xvEPA
United States
Environmental Protection
Agency
Office of Health and
Environmental Assessment
Washington DC 2O46O
EPA/600/S-82/006FA
June 1987
External Review Draft
Research and Development
Addendum to the
Health Assessment
Document for
Trichloroethylene:
Updated
Carcinogenicity
Assessment for
TrichloroethyDene
Review
Draft
(Do Not
Cite or Quote)
NOTICE
This document is a preliminary draft. It has not been formally
released by EPA and should not at this stage be construed to
represent Agency policy. It is being circulated for comment on its
technical accuracy and policy implications.
-------
t
.1
*
-------
DRAFT ! EPA/600/8-82/006FA
DO NOT QUOTE OR CITE June 1987
Review Draft
ADDENDUM TO THE HEALTH ASSESSMENT DOCUMENT
FOR TRICHLOROETHYLENE
Updated Carcinogenlcity Assessment
for Trichloroethylene
NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally released by
the U.S. Environmental Protection Agency and should not at this stage be
construed to represent Agency policy. It is being circulated for comment
on its technical accuracy and policy implications.
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C.
-------
DISCLAIMER
This document is an external draft for review purposes only and does not
constitute Agency policy. Mention ofj trade names or commercial products does
i
not constitute endorsement or recommendation for use.
The Health Assessment Document for Trichloroethylene (July 1985;
EPA/600/8-82/006F) is available from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 70,3/487-4650
Order No.: PB85-249696
Cost : $ 30.95 (subject to change)
ii
-------
CONTENTS
Tab! es v
Figures . viii
Preface ix
Abstract ........... ..... x
Authors, Contributors, and Reviewers. ..... xi
1. EXECUTIVE SUMMARY. . . . . . . . ... . . .... . 1-1
2. BACKGROUND INFORMATION: 1985 HEALTH ASSESSMENT DOCUMENT FOR
TRICHLOROETHYLENE. . 2-1
2.1. USE AND EMISSIONS OF TCI 2-1
2.2. CANCER BIOASSAY AND RELATED DATA 2-2
2.3. METABOLISM AND PHARMACOKINETIC CONSIDERATIONS ...... 2-8
2.4. CANCER RISK DERIVATION 2-14
3. ONCOLOGY AND TOXICOLOGY 3-1
3.1. ANIMAL BIOASSAYS . 3-2
3.1.1. NTP (1987) Rat Study; Oral Exposure 3-2
3.1.2. Henschler (1984) Mouse Study; Oral Exposure .... 3-5
3.1.3. Herren-Freund (1986, 1987) Mouse Study; Oral Exposure . 3-8
3.1.4. Fukuda et al. (1983) Rat and Mouse Study;
Inhalation Exposure 3-10
3.1.5. Maltoni et al. (1986) Rat and Mouse Study; Oral and
Inhalation Exposure . ..... 3-12
3.1.5.1. Rat Study; Oral Exposure 3-12
3.1.5.2. Rat Study; Inhalation Exposure 3-14
3.1.5.3. Mouse Study; Inhalation Exposure. .... 3-22
3*2. EPIDEMIOLOGIC STUDIES 3-28
3.3. GENOTOXICITY STUDIES 3-32
3.4. DATA INTERPRETATION: DISCUSSION 3-33
4. ONCODYNAMICS 4-1
4.1. SELECTION OF EFFECTIVE DOSE 4-2
4.1.1. IDENTIFICATION OF THE ACTIVE AGENT(S) 4-3
4.1.2. DOSE OF THE ACTIVE AGENT(S) 4-8
4.2. ESTIMATION OF EFFECTIVE DOSE IN HUMANS 4-9
4.3. ESTIMATION OF EFFECTIVE DOSE FROM ANIMAL DATA 4-13
4.4. ESTIMATION OF HUMAN EQUIVALENT DOSE FROM ANIMAL DATA .... 4-23
4.5. SELECTION OF ANIMAL DATA SETS 4-25
4.6. EXTRAPOLATING AND SCALING ..... 4-27
m
-------
CONTENTS | (continued)
4.6.1. Route-to-Route . 4-29
4.6.2. Intra-Species 4-30
4.6.3. Shorter to Longer Duration of Exposure 4-32
5. QUANTITATIVE EVALUATION: UNIT RISK DERIVATION 5-1
5.1. DATA USED FOR THE DOSE-RESPONSE CALCULATION ....... 5-1
5.2. SLOPE (POTENCY) CALCULATION ; 5-6
5.3. RISK ASSOCIATED WITH 1 yg/m3 OF TCI IN AIR . . . . . . . 5-10
5.4. DISCUSSION 5-13
5.5. SUMMARY 5-14
6. SUMMARY AND CONCLUSIONS. 6-1
6.1. QUALITATIVE 6-1
6.1.1. Background ... 6-1
6.1.2. Weight of Evidence: j Likelihood of Human
Carcinogenic Potential . . ... 6-2
6.2. QUANTITATIVE 6-8
6.2.1. Estimate of Carcinogenic Potency (Unit Risk) .... 6-8
6.2.2. Unit Risk Characterisation 6-10
7. REFERENCES \ . 7-1
-------
TABLES
2-1 TCI carcinogenicity bioassays in animals. .......... 2-3
3-1 Percent survival at termination and mean midstudy body
weights of four strains of rats treated by gavage with TCI. . 3-4
3-2 Incidences of renal cortical lesions in AC I, August,
Marshall, Osborne-Mendel, and F344/N rats in the 2-year
gavage studies of TCI 3-6
3-3 Tumors and preneoplastic liver nodules from exposure to
TCI with and without stabilizers to ICR/Ha Swiss mice .... 3-9
3-4 Liver cancer response in B6C3F1 mice following exposure
to TCI and metabolites in the drinking water for 61 weeks . . 3-11
3-5 Incidence of lung tumors in female ICR mice exposed to
airborne concentrations of TCI 7 hours/day, 5 days/week
for 107 weeks 3-13
3-6 Incidence of renal meganucleocytosis in male rats treated
with TCI 4 to 5 days/week by gavage for 52 weeks 3-15
3-7 Incidence of leukemia and immunoblastic lymphosarcomas in
male rats treated with TCI 4 to 5 days/week by gavage for
52 weeks 3-15
3-8 Inhalation treatment regimen for rats 3-17
3-9 Incidence (%) of tumors among Sprague-Dawley rats exposed
to airborne concentrations of TCI 5 days/week for 8 weeks . . 3-19
3-10 Incidence of renal meganucleocytosis in male Sprague-Dawley
rats exposed to airborne concentrations of TCI 5 days/week
for 108 weeks 3-19
3-11 Incidence (%) of tumors among 90 Sprague-Dawley rats of
each sex exposed to airborne concentrations of TCI
7 hours/day, 5 days/week for 104 weeks (except controls). . . 3-20
3-12 Incidence (%) of tumors among 130 Sprague-Dawley rats
of each sex exposed to airborne concentrations of TCI
7 hours/day, 5 days/week for 104 weeks (except controls). . . 3-21
3-13 Dosing regimen for inhalation exposure of mice 3-23
3-14 Incidence (%) of tumors among Swiss mice exposed to airborne
concentrations of TCI 7 hours/day, 5 days/week for 8 weeks . . 3-24
-------
3-15
3-16
3-17
3-18
3-19
3-20
3-21
4-1
4-2
4-3
4-4
TABLES
Incidence (%) of tumors
concentrations of TCI 7
Incidence (%) of tumors
airborne concentrations of
for 78 weeks
(continued)
among Swiss mice exposed to airborne
hours/day, 5 days/week for 78 weeks ,
among B6C3F1 (NCI) mice exposed to
TCI 7 hours/day, 5 days/week
Incidence (%) of tumors amohg B6C3F1 (CRL) mice exposed
to airborne concentrations ,of TCI 7 hours/day, 5 days/week
for 78 weeks I
i
Analysis of tumors among Sprague-Dawley rats exposed to
airborne concentrations of TCI 7 hours/day, 5 days/week
from the studies (BT304, BT304bis) of Maltoni et al* (1986)
Analysis of statistically significant tumor incidence
among Swiss mice exposed byi Maltoni et al. (1986) to
airborne concentrations of jTCI 7 hours/day, 5 days/week
for 78 weeks. j . .
Analysis of statistically significant tumor incidence
among B6C3F1 mice exposed by Maltoni et al. (1986) to
airborne concentrations of TCI 7 hours/day, 5 days/week
for 78 weeks i .
t
Analysis of the incidence of lung tumors in female ICR
mice exposed by Fukuda et al. (1983) to airborne concen-
trations of TCI 7 hours/dayi, 5 days/week for 107 weeks.
Estimation of the TTCIM (mgj) formed at rest and at work in
men exposed to airborne TCI' concentrations (140 ppm - 750
mg/m3 and 70 ppm - 376 mg/m&) recalculated from Monster et
al. (1976a)
Summary of data used in thejanalysis of enzyme kinetics
in the rat
Summary of data used in the analysis of enzyme kinetics
in the mouse . .
i
Summary of estimated daily dose in terms of TTCIM at the
airborne concentrations and(durations of daily exposure
used in the bioassay with Sprague-Dawley rats (Maltoni
et al., 1986) and the resulting tumor incidence, plus
the HED | . . .
3-26
3-27
3-29
3-35
3-38
3-39
3-41
4-11
4-15
4-16
4-26
VI
-------
TABLES (continued)
4-5 Summary of estimated daily dose in terms of TTCIM at the
airborne concentrations and durations of daily exposure used
in the bioassay with Swiss mice and female B6C3F1 mice
(Maltoni et al., 1986) and with female ICR mice (Fukuda
et al., 1983), and the resulting tumor incidence, plus
: the HED 4-28
4-6 Summary of estimated daily dose in terms of TTCIM at the
applied doses,(oral-mg/kg [NCI, 1976] and inhalation-ppm
[Maltoni et al., 1986]) in the bioassays with female B6C3F1
mice, and the resulting tumor incidence, plus the HED .... 4-31
4-7 Summary of estimated daily dose in terms of TTCIM at the
airborne concentrations and durations of daily exposure used
in the bioassays with Swiss male mice (Maltoni et al., 1986)
and with Sprague-Dawley male rats (Fukuda et al., 1983), and
the resulting tumor incidence, plus the HED . ... . . . . . 4-33
4-8 Summary of estimated daily dose in terms of TTCIM at the
airborne concentrations and durations of daily exposure used
in the bioassays with Swiss mice and female B6C3F1 mice
(Maltoni et al., 1986) and with female ICR mice (Fukuda
et al., 1983), and the resulting tumor incidence plus
the HED and the human lifetime-adjusted HED 4-36
5-1 Tumor reponses and the corresponding metabolized doses
from the Maltoni et al. (1986) study in rats 5-2
5-2 Tumor reponses and the corresponding metabolized doses
from the Maltoni et al. (1986) study in Swiss mice 5-3
5-3 Tumor reponses and the corresponding metabolized doses
from the Maltoni et al. (1986) study in B6C3F1 mice 5-4
5-4 Tumor reponses and the corresponding metabolized doses
from the Fukuda et al. (1983) study in ICR mice 5-5
5-5 Slope estimates per (mg metabolized dose/kg/day)
calculated on the basis of different data sets and
under different dose-equivalence assumptions 5-7
5-6 Potency slope per (mg metabolized dose/kg/day) derived
from different animal species (rats and mice) and under
two different dose-equivalence assumptions . . . 5-9
5-7 Summary of unit risk estimates derived from different animal
species and under different dose-equivalence assumptions. . . 5-12
6-1 Summary of the relationship between sufficient weight-of-
evidence factors and bioassay results 6-7
-------
FIGURES
4-1 The metabolic pathways of TCI 4-7
4-2 Relationship between a single human (at rest) 8-hour
exposure to airborne concentrations (mg/m3) of TCI and the
estimated amount of total TCI metabolized (TTCIM mg). . . . 4-12
4-3 Relationship in the rat between reciprocal doses of TCI
(po mg/kg: inhalation mg eq/kg) and the reciprocal of the
amount of the total TCI metabolized (TTCIM mg eq) . . . . . 4-17
i
4-4 Relationship in the rat of the doses doses of TCI (po mg/
kg: inhalation mg eq/kg) and the amount of the total TCI
metabolized (TTCIM mg eq) as predicted by the regression
in Figure 4-3 \ 4-18
4-5 Relationship in the mouse of! the doses of TCI (po mg/kg:
inhalation mg eq/kg) and the amount of total
TCI metabolized (TTCIM mg eq) 4-19
4-6 Relationship in the rat between airborne concentrations
of TCI and total amount of TjCI metabolized (TTCIM mg eq)
obtained by computer simulation 4-21
4-7 Relationship in the mouse between airborne concentrations
of TCI and total amount of TCI metabolized (TTCIM mg eq)
obtained by computer simulation ... 4-22
4-8 Log plot (shown as a line) of experimental rat data at
10 and 600 ppm and simulated TTCIM (mg eq) (shown as
points) ......... 4-24
viii
-------
PREFACE
The Office of Health and Environmental Assessment has prepared this
addendum to serve as a source document for EPA use. The addendum updates
EPA's July 1985 Health Assessment Document (HAD) for Trichloroethylene (TCI)
by providing a review and interpretive analysis of cancer bioassays that have
become available since the completion of the HAD. Similarly, recent literature
on metabolism, pharmacokinetics, and short-term tests are also reviewed. The
addendum discusses the impact of the newer data upon the weight of evidence
for the carcinogenicity of TCI, and uses relevant data to develop a revised
unit risk estimate for inhalation exposure to TCI.
The literature search suppporting this document is current to December
1986; however, several documents published since this date have been included
in the interest of completeness.
-------
ABSTRACT
This addendum updates the Health Assessment Document (HAD) for Trichloro-
ethylene (TCI) published in July 1986.: The addendum focuses on the inhalation
and oral carcinogenicity bioassays that have become available since the 1985
HAD was completed. The 1985 HAD concluded that the available carcinogenicity
i
evidence for TCI was, according to EPAj's Guidelines for Carcinogen Risk Assess-
t
ment, inadequate in humans, and sufficient in animals, thus equivalent to a
weight-of-evidence classification of Group B2. This conclusion was based pri-
marily on mouse liver tumor responses, with support from metabolism data and
i
other positive but compromised bioassaV studies. This addendum identifies
positive findings in rats, a second species, and positive findings by inhala-
tion exposure in rats and mice with both similar and different tumor sites
compared to the oral studies. A quantitative analysis of the recently reported
inhalation studies shows that the uppe'r-bound estimate of risk is very close to
the earlier inhalation estimate in the| 1985 HAD derived from the oral studies.
[
The recommended upper-limit incrementafl unit risk for humans exposed for a
70-year lifetime to a 1 yg/m3 airborne! concentration of TCI is 1.7 x 10"6. In
i
comparison with the potency of 58 other compounds evaluated by the Carcinogen
Assessment Group, TCI ranks in the lowest quartile. Using EPA's classification
i
criteria, the weight-of-evidence for TJSI is clearly equivalent to a Group B2
classification, meaning that TCI shoulfl be considered a probable human carcino-
gen.
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
This document was prepared by the following members of the Carcinogen
Assessment Group, Office of Health and Environmental Assessment.
AUTHORS
Robert P. Beliles, Ph.D., D.A.B.T.
Chao W. Chen, Ph.D.
Herman J. Gibb, B.S., M.P.H.
Jean C. Parker, Ph.D.
Charles H. Ris, M.S.S.E.
Chapters 2, 3,. 4, and 6
Chapters 5 and 6
Chapters 3 and 6
Chapters 2, 3, 4, and 6
Chapter 6
CONTRIBUTORS
I.W.F. Davidson, Ph.D.
The Bowman Gray School of Medicine
Wake Forest University
Winston -Sal em, NC
Chapter 2
REVIEWERS
The* following individuals reviewed earlier drafts of this document and
provided valuable comments,
Karl P. Baetcke
Office of Toxic Substances
Office of Pesticides and Toxic Substances
Jerry N. Blancato
Exposure Assessment Group
Office of Health and Environmental Assessment
(Chairman of the Hazard/Risk Assessment Committee of the
Integrated Chlorinated Solvents Project)
Murray Cohn
U.S. Consumer Product Safety Commission
(Member of the Hazard/Risk Assessment Committee of the
Integrated Chlorinated Solvents Project)
xi
-------
Charalingayya B. Hiremath
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Lorenz Rhomberg
Carcinogen Assessment Group J
Office of Health and Environmental Assessment
(Member of the Hazard/Risk Assessment Committee of the
Integrated Chlorinated Solvents Project)
XI1
-------
1. EXECUTIVE SUMMARY
This addendum updates the Health Assessment Document (HAD) for Trichloro-
ethylene (TCI) published in July 1986. The addendum focuses on the inhalation
and oral carcinogenicity bioassays that have become available since the 1985
HAD was completed. The recently reported inhalation bioassays are the first
studies that are adequate for quantitative risk analysis. In addition, recent-
ly published studies concerning TCI metabolites also warrant evaluation.
The 1985 HAD concluded that the available carcinogenicity evidence for TCI
was, according to EPA's Guidelines for Carcinogen Risk Assessment, inadequate
in humans, and sufficient in animals, thus equivalent to a weight-of-evidence
classification of Group B2. This conclusion was based primarily on mouse liver
tumor responses, with support from metabolism data and other positive but com-
promised bioassay studies.
This addendum identifies positive findings in rats, a second species, and
positive findings by inhalation exposure in rats and mice with both similar
and different tumor sites compared to the oral studies. Current literature
further supports the hypothesis that TCI metabolites are the active carcino-
genic agents. Considering both the earlier and more recent animal studies,
there is a "sufficient" level of evidence that TCI is a carcinogen in animal
test systems. While a new epidemiologic study was recently published, the
available epidemilogic data remain inadequate to refute or demonstrate a human
carcinogenic potential, as was the case in the 1985 HAD.
Overall, using EPA's classification criteria, the weight-of-evidence for
TCI is clearly equivalent to a Group B2 classification. Group B2 means that
TCI should be considered a probable human carcinogen.
1-1
-------
A quantitative analysis of the recently reported inhalation studies shows
t
that the upper-bound estimate of risk [is very close to the earlier inhalation
i
estimate in the 1985 HAD derived from the oral studies. The recommended upper-
limit incremental unit risk for humans exposed for a 70-year lifetime to a
1 yg/m3 airborne concentration of TCI :is 1.7 x 10~6. In comparison with the
,
potency of 58 other compounds evaluated by the Carcinogen Assessment Group,
TCI ranks in the lowest quartile.
1-2
-------
2. BACKGROUND INFORMATION: 1985 HEALTH,ASSESSMENT DOCUMENT
FOR TRICHLOROETHYLENE (TCI)
This chapter summarizes the carcinogenicity and related sections of the
1985 Health Assessment Document (HAD) for Trichloroethylene (U.S. EPA, 1985a).
Several topics have been expanded in this summary to set the stage for the
updating and reanalysis that are covered in later chapters. Some specific
references and related data have been provided for clarification and emphasis.
In addition, some publication dates reported in the 1985 HAD been updated to
reflect the final publications.
2.1. USE AND EMISSIONS OF TCI
Trichloroethylene (1,1,2-trichloroethylene) (TCI) is a multimedia environ-
mental pollutant. It is a highly volatile solvent widely used in the indus-
trial degreasing of metals. It has no known natural sources. Current U.S.
production, as identified in the 1985 HAD, was estimated at about 130,000
metric tons per year. Of the TCI used in the United States, 80% to 95% is used
in the degreasing of fabricated metals, with the remaining 5% to 20% divided
between exports and miscellaneous applications. In addition to the workplace,
TCI has been detected in a variety of urban and nonurban areas of the United
Stat§§ gftd other regions of the world. It has been measured in ambient air4 artd
water. An average ambient air concentration of about 1 part per billion (ppb)
Wduld be expected for some large urban centers. Concentrations as high as 32
ppb have been measured in urban centers in the United States, and 47 ppb has
been measured in urban Tokyo. Ambient air concentrations of TCI are greatly
influenced by the rate and geographic distribution of emissions and the rate
of decomposition and transposition in the atmosphere. The average mixing ratio
in the troposphere of the northern hemisphere is 11 to 17 parts per trillion
2-1
-------
(ppt). Reaction with hydroxyl radicals is the principal mechanism by which TCI
is scavenged from the atmosphere.
2.2. CANCER BIOASSAY AND RELATED DATA
The evidence reviewed in the 1985 HAD for the carcinogenicity of TCI in
experimental animals includes: increased incidence of hepatocellular carcin-
[
omas in male and female B6C3F1 mice (tjhree studies) by gavage; malignant lym-
t
phomas in female HantNMRI mice by inhalation; and renal adenocarcinomas in male
i
Fischer 344 rats by gavage. Table 2-1 presents an overview of the bioassays
reviewed in the HAD, as well as the recent bioassays reviewed in this addendum.
Statistically significant increases of malignant liver tumors were ob-
served in both male and female B6C3Fljmice. A gavage study of purified TCI in
both sexes of B6C3F1 mice was carried [out with a design similar to that for the
gavage study of technical grade TCI in male and female B6C3F1 mice to evaluate
i
the role of epoxide stabilizers in TCI in the induction of hepatocellular
'
carcinomas. Similar carcinogenic responses were observed for the purified
epoxide-free and for stabilized TCI. JThe other studies provide some additional
i
support to the body of evidence, particularly since two of them were carried
out in a different species or strain.! However, a third study in B6C3F1 mice,
art inhalation study in which an increase in liver tumors was seen, was weakened
by deficiencies in its conduct. In the inhalation study in Han:NMRI mice, a
30% incidence of spontaneous lymphomah'n control mice, and the possibility of
immunosuppression having a role affecting the carcinogenic response in treated
|
mice, made these results difficult to interpret. A small incidence of renal
adenocarcinomas in high-dose males at terminal sacrifice was observed in the
i
gavage study of Fischer 344 rats with[purified TCI. The incidence was statis-
tically significant by life table analysis. A gavage study with male and
i
female Osborne-Mendel rats was negative; however, this study may be considered
2-2
-------
00
<:
:e
* i
'^L
^
z:
i t
oo
CO
i 3
o
CO t-
co
O
Q
I/)
CU
r-
O
CU
; CL
OO
re
o
CO -r-
x: t-
O 3
CL
t i
O
h-
>*
-O
3
4->
OO
et
c_ c
re re
r
3 in
i CU
i r
eu re
0 E
0
P C
re T-
O.
CU CO
x: re
E
<4- O CO
o c cu
i i o re
» i c« E
oo re cu
oo o 14-
>,
"O re
*>» "O
co*^»
j>£ co
~"^. .si
E"O>"S
cnEco
co en i
CO 00
CM r--. «
i-H CU
* CO
en * re
t en >
« 1 10 re
r-i oo en
cu
o
r ~O
E c
re ** ~^
t 1 CO
U_ CO i
oo r re
cj re E
VO E CO
CO 14-
s*
en
CO T *
-o
to o
t_
en co
c
r a* re
reen -P
oen 3
c A|>>
sz x
o> < o
COO Q.
1 1 CO
^>O
j
en
r-l
A
1 1
O
z:
CO
r-
re
en
^
en
_si
Si
CO
1^- CO
en
O "
i t CU
en
r, «
en >
^i" ^3
LO en
r- -o
"cu c
o re
c
cu co
S. i
i re
co s- -
c co
o in re
-Q -u e
10 re co
o t_ ) ^
-o S
CO
^.rH
IP CO
CO
o re
o >
o re
« i en
CO
o
f" "O
E c
re ** ^
« 1 CU
U_ CU i
co i re
o re E
Co E cu
CO M-
T3
cu
'j
^
t
3
O_
' «*
CM
CO
en
r-4
re
oTco
i en
z: i i
i)_
O m
re
cu E *
O O co
c c cu
CO !- t
-o o re
r- t- E
o re
c o cu
i- O CO
c o
T3 CO T3
CO T3 1
(& C7)
CO i -r-
c_ re x:
o c
c cu c
i i t_ .|
£
-a
**»^
en
-N^ >!f*
"5^ 2
ECO
o
Oft
O
0 «
t 1 CO
CO
ffi re
o >
o re
LO en
^* **~^
*f cu cu
CO i r
re re
f- e e
CO CO
O CO
m +J TD
r- re c
u. (- re
<4_
0
CO
u in
c eu
cu
HO (O
r- £
0
c c
-o re
£'i
re eu
CO 3
i CO
UJ i
_L?"s
CO CM
E LO
O >
LO CO
CM en
re
o re
LO CO
>>-o
co c
i re
re cu
Q r
i re
CO E
3- CO
CO r-
re in re
«- 4-> E
cu re co
00 t_ <4-
"0
CO
<*-
r
£_
3
a.
10
oo
en
r-l
M
r*
er
O
re
IE
1
] ^
M -0
c re 10 co
r- t_ +J C
10 co re -i-
co t- N >
CO O T- CO CO
re E i i co
CO 3 -i- CO T-
C- +-> JQ O -t-J
u re co c o
c JK +j e- cu 3
r O CO O T- "O
re E o c
-o E o 3 -i- o
CO O 4-> +J <»- o
> -P CO
c_ en -o t- -a >,
cu eu eu co -a
CO C- +J > 4-> 3
.a o 3 M- 3 +j
O M- J3 1 XI CO
-* IB
"X - * 00
o oo f-
COLO O CO O
^t en c >
*^x: t- cu
en -|j « o M co-
E «r en 4 o re
S c >
o ^4 co j*: 5 co
^1" S O ~--
CM -O COO >
o E r-i en
««* o c
o c o x: T-
O t- O 4J CO
OO O 4- CM M- O
«-H >4- O r-4 S -O
-o
c
co re
CO
i- CO
5l r**"*
oo re
re co
~~. cu "re
Oi O E
O !- CO
T3
c -a
re cu
Nl
-a r~
CO i
*j o
r- re
i- -t->
3 CO
a.
