\ UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
'/ WASHINGTON, D,C. 20460
March 9» 1988
SAB-EHC-88-011
Hon. Lee M. Thomas
Administrator OFFICE OF
THE APMiisj *STR A TOR
U»S. Environmental Protection
Agency
401 M Street, SW
Washington, D.C. 20460
Gear Mr. Thomas T ,, ,-;eA"ri
$'f
Thank you for your thoughtful response of August J to the Science Advisory
Board's review of scientific evidence associated with exposure to perchloroethy-
lene. In your letter you asked the Board to provide further scientific advice
on three issues that will subsequently bear on your risk management decision
for this compound. The Board appreciates this opportunity for further scientific
dialogue on these issues and hopes that its views in this letter can better
promote consensus on the scientific issues under review.
As noted in your letter, the assessment of the scientific evidence frcnt
experimental animal studies centers on the relative significance for humans
of the production of rat kidney and mouse liver tumors. This question is
applicable to a broad range of chlorinated hydrocarbon compounds—including
dichlorcfliethane, para-dichlorobensene, trichloromethane, and triehloroethylene—
which produce tumors of the rat kidney and mouse liver under some experimental
conditions. While recognizing the implications of such issues to these and
other compounds, this letter is directed specifically to an assessment of
perchloroethylene,
in responding to your letter, the Board's Environmental Health Committee
and its Halogenated Qrganics Subcommittee organized a scientific workshop on
August 12, 1987 to explore these and other issues with leading researchers in
the field, SPA staff and members of the public. An agenda of the workshop is
attached. The Board has utilized the Information obtained in this workshop,
and the discussions among Committee and Subcommittee members, to respond to
your August 3 letter and also to advise the Agency on health effects evaluated
in its Draft Health Assessment Document Addenda for Dichloromethane and Tri-
ehloroethylene. The Board's findings and recommendations on these latter two
compounds will be transmitted to you in separate letters. Our response to your
specific questions follows.
Question^1: Assuming that not all animal tumors are of equal significance to
evaluating human hazard, what is the Science Advisory Board's current consensus
position, based on scientific evidence or professional judgment, of the relative
significance of male rat kidney or mouse hepatocellular tumors for human risk
assessment?
_: In general, the Board's consensus conclusion on the significance
of male rat kidney tumors stems from recent research (not yet published, but in
-------
-2-
press) that indicates that for many halogenated organics, probably including
perchloroethylene, the mechanism producing these type of tutors is probably not
operative in humans and, therefore, may not be relevant for human risk assessment.
This mechanism involves the metabolism of the compound in the liver and the bind-
ing of a protein (alpha-2u-globulin) with the metabolite as a conjugate molecule.
This molecule is filtered and accumulates in the kidney. One hypothesis is that the
conjugate is more difficult to metabolize than the alpha-2u-globulin alone. This
protein then accumulates and is injurious to the cell. Repair is followed by a
cancerous formation at the site in a low percentage of cases. Prom available scien-
tific evidence, this mechanism appears to be unique to male rats.
Thus far, thirteen substances have been demonstrated to produce renal
tumors in male rats through this mechanism including perchloroethylene, para-
dichlorobenzene and unleaded gasoline. Trichloroethylene, on the other hand,
appears to produce renal tumors in male rats through a different (unknown)
mechanism, thus creating important implications for human health risk assessment-
The Board's consensus on the significance of mouse liver tumors is that
mechanistic explanations are not sufficiently well developed and validated at
this time to change EPA's present approach expressed in its risk assessment
guidelines for carcincgenicity. It concludes that the generation of mouse
liver tumors by chemicals is an important predictor of potential risks to
humans. Of the several mechanistic models under consideration (including
regenerative hyperplasia, oncogene activation and tri-halcmethyl radical
formation) , the one most promising for immediate application to risk assess-
ment is characterized by proliferation of peroxisomes, an intracellular organ-
elle, in the liver.
Peroxisone proliferation may be important for compounds such as perchloro-
ethylene, but liver tumors observed after exposure to chlorinated solvents
may involve different mechanisms. The importance of understanding the biologi-
cal mechanisms is that they may provide a basis other than the bioassay
statistical analysis for low-dose risk estiination. A plausible mechanism
(peroxiseme proliferation or something else) may imply low-dose nonlinearity
for some substances that induce mouse liver tumors. However, different {pre-
sumably linear) mechanisms may operate for other substances, and these mechan-
isms may be consistent with linearity at low doses or a linear relationship to
dose. These distinctions in low dose risk estimation should be explicitly
included in the quantitative estimate of human risk.
Several substances that induce peroxisome proliferation in rodent livers,
such as hypolipidemic drugs and the plasticizer di-ethylhexylphtalate ( EEHP) ,
also produce liver tumors in rodents. In summary, however, a causal relation-
ship for this mechanism is plausible but unproven.
Some scientists have reported the detection of onccgenes after administra-
tion of presumably non-genotoxic agents-1
Steven H. Reynolds, Shari J. Stbwers, Rachel M. Patterson, Robert R- MaronpoT
Stuart A. Aaronson, Marshall W. Anderson, "Activated CTtcogenes in B6C3F1 Mouse
Liver Tumors: Implications for Risk Assessment," Science Vol. 237 (September
11, 1987), pp. 1309-1316.
