UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
April 10, 1985
Honorable Lee M. Thomas
Admini strator
U.S. Environmental Protection Agency OFFICE OF
401 M Street, S.W, TMIE ADMINISTRATOR
Washington, D*C. 20460
Dear Mr. Thomas:
The Environmental Health Committee of IPA's Science Advisory Board has
completed its review of the Agency's draft lealth Assessment Document for
Chloroform. The stated purposes of the document are to serve as a source
document for Ageneywlde use and to serve as a scientific basis for decision
making on hazardous air pollutants by the Office of Air and Radiation.
The Committee found the Chloroform document to be of better overall
scientific quality than draft health assessment documents for other sub-
stances that the Committee has reviewed during the past year* In addi-
tion, the Committee appreciated the responsiveness of Agency scientists in
responding to its questions during the review* The Committee agrees with
the conclusions stated in the document that! (1) chronic exposure to Chloro-
form is associated with renal, cardiac, neurological and hepatotoxle effects,
and (2) sufficient pharmacoklnetic data exist for Chloroform to incorporate
this information in the quantitative risk estimates. Before the final version
is printed, the Committee requests that the Agency perform additional work to
improve the document in three areas. These include carcinogen!city, mutagenicl-
ty and teratology. Detailed comments on these and other Issues are provided in
the attached technical report. In summary, the Committee concludes that the
Health Assessment Document for Chloroform is scientifically adequate for its
stated purposes* Unless the Agency requests additional advice from the Committee,
the current draft should not require further review.
We appreciate the opportunity to comment on this public health issue and
stand ready,to provide any further scientific advice. We request a written
response to our advice.
Sincerely,
Richard A. Grieaemer, B.V.M., Ph.D.
Chair» Environmental Health Committee
Norton Nelson, Ph*D.
Chair, Executive Committee
Enclosure
-------
REPORT OF THE ENVIRONMENTAL HEALTH COMMITTEE OF SPA'S SCIENCE ADVIS01Y
BOARD REGARDING A DRAFT HEALTH ASSESSMENT DOCUMENT FOE CHLOROFORM
INTRODUCTION
On December 20, 1984, the Chlorinated Qrgaaics Subcommittee of the
Environmental Health Committee reviewed a draft Health Assessment Document
for Chloroform [EPA-600/ 8-84-QQQ4A; March, 1984; External Review Draft].
The doeunent was prepared by the Office of Health and Environmental Assess-
ment (OilA) In the Office of Research and Development* The Subcommittee
report* signed by the Chair, Dr. John Doull, Is separately available. Sub-
sequently, the draft document was reviewed by the full Committee. The
Committee's najor conclusions and technical comments are presented below.
The draft document, Including revisions proposed by OHBA staff, gen-
erally was well-written, In that*.
(1) the relevant literature was analyzed critically,
(2) Interpretation of the literature was organized about a central
point of view,
(3) the document exhibited an open-minded approach to the data,
(4) the different chapters reached scientifically reasonable con-
clusions , and
(5) an effort was made to interrelate the results of different
studies within chapters and to Integrate the conclusions of
different chapters.
In addition, the scientists In charge of revising different portions of
the draft document cane to the meeting prepared to comment on advice about
closely related issues from other Committee reviews. Therefore, a produc-
tive dialogue ensued.
MAJOR COMMENTS
The Committee requests that OHEA scientists do further concerted work
on three chapters: carcinogen!city, mutagenlcity and teratology. The pri-
mary reasons for the Cocunittee's requests are discussed in this section.
More detailed comments are discussed in the appropriate sections below*
these comments should not be misconstrued as indicating a desire by the Com-
mittee to review the draft further.
Using the criteria developed by the International Agency for Research
on Cancer (IARC), the document concludes that sufficient animal evidence of
the carcinogeniclty of chloroform exists, but that the epldemlologlc evi-
dence is Inadequate. Consistent with the IARC criteria, chloroform would
be placed into Category 2B, meaning that it is probably carcinogenic to
humans. The Committee agrees with the conclusion regarding the animal evi-
dence, but requests that OHIA staff further analyse the conclusions regard-
ing eptdemlologic evidence. The Committee understands that the dividing
line between the IARC definitions of "Inadequate" and "limited" evidence
-------
-2-
1s a fine one. However, the criterion of limited evidence is that "a.
causal explanation is credible," and the evidence regarding chlorinated
drinking water nay appear to meet this standard.
