United States EPA Science Advisory EPA-SAB-EHC-03-002
Environmental Board (1400A) December 2002
Protection Agency Washington, DC wwiv.epa.gov/sab
4>EPA Review of Draft
Trichloroethylene
Health Risk
Assessment: Synthesis
and Characterization:
An EPA Science
Advisory Board Report
A REVIEW BY THE TCE
REVIEW PANEL OF THE
ENVIRONMENTAL HEALTH
COMMITTEE OF THE US EPA
SCIENCE ADVISORY BOARD
(SAB)
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
December 2, 2002
OFFICE OF THE ADMINISTRATOR
SCIENCE ADVISORY BOARD
EPA-SAB-EHC-03-002
The Honorable Christine Todd Whitman
Administrator
U.S. Environmental Protection Agency
1200 Pennsylvania Avenue, NW
Washington, DC 20460
Subj ect: Review of Draft Trichloroethylene Health Risk Assessment:
Synthesis and Characterization: An EPA Science Advisory Board Report
Dear Governor Whitman:
A Panel (Trichloroethylene Health Risk Assessment: Synthesis and Characterization
Review Panel) of the U.S. EPA Science Advisory Board's Environmental Health Committee met
on June 18-19, 2002 to review the Agency's draft human health assessment, "Trichloroethylene
Health Risk Assessment: Synthesis and Characterization."
The Board advises the Agency to move ahead to revise and complete this important
assessment. The assessment addresses a chemical, trichloroethylene (TCE), significant for being
a nearly ubiquitous environmental contaminant in both air and water, being a common
contaminant at Superfund sites, and because it is "listed" in many Federal statutes and
regulations. The draft assessment is also important because it sets new precedents for risk
assessment at EPA. We believe the draft assessment is a good starting point for completing the
risk assessment of TCE. The Panel commends the Agency for its effort and advises it to proceed
to revise and finalize the draft assessment as quickly as it can address the advice provided in this
report.
The Board commends the Agency for its groundbreaking work in this draft assessment in
several important new areas in risk assessment: a) risk to children and other susceptible
populations; b) cumulative risk; c) examination of multiple kinds of evidence including evidence
about physiological and molecular modes of action; d) the assessment of the health risks
associated with the many metabolites of TCE; e) the use of biologically-based modeling; f) the
explicit recognition and acknowledgment of uncertainties in the risk analysis; and g) the
consideration of multiple data sets from animal and human studies to derive cancer slope factors.
Although the Board welcomes this effort, it also cautions the Agency that the new areas
explored involve considerable uncertainty. Progress in reducing these uncertainties will be an
evolutionary process that will necessitate advancements in scientific research and analysis. The
Board also notes that there is a need for Agency-wide guidance in many of the areas explored in
the draft assessment. The Agency should develop consistent policies across program areas on
protection of children and other vulnerable populations, cumulative risk, and aggregate risk
Because the draft assessment breaks ground in several areas and sets important
precedents, there is a need to strengthen the rigor of the discussion in the revised assessment so
that the basis for all derived values is transparent and clearly supported by the available data.
The Board notes that public comments have raised valid concerns the Agency should carefully
address. The Panel urges the Agency to review and address the public comments it received on
-------
the review draft, especially those from experts who had conducted research related to the
assessment of TCE's health risks and whose reviews had been published in a supplemental issue
of Environmental Health Perspectives (Volume 108, Supplement 2, May 2000).
The Board notes five key substantive areas for the Agency to address: a) the need to
strengthen and expand the use of epidemiology data to update the uncertainty analysis, to
incorporate new studies, and to focus on first tier studies and case-control studies that
specifically address TCE exposure; b) the need to develop a more formal method for selecting
and weighing evidence and communicating those decisions, when information comes from
multiple lines of evidence; c) the need for a more detailed explanation for the Agency's treating
cancer mode of action in a linear way, even as the Board notes that the Agency had provided
clear criteria for the choice of a linear approach; d) the need to explain the derivation of the RfD
and RfC study-by-study, endpoint-by-endpoint; e) the need to quantify and provide more
explicit justification for the background exposures to be included in uncertainty factors and
incorporated in the TCE assessment; and f) the need to be explicit about the assumptions
underlying its analyses.
In addition, to strengthen and provide confidence in the Agency's assessment, the Board
has identified several needs. There is the need for a summary paragraph in each section
describing the Agency position/conclusion and a clear description in each section of the
scientific basis for those choices and other alternatives considered. The Agency should develop
a new children's chapter to discuss the pharmacokinetic, pharmacodynamic, and risk assessment
conclusions in a comprehensive unified chapter focused on children's health. The new children's
chapter should offer a model for other draft assessments to follow and would integrate
information about specific aspects of risks to children's health as discussed in the other sections
of the assessment. And finally, and most importantly, there is a need to enable others to
reproduce the calculations and models on which the assessment's conclusions are based.
Therefore, we strongly advise the Agency to reference original papers on key issues, not only
review articles, and to provide access to data, documentation, and results of intermediate
calculations from which the Agency's results can be recreated.
In closing, the Board greatly appreciates the efforts of the Agency's staff developing this
groundbreaking assessment. We look forward to your response to this report.
Sincerely,
/Signed/ /Signed/
Dr. William Glaze, Chair Dr. Henry Anderson, Chair
EPA Science Advisory Board Environmental health Committee, and
Trichloroethylene Health Risk Assessment: Synthesis
and Characterization Review Panel
EPA Science Advisory Board
-------
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.
Distribution and Availability: This EPA Science Advisory Board report is provided to the EPA
Administrator, senior Agency management, appropriate program staff, interested members of the
public, and is posted on the SAB website (www.epa.gov/sab). Information on its availability is
also provided in the SAB's monthly newsletter (Happenings at the Science Advisory Board).
Additional copies and further information are available from the SAB Staff [US EPA Science
Advisory Board (1400A), 1200 Pennsylvania Avenue, NW, Washington, DC 20460-0001; 202-
564-4533.
-------
ABSTRACT
A Panel of the U.S. EPA Science Advisory Board's Environmental Health Committee
reviewed the Agency's draft human health assessment, "Trichloroethylene Health Risk
Assessment: Synthesis and Characterization." It responded to Agency charge questions that
touched on how the document addressed several important new areas in risk assessment: a) risk
to children and other susceptible populations; b) cumulative risk; c) examination of multiple
kinds of evidence including evidence about physiological and molecular modes of action; d) the
assessment of the health risks associated with the many metabolites of TCE; e) the use of
biologically-based modeling; f) the explicit recognition and acknowledgment of uncertainties in
the risk analysis; and g) the consideration of multiple data sets from animal and human studies to
derive cancer slope factors.
The Board commends the Agency for its groundbreaking work in this draft and advises
the Agency to move ahead to revise and complete this important assessment.
The Board also cautions the Agency that the new areas explored involve considerable
uncertainty. Progress in reducing these uncertainties will be an evolutionary process that will
necessitate advancements in scientific research and analysis. The Board also notes that there is a
need for Agency-wide guidance in many of the areas explored in the draft assessment. The
Agency should develop consistent policies across program areas on protection of children and
other vulnerable populations, cumulative risk, and aggregate risk. Because the draft assessment
breaks ground in several areas and sets important precedents, there is also a need to strengthen
the rigor of the discussion in the revised assessment so that the basis for all derived values is
transparent and clearly supported by the available data. The Board identifies five key
substantive areas for the Agency to address and also ways to revise the document to strengthen
and provide confidence in the Agency's assessment.
Keywords: TCE, Risk Assessment, Human Health, Children's Health, Pharmacokinetic
Modeling
11
-------
U.S. Environmental Protection Agency
Science Advisory Board
Environmental Health Committee
Trichloroethylene Health Risk Assessment:
Synthesis and Characterization Review Panel*
CHAIR
Dr. Henry Anderson, Wisconsin Division of Public Health, Madison, WI
Also Member: Integrated Human Exposure Committee
Executive Committee
EHC MEMBERS
Dr. Dale Hattis, Clark University, Worcester, MA
Dr. David Hoel, Medical University of South Carolina, Charleston, SC
Dr. George Lambert, Robert Wood Johnson Medical School/ University of Medicine and
Dentistry of New Jersey, Piscataway, NJ
Dr. Grace LeMasters, University of Cincinnati, Cincinnati, OH
Dr. Abby Li, Monsanto Company, St. Louis, MO
Dr. Ulrike Luderer, University of California at Irvine, Irvine, C A
CONSULTANTS
Dr. Susan J. Borghoff, CUT Centers for Health Research, Research Triangle Park, NC
Dr. Lutz Edler, German Cancer Research Center, Heidelberg, Germany
Dr. Michael McClain, McClain Associates, Randolph, NJ
Dr. Gina Solomon, Natural Resources Defense Council, San Francisco, CA
Dr. Robin Whyatt, Columbia University, New York, NY
Dr. Raymond Yang, Colorado State University, Fort Collins, CO
FEDERAL EXPERTS
Dr. Aaron Blair, National Institutes of Health, Bethesda, MD
Dr. Ronald Melnick, National Institute of Environmental Health Sciences, Research Triangle
Park, NC
EPA SCIENCE ADVISORY BOARD STAFF
Dr. Angela Nugent, Washington, DC
Ms. Diana Pozun, Washington, DC
Ms Betty Fortune, Washington, DC
in
-------
TABLE OF CONTENTS
1. EXECUTIVE SUMMARY OF RESPONSES TO CHARGE QUESTIONS 1
2. INTRODUCTION 4
2.1. Background 4
2.2. Process for Developing this Report 5
3. CHARGE QUESTION 1 6
3.1. Agency Charge Question and Suggested Areas for Inquiry 6
3.2. Panel Response 6
4. CHARGE QUESTION 2 10
4.1. Agency Charge Question and Suggested Areas for Inquiry 10
4.2. Panel Response 10
4.2.1. Cancer Classification for TCE 10
4.2.2. Human Epidemiological Studies 11
4.2.3. Animal Toxicology 13
5. CHARGE QUESTION 3 16
5.1. Agency Charge Question and Suggested Areas for Inquiry 16
5.2. Panel Response 16
6. CHARGE QUESTION 4 18
6.1. Agency Charge Question and Suggested Areas for Inquiry 18
6.2. Panel Response 18
6.2.1. Multiple Critical Effects: Does the draft assessment adequately
characterize the data at each site of toxicity and focus on an
appropriate subset of critical effects? 18
6.2.2. Modes of Action of TCE Toxicity 20
6.2.3. Uncertainty Factors - Areas of Agreement and Differences Within
the Panel 20
7. CHARGE QUESTION 5 25
7.1. Agency Charge Question and Suggested Areas for Inquiry 25
7.2.1. Panel Response 25
7.2.1.1 Clarification of the Cancer Slope Factors 25
7.2.1.2 Suitability and Use of the Cancer Slope Factors 26
7.2.1.3 Improved Mediation of the Cancer Slope Factors 27
7.2.2. Further Studies to Be Included 28
7.2.3. Linear or Nonlinear Approach 28
7.2.4. Sensitive Populations 29
7.2.5. For Further Consideration 29
8. CHARGE QUESTION 6 30
8.1. Agency Charge Question and Suggested Areas for Inquiry 30
8.2. Panel Response 30
8.2.1. Modeling 31
8.2.2. Uncertainty Analysis 32
8.2.3. Data Availability 33
8.2.4. Markov-Chain Monte Carlo (MCMC) 34
IV
-------
9. CHARGE QUESTION 7 35
9.1. Agency Charge Question and Suggested Areas for Inquiry 35
9.2. Panel Response 35
10. CHARGE QUESTION 8 38
10.1. Agency Charge Question and Suggested Areas for Inquiry 38
10.2. Panel Response 38
11. CHARGE QUESTION 9 39
11.1. Agency Charge Question and Suggested Areas for Inquiry 39
11.2. Panel Response 39
11.2.1. Major Summary Consensus Points of the Panel 39
11.2.2 Background to the Panel's Conclusions 40
11.2.3. How the Draft Assessment Can Be Improved 44
REFERENCES R-l
APPENDIX A
SPECIFIC PANEL COMMENTS ON THE AGENCY'S ASSESSMENT OF
NONCANCER ENDPOINTS A-l
1. Specific Comments on Hazard Characterization for Noncancer Endpoints A-l
1.1. Liver Effects A-l
1.2. Kidney Effects A-l
1.3. Developmental Effects A-2
1.4. Neurotoxicity Effects A-3
1.5. Endocrine System Effects and Reproductive Toxicity Effects A-3
2. Specific Comments on Uncertainty Factors for NonCancer Endpoints A-4
2.1. Human Variation A-4
2.2. Animal-to-Human Uncertainty A-5
2.3. Subchronic-to-Chronic Uncertainty A-5
2.4. LOAEL-to-NOAEL Uncertainty A-6
2.5. Other Factors A-9
APPENDIX B
RECENT RESEARCH ON GENERAL PHARMACOKINETIC DIFFERENCES
BETWEEN CHILDREN OF VARIOUS AGES AND ADULTS B-l
APPENDIX C
Biosketches of Members of the US EPA Science Advisory Board (SAB)
Trichloroethylene Health Risk Assessment: Synthesis and Characterization
Review Panel (TCE Review Panel) C-l
-------
1. EXECUTIVE SUMMARY OF RESPONSES TO CHARGE QUESTIONS
a) Charge Question 1: Does the assessment adequately discuss the likelihood that
trichloroethylene (TCE) acts through multiple metabolites and multiple modes of action?
Panel's Response: EPA should be commended for its efforts to date to evaluate a wide
variety of hypotheses for the carcinogenic and other toxic effects of TCE. The draft assessment
could be enhanced by including additional selective quantitative analysis and further evaluation
of dose-response relationships, especially relationships that may clarify the role of different
metabolites in the toxicity of TCE at human exposure levels. Panel members offered several
examples of quantitative analyses that the Agency could use to improve the document.
b) Charge Question 2: Is the cancer weight-of-evidence characterization adequately
supported?
Panel's Response: The Panel commends EPA for compiling an extensive array of
scientific literature that included over 80 epidemiological studies and hundreds of toxicological
and mechanistic studies and for characterizing the evidence relatively clearly and cohesively.
The Panel feels that the Agency's overall qualitative cancer risk characterization was reasonable
based on: a) significant experimental evidence showing tumors at multiple sites in two species
(rats and mice); b) epidemiologic evidence in humans showing associations between TCE
exposure and several cancers including at several of the same sites seen in animal bioassays; and
c) mechanistic data indicating relevance of experimental findings to humans. There was a
suggestion that EPA clarify more explicitly for the public what is meant by a "weight-of-
evidence characterization."
The Panel also advises the Agency to improve the characterization of the cancer weight-
of-evidence by evaluating human and animal studies more rigorously, explicitly using Agency
criteria for evaluating those studies. On a related issue, several panel members noted that the
assessment could be strengthened if the Agency emphasized and integrated information about
dose-response and if it presented a coherent analysis of quantitative information available on the
mode-of-action of metabolites. The Panel also advises the Agency to provide a more thorough
discussion and critical evaluation of the conflicting human epidemiological evidence for kidney
tumors.
The Panel suggests that it is important to provide both upper confidence limit estimates
and the mean "expected value" estimate when developing risk ranges to give users confidence
intervals. It notes that it will also be important for the Agency to provide risk management
guidance that might indicate when it is most appropriate to use mean values, or when to use high
end values with confidence intervals
c) Charge Question 3: A new feature of the cancer database is molecular information on
the von Hippel-Lindau tumor suppressor gene. Is this information adequately discussed and are
the conclusions appropriate.
Panel's Response: The consensus of the panel is that the discussion in the draft
assessment is generally appropriate. The panel generally agrees that EPA is wise not to regard
the evidence as entirely conclusive pending independent confirmation. The discussion in the
draft assessment might be improved by including some additional comparative observations
from kidney cancers not in the TCE exposed workers.
-------
d) Charge Question 4: Does the assessment adequately discuss the use of multiple critical
effects in developing an oral reference dose (RfD) and inhalation reference concentration (RfC)
for effects other than cancer? Are the uncertainty factors well discussed and well supported?
Panel's Response: The Panel commends the Agency for consideration of multiple
noncancer endpoints in both the general discussion and in the derivation of the RfD and RfC.
TCE clearly has important hepatotoxic, nephrotoxic, neurotoxic, immunologic, developmental
and reproductive effects that should be considered in the derivation of the RfD and RfC. The use
of multiple critical effects increases one's confidence that the point of departure dose is at the
low end of doses at which adverse effects can be observed.
Some panel members suggested that the characterization of the data at each site of
toxicity could be strengthened considerably. The Panel recognizes that a lengthy dissertation of
each study cited by EPA would be counterproductive. However, in the opinion of some
panelists, the current discussion lacks the type of critical analysis and discussion of the weight-
of-evidence that is necessary to understand the Agency's rationale for selection of endpoints,
level of concern, dose-response extrapolation, effect of time-duration on key endpoints, and
application of uncertainty factors. Other panel members thought the discussion of non-cancer
effects and discussions surrounding the development of the RfD and RfC were quite good and
that with relatively limited additional clarification the non-cancer section of the draft assessment
would be complete.
e) Charge Question 5: Does the assessment adequately discuss the derivation of a range
of estimates for the cancer risk? Are there any studies that should/should not have been
included?
Panel's Response: The Panel commends EPA for the derivation of a set of cancer risk
estimates or cancer slope factors (CSF) for TCE in the draft assessment. The presentation of a
range of estimates is a step forward for EPA towards a more explicit and more quantitative
representation of the substantial uncertainties in estimates of cancer risks.
The Panel identifies a key study (Hansen et. al., 2001) that should be included in the
revision of the draft assessment. The Panel advises that, where epidemiological studies are the
basis of risk estimates, EPA should select the broadest possible array of studies for each endpoint
taking into consideration study design, availability of exposure estimates, and the goal of
protecting health. The Panel commends the Agency for providing sections on sensitive
populations and cumulative risks and added several suggestions for strengthening quantitative
aspects of the risk assessment methodology that are important for the refinement of the risk
assessment of TCE.
f) Charge Question 6: Please comment on the use of calibrated models and uncertainty
analysis to address the question of pharmacokinetic model uncertainty.
Panel's Response: The Panel commends the Agency's inclusion of physiologically-based
pharmacokinetic models and its explicit recognition of model uncertainty in the draft assessment.
It advises the Agency to explain the modeling methods more clearly, and to make the models,
data, and assumptions used available, so that the Agency's results can be reproduced. The Panel
calls on the Agency to compare the two calibrated models used and to show how the models and
analyses compare and relate to one another. The Panel advises the Agency to highlight the
impact of the uncertainty analysis on the dose estimates of the different models and on the dose-
response analysis and to explain the differences between the models and the ranges of
uncertainty.
-------
g) Charge Question 7: Is it appropriate to consider background exposures and other
characteristics of an exposed population as modulating the risk of TCE exposure in that
population?
Panel's Response: The Panel is pleased that the Agency has taken the first steps of
including the issue of cumulative risk in a health risk assessment. Although there was agreement
that background exposures to TCE and/or metabolites is a very important issue, there was
disagreement about whether the RfD, and the uncertainty factor used to derive it, should be the
method by which this background exposure is addressed. The Panel agrees, however, that
regardless of EPA's final policy decision on whether or not to include an additional uncertainty
factor in the RfD for background exposure, more attention and detail is needed to provide a
rationale for the Agency's use of such an uncertainty factor.
h) Charge Question 8: Do the data support identifying risk factors that may be associated
with increased risks from TCE exposure? Are there any risk factors that should/should not have
been included?
Panel's Response: The Panel finds that the data support identifying numerous risk factors
that may be associated with increased risks to susceptible subpopulations from TCE exposure.
The EPA draft assessment has done a good job identifying the general areas of concern related to
prenatal, reproductive and developmental risks associated with TCE exposure, especially given
the level of information known to date. The Panel agrees with the draft risk assessment's
identification of multiple background exposures to ethanol, TCE, and its metabolites, and other
chemical solvent mixtures as factors that may be associated with increased risks.
i) Charge Question 9: Do the data support the possibility that TCE can affect children and
adults differently? Should this be reflected in the quantitative assessment?
Panel's Response: The Panel reached consensus on the following conclusions related to
this charge question: a) the data presented suggest that TCE can affect children differently than
adults, although there is a very limited database of TCE in children due to lack of directly
applicable studies; b) the draft does not explicitly discuss whether or not the uncertainty factors
adequately address risk to children or attempt to develop toxicity values that take children into
consideration; c) the Panel advises the Agency to develop a stand-alone comprehensive
children's chapter that discusses all the children's issues, including exposure, susceptibility
during pregnancy, pharmacokinetics, and pharmacodynamics, in addition to discussing
developmental animal and children data in every section; and d) the Panel advises the Agency to
support statements about differences between children and adults with a quantitative discussion,
whenever possible. Although the Panel differs on the question of whether the Agency should
add a quantitative uncertainty factor to protect children above the composite uncertainty factor
already in the draft assessment, it did advise the Agency to address this issue explicitly and to
clarify how such a factor would relate to other uncertainty factors used.
-------
2. INTRODUCTION
2.1. Background
The purpose of this report is to provide advice to the Agency in developing a final health
risk assessment for TCE. The SAB formed the Trichloroethylene Health Risk Assessment:
Synthesis and Characterization Review Panel to review a draft assessment dated August 2001
that EPA's Office of Research and Development (ORD) provided for external public comment
on September 19, 2001.l TCE is a major contaminant of concern in EPA's air, water, and waste
programs. EPA's regulatory program and regional offices have identified TCE as among the
highest priorities for a new assessment.
The Agency has noted, and the Panel acknowledges, that the draft assessment submitted
for review was shaped by several new developments in risk assessment. The practice of risk
assessment has been evolving from a focus on a single toxic effect of one pollutant in one
environmental medium toward integrated assessments covering multiple effects and multiple
media and incorporating information about mode of action, uncertainty, human variation, and
cumulative effects of multiple pollutants in different media. This evolution has responded to
recommendations of the National Research Council, whose recommendations have been
embraced in EPA's proposed cancer guidelines.
The TCE draft assessment breaks new ground in addressing the new dimensions of risk
assessment that EPA and others have advocated. The draft assessment discusses the possibility
that children, infants, and the developing fetus may differ from adults with respect to
susceptibility to TCE's toxic effects. The assessment also addresses cumulative risks by
discussing the implications of other chlorinated solvents and agents that have metabolic
pathways, potential modes of action, and toxic effects similar to TCE. The assessment
implements principles of the proposed cancer guidelines by emphasizing characterization
discussions, and by using information on mode-of-action and information on susceptible
populations to derive cancer slope factors and RfD and RfC values
The issues surrounding TCE are quite complex, with extensive information in some areas
and relatively little information in others. EPA's ORD initiated development of 16 peer-
reviewed state-of-the-science papers that were published in a special supplementary issue of the
journal Environmental Health Perspectives (May 2000). The Agency acknowledged that it drew
on those papers, plus some other key references, as scientific support for the draft assessment.
In the fall of 2001, EPA asked the SAB to convene a Panel to address the following draft
charge questions:
a) Does the assessment adequately discuss the likelihood that trichloroethylene
(TCE) acts through multiple metabolites and multiple modes of action?
b) Is the cancer weight-of-evidence characterization adequately supported?
1 U.S. Environmental Protection Agency, Trichloroethylene Health Risk Assessment: Synthesis and
Characterization, External Review Draft, August 2001, EPA/600/P-01/002A
-------
c) A new feature of the cancer database is molecular information on the von Hippel-
Lindau tumor suppressor gene. Is this information adequately discussed and are the conclusions
appropriate?
d) Does the assessment adequately discuss the use of multiple critical effects in
developing an oral reference dose (RfD) and inhalation reference concentration (RfC) for effects
other than cancer? Are the uncertainty factors well discussed and well supported?
e) Does the assessment adequately discuss the derivation of a range of estimates for
the cancer risk? Are there any studies that should/should not have been included?
f) Please comment on the use of calibrated models and uncertainty analysis to
address the question of pharmacokinetic model uncertainty.
g) Is it appropriate to consider background exposures and other characteristics of an
exposed population as modulating the risk of TCE exposure in that population?
h) Do the data support identifying risk factors that may be associated with increased
risks from TCE exposure? Are there any risk factors that should/should not have been included?
i) Do the data support the possibility that TCE can affect children and adults
differently? Should this be reflected in the quantitative assessment?
2.2. Process for Developing this Report
The SAB formed a special panel to address the Agency's charge questions. It was
composed of members of the SAB's Environmental Health Committee, augmented to provide
additional expertise needed to address the charge and to provide breadth of viewpoints on issues
key to the review. Panel members were added to provide expertise in the following areas: TCE
epidemiology; pharmacokinetic modeling; cancer toxicity biostatistics, and modeling; modes of
action at the molecular level; modes of action at the physiological level; differing perspectives
on how the toxicology database of information on TCE can be understood; and risk assessment
expertise. Biosketches of panel members can be found in Appendix C.
The Panel reviewed the Agency's draft assessment, along with supplementary
background information that included: a) Environmental Health Perspectives (May 2000); b)
Summary of Public Comments for EPA's Science Advisory Board; c) Log of public comments
for "Trichloroethylene (TCE) Health Risk Assessment Synthesis and Characterization;" and d)
over 800 pages of public comments.
The Panel held a public planning teleconference on June 5, 2002. At that meeting it
considered "areas of inquiry" suggested by the Agency to help guide the Panel's discussion of the
nine charge questions. These "areas of inquiry," suggested by the Agency, are included in the
different sections of the Panel's report below.
The Panel held a face-to-face public meeting in Washington D.C. on June 18-19, 2002,
and a public teleconference to discuss this report in draft form on July 18, 2002.
-------
3. CHARGE QUESTION 1
3.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Does the assessment adequately discuss the likelihood that
trichloroethylene (TCE) acts through multiple metabolites and multiple modes of action?
Suggested Areas for Inquiry: A prominent issue is the role of the metabolite DCA. One
view is that DCA has little or nothing to do with TCE's toxicity; another view is that more than
one metabolite (both TCA and DCA) could be responsible for TCE's effects in the liver. The
draft assessment discusses the evidence supporting both positions, as well as potential modes of
action and metabolites involved at other sites of toxicity. Does the draft assessment adequately
consider and characterize such information?
3.2. Panel Response
EPA should be commended for its efforts to date to evaluate a wide variety of hypotheses
for the carcinogenic and other toxic effects of TCE and its different metabolites. The draft
assessment could be enhanced by considering additional selective quantitative analysis and
further evaluation of dose-response relationships. Panel members offered several examples of
quantitative analyses that the Agency could use to improve the document.
a) The discussion of the potential roles of TCA vs. DCA in the causation of liver cancers
can be usefully informed by comparing the observed liver cancers in existing animal bioassays
with those that would be predicted based on:
1) calculations of the potency of these metabolites when administered separately
in similar animal model systems, and
2) pharmacokinetic modeling estimates of how much of each metabolite would be
produced from TCE under the bioassay conditions.
The Panel advises the Agency to base potency estimates of the metabolites resulting from direct
exposure, versus exposure from TCE metabolism on time-dependent liver dosimetry.
Some analysis along these lines seems to be contained in the state-of-the-science paper
by Chen (2000). In the paper by Bull et al. (2000), there is a discussion concerning the effective
level of TCE metabolites in blood following administration of TCE, TCA, and DCA that causes
liver tumors. It should be recognized that estimations of blood metabolite levels would not be
adequate for trans-species extrapolations, because of possible differences in the partition
coefficients in rodents versus humans.
Critical review of this information could be incorporated into Section 3.5 of the Agency's draft
assessment to help evaluate the likely contributions of the different metabolites to the liver tumor
observations.
b) TCA is metabolized to DCA yet the liver tumors induced by DCA display different
characteristics compared to the tumors induced by TCA (section 3.5.1.2). Was a comparison of
these tumor characteristics induced by TCA and DCA ever made to similar tumors produced by
TCE? Again, a quantitative discussion of the levels of these metabolites present following
administration of the metabolite or the parent TCE would be useful.