-------
-3-
Also, as you are aware, our increasing knowledge of the role of mechan-
isms of promotion (later events in the carcinogenic process) may well clarify
our understanding of cancer induction; certainly this is the cage with dioxin
and may relate to the halogenated hydrocarbons.
CXiest_ipn__2; Wtat is the Board's view of the approach taken by EPA in using
its guidelines to infer human carcinogenic potential from the total body of
scientific evidence on perchloroethylene?
_SAB_ Response; The issues regarding the application of the risk assessment
guidelines appear not to represent disagreanent among scientists about scien-
tific evidence but, rather, the consequence of attempting to fit the weights
of evidence into necessarily arbitrary categories of risk. Since the weights
of evidence, and uncertainties associated with such evidence, for perchloro-
ethylene and other compounds fall within a range of scientifically defensible
choices, it may not be possiblef in some instances, to fit them neatly into
only one risk category. Moreover, the more incomplete the data, the less
precision one can expect in classifying a compound within EPA's cancer guide-
lines. In additionr the type of evidence that places a compound in a particular
category may vary considerably from substance to substance within that category*
For perehloroethylene, as with trichloroethylene, the Science Advisory Board
concludes that the overall weight of evidence lies on the continuum between the
categories 83 and C of EPA's risk assessment guidelines for cancer-
As perehloroethylene illustrates, the distinction between the 82 and C
categories can be an arbitrary distinction on a continuum of weight of evidence.
The "black-white interpretation" that you referred to in your letter is indeed
troubling. Frcm a scientific point of view, it seems inappropriate for EPA
and other agencies to regulate substances that are classified 82 and not to
consider regulation of compounds classified as C, regardless of the level of
human exposure. In the case of %i, BJ. or even A categorized compounds where
exposure levels are lew, EPA may, with scientific justification, decline to reg-
ulate because the potential health effects appear to be trivial in magnitude.
A substance classified as C (limited evidence in animals) for which human
exposure is high may represent a much greater potential threat to human health.
EPA and other agencies (including those in state governments) may, there-
fore, wish to take steps to reduce high exposures to substances in the C cate-
gory whenever there appears to be a potentially significant threat to human health
(in the sense that the plausible upper bound estimate of potency times lifetime
exposure is above the threshold where regulation may be judged appropriate).
indoor exposure to perehloroethylene, such as might be found in dry cleaning
establishments not using the equivalent of good industrial hygiene practices,
could merit action under this criterion. So might high levels of exposure to
other solvents, pesticides or industrial chemicals that have been considered by
the public as "safe" in the absence of sufficient evidence of carcinogenicity
in animals. In many instances, this appearance of safety results from not yet
having the results from well-designed bioassays such as those conducted by the
National Toxicology Program.
-------
Finally, you noted the evaluation of perehloroethylene by the Internation-
al Agency for Research on Cancer (IARC) as an exanple of evolving terminology
for classifying potential carcinogens, m general, the Board believes that
public understanding of oenplex scientific issues is enhanced when scientists
and regulators can speak with a cannon voice, lit view of its own experience
of using the cancer risk assesanent guidelines and, in particular, having to
address the issue of the scientific uncertainty that exists among and within
guideline categories, ERA should re-evaluate its labeling system and methods
for characterizing uncertainty. It should also review whether to be more con-
sistent with lARC's terminology.
Question 3; Is there research underway or anticipated that will clarify these
rodent tumor responses and their relationship to human health risk assessment?
vfcat additional research should be undertaken?
SflB Response; Current research undertaken in various laboratories, including
the National institutes of Health, can reduce some of the uncertainties assoc-
iated with rodent tumor responses. Research results and hypotheses presented
at the Board's August 12 workshop has served to clarify our understanding of
male rat kidney tumors and their significance for human risk assessment. In
addition, the Reynolds et. al. paper supplements our knowledge of activated
oncogenes in mouse livers tumors.
Several research efforts should be initiated to further narrow scientific
uncertainty for perchloroethylene and structurally related compounds. These
Include:
o Validation of mechanistic models for the rat kidney and mouse liver
tumors through experimentation with selected known carcinogens and
non-carcinogens.
o Development of improved methods for assessing low-dose response to
environmental pollutants that induce peroxiscnte proliferation,
Once again, we are pleased to have this opportunity to present the views
of the Science Advisory Board on these in^ortant scientific issues. He hope
that the consensus stated above -assists you in making the difficult risk manage-
ment decisions on perchloroethylene and other compounds.
