United States Science Advisory Board EPA-SAB-04-003
Environmental Staff Office (1400A) March 2004
Protection Agency Washington DC www.epa.gov/sab
Review of EPA's Draft
Supplemental Guidance For
Assessing Cancer
Susceptibility From Early-Life
Exposure to Carcinogens
A Report By
The Supplemental Guidance For Assessing
Cancer Susceptibility Review Panel
Of The EPA Science Advisory Board
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON D.C. 20460
OFFICE OF
THE ADMINISTRATOR
EPA SCIENCE ADVISORY BOARD
March 3, 2004
The Honorable Michael Leavitt
Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, N.W.
Washington, D.C. 20460
Subject: Review of EPA's Draft Supplemental Guidance for Assessing Cancer Susceptibility
from Early-Life Exposure to Carcinogens
Dear Administrator Leavitt:
A Review Panel of the EPA Science Advisory Board (SAB) met on May 12-14, 2003 to
review the Agency's Draft Supplemental Guidance for Assessing Cancer Susceptibility from
Early-Life Exposure to Carcinogens (Supplemental Guidance). The SAB Review Panel, known
as the Supplemental Guidance for Assessing Cancer Susceptibility (SGACS) Review Panel
(hereinafter, Review Panel), was composed of members of the SAB Environmental Health
Committee (EHC) and Radiation Advisory Committee (RAC) along with members of the Federal
Insecticide, Fungicide, and Rodenticide Act (FIFRA) Scientific Advisory Panel (SAP) and the
Children's Health Protection Advisory Committee (CHPAC).
The Supplemental Guidance represents an effort by the Agency to be responsive to the
previous SAB recommendations regarding the EPA's revision of the Guidelines for Carcinogenic
Risk Assessment. A key SAB recommendation was the consideration of age-dependent
susceptibility when assessing cancer risk. The Supplemental Guidance provides a proposed
approach for assessing cancer susceptibility from early-life exposure to carcinogens. The
Agency concludes that cancer risks generally were higher from early-life exposure to carcinogens
that act through a mutagenic mode of action than from similar exposure durations later in life.
Accordingly, in the absence of chemical specific data on early-life exposure, the Agency
proposes to use a default approach to account for differential susceptibility from early-life
exposure. Adjustments to the cancer slope factor typically derived from adult exposure will
depend on the age group:
• A 10-fold (lOx) adjustment for exposures before 2 years of age.
• A 3-fold (3x) adjustment for exposures between 2 and 15 years of age.
• No adjustment for exposures after 15 years of age.
-------
We appreciate the Agency's consideration of the SAB's previous recommendations. In
this review activity, the Agency sought the SAB's evaluation of the soundness of the Agency's
analysis of the underlying scientific information that supports the proposed guidance for
assessing cancer susceptibility from early-life exposures to carcinogens. The Review Panel
concurs with the Agency's conclusions and the overall approach adopted by the Agency of using
adjustment factors to account for increased susceptibility due to early-life exposure. The Review
Panel also agrees that the values chosen for the cancer slope adjustment factors in the
Supplemental Guidance appear to be reasonable from consideration of the literature. However,
the Review Panel suggests that the Agency improve the statistical analysis of the data and
provide a more extensive discussion of how the Agency arrived at the choice of the lOx and 3x
adjustment factors. The Review Panel also suggests that the Agency emphasize the use of
default adjustment factors only when no chemical-specific data are available to directly assess
cancer susceptibility from early-life exposure to a particular carcinogen. The Agency should
consider conducting additional research to address this issue as discussed in the report.
SUMMARY OF RECOMMENDATIONS
• The Review Panel agrees with the Agency that the science supports the conclusion that
early-life exposures result in increased susceptibility to carcinogens that act through a
mutagenic mode of action as compared to adult exposures. The Review Panel notes that a
broader look at the scientific literature beyond the studies included in the Supplemental
Guidance analysis would strengthen that conclusion.
• The Review Panel notes that for certain groups of non-mutagenic chemicals with known
modes of action (e.g., estrogen receptor agonist/antagonist) there is sufficient evidence
supporting increased susceptibility to cancer with early-life exposure. The Review Panel
suggests the Agency include a discussion of these agents in the Supplemental Guidance.
Non-mutagenic carcinogens with known modes of action should be assessed on a case-
by-case basis as suggested by the Agency.
• The Review Panel supports the use of slope factor adjustments in developing default
approaches. Application of an adjustment to the adult cancer slope factor seems to be the
most transparent and practical approach for risk assessment.
• The Review Panel reviewed age-specific human vulnerabilities and concludes that it
would be useful to include an additional age grouping (age 9 -15) to recognize the
potentially important vulnerabilities during puberty. Thus, four age groupings would be
appropriate (0-2, 3-8, 9-15, 15+) to represent critical periods of human growth and
development.
• The Review Panel suggests that the Agency consider alternative analyses that might allow
them to use more of the available data and directly test hypotheses concerning the
appropriateness of the adjustment values for predicting the dose-response from early-life
exposure.
• The Review Panel recommends that a priority for the near term would be the
development of mode of action approaches for endocrine disrupters, beginning with
estrogenic agents.
ii
-------
• The Review Panel cannot recommend at this time a feasible method for incorporating
transplacental or in utero exposure data. However, the Review Panel believes this to be
an important issue that requires further research.
• The Review Panel recommends that the Agency work more closely with the research
community to encourage the evaluation of early-life stage susceptibilities. For chemical
agents that are known to increase cancer risk, carcinogenic potency and the extent of
exposure should be used in deciding which chemicals to study first.
• Certain groups of non-mutagenic carcinogens with known modes of action serve as
important examples in support of applying a default factor to non-mutagenic carcinogens
when the mode of action is unknown. The Review Panel suggests that the Agency
reconsider limiting the application of adjustment factors only to mutagenic agents and
instead apply a default approach to both mutagenic and to non-mutagenic chemicals for
which mode of action remains unknown or insufficiently characterized.
In closing, the SAB appreciates the Agency's development of the Supplemental Guidance
as a stand-alone document. Because many parts of the Cancer Guidelines provide the
background for the Supplemental Guidance, issuance of the Supplemental Guidance before the
Guidelines could be confusing. The Review Panel encourages the Agency to rapidly finalize the
Guidelines, and the Supplemental Guidance soon after, if not concurrently. We wish to
commend the Agency for the hard work reflected in the Supplemental Guidance and look
forward to your response to this report.
Sincerely,
/Signed/ /Signed/
Dr. William Glaze, Chair Dr. Henry Anderson, Chair
EPA Science Advisory Board SGACS Review Panel
EPA Science Advisory Board
in
-------
NOTICE
This report has been written as part of the activities of the EPA Science Advisory Board, a public
advisory group providing extramural scientific information and advice to the Administrator and
other officials of the Environmental Protection Agency. The Board is structured to provide
balanced, expert assessment of scientific matters related to problems facing the Agency. This
report has not been reviewed for approval by the Agency and, hence, the contents of this report
do not necessarily represent the views and policies of the Environmental Protection Agency, nor
of other agencies in the Executive Branch of the Federal government, nor does mention of trade
names or commercial products constitute a recommendation for use. Reports of the EPA
Science Advisory Board are posted on the EPA website at http://www.epa.gov/sab.
IV
-------
U.S. Environmental Protection Agency
Science Advisory Board
Supplemental Guidance for Assessing Cancer
Susceptibility (SGACS) Review Panel
SCIENCE ADVISORY BOARD (SAB)
Dr. Henry Anderson (Review Panel Chair), Chief Medical Officer, Division of Public Health,
Wisconsin Division of Public Health, Madison, WI
Dr. James E. Klaunig, Professor and Director, Department of Pharmacology and Toxicology,
School of Medicine, Indiana University, Indianapolis, IN
Dr. Ulrike Luderer, Assistant Professor, Department of Medicine, Center for Occupational and
Environmental Health, University of California at Irvine, Irvine, CA
Dr. Anne Sweeney, Associate Professor, Department of Epidemiology/Biostatistics, Health
Science Center, School of Rural Public Health, Texas A&M University, Bryan, TX
Dr. Richard J. Vetter, Head, Radiation Safety Program, Mayo Medical School, Mayo Clinic,
Rochester, MN
CHILDREN'S HEALTH PROTECTION ADVISORY COMMITTEE (CHPAC)
Dr. Daniel A. Goldstein, Director, Medical Toxicology, Monsanto Company, St. Louis, MO
Dr. Melanie Marty, Chief, Air Toxicology and Epidemiology Section, California EPA Office of
Environmental Health Hazard Assessment, Oakland, CA
FEDERAL INSECTICIDE, FUNGICIDE, and RODENTICIDE ACT SCIENTIFIC
ADVISORY PANEL (SAP)
Dr. Stuart Handwerger, Director, Division of Endocrinology, Cincinnati Children's Hospital
Medical Center, University of Cincinnati, Cincinnati, OH
Dr. Steven G. Heeringa, Director, Statistical Design and Analysis, Institute for Social Research,
University of Michigan, Ann Arbor, MI
Dr. Christopher J. Portier, Director, Environmental Toxicology Program, National Institute of
Environmental Health Sciences, Research Triangle Park, NC
SCIENCE ADVISORY BOARD STAFF
Dr. Suhair Shallal, Designated Federal Officer, Washington, DC
-------
TABLE OF CONTENTS
1. Introduction 1
2. Charge to the Review Panel 3
3. Response to the Charge Questions 5
3.1. Response to Charge Question 1 5
3.2. Response to Charge Question 2 9
3.3. Response to Charge Question 3 12
3.4. Response to Charge Question 4 12
3.5. Response to Charge Question 5 15
3.6. Response to Charge Question 6 17
3.7. Response to Charge Question 7 19
3.8. Response to Charge Question 8 20
4. Miscellaneous Comments 22
5. References 25
Appendix 1: Suggested Additional Studies for Quantitative Analysis 31
Appendix 2: Initiation-Promotion Studies in Neonatal Mice 33
VI
-------
1. INTRODUCTION
In 1996, EPA published for public comment the Agency's proposed revisions to EPA's
1986 Guidelines for Carcinogen Risk Assessment (61 FR 17960, Apr. 23, 1996). In February
1997, the Science Advisory Board's (SAB) Environmental Health Committee (EHC) reviewed
the proposed revisions (http://www.epa.gov/sab/pdf/ehc9710.pdf). In January 1999, the SAB's
EHC met again to consider selected sections of the draft Guidelines that were revised to address
public comments and SAB recommendations on the 1996 proposed revisions. The revisions
included: new hazard descriptors and example narrative summaries; the expanded guidance on
the use of Mode of Action information; the use of departure points for the dose-response
analysis; and the approach to the Margin of Exposure analysis
(http://www.epa.gov/sab/pdf/ecl5.pdf). The SAB's EHC met for a third time in July 1999 to
provide advice and comment to the EPA on issues related to applying the provisions of EPA's
proposed revised guidelines for children (http://www.epa.gov/sab/pdf/ec0016.pdf). In that report
(p. 34), the SAB suggested, " Quantitatively analyzing the available experimental and
epidemiological literature on age dependence in carcinogenesis, in a comprehensive and
systematic review, would be very helpful." The SAB review suggested the possibility of
incorporating age-dependent susceptibility through age-specific adjustment factors for potency or
response to exposures.
In 2003, the Agency published the Draft Final Guidelines for Carcinogen Risk
Assessment (Cancer Guidelines) and Draft Supplemental Guidance for Assessing Cancer
Susceptibility from Early-Life Exposure to Carcinogens (Supplemental Guidance) (see USEPA,
2000a; USEPA 2000b). Concurrently, the Agency requested that the SAB conduct a peer review
of the Supplemental Guidance and utilize the expertise of two other EPA advisory committees,
the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) Scientific Advisory Panel
(SAP) and the Children's Health Protection Advisory Committee (CHPAC). By including
members of these three EPA advisory bodies in the review of this guidance, the Agency hoped to
benefit from their unique expertise in children's risk assessment.