V
re
| ^
CO
f
co
^
o
CO *
c oo
cu en
Z r-l
i
c
r- <4-
o
CU CU
m o m
3 c re
re cu E
o ~a o
r- C
0 t- -r-
1 0 r-
J^ CO
r-i en
to E
> 0
l_ 0
eu coo
re ~~-
co «c
ceo o
r- O
c *»
r i I ^f
£. O O
-01 I
CO
o
E
U- CU
CO r
10 E
CO
T3
o <-
re o
re
0
r *r- O
re -P T^
r- CO 4J
t- o ai
-p re o
in o re
CO 3 t_ O
t- -a o «-
3 C r O
CL-r- x; i
o x:
* i i i *i u
O CJ t- !-
oo
en
"O «
300
cu en
£_ rH
1 1
1 »
c
CO i
c_ re
CO 4->
3C CU
CO
Cn
re
CL
CO
c
o
_
o
CO
rr
P
o
T5
CO
C
+0
c
o
o
2-3
-------
*"**
o
CD
3
C
r
4J
C
O
*«^
m
r-4
CM
UJ
Results
^
CO
CD
> Q)
CD 4-»
O
CD £-
O
O
C/)
CD
P
O
CD
CX
in
l_
CO
o
'i 4^*
CD *P"
"5 3
C3u
»-H
o
r-
f
4J
*
o.
I
£
o
(*t
to
40
a.
a>
o
u
CO
4J
O
i,^^
CO
13
en
*^^
rj
E
O
0
o
i-H
cT
o
p-"
a>
13
c
CD
^~
QJ
C
(_
O
to
CO
-a
a>
r
r"
C_
3
a.
i
oo
^
(
|
r-
O t_
C CO *
-5^
13 J3 O
CD 3
CO 4-> C
CO »r-
OJ i
C- CO IO
O C CO
C CD E
i- t- O
C
13 c+_ O)
O) O 13
> CO
t- a)
ni cii
o -a o
^
<*
A
a>
en
CO
en
* fv
co t***
* o
ur> co
CO
i CM
CO 1 ^
je to
to
c_
<4- O
0 E
CD 4->
U
CD p^
0 0)
i- O
o
c: i
t- tO
13 4J
CD -P-
increas
interst
in MAR.
13
CO
CD
CO ^""**
E CD
10 tO 3 CO
4j c_ cni < E
CO CO 3 O CD
£-3=1 CO T3
O CX r
c CD c: i
i
« i
E ^5 f^
CX E 3 C
ex to o
CO CvJ ^3 4-5
O CD tO
IO « > CD
C V- 3
> o cu cr
CD .r- O
o 4-> a> o
CO CO t. 4J
1*
o .c to o
o c o
CD
o
i- -a
E c
CO *-""*
. < CD
u_ a) r
CO i CO
CO *^*^ **-~
a
t_ ir>
en «-H
g
r *% O O
CO CT> C.
O CTl <3 C
C A E tl
f o XJ
U I-H T- >,
<1) O i~» f
h- i CD
*
i CO
co r-.
cr>
a>
O) 3C
CO
a>
CO
en
CD
E o
ex E
a.
O C\J
0
vo .
c
0°
0 4->
CO CO
!_
« CO
O J=
C3 C
r- 1 -p-
c_
CD
r~" OJ
Oi r
to E
CD^_-
c- to
CO 4->
si to
0 t-
O)
CO
£
(D
CD
CD cr
in ce co
co c: E
CD en CD
t_ -r- M-
O
c co c:
*^
^
co
[^
E
Q. >
a. c:
o
O 'P-
0 4->
in co
p
> CO
O J=
0 C
r 1 -r-
O)
O
1
13
l < C
cni to - »
y CD
Z CD P
i CO
f= co E
CO E CD
3= > ,
-C 4->
p-
CO4^
E -O
c: 4J
o a.
i~* OJ
to o
to to
O) 3
C- tO
a.
3 -a
immunos
enhance
CD o>
CO CO
en en
CD CD
z. z=
r>^ ^s<
!£ 3
CO 00
r«. t-~
E E
cx " a. «
a. c: cx c
0 0
O -r- O -r-
O 4-> O 4->
LO CO LO CO
p P
« CO > CO
0 j= 0 JC
o c o c
I 4 «p- rH »r
to
to CD
4 ' 4~*
CO tO
£_ E
13 CO 13
4J C J= C
to CO x-"x co ""**
p- CD C O)
3 CD i CO CD i
p CO *P P-~ CO
cz co E t- co E
ce E a) >, E a>
^
CD
en
CO
cx
en
<=
§
P
o
D
5
13
CD
P
O
O
vv'
2-4
-------
''
^ X
^
Ol
3
C
C
o
u
t 1 ;
1
LU
_J
CQ
<£.
r-
CO
^>
r
3
CO
QJ
o:
A
CO
r-
O)
> OJ
O) 4->
i 3
0
QJ t-
o
Q
*
E -P
O) M-
-C t_
O 3
CL
O
1
>~
T3
3
00
.
CO
P <*-
re o QJ
c_ (
QJ re -p
O) o E co
i C QJ QJ
re o) <»- j=
E -o en
QJ - C -r-
<4- O -i- .C
c~
C -i- CO O
r- L jS
-O O -P
0) 0) E
> CO 3 +J
r re -p re
I * QJ t/J
re t- en oj QJ
en o c: o co
QJ C 3 -i- O
"z. I-H i E *o
-^
E
Q.O
i-H
O
Cf) *
«tf- C
O
f"^ [ *
LO re
t-H 1
re
« jc
O c:
Lf) -r-
>k
OJ- ^
QJ
^ r
(T3 (13
Q e
1 QJ E fO
ro to E
C. 4J fV ^J
CL (C O *i-
,J
^_>
OJ
re
3 OO
^ oo
3 cr>
Ll- T-H
i
3
QJ C
, i r -O !-
re c
it- c: re co
O QJ I-
t_ « o
T3 CO E
c: 4- re 3
Q) 0 E -P
t. 0
.p I i C i
1 -I r
T3 OO O OJ
QJ OO I- O
co re
re > o en
QJ re o -r- co
t-'-r- C T3 QJ
O E QJ >>
c QJ "o oj re
HH ^ re _J E
E 5
Q.
Q *d"
"o
O rH
O
IO >
C
., o
O !-
O ^->
oo re .
r»-
«t n3
o s=.
o c
I 1 !-
>j "O
QJ C
5
re QJ
f-*v ^
I (O
a) e-~
3 ' 0)
O> ^~
re co re
C. +J E
Q. re QJ
00 t- t«-
-o
C TJ
re Q)
Nl
T3 T-
0) ^-
r *^
z: r-»
T3 O) -O
C f <4- C
re QJ M re o re
i 0 E QJ
co re QJ i < o) o
re E "-H M I-H p !
E I-H oo re E
o c oo c: oo E
-p - OO -r- il
re » c u_
CL co > co QJ T- OO
OJ t_ OJ £_ O O
-C O O O i- CO VO
E -r- E E re oa
en
QJ OJ
-o <
-o QJ
t i .p
oo * i re
Lf) i 1 O P "O
/-> ^j
" « t- O CVJ
J>i jx; o re lf>
S 5 XC *!j"
~~. *~- CL QJ
X X 4J >
oo i oo en re r
re 3. -p re
> U > CO O
E CL E in tl a.
0 . >, 0
QJ
0
E OJ QJ
re r r
co 3: re re
co --~ E E
r^ Qi QJ QJ
» O »-
00 I-H ^ -
"O
OJ
1
(4
r
f
3
ca-
r-
re
OJ
c oo
QJ 00
t_ en
3 --H
3
Q >
O1*
C t--
re 01
> i i
a?
en
re
CL
cn
c
U
o
r-~
O
4-
O)
JC
c:
o
T3
QJ
3
C
C
O
O
2-5
-------
o
cu
3
C
O
1
UJ
CO
cu
CO 4->
^ ;3
o
cu t-
t/>
0
n
in
cu
r
O
cu
CL
to
75
o
>r« ^.^
F; 4 J
CU -r-
x: t_
O 3
CL
* i
O
1 I
^
3
CO
CU CU i
> > cu cu cu
"re re 4j! '43 '£!
CT) o> re? re re
CU CU CJ} U) C7^
-a «
cu t- cu -a
CM > CU « "O O T3
CM O) 4J i -r 4 -t (~-
10^^ -orere x x«*
5 -i- 4J U Oi O UO
« "O xoim-i- a_re CL
\x - I .)-> 1 >
> > O r O) ^
CT) ^71 A) H~" * O 4-^ 5^ * ^^ 5
E Ez C3^ ^^J f^ fQ "T3 £ sx Q |S
ro O> 5 £- 4J ^ LO > E ^ O OJ CM Lf) --s.*O LO O
»rtJ X.COLO »XCVi «C
o o en r 4 co CL co o o
_ -g _ _ _
cu re * - cu cu cu
I CU r r r
re cu ^~ re r re re
E f re E E E
cu re e OM cu cu
4 E CU 4 4 M
^ 3
3 --1 «=
3 ^>
c: r o
re cr> o
cu
o
cr
cu
-o
1
o
c
T3
cu
re
cu
o
c
r
^
C
re
o
<*-
r-
C
OT
^
"re
0
'5
r
re
in
II
H- H
I-H
CO
re
2-6
-------
inconclusive because of high mortality in the treatment groups. Exposure to
airborne concentrations of TCI did not result in a carcinogenic effect in
Han:Wist rats and Syrian hamsters, although higher dose levels of TCI probably
could have been given to the test animals. Several other animal studies did
not suggest a carcinogenic potential for TCI but are inadequate for refuting a
carcinogenic potential for TCI for one or more reasons. These studies were not
specifically designed to evaluate carcinogenicity, involved only small treat-
ment groups, lacked adequate doses of TCI, or were of insufficient duration
because of the expected latency period for cancer induction.
TCI induced malignant tumors of the liver in both male and female B6C3F1
mice in multiple oral studies. This constitutes evidence that TCI might be
carcinogenic in humans. The kidney tumors observed in the NTP F344 rat study
(NTP, 1986b) were not considered to be a strong indication of a response in
a second species because of the small number of animals responding (3 of 49
animals) and because of the high mortality, although the statistical signifi-
cance after mortality corrections suggests that a carcinogenic effect may be
taking place.
The 1985 HAD for TCI reviewed three cohort studies of workers exposed to
TCI, one malignant lymphoma case-control study which evaluated exposure to TCI,
and two surveys of liver cancer cases for TCI exposure. These studies suffered
from one or more of the following deficiencies which limit their usefulness
and/or diminish the sensitivity to detect a human carcinogenic response: small
sample size, lack of analysis by tumor site, problems with exposure definition,
and problems with length of exposure. Overall, the epidemiologic data were
judged to be inadequate for evaluating the carcinogenic potential of TCI.
The U.S. Environmental Protection Agency's Guidelines for Carcinogen
Assessment (1984 proposed and 1986 final) take the position that the mouse-
2-7
-------
liver-tumor-only response, when other
conditions for a classification of "suf-
ficient" evidence in animal studies are met, should be considered as "suffici-
ent" evidence of carcinogenicity. According to the guidelines, a variety of
i
factors could downgrade this evidencejfrom a "sufficient" to "limited" category.
i
However, the downgrading does not occur with the TCI data base. For example, a
statistically significant increase of[predominantly malignant tumors occurred
in both sexes of mice, metastatis to the lung was noted, and other tumor end
points, although marginal, were observed in different studies. Additional sup-
porting considerations include mutagenicity, cell transformation, and metabolism
data. Cell transformation activity by TCI in Fischer rat embryo was found, but
it is not certain whether the effect occurred in the host genetic material.
Currently available data provide suggestive evidence that commercial grade TCI
is weakly active as a mutagen requiring metabolic activation. The data avail-
I
able for pure TCI do not allow a conclusion to be drawn about its mutagenic
potential. There is evidence that TC[l metabolites can react with cellular
protein macromolecules and with exogenous DNA and RNA in vitro. There are no
known differences among mammalian species, including humans, with regard to
metabolic pathways or profiles for TCI. As reported in the 1985 HAD, the cumu-
i
latiVe evidence for TCI resulted in a| classification of B2, i.e., a probable
human carcinogen.
i
2.3. METABOLISM AND PHARMACOKINETIC [CONSIDERATIONS
I
The pharmacokinetics and metabolism of TCI have been studied in man as
' ' 1
well as in experimental animals. InHalation of the vapor (1 ppm = 5.38 mg/m°
at STP) is a route of concern by which TCI enters the body. Ingestion of
drinking water contaminated with TCI [is another important concern. The extent
of TCI absorption after oral ingestion is nearly complete; with inhalation
exposure, the amount of TCI absorbed increases in proportion to its concentra-
2-8
-------
tion in inspired air, the respiratory rate, and duration of exposure.
The information regarding the oncodynamics of TCI as presented in the HAD
indicate that TCI is readily absorbed into the body through the lungs and gas-
trointestinal mucosa, the portals of entry for exposure to contaminated air,
water, and foods. TCI distributes throughout the body tissues, concentrating
in adipose tissue, with a relatively long half-time of residence (4 to 6
hours). Elimination of TCI involves two major processes: pulmonary excretion
of unchanged TCI, and biotransformation to mostly urinary metabolites. The TCI
epoxide, an intermediate metabolite, is judged to be the most likely causative
agent for inducing a carcinogenic response, although this has not been demon-
strated nor have the other metabolites been ruled out. Most of the metabolism
is thought to go through the epoxide. The quantitative risk assessment-in the
HAD based its estimate of effective dose on the total amount metabolized since
there was no way to estimate the epoxide amount only.
Although the principal site of metabolism of TCI in mammals is the liver,
lung, and the kidney, other tissues also metabolize TCI. Humans appear to
metabolize TCI extensively, more so than shown in animal test systems. Both
rats and mice also have a considerable capacity to metabolize TCI, and the
maximal capacities of the rat versus the mouse appear to be more closely
related to relative body surface areas (bw2/3) than to direct body weights.
Metabolism is nearly linearly related to dose at lower doses, becoming satu-
rated at higher doses, and is probably best described by Michaelis-Menten
kinetics.
Major end metabolites are trichloroethanol, trichloroethanol-glucuronide,
and trichloroacetic acid. The oxidation of chloral hydrate, an intermediate,
to trichloroacetic acid occurs mainly in the liver and kidney. Most tissues,
including the brain, can also reduce chloral hydrate to trichloroethanol which
2-9
-------
is conjugated by glucouronic acid.
Metabolism also produces several reactive intermediate metabolites, in-
cluding chloral, TCI-epoxide, dichloroacetyl chloride, formyl chloride, and
i
chloroform. Chloroform is characterized as being further metabolized to phos-
gene, which may covalently bind extensively to cellular lipids and proteins,
and, to a much lesser degree, to DNA or! other nucleic acids. The formation of
* i
i
chloroform may be an artifact of the analytical methodology and needs further
confirmation. |
The binding intensity of TCI metabolites is greater in the liver than in
I
the kidney. Comparable biotransformation studies of TCI in rats and mice
failed to detect any major species or strain differences in their metabolic
profiles.
Persuasive experimental evidence that the metabolic pathways for TCI are
qualitatively similar in mice, rats, and humans was presented in the HAD. In
these species, the principal urinary metabolites of TCI that have been identi-
fied are trichloroethanol (TCE),- and its glucuronide, and trichloroacetic acid
(TCA), Minor metabolites have also been identified for each of these species.
In the rodent liver, TCI is first metabolized to a reactive epoxide (TCI oxide)
by a microsomal P450 system. Reactive[intermediate metabolites, such as TCI
i . ,
epoxide, covalently bind to cellular macromolecules, principally protein, and
to a much lesser extent, DNA.
i
Quantitative differences in metabolism across species probably result from
differences in metabolic rate and enteifohepatic recirculation of metabolites.
The proportion of the TCI dose metabolized in rodents has been demonstrated to
decrease at higher doses, with the metabolism becoming nonlinear and approach-
ing saturation. However, metabolism i^ linearly proportional to the inhaled
concentration dose in humans up to aboijt 300 ppm (1,614 mg/m3). There is no
2-10
-------
evidence regarding the doses at which metabolism in humans is influenced by
saturation.
As documented in the HAD, the animal metabolism data correlate approxi-
mately with a dose equivalence scaling factor of bw2/3 between species in the
range of the NCI and NTP bioassay doses. This bw2/3 correlation provides a
basis for making estimates of daily animal exposure (fractional dose metab-
olized) in relation to the tumori genie response in the bioassays, and provides
a basis for scaling from rodents to humans to obtain an equivalent dose value
(amount of metabolite); this is known as a human equivalent dose. For example,
at oral doses of 500 to 2,000 mg/kg (covering the range used in the NTP and NCI
mouse bioassays, based on the data of Green and Prout (1985), the ratio of the
amount metabolized in the rat versus the amount metabolized in the mouse is as
follows:
Dose Ratio: rat /mouse / bw rat \
(mg/kg) (mg metabolized/animal) bw mouse
500 (55.40/13.73) = 4.03 3.6
1,000 (79.00/23.28) = 3.39 3.6
MOO ; (80.4/46.90) =1.71 3.6
At the highest dose, the ratio of the amount metabolized in the rat versus the
mouse was smaller than it was at lower doses, because of saturation of the
metabolic pathway in the rat. At the lowest dose, closer to 100% of the assi-
milated dose is metabolized in both rats and mice. These data indicate that,
for the dose range used in the cancer bioassays,, the metabolism and hence the
fractional TCI doses metabolized for rats and mice are more consistent with a
bw2/3 base (surface area) than a bw1 base (mg/kg).
2-11
-------
The use of a metabolism scaling based on bw^/3 is also supported by Dekant
i
et al. (1984), using balance studies comparing the metabolism of TCI in female
rats (Wistar) and female mice (NMRI). [ The 14C-TCI was administered as a single
dose by gavage in corn oil at 200 rag/kg. Radioactivity was determined in
urine, feces, carcass, and exhaled air for 72 hours post-administration. Total
I
recovery of radioactivity was 93% to 98%. As shown in the HAD (p. 8-92),
nearly complete oral absorption occurred with less than 1% in feces. For this
I
single dose, the rats exhaled 52% of the TCI dose through the lungs unchanged.
The mice exhaled 11% of the dose unchanged. The ratio of the amount of TCI
metabolized by the rat .versus that metabolized by the mouse was as follows:
Dose Ratio: Hat/mouse / bw rat \
(mg/kg) (mg metabolized/animal) bw mouse
200 (23.04/4.56)} = 5.05 4.5
j
i
i
The comparative surface area ratio of the two species, as calculated from their
experimental body weights, is (240/25.p)°*67 = 4.46. At this dose, the amounts
of TCI metabolized by the rat and mousle are closely proportional to their sur-
face areas. i
Green and Prout (1985) dosed rats' and mice orally for 180 days with TCI in
corn oil at a level of 1,000 mg/kg/day. The relative proportions of urinary
\
metabolites (TCA and TCE-glucuronide) were unaffected by the species difference.
The amount of the daily dose metabolized stayed constant, although the ratio of
i
TCA to TCE-glucuronide increased. TheSe observations suggest that (1) even at
a high dosage level, TCI does not accumulate significantly with daily dosing,
i.e., the total body half-life is sufficiently short (1 to 2 hours) so that
each daily dose is cleared (metabolized, exhaled, etc.) within the 24-hour
2-12
-------
dosage interval; and (2) the increased ratio of TCA to TCE-glucuronide suggests
a possible adaptation of metabolic pathways, which may occur simply because TCA
has a longer half-time in the body (T 1/2) than TCE. Prout et al. (1985)
estimated the T 1/2 for TCI for mice of 1.75 hours; therefore, 98% elimination
occurs in 10.5 hours (T 1/2 x 6), and for some 14 hours of each dose interval
(24 hours), exposure to the highly reactive intermediates formed from TCI is
practically nil. For the rat or human with longer metabolite tissue half-
times, proportionally less time free from exposure to the active chemical
species is available (T 1/2 = 4 to 6 hours for humans).
The toxicities (except possibly for the acute cardiac response) associated
with TCI exposure were considered to result from the presence of TCI's reactive
metabolites. For example, the observed chronic hepatoxicity of TCI in mice is
thought to be a function of TCI metabolism in the liver. Any mutagenic and
carcinogenic potential of TCI is similarly thought to be related to the reac-
tive intermediate biotransformation products instead of the parent TCI molecule
itself, although the biological mechanism(s) by which TCI metabolites influence
the carcinogenic activity observed in experimental animals are not known.
The observed behavioral and psychological effects, particularly as they
affect psychomotor performance, have been reported at levels of 200 ppm
(1»076 (tig/m3) in some, but not all, experimental and epidemiologic studies.
fhefe1 !§ ho information that such impairment, suggesting dysfunction of the
central nervous system (CNS), would occur during chronic, low-level exposures.
At levels found or expected in the ambient environment (< 200 ppm), such an
effect would be unlikely.
Teratogenicity is another health end point for which available data from
experimental animals suggest that the conceptus is not uniquely susceptible to
TCI. Exposure of rats, mice, and rabbits, during gestation, to levels (300 ppm
2-13
-------
or 1,614 mg/m3) greater than those generally found in the environment, has not
resulted in any observation of teratogenic effects. The teratogenic potential
of TCI for humans cannot be directly extrapolated from the observations in the
animal studies. However, the animal studies do not indicate that TCI is toxic
to the fetus at levels below the maternally toxic level. Alsos no definitive
clinical evidence of fetotoxicity or teratogenicity from TCI exposure has been
reported.
Metabolism of TCI is enhanced by Hrugs (e.g., barbiturates, oral anti-
diabetic agents) and by other xenobiotfics (e.g., PCBs) which induce the hepatic
microsomal P450 metabolizing system. Heightened toxicity can result from drug
interactions known to occur with ethanol , barbiturates, disulfiram, and Warfarin.
2.4. CANCER RISK DERIVATION
Four sets of gavage bioassay data| that show hepatocellular carcinomas in
male and female mice provide a basis t|o calculate an upper-bound carcinogenic
slope estimate for TCI using a linearijzed multistage model. The responses and
human equivalent doses (HED) are presented in Table 2-2.
The HED was used to calculate the1 qt, and the risk was adjusted to reflect
i x
the risk at an applied dose of 1 mg/kg7day. Based on the NCI data, the risk
Was also adjusted from only 90 weeks of exposure as compared to a lifetime
exposure (104) weeks. Since the upperi-bound potency estimates that were ca1 -
culated based on these data sets were icomparable, ranging from 5.8 x 10"3 to
I
1.9 x 10~2 mg/kg/day, a geometric mean, q£ = 1.1 x 10"2 mg/kg/day was selected
and used to develop the incremental lifetime cancer risk (i.e., unit risk) of a
i
unit exposure of TCI in drinking water and in air.
i
The upper-bound estimate of the cancer risk from 1 yg/L of TCI in drink-
7
ing water is 3.2 x 10"'. The upper-bound estimate of the cancer risk from
3 of TCI in air is 1.3 x 10"6. This was derived from the geometric
2-14
-------
TABLE 2-2. DOSE CONVERSION FOR THE B6C3F1 MOUSE BY THE ORAL ROUTE
Oral experimental
dose (time-weighted)
mg/kg/day mg/day
Animal Human equivalent dose
metabolized dose (lifetime average exposure)
mg/day
_ Incidence of
mg/day mg/m'/day* mg/kg/day liver tumors
NTP (1982)
Males
0 0
1000 40
Females
0 0
1000 35
0
. 31.98
0
28.17
00 0 8/48
3317. 23° 1793.10 " 47.39 30/50
0 0 0 2/49
3194.06b 1726.52 45.62 13/49
NCI (1976)
Hales
0 0
1169 38.58
2339 77.19
Females
0 0
869 22.59
1739 45.21
;.. ,
0
30.90
58.77
(39.96)d
0
18.49
35.89
(34.45) EPA, 1985a.
2-15
-------
* 9
mean c\i = 1*3 x 10~^ using the HED (metabolized) and the liver tumor incidence
from the NTP and NCI studies together with an estimate of the amount humans
metabolize daily (9.9 x 10"5) based On the work of Monster et al. (1976a) if
i
exposed to 1 vg/m3 daily. The upper-abound nature of these estimates is such
that the true risk is not likely to exceed this value and may be lower. The
i
carcinogenic potential of TCI is generally considered to result from the pre-
i
sence of cellular reactive intermediate metabolites, and therefore metabolic
and pharmacokinetic factors have been; used in the calculation of the drinking
water and air unit risks.
Several other data sets were analyzed to determine whether derived unit
i
risk estimates would be comparable arid to investigate alternative bases for
I
risk estimation. An alternate analysis was made using the incidences of hepa-
i
tocellular carinoma in male mice exposed to an airborne concentration of TCI
from the study of Bell et al. (1978)i
i
Airborne concentration Lifetime exposure Liver carcinoma
(ppm) ' (ppm) incidence
0 : 0 18/99
i
100 | 17.9 28/95
300 53.6 31/100
600 107.1 43/97
I
This analysis gave an estimate of risk of
(animal) = 4.8 x 10~3/ppm or 8.8 x 10"7 yg/m3
Under the assumption that the amount metabolized per body surface area is
2-16
-------
equivalent'between mice and humans, the human concentration C(ng/m3) must
satisfy the relationship
(C x 20 m3/day)/70 kg0-67 = (1 x 0.043 m3/day)/0.035 kg0-67
where 20 m3/day and 0.043 m3/day are assumed to be, respectively, the volumetric
metric breathing rates for a human weighing 70 kg and a mouse weighing 0.035
kg. As a consequence, C = 0.34 yg/m3. That is, the exposure concentration,
0.34 yg/m3, for humans is considered equivalent to that for mice at 1 yg/m3.