Sincerely,
Nelson, Chairman
Executive Oonnittee
Richard A. "Griesaner,
, Chair
lenated Ctgaptfcs Subcomittee
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD
ENVIRONMENTAL HEALTH COMMITTEE/HALOGENATED ORGANICS
SUBCOMMITTEE (COMBINED ROSTER)
CHAIRMAN
Or. Richard A. Griesemer, Director, Biology Division, Oak Ridge National
Laboratory, Martin Marietta Energy Systems, Inc., P.O. Box Y, Oak Ridge,
Tennessee 37831
CHAIRMAN OF THE HALOGENATED ORGANICS SUBCOMMITTEE
Dr. John Ooull, Professor of Pharmacology and Toxicology, University of Kansas
Medical Center, Kansas City, Kansas 66103
MEMBERS
Dr. Seymour Abrahamson, Professor of Zoology and Genetics, Department of
Zoology, University of Wisconsin, Madison, Wisconsin 53706
Dr* Linda Birnbaum, National Institute of Environmental Health Sciences,
P.O. Box 12233, Research Triangle Park, North Carolina 27709
Dr. George T, Bryan, Department of Human Oncology, University of Wisconsin,
K»4, Room 528, 608 Clinical Science Center, 600 Highland Ave., Madison,
Wisconsin 53792
Dr. James Bus, Pathology and Toxicology Research, Upjohn Company,
Kalamazoo, Michigan 49001
Dr. Gary Carlson, Department of Pharmacology and Toxicology, School of
Pharmacy, Purdue University, West Lafayette, Indiana 47907
Dr, Robert Dedrick, Chief, Chemical Engineering Section, National Institutes
of Health, Bldg. 13, Room 3W13, Bethesda, Maryland 20892
Dr. Philip Enterline, Department of Biostatistics, Graduate School of Public
Health, University of Pittsburgh, 130 Desoto Street^ Pittsburgh, Pennsylvania
15261
Dr. David Gaylcr, National Center for Toxicological Research, Jefferson,
Arkansas 72079
Dr. Ronald D. Hood, Professor and Coordinator, Cell and Developmental Biology
Section, Department of Biology, The University of Alabama and Principal
Associate, R.D. Hood and Associates, Consulting Toxicologists, P.O. Box 1927,
University, Alabama 35486
Dr. K. Roger Hornbrook, Department of Pharmacology, P.O. Box 26901, University
of Oklahoma, Oklahoma City, Oklahoma 73190
-------
Dr. Curtis Klaassen, Professor of Pharmacology, Man 1638, University of
Kansas Medical Center, 39th and Rainbow 81vd,, Kansas City, Kansas 66103
Or, D. Warner North, Principal, Decision Focus Inc., Los Altos Office Center,
Suite 200, 4984 El Camino Real, Los Altos, California 94022
Dr. Karl K, Rozman, Department of Pharmacology, Toxicology and Therapeutics,
University of Kansas, 39th and Rainbow Blvd., Kansas City, Kansas 66103
Dr. Stephen Safe, Department of Veterinary, Physiology & Pharmacology Texas
ASM University, College of Veterinary Medicine, College Station, Texas
77843-4466
Dr. Robert Squire, 1515 Label1e Avenue, Ruxton, Maryland 21204
Dr. Thomas Starr, CUT, P.O. Box 12137, Research Triangle Park, North Carolina
27709
Dr. Robert Tardiff, Principal, Environ Corporation, The Flour Mill,
1000 Potomac St. N.W., Washington, O.C. 20007
Dr. Bernard Weiss, Professor, Division of Toxicology, P.O. Box RBB, University
of Rochester, School of Medicine, Rochester, New York 14642
Dr. Ronald Wyzga, Electric Power Research Institute, 3412 Hillview Avenue, P.O.
Box 1041, Palo Alto, California 94303
EXECUTIVE SECRETARY
Dr. C. Richard Cothern, Executive Secretary, Science Advisory Board
[A-1Q1F] U.S. Environmental Protection Agency, Washington, D.C. 20460
NOTE: This combined roster only include those individuals attending the
meeting.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
.WASHINGTON, D.C 20460
THE ADMINISTRATOR
3 !98T
DC* Norton Nelson
Institute of Environmental Medicine
New York. University Medical Center
550 First Avenue
New York, New York 10016
Dear Dr. Nelsons
Thank you for the letter that you and Dr. Richard Griesenee- sent me
on January 27, 1987 submitting the report of the Halogenated Organics
Subcommittee of the Science Advisory Board's Environmental Health Cotwaittee
following its review of the Draft Health Assessment Document Addendum for
Tetrachloroethylene (Perchloroethylene). Your letter has stimulated a
wide and fruitful discussion on a number of scientific issues among EPA
staff scientists and between the 'staff and senior policy officials,
including myself. The thoroughness of these discussions accounts for the
delay in responding to your letter.
The recommendation of .the Halogenated Organics Subeonmittee that
perchloroethylene be classified in Category C of EPA's Risk, Assessment
Guidelines for Carcinogenicity.. disagrees with the 'position taken in 'the
draft assessment document addendum that this compound meets the criteria'
established in the guidelines for a classification of B2. As you know
from having participated in the review of the guidelines, they are not
intended to be used in a rote like fashion but, rather, to represent the
flexible use of the best scientific judgment based upon the weight of the
evidence. Your letter has encouraged EPA scientists and nianagers to re-
examine many of the assumptions and judgments applied both in the guidelines
and in the assessment of perchloroethylene.