The Supplemental Guidance recognizes that the standard methodology to calculate cancer
risk utilizes the lifetime average daily dose and accounts for differences between adults and
children with respect to exposure factors, such as eating habits and body weight. However,
susceptibility differences with respect to early-life stages are not currently taken into
consideration because the cancer slope factors are based on effects observed following adult
exposures. The purpose of the Supplemental Guidance is to provide a possible approach for
assessing cancer susceptibility from early-life exposure to carcinogens. Since a much larger
database exists for chemicals inducing cancer in adult humans or animals following mainly adult
exposures, an analysis was undertaken to determine if adjustment of adult-based cancer slope
factors would be appropriate when assessing cancer risks from exposures early in life. The
analysis undertaken addresses this issue, focusing upon studies that define the potential duration
and degree of increased susceptibility, if any, arising from childhood (or early postnatal and
juvenile animal) exposures.
According to the Supplemental Guidance, children's cancer risk includes early-life
exposures that may result in both the occurrence of cancer during childhood and cancers that
-------
occur later in life. The relative rarity of childhood cancers and a lack of animal testing guidelines
with perinatal exposure impede a full assessment of children's cancer risks from exposure to
chemicals in the environment. "Perinatal" was defined as the time around birth and may include
both prenatal (prior to birth) and postnatal (after birth) periods. The focus of the Supplemental
Guidance is on childhood exposures resulting in cancer later in life.
The analysis was conducted to ascertain whether there are quantitative scientific data that
would inform risk assessment policy choices for adjusting cancer slope factors based upon adult
human epidemiology or standard chronic adult rodent bioassays in the assessment of cancer risk
from childhood exposures. Thus, the critical data required are either human epidemiological data
on childhood exposures resulting in adult cancer or research studies with rodents involving early
postnatal exposures.
The Agency's review of the literature identified 21 studies (see Tables 4, 5, and 6 of the
Supplemental Guidance) that directly provided quantitative data on carcinogenesis following
early postnatal exposures and adult exposures to chemicals in animals. The carcinogenesis
studies utilized 16 chemicals. Studies included in this analysis were those that reported tumor
response from experiments that included both early-life and adult exposures. In addition, studies
were identified for five other chemicals that showed early life-stage sensitivity with early
postnatal exposure that were not evaluated quantitatively due to confounding factors related to
experimental design.
The major available human data on early-life exposures to mutagens are from
epidemiological studies on the effects of radiation, with very limited data available for humans
exposed during childhood to chemicals. A supporting role was assigned to the available human
radiation data, where cancer incidences in adults who were children at the time of the atomic
bomb (A-bomb) exposure were compared with cancer incidences in adults who were older at the
time of exposure. Although there are recognized differences in the mechanism between radiation
and mutagenic chemicals, the data on A-bomb survivors provide information in humans on many
different cancer sites with a single exposure involving all ages. In addition to the richness of the
data, a number of national and international committees of experts have analyzed and modeled
these data to develop risk estimates for various specific applications.
The Agency concluded that analysis of the available data supports higher cancer risks
from exposures to mutagenic carcinogens that occur early in life compared to the same exposures
during adulthood. Consequently, in the absence of early-life studies on a specific agent under
consideration, the Agency generally should use linear extrapolation to lower doses since
mutagens, based on mode-of-action data, can give rise to cancers with an apparently low-dose-
linear response. Risk estimates that pertain to childhood exposure should be adjusted since risk
estimates based on a lifetime-average daily dose do not consider the potential for higher cancer
risks from early-life exposure. The following adjustments to the cancer slope factor typically
derived from adult exposure represent a practical approach that reflects the results of the analysis
presented in the Supplemental Guidance, which concluded that cancer risks generally were
higher from early-life exposure than from similar exposure durations later in life:
-------
• For exposures before 2 years of age, a 10-fold adjustment.
• For exposures between 2 and 15 years of age, a 3-fold adjustment.
• For exposures after 15 years of age, no adjustment.
The draft Supplemental Guidance concludes that, with regard to modes of action other
than mutagenicity, there is insufficient information currently available to determine a general
adjustment; consequently, no general adjustment was recommended at this time even though the
available science indicates that higher cancer risks sometimes result from early-life exposure.
The Agency expects that as other modes of action become better understood, this information
will include data on quantitative differences between children and adults, and these differences
will be reflected in risk estimates for childhood exposure. The Agency expects to expand the
Supplemental Guidance to include other modes of action as they are understood and used in risk
assessments.
When the mode of action cannot be established, the current practice of using linear
extrapolation to lower doses such that risk estimates are based on a lifetime-average daily dose
without further adjustment should be continued and no general adjustment is recommended at
this time by the Agency. The result would be expected to produce risk estimates that generally
are protective, based on the use of linear extrapolation as a default in the absence of information
on the likely shape of the dose-response curve.
2. CHARGE TO THE REVIEW PANEL
The Agency sought the SAB's review of the soundness of the Agency's position that the
Agency's analysis and the underlying scientific information support the conclusion that there is
greater susceptibility for the development of tumors as a result of exposures in early life-stages
as compared with adults to chemicals acting through a mutagenic mode of action.
Question 1
Please comment on whether the Agency's analysis as applied to chemicals acting through a
mutagenic mode of action is accurate, reliable, unbiased and reproducible. Likewise, please
comment on whether the underlying scientific information used to develop the guidance is
accurate, reliable, unbiased and reproducible. Are there any key studies that the Agency has
overlooked in reaching this conclusion?
Question 2
For chemicals acting through non-mutagenic modes of action, the Agency concludes that a
range of approaches needs to be developed over time for addressing cancer risks from
childhood exposures. Please comment on the Agency's conclusion that the scientific
knowledge and data are insufficient at this time to develop generic guidance on how to
address these chemicals and that a case-by-case approach is more suitable. Is the SAB aware
of any additional data for chemicals acting through non-mutagenic modes of action relevant
to possible early life-stage sensitivity?
-------
Question 3
Assuming that it is appropriate to conclude that there is differential life-stage susceptibility to
chemicals acting through a mutagenic mode of action, the Agency's guidance uses a default
approach that adjusts cancer slope factors (typically from conventional animal bioassays
and/or epidemiologic studies of adult exposure) to address the impact of early life-stage
exposure. Please comment on whether the approach is justified by the available data? Can
the SAB suggest other approaches that might be equal or more appropriate?
Question 4
When considering differential susceptibility, the Agency's guidance separates the potential
susceptible period into two age groups, 0-2 years and 2-15 years. These groupings were
based on biological considerations rather than exposure considerations. The first grouping, 0
- 2 years of age, is meant to encompass a period of rapid development and the second
grouping, 2-15 years of age, was selected to extend through middle adolescence
approximately following the period of rapid developmental changes during puberty. Please
comment on the scientific rationale that was used to justify these age groupings. Can the
SAB suggest other plausible ways to make these groupings?
Question 5
The guidance provides a quantitative approach to account for the greater susceptibility of
early-life exposure to chemicals that act through a mutagenic mode of action. An adjustment
factor of 10 is applied to the cancer slope factor (derived from animal or epidemiology
studies) for exposures before 2 years of age, a factor of 3 is applied for ages between 2 and
15 years, and no adjustment is applied after the age of 15. Please comment on whether the
data and EPA analysis are scientifically sufficient to support these adjustment factors. Are
sufficient data, including breadth of chemicals, available to make these determinations?
Question 6
The Agency recognizes that consideration of children's risk is a rapidly developing area and,
therefore, the Agency intends to issue future guidance that will further refine the present draft
guidance and possibly address other modes of action as data become available. The Agency
welcomes the SAB's recommendations on other modes of action that may be most fruitful to
assess in similar future analyses.
Question 7
The analysis presented in the current Guidance relies on postnatal studies. Can the SAB
recommend how to best incorporate data from transplacental or in utero exposure studies into
future analyses?
Question 8
The Agency welcomes the SAB's recommendations on critical data needs that will facilitate
the development of future guidance addressing differential life-stage susceptibility.
-------
3. RESPONSE TO THE CHARGE QUESTIONS
The Review Panel concurs with the overall approach adopted by the Agency of using
default adjustment factors to account for increased susceptibility due to early-life exposure, and
the Review Panel agrees that the values chosen for the cancer slope adjustment factors in the
Supplemental Guidance appear to be reasonable from consideration of the literature. The
Review Panel, however, suggests that the Agency improve the statistical analysis of the data and
provide a more extensive discussion of how the Agency arrived at the choice of the lOx and 3x
adjustment factors. The Agency should also make clear that these default adjustment factors
would be used only when no data are available to directly assess cancer susceptibility from early-
life exposure to a particular chemical carcinogen. The Agency should consider conducting
additional research to address this issue directly as suggested by several public presenters. After
considering all relevant materials, both written and oral, the Review Panel provides below its
comments and recommendations for each charge question individually.
3.1. Response to Charge Question 1
Overall, the specific information and data selected, presented, and analyzed by the
Agency on the mutagenic mode of action appear accurate and reliable, and the presentation on
the mutagenic agents was clear and concise. The Tables were for the most part self-explanatory.
While quantification of the differences in potency across life stages is difficult, the steps taken by
the Supplemental Guidance - namely 1) the default assumption that early-life represents periods
of increased susceptibility to mutagenic carcinogens, and 2) the quantification of the potency
slope adjustment are reasonable given the available data. It should be pointed out that this
statement is made with the knowledge that the procedure established in the Supplemental
Guidance for weighting carcinogens for early-life exposure is a default procedure to be used in
the absence of chemical-specific information relevant to risk assessment following early-life
exposure. As noted in the Agency's carcinogen risk assessment guidelines, when there are
chemical-specific data on early-life susceptibility (or lack thereof), that information should be
used in the risk assessment of the specific carcinogen.
The assumption that mutagenic carcinogens are likely to be more potent when exposure
occurs early in life is supported by a number of additional lines of inquiry not explicitly noted in
the Supplemental Guidance. Indeed, the neonatal mouse model, used for decades, is known to
be useful for detecting carcinogens with a mutagenic mode of action (McClain et al., 2001;
Flammang et al., 1997). Studies have also shown elevated DNA-adduct formation in tissues
from young animals exposed to mutagenic carcinogens relative to older animals (e.g., for vinyl
chloride) (Laib et al., 1989; Morinello et al., 2002).
There are a large number of studies looking at the impacts of early-life exposure to
carcinogens. Many of these studies, as well as the basic theories of carcinogenesis, point to the
potential for early-life stages to be especially susceptible to chemicals acting through a
mutagenic mode of action. Factors that contribute to this phenomenon may include, but are not
limited to, differences by age in: 1) cell division rate, 2) DNA repair capability, 3) state of
differentiation and presence of stem cells, and 4) metabolic activating and detoxifying capability
of tissues. These important factors differ in a growing and differentiating organism from a
-------
mature one, and differ at different stages of development. As noted by Swenberg et al. (1992),
Anderson et al. (2000), Ginsberg (2003) and others, a major factor in early-life sensitivity to
carcinogens is believed to be rapid cell division in growing and differentiating organisms.
Mutations caused by carcinogens may be propagated if DNA repair does not occur before the
cell divides. The rapid tissue growth and concomitant cell division can result in clonal expansion
of initiated cells followed by promotion/progression to tumor formation. It has been observed
that actively transcribing DNA is more prone to adduct formation (Thomale et al., 1994). DNA
repair can be deficient in fetal and neonatal tissues for some repair enzymes relative to adult
organisms. This appears to be the case for alkyl-guanine alkyltransferase in neuronal tissues and
likely plays a major role in the production of nervous system tumors by alkylating agents when
exposure occurs early in life but not later in life (Rice and Ward, 1982; Naito et al., 1981).
McConnell (1992) noted that perinatal exposure in conjunction with adult exposure usually
increases the incidence of neoplasms and reduces the latency to tumor formation. Interestingly,
this has also been observed for some non-mutagenic carcinogens.