A human q^ was estimated as follows:
qj = 8.8 x 10-7/0.34 = 2.6 x lO'6
This estimate was not adopted because of the problems in the quality of the
underlying bioassay.
While the six epidemiologic studies or surveys are considered inadequate
to evaluate the likelihood of TCI being a human carcinogen, subcohort data
from Axelson et al. (1978) were used to develop an alternative estimate of the
upper limit of risk from a negative study. This was done to provide a basis
for risk value comparisons with the animal based risk estimates. The calcula-
ted human estimate was higher than the estimate from animal data, and thus did
not contradict the risk estimates calculated from the animal data.
Expressed as relative potency, the unit risk estimates derived from the
mice data would place TCI in the lowest quartile among the 54 suspect or known
human carcinogens evaluated by EPA's Carcinogen Assessment Group.
2-17
-------
-------
3. ONCOLOGY AND TOXICOLOGY
Since the July, 1985 publication of the TCI Health Assessment Document
(HAD) (U.S. EPA, 1985a), two major sources of animal bioassay data have been
published in the scientific literature. The main goal of this revaluation of
the carcinogenicity of TCI is to consider the impact of the oral and inhalation
bioassays reported by Maltoni et al. (1986) of TCI in rats and mice and the
inhalation assay by Fukuda et al. (1983) on the assessment of the risk of TCI
as an environmental airborne pollutant. The EPA was aware of the investigations
of Fukuda et al. (1983) on the inhalation toxicity of TCI in mice and rats, but
these investigations were not considered in the 1985 HAD because the results
were published after the Science Advisory Board review of the HAD document.
Secondarily, the impact of new "long-term bioassays involving oral administra-
tion of TCI is evaluated from the perspective of their impact on the weight-of-
evidence characterization of the likelihood for human cancer potential. The
National Toxicology Program (NTP, 1987) reported on toxicology and carcinogen-
esis studies of TCI by the oral route (gavage) in four strains of rats (ACI,
August, Marshall, and Osborne-Mendel). The NTP (1987) report also included a
review of the toxicology of TCI and new information on the mutagenicity of TCI.
Maltoni et al. (1986) also reported on the response to TCI gavage in Sprague-
Dawley rats. The work of Henschler et al. (1984), who studied the potential
carcinogenic effects from oral administration of TCI with and without epoxide
stabilizers, will be reviewed. This investigation was known to the Agency when
the 1985 HAD was published, but documentation was not available at that time.
The EPA-sponsored work by Herren-Freund et al. (1986, 1987) on the induction of
cancer in mice by TCI and by either of two end metabolites when administered in
the drinking water is also included in this update. Lastly, a new epidemiologic
3-1
-------
study (Shindell and Ulrich, 1985) wil;l be reviewed and the findings integrated
into the overall characterization of the human data.
3.1. ANIMAL BIOASSAYS
1 |
3.1.1. NTP (1987) Rat Study; Oral Exposure
The NTP (1987) chronic oral rat Ibioassays using the ACI, August, Marshall,
and Osborne-Mendel strains of rats were conducted at Papanicolaou Cancer Insti-
tute, Miami, Florida. The study desijgn was similar to that used for the study
(NTP, 1986a) on the F344 rat and reported in the HAD from a 1982 draft.
i
TCI (epichlorohydrin-free) was a'dministered 5 days/week by gavage (in corn
i
oil) for 2 years to four strains of rats. There were 50 rats of each sex in
each of two dose groups (500 and 1,000 mg/kg) plus a similar number in the
respective control groups, vehicle arid untreated. The volume of the dose was
1.0 ml/kg except in the case of the Osborne-Mendel rats, which received 5 mL/kg
for the first 7 weeks. Selection 'of [the doses was based on the results (minimal
body weight difference from controls jand survival) of 13-week studies or a
previous chronic study in the case of the Osborne-Mendel strain. The animals
were housed five per cage and provided food and water ad libitum. They were
6.5 to 8 weeks of age when the dosing began.
the animals were observed twice [daily. Body weights were determined
weekly for the first 12 through 15 weeks and monthly thereafter. Clinical
pathologic evaluations were not performed during the course of the study.
i
Necropsy was performed on all animals[. Gross lesions and tissue masses, plus
sections of the skin, mesenteric lymph node, mammary gland, salivary gland,
thigh muscle, lungs and main stem bronchi, heart, thyroid gland, parathyroids,
esophagus, stomach, duodenum, ileum, colon, liver, vertebrae with bone marrow,
i
thymus, larynx, trachea, pancreas, spleen, kidneys, adrenal glands, urinary
bladder, brain, eyes, pituitary gland;, spinal cord, and seminal vesicle/pro-
3-2
-------
state/testis or ovaries/uterus were histopathologically evaluated. The patho-
logic diagnoses were reviewed by the NTP Pathology Working Group. While per-
formance under good laboratory practice was not required, these studies were
the subject of data audits.
The percent of the rats surviving at the termination of the study and the
mean body weights at midstudy are tabulated in Table 3-1. The survival of all
treated groups was lower than that of the controls. The survival difference
between the untreated (unt) and vehicle (veh) groups was not remarkable. The
midstudy mean body weights of the untreated control groups tended to be slight-
ly higher than those of the vehicle controls. The terminal mean body weights
of all dosed groups except for the high-dose female Osborne-Mendel rats were
somewhat lower than those of the vehicle controls. Signs of central nervous
system toxicity (ataxia, lethargy, hindlimb paralysis, convulsions) were ob-
served sporadically in all dose groups.
There was a decreased incidence of pheochromocytomas of the adrenal glands
in both sexes of Osborne-Mendel rats, among the female August and Marshall rats,
and among the male ACI rats. An increased incidence of interstitial cell tu-
mors of the testis was observed in the high-dose male Marshall rats (unt-16/46,
veh-17/46, 500 mg/kg-21/48, 1,000 mg/kg-32/48). TCI oral administration was
also associated with an increased incidence of renal tubular cell adenomas and
adenocarcinomas. However, the NTP concluded that the study was inadequate for
assessing either the absence or presence of carcinogenicity because of reduced
survival, toxicity, and deficiencies in conduct of the study. The NTP (1987)
indicated that while not acceptable for carcinogenic evaluation, the studies
did, however, indicate renal toxicity. They reviewed the results regarding
renal toxicity of these studies along with those reported in the F334 rat (NTP,
1986a). The histopathologic evaluations of the kidneys of all five strains was
3-3
-------
TABLE 3-1. PERCENT SURVIVAL AT TERMINATION AND MEAN MIDSTUDY BODY
HEIGHTS OF FOUR STRAINS OF| RATS TREATED BY GAVAGE WITH TCI
Strain
AC I
Marshall
August
Osborne-Menclel
Dose
Sex (ing/kg)
M unt
veh
500
1,000
F unt
veh
500
1,000
M unt
veh
500
1,000
F unt
veh
500
1,000
M unt
veh
500
1,000
F unt
veh
500
11,000
M unt
! veh
500
it 000
F unt
veh
500
iLooo
Survival
%
78
76
38
22
74
70
40
38
64
52
24
12
62
60
24
20
48
42
26
32
52
46
52
50
42
44
34
30
38 '
40
22
14
Mean
body weight
(g)
350
327
307
217
212
215
311
312
304
235
232
222
407
382
364
212
206
208
502
481
446
301
298
292
SOURCE: Summarized from NTP, 1987.
3-4
-------
summarized as shown in Table 3-2. The NTP concluded that the toxic nephropathy
was clearly not the spontaneous lesion occurring in aging rats; rather, the
lesions were characterized by cytomegaly, karyomegaly, and toxic nephrosis of
the tubular epithelial cells in the inner renal cortex. Cytomegaly was diag-
nosed in rats that died as early as 26 weeks and was directly proportional in
severity to the duration of dosing.
The NTP (1987) reported that there was a statistically significant increase
in renal adenomas among the males of the Osborne-Mendel strain as well as an
increase in renal adenocarcinomas among male F334 rats as previously reported
(NTP, 1986a). The NTP also noted that one tubular cell adenocarcinoma had been
observed in the low-dose (549 mg/kg) male Osborne-Mendel rat in a previous
study (NCI, 1976).
3.1.2. Henschler (1984) Mouse Study; Oral Exposure
Henschler et al. (1984) studied the effects of oral administration of TCI
with and without epoxide stabilizers, (epichlorohydrin [EPC] and 1,2-epoxybu-
tane [BO]) on ICR/Ha Swiss mice (50/sex/dose group). The mice were 5 weeks of
age at the start of dosing. The test mixtures were administered at doses of
2,400 and 1,800 mg TCI/kg to males and females, respectively, by gavage in corn
oil 5 days/week. It was intended that the mice be treated for 18 months, with
Observation until the 24th month. The test mixtures incorporated into the cord
oil were as follows:
TCI pure - stabilized with 0.0015% triethanolamine
TCI industrial (99.4% pure) - (contained 0.603% EPC and 0.11% EPC)
TCI + 0.8% EPC (epichlorohydrin)
TCI + 0.8% BO (1,2-epoxybutane)
TCI + 0.25% EPC + 0,025% BO
3-5
-------
CO
H-
s
as
*>.
«=»
CO
CO
u,
g
eC
_T
UJ
0
2=
UJ
i^
UJ
2S
O
CO
CO
o
1
M«J
3= O
CO r
< u_
£ 0
>co
(_ UJ
CO 1-4
ID =
=3 t
a co
i-i CD
< >
as
CO fC.
2= UJ
0 >-
1-4 1
co CM
UJ
_1 UJ
J I
0 2=
> « 14
i
r*
O
O
3
2=
UJ
U.
0
CO
UJ
O
2=
/S
t 4
y
1-4
CM
I
CO
UJ
CQ
a;
1-
03 ,
r- E
i O
O) C
0 T-
o
C_ L.
CO (O
i O
3 O
.0 C
3 d)
I X)
^
r»
O)
O 03
E
C_ 0
re c
r O)
3-0
3
t
r
r rtJ
O) !-
O t/>
CO
C- i
^3 ^,
3 O)
-Q O.
3 >,
h- J=
>>
ic^
4-^
O rtJ
S§-
o c_
1 ^:
n.
OJ
o.
3
O
r_
CD
CTl O CTl CTl
«* ID "i" «d"
O 0 r-H O
CTl O CTl CTl
^- ID ^" *^
"s^» "*^^ "^^w *"*^»
o o o o
CTl O CTl CTl
*i- ID *3" «d"
0 O 0 O
CTl O CTl CTl
«* ID «d- <*
"^*v ^^*. *^*^ "*"*«
o o co co
r-4 r-l
*O
£_ ,
+J O
C £-
^3 ^-^
O c: cn
O -^
*o O cn*^^.
oi jw; cn
+-> o) "^ E
tt) "o E o
t_ T- O
!-> J= O O
C O) O >
=) >. ID r-l
CO
O)
r I t
03 O
y cc
O
ID
O O CTl
D ID «*
0 0 r-4 0
O
ID
*^^
=1 O CTl
D ID «*
"****. ****«« ****.
O O r-4 r-4
O
ID 1
O
z> a cn
D ID «d-
0 0 r-4
O O O CTl
ID ID ID «*
**«»^ *
O <
"o
C- r
) t
d
0
O
o
0)
rd i
Ol
C-
4-> .
c
"**«» *"**. "*"*»«
O O r-4
r-H CO
O
t_
»->
c cn
o jx:
o cn*^^.
01 ^t i*
o e o
r- 0
COO
O) O >
=5 > ID r-4
4J
CO
3
cn
3
CTl CTl C3 f^-
^J- ^J- LD *^"
O O O t-4
CTl CTl CD r*-*
* «^ ID «d-
^**^ "^*. *^*«t ^"*»*
CM O r-4 O
CTl CTl O I"*-
^ «* LD «*
O O r-l O
CTl CTl O f-.
«* «d- ID <*
**^^ -^*« ^^x. ^**«.
O O CO CO
r-4 CM
"o
t- 1
4-> O
c t_
o +->
u c cn
o ^
XJ o cn**^.
O) -^ cn
4-> a) ^-. e
fO r Cn
O) O E: O
t-.r- 0
-P J= O O
c: o) o >
=» ID r-l
^n
03
-C
co
c_
03
o o o o
ID ID ID ID
o a a r-<
o o o o
ID ID ID ID
*^*«« ^v^ *»*^ ^.^^
O O VO r-4
O O O O
LD LD ID ID
O O LD CO
O O O O
ID ID ID ID
*"*"' "****. """"""Hi "****>
O O CTl ID
CO CO
"o
t_ 1
4-> O
C t-
O +J
o e cn
o jx;
*O O cn *,
O) ^ cn
4-> o> ~^- E
03 r cn
O) o e o
l_ T- O
-4-> ^r o o
c o> o «
ZD > ID r-4
'a;
a
c=
a>
s:
O)
c:
t_
o
co
C3
CTl OO CTl CTl
O O O CO
CTl OO CTl CTl
^^ ^" *5t* CT>
"***«* *Sii*i. "'*** ^*^»
O O OO CO
^J" ^*
'o
£_ ,
4J 'O
C t-
o +->
o c cn
O ^^
a o cn--.
O) -iyi cn
4->
Z3 S» ID r-4
2:
>.
^«
«5j-
CO
u.
^-~.
ai
cn
fO
d.
cn
c
3t
6
_
o
14-
a>
%
o
-a
0)
3
C
> ^
C
0
o
~~*
3-6
-------
*~ »
T3
CD
C
f
4_>
C
o
CM
CO
LU
1
CO
-- E
O
CO c:
O !
o
t_ J_
<0 10
r O
3 O
_o c
3 01
p ~O
re)
~
r-~
(11
O g
t- O
re) c:
"3 -o
JQ re)
3
h-
11
CO -t-
CJ CO
(0
<- 1
re) o.
, ^
. 3 CO
.£> 0.
3 >>
1 -E
>>
JCZ
1 N
o re)
o t_
o.
CO
c
jQ
cn 00 r- co
«d- «=t- «* «*
"""i ****** "*''"» """"^i
O O i-l i-l
cn oo i- co
«****
O O CM O
cn oo i-- co
^- «^- ^j- ^-
o o o o
cn co r~ co
«* ** «=& -=t-
""""i. ****^. *"""» "***«*
O o i-i cn
CM t 1
"o
£_ ( "
4-9 O
C C_
O 1 --*
0 c cn
o -*:
-o o cn\
co \.if cn
+-> co ---..£
re) r1 cn
CO 0 EO
t- -i O
4-> .C O O
c cu o >
=> > LO ^H
CO
d)
re)
E I-H
0
C t-
O +->
o c cn
O -^
T3 O Cn-- .
CO N^ cn
4-9 CO " ^. E
re) ^~ cn
CO O E O
£_ -i- 0
4-> -E C3 O
cr a) o «
=> > LO ^H
CO
3
cn
3 '
cn o oo ^*
* ^«. *^"^ ^
O O 1-1 1-1
cn o oo «*
*5j* LO -^ ^*
r 1 i 1 i 1 O
cn o oo «*
«* LO -=J- «*
i 1 r 1 O i 1
cn o oo «*
«* LO =! ^-
**""^ **^i. "x*^« "*111x.
o o o o
CO CO
'o
C_ i
4-9 O
E t-
0 4->
o E cn
o *^
T3 o cn
0) -*: cn
re) i cn
CO O E O
t- T- O
4-> -C O O
E CO O >
=> 5> LO r-1
r__
^-
j2
CO
i_
re)
CD o o cn
LO LO LO **
*^^ ^ ^*^» *».
o o o o
o o o cn
LO LO LO *
^ 0 0 i-H
o o o cn
LO tO LO «d-
O i-H i 1 CO
O o O cn
LO LO LO «3"
* . ^ «. ^ ^ .
cs o o cn
CO CO
'o
t_ r~~
4-> O
c t-
O 4->
o E cn
O -l£
o o cn---.
CO -^ cn
re> i cn
CO 0 E 0
£- -i O
4-9 .E O O
E 0) O >
=3 > LO i-l
"co
-o
, E
*
CU
E
O
^3
10
O
cn o cn oo
"""*»* "**^ *Xl** "*"1^«
O O O i-l
cn o cn oo
* LO -3- »*
o o o o
cn o CM oo
«* LO <* «^-
o o o o
cn o cn oo
"**«» "fc,^^^^ "i*^^ "n^^^
O CD OO OO
*0
C_ i
4-9 O
E C-
O 4-9
o E cn
O x>*
"O o cn*^^,
co j«j cn
4-9 CO ^ E
re) i cn
CO O E O
S--r- O
4-> -E O O
E CO O >
=3 > LO i-l
^.
ci"
<^-
co
U-
-o
CO
re)
cn
CO
» o
re) O E
re) co t3
CZ to <£
re) _o
00
cn
r 1
w
Qu
^
UJ
cm
0
oo
3-7
-------
A sixth group was administered corn oil only and served as a control. The
females were housed five per cage and the males one per cage. Food and water
were provided ad libitum. Body weights were determined weekly. The animals
were observed twice daily. Because of toxicity, gavage dosing was stopped
during weeks 35 through 40, 65, and 69 through 78. All doses were halved at
[
the 40th week. The daily doses over the course of the study averaged about 30%
of the initial intended doses.
Necropsy was performed on all animals. Gross lesions and tissue masses,
i
plus sections of most tissue, were evaluated histopathologically. The patho-
logical diagnoses were done by two pathologists. These experiments began in
April 1977.
The authors (Henschler et al., 1984) concluded that there was no signifi-
cant increase in tumors except in the groups treated with TCI containing 0.8%
EPC or BO and TCI + 0.25% EPC + 0.025% BO. In these groups there was an in-
crease in forestomach cancers. The authors attributed this increase to the
direct alkylating properties of EPC and BO.
i
The authors reported central nervous system effects occurring among all
groups dosed with TCI. Growth and survival were decreased by all treatments as
compared to the controls.
The hepatocellular tumors and preneoplastic nodules reported in this study
are listed in Table 3-3.
3.1.3. Herren-Freund (1986, 1987) Mouse Study; Oral Exposure
i
Recently, Herren-Freund et al. ,(1986, 1987), as sponsored by EPA, reported
i
on the comparative carcinogenic potejncy of TCI and its metabolites, trichloro-
acetic acid (TCA) and dichloroacetic acid (DCA), in the mouse liver. The test
materials were administered in the drinking water for 61 weeks to male B6C3F1
i ' '
mice beginning on day 28 of life. M|ice into whose drinking water salt (NaCl)
3-8
-------
TABLE 3-3. TUMORS AND PRENEOPLASTIC LIVER NODULES
FROM EXPOSURE TO TCI WITH AND WITHOUT STABILIZERS
TO ICR/Ha SWISS MICE
Type of liver tumor
Carcinoma
Treatment
(mice autopsied)
Control (m50/f50)
TCI pure (m49/f50)
TCI industrial (m49/f50)
TCI + EPC (m49/f50)
TCI + BO (m49/f48)
TCI + EPC + BO (m49/f50)
M
0
0
0
0
0
0
F
0
0
..0
0
1
0
Adenoma
M
3
5
8
3
4
4
F
0
1
1
1
1
0
Nodule
M
0
1
1
0
0
0
F
0
0
0
0
0
0
SOURCE: Adapted from Henschler et al., 1984.
3-9
-------
and phenobarbital had been incorporatjed were used as controls. Some groups had
been pretreated on day 15 of life with i.p. 2.5 or 10 y/g of ethylnitrosourea
i
(ENU). After 61 weeks of treatment by incorporation of the test agents into
i
the drinking water, all surviving mice were killed. The livers were evaluated
histopathologically.
TCA and DCA at 5,000 mg/L slightly reduced the body weight. The dosing
regimen and the results of the pathollogic evaluation are presented in Table
3-4. The percentage of mice with liver carcinomas was increased, as were the
I
mean number of adenomas per mouse and the mice with adenomas. The TCI-treated
i
mice had an increased response, but the increase was not statistically signifi-
cant, possibly because of the low dos|e of TCI received and metabolite formed.
The authors concluded that TCA and DCA caused cancer without prior initi-
ation with ENU. The authors pointed iout that while DCA and TCA have not been
shown to be genotoxic, they acted as 'complete carcinogens in this assay.
3.1.4. Fukuda et al. (1983) Rat and [Mouse Study; Inhalation Exposure
Fukuda et al. (1983) reported the results of bioassays with female mice
and rats exposed to airborne concentrations (50, 150, and 450 ppm) of TCI.
Groups of 49 to 51 Sprague-Dawley rats and ICR mice were exposed to graded con-
centrations of TCI (0.019% EPC) 7 hoiirs/day, 5 days/week for 107 weeks. (The
i
animals were purchased from the Charles River Laboratory; thus, the designa-
tions CD-I and CrL-CD [SD] might also be used.) A similar number of female
animals served as controls. The experimental animals were 7 weeks of age when
I
the exposures started. The rats were housed 8 per cage and the mice 12 per
cage. Food and water were provided ad libitum. The exposures were conducted
I
in dynamic exposure chambers measuring 1.96 cubic meters. The chamber con-
centrations were monitored at 30-minlite intervals using gas chromatography.
It is assumed that the majority of tissues from most animals were evaluated
3-10
-------
TABLE 3-4. LIVER CANCER RESPONSE IN MALE B6C3F1 MICE
FOLLOWING EXPOSURE TO TCI AND METABOLITES
IN THE DRINKING WATER FOR 61 WEEKS
Agent
TCI
TCI
TCI
TCI
DCA
DCA
DCA
TCA
TCA
TCA
TCA
NaCl
NaCl
NaCl
PB
PB
Concen-
tration
(mg/L)
40
40
3
40
5,000
2,000
5,000
5,000
5,000
2,000
5,000
2,000
2,000
2,000
500
500
Dose
(mg/kg)
8
8
0.6
8
1,000
400
1,000
1,000
1,000
400
1,000
ENU
(i.p., yg/g)
10
2.5
2.5
0
2.5
2.5
0
10
2.5
2.5
0
10
2.5
0
2.5
0
Number
of mice
19
25
27
31
32
29
26
28
23
33
22
23
21
22
16
22
Liver
carcinomas
(%)
37
4
11
3
72
66
81
54
48
48
32
39
5
0
0
0
Mean
number/
mouse
0.58
0.08
0.11
0.01
1.47
1.17
1.69
0.93
0.57
0.64
0.50
0.57
0.05
SOURCE: Adapted from Herren-Freund et al., 1986, 1987.
3-11
-------
histopathologically, since the author makes no comment.
The authors (Fukuda et al., 1983) reported that the body weight changes
were normal and that survival was only adversely affected amo.ng the rats at 85
weeks and after 100 weeks. There wasino statistically signficant increase in
I
tumors among the rats; however, the authors indicated that there was one renal
clear cell carcinoma among the rats in the high-dose group. Among the mice
i
there was a statistically signficant increase in lung adenocarcinomas, but not
adenomas or combined tumors, as indicated in Table 3-5. The average number of
tumors per mouse was increased at theitwo highest exposure concentrations.
i
The authors concluded that the increased incidence (statistically signif-
i
icant at 450 and 150 ppm) was caused by TCI and that the findings suggest that
TCI exposure prompted the transformation of adenomas into carcinomas. P values
were not provdided by the authors, bull see discussion section for additional
analysis.
3.1.5. Maltoni et al. (1986) Rat and Mouse Study; Oral and Inhalation Exposure
I
The 1985 HAD refers to a preliminary report by Maltoni et al. (1979) in
I
which Sprague-Dawley rats were given TCI by gavage. The series of investiga-
tions published by Maltoni et al. (1986) provides an expanded and final report
of this and other bioassays from their laboratory.
3.1.5.1. Rat Study; Oral ExposureT(tl (epoxide-free and stabilized with butyl-
j
hydroxytoluene) was administered 4 or!5 days/week by gavage (in olive oil) for
52 weeks to Sprague-Dawley rats. After dosing, the rats were maintained until
they died. The experimental rats were bred in these investigators' laboratory.
The rats were 13 weeks of age when the dosing began. There were 30 rats of
i
each sex in each of two dose groups (250 and 50 mg/kg), plus a similar number
in the vehicle control group. The animals were housed five per cage and pro-
vided food and water ad libitum. The animals were observed three times daily.
3-12
-------
TABLE 3-5. INCIDENCE OF LUNG TUMORS IN FEMALE ICR MICE
EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK FOR 107 WEEKS
Mice with tumors/mice examined (average tumors/mouse)
Airborne
concentration (ppm) Adenocarcinomas Combined
450 7/46 11/46 (0.39)
150 8/50 13/50 (0.46)
50 3/50 5/50 (0.10)
0 1/49 6/49 (0.12)
SOURCE: Adapted from Fukuda et al., 1983.
3-13
-------
I
Body weights were determined biweekly [during treatment and every 8 weeks there-
after.
Necropsy was performed on all aniimals. Histopathologic evaluations were
i
performed on gross lesions and tissue masses, plus sections of the skin; sub-
cutaneous, mediastinal, and mesentric |lymph nodes; mammary gland, Zymbal gland,
salivary gland, Harderian gland, eyes, tongue, thyroid gland, bone marrow
(femur), thymus, larynx, lungs, heart; aorta, esophagus, diaphragm, liver kid-
ney, pancreas, spleen, adrenal glandsj intestine (three levels), urinary blad
i
der, brown fat, brain, seminal vesicle/epididymes/prostate/testis or ovaries/
I
uterus, right thigh muscle, and pituitary gland. The pathologic diagnoses of a
i
junior pathologist were reviewed by a senior pathologist. These experiments
began as early as 1976, and hence not all were conducted strictly according to
good laboratory practices, but the methodological protocol was said (Maltoni
[
et al., 1986) to have met the requirements of the "Good Laboratory Practices
Act."