-------
c) Another example of a quantitative analysis suggested by a panel member relates to the
peroxisome proliferator hypothesis for the mode of action of TCE and/or its metabolites. For
this hypothesis for the mode of action, there should be a reasonably good quantitative correlation
between the peroxisome proliferation potency and the apparent carcinogenic potency of TCE and
possibly its relevant metabolites across species, genders, etc. From the Bull (2000) report in the
state-of-the science papers that TCE induces peroxisome proliferation in rats but not liver
tumors, and some other comments there, it appears that the correlation may not be very good. If
so, a quantitative comparative analysis will reveal that.
In the absence of such a demonstrated correlation between the potencies for causing
peroxisome proliferation and the potencies for causing liver cancer, several members of the
Panel do not completely understand the relatively favorable attention given to this possibility in
the draft assessment. The Panel believes the potential mode of action discussion should include
this hypothesis in the review. However some panel members suggested that EPA should
consider giving it somewhat less weight than the possibility that increases in cell replication
rates interact with some amount of classical mutagenic/clastogenic activity by highly reactive
intermediates such as the TCE-epoxide or metabolites of the metabolites TCA and DC A (see
suggestion below that additional quantitative analysis of genotoxic hypotheses be included to the
extent possible based on available data). One panelist suggested that the open question of the
possible contribution of the release of free TCE-epoxide and resulting DNA reactions be at least
mentioned in the revised draft assessment. It also should be noted that the mechanisms by which
TCE metabolites affect cell cycling have not been sufficiently worked out so as to provide a
clear indication on how these changes in relation to metabolite dosimetry and as a function of
age relate to the probability of liver tumor development.
d) In section 3.5.1.2, the cell-signaling mode of action for TCA and DCA is discussed.
The only paper that is cited that supports this mode of action is by Bull (2000). Are there other
studies in which this potential mode of action is supported? If so, they should be included. The
responses observed for TCA and DCA are not compared specifically to what occurs following
doses of TCE that cause observed liver tumors. Are these data available? If not, the Panel
advises EPA to present the doses of TCA and DCA administered to cause these responses along
with a prediction [utilizing physiologically-based pharmacokinetic (PBPK) models] of the
dosimetry of these metabolites in the liver following administered dose of the metabolite or TCE.
e) A more quantitative discussion of the genotoxic data relative to cytotoxicity and tumor
formation is especially important in clarifying the genotoxic contributions to carcinogenic action
and in understanding the shape of the dose-response curve at human exposure levels. The
Agency's characterization of TCE and its major metabolites as "weak" mutagens/clastogens in
various test systems implies that the observed rather low potency for carcinogenic activity by
TCE is somehow not consistent with a primary genetic mode of action because TCE and its
metabolites do not have what might be called "strong" potency for mutagenic endpoints. The
use of terms such as "weak" and "strong" in this context is vague and open to misinterpretaton.
If arguments of this sort are to be used by the Agency, the Panel advises EPA to base those
arguments on quantitative analysis of the correlation between mutagenic potency in specific test
systems used for TCE and its metabolites and carcinogenic potencies for chemicals (perhaps
related chemicals) as conventionally determined by EPA procedures. Readers can then assess
the usefulness of such correlations for supporting inferences about the mutagenic/clastogenic
potency of major TCE metabolites and the potential for a genetic mode of action. The general
impression of some panel members is that such correlations, while present, are imprecise. They
will not support a strong inference that relatively low mutagenic/clastogenic potency by major
TCE metabolites is inconsistent with a genetic mode of action for the relatively low carcinogenic
-------
potency of TCE, compared with other related small-molecular-weight organic chemical
carcinogens.
An important issue concerned the mechanism for liver cancer. The Panel advises the
Agency to explore hypotheses for both non-genotoxic and genotoxic mechanisms for liver
cancer, considering effects at different dose levels. The Panel advises the Agency to consider
the information and hypotheses provided in the state-of-the-science papers (e.g., Bull, 2000) and
other relevant literature.
Some other comments by panel members suggested various types of summaries to
strengthen the draft assessment. In particular,
a) It would be useful to include a table summarizing evidence for and against the
potential modes of action by which TCE causes liver, kidney and lung cancer. This would
include which metabolite has been demonstrated or is suspected to operate through which
specific mode(s) of action and at what exposure levels.
b) In discussing the various metabolites and their related reactions, a comprehensive
metabolic pathway should be given. There was the mention of Figure 2-1 (page 2-1, line 2) in
which TCE metabolic pathways and those from chemicals sharing some TCE metabolites were
presumably given. However, Figure 2-1 is missing from the draft assessment. This should be
rectified. Figure 3-1 on page 3-57 achieved part of that purpose; however, the sharing of part of
the metabolic pathway by other chemicals should be explicitly indicated.
A summary sentence or paragraph would often be helpful at the end of long text
discussions of alternative hypotheses in order to clarify EPA's overall conclusion about likely
modes of action for different metabolites and toxic effects.
The Panel notes that the current draft assessment, in discussing the effects of metabolic
interactions, emphasizes the possibility that these interactions will increase toxicity. This is not
necessarily true. Various alternatives should be discussed. Since it is commonly accepted that
TCE metabolites are the more toxic species, the overall TCE toxicity might decrease to the
extent that this would lead to some increase in the exhalation or urinary excretion of unchanged
TCE. On the other hand, if TCE concentration increases in the blood because the P450 pathways
are partially saturated, the amount of TCE processed via the GST pathway would be expected to
increase, leading to greater internal exposure to renal toxic metabolites. The Panel advises the
Agency to evaluate where evidence exists regarding such competitive inhibition of TCE.
The Panel also notes that, although the draft assessment did a very good job of exploring
different pathways, there may have been too great a tendency to focus on the individual actions
of particular metabolites, in part because key informative experiments have most often been
done by administering either TCE itself or specific metabolites by themselves. The draft
assessment, however, does usefully note that no one metabolite may be totally responsible for
specific toxic actions. Therefore, there should be continuing research both on pharmacokinetic
and pharmacodynamic interactions between TCE, TCE metabolites, and possibly other
environmental toxicants.
The Panel recognizes that, given the current state of knowledge, there is uncertainty
about which metabolites cause specific adverse effects and in the sequence of biological changes
that lead to tumor development. Consequently, it is recognized that there would be substantial
uncertainties associated with extrapolations of hypothetical mechanisms (or modes of action
hypotheses) across species.
-------
In conclusion, the Panel reiterates its commendation of EPA for an extensive evaluation
of different modes of actions in the current document. However, the Panel strongly advises the
Agency to add a more thorough quantitative evaluation of dose response relationships and
dosimetry to its discussion of the role of different metabolites and multiple modes of action.
Quantitative data are available in the literature that can improve characterization of mode of
action in terms of cross-species and low dose extrapolation. More extensive use of this
information can improve scientific understanding and strengthen the basis for decisions on
cancer classification and the final risk assessment approach.
-------
4. CHARGE QUESTION 2
4.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Is the cancer weight-of-evidence characterization adequately
supported?
Suggested Areas for Inquiry: The cancer characterization is based on both
epidemiological and animal studies. Does the draft assessment adequately characterize the
strength of the epidemiologic evidence and adequately address questions concerning the analysis
by Wartenberg et al. (EHP 2000) and its inclusion of the Henschler study? Does the draft
assessment adequately present and consider the animal evidence of tumors at multiple sites and
its relevance to humans?
4.2. Panel Response
The Panel commends EPA for compiling an extensive array of scientific literature that
included over 80 epidemiological studies and hundreds of toxicological and mechanistic studies
and for characterizing the evidence relatively clearly and cohesively. The Panel feels that the
Agency's overall qualitative cancer risk characterization was reasonable based on: a) significant
experimental evidence showing tumors at multiple sites in two species (rats and mice); b)
epidemiologic evidence in humans showing associations between TCE exposure and several
cancers including at several of the same sites seen in animal bioassays; and c) mechanistic data
indicating relevance of experimental findings to humans. There was a suggestion that EPA
clarify more explicitly for the public what is meant by a "weight-of-evidence characterization."
The Panel also advises the Agency to improve the characterization of the cancer weight-
of-evidence by evaluating human and animal studies more rigorously using established Agency
criteria for evaluating studies of those types. On a related issue, several panel members noted
that the assessment could be strengthened if the Agency emphasized and integrated information
about dose-response and if it presented a coherent analysis of quantitative information available
on the mode-of-action of metabolites. The Panel also advises the Agency to provide a more
thorough discussion and critical evaluation of the conflicting human epidemiological evidence
for kidney tumors.
4.2.1. Cancer Classification for TCE
The Panel did not receive a charge question from the Agency regarding the cancer
classification for TCE, and thus it did not discuss that topic at length or agree on the cancer
classification. Instead, the Panel agrees on the need for the Agency to evaluate the different lines
of evidence for each tumor type in a more integrated manner. Some members advised that the
Agency give special attention to more rigorous quantitative evaluation of exposure levels in
human and animal experiments. Specific recommendations are presented in the subsections of
this report below.
The Panel also notes that it was informed by Agency Staff at the panel meeting on June
18-19, 2002, that EPA was currently revising its cancer guidelines and that the Agency intends
to apply to the final TCE assessment whatever guidance exists at that time for cancer
characterization.
10
-------
In the course of discussion of several charge questions, panel members briefly discussed
their different individual views of the cancer classification, and acknowledged there was a large
policy component to the question of cancer classification. Panel members differed in their
interpretation of how to apply the draft revised cancer classification guidelines and some
requested clarification of the EPA cancer guidelines classification scheme before they could
form a personal opinion. Several panel members characterized the weight-of-evidence as "very
strong" and spoke in support of the Agency's proposed designation of TCE as "highly likely to
be carcinogenic to humans." Several members, however, also suggested that the chemical could
come closer to being classified as "known to be carcinogenic to humans." These panel members
based their views on the animal cancer data showing tumor induction at multiple sites in two
species, epidemiological data showing associations between exposure to TCE and excess cancer
incidence among occupational and environmental cohorts, including several sites where tumors
were induced in experimental animals, and mechanistic data showing similar metabolic
pathways in animals and humans and a high incidence of a specific point mutation in the von
Hippel-Lindau tumor suppressor gene in kidney cancer patients who had been exposed to TCE.
In addition, Panel members noted that there are no data showing that the shape of the dose-
response curve is different at environmental exposures versus at occupational exposures that are
associated with increased cancer risks.
Another view expressed was that the Agency's characterization was not adequately
supported by quantitative data and that a more appropriate characterization was "likely to be
carcinogenic in animals at high doses, not likely to be carcinogenic in humans at lower doses."
More specifically, in this view, TCE was seen as not likely to be carcinogenic in humans at
lower doses for liver tumors and likely or suggestive to be carcinogenic for kidney tumors.
4.2.2. Human Epidemiological Studies
Among the epidemiological studies, the data appear strongest overall for liver cancer and
also, to some degree, for lymphoma. All Tier 1 studies show excesses of liver cancer, yet two
key German studies (Henschler, 1995 and Vamvakas, 1998) show excesses for kidney cancers,
but did not address liver or lymphoma. There was a request for more data on exposure-response
relationships for liver cancer from the Tier 1 studies and the Vamvakas study, and for more
discussion of the liver cancer endpoint generally, since it was felt to be so strong. Tier 1 studies
show excesses for Hodgkin's disease, non-Hodgkin's lymphoma, and multiple myeloma. The
addition of a recently published study by Hansen et al., significantly adds to the weight-of-
evidence for lymphoid tumors.
The Panel does have some concerns about the strength of the evidence for the kidney
cancer endpoint based on the available epidemiological data. The uncertainty focused primarily
on the study by Henschler et al. Concerns included: a) this study originated as a cluster
investigation, thereby potentially introducing bias; b) the variability of underlying population
rates for kidney cancer in German and Danish cancer registries; c) the magnitude of the indicated
risk which was far out of proportion to risks observed in most other studies; and d) the
significance of the Henschler study in light of the whole epidemiology database.2
2 Further analysis by a panel member showed that if one deletes the two early cluster cases then there are 2 RCCs
and 1 renal pelvis cancer. In the US 10% of SEER ICD 189 cases are renal pelvis and thus the expected number of
cases of RCC is about 1.3-0.13=1.2 for the current GDR registry (1988-89). This yields a non significant SMR of
2/1.2=1.7 . The inclusion of the 2 early cluster cases takes us to abiased high SMR estimate of 3.3 which is not
significant either. If one accepts that the Germans are more like Danes the expected number of cases is less and the
SMR using the cluster cases becomes significant. Finally it should be mentioned that the Henschler study employed
11
-------
There was agreement that the inclusion of the Henschler study introduces significant
heterogeneity onto the overall meta-analysis of the renal cancer endpoint. In addition, the Panel
notes that the Henschler study did not report any excess in liver cancer, thereby making the study
less credible in light of the rest of the epidemiological literature. In defense of the Henschler
study, several panel members raised the following points: a) in Germany, people frequently
remain in the same workplace for many decades, thereby potentially increasing the total lifetime
exposure in this cohort; b) from the description of the workplace, the exposures may have been
very high compared to other studies; c) the results were directionally consistent with the
Vamvakas study (1998) although not as strong, somewhat consistent with the recently published
study by Pesch et al., and in line with the animal toxicology and mechanistic data (including the
apparent over 10-fold increased risk reported among exposed workers with active forms of two
GST genes) (Briining et al, 1997a); and d) the very high TCE exposure may saturate P450
pathways thereby shunting the TCE toward the GST pathway and if that is the case, it is likely
that kidney tumor may predominate.
Some members pointed to the history of cancer clusters in identifying other significant
occupational carcinogens (e.g. asbestos, bis-chloromethyl ether, vinyl chloride). In these
instances, the increased incidence in tumors associated with these other chemicals in one
epidemiological study was replicated in other epidemiology studies. In the end, despite the
uniqueness of the Henschler study, the Panel does not advocate omitting it from the draft
assessment.
Several panel members said that the Henschler study should be omitted from
consideration only if it were to become fairly clear that there is some other factor that would
explain the elevated incidence and mortality from renal cancer in this cohort. In that case, the
Agency could use the Finnish and Danish studies with liver and lymphoma as end points, and
then simply discuss kidney, prostate and cervix without quantification of risk. The Panel
recommends EPA include this study in the overall weight-of-evidence, taking into account
information on exposure levels, and to address concerns regarding the Henschler study (1995a)
and to discuss the results in light of competing lines of evidence including the recently published
Danish study.
The Panel advises EPA to identify more clearly and then explicitly apply criteria for the
selection of studies. EPA should select the broadest possible array of studies for each endpoint
meeting those criteria, taking into consideration study design, availability of exposure estimates,
and the goal of protecting health.
Panel members endorsed the division of the cohort studies into three tiers, and
recommended that EPA explicitly weight the Tier 1 cohort studies (and case control studies that
specifically focus on TCE) more strongly than the other studies that involved exposure to a
variety of chemicals. Although there was some debate about the relevance of the Tier 3 dry
cleaner/laundry worker studies, there was support for continuing to include these studies because
many of the metabolites of PCE and TCE are the same. These dry cleaner studies, however,
should not be weighted heavily in the overall weight-of-evidence assessment.
Several members suggested including discussions about prostate cancer and childhood
leukemia, as there is limited epidemiological evidence to support both of these endpoints. In the
case of prostate cancer, all the Tier 1 cohort studies showed slight increases in relative risk,
cohort screening using abdominal sonography which should yield a higher incidence of tumor than expected from a
population based cancer registry.
12
-------
leading to questions about a possible weak tumorigenic effect in humans. In the case of
childhood leukemia, this disease has shown up in numerous community-based studies. In case
control studies of childhood leukemia, parental occupation in a solvent-exposed industry is a
consistent positive association (although the link specifically with TCE is unclear). Childhood
leukemia would not be expected to show up in the occupational cohort studies because these
studies did not evaluate offspring. In addition, the Panel recommends EPA add a discussion of
the Hansen (2001) study, which adds scientific weight to the lymphoma and cervical cancer
endpoints, but not to the kidney cancers, and of the recent Pesch study that found a slightly
increased risk of kidney cancers. There was also a recommendation for a discussion in this
section of glutathione S-transferace (GST) polymorphisms and their possible role in creating
susceptible subpopulations for kidney cancer (Briining et al., 1997).
Some panel members had criticisms and concerns about the Wartenberg review article
(2000). In particular, the following points were raised. Wartenberg et al. adjusted the lower
bounds of the confidence intervals of the reviewed studies; such an adjustment can be misleading
and lead to results that appear to be more significant than they are (if the results need to be
symmetrical on the log scale, the upper confidence bound could be lowered instead, or the
confidence intervals could simply have been presented as published by the original authors
without impeding the ability to calculate the variance of the log of the SMR). The Wartenberg
review included blanks (dashes in the tables) for some endpoints where no cases occurred in the
original studies even though some were expected; this could bias the overall analysis toward
finding an effect. In at least one case (Henschler brain tumors), risk numbers for the exposed
group were reported even though the risks in the unexposed group in the same study were
actually higher. Finally, the separate calculations of incidence and mortality resulted in some
cancers being counted twice, once in the incidence summary statistics, and again in the mortality
summary statistics.
The Panel agrees that some of the key underlying studies need to be directly examined by
the Agency and potential biases and errors should be addressed and corrected. Epidemiological
data merit special attention because they may be potentially important in terms of population-
attributable risk. In that context, the burden is on EPA to make decisions about which studies to
weight most heavily with an eye toward justifying the decisions scientifically and protecting the
public health.
There was a discussion of the pros and cons of performing a formal meta-analysis of the
TCE cancer studies. While several panel members advocated a more traditional meta-analysis of
the Tier 1 studies, because it would convey more clearly how individual studies were weighted,
others pointed out that the study designs were perhaps too disparate, and that a meta-analysis
would simply demonstrate a lot of uncertainty and would add very little information. The Panel
advises the Agency to use criteria in weighing the overall epidemiological evidence and to
provide tables summarizing critical information for each key epidemiology study including type
of study, number of subjects, sources of exposure information, years and estimated levels of
exposure, and basis for the estimated exposure levels.
4.2.3. Animal Toxicology
The Panel agrees that TCE is an animal carcinogen, although its potency is relatively
weak. Tumors are observed in multiple organs of multiple strains of two species at relatively
high doses. In the rat, TCE has been observed to cause a low incidence of rare kidney tumors at
high doses (in the presence of renal toxicity), which should be considered treatment-related.
TCE also causes Ley dig cell tumors in the male rat, another effect that is likely to be treatment-
related. In the mouse, liver tumors occurred primarily by gavage rather than by inhalation, and
13
-------
these are considered to be treatment related, as are the lung tumors after exposure by the
inhalation route.
Although human studies show excess for lymphomas, lymphomas in the mouse are more
problematic. These were increased in 3 out of 6 studies, but may only be treatment-related in
one study, because the NTP study showed the control was somewhat low and the incidence in
the treatment group fell within the range of historical controls. Other panelists pointed out that
while historical controls are a useful comparison, the study controls were in an acceptable range
and that the elevated rate of lymphoma in the treated animals should not be dismissed. There
was a suggestion that EPA reexamine the mouse lymphoma data to see if the endpoint is
treatment-related. If EPA decides the mouse lymphoma endpoint is treatment-related, then a
discussion reflecting the issues with the NTP studies should be added. In general, many of the
carcinogenicity studies that were considered negative are not included in the tables and all
studies for each tumor type should be included. In terms of site concordance, none of the tumors
observed in rats were observed in mice and vice-versa.
There was discussion about the genotoxicity issue with multiple viewpoints expressed on
the likelihood of genotoxicity. Most panelists agreed that genotoxic mechanisms are plausible
for some metabolites.
Many panel members supported the conclusions concerning the genotoxicity of TCE and
its metabolites, as discussed by Moore and Harrington-Brock (2000). They concluded: the
weight-of-evidence argues that chemically induced mutation is unlikely to be a key event in the
induction of human tumors that might be caused by TCE or its metabolites. This conclusion
draws from the fact that these chemicals require very high doses to be genotoxic. The exception
is DCVC, the glutathione conjugate of TCE. Moore and Harrington-Brock (2000, Page 221, 1st
paragraph) conclude that the potency of DCVC is unknown "because there are no data for
mammalian cells from in vitro or in vivo experiments." Thus, while the weight-of-evidence
indicates that most of the tumors induced by TCE are unlikely to be due to a mutation event, it is
not possible to exclude this possibility.
Other panel members were concerned about the remaining possibility of genotoxicity.
One panel member raised the issue of a transient epoxide intermediate that may be formed in the
first metabolic step in the oxidative pathway because the oxidative metabolites TCA and DCA
are derived from this intermediate. There is some evidence that this may occur, and the issue is
discussed in the Lash paper (2000), but this issue is not even mentioned in the EPA draft
assessment. Another panel member raised the issue of the mutagenicity of chloral hydrate as
another controversy that bears more discussion in the draft assessment. Several panel members
advocated strengthening the discussion of genotoxicity in the draft assessment.
Another view expressed was that the genotoxic hypothesis needed to be more carefully
and quantitatively examined by the methodology described in the response to Question 1. Panel
members overall, held somewhat different views on the weight that should be given to genotoxic
modes of action for the liver tumors Panel members felt it was important to keep evaluation of
the genotoxic hypothesis in the mix of possibilities.
For liver tumors, several hypotheses are discussed in the EPA draft assessment,
including: peroxisome proliferation, disturbances in cell signaling and carbohydrate metabolism,
and DNA damage. TCA and DCA likely account for the liver tumor response to TCE, but have
different characteristics. TCA is more potent than DCA and has a greater sustained peroxisome
proliferation effect. TCA and DCA, when given individually, result in tumors that are
phenotypically different. An explanation is that DCA and TCA may selectively modify the
14
-------
growth rates of different clones of cells via altered cell replication and apoptosis rates. Many
scientists agree that TCA and DCA are hepatic tumor promoters (a term that is avoided in the
draft assessment) that exert their effects through cell proliferation and death. Different tumor
promoters select different subpopulations of hepatocytes that are clonally expanded. In the draft
assessment, there is considerable discussion of effects of TCE and its metabolites on
carbohydrate metabolism. One panel member viewed this discussion as speculative, and felt that
the significance of this mode of action is not really clear. One of the problems with DCA is that
it is not measurable after the administration of TCE, but modeling studies indicate that there is
sufficient exposure. The conclusion is that both TCA and DCA contribute to the
hepatocarcinogenicity of TCE in mice at high dose levels and that chloral may also be involved.
For rat kidney cancers, there was agreement that the draft assessment characterized
alternative modes of actions. There were differences among panel members about whether the
scientific evidence was sufficiently strong to support a mode of action for kidney tumor
formation based on the GST metabolite. Some panel members believed that the mode of action
of the kidney cancers was likely to be primarily through the GST metabolite, DCVC, a known
mutagen and cytotoxic chemical, and that this mode of action was relevant to humans. In this
view, peroxisome proliferation and alpha 2u-globulin were unlikely modes of action in the
kidney. From this perspective, the cytotoxic and mutagenic effects of DCVC following its
activation by renal beta-lyase was the likely mode of action for kidney tumor formation. The
finding of DCVC in blood of human volunteers exposed to TCE and the presence of beta-lyase
in the human kidney provided additional evidence that the GSH-mediated pathway was operable
in humans. An alternate view on the panel was that the EPA document did not include critical
information to fully characterize the weight-of-evidence for kidney tumor formation in rats.
Information that would be helpful would include the extent of kidney toxicity and mortality
observed at dose levels causing the low incidence of kidney tumors and the dose-response
relationship of TCE and key metabolites relevant to the different modes of action. The EPA
document could benefit by including consideration of the analytical discussion provided in Lash
et al. (2000) that weighs the evidence regarding different possible modes of actions in relation to
the dose-response relationship at likely human exposure levels of TCE.
There was agreement that EPA appropriately characterized a need for further research on
mouse lung tumors. One hypothesis is that mouse lung tumors are due to the formation and
accumulation in the Clara cells of Chloral Hydrate (CH) formed via CYP2E1. CYP2E1 and
Clara cells are higher in mice as compared to rats and humans suggesting a species difference in
sensitivity. There is a need to demonstrate that CH is responsible for lung tumor development,
since CH is clearly clastogenic and mutagenic at high doses, suggesting that both genotoxicity
and cytotoxicity may be involved. There is debate about the relevance to humans of the rat
Leydig cell tumors, which might be expanded and improved by a fuller discussion, including a
discussion of rat Leydig cell tumors in the draft assessment, with a reference to the paper by
Cook et al., which discusses the basis for differences in responsiveness between rats and
humans.
The overall toxicological assessment was that the rat kidney cancers are relevant to
humans. There was some scientific difference of opinion about the human relevance at low
doses of the lung and Leydig cell tumors. Regarding the mode of action, panel members agreed
that a genotoxic mode of action is plausible for kidney tumor formation based on the currently
available evidence, although there was a difference of opinion about the likelihood that a
genotoxic mode of action exists for TCE. Overall, the Panel recommends that EPA expand the
discussions about some of the main scientific controversies, with some more references to the
primary scientific literature in selected cases involving critical studies.
15
-------
5. CHARGE QUESTION 3
5.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: A new feature of the cancer database is molecular information on the
von Hippel-Lindau (VHL) tumor suppressor gene. Is this information adequately discussed and
are the conclusions appropriate?
Suggested Areas for Inquiry: Does the draft assessment adequately present alternative
interpretations of the von Hippel-Lindau findings and identify this as a research area that would
help resolve an open question about TCE and kidney cancer?
5.2. Panel Response
The consensus of the Panel is that the discussion in the draft assessment is generally
appropriate. There was a recommendation to strengthen the description from the "suggestive
evidence" referred to on line 8 of page 3-38. The Panel generally agreed that EPA is wise not to
regard the evidence as entirely conclusive pending independent confirmation. The discussion in
the draft assessment might be improved by including some additional comparative observations
from kidney cancers not in the TCE exposed workers. Panel members offered the following
enhanced discussion as a starting point for EPA consideration.
Mutations in the VHL tumor suppressor gene (both germline and somatic) have been
associated with increased risk of renal cell carcinoma (RCC). Recent studies by Bruning et al.
(1997) and Brauch et al. (1999) provide evidence that TCE exposure may be associated with
VHL mutations among RCC patients. Specifically, Bruning et al. examined VHL mutation by
single strand conformation polymorphism (SSCP) in renal tissue, initially from 23 RCC patients
with documented high occupational TCE exposure. All (100%) of this first set evidenced VHL
mutation, which the authors concluded was higher than the background frequency of 33% - 55%
among TCE-unexposed RCC patients. In a follow-up, Brauch et al. determined VHL mutation
frequencies by SSCP and direct sequencing of mutations in renal tissue from 44 TCE-exposed
RCC patients. 75% of TCE-exposed patients had mutations in VHL and 39% had a "C to T"
mutation at nucleotide 454. (All "C toT" transitions in the control renal cell carcinoma patients
were evidently relatively rare at about 6% total incidence based on combined data from several
authors.) VHL mutations were detected by Brauch et al. in workers with medium and high but
not low TCE exposure. However, only 3 patients were classified as having low exposure.
Overall, these data indicate a highly significant association (p=0.0006) between TCE exposure
and multiplicity of VHL mutations.
The authors of this paper did not measure total VHL mutation frequency among TCE-
unexposed RCC patients, but used a restriction endonuclease-based assay to evaluate the specific
"C to T" mutation at nucleotide 454 among 107 unexposed patients. None of the TCE-
unexposed RCC patient had the "C to T" mutation at nucleotide 454, indicating a mutational hot
spot in VHL associated with TCE-exposure. The apparent elevation in the frequency of renal
cancers with this specific mutation is indicated to be at least six-fold, and is likely to be 40-fold
or more.