There are many studies evaluating carcinogenesis after preconceptional exposure,
transplacental exposure, lactational exposure, and early postnatal exposure to mutagenic
carcinogens that are not cited in the Supplemental Guidance (see Anderson et al., 2000).
Although most of these investigations did not expose adults and juveniles in the same study, the
data generally indicate increased early-life sensitivity when compared to results of studies in
which exposure starts at maturity. This is manifested as higher tumor yield, shorter latency, and
in some cases different tumor sites. At a minimum, one can say that these studies provide
supporting evidence for use of a cancer slope adjustment factor for early-life exposure to
mutagenic carcinogens.
For some mutagenic chemicals the highest tumor yields may be from prenatal exposure,
early postnatal exposure, and from adult exposure (Anderson et al., 2000). In general, the studies
reviewed by McConnell (1992) and Anderson et al., (2000) indicate that early-life exposure to
mutagenic agents appears to result in higher tumor yield and shorter latency relative to later-life
exposures alone. It should be noted that many studies also reported higher tumor incidence from
exposure to non-mutagenic carcinogens when exposure starts early in life (e.g., DES, dieldrin,
estragole, dioxin), and particularly when exposure continues through adulthood (Newbold et al.,
1982, 1990, 1998, 1995; Okashaetal., 2002).
Many carcinogens require metabolic activation. The xenobiotic metabolizing enzymes of
the liver and presumably other tissues have a generally lower level of activity and different
isoforms prenatally as well as for some time postnatally (Cresteil et al., 1998; Milsap and Jusko,
1994; Snodgrass, 1992). Despite the apparently lower potential for metabolic activation in early-
life, the susceptibility to carcinogenesis can be elevated in early life even when metabolic
activation is required (e.g., benzo(a)pyrene).
Many investigations focused on prenatal exposure to carcinogens in order to shed light on
mechanisms of carcinogenicity and the relationship between development and carcinogenesis.
Relatively fewer studies evaluated early-life postnatal exposures and adult exposures in the same
study or series of studies. Increased susceptibility in post-natal early-life to mutagenic
carcinogens relative to adult exposures conducted in the same animal studies has been
demonstrated for a number of compounds and agents including N-ethyl-N-nitrosourea (ENU),
-------
some polycyclic aromatic hydrocarbons, vinyl chloride, urethane, some nitrosamines,
azoxymethane, amitrole, benzidine, and various types of radiation (see review by Anderson et
al., 2000). Most of the key studies are cited in the Supplemental Guidance. Additional studies,
not cited in the Supplemental Guidance, which may describe relevant data useful for quantifying
the adjustment factor are provided in Appendix 1.
Available human data indicate that exposure to ionizing radiation early in life results in
higher incidences of cancer relative to adult exposure for some tissues (thyroid, bone marrow,
stomach, colon, lung, breast) (see Japanese survivor studies cited in the Supplemental Guidance;
Miller, 1995), with evidence of specific windows of susceptibility (e.g., puberty for breast cancer
risk from radiation treatment for Hodgkin's lymphoma, as reported by Bhatia et al., 1996). Two
other examples should also be noted because they illustrate the complicated interactions of
radiation damage and life stages. Those examples include the data on radiation treatment of
enlarged thymus in infancy and breast cancer risks and the risk of these cancers in childhood and
young women receiving repeat fluoroscopy for tuberculosis (Carmichael et al., 2003; Hildreth et
al., 1989; Ron, 2003).
In addition, there are several studies not cited in the Supplemental Guidance that have
utilized neonatal mice in an initiation-promotion protocol (see Appendix 2). These studies have
demonstrated distinct gender, age, strain, and compound-related differences in the liver tumor
promoting response in neonatal mice. These data suggest a different mode of action for liver
neoplasms in the treated neonatal mouse compared to the adult treated mouse. The Agency
should expand the discussion of these data in the Supplemental Guidance as they illustrate a
potential difference in the biology of the lesions induced in the neonatal mouse versus those
induced in the adult mouse. If the lesions are different in their biology then they may infer a
different mode of action. If this were the case, additional guidance from the Agency would be
useful.
Need for Better Explanation of Inclusion/Exclusion Criteria
As emphasized by some of the public commenters, the criteria for inclusion/exclusion of
specific data in the analyses need clarification. The contexts in which data are collected to
address a specific question define the bounds one must put on the interpretation of the results of
the analysis using the data. In very broad terms, data can fall into four specific areas: anecdotal,
selective, comprehensive and representative. Representative data is the ultimate scientific goal
in that an analysis of representative data, when done properly, should provide information on the
distribution of possible outcomes in the general population of outcomes that can conceivably
occur. Medians, means, and percentiles have meaning relative to the general population.
Comprehensive data would encompass the collection of all possible data relating to an issue,
which match some clearly defined criteria for what constitutes acceptable data. Comprehensive
data are more difficult to interpret than representative data, but still provide distributional
information that would be of value. Selective data refer to situations in which you select certain
pieces of information because you feel they would give you some information on the range of
possible outcomes that might occur. As such, selective data can be informative to the range of
outcomes but are unlikely to inform the probability of a certain outcome occurring in the entire
range of possible outcomes. Finally, anecdotal evidence can inform about the possibilities of a
-------
certain outcome, but gives only a very crude estimate for the possible range of outcomes. The
toxicological data used by the EPA in the analysis of the factor to use in adjusting the slope for
perinatal/childhood exposure is somewhere between anecdotal and selective and one must
consider this in interpreting the findings from the evaluation.
As described in Section 2.1 of the Supplemental Guidance, the Agency chose to utilize
studies in which exposures occurred during various life-stages in the same study. The reason
being that such studies exclude problems with inter-study comparison which is a valid concern.
While this is a sound reason for including the studies that were analyzed, more effort should
have been made to evaluate some of the excluded studies. There are studies not used in the EPA
analysis in which exposures of juvenile and mature animals to carcinogens occurred in the same
study (see Appendix 1). The reason for exclusion of these studies is not always apparent.
The decision to select studies that compared tumor incidence between early-life and adult
exposures (p. 11, par. 1 of the Supplemental Guidance) yielded a more consistent database for
the mutagenic, complete carcinogens examined. Other studies that used neonatal and newborn
exposure and measured neoplasm formation have been excluded by design. Reliance on selected
references provides a less complete data set to examine the hypothesis that the young are more
sensitive than adults to carcinogens than if all infant treatment papers were included. The
database on which the mutagenic mode of action analysis was based came from predominantly
one research group working with a mouse model. This might lead some to presume that the
conclusions derived from the analysis are not generalizable. The inclusion of additional studies
would address this issue.
The criteria used by the Agency to select studies did not allow the use of data on
mutagenic carcinogens for which exposure occurred at different life-stages in the same species in
multiple investigations. Extending the presentation of some of these data would help the
argument that mutagenic carcinogens are likely to be more potent when exposure occurs early in
life. If tumor incidence data following exposures at different life stages are available from
different studies in the same strain, it would be reasonable and possible to use those data in the
adjustment factor analysis.
Interstudy Comparisons
Dosing Regimen
It appears that some studies were excluded from the analysis because the dose regimens
at the early-life and mature stages were different. For instance, the data for tamoxifen-induced
tumors in Wistar rats (Carthew et al., 1996, 2000), which demonstrated higher potency when
given to juvenile rats relative to adult rats, were not used because of dose differences in the
immature versus mature rats. It seems that an approach could be taken to evaluate these data as
part of the analysis on appropriate adjustment factors (see, response to Question 5).
The Supplemental Guidance (e.g., see p. 15) states that weekly food consumption rates
and body weights generally were not available to allow more precise expression of the doses in
terms of mg/kg for studies in which the carcinogen was dosed via the feed or drinking water.
-------
One could assume that the exposure itself did not affect food consumption or weight gain and
use standard available data on typical values for the species in question. This might allow use of
more of the available data for the analysis of the potency slope adjustment factors.
Different Tumors at Different Age-of-Exposure
It also appears that some studies were excluded from the analysis because different types
of tumors resulted from exposure to carcinogens at different life stages. The Supplemental
Guidance (p. 22, last par.) indicates that early-life is a time of increased susceptibility to
urethane-induced lung adenomas, and that these tumors do not occur following exposure of adult
animals (Rogers, 1980; Liebert et al., 1964). However, urethane does induce other tumor types
in adults. Many times there is little site concordance between species or within species of
different life stages. Standard risk assessment practice is to use the most sensitive site and sex as
the basis for calculating cancer slope factor. The Agency could consider evaluating the ratios of
the dose that produced an early-life specific tumor type to the ratio for a later-life but different
tumor type. This would be particularly appropriate if the most sensitive site in the early-life
exposure in terms of potency is the site that does not develop tumors when exposure starts at
maturity.
3.2. Response to Charge Question 2
The Review Panel agrees with the Agency's conclusion that approaches need to be
developed for agents with a known mode of action that is non-mutagenic (Tier 2b, Fig. 3 of the
Supplemental Guidance). The Review Panel disagrees with the Agency's conclusion that
approaches and data are insufficient at this time to develop guidance on how to address non-
mutagenic chemicals with an unknown mode of action (Tier 3, Fig. 3 of the Supplemental
Guidance). The Review Panel believes the data set for the non-mutagenic carcinogens to be
qualitatively similar to that for the mutagenic carcinogens, although there are obvious
deficiencies in both data sets, including small numbers of tumors overall and non-significant
differences between adult and juvenile tumor incidences for some of the chemicals presented in
the non-mutagenic data set. Although the non-mutagenic carcinogens differ widely in
mechanism of action, the patterns of effects and the magnitudes of the ratios of juvenile versus
adult incidences in the non-mutagenic data set do not differ appreciably from those in the data set
for chemicals with a mutagenic mode of action. Therefore, the Panel believes that the Agency
should consider the development and application of default adjustment factors for chemicals that
are carcinogenic through an unknown mode of action (Tier 3, Fig. 3 of the Supplemental
Guidance).
Support for the proposition that early-life exposure to carcinogens, regardless of the
mode of action, results in increased incidence of tumors comes from the application of the time-
dependent version of all multistage models of carcinogenesis. Assuming life expectancy is not
dramatically affected, exposure for a fixed period early in life to a carcinogenic agent, compared
to the same exposure later in life, provides a longer time window for any early stage effects to
present themselves as detectable tumors. For example, early-life exposure to a carcinogen
provides more time for tumors to be expressed, particularly if the agent in question has a long
latency period (see Figure below). This difference in latency is not currently incorporated into
-------
the Agency's guidelines. The slope adjustment factors chosen by the Agency will help to
address these limitations in current risk assessment.
Exposure period
Long Latency
Exposure period
^(£
Shoi
Birth
Tumor Onset
The Review Panel notes that for certain groups of chemicals that act by non-mutagenic
modes of action, there is enough evidence supporting increased susceptibility to cancer with
early life exposure that the Agency should include a discussion of these agents in the
Supplemental Guidance. Although these chemicals may not be amenable to the quantitative
analysis performed by the Agency, they serve as important examples in support of applying a
default factor to non-mutagenic mode of action carcinogens when the mechanism of action is
unknown.
According to the Supplemental Guidance (p. 18, par. 1), chemicals that are estrogen
receptor agonists or antagonists, such as DBS and tamoxifen, were not subjected to quantitative
analysis by the Agency because no studies were available in which both juvenile and adult
dosing occurred. However, multiple studies have been performed with both of these compounds,
which observed increased reproductive tract tumors in rodents treated prenatally or during the
neonatal period compared to an absence of such tumors with treatment during adulthood. For
example, uterine, vaginal, and cervical cancers were observed with prenatal and neonatal
exposure of mice to DES (McLachlan et al., 1980; Newbold and McLachlan, 1982; Newbold et
al., 1990), whereas no such tumors were observed with lifetime exposure of adult mice
(Highman et al., 1978). Although these observations come from different studies using different
strains of mice, a review paper by Newbold (1995) cites unpublished data from her laboratory
showing that acute treatment of adult mice does not result in uterine adenocarcinoma, whereas a
similar treatment regimen during the neonatal period does cause adenocarcinoma. Presumably
these studies would have been done in the same strain of mouse. The human data for DES
support the animal data in that women who took DES did not develop vaginal adenocarcinoma
or other cancers, but their daughters who were exposed in utero did develop vaginal
adenocarcinoma. Other estrogen receptor agonists, including 17beta-estradiol (Newbold et al.,
1990) and genistein (Newbold et al., 2001), have also been shown to induce uterine
adenocarcinoma with treatment during the neonatal period. Perhaps with some minimal effort,
the Agency may be able to obtain these expanded data as they move forward with known non-
mutagenic modes of action.