There was a slight increase in tlie deaths among the treated female rats.
The average body weights were not affected. The average body weights after 1
year of treatment were 600 g for the males and 350 g for the females. Survival
time for the last rat was 140 weeks. The investigators (Maltoni et al., 1986)
reported an increase in renal meganucjeocytosis (cytokaryomegaly) among the
treated male rats, as indicated in Table 3-6. The tumor response was limited
to a slight increase in immunoblastic lymphosarcomas, a form of leukemia, among
the treated males, as shown in Table 3-7.
3,1.5.2. Rat Study; Inhalation ExposureMaltoni et al. (1986) also reported
1
the results of several bioassays with I mice and rats exposed to airborne concern
trations of TCI. Sprague-Dawley rats were exposed to graded concentrations of
TCI (epoxide-free stabilized with butyl-hydroxytoluene) 7 hours/day,, 5 days/
! 3-14
-------
TABLE 3-6. INCIDENCE OF RENAL MEGANUCLEOCYTOSIS IN MALE RATS
TREATED WITH TCI 4 TO 5 DAYS/WEEK BY 6AVAGE
FOR 52 WEEKS
Dose
(mg/kg)
% of rats (no. of 30) with lesions
250
50
Control
46.7 (14)
23.3 (7)
0
SOURCE: Adapted from Maltoni et al., 1986.
TABLE 3-7. INCIDENCE OF LEUKEMIA AND IMMUNOBLASTIC LYMPHOSARCOMAS
IN MALE RATS TREATED WITH TCI 4 TO 5 DAYS/WEEK BY GAVAGE
FOR 52 WEEKS
Dose
(mg/kg)
250
50
Control
Leukemia
10
6.7
3.3
% of 30 rats with
3.
6.
0
lesions
Immunoblastic
lymphosarcomas
3 (1/25 - 67.3 weeks
7 (2/28 - 70 weeks)
)a
^Number affected/number alive at first leukemia - average latency.
SOURCE: Adapted from Maltoni et al., 1986.
3-15
-------
week using three different exposure regimens, as indicated in Table 3-8. The
i
authors indicated that the exposure concentrations were selected because of
their similarity to the occupational exposure limits. After 8 or 104 weeks of
I
exposure, the rats were maintained unt|il they died. The experimental rats were
bred in the investigators' laboratory. The rats of the 8-week exposure were 13
weeks of age when exposure was begun, while those exposed for 104 weeks were 12
weeks of age when exposure was begun. The animals were housed 10 per cage
during the inhalation phase and five pbr cage after the exposure period, and
were provided food and water ad libitum. The animals were observed three times
daily. Body weights were determined bjiweekly during treatment and every 8
weeks thereafter.
The exposures were conducted in dynamic chambers measuring 8.5 cubic
meters, with an air flow providing 12 to 15 changes per hour. The chamber
concentrations were monitored continuously using gas chromatography.
Necropsy was performed on all animals. Histopathologic evaluations were
made on gross lesions and tissue masses, plus sections of the skin; subcutane-
ous, mediastinal, and mesenteric lymph! nodes; mammary gland, Zymbal gland,
i
salivary gland, Harderian gland, eyes,!tongue, thyroid gland, bone marrow
(femur), thymus, larynx, lungs, heart, aorta, esophagus, diaphragm, liver
kidney, pancreas, spleen, adrenal glands, intestine (three levels), urinary
bladder, brown fat, brain, seminal vesiicle/epididymes/prostate/testis or ova-
ries/uterus, right thigh muscle, and pjituitary gland. These experiments began
as early as 1976, and hence not all wehe conducted strictly according to good
i
laboratory practices, but the methodological protocol was said (Maltoni et al.,
i
1986) to have met the requirements of the "Good Laboratory Practices Act."
BT302Results of 8-week exposure^ There was no change in survival or
mean body weight among the rats exposed to airborne concentrations of TCI as
3-16
-------
TABLE 3-8. INHALATION TREATMENT REGIMEN FOR RATS
Laboratory
designation
BT302
BT304
BT304bis
Concentration
(ppm)
600
100
0
600
300
100
0
600
300
100
0
Duration
of exposure
(weeks)
8
8
8
104
104
104
104
104
104
104
104
Number
Male
72
,60
90
9.0
90
90
95
40
40
40
40
of rats
Female
72
60
90
90
90
90
105
40
40
40
40
SOURCE: Adapted from Maltoni et al., 1986.
3-17
-------
compared to the controls. The average body weight at 64 weeks of age was about
600 g and 375 g for males and females, respectively. There was no increase in
the incidence of renal tubuli meganucleocytosis (or cytokaryomegaly) among the
male rats exposed to TCI for 8 weeks and observed for 164 weeks.
The authors (Maltoni et al., 1986) concluded that neither the number of
i
tumor-bearing animals nor the incidence of the most frequently expected tumors
(mammary tumors, leukemias, pheochromocytomas, and pheochromoblastomas) was
I
increased by 8 weeks of exposure to Tpl at 600 or 100 ppm. These incidences
are given in Table 3-9. ;
BT304 and BT304bisResults of 104-week exposure; There was no change in
I
survival or mean body weight among this rats exposed to airborne concentrations
of TCI as compared to the controls. The average body weight at 64 weeks of age
was about 620 g (575 g - 304bis) and 390 g (360 g - 304bis) for males and
females, respectively.
There was an increased incidence of renal tubuli meganucleocytosis among
the male rats exposed to TCI for 104 weeks and observed for 159 weeks, as indi-
cated in Table 3-10. The first lesion of this type was noted at 47 weeks.
The authors (Maltoni et al., 1986) concluded that neither the number of
tumof^bearing animals nor the incidence of some frequently expected tumors
(mammary tumors, pheochromocytomas, and pheochromoblastomas) was increased by
104 weeks of exposure to TCI at 600, 300, or 100 ppm. These incidences are
given in Table 3-11.
Maltoni et al. (1986) reported a-slight increase in pituitary adenomatous
growth in experiment BT304 that was npt confirmed in experiment BT304bis after
I
an observation period of 159 weeks. When the two 104-week studies were com-
bined, the incidences reported were as shown in Table 3-12.
3-18
-------
TABLE 3-9. INCIDENCE (%) OF TUMORS AMONG SPRAGUE-DAWLEY
RATS EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI 5 DAYS/WEEK FOR 8 WEEKS
TCI air concentration (jDpm)/sex
600
Parameter
Tumor-bearing rats
Mammary tumors
Leukemias
Pheochromocytomas
Pheochromobl astomas
M
53
14
4
19
6
F
74
57
3
18
0
300
M
30
7
5
10
3
F
85
67
7
15
0
0
M
60
14
19
20
4
F
74
60
9
14
1
SOURCE: Adapted from Maltoni et al., 1986.
TABLE 3-10. INCIDENCE OF RENAL MEGANUCLEOCYTOSIS IN MALE
SPRAGUE-DAWLEY RATS EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI 5 DAYS/WEEK
FOR 108 WEEKS
Airborne concentration
(ppm)
Incidence of rats bearing lesions
600
300
100
0
101/130
22/130
0/130
0/130
SOURCE: Adapted from Maltoni et al., 1986.
3-19
-------
TABLE 3-11. INCIDENCE (%) OF TUMORS AMONG 90 SPRAGUE-DAWLEY
RATS OF EACH SEX EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK F0R 104 WEEKS (EXCEPT CONTROLS)*
Air concentration (ppm)/sex)
Parameter
Tumor-bearing rats
Mammary tumors
Pheochromocytomas
Pheochromobl astomas
Leukemias^
Ley dig cell tumors
Renal adenocarcinomas
600
M
72
20
24
3
10(0)
24
3
300
F
77
56
10
0
8(0)
1
M
i
! 61
18
19
1
12(3)
27
0
F
80
67
8
0
0(0)
~
0
100
M
60
20
16
6
11(4)
12
0
F
67
56
4
0
7(2)
.
0
0
M
73
19
32
1
6(1)
5
0
F
80
54
16
1
7(0)
0
aFor controls, 95 males and 105 females.
bNumbers in parentheses represent immunoblastic lymphosarcomas,
SOURCE: Adapted from Maltoni et al., 1986.
3-20
-------
TABLE 3-12. INCIDENCE (%) OF TUMORS AMONG 130 SPRAGUE-DAWLEY
RATS OF EACH SEX EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK FOR 104 WEEKS (EXCEPT CONTROLS)a
Air concentration (ppm)/sex
600
Parameter
Tumor-bearing rats
Mammary tumors
Pheoch romocytomas
Pheochromobl astomas
Leukemias^
Leydig cell tumors
Renal adenocar-
cinomas
M
72
22
22
3
12(2)
24
3
F
7,8
59
9
0
9(1)
__
1
300
M
63
18
21
2
11(3)
23
0
F
75
64
9
0
2(1)
_.
0
100
M
62
21
19
5
10(4)
12
0
F
68
57
5
1
7(3)
0
0
M
67
13
28
1
13(1)
4
0
F
82
60
22
1
3(0)
0
j^For controls, 135 males and 145 females.
bNumbers in parentheses represent immunoblastic lymphosarcomas,
SOURCE: Adapted from Maltoni et al., 1986.
3-21
-------
The investigators (Maltoni et al|., 1986) did note a slight increase in
I
leukemias, particularly of immunoblasjtic lymphosarcomas, and renal adenocar-
cinomas among the TCI-exposed male ra|ts. The researchers pointed out that
i
renal adenocarcinomas had not been observed in this strain of rats in their
colony. They claim a history of over| 50,000 rats. Leydig cell tumors were
I
increased in males of all exposure groups. One Leydig cell tumor had metas-
tasized to the heart.
3.1.5.3. Mouse Study; Inhalation Exposure Using essentially the same experi-
f
mental design as detailed for the Sprague-Dawley rat bioassays, Maltoni et al.
(1986) studied the effects of exposure to airborne concentrations of TCI on
Swiss mice from their laboratory and on B6C3F1 mice obtained from the National
Cancer Institute (NCI) or from the Charles River Laboratory (CRL). At the
j
start of exposure, the Swiss mice were 11 weeks of age and the B6C3F1 mice were
12 weeks of age. Four different minor variations in experimental design were
used, as indicated in Table 3-13. The BT306bis study was initiated because of
early deaths in the BT306 study among male mice due to fighting.
BT303/SwissResults of 8-week exposure: There was no change in survival
i
or mean body weight among the mice exposed to airborne concentrations of TCI
for 8 Weeks as compared to the controls. The total observation period was 134
i
weeks. The average body weight at 64 weeks was 44 g and 40 g for males and
females, respectively. '
!
The authors concluded that neither the number of tumor-bearing animals nor
the incidence of the most frequently expected tumors (mammary carcinomas,
leukemias, pulmonary tumors, and hepatomas) was increased by 8 weeks of expo-
sure to TCI at 600 or 100 ppm. However, a slight but nonsignificant increase
in hepatomas was evident in the TCI-exposed males. These incidences are given
in Table 3-14.
3-22
-------
TABLE 3-13. DOSING REGIMEN FOR INHALATION EXPOSURE OF MICE
Number of mice
Laboratory
designation/strain
BT303/Swiss
BT305/Swiss
BT306/B6C3F1 (NCI)
BT306bis/B6C3Fl (CRL)
Exposure Concentration
duration (weeks) (ppm)
8 600
100
0
78 600
300
100
0
78 600
300
100
0
78 600
300
100
0
M
72
60
100
90
90
90
90
90
90
90
90
90
90
90
90
F
72
60
100
90
90
90
90
90
90
90
90
0
0
0
0
SOURCE: Adapted from Mai torn" et al., 1986.
3-23
-------
TABLE 3-14. INCIDENCE (%) OF TUMORS AMONG SWISS MICE
EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK FOR 8 WEEKS
! TCI ai
600
Parameter
Tumor-bearing mice
Mammary carcinomas
Pulmonary tumors
Leukemias
Hepatomas
M
14
0
4
1
6
F
22
4
4
4
0
i
r concentration (ppm)/sex
100
M
18
0
12
0
5
F
22
3
7
8
2
0
M
9
0
5
1
1
F
22
4
7
8
1
SOURCE: Adapted from Maltoni et al.,' 1986.
3-24
-------
BT305/Swiss--Results of 78-week exposure: There was no change in survival
or mean body weight among the Swiss mice exposed to airborne concentrations of
TCI for 78 weeks as compared to the controls. The average body weight at 64
weeks of age was about 47 g and 40 g for males and females, respectively. The
total observation period was 145 weeks.
The neoplastic incidences of interest in the Maltoni et al. (1986) study
are given in Table 3-15.
There was an increase in hepatomas and pulmonary tumors among male mice.
Maltoni et al. (1986) indicated that statistical significance was only at the
high- and mid-exposure concentrations for the lung lesions and only at the high
concentration for the liver lesions. The authors further classified the pul-
monary tumors as adenomatous hyperplasia-early adenomas, adenomas, or adenocar-
cinomas. The percentages of mice affected in these sub-classifications are
included in Table 3-15.
BT306/B6C3F1(NCI) Results of 78-week exposure: There was no change in
survival or mean body weight among the mice exposed to airborne concentrations
of TCI for 78 weeks as compared to the controls. The average body weight at 64
weeks of age was about 35 g^ and 32 g for males and females, respectively. The
total observation period was 154 weeks.
The neoplastic incidences, as indicated by Maltoni et al. (1986), are
given in Table 3-16.
There was an increase in hepatomas and pulmonary tumors. The authors
(Maltoni et al., 1986) further classified the pulmonary tumors as adenomatous
hyperplasia-early adenomas, adenomas, or adenocarcinomas. The percentage of
mice affected in this sub-classification is also shown in Table 3-16. The
authors indicated that when taken together there was a slight increase in
tumor-bearing female mice. The increase in total number of malignant tumors
3-25
-------
TABLE 3-15. INCIDENCE (%) OF TUMORS AMONG SWISS MICE
EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK FOR 78 WEEKS
iTCI air
600 !
Tumor site
Tumor-bearing mice
Mammary carcinomas
Leukemias
Hepatomas
Pulmonary tumors
Adenomatous
hyperplasia-
early adenomas
Adenomas
Adenocarcinomas
M
29
0
6
14
30
17
12
1
F
60
16
28
' 1
22
4
16
2
concentration (ppm)/sex
300
M
30
0
1
9
26
12
13
0
F
51
19
17
0
14
1
13
0
100
M
19
0
4
2
12
7
6
0
F
48
11
17
0
17
3
13
0
0
M
26
0
3
4
11
0
11
0
F
60
14
23
0
17
0
14
2
SOURCE: Adapted from Maltoni et al.,|1986.
3-26
-------
TABLE 3-16. INCIDENCE (%) OF TUMORS AMONG B6C3F1 (NCI) MICE
EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK FOR 78 WEEKS
Air concent rat i
; 600
Tumor site
Tumor-bearing mice
Mammary carcinomas
Leukemias
Hepatomas
Pulmonary tumors
Adenomatous
hyperplasia-
early adenomas
Adenomas
Adenocarci nomas
M
9
0
1
7
1
0
1
0
F
77
7
42
10
17
1
16
0
300
M
6
0
o
3
3
1
2
0
F
67
4
40
4
8
0
8
0
on (ppm)/sex
100
M
4
2
1
1
2
0
2
d
F
64
7
37
4
7
0
6
1
0
M
12
0
7
1
2
1
1
0
F
57
2
37
3
4
2
2
0
SOURCE: Maltoni et al.t 1986.
3-27
-------
among the females was statistically significant at all exposure concentrations.
There was an increase in hepatomas among the males and females of the highest
exposure group. Pulmonary tumors werb increased significantly among females of
the high-exposure group.
BT306bis/B6C3Fl (CRL) malesResjilts of 78-week exposure: There was no
change in survival or mean body weighjt among the mice exposed to airborne
concentrations of TCI for 78 weeks asj compared to the controls. The average
body weight at 64 weeks of age was ab|out 34 g. The total observation period
was 135 weeks.
The neoplastic incidences discusjsed in detail by Maltoni et al. (1986) are
given in Table 3-17. The tumor responses across doses are not elevated com-
pared to controls.
3.2. EPIDEMIOLOGIC STUDIES
Under the sponsorship of Warner Brake and Clutch Company, Shindell and
Ulrich (1985) reported the results of a cohort study in a manufacturing plant
that used TCI as a degreasing agent. The plant began operations in 1957; the
exact date was not reported. When the plant began operation, all manufacturing
operations that had formerly been conducted in another facility several miles
away were transferred to the new plant. Production workers were included in
the cohort if they worked 3 months in either facility between January 1, 1957
and July 31, 1983. Office workers were included if they worked 3 months or
more at the new plant between its opening and July 31, 1983.
The authors stated that "while tihe use of trichloroethylene in the plant
is under controlled conditions (i.e., in degreasing machines designed to con-
trol the vapors) small amounts of this highly volatile substance tend to escape
into the atmosphere. During the reclamation process and while maintenance
activities were being performed on these machines, additional possibilities
3-28
-------
TABLE 3-17. INCIDENCE (%) OF TUMORS AMONG MALE B6C3F1 (CRL) MICE
EXPOSED TO AIRBORNE CONCENTRATIONS OF TCI
7 HOURS/DAY, 5 DAYS/WEEK FOR 78 WEEKS
Tumor site
Tumor-bearing mice
Leukemias
Hepatomas
Pulmonary tumors
Adenomatous
hyperplasia-
early adenomas
Adenomas
Adenocarcinomas
600
40
7
23
10
2
7
1
Air concentration
300
51
7
30
13
6
8
o
(ppm)
100
33
11
21
10
3
7
0
0
57
13
19
18
4
13
0
SOURCE: Adapted from Maltoni et al., 1986.
3-29
-------
existed for exposure of workers engageid in these activities." The authors fur-
i
ther stated that quantification of the1 amount of the exposure for the entire
period of the plant was not possible, but that monitoring of the levels to which
"current employees" have been exposed revealed conformance to the Occupational
Safety and Health Administration (OSHA|) standards. The OSHA Permissible Expo*
sure Limits (PEL) adopted in 1971 for TCI are 100 ppm as an 8-hour time-weighted
average, 200 ppm as a ceiling, and 300 ppm as a maximum peak (5 minutes in any
3 hours). The authors also stated that the workers drank water at the plant
containing 43 ppb of TCI, which they stated is comparable to levels of TCI
contamination of public water supplies.
The cohort of 2,646 included 2,140 white males, 76 nonwhite males, and
430 females (the females were not specified by race). As indicated above, the
cohort included both production workers and office employees; however, the
authors did not indicate how many of each were included. There were 16,332
person-years of employment and 38,052 person-years of follow-up. The authors
did not indicate at what point follow-up began. By the end of the follow-up
period on July 31, 1983, 618 persons Were employed and 2,028 were former
\
employees of whom 141 were determined to have died. Of the 1,887 not known
to be deceased, 52 could not be confirmed as alive or dead.
i
Overall, the vital status of 98.03% of the cohort could be accounted for
by the end of the study period. The 52 individuals whose status could not b£
determined during the study were reported to be "largely short-term employees
with 1.4 years of employment on the average and generally were relatively young
persons at the time the contact was attempted (x = 41.7 years)."
Expected mortality for the group!was calculated based on the national
mortality experience. There were fewer deaths than expected overall, and there
was a significant deficit among white males (131 observed, 165.5 expected, p <
3-30
-------
0.01). Also, nonrespiratory cancer deaths (21 observed, 38.5 expected, p <
0.05) were significantly less than expected among white males. The employees
having the greatest opportunity for occupational exposure to TCI were the
"assemblers" whose mortality was reported to generally have conformed to the
expected values for all cancers." The size of the assembler group was not
reported.
The 2,453 employees known to be alive were contacted and asked to report
any medical condition for which they currently (or periodically) required
medical treatment. Such information was obtained from 2,188 of the 2,453
employees. Of those responding, 84.3% reported no health problems, which, the
authors claimed, was a more favorable morbidity status than was found in a
reference group of some 9,500 employees of eight industrial plants studied
between 1978 and 1983. The authors did not cite a reference for the survey
of 9,500 employees. The primary reason for the better morbidity status of the
TCI-exposed group was the less-than-expected incidence of hypertension and
cardiovascular disease. These two conditions accounted for only 3.6% of the
reported conditions, whereas normally, the authors reported, they would be
expected to account for 10% to 15%. The authors did not indicate what health
problems were found among the 15.7% who did report health problems.
This study examined the mortality and morbidity experience of an indus-
trial cohort exposed to TCI, and did not find any cause-specific excess mortal**
ity. ; It is unknown to what extent the workers were exposed, however. Office
employees were included in the cohort along with production workers. Further-
more, the cohort included anyone who had worked at the manufacturing plant for
as little as 3 months. Although one can calculate from the data presented that
the average duration of exposure was 6.17 years, one does not know how the
duration of exposure was distributed among the cohort. Similarly, although one
3-31 ,
-------
can calculate from the data presented iby the authors that the average duration
of follow-up was 14.4 years, one does jnot know how the length of follow-up was
distributed. Thus, despite the fact t|hat the cohort was relatively large
(2,646 persons), it is possible that a sizable proportion of the cohort was
composed of persons with little or no ^exposure and/or who had worked a rela-
tively short time and/or had a relatively short follow-up.
i
It is also possible that a sizable proportion of workers may have moved
I
from the previous plant to the new plant when it opened in 1957. Thus, there
may have been a survival effect (Fox and Collier, 1976) since workers from the
old plant would have had to survive to 1957, the beginning date of employment
at the new plant, to have entered into the cohort. This survival effect would
have underestimated the true mortality risk, if any, due to TCI.
j
In conclusion, this study does not demonstrate elevated cancer mortality
among a group of workers reported to be exposed to TCI. The data, however,
are inadequate to fully evaluate the observed absence of an effect. Thus, this
study does not have any additional interpretive value to the human studies data
base contained in the 1985 HAD on TCI *
3.3. GENOTOXICITY STUDIES
The NTP (1987) report included m(itagenic evaluations using several proto-
cols, as follows:
Ames test (plate assay - four strains without S9).
Ames test (plate assay - four strains with rat S9)
l
Ames test (plate assay - four strains with hamster S9)
Mouse lymphoma cells without S9
Mouse lymphoma cells with S9
! 3-32
-------
Chromosomal aberrations in Chinese hamster ovary cells
Chromosomal exchange in Chinese hamster ovary cells
All assays were negative except for the sister chromatid exchange, which
was considered equivocal, and the mouse lymphoma cells with Aroclor-induced
male rat liver S9, which was considered positive.
3.4. DATA INTERPRETATION: DISCUSSION
Neither the epidemiologic study by Shindell and Ulrich (1985) nor the NTP
(1987) genotoxicity studies provide a basis for changing the TCI weight-of-evi-
dence classification for caretnogenicity. According to the NTP, the study
(NTP, 1987) of the effects of TCI on four strains of rats (Marshall, Osborne-
Mendel, AC1, August) is not acceptable for assessing carcinogenic potential
because of reduced survival, toxicity, and deficiencies in the conduct of the
study, even though increased incidences of renal tubular cell adenomas and
adenocarcinomas were observed. The study did show, however, a consistent toxic
response in the kidneys.
As discussed in the 1985 HAD, some scientists consider the rare renal
cancer response in the F344 rat (NTP, 1986a) to be an indication of carcino-
genicity induced by TCI. Indeed, Dekant et al. (1986) recently identified a
minor Urinary metabolite (N-acetyl-DCVC), the precursor (DCVC) for which, they
irtdicate, may have a role in the induction of renal cancer.
The corn oil gavage study in mice by Henschler et al. (1984) was not con-
ducted in accordance with good laboratory practices. The maximum tolerated
dose was exceeded. While the liver tumors were clearly elevated among all
TCI->treated groups, even the highest incidence (9/49) as compared to the con-
trols had a p value of only 0.06. Perhaps the duration of treatment was not
sufficient. This study can only be considered suggestive with regard to any
3-33
-------
carcinogenic effect.
Maltoni et al. (1986) reported a slight increase in immunoblastic lympho-
sarcomas, a form of leukemia, in male Sprague-Dawley rats gavaged with TCI at
doses of 50 and 250 mg/kg for 52 weeks;and held until death. The less-than-
lifetime dosing in this study may not have been adequate for optimum bioassay
i
SensUi'V'ity, although the observation period was adequate to allow for tumdf
expression. The maximum tolerated dosJ2 (MTD) was probably not reached unless
one considers the noncarcinogenic renajl lesions, which did not affect survival,
i
to be indicative of an MTD. On the other hand, 500 mg/kg in the four strains
of rats tested by NTP appeared to be a'maximum tolerated oral dose. Despite
the nonstatistical nature of the leukemia response, it should not be totally
dismissed, since a similar finding wasi made in the inhalation bioassay of
Maltoni et al. (1986), and also becausfe of the finding of malignant lymphomas
in female HanrNMRI mice, as detailed in the 1985 EPA TCI assessment. In addi-
tion, the NTP (1986b) reported that lebkemia and renal cancer among male rats
exposed by inhalation to perch!oroethylene (PERC), a related chlorinated sol-
vent, constituted clear evidence of carcinogenicity.