I personally welcome this continuation of the scientific dialogue. I
would like to use this opportunity of responding to your letter not only to
build consensus with the scientific cornrounity that the Science Advisory
Board represents, but also with the Congress and the public. We all need
to work together towards a better understanding of the issues involved in
decision making under conditions of uncertainty.
At least two sets of issues are involved in the Agency's evaluation
of your letter, both of which emphasize the need for additional consul-
tations. The first concerns the assessment of the scientific evidence
for perchloroethylene and, the second addresses the use of the results
of the assessment to reach a decision on the hazard of perchloroethylene
by applying the classifications expressed in the risk assessment guidelines
for carcinogenicity. An additional set of issues pertains to how defensible,
fron a scientific point of view, a particular classification decision may
be. This, in turn, raises the issue of how flexible the guidelines are
in evaluating a compound such as perchloroethylene, and their utility in
regulatory decision making in their current form.
-------
- 2 -
More than one scientifically defensible position may exist on
perehloroethylene (or other substances) because of the' uncertainties in
the available scientific data and because of the difficulties in extrapo-
lating effects reported in test animals (such as mouse liver tumors and
certain rat kidney tumors) to humans. The EPA cancer guidelines take the
position that there is a continuum, or ladder, of hazard infonnation depend-
ing on the available evidence. The Science Mvisory Board supported tne
classification system developed in the guidelines, and the evidence needed
to evaluate—both qualitatively and quantitatively—'the risk posed by a
compound. The Halogenated Organics Subcommittee report on perchloroethylene
reminds us that more than one interpretation of the same scientific data
is possible under the guidelines.
I have enclosed the response of EPA staff scientists to the Subcom-
mittee's report. For reasons that are provided in the response, the staff
believe that the scientific evidence seems to more closely support a B2
classification of perchloroethylene. After your review of the EPA. staff
response and before I reach a final decision in the coming weeks, however,
I would like to have the benefit of further scientific advice from you and
other members of the Science Advisory Board on some important issues that
will bear on my decision, Ihese include:
o Assuming that not all 'animal, tumors are of equal significance ,to ,
evaluating human hazard, what is the Science Mvisory Board's current
consensus position, based on scientific evidence or professional -
judgment, of the relative significance of male rat kidney or mouse
hepatocellular tumors for human risk assessment?
o What is the Board's view of the approach taken by EPA in using its
guidelines to infer human carcinogenic potential from the total
body of scientific evidence on perchloroethylene? I note, for exam-
ple, the recent decision by the International Agency for Research
on Cancer to classify perchloroethylene as a 2B carcinogen, based
upon a determination of "sufficient" evidence of carcinogenicity
in animals, under its new classification system but to label 2B as
a category of "possible" human carcinogens.
o Is there research underway or anticipated that will clarify these
rodent tumor responses and their relationship to human health risk
assessment? Wrtat additional research should be undertaken?
It is important to understand that a decision on the classification of
any compound under the cancer guidelines is not an EPA decision to regulate
that compound? however, it does weigh heavily on the type or extent of the
possible regulation, especially under certain environmental statutes. A
decision to regulate a compound represents a statement of potential hazard
in the absence of other factors such as exposure* A regulatory decision
by EPA on whether to control the sources of a specific ecnpound, and the
degree of control, mist necessarily weigh hazard, potency, exposure and
other factors. It is clear, however, that SPA'S classification of a
-------
compound has major ramifications beyond its use in EPA's own decision
making process. Rightly or wrongly, state environmental decisions and
public perceptions of risk are often triggered by an EPA determination
to classify a compound as a B2 carcinogen. This black-white interpreta-
tion of the classification system is troubling. '
I note with interest the initiative taken by the Board in organizing
a workshop on specific rodent tumors on August 12. I applaud your leadership
in assembling scientific experts from a number of agencies and organizations
to broaden the level of understanding on these issues. The workshop
furthers my desire to continue a scientific dialogue on issues related
to risk assessment and clearly addresses the issues I must decide relative
to perchloroethylene.
Given the importance of these issues and their relationship to the
regulatory decisions that I Bust make, I would appreciate your expeditious
response to the issues I have identified, or others that you believe are also
significant. Again, I thank you and other SAB members for your continuing
efforts to improve the scientific basis for decision making in EPA,
"/*. Sincenely,
Lee M, Thomas
-------
30198?
EPA Staff Comments on Issues Regarding the
eareinogenicity of Perch!oroethylene (Perc) Raised by the SAB
Introduction
Environmental Protection Agency staff scientists have reviewed the
comments of the SAB and Its Halogenated Organics Subcommittee and has prepared
this detailed response to issues raised. These include points raised by the
SAB about the NTP data and their Interpretation by both the NTP and the Agency.
As discussed in this document, some of these issues may be subject to alternative
scientific interpretation. It would also appear that additional discussion
regarding the Agency's Interpretation of its Risk Assessment Guidelines would
be valuable* Agency staff will be pleased to engage in such discussions with
the SAB, Both quality of science and consistency of approach to data interpre-
tation are important Issues for the'Agency in its dealings with the NTP, with
its advisory groups, and with the public. The SAB's comments on the March,
1986 perchloroethylene addendum and the responses that follow suggest the need
to focus closely on a number of these issues.
