DRAFT - DO NOT CITE OR QUOTE EPA/635/R-08/011A
External Review Draft
f/EPA
TOXICOLOGICAL REVIEW
OF
Tetrachloroethylene
(Perchloroethylene)
(CAS No. 127-18-4)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
June 2008
NOTICE
This document is an External Review Draft. This information is distributed solely for the
purpose of pre-dissemination peer review under applicable information quality guidelines. It has
not been formally disseminated by EPA. It does not represent and should not be construed to
represent any Agency determination or policy. It is being circulated for review of its technical
accuracy and science policy implications.
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of pre-dissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and
should not be construed to represent any Agency determination or policy. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 ii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTSCTOXICOLOGICAL REVIEW for TETRACHLOROETHYLENE
(PERCHLOROETHYLENE) (CAS No. 127-18-4)
LIST OF TABLES ix
LIST OF FIGURES xii
LIST OF ABBREVIATIONS AND ACRONYMS xvi
FOREWORD xix
AUTHORS, CONTRIBUTORS, AND REVIEWERS xx
1. INTRODUCTION 1
REFERENCES FOR CHAPTER 1 3
2. BACKGROUND 1
2.1. USES AND PHYSICAL/CHEMICAL PROPERTIES 1
2.2. OCCURRENCE AND EXPOSURE 1
2.2.1. Air 1
2.2.2. Water 4
2.2.3. Food 5
2.2.4. Breast Milk 6
2.2.5. Direct Ingestion 7
REFERENCES FOR CHAPTER 2 8
3. TOXICOKINETICS 1
3.1. ABSORPTION 1
3.1.1. Inhalation 1
3.1.2. Oral 2
3.1.3. Dermal 2
3.2. DISTRIBUTION AND BODY BURDEN 2
3.3. METABOLISM 4
3.3.1. Introduction 4
3.3.2. Extent of Metabolism 4
3.3.3. Pathways of Metabolism 5
3.3.3.1. Cytochrome P450-Dependent Oxidation 6
3.3.3.2. Glutathione (GSH) Conjugation Pathway 12
3.3.3.3. Relative Roles of the Cytochrome P450 (CYP) and Glutathione
(GSH) Pathways 18
3.3.4. Susceptibility 19
3.3.5. Comparison of Tetrachloroethylene Metabolism with Trichloroethylene
Metabolism 20
3.4. EXCRETION 20
3.5. PHYSIOLOGICALLY BASED AND OTHER TOXICOKINETIC MODELING 23
3.5.1. Various Physiologically Based Pharmacokinetic (PBPK) Models 23
3.5.2. Variability and Uncertainty 28
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 iii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
3.5.3. Animal-to-Human Extrapolation Using a Physiologically Based
Pharmacokinetic (PBPK) Model 33
3.5.3.1. Choice of Physiologically Based Pharmacokinetic (PBPK) Model 33
3.5.3.2. Implementation of Physiologically Based Pharmacokinetic
(PBPK) Models 36
3.5.4. Comparison of Physiologically Based Pharmacokinetic (PBPK)
Simulations With Experimental Data 39
3.5.5. Physiologically Based Pharmacokinetic (PBPK) Model Comparisons
and Interspecies Differences 47
3.5.6. Metabolic Interactions With Other Chemicals 56
APPENDIX FOR CHAPTER 3: COMPARISONS OF TETRACHLOROETHYLENE
METABOLISM WITH TRICHLOROETHYLENE METABOLISM 58
REFERENCES FOR CHAPTER 3 62
4. HAZARD IDENTIFICATION 1
4.1. OVERALL APPROACH 1
4.2. OVERVIEW OF TETRACHLOROETHYLENE METABOLISM 1
4.3. GENOTOXICITY 3
4.4. LIVER TOXICITY 7
4.4.1. Human Effects 7
4.4.1.1. Liver Damage 7
4.4.1.2. Liver Cancer 11
4.4.2. Animal Studies 11
4.4.2.1. Liver Toxicity 11
4.4.2.2. Liver Cancer 13
4.4.3. Summary of Liver Effects in Humans and Animals 15
4.4.4. Mode of Action for Liver Toxicity 16
4.4.4.1. Background 17
4.4.4.2. Relationship of Metabolism to Potential Mode of Action and
Organ Toxicity 18
4.4.4.3. Description of a Hypothesized Mode of Action (MOA):
Peroxisome Proliferator-Activated Receptor (PPAR) Mediated
Hepatocarcinogenesis 19
4.4.4.4. Effects That Could Be Related to Other Potential Modes of Action. ...32
4.4.4.5. Summary and Conclusions 34
4.5. KIDNEY TOXICITY 35
4.5.1. Human Studies 35
4.5.1.1. Kidney Toxicity in Humans 35
4.5.1.2. Kidney Cancer 37
4.5.2. Animal Studies 39
4.5.2.1. Kidney Toxicity in Animals 39
4.5.2.2. Kidney Cancer in Animals 40