Tamoxifen, an estrogen receptor agonist/antagonist, causes uterine adenocarcinoma when
administered gestationally (Diwan et al., 1997) and neonatally (Carthew et al., 1996; Newbold et
10
-------
al., 1997) in rats and mice, whereas adult treatment (Carthew et al., 1996) does not. The
Carthew et al. studies (1996, 2000) are cited in the Supplemental Guidance (p. 18, par. 1) as
being inappropriate for quantitative analysis because of the very different doses used for adult
and neonatal treatment (42mg/kg/d in adult rats versus Img/kg/d in neonatal rats). This seems to
be missing the obvious point that uterine cancers were induced by dosing with a much lower
dose for a much shorter interval in neonatal animals. However, the Carthew et al. study (1996),
states that the dose was actually 420 mg/kg of feed, whereas the Carthew et al. study (2000) used
gavage dosing. If the Agency estimated the daily dose based on average feed intake this should
be stated in the Supplemental Guidance (this would imply a food intake of 100 g/day, which
seems high). The Supplemental Guidance also states that "the adult dosing period of only three
months in the tamoxifen study potentially results in an overestimate of the early susceptibility
compared with other adult studies with chronic dosing." (see p. 18, par. 1). This would seem to
be incorrect for two reasons. First, the calculation of incidence per unit time of dosing
presumably adjusts for this. Second, there were two adult dosing regimens used in this study,
daily dosing for 3 months in rats or daily dosing from 8 weeks until 24 months in mice (Carthew
et al., 1996). The authors report 4/24 animals with uterine tumors (two deciduomas, one
hemangioma and one leiomyoma, but no adenocarcinomas) at 20 months age with the 3-month
dosing regimen in rats and no tumors with the 24-month regimen in mice. These tumors
occurred with adult only treatment and may not be treatment-related. The Review Panel offers
the studies cited above as additional support for the assertion that there may be greater
susceptibility to cancer development from early life-stage exposure to chemicals that act as
estrogen receptor agonists than from adult exposure.
Dioxins and related compounds comprise another class of compounds about which more
could be said in the Supplemental Guidance. Dioxins are known human carcinogens (IARC,
1997; USEPA, 2001). A recent publication on the Seveso cohort of humans exposed to dioxin as
a result of an industrial explosion showed a significantly increased risk for breast cancer with
increasing serum dioxin concentration obtained soon after the time of the explosion in 1976
(Warner et al., 2002). Animal bioassays have not shown increased mammary cancer with adult
dioxin treatment (reviewed in USEPA, 2001), but a recent study by Brown et al. (1998) found
that gestational day 15 treatment with 1 |ig/kg TCDD resulted in enhanced susceptibility to
DMBA-induced mammary tumors. Similarly, neonatal treatment with 2.5 |ig/kg TCDD on
postnatal day 18 was shown to enhance susceptibility to methylnitrosourea-induced mammary
tumors (Desaulniers et al., 2001). Unfortunately, neither study evaluated a group treated only
with TCDD perinatally for development of mammary tumors. Nonetheless, the data suggest that
perinatal exposure to TCDD may increase susceptibility to the development of mammary cancers
when compared with treatment only during adulthood.
In summary, the Review Panel agrees that the need for adjustment for early life-stage
susceptibility for carcinogens acting through a known, non-mutagenic mode of action (Tier 2b,
Fig. 3 of the Supplemental Guidance) should be evaluated by the Agency on a case-by-case
basis. The Review Panel recommends that among this group of carcinogens, the Agency should
consider developing guidance for carcinogens acting via estrogen receptor binding or other
mechanisms that impact hormonally responsive tissues early in life. Particular consideration
should be given to agents that may produce a persistent increase in susceptibility to cancer across
multiple life stages following early life exposure. Finally, when the agent is non-mutagenic and
11
-------
the mode of action is unknown (Tier 3, Fig. 3 of the Supplemental Guidance), the Agency has
decided to implement a linear approach identical to that used for mutagenic agents. Because the
data for non-mutagenic agents are qualitatively similar to the data seen for mutagenic agents and
because the modeling approaches are identical, the Review Panel suggests that the Agency
reconsider the decision not to apply a default adjustment factor for the unknown mode, non-
mutagenic agents.
3.3. Response to Charge Question 3
The available studies analyzed adequately support a determination of increased early-life
susceptibility to carcinogens. Despite the large number of carcinogens and considerable testing,
the data available to allow quantification of any differential risk either broadly or for specific
tumors in humans is limited. Increased risk will likely depend upon the cancer type. Simple
multistage cancer models also predict that early-life exposures to early-stage carcinogens should
increase total lifetime risk relative to later-life exposures. For later-stage carcinogens the models
suggest the opposite.
Because many carcinogens lack a comprehensive early-life data set, the need exists for a
default approach that in the absence of agent specific information adjusts for potentially
increased early-life susceptibility. The data are strongest for mutagenic carcinogens largely
because that database is more extensive, but are hard to distinguish from the general pattern seen
for the non-mutagenic agents included in the analysis. The data set analyzed was restricted to
chemicals for which multiple exposures in different life stages were available. However there is
a wealth of other individual chemical studies that support the basic premise of early life
differences but do not allow a quantification of the differences. These include DBS and
tamoxifen, as has been discussed earlier, and others that can be found in a review by Anderson et
al. (2000). Thus, there is broader scientific support for differential susceptibility than reflected in
the Supplemental Guidance. In recognition of this differential susceptibility, application of an
adjustment to the adult cancer slope factor seems to be the most transparent and practical
approach for risk assessment. One other approach would be to evaluate chemicals on a case-by-
case basis, however, the Review Panel believes that the data for increased susceptibility to cancer
with early-life exposure are sufficiently compelling that this approach could be rejected.
3.4. Response to Charge Question 4
The Agency is proposing to adjust the risk estimates for adult cancer risks from early-life
exposures by incorporating two age groupings intended to capture increased periods of
susceptibility: 0-2 years of age, and 2-15 years. The first group encompasses the period of most
rapid growth and development (Gokhale and Kirschner, 2003; Okasha et al., 2002). The second
group was selected to "represent middle adolescence appropriately following the period of rapid
developmental changes during puberty." These recommendations were based on experimental
data that compared the early-life only versus adult only and lifelong versus adult only exposure
periods.
The Panel believes that the Supplemental Guidances would be strengthened by including
more precise definitions of selected terms. The age categories need to be defined so that they are
12
-------
mutually exclusive. In addition, "adult" cancer risk is not well defined, other than to say that the
focus of the Supplemental Guidance is on "... childhood exposures resulting in cancer later in
life." (p. 6 of the Supplemental Guidance).
Although there are significant physiological differences between pre-pubertal and
pubertal children, there are limited data to indicate that the risk for development of cancers may
be different in the two groups. Individuals during puberty may be more susceptible to some
carcinogens than individuals at other life stages; consequently, the Review Panel concludes that
there should be a separate adjustment factor for the 9-15 year old group.
There has been a great deal of interest in the identification of critical windows of
exposure as related to health outcomes in both children and adults. Several recent publications
describe investigations of growth and development characteristics in childhood ("childhood
exposures") and adult health outcomes, including cancer. Many of the studies assessing the
impact of growth on subsequent health status have categorized growth into three phases, based
on a model proposed by Karlberg et al. (Karlberg et al., 1987). Although the cut points used to
define these three groupings vary somewhat across studies, generally the categories are defined
as: 1) Infancy - from midgestation to age 2-3 years; 2) Childhood - from 3 years until
"puberty"; and 3) Puberty (Gokhale and Kirschner, 2003; Okasha et al., 2002; Hilakivi-Clarke et
al., 2001; De Stavola et al., 2000). Moreover, the importance of growth velocity with respect to
risk of subsequent adverse health outcomes, rather than absolute height and weight attained, is
stressed in these investigations (Gokhale and Kirschner, 2003; Okasha et al., 2002; De Stavola et
al., 2000; Lofqvist et al., 2001). The relevance of these growth-related changes during each
interval is described below. In order to better understand the implications of the rodent data, it
would be helpful for the Agency to include a discussion of the relationships between
developmental events in rodent species and humans. This would also allow for a closer
comparison of the exposure and dose and effect data from rodent to human when available.
The Birth to Less Than Two Years of Age Category
Growth occurs more rapidly during infancy than at any other interval over an individual's
lifetime. Physiologic characteristics of importance relative to assessing risk for adult cancers are
pronounced in infancy. During this period, there is a marked increase in linear growth and in the
growth of all organs. For example, there is a significant increase in neuronal proliferation and
maturation. The developing immune system may have a great impact on the ability to withstand
environmental insults during this period (Klinnert et al., 2001).
The 2-8 Years of Age Category
The 2-8 year old group represents a pre-pubertal period during which children grow at a
linear rate of 5-6 cm per year (Grumbach, 2002). The rate of growth during the childhood phase
is steady, although girls tend to grow in height and weight at a quicker pace than do boys and
achieve puberty earlier than their male counterparts. Hormonal influences on growth and
development are of special interest in attempting to identify appropriate age groupings for risk
assessment. Growth hormone stimulates both somatic and skeletal growth, particularly growth
of the leg bones (Karlberg et al.,1987). Insulin-like growth factors (IGF-I) and thyroid hormones
13
-------
have also been shown to influence growth during this period (Robson et al., 2002; Lofqvist et al.,
2001).
The 9-15 Years of Age Group
The 9-15 year old age group represents the period of pubertal development during which
dramatic increases in hormone levels result in growth and maturation of reproductive and other
organs. The rate of linear growth and organ growth is much greater during this period than in the
2 to 8 year age group. It is acknowledged that there is variability both within and between
genders with regard to the onset of puberty, emphasizing the differences in hormonal functioning
according to age and gender. Other factors known to influence the age at onset of puberty
include race/ethnicity and body mass index (Anderson et al., 2003; Karlberg, 2002; Rosenfield).
In males, there is very little secretion of gonadotropins by the pituitary gland until the age
of 10 years, when secretion begins to increase steadily with the onset of puberty occurring at
approximately 8-10 years of age (Grumbach, 2002). In females, the pituitary begins secreting
progressively larger amounts of gonadotropic hormones at approximately eight years of age,
with menarche occurring between ages 11 and 15 years, approximately two years after the onset
of puberty.
Peak height velocity coincides with the onset of puberty in girls (around eight years of
age) and in boys (around ten years of age) (Gokhale and Kirschner 2003; Grumbach and Styne
2002). Linear growth in young females continues but at a slower pace following menarche, with
puberty ending when the breasts have reached the adult maturation stage; there is little continued
gain in height after this period. In young males, puberty continues until age 18-20 years.
Growth and development for both sexes is regulated by growth hormone and sex hormones; the
marked increase in sex steroid secretion early in adolescence results in significant physiologic
changes, including induction of serum binding proteins and detoxification enzymes (Grumbach,
2002).
The observation that puberty is a window of susceptibility for mammary tissue has been
noted for ionizing radiation in the Japanese survivors and also in treatment for Hodgkins with
radiation and chemotherapy during puberty (Bhatia et al., 1996) and possibly for tobacco smoke
(Lash and Aschengrau, 1999; Morabia, et al., 2000). The Supplemental Guidance itself
describes this phenomenon on page 23 for mammary tumors induced by DMBA in rats (Meranze
et al., 1969; Russo et al., 1979). Increasing the slope adjustment factor for 9-15 year olds for
reproductive organ and mammary gland carcinogens follows the logic in identifying early-life as
a period of potentially increased susceptibility due to rapid cell proliferation and the associated
increased potential for clonal expansion of initiated cells.