Quality Assurance of the Bioassay
The Agency staff agrees with the opinion of the subcommittee that the
quality assurance Information available on the NTP bioassay suggest that this
was a scientifically acceptable, well-run study. The NTP 1s convinced that the
study data» documents, and materials in the NTP archives support the data
and interpretations presented in their technical report. The Agency will
incorporate additional statements regarding the implications of the quality
assurance information in the final version of the addendum.
-------
based on a Life Table analysis. The Fisher Exact test values were inadvertantly
replaced by these values in the preparation of the document. The appropriate
valuest as calculated by the NTP and confirmed by Agency statisticians ape
as follows:
Control
Overall Rate 1/49 (2%)
Life Table P = 0.054
Fisher Exact (one-tailed)
200 PPM
3/4§ (6%}
P = 0.259
P * 0.309
400 PPM
4/50 (8%)
P = 0,070
P = 0.187
The correct values for the Fisher Exact test will appear in the final version of
the addendum. In addition, a discussion of the results of the trend and Life
Table analysts included in the NTP report will be presented^
3) The SAB raises questions about the rat kidney response seen in this
study. The NTP indicates in the technical report that because these lesions
appeared consistently in dosed animals but not in controls and are considered
uncommon tumors, they are considered to be caused by perchloroethylene. As of
August, 1985, the historical Incidences for adenomas in lattelle PNL chamber
controls and in NTP program untreated controls are 1/249 (0.4$) and 4/1,968
(0.21) respectively. No malignant renal tubular cell tumors have ever been
observed in control groups. The latest information available on these control
rates will be Included in the final addendum, iiven that no malignant renal
tumors have been seen in control animals, the probability of finding 2
malignancies 1n SO animals by chance with a historical control incidence of
0/1,968 is highly unlikely (p< .001, Fisher Exact test, one-tailed).
-------
4) The Agency staff can find little support for the subcommittee view
that renal tubular cell neoplasia in control groups tends to be under-reported.
In this specific case, while the subcommittee did receive public comment that
review of control groups from latelle showed undiagnosed renal tumors {1 in
each of 3 groups examined), follow-up review by an NTP panel of pathologists
failed to confirm this discrepancy. The Agency has contributed to the development
of a well designed and thorough review process for NTP studies and staff has
considerable confidence in them. Of course, NTP data, as with all other data
considered by the EPA, should be subject to further review by both Agency and
other scientists.
5} Although the kidney response is elevated but not significant at
a p « 0.5 level compared to the concurrent control group, it does represent an
increased incidence of a rare carcinogenic event apparently attributable to
perch!oroethylene. While not reason enough on its own to suggest a strong
positive carcinogenic response in rats, this rare event does add to the weight
of evidence. Although the relevence of such a response to humans, especially
at low doses has been questioned, there is little doubt that the kidney is a
target organ for perchloroethylene and othtr chlorinated ethanes and ethylenes
in mammalian species and this also contributes to our concern.
o Mononuclear Cell Leukemia in Rats
1) The SAS disagreed with the Agency's (and the NTP's) position that a
marginally statistically significant increase in mononuclear cell leukemia is
seen in both sexes of rats exposed to perchloroethylene 1n the NTP study. Our
analysis of these data does not support the supposition that "faulty pathological
diagnoses or some unusual circumstances in the rat colony at the time" are the
basis for this response.
-------
2) The SAB has stated that "the pathology of these tumors Is not well
understoodt and little background Information is available in the literature."
However, Agency staff concur with NTP scientists on the point that rat mononuclear
cell leukemia is a well characterized disease. There appears to be a substantial
data base characterizing the disease both biologically and pathologically. The
literature speaks to origin, time of onset, progression, and clinical pathology
of this disease. References to this work will be included in the discussion of
mononuclear cell leukemia in the final version of the addendum.
3) The SAE suggests that "the results of the statistical analysis are not
convincing" and takes issue with the use of "staging in the assessment of this
response," The Agency's position with regard to the relationship between the
statistical analysis of the leukemia response and the staging of that response
will be clarified in the final document. The Fisher Exact test shows a marginally
statistically significant response in both sexes. In contrast to the somewhat
equivocal findings In male rats, the leukemias in the female rats seem to be
clearly elevated by perchloroethylene. Discussions with the NTP confirm this
fact and suggest that the staging was used as a supplemental analysis to determine
whether the statistical difference might be due to differential diagnosis of
the lesions. This situation might have been suspected had the difference been
due primarily to early stage leukemia in the dosed groups. However, this was
not the case. Considering only advanced (stage 3 as defined by NTP criteria)
cases resulted in slightly less significance to the response in males and
slightly wore in females compared to the overall incidence. Staging did not,
therefore, strongly influence the significance of this response. As a comparative
-------
diagnostic test, staging seems quite appropriate and the similarities or
differences with the staging of human leukemias as a measure of leukemic
progression seems to be a moot point.