Applying a Poisson distribution to the occurrence of this specific mutation in the studied
groups, the finding of zero cases in 42 VHL mutation bearing tumors in the control group can be
used to rule out a true incidence of as much as 3 cases in 42 (or about seven percent) with 95%
confidence. Therefore, a conservative estimate of the relative enhancement of the frequency of
16
-------
this specific mutation in the renal cancers from the trichloroethylene exposed workers is
39%/7% or just over 5.5-fold. If we take as a plausible but very tentative "best estimate"
incidence of 0.5 of these mutations in 42 mutation-bearing controls examined (about 1%), then
the indicated enrichment of the specific mutation is about 39%/l% or just under 40-fold. The
true enhancement could, of course, be even larger than this, but that could only be determined by
observations with a much larger sample size.
Follow-up studies are needed to confirm the association between TCE and mutations at
nucleotide 454 and to compare total mutation frequencies in VHL gene among RCC patients
with and without TCE exposure, as there is uncertainty over the background rate of VHL
mutations in RCC. Nonetheless, the Agency's conclusion that these findings appreciably
"augment" the characterization of TCE as highly likely to produce cancer in humans (3-51) is
appropriate.
The importance of these observations is reinforced by finding an association of the VHL
mutations with loss of heterozygosity at the VHL locus (Brauch et al., 1999), making a strong
analogy with the classic case of changes in both copies of the retinoblastoma gene in the
causation of retinoblastoma. If the hot-spot and other VHL mutations found in the worker
studies are in fact inactivating mutations coupled with loss of heterozygosity indicating
inactivation of the homologous VHL gene on the opposite chromosome, then the inference must
be that these mutations are not just indicators of TCE exposure, but are likely to be directly on
the causal pathway for the kidney cancers. The Panel suggests that EPA address this issue in the
draft assessment.
Another suggestion for the Agency to consider is that the discussion could be improved
by more clearly defining the "alternative" hypothesis (lines 26-28 on p. 3-38) about some
selection mechanism that could account for these observations without involving mutagenesis by
a TCE metabolite. Certainly it is conceivable that the overtly toxic conditions of high level TCE
exposure could lead to differential growth vs. death/differentiation rates for cells with particular
mutations. But the mutations must be present in the cell population in order to be selected and
the toxic metabolite of TCE that is formed in the kidney is also reactive with DNA, suggesting
the potential involvement of mutational events.
To date, there is no supplementary information on the implications of this particular
mutation (other than probably repressing VHL gene function, which is likely to be a property of
many other tumor-associated mutations as well) that supports the idea that cells possessing this
mutation would have a selective advantage over cells possessing other VHL-inactivating
mutations. The most likely alternative hypothesis is that somehow cells with this particular VHL
mutation have an enhanced selective advantage in the TCE-influenced kidney environment
before the final mutagenic steps leading to fully developed tumors have occurred. The
differential selective advantage might lead to a larger clone of precursor cells (relative to
precursor cells with other VHL mutations) in which the final mutagenic steps leading to cancer
can occur.
17
-------
6. CHARGE QUESTION 4
6.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Does the assessment adequately discuss the use of multiple critical
effects in developing an oral reference dose (RfD) and inhalation reference concentration (RfC)
for effects other than cancer? Are the uncertainty factors well discussed and well supported?
Suggested Areas for Inquiry: The RfD and RfC were developed after considering both
human and animal studies. Does the draft assessment adequately characterize the data at each
site of toxicity and focus on an appropriate subset of critical effects? A key issue is the
application of uncertainty factors. Alternative views range from use of fewer uncertainty factors
to use of additional uncertainty factors to reflect studies showing reproductive effects and
enzyme differences between children and adults. Does the draft assessment adequately discuss
and characterize the evidence supporting alternative positions as it arrives at an RfD and RfC?
6.2. Panel Response
6.2.1. Multiple Critical Effects: Does the draft assessment adequately characterize the data
at each site of toxicity and focus on an appropriate subset of critical effects?
The draft assessment summarizes many epidemiological and experimental studies and
identifies the Lowest-observed-adverse-effect levels (LOAELs )/ no-observed-adverse-effect
levels (NOAELs) for multiple critical effects to aid in development of a single point of
departure. The Agency's consideration of multiple noncancer endpoints in both the general
discussion and in the derivation of the RfD and RfC is commendable. TCE clearly has important
hepatotoxic, nephrotoxic, neurotoxic, immunologic, developmental, and reproductive effects that
should be considered in the derivation of the RfD and RfC. The use of multiple critical effects
increases one's confidence that the point of departure dose is at the low end of doses at which
adverse effects can be observed.
Some panel members suggested that the characterization of the data at each site of
toxicity could be strengthened considerably. In the opinion of some panelists, the current
discussion lacks the type of critical analysis and discussion of the weight-of-evidence that is
necessary to understand the Agency's rationale for selection of endpoints, level of concern, dose-
response extrapolation, effect of time-duration on key endpoints, and application of uncertainty
factors. Other panel members thought the discussion of non-cancer effects and discussions
surrounding the development of the RfD and RfC were quite good and that with relatively
limited additional clarification the non-cancer section of the draft assessment would be complete.
The general recommendations for improvement of Section 3.4. of the draft assessment
are identified immediately below. Detailed comments on specific endpoints that are important
for EPA to consider can be found in Appendix A to this report.
a) The Panel advises the Agency to discuss the key studies in Section 4 and listed in
Table 4.2 in greater detail to outline the scientific basis for selection of endpoints for derivation
of the RfD and RfC. At present, some of the critical studies in Table 4.2 that are used to derive
the RfD/RfC are not discussed at all in Section 3.4; others are given only a cursory mention.
This is a striking deficiency in the draft assessment that prevents EPA's rationale for deriving the
RfC and RfD from being clearly understood.
18
-------
b) The discussion of the specific toxicity endpoints in Section 3.4 does not provide the
essential information that is necessary to understand the Agency's rationale for selection of point
of departure, level of concern, or uncertainty factors. Although a lengthy detailed examination of
the different studies is not appropriate, a more critical evaluation of the data is needed. At a
minimum, scientific data such as exposure levels, severity and nature of effects, methods used to
detect effects would be useful. The Panel suggests that EPA include tables of studies discussed
for each organ toxicity discussion. These tables could include information on species, number of
subjects/animals, doses used, route and duration of exposure, and type of effect noted at each
dose. Additional information useful for human studies includes type of study (e.g. cohort, case
control, cross sectional, ecologic, prevalence), source of control population, method of
establishing exposure levels, and possible exposures to other chemicals. Finally, discussion of
data, especially pharmacokinetic and pharmacodynamic data that assist in understanding relative
sensitivity of humans and animals, human variability, and responses relative to duration of
exposure could be highlighted in order to provide a stronger scientific foundation for later
discussions on uncertainty factors. This type of balanced critical evaluation would strengthen
the scientific data necessary to support the Agency's derivation of the RfD and RfC.
c) An improved critical analysis and discussion (as outlined in point 2) of key
developmental studies and other studies evaluating different life-stages should be integrated into
each toxicity section. It will be important to highlight any data and relevant mechanistic data
that will aid in understanding relative sensitivity of different life-stages. These discussions will
then provide a strong scientific basis for a separate section devoted to summarizing and
integrating discussion of the different developmental effects.
d) The Panel notes that although EPA has evaluated information on metabolism,
pharmacokinetic and mode-of-action data for TCE related to cancer endpoints, it has not
extended this scientific discussion to non-cancer endpoints of TCE toxicity. For example,
several toxicity sections include in vivo and in vitro studies on TCE metabolites, but fail to
provide the essential information on doses used in these studies compared to occupational or
environmental human exposure levels. This information would be helpful in understanding the
appropriate use of this data for dose-response analysis and in assessing TCE toxicity at relevant
human exposure levels.
e) The mode-of-action discussion is focused primarily on cancer endpoints. A more
rigorous discussion of background would be helpful for each critical endpoint as it relates to
mode of action. Each endpoint may have different modes of action involving potentially
different key metabolites that need to be taken into account separately in considering background
cumulative exposures. We note that Table 2-1 lists data sources for estimated adult exposures,
but the references cannot be found based on the numbering system provided.
f) The Panel advises EPA to develop its assessment from the set of well-conducted
studies that make a difference to the weight-of-evidence. Evaluation of the quality of the study
using criteria developed by EPA for adequacy of studies (U.S. Environmental Protection
Agency, 2002b, page 4-6 to 4-8) is especially important to understand the strengths and
weaknesses of different papers cited. Additional studies that should be integrated into the
critical evaluation of the weight-of-evidence are by Fisher et al. (2001) and Albee, R. (1993,
1994). Any other well-conducted studies on TCE, especially those conducted under Good
Laboratory Practices that have been submitted to EPA, need to be evaluated by these criteria.
The Panel is providing these general comments as major areas that should be addressed
for each of the noncancer endpoints. Specific comments on noncancer endpoints are provided in
19
-------
Appendix A to this report. Section 1 of Appendix A addresses one of EPA's proposed areas of
inquiry: Does the draft assessment adequately characterize the data at each site of toxicity?
6.2.2. Modes of Action of TCE Toxicity
The mode-of-action discussion in the draft assessment focuses on cancer. However, the
discussion is also important to the non-cancer toxicity. The draft assessment needs to provide a
balanced discussion of the role of TCE and metabolites in the mode of action of TCE toxicity.
Bull (2000) concluded that DCA is unlikely to contribute to the induction of peroxisome
synthesis at levels that are produced by the metabolism of TCE. Barton and Clewell (2000)
conclude that there are two major hypotheses for the mode of action of TCE in the causation of
neurological effects: the activity of parent TCE, or the metabolite, TCOH. In their opinion, DCA
is not considered to play a role. While some panelists believe the role of DCA in non-cancer
endpoints such as neurotoxicity is overstated in the draft assessment, given these conclusions of
the state-of-the-science papers, others were not convinced that the state-of-the-science papers are
correct on this point and found it reasonable for EPA to consider studies involving DCA. In
addition, the discussion of TCE metabolism needs to include the impact of exposure levels on
the kinetics of TCE metabolism and compare and contrast what is known about the kinetics at
the high doses studied, compared to environmentally relevant levels.
6.2.3. Uncertainty Factors - Areas of Agreement and Differences Within the Panel
The Agency suggested that an area for inquiry for Charge Question 4 was the use of
uncertainty factors. The Agency asked: "Does the draft assessment adequately discuss and
characterize the evidence supporting alternative positions as it arrives at an RfD and RfC?"
The Panel advises the Agency to explain more clearly in the draft assessment how
uncertainty factors were derived for the RfD and RfC. The draft assessment does not adequately
discuss and characterize the evidence supporting alternative positions in deriving these values.
The assessment should address the public comments from the authors of the state-of-the-science
papers regarding these alternative positions.
The Panel also agrees that discussion of uncertainty factors in the document is complex.
Multiple factors were chosen by the Agency for the RfD and the RfC, and several charge
questions were posed to the Panel related to these factors. The Agency's draft factors and related
charge questions are summarized in the table below.
As the Table notes, several Charge Questions, and therefore several parts of this Panel
report relate to the issue of uncertainty. The Panel finds it difficult to isolate separate aspects of
the uncertainty issue not only because they are logically related, but also because the composite
uncertainty factors derived from their product will be used to calculate the RfD and RfC and will
have a major impact on risk management decisions. The Panel also notes the policy constraints
mentioned in the draft assessment (p. 4-10), which stated that EPA has limited the RfDs
calculated using conventional 10-fold uncertainty factors to 3,000 when human-equivalent doses
are used. EPA in its draft assessment then limited the composite uncertainty factor for TCE to
3,000.
20
-------
Table 1
Summary of Uncertainty Factors in the Draft Assessment and Related
Charge Questions Posed to the SAB
Uncertainty Factor
Human Variation
Children's Uncertainty
factor3
Sub chrom c-to-Chrom c
uncertainty
LOAEL-to-NOAEL
uncertainty
Other factors4
Composite Uncertainty
Factor
Draft Value lor
theRfD
50
io-5
io-5
IO5
3,000
Draft Value lor
the RfC
10
10
10
No factor used
1,000
Related SAB Charge
Question and Related
Section of this Report
Question 8; Section 10
and Appendix A
Question 9; Section 1 1
Question 4; Section 4
and Appendix A
Question 4 Section 4
and Appendix A
Question 7; Section 9
and Appendix A
Given this policy constraint and the Agency's application of standard uncertainty factors
(draft assessment, p. 4-7) to TCE, the Panel found itself—and the Agency— far from an ideal
situation for the risk assessment. In an ideal situation, there would have been data for different
endpoints and different populations or different pharmacokinetic variability "uncertainty factors"
for different endpoints and different populations. The Panel observes that this would be a useful
topic for future research. Ultimately, the whole system of uncertainty factors could be usefully
revisited and defined in terms of an objective of achieving x level of risk for the yth percentile of
the variable human population with z degree of confidence.
Given the limitations of current data and methods, there were several alternative views
held within the Panel about the uncertainty factors chosen and their relationship to composite
uncertainty factors. These views are discussed immediately below and also discussed in the
sections referenced in Table 1.
One concern involved the derivation of the RfD and the RfC. In deriving those values,
EPA chose to apply uncertainty factors to "point-of-departure" doses. These were derived as
composite values, representing the lowest human equivalent doses at which several of the critical
studies identified adverse effects and including uncertainty factors based on different endpoints.
For example, the RfD was based on liver-to-body weight changes, but subchronic to chronic
3 Although a Children's uncertainty factor is listed separately in this table to highlight the issue of whether EPA
should explicitly develop a quantitative uncertainty factor for children, the reader should note that the Panel was
informed by EPA during the TCE Public Teleconference on July 18, 2002 that the Uncertainty Factor for Human
Variation was designed to include variation due to different life stages. This is consistent with the discussion of
uncertainty factors in footnote 61, page 4-7 of the draft assessment
4 Other factors to reflect professional assessment of scientific uncertainties not explicitly treated above, including
completeness of the overall database, minimal sample size, or poor exposure characterization. In the case of TCE, a
modifying factor of 105 was set to account for the difference between human background exposures to TCE and its
metabolites compared to background exposures in test animals.
21
-------
uncertainty was derived from duration response trends for decreased testosterone, central
nervous system toxicity, and other effects.
Several panel members felt that EPA should not apply uncertainty factors to composite
"point of departure" doses, but should determine which uncertainty factors are scientifically
justified based on the proposed mode of action for each of the critical effects separately. Then,
these factors should be applied to the human equivalent dose for that particular critical effect; the
most protective value resulting from these calculations could then be chosen as the RfD or RfC.
These panel members also felt that evaluating key studies separately as outlined in Barton and
Clewell (2000) was also an important exercise in making sure that decisions are internally
consistent with the experimental study from which the NOAELs/LOAELs were derived. For
example, applying an uncertainty factor for LOAEL to NOAEL conversion may be scientifically
justified for the critical study that identified a LOAEL (Buben and O'Flaherty, 1985), but not for
the critical studies that identified NOAELs (Tucker et al., 1982; Sanders et al., 1982; Maltoni et
al., 1986). The scientific data available for TCE allow EPA to take a more systematic and
rigorous approach to risk assessment as outlined by Barton and Clewell (2000)5. Specific
recommendations regarding different uncertainty factors are discussed in Appendix A.
Another set of concerns involved children's susceptibility. One concern was that
children's susceptibility was not included explicitly in the derivation of the uncertainty factors.
Panelists felt that it made no sense to include in the document a discussion of ways that fetuses
and children are more susceptible to TCE and then fail to account for this susceptibility
quantitatively.
Many panelists advised the Agency to develop a children's uncertainty factor to account
for this difference in susceptibility. Several panelists expressed concern that EPA initially
derived a composite uncertainty factor for the RfD of 5,000, but reduced the factor to 3,000,
because of policy precedents in setting RfDs. It was unclear which of the uncertainty factors
discussed above was eliminated in order to come up with a total of only 3,000 instead of 5,000,
and these panel members asked that this be clarified in the draft assessment. They considered it
unscientific and inconsistent to build a composite uncertainty factor and then arbitrarily decrease
it to bring it down below a level that the Agency considers acceptable. These panelists urged the
Agency to use the full 5,000-fold uncertainty factor if all the components can be justified.
An alternative view was expressed by other panelists, who noted that a 3.5-fold
difference between adults and infants for metabolism of TCE should have been included in the
EPA analysis (Renwick, 1998). Given the large uncertainty factor of 3,000 already established
for the RfD for TCE, these panelists suggested that this additional 5-fold uncertainty factor
should be retained as one part of the existing "Other" uncertainty factor (either at the 105 value
for this factor set in the draft assessment or another value to be set after reexamination of the
supporting information) already allocated in the Agency's draft assessment to reflect
"professional assessment of scientific uncertainties not explicitly" covered by other factors used
(see additional discussion in Section 11 of this draft report). Another view was that the 50-fold
human variability uncertainty factor that appears in the draft assessment for the RfD was
probably large enough that an additional factor for children is unnecessary. However, an
additional uncertainty factor for children may be warranted for the RfC, which is based largely
5 In this state-of-the-science paper on TCE, Barton and Clewell provide a framework within which to make
consistent scientific judgments regarding selection of critical studies, internal dose metrics, pharmacokinetic models,
approaches for interspecies extrapolation of pharmacodynamics, and uncertainty factors for each critical effect.
22
-------
on central nervous system effects in health adults and currently only includes a 10-fold factor for
human variability.
Yet another view was that the discussion of children's uncertainty factor needed to be
more closely linked to the critical endpoint under consideration. They recommended the
approach of Barton and Clewell. Based on the critical endpoint of concern, decisions could be
made regarding whether or not there is residual concern regarding risks to children. According
to this view, these decisions should not be based on differences in exposure that will be routinely
accounted for in the exposure assessment, but should be based on pharmacodynamic concerns
specific to children based on the endpoint of concern. For example, the decision might be
different for liver-to-body weight changes than for neurologic effects, depending on the severity
and nature of the effects.
A similar set of alternative views emerged relating to uncertainty factors for background
exposures. In light of the prevalence of such background exposures in the general population,
many panel members thought it prudent to apply a modifying factor to the RfD, as proposed in
the draft assessment. These panel members felt that in special cases, where data are available for
estimating these co-exposures for all relevant populations, the modifying factor could be
omitted. Some panelists agreed with the Agency's argument that, unlike the RfD, the RfC was
largely derived from human studies, and thus already incorporated these background exposures
(because the study subjects likely had similar background exposures as the general population).
Therefore, it was a reasonable argument for not applying the modifying factor.
In contrast, other panel members did not agree with the application of a modifying factor
for background exposures for either the RfD or RfC. Some expressed the view that it was
inappropriate to use such a new groundbreaking approach for an individual chemical, prior to the
Agency's finalizing a cumulative risk assessment approach (US EPA, 2002a) that provides a
framework for taking background exposure into account. Since RfDs are often compared for
priority setting, these panelists felt that a consistent approach to handling background be taken.
A panelist noted that such an approach separate from the RfD approach has been successful for
EPA's Office of Pesticide Programs. (See additional discussion of an uncertainty factor for
background exposures in section 9 of this report).
Additional general comments relating to derivation of the RfD and RfC and LOAEL to
NOAEL uncertainty for the RfD and RfC follow immediately below. More detailed comments
with recommendations for improving the Agency's discussion and characterization of the
evidence supporting alternative positions for use of uncertainty factors for deriving the RfD and
RfC can be found in Appendix A.
The Panel provides the following general advice to the Agency in its discussion of and
development of the RfD and RfC in a revised assessment for TCE.
a) The Panel advises EPA to clarify more fully the reasons for the differences in the
uncertainty factors used in the RfD and for the RfC.
b) The Panel advises EPA to describe explicitly its underlying assumptions when
choosing uncertainty factors based on knowledge of mode of action and to describe evidence for
alternative mechanisms.
In the opinion of some panelists, EPA departed from common practice in applying
uncertainty factors to point-of-departure doses that are composite doses derived from the
23
-------
LOAELs and NOAELs of multiple critical studies with different endpoints and different study
designs. Several members of the Panel expressed strong disagreement with EPA's approach and
recommended instead the approach used by Barton and Clewell (2000), which considers each
critical endpoint and applied uncertainly factors based on scientific knowledge of the mode of
action for that endpoint. This approach would allow the RfD to be supported by all the available
data, and in a transparent manner.
Several other panelists felt that the Agency's approach was justified on the basis of the
currently available scientific evidence and for the purpose of ensuring adequate protection. They
noted that one would not necessarily expect that all non-cancer health effects would occur at
similar doses. In fact, normally some of the effects would be high dose, and others could occur
at lower doses.
c) Uncertainty Factors for NOAELS and LOAELs. In regard to presentation and
communication of the Agency's justification for choosing NOAELs and LOAELs, the Panel
advises EPA to show step-by-step how NOAELs and LOAELs of the key studies were converted
to human equivalent doses. The Panel advises the Agency to articulate clearly the scientific
rationale for selecting the key studies. This is especially needed because these studies are not
standard toxicity studies of duration and design that are typically used to set chronic RfDs and
RfCs. Such a discussion would make the selection of the point of departure more transparent.
Improved tabulation of studies in Table 4.2. of the draft assessment, as described in
Appendix A, Section 2.4 of this report will go a long way towards making the EPA's decisions
more transparent. It will also help the EPA identify areas that require further discussion in the
text.
In regard to the Agency's choice of NOAELs and LOAELs, the Panel represents a
spectrum of views about: 1) whether an uncertainty factor for a LOAEL is needed; 2) the size of
the uncertainty factor chosen; 3) the rationale for a difference in factors chosen for the RfD and
RfC; and 4) why an additional uncertainty factor was added to the LED 10 that was calculated
using the benchmark dose approach. These issues and views are discussed in Section 2.4 of
Appendix A. The Panel advises the Agency to provide a clearer discussion in the assessment of
the spectrum of scientific views on these issues and a clearer justification for the decision made
by the Agency.
There was, however, general agreement that if EPA decides to retain the uncertainty
factors described for TCE in the draft assessment, it is important for EPA to spend the time and
effort to develop the scientific rationale for these additional factors, so that this new approach is
not regarded as arbitrary. For example, the available scientific data on pharmacokinetics may
help to quantify the impact of background exposures at relevant human occupational and
environmental exposure levels. Such justification would strengthen EPA's decision to include
such derived uncertainty factors before the draft Framework for Cumulative Risk Assessment
(U.S. Environmental Protection Agency, 2002a) is finalized.
24
-------
7. CHARGE QUESTION 5
7.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Does the assessment adequately discuss the derivation of a range of
estimates for the cancer risk? Are there any studies that should/should not have been included?
Suggested Areas for Inquiry: There is some question about whether the draft assessment
should condense the range into a single "point" estimate. Does the draft assessment adequately
present the case for a range of estimates? A key question for each tumor site is the choice of a
linear or a nonlinear approach based on the mode(s) of action at that site. Does the draft
assessment adequately describe how the available data on mode of action would support either a
linear or a nonlinear approach?
7.2.1 Panel Response
The Panel commends EPA for the derivation of a set of cancer risk estimates or cancer
slope factors (CSF) for TCE in the draft assessment. The presentation of a range of estimates is
a step forward for EPA towards a more explicit and more quantitative representation of the
substantial uncertainties in estimates of cancer risks.
The Panel does not recommend reducing the range of CSFs into a single "point" estimate,
either a geometric mean or another measure of central tendency. Reasons for not reducing the
range of the CSFs into a single number have been given in the draft assessment sufficiently
clearly and exhaustively.
On the surface, the estimation and the rationale underlying the presentation of a range of
cancer estimates appear to be reasonable and informative. Further investigation of this new
approach for risk assessment is, however, warranted
The Panel recommends that the Agency improve its presentation of the cancer slope
factors and discussion of the scientific rationale for choices made. Several suggestions for
improvement appear below.
7.2.1.1 Clarification of the Cancer Slope Factors
The Panel discussed suggestions for clarification. They directly concern improving the
presentation of the derivation of the risk estimates using the data available for TCE.
a) One obvious addition to the present content of the draft assessment should be a
detailed presentation of how each study contributed to the set of cancer slope factors. All
available information should be provided to enable other risk assessors to reproduce the
derivation of the individual cancer slope factors used in the risk assessment procedure for TCE.
The reasons for choosing a study should be explained and the type and wealth of data used for
the cancer slope factors calculation should be described. The derivation of exposure estimates
should be given and their strength and weaknesses should be discussed on a study-by-study
basis. More detail on the exact calculations used would be helpful in a technical appendix. With
such a presentation, the Agency would elaborate as transparently and as explicitly as possible the
approach it has used for the derivation of each of the cancer slope factors in the draft assessment.
25
-------
This presentation should include several specific revisions of the:
1) Introductory statement, "Several cancer slope factors were developed, with
most between 2x10-2 and 4x10-4 mg/kg-d."
2) Definition of the slope factor (pages I -7 and 4-15)
3) Explanation of the overall approach of deriving this set of alternative estimates
(e.g., pages 4-1, 4-2).
EPA's own responses provided to the Panel in its "Summary of Public Comments for
EPA's Science Advisory Board" with regard to Question 5 already contain valuable information
that could be used in such a presentation.
b) It appears that calculations of exposures for the Finnish cohort (Anttila, et al., 1995)
were based only on the urinary TCA measurements for the cancer cases (see page 4-16, line
13pp). If this is what was done, the exposure estimates should be revised to include the full
group(s) of workers who were at risk - not just those who ultimately developed the cancers.
Also, for the occupational epidemiological study-based calculations as a whole, it is not clear
that there was a correction for the healthy worker effect (or the healthy worker survivor effect,
whereby healthier lower-background-risk people tend to stay on jobs longer and receive greater
exposures) (see page 4-16, line 26).
c) Public comments raised on the correct use of the clonal two-stage carcinogenesis
model (Moolgavkar-Venzon-Kundsen two-stage model of carcinogenesis) for biologically based
mechanistic modeling should be considered very carefully. The Panel advises the Agency to
carefully document the description of the modeling itself and make the data used for this
modeling available so that other risk assessors can replicate those evaluations. The Panel
encourages research to improve biologically based mechanistic carcinogenesis modeling for
TCE.
d) The Panel advises the Agency to discuss in the assessment (section 4.5.1.3) whether
problems exist when the New Jersey Drinking Water Study (Cohn et al., 1994) is used.
e) The Panel finds that footnotes in Chapter 1 are very helpful for readers. More
footnotes in other chapters would improve the discussion. For instance, there is a need to
provide a fuller discussion of several points regarding Cancer Slope Factors, specifically:
1) p. 4-16, line 16, the number of 2.956.Where did it come from?
2) p. 4-18, line 28, genotype GSTM/GSTT
3) Table 4-1, the ratio of 97.5 percentile/2.5 percentile equals the span
of the 95% confidence interval. What is meant exactly by this?
These are only some examples.
7.2.1.2 Suitability and Use of the Cancer Slope Factors
Several further issues need to be addressed, clarified or discussed for making the current
derivation of the range of cancer slope factors better suitable for the risk assessment of TCE.
26
-------
a) The meaning of the "upper bound" figures quoted needs to be clarified (e.g. page 4-19,
line 27-28). Are these 95% confidence upper bounds considering only the statistical sampling
error? What additional uncertainty would be expected from the uncertainty in the estimates of
long term exposures? Can upper bounds on the cancer slope factors be defined for human
exposure?
b) Estimates of cancer risks for different target organs should not be seen as alternatives
to one another; but such risks should be probabilistically additive for the typical person who has
a full complement of a liver, kidneys, and other potentially at-risk organs. Competing causes of
cancer mortality or complementary risks from different cancer sites should be taken into
consideration when presenting site-specific cancer slope factors.
c) Adequacy of the PBPK dose estimates for use to derive cancer slope factors should be
discussed in each case. The Agency should investigate how far uncertainties identified during
PBPK-modeling induce uncertainties on the cancer slope factors. The draft assessment should
explain in more detail the differences between the Fisher (2000) model and the Clewell et al.