In summary, the Review Panel recommends that the 2-15 year age group be divided into
pre-pubertal (age 2-8 years) and pubertal period (age 9-15 years). Since the risk for some tumors
increases with exposure to carcinogens during puberty, the Agency should consider increasing
the adjustment factor during this period.
14
-------
3.5. Response to Charge Question 5
The values chosen for the cancer slope adjustment factors in the Supplemental Guidance
appear to be reasonable from consideration of the literature. The Review Panel also suggests
that the Agency improve the statistical analysis of the data (as discussed below) and provide a
more extensive discussion of how they arrived at the choice of the lOx and 3x adjustment
factors.
The Data Used in Support of the Default Adjustment Factors
Considering first the 10-fold slope adjustment factor for age 0-2 exposures, the data
summarized in the Supplemental Guidance (see Table 4, Figs. 1 and 2) (n=l 1 studies for chronic
exposures) show that the median slope ratio for the linear prevalence vs. dose model is 10.0 with
a range in ratios across the 11 individual studies of 0.3 to 65.0. Whether the median value of the
distribution of 11 independent study results is an appropriate adjustment factor for modeling 0-2
age-specific exposure risks for mutagenic compounds is not clear. The public commenters have
pointed to some unique features of the collection of studies that influence the derivation of this
median value — many by a single investigator, common tumor sites (liver), the largest ratios are
all obtained from studies that use male mice. By its nature as an estimate of central tendency in
outcomes for the observed study data, it is a plausible value in the absence of actual age-specific
dose-response data for a new compound.
The choice of a 3x multiplier for the slope adjustment factor for exposures during the age
2-15 year interval is derived entirely from a crude interpolation between the Ix factor for adults
age 15+ and the lOx factor for infants age 0-2. Again this is a plausible factor given the study
data that are available but other than conforming to intuitive, if not scientifically-substantiated
bounds, there is no scientific basis in the analysis for choosing the factor of 3 over alternative
values in this bounded range.
The Supplemental Guidance uses estimates of average excess relative risk (ERR) from
atomic bomb survivor studies (Life Span Study) to support the premise of a life stage effect for
mutagenic chemicals. These data strongly support this premise. For many types of cancer
identified in the Life Span Study, estimates of ERR show an inverse relationship between
exposure and age at the time of exposure, i.e. younger people have a higher risk of cancer than
older people. However, these estimates vary considerably with age among the various types of
cancer. In some cases the 95% CI is large enough to include zero for all age categories (see
mortality data in UNSCEAR 2000 Annex I). Thus, precise adjustment factors for younger age
groups may be somewhat misleading without a discussion of uncertainties and limitations.
Discussion should include the error associated with incidence data used to estimate ERR among
the age groupings and the variation in ERR with age among the different types of cancer. For
example, Table 9 in the Supplemental Guidance provides average ERR for four age groups. The
trend clearly supports the premise that younger people have a higher risk of thyroid cancer, but
the number of cases is small, and there is no indication of variance.
The ERR estimates cited in the Supplemental Guidance (see Tables 8, 9, and 10) are
based on cancer observed in populations exposed to large doses of radiation delivered at a high
15
-------
dose rate (IMSCEAR 2000). The original ERR estimates were based on a linear model applied
through the entire dose range even though incidence data clearly are not linear over the entire
dose range. Thompson et al. (1994) shows a large increase in incidence rate for all cancers in the
>1 Sv cohort (mean dose of 1.6 Sv) but a small increase in incidence rate in the 0.01-0.99 Sv
cohort (mean dose of 0.16 Sv). When broken down by cancer type, the number of cases per
cohort per cancer type is very small, even zero in some cohorts. The Supplemental Guidance
also ignores dose rate considerations. BEIR V provides a discussion of dose rate effectiveness
factors for radiation. BEIR V appears in the reference list but does not appear to have been used
in the text. Dose rate clearly affects risk. Consequently, the Supplemental Guidance should
include a discussion of the impact of dose and dose rate on the uncertainty associated with these
risk estimates.
Thompson et al. (1994) provide incidence rates among six age groupings for various
types of cancer. However, the number of cases within many of the cohorts (including those for
thyroid cancer) is very small; several of them have zero cases. This is particularly problematic
for the high dose (>1 Sv) cohorts. Thompson et al. (1994) estimated ERR at 1 Sv for each type
of cancer by sex and age at exposure, but use of these estimates in the Supplemental Guidance
needs to be accompanied by a discussion of the uncertainties. For example, the ERR for thyroid
cancer in the 0-9 age group was 9.46 and for the 10-19 age group was 3.02. However, the 95%
confidence interval for all ages was 0.48 - 2.14, once again pointing out the significance of
uncertainty in the estimates.
Are the Analyses Used to Derive the Adjustment Factor Values Appropriate?
The analyses presented in the Supplemental Guidance are descriptive and use no formal
statistical evaluations to test the selected adjustment values. Formal statistical procedures could
have been used to more appropriately analyze individual study data; one such method is
described in Halmes et al. (2000). This analysis corrects for survival differences and differences
in observation time, something not done in the EPA analysis and something which is likely to
change the observed ratios. EPA is interested in whether the pattern of dose-response resulting
from curve-fitting of the adult exposure data will, with their dosing correction and an appropriate
factor change on the slope of the dose-response curve, predict the dose-response seen from early-
life exposure. This is readily analyzed through direct statistical methods rather than a focus on
only paired exposure groups. For example, EPA could apply their model choice to the combined
perinatal/adult dose-response data and simply evaluate how often this hypothesis is rejected.
However, given the limitations of the current data set, such an analysis is unlikely to
substantially alter the general range of ratios seen in the supplementary guidance unless
additional data could be used.
In the Halmes et al. (2000), the majority of strictly early adult-life exposures, when
averaged over the lifespan of the animals, produced greater risk than predicted by the chronic
exposure dose groups and no apparent difference existed between mutagenic and non-mutagenic
exposures. While these analyses were done for data with early adult exposure rather than
perinatal exposure, these findings support EPA's use of a slope adjustment in the perinatal period
and suggest that non-mutagenic agents of unknown mode of action could also use a slope
adjustment in early-life.
16
-------
Even assuming a full analysis as done by Halmes et al. (2000) is not used here, the
computation of the relative slope coefficients for juveniles and adults could have been done on
the log-scale rather than the arithmetic scale. Since most models for cumulative incidence for
tumor onset assume a functional form that includes an exponentiated dose function, changes in
the point-of-departure for a fixed risk would better be reflected by a comparison of log-
transformed data. The math is as follows:
P(dose)=l-[l-P(0)]exp(-slope*dose) [1]
Hence
(log[l-P(0)]-log[l-P(dose)]}/dose=slope [2]
This equation then implies that the ratio of the slopes would be the ratio of equation 2 for
juveniles divided by equation 2 for adults. For small P(dose) and small P(0), the EPA formula is
approximately equal to [2]; for medium range P(dose) as we have here, the equations are not the
same. This transformation is nonlinear so the resulting ratios will be different.
If the EPA uses the analysis as presented in the Supplemental Guidance, the exclusion of
cases where the adults had no tumor and the juveniles had some tumors biases the median
estimate of the resulting adjustment factors downward. The cases represent more than 10% of all
tumors cited in the EPA data. This bias is likely to be in the direction of smaller ratios for
medians, etc. Treating the division by zero as a big number, medians can still be calculated.
3.6. Response to Charge Question 6
Lifetime risk assessment appears to be little affected by changes in susceptibility that are
limited in duration to the period of childhood itself, relative to the extant uncertainties and to the
conservative assumptions made. This is not surprising in view of the relatively short duration of
childhood vs. adult life. Effects in childhood that cause persistently elevated susceptibility
throughout much or all of later life are likely to produce greater impacts on lifetime risk
assessment and would be an appropriate focus for future research efforts. Further research needs
to be undertaken to understand the circumstances under which early exposures to environmental
agents may "re-program" (this term is intended to cover a diversity of mechanisms) cellular or
organismal function(s) in a manner which increases future risk independent of ongoing exposure
to the agent in question. While this mechanism may appear to be particularly relevant to
hormonally active materials, it could result from other mechanisms such as the induction of long-
term changes in cytochrome activity, alterations in cell population size, changes in cellular
turnover rate, etc.
It is likely that early-life stages have windows of susceptibility to carcinogens acting
through endocrine disruption. There are a number of studies that demonstrate susceptibility of
early-life stages to carcinogenesis by estrogen agonists/antagonists. Some of the studies on
tamoxifen cited in the Supplemental Guidance are an example. Diethyl stilbestrol exposure in-
utero produces female reproductive tract cancers in human offspring, without apparently
17
-------
increasing the risk of cancer in the mothers. Likewise, in animal models, both transplacental and
in utero exposure to DES causes increased uterine adenocarcinomas and/or cervical cancers
(Newbold et al., 1990). In addition, preconceptional exposure resulted in uterine cancers in the
offspring (Newbold et al., 1998). In Newbold et al., 1990, the investigators tested other
estrogenic compounds including hexestrol, trifluorodiethylstilbestrol and 17|3 -estradiol. The
authors note that when the incidences of hyperplasia and adenocarcinoma were combined, the
induction of these tumors and lesions followed the estrogenic potency of the compounds. The
tumors were dependent on estrogen for growth in this study, as mice ovariectomized prior to
puberty did not develop the tumors. Thus there is interplay between early-life exposure to
estrogenic compounds and later pubertal development in terms of carcinogenesis.
Additional studies have evaluated the potential for carcinogenesis following perinatal
exposure to tamoxifen, an estrogen antagonist in breast tissue but an estrogen agonist in uterine
tissue. In addition to reproductive tract abnormalities, tamoxifen induced uterine
adenocarcinomas and focal hyperplasias in mice following exposures the first five days after
birth (Newbold et al., 1997). Induction of uterine tumors in adult mice was not observed in
another study (Carthew et al., 1996). The soy phytoestrogen genistein is also capable of
inducing uterine adenocarcinoma in mice following postnatal exposure on days 1-5 (Newbold et
al., 2001). Studies of tamoxifen effects following neonatal and adult exposures of Wistar rats
indicate that the pups were more susceptible to uterine cancer induced by tamoxifen than the
adult animals (Carthew et al., 1996; 2000). It should be noted that tamoxifen may be acting by
multiple mechanisms as DNA-adducts in liver have been observed in rodent studies, and
tamoxifen exposure to adult rats results in hepatocellular carcinoma. An additional example
would be that of juvenile exposures to dioxin possibly increasing the potency of DMBA as a
mammary tumorigen (see Response to Charge Question 2, p. 11).
In summary, there is reason to believe that hormonal agents can be more potent
carcinogens when exposure occurs in early-life stages than in later-life stages alone. This area is
important to explore and the Agency should in future revisions of the Supplemental Guidance
conduct an analysis of the differences in potency by age when data become available. As noted
in the Supplemental Guidance, three estrogen active agents are currently in test at the National
Toxicology Program (NTP) in multigenerational studies, and the results of those studies should
shed light on early-life stage susceptibility. The Review Panel would also encourage the Agency
to look at clinical data with secondary tumors arising from primary chemotherapy in children
versus adults.
The proper approach for addressing other modes of actions for young and infant animals
will be dictated by the effects of the particular chemical or physical carcinogen. Since this is still
a developing area of research investigation for adult animals, the application and relevance to
young and infant animals also requires additional research investigations. These investigations,
just like those involving adult animals, should employ multiple doses to develop well-defined
dose response characteristics for each chemical/physical agent.
The Agency might also look at the data on gene-environment interactions as they relate to
polymorphisms in genes associated with xenobiotic metabolism and the critical windows of
susceptibility. This may greatly enhance our understanding of these exposures and their
18
-------
relationship to cancer (in both childhood and adulthood) from a mechanistic point of view. A
careful review of this literature linked to expression levels of the same enzymes compared
between early-life versus late-life may be helpful in setting defaults for specific classes of agents.
