods in Toxicology: Female Reproductive Toxicology. 
Academic Press, San Diego. pp. 92-109.
    Perreault, S.D., Jeffay, S., Poss, P., Laskey, J.W. (1992) Use 
of the fungicide carbendazim as a model compound to determine the 
impact of acute chemical exposure during oocyte maturation and 
fertilization on pregnancy outcome in the hamster. Toxicol. Appl. 
Pharmacol. 114:225-231.
    Peters, J.M., Preston-Martin, S., Yu, M.C. (1981) Brain tumors 
in children and occupational exposure of the parents. Science 
213:235-237.
    Plowchalk, D.R., Smith, B.J., Mattison, D.R. (1993) Assessment 
of toxicity to the ovary using follicle quantitation and 
morphometrics. In: Heindel, J.J., Chapin, R.E. Methods in 
Toxicology: Female Reproductive Toxicology. Academic Press, San 
Diego. pp. 57-68.
    Qiu, J., Hales, B.F., Robaire, B. (1995) Damage to rat 
spermatozoal DNA after chronic cyclophosphamide exposure. Biol. 
Reprod. 53:1465-1473.
    Ratcliffe, J.M., Clapp, D.E., Schrader, S.M., Turner, T.W., 
Oser, J., Tanaka, S., Hornung, R.W., Halperin, W.E. (1986) Semen 
quality in 2-ethoxyethanol-exposed workers. Health Hazard evaluation 
report, HETA 84-415-1688. Department of Health and Human Services, 
National Institute for Occupational Safety and Health, Cincinnati, 
Ohio.
    Ratcliffe, J.M., Schrader, S.M., Steenland, K., Clapp, D.E., 
Turner, T., Hornung, R.W. (1987) Semen quality in papaya workers 
with long term exposure to ethylene dibromide. Br. J. Ind. Med. 
44:317-326.
    Ratcliffe, J.M., Schrader, S.M., Clapp, D.E., Halperin, W.E., 
Turner, T.W., Horning, R.W. (1989) Semen quality in workers exposed 
to 2-ethoxyethanol. Br. J. Ind. Med. 46:399-406.
    Redi, C.A., Garagna, S., Pellicciari, C., Manfredi-Romanini, 
M.G., Capanna, E., Winking, H., Gropp, A. (1984) Spermatozoa of 
chromosomally heterozygous mice and their fate in male and female 
genital tracts. Gamete Res. 9:273-286.
    Rier, S.E., Martin, D.C., Bowman, R.E., Dmowski, W.P., Becker, 
J.L. (1993) Endometriosis in rhesus monkeys (Macaca mulatta) 
following chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. 
Fundam. Appl. Toxicol. 21: 433-441.
    Robaire, B., Smith, S., Hales, B.F. (1984) Suppression of 
spermatogenesis by testosterone in adult male rats: effect on 
fertility, pregnancy outcome and progeny. Biol. Reprod. 31:221-230.
    Rosenberg, M.J., Wyrobeck, A.J., Ratcliffe, J., Gordon, L.A., 
Watchmaker, G., Fox, S.H., Moore, D.H. (1985) Sperm as an indicator 
of reproductive risk among petroleum refinery workers. Br. J. Ind. 
Med. 42:123-127.
    Rothman, K.J. (1986) Modern epidemiology. Little, Brown, Boston.
    Rowland, A.S., Baird, D.D., Weinberg, C.R., Shore, D.L., Shy, 
C.M., Wilcox, A.J. (1992) Reduced fertility among women employed as 
dental assistants exposed to high levels of nitrous oxide. N. Engl. 
J. Med. 327:993-997.

[[Page 56319]]