(2000) PBPK models (see also the Panel response to Charge Question 6.) The discussion of
those differences can then be related to Bois's (2000a,b) uncertainty analysis results.
d) The Agency should show how the conversion of TCA and DC A area under the curve
metric between the chronic oral route and chronic inhalation route of exposure can be / has been
performed (Section 4.2.3 of the draft assessment). Has there been a model applied? Which
assumptions are used?
e) The set of cancer slope factors describing the range may be characterized by different
uncertainties (sampling errors in the case of cancer slope factors derived from animal
experiments, more complicated variability in the case of cancer slope factors derived from
human studies) that should be addressed when presenting the range. Such characterizations are
appropriate to help understand the role of the range of the cancer slope factors and prevent this
range being naively interpreted as a range of high probability of the location of a true cancer
slope factor.
f) The reader is not provided with an integrated treatment of comparable uncertainties for
the different data inputs and routes to calculation. For the animal-based estimates, it appears that
linear projections have been made from LED 10s. Was some upper confidence limit calculated in
the derivation of cancer potency factors from the human studies? Why should the statistical
uncertainties built into the LED 10s be treated differently from the uncertainty resulting from the
use of different data sets, different estimates of exposure for the human studies, different
pharmacokinetic models, or even different assumptions about the relevant dosimeter for
particular endpoints? Sections 4.5.3-4.5.5 should be revised by taking those considerations and
questions into account.
g) There is no weighting of the different bases (animal vs. human studies or cancer organ
site) for estimating TCE cancer risks. Implicitly all values within the selected range are treated
as equally likely, and there is no representation of the likelihood that the true population risk
could lie outside the summary range provided.
7.2.1.3 Improved Mediation of the Cancer Slope Factors
For a better understanding of the role of the range of cancer slope factors for assessment
of TCE risks and for using the range for guidance in deriving exposure limits in specific risk
27
-------
management problems, the strengths and limitations of this new approach should be discussed in
the draft assessment. This discussion should address the following points:
a) The range is not an interval estimate in the sense of a statistical confidence interval and
its concept is not grounded in a sampling model for populations. This reasoning provides
actually another argument for not calculating a central tendency value (see also above, section
7.2.1.2 g).
b) The stability and robustness of the range as it is derived at present should be discussed
in light of the perspective that a new study may appear exhibiting estimates at the lower or at the
higher end of the current range. The sources of disturbance against which robustness is lacking
should be explained in more detail, e.g., referring to the factors reflecting human variation,
sensitive populations, and susceptibility, as specified in sections 1.6 and 3.3 of the draft
assessment.
7.2.2. Further Studies to Be Included
The question posed to the Panel about the inclusion or exclusion of specific studies raised
three issues.
a) The study of Hansen et al. (2001) should be included for the risk assessment of TCE
and in particular for the derivation of cancer slope factors specific for the dose-response data
available for this study. New valuable information for TCE cancer risk is expected for non-
Hodgkin lymphomas, esophagus cancer and cervix cancer. The Panel also advises the Agency to
review and consider the Pesch et al. study (2000) for inclusion.
b) Where epidemiological studies are the basis of risk estimates, EPA should review and
consider all studies (positive and negative), and then make a separate determination about which
studies to use to calculate the cancer slope factor.
These studies should be the ones, among the studies that are well designed, that would generate
the most health-protective number. To select only studies with "statistically significant" results
risks introducing a bias that good meta-analysis practice would avoid.
c) Exposures in most of the epidemiologic studies could be estimated. For example, the
Hill Air Force Base study (Blair, 1998) did not report exposures in ppm, but there are monitoring
data that would provide as accurate a guideline as using toxicologic data with a 10-20-fold
adjustment.
7.2.3. Linear or Nonlinear Approach
The choice of a linear or a nonlinear approach for each tumor site is based on the mode(s)
of action at that site. The Agency has clearly explained in the draft assessment the criteria for
the choice of the linear or the nonlinear approach separately by tumor site. It has also described
the key limitations of the nonlinear analyses, which are the uncertain identity of the active
metabolites and the key events involved in TCE-induced cancers in humans and in animals.
Whereas the linear approach represents a best estimate for the case where one believes the mode
of action is direct or indirect interaction with DNA, and human inter-individual variability is not
very large (see Hattis and Barlow, 1996), the nonlinear approach is used to quantify the extent of
uncertainty and to incorporate this into the determination of RfD and RfC estimates.
The draft assessment does describe in Section 3.5 how the available data on mode of
action would support either a linear or a nonlinear approach. However, the critical reasons for
28
-------
choosing the nonlinear extrapolation should be explained in more detail. The Panel advises EPA
to provide the type of quantitative analysis supporting nonlinear modes of action that was
included in some of the state-of-the-science papers (e.g., Bull, 2000)
The compilation of the cancer estimates in Figure 4-3 provides an excellent overview on
the linear and nonlinear results as far as they were obtained from available data. The Panel also
notes that nonlinear projections were made from LED 10s for liver and testes.
There was disagreement on whether the database for TCE was sufficient to provide
confidence in the risk estimates derived from nonlinear models. Ambiguities in the
determination of the uncertainty factors in the nonlinear extrapolation complicate a comparison
of risk estimates obtained by both methods (the linear and the nonlinear) for one tumor site.
7.2.4. Sensitive Populations
The Panel commends the Agency for providing the sections on sensitive populations and
cumulative risks (pages 4-29, 4-30). It notes, however, that the draft assessment incorrectly
suggests that the different slope factors apply to different characteristics of the exposed
populations (page 4-30, lines 1-6). The Panel advises that Section 4.5.6.4 of the draft assessment
be revised to indicate more clearly whether and how the various slope factors can be based on
variability.
7.2.5. For Further Consideration
Finally, the Panel provides two suggestions that go beyond the present charge question,
but are considered important for the refinement of the risk assessment of TCE.
a) Ultimately, EPA needs to use the TCE and other complicated cases to develop an
integrated probabilistic methodology (e.g. using Bayesian methods) that will weight the different
sources of information bearing on risks appropriately, fairly represent a fuller array of
uncertainties, and systematically derive risk descriptors that are needed for different types of risk
management analyses and decisions.
b) In this, it is important to provide both upper confidence limit estimates and the mean
"expected value" estimate when developing risk ranges, to give users confidence intervals. It is
also important to provide risk management guidance that might indicate when it is most
appropriate to use mean values, or when to use high end values with confidence intervals.
Arithmetic mean estimates may be particularly needed for use in juxtaposing costs and health
benefits of different measures to reduce. Upper confidence limit estimates are needed for
decisions under regulatory programs that seek to redistribute the burden of reducing uncertainty
in risk on economic responsible parties who can make choices either to bring about risk
reductions or fund research projects to reduce the persisting uncertainty of present risk estimates.
29
-------
8. CHARGE QUESTION 6
8.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Please comment on the use of calibrated models and uncertainty
analysis to address the question of pharmacokinetic model uncertainty.
Suggested Areas for Inquiry: The calibrated models (Bois, EHP 2000a, 2000b) build on the
pharmacokinetic models (Fisher, EHP 2000; Clewell et al, EHP 2000) by fitting them to
additional datasets. Is the draft assessment's use of the calibrated models adequately discussed
and supported? In addition, Bois's uncertainty analyses indicate the extent of uncertainly in the
dose estimates calculated for the liver, lung, and kidney. Is the draft assessment's use of these
uncertainly analyses to characterize pharmacokinetic uncertainty adequately discussed and
supported?
8.2. Panel Response
The Agency is commended for including PBPK modeling and its uncertainty analysis
into the risk assessment of TCE. In doing so, the Agency steps forward to meet recent requests
that it use applicable methods for characterizing uncertainty in risk assessment. The Panel
encourages EPA to proceed further in this direction and to include PBPK modeling into the risk
assessment process with identification of the uncertainty in model structure and parameters.
Therefore, an uncertainty analysis should remain in the draft assessment. The Panel
advises the Agency to show that uncertainty has been assessed on the best available scientific
knowledge of the pharmacokinetics and pharmacodynamics of TCE. The use of an uncertainty
analysis for the derivation of different dose metrics is useful for a more realistic dosimetry (the
pharmacokinetic part) for the dose-response assessment. The Agency might, however, also
explore formally including the pharmacodynamic elements in a more comprehensive uncertainty
analysis (uncertainty of causal effect models).
The issue of uncertainty between the two models of Fisher (2000) and Clewell et al.
(2000) is a natural evolution of a relatively new area. In this case, Dr. Bois's application of
statistical methods (2000a,b) to estimate the level of uncertainty helped to strengthen the
argument for application of PBPK modeling in this risk assessment process. Presently, there are
at least 700 papers on PBPK modeling and the area is more than mature enough to be utilized by
the Agency to meet its commitment in its cancer risk assessment guidelines to use biologically
based modeling (USEPA 1999). In many ways, this risk assessment of TCE serves as a role
model for the next generation of risk assessment from EPA.
The state-of-the-science papers on the uncertainty of PBPK models (Bois, 2000a,b) have
been useful to aid in how to credibly apply the results of the Fisher and Clewell's
pharmacokinetic models. Obviously, PBPK modeling should be accompanied by a realistic
uncertainty analysis. However, one has to be aware that in the present case this is a multi-step
approach based on assumptions and further uncertainties. The risk estimates and uncertainties
reported in the draft assessment result after a three-step procedure extracted from the state-of-the
science papers in the Environmental Health Protection Supplemental Volume of 2000, namely:
a) the basic but different Fisher (2000) and Clewell (2000) models
b) the Bois (2000a,b,) uncertainty analysis
30
-------
c) the use of the Bois (2000a,b,) results in the Rhomberg paper (2000).
Therefore, a transparent explanation of the model definition and the intended usage of the
model output are crucial for understanding and applying the modeling and the model outcomes
for the assessment of TCE. Substantially more explanation should appear in the revision of the
draft assessment (using footnotes and/or appendices).
Specific issues related to modeling and uncertainty analysis that the Panel thinks are
necessary for the clarification and the improvement of the Agency's draft assessment are
explained below.
8.2.1. Modeling
Using pharmacokinetic models that estimate the target tissue dose for the key metabolites
identified in the toxicity and carcinogenicity of TCE is an extremely useful tool. The Agency
has recognized this by using these models in the draft assessment of TCE. Although these
models in the assessment have their limitations in describing the complex pharmacokinetic
reality, they are now evolving, perhaps to a point at which they can be used to identify
experimental data to further verify the ability of the current models to predict levels of TCE and
metabolites in the various target tissues.
As described in the draft assessment there are two PBPK models for TCE: one published
by Fisher (2000) and one by Clewell et al., (2000). The basic structural differences between
these models are briefly reviewed in the draft assessment (Section 4.2.1) but not the differences
in estimation of key parameter values. Sources of the differences between the Fisher and
Clewell models and their implications on the central estimates and on the output of the models
and how they translate in dose and risk estimates should be provided. It is important to see how
values are fed from one model into another. A critical evaluation of both these models with
respect to structural features, assumptions, and parameter estimates needs to be included in this
section prior to a discussion of the recalibration of these models by Bois (2000a,b). This
discussion is necessary to understand the value and limits of each individual model as well as the
parameter uncertainty associated with that model. The new discussion should explain the basic
features of model building, prior information, new data, and posterior information of the
calibrating models of Bois (see also a further discussion in section 8.2.2. of this report). It is
also necessary to outline the differences between the simulations performed by Clewell et al.
(2000) and the Bayesian hierarchical modeling of Bois (2000b).
The discussion should also seek to identify and describe the reasons why the Fisher and
Clewell models seem to make divergent dosimetric predictions, and the likely sources of the
residual differences between the model predictions after the Bois recalibration. The draft
assessment mentions on page 4-3 that a Bayesian statistical framework and Markov-Chain
Monte-Carlo (MCMC) simulation was used to refine the Fisher (2000) and Clewell et al. (2000)
model by Bois (2000) using more data sets to estimate each model's parameters. It is also stated
that the result is a set of calibrated models that better fits a wider range of experimental data.
This discussion raised several issues in the Panel.
a) Bois (2000b) used the Clewell et al. (2000) model with three modifications: 1) one
compartment was added to describe the gas uptake data from the work of Fisher et al. (see Bois
200b); 2) the volume of the poorly perfused compartment was changed; and 3) the computation
of the blood flow to the richly perfused compartment was revised, but it is not clear in Bois
(2000b) which PBPK parameter values were changed and why exactly those changes were made.
It would be useful for the Agency to include in this section a discussion of the specific data sets
31
-------
that were used by Dr. Bois to update these models and how these specific data help in parameter
estimation. In reviewing Bois's model (2000b) it appears that gas uptake data may have been
the only new data set used for model calibration and estimation of variability and uncertainty.
The Agency needs to include in the assessment a discussion concerning the limitations of only
using gas uptake data for this purpose since they are an indirect measure of metabolism.
b) The characterization of the uncertainty and the use of PBPK models and their
uncertainty analysis models should be fully described in the draft without requiring extensive
consultation of the state-of-the-science papers in the EHP supplement issue (e.g., through an
electronic appendix, including data and programs if available). Full documentation of the
original data is recommended. The Panel also considers it important to identify which data were
selected for the assessment of uncertainty, and which were not, and the rationale for the choices
that were made in this respect.
c) Bois's modeling is a comprehensive modeling of all aspects of the pharmacokinetic
modeling of derived dose metrics. Reproducibility of the methods used by Dr. Bois, however, is
a question of concern. All assumptions going into the model should be made clear. The Panel
asks that the models be made publicly available. If the assessment of TCE is based on the state-
of-the science papers, it is necessary for the model to be available for independent researchers to
reevaluate the analysis.
8.2.2. Uncertainty Analysis
Parameter uncertainty arises from many sources such as measurement errors or the use of
surrogate data (indirect measurements of parameter values), as in the case of gas uptake data
where changes in parent compound in a closed chamber are measured instead of observations of
concentrations or generation rates of specific metabolites. Concern for parameter uncertainty
should be discussed in this section of the assessment.
Model uncertainty arises due to gaps in understanding the specific mechanism(s) that
affect both the kinetics and the dynamic actions of the compound in question. In Section 4.5.7.1,
the draft assessment states that the full extent of model uncertainty cannot be quantified, only the
models that have been analyzed This is an appropriate warning to the reader that the extent of
model uncertainty that can be quantitatively assessed is limited by the analysts' creativity and
the resources available for examination of alternative conceivable model structures consistent
with available biological understanding. There appears to be model uncertainty with respect to
the Fisher and Clewell models. The performances of different models with the same data sets
can be compared on the basis of a common measure of goodness of fit. It is not made clear how
the Bayesian method contributes to the analysis and how all the available information is used to
judge the relative likelihoods that different models are right.
Several areas for improvement were identified. First, the type of uncertainty covered by
the pharmacokinetic models of Bois (2000a,b) on the basis of the Fisher (2000) and the Clewell
et al. (2000) models should be explained in much more detail within the draft assessment and in
a way that enables the reader to assess the benefits and the limitations of these analyses (e.g.,
further explanation in terms of lack of knowledge and variation between individuals). The
impact of all these modeling exercises on the dose estimates should be exhibited to a larger
extent and discussed critically before using the specific modeling result [e.g., that of Bois
(2000a,b)] to define the uncertainty and variability in the dose metric. The Panel advises that
this be done before any dose metric is used for the derivation of cancer slope factors and the
consequent uncertainty and variability in the cancer slope factors themselves.
32
-------
Another part of the model uncertainty that can be quantified concerns the differences in
expected risks that are produced by assuming that one dose metric is the correct predictor of the
risk of cancer at a particular site, relative to the risk produced by selecting another dosimeter. In
particular, the impact of the uncertainty analyses on dose responses in humans should be
addressed. Furthermore, median results taken only as dose estimates are not enough. It is
necessary to use other percentiles of the distribution (or preferably a representative sampling of
different outputs from the distributions of each plausible dosimeter) and define the effects on the
distribution of risks.
The second area for improvement involves sources of uncertainty revealed by this
modeling approach. These sources, assumptions about the parameter values and their variation,
transfer of a number of parameter estimates between species, should be discussed and related to
the gain in precision obtainable through this calibration. Skepticism about the posterior
distribution, the comparability of the prior and the posterior parameters, and the sensitivity of the
Bois model should be addressed by reviewing the weights that the Bois model places on the new
calibrating information relative to the prior distributions. Prior information may not be given the
weight it deserves or new calibrating information may be evaluated as having less uncertainty
than it should.
The Panel notes several concerns related to the characterization of uncertainty.
a) In Section 4.5.7.1, the draft assessment states that the two pharmacokinetic models
initially led to risk estimates that differed by 15-fold (see Table 4-4). It was stated that to reduce
this uncertainty the models were fitted to additional data sets that improved the models and made
them more compatible reducing model uncertainty. In reviewing Table 4-4 it is not at all clear
what is being compared to get the 15-fold difference and how the calibrated models improve
this. Also, based on what is presented throughout the draft assessment is it even appropriate to
discuss DCA area under the curve as a dose metric? Notice that this clarification has a direct
impact on the Summary and Conclusions section 1.5.1 on page 1-9).
b) The information presented in Table 4.1 needs further explanation. At what dose levels
of TCE were these dose metrics examined? A discussion (in Section 4.2.2) of the information
content of the values in Table 4-1 and the appropriateness of the use of the span of the 95 %
confidence interval would help judging the role of this approximate uncertainty analysis which
has been used to choose dose metrics for the risk assessment.
c) For the benefit of research planning, uncertainty needs to be discussed separately for
each metabolic or follow-up product and it should also account for the difficulties in obtaining
experimental estimates of model parameters.
d) The difference of the pharmacokinetic models for male and females needs in-depth
discussion. Is it intended to derive different estimates for male and females or is this difference
just one aspect of population variation? Differences between males and females, if stated,
should be explained also on the basis of the original data.
8.2.3. Data Availability
Questions have been raised on the cleanness or completeness of the data used in this
modeling (e.g., DCA values reported to contain errors). It was also noted that Dr. Bois was
given an early version of the Fisher model that was changed by the time the mice and human
model reached publication (Fisher, 2000) and that there are new mouse and human data available
from Dr. Fisher. This should be checked and if there are errors or unexplainable inconsistency, a
33
-------
re-analysis should be performed on order to get appropriately revised dose estimates, even if
those errors were of minor influence on the final risk estimate. Therefore, it is strongly
recommended to disclose the sources of the data which were used for this TCE risk assessment
and to describe their availability. The full power of the Bois (2000a,b) modeling is obtained
only if all available data are used.
8.2.4. Markov-Chain Monte Carlo (MCMC)
The section where the statistical analyses of the pharmacokinetic models are described
(Section 4.2.1, 3rd paragraph) should be completely rewritten to describe very clearly and
precisely the methods applied by Bois (2000a,b). Since this method has been discussed widely,
the Panel advises the Agency to provide (e.g., in an appendix) a comprehensive summary of this
methodology with a few key references. Without going too much into the details, the basic
concept of the Bayesian hierarchical modeling should be outlined and the role of the MCMC
method within the use of the Bayesian hierarchical modeling clarified, namely for the calculation
of posterior distributions and the numerical integration necessary to achieve this calculation.
34
-------
9. CHARGE QUESTION 7
9.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Is it appropriate to consider background exposures and other
characteristics of an exposed population as modulating the risk of TCE exposure in that
population?
Suggested Areas for Inquiry: The draft assessment discusses the case that TCE's toxicity
can be modulated by background exposures to TCE's metabolites. A modifying factor is
proposed because the data for estimating the effect of co-exposures may not be available to risk
assessors in the field, but the potential for modification of TCE's toxicity is present. How can
the potential effects of co-exposure be best addressed?
9.2. Panel Response
The Panel is pleased that the Agency has taken the first steps of including the issue of
cumulative risk in a health risk assessment. Although there was agreement that background
exposures to TCE and/or metabolites is a very important issue, there was disagreement, as noted
in section 6.2.3 of this report, about whether the RfD should be the method by which this
background exposure is addressed.
In light of the prevalence of some ubiquitous background exposures in the general
population, some panel members thought it prudent to apply a modifying factor to the RfD. This
factor would address exposures shared by all and not simply due to site-specific scenarios best
addressed by risk managers.
Other panel members argued against including an uncertainty factor for background
exposure, because EPA is in the process of finalizing the cumulative risk assessment approach
that provides a framework for taking background exposure into account separately from the
derivation of the RfD. Since RfDs are often compared for priority setting, it is essential that a
consistent approach to handling background be taken. In addition, total background exposures
may best be taken into account through a thorough aggregate/cumulative exposure assessment
for a specific scenario that needs to be addressed.
Another view noted that, unlike the RfD, the RfC was largely derived from human
studies, and thus already incorporated these background exposures (because the study subjects
likely had similar background exposures as the general population). Therefore, it was a
reasonable argument for not applying the modifying factor.
If EPA decides to include an uncertainty factor for background, it will be helpful for EPA
to include a rigorous discussion of the evidence that a cumulative effect is expected at human
exposure levels based on the modes of actions proposed. Some members of the panel proposed
that EPA utilize the available quantitative data and model the cumulative effect based on general
background for human exposure levels.
The Panel agrees that regardless of EPA's final policy decision on whether or not to
include an additional uncertainty factor in the RfD for background exposure, the Agency should
detail more completely its reasons for choosing, or not choosing, such an uncertainty factor.
Specific comments related to this issue follow.
35
-------
Humans are never exposed to only one agent in isolation. Particularly insofar as multiple
exposures may share modes of action, common metabolic pathways, etc., it seems appropriate to
consider background exposures of an exposed population as modulating the risk of TCE
exposure. It is highly appropriate for EPA to consider background exposures to TCE's
metabolites and to other compounds that produce the same metabolites, because the range of
estimated adult doses for the general population for some of these compounds are comparable to
or even exceed the range of estimated doses for TCE (Table 2-1 of the TCE draft assessment).
For example, tetrachloroethylene, which produces the same metabolites as TCE, is present in the
ambient air at levels 10 times higher than TCE. The same is true for the presence in water of the
chlorination byproducts, DC A, and TCA, both of which are metabolites of TCE. DC A and TCA
may be the active metabolites for some of the TCE's adverse effects, such as hepatomegaly and
hepatic carcinogenesis. Thus, background exposures to these compounds clearly have relevance
to the risk of exposure to TCE.
Understanding and measuring these background exposures presents many
methodological issues. The human studies that state TCE is a primary exposure are still dealing
with mixtures. For example, TCE was a primary exposure at Hill AFB but subjects were
generally exposed to jet fuels such as JP- 4 (now JP-8) on a daily basis as well as other solvents
besides TCE (LeMasters et al. 1997,1998; Stewart 1991). The Wilson et al. (1998) cardiac
malformation study is a prime example of the challenge that this draft assessment faces with
human studies having TCE as the primary exposure. In Section 3.4.5.1 of the draft assessment, it
was reported that "women exposed to degreasing solvents, including TCE have reported elevated
risks for cardiac anomalies in their offspring ... with an attributable risk of 4.6%" for hypoplastic
left heart anomalies. The investigators, however, only asked the parents regarding their exposure
to "solvents/degreasing compounds" but no specific mention of TCE was in the entire study.
Further, it was not clear whether or not the mother or father was exposed, but what was known is
that for 98% of the cases, the mother was interviewed and in 20% of the cases the father was
present. In fact, generally it is unlikely the individuals know the exact compounds contained in
degreasing or solvent exposure. This has been a common experience from interviewing
numerous men and women at sites such as Air Forces Bases (see Hill AFB and articles by
LeMasters et al., 1997, 1999). This suggests that based on the human studies, we cannot
specifically implicate TCE, but can only use these studies as supportive evidence. The Agency
needs a rigorous way of interpreting these studies and incorporating them into its assessment of
background exposures.
In light of the prevalence of such exposures in the general population, some panel
members thought it prudent to apply a modifying factor to the RfD, as is argued in Section 4.3.3
of the draft analysis. These panel members felt that in cases where data are available for
estimating these co-exposures for all relevant populations, the modifying factor could be
omitted. Some other panel members thought it reasonable not to apply the modifying factor to
the RfC, and agreed with the Agency's argument that, unlike the RfD, the RfC was largely
derived from human studies, and thus already incorporated these background exposures (because
the study subjects likely had similar background exposures as the general population). It should
also be noted that the cancer risk estimates based on animal data were not adjusted for
background exposures. Yet another view was that the application of a modifying factor for
background exposures was not appropriate.
The Panel feels that Tables 2-1 should include data on the estimated TCE metabolite
levels derived from the TCE-related compounds. In particular, there is the need to estimate
quantitatively how these background exposures would affect the risk of TCE. This should be
used in a justification of a 3-fold factor (or some other factor) applied to the RfD for background
and co-exposures.
36
-------
Besides background levels of TCE, its metabolites, there are lifestyle exposures and other
co-exposures that will theoretically modulate TCE metabolism, utilize the same metabolic
pathways, or share targets of toxicity with TCE. Examples are acetaminophen and ethanol,
which can theoretically alter susceptibility to TCE effects by influencing CYP2E1 activity. This
leads to the recommendation that a table be developed providing a list of relevant exposures that
modulate CYP2E1 with information that can be used to estimate the impact on TCE risk. In
particular, this table should show how these exposures can be used to justify the choice of the 3-
fold factor applied to the RfD.
Finally, the Panel feels that this important area of cumulative risk required more detailed
treatment as it especially relates to TCE.
37
-------
10. CHARGE QUESTION 8
10.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Do the data support identifying risk factors that may be associated with
increased risks from TCE exposure? Are there any risk factors that should/should not have been
included?
Suggested Areas for Inquiry: Does the draft assessment adequately present and consider
the data supporting identification of potentially susceptible populations, including the role of
differences in enzyme activity to affect TCE's metabolism and toxicity?
10.2. Panel Response
Yes, the data support identifying numerous risk factors that may be associated with
increased risks to susceptible subpopulations from TCE exposure. The EPA draft assessment
has done a good job identifying the general areas of concern related to prenatal, reproductive,
and developmental risks associated with TCE exposure, especially given the level of information
known to date.
A major issue is related to multiple exposures and routes of exposure to susceptible
groups from background exposures to ethanol, TCE and its metabolites, chemical solvent
mixtures and the limited data available for perhaps the most susceptible population, the
embryo/fetus, infant, and child, given the data found in Table 2-1. The Panel notes that that
none of these exposures are voluntary to the fetus, newborn and infant. The implication of levels
of exposure found in the ambient environment and relevance of exposure that might be found in
breast milk, the fetal compartment or in other areas where infants are exposed such as with
preparation of formula with TCE contaminated water supplies are basic areas of extrapolation.
This approach could then serve as a long-term guide for future agents toward evaluating
reproductive, prenatal, and childhood environmental exposures related to age specific effects.
Numerous other potentially susceptible populations were identified and discussed, including
individuals with underlying diseases that alter their metabolism of TCE, individuals on
medications that alter CYP2E1, and individuals with diseases that put them at higher risk for
developing kidney cancer, liver cancer, lymphoma, and other diseases. Concerns were expressed
for diseased individuals (diabetes, hepatitis, HIV positive, etc.), who may be especially
susceptible to TCE exposure. It is not clear whether their increased risk will fall within the 10-
fold RfC population variability factor. Some further discussions on these potential high-risk
individuals would be helpful.
The Panel advises the Agency to discuss other potential risk factors, including: reduced
immune function in children and the elderly; genetic traits that result in variations in
metabolizing genes; DNA repair genes; and inherited mutations in genes that predispose to
particular diseases.
38
-------
11. CHARGE QUESTION 9
11.1. Agency Charge Question and Suggested Areas for Inquiry
Charge Question: Do the data support the possibility that TCE can affect children and
adults differently? How can this be reflected in the quantitative assessment?