3.7. Response to Charge Question 7
The Review Panel cannot recommend a method to incorporate data from transplacental or
in utero exposures at this time. However, the Review Panel believes that this is an extremely
important issue. It is clear from both human and animal studies that carcinogens can be
transported across the placenta and induce tumor formation in the offspring. Clearly, use of DES
as a therapeutic agent during pregnancy resulted in vaginal cancers in daughters. Incorporating
data from transplacental carcinogenesis studies is difficult but potentially important.
Studies that exposed animals prenatally and as adults have shown early-life sensitivity
from in utero exposure to a number of mutagenic carcinogens including radiation (Delongchamp
et al., 1997), benzene (Maltoni et al., 1989), vinyl chloride (Maltoni and Cotti, 1988), AZT
(Olivera et al., 1997; Diwan et al., 1999), dibenzanthracene (Law, 1940), benzo(a)pyrene (Urso
and Gengozian, 1982), arsenic (Waalkes, 2003), and a host of others (reviewed in Anderson et
al., 2000).
DNA adducts have been measured in both animal embryos and human fetuses exposed to
mutagenic carcinogens including poly cyclic aromatic hydrocarbons (PAHs) (Arnould et al.,
1997; Klopov, 1998; Autrup et al., 1995; Whyatt et al., 1998), vinyl chloride (Laib et al., 1989),
ENU, and others. DNA adducts in the liver are higher after perinatal exposure to vinyl chloride
than after exposure at maturity (Swenberg et al., 1992). In at least one study, PAH-DNA adduct
levels were higher in white blood cells in the newborn human than the mother (Whyatt et al.,
1998).
One possible approach to incorporating prenatal exposures in evaluating early-life
sensitivity to carcinogenesis is to assess studies where both in utero and adult exposures were
investigated in the same study. The review by Anderson et al. (2000) that is cited in the
Supplemental Guidance cites a number of papers that could be used in this type of analysis.
Since the time of peak early-life sensitivity can be either pre- or postnatal, studies that evaluated
repeated prenatal, postnatal, and adult exposures would be the most useful for quantitative
analysis of an adjustment factor for early-life exposure. Focusing on those studies might enable
one to define the most sensitive period more clearly. However, quantifying the dose to the pups
is difficult in these studies; that in turn makes quantitative evaluation of early-life susceptibility
difficult. Thus, it seems unlikely that such studies will contribute data directly useful for
quantitative risk assessment unless and until a marker or model of systemic exposure to the
relevant material within the fetal compartment can be developed and validated. Application of
physiologically-based pharmacokinetic (PBPK) modeling of transplacental transfer may prove
fruitful although the models themselves are relatively undeveloped and require use of
assumptions as much of the necessary data are unavailable. The Agency should, despite these
difficulties, invest some effort in evaluating the prenatal studies as they may provide better
evidence of peak developmental susceptibility. The evaluation could initially be qualitative and
19
-------
move over time towards a quantitative assessment as models are developed and new data are
obtained.
The Agency may wish to give early consideration to the manner in which such data are to
be utilized. Specifically, such data could be used either on a chemical specific basis to establish
individual chemical risks, or could be used to obtain a better understanding of the appropriate
application of adjustments to exposure data obtained in later-life exposures. Because of
differences in, for example, metabolic ontogeny between rodents and humans, it is not clear that
early-life exposure is, on a chemical by chemical basis, an appropriate model for quantitative
human risk assessment. A more accurate and appropriate risk assessment may well be achieved
by the application of biological understanding and quantitative adjustments obtained in
controlled, early-life experiments to later-life exposure data, as described in the Supplemental
Guidance.
3.8. Response to Charge Question 8
There are rather large data gaps that need to be filled for the myriad of carcinogens that
the Agency is charged with regulating. The majority of carcinogens have not been adequately
tested in terms of early-life susceptibility. The Agency could work more closely with the
research community to encourage the evaluation of early-life stage susceptibilities on a routine
basis. Prioritization of carcinogens in the environment in terms of potency and extent of
exposure would aid in deciding which chemicals to study first. The Agency should also partner
with other federal agencies such as the Centers for Disease Control (to evaluate human exposures
using monitoring data in order to inform the prioritization of chemicals for study) and Food and
Drug Administration (which may have animal carcinogenicity studies on pharmaceuticals
pertinent to the issue). Finally, the Agency could provide more resources to support the study of
appropriate protocols for testing for early-life susceptibility to carcinogens with varying
mechanisms of action.
Specific Suggestions (Not in Priority Order)
The Supplemental Guidance relegates data on ionizing radiation to a supportive role.
There is a large amount of published information, some of which EPA itself has
reviewed, from human data on the Japanese bomb survivors that could possibly be used
to improve the analysis. Since these analyses are of humans exposed to radiation, the
uncertainty of inter-species extrapolation does not exist. Further, pharmacokinetic issues
are moot for radiation exposures so these studies may provide a clearer view of the
importance of pharmacodynamic factors. The data in Tables 8 and 9 of the Supplemental
Guidance indicate that amongst the Japanese survivors of the atomic bomb the younger
age groups were more sensitive than the adult age groupings to the induction of a number
of cancers including thyroid, bone and connective tissue, skin, breast, and leukemia. The
Agency should consider folding these data on ionizing radiation into the potency slope
adjustment factor analysis and weighting them quantitatively.
Additional research on adaptive responses in both adult and young is needed. Study of
possible hormesis effects - protective effects at low dose - if known for the young should
20
-------
be explored. The state of the science in this field especially as it relates to
infant/perinatal exposure should be incorporated in the Supplemental Guidance.
There is a clear need to develop a better understanding of the biology and physiology of
rodents typically used in carcinogenesis bioassays as they relate to similar phenomena in
humans. The impacts of life-stage, gender, and related underlying physiological
differences in the animal models need to be related to similar changes in humans. Use of
primate models, which more closely mimic lifestages in humans, may further the
understanding of early-life stage physiology and biology. In addressing life stage
changes in physiology, key areas to address include the influence of hormonal levels and
of phase I and phase II metabolic enzymes.
Research is needed to better integrate our molecular understanding of carcinogenesis with
life stages in humans and laboratory animal models. The use of genomics and
proteomics in conjunction with bioinformatics holds promise for elucidating the many
changes occurring in the cell/tissue/organ/organism during carcinogenesis as well as
during development.
There is a clear need for studies that address dosimetry issues. Studies using some of the
compounds for which there appears to be evidence of increased early-life stage sensitivity
which are specifically designed to take into account the need for dose quantification and
tumor latency could be performed, at least as related to postnatal exposure. Such studies
would probably require less-than-lifetime dosing during younger and older life stages,
with multiple and similar times of sacrifice after onset of exposure to assess latency
issues. As noted in the Supplemental Guidance, one would like to have studies with
excellent quantitative data on tissue levels of test compound and its active metabolites in
both exposed embryos/fetuses and exposed adults so that, following in utero exposures
and adult exposures resulting in known target organ doses, the subsequent development
of cancers can be compared. Improved PBPK models would also be very useful in
extrapolating internal doses.
The Agency needs to look more towards models applicable to groups of chemicals
related either structurally or by mechanism. Studies of prototypes of such groupings
would be informative.
Planning efforts currently underway by the National Institute of Child Health and Human
Development, EPA and National Institute of Environmental Health Sciences for the
prospective National Children's Study (NCS) are directly relevant to the questions being
posed here. If the NCS becomes a reality, there may be opportunity to examine
physiological and biochemical changes that might relate to cancer susceptibility and
improve the current Supplemental Guidance.
In the future, the Agency should attempt to evaluate chemicals that are structurally
similar to those chemicals that only produced tumors when exposure occurs early in life.
These chemicals, while likely few in number, would be of great concern because the
standard bioassay or typical occupational epidemiological study would not pick them up
21
-------
as carcinogens. Hence, such chemicals would not be treated as carcinogens by risk
assessors. Perhaps the Agency can work towards identifying environmental chemicals
that are structurally similar to the chemicals that only produce tumors when exposure
occurs early in life for the risk assessor to consider. The Panel recommends that a more
systematic effort be made to identify such chemicals and to define their characteristics.
4. MISCELLANEOUS COMMENTS
Clarification of the Terms and Definitions Used in the Supplemental Guidance
Many of the terms used in the Supplemental Guidance (i.e., mutagenesis, DNA reactive,
genotoxic, nongenotoxic) should be clearly defined. This could be accomplished by including a
glossary or appendix section with the definitions used. In addition, the term "mutagenic mode of
action" should be more clearly defined, and consideration should be given to utilization of this
term in the main guidance document to assure that either the usage is identical or that any
differences in intended usage are made clear. It appears that the draft Supplemental Guidance
considers a mutagenic mode of action if a chemical is carcinogenic and it is mutagenic in short-
term bioassays. Several questions should be addressed: Does DNA binding in vivo infer
mutagenicity? Are the terms mutagenic, DNA reactive, and genotoxic used interchangeably in
the Supplemental Guidance? Each of these three terms has a specific identity associated with it
and a specific mechanism and result. How will indirect mutagens, i.e. oxidative damage, be
considered? Along this line, with the DNA reactive carcinogens, mutation is not the only
component of the mode of action involved in the neoplasm formation. Modulation of cell
proliferation, apoptosis, and gene expression also participate in the development of the observed
cancers and need to be considered and addressed in proposed modes of action for these
chemicals.
Data for Use in Determining the Mode of Action
The Supplemental Guidance should explicitly state the criteria for deciding that there are
sufficient data to determine a particular agent's mode of action both in infant and adult animals
(or at least refer back to the Cancer Risk Guidelines where these criteria are stated) Along these
same lines, the Supplemental Guidance should comment on the quantity and quality of
experimental evidence needed before a default approach would be applied.
Tables
The tables do not indicate the reason for animal death in each study. Was the death due
to chemically induced carcinomas or due to other organ failure? For example, Nitrosamines
produce cirrhotic and general liver and kidney cytotoxicity in mice.
Was the tumor incidence expressed in the tables based on adenomas, carcinomas,
combined adenomas and carcinomas? The tumor incidence values should specify the type of
each tumor induced.
22
-------
In several of the studies cited (see Tables on pp. 60, 62, 63, 64 of the Supplemental
Guidance) no control groups were apparently utilized in the studies, making interpretation of the
results difficult. This is a particular problem in trying to assess dose-response characteristics and
threshold dose levels for the studies involved. Both parameters are needed in developing strong
mode of action evaluations.
Tables 4 and 5
There are several errors. EPA should recheck data in tables against original papers and
recalculate distribution of ratios; the errors found would probably not change the analysis
significantly, although at least one ratio was off by a factor of 3 (in Table 4, DEN 6 ug/kg male
mouse liver ratio should be 4.6, not 1.8).
Rounding should take place at the end, not the beginning. EPA was inconsistent in doing
this, sometimes rounding the percent incidences prior to calculating the ratios and sometimes
not. One can get different calculated ratios, of course, when rounding at the beginning rather
than the end.
Some of the citations are missing from the bibliography (e.g., Vesselinovitch et al.,
1983). Another citation, Maekawa et al., 1990, should really be Druckrey, 1970 as cited in
Maekawa and Mitsumori, 1990. Also, the Maekawa and Mitsumori 1990 citation is missing
from the bibliography.
In the study by Meranze et al., 1969, exposures were evaluated in neonatal rodents, 5-8
week rodents and adults. The most sensitive period for mammary tumors occurred during the 5-
8 week old period and undoubtedly represents development of the mammary gland during
puberty in these animals. Ratios were calculated from data for both the neonatal compared to
adult and for the adolescent compared to adults for total tumors and for mammary tumors in the
female animals. It is not clear whether all those ratios were included in the analysis of the
adjustment factors. In one Panelist's opinion, only the higher ratio for the female animals
exposed at 5-8 weeks of age makes sense to include as that represents exposure during the more
sensitive postnatal time period for the females. To include the total tumor ratio as well actually
dilutes the difference between adolescence and adult exposures for this tumor site.