4) The SAB states that "18 versus 30 is not a striking observation" with
regard to the leukemia incidence in control and female low dose groups respec-
tively. Based on evidence accumulated by the NTP. however, 19/50 is the highest
incidence ever setn in an untreated control group 1n some 2000 animals which
have been examined in the Program. Dr. Joseph Haseman of tht NTP has addressed
the question of a chance finding of such results if two of six experimental
groups are drawn at random. In in internal memorandum responding to the SA8
comments. Dr. Haseman states "the answer 1s that it 1s extremely unlikely (exact
p-value difficult to determine precisely, but at least p
-------
6) Agency scientists believe that there is scientific support for the
conclusion that perchloroethylene exposure is associated with increased frequency
of mononuelear cell leukemia in rats, based on the NTP data. While scientists
may disagree on the strength of the evidence, it is generally felt that these
data show "some evidence" of Perehloroethylene-related cardnogenicity if not
"clear evidence", to use NTP's terms. For an indication of the range of opinions
among the NTP's Board of Scientific Counselors Review Panel, see p. 14-15 of
the NTP Technical Report (NTP, 1986). it is reasonable to conclude that these
results add to the weight of evidence for a carcinogenic response in rats. At
present, there is no satisfactory explanation as to why these results were not
seen in previous studies. Shortcomings of the earlier studies and the fact
that a different route of exposure was used may have played a role in the
Inconsistency of reported response. Exposure via the oral route was more
acutely toxic and caused significantly higher mortality in all three dose
groups in both sexes in the NCI study (1977). In the only other rat bioassay
of Perc, survial was significantly reduced in the high dose groups of both
sexes (Rampy et a!.» 1978). Since mononuclear cell leukemia 1s a disease of
old age, high non-tumor mortality in treated groups in both of these studies
could have masked late appearing cancer effects. The results of the completed
NTP Savage Study of four rat strains may help to clarify this issue, A determination
as to the utility of these studies for assessing the potential carcinogen!city
of perchloroethylene in rats will be made once the resolution of audit problems
is completed by the NTP. A decision in this regard is expected by September,
1987.
§ Liver Tumors in Mice
1) The SAB suggests that, adenomas and carcinomas be analyzed both
separately and combined. Such an analysis has been presented on p. 3-14 of the
7
-------
draft Addendum. The statement by the SAB that "It should be remembered that
many mice with hepatic carcinomas also Nave adenomas that have not Been Included
in the summary tallies" Is also answered in tne draft Addendum, Mice with both
lesions are enumerated in the footnotes in Table 3-6 on p. 3-14, In no dose
group was the incidence of mice with both lesions more than 10%.
2) Despite acknowledging that the results of this study confirm previous
results and predictions based on route to route extrapolation, the SAB concludes
that "no new, dispositive information has been gained". The fact that a strong
carcinogenic response has been demonstrated in two separate experiments, in
different laboratories, using different routes of exposure, producing similar
dose-related responses, increases the weight of the evidence that the response
is indicative of a carcinogenic response in animals. While this Interpretation
can be debated because the response is seen in mouse liver and is accompanied
by some non-neoplastie pathology, the confirmatory finding as well as the nature
of the response is viewed by many in the science community as "sufficient
evidence" of an animal carcinogenic response, as is stated in the Agency's
Carcinogen Risk Assessment Guidelines. Additional support for this view comes
from the recent deliberations on the classification of perchloroethylene by
IARC. The final statement of the IARC review group is expected to show an
upgrade of the animal evidence of eardnogenicity of perch!oroethylene from
"limited" to "sufficient" based strongly on the new confirmatory data on mice
from the NTP study. It is the position of the Agency, therefore, that the new
inhalation liver tumor data from the NTP study should add to the weight-of-the-
evidence determination for perchloroethylene.
-------
Other Data from the NTP Bloassay
The SAB has requested that "Agency staff fully assess all of the Information
available from the National Toxicology Program study.11 The Agency agrees that
additional discussion of other data from the NTP bloassay should bi included in
the final Addendum, Although not statistically significant, increases 1n
pathology are noted in the kidneys, respiratory tract and brain in both sexes
and in the testes 1n males. Adequate discussion of these responses and their
potential Implications for an analysis of neoplasia in these animals 1s presented
in the NTP report and will be summarized in the Addendum. A discussion of
mortality outcomes including a time-to-death analysis will be provided to
assess the hypothesis that the kidney is also a target organ for chronic
toxicity in the mice,
Stati stlcal jnalysjs of theINgassajy Results
The Agency has used one-tailed statistical analyses 1n this Addendum and
will state this as the SAB requests. Agency statisticians as well as those at
the NTP feel that this is an appropriate approach when trying to detect a
carcinogenic response. While recognizing that chemicals may influence tumor
incidence 1n either a positive or negative way» in this case* no evidence of
decreased tumor incidence 1s found, nor would a two-tailed test radically
change the significance of the results. The fact that the rat leukemia data
is shown to be significant even by a two-tailed analysis might be viewed as
adding additional weight to the significance of the response. This point will
be added to the discussion.