Maltoni et al. (1986) reported a [slight increase in pituitary adenomatous
growth among the rats in the BT304 inhjalation bioassay, although this finding
was not confirmed in experiment BT304b|is. When the two 104-week rat studies
(BT304, BT304bis) were combined, as Malltoni et al. chose to do, the noteworthy
incidences are as shown in Table 3-18.| Because of a lack of time-to-tumor
data, although the average latencies do not suggest a marked reduction, the
statistical analysis (Fisher Exact Test) used the number of animals at risk at
the time of the first tumor. This procedure was used to eliminate any problems
that led to early deaths related to the beginning of the study. The adjustments
were minimal. This analysis did not confirm statistical differences (p = 0.05
3-34
-------
TABLE 3-18. ANALYSIS OF TUMORS AMONG SPRAGUE-DAWLEY RATS EXPOSED
TO AIRBORNE CONCENTRATIONS OF TCI 7 HOURS/DAY, 5 DAYS/WEEK, FROM THE
STUDIES (BT304, BT304bis) OF MALTONI ET AL. (1986)
Air concentration (ppm)
Parameter
Males
Leydig cell tumors
Leukemias
Immunoblastic lympho-
sarcomas (a kind of
leukemia)
Renal adenocarcinomas
Adenomas
Combined
Females
Leukemias
IfntnUhoblastic lympho-
Saf comas (a kind of
leukemia)
Renal adenocarcinomas
Adenomas
Combined
600
^
31/1293
p=0. 00003
(110)b
15/129
p=Q.117
2
(92.6)
4/129
1
5
p=0.026
(115)
11/130
p=0.16
(84.1)
1
1/130
0
1
(123)
300
30/130
p=0. 00006
(113.2)
14/130
4
(101.8)
0/130
0
0
2/130
(101)
1
0/130
0
. 0
100
16/130
p =0.01 7
(109.7)
13/130
5
p=0.09
(100.5)
0/130
1
1
(84)
9/130
p=0.314
(100)
4
p=0.48
0/130
I
(116)
0
6/135
(113.2)
9/135
p=0.17
(trend)
1
(72.4)
0/135
0
0
7/145
(92.4)
0
0/145
0
0
^Tumor-bearing rats/rats alive at the first tumor.
"Average latency (weeks).
3-35
-------
or less) in any TCI exposure groups with regard to leukemiias.
The increase in renal tumors among the (Maltoni et al., 1986) TCI high-
exposure Sprague-Dawley male rats was statistically significant (p = 0.026).
The authors pointed out that renal adenocarcinomas had not been previously
observed in this strain of rats in their colony. The EPA Guidelines permit the
use of rare tumor data as reason to consider a higher weight-of-evidence clas-
sification. The World Health Organization (WHO, 1985) considered this response
along with the renal cancer response following oral administration in the F344
rat as reported by NTP (1986a), to be "some" evidence of carcinogenicity.
Both the report from the NTP on the chronic toxicity of TCI in four strains
of rats and that from Maltoni et al. (1986) confirm the causal relationship be-
tween toxic nephrosis and TCI exposure. The lesions described by Maltoni et al.
(1986) were judged to be the same lesions grouped by the NTP (1987) as toxic
nephrosis. However, these latter investigators (Maltoni et al., 1986) noted a
lower frequency and did not observe the lesion in the females, perhaps because
of the shorter duration of treatment. The NTP noted that TCI had also produced
the same type of changes in the B6C3F1 mouse (NTP, 1986a) and in an earlier
study of Osborne-Mendel rats (NCI, 1976). They also noted that the identical
lesibtl occurred (NTP, 1986b) as a result of exposure to airborne concentrations
of PERC in rats and mice, as well as after gavage exposure to pentachloroethane
(PENTA) in rats and mice (Mennear et al., 1982; NTP, 1983). The high incidence
of this type of toxic nephrosis and the low or nonexistent incidence of renal
cancer are not consistent with the hypothesis offered by some individuals that
a causal relationship exists between this toxicity and renal cancer,,
Following inhalation exposure, Leydig cell tumors were significantly in-
creased among male Sprague-Dawley rats of all exposure groups (Maltoni et al.,
1986). While this finding alone is clear evidence of carcinogenic activity, it
3-36
-------
is further supported by the NTP finding of an increased incidence of intersti-
tial cell tumors of the testis (the same tumor type, but a different nomencla-
ture) as observed in high-dose male Marshall rats (unt-16/46, veh-17/46, 500
mg/kg-21/48, 1,000 mg/kg-32/48).
The cumulative evidence regarding the carcinogenic potential of TCI in
rats is judged to be sufficient, to indicate that TCI produces cancer (leu-
kemias, renal carcinomas, and Leydig cell tumors) among male rats of up to
three strains following chronic oral and inhalation exposure.
In Swiss mice exposed to TCI by inhalation for 78 weeks by Maltoni et al.
(1986), the incidence of statistically significant tumors was as shown in Table
3-19. The increased hepatomas among the male Swiss mice is in keeping with
similar lesions reported following oral administration of TCI or oral and
inhalation exposure of related compounds. The pulmonary tumors, as well as the
hepatomas, were also observed by Maltoni et al. (1986) in B6C3F1 mice.
In B6C3F1 mice exposed to TCI by inhalation for 78 weeks, the incidence of
statistically significant tumors was as shown in Table 3-20. In female B6C3F1
mice, the incidence of hepatomas and pulmonary tumors was increased in a dose-
related manner; the high dose-response being marginally statistically signifi-
cant at p = 0.06. The response among the males was not statistically signifi-
cant with regard to liver tumors except when the bioassays are combined in the
300 ppm group. These two bioassays were not combined in Maltoni's analysis.
The incidence of liver tumors in the BT306 study was unusually low for this
strain of mouse, and the marked difference in the control response may be the
result of different sources of test animals. It was judged by Maltoni et al.
that the response in the males of this strain should not be considered suit-
able for assessment of risk rate, possibly because the BT306bis study had such
a random tumor occurrence compared to BT306. The mouse studies in which the
3-37
-------
TABLE 3-19. ANALYSIS OF STATISTICALLY SIGNIFICANT TUMOR INCIDENCE
AMONG SWISS MICE EXPOSED BY MALTONI ET AL. (1986)
TO AIRBORNE CONCENTRATIONS OF |TCI FOR 7 HOURS/DAY, 5 DAYS/WEEK
FOR 78 WEEKS
Air concentration (ppm)
600
300
100
Hales
Hepatomas
Pulmonary tumors
Females
Hepatomas
Pulmonary tumors
13/90a
p=0.02
(72.1)b
27/90
p=0.002
(73.7)
1/89
(85)
20/89
p>0.05
(94.4)
8/89
p>0.05
(69.9)
23/89
p=0.01
(67.9)
0/89
13/89
(102.4)
2/89
(69)
11/89
(85.4)
0/89
15/89
(102.5)
4/88
(74)
10/88
(80.9)
0/90
15/90
(97.1)
^Tumor-bearing mice/mice alive at the first tumor.
"Average latency (weeks).
3-38
-------
TABLE 3-20. ANALYSIS OF STATISTICALLY SIGNIFICANT TUMOR INCIDENCE
AMONG B6C3F1 MICE EXPOSED BY MALTONI ET AL. (1986)
TO AIRBORNE CONCENTRATIONS OF TCI FOR 7 HOURS/DAY, 5 DAYS/WEEK
: FOR 78 WEEKS
Air concentration (ppm)
600
300
^Tumor-bearing mice/mice alive at the first tumor.
b Average latency (weeks).
100
Males
Hepatomas
306bis
306
Females
Hepatomas
Pulmonary tumors
27/178a
p-0.1
(72.1)b
21/90
(110. 3)b
6/88
p=0.06
(71.2)
9/87
p=0.058
. (107.3)
14/87
p=0.001
(111.3)
30/178
p=0.0496
(69.9)
27/90
(103)
3/88
(63.3)
4/89
(111.2)
7/89
p=0.08
(112.6)
20/175
(69)
19/89
(117.8)
1/86
(85.5)
4/90
(108.5)
6/90
p=0.13
(119)
,
18/175
(74)
17/90
(113)
1/85
(81)
3/90
(102.3)
2/90
(110)
3-39
-------
inhalation exposure duration was 78 weeks are judged by EPA to be sufficient to
demonstrate that TCI produces a carcinogenic response in mice by inhalation
exposure.
The finding of TCI-induced lung tumors in female ICR mice by Fukuda et al .
i
(1983) supports the similar finding by[ Maltoni et al . (1986). The incidence
x
reported by Fukuda et al . is shown in [Table 3-21. The findings of Fukuda et
i
al . (1983) and Maltoni et al . (1986) are remarkably similar, considering the
r
different duration of exposure, differing mouse strains, and the possibility of
different pathologic interpretations. The Maltoni pathologic diagnosis shows
mostly benign tumor increases, while the Fukuda results report mostly carcino-
mas as well as benign lesions.
The recent reports of Herren-Freund et al . (1986, 1987) are important in
that known TCI metabolites (TCA and DGA), at present judged to be nongenotoxic,
nevertheless produce liver carcinomas in mice, according to the authors, by
acting as complete carcinogens. One of the unusual features of this investiga-
tion is the very young age at which the treatment with TCA and DCA was started.
The appearance of the cancers within such a short period of time and at a rela-
tively high rate suggests that young animals may be much more sensitive than
Older dfies. Additionally, some might jsuggest that the appearance of these c'att*
cer-inducing agents (TCA and DCA) in tihe urine of humans should be considered
prima facie evidence of human risk. This consideration is supported by the
.
work of Bergman (1982, 1983a), who indicated l,N6-ethanoadenine,, a DNA adduct
in mouse liver resulting from TCI exposure (probably as a result of metabo-
lites) to be one of the same adducts associated with exposure to monochloro-
ethylene (vinyl chloride), a structurally related known human carcinogen.
Six human studies related to the jissue of TCI exposure and cancer were
reviewed in the HAD, and one additional study is critically reviewed in this
13-40
-------
TABLE 3-21. ANALYSIS ,OF THE INCIDENCE OF LUNG TUMORS
IN FEMALE ICR MICE EXPOSED BY FUKUDA ET AL. (1983)
TO AIRBORNE CONCENTRATIONS OF TCI FOR 7 HOURS/DAY, 5 DAYS/WEEK
FOR 107 WEEKS
Mice with tumors/mice examined
Concentration
(ppm) Adenocarcinomas Combined
450 7/46 (p = 0.024) 11/46 (p =0.10)
150 8/50 (p =0.026) 13/50 (p =0.06)
50 3/50 (p = 0.31) 5/50
0 1/49 6/49
3-41
-------
document. These studies suffered from one or more of the following deficien-
cies: small sample size, lack of analyisis by'tumor site, problems with expo-
sure definition, and problems with length of exposure. According to EPA's
Carcinogen Risk Assessment Guidelines, the epidemiologic data are judged
inadequate for demonstrating or refuting the carcinogenic potential of TCI.
The mutagenic data presented in this addendum do not suggest a change in the
characterization of TCI's mutagenic activity as presented in the 1985 HAD. The
cumulative information regarding the carcinogenic response in rats provides
sufficient evidence based on multiple-jsite tumor induction by two routes of
exposure. The new mouse data presente|d in this addendum support the 1985
finding of a positive liver hepatocellular carcinoma and hepatoma response by
oral exposure, because liver tumors were observed following exposure to air-
i
borne concentrations in Swiss as well as B6C3F1 mice. In addition, the mouse
inhalation data in three strains from two laboratories provides another tumor
site, the lung. Thus, based on EPA's cancer guidelines, the cumulative evi-
dence for TCI remains in classification B2, i.e., a probable human carcinogen.
The 1985 HAD B2 classification is markedly strengthened with this addendum by
virtue of a second exposure route (inhalation) showing carcinogenic activity in
rats and mice and the diverse tumor sites identified.
In terms of the weight-of-evidence classification and the ultimate risk
characterization, it is noteworthy tha!t two of TCI's metabolites have been
shown by a recent EPA-sponsored study to act as complete carcinogens in mice
and have also been identified in the urine of TCI-exposed humans. Furthermore,
it has been shown that TCI-exposed mice develop a DNA adduct in the liver which
is identical to one induced by vinyl chloride, a known human carcinogen. If
i
the epidemiologic data base were more substantial, and otherwise capable of
i
detecting a relatively low-potency carcinogenic effect such as is possibly pro-
3-42
-------
vided by TCI exposure, the direct human evidence could be used to demonstrate
or refute more clearly the likelihood of TCI's being a human carcinogen.
3-43
-------
-------
4. ONCODYNAMICS
The estimation of cancer risk related to chemical exposure involves the
mathematical modeling of generally high experimental doses and corresponding
observed animal responses to estimate the risk (probability of response minus
background) at levels lower than the experimental doses. The doses and response
data are fitted to a curve and., by extrapolation, the additional risk at a
lower dose is estimated. The "dose" used in these calculations is assumed to
be the effective dose. This may vary from the concentration of the material in
the environment or the amount of the chemical administered to the amount of the
reactive agent(s) reaching the target tissues, the latter being a better esti-
mate of the actual effective dose.
Regarding cancer risks, special consideration should be given to the dose
of reactive molecule(s) involved with tumor initiation, promotion, and progres-
sion or other mechanisms or stages. Consideration of genotoxicity, in vivo DNA
binding, toxicity, and the toxicology and pharmacokinetics of related compounds
can be used in selecting the appropriate end point and reactive chemical spe-
cies on which to base further evaluation.
The doses are scaled across species to account for differences in size,
fteMbolic rate, and other variables. Evidence supporting scaling using body
weight to the 2/3 power for metabolic variations in the case of TCI was pro-
vided in the 1985 HAD, and is in agreement with the recent review of Davidson
et al, (1986) on allometric methods for scaling across mammalian species.
In many instances, differences in the duration of exposures and the per-
cent of lifetime exposed should be considered. There is little experience in
dealing with other than lifetime risks when dichotomous data are used. Some-
times the less-than-lifetime exposure is adjusted for by use of the ratio of
4-1
-------
the intended time of the study (lifetime) divided by the actual duration of
the study to the third power (U.S. EPA, 1985b). For example, if the test
animals are killed after 61 weeks of la lifetime (104 weeks) study, the esti-
i ,
mated risk for a given dose at termination (61 weeks) would be adjusted by
4.96 (104/61^). An alternative is to adjust the risk proportionally for the
fraction of lifetime. With the exposure concentrations producing a lifetime
risk of 1/1,000, half of the risk from lifetime exposure would be 1/2,000. It
i
should be noted that in upward concaye dose-response curves the resultant risk
will be less than if the time-scaling factor were applied to dose before sub-
stitution into the model, if the factor is greater than 1. Normally, one
considers that a rodent inhalation exposure of 6 hours/day, 5 days/week for 2
years is approximately equal to a lifetime occupational exposure, and that
daily exposures for 2 years by other1routes are equivalent to a human lifetime.
Recently, in most instances, th6 EPA, in order to estimate lifetime con-
tinuqus exposure risk, has adjusted the dose to account for doses applied 5
days/week, instead of 7 days/week, and adjusted the risk if the observation
period was less than a full lifetime* Two years (at a minimum, 90 weeks) is
generally considered to be a sufficient observation period to approximate a
rodent lifetime.
In addition, adjustments in dose calculations because of dose-dependent
pharmacokinetics may be required when extrapolating from high experimental
i
exposures to the lower exposures of interest. These exposures also are some-
times of a different dosing regimen.
4.1. SELECTION OF EFFECTIVE DOSE
The selection of the dose to be used in the assessment of risk should be
based on the use of pharmacokinetic, biochemical, and toxicologic information
to determine, when possible, the cancer-causing moiety. With some compounds,
i
! 4-2
-------
the agent itself may be a complete carcinogen, while with other compounds a
metabolite may be the reactive molecule. It is also possible that some com-
bination of metabolites may be responsible for the induction of cancer.
In addition to the identification of the active agent(s), it may also be
possible to incorporate the time of residence in the blood or tissue for the
active agent(s); when this is done, the integrated dose (area under the con-
centration curve with time) may be used. If compounds have long-half times,
it may be appropriate to consider the steady-state level or the average body
concentration over time. Thus, the selection of the effective dose involves
identification of the active agent(s) as well as consideration of the dose con-
figuration in terms of impact on the target tissues.
4.1.1. Identification of the Active Agent(s)
It is generally considered that TCI metabolite(s) are the active .cancer-
causing agents. This is supported by short-term tests involving some of the
intermediate and end metabolites, as well as a cancer bioassay using two of the
metabolites.
The assessment of risk rate 1n the HAD by the oral route was based on the
use of response data from gavage studies in mice showing an increase in liver
neoplasia. The doses used in the assessment were based on the amount of the
SdftHrHsleTed dose absorbed and metabolized to the active chemical species.
Barinerjee and Van Duuren (1978) considered covalent binding of TCI, although
slight, to be consistent with the formation and alkylating activity of TCI-
epoxide, the transient epoxide (tl/2 =3.5 minutes). All TCI would be metab-
olized through the epoxide, if the scheme proposed by Dekant et al. (1984) were
correct.
Elcombe et al., (1982) proposed that TCA is the active agent associated
with peroxisome proliferation and cell proliferation associated with exposure
4-3
-------
to TCI. Elcombe et al. (1985) suggested that peroxlsome proliferation and
cell proliferation may play a role in! the development of liver cancer in mice.
At 1,000 mg/kg of TCI for 10 days, thfey found that the peroxisome activity (%
i
cytoplasmic volume) was 130% of the cbntrol in rats and 1,116% in mice. Doses
of 1,500 mg/kg to mice, however, only; produced an increase of 1,066%. Others
have indicated that the perioxisome proliferation is an example of an epigenetic
mechanism. Elcombe et al. (1982) indicated that the greater increase in perox-
isome proliferation in the mice as.compared to the rats is due to the faster
formation of TCA in the mouse. These investigators (Elcombe et al., 1985)
stated that "the relevance, of these observations to the possible human health
hazard is unclear." They suggested, however, that TCI does not present a
significant hepatocarcinogenic hazard to humans since most nonrodent species
are nonresponsive (or at least less susceptible) to peroxisome proliferators.
However, Reddy et al. (1980) reported an increase in peroxisomes among primates
after administration of hypolipidemic: agents.
Reddy et al. (1980) first suggested a correlation between peroxisome
proliferation and hepatocellular tumors in rodents. They postulated that
increases in catalase and peroxisomalj hydrogen peroxide-generating oxidase
cause an increase in the steady-state concentration of intracellular hydrogen
peroxide. This or other reactive oxylgen species leads to DNA damage, and thus
to a mutagenic change and cancer.
El combe et al. (1985) postulated that the increased cellular proliferation
observed after TCI administration may lead to promotion of "reactive oxygen-
Initiated" cells. Grisham et al. (1984) pointed out, however, that alkylating
agents increase cell turnover. Thus, TCI-epoxide, not TCA alone, formed in.the
i s ,
liver could play a role in the development of the tumors from the "reactive
oxygen-initiated" cells by increasing cell turnover. This possibility is sup-
4-4
-------
ported by the recent findings of Goldsworthy and Popp (1987) of CUT, who
reported that TCI is more potent than TCA in increasing cyanide-insensitive
pal mi toy! CoA activity (PCO), which they used as a measure of peroxisome pro-
liferation response.
These authors (Goldsworthy and Popp, 1987) also reported an increase in
rat-liver peroxisome proliferation activity (1.8- to 3.2-fold), in contrast to
the work of Elcombe et al . (1985), who failed to find such a large increase
in rats;. Goldsworthy and Popp (1987) also reported an increase in rat renal
PCO activity following administration of TCI, but not after administration
of tetrachloroethylene (PERC) and pentachloroethane (PENTA). With regard to
mouse liver tumors, these investigators (Goldsworthy and Popp, 1987) indicated
that while an association may exist between peroxisome proliferation and the
development of TCI- and PERC-induced hepatocellular tumors, the weak response
observed following administration of PENTA suggests that other factors are
involved.
The World Health Organization (1985) recommended further research on the
role of TCA-peroxisome proliferation, which they indicated has been implicated
in the induction of hepatocellular carcinomas in mice and rats by an "epigene-
tic" mechanism. The classification of the mechanistic mode of TCA-peroxisome
proliferation is still in doubt. Williams and Weisburger (1986) did not con-
sider peroxisome proliferators as either genotoxic or epigenetic, but referred
to them as unclassified. However, these authors indicated that if the forma-
tion of reactive oxygen species that have the capability to damage DNA and
initiate cancer were substantiated, they would classify such agents as epi-
Dekant et al . (1986) identified S-l,2-dichlorovinyl-N-acetyl-cysteine
(N-acetyl-DCVC), a minor metabolite of TCI, in the urine of rats. They indi
4-5
-------
cated that this finding establishes the existence, besides the well-known P-450
oxidative metabolism, of a novel pathway of biotransformation. This pathway,
as well as the classical pathway through the epoxide, is shown in Figure 4-1.
Dekant et al. (1986) indicated that jbhe first step in the formation of this
urinary metabolite probably takes place in the liver and is conjugated with
gluthathione. They indicated that a precursor, dichlorovinyl-cysteine (DCVCr),
t
I
of the urinary metabolite exerts nephrotoxic as well as genotoxic effects. In
.
renal tubuli, the C-S bond is cleaved by enzymes, resulting in the reformation
of DCVC. This process is similar to the sequence of events leading to nephro-
toxicity postulated by Rush et al. (1984) for several halogenated hydrocarbons.
The nature of the relationship between the metabolites and the observed
Leydig cell tumor response in rats is not clear. Since adverse effects in
the testis of mice have been reported after exposure to commercial-grade TCI,
the rat tumors might be a direct effect of one or more metabolites. On the
other hand, it could be suggested that the TCI metabolite, TCES may have a CNS
effect through hypothalamic stimulation on Leydig cell tumor development.
The recent work of Herren-Freun|d et al. (1986, 1987) indicates that the
f
metabolites, TCA and DCA, are both carcinogens. This finding, in addition to
the previously discussed information, supports the use of the total amount of
, I
TGt metabolized (TTCIM) as the effective dose in the cancer risk assess-
ment of TCI. It is impossible to identify only one active metabolite as the
carcinogenic agent. On the cellular level, it is probable that some combina-
tion of the products of TCI metabolism (including TCI-epoxide, TCA [in part by
i
peroxisome proliferation], DCA, or DCVC) may produce genotoxic effects as well
as changes considered to be consistent with non-DNA-reactive effects (including
the facilitation of DNA changes, the increase in DNA-altered cells, and the
progression, influenced by immunosuppression, of these cells to cancer). The
4-6
-------
ft u
o o
O
>* §
u
I
>-
Iti
o
<
o
0
4-7
-------
existence of multiple reactive metabolites having several probable cellular
activities related to cancer induction is consistent with the multistage theory
of cancer development.
4.1.2. Dose of the Active Agent(s)
i
Since the clearance of the metabolites in experimental animals is complete
within the non-dosing period in the treatment regimen, a buildup is unlikely to
occur, as indicated in the HAD. Therefore, the modeling of dose on the basis
of average concentration or steady state is inappropriate. Because several
i
responses will be used in the assessment of the risk of TCI from airborne ex-
i
posure, and because metabolite formation is likely to occur in several organs,
physiologic compartmental modeling seems impossible given the lack of organ-
i
specific metabolite data which may include a synergistic role for metabolites.
Thus, the total amount metabolized resulting from a single daily exposure is
i
the best configuration to use in effective dose estimation.
i
Recently, the National Research Council (NRC, 1986) examined the pharmaco-
I
kinetics of TCI with regard to the intake via water and by inhalation of vari-
ous concentrations in rats, and the resulting AMEFF (effective concentration
of reactive metabolite formed in a compartment of specific volume).; Their
simulation was based on in vitro data. They imply that the AMEFF derived from"
inhalation studies may be useful in determining the risk in drinking water.
The different cancer sites by the oral as compared to the inhalation exposure
route, detailed previously, suggest that this may not be the case. In addi-
tion, their simulation suggests that single oral doses would produce a greater
AMEFF in both rats and humans than repeated smaller doses. The latter would be
similar to the animal and human situation when a toxicant is in the drinking
water.
In a paper soon to be pub!ished,i Koizumi et al. (1987) examined these phe-
4-8
-------
nomena using data they generated with regard to oral intake of TCI and total
Metabolite formed. These authors used the data from Stott et al. (1982) for
their comparison to inhalation exposure. Based on modeling using Michaelis-
Menten type constants that fit the inhalation data of Stott et al. (1982),
these authors indicated that the inhalation data would be useful in modeling
by summing the various compartments to estimate the total amount of TCI-metab-
olite formed as the result of oral administration. Contrary to the impression
left by the NRC, the modeling and data of Koizumi et al. (1987) confirmed that
gavage dosing or intake in the drinking water at approximately equal doses
produced the same percent metabolized. Furthermore, Koizumi et al. (1987)
concluded that if the metabolism is not saturated, the metabolic fate of TCI is
independent of the route of administration, and that toxicity data can be used
to predict the potential hazard of low doses of TCI administered in the drink-
ing water.
4.2. ESTIMATION OF EFFECTIVE DOSE IN HUMANS
Generally, either human pharmacokinetic data are not available for use
in risk analysis or the available data are sparse and not very useful, fre-
quently dealing only with the parent compound and not the reactive metabolites.
Such is not the case with TCI.
In order to estimate the human risk, the amount of TTCIM formed as a
result of various airborne concentration needs to be estimated. The estima-
tion of the amount of a reactive metabolite formed (referred to as TTCIM) as
a result of airborne exposures to TCI is generally more difficult in humans
than in rodents because more complete studies can be conducted in the rodent.
In humans exposed to TCI, almost 8 hours is required to reach steady state,
although in 4 hours the ratio of alveolar air/tissue partial pressure is about
,0.9 (Fernandez et al., 1977). Secondly, the percent recovery in humans tends
4-9
-------
to be much lower than in animals, beqause of the use of nonlabeled materials as
well as the longer clearance time of the metabolites (for example, TCA, which
binds to plasma protein).