The Committee agrees with the statement: In the Executive Summary that
chloroform is a potential Inhalational teratogen. However, the supporting
text interprets only teratogenic effects. The corresponding evidence for
» fetotoxlc effect of chloroform Is described but not evaluated. The Com-
mittee suggests that OHM analyze both teratogenlclty and fe tot oil city
data for levels at which no effects are observed.
The Committee believes that the conclusion that chloroform m&y be a
weak autagen is not veil documented. It Is equally plausible that this
substance is a nonmutagen and that false positive results have been obtain-
ed in a few tests. While more or newer tests night detect mutagemlc activ-
ity at lover concentrations, current testing Methods provide the state-of-
-the-art definition of a mutagen.
EXPOSURE
For the Committee's background Information, the Office of Mr Quality
Planning an4 Standards provided a written "Summary of Exposure Information
for Chloroform (Trichloroaethane)", which was helpful in evaluating the
health effects information in the draft document. The health assessment
document also presents information regarding the atmospheric chemistry of
chloroform, which lies beyond the expertise of the current members of the
Committee to review. The Committee understands, however, that OBEA has
obtained expert peer review of this section elsewhere.
Data exist on the pharmacokinetlcs of chloroform in three mouse
strains, as well as rat, monkey and man. These data suggest that chloro-
form Is rapidly and completely absorbed. The unchanged portion of chloro-
form principally is excreted through the lung. Ho evidence exists for
different metabolic pathways. Therefore, the Committee suggests that OEEA
can reasonably assume that In different species the qualitative metabolic
pathways for metabolism do not differ among mammalian species.
Chloroform metabolism Is saturated (or close to saturated) at bloaasay
doses* This Implies that correction to internal dose will affect the
slope of potency estimates. Chloroform is metabolized to reactive Inter-
mediates, which will bind covalently to cellular macromolecules. Evidence
Is consistent with an increase la covalent binding to a point of saturation.
One reactive metabolite, phosgene, is a causative factor in the mechanism
of renal and hepatic toxic effects of chloroform.
The time period of 7 to 8 days (cited on page 3 of the Carcinogen
Assessment Group's "Replies to Public Comments" on the draft document,
dated October 10, 1984} will appply at any given exposure, due to the na-
ture of first order kinetics. The Committee Is curious whether any addi-
tional data exist regarding the kinetics of elimination at different
exposures. If the kinetics are hiexponential, then unusual kinetic
-------
-3-
phenonema can occur, but the data suggest a one compartment model for
chloroform. Thus, acute versus chronic exposure will not change the
elimination rate. The Committee also suggests that the chapter on pnarmaeo-
kinetics Illustrate glueuronlde formation in Figure 4-6 (see page 4-36),
OHEA should correct the carbon monoride-anaerobic pathway, as discus-
sed at the neeting* Metabolism of chloroform to phosgene may not be the
only active product involved. Electrophilic halogens also may be produced.
The February 1984 issue of Treads, in Pharmacological Sciences contains an
article on this subject**
G1HERAL TOOCOLOGT
The acute toilcity section is well-written and is a scientifically
defensible statement of the literature on this substance. The Committee
agrees with the conclusions reached that chronic exposure to specific
concentrations of chloroform elicts renal, cardie, neurological and hepato-
toxic effects*
In addition to the draft health assessment document, the Committee
received a revised chapter on mutagenieity prior to the meeting from
OHEA's Reproductive Effects Assessment Group. The Committee believes that
this chapter attempts to interpret critically an extensive and somewhat
contradictory data base*
The revised chapter concludes that the available information supports
the finding that chloroform may be a "weak" (or low potency) mutagen. The
Committee does not agree with this finding and suggests instead that it is
possible that extensive testing for mutagenicity of chloroform may have
generated some false positive results with a nonmutagen. The conclusions
regarding mutagenicity of chloroform do not appear to reflect in full the
weight of the evidence and need further amplification, in part because new
and unproven test methods were applied to elicit some weakly positive re-
sults. Within the scientific community the current working definition of
a "mutagen" depends on test results obtained with standardized protocols.