Met abo1jsm and Pharmacok1net i cs
The Agency appreciates the SAB's insights on some of the assumptions made in
the pharmacokinetic modeling of perchloroethylene. Much has transpired since
the subcommittee meeting with regard to the physiologically-based pharmaeokirietit
-------
modeling of chlorinated solvents. The final addendum will present the results
of an interagency effort to reach consensus on aspects of modeling these
compounds. This effort will not substantially change the presentation as
reviewed by the SAB, but will provide some alternative approaches and their
respective strengths and weaknesses given the current state of our knowledge on
this chemical and this field. Specific assumptions related to a hypothetical
genotoxic metabolite will be revisited in the final addendum.
Mechanism
The SAB "hypothesizes that, operationally, perchloroethylene may be an
indirect acting carcinogen or carcinogenic promoter of low potency." Of the
six pieces of "evidence" laid out by the SAB to support this contention, none
addresses the Issue of promotion directly. The facts that perchloroethylene
is not mutagenic 1n routine'testing and binds to DNA only minimally does not
indicate a promoting compound but suggests interpretation of the bioassay-data
as promotion by default. The fact that perchloroethylene induces liver carcinoma
in B6C3F1 mice does not argue strongly for either an indirect or a promoting
carcinogen since this does not appear to be the only species or organ site
affected by perchlorotthylene and direct-acting, gtnotoxic carcinogens also
induce liver tumors in mice. The fact that perchloroethylene is a peroxisome
proliferator still does not fully explain its carcinogenic mechanism nor confirm
its promotional capability. While perchloroethylene produces trichloroacetic
acid (TCA) which is a potent peroxisome proliferator, TCA also acts as a complete
carcinogen, perhaps Indirectly, in the mouse liver (Herren-Freund et al, 1987
in press), and specifically does not enhance the response in ethyl nitrosourea
pretreated animals. Perch!oroethylene's consistent carcinogenic behavior when
compared with other halo-su&stituted ethylenes argues neither for nor against a
10
-------
specific mechanism. In summary, data do not support the contention that
perch!oroethylene 1s solely a promoter, and the evidence for an indirect mechanism
of carc1nogen1e1tyf while accumulating for certain tumor sites, is far from
certain. The mechanism(s) by which chlorinated hydrocarbons, and specifically
perch!oroethylene produce a carcinogenic response remains unknown. While
peroxisome proliferation may play a role in the mouse liver response, this
finding is not sufficient to explain the rat kidney response (See Goldsworthy
and Popp, 1987, for a discussion of this issue). These issues will only be
resolved with additional research. At the present tlmt^ A|ency staff have
maintained a position of concern for a potential for carcinogenic response in
humans exposed to chemicals producing such responses 1n animals. Qualitatively,
no strong scientific argument can yet be made for the Irrelevance of these
tumors to man on a mechanistic, basis, although differences of opinion on their
relevance are recognized by the scientific community.
Epidemiology
The Agency will Include a discussion of the latest informition on epidemiologic
studies related to perchloroethylene exposure. Given the projected potency of
carcinogenic response based on animal data, 1t is unlikely that studies with
sufficient power to determine a positive response will be forthcoming.
Weight of Evidence Category
While recognizing that the available data may be subject to alternative
interpretations, the staff believes that perch!oroethylene should fall more
readily Into the B2 category* probably carcinogenic to humans, based on sufficient
animal but Inadequate human evidence of eardnogenielty. The Agency's interpre-
tation of the bioassay dita described above leads to the conclusions that l) a
11
-------
statistically significant tumor response related to perchloroethylene exposure
is seen In. both mice and rats; 2) three tumor sites ire identified (rat leukemia
and kidney and mouse liver); 3} positive results are found in the mouse by two
routes of administration (gavage and inhalation); 4} a shortening of time-to-tumor
is noted In both species at two different sites (mouse liver and rat leukemia);
5) malignant mouse liver tumors art seen in both sexes and at both doses; 6)
renal tumors 1n the rat show a dose-related trend; 7) metabolism of perchloro-
ethylene appears to be similar in rodents and man; and 8) the mechanism of
carcinogenic action of perchloroethylene is currently unknown although several
mechanisms have been hypothesized. Both the rat data 1n_ toto when considered
with the mouse data, or the mouse data alone, when viewed as described abovet
support this classification under the current EPA Guidelines. The Agency
currently takes the position that these rodent responses are generally considered
to be relevant to man in the context of hazard identification, that is, there
is a potential for producing a carcinogenic response at some dose in humans.
While the Agency has considered arguments against the use of each of these
animal responses individually to assess the hazard to humans, the overall
weight of the evidence is the more pertinent issue. The sum of the carcinogenic
responses and the confirmatory nature of the mouse data would seem to override
the position adopted by the SAB that the animal evidence is only limited. The
decision of IARC, when published, is expected to further support the view that
results of the NTP study raise the weight of the evidence of animal carcinogenicity
of perch!oroethyltne from limited to sufficient.
Unless downgraded on the basis of data Indicating a species-specific
response, sufficient evidence in animals with inadequate evidence in humans is
indicative of a probable human carcinogen under EPA's current Guidelines for
12
-------
Carcinogen Risk Assessment, According to Agency representatives to an IARC
committee drafting a new IARC Monographs preamble, the IARC will also use the
terras "probable" and "possible" huraan carcinogen in Its overall evaluation of
carcinogens. Since the classification schemes are not exactly comparable
sufficient animal evidence will place perch!oroethylene in Category 28 which
the IARC will term "possible" carcinogenic to man. The implications of use of
the terms "possible" and "probably" in EPA's risk assessments in light of this
development should be a point for discussion among Agency scientists and the
SAB.