    Rubin, H.B., Henson, D.E. (1979) Effects of drugs on male sexual 
function. In: Advances in Behavioral Pharmacology. Academic Press, 
New York. pp. 65-86.
    Russell, L.D. (1983) Normal testicular structure and methods of 
evaluation under experimental and disruptive conditions. In: 
Clarkson, T.W., Nordberg, G.F., Sager, P.R. Reproductive and 
Developmental Toxicity of Metals. Plenum Publishing Co., New York. 
pp. 227-252.
    Russell, L.D., Malone, J.P., McCurdy, D.S. (1981) Effect of 
microtubule disrupting agents, colchicine and vinblastine, on 
seminiferous tubule structure in the rat. Tissue Cell 13:349-367.
    Russell, L.D., Ettlin, R., Sinha Hikim, A.P., Clegg, E.D. (1990) 
Histological and Histopathological Evaluation of the Testis. Cache 
River Press, Clearwater, FL.
    Safe, S.H. (1995) Modulation of gene expression and endocrine 
response pathways by 2,3,7,8-tetrachlorodibenzo-p-dioxin and related 
compounds. Pharmacol. Ther. 67:247-281.
    Sakai, C.N., Hodgen, G.D. (1987) Use of primate folliculogenesis 
models in understanding human reproductive biology and applicablity 
to toxicology. Reprod. Toxicol. 1:207-222.
    Samuels, S.J. (1988) Lessons from a surveillance program of 
semen quality. Reprod. Toxicol. 2:229-231.
    Savitz, D.A., Harlow, S.D. (1991) Selection of reproductive 
health end points for environmental risk assessment. Environ. Health 
90:159-164.
    Savitz, D.A., Sonnenfeld, N.L., Olshan, A.F. (1994) Review of 
epidemiologic studies of paternal occupational exposure and 
spontaneous abortion. Am. J. Ind. Med. 25:361-383.
    Scala, R.A., Bevan, C., Beyer, B.K. (1992) An abbreviated repeat 
dose and reproductive/developmental toxicity test for high 
production volume chemicals. Regul. Toxicol. Pharmacol. 16:73-80.
    Schardein, J.L. (1993) Chemically Induced Birth Defects. Marcel 
Dekker, New York.
    Schrader, S.M., Chapin, R.E., Clegg, E.D., Davis, R.O., 
Fourcroy, J.L., Katz, D.F., Rothmann, S.A., Toth, G., Turner, T.W., 
Zinaman, M. (1992) Laboratory methods for assessing human semen in 
epidemiologic studies: a consensus report. Reprod. Toxicol. 6:275-
279.
    Schrag, S.D., Dixon, R.L. (1985a) Occupational exposures 
associated with male reproductive dysfunction. Ann. Rev. Pharmacol. 
Toxicol. 25:567-592.
    Schrag, S.D., Dixon, R.L. (1985b) Reproductive effects of 
chemical agents. In: Dixon, R.L. Reproductive Toxicology. Raven 
Press, New York. pp. 301-319.
    Schwetz, B.A., Rao, K.S., Park, C.N. (1980) Insensitivity of 
tests for reproductive problems. J. Environ. Pathol. Toxicol. 3:81-
98.
    Scialli, A.R., Clegg, E.D. (1992) Reversibility in Testicular 
Toxicity Assessment. CRC Press, Boca Raton.
    Scommegna, A., Vorys, N., Givens, J.R. (1980) Menstrual 
dysfunction. In: Gold, J.J., Josimovich, J.B. Gynecologic 
Endocrinology. Harper and Row, Hagerstown, MD.
    Seed, J., Chapin, R.E., Clegg, E.D., Darney, S.P., Dostal, L., 
Foote, R.H., Hurtt, M.E., Klinefelter, G.R., Makris, S.L., Schrader, 
S., Seyler, D., Sprando, R., Treinen, K.A., Veeranachaneni, R., 
Wise, L.D. (1996) Methods for assessing sperm motility, morphology, 
and counts in the rat, rabbit and dog: a consensus report. Reprod. 
Toxicol. 10:237-244.
    Selevan, S.G. (1980) Evaluation of data sources for occupational 
pregnancy outcome studies. Thesis. University of Cincinnati.
    Selevan, S.G. (1981) Design considerations in pregnancy outcome 
studies of occupational populations. Scand. J. Work Environ. Health 
7:76-82.
    Selevan, S.G. (1985) Design of pregnancy outcome studies of 
industrial exposure. In: Hemminki, K., Sorsa, M., Vainio, H. 
Occupational Hazards and Reproduction. Hemisphere, Washington, DC. 
pp. 219-229.
    Selevan, S.G. (1991) Environmental exposures and reproduction. 
In: Keily, M. Reproductive and Perinatal Epidemiology. CRC Press, 
Boca Raton. pp. 115-130.
    Selevan, S.G., Lemasters, G.K. (1987) The dose response fallacy 
in human reproductive studies of toxic exposure. J. Occup. Med. 
29:451-454.
    Selevan, S.G., Edwards, B., Samuels, S. (1982) Interview data 
from both parents on pregnancies and occupational exposures. How do 
they compare? Am. J. Epidemiol. 116:583.
    Sever, L.E., Hessol, N.A. (1984) Overall design considerations 
in male and female occupational reproductive studies. In: Lockey, 
J.E., Lemasters, G.K., Keye, W.R. Reproduction: The New Frontier in 
Occupational and Environmental Research. Alan R. Liss, Inc., New 