Suggested Areas for Inquiry: Given the potential for differences between children and
adults, does the draft assessment develop toxicity values that are protective of children, including
minimization of exposure through human milk? Does the draft assessment adequately consider
the information on differences in metabolism and clearance between children and adults and
appropriately characterize the potential for differences in response? With the data at hand, are
there ways to make the characterization more quantitative? Does the TCE database warrant an
explicit uncertainty factor to reflect data gaps concerning the potential risks to children?
11.2. Panel Response
11.2.1. Major Summary Consensus Points of the Panel
The Panel reached consensus on the following conclusions related to this charge
question:
a) The data presented suggest that TCE can affect children differently than adults,
although there is a very limited database of TCE in children due to lack of directly applicable
studies. Based on the TCE database, children appear to be at greater risks than adults from TCE
exposure, due to possible differences in exposure, metabolism, and clearance. In regard to end
organ susceptibility, data from other solvents and neurotoxicants, in general, would indicate that
the child's central nervous system function is potentially more susceptible to TCE than the adult.
b) The draft assessment does not explicitly discuss whether or not the uncertainty factors
address risk to children or attempt to develop toxicity values that take children into
consideration.
c) The Panel advises the Agency to provide a more complete discussion of the key
articles and information relevant to the issue of differences between children and adults. The
Panel recommends that there be a stand-alone comprehensive children's chapter that discusses
all the children's issues, including exposure, susceptibility during pregnancy, pharmacokinetics,
and pharmacodynamics, in addition to discussions of developmental animal and children data in
every section.
The Panel advises the Agency to include in that chapter a discussion of the need for, or
lack of need for, an additional quantitative children's uncertainty factor.
d) The Panel advises the Agency to support statements about differences between
children and adults with a quantitative discussion, whenever possible. The Panel recognizes that
assessment of children's end organ susceptibility will be one aspect especially difficult to
quantify. The Agency could, however, examine developing uncertainty factors based on the
known distributional estimates of quantitative differences in the pharmacokinetics, and
pharmacodynamics of TCE and its metabolites in children of various ages as compared to adults
(Ginsberg et al., 2002).
39
-------
In regard to the issue of an additional uncertainty factor for children, many members of
the Panel feels that the data supporting the possibility that TCE can affect children differently
than adults led to the conclusion that it would be prudent for EPA to add an additional
uncertainty factor to protect children. They believed this additional factor was merited, based on
heightened concern for children, given existing information, and also based on the uncertainty of
the TCE risk assessment, given the limited developmental toxicity data available.
Others agreed that children are at possibly greater risk, but felt that the existing
composite uncertainty factor of 3,000 for the RfD was already large and adequately protective of
children. They suggested that a component of the "other uncertainty factor" (see section 6.2.3. of
this Panel report) established to cover a wide range of uncertainties involving susceptibility and
background could be explicitly identified as a children's uncertainty factor. Yet one other view
was that the 50-fold human variability uncertainty factor for the RfD in the draft assessment was
probably large enough that an additional factor for children was unnecessary. However, an
additional uncertainty factor for children may be warranted for the RfC, which currently only
includes a 10-fold factor for human variability.
Yet another view was that the discussion of children's uncertainty factor needed to be
more closely linked to the critical endpoint under consideration. This view recommended the
approach of Barton and Clewell. Based on the critical endpoint of concern, decisions could be
made regarding whether or not there is residual concern regarding risks to children. According
to this view, these decisions would then not be based on differences in exposure that will be
routinely accounted for in the exposure assessment, but be based, instead, on pharmacodynamic
concerns specific to children based on the endpoint of concern. For example, the decision might
be different for liver-to-body weight changes than for neurologic effects, depending on the
severity and nature of the effects.
The Panel's discussion of the issue of an additional children's uncertainty factor
emphasizes the importance of the inter-relationships among various components of the Agency's
risk assessment (again see section 6.2.3. of this report). It also underlines the importance of the
Panel's advice that the Agency explicitly address the need for, or lack of need for, an additional
quantitative uncertainty factor for children and clarify how that factor relates to other uncertainty
factors used.
In regard to cancer and children's susceptibility, the Panel notes another issue concerning
the complexity of the TCE assessment. The Panel notes that the issue of children's susceptibility
is closely linked to the Agency's overall risk assessment approach. In response to Charge
Questions 5 and 8, the Panel raises concerns about childhood leukemias and lymphomas
associated with drinking water contamination (New Jersey Drinking Water Study). It should be
noted that this is the only data set used by EPA to address children's cancer risk differently from
adults. Thus, if EPA were to decide not to include that study in its determination of cancer risk,
then an adjustment of the cancer slope factor would be needed to address the children's cancer
risk issue.
11.2.2 Background to the Panel's Conclusions
The critical studies in the human population have not been completed. However, the
pharmacokinetic and pharmacodynamic information that is available on TCE and its metabolites
in children and developing animals present a strong case that the developing child may be more
susceptible to adverse effects from TCE than adults. Until adequate data exist to determine the
exact risk, the Panel advises the Agency to address explicitly how it will factor in protection for
children into its quantitative risk assessment.
40
-------
Generally accepted knowledge of the pharmacokinetics and pharmacodynamics of TCE,
its metabolites (Fisher, 1989), solvents in general, and many xenobiotics support the overall
conclusion that children, as compared to adults, are potentially at greater risk from TCE and its
metabolites (Ginsberg, 2002; Hattis, 2002 in press; and Renwick 1998). If one takes an
approach based on pharmacokinetics and pharmacodynamics, the embryo, fetus, infant, child,
and adolescent (referred from here forward only as the "child"), as compared to the adult, have
altered TCE exposure, absorption, metabolism, clearance, and potentially end organ
susceptibility.
In regard to exposure and absorption, the Panel notes that the child is exposed to TCE
and its metabolites transplacentally, via breast milk, and through the same routes as the adult.
Since the infant has greater skin surface area by a factor of 2-3 on a per kilogram basis, they
should absorb more TCE from the transcutaneous route than the adult. Transcutaneous
absorption of TCE may also be increased in the human premature infant as compare to the adult
due to increase permeability of the premature's skin as compared to the adult (Reed, 1996).
Children drink more water and breathe more air than the adult, so exposure would be greater.
Considering the large number of drinking supplies that have TCE, these exposure-related factors
are of great concern for children (ATSDR, 1997. Fisher et al., 1989).
While there was agreement that differences in exposure assessment should be discussed
more rigorously, including data that quantify these differences, there was no agreement on
whether this difference in exposure should be the basis for additional uncertainty factors.
Several panel members felt that the increased exposure of children to TCE can be determined
and specifically factored in when the overall determination for human TCE uncertainty factors
are calculated. This specific quantitative assessment and adjustment for TCE uncertainty factors
due to children's increased exposure would then preclude additional nonquantitative assignment
of uncertainty factors to be considered in the overall assignment of uncertainty factors of TCE.
In developmental biology, certain additional factors should be considered in risk
assessment. One such factor is the fetal/maternal distribution of the chemicals. In one animal
study (Ghantous et al., 1986), a TCE metabolite, TCA, appears to accumulate in the amniotic
fluid after exposure of the mother to TCE. This accumulation may be due to several factors
including fetal/maternal differences of protein binding, acid/base balance, or metabolism. The
accumulation will result in higher body burdens in the fetus as compared to the mother.
The metabolism and clearance of TCE and its toxic metabolites have been examined by
various studies (Fisher, 2000 and Clewell et al., 2000). Although there is no study that has
addressed the clearance of TCE in children, general pharmacokinetic differences between
children of various ages and adults have been explored by Ginsberg et al. (2002) and Hattis et al.
(2002, In Press). Specifically, for pharmacokinetic parameters including elimination half-life,
clearance, and volume of distribution, these workers assembled a database of human
observations, principally from the pharmaceutical literature, and created a series of combined
analyses from which reasonable assumptions about clearance of TCE by infants and children can
be derived. This research is described in Appendix B to this report and the authors concluded
that the clearance of TCE and metabolites are reduced in children, as compared to adults.
The Panel also notes that there is evidence that the clearance of many TCE metabolites,
which are also toxic, is delayed. Delay of the clearance of chloral hydrate, TCA, TCOH (Mayers
et al., 1991), and TCE-G (Gorecki et al 1990) have been reported. These reductions in
clearance have resulted in the clinical recommendation that the prescribed drug, chloral hydrate
not be used on a repetitive basis in the newborn due to delayed clearance of the choral hydrate
and the metabolites (American Academy of Pediatrics Statement, 1993).
41
-------
Also, from what is known from the metabolism of xenobiotics in the fetus and newborn,
most of the major enzymes responsible for the metabolism of TCE and its metabolites are
reduced in the fetus as compared to the adult (Tateishi et al., 1997). The reduced
metabolism/clearance for many pathways continues through the newborn period and into
infancy. Since the toxic effects of TCE and the metabolites are dose-dependent, the best
indication is that TCE itself and chloral hydrate or TCOH are probably all toxic and have
decreased clearance in the developing human, placing the child at increased risk. The
metabolism of TCOH in the premature newborn has been shown to be decreased by as much as
5-fold (Mayers et al., 1991) and this compound is considered to be more toxic than some of the
parent compounds. Some of the metabolites, such as TCA, were still increasing 164 hours after
the single administration of chloral hydrate (Gorecki, 1991).
This decreased clearance is due to the decreased enzyme activity responsible for
metabolism and clearance of TCE and the metabolites. As recently reviewed (Hattis 2002, Hines
and McCarver 2002), the enzymes that are known to be greatly altered during human
development would include at least the following important enzymes: P450 2E1, ADH, UDP-
glucuronosyl transferases, GSTs, and P450 1A2 (fetus, newborn, and infant). The situation is
even more complicated than just expression of specific enzymes in that select P450 and GST
isozymes are present in the fetus and not to any great extent in the adult and visa versa. In
addition, renal clearance of phase II products is decreased in the fetus and newborn. Finally, it is
known that the enterohepatic circulation of glucuronidated substrates such as bilirubin is
enhanced in the newborn resulting in increased body burdens of glucuronidated substrates
excreted into the biliary tree.
Another major area of concern is TCE-related adverse effects. The Panel advises the
Agency to consider the weight of evidence systematically, endpoint-by-endpoint and overall in
regard to adverse effects of TCE and its metabolites on the developing fetus, infants, and
children.
The Panel advises the Agency to discuss and assess studies of the cardiac teratogenicity
of TCE and its TCE metabolites that are not included in the draft assessment (Fisher et al.,
2001) as well as to review systematically the literature on cardiac teratogenicity (Johnson et al.,
1998a, 1998b, Dawson et al., 1993).
Neurobehavioral toxicity also deserves more attention in the draft. The Panel notes, from
animal studies of TCE and its metabolites, that the developing rodent may be more susceptible to
altered neurobehavioral function. The Panel observes, however that a larger uncertainty factor
for human variation was applied to the calculation of the RfD (50) than for the calculation of the
RfC (10), even though the human data supportive of the RfC were based largely on central
nervous system effects in healthy adults and did not address the potential greater sensitivity of
children. The Panel advises the Agency to include in the draft assessment a structured
evaluation of the literature concerning the neurotoxicological effects of TCE and its metabolites
on children. Some specific suggestions for that discussion are included below.
The only developmental neurotoxicity study of TCE is the Taylor et al. (1985) study,
which did not compare dosing during adulthood with gestational dosing and which did not report
sufficient data to be able to determine a LOAEL in mg/kg. The authors describe
neuropathological changes in the animals treated during development. Taylor (1985), looking at
TCA, found the observed neurobehavioral effects may be permanent, if the exposure occurred
throughout gestation and weaning, unlike other studies in adults where the effects appear
transient. A study by Moser et al. (1999) reported that DCA produced neuromuscular toxicity
including limb weakness and deficits in gait and righting reflex. Weanling rats were more
42
-------
sensitive than adults. It is, however, unlikely that the levels of DC A associated with
developmental neurotoxicity would be achieved with even very high exposures to TCE. Thus,
there are not sufficient data to conclude definitely that the central nervous system of the
developing organism is more sensitive to the effects of TCE than is the adult central nervous
system. This is clearly an important research need.
There are no neurodevelopmental studies examining the offspring of children born to
mothers exposed to TCE in the workplace. Also there are no neurobehavioral studies comparing
children exposed to drinking water with different levels of TCE. These studies would be most
difficult and expensive to conduct. Nevertheless, the Panel advises that the studies be conducted
in order to assess the neurobehavioral risk TCE for children more accurately.
Due to the paucity of developmental, neurobehavioral, and neuropathological studies, it
may be useful to look at studies of other chemicals and humans during development. In general,
the finding that the developing mammal, including the human, is more susceptible during
development to central nervous system toxic chemicals is generally well accepted. This is easily
shown with ethanol in the Fetal Alcohol Syndrome seen in children exposed to ethyl alcohol in
utero (Sood et al, 2001; Hannigan and Armant, 2000; Streissguth and O'Malley, 2000), and also
with mercury intoxication (Marsh et al, 2001), and poly chlorinated biphenyl exposure (Guo et
al., 1996, Grandjean et al., 2001). The effects for example in Fetal Alcohol Syndrome are not
only dramatic but also apparently life-long. TCE and its metabolites could almost be anticipated
to cause greater harm in the child than the adult, and the harm may even cause permanent
alteration in children
There are other toxicities of concern discussed in the draft assessment, such as birth
defects (cardiac, eye, and central nervous system), endocrine disruption, hepatic toxicity,
immune dysfunction, and cancer. In the human newborn, there has been some evidence that
chloral hydrate is associated with hepatic dysfunction after a few weeks of exposure (Reimsche
et al., 1989; Lambert et al., 1990). In regard to birth defects, there are animal and human
population data that suggest that TCE and/or its metabolites may be associated with these birth
defects. Studies in animals show some strong indications that TCE can cause birth defects
(Dawson et al., 1993; Johnson et al., 1998a, 1998b). There are, however, some negative and
ambiguous animal studies (Cosby and Dukelow, 1992; Fisher et al., 2001). There are also data
from human studies that do not support the findings of specific birth defects, and in the studies
that reported effects, some studies do not establish that the effects were due to TCE (Bove et al.,
2000; Bove et al., 1995; Rodenbeck et al., 2000). The Panel advises the Agency to review and
discuss this literature and to draw conclusions from it related to differential susceptibility of
children.
An additional area of concern that has not been addressed in any developmental study is
endocrine disruption. This is a potentially important area that may have significant impact.
All of these concerns demonstrate increased uncertainty and areas where more research is
needed. In addition, more research is needed to determine whether toxicities to these organ
systems occur in developing mammals at lower doses than in adults. The relevant questions
concerning susceptibility are not only whether toxicities to these organ systems occur in
developing mammals, but also whether they occur at lower doses than in adults and if the toxic
effects are more severe and long lasting. The data on TCE are too limited to conclude either
way, but the weight-of-evidence supports that the developing mammal may be at greater risk.
Another great area of concern is cumulative risk and in particular exposure to alcohol
during gestation. The potential for chemical-to-chemical interaction between TCE and ethyl
43
-------
alcohol resulting in greater toxicity of both to the human fetus is substantial and should be
carefully addressed in the draft assessment. This is particularly important as many human
fetuses are exposed to TCE and alcohol.
In summary, the developing mammal, including the human, may be at greater risk from
TCE than the adult. It would be a worthwhile effort to try to quantitate the potential differences
observed in clearance, body burdens, and susceptibility. This task is very difficult to
accomplish, since the clearance and toxicity of the TCE is different from the metabolites.
Despite these difficulties it would be a worthwhile exercise.
11.2.3. How the Draft Assessment Can Be Improved
Overall, the review draft assessment is well prepared. The data on children are presented
throughout the draft assessment. This is appropriate, but the overall review and understanding of
the children's issues would be greatly improved if the overall pharmacokinetics and
pharmacodynamics would be sequentially discussed in one section and the draft assessment
included a comprehensive chapter on children.
A discussion of the various adverse outcomes (cancer, neurobehavioral dysfunction,
cardiac anomalies, endocrine disruption and liver toxicity) and the potential mechanism of action
need to be discussed for TCE and each of the metabolites. It is possible or even likely that there
is not one mechanism of action, but many. It is possible that due to altered metabolism during
development that the potential toxicity mechanisms may change.
It would be helpful for the Agency to discuss the pharmacokinetics and
pharmacodynamics for each metabolite and the effects of each metabolite of TCE on the
metabolism, clearance, adverse effects, and mode of action. The Panel advises the Agency to
examine how age would alter the kinetics and dynamics and how TCE and the metabolites can
interact with one another. There is a large amount of information for chloral hydrate that was
not fully discussed.
The other issue is cumulative risk. A discussion of TCE's potential and probable
interaction with maternal alcohol intake and other chemicals would be important and of high
priority.
One area needing further clarification in the draft assessment is related to the timing of
exposure to TCE during development. It is important to identify exposure scenarios through
various routes and doses to the developing and growing fetus/child.). The Panel notes that mean
exposure to TCE in the urban air is at about 0.3 ppb and water contamination is at 50 ug/L or
less (Wu et al. 2000) and that Agency for Toxic Substances and Disease Registry estimates that
up to one-third of drinking water supply sources have some TCE contamination (ATSDR, 1997)
and advises the Agency to clarify if and how prenatal exposure to these levels is likely to be of
concern. Further, what, if any, are the implications of fat storage during pregnancy and then
mobilization of the fat during the last trimester? Exposures to the fetus and neonate, given these
potential internal sources of exposure and given the considerable pharmacokinetic and
pharmacodynamic uncertainty for children, need more detailed estimations. Breast milk
exposure is of special concern, since TCE is lipophilic and is measured in breast milk. The
question of a child's body burden, given these multiple exposures needs to be estimated and
probably can be estimated fairly accurately. The footnote on page 1-15 of the draft assessment
provides a critical example how the various scenarios could be developed showing the unit risk
for water consumption.
44
-------
The uncertainty of the quantitative risk assessment is very high due to lack of data.
Critical longitudinal studies after intrauterine and new born/infant exposure have not been done.
Of particular concern are neurobehavioral effects, endocrine function, reproductive function,
birth defects, and cancer. These studies will have to be completed in the human to improve the
overall risk assessment for children. Regarding additional research needed, the Panel
recommends that a study be done to evaluate the body burden of newborns and infants who are
likely exposed due to air and/or water contamination. This may be accomplished via a
combination of breath, urine, and possibly occasional blood sampling. Also studies addressing
TCE interaction with other chemicals and in particular alcohol are important.
The key studies should be discussed at greater depth and length. Finally, when children's
risks are discussed, the pharmacokinetics and pharmacodynamics of the parent compound and
the metabolites for the pregnant female would improve the overall evaluation of the study.
In regard to cancer, the draft assessment would be improved with a discussion and
examination of vinyl chloride exposure during development and cancer. In particular, the draft
assessment might benefit with a comparison to potential TCE cancers in children, by comparing
sensitivity and expression of the induced cancers.
One last comment regarding this and other EPA draft assessments, which state that
children generally metabolize chemicals faster. This is not true in the fetus and is rarely true for
the newborn and young infant (Hines and McCarver, 2002). A general statement should not be
made. With the realization that the fetus may express different enzymes than the adult, the idea
of just looking at expression of enzymes and not how they function in the metabolism of
chemicals by the fetus may be too simplistic.
45
-------
REFERENCES
Adibi, J. et al., In press. Personal air monitoring of6phthalate diesters in two urban
populations. Indoor Air
Agency for Toxic Substances and Disease Registry. 1997. Toxicological Profile for
Trichloroethylene (TCE), CAS# 79-01-6. Available from the World Wide Web:
(http://www.atsdr.cdc.gov/toxprofiles/tpl9.html)
Albee, R., et al., 1993. Neurotoxicological examination of rats exposed to trichloroethylene
vapor for 13 weeks. Unpublished Report
Albee, R. R., Spencer, P. J., Johnson, K. A., and Bradley, G. H., 1994. 13-week trichloroethylene
(TCE) vapor exposure of rats caused ototoxicity but no effects on the trigeminal nerve.
The Toxicologist 14: 351
Allen, B. C., Kavlock, R. J., Kimmel, C. A., Faustman, E. M., 1994. Dose-response assessment
for developmental toxicity. II. Comparison of generic benchmark dose estimates with no
observed adverse effect levels. Fundam Appl Toxicol 23(4): 487-95
American Academy of Pediatrics. 1993. Statement: Use of chloral hydrate for sedation in
children. PEDIATRICS 471-3
Anderson, L.M., Diwan, B.A., Fear, N.T., 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: 573-94
Anderson, L. M., Jones, A.B., Miller, M.S., and Chauhan, D. P. 1989. Metabolism of
transplacental carcinogens. In: Perinatal andMultigeneration Carcinogenesis, (N.P.
Napalkov, J. M. Rice L. Tomatis H. Yamasaki, (Eds.)). Lyon, France: International
Agency for Research on Cancer
Anttila, A. et al., 1995. Cancer incidence among Finnish workers exposed to halogenated
hydrocarbons. J Environ Occup Med 37: 797-806
Arito, H. et al., 1994. Effect ofsubchronic inhalation exposure to lowe-level trochloroethylene
on heart rate andwakefulness-sleep in freely moving rats. Jpn J Ind Health 36: 1-8
Bae, D. S., Gennings, C., Carter, Jr., W. H., Yang, R. S. H., and Campain, J. A. 2001.
Toxicological interactions among arsenic, cadmium, chromium, and lead in human
keratinocytes. Toxicol. Sci. 63:132-142
Barton, H. A., and Clewell III, H. J., 2000. Evaluating noncancer effects of trichloroethylene:
dosimetry, mode of action, and risk assessment. Environ. Health Perspect 108 (suppl):
323-334
Berman, Ezra et al., 1995. A Multidisciplinary Approach to Toxicological Screening: I.
Systemic Toxicity. Journal of Toxicology and Environmental Health 45: 127-143
R-l
-------
Blair, A., Hartge, P., Stewart, P. A., McAdams, M., Lubin, J., 1998. Mortality and cancer
incidence of aircraft maintenance workers exposed to trichloroethylene and other
organic solvents and chemicals: extended follow-up. Occup Environ Med 55:161-71
Bloemen, L. J. et al., 1995. Letter to the Editor Re: "Increased incidence of renal cell tumors in
a cohort of cardboard workers exposed to trichloroethene." Arch. Toxicol 70:129-130
Bogdanffy, M.S. etal., 2000. Development of a Physiologically-based Pharmacokinetic Model
for a Biologically-Based Dose-Response Assessment for Dimethyl Sulfate. Ab stract 859
from SOT Annual Meeting
Bois, F.Y., 2000a. Statistical analysis of Fisher et al. PBPK model of trichloroethyelene kinetics.
Environ Health Perspect 108(suppl 2): 275-282
Bois, F. Y., 2000b. Statistical analysis ofClewell et al. PBPK model of trichloroethyelene
kinetics. Environ Health Perspect 108(suppl 2): 307-316
Bove, F., Shim, Y., Zeitz, P., 2002. Drinking water contaminants and adverse pregnancy
outcomes: a review. Environ Health Perspect. 110 Suppl 1:61-74
Bove, F.J., Fulcomer, M.C , Klotz, J.B., Esmart J., Dufficy, E.M., Savrin, I.E., 1995. Public
drinking water contamination and birth outcomes. Am J Epidemiol. 141(9):850-62
Boyes, W. K. et al., (2000) Neurotoxic and phrmacokinetic responses to trichlorethylene as a
function of explosures scenario. Environ. Health Perspect 108(suppl 2): 317-322
Brauch, H., Weirish, G., Hornauer, M. A., Storkel S., Wohl, T., Bruning, T. 1999.
Trichlorethylene exposure and specific somatic mutations in patients with renal cell
carcinoma. JNCI 91: 854-861
Brodkin C.A. et al., 1995. Hepatic ultrasonic changes in workers exposed to perchloroethylene.
Occup Environ Med 52:679-685
Bruning T., Weirich G, Hornauer, M. A., Hofler H., Brauch H., 1997. Renal cell carcinomas in
trichlorethylene (TRI) exposed per sons are associated with somatic mutations in the von
Hippel-Lindau (VHL) tumor suppressor gene. Arch Toxicol 71:332-335
Buben, J. A. and O'Flaherty, E. J., 1985. Delineation of the role of metabolism in the
hepatoxocity of trichlorethylene andperclorethylene: a dose-effect study. Toxicol. Appl
Pharmacol. 78: 105-122
Bull, R.J., 2000. Mode of action of liver tumor induction by trichloroethylene and its
metabolites, trichloracetate anddichloracetate. Environ Health Perspect 108: 241-259
Burg, J. A. and Gist, G. L., 1997. The Potential Impact on Women from Occupational
Exposures. J Women's Health 6: 159-161
Chen, C.W., 2000. Biologically-based dose-response model for liver tumors induced by
trichloroethylene. Environ Health Perspect 108(suppl 2): 335-342
Chia, S. E. et al., 1996. Semen parameters in workers exposed to trichloroethylene. Reprod
Toxicol 10: 295-299
R-2
-------
Chia, S.E., et al., 1997. Endocrine profiles of male workers with exposure to trichloroethylene.
Am JIndMed 32: 217-222
Cherrie, John W. et al., Letter to the Editors: The Importance of reliable exposure estimates in
deciding whether trichloroethylene can cause kidney cancer. J Cancer Res Clin Oncol
127: 400-402
Chia, S. E., et al., 1997. Endocrine Profiles of Male Workers With Exposure to
Trichloroethylene. American Journal of Industrial Medicine 32: 217-222
Clewell III, H. J., 2002. Public comments presented to the SAB TCE Review Panel on June 18,
2002
Clewell III, H. J. et al., 2000. Development of a physiologically based pharmacokinetic model
of trichloroethylene and its metabolites for use in risk assessment. Environ Health
Perspect 108(suppl 2): 283-305
Cohn, P. et al., 1994. Drinkingwater contamination and the incidence of leukemia andnon-
Hodkin's lymphoma. Environ Health Perspect 102: 556-561
Cook, G. S. etal., 1999. Rodent Leydig cell tumor agenesis: a review of the physiology,
pathyologt, mechanisms and relevance to huamans. Crit Rev Toxicol. 29: 169-261
Cosby N.C. AND Dukelow WR., 1992. Fundam Appl Toxicol 19(2):268-74
Dawson, B.V. et al., 1993. Cardiac teratogenesis of halogenatedhydrocarbon-contaminated
drinking water. J Am Coll Cardiol 21:1466-1472
Dobrev, I, Andersen, M. E., and Yang, R. S. H., 2001. Assessing interaction thresholds for
trichloroethylene, tetrachloroethylene, and 1,1,1-trichloroethane using gas uptake
studies andPBPKmodeling. Arch. Toxicol. 75:134-144
Dobrev, I, Andersen, M. E., and Yang, R. S. H., 2002. In silico toxicology: Simulating
interaction thresholds for human exposure to mixtures of trichloroethylene,
tetrachloroethylene, and 1,1,1-trichloroethane. Environ. Health Perspect. In press
Dorfmueller, M.A. etal., 1979. Evaluation ofteratogenicity and behavioral toxiciety with
inhalation exposure of maternal rats to trichlorethylene. Toxicology 14: 153-166
Environmental Health Perspectives. 2000. Tricholorethylene Health Risks. 108 Supplement 2
El-Masri, H. A. and Mumtaz M. M., 2000. A test for reasonableness of default values for
environmental chemicals using Monte-Carlo simulations and physiologically based
pharmacokinetic (PBPK) models. Toxicol Sci 54: 183
Executive Order on Protection of Children from Environmental Health Risks and Safety risks.
April 21, 1997. http://www.epa.gov/children/executiv.htm
Fisher, J. W., 2000. Physiologically based pharmacokinetic models for trichloroethylene and its
oxidative metabolites. Environ Health Perspect 108(suppl 2): 265-273
R-3
-------
Fisher J. W., Channel S. R., Eggers, J. S., Johnson P. D., MacMahon KL, Goodyear C. D.,
Sudberry G. L., Warren D. A., Latendresse J. R., and Graeter L. J., 2001.