While the evaporation of chloroform may result in negative results with
some procedures, all of the tests with which the chapter appears to have
interpretational difficulty also have negative results. These negative
results come from reliable methods, whereas the interpretation^! difficulty
comes from a bias towards positive results*
We request that OHEA reexandne the testing evidence from the point of
view that the hypothesis of false positive results needs to be refuted and
that the validity of the positive evidence needs to be investigated equally
with the negative.
A recent book by DeSerres and Ashby describes results fron thirty-
t L,R. POHL and B-A. MICO, Ilectrophillic Halogens as Potentially
Metabolites of Halogenated Compounds. Trends in Pharmacological Sd
-------
seven highly qualified laboratories which have conducted short-term testing
on chloroform. These findings are discounted la the document. Seven of
forty studies in this book were positive Cone of six DNA repair teats, two
of nineteen bacterial mutation tests, four of thirteen eukaryote tests,
and neither of the two whole animal tests). These tests were performed
with well-defined, standardized protocols,, and the results futher emphasize
the possibility of false positive results with a nonmutagen,
"Insensitivity" in the DeSerres and Ashby reference was defined with
respect to the detection of a potential carcinogen, not a mutagen. When
insensitivity is discussed in the draft health assessment document, however,
the reference is to detection of a potential mtagen. Where the chapter
on mutagenicity of chloroform describes the conclusions from DeSerres and
Aahby book, the definition of a false negative result (in relation to a
potential carcinogen) should be used.
The conclusion regarding the mutagenicity of chloroform has important
implications for the qualitative findings for carcinogen!city, if chloro-
form is a not a nitagen but rather a carcinogen, some scientists will
treat chloroform as a presumptive promoter, although some Committee members
question the adequacy of mutagenic testa to conclude that chloroform is a
"promoter"* The Committee requests that the final document further discuss
this issue in detail.
The Committee also advises that the use of qualifiers such as "weak"
and "strong" to describe mutagens is potentially misleading, because poten-
cy can be confused with either the qualitative weight of evidence for muta-
genic effects or the severity of mutatlonal effect. Use of the terminology
suggested for weight-of-the-evldence categories in the Agency's proposed
guidelines for mutagenicity risk assessment may partially alleviate this
problem. Perhaps potency can be described better through reference to
other agents or to the actual quantitative information.
Discussions of the papers by Direnzo and by Castro are Inconsistent
with each other. The Castro study had the same specific activity l*C-la-
belled chloroform as the Direnzo study. Therefore, specific activity of
the isotope cannot be an adequate basis to dismiss one study and accept
the other. In addition, the numerical results reported are in disagreement
between the two studies and require further comment.
Inconsistency also occurs in the discussion of microsomal preparations
in these two papers. Castro used mouse microsomes and found no DNA binding,
whereas' Direnzo used rat microsomes and found DM4 binding. The document
should carefully and consistently Implement the conclusions regarding Incor-
poration of chloroform into DMA in all chapters. The Castro study consists
of two papers, and the data from each paper needs to be Interpreted la the
light of the other. Castro found lipid and protein binding in one paper.
This binding shows that activation of chloroform took place with mouse
microaomal preparations.
CARCINOSENlCirg
The staff of OEM's Carcinogen Assessment Group distributed a revised
Carcinogenic!ty chapter at the meeting on which some of their oral comments
-------
-5-
were based. Much of the discussion below reflects Che oral presentation
aade by OHEA scientists at the Chlorinated Qrganles Subcommittee meeting.