Thus, among the available human data, the best estimation of TTCIM may
be obtained by subtracting the amount exhaled after exposure from the amount
retained during exposure. The work of Monster et al. (1976a) best meets these
preferences. These investigators exposed four volunteers (mean body weight
r
69.8 kg) at rest to two (70 and 140 ppm) concentrations (c) of TCI for 4 hours
(240 minutes), and calculated the retention rate (r fraction retained), minute
i
volume (MV), amount of TCI exhaled after exposure, and urinary metabolites.
They were able to account for only about 70% of the TCI. The technique used to
calculate the percentage of TCI exhaled produces the most consistent results
and is judged to be more reliable than recovery of metabolites. Monster et al.
(1976b) and Monster (1979) reported 10% at a 70-ppm TCI exposure and up to 95%
for PERC in other papers. This procedure was also carried out for two 30-min-
ute work periods within 4 hours. The values calculated from the averages are
presented in Table 4-1. The retention percentages are in keeping with a mean
i
of 45% calculated for 30-minute exposures to TCI at 100 and 200 ppm from the
data of Astrand and Ovrum (1976).
As summarized in the HAD based on the work of others (Nomiyama and Nomi-
i
I
yama, 1977; Ikeda, 1972), the metabolism is not saturated in humans; at these
concentrations (70 and 140 ppm). Since an 8-hour occupational exposure is more
likely to occur, the data were converted to an 8-hour exposure, and these data
are linear through zero, as shown iniFigure 4-2. It should be noted that the
human data are given for humans at rest. The formation of TTCIM by workers is
somewhat greater because of an increase in minute volume (10 liters; at rest to
i
30 liters at 100-watt work). The use of about 4 hours of work and 4 hours at
4-10
-------
TABLE 4-1. ESTIMATION OF THE TTCIM (mg) FORMED AT REST AND AT WORK
IN MEN EXPOSED TO AIRBORNE TCI CONCENTRATIONS
(140 ppm - 753 mg/m3 AND 70 ppm - 376 mg/m3)
RECALCULATED FROM MONSTER ET AL. (1976a)
Concentration
in mg/m3
(minutes-
activity)
TCI
inhaled
(t x Mv x c)
(mg)
Fraction TCI
retained retained
(%) (mg)
Exhaled as
% TCI
retai ned
TTCIM
753 (240-rest)
753 (60-work)
376 (240-rest)
376 (60-work)
1,861
1,355
875
677
41
38
46
38
763
515
402
258
8
10
11
10
702
463
358
232
4-11
-------
O
d o d d o ci o d o
a>
I N
t_ -r-
5 "o
J2
O «3
4J ^_j
0)
a,.
3 -<
(/» O
O H-
a.
X i
0) «J
M
c- o
3 4->
O
J= 4-
I O
00
^->
~* c
4J 3
O
a)
-a
a»
(O !-
e -u
3 V)
JC H-
0)
!£'*-
4-> O
ai
,
o
r- tfl
0) 4-)
C£ rt)
e.
-------
rest will give slightly less total volume inhaled than the 10 m3 normally used,
but may be a good estimate of moderate work. (The International Commission
on Radiological Protection [1977] uses an assumption of an 8-hour moderate
work period in which 10 m3 are breathed to arrive at their estimate of 20 m3
breathed per day. Therefore, the total TCI metabolite formed during 8 hours of
moderate work at any given concentration would be about 1.8-fold greater than
is indicated in Figure 4-2.) F:or 24-hour continuous exposure, it is estimated
that the TTCIM would be about 3.8 times greater than at rest, assuming two
8-hour periods of moderate work or activity and 8 hours of rest. It should be
noted that the animal data on which pharmacokinetic simulations are based come
from resting animals, the most common situation during chronic inhalation expo-
sure.
4.3. ESTIMATION OF EFFECTIVE DOSE FROM ANIMAL DATA
Stott et al. (1982) explored the formation of total metabolite after
inhalation exposure for 6 hours to 10 and 600 ppm of radio!abeled TCI. While
it is unfortunate that Stott et al. did not explore additional airborne concen-
trations, their investigation seems to provide the best rodent inhalation
pharmacokinetic information available. Prout et al. (1985) performed balance
studies following oral administration of single doses of radiolabeled TCI in
corn oil. These are described in detail in the 1985 HAD. The results of Prout
et al. (1985) showed no marked difference in the total amount of metabolite
formed in two strains of rats and mice at oral TCI doses of 10, 100, 500, and
1,000 mg/kg; therefore, the data were combined as shown in the HAD. The tech-
niques used by Prout et al. (1985) are quite similar to those used by Stott et
al. (1982). It was felt that the Prout et al. (1985) data could be used to
supplement the inhalation data of Stott et al. (1982) for examination of the
enzyme kinetics of TCI.
4-13
-------
In Tables 4-2 and 4-3 these data are summarized and the mg eq/animal from
Stott et al. (1982) is normalized forj the weights of animals used by Prout et
al. (1985). The fact that the data from the rats are not linear suggests the
saturation of metabolism. The regression [0.01213 + 4.47622 (1/X) = 1/Y, r? =
0.99897] of the reciprocals of the dose (mg/kg or mg eq/kg) and the TTCIM is
shown in Figure 4-3. The non-reciprocal plot is illustrated in Figure 4-4.
The data from the mice appear linear at 2,000 mg/kg or less, as would be
expected from information related toienzyme kinetics in the mouse. The mouse
data from the oral and inhalation exposures have been fitted to a linear
regression (0 + 0.02474 x = y, r2 = 0.98), and are plotted in Figure 4-5. The
Lineweaver-Burk solution was 0.00007;+ 34 (1/X) = 1/Y.
When the rodent inhalation data!are examined, it is clear that one is
dealing with a two-compartment model. The second compartment is the first-
order enzymatic degradation of the parent compound into its metabolites, as
discussed previously. The first compartment involves the inhalation of TCI
vapor and its retention. The amounts inhaled was estimated using the allometric
expressions provided by Anderson et |al. (1983) for amount of air inhaled as
follows:
Rats 0.105 (bw/0.113 kg)j°-67 = .m3 inhaled in 24 hours
Mice 0.0345 (bw/0.025 kg)0-67 = m3 inhaled in 24 hours
The percentages retained at 10 and 600 ppm, based on information from the work
of Stott et al. (1982), are neitheriproportional to the airborne concentration
nor the same between species at equal airborne concentrations.
I
In order to determine the amount of TTCIM produced from inhalation of
i
various airborne concentrations, several mathematical transformations were
4-14
-------
TABLE 4-2. SUMMARY OF DATA USED IN THE ANALYSIS OF ENZYME KINETICS
IN THE RAT
Applied dose (TCI)
ppm
From Stott
10
600
From P rout
mg or
mg eq/kg
et al., 1982
4.7
141.24
et al., 1985
10
500
1,000
2,000
mg or
mg eq/rat
1.18
35.31
2
100
200
400
Exhaled
mg eq
0.03
7.46
0.3
43
112
311
Metabolized
mg eq
1.15(0.9802)*
27.82(23.796)9
1.93
55.4
79
80.4
Adjusted (bw 2/3 power) to give the TTCIM expected from a 200-gram rat.
4-15
-------
TABLE 4-3. SUMMARY OF DATA USED IN THE ANALYSIS OF ENZYME KINETICS
IN THE MOUSE
Applied dose (TCI) I
ppm
From Stott
10
600
From Prout
mg or
mg eq/kg
et al., 1982
10.3
412
et al., 1985
10
500
1,000
2,000
i
mg or j
mg eq/rat
0.36
14.4
0.30
15
30
60
Exhaled
mg eq
0.003
0.346
0.01
0.9
5.3
8.2
Metabolized
mg eq
0.36(0.
14.0(12.
0.28
13.7
28.3
46.9
32)a
5)a
aAdjusted (bw 2/3 power) to give the JTCIM expected from a 30-gram mouse.
4-16
-------
CM
C ID
O I--
r- «5j-
4J .
IO O) "=>
i JS
fO h- +
c co
O)
O
o>
« -« 3:
o o
0.1 C
'I O
t i t/J
<_) -o tn
j re) +J J=
O O F
a E
(O O CM
O H- I
o <*
Q. (O CO
r- -(J r
O O JQ
CO )-> (O
i. I
ja 4J -r-
£= O)
«-> =J
« o c
e o
(O -r
S_
o
«!-««-
O £=
J= ra C
CO O O *
CO O>
O £- t3 CTl
i Q. CO
+J !«/» O
CO (O -~-
I -- 4-> r-t
^r CT> rtj
*t -a it
CO -^
S. CT+J'-^
3 CO =J X
en Q.--*.
4-17
-------
CD
E <"
en
e
a> o
o g
c
r- i
o
Q.I
in (O
C 4J
O O
(O »l-
r O
(U
o: 4J
c
3
Bui '
d)
<0 ^- ft)
t_ .p u
o>
i C >p-
U_ (O U.
4-18
-------
01
i
rfl
10
I
O
I
o
o
CM
I
O
CM
CU
en
O "O
r- CU .
4J ,
Jus T3
II
CM
n
O
O)
S
CO ^^
-d I-
3 X
Q.
o c r*.
Q.-I- (*
CM
CU O
010
M-
O ~»O
cr
to cu to
CU it)
v) en 36
os
o c:
CU "' -I
£Z O V)
-M I W)
I- CU
tt) OJ C-
LO N)
3 -I- £-
O i (0
E O Q)
-Q C
Ol i
4-> O)
£ O)
r- >I f
O
OJ
d
jc r co
(/) fO I
C +J «*
O O
r- 4-> r
(O <*- -Q
O (O
a; h-
Q£ 4J
c c:
3 .r-
o
6m
«* >
-o o
r- C T-
4-19
-------
tried. The overall solution required'that the solution to the first compart-
ment be similar in form for both species, while the solutions to the second
compartments were as for the enzymatic degradation. The overall optimum
solutions are illustrated in Figures 4-6 and 4-7 for rats and mice, respec-
tively, and turn out to be a log-log solution in the first compartment and the
Lineweaver-Burk solution for the Mich&elis-Menten equation. As can be seen,
nonlinearity for the rats (200 g) is still slightly evident in the range of
airborne concentrations (6-hour exposures) used in the simulation. The simu-
lation from the mice (30 g) is nearly linear. Over the TCI exposure concen-
trations used in the inhalation bioassays, the daily TTCIM predicted for a
6-hour exposure were as follows:
Airborne TCI Predicted TTCIM (mg)
concentration
(ppm) Rat (200 g) Mouse (30 g)
600 21.81 10.2
450 18.19 7.87
300 13.85 5.45
150 8.83 2.91
100 5.85 1.74
50 3.57 1.08
10 0.96 0.25
It should be noted that the simulation of TTCIM in the mouse resulted in
slightly smaller amounts than the actual mean measurements reported by Stott et
.
al. (1982), even when those figures are scaled downward for a smaller animal.
The modeling techniques used here give 105% and 91% of the observed values
at 10 and 600 ppm in the rat. This result is certainly within the range of
4-20
-------
CL
a
M
O
h-
03
P
O
o
c
re
o
o o
C <0
Or-
i- =J
-(-> E
fO *r~
£_ 00
4->
C «-
(U OJ
04J
C 3
O 0.
O E
O
O) U
o
J3
S- T3
r- O)
(T3 C
cr
O)
CDS
CH-
0.-0
r- O)
C i
0 2
i- .O
i i i r i
o. e» co. N> O
LL. IQ
4-21
-------
fd
4->
o
a
a
M
O
a
c
CO
O
f-
o
r- IO
±3 r^-
Si
O 3
Q.
)
t. XI
r"
as -a
O)
c c
<1J -r-
ai
S E
a> s:
C H-
T- O
r- O
at
£ N
(/» -r-
o "o
r O)
O) E
a:
.
4-
O
(- C
3 3
en a
4-22
-------
experimental variation. In comparison, the body burdens (mg eq/kg) for rats
exposed to 10 and 600 ppm predicted by Koizumi et al. (1987), using a summed-
compartmental technique, were 5.32 mg eq/kg and 200 mg eq/kg, as compared to
4.70 and 141 mg eq/kg from the data of Stott et al. (1982). This is 113% and
142% at 10 and 600 ppm, respectively.
The solution of the first compartment indicated that the log of the per-
cent retention decreased as the log of the concentration increased. A similar
situation was noted by Bond et al. (1986) for 1,3-butadiene and by Beliles and
Parker (1985) from the work of Tyler and McKelvey (1980) on ethylene oxide.
Parker and Davidson (1985) showed a log-log linear relationship below Vmax be-
tween the external concentration of vinyl chloride and the amount metabolized.
In Figure 4-8, the plot of the log TCI concentration and log TTCIM between 10
and 600 ppm for rats is shown as a straight line. The output of the simulation
is shown as the points. The computer simulation for TCI clearly produces a
solution that is similar to that found to exist for some other hydrocarbons.
4.4. ESTIMATION OF HUMAN EQUIVALENT DOSE FROM ANIMAL DATA
From these simulations, the TTCIM at the daily airborne concentrations
used in the various bioassays can be scaled to the best estimate of the weights
of the bioassay animals and to the time of daily exposure, if it varies from
6 hl5iJrUs as in the pharmacokinetic studies. Then, TTCIMs will be scaled to a
daily dose in humans. This scaling results in the human equivalent dose (HED),
which is the daily dose required to produce the same response in humans as seen
in animals. When humans are exposed to airborne concentrations approaching and
above metabolic saturation, a longer duration of exposure would be necessary
for the production of this estimated amount of TTCIM. The use of rat simula-
tion data and scaling of 10 ppm to a 70-kg human for an 8-hour exposure gives
65 mg of TTCIM. This compares to an estimated TTCIM of 102 mg in humans at
4-23
-------
1 2 34567891 2 34567
s
»
o
H
)
1
x
,
x
/!
^/
x
/
i
/
y
/
/
' 9
/
X
/,
8
1
fi
4
1
9
R
P
4
1
3
7
1
0 100 1000
TCI, ppm
Figure 4-8. Log plot (shown as a line) of experimental
rat data at 10 and 600 ppm and simulated TTCIM (mg eq)
(shown as points). ;
4-24
-------
10-ppm exposure for 8 hours from the human data as shown in Figure 4-2. The
variation in the metabolite estimates could be due to the different activity
levels of the rats compared to the humans.
4.5. SELECTION OF ANIMAL DATA SETS
With the variety of tumors in the series of bioassays available, it would
be useful to have the number (of significance) of tumor-bearing animals.
Combining the reported tumor incidence will lead to a possible overestimate of
response, if more than one tumor occurred in a single animal.
As previously indicated, Prout et al. (1985) showed that there is little
difference between the metabolism of TCI in the Swiss or B6C3F1 mouse and that
which occurs in the Osborne-Mendel or Alderley Park (Wistar-derived) rat. The
mouse strains used were those used by Maltoni et al. (1986) in their series
of bioassays. The Sprague-Dawley rat used by Maltoni et al. (1986) origina-
ted from the Wistar strain, but now tends to be a little larger in adulthood.
Other than this difference, which will lead to a slightly lower metabolic rate,
it seems reasonable to assume that there is little difference between the data
on the Sprague-Dawley rats used by Maltoni et al. (1986) and the Osborne-Mendel
or combined rat data which serves as the basis of the pharmacoklnetic informa-
tion derived from Stott et al. (1982).
The tumor incidence in rats with respect to different end points from the
Maltoni ;et al. (1986) assays, in which rats were exposed to graded airborne
TCI, is shown in Table 4-4, along with the calculated daily dose in terms of
TTCIM. The incidence data are expressed in terms of animals at risk at the
appearance of the first tumor, as previously discussed. The renal tumors,
although of biological interest, were not used because of the low response
rate.
The use of animal inhalation bioassays to predict the risk to humans from
4-25
-------
TABLE 4-4. SUMMARY OF ESTIMATED DAILY DOSE IN TERMS OF TTCIM
AT THE AIRBORNE CONCENTRATIONS AND DURATIONS OF DAILY EXPOSURE
USED IN THE BIOASSAY WITH SPRAGUE-DAWLEY RATS (MALTONI ET AL., 1986)
AND THE RESULTING TUMOR INCIDENCE, PLUS THE HED
TCI (ppm)
Males (580 g)
600
300
100
0
Females (375 g)
600
300
100
0
Daily TTCIM
Bioassay
52.0
33.0
13.9
38.8
24.6
10.4
(mg)
HED
1,289
818
346
1,289
818
346
Tumor
Leydi g
cell
31/129
30/130
16/130
6/135
incidence
Leukemias
15/129
14/130
13/130
9/135
11/130
2/130
9/130
7/145
4-26
-------
ambient airborne concentrations certainly has more appeal than trying to pre-
dict the risk from studies utilizing a different exposure route. In addition,
the use of the mouse pulmonary tumors is judged important and consistent with
the capability of lung tissue to partly metabolize a fraction of the inhaled
TCI. Dalbey and Bingham (1978) showed, using isolated perfused lung prepar -
ations, that pulmonary tissue can metabolize TCI to TCE and TCA. Chloral was
not isolated in their preparations. Bergman (1983a, b) illustrated the binding
of TCI or its metabolites in lung and bronchi after oral or inhalation exposure
to radiolabeled TCI. Interestingly, this author indicated that the binding
activity of TCI after oral administration to RNA in the lung exceeds that seen
in the kidney or the liver.
The tumor incidence in Swiss and B6C3F1 mice from the Maltoni et al.
(1986) assays resulting from exposure to graded airborne concentrations of TCI,
and the tumor incidence in the ICR mouse from the Fukuda et al. (1983) inves-
tigation, are shown in Table 4-5, along with the calculated daily dose in terms
of TTCIM (mg) for the bioassay animals, with the estimated daily HED. Because
of the variation in response between the two types of male mice of the B6C3F1
strain from different suppliers, it would appear appropriate to eliminate those
from consideration. Except for the ICR female mice, the mid-study animal body
Weights are used for calculations. For the ICR female mice, the mid-term body
weight was assumed to be the same as for the Swiss female mice.
4.6. EXTRAPOLATING AND SCALING
The variations in experimental design in terms of route of exposure and
duration of exposure, the similarity of response within species, and an appa-
rently good pharmacokinetic model with regard to the oncogenic agent, TTCIM,
allow an unusual opportunity to explore certain assumptions which normally are
adopted when extrapolating from route to route, or from shorter (less-than-
4-27
-------
TABLE 4-5. SUMMARY OF ESTIMATED DAILY DOSE IN TERMS OF TTCIM AT THE AIRBORNE
CONCENTRATIONS AND DURATION OF DAILY EXPOSURE USED IN THE BIOASSAYS WITH SWISS
MICE AND FEMALE B6C3F1 MICE (MALTONI ET AL.f 1986) AND WITH FEMALE ICR MICE
(FUKUDA ET AL., 1983), AND THE RESULTING TUMOR INCIDENCE, PLUS THE HED
TCI (ppm)
Swiss niale mice (47 g;
600
300
100
0
Daily TTCIM (mg)
Bioassay HED
78-week exposure)
16.1 2148
8.59 1148
2.74 367
Tumor incidence
Liver Lung
13/90 27/90
8/89 23/89
2/89 11/89
4/88 10/88
Swiss female mice (40 g; 78-week exposure)
600
300
100
0
B6C3F1 female mice (32
600
300
100
0
ICR female mice (40 g;
450
150
50
0
14.4 2148
7.71 1148
2.46 367
g; 78-week exposure)
12.4 2148
6.64 1148
2.12 367
107-week exposure!
11.1 1658
4.12 613
1.53 227
20/89
13/90
15/89
15/90
9/89 14/87
4/89 7/89
4/90 6/90
3/90 2/90
11/46
13/50
5/50
6/49
4-28
-------
lifetime) to longer (lifetime) studies and across species. In order to explore
these assumptions, an occupational scenario of 10 ppm for 8 hours or 1 pg/m^
for an environmental 24-hour exposure is used. This scenario will allow the
use of the daily estimated human doses of 184 or 0.00722 mg of TTCIM for the
occupational or environmental exposures, as well as the HEDs previously calcu-
lated for the various bioassays.
4.6.1. Route-To-Route
The apparent good fit of the TCI metabolic data from both the oral bolus
dosing and the inhalation dosing in balance studies would seem to allow predi-
cation of the inhalation response in mice from oral gavage response with regard
to the same end point (liver tumor). However, by the oral route some of the
parent compound is degraded in the liver during the first tissue exposure (the
first-pass effect). Thus, the remainder of the organs are essentially exposed
to a less potent TTCIM mixture than is the liver tissue, assuming that the
epoxide is the more reactive metabolite. In mice exposed by inhalation, pul-
monary tumors were the primary tumors, while in the oral studies, tumors at
this site were not reported to be significantly increased. The observation of
pulmonary tumors following inhalation exposure is consistent with the metabolic
capability of lung tissue to partly metabolize TCI through the epoxide.
To evaluate the exposure route equivalence consideration, the response
in the female B6C3F1 mice treated by gavage for 78 weeks (NCI, 1976) and the
response in the B6C3F1 female mice exposed to airborne concentrations for 78
weeks (Maltoni et al., 1986) were selected on the basis of similarity of dura-
tion of exposure. The related information is presented in Table 4-6. Using
these HEDs and the response, the risks equivalent to that for 3/4 of an occupa-
tional lifetime exposure of 10 ppm (estimated human dose, 184 mg TTCIM) were
calculated using a multistage model (polynomial degree equal to dose groups
4-29
-------
zero) (Howe et al., 1986). The oral-jbased input (HED and response) resulted in
i
an estimated extra human risk of 0.0015 for the occupational exposure sceario,
while the inhalation data using lung tumors produced an estimated risk of 0.010.
The inhalation data based on liver tumors yielded a risk of 0.0003. These
results are consistent with probable variations due to route-related pharmaco-
kinetics. The use of the liver response in the inhalation .study and the oral
i
liver response gave, when the responses from both routes were examined, a risk
i
of 0.004 and a better fit of the response data (shown in Table 4-6) than any of
i
the other combinations (po liver-inhalation lung or po liverinhalation liver
and lung). This would seem to imply [that in the case of the same end point,
the oral data can be used to predict |the response via inhalation.
4.6.2. Intra-Species
In the case of TCI, clearly thejuse of the applied dose as the effective
dose would not be appropriate for estimation of risk, particularly when con-
sidering the response associated with higher doses. This is because of the
marked difference in enzyme kinetics;, as indicated by a Vmax of 1S429 in the
I
mouse and 82 in the rat from the Lineweaver-Burk solutions.
When extrapolating across species for the purpose of assessing risk, it
is generally assumed that an organ-specific response in one species is predic-
tive of carcinogenic risk in another!species (although not necessarily the same
type of tumor or at the same site). |Differences from species to species in the
site of tumor formation are not well understood. For example, Garman et al.
(1985) noted that brain tumors resulted only from exposure of rats to alkylat-
ing agents, and that this organ response was noted in rats and not in mice, in
the NTP/NCI bioassay program. These authors were interested in tumor develop-
ment resulting from airborne exposure to ethylene oxide, an alkylating agent,
which in male rats produced leukemias, mesotheliomas, and brain tumors, and
4-30
-------
TABLE 4-6. SUMMARY OF ESTIMATED DAILY DOSE
IN TERMS OF TTCIM AT THE APPLIED DOSES
(ORAL-mg/kg [NCI, 1976] AND INHALATION-ppm [MALTONI ET AL., 1986])
IN THE BIOASSAYS WITH FEMALE B6C3F1 MICE,
AND THE RESULTING TUMOR INCIDENCE, PLUS THE HED
Tumor incidence
Applied
dose
Mouse TTCIM
(mg)
HED TTCIM
(mg)
Liver
(expected)
Lung
B6C3F1 female mice (26 g; 78«week exposure; killed at 90 wks)
1,739 mg/kg 43.0 8548 11/47 (0.232)3
839 mg/kg 21.0 4272 4/50 (0.115)
0 0/20 (0.032)
B6C3F1 female mice (32 g; 78-week exposure; lifetime observation)
600 ppm 12.4 2148 9/87 (0.074) 14/87
300 ppm 6.6 1148 4/89 (0.055) 7/89
100 ppm 2.12 367 4/90(0.039) 6/90
0 3/90 (0.032) 2/90
aTumor incidence expected, based on multistage model analysis for combined
studies.
4-31
-------
produced a variety of tumors in mice, Including lung tumors but not brain
tumors. ' ' '' 'X :,:.-;:; '^ -\-; :''''...
For the purpose of examining across-species extrapolation, the tumor re-
sponses as a result of exposure to airborne, concentrations of TCI from Maltoni
i
et al. (1986) in male Sprague-Dawley irats and Swiss mice were selected because
,
of the similarity in experimental design. In the rats the significant tumor
i *
was Leydig cell tumors, while in the [mice the significant tumors were in the
lung. The pertinent information is summarized.in Table 4-7.
Using these HEDs and the corresponding responses, the human risk equiva-
lent to that for 3/4 of occupational ilifetfme exposure of 10 ppm (TTCIM 184 mg)
were estimated using the multistage model. Given this occupational scenario,
the analyses indicated an extra human risk of 0.037 based on the rat data
(Leydig cell tumors), a risk of 0.022 based on lung tumors among the mice, and
a risk of 0.029 when the data were combined. From this one may infer that the
possibility for reasonable cross-species extrapolation is good, if selection of
the active agent(s) and pharmacokinetic modeling are appropriate. It should be
noted that the estimated risk from the mouse lung tumor incidence is similar to
that predicted from the mouse data in the previous section, although data from
the opposite sex and a different strain of mouse had been used.