The Agency should re-examine the way they utilized the data from Hard (1979). This
study exposed rats to DMN at 3 weeks of age (earliest in this study), and at 4 weeks of age, as
well as at 1.5, 2, and 3 months of age. The paper itself describes the 4-week old animals as
juveniles (4 week old rats are still in adolescence), but the Agency treated them as adults in
calculating the ratios used in the weighting analysis. The highest tumor incidences occurred in
the 4 week old rats. If the ratio is recalculated treating these animals as juveniles, which is
appropriate, then one gets slightly higher ratios when comparing the 6-week and older age
groups.
A similar problem occurs when evaluating the data from Naito et al., 1981, although it is
harder to "fix." In Naito et al., 1981, ENU was given to 1-day old, 1-week, 2-week, 3-week, and
23
-------
4-week old rats. So, 4 weeks was the oldest animal group in this study, but the rats are still
adolescents. Thus, the ratios calculated comparing the earlier age rodents with the 4-week old
rodents may slightly underestimate the difference between immature and fully mature rodents in
response to ENU with respect to neurological tumors. It is likely, though, that the underestimate
would be slight because the induction of nervous tissue tumors by ENU appears to peak with
prenatal exposures and drop fairly rapidly postnatally (see Naito et al., 1981). This may have
been recognized by the Agency and thus provides validity to the use of these data in the analysis
of adjustment factors.
The proposed method of analysis does not take into account differences in multiplicity of
tumors from early-life exposure. A number of studies have shown large differences in tumor
multiplicity depending on the developmental stage of an organ in relation to timing of exposure
(e.g., breast tumors in Meranze et al., 1969; lung tumors in a number of studies with urethane,
nerve tissue tumors in a number of studies with ENU). Multiplicity of tumors in an organ is
another indicator of susceptibility and would certainly be expected to influence disease outcome
in both animals and humans. Thus, while it may be difficult to quantitatively weight
multiplicity, it is certainly important to severity of disease, and an attempt should be made to
weight multiplicity in future analyses.
Table 6
The reference by Vessilinovitch et al. (1983) on amitrole is not included in the list of
references.
The adult tumor incidence per time for ETU-induced thyroid tumors in female mice is
incorrectly calculated as 0.02 due to an incorrect incidence rate in the control females being
subtracted. The correct incidence/time is 4/96=0.04. This decreases the ratio from 10 to 5.
For PBB induced liver tumors in female mice, the adult dosing incidence for the 0:10
dosing regimen of 42/50 is used. For the male mice and female juvenile exposures the 30 ppm
dose is used. The incidence from the 0:30 dosing regimen of 47/50 should be used instead,
which would increase the adult incidence per time to 0.875 and reduce the ratio to 3.3.
24
-------
5. REFERENCES
Anderson SE, Dallal GE, Must A. 2003. Relative weight and race influence average age at
menarche: results from two nationally representative surveys of US girls studied 25 years apart.
Pediatrics. 111( 4 Pt 1):844-50.
Anderson LM, Diwan BA, Fear NT, and Roman E. 2000. Critical windows of exposure for
children's health: Cancer in human epidemiological studies and neoplasms in experimental
animal models. Environ Health Perspect. 108(Suppl 3):573-594.
Arnould JP, Verhoest P, Bach V et al. 1997. Detection of benzo(a)pyrene-DNA adducts in
human placenta and umbilical cord blood. Human and Experimental Toxicology 16(12):716-
721.
Autrup H, Vestergaard AB, Okkels H. 1995. Transplacental transfer of environmental
genotoxins:polycyclic aromatic hydrocarbon-albumin in nonsmoking women, and the effect of
maternal GSTml genotype. Carcinogenesis 16(6): 1305-9.
Bhatia S, Robison LL, Oberlin O, et al. 1996. Breast cancer and other second neoplasms after
childhood Hodgkin's disease. NEJM334:745-751.
Brown NM, Mansolillo PA, Zhang J-X, Wang J, Lamartiniere CA. 1998. Prenatal TCDD and
predisposition to mammary cancer in the rat. Carcinogenesis 19:1623-1629.
Carmichael A, Sami AS, Dixon JM, 2003. Breast cancer risk among the survivors of atomic
bomb and patients exposed to therapeutic ionizing radiation. Eur JSurg Oncol. 29(5):475-9.
Carthew P, Edwards RE, Nolan BM, Martin EA, Smith LL. 1996. Tamoxifen associated uterine
pathology in rodents:relevance to women. Carcinogenesis 17:1577-1582.
Carthew P, Edwards RE, Nolan BM, Martin EA, Heydon RT, White IN, Tucker MJ. 2000.
Tamoxifen induces endometrial and vaginal cancer in rats in the absence of endometrial
hyperplasia. Carinogenesis 21:793-797.
Cresteil T. 1998. Onset of xenobiotic metabolism in children: toxicological implications. Food
Additiv Contamin 15 (suppl):45-51.
Delongchamp RD, Mabuchi K, Yoshimoto Y, and Preston DL. 1997. Cancer mortality among
atomic bomb survivors exposed in utero or as young children, October 1950-May 1992. Radiat
Res 147:385-395.
Desaulniers D, Leingartner K, Russo J, Perkins G, Chittim BG, Archer MC, Wade M, Yang J.
2001. Modulatory effects of neonatal exposure to TCDD, or a mixture of PCBs, p,p'-DDT, and
p-p'-DDE, on methylnitrosourea-induced mammary tumor development in the rat. Environ
Health Perspect 109:739-747.
25
-------
De Stavola BL, Hardy R, Kuh D, dos Santos Silva I, Wadsworth M, Swerdlow AJ. 2000.
Birthweight, childhood growth and risk of breast cancer. Br J Cancer 83:964-68.
Diwan BA, Anderson LM, Ward JM. 1997. Proliferative lesions of oviduct and uterus in CD-I
mice exposed prenatally to tamoxifen. Carcinogenesis 18:20092014.
Diwan BA, Riggs CW, Logsdon D, Haines DC, Olivera OA, Rice JM, Yuspa SH, Poirier M, and
Anderson LM. 1999. Multiorgan transplacental and neonatal carcinogenicity of 3'-azido-3'-
deoxythymidine in mice. Toxicol ApplPharmacol 161:82-99.
Flammang TJ, Von Tungeln LS, Kadlubar FF, and Fu PP. 1997. Neonatal mouse assay for
tumorigenicity: Alternative to the chronic rodent bioassay. Reg Toxicol Pharmacol 26:230-240.
Ginsberg, G. 2003. Assessing cancer risk from short-term exposures in children. Risk Analysis
23:19-34.
Gokhale R and Kirschner BS. 2003. Assessment of growth and nutrition. Best Practice and
Research Clinical Gastroenterology. 17:153-62.
Grumbach, MM, 2002. The neuroendocrinology of human pubert revisited. Hormone Research
57, Suppl 2: 2-14.
Grumbach, MM and Styne, DM, 2002. Puberty: Ontogeny, Neuroendocrinology, Physiology and
Disorders. In Williams Textbook of Endocrinology, P.R. Larsen, H.M. Kronenberg, S. Melmed
and K. Polansky, editors Saunders, Philadelphia, pages 1115-1286.
Halmes NC, Roberts SM, Tolson JK, and Portier, CJ. 2000. Reevaluating cancer risk estimates
for short-term exposure scenarios. Toxicological Sci 58:32-42.
Hard G. 1979. Effect of age at treatment on incidence and type of renal neoplasm induced in the
rat by a single dose of dimethylnitrosamine. Cancer Research 39:4965-4970.
Hilakivi-Clarke L, Cho E, deAssis S, Olivo S, Ealley E, Bouker KB, Welch JN, Khan G, Clarke
R, Cabanes A. 2001. Maternal and prepubertal diet, mammary development and breast cancer
risk. Am Soc Nutr Sci JNutr. 131:154S-57S.
Hildreth NG, Shore RE, Dvoretsky PM, 1989. The risk of breast cancer after irradiation of the
thymus in infancy. NEnglJMed. 321(19):1281-4.
Highman B, Norvell MJ, Shellenberger TE, 1978. Pathological changes in female C3H mice
continuously fed diets containing diethylstilbestrol or 17B-estradiol. J Environ Pathol Toxicol
1:1-30.
IARC 1997. Poly chlorinated dibenzo-para-dioxins and poly chlorinated dibenzofurans. IARC
Monogr Eval Carcinog Risk Hum 69:3 3 -3 44.
26
-------
Karlberg J, Engstron I, Karlberg P, Fryer JG. 1987. Analyses of linear growth using a
mathematical model. AdaPaediatrica Scandinavica 76(Supplement):478-88.
Karlberg J. 2002. Secular trends in pubertal development. Norm Res 57(suppl 2): 19-30.
Klinnert MD, Nelson HS, Price MR, Adinoff AD, Leung DY, Mrazek DA. 2001. Onset and
persistence of childhood asthma: predictors from infancy. Pediatrics 108:E69.
Klopov VP. 1998. Persistent organic compounds in women residing in the Russian arctic.
InternationalJoumalof CircumpolarHealth 57 Suppl 1:555-60.
Laib RJ, Bolt HM, Cartier R, and Bartsch H. 1989. Increased alkylation of liver DNA and cell
turnover in young versus old rats exposed to vinyl chloride correlates with cancer susceptibility.
ToxicolLett 45:231-239.
Lash TL and Aschengrau A 1999. Active and passive cigarette smoking and the occurrence of
breast cancer. AmJEpidemiol 149:5-12.
Law LW. 1940. The production of tumors by injection of a carcinogen into the amniotic fluid of
mice. Science 91:96-97'.
Liebert RA, Liebert AG, and Lane M. 1964. Hormonal influences on urethan carcinogenesis in
C3H/f mice. Cancer Research 24:186901879.
Lofqvist C, Andersson E, Gelander L, Rosberg S, Blum WF, Wikland KA. 2001. Reference
values for IGF-I throughout childhood and adolescence: A model that accounts simultaneously
for the effect of gender, age, and puberty. EndocrinolMetab 86:5870-76.
Maekawa A and Mitsumori K. 1990. Spontaneous occurrence and chemical induction of
neurogenic tumors in rats - influence of host factors and specificity of chemical structure. Clin
Rev Toxicol. 20:287-310.
Maltoni C, Cotti G. 1988. Carcinogenicity of vinyl chloride in Sprague-Dawley rats after
prenatal and postnatal exposure. Ann NYAcadSci 534:145-159.
Maltoni C, Ciliberti A, Cotti G, Conti B, Belpoggi F. 1989. Benzene, an experimental
multipotential carcinogen: results of the long-term bioassays performed at the Bologna Institute
of Oncology. Environ Health Perspect 82:109-124.
McClain RM, Keller D, Casciano D, Fu P, MacDonald J, Popp, J, and Sagartz J. 2001. Neonatal
mouse model: Review of methods and results. Toxicologic Pathol 29(Supp\.): 128-137.
McConnell EE. 1992. Thyroid follicular cell carcinogenesis: results from 343 2-year
carcinogenicity studies conducted by the NCI/NTP. Regul Toxicol Pharmacol. 16(2): 177-88.
27
-------
McLachlan JA, Newbold RR, Bullock BC. 1980. Long-term effects on the female mouse genital
tract associated with prenatal exposure to diethylstilbestrol. Cancer Res 40:3988-3999.
Meranze DR, Gruenstein M, and Shimkin MB. 1969. Effect of age and sex on the development
of neoplasms in Wistar rats receiving a single intragastric instillation of 7,12-
dimethylbenz(a)anthracene. IntJ Cancer 4:480-486.
Miller RW. 1995. Special susceptibility of the child to certain radiation-induced cancers. Environ
Health Perspect 103(Suppl 6):41-44.
Milsap RL and Jusko WJ. 1994. Pharmacokinetics in the infant. Environ Health Perspect
102(Suppl 11): 107-110.