Neither the Subcommittee nor the Committee has had the resources then (or
subsequently) to review the revised chapter in detail or to collate the
revised chapter with the oral presentation. Reference to the oral presen-
tation my be obtained through the transcript of the meeting, which is
available through EPA'a Committee Management Office (PM-213, loom 2515
Mall]. However, the Committee regards the changes suggested by staff
favorably and desires to see then implemented in the final version of this
chapter*
The Committee agrees with the qualitative conclusion in the document
that sufficient animal evidence of the carcinogenicity of chloroform exists
from replicated studies. The animal studies are hiatologically and bio-
logically sound, and they have been adequately interpreted. According to
IAKC criteria, this information will place chloroform into Category 2,
which is defined as "the chemical... probably is carcinogenic to humans.1*
However, as described above under "Major Comments," the Committee questions
whether the human evidence is inadequate or limited. Therefore, the Com-
mittee requests that OHEA's scientists reexamioe the epidemlological evi-
dence regarding carcinogenic effects of chlorinated drinking water, to
confirm or modify the conclusion regarding the weight-of-the-evidence for
humans in the light of the comments below.
The Committee believes that the terminology used by IARC is "inadequate*
evidence, not "insufficent* evidence. (See page 1-10 of the draft docu-
ment*} The criteria of inadequate versus limited evidence are very close
together, and the Committee appreciates the difficulty facing OHEA's scien-
tists in applying these definitions in the case of chloroform In drinking
water. However, the criterion of "limited" evidence is that ** a causal
explanation is credible." The comparison of the quantitative estimates of
risk based on animal versus human information suggests that the animal and
human data are not inconsistent, although the quantitative estimates are
highly uncertain. (See "Quantitative Estimates of Bisk" below,) The
Committee looks at the evidence as follows:
(1) The statistical power of any of of the individual studies is weak.
The ratios range from 1*1 to 2.0* The middle of this range, however,
represents a 50% Increase in the incidence of common human cancers, which
la substantial.
(2) The specificity of the cancers found in the different studies are
Identical; that Is, drinking chlorinated water is associated consistently
with cancers of the rectum, bladder and colon.
(3) The association between chloroform In drinking water and human can-
cers has biological plauslblity. The route of ingestion and excretion Is
rationally related to the organs with which cancer la associated. Be-
cause chloroform causes tumors in animals, it is reasonably suspect in
the human surveys.
-------
-6-
(4) Ho dose-response relationship was found.
(5) The different studies are consistent with each other. A positive
association with the same cancers was found In all studies.
While a cautious interpretation Is warranted, it will be worthwhile to
reexanine all of the data in the light of the IMC criteria. Technically,
the Coralttee suggests that OHEA try the method of residues. The data
on leukeoias could he extracted from the existing data, and the residues
reanalyzed for evidence of associations.
Once OHEA, has reinterpreted the epidemiologies! evidence for chloroform,
the Committee further request* some editing of the draft, regardless of
the qualitative conclusion. The analyses of individual studies of chlor-
inated drinking water each close with strongly worded summaries. These
summaries appear inconsistent with the overall summary of the human evi-
dence. The utility of the fifth paragraph on page 8-36 is questionable.
The logic fails in many analogies* For example» dioxins are minor con-
taminants of Agent Orange, but they are thought responsible for moat of
the toxic properties of this pesticide.
A new study of chloroform carcinogen!city in animals is forthcoming,
which was conducted by SSI International. A review of this study could be
added to the draft document or later placed into an addendum. The results
of this bloassay were not reviewed In detail by the Subcommittee; they
could affect the quantitative conclusions but are not likely to change the
qualitative findings. The evidence from this and other studies of chloro-
form tends to pinpoint some general issues about bloassays. These issues
reflect the state-of-the-art in toxicology, and it may be useful to the
Agency to comment on them in the context of the new study*
In the SH bioassay, no response was obtained when the C57B16 mouse
was given chloroform In drinking water, while chloroform administered in
corn oil yielded positive results. Thus* the risk of cancer in alee cannot
be related to chloroform alone in this assay. In contrast, approximately
the same results were obtained with both vehicles in the rat.
The SRI results raise two significant issues. First, complete absorp-
tion occurs with both vehicles. Absorption is faster when chloroform is
administered in drinking water, but the animals sip water throughout the
day. Therefore, the vehicle dependence of the findings can not be explain-
ed by incomplete absorption with water In ccaparison to corn oil. Second,
a bolus effect occurs with corn oil, since corn oil Is employed in gavage
studies. A pharmacoklnetic analysis may be required to explain the
effects of vehicle on the carclnogenlcity of chloroform for the mouse.