4.5.3. Summary of Kidney Effects in Humans and Animals 41
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 iv DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
4.5.4. Mode of Action for Kidney Toxicity and Carcinogenicity 42
4.5.4.1. Background 42
4.5.4.2. Summary Description of a Postulated Mode of Action—
alpha-2u-globulin Accumulation 43
4.5.4.3. Other Modes of Action for Tetrachloroethylene-Induced Renal
Tumors in Rats 48
4.5.4.4. Summary 51
4.6. NEUROTOXICITY 52
4.6.1. Human Studies 52
4.6.1.1. Environmental Chamber Studies 58
4.6.1.2. Chronic Exposure Studies 63
4.6.1.3. Summary of Neuropsychological Effects in Low- and Moderate-
Exposure Studies 96
4.6.2. Animal Studies 104
4.6.2.1. Inhalation Studies 104
4.6.2.2. Oral andIntraperitoneal Studies Ill
4.6.3. Summary of Neurotoxic Effects in Humans and Animals 114
4.6.4. Mode of Action for Neurotoxic Effects 117
4.7. DEVELOPMENTAL/REPRODUCTIVE STUDIES 118
4.7.1. Human Studies 118
4.7.2. Animal Studies 124
4.7.2.1. Summary of Animal Studies 134
4.7.3. Summary of Human and Animal Developmental/Reproductive Studies 134
4.7.4. Mode of Action for Developmental Effects 137
4.8. TOXIC EFFECTS IN OTHER ORGAN SYSTEMS 138
4.8.1. Human Studies 138
4.8.1.1. Noncancer Effects 138
4.8.1.2. Cancer 146
4.8.2. Animal Studies 156
4.8.2.1. Noncancer Effects 156
4.8.2.2. Cancer Effects 158
4.8.3. Summary of Immunotoxicologic Effects in Humans and Animals and
Potential Mode of Action 163
4.9. SUSCEPTIBLE POPULATIONS 164
4.9.1. Life Stages 165
4.9.1.1. Life Stage-Specific Exposures 165
4.9.1.2. Early Life Stage Effects 166
4.9.1.3. Later Life Stages 173
4.9.2. Other Susceptibility Factors 173
4.9.2.1. Health and Nutritional Status 173
4.9.2.2. Gender 174
4.9.2.3. Race/Ethnicity 175
4.9.2.4. Genetics 175
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 v DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
4.9.3. Multiple Exposures and Cumulative Risks 176
4.9.4. Uncertainty of Database for Susceptible Populations 178
4.9.4.1. Uncertainties of Exposure 178
4.9.4.2. Uncertainties of Effects 180
4.9.5. Conclusions on Susceptibility 180
4.10. SUMMARY OF HAZARD IDENTIFICATION 182
4.10.1. Description of Effects and Exposure Levels at Which They Occur 182
4.10.1.1. Summary of Effects in Humans 182
4.10.1.2. Summary of Effects in Animals 187
4.10.1.3. Summary of Effect Levels 189
4.10.2. Characterization of Cancer Hazard 190
4.10.2.1. Background 190
4.10.2.2. Hazard Characterization for Tetrachloroethylene 191
4.10.3. Mode-of-Action Summary 193
4.10.4. Rationale for Select!on of Dose Metric 201
4.10.4.1. Liver 201
4.10.4.2. Kidney 201
4.10.4.3. Hematopoietic Target Organ 202
4.10.4.4. Central Nervous System 203
APPENDIX 4A: CONSISTENCY OF TETRACHLOROETHYLENE AND
TRICHLOROACETIC ACID HEPATOCARCINOGENICITY 204
APPENDIX 4B: HUMAN STUDIES OF CANCER 215
REFERENCES FOR CHAPTER 4 219
5. DOSE-RESPONSE EVALUATION 1
5.1. INHALATION REFERENCE CONCENTRATION (RFC) 1
5.1.1. Choice of Principal Study and Critical Effect 1
5.1.2. Method of Analysis 10
5.1.3. Reference Concentration (RfC) Derivation, Including Application of
Uncertainty Factors 11
5.1.4. Supporting Studies 15
5.1.5. Previous Inhalation Assessment 19
5.2. ORAL REFERENCE DOSE (RFD) 20
5.2.1. Choice of Principal Study and Critical Effects 20
5.2.2. Methods of Analysis, Including Models 21
5.2.3. Reference Dose (RfD) Derivation, Including Application of Uncertainty
Factors 24
5.2.4. Supporting Studies 27
5.2.5. Previous Oral Assessment 29
5.3. UNCERTAINTIES IN INHALATION REFERENCE CONCENTRATION
(RFC) AND ORAL REFERENCE DOSE (RFD) 30
5.3.1. Point of Departure 32
5.3.2. Extrapolation from Laboratory Animal Studies to Humans 32
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 vi DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
5.3.3. Human Variation 33
5.3.4. Database Uncertainties 33
5.4. CANCER DOSE-RESPONSE ASSESSMENT 34