4.6.3. Shorter to Longer Duration of Exposure
It is usually assumed that a 2-year exposure represents a lifetime expo-
sure in rodents. However, at the end of a 2-year exposure study the test
i . ;
animals may actually be 26 months of iage, allowing for nursing, shipping, and
acclimatization. In addition, Solleveld et al. (1984) indicated that 28 months
is the median lifespan of the F344 rat. On the other hand, these authors
indicated that animals of 116 weeks of age or older were not characterized by
unique lesions. In recent NTP studies, the survival in the control B6C3F1 mice
4-32
-------
TABLE 4-7. SUMMARY OF ESTIMATED DAILY DOSE IN TERMS OF TTCIM
AT THE AIRBORNE TCI CONCENTRATIONS AND DURATIONS OF DAILY EXPOSURE
USED IN THE BIOASSAYS WITH SWISS MALE MICE (MALTONI ET AL., 1986)
AND WITH SPRAGUE-DAWLEY MALE RATS (FUKUDA ET AL., 1983),
AND THE RESULTING TUMOR INCIDENCE, PLUS THE HED
Airborne Animal
concentration TTCIM HED TTCIM
(PPm) (mg) (mg) Tumor incidence
Swiss male mice (47 g; 78~week exposure)
600 16.1 2148 27/90
300 8.59 1148 23/89
100 2.74 367 11/89
0 10/88
Sprague«Daw1ey male rats (580 g; 104-week exposure)
600 52.0 1289 31/129
300 33.0 818 30/130
100 13.9 346 16/130
0 6/135
4-33
-------
is most often greater than 50%; thus, a similar assumption would seem to be
appropriate for mice.
The assumptions made concerning a maximum occupational exposure are that
j
worker exposure occurs 40 hours/week,! 48 weeks/year (10 days vacation and 10
days holiday and sickness) for 45 years, and that the total lifetime is 70
years. Thus, a maximum occupational lifetime exposure represents an exposure
of only 14% of the total lifetime (70 x 365 x 24 = 613,200 hours). Using the
estimate of 28 months for a total rodent lifetime, the percentages of lifetime
actually exposed (7 hours/day, 5 days|/week) in the TCI inhalation bioassays
were as follows:
Weeks of bioassay
exposure
107
104
78
52
% equivalent
lifetime exposure
18.3
17.8
13.3
8.1
Thus, the 78-week exposure of rodents is the nearest to a maximum occupational
exposure when the percentage of lifetime is considered. When this considera-
tion is applied to the risk based on the rat data (104-week exposure) in the
previous risk estimation, the risk from a maximum occupational exposure would
be 0.029. The adjustment of the HED Iby 17.8/14 also gives a risk of 0.029.
This may then be compared to the risk of 0.010 from the female B6C3F1 mouse
data or 0.022 from the male Swiss mouse data.
Recently, The Netherlands (1987) has used the 1/4 factor (40 hours of
occupational exposure versus 168 hours in a 7-day week) to convert from occu-
pational to lifetime exposure. The calculation that a maximum occupational
exposure is only 14% of the lifetime indicates a discrepancy of about 10% when
4-34
-------
compared to the 1/4 factor.
In order to evaluate the approach used to estimate percent of lifetime as
applied to environmental risk of TCI, the HEDs from various mouse bioassays
were converted to lifetime equivalents by use of the percent lifetime of the
dose together with the lung tumor response in mice, as indicated in Table 4-8.
The extra human cancer risk estimated is 3.9 x 10"6 (upper confidence limit
5.3 x 10"6) based for TTCIM produced by a continuous (24-hour) lifetime expo-
sure at 1 yg TCI/m3 (TTCIM 0.00722 mg/day). The exclusion because of a
difference in apparent background (control) rate for lung tumors among the
female B6C3F1 mice gave a slightly better fit, although the risk estimate was
similar (4.1 x 1Q-6).
4-35
-------
TABLE 4-8. SUMMARY OF ESTIMATED DAILY DOSE IN TERMS OF TTCIM AT
THE AIRBORNE TCI CONCENTRATIONS AND DURATIONS OF DAILY EXPOSURE USED IN THE
BIOASSAYS WITH SWISS MICE AND FEMALE B6C3F1 MICE (MALTONI ET AL., 1986)
AND WITH ICR FEMALE MICE (FUKUDA ET AL., 1983)
AND THE RESULTING TUMOR INCIDENCE PLUS THE HED
AND THE HUMAN LIFETIME-ADJUSTED HED
TCI (ppm)
Swiss male mi
600
300
100
0
Swiss female
600
300
100
0
B6C3F1 female
600
300
100
0
ICR female mi
450
150
50
0
Daijy TTCIM (rngj
Bioassay HED (adjusted)
ce (47 g, 78-week exposure)
16.1 2148(286)
8.59 1148(153)
2.74 367(49)
mice (40.g, 78-week exposure)
14.4 2148(286)
7.71 1148(153)
2.46 . 367(49)
mice (32 g, 78-week exposure)
12.4 2148(286)
6.64 1148(153)
2.12 !367(49)
ce (40 g, 107-week exposure)
11.1 1658(303)
4.12 613(112)
1.53 227(42)
Lung tumor
Observed
27/90
23/89
11/89
10/88
20/89
13/90
15/89
15/90
14/87
7/89
6/90
2/90
11/46
13/50
5/50
6/49
incidence
Predicted3
0.23
0.17
0.13
' 0.10
0.23
0.17
0.13
0.10
0.23
0.17
0.13
0.10
0.24
0.15
0.12
0.10
aBased on multistage model analysis for combined studies,
4-36
-------
5. QUANTITATIVE EVALUATION: UNIT RISK DERIVATION
This chapter contains a quantitative analysis of the inhalation bioassay
data as reported by Maltoni et al. (1986) and Fukuda et al. (1983). This
chapter is an update to the unit risk derivation that was part of the 1985
HfcSltfi Assessment Document (HAD) for Trichloroethylene (U.S. EPA, 1985a). 1ft
the 1985 HAD, the inhalation cancer risk estimates were developed from gavage
data, since the available inhalation studies were not suitable for analysis.
Given the the more recent inhalation data, estimates for cancer risk can be
developed based on exposure-route-specific animal bioassay data, with the
previously-developed gavage-based risk estimates providing a basis for com-
parison.
The unit risk estimate for an air pollutant is defined as the increased
lifetime cancer risk for an individual who is continuously exposed to an air
pollutant from birth to death (assumed 70-year lifetime) at a concentration of
1ng/m3 of the agent in air. In the 1985 HAD, the unit risk for TCI in air
was calculated to be 1.3 x 10-6 on the bas1s of liver tumor response 1n mice
from gavage studies, using relevant information on metabolism and pharmaco-
kinetics for TCI.
5.1. DATA USED FOR THE DOSE-RESPONSE CALCULATION
Tables 5-1 through 5-4 give seven sets of dose-response data on mice and
rats from Maltoni et al. (1986) and Fukuda .et al. (1983). The dose-response
data and the derivation of the metabolized dose have been discussed in the pre-
vtoils chapter.
Two dosimetries are given in these tables: the animal metabolized dose
and the human equivalent dose. Both of these dosimetries are demonstrated in
the risk calculations. The animal metabolized dose is used when the dose, in
5-1
-------
TABLE 5-1. TUMOR RESPONSES AND THE CORRESPONDING METABOLIZED
DOSES FROM THE MALTONi; ET AL. (1986) STUDY IN RATS
Inhalation
Sex
Male
(580 g)
dose
(ppm)
0
100
300
600
Animal
metabol i zed
dose
(mg/kg/day)a
0
17.1
40.6
64.0
Human
1 equivalent
dose
(mg/kg/day )b
i
0
3.5
8.2
13.0
Tumor incidence
Ley dig cell
6/135
16/130
30/130
31/129
aAnimal metabolized dose (mg/kg/day) = (daily bioassy dose metabolized, mg)/W
x 5/7, where 5/7 reflects that animals were exposed 5 days/week and W is the
body weight in kg. The daily dose metabolized (TTCIM) is presented in Chapter
4.
bHuman equivalent dose = (animal metabolized dose, mg/kg/day) x (Wa/WH)1/3,
where Wa/WH is the ratio of animal and human body weights. This value can
also be obtained by dividing the HED in Table 4-4 by 70 kg (the human body
weight) and multiplying by 5/7. It is assumed in this calculation that
mg/surface area is equally potent among species.
5-2
-------
TABLE 5-2. TUMOR RESPONSES AND THE CORRESPONDING METABOLIZED
DOSES FROM THE MALTQNI ET AL. (1986) STUDY IN SWISS MICE
Sex
Male
(47 g)
Female
(40 g)
Inhalation
dose
(ppm)
0
100
300
600
0
100
300
600
Animal
metabolized
dose
(mg/kg/day)a
0
31.2
97.9
183.5
0
32.9
103.3
192.9
Human
equivalent
dose
(mg/kg/day)b
0
2.7
8.6
16.1
o-
2.7
8.6
16.0
Tumor incidence
Liver
4/88
2/89
8/89
13/90
Lung
10/88
11/89
23/89
27/90
15/90
15/89
13/90
20/89
aAnimal metabolized dose (mg/kg/day) = (daily bioassay dose metabolized, mg)
x 5/7 x (78 weeks/104 weeks)/W, where W is the body weight in kg; 5/7 and
78/104 reflect that animals were exposed 5 days/week for 78 weeks out of the
assumed lifespan of 104 weeks,,
bHuman equivalent dose = (animal metabolized dose, mg/kg/day) x (Wa/WH)1/3,
where Wa/WH is the ratio of animal and human body weights. This value can
also be obtained by using the HED in Table 4-5 as follows:
It is assumed
species.
HED x (78/104) / 70 kg x 5/7
in this calculation that mg/surface area is equally potent among
5-3
-------
TABLE 5-3. TUMOR RESPONSES AND THE CORRESPONDING METABOLIZED
DOSES FROM THE MALTONI ET AL. (1986) STUDY IN B6C3F1 MICE
Sex
Female
(32 g)
Inhalation
dose
(ppm)
0
100
300
600
Animal
metabolized
dose
(mg/kg/day)a
0
35.5
111.9
207.6
Human
equivalent
dose
(mg/kg/day )b
0
2.7
8.6
16.0
Tumor incidence
Liver Lung
2/90 2/90
4/90 6/90
4/89 7/89
9/89 14/87
aAnimal metabolized dose (mg/kg/day) = (daily bioassay dose metabolized, mg)
x 5/7 x (78 weeks/104 weeks)/W, where W is the body weight in kg; 5/7 and
78/104 reflect that animals were exposed 5 days/week for 78 weeks out of the
assumed lifespan of 104 weeks.
"Human equivalent dose = (animal metabolized dose, mg/kg/day) x (Wa/WH)l/3,
where Wa/WH is the ratio of animal and human body weights. This value can
also be obtained by using the HED in Table 4-5 as follows:
HED (mg) x (78/104)/70 kg x 5/7
It is assumed in this calculation that mg/surface area is equally potent among
species.
5-4
-------
TABLE 5-4. TUMOR RESPONSES AND THE CORRESPONDING METABOLIZED
DOSES FROM THE FUKUDA ET AL. (1983) STUDY IN ICR MICE
Sex
Female
(40 g)
Inhalation
dose
(ppm)
0
50
150
450
Animal .
metabolized
dose
(mg/kg/day )a
0
27.4
73.6
193.2
Human
equivalent
dose
(mg/kg/day )b
0
2.3
6.1
16.1
Tumor incidence
Lung
6/49
5/50
13/50
11/46
aAmmal metabolized dose (mg/kg/day) = (daily bioassay dose metabolized, mg)/W
x 5/7, where 5/7 reflects that animals were exposed 5 days/week and W is the
body weight in kg.
bHuman equivalent dose = (animal metabolized dose, mg/kg/day) x (Wa/WH)1/3,
where Wa/WH is the ratio of animal and human body weights. This value may
also be obtained by using the HED in Table 4-5 as follows:
HED (mg)/70 kg x 5/7
It is assumed in this calculation that mg/surface area is equally potent among
species.
5-5
-------
mg/kg/day, is assumed to be equivalent (i.e., equally potent) among species.
This is referred to as the body-weight dose equivalence assumption. When the
equivalent dose among species is assumed to be correlated with mg/surface area/
day (referred to as the body-surface dose equivalency), the human equivalent
dose is used in the risk calculations. Both body-surface and body-weight dose-
equivalence assumptions have support; depending on the chemical under review.
According to EPA's Guidelines for Carcinogen Risk Assessment, the body surface
(bw^/3) scaling assumption is preferred for interspecies variability unless
there is evidence to support an alternative assumption, there being no such
evidence for TCI. Interestingly, when the animal-based risk estimates are
compared to the cancer incidence observed directly from the human population
for twenty-odd chemicals, both dose-equivalence assumptions are shown to be
reasonable (although pharmacokinetics are not taken into account in any of the
risk value derivations), as demonstrated in the EPA/DOD co-sponsored project,
"Investigation of Cancer Risk Assessment Methods" (U.S. EPA, 1987).
5.2. SLOPE (POTENCY) CALCULATION
On the basis of the seven dose-response data sets given in Tables 5-1
through 5-4, the slope estimates, q£ (also called potency values), for TCI are
calculated using the linearized multistage model programmed as GLOBAL86, as
shown in Table 5-5. The calculation procedure, the interpretation, and the
utility of the unit risk concept were given in the 1985 HAD. The q^ value is
the 95% upper-bound estimate of the 'linear parameter in the multistage model,
and is used in the derivation of the actual unit risk value. The slope esti*-
4* :
mates (q^) based on metabolized dose calculated from the seven data sets range
from 7.1 x 10~3 to 2.7 x 10~2 using the body-surface dose-equivalence assump-
tion, and from 5.5 x 10~4 to 5.3 x lO"3 using the body-weight dose-equivalence
assumption. The potency estimates appear more comparable when the body-surface
5-6
-------
TABLE 5-5. SLOPE ESTIMATES PER (mg metabolized dose/kg/day)
CALCULATED ON THE BASIS OF DIFFERENT DATA SETS AND UNDER DIFFERENT
DOSE-EQUIVALENCE ASSUMPTIONS
Data base
Dose-equivalence assumption
Body-weight basis3
(animal metabolite dose)
Body-surface basisb
(human equiv. dose)
Maltoni et al., 1986
Male rats
Leydig cell
Swiss male mice
Li ver
Lung
Swiss female mice
Lung
B6C3F1 female mice
Li ver
Lung
Fukuda et al.. 1983
ICR female mice
Lung
198.5 HAD estimate
Mouse cjavage studies
5.3 x lO-3
9.2 x 10-4
2.1 x ID"3
7.4 x 10-4
5.5 x 10-4
1.0 x lO-3
2.0 x
2.7 x lO-2
1.1 x lO-2
2.4 x 10-2
9.0 x lO-3
7.1 x ID'3
1.3 x 10-2
2.4 x
1.3 x
aThe third columns in Tables 5-1 through 5-4 are used in dose-response calcu-
lations.
The fourth columns in Tables 5-1 through 5-4 are used in dose-response calcu-
lations.
5-7
-------
dose-equivalence assumption is made than with the alternative body weight dose-
equivalence assumption. The geometric means of the q^ values calculated on
the basis of mouse liver and lung incidence data are 8.7 x 10~3 and 1.7 x 10~2,
respectively, under the body-surface dose-equivalence assumption, and 7.0 x
10~4 and 1.3 x 10~^, respectively, under the body-weight dose-equivalence
1
assumption. These values are comparable to the slope value calculated on the
basis of the NTP/NCI mice gavage studies (U.S. EPA, 1985a). Under the body-
surface dose-equivalence assumption, the potency slope was calculated to be
1.3 x 10~2/mg metabolized dose/kg/day, using the data from the gavage studies.
When mice were exposed to TCI by gavage, only liver tumors were induced,
while with inhalation exposure both liver and lung tumors were induced. There-
i
fore, an ideal approach for calculating the slope value on the basis of the
inhalation data would be to use the proportion of animals with liver and/or
lung tumors as the tumor response rate rather than considering each of the two
tumor types separately. This cannot be done at present, however, because the
individual animal data are not available. Since both lung and liver tumors
were induced in mice, the potency calculated on the basis of the combined
incidence,data would be greater than that calculated on the basis of either
i
lung or liver incidence separately, but less than the sum of these two indi-
vidual potency slopes. Therefore, where animals were observed to develop mul-
tiple tumor types, it is more appropriate to adopt the maximum value among
different slope values calculated on the basis of each of the individual tumor
types, rather than to take the average of these estimates across tumor types.
i
Table 5-6 gives the potency estimates calculated from each of the two animal
species. These potency values, which reflect the diversity of animal species,
are subsequently used to calculate the unit risk for TCI in air.
5-8
-------
TABLE 5-6. POTENCY SLOPE PER (mg metabolized dose/kg/day)
DERIVED FROM DIFFERENT ANIMAL SPECIES
(RATS AND MICE)
AND UNDER TWO DIFFERENT DOSE-EQUIVALENCE ASSUMPTIONS
Dose-equivalence assumption
Body-weight basis Body-surface basis
Species (animal metabolite dose) (human equiv. dose)
Rats 5.3 x 10'3 2.7 x 10'2
Mice 1.3 x ID'3 1.7 x 10'2
5-9
-------
5.3. RISK ASSOCIATED WITH 1 yg/m3 OF TCI IN AIR ,
In order to derive a unit risk using the potency slope, ql9 which is
expressed in terms of the metabolized dose, it is necessary to estimate the
amount metabolized when a person is exposed to 1 yg/m3 of TCI in air. As dis-
cussed previously, the study by Monster et al. (1976a) can be used for this
purpose. Using the Monster et al. data, and assuming that a person works
strenuously (equivalent to 100 watts
-------
weight dose-equivalence assumption, are 1.3 x 10~7 and 5.3 x 10"7, derived from
the mouse lung and the rat Leydig cell data, respectively.
Another approach that does not require the use of a human metabolized
dose corresponding to the ambient air concentration (e.g., 1 wg/m3) of TCI
is to assume that the air concentration is equally potent among species. That
is, the risk for humans due to 1 yg/m3 of TCI in air is the same as that for
animals.' The risks for rats and mice due to 1 ng/m3 of TCI in air are cal-
culated below.
The amounts metabolized corresponding to 1 yg/m3 of TCI in air are esti-
mated to be 2.9 x ICT4 mg/kg/day for rats and 6.5 x 10"4 mg/kg/day for mice.
These values are derived by interpolating linearly from the empirical obser-
vations that rats and mice exposed to TCI for 7 hours at 10 ppm (54,750 yg/m3)
produced 4.7 mg/kg and 10.3 mg/kg of the metabolized dose, respectively. By
multiplying the animal slope value, which is derived under the body-weight
dose-equivalence assumption (i.e., the values in the last column of Table 5-6)
and the amount metabolized, the risk estimates for animals are 1.5 x 10~6 for
rats and 8.6 x 1Q-7 for mice. Therefore, under the assumption that an ambient
air concentration is equally potent between animals and humans, the unit risk
for.TCI in air would range from 0.9 x 10'6 to 1.5 x 1Q-6. This estimate is
very close to the range previously calculated using the body-surface dose-
equivalence assumption. It should be noted that the procedure used above dif-
fers frdrti the approach that assumes dose equivalence on the basis of air con-
centration but does not adjust for the metabolism. The adjustment for the
metabolism somewhat improves the extrapolation of the risk from high to low
doses because of the metabolic saturation at high doses used in the bioassays.
Table 5-7 summarizes the inhalation unit risk estimates calculated from differ-
ent animal species and under different dose-equivalence assumptions.
5-11
-------
TABLE 5-7. SUMMARY:OF UNIT RISK ESTIMATES
DERIVED FROM DIFFERENT ANIMAL SPECIES
AND UNDER DIFFERENT DOSE-EQUIVALENCE ASSUMPTIONS
Dose-equivalence assumption
Data base
(species)
Body-weight
basis
Body-surface
basis
Air concentration
basis
Rats
Mice
5.3 x 10-7
1.3 x 10-7
2.7 x 10-6
1.7 x 10-6
1.5 x 10~6
0.9 x 10-6
5-12
-------
5.4 DISCUSSION
Although the available TCI metabolism data (total amount of metabolites)
are used 1n the risk calculation presented here, the analysis of metabolism
factors is far from ideal because of the lack of various pharmacokinetic data
that are necessary for organ-specific analysis as might be accomplished with a
physiologically-based pharmacokinetic (PB-PK) model. We are aware of the PB-PK
(Andersen) model proposed for TCI (NRC, 1986) which assumes that all the metab-
olites are produced in the liver. The assumption that the liver is the only
major metabolizing tissue is not adequate, given that EPA already knew (HAD,
1985a) that the lung and kidney were metabolizing tissues. Recent inhalation
bioassays further support this hypothesis. Based on the fact that the metabo-
lized doses for both animals and humans were interpolated from the observed
data, the risk value for TCI is not expected to be significantly different even
if and when a PB-PK model is used to predict the amount metabolized, as long as
the total metabolites are used as the dose in the risk calculation. However,
with the appropriate kinetic information, such as the metabolic activities of
lung and liver and knowledge of the metabolite synergism, or lack thereof, in
the various tissues, the PB-PK modeling approach could help to clarify some
aspects of the carcinogenic mechanism and to eventually improve the overall
characterization of the probable public health hazard.
One aspect of uncertainty in the present risk calculation relates to inter-
species dose conversion. Three options for species conversion have been consi-
dered. They are referred to as the body-surface dose-equivalence (bw2/3), the
body-weight dose-equivalence, (mg/k'g)1, and the air concentration equivalence
assumptions. Although both the body-weight and the body-surface dose-equiva-
lence assumptions have some support, we recommend that the risk estimates
derived using the body-surface assumption be used as recommended in the guide-
5-13
-------
lines and as further supported by the fact that these estimates are also con-
sistent with those derived under the assumption that the ambient air concen-
tration (pg/m3) is equivalent among species, an assumption that Is often
used in the risk assessment analysis of inhalation data. The fact that one
interspecies variability factor, metabolite production, correlates with body-
surface dose equivalence may or may not account for the full range of the
interspecies variability.
Because of the conservative approaches taken in the various steps of the
risk calculations, the estimated risk-for TCI should be considered to be an
upper bound of the true risk, as oppoked to a "best" or "true" estimate; that
is, the true risk could be considerably lower than the calculated upper-bound
value. The risk estimates presented fray be factored into regulatory decisions
to the extent that the concept of an upper-bound estimate of risk is considered
to be useful.
5.5. SUMMARY :
The inhalation unit risk estimates calculated directly on the basis of
several inhalation studies in rats and mice are comparable to those previously
derived from gavage studies in mice. The upper-bound estimate of the incre-
mental cancer risk due to lifetime exposure to 1 yg/m3 of TCI in air is esti-
mated to range from 1.7 x 1CT6 (on ttje basis of mouse data) to 2.7 x 1CT6 (on
a *5
the basis of rat data). The upper-bound unit risk value of 1.7 x 10-%ig/m
I
is preferred because of its relationship to first-pass effects resulting from
inhalation exposure (mouse lung), and the broadness of the data base in terms
of multiple mouse strains and both sexes. The unit risk based on the rat,
while slightly more potent, is nevertheless based on a single study and one
sex. The previous unit risk estimate, which was derived from the liver tumor
responses in the mouse by gavage exposure, is 1.3 x ICT6.
5-14
-------
No new epidemiologic data are suitable for risk analysis. Based on the
non-positive response observed in the epidemiologic study by Axelson et al.
(1978), an upper-bound estimate of risk due to 1 yg/m3 of TCI in air was
estimated to be approximately 1.7 x 10~5 (U.S. EPA, 1985a). This estimate pro-
vides no basis for postulating that the risk values derived from animal bio-
assay data seriously overestimate the human risk, this being the usual reason
for comparing risk derivations from both human and animal data. Since the
estimate from the human data is, in fact, the highest estimate in which several
major assumptions were made, the unit risks derived from the animal studies are
thought to provide a better basis for estimation of human risk.
5-15
-------
-------
6. SUMMARY AND CONCLUSIONS
6.1. QUALITATIVE
6.1.1. Background
This addendum updates a comprehensive health hazard review that was pre-
pared in the 1984-1985 period and was ultimately printed and distributed as a
Health Assessment for Trichloroethyiene in July, 1985. The addendum was under-
taken because several additional carcinogenicity bioassays have become avail-
able since the HAD was prepared. Notably, the inhalation animal experiments
recently reported are the first inhalation studies that are adequate for quan-
titation of risk. New oral exposure bioassays are also now available, thus
supplementing the oral bioassay data that was reported in the HAD. From a risk
estimation perspective, this addendum focuses on risk estimates for inhalation
exposure which supersede the 1985 HAD discussion of inhalation risks based on
extrapolation from gavage studies. The weight-of-evidence analysis for the
likelihood of TCI being a human carcinogen (e.g., classified as Group B2 in the
1985 HAD) is also reviewed in this addendum.
The evidence for the carcinogenicity of TCI in animals is quite broad,
as shown by tumor induction in male rats and in mice of both sexes and by
inhalation and oral exposure. The available metabolism and pharmacokinetic
information is sufficient to provide reasonable hypotheses for the carcinogenic
activity of TCI and to provide a basis for reducing some of the uncertainty
associated with dose-response extrapolation to humans. The available data and
interpretations from short-term tests and metabolism and kinetic studies are
consistent with the biological responses observed in the animal bioassays.
In terms of the weight of evidence for evaluating the likelihood of TCI
being a human carcinogen, the available data have been interpreted using the
6-1
-------
assessment criteria contained in EPA's Guidelines for Carcinogen Risk Assess-
ment (U.S. EPA, 1986). Similarly, the dose-response assessment and related
estimation of upper-bound cancer unit risk have been prepared in conformance
with the guidelines. As has been the case over the last few years, a specific
effort has been made to factor relevant metabolic and pharmacokinetic data into
the risk estimation process.