A discussion of the issue of liver damage in relation to carcinogen-
ic! ty will improve the chapter. The issue of liver damage relates to the
possibility of chloroform acting as a carcinogen through promoter effects.
OHEA scientists discussed how the experimental data support several pos-
sible mechanisms of carcinogenesis. While this information is conjectural,
the Committee believes it will improve the document to discuss the possible
mechanisms in the chapter. The issue also Is important in relation to
bolus effects.
-------
The Committee thinks that an Ideal study of earelnogenlcity will have
evidence for or against the presence of tumors at doses below a maximally
tolerated dose fox non-carcinogenic toxic effects. The difficulty with
this Ideal approach Is that dose-related toad.city studies do not often
yield a clear-cut indication of the maximally tolerated dose. It will be
useful to discuss this problem in the document within the context of the
available chloroform bloassays* since there is a growing awareness In the
scientific community of this problem with bioassays.
The question of a correlation between hepatotoxlcity and hepatic
carcinogenesia is of particular interest. The sane mechanism probably is
Involved in the early stages of both toxic effects* Therefore, a correla-
tion between the two endpolnts is expected. Perhaps it is more unusual
why chloroform is not more carcinogenic. The hepatotoxlc and hepatic car-
cinogenic effects of chloroform can not be totally separated, as the data
on page 8-19 show. After liver necrosis has occurred, if chloroform ad-
ministration is stopped, then tumor formation also stops* Emanuel Farber
has described a model of hepatic carcinogenesia through toxic effects on
different populations of liver cells. His model may be of use to OHEA.
While the Committee recommends the addition of competing theories of
chloroform-induced cancer to the discussion, these optional mechanisms
should relate specifically to chloroform. We recommend that "boilerplate"
language not required for the analysis of chloroform be removed from the
text.
TElATQGSNICm ABO FETOTQXICITY
Prior to its meeting, the Committee received a revised chapter on
teratogenleity and reproductive effects from the Reproductive Effects
Assessment Group. Even with the proposed revisions, the chapter on terato-
genleity and reproductive effects merits further changes.
There are a number of reasons for this recommendation. The evidence
for a fetotoxic effect of chloroform Is described but not evaluated in the
text. With the existing data for inhalation*! and ingestational routes
of administration, the Committee believes that it is possible to derive
no-observed-effeet-levels for teratogenic effects of chloroform but that
no evidence exists for either route of administration that levels have
been found at which fetotoxic effects do not occur* There is no conclusion
in the Executive Summary regarding fetotoxlc effects of chloroform. The
discussion in the revised chapter is focused only on teratogenic effects*
The Committee suggests that OHEA analyze the levels at which no effects
are observed for both of the endpoiuts, teratogenleity and fetotoxicity.
The Committee agrees with the statement in the Executive Summary that
chloroform is a potential inhalational teratogen. However, the discussion
in the chapter on teratology partially contradicts this finding. In ad-
dition, a number of technical errors In the chapter should be revised.
The Committee has supplied detailed comments on this subject directly to
OHEA.
-------
During the discussion with the Committee, OHEA scientists suggested
the possibility of conducting a risk assessment for teratogenic or feto—
toxic effects of chloroform* The Committee agrees that this subject Is
worthy of pursuit. Fetotoxic effects apparently occur at doses similar to
those that elicit earclnogejilclty but with much shorter exposures. The
Committee believes* however, that it Is not reasonable to extrapolate to
low doses sod expect a proportionally low Incidence of terata or lost con-
ceptuses. The concept of a threshold for teratogenic effects is based on
the capacity of the developing fetus to repair damage* A "one-hit" assump-
tion is not reasonable*
Scientists fron the Carcinogen Assessment Group also addressed the
problem of Scaling the absorbed dose between species* Sufficient data
exist regarding the allonetrle relationship for chloroform to use a cor-
rection factor to obtain human equivalent doses (not necessarily human
equivalent responses). The Committee believes that the approach taken
In these calculations Is appropriate, and that it responds to concerns the
Science Advisory Board has raised with the hazard assessments of other
substances*
A calculation also was orally presented that explored the question of
sensitivity of response to dose units. The Committee felt particularly
gratified to Inspect this work, since it responds to recommendations made
by the Environmental Health Coamittee in its review of previous health
assessment documents* The Committee found the approach taken in these
calculations to be scientifically reasonable. In addition, they tend to
make explicit the information needed to understand the significance of
the assumptions used in deriving the unit risk factor.