Trichloroethylene, trichloroacetic acid, and dichloroacetic acid: do they affect fetal rat
heart development? International Journal of Toxicology 20: 257-267
Fisher J.W., Whittaker T.A., Taylor D.H., Clewell H.J., 3rd, Andersen M.E., 1989.
Physiologically based pharmacokinetic modeling of the pregnant rat: a multiroute
exposure model for trichloroethylene and its metabolite, trichloroacetic acid. Toxicol
Appl Pharmacol 99(3):395-414
Ginsberg, G., Hattis, D., Sonawane, B., Russ, A., Banati, P., Kozlak, M., Smolensk!, S., and
Goble, R., 2002. Evaluation of Child/Adult Pharmacokinetic Differences from a
Database derived from the Therapeutic Drug Literature. Toxicological Sciences, 66:
185-200
Gennings, C., Carter, Jr., W. H., Campain, J. A., Bae, D. S., and Yang, R. S. H., 2002.
Statistical analysis of interactive cytotoxicity in human epidermal keratinocytes following
exposure to a mixture of four metals. J. Agric. Biol. Environ. Stat. 7: 58-73
Ghantous H., Danielsson B.R., Dencker L., Gorczak J., Vesterberg O., 1986. Trichloroacetic
acid accumulates in murine amniotic fluid after tri- and tetrachloroethylene inhalation.
Acta Pharmacol Toxicol (Copenh). 58(2): 105-14
Ginsberg, G., Hattis, D., Sonawane, B., Russ,A., Banati, P., Kozlak,M., Smolensk!, S., and
Goble, R., 2002. Evaluation of Child/Adult Pharmacokinetic Differences from a
Database derived from the Therapeutic Drug Literature. Toxicological Sciences 66: 185-
200
Goh, Victor Hng-Hang et al., 1998. Effects of Chronic Exposure to Low Doses of
Trichloroethylene on Steroid Hormone and Insulin Levels in Normal Men.
Environmental Health Perspectives 106: 41-44
Gorecki, D.K.J., Hindmarsh, K.W.,. Hall C.A., and Mayers D.J. 1990. Sterminatin of chloral
hydrate metabolism in adult and neonate biololgicalfluids after singe-dose
administration. Journal Chrom 528: 333-341
Grandjean, P., Weihe, P., Burse, V.W., Needham, L.L., Storr-Hansen, E., Heinzow, B., Debes,
F., Murata, K., Simonsen, H., Ellefsen, P., Budtz-Jorgensen, E., Keiding, N., White, R.F.,
2001. Neurobehavioral deficits associated with PCB in 7-year-old children prenatally
exposed to seafood neurotoxicants. Neurotoxicol Teratol 23(4):305-17
Green, Laura C. and Lash, Timothy L., 1999. Letter to the Editors Re: "Renal cell cancer
correlated with occupational exposure to trichloroethene." J Cancer Res Clin Oncol
125: 430-432
Guo, Y.L., Lambert G.H, Hsu, C.C. 1995. Growth abnormailites in the population exposed in
utero and early postnatally to polychlorinated biphenyls and dibenzofurans. Environ
Health Perspect, 103 suppl 6:117-122
Hannigan, J.H. and Armant, D.R., 2000. Alcohol in pregnancy and neonatal outcome. Semin
Neonatol. 5(3):243-54
R-4
-------
Hansenetal., 2001. Cancer incidence among Danish workers exposed to trichloroethylene. J
Occup Environ Med 42, 133-139
Hardinetal., 1981. Testing of'selected'workplace chemicals for teratogenic potential. Scand J
Work Environ Health 7: 66-75
Hattis, D., Ginsberg, G., Sonawane, B., Smolensk!, S., Russ, A., Kozlak, M., and Goble, R.,
2002, In Press. Differences in Pharmacokinetics Between Children and Adults—II.
Children's Variability in Drug Elimination Half-Lives and in Some Parameters Needed
for Physiologically-BasedPharmacokinetic Modeling. Risk Analysis
Hattis, D. and Barlow, K., 1996. Human Interindividual Variability In Cancer Risks—Technical
And Management Challenges. Human and Ecological Risk Assessment 2: 194-220
Hattis, D. and Swedis, S., 2001. Uses OfBiomarkers For Genetic Susceptibility And Exposure
In The Regulatory Context. Jurimetrics 41(2): 177-194, Winter
Healy, T. E. etal., 1982. Rat Fetal Development and Maternal Exposure to Trichlorethylebe
100ppm. Br. J. Anaesth 54:337-341
Henschler, D. et al., 1995. Increased incidence of renal cell tumors in a cohort of cardboard
workers exposed to trichloroethene. Arch Toxicol 69:291-299
Henschler, D. et al., 1995. Reply: Increased incidence of renal cell tumors in a cohort of
cardboard workers exposed to trichloroethene. Arch Toxicol 131-133
Hines R. and, McCarver, D.G. 2002. The ontogeny of human drug-metabolizing enzymes: Phase
I oxidative Enzymes, Phase II conjugation enzymes and regulatory mechanisms. JPET
300: 355-366
Isaacson, L.G., Spohler, S.A., Taylor, D. H., 1990. Trichloroethylene affects learning and
decreases myelin in the rat hippocampus. Neurotoxicology and Teratology 12: 375-381
Johnson, P.D., Dawson, B.V., and Goldberg, S.J., 1998a. A review: trichloroethylene
metabolites: potential cardiac teratogens. Environ Health Perspect 106 Suppl 4:995-9
Johnson, P.D., Dawson, B.V., and Goldberg, S.J., 1998b. A review: trichlorethylene
metabolites: potential cardiac teratogens. Environ Health Perspect 106: 995-199
Kjellstrand et al., 1983. Trichlorethylene: further studies of the effects on body and organ
weights and plasma betyrylcholinesterase activity in mice. Acta Pharmacol Toxicol 53:
375-384
Lambert, G.H., Muraskas, J., Anderson, C.L., and Myers, T.F. 1990. Direct hyperbilirubinemia
associated with chloral hydrate administratoin in the newborn. Pediatrics 86: 277-281
Lash, L. H., Parker, J. C., and Scott, C. S., 2000. Modes of action of trichlorethylene for kidney
tumor genesis. Environ Health Perspect 108: 225-240
Lemasters G. K. et al., 1997. Genotoxic changes after low level solvent and fuel exposure on
aircraft maintenance personnel. Mutagenesis 12 :23 7-243
R-5
-------
Lemasters, G. K. et al., 1999a. Male reproductive effects of solvent and fuel exposure during
aircraft maintenance. Reproductive Toxicology 13 155-166
Lemasters G.K , Lockey, J. E., Olsen, D. M., Selevan, S. G. et al., 1999b. Comparison of
Internal Dose Measures of Solvents in Breath, Blood and Urine and Genotoxic Changes
in Aircraft Maintenance Personnel. Drug and Chemical Toxicology 22 : 181-200
Maltoni, C. et al., 1986. Experimental research on trichloroethylene carcinogenesis. In
Archives of Research on Industrial Carcinogenesis. Malton, C. and M.A. Mehlman, eds.
Princeton: Princeton Scientific Publishing Co. 1-393
Mandel, Jack S., 2001. Letter to the Editors Re: "Renal cell cancer correlated with
occupational exposure to trichloroethene." J Cancer Res Clin Oncol 127: 265-266
Mayers, D.J., Hindmarsh, K.W., Sankaran, K., Gorecki, O.K., Kasian, G.F., 1991. Chloral
hydrate disposition following single-dose administration to critically ill neonates and
children. Dev Pharmacol Ther. 16:71-7
Moore, M. M. and Harrington-Brock, K., 2000. Mutagenicity of trichlorethylene and its
metabolites: implications for the risk assessment of trichloroethylene. Environ Health
Persp 108(Suppl 2): 215-223
Moser, Virginia et al., 1995. A Multidisciplinary Approach to Toxicological Screening: II.
Neurobehavioral Toxicity. Journal of Toxicology and Environmental Health 45: 173-
210
Moser et al., 1999. Behavioral evaluation ofthe neurotoxicity produced by dichloroacetic acid
in rats. Neurotoxicol Teratol 21:719-731
Moser, V.C., Phillips, P.M., McDaniel, K.L., and MacPhail, R.C., Unpublished Slides and
Abstract. Evaluation ofthe Neurotoxicity Produced by Dichloracetic Acid in Rats.
Information provided to the SAB TCE Review Panel by Dr. V. James Cogliano, June
2002.
Narotsky, M.G. and Kavlock, R. J., 1995. A multidisciplinary approach to toxicological
screening: II. Developmental toxicity. J Toxicol Environ Health 45: 145-171
National Academy of Sciences., 1993. Pesticides in the diets of infants and children.
Washington, D.C.: National Academy Press
Okawa, M.T. and Bodner, A., 1973. Health Hazard Evaluation of Western Electric Company,
Inc. HUE 72-72-51. National Institute for Occupational Safety and Health
Perera, P.P., 1997. Environment and cancer: Who are susceptible? Science 278: 1068-1073
Perera, P.P., Jedrychoski, W., Rauh, V., Whyatt, R. M., 1999. Molecular epidemiologic
research on the effects of environmental pollutants on the fetus. Environ Health Perspect
107:451-460
Perera PP., 2000. Molecular epidemiology: on the path to prevent!on? J Natl Cancer Inst 92:
602-12
R-6
-------
Perera, F. P. et al., 2002. The challenge of preventing environmentally related disease in young
children: community-based research in New York City. Environ Health Perspect 110:
197-204
Pesch et al., 2000. Occupational risk factors for renal cell carcinoma: agent-specific results
from a case-control study in Germany. Int JEpidemiol 29: 1014-1024
Pinkerton, K., 2002. California Environmental Protection Agency, Office of Environmental
Health Hazard Assessment. Environmental factors affecting respiratory/neural function
during development. Proceedings: Children's environmental health. Developing a risk
assessment framework . Sacramento, California : May 30-31, 2002
Rasmussen, K. et al., 1993. Clinical neurologicalfindings among metal degreasers exposed to
chlorinated solvents. ActaNeurol Scand 87: 200-204
Reed, M. 1996. Principles of Drug Therapy. In Textbook of Pediatrics. Edited by Behrman,
Kleigman, and Arvin p. 29298. WB Saunders, Phil.
Reimche, L.D., Sankaran K., Hindmarsh, K.W., Kasian, G.F., Gorecki, D.K., and Tan, L. 1989.
Chloral hydrate sedation in neonates and infants—clinical andpharmacologic
considerations. Dev Pharmacol Ther. 12:57-64
Renwick, A.G, 1998. Toxicokinetics in infants and children in relation to the ADI and TDI.
Food Add Contam 15(suppl): 17-35.
Reutman S. R, Lemasters, G. K., Knecht, E. A., et al. 2002. Evidence of reproductive
endocrine effects in women with occupational fuel and solvent exposures. Env. Health
Persp. August 2002
Rhomberg, Lorenz., 2000. Dose-Response Analysis of the Carcinogenic Effects of
Trichloroethylene in Experimental Animals. Environmental Health Perspectives 108
(Suppl 2): 343-358
Rodenbeck, S.E., Sanderson, L.M., Rene, A., 2000. Maternal exposure to trichloroethylene in
drinking water and birth-weight outcomes. Arch Environ Health 55(3): 188-94
Ruitjen, M.W.M.M. et al. 1991. Nerve function in workers wth long term exposure to
tricloroethlene. Br J Ind Med 48:87-92.
Sanders, Virginia M. et al., 1982. Humoral and Cell-Mediated Immune Status inMice Exposed
to Trichloroethylene in the Drinking Water. Toxicology and Applied Pharmacology 62:
358-368
Sood B., Delaney-Black. V., Covington, C., Nordstrom-Klee, B., Ager, J,, Templin, T., Janisse,
J., Martier, S., and Sokol, RJ,. 2001. Prenatal alcohol exposure and childhood behavior
at age 6 to 7. dose-response effect. Pediatrics 108: E34
Stewart P. A., Lee, J. S., Marano De et al., 1991. Retrospective cohort mortality study of
workers at an aircraft maintenance facility: II. Exposures and their assessment. Br J Ind
Med 48:531-537
R-7
-------
Streissguth, A.P. and O'Malley, K., 2000. Neuropsychiatric implications and long-term
consequences of fetal alcohol spectrum disorders. Semin Clin Neuropsychiatry
5(3): 177-90
Swaen, G.M.H., 1995. Letter to the Editor Re: "Increased incidence of renal cell tumors in a
cohort of cardboard workers exposed to trichloroethene. " Arch Toxicol 70:127-128
Tateishi, T, Nakura, H., Masoh, Watanabe, M., Tanaka, M., Kumai, M., Takashima, S., Imaoka,
S, Funae, Y., Tabusaki Y, Kamataki, Kobayashi, Y., 1997. A comparison of hepatic
cytochrome P 450 protein expression between infancy andpostinfancy. Life Sciences
61:2567-2
Taylor, D.H., Lagory, K. E., Zaccaro, D. J., Pfohl, R. 1, Laurie, R. D., 1985. Effect of
trichloroethylene on the exploratory and locomotor activity of rats exposed during
development. Sci Tot Environ 47: 415-420
Tucker, Anne N. et al., 1982. Toxicology of Trichloroethylene in the Mouse. Toxicology and
Applied Pharmacology 62: 351-357
U.S. Environmental Protection Agency, 2002a. Framework for Cumulative Risk Assessment
(External Review Draft)
http://oaspub.epa.gov/eims/eimsapi.detail?deid=29570&partner=ORD-NCEA
U.S. Environmental Protection Agency. 1999. Guidelines for Carcinogen Risk Assessment.
SAB Review Draft. NCEA-F-0644.
U.S. Environmental Protection Agency, 2001. Supplementary Guidance for Conducting Health
Risk Assessment of Chemical Mixtures.
http://oaspub.epa.gov/eims/eimsapi.detail?deid=20533&partner=ORD-NCEA
U.S. Environmental Protection Agency. 2002b. A Review of the Reference Dose and Reference
Concentration Processes. External Review Draft.
U.S. Environmental Protection Agency, 2002c. Summary of Public Comments for EPA's
Science Advisory Board
Vamvakas, S., 1998. Renal cell cancer correlated with occupational exposure to
trichloroethene. J Cancer Res Clin Oncol 124: 374-382
Vamvakas, S. et al., 2000. Letter to the Editosr Re: "Renal cell cancer correlated with
occupational exposure to trichloroethene. " J Cancer Res Clin Oncol 126: 178-180
Vamvakas, S. etal., 2001. Exposure estimates in studies on workplace exposures to
trichloroethylene in relation to renal cell cancer formation - a reply. J Cancer Res Clin
Oncol 127: 403-404
Vandervort, R. et al., 1973. Health Hazard Evaluation of Dunham-Bush, Inc. HHE 72-84-31.
National Institute for Occupational Safety and Health
Wartenberg, D. et al., 2000. Trichloroethylene and cancer: epidemiologic evidence. Environ
Health Perspect 108(suppl 2): 161-176
R-8
-------
Whyatt, R. M. et al., 2001. Biomarkers ofpolycyclic aromatic hydrocarbon-DNA damage and
cigarette smoke exposures in paired maternal and newborn blood samples as a measure
of differential susceptibility. Cancer Epidemiol Biomarker Prev 10: 581-588
Whyatt, R. M. et al., 2002. Residential pesticide use during pregnancy among a cohort of
urban minority women. Environ Health Perspect 110: In press
Whyatt, R. M. et al., 2000. Association between poly cyclic aromatic hydrocarbon-DNA adduct
levels in maternal and newborn white blood cells andglutatione S-transferase PI and
CYP1A1 polymorphisms. Cancer Epidemiol Biomarkers Prev 9: 207-212
Wilson, P. D. et al., 1998. Attributable fraction for cardiac malformations. Am J. Epidemiol
184: 414-423
Wu, C., and Schaum, J. et al., 2000. Exposure assessment of trichloroethylene. Envir Health
Perspect 108(suppl 2): 359-363
Zenick, H. et al., 1984. Effects of trichloroethylene exposure on male reproductive function in
rats. Toxicol 31: 237-250
R-9
-------
APPENDIX A
SPECIFIC PANEL COMMENTS ON THE AGENCY'S ASSESSMENT OF
NONCANCER ENDPOINTS
1. Specific Comments on Hazard Characterization for Noncancer Endpoints
This section considers whether the draft assessment adequately characterizes the data at
each site of toxicity.
1.1. Liver Effects
This endpoint requires significant attention because the liver weight to body weight
change (LW/BW) is the key endpoint used to establish the oral RfD. There are scientific data
available on mode of action, species differences, case studies (both positive and negative)
following oral ingestion, severity of effect, and the relationship between effect and duration of
exposure that can greatly inform the selection of uncertainty factors. The Panel advises the
Agency to include a critical evaluation of the key studies used for determining the oral RfD [e.g.
Tucker et al., (1972), Buben and O'Flaherty (1985), Berman (1995)] in this section. It would be
helpful to discuss the quantitative changes in LW/BW ratios and to determine if the National
Toxicology Program (NTP) toxicology and carcinogenesis studies of TCE in four strains of rats
(CAS No. 79-01-06) report any LW/BW changes either in the report or raw data base and how
these changes relate to liver histopathology. Understanding how effects on LW/BW progress
with increasing duration of exposure will provide a scientific basis for extrapolating from
subchronic to chronic exposures.
The Agency's discussion of this endpoint refers to the state-of-the-science reviews by
Bull (2000) and Barton and Clewell (2000) that summarize the effects of TCE, TCA, DCA and
CH on increased liver size, but fails to also include the discussions on relevance of these findings
to TCE toxicity based on the dose level and cytotoxicity. As discussed in the general comments,
it would be helpful to accompany any discussion of metabolites, where possible, with a more
quantitative critical evaluation of the data.
The Agency's discussion of this endpoint mentions the results of several animal studies
that examined the hepatotoxicity of metabolites of TCE. Human studies that have examined
hepatotoxicity associated with TCE metabolites or with compounds that have the same
metabolites, such as PCE, should also be summarized and quantitatively analyzed here. An
example is the study by Brodkin et al. (1995) that found significant hepatic ultrasound
abnormalities in dry cleaning workers exposed to PCE, even in the absence of significant effects
on routine liver function studies. Such a review should include a discussion of exposure levels.
Sub-section 3.4.2.2 of the draft assessment refers to Tier I, II, III studies, but the meaning
of these terms is not explained until later in the text (footnote 46 to Section 3.6.2). The footnote
should be moved to the first mention of the terms.
1.2. Kidney Effects
For TCE, the Agency notes that an RfD can be based on critical effects in the liver,
kidney and developing fetus. There is very little information presented in the draft assessment
(section 3.4.3.1) describing the kidney toxicity that was stated to occur in humans exposed to
TCE and in rodents. There should be a more thorough summary of the human toxicity studies
since some are negative and some report changes in urinary proteins reflective of damage. This
A-l
-------
does not come across in this draft assessment. It appears the RfD review was based on the
kidney effects reported by Maltoni et al., (1986) where the specific effects in the kidney were not
described in any detail in this draft assessment. Going back to this original study it appears that
Sprague-Dawley rats were dosed orally for 52 weeks and then followed until natural death. The
response of kidney meganucleocytosis was only observed in male rats at 250 mg/kg. It is not
clear from what is presented in the Maltoni et al. publication or in the draft assessment whether
this lesion is associated with normal aging or is thought to be a response to chemical exposure.
There is a National Toxicology Program (NTP) report (1983: NTP TR 243) which reports
karyomegaly of the renal tubular cells in male rats at 2000 mg/kg TCE and female rats at 1000
mg/kg TCE (13 week study). The Panel advises the Agency to reevaluate both of these studies
and their findings prior to setting of the kidney NOAEL. It is also not clear why the Berman et al
(1995) paper was not used by either Barton and Clewell or in the draft assessment to derive a
LOAEL for kidney toxicity of 50 mg/kg/d with 14 day dosing. This would have lowered the
human effective dose (HED) for nephrotoxicity and should therefore be explicitly addressed in
the draft assessment.
1.3. Developmental Effects
The draft does discuss multiple effects on children and the developing fetus. The draft
assessment discusses the evidence that TCE may be a cardiac (and possibly ophthalmologic)
teratogen. This issue deserves greater attention and critical analysis. Cardiac teratogenesis has
been reported in community-based epidemiological studies in which TCE was a contaminant.
Four studies in the rat model (on TCE, TCA and DC A) have revealed significant excesses of
cardiac defects. A relatively recent mechanistic study using chick embryos cultured on collagen
gel (Boyer et al., 2000) has identified a possible mode of action by which TCE may cause
cardiac defects - dose dependent inhibition of mesenchymal cell transformation. A more recent
study by Fisher et al, (2001) failed to identify cardiac defects in the rat. Differences between this
study and the other studies should be evaluated relative to sample size route of dosing, duration
and timing of exposure, maternal toxicity, and relevance to humans.
The Panel advises the Agency to improve its discussion of the referenced studies and
especially of the critical studies referenced in Table 4-2 for developmental toxicity (Narotsky et
al., 1995a,b and Dawson et al,, 1993). The second paragraph superficially summarizes a number
of positive developmental studies of TCE, TCA, and DCA without stating which study evaluated
which compound. The studies that evaluated the developmental toxicity of TCE should be
summarized and critiqued first. Data from the studies of the TCE metabolites, TCA and DCA,
should be summarized separately and evaluated as to their consistency with the TCE studies and
pharmacokinetics of TCE.
The Panel advises the Agency to discuss the strengths and weaknesses of the studies,
based on dose-response relationships extrapolated to relevant occupational and environmental
exposure levels, and based on relevance of methods used (e.g. chick embryos cultured on
collagen gel; osmotic mini pumps delivering TCE directly to the uterine lumina). The section
could by improved by discussing more several negative developmental studies that used
inhalation exposure to TCE at relevant exposure levels. These studies are not referenced
individually, but rather the reader is referred to Barton and Clewell (2000 ) [the actual referenced
papers are Healy et al.(1982); Dorfmueller et al. (1979); and Hardin et al. (1981)]. These
negative as well as the positive studies should be summarized and critically evaluated. Special
attention should be given to the Fisher et al. (2001) paper that used 19-20 litters, high oral
gavage doses, exposure duration of GD 6-15 which spans the critical periods of heart
development in rats, and sensitive techniques conducted blind in collaboration with an
investigator who initially detected an effect in an earlier study (Johnson et al., 1998).
A-2
-------
At times, the draft assessment makes overgeneralizations that make detailed and specific
interpretations difficult. One specific example is in section 3.4.4.2. One study (Cohn et al.,
1994) is reported related to childhood leukemia with an "observed very strong association" with
exposure during pregnancy and an exposure response gradient with drinking water
contamination as the etiologic agent with TCE "often the chemical found in highest
concentration." Though the odds ratio is 13.2, the confidence interval includes 1 ranging from
0.9 to 205.2. Thus, the significance of this finding may be over-stated.
1.4. Neurotoxicity Effects
Several of the critical studies in Table 4-3 for neurotoxicity are human occupational
studies. The Panel advises the Agency to discuss whether the subjects had other exposures
besides TCE. For example, the study by Rasmussen et al. included workers with concomitant
exposures to CFC113 as well as other unspecified solvents. The Arito et al. (1994) study was
not discussed in Section 3.4.1. and should be discussed because it may be one of the most
sensitive endpoints for setting RfCs based on neurotoxicity.
The text in section 3.4.1 refers to a study by Moser et al. from 1999 that used DCA as the
test substance but does not discuss the dose-response relative to quantitative levels of DCA
following TCE exposure to rats and to known relevant human exposure levels. This type of
discussion is essential in order to understand the relevance of these studies to the TCE risk
assessment. It is unlikely that DCA plays a role in TCE neurotoxicity, based on the discussions
in the state-of-the-science papers (Lash et al., 2000; Barton and Clewell, 2000). The 1995 paper
by Moser et al. that tested the neurobehavioral effects of 1- and 14-day exposures to TCE
(referenced in Table 4.2), among other compounds, should be discussed in section 3.4.1. and
included in the reference list. The Boyes et al. (2000) paper studying acute peak vs. repeated
exposure, and the Moser et al. (1995) comparison of 1 and 14-day exposure should be evaluated
in terms of providing information in extrapolation from shorter to longer duration exposures.
1.5. Endocrine System Effects and Reproductive Toxicity Effects
The relevant section of the draft assessment needs to be discussed more carefully because
endocrine effects raise particular concerns for potential effects on the developing fetus, including
concerns about vulnerable life stages that should be considered in the discussion about
uncertainty factors and children's vulnerability. Therefore it is important that a more critical
evaluation and balanced presentation of the data be given in this section. Additional specific
comments on the human endocrine toxicity studies used in the derivation of the RfC are provided
in section 2 of Appendix A of this report. Both negative and positive data should be reported
including the results of the 2-generation reproductive studies with mice and rats using
microencapsulated TCE in feed discussed in the state-of-the-science review by Barton and
Clewell (2000).
In terms of reviewing the epidemiological literature, section 3.3.1.3 of the draft
assessment reports excess risks of cervical cancer in occupationally or environmentally exposed
women. More specific details of these studies are needed in order to bring clarity to the
discussion, as specifics of the actual study findings are sparse. The report states "TCE exposure
has been associated with excess risks of cervical cancer in occupationally or environmentally
exposed women (Blair et al., 1998; Anttila, 1997; and Burg, 1997 cited). These studies typically
cannot account for possible confounding from lifestyle factors. According to public comments
provided to the Panel, the Anttila et al. (1995) study found a significant increase in cervical
cancer for women exposed to TCE for less than ten years, but not for longer than 10. Blair
showed non-significant but elevated breast cancer mortality rate ratios in the low level
A-3
-------
intermittent or continuous exposures (3.1-3.4) that were higher than those reported in those
having frequent peaks RR=1.4. In the Blair et al. study (1998), cervical cancers had rate ratios
of 1.8, which was not significant with confidence intervals of 0.5-6.5. Prostate cancers in men
also were not significant. A more thorough description of the actual findings, limitations, and
implications is needed.
In terms of male reproductive effects, the data are mixed. When evaluating cytotoxic
effects of exposure to solvents and fuels containing TCA at Hill AFB, effects were observed
related to an increase in micronuclei and sister chromatic exchanges (Lemasters, 1997, 1999a,
1999b). Although the epidemiological literature provides evidence for reproductive effects of
TCE in men, but not in women, this is primarily because there are virtually no studies of
reproductive function in TCE-exposed women. Thus, the human reproductive toxicity data
cannot be used to determine whether TCE toxicity is modulated by gender.
In an NTP CD-I mice study a 45% and 18% reduction in sperm motility was observed in
the baseline and first generation of males, even though no effect was observed on mating,
fertility, or reproductive performance. In contrast, the primary finding of the Zenick et al. 1984
study on male rats showed that TCE-related effects were seen primarily in the 1000 mg/kg group
related to impaired copulatory behavior. The copulatory functions had returned to normal by the
fifth week of exposure and essentially no effects on sperm parameters were observed. The
conclusion from this latter study was that "TCE exerts minimal direct effects on the male
reproductive system in terms of spermatotoxicity," but TCOH showed a 3-7 fold increase in the
testis, prostrate, seminal vesicle, fat, liver, kidney and lung. The Zenick et al. study
demonstrates the ability of the reproductive organs to concentrate TCE and its metabolites with
increasing dose.
2. Specific Comments on Uncertainty Factors for NonCancer Endpoints
2.1. Human Variation
In the view of some panel members, the current draft assessment makes a potential error
in treating the pharmacokinetic uncertainty as if it were a measure of human inter-individual
variability. Some panel members nonetheless argued that EPA's choice of a 50-fold uncertainty
factor could be justified on other grounds. These panel members felt that inter-individual
variability in metabolism and response to TCE is likely to be large, especially when children are
considered. The El-Masri (2000) analysis would appear to indicate that the data-derived factor
for human variation and animal-to-human extrapolation combined should actually be 625x.