Morabia A, Bernstein MS, Antonini S. 2000. Smoking, dietary calcium and vitamin D
deficiency in women: a population-based study. Eur J Clin Nutr. 54(9):684-9.
Morinello EJ, Ham A-JL, Ranasinghe A, Nakamura J, Upton PB, and Swenberg JA. 2002.
Molecular dosimetry and repair of N2,3-ethenoguanine in rats exposed to vinyl chloride. Cancer
Research 62:5189-5195.
Naito M, Naito Y, and Ito A. 1981. Effect of age at treatment on the incidence and location of
neurogenic tumors induced in Wistar rats by a single dose of N-ethyl-N-nitrosourea. Gann
72(4):569-77.
Newbold RR, Padilla-Banks E, Bullock B, and Jefferson WN. 2001. Uterine adenocarcinoma in
mice treated neonatally with genistein. Cancer Research 61:4325-4328.
Newbold RR, Jefferson WN, Padilla-Burgos E, Bullock BC. 1997. Uterine carcinoma in mice
treated neonatally with Tamoxifen. Carcinogenesis 18(12):2293-2298.
Newbold RR, Bullock BC, McLachlan JA. 1990. Uterine adenocarcioma in mice following
developmental treatment with estrogens: a model for hormonal carcinogenesis. Cancer Res.
50(23):7677-81.
Newbold RR, Hanson RB, Jefferson WN et al. 1998. Increased tumors but uncompromised
fertility in the female descendants of mice exposed developmentally to diethylstilbestrol.
Carcinogenesis 19:1655-1663.
Newbold R. 1995 Cellular and molecular effects of developmental exposure to diethylstilbestrol:
Implications for other environmental estrogens. Environ Health Perspect 103(Supp 7):83-87.
Newbold RR, and McLachlan JA. 1982 Vaginal adenosis and adenocarcinoma in mice exposed
prenatally or neonatally to diethylstilbestrol. Cancer Res 42:2003-2011.
Okasha M, Gunnell D, Holly J, Smith GD. 2002. Childhood growth and cancer. Best Practice
and Research Clinical Gastroenterology. 16:225-41.
28
-------
Olivera OA, Anderson LM, Diwan BA, Haines DC, Harbaugh SW, Moskal TJ, Jones AB, Rice
JM, Riggs CW, Logsdon D, Yuspa SD, and Poirier MC. 1997. Transplacental effects of 3'-
azido-3'-deoxythymidine (AZT): Tumorigenicity in mice and genotoxicity in mice and monkeys.
JNCI 89:1602-08.
Rice, JM and Ward, JM 1982. Age dependence of susceptibility to carinogenesis in the nervous
system. Ann NYAcadSci 381:274-89.
Robson H, Siebler T, Shalet SM, Williams GR. 2002. Interactions between GH, IGF-I,
glucocorticoids, and thyroid hormones during skeletal growth. Pediatr Res. 52:137-47.
Rogers S, 1950. Age of the host and other factors affecting the production with urethane of
pulmonary adenoms in mice. JExperimentalMedicine 93:427-49.
Ron E, 2003. Cancer risks from medical radiation. Health Phys. 85(l):47-59.
Rosenfield RL, Puberty in the Female and its Disorders. Pediatric Endocrinology, M. Sperling,
editor Saunders, Philadelphia, pages 455-518.
Russo J, Wilgus G, and Russo IH. 1979. Susceptibility of the mammary gland to carinogenesis.
I. Differentiation of the mammary gland as a determinant of tumor incidence and type of lesion.
Am JPathol 96:721-736.
Snodgrass SR. 1994. Hepatic toxic effects associated with valproate therapy. J Pediatr.
124(2):334.
Swenberg JA, Fedtke N, Fishbein L 1992. Age-related differences in DNA adduct formation and
carcinogenesis of vinyl chloride in rats. Similarities and Differences Between Children and
Adults: Implications for Risk Assessment. P.S. Guzelian, C.J. Henry, and S.S. Olin, Eds. ILSI
Press, Washington D.C.
Thomale J, Hochleitner K, and Rajewski M.F. 1994. Differential formation and repair of the
mutagenic DNA alkylation product O6-alkylguanine in transcribed and nontranscribed genes of
the rat. JBiol Chem 269:1681-86.
Thompson DE, Mabuchi K, Ron E, Soda M, Tokunaga M, Ochikubo S, Sugimoto S, Ikeda T,
Terasaki M, Izumi S, et al. 1994. Cancer incidence in atomic bomb survivors. Part II: Solid
tumors, 1958-1987. RadiatRes. 137(2 Suppl):S17-67.
UNSCEAR 2000 Report, Vol. 2, 2000.SOURCES AND EFFECTS OF IONIZING
RADIATION, Volume II: EFFECTS, United Nations Scientific Committee on the Effects of
Atomic Radiation UNSCEAR 2000 Report to the General Assembly, with scientific annexes.
http://www.unscear.org/reports/2000_2.html.
29
-------
Urso P, and Gengozian N. 1982. Alterations in the humoral immune response and tumor
frequencies in mice exposed to benzo(a)pyrene and X-rays before or after birth. J Toxicol
Environ Health 10:817-835.
USEPA. 2003a. Draft Final Guidelines for Carcinogen Risk Assessment. United States
Environmental Protection Agency, http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=55445.
USEPA. 2003b. Draft Supplemental Guidance for Assessing Cancer Susceptibility from Early-
Life Exposure to Carcinogens. United States Environmental Protection Agency.
http://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=55446.
USEPA. 2001. Draft Dioxin Reassessment, Draft Exposure and Human Health Reassessment of
2,3,7,8-Tetrachlorodibenzo:p-Dioxin (TCDD) and Related Compounds,
http://cfpub.epa.gov/ncea/cfm/partland2.cfm?ActType=default.
Vesselinovitch SD, Mihailovich N. 1983. Kinetics of diethylnitrosamine hepatocarcinogenesis
in the infant mouse.Cancer Res. 1983 Sep;43(9):4253-9.
Waalkes MP, Ward JM, Liu J, Diwan BA. 2003. Transplacental carcinogenicity of inorganic
arsenic in the drinking water: induction of hepatic, ovarian, pulmonary, and adrenal tumors in
mice. ToxicolAppl Pharmacol 186:7-17.
Walters MA. 1966. The induction of tumours by the injection of 9,10-dimethyl-l,2-
benzanthracene (DMBA) into newborn sucking and young adult mice. A dose response study.
BrJCancer 20:148-160.
Warner M, Eskenazi B, Mocarelli P, Gerthoux PM, Samuels S, Needham L, Patterson D,
Brambilla P. 2002. Serum dioxin concentration and breast cancer risk in Seveso Women's
Health Study. Environ Health Perspect 110:625-628.
Whyatt RM, Bell DA, Jedrychowski W, Santella RM, Garte SJ, Cosma G, Manchester DK,
Young TL, Cooper TB, Ottman R, Perera FP. 1998. Polycyclic aromatic hydrocarbon-DNA
adducts in human placenta and modulation by CYPlal induction and genotype. Carcinogenesis
19(8): 1389-92.
30
-------
APPENDIX 1
Suggested Additional Studies for Quantitative Analysis
Drew RT, Boorman GA, Haseman JK, McConnell EE, Busey WM, and Moore JA. 1983. The
effect of age and exposure duration on cancer induction by a known carcinogen in rats, mice, and
hamsters. ToxicolAppliedPharmacol 68:120-130.
Druckrey H and Lange A. 1972. Carcinogenicity of azoxymethane dependent on age in BD rats.
FedProc 31(5): 1482-4.
Druckrey H. 1973. Chemical structure and action in transplacental carcinogenesis and
teratogenesis. IARC Sci Publ 4:45-58.
Gray R, Peto R, Brantom P, and Grasso P. 1991. Chronic nitrosamine ingestion in 1040 rodents:
the effect of the choice of nitrosamine, the species studied, and the age of starting exposure.
Cancer Res (23 Pt 2): 6470-91.
Jurgelski W Jr., Hudson P, Falk HL. 1979. Tissue differentiation and susceptibility to embryonal
tumor induction by ethylnitrosourea in the opossum. NCIMongraph 51:123-158.
Neal J and Rigdon RH. 1967. Gastric tumors in mice fed benzo(a)pyrene: A quantitative study.
Texas Reports on Biology and Medicine 25:553-557.
O'Gara RW and Kelly, MG. 1963. Comparative susceptibility of newborn, weanling, and adult
mice to tumor induction by 3-methylcholanthrene and dibenz(ah)anthracene. Proceedings of the
Association of Cancer Research 4:49.
Peto R, Gray R, Brantom P and Grasso P. 1984. Nitrosamine carcinogenesis in 5120 rodents:
chronic administration of sixteen different concentrations on NDEA, NDMA, NPYR, and NPIP
in the water of 4440 inbred rats, with parallel studies on NDEA alone of the effect of age of
starting (3,6, or 20 weeks) and of species (rats, mice, or hamsters). IARC Sci Publ 57:627-665.
Pietra G, Rappaport H, Shubik P. 1961. The effects of carcinogenic chemicals in newborn mice.
Cancer 14:308-317.
Rice, JM and Ward, JM 1982. Age dependence of susceptibility to carinogenesis in the nervous
system. AnnNYAcadSci 381:274-89.
Rice IM. 1979. Problems and perspectives in perinatal carcinogenesis: a summary of the
conference. NCI Monographs 51:271
Toth B. 1968. A critical review of experiments in chemical carcinogenesis using newborn
animals. Cancer Res 28(4):727-38.
-------
Vesselinovitch SD, Rao KV, and Mihailovitch N. 1975. Factors modulating benzidine
carinogenicity bioassay. Cancer Res 35(10):2814-19.
Vesselinovitch SD, Rao KV, Mihailovich N, Rice JM, and Lombard LS. 1974. Development of
broad spectrum of tumors by ethylnitrosourea in mice and the modifying role of age, sex, and
strain. Cancer Res 34(10):2530-38.
32
-------
APPENDIX 2
Initiation-Promotion Studies in Neonatal Mice
Klaunig JE. PereiraMA. Ruch RJ. 1988.Weghorst CM. Dose-response relationship of
diethylnitrosamine-initiated tumors in neonatal balb/c mice: effect of phenobarbital promotion.
Toxicologic Pathology. 16(3):381-5.
Lee GH. Ooasa T. Osanai M. 1998. Mechanism of the paradoxical, inhibitory effect of
Phenobarbital on hepatocarcinogenesis initiated in infant B6C3F1 mice with diethylnitrosamine.
Cancer Research. 58(8): 1665-9.
Siglin JC. Weghorst CM. Rodwell DE. Klaunig JE. 1995. Gender- dependent differences in
hepatic tumor promotion in diethylnitrosamine initiated infant B6C3F1 mice by alpha-
hexachlorocyclohexane. Journal of Toxicology & Environmental Health. 44(2):235-45.
Weghorst CM. DevorDE. 1994. Henneman JR. Ward JM. Promotion of hepatocellular foci
and adenomas by di(2-ethylhexyl) phthalate and phenobarbital in C3FI/HeNCr mice following
exposure to N-nitrosodiethylamine at 15 days of age. Experimental & Toxicologic Pathology.
45(7):423-31.
Weghorst CM. PereiraMA. Klaunig JE. 1989. Strain differences in hepatic tumor promotion
by phenobarbital in diethylnitrosamine- and dimethylnitrosamine-initiated infant male mice.
Carcinogenesis. 10(8): 1409-12.
Weghorst CM. PereiraMA. Klaunig JE. 1989. Strain differences in hepatic tumor promotion
by phenobarbital in diethylnitrosamine- and dimethylnitrosamine-initiated infant male mice.
Carcinogenesis. 10(8): 1409-12.
Weghorst CM. Klaunig JE. 1989. Phenobarbital promotion in diethylnitrosamine-initiated
infant B6C3F1 mice: influence of gender. Carcinogenesis. 10(3):609-12.
33
------- |