York. pp. 15-48.
    Sharpe, R.M. (1994) Regulation of spermatogenesis. In: Knobil, 
E., Neill, J.D. The Physiology of Reproduction. Raven Press, New 
York. pp. 1363-1434.
    Sheehan, D.M., Young, J.F., Slikker, W., Gaylor, D.W., Mattison, 
D.R. (1989) Workshop on risk assessment in reproductive and 
developmental toxicology: addressing the assumptions and identifying 
the research needs. Regul. Toxicol. Pharmacol. 10:110-122.
    Shepard, T.H. (1986) Human teratogenicity. Adv. Pediatrics 
33:225-268.
    Silverman, J., Kline, J., Hutzler, M. (1985) Maternal employment 
and the chromosomal characteristics of spontaneously aborted 
conceptions. J. Occup. Med. 27:427-438.
    Skett, P. (1988) Biochemical basis of sex differences in drug 
metabolism. Pharmacol. Ther. 38:269-304.
    Slott, V.L., Perreault, S.D. (1993) Computer-assisted sperm 
analysis of rodent epididymal sperm motility using the Hamilton-
Thorne motility analyzer. In: Chapin, R.E., Heindel, J.J. Methods in 
Toxicology: Male Reproductive Toxicology. Academic Press, San Diego. 
pp. 319-333.
    Slott, V.L., Suarez, J.D., Simmons, J.E., Perreault, S.D. (1990) 
Acute inhalation exposure to epichlorohydrin transiently decreases 
rat sperm velocity. Fundam. Appl. Toxicol. 15:597-606.
    Slott, V.L., Suarez, J.D., Perreault, S.D. (1991) Rat sperm 
motility analysis: methodologic considerations. Reprod. Toxicol. 
5:449-458.
    Slott, V.L., Jeffay, S.C., Suarez, J.D., Barbee, R.R., 
Perreault, S.D. (1995) Synchronous assessment of sperm motility and 
fertilizing ability in the hamster following treatment with alpha-
chlorohydrin. J. Androl. 16:523-535.
    Smith, C.G. (1983) Reproductive toxicity: hypothalamic-pituitary 
mechanisms. Am. J. Ind. Med. 4:107-112.
    Smith, C.G., Gilbeau, P.M. (1985) Drug abuse effects on 
reproductive hormones. In: Thomas, J.A., Korach, K.S., McLachlan, 
J.A. Endocrine Toxicology. Raven Press, New York. pp. 249-267.
    Smith, E.R., Davidson, J.M. (1974) Luteinizing hormone releasing 
factor in rats exposed to constant light: effects of mating. 
Neuroendocrinology 14:129-138.
    Smith, S.K., Lenton, E.A., Landgren, B.M., Cooke, I.D. (1984) 
The short luteal phase and infertility. Br. J. Obstet. Gynaecol. 
91:1120-1122.
    Snell, L.M., Little, B.B., Knoll, K.A., Johnston, W.L., et al. 
(1992) Reliability of birth certificate reporting of congenital 
anomalies. Am. J. Perinatol. 9:219-222.
    Sonawane, B.R. (1995) Chemical contaminants in human milk: an 
overview. Environ. Health. Perspect. 103:197-205.
    Sonawane, B.R., Yaffe, S.J. (1983) Delayed effects of drug 
exposure during pregnancy: reproductive function. Biol. Res. 
Pregnancy 4:48-55.
    Starr, T.B., Dalcorso, R.D., Levine, R.J. (1986) Fertility of 
workers: a comparision of logistic regression and indirect 
standardization. Am. J. Epidemiol. 123:490-498.
    Stein, A. and Hatch, M. (1987) Biological markers in 
reproductive epidemiology: prospects and precautions. Environ. 
Health 74:67-75.
    Stein, Z., Kline, J., Shrout, P. (1985) Power in surveillance. 
In: Hemminki, K., Sorsa, M., Vainio, H. Occupational hazards and 
reproduction. Hemisphere, Washington, DC. pp. 203-208.
    Steinberger, E., Lloyd, J.A. (1985) Chemicals affecting the 
development of reproductive capacity. In: Dixon, R.L. Reproductive 
Toxicology. Raven Press, New York.
    Stevens, K.R., Gallo, M.A. (1989) Practical considerations in 
the conduct of chronic toxicity studies. In: Hayes, A.W. Principles 
and Methods of Toxicology. Raven Press, New York. pp. 237-250.
    Stiratelli, R., Laird, N., Ware, J.H. (1984) Random-effects 
models for serial observations with binary responses. Biometrics 
40:961-971.
    Stoker, T.E., Goldman, J.M., Cooper, R.L. (1993) The 
dithiocarbamate fungicide thiram disrupts the hormonal control of 
ovulation in the female rat. Reprod. Toxicol. 7:211-218.
    Sweeney, A.M., Meyer, M.R., Aarons, J.H., Mills, J.L., LaPorte, 
R.E. (1988) Evaluation of methods for the prospective identification 
of early fetal losses in environmental epidemiology studies. Am. J. 
Epidemiol. 127:843-850.
    Tanaka, S., Kawashima, K., Naito, K., Usami, M., Nakadate, M., 
Imaida, K., Takahashi, M., Hayashi, Y., Kurokawa, Y., Tobe, M. 
(1992) Combined repeat dose and reproductive/developmental toxicity 
screening test (OECD): familiarization using cyclophosphamide. 
Fundam. Appl. Toxicol. 18:89-95.