EPA's analysis would be strengthened by explaining the range of data and methods available for
assessing inter-individual variability and describing the rationale more clearly for the approach
taken.
In addition, the 3.5-fold difference between adults and infants for metabolism of TCE
should have been included in the EPA analysis. This could be addressed by specific modeling of
human variability distributions (Renwick, 1998).
Pharmacodynamic variability is another matter. Different endpoints may be causally
related to different metabolites and different dosimeters related to those metabolites [e.g.
maximal concentration (Cmax) vs. area under the curve in relevant locations in the body].
Therefore, there may be different pharmacokinetic variability "uncertainty factors" for different
endpoints. This would be a useful question for future research. A related issue concerns
appropriate estimates of the pharmacokinetic portion of human inter-individual variability. The
Panel encourages the Agency to explore deriving them by exercising the various
A-4
-------
pharmacokinetic models using the population variability of various pharmacokinetic parameters
estimated by Dr. Bois, together with the dependencies (a more general word than "correlations")
among the values of these parameters. Obtaining these inputs for variability simulations may
require consultation with Dr. Bois. Variability should be calculated separately for different
dosimiters putatively related to different adverse effects (e.g. areas under the curve vs. peak
levels of key metabolites hypothesized to be involved in causing specific effects).
Ultimately, the whole system of uncertainty factors could be usefully revisited and
defined in terms of an objective of achieving x level of risk for the yth percentile of the variable
human population with z degree of confidence.
2.2. Animal-to-Human Uncertainty
In regard to the RfD, the Agency's RfD is based on effects on LW/BW ratio. The EPA
document explores many different modes of action for TCE's effects on the liver. The state-of-
the-science papers (Bull, 2000, Barton and Clewell, 2000) critically and quantitatively evaluate
the likelihood of different modes of actions to be relevant. Barton and Clewell conclude that
LW/BW alterations involve peroxisome proliferator-activated receptor (PPAR) and that the data
do not support the standard default assumption that humans are more sensitive than the most
sensitive rodents. The EPA document also acknowledges that humans in general have lower
expression of PPAR alpha compared to mice and that these "quantitative differences have import
to the dose response analysis of the mouse liver tumors (pp. 3-27). The EPA document should
discuss this scientific evidence in their discussion of the selection of uncertainty factors. If
different endpoints are used to derive temporary RfDs, as recommended by the EPA's SAB, then
different considerations should be made depending on the most likely modes-of-action
In regard to the RfC, and in light of the supportive data for effects at similar exposure
levels in the human studies, it seems appropriate to omit this factor.
2.3. Subchronic-to-Chronic Uncertainty
Regarding the RfD and RfC, some panel members felt that it was not appropriate to apply
an uncertainty factor for sub-chronic to chronic effects to the "point of departure" dose because
that dose was derived from chronic, as well as from subchronic, dosing studies. These panel
members felt that EPA should not apply any uncertainty factors to composite "point of
departure" doses, but should determine which uncertainty factors are scientifically justified for
each of the critical studies. Then, these factors should be applied to the human equivalent dose
for that particular study; the most protective value resulting from these calculations could then
be chosen as the RfD or RfC.
It was unclear to some panelists why a full ten-fold factor was used for subchronic to
chronic in the derivation of the RfC, but only a three-fold factor was used for the RfD. This point
requires clarification in the draft assessment.
The liver endpoints bring up yet another issue, which is the definition of "chronic."
Some panelists considered the 6-month study by Tucker et al. (1982) to be adequate to establish
a chronic NOAEL. However, because the mouse lifespan is greater than 2 years, others felt that
a subchronic-to-chronic uncertainty adjustment was needed, The Berman et al. study (1995)
reported dose-dependent increases in liver weight after 14 days of dosing, and the Buben and
O'Flaherty (1985) study after 6 weeks of dosing. One would like to be able to assess whether
the effect was greater in the studies with longer dosing durations to support the use of an
A-5
-------
uncertainty factor for subchronic-to-chronic adjustment for the two shorter studies.
Unfortunately, the Tucker et al. study did not report the liver weight values. Barton and Clewell
(2000) argued that changes in relative liver weight are early events that are sensitive indicators
of potential liver effects observed at later times, and therefore no adjustment should be made for
exposure duration. This argument would be supported if there were no evidence of duration-
response trends in liver weight or other aspects of liver toxicity.
The use of an uncertainty factor for sub-chronic to chronic dosing makes more sense for
some of the endpoints than others. As the draft risk assessment points out, there is evidence for
duration response trends for neurotoxicity from the human inhalation exposure studies. There
are fewer effects from acute high exposures, compared to longer duration exposure [Moser
(2000) and Boyes et al. (2000)]. The application of a subchronic-to-chronic uncertainty factor to
developmental toxicity studies, such as Dawson et al. (1993) and Narotsky et al. (1995) does not
make sense from a scientific point of view, because pregnancy is not a chronic state. The
application of uncertainty factors is somewhat endpoint-specific and combining uncertainties
from different endpoints may be difficult to justify scientifically.
In regard to the RfC, the Panel notes that the RfC includes a 10-fold factor for
subchronic-to-chronic. However, it could be argued that the critical human studies cited in Table
4-3 of the draft assessment as being of "subchronic" duration should be considered chronic
studies because the mean exposure durations ranged from 5 to 16 years. The draft assessment
should explain how this might impact such a factor. The Arito et al. neurotoxicity study was a 6-
week study. Some panelists said that if duration-response effects were observed in the
supportive human neurotoxicity studies, it seems appropriate to apply this factor to this study.
For the liver weight effects observed in the 30 day Kjellstrand (1983) study similar arguments
can be made as for the liver effects in the oral dosing studies. Applying this factor should
depend on whether there is any evidence of duration-response trends for this endpoint. Other
panelists supported the use of a ten-fold factor for subchronic-to-chronic and urged the Agency
to also apply the full ten-fold factor to the RfD.
2.4. LOAEL-to-NOAEL Uncertainty
In regard to presentation and communication of the Agency's justification for choosing
NOAELs and LOAELs for the RfD, Table 4-2 needs to explicitly identify which
NOAEL/LOAEL goes with which effect. In Table 4-2, first row, the NOAEL of 18 mg/kg/d for
males seems to be from the Tucker et al. (1982) study for liver weight, not the Sanders et al.
(1982) study as indicated in the table. In contrast, the 18-mg/kg/d dose represents a LOAEL for
cell-mediated immune response to sheep erythrocytes in females from the Sanders et al. study.
The 217 mg/kg/d NOAEL in males is identified as coming from the Sanders et al. study, and
seems to refer to the humoral response to sheep erythrocytes after 6 months of TCE exposure.
Other effects were seen in males at lower doses, and it is not clear why those were not chosen as
the critical effect. For example, the recruitment of peritoneal cells showed a dose-dependent
decline with a LOAEL of 18 mg/kg/d in males at 4 months. For several of the other parameters,
there was not a clear dose-response in males. The female NOAEL/LOAEL of 193 mg/kg/day
could not be found in either of those studies of 193 mg/kg/d. The NOAEL for liver weight in
females in the Tucker et al. (1982) study was 437 mg/kg/d according to the text. Tables 3 and 4
from Barton and Clewell (2000), which is referenced as a source of the experimental doses listed
in Table 4-2 of the TCE draft assessment, identify a NOAEL for immunotoxicity from the
Sanders et al. study as 200 mg/kg/d. This is because they discount the LOAEL of 18 mg/kg/d in
females for the cell-mediated response to Sheep Red Blood Cells and in male for peritoneal
macrophage recruitment. Barton and Clewell cite the difference between the naive and vehicle
controls for the former and do not mention the latter. Regarding the effect on peritoneal
A-6
-------
macrophage recruitment, Sanders et al. state that the vehicle control levels in the males were
higher than their historical controls, so that may be the reason for not considering that a critical
effect.
Use of the statistically significant effects that were observed in the Sanders et al. study at
18 mg/kg/d as the critical effects would have resulted in a HED NOAEL for immune effects of
less than 1 mg/kg/d, so this is an important issue.
In Table 4-2, row 3, the 50-mg/kg/d dose is a LOAEL for liver weight and kidney weight
according to Table 4 of the original Berman et al. (1995) paper. The text and Table 3 of the
Berman paper list 150mg/kg/d as the LOAEL in the multivariate ANOVA, which included
serum liver function test values and liver histopathology, as well as liver weight. Barton and
Clewell (2000) considered 50 mg/kg/d to be a LOAEL for liver toxicity in the Berman et al.
study. It is not clear why the Berman et al. paper was not used by either Barton and Clewell or
in the draft assessment to derive a LOAEL for kidney toxicity of 50 mg/kg/d with 14 day dosing.
This would have lowered the HED for nephrotoxicity and should therefore be explicitly
addressed in the draft assessment. The Maltoni et al. study (1986) from which the HED in Table
4-2 is derived dosed male Sprague-Dawley rats for 52 weeks and then followed the animals until
their natural deaths, so some recovery from nephrotoxicity might have occurred before the
kidneys were evaluated. In the Berman et al. study, female Fisher 344 rats were dosed for 14
days and the kidneys were evaluated the 24th day after the last dose. Thus there were sex, strain,
and experimental differences between the two studies.
In Table 4-2, row 3, the 150 mg/kg/d dose is a NOAEL for neurotoxicity following 14d
of exposure by gavage (Moser et al., 1995). Interestingly, the neurotoxicity of DC A, a
metabolite of TCE, was found in another study by Moser et al. (1999) to be significantly greater
by the drinking water than the gavage route, to be duration-dependent, and to be greater when
dosing was started at weaning than in adulthood. Although the latter study tested DCA rather
than TCE, it raises considerable uncertainty about the extrapolatability of the 150-mg/kg/d
neurotoxicity NOAEL to chronic situations, to susceptible subpopulations (the young), or to
drinking water exposure (the more relevant route in humans). The observation that, for chronic
oral dosing with DCA, bolus administration of the entire daily dose by gavage results in less
toxicity than gradual administration of the same dose over the 24 hour period in the drinking
water suggests that one should not assume that gavage dosing with TCE is equivalent to dosing
via the drinking water. This suggests a research need for direct comparison of the effects of
gavage and drinking water dosing with TCE on various critical endpoints.
The discussion above points out additional issues that EPA may need to consider.
However, the final decision on selection of endpoints should be based on the weight-of-evidence
and consistency of findings across well-conducted studies. For example, the Berman et al. study
was a research study designed to develop screening methods and used 8-animals/dose group and
exposure was only 14 days. Certain endpoints measured in this study may have been better
evaluated in other more robust subchronic studies. Careful critical evaluation of the strengths
and weaknesses of studies is essential in the selection of key studies and critical endpoints. As
mentioned in general comments, tabulation of studies will go a long way towards making the
EPA's decisions transparent and help the EPA identify areas that require further discussion in the
text.
In regard to presentation and communication of the Agency's justification for choosing
NOAELs and LOAELs for the RfC, in Table 4-3, row 2 of the draft assessment, the 30 ppm
LOAEL is given as the mean time-weighted average exposure for the workers in the Chia et al.
(1997) and Goh et al. (1998) study and the mean exposure duration is stated to be 5 years. The
A-7
-------
30-ppm figure was obtained by personal breathing zone monitoring on 12 of 85 individuals who
participated in the study. There is no indication given in the articles as to whether it is reasonable
to assume that exposure levels in the factory had not changed significantly over the past 5 years.
This would be important to know to assess whether it is appropriate to use the 30-ppm level as
the mean chronic exposure level. The articles also do not mention whether the workers had any
other concomitant exposures. It is also not clear why the authors chose to use analysis of
variance, with years of TCE exposure as a categorical variable, rather than linear regression,
with years of TCE exposure as a continuous variable, as the method of analysis. In fact, the
categories into which TCE exposure duration is grouped are different in the two papers even
though the data are the same. Another shortcoming is the absence of an unexposed control
group. Despite these shortcomings, the authors did find significant relationships between some
serum hormone concentrations and years of exposure to TCE using analysis of covariance with
adjustment for age, smoking, and testicular size. There were significant inverse relationships
between TCE duration and serum FSH, sex hormone binding globulin, and insulin levels. There
was a significant positive relationship between dehydroepiandrosterone (DHEAS) levels and
TCE duration. There was also a significant negative correlation between years of TCE exposure
and testosterone levels, but this relationship became non-significant after adjustment for age,
smoking, and testis size. The argument can be made that the authors should not have adjusted
for age or testes size because neither one meets the criteria for a potential confounding variable
of being significantly associated with the exposure variable and the outcome variable. Age
clearly is a surrogate for years of TCE exposure in that older workers are likely to have more
years of exposure. Although testosterone levels tend to decrease with age, this is typically not
observed until after the age of 50. In this young cohort it is very unlikely that the significant
negative correlation observed between age and testosterone levels represents the effect of age on
testosterone. It is more likely that age is acting as a surrogate for exposure and, thus, that
adjusting for age will falsely reduce the relationship of interest (i.e., the relationship between
exposure and testosterone). For testes size, there is no reason to believe that this variable would
be independently associated with exposure, and therefore, it does not make sense to adjust for it.
The LOAEL of 40-60 mg/L urinary TCA (20 ppm TCE) listed in Table 4-3, row 5 for the
Rasmussen et al. (1993) study is better justified than the LOAEL in the Chia (1996,1997) and
Goh (1998) papers. The authors use historical exposure monitoring data from the Danish Labor
Inspection Service for the period 1947-87 to establish that this was the typical TCE exposure
range during that historical period. One flaw of the exposure assessment of this study is that a
significant subset of the subjects (25 of 99) were primarily exposed to CFC 113 rather than TCE,
but these two groups were lumped together in the calculation of the cumulative solvent-exposure
index. Nonetheless, there is a convincing dose-response relationship between increasing solvent
exposure and number of abnormal coordination tests.
Although the LOAELs for the human neurotoxicity and endocrine toxicity studies have a
good bit of uncertainty associated with them, as illustrated by the above discussions, the
estimated HED LOAELs from 5 different human studies fall within a remarkably narrow range
(7-16 ppm). Additional confidence in using these levels as the point of departure is provided by
the subchronic rat neurotoxicity study by Arito et al. (1994) from which a HED LOAEL of 9
ppm was derived, and the LOAEL of 12 ppm for liver toxicity in the subchronic mouse study by
Kjellstrand et al. (1983). It may be more appropriate to use the animal studies as the critical
studies, with the other human studies as supportive studies. This is mainly because none of the
human studies had long-term exposure data on the individual level. The study by Ruitjen et al.
(1991) had the best exposure information (area samples from the plant spanning several decades
and specific information about when changes such as installation of exhaust ventilation
occurred) and chronic exposure indices based on the monitoring data were calculated for each
subject. The Vandervort and Pelakoff (1973) and Okawa and Bodner (1973) studies were
A-8
-------
National Institute for Occupational Safety and Health (NIOSH) Health Hazard Evaluations that
had good current exposure data in the form of personal air samples and urinary TCA. Both
studies lacked information about chronic exposure conditions, and the Okawa study lacked a
control group or any analysis of an exposure effect relationship.
In regard to the Agency's choice of NOAELs and LOAELs for the RfD, the Panel does
not agree on whether an uncertainty factor for duration is needed, given that the LW/BW is
considered more of an early event in the toxicity process and a sensitive indicator of potential
liver effects observed at later times. The sizes of the changes are small (12 % -7%) (see Barton
and Clewell, 2000). Other panelists supported the use of a ten-fold uncertainty factor. Several
panelists questioned why EPA uses a full ten-fold uncertainty factor in the derivation of the RfC,
but decreases the factor to only three-fold for the RfD. The Panel suggests that EPA clarify this
issue in the draft assessment.
Another view cautioned EPA in treating the LED 10, derived from a benchmark dose
analysis, as a LOAEL. The original basis for the selection of LED 10 is that it most closely
estimated NOAELs for large numbers of developmental studies (Allen, 1994). Some panelists
believed the benchmark dose should be reducing uncertainty, not requiring more uncertainty to
be added. Other panelists argued that the LED 10 is by definition an effect level and that
therefore it more closely approximates a LOAEL. The benchmark dose approach is considered a
more preferred approach to estimate the point of departure that takes into account the
experimental variability.
In regard to the Agency's choice of NOAELs and LOAELs for the RfC, Barton and
Clewell argued that the LOAEL of the Arito et al. (1994) study reflected a minimal, though
statistically significant effect, and should be evaluated to determine if clinically significant
effects were seen at the lowest dose level. Therefore, the application of a 10-fold uncertainty
factor for LOAEL to NOAEL extrapolation is justified if effects are noted in the Arito et al.
neurotoxicity study and are consistent with the weight-of-evidence from other carefully
conducted neurotoxicity studies. The Kjellstrand et al. (1983) study also identified a LOAEL
(for liver weight), and the application of an additional factor may be justified depending on the
duration-response trends for this endpoint. On the other hand, some panelists suggested that if
the benchmark dose (derived by Barton and Clewell) is used, then an uncertainty factor should
not be automatically applied. There was a question in the minds of some panelists about why
only a three-fold uncertainty factor is used for the LOAEL to NOAEL extrapolation in the
derivation of the RfD, when a ten-fold factor is used here. They suggested that the Agency
improve the consistency of the draft assessment by applying a standard ten-fold factor
throughout. Other panelists disagreed and felt that the use of different uncertainty factors for the
RfD and RfC could be appropriate if justified by the scientific evidence.
2.5. Other Factors
Some panelists believed that use of medications and presence of diseases is part of the
default human variability factors and should not be double counted, while others pointed out that
these might further extend the range of variability and would require an additional uncertainty
factor. The Panel advises EPA to consider a point raised by Dr. Clewell in his public comments
submitted to the Agency, namely, that induction of CYP2E1 is not likely to have a major impact
on increasing oxidative metabolism at environmentally relevant concentrations of TCE. This
argument may or may not be correct and would be a good topic for future research
A-9
-------
APPENDIX B
RECENT RESEARCH ON GENERAL PHARMACOKINETIC DIFFERENCES
BETWEEN CHILDREN OF VARIOUS AGES AND ADULTS
General pharmacokinetic differences between children of various ages and adults have
been explored by Ginsberg et al. (2002) and Hattis et al. (2002, in press). Specifically, for
pharmacokinetic parameters including elimination half-life, clearance, and volume of
distribution, these workers assembled a database of human observations, principally from the
pharmaceutical literature, and created a series of combined analyses combined analyses using a
regression-type approach. This analytical technique allowed them to bring data from many
different chemicals together to assess geometric mean ratios of the values seen for children of
particular ages in relation to adults.
Table 2 shows antilog (geometric mean) results from this type of analysis for elimination
half lives, together with ± 1 standard error uncertainty ranges, for the database as a whole, and
for drugs sorted by their major elimination. As indicated by the examples in Table 2, dividing
the database into subsets according to either pharmacokinetic processing categories or
pharmacodynamic modes of actions can allow inferences to be made about different age-related
dosimetric adjustment factors that may be applied for members of those groupings. Using this
same framework, analyses were also made of the distribution of residuals from the
models—yielding an indication of how often the results for individual chemicals would be
expected to differ by various amounts from the geometric mean within each group.
It can be seen in Table 2 that overall premature infants show about a 4-fold prolongation
of elimination half life; and infants under 2 months of age have about double the half life of
adults. The 6-month to 2-year age group shows, if anything, a slightly shorter geometric mean
half life than in comparable adult studies. The departures of individual findings from these
overall tendency tend to show much greater inter-individual pharmacokinetic variability in the
time periods from 1 week to 2 months when the switch from early infant patterns of metabolism
to more mature patterns takes place (Hattis et al., 2002 in press).
Based on these data, one can suspect: a) greater internal exposure per unit external
exposure to unchanged trichloroethylene in the youngest infants; b) greater persistence of any
activated metabolites that are generated in the youngest infants (although the rates of generation
may be slower); and c) perhaps some more rapid metabolism and generation of activated
intermediates per unit of external exposure for babies in the 2-6 month age range. More definite
conclusions would of course require more specific data on the patterns of age-related changes in
the several activation vs detoxification steps that trichloroethylene undergoes in people. No
specific data are available at present about developmental changes in the rates of generation of
the glutathione-transferase metabolites, the balance between glutathione and P450-metabolic
intermediates, or the rates of elimination of the activated metabolites generated by these two
competing pathways.
B-l
-------
Table 2. Geometric Mean Ratios of Child/Adult Elimination Half-Lives.6
Pa
Major Elimination
All pathways
All CYP (P450
All Non-CYP
Unclassified
Premature
3.89
4.52
3.43
Full term
1.96
1.83
1.80
More detailed classification:
CYP1A2
Renal
Glucuronidation
CYP3A
CYP2C9
Other, mixed CYP's
Other Non-CYP's (not
2.74
4.40
5.28
0.41
9.45
2.78
2.98
2.08
2.19
1.27
1.22
Iwk- 2
1.93
3.51
1.46
4.29
2.75
2.15
1.91
1.05
2-6 mo
1.17
1.22
1.06
1.24
1.15
0.98
0.77
6 mo - 2 yr
0.79
0.51
0.98
1.00
0.57
0.81
1.19
0.41
0.55
2-12yr
0.98
0.61
0.92
0.94
0.54
0.60
1.36
0.61
0.77
1.08
1.24
12- 18 yr
1.11
0.73
1.11
1.13
1.47
0.73
1.41
rentheses show the ± 1 standard error range.
6 Data represent regression results from 135 data groups for 41 drugs, Log(Arithmetic Mean Half-Life) Data. Please
consult Hattis et al. (2002) for information about the equation fit.
B-2
-------
APPENDIX C
Biosketches of Members of the US EPA Science Advisory Board (SAB) Trichloroethylene
Health Risk Assessment: Synthesis and Characterization Review Panel (TCE Review
Panel).
Anderson, Henry: Wisconsin Division of Public Health, Proposed Chair of the TCE
Review Panel and Current Chair of the SAB's Environmental Health Committee. Also a
current member of the SAB Executive Committee.
In 1980 Dr. Anderson joined the Wisconsin Department of Health and Social Services as
the State Environmental and Occupational Disease Epidemiologist. In 1991 he also assumed the
duties of Chief Medical Officer. Among his duties for the State of Wisconsin has been the
development of the scientific support draft assessments for Wisconsin's Groundwater
Enforcement Standards. One standard promulgated was for TCE. He was also responsible for
state fact sheets on TCE in air and at hazardous waste sites.
He received his MD degree in 1972 and entered an Internal Medicine internship and then
an occupational medicine residency. He was certified in 1977 by the American Board of
Preventive Medicine with a sub-specialty in occupational and environmental medicine and in
1983 became a fellow of the American College of Epidemiology. He holds adjunct
Professorships at the University of Wisconsin - Madison, Department of Preventive Medicine
and the University of Wisconsin Institute for Environmental Studies, Center for Human Studies.
He has published over 160 scientific articles on a broad spectrum of environmental, occupational
and public health topics. He is principal investigator on nine active grants and cooperative
agreements from federal government agencies including the U.S. EPA. None of these focus
upon TCE, although the ATSDR Superfund Site Assessment Cooperative Agreement has
evaluated sites contaminated with TCE and conducted exposure assessments.
His US EPA funded research grants address children's health issues, such as reproductive
and endocrine function of frequent Great Lakes sport fish consumers and evaluation of women's
awareness of mercury toxicity and sport fish consumption advisories. Other current research
includes, childhood asthma, lead poisoning, arsenic in drinking water, youth occupational health,
occupational fatalities and bioterrorism response. His expertise includes public health,
preventive, environmental and occupational medicine, respiratory diseases, epidemiology,
human health risk assessment and risk communication.
He was a founding member of the Agency for Toxic Substances and Disease Registry
(ATSDR) Board of Scientific Councilors (1988-1992). He served on National Academy of
Sciences/Institute of Medicine (NAS/IOM) committees that developed the reports "Injury in
America" and "Nursing, Health & Environment." He was a member of the Armed Forces
Epidemiology Board. He is current chair of the Environmental Health Committee of the USEPA
Science Advisory Board and past chair of the SAB Integrated Human Exposures Committee. He
serves on the USEPA SAB Executive Committee. He serves on several other FACA committees
including the Director's Advisory Board for the National Center for Environmental Health,
Centers for Disease Control and Prevention, the Hanford Health Effects Subcommittee for
ATSDR and is a member of the NIOSH Advisory Board on Radiation and Worker Health. He is
a fellow of the Collegium Ramazzini and the American Association for the Advancement of
Science. He is associate editor of the American Journal of Industrial Medicine and serves on the
editorial board of Cancer Prevention International.
C-l
-------
Blair, Aaron: National Cancer Institute, National Institute of Health
Dr. Blair is Chief of the Occupational Epidemiology Branch of the Division of Cancer
Epidemiology and Genetics, National Cancer Institute. His research has focused on cancer risks
from agricultural exposures, industrial chemicals, physical inactivity, occupational exposures
among women, and methodolologic issues in occupational epidemiology. He has over 250
publications. He has evaluated the risk of non-Hodgkin's lymphoma, leukemia, and multiple
myeloma among farmers in the first case-control studies to obtain detailed information on
pesticide used and application practices. This work has culminated the development of the
Agricultural Health Study, a long-term prospective study of 90,000 farmers and their spouses in
Iowa and North Carolina. His studies of cancer mortality among workers exposed to the
important industrial chemicals formaldehyde and acrylonitrile were among the first to employ
sophisticated algorithms to develop quantitative estimates of exposure in multi-company studies.
He has evaluated cancer risks among women in studies of dry cleaners and aircraft maintenance
workers, who have significant exposures to various organic solvents including
tetrachloroethylene and trichloroethylene. Methodologic studies have focused on confounding,
meta-analysis, and misclassification in exposure assessment.
Dr. Blair has served on: IARC Monograph Working Groups; Environmental Protection
Science Advisory Panel Subgroup on Atrazine; Federal Panel on Formaldehyde; National Center
for Toxicologic Research Consensus Conference on Formaldehyde; IARC Workshop on
Priorities for Epidemiologic Studies on Occupational Cancer; Advisory Committee to Trans-
Canadian Study of Lymphatic and Hematopoietic Cancers; Task Force on Environmental Cancer
and Heart and Lung Disease; Advisory Panel to Bureau of Chronic Disease, Health and Welfare,
Canada on Future Research Directions; Farmers Study Advisory Committee, Health and
Welfare, Canada; Advisory Group for Canadian Environmental Health Survey, Health and
Welfare, Canada; NIH Inter-Institute Breast Cancer Working Group; Science Advisory
Committee for the Lower Mississippi River Interagency Cancer Study; Louisiana State
University Medical School; DHHS Environmental Health Policy Committee Subcommittee of
Data Needs; Expert Panel on Domestic Use of Pesticides, National Cancer Institute of Canada;
NCI Program Review Group on Leukemia, Lymphoma, and Multiple Myeloma; Cancer
Research Methods Team; National Occupational Research Agenda, NIOSH; NCI Intramural
Advisory Board; National Toxicology Board of Scientific Counselors; and on Organizing
Committees for Conferences on Assessment of Smoking in Occupational Studies, Exposure
Assessment in Occupational Investigations, and Physical Activity and Cancer.
He has served on Editorial Boards of the American Journal of Epidemiology,
Scandinavian Journal of Work, Environment and Health, and the Journal of Agricultural Safety
and Health. Dr. Blair is a member of the American Epidemiologic Society and a Fellow and
Board Member of the American College of Epidemiology.
Academic Degrees: B.A., Kansas Wesleyan University 1965 Biology; M.S. North
Carolina State University, 1967, Botany; Ph.D .North Carolina State University, 1970, Genetics;
M.P.H., University of North Carolina, 1976, Epidemiology.