Quant itat i ye Pot ency Est1mate
The Agency agrees with the SAB that the basis for the range of upper bounds on
the risk should be very clear in the final Addendum. The,influence of various
approaches and assumptions, including tht use of physiologically based pharmaco-
kinetic modeling, on the risk estimates will be described in greater detail or
clarified in the final Addendum. The implications for Agency-decision making
of both the qualitative and quantitative assessments for perch!oroethylene
are recognized and both points were explicitly included in the current addendum
summary on pages 1-2 and 1-5, that is while being ranked as a 82 carcinogen
(qualitatively), its relative potency is among the lowest evaluated by the
Agency (quantitatively).
Miscel1aneous Issues
1) A discussion of research needs will be added to the final addendum.
Recent discussions in the Inter-Agency Health Risk Assessment Committee of
the Integrated Solvents Project will provide the basis for much of this
discussion. Issues relating to metabolism, pharmacokinetics and pharmaco-
dynamics are of particular interest to this group.
13
-------
2) In response to the SAB's request, the Agency willsupply the SAB with
a recent report from a jointly funded project entitled, "Investigation of
Cancer Risk Assessment Methods" prepared by K.S. Crump and Company, This
report is the result of i two year study to examine the assumptions, other
thin those involving low dose extrapolation, used in quantitative cancer
risk assessment. The study was funded by the Department of Defense [through
an interagency transfer of funds to the Environmental Protection Agency (EPA)],
the EPA, the Electric Power Research Institute and, in its latter stages, by
the Risk Science Institute, The objectives of the study were as follows;
1. To identify and express quantitatively uncertainties that are involved
in the process of risk estimation, excluding the uncertainties in the
low dose extrapolation model;
2. To examine the impact of the different assumptions that are made in"
risk estimation;
3. To compare results calculated from human and animal data, including
the identification of the assumptions that produce the best correlation
of risk estimates between humans and animals;
4. To develop guidelines for presenting a range of risk estimates based
on different but scientifically acceptable assumptions or assumptions
that have considerable backing in the scientific community.
These objectives are pursued using empirical methods in which cardnogenicity
data for 44 chemicals are analyzed systematically- in a variety of ways, particular
attention is placed on those 23 chemicals for which there exist data from both
animal and epidemlological studies suitable for making quantitative comparisons.
14
-------
Conclusion
The Agency's staff scientists have responded to the questions and Issues
raised by the SAB, focusing mainly on the evidence of carcinogenesis In animals
provided by the NTP study* The Agency has presented evidence for the appropriate-
ness of using the mononuclear cell leukemia response in the rats and liver
tumors In mice in a weight-of-evidence approach and has presented additional
interpretive considerations for the rat kidney tumor response. It concludes
that, despite uncertainties with extrapolation from each of these endpoints
taken individually, the body of evidence for carcinogenicity in both rats and
mice is sufficient which leads to the conclusion that perch!oroethylene would
probably be carcinogenic to humans at some dose: a 62 carcinogen under £PA
Guidelines, We hope that this response to comments will strengthen the
articulation of the Agency's position on the potential carctnoienicHy of -
perch!oroethylene as will be found 1n the final Addendum.
15
-------
References
Goldsworthy, T.L.; Popp, J.A. (1987). Chlorinated Hydrocarbon - Induced Peroxisomal
Enzyme Activity 1n Relation to Species and Organ Carcinogenicity. Toxicol.
Appl. Pharmaco!. 88: 225-233.
Baseman, J, (1987). Intirnal memorandum to the Director, NIEHS concerning the
Report of the Halogenated Qrginlcs Subcommittee of the Environmental Health
Committee on the NTP Perch!oroethylene Study. Memorandum dated March 13» 1987,
Herren - Freund, S,U; Pereira, H.A.; Qlsen, G, (1987). The Carcinogenicity of
THchloroethylene and Its Metabolites* Tr1chloroacet1c Add and Dlchloroacetic
Acid, 1n Mouse liver. Toxicol. Appl. Pharmacol., in press.
McConnell, E.; Solltveld, H.; Swenberg, J.; Boortnan, Q. (198&). Guidelines for
combining neoplasms for evaluation of rodent carclnogenesis studies. J. Natl.
Cancer Inst. 76; 283-289.
National Toxicology Program (NTP). (1986). Toxicology and Carclnogenesis
Studies of Tetrachloroethylene (Perchloroethylene) in F344/N flats and
Mice (Inhalation Studies). Technical Report 311.
Rampy, L.; Quast, J; Leong, Br; iehrlng, P. (1978). Results of a long-term
inhalation toxicity study on rats of a perch!oroethylene (tetrachloroethylene)
formulation. Toxicol. Res. Lab., Health and inviron. Res,, Dow Chemical USA,
Midland, MI.
16
------- |