[[Page 56320]]

    Terranova, P.F. (1980) Effects of phenobarbital-induced 
ovulatory delay on the follicular population and serum levels of 
steroids and gonadotropins in the hamster: a model for atresia. 
Biol. Reprod. 23:92-99.
    Thomas, J.A. (1981) Reproductive hazards and environmental 
chemicals: a review. Toxic Subst. J. 2:318-348.
    Thomas, J.A. (1991) Toxic responses of the reproductive system. 
In: Amdur, M.O., Doull, J., Klaassen, C.D. Casarett and Doull's 
Toxicology. Pergamon Press, New York. pp. 484-520.
    Tilley, B.C., Barnes, A.B., Bergstrahl, E., Labarthe, D., 
Noller, K.L., Colton, T., Adam, E. (1985) A comparision of pregnancy 
history recall and medical records: implications for retrospective 
studies. Am. J. Epidemiol. 121:269-281.
    Toppari, J., Larsen, J.C., Christiansen, P., Giwercman, A., 
Grandjean, P., Guillette, L.J., Jegou, B., Jensen, T.K., Jouannet, 
P., Keiding, N., Leffers, H., McLachlan, J.A., Meyer, O., Muller, 
J., Rajpert-De Meyts, E., Scheike, T., Sharpe, R., Sumpter, J., 
Skakkebaek, N. (1995) Male Reproductive Health and Environmental 
Chemicals with Estrogenic Effects. Miljoprojekt nr. 290. Danish 
Environmental Protection Agency.
    Toth, G.P., Stober, J.A., Read, E.J., Zenick, H., Smith, M.K. 
(1989a) The automated analysis of rat sperm motility following 
subchronic epichlorohydrin administration: methodologic and 
statistical considerations. J. Androl. 10:401-415.
    Toth, G.P., Zenick, H., Smith, M.K. (1989b) Effects of 
epichlorohydrin on male and female reproduction in Long-Evans rats. 
Fundam. Appl. Toxicol. 13:16-25.
    Toth, G.P., Stober, J.A., George, E.L., Read, E.J., Smith, M.K. 
(1991a) Sources of variation in the computer-assisted motion 
analysis of rat epididymal sperm. Reprod. Toxicol. 5:487-495.
    Toth, G.P., Stober, J.A., Zenick, H., Read, E.J., Christ, S.A., 
Smith, M.K. (1991b) Correlation of sperm motion parameters with 
fertility in rats treated subchronically with epichlorohydrin. J. 
Androl. 12:54-61.
    Toth, G.P., Wang, S.R., McCarthy, H., Tocco, D.R., Smith, M.K. 
(1992) Effects of three male reproductive toxicants on rat cauda 
epididymal sperm motion. Reprod. Toxicol. 6:507-515.
    Treloar, A.E., Boynton, R.E., Borghild, G.B., Brown, B.W. (1967) 
Variation in the human menstrual cycle through reproductive life. 
Int. J. Fertil. 12:77-126.
    Tsai, S.P., Wen, C.P. (1986) A review of methodological issues 
of the standardized mortality ratio (SMR) in occupational cohort 
studies. Int. J. Epidemiol. 15:8-21.
    Tucker, H.A. (1994) Lactation and its hormonal control. In: 
Knobil, E., O'Neill, J.D. The Physiology of Reproduction. Raven 
Press, New York. pp. 1065-1098.
    Tyl, R.W. (1987) Developmental toxicity in toxicologic research 
and testing. In: Ballantyne, B. Perspectives in Basic and Applied 
Toxicology. John Wright, Bristol. pp. 203-208.
    U.S. Congress (1985) Reproductive Health Hazards in the 
Workplace. Office of Technology Assessment, OTA-BA-266, U.S. 
Government Printing Office, Washington, DC.
    U.S. Congress (1988) Infertility: Medical and Social Choices. 
Office of Technology Assessment, OTA-BA-358, U.S. Government 
Printing Office, Washington, DC.
    U.S. Environmental Protection Agency (1982) Reproductive and 
Fertility Effects. Pesticide Assessment Guidelines, Subdivision F. 
Hazard Evaluation: Human and Domestic Animals. Office of Pesticides 
and Toxic Substances, Washington, D.C. EPA-540/9-82-025.
    U.S. Environmental Protection Agency (1985a) Hazard Evaluation 
Division Standard Evaluation Procedure. Teratology Studies. Office 
of Pesticide Programs, Washington, DC. pp. 22-23.
    U.S. Environmental Protection Agency (1985b) Toxic Substances 
Control Act Test Guidelines: Final Rules. Federal Register 50 
(188):39426-39436.
    U.S. Environmental Protection Agency (1986a) Guidelines for 
Carcinogen Risk Assessment. Federal Register. 51(185):33992-34003.
    U.S. Environmental Protection Agency (1986b) Guidelines for 
Estimating Exposures. Federal Register 51(185):34042-34054.
    U.S. Environmental Protection Agency (1986c) Guidelines for 
Mutagenicity Risk Assessment. Federal Register 51(185):34006-34012.
    U.S. Environmental Protection Agency (1987) Reference Dose 
(RfD): Description and Use in Health Risk Assessments. Integrated 
Risk Information System (IRIS): Appendix A. Integrated Risk 
Information System Documentation, Vol. 1. EPA/600/8-66/032a.
    U.S. Environmental Protection Agency. (1991) Guidelines for 