6.1.2. Weight of Evidence: Likelihood of Human Carcinogenic Potential
The available epidemiologic data remain inadequate either to demonstrate
or refute a carcinogenic potential for, TCI in humans. A new cohort study
reviewed in this addendum, Shindell and Ulrich (1985), is deemed inadequate
to refute or demonstrate a carcinogenic potential because the data do not
i
allow a validation of the reported absence of an effect in the cohort. The
1985 HAD concluded that six other epidemiologic studies or surveys were also
inadequate for characterizing the carcinogenic potential according to criteria
in the Guidelines for Carcinogen Risk [Assessment.
The currently available positive janimal data base is expanded relative to
i
the 1985 HAD with the addition of recently published inhalation and oral
studies. The highlights of the recent and older positive data include the
following information:
* Bioassays
Inhalation Exposure - Recent
Sprague-Dawley: male -SSII'of Leydig cell at all dose levels
(Maltoni et male - SSII of renal adenomas and carcinomas com-
al., 1986) bined, at high dose
male - elevated incidence of leukemia
female - slight increase in renal adenocarcinomas at
highest dose
Swiss mice: male - SSII of hepatomas at high dose
(Maltoni et male - SSIIof lung adenomas at two highest doses
al., 1986) female - elevated incidence of lung adenomas at high dose
6-2
-------
B6C3F1:
(Maltoni et
al., 1986)
ICR:
(Fukuda et
al., 1983)
male - SSII of hepatomas at mid dose, dose-related
increase at lower doses
female - marginally SSII of hepatomas at high dose,
dose-related increase
female - SSII of lung adenomas at high dose
female - SSII of lung adenocarcinomas at two highest
doses
Inhalation Exposure: 1985 HAD
B6C3F1:
(Bell et
al., 1978)
Han:NMRL:
(Henschler
et al., 1980)
male & r SSII of hepatocellular carcinomas at all dose
female levels, some uncertainty about administered
dose levels
female - SSII of malignant lymphomas in low and high
dose groups, role of TCI has some uncertainty
Oral Exposure - Recent
B6C3F1: male
(Herren-Freund
et al., 1986, 1987)
Rats-4 strains:
(NTP, 1987)
-SSII of liver carcinomas for DCA, TCA
metabolites
- deemed not acceptable bioassay by NTP
(although increased incidences of renal
tubular cell adenomas and carcinomas were
reported in male Osborne-Mendel rats, and
increased incidence of interstitial cell
tumors of testis in high-dose Marshall rats)
Oral Exposure - 1985 HAD
B6C3F1:
(NCI, 1976)
male &
female
Sprague-Dawley: male
(Maltoni, 1979;
Maltoni et al., 1986)
F344 Rat:
(NTP, 1982,
1986a)
B6C3F1:
(NTP, 1982,
1986a)
male
male &
female
-SSII of hepatocellular carcinomas
elevated incidence of leukemia (immunoblastic
lymphosarcomas)
-SSII of renal adenocarcinomas in high-dose
males (NTP said bioassay provides some evi-
dence)
- SSII of hepatocellular carcinomas
Human metabolic profile is shown to be qualitatively the same as test
animals. TCI is largely absorbed and metabolized. Some metabolites are
shown to be the active agents; others are thought to be active. Avail-
able metabolism data in both rodents and humans allows the estimation of
6-3
-------
an effective metabolite dose in animals and humans and explains some of
the observed species variability in tumor response. Four metabolites
of TCI are likely to be involved in cancer development in animals. Two
of the four, TCA and DCA, have been shown to be complete carcinogens in
male B6C3F1 mouse liver and also have been found in urine of humans ex-
posed to TCI.
An understanding of TCI's carcinogenic mechanisms is limited,,
Pour metabolites: TCI epoxide, DCA, TCA, DCVC may act in some combina-
tion or separately to induce cancer in animals.
i
Covalent binding of active TCI metabolites has been shown to occur and
is proportional to metabolism.!
TCI produces the same adduct in mouse liver as does the known human (and
animal) carcinogen vinyl chloride.
Cell transformation activity has been observed for TCI in rat embryo
cells, and for TCI oxide in hamster embryo cells.
The current knowledge of TCI metabolism and the acute cellular toxicity
of reactive metabolites suggest several cellular processes that may lead
to carcinogenic activity. Several mechanisms have been proposed. A
direct genotoxic mechanism for TCI is only weakly demonstrated in short
term assays which measure mutagenic or chromosomal responses.
The accumulated animal data, while categorized in the 1985 HAD as "suffici*
ent," i.e., Group B2, is significantly reinforced by the additional findings
in this addendum. A summary of the positive bioassay data by species is as
follows:
Rat
Inhalation and oral exposure produces both statistically significant and
marginally significant increases of renal adenocarcinomas and adenomas
in two strains of male rats: F344, Sprague-Dawley. (According to the
NTP [1986a], some evidence in F344; Osborne-Mendel is inconclusive,
although a response was observed [NTP, 1987].)
I .
Inhalation exposure produces statistically significant increases of
Leydig cell tumors in Sprague-Dawley rats.
Oral exposure produces Leydigi(interstitial) cell tumors in Marshall
rats, although the NTP (1987)[considers overall bioassay to be inade-
quate. |
Inhalation and oral exposure produces an elevation of leukemia in male
Sprague-Dawley rats. ;
: 6-4
-------
Mouse
« Inhalation exposure produces hepatomas and hepatocellular carcinomas in
two mouse strains, B6C3F1 and Swiss.
Oral exposure produces hepatomas and hepatocellular carcinomas in
multiple studies of B6C3F1 mice.
» Inhalation exposure produces primarily lung adenomas in B6C3F1 and
Swiss mice.
« Inhalation exposure produces lung adenocarcinomas in female ICR mice.
Inhalation exposure produces malignant lymphomas in female HAN:NMRL
mice.
« Oral exposure to two TCI metabolites produces hepatocellular carcinomas
in B6C3F1 mice.
Some negative TCI bioassays have been reported in rats and hamsters by inhala-
tion exposure and in rats and mice by oral exposure. Some of the negative
studies have limitations in dosing or conduct which limit their sensitivity to
detect carcinogenic activity. A positive response in one species/strain/sex is
not generally negated by negative results in other species/strain/sex unless
the negative studies are essentially identical in design and conduct.
The weight of evidence that an agent is potentially carcinogenic for
humans increases (1) with the increase in number of tissue sites affected;
(2) with the increase in number of animal species, strains, sexes and number of
experiments and doses showing a carcinogenic response; (3) with the occurrence
of cl^ar-cut dose response relationships as well as a high level of statistical
significance of the increased tumor incidence in treated compared to control
groups; (4) when there is a dose-related shortening of the time-to-tumor; and
(5) when there is a dose-related increase in the proportion of malignant tumors.
Sufficient evidence for experimental animals, weight-of-evidence classi-
fication Group B2, is shown when there is a statistically significant increased
incidence of malignant tumors or combined malignant and benign tumors (1) in
6-5
-------
multiple species or strains; (2) in multiple experiments, e.g., with different
routes of exposure or using differentjdose levels with the same route of expo-
sure; or (3) to an unusual degree in a single experiment with regard to high
i
incidence, unusual site or type of tumor, or early age at onset.
Table 6-1 provides a tabular comparison that matches the weight-of-evidence
criteria to the available positive animal evidence.
The Guidelines also take the posjition that when the only tumor response
is in the mouse liver and when other conditions for a classification of "suf-
ficient" evidence are met, the mouse [liver data should be considered as "suf-
I
ficient" evidence unless the following factors are observed, in which case
the mouse liver only data should be downgraded to "limited" [condition: data
observed],
i
(1) Increased incidence of tumors only in highest dose group and only
at end of study [dose-related trend in most dose groups; some sta-
tistically significant]
(2) No substantial dose-related; increase in the proportion of malig-
nant tumors [was a substantial dose-related increase in oral
experiments]
(3) Occurrence of tumors that are predominantly benign [higher percent-
age of carcinomas, e.g., NTP (1982) 27 (30%) adenomas, 93 (44%)
carcinomas]
(4) No dose-related shortening of the time to appearance of first
tumors [shortening shown in NCI, (1976) not shown in NTP (1982) oral
studies]
(5) Negative or inconclusive results from a spectrum of short-term
tests for mutagenic activity [parent compound weak to inconclusive
evidence; metabolites show cellular activity]
(6) The occurrence of excess tujnors only in a single sex [excess tumors
in both sexes]
The criteria for a sufficient leyel of animal evidence are satisfied for
both the mouse liver data alone and the broader data base. Pre-1985 and more
6-6
-------
TABLE 6-1. SUMMARY OF THE RELATIONSHIP BETWEEN SUFFICIENT
WEIGHT-OF-EVIDENCE FACTORS* AND THE BIOASSAY RESULTS
Weight-of-evidence
factorb
Bioassay results
In multiple species
In multiple strains
In multiple sexes
Male rats - Leydig cell, kidney tumors, and
suggestion of leukemia
Male and
female mice - Liver tumors, lung tumors,
lymphomas
Evidence of kidney tumors in 2 or 3 strains
of male rats (F344, Sprague-Dawley, Osborne-
Mendel); liver tumors in 2 strains of mice
(Swiss, B6C3F1); lung tumors in 3 strains of
mice (Swiss, B6C3F1, ICR); Leydig cell tumors
in 1 or 2 strains of rat (Sprague-Dawley,
Marshall)
Male and female - liver tumors in mice
Male and female - lung adenomas in mice
In multiple experiments
In experiments with
different routes of
exposure
Same route of exposure
with different dose
levels
Liver tumors in mice by gavage, inhalation
Liver tumors in mice; kidney tumors in
rats; elevation of leukemia in rats
Liver tumors in mice by inhalation
To an unusual degree with
regard to high incidence
To an unusual degree with
regard to site or type of
tumor
To an unusual degree with
regard to early age of onset
Some evidence of kidney tumors in male
rats
Some evidence of kidney tumors in male
rats
Liver tumors in male B6C3F1 mice with TCA,
DCA (metabolites of TCI)
*EPA Guidelines for Carcinogen Risk Assessment.
DSufficient evidence exists in animals when there is a statistically signifi-
cant increased incidence of malignant tumors or combined malignant and benign
tumors within any one of these factor groupings. Mouse liver tumor only
responses are weighted in a slightly different manner.
6-7
-------
recent information about TCI's metabolites and their reactivity is supportive
of the observed tumor responses and the "sufficient" ranking. There would seem
to be some structure-activity relationships emerging between. TCI and related
chlorinated hydrocarbons. Given the inadequacy of the epidemiologic data, the
criteria for a Group B2 weight-of-evidence classification are clearly met. A
Group B2 classification means that the likelihood of TCI being a human carcino-
gen is considered to be "probable." In the weight-of-evidence ranking scheme
the only higher classification is Group A which applies to a situation where
the evidence is considered to be "definite," while the next lower classifica-
tion is Group C, which means that the human evidence is inadequate, the animal
evidence is limited, and the overall likelihood of a human carcinogenic poten-
tial is considered to be "possible."
6.2. QUANTITATIVE
I
6.2.1. Estimate of Carcinogenic Potency (Unit Risk)
The dose-risk evaluation in this jaddendum is focused entirely on inhala-
tion exposure. Recently published bioassay reports for the first time provide
a basis for estimating the carcinogenijc risk of TCI to humans directly from
animal inhalation experiments, whereas, in the 1985 HAD, reliable positive data
were not available and inhalation risl< estimates were based on gavage exposure
responses to TCI.
There are seven animal inhalation bioassay data sets that lend themselves
to dose-risk evaluation with subsequent scaling and extrapolation to humans.
These data include Leydig cell tumors in male rats (Maltoni et al., 1986)s
lung and liver tumors in male Swiss mice and lung tumors in female Swiss mice
(Maltoni et al., 1986), lung and liver tumors in female B6C3F1 mice (Maltoni et
al., 1986), and lung tumors in female ICR mice (Fukuda et al., 1983). While
the kidney responses across several bioassays are thought to be biologically
6-8
-------
significant for qualitatively assessing the likelihood of human cancer, the
strength of the response data is less than for other tumor sites, and therefore
the kidney response is not included in the risk analysis.
The available information on metabolism and oncodynamics for animals and
humans is very useful in explaining some of the observed species variability in
tumor response, as well as providing a basis for defining an effective dose of
reactive agent in animals and an equivalent dose in humans. The metabolized
dose analysis is far superior to the more typical situation in which metabolism
in humans is dealt with solely by assumption.
Two scaling approaches are considered in the dose-risk analysis from the
animal data. One uses the body surface area dose equivalence (bw2/3) and the
other uses a body-weight dose equivalence (mg/kg)1. The body surface scaled
risks are about 13 and 6 times higher than the body weight scaled risks for the
mouse and rat, respectively. The body surface area approach for species-to-
species extrapolation is preferred for TCI since the Guidelines for Carcinogen
Risk Assessment recommend using this type of scaling if there is no scientific
evidence to justify using other scaling techniques; no such evidence exists for
TCI. In fact, for TCI the available metabolism data does correlate reasonably
well with the bw2/3 approach in the range of the bioassay doses. The fact that
one interspecies variability factor, metabolite production, correlates reason-
ably well with body surface dose equivalence may or may not account for the
full range of true interspecies variability.
The various potency values derived from the new inhalation data, together
with the value derived in the HAD from oral route extrapolation, are notably
consistent, the difference being less than a factor of four between the highest
(2.7 x 10~2) and the lowest (0.7 x 10~2). Given the narrow range of the qjf
values in mice for the lung and liver responses, a geometric mean for each tumor
6-9
-------
type is calculated across the different mouse strains. The one rat data set
with Leydig cell responses requires no averaging. Based on this aggregation,
the mouse lung and the rat Leydig celjl data are found to represent the most
sensitive tumorigenic responses among the species.
The actual unit risk is calculated by multiplying the qj (based upon
amount of TCI metabolized) by the amount of metabolite produced in a 70-kg
human at 1 ug/m3 of exposure to TCI. The resulting unit risk values, 1.7 x
10-6/pg/m3 based on mouse lung data, and 2.7 x 10-6/pg/m3 based on rat data,
are comparable to the inhalation unit risk of 1.3 x 10-6 based on data from
mouse gavage studies, as shown in the 1985 HAD.
6.2.2. Unit Risk Characterization :,
While the available risk estimates describe a narrow range and which could
be employed in future public health impact evaluations, the unit risk from the
mouse lung data, 1.7 x 10-6/pg/m3, is preferred. The risk of 1.7 x 10'6
reflects the impact of first-pass activation in pulmonary tissue, and is de-
rived from multiple mouse studies, strains, and sexes, whereas the rat estimate
of 2.7 x ID"6 is based on results from one study and in one sex.
The value of 1.7 x 10-6/mg/m3 p'f TCI is.an incremental risk that is an
upper-limit estimate derived from a linearized multistage extrapolation model,
currently programmed as part of GLOBAL86. Inherent in this analytical approach
is a question as to whether the upper-limit estimate is reasonable. Normally,
unit risk estimates based on animal jdata are thought to be plausible estimates
of upper-limit risk with a data base that is insufficient to identify specific
I
reasons for uncertainty. In the case of the inhalation unit risks for TCI,
the inability to group the mouse response data so that mice with more than one
tumor type can be incorporated in the dose-response analysis, suggests that the
unit risk of 1.7 x lO'6 based on lung tumors alone may somewhat underestimate
' 6-10
-------
the risk. From a different perspective, the use of animal and human metabolism
data reduces some of the uncertainty typically associated with dose-risk extrap-
olation. The HAD presented a "what-if" unit risk derivation from a negative
epidemiologic study in order to see if the negative human data, together with
exposure assumptions, would provide a unit risk that was comparable to the
animal estimates. The estimate from the human data of 1.7 x 10"5 per mg/m3 is
the highest unit risk even when compared to the newer inhalation risk estimates
contained in this addendum. Thus, as reported in the HAD, the human data
provide no evidence that the carcinogenic potency of TCI is less for humans
than would be predicted from animal studies.
The uncertainty associated with the dose equivalence between species is,
in general, debatable. The Guidelines recommend using the body surface dose
equivalence for interspecies dose equivalence if there is no scientific evi-
dence to justify using another approach. The available TCI data in fact
provide support for using the body surface equivalence to account for size and
metabolic equivalence.
The inhalation unit risk value of 1.7 x lO"6 is an upper-limit estimate of
the true human risk. A better estimate for human risk could only be provided
by having adequate human data to directly estimate a unit risk, or otherwise
providing a strong basis for checking the reasonableness of the animal-based
estimate* The use of alternative animal-to-human extrapolation models and
maximum likelihood estimates is not considered to have merit in describing a
better estimate of the true human risk.
The risk estimate for TCI is best viewed as a reasonable upper-limit
estimate for which the actual human risk, while not identifiable, would not be
expected to exceed the upper limit and, in fact, may be lower.
6-11
-------
-------
7. REFERENCES
Anderson, E.L.; Carcinogen Assessment Group, U.S. Environmental Protection
Agency. (1983) Quantitative approaches in use to assess cancer risk.
Risk Analysis 4:277-295.
Astrand, I.; Ovrum, P. (1976) Exposure to trichloroethylene. I. Uptake and
distribution in man. Scand. J. Work Environ. Health 4:199-211.
Axelsdn, 0.; Anderson, K. ; Hogstedt, C.; Holmberg, B.; Molina, 6.; de Verdier,
A. (1978) A cohort study on trichloroethylene exposure and cancer mor-
tality. J. Occup. Med. 20:194-196.
Banerjee, S. ; Van Duuren, B.L. (1978) Covalent binding of the carcinogen
trichloroethylene to hepatic microsomal proteins and to exogenous DNA
in vitro. Cancer Res. 38:776-780.
Beliles, R.P.; Parker, J.C. (1985) Oncodynamics and risk assessment of ethyl-
ene oxide. The lexicologist 6:280.
Bell,, Z.G.; Olson, K.H. ; Benya, T.J. (1978) Final report of audit findings
of the Manufacturing Chemists Association: administered trichloroethylene
chronic inhalation study at Industrial Biotest Laboratories, Inc.,
Decatur, IL. Unpublished.
Bergman, K. (1982) Reactions of vinyl chloride with RNA and DNA of various
mouse tissues in vivo. Arch. Toxicol. 49:117-129.
Bergman, K. (1983a) Interactions of trichloroethylene with DNA in vitro and
with RNA and DNA of various mouse tissues in vivo. Arch. ToxTcol I. 54:182-
193.
Bergman, K. (1983b) Applications of whole-body radiography in distribution
studies of organic solvents. Crit. Rev. Toxicol. 12:59-118.
Bond* JkA.; Dahl , A.R. ; Henderson, R.F.; Dutcher, J.S. ; Mauderly, J.L.;
Birfibaum, L.S. (1986) Species differences in the depositon of inhaled
butadiene. Toxicol. Appl . Pharmacol . 83:617-627.
Bubertj J. A.; 0 'Flaherty, E.J. (1985) Delineation of the role of metabolism
in the hepatotoxicity of trichloroethylene and perchloroethylene: a
dose-effect study. Toxicol. Appl. Pharmacol. 78:105-122.
Dalbey, W.; Bingham, E. (1978) Metabolism of trichloroethylene by the iso-
lated perfused lung. Toxicol. Appl . Pharmacol . 43:267-277.
Davidson, I.W.F.; Parker, J.C.; Beliles, R.P. (1986) Biological basis for
extrapolation across mammalian species. Reg. Toxicol. Pharmacol.
6:211-237.
7-1
-------
Dekant, W.; Metzler, M.; Henschler, D., (1984) Novel metabolites of tri-
chloroethylene through dechlorination reactions in rats, mice, and
humans. Biochem. Pharmacol. 33:2021-2027.
Dekant, W.; Metzler, M.; Henschler, D. (1986) Identification of S-1,2-
dichloro-N-acetyl-cystein as a urinary metabolite of trichloroethylene:
a possible explanation for its nephrocarcinogenicity in male rats.
Biochem. Pharmacol. 35:2455-2458.
Elcombe, C.R.; Pratt, I.S.; Green, T. (1982) The rate of trichloroacetic
acid (TCA) formation determines the species differences in hepatic
peroxisome proliferation due to trichloroethylene (TCI). The Toxicol-
ogist 2:173.
Elcombe, C.R.; Rose, M.S.; Pratt, I.SJ (1985) Biochemical, histological, and
ultrastructural changes in rat and mouse liver following the administra-
tion of trichloroethylene: possible relevance to species difference in
hepatocarcinogenicity. Toxicol. Appl. Pharmacol. 79:365-376.
Fernandez, J.G.; Droz, P.O.; Humbert, B.E.; Caperos, J.R. (1977) Trichloro-
ethylene exposure: simulation of'uptake, excretion, and metabolism using
a mathematical model. Br. J. Ind. Med. 34:43-55.
Fox, A.O.; Collier, P.F. (1976) Low mortality rates in industrial cohort
studies due to selection for work and survival in the industry. Br. J.
Social and Preventive Medicine 36:225-230.
Fukuda, K.; Takemoto, K.; Tsuruta, H. (1983) Inhalation carcinogenicity of
trichloroethylene in mice and rats. Ind. Health 21:243-254.
Carman, R.H.; Snellings, W.M.; Maronpot, R.R. (1985) Brain tumors in F344
rats associated with chronic inhalation exposure to ethylene oxide.
Neurotoxicology 6:117-138.
Goldsworthy, T.L.; Popp, J.A. (1987) Chlorinated hydrocarbon-induced perox-
isomal enzyme activity in relation to species and organ carcinogenicity.
TAP (in press).
Green, T.; Prout, M.S. (1985) Species differences in response to trichloro-
ethylene. II. Biotransformation in rats and mice. Toxicol. Appl.
Pharmacol. 79:401-411.
Grisham, J.W.; Smith, G.J.; Bellies, K.E. (1984) Comparison of analysis and
quantification of cell death in Vivo and in vitro. Proceedings of the
15th Conference on Environmental toxicolology., pp. 158-168.
Henschler, D.; Romen, W.; Elsasser, H.M.; Reichert, D.; Eder, E..; Radwan, Z.
(1980) Carcinogenicity study of trichloroethylene by long-term Inhala-
tion in the animal species. Arch. Toxicol. 43:237-248.
Henschler, D.; Elsasser, H.; Romen, W;.; Elder, E. (1984) Carcinogenicity
study of trichloroethylene, with and without epoxide stabilizers, in
mice. J. Cancer Res. Clin. Oncol. 104:149-156.
7-2
-------
Herren-Freund, S.L.; Pereira, M.A.; 01 sen, G. (1987) The carcinogenicity of
trichloroethylene and its metabolites, trichloroacetic acid, dichloro-
acetic acid in the mouse liver, toxicol. Appl. Pharmacol. (in press).
Herren-Freund, S.L.; Pereira, M.A.; Olsen, G.; DeAngelo, A.B. (1986) The
carcinogenicity of trichloroethylene and its metabolites, trichloroacetic
acid, dichloroacetic acid in the mouse liver. Proc. Am. Assoc. Cancer
Res. 27:91.
Howe, R.B.; Crump, K.S.; Landingham, C.V. (1986) GLOBAL86: a computer pro-
gram to extrapolate animal toxicity data to low doses. Prepared for U.S.
Environmental Protection Agency under subcontract to Research Triangle
Institute. Contract No. 69-01-6826.
Ikeda, M.; Ohtsuji, H.; Imamura, T.; Komoike, Y. (1972) Urinary excretion
of total trichloro-compounds, trichloroethanol, and trichloroacetic acid
as a measure of exposure to trichloroethylene and tetrachloroethylene.
Brit. J. Ind. Med. 29:328-333.
International Commission on Radiological Protection (ICRP). (1977) Recommen-
dation of the International Commission on Radiological Protection. Pub.
No. 26, adopted Jan. 17, 1977. Oxford, England: Pergamon Press.
Koizumi, A.; Kastl, P.E.; Reitz, R.H.; Nolan, R.J. (1987) Fate of 14C-tri-
chloroethylene administered to rats in drinking water. In preparation.
Maltoni, C. (1979) Results of long-term carcinogenicity bioassays of tri-
chloroethylene experiments by oral administration on Sprague-Dawley
rats. Unpublished.
Maltoni, C.; Lefemine, G.; Cotti, G. (1986) Experimental research on tri-
chloroethylene carcinogenesis. In: Maltoni, C.; Mehlman, M.A., eds.
Archives of research on industrial carcinogenesis, Vol. V. Princeton,
NJ: Princeton Scientific Publishing Co.
Mennear, J.; Haseman, J.; Sullivan, D.; Bernal, E.; Hildebrandt, P. (1982)
Studies on the carcinogenicity of pentachloroethane in rats and mice.
Fundam. Appl. Toxicol. 2:82-87. :
Monster, A.C. (1979) Difference in uptake, elimination, and metabolism in
exposure to trichloroethylene, 1,1,1-rtrichloroethane, and tetrachloro-
ethylene. Int. Arch. Occup. Environ. Health 42:311-317.
Monster, A.C.; Boersman, G.; Duba, W.C. (1976a) Pharmacokinetics of trichlro-
ethylene in volunteers: influence of workload and exposure concentration.
Int. Arch. Occup. Environ. Health 38:87-102.
Monster * A.C.; Boersman, G.; Duba, W.C. (1976b) Kinetics of trichloroethylene
1h repeated exposure of volunteers. Int. Arch. Occup. Environ. Health
42:283-292.
7-3
-------
-------
-------
li
0) o
en
fl)
me
(033
CD < y
3 5 »
a I
(B 0)
2
------- |