Borghoff, Susan J.: CUT Centers for Health Research
Dr. Susan Borghoff has been a Staff Scientist at CUT Centers for Health Research in the
Research Triangle Park, North Carolina since 1989 following her postdoctoral fellowship. Prior
to her position at CUT, Dr. Borghoff was a graduate student at the University of North Carolina
and conducted her research at the National Institute for Environmental Health Sciences
(NIEHS).
Along with Dr. Borghoff s research program at CUT she is also the Director of Education
Programs which involves oversight of the pre- to post- graduate training programs and K-12
educational outreach activities. Her research interests have focused on understanding the mode-
C-2
-------
of-action by which specific chemicals cause kidney toxicity and cancer in rats with a view to
understanding the relevance of this response for human risk assessment. Her research has also
focused on understanding the metabolism and pharmacokinetics of various chemicals with
emphasis on the development of physiologically based pharmacokinetic models that can be used
for risk assessment.
Currently Dr. Borghoff s research is focused on the developmental pharmacokinetics of
estrogen-like compounds such as genistein. CUT Centers for Health Research is a not-for-profit
research institution in which the major core funding is a grant from the American Chemistry
Council Long-Range Research Initiative. Other financial support comes from government
agencies (U.S. Environmental Protection Agency (USEPA) and NIEHS), independent research
organizations, trade associations, and corporations. Dr. Borghoff s research projects have been
funded both by the Core research program and through specific research grants from Oxygenated
Fuels Association, American Petroleum Institute, American Chemistry Council and ARCO (now
Lyondell) Chemical Company. She has recently accepted an opportunity to consult for
Huntsman Chemical Company which involves conducting a literature review on what is known
on the health effects of methyl tertiary butyl ether.
Edler, Lutz: German Cancer Research Center
Dr. Edler is the Head of Biostatistics at the Research Programme Genome Research and
Bioinformatics of the German Cancer Research Center in Heidelberg Germany. He holds a Dipl.
Math (M.S.) Mathematics, Physics from the Albert-Ludwigs-University, Freiburg, FRG and a
Dr. rer. nat (Ph.D.) Mathematics from Johannes-Gutenberg-University, Mainz, FRG. His major
areas of research are: Mathematical-statistical modeling of carcinogenesis and risk assessment;
Pharmacokinetics and development of methodology for clinical oncology with a strong emphasis
on the application computational statistics; Statistical Computing; Biostatistical Methods in
Design and Analysis of Experiments; Mathematical and Statistical Modeling in Oncology; and
Survival Analysis and Clinical Trials.
From 1990-1991 he was a Visiting Scientist, National Institute of Environmental Health
Sciences, Division of Biometry and Risk Assessment, Research Triangle Park, U.S.A.
He has listed the following "Expert Meetings" in which he has participated: (1994)
DAAD, Bad Godesberg; (1994) Human PBPK Models for TCDD, NIEHS, Research Triangle
Park, USA; (1994, 1998) EUROSTAT, Luxembourg; (1998) Risk Assessment of
Electromagnetic Waves, US NIEHS, Tucson, AZ, USA; (2000) Risk Assessment of Dioxin, US
EPA, Fort Collins, USA; 5th Framework Program, EU, Brussels; (1998) Rapporteur at EMF
Science Review Symposium of the NIEHS, Phoenix, AZ; and (2002) Working Group of US-
Vietnam Scientific Conference on Human Health and Environmental Effects of Agent
Orange/Dioxins, March 2002, Hanoi, Rep Vietnam.
He is a member of the following professional societies: American Statistical Association
(ASA); Drug Information Association (DIA); International Biometric Society, German Region
(IBS.DR); International Society for Clinical Biostatistics (ISCB); International Association for
Statistical Computing (IASC); Bernoulli Society; Deutsche Krebsgesellschaft (DKG);
Gesellschaft fuer Medizinische Dokumentation und Statistik (gmds); and International Statistical
Institute (ISI, elected).
Professional Activities include: (1991-1995) Scientific Secretary International
Association for Statistical Computing (IASC); (1995-1997) Vice President of the International
Association for Statistical Computing (IASC); (1999 -2001) President of the International
Association for Statistical Computing (IASC); (1993-1997) Member of the Council of the
German Region International Biometric Society; (1998-2002) Member of the Council of the
International Biometric Society; and (1993- now) Member of the Animal Protection Commission
at the RegPr. Karlsruhe. Currently he is 2002 Co-Organizer of the Session 'Clinical Trial" at the
International Biometric Conference, Freiburg, Germany; 2002 Coorganizer of the Session
C-3
-------
Tharmacogenetics and Pharmacogenomics Data Analysis Methods in Future Clinical Trials',
38th DIA Annual Meeting, Chicago; 2003 Chair of the International Organizing Committee of
the International Conference on Carcinogenesis Risk Assessment (ICCRA), Athens, Greece; and
2004 Co-Chair of the Local Organizing Committee of the Biometrical Colloquium of the
German Region of the International Biometric Society, Heidelberg, Germany.
His grants include: (Feb. 1991) Visitor at the Universidad Nacional de Colombia at
Bogota, Columbia; (1990) DFG Travel Grant for 48th Session of the ISI in Cairo, Egypt; (June,
1993) DAAD Travel Grant for a visiting lectureship in Columbia; (1995) DFG Travel Grant for
50th Session of the ISI in Beijing, China; (Sep-Dec 1995) Consulting National Institute of
Statistical Sciences (NISS), Res.Triangle Park; and (Aug-Sep 2001) KOSEF-DFG Visiting
Scientist Grant, Yonsei University, Seoul, South-Korea.
He serves on the following committees and Advisory Boards: Advisor for the
Collaborative Project on Knowledgebased Systems in Medicine; Reviewer for the Government
Department of Research and Technology Funding Programme; Reviewer for the DFG; -
Extramural Review Board of the AIO (German Cancer Society); Project Assessment Committee
of the Phase I/II Study Group of the AIO; Independent Safety Committee for Boehringer
Mannheim Co.; and Reviewer for the German Cancer Society and Krebshilfe.
Currently his editorial tasks include: (since 1993) Associate Editor of Computational
Statistics and Data Analysis (CSDA) and Associate Editor of ONKOLOGIE; (since 1994)
Associate Editor of the Biometrical Journal (Biometrische Zeitschrift); (since 1999) Associate
Editor of Journal of Cancer Research and Clinical Oncology; and (since 2002) Editor of the
Virtual Online Journal "Biostatistics" (Elsevier, Publ.)
Hattis, Dale: Clark University (Current member, SAB Environmental Health Committee)
Dale Hattis is Research Professor with the Center for Technology Environment and
Development (CENTED) of the George Perkins Marsh Institute at Clark University. For the
past twenty-seven years he has been engaged in the development and application of methodology
to assess the health ecological and economic impacts of regulatory actions. His work has
focused on the development of methodology to incorporate interindividual variability data and
quantitative mechanistic information into risk assessments for both cancer and non-cancer
endpoints.
Specific studies have included quantitative risk assessments for hearing disability in
relation to noise exposure renal effects of cadmium reproductive effects of ethoxyethanol
neurological effects of methyl mercury and acrylamide and chronic lung function impairment
from coal dust four pharmacokinetic-based risk assessments for carcinogens (for
perchloroethylene ethylene oxide butadiene and diesel particulates) an analysis of uncertainties
in pharmacokinetic modeling for perchloroethylene and an analysis of differences among species
in processes related to carcinogenesis.
He has recently been appointed as a member of the Environmental Health Committee of
the EPA Science Advisory Board and for several years he has served as a member of the Food
Quality Protection Act Science Review Board. Currently he is also serving as a member of the
National Research Council Committee on Estimating the Health-Risk-Reduction Benefits of
Proposed Air Pollution Regulations.
The primary source of his recent cooperative agreement support is the U.S.
Environmental Protection Agency and specifically the Office of Research and Development's
National Center for Environmental Assessment. This research includes: (1) Age related
differences in susceptibility to carcinogenesis; towards a quantitative analysis of empirical data.
Instrument number (Term: April 2002-Sept 2003); (2) Methods for evaluating human
interindividual variability regarding susceptibility to particulates (Term Sept 98—September
2002); and (3) also funding from the State of Connecticut to work on Child/Adult differences in
pharmacokinetic parameters, as a subcontractor as part of a cooperative agreement.
C-4
-------
He has been a councilor and is a Fellow of the Society for Risk Analysis and serves on
the editorial board of its journal Risk Analysis. He holds a Ph.D. in Genetics from Stanford
University and a B.A. in biochemistry from the University of California at Berkeley.
Hoel, David: Medical University of South Carolina (Current member, SAB Environmental
Health Committee)
David G. Hoel, Ph.D., is a Distinguished University Professor at the Medical University
of South Carolina. Dr. Hoel received his A.B. degree in Mathematics and Statistics from the
University of California at Berkeley and his Ph.D. from the University of North Carolina at
Chapel Hill and has more than 25 years of experience as a biostatistician, toxicologist and
environmental health researcher.
Dr. Hoel's research specialties include: environmental causes of cancer, risk assessment
models; statistical and mathematical applications in biology and medicine; epidemiology; and
radiation health effects. Dr. Hoel is widely published, having authored or co-authored over 160
journal articles and co-editor of several books and journals. He serves on a variety of national
association committees and panels, such as a member of the Institute of Medicine, Agent Orange
Committees, EPA's Science Advisory Board.
He is a member of the National Academy of Sciences Institute of Medicine, is a National
Associate of the National Academy of Sciences and National Research Council and a Fellow for
the American Association for the Advancement of Science. Before joining the faculty at the
Medical University Dr. Hoel was a division director at the NIEHS of NIH. This division was
made up of four branches with responsibility for the Institute's program in biostatistics,
epidemiology and molecular toxicological risk assessment.
Sources of recent grant and/or contract support: include: (1) Savannah River Site Former
Production Workers Medical Surveillance Program - Phase II Year Continuation (funded by the
Department of Energy)--the goal of this project is to assess occupational exposures reviewed by
former DOE workers at SRS and conduct appropriate medical examinations in order to evaluate
work related illness and risk.; (2) "Low Dose Radiation Project" (funded by the Department of
Energy, Environmental Biosciences Program); the goal of this project is to develop methods for
estimating cancer risks from low dose and low dose rate ionizing radiation; (3)."Radiation
Leukemogenesis: Applying Basic Science to Epidemiology Estimates of Low Dose Risks and
Dose-Rate Effects" (funded by the Department of Energy)--the goal of this project is to
incorporate biological information into mathematical models of radiation induced leukemias; and
(4)"Radiation Risk Analysis: Model Issues and Interspecies Extrapolation"(funded by the
National Opinion Research Center/NASA)—the goal of this project is to use and evaluate
experimental animal data for estimation of human health risks from radiation.
Lambert, George: Robert Wood Johnson Medical School/ University of Medicine and
Dentistry of New Jersey (Current member, SAB Environmental Health Committee)
Dr. Lambert is Associate Professor of Pediatrics, Director Division of Pediatric
Pharmacology and Toxicology at the University of Medicine and Dentistry of New Jersey,
Robert Wood Johnson Medical School - Piscataway/New Brunswick. He is also the Director of
the NIEHS/EPA Center for Childhood Neurotoxicology and Exposure Assessment, which is
located at the Environmental and Occupational Health Sciences Institute, a jointly sponsored
institute of Rutgers, The State University of NewNJ and UMDNJ-Robert Wood Johnson
Medical School
He holds a B.S. in zoology from University of Illinois, Champaign-Urbana (1968) and an
M.D. from the University of Illinois, Chicago, IL (1972).
Recent grants and other outside funding sources include the following: (1) a grant to
study the Reproductive Outcomes of the World Trade Center Tragedy (funded by National
C-5
-------
Institute of Environmental Health Sciences) (2) a grant to determine the influences of
environmental exposure to neurotoxicants on child neurological health and development with
special emphasis on autism and related disabilities (funded jointly by the National Institute of
Environmental Health Sciences and the Environmental Protection Agency) (3) a grant to study
the effects of Herbal Phytoestrogens & Prostate Cancer (funded by the Cancer Commission of
New Jersey); (4) Effects of eating Crabs with PCBs and Dioxin Laden on Human Health
(funded by the New Jersey Department of Environmental Regulations); (5) a grant to study the
role of gene polymorphisms in Birth Defects, (funded jointly by the Centers for Disease Control
and the NJ State Birth Defects Registry); and (6) the correlation between hypospadism and
xenoestrogens (funded jointly by the Centers for Disease Control and the New Jersey
Department of Health).
Lemasters, Grace: University of Cincinnati (Current member, SAB Environmental Health
Committee)
Dr. Lemasters is a Professor in the Division of Epidemiology and Biostatistics
Department of Environmental Health, College of Medicine, University of Cincinnati and former
head of Epidemiology and Biostatistics in the Department of Environmental Health, College of
Medicine.
She holds a Ph.D., Department of Environmental Health, College of Medicine,
University of Cincinnati, Epidemiology and Environmental Health Science; M.S.N, University
of Cincinnati; and aB.S.N., Indiana University.
For almost three decades she has conducted researched in occupational and
environmental epidemiology and investigating health effects including ergonomics and
musculoskeletal research, respiratory disease, cytogenetic effects, and childhood allergy and
asthma. Dr. LeMasters is a national and international expert in occupational and environmental
health studies and has published numerous scientific articles and book chapters book in the areas
of exposures and health effects and study design methodologies.
She has conducted research on men and women in the military for over 15 years
examining the effects of exposures to fuels and solvents on cytogenetics, female hormones, male
reproduction and neurological effects. Other areas of research include a 15-year pulmonary
longitudinal study of the health effects of refractory ceramic fiber exposure (substitute for
asbestos) and lung cancer and lung disease. She has recently received funding as the principle
investigator on a 5-year study on diesel exposure and atopy and respiratory disorders in children.
Other current research includes the following: caffeine effects on female hormones during early
pregnancy, occupational risk factors related to falls, and exposures of women in the military to
jet fuel and hormonal changes.
Among her service on Committees and Associations she lists: Federal Advisory
Committee on Children's Health NICHD (2002- ); Armed Forces Epidemiological Board (2001-
present); Reviewer Department of Defense PRMRP (July 11-13, 2001; Member National
Toxicology Program Board of Scientific Counselors of the Office of the Assistant Secretary and
Surgeon General (1999-2002); Editorial Board: Occupational and Environmental Medicine
(1996-2001); Editorial Board: Journal of Reproductive Toxicology (1991-); Fellow, American
College of Epidemiology; Member, Society for Epidemiology Research; and Member: Sigma
Theta Tau Alpha and Beta Honors Chapters.
Current sources of recent grant and/or contract support are the: Environmental Protection
Agency; NIH-CDC/NIOSH; NIH-NIEHS; and the Refractory Ceramic Fiber Coalition.
C-6
-------
Li, Abby: Monsanto Company (Current member, SAB Environmental Health
Committee)
Dr. Abby Li received her Ph.D. from the University of Chicago in pharmacology and
physiology. She is currently a Senior Science Fellow at Monsanto. She is a toxicologist in the
Department of Toxicology and Human Health Risk Assessment. She has specialized expertise in
neurotoxicology as well as product stewardship responsibilities involving general toxicology,
exposure and risk assessment issues. Dr. Li has conducted numerous studies primarily for
regulatory submission in neurotoxicology in adult and developing rats, in humans and in vitro
systems.
She was Monsanto's Neurotoxicology Team Leader responsible for developing testing
capabilities at Monsanto including motor activity, schedule controlled operant behavior,
functional observational battery, auditory startle habituation, learning and memory and
neuropathology. She has also conducted in vivo pharmacokinetic studies (ADME studies) and in
vitro metabolism studies. Dr. Li served on the Editorial Board of Neurotoxicology from 1995-
2001.Dr. Li was invited by the US EPA country representative to serve on the US team of
experts to develop international OECD guidelines on neurotoxicity (1995 - 1998) and
developmental neurotoxicity (1996-2000). Dr. Li is the Chair of the Neurotoxicology Technical
Panel of the American Chemistry Council's Long Range Initiative (ACC LRI) responsible for
funding research to advance the field of neurotoxicology in focus areas such as susceptible
populations, and in developing new methods for hazard and exposure assessment. She served as
Co-Chair of Crop Life America's Developmental Neurotoxicology Working Group in 2000 and
is currently a member of this group. She is a member of the EPA's Science Advisory Board's
Environmental Health Committee and reviewed the EPA's 1999 draft cancer guidelines, the RfC
Methods Case Studies, and the Lead 403 Rule among other draft assessments. Dr. Abby Li was
a peer consultant to the September 10-11, 1996 EPA Benchmark Dose Peer Consultation
Workshop
Luderer, Ulrike: University of California at Irvine (Current member, SAB
Environmental Health Committee)
Dr. Ulrike Luderer is Assistant Professor of Medicine in the Division of Occupational
and Environmental Medicine at the University of California at Irvine. She also holds joint
appointments in the Departments of Developmental and Cell Biology and Environmental
Toxicology. Dr. Luderer's research focuses on mechanisms of action of reproductive toxicants
and on protective mechanisms against those toxicants. She is a recipient of a National Institute
of Environmental Health Sciences research grant (2002-2007) entitled "Glutathione:Protecting
Ovarian Follicles from Oxidant Injury" and a co-investigator on an EPA grant "Latent Effects of
Gestational Exposure to Heptachlor" She has published peer-reviewed journal articles and book
chapters and presented research at national and international scientific conferences on such
topics as the effects of toluene exposure on reproductive endocrine function, the functions of and
regulation of glutathione in the ovary, the differential regulation of follicle-stimulating hormone
and luteinizing hormone secretion, and reviews of reproductive and developmental and
endocrine toxicology. She has served on the National Toxicology Program/NIEHS Center for
the Evaluation of Risks to Human Reproduction Expert Panel on 1- and 2-Bromopropane and on
the National Research Council subcommittee on methyl bromide. She is currently a member of
the EPA SAB's Environmental Health Committee. Dr. Luderer has a Ph.D. in reproductive
endocrinology and an M.D. from Northwestern University and is board-certified in Internal
Medicine and in Occupational and Environmental Medicine. She has a Sc.B. in biomedical
engineering from Brown University.
C-7
-------
McClain, Michael: McClain Associates
Dr. R. Michael McClain is currently an Adjunct Professor University of Medicine and
Dentistry of NJ and now works primarily as a consultant in toxicology. He was formerly a
Distinguished Research Leader and Director of Toxicology, Hoffmann-La Roche, Inc. Dr.
McClain received his Ph.D. from the Department of Pharmacology at the University of Iowa and
B.S. and M.S. degrees from Duquesne University. Dr. McClain is a Diplomate of the American
Board of Toxicology and a Fellow of the Academy of Toxicological Sciences. He has worked in
the pharmaceutical industry for over 30 years in the areas of teratology and reproductive
toxicology, general toxicology and carcinogenicity testing. His research activities are involved
primarily in mechanisms of chemical carcinogenesis for thyroid, liver and adrenal and regulatory
aspects for cancer risk assessment. He has been active in the Pharmaceutical Research and
Manufactures Association and PhRMAs efforts on harmonizing international guidelines for drug
development (ICH). He has been involved with the ILSI organization and served as President of
the ILSI's Health and Environmental Science Institute (HESI) and as a member of ILSI's Board
of Trustees. Dr McClain is a member of the National Advisory Environmental Health Sciences
Council for NIEHS. Dr. McClain is also active in the Society of Toxicology having served a
term as Treasurer and as President of the Society in 1998
Melnick, Ronald: National Institute of Environmental Health Sciences
Dr. Melnick is a Senior Toxicologist and Director of Special Programs in the
Environmental Toxicology Program at the National Institute of Environmental Health Sciences
(NIEHS), National Institutes of Health in Research Triangle Park, North Carolina. Prior to this
position he was Group Leader of the Toxicokinetic and Biochemical Modeling Group in the
Laboratory of Computational Biology and Risk Analysis at NIEHS. Dr. Melnick obtained his
B.S. degree from Rutgers University and his Ph.D. in food science/biochemistry from the
University of Massachusetts at Amherst. He was a postdoctoral research fellow in the
Department of Physiology-Anatomy at the University of California in Berkeley and then an
assistant professor of life sciences at the Polytechnic Institute of New York. At NIEHS he has
been involved in the design, monitoring and interpretation of NTP toxicity and carcinogenesis
studies, as well as mechanistic studies to characterize the behavior of environmental carcinogens.
He spent one year as an agency representative to the White House Office of Science and
Technology Policy to work on interagency assessments of health risks of environmental agents
and on risk assessment research needs in the Federal government. Dr. Melnick has organized
several national and international symposiums and workshops on health risks associated with
exposure to environmental and occupational toxicants. He has also served on numerous
scientific review and advisory panels, including the working group of the International Agency
for Research on Cancer (1995) that classified trichloroethylene as probably carcinogenic to
humans. Dr. Melnick has served on several committees at NIEHS, including Chair of the
Toxicokinetic Faculty and member of the NIEHS review group for the NTP Report on
Carcinogens. The latter group reviewed data on trichloroethylene for listing in the Report on
Carcinogens. Dr. Melnick is a Fellow of the Collegium Ramazzini. As a federal employee, he
does not receive any grant or contract support.
Solomon, Gina: Natural Resources Defense Council
Dr. Gina Solomon is a Senior Scientist at the Natural Resources Defense Council in San
Francisco and an Assistant Clinical Professor of Medicine at the University of California at San
Francisco. Dr. Solomon is a specialist in internal medicine, preventive medicine, and
occupational and environmental medicine. Her work has focused on environmental and
occupational threats to reproductive health and child development. She attended medical school
-------
at Yale and underwent post-graduate training in medicine and public health at Harvard. Dr.
Solomon served on the U.S. EPA's Federal Advisory Committee on endocrine disrupting
chemicals and is a scientific advisor to numerous organizations including the California
Department of Health Services Environmental Epidemiology Section and the Pediatric
Environmental Health Specialty Unit at U.C. San Francisco. Dr. Solomon has published peer-
reviewed articles on various topics, including solvents and miscarriage, endocrine disrupters,
diesel exhaust and asthma, and contaminants in breast milk. She is a co-author of the book,
Generations at Risk: Reproductive Health and the Environment, published by MIT Press in
1999.
Whyatt, Robin: Department of Environmental Health Sciences
Dr. Robin Whyatt is Deputy Director of the Columbia Center for Children's
Environmental Health and is Assistant Professor in the Department of Environmental Health
Sciences at the Mailman School of Public Health, Columbia University. Dr. Whyatt's research
focus is on the effects of environmental exposures on women and children, including the
developing fetus. Prior to coming to Columbia in 1991, she evaluated the extent of pesticide
exposure in the preschooler's diet as Senior Staff Scientist at the Natural Resources Defense
Council (NRDC). Her research at Columbia University has used biologic markers to study
effects of environmental exposures during pregnancy. This has included a molecular
epidemiologic study of prenatal exposures to ambient air pollution and cigarette smoking in
Poland. Dr. Whyatt's is currently collaborating on a comprehensive community-based study of
environmental risks to African American and Dominican mothers and newborns in Northern
Manhattan and the South Bronx. The prospective cohort study is evaluating effects of
environmental exposures on fetal growth, neurocognitive developmental and asthma risk. Dr.
Whyatt's focus is on the extent of exposure to non-persistent pesticides (organophosphates,
carbamates and pyrethroids) during pregnancy among this minority population. Dr. Whyatt is
also collaborating with the Center for Disease Control on the validation of biomarkers of
exposure to contemporary-use pesticides during pregnancy. Dr. Whyatt has published widely on
the application of biologic markers to studies of environmental risks to infants and children and
on the effects of environmental exposures during fetal development. She is currently principal
investigator on three grants: a U.S. EPA STAR grant to validate the measurement of non-
persistent pesticides in postpartum meconium as a biomarker of fetal exposure; a NIEHS RO1
grant to validate a battery of biomarkers of prenatal exposure; and on an intervention grant from
the Speaker's Fund for Public Health Research to reduce residential pesticide exposures during
pregnancy. Dr. Whyatt served on the U.S. EPA Workshop, Critical Windows of Exposure for
Children's Health, and on the U.S. EPA Workshop, Technical Workshop on Issues Associated
with Considering Developmental Changes in Behavior and Anatomy when Assessing Exposure
to Children. She was Co-chair of the Symposium on Alternative Human Matrices for
Biomonitoring, at the 2001 International Agency for Exposure Assessment, Charleston, South
Carolina and is currently serving on the Exposures to Chemical Agents Work Group of the
National Children's Longitudinal Cohort Study. Dr. Whyatt received her Doctorate in Public
Health (Dr.P.H.) from Columbia University with honors in 1995 and her Masters in Public
Health (M.P.H) from Columbia University in 1985.
Yang,Raymond: Colorado State University
Raymond S. H. Yang is presently Professor of Toxicology and Director of Center for
Environmental Toxicology and Technology, one of 14 Programs of Research and Scholarly
Excellence at Colorado State University (CSU). Between July 1990 and June 1995, Dr. Yang
served as the Head, Department of Environmental Health, College of Veterinary Medicine and
Biomedical Sciences, CSU, Fort Collins, CO. Prior to joining CSU in 1990, Dr. Yang spent
C-9
-------
seven years each in chemical industry (Bushy Run Research Institute, Union Carbide - Mellon
Institute, 1976 - 1983) and in the federal government [National Institute of Environmental Health
Sciences/National Toxicology Program (NIEHS/NTP), 1983 - 1990].
Dr. Yang received his B.S. in Biology from the National Taiwan University in 1963;
M.S. and Ph.D. in Toxicology/Entomology from North Carolina State University in 1967 and
1970, respectively. Between 1970 and 1973, he was a postdoctoral fellow at Cornell University.
Between 1973 and 1976, he was Research Associate and then Assistant Professor at the Institute
of Comparative and Human Toxicology, Albany Medical College. Dr. Yang had also been
appointed Adjunct Associate Professor at University of Pittsburgh and Adjunct Professor at
North Carolina State University.
Dr. Yang's research expertise and interests cover many subdisciplines in toxicology,
including toxicology of chemical mixtures, toxicologic interactions, physiologically based
pharmacokinetic/pharmacodynamic (PBPK/PD) modeling, biologically based dose-response
(BBDR) modeling, carcinogenesis and neuro-developmental toxicology. Between 1992 and
2000, he served as the Program Director of the NIEHS Superfund Basic Research Program
Project at CSU and since the summer of 1999 he has been the Program Director for an NIEHS
Quantitative Toxicology Training Grant. Since 1990, Dr. Yang has been developing an
interdisciplinary research program on Quantitative and Computational Toxicology using the
central theme of PBPK/PD, BBDR, and reaction network modeling of chemicals and chemical
mixtures at CSU.
Dr. Yang's committee work includes serving as a Committee or Expert Panel Member for
the following Committee/Panel or organizations: National Academy of Sciences/National
Research Council Safe Drinking Water Subcommittee on Mixtures; USEPA/Environmental
Criteria Assessment Office (ECAO); Screening and Testing Work Group of the Endocrine
Disrupter Screening and Testing Advisory Committee, USEP A; Electric Power Research
Institute (EPRI); Expert Panel Member, Risk Assessment for Mixtures of Drinking Water
Disinfection-Byproducts, International Life Sciences Institute/USEPA; Institute of Medicine,
National Academy of Sciences Committee to Study the Interactions of Drugs, Biologies, and
Chemicals in Deployed U. S. Military Forces; Chair for a Chemical Mixture Workshop Agency
for Toxic Substances and Disease Registry (ATSDR); Health Council of the Netherlands;
Society of Toxicology Expert Panel on Chemical mixtures; Chemical Mixture Committee
member to National Occupational Research Agenda, NIOSH; and NIEHS Environmental Health
Sciences Review Committee. Dr. Yang's research support came principally from the National
Institute of Health (NIH), U.S. Air Force, U.S. Environmental Protection Agency (EPA),
ATSDR, and Center for Disease Control and Prevention (CDC)/National Institute of
Occupational Safety and Health (NIOSH).
C-10
------- |