Developmental Toxicity Risk Assessment. Federal Register 
56(234):63798-63826.
    U.S. Environmental Protection Agency (1992) Guidelines for 
Exposure Assessment. Federal Register 57(104):22888-22938.
    U.S. Environmental Protection Agency (1995a) Proposed Guidelines 
for Neurotoxicity Risk Assessment. Federal Register 60(192):52032-
52056.
    U.S. Environmental Protection Agency (1995b) The Use of the 
Benchmark Dose Approach in Health Risk Assessment. EPA/630/R-94/007.
    U.S. Environmental Protection Agency (1996a) Health Effects Test 
Guidelines OPPTS 870.3800: Reproduction and Fertility Effects 
(Draft). Federal Register 61(43):8282-8283.
    U.S. Environmental Protection Agency (1996b) Proposed Guidelines 
for Carcinogen Risk Assessment. Federal Register 61(79):17960-18011.
    U.S. Environmental Protection Agency (1996c) Benchmark Dose 
Technical Guidance Document. EPA/600/P-96/002A.
    Van Waeleghem, K., De Clerq, N., Vermeulen, L., Schoonjans, F., 
Comhaire, F. (1996) Deterioration of sperm quality in young healthy 
Belgian men. Hum. Reprod. 11:325-329.
    Vierula, M., Niemi, M., Keiski, A., Saarikoski, M., Suominen, J. 
(1996) High and unchanged sperm counts of Finnish men. Int. J. 
Androl. 19:11-17.
    Wade, G.N. (1972) Gonadal hormones and behavioral regulation of 
body weight. Physiol. Behav. 8:523-534.
    Walker, R.F. (1986) Age factors potentiating drug toxicity in 
the reproductive axis. Environ. Health 70:185-191.
    Walker, R.F., Schwartz, L.W., Manson, J.M. (1988) Ovarian 
effects of an anti-inflammatory-immunomodulatory drug in the rat. 
Toxicol. Appl. Pharmacol. 94:266-275.
    Waller, D.P., Killinger, J.M., Zaneveld, L.J.D. (1985) 
Physiology and toxicology of the male reproductive tract. In: 
Thomas, J.A., Korach, K.S., McLachlan, J.A. Endocrine Toxicology. 
Raven Press, New York. pp. 269-333.
    Wang, G.H. (1923) The relation between the ``spontaneous'' 
activity and the oestrous cycle in the white rat. Comp. Psychol. 
Monographs 2:1-27.
    Warren, J.C., Cheatum, S.G., Greenwald, G.S., Barker, K.L. 
(1967) Cyclic variation of uterine metabolic activity in the golden 
hamster. Endocrinology. 80:714-718.
    Weinberg, C.R., Gladen, B.C. (1986) The beta-geometric 
distribution applied to comparative fecundability studies. 
Biometrics 42:547-560.
    Weinberg, C.R., Baird, D.D., Wilcox, A.J. (1994) Sources of bias 
in studies of time to pregnancy. Stat. Med. 13:671-681.
    Weir, P.J., Rumberger, D. (1995) Isolation of rat sperm from the 
vas deferens for sperm motion analysis. Reprod. Toxicol. 9:327-330.
    Welch, L.S., Schrader, S.M., Turner, T.W., Cullen, M.R. (1988) 
Effects of exposure to ethylene glycol ethers on shipyard painters: 
II. Male reproduction. Am. J. Ind. Med. 14:509-526.
    Welch, L.S., Plotkin, E., Schrader, S. (1991) Indirect fertility 
analysis in painters exposed to ethylene glycol ethers: sensitivity 
and specificity. Am. J. Ind. Med. 20:229-240.
    Whorton, D., Milby, T.H. (1980) Recovery of testicular function 
among DBCP workers. J. Occup. Med. 22:177-179.
    Whorton, D., Krauss, R.M., Marshall, S., Milby, T.H. (1977) 
Infertility in male pesticide workers. Preliminary communication. 
Lancet 2(8051):1259-1261.
    Whorton, D., Milby, T.H., Krauss, R.M., Stubbs, H.A. (1979) 
Testicular function in DBCP exposed pesticide workers. J. Occup. 
Med. 21:161-166.
    Wilcox, A.J. (1983) Surveillance of pregnancy loss in human 
populations. Am. J. Ind. Med. 4:285-291.
    Wilcox, A.J., Weinburg, C.R., Wehmann, R.E., Armstrong, E.G., 
Canfield, R.E., Nisula, B.C. (1985) Measuring early pregnancy loss: 
laboratory and field methods. Fertil. Steril. 44:366-374.
    Wilcox, A.J., Weinberg, C.R., O'Connor, J.F., Baird, D.D., 
Schlatterer, J.P., Canfield, R.E., Armstrong, E.G., Nisula, B.C. 
(1988) Incidence of early pregnancy loss. N. Engl. J. Med. 319:189-
194.
    Williams, J., Gladen, B.C., Schrader, S.M., Turner, T.W., 
Phelps, J.L., Chapin, R.E. (1990) Semen analysis and fertility 
assessment in rabbits: statistical power and design considerations 
for toxicology studies. Fundam. Appl. Toxicol. 15:651-665.
    Wilson, J.G. (1973) Environment and Birth Defects. Academic 
Press, New York.
    Wilson, J.G. (1977) Embryotoxicity of drugs in man. In: Wilson, 
J.G., Fraser, F.C. Handbook of Teratology. Plenum Press, New York. 
pp. 309-355.
    Wilson, J.G., Scott, W.J., Ritter, E.J., Fradkin, R. (1975) 
Comparative distribution