The Committee suggests that the term "metabolized dose" Instead of
"body burden" be used In the discussion, to prevent confusion with the
concept Of amounts of chemical accumulated on chronic administration. If
a correction for metabolized dose is made, then new potency estimates are
found which do not differ much from the previous, uncorrected estimate for
chloroform* This calculation shows that the potency estimate for chloro-
form is not sensitive to assumptions regarding metabolized dose* Because
data are available for only one dose for this calculation, the Committee
finds It Impossible to comment on the question of nonltnearity of response
with metabolized dose*
Carcinogen Assessment Group staff also compared the potency estimate
derived from animal data to the available epidemiology information, by
means of a "what-if* calculation. Given appropriate caveats, the Committee
believes that the display of the calculation Is a good way to illustrate
some of the uncertainties and consequences of the potency estimates for
decision-makers. Studies of the association between drinking water and
cancer Incidence yield a range of relative risks* Chloroform, thought to
arise from the chlorination of organic material Itt drinking,water, was
present In varying concentrations in these studies*
-------
-g-
Given some assumptions, the "what-if" calculation shows Chat the
uncertainty in the unit risk estimate might be of three to fours orders of
magnitude (1,000-10,000). These calculations aye the sane in principle
as estimates prepared "by the National Research Council's Committee on the
Biological Effects of Ionizing Radiation. The Committee also suggests
that it would be reasonable to illustrate an hypothesis derived by back-
calculating from the tf.S. incidence of bladder cancer, assuming that all
bladder cancer is due to the presence of chloroform in water* This cal-
culation could also be treated as a doubling dose calculation.
The terminology by which the unit risk estimate is implemented is
inconsistent in the draft document* For example, on page 8-79 the upper
bound nature of the estimate Is not stated accurately.
RANKING._OF..KEIAt_I_yE POTENCIES
OH1A has modified the table (pages 8-82 to 84) of relative potencies
from previous documents to Include a new column that lists IMC categories
of the substances under comparison* The accompanying histogram (page 8-81)
that illustrates the relative potencies of substances previously reviewed
by the Carcinogen Assessment Group has not been changed and does not in**
corporate IMC categories. The Committee does not believe that Insertion
of IAHC categories removes the concerns» as expressed in previous reviews
of other health assessment documents, regarding the potential confusion
between potency and severity. The problem is no different in principle
than the distinction made between potency and efficacy with pharmacological
agents. Describing only potency overemphasizes the lowest dose that might
have an effect.
If It is crucially important to retain the table, the Committee sug-
gests that QHEA also add columns for factors such as;
(1) the fraction of the maximally tolerated dose (or the dose of some
other Indicator of toxielty) at which a carcinogenic response is
seen,
(2) the percent of animals which had tumors at the dose at which maximum
tumor incidence was observed,
(3) the number of species in which tumors were observed, and
• (4) the degree of malignancy of the tumors observed.
RESEARCH STUDIES
The draft document calls for additional research studies in several
areas, particularly epidemiology* The Committee agrees that the proposals
probably will Improve our knowledge, but the proposals would benefit from
greater explanatory detail in the text. Are some of the projects key
studies? More extensive descriptions would have a value beyond the Agency's
immediate needs.
-------
-10-
IJTERATP1E
The Committee requests that all future health assessment documents
Identify the date at which the literature review was completed* Only
citations which substantially affect critical Issues should be introduced
beyond this date (perhaps in an addendum). Some of the references appear
to need an update. For example, some references on metabolism are cited
as "in press* or "submitted" in 1982. Further, the Committee requests
that future documents describe the general nature of the information cited,
such as peer reviewed articles, primary data from Industry-sponsored
toxicity studies, and so forth.
------- |