[[Page 56321]]

and embryotoxicity of hydroxyurea in pregnant rats and rhesus 
monkeys. Teratology 11:169-178.
    Wilson, J.G., Ritter, E.J., Scott, W.J., Fradkin, R. (1977) 
Comparative distribution and embryotoxicity of acetylsalicylic acid 
in pregnant rats and rhesus monkeys. Toxicol. Appl. Pharmacol. 
41:67-78.
    Witorsch, R.J. (1995) Reproductive Toxicology. Raven Press, New 
York.
    Wolff, M.S. (1993) Lactation. In: Paul, M. Occupational and 
Environmental Reproductive Hazards. Williams and Wilkins, Baltimore. 
pp. 60-75.
    Wong, O., Utidjian, H.M.D., Karten, V.S. (1979) Retrospective 
evaluation of reproductive performance of workers exposed to 
ethylene dibromide. J. Occup. Med. 21:98-102.
    Working, P.K. (1988) Male reproductive toxicity: comparison of 
the human to animal models. Environ. Health 77:37-44.
    Working, P.K. (1989) Toxicology of the Male and Female 
Reproductive Systems. Hemisphere, New York.
    World Health Organization (1992) WHO Laboratory Manual for the 
Examination of Human Semen and Sperm-Cervical Mucus Interaction. 
Third edition. Cambridge University Press, Cambridge.
    Wright, D.M., Kesner, J.M., Schrader, S.M., Chin, N.W., Wells, 
V.E., Krieg, E.F. (1992) Methods of monitoring menstrual function in 
field studies: attitudes of working women. Reprod. Toxicol. 6:401-
409.
    Wyrobek, A.J. (1982) Sperm assays as indicators of chemically-
induced germ cell damage in man. In: Mutagenicity: New Horizons in 
Genetic Toxicology. Academic Press, New York. pp. 337-349.
    Wyrobek, A.J. (1984) Identifying agents that damage human 
spermatogenesis: abnormalities in sperm concentration and 
morphology. In: Monitoring human exposure to carcinogenic and 
mutagenic agents. Proceedings of a joint symposium held in Espoo, 
Finland. Dec. 12-15, 1983. International Agency for Research on 
Cancer, Lyon, France.
    Wyrobek, A.J., Bruce, W.R. (1978) The induction of sperm-shape 
abnormalities in mice and humans. In: Hollander, A., de Serres, F.J. 
Chemical Mutagens: Principles and Methods for Their Detection. 
Plenum Press, New York.
    Wyrobek, A.J., Gordon, L.A., Burkhart, J.G., Francis, M.W., 
Kapp, R.W., Letz, G., Malling, H.V., Topham, J.C., Whorton, D.M. 
(1983a) An evaluation of the mouse sperm morphology test and other 
sperm tests in nonhuman mammals. Mutat. Res. 115:1-72.
    Wyrobek, A.J., Gordon, L.A., Burkhart, J.G., Francis, M.W., 
Kapp, R.W., Jr., Letz, G., Malling, H., V, Topham, J.C., Whorton, 
D.M. (1983b) An evaluation of human sperm as indicators of 
chemically induced alterations of spermatogenic function. Mutat. 
Res. 115:73-148.
    Wyrobek, A.J., Watchmaker, G., Gordon, L. (1984) An evaluation 
of sperm tests as indicators of germ-cell damage in men exposed to 
chemical or physical agents. In: Lockey, J.E., Lemasters, G.K., 
Keye, W.R. Reproduction: The New Frontier in Occupational and 
Environmental Health Research. Alan R. Liss, New York. pp. 385-407.
    Yeung, C.H., Oberlander, G., Cooper, T.G. (1992) 
Characterization of the motility of maturing rat spermatozoa by 
computer-aided objective measurement. J. Reprod. Fertil. 96:427-441.
    Zeger, S.L., Liang, K.Y. (1986) Longitudinal data analysis for 
discrete and continuous outcomes. Biometrics 42:121-130.
    Zenick, H., Blackburn, K., Hope, E., Baldwin, D.J. (1984) 
Evaluating male reproductive toxicity in rodents: a new animal 
model. Teratogenesis Carcinog. Mutagen. 4:109-128.
    Zenick, H., Clegg, E.D., Perreault, S.D., Klinefelter, G.R., 
Gray, L.E. (1994) Assessment of male reproductive toxicity: a risk 
assessment approach. In: Hayes, A.W. Principles and Methods of 
Toxicology. Raven Press, New York. pp. 937-988.
    Zinaman, M.J., Clegg, E.D., Brown, C.C., O'Connor, J., Selevan, 
S.G. (1996) Estimates of human fertility and pregnancy loss. Fertil. 
Steril. 65:503-509.
    Zuelke, K.A., Perreault, S.D. (1995) Carbendazim (MBC) disrupts 
oocyte spindle function and induces aneuploidy in hamsters exposed 
during fertilization (meiosis II). Mol. Reprod. Dev. 42:200-209.

Part B. Response to Science Advisory Board and Public Comments

I. Introduction

    A notice of availability for public comment of these Guidelines was 
published in the Federal Register (FR) in February 1994. Seven 
responses were received. These Guidelines were presented to the 
Environmental Health Committee of the Science Advisory Board (SAB) on 
July 19, 1994. The report of the SAB was provided to the Agency in May 
1995, with further communication from the SAB Executive Committee 
provided in December 1995.
    The SAB and public comments were diverse and represented varying 
perspectives. Many of the comments were favorable and expressed 
agreement with positions taken in the proposed guidelines. A number of 
the comments addressed items that were more pertinent to testing 
guidance than risk assessment guidance or were otherwise beyond the 
scope of these Guidelines. Some of those were generic issues that are 
not system specific. Others were topics that have not been developed 
sufficiently and should be viewed as research issues. There were 
conflicting views about the need to provide additional detailed 
guidance about decision-making in the evaluation process as opposed to 
promoting extensive use of scientific judgment. Also, comments provided 
specific suggestions for clarification of details.

II. Response to Science Advisory Board Comments

    In general, the SAB found ``the overall scientific foundations of 
the draft guidelines' positions to be generally sound.'' However, 
recommendations were made to improve specific areas.
    The SAB recommended that EPA retain separate sections for 
identification and dose-response assessment in the draft guidelines. In 
subsequent meetings involving the SAB Executive Committee, members of 
the Clean Air Scientific Advisory Committee, and the Environmental 
Health Committee, this issue was explored further. After discussion, 
the SAB agreed with expanding the hazard identification to include 
certain components of the dose-response assessment. The resulting 
hazard characterization provides an evaluation of hazard within the 
context of the dose, route, timing, and duration of exposure. The next 
step, the dose-response analysis, quantitatively evaluates the 
relationship between dose or exposure and severity or probability of 
effect in humans. EPA has revised these Guidelines to reflect that 
position which is consistent also with the 1994 NRC report, Science and 
Judgment in Risk Assessment. The SAB suggested an alternative scheme 
for characterizing health effects data in Table 5. The Agency's intent 
for Table 5 is not to characterize the available data, but rather to 
judge whether the database is sufficient to proceed further in the risk 
assessment process. The text has been modified to clarify the intended 
use of this table and to ensure that it is consistent with the 
reorganization of the Guidelines into separate hazard characterization 
and quantitative dose-response analysis sections.
    The SAB supported the concept of using a gender neutral default 
assumption, but indicated that more discussion to support this 
assumption was needed. In particular, the Committee indicated that a 
fuller discussion is needed on ``information to the contrary'' (to 
obviate the need for making this default assumption), as well as 
additional guidance for using this and other default assumptions in 
risk characterization. The Agency agrees with this recommendation and 
provides further guidance on the use of the gender neutral default 
assumption. In keeping with recent Agency guidance on risk 
characterization, discussion on the use of default assumptions has been 
expanded in the risk characterization section of these Guidelines.
    The SAB in its reviews of the reproductive toxicity and 
neurotoxicity risk assessment guidelines discussed assumptions about 
the behavior of the dose-response curve. The SAB's advice has been that 
the Agency examine available data first, and only use

[[Page 56322]]

nonlinear behavior as a default if available data do not define the 
dose-response curve. The SAB also recommended that the benchmark dose 
method be considered as a possible alternative to the NOAEL/LOAEL 
approach. The Agency agrees.
    The SAB recommended that more discussion be devoted to the issue of 
disruption of endocrine systems by environmental agents. The section on 
Endocrine Evaluations has been expanded to include endocrine disruption 
of the reproductive system during development in addition to effects on 
adults.
    The SAB supported the principle in the Guidelines that more than 
one negative study is necessary to judge that a chemical is unlikely to 
pose a reproductive hazard. That principle has been retained and, as 
recommended by the SAB, an explicit statement included that data from a 
second species are necessary to determine that sufficient information 
is available to indicate that an agent is unlikely to pose a hazard.
    The SAB recommended that the topic of susceptible populations be 
expanded and that the Guidelines should indicate that relevant 
information be incorporated into risk assessments when possible. To 
address this issue, the Agency has emphasized potential differences in 
risks in children at different stages of development, females 
(including pregnant and lactating females), and males, and indicated 
that relevant information on differential risks for susceptible 
populations should be included in the risk characterization section 
when available. When specific information on differential risks is not 
available, the Agency will continue to apply a default uncertainty 
factor to account for potential differences in susceptibility.
    The SAB recommended that the Agency provide more specific guidance 
for exposure assessment issues that arise when characterizing exposure 
for reproductive toxicants. The Agency agrees and has indicated that an 
exposure assessment: include a statement of purpose, scope, level of 
detail, and approach used; present the estimate of exposure and dose by 
pathway and route for individuals, population segments, and populations 
in a manner appropriate for the intended risk characterization; and 
provide an evaluation of the overall level of confidence (including 
consideration of uncertainty factors) in the estimate of exposure and 
dose and the conclusions drawn. The SAB recommended that the MOE 
discussion be modified to address specific circumstances where the 
administered dose and the ``effective dose'' are known to be different. 
The discussion has been modified to emphasize that pharmacokinetic 
data, when available, be utilized to address such instances.
    The SAB recommended that the Agency expand substantially the 
discussion of overall strategy to evaluate exposure from mixtures, 
exposures to multiple single agents, and exposures to the same agent 
via different routes. It is anticipated that this type of information 
will be addressed in the Agency's upcoming revisions to the chemical 
mixture guidelines.

III. Response to Public Comments

    In addition to numerous supportive statements, several issues were 
indicated although each issue was raised by a very limited number of 
submissions. Use of the benchmark dose was supported along with the 
suggestion that the amount of text could be reduced on that subject. 
The text has been reduced and reference made to the report, The Use of 
the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 
1995b). A request was made for increased emphasis on paternally 
mediated effects on offspring. The text in that section has been 
expanded to provide additional discussion and references. Concern was 
expressed about the existence of constraints on the use of professional 
judgment in the risk assessment process, particularly in determining 
the relevance and sufficiency of the database, in evaluating biological 
plausibility of statistically different effects, and in the 
determination of uncertainty factors. Requests also have been made to 
provide additional criteria for when and under what conditions the risk 
assessment process will be used. These Guidelines emphasize the 
importance of using scientific judgment throughout the risk assessment 
process. They provide flexibility to permit EPA's offices and regions 
to develop specific guidance suited to their particular needs. The 
comment was made that the exposure assessment and risk characterization 
sections were not developed as well as the rest of the document. In 
1992, EPA published Guidelines for Exposure Assessment (U.S. EPA, 1992) 
that were intended to apply generically to noncancer risk assessments. 
These Guidelines only address aspects of exposure that are specific to 
reproduction and have been developed sufficiently. The risk 
characterization section has been expanded substantially to reflect the 
recent guidance provided within EPA for application in all risk 
assessments.

[FR Doc. 96-27473 Filed 10-30-96; 8:45 am]
BILLING CODE 6560-50-P