5.4.1. Choice of Study/Data with Rationale and Justification 35
5.4.2. Dose-Response Data 36
5.4.2.1. Liver Tumors in Mice 36
5.4.2.2. Mononuclear Cell Leukemia in Rats 41
5.4.2.3. Other Tumor Sites in Male Rats 45
5.4.3. Estimation of Dose Metrics for Dose-Response Modeling 46
5.4.3.1. Dose Metric for Hepatocellular Carcinogenicity 46
5.4.3.2. Dose Metric for Rat Leukemias and Kidney Tumors 48
5.4.3.3. Dose Metric for Sites Not Addressed by Physiologically Based
Pharmacokinetic (PBPK) Modeling 49
5.4.4. Extrapolation Methods 49
5.4.4.1. Dose-Response Models and Extrapolation to Low Doses 49
5.4.4.2. Extrapolation to Human Equivalent Environmental Exposure 51
5.4.5. Cancer Risk Values 55
5.4.5.1. Dose-Response Modeling Results 55
5.4.5.2. Recommended Inhalation Unit Risk 71
5.4.5.3. Recommended Oral Slope Factor 73
5.4.5.4. Quantitative Adjustment for Sensitive Populations 73
5.4.6. Discussion of Uncertainties in Cancer Risk Values 75
5.4.6.1. Sources of Uncertainty 75
5.4.6.2. Summary and Conclusions 85
APPENDIX 5A: BENCHMARK DOSE MODEL RESULTS 86
APPENDIX 5B: PROBABILITY DISTRIBUTIONS OF CANCER RISK ESTIMATES 96
REFERENCES FOR CHAPTER 5 102
6. CHARACTERIZATION OF HAZARD AND DOSE-RESPONSE 1
6.1. SUMMARY OF HUMAN HAZARD POTENTIAL 1
6.1.1. Exposure 1
6.1.2. Absorption, Metabolism, Distribution, and Excretion 1
6.1.3. Noncancer Toxicity in Humans and Laboratory Animals 4
6.1.4. Carcinogenicity in Humans and Laboratory Animals 5
6.1.5. Mode-of-Action Information 7
6.1.5.1. Liver Mode-of-Action Information 7
6.1.5.2. Kidney Mode-of-Action Information 8
6.1.5.3. Mode-of-Action Information for Other Targets of Toxicity 8
6.1.5.4. Mode-of-Action Conclusions and Implications for Dose-Response
Analyses 9
6.1.6. Weight-of-Evidence Descriptor for Cancer Hazard 9
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 vii DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS (continued)
6.2. DOSE-RESPONSE CHARACTERIZATION 11
6.2.1. Noncancer Toxicity (Reference Concentration [RfC]/Reference Dose
[RfD]) 11
6.2.1.1. Assessment Approach Employed 12
6.2.1.2. Impact of Assumptions, Uncertainties and Alternatives on
Reference Concentration and Reference Dose 15
6.2.2. Cancer Risk Estimates 21
6.2.2.1. Assessment Approach Employed 22
6.2.2.2. Impact of Assumptions, Uncertainties and Alternatives on Unit
Risk Estimates 24
6.2.2.3. Quantitative Analysis of Multiple Uncertainties on Cancer Unit
Risk 32
6.2.2.4. Conclusions 35
REFERENCES FOR CHAPTER 6 40
APPENDIX A: SUMMARY OF EARLIER ASSESSMENTS A-l
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 viii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES
2-1. Physical and chemical properties of tetrachloroethylene 2-2
3-1. Comparison of Vmax/Km ratios 3-25
3-2. Variation in values of metabolic parameters for tetrachloroethylene, as seen
in the literature 3-30
3-3. Ratio of average daily dose at various life stages to the average daily dose
for a 25-year-old adult: PBPK simulations 3-31
3-4. Parameters for tetrachloroethylene PBPK modeling 3-40
3-5. Comparison of venous blood tetrachloroethylene concentrations: PBPK
simulations and Altmann et al. (1990) study 3-46
4-1. Summary of studies of human liver toxicity 4-10
4-2. Summary of rodent liver toxicity studies 14
4-3. Summary of human kidney toxicity marker studies in dry cleaners 4-38
4-4. Summary of human neurotoxicology studies 4-54
4-5. Summary of neuropsychological effects of tetrachloroethylene in humans 4-97
4-6. Summary of animal inhalation neurotoxicology studies 4-109
4-7. Summary of oral neurotoxicity animal studies 4-115
4-8. Developmental/reproductive studies in humans 4-125
4-9. Exposure concentrations (ppm) at which effects occurred in a two-
generation study 4-132
4-10. Summary of animal developmental/reproductive studies for
tetrachloroethylene, in chronological order 4-135
4-11. Immune-related conditions in studies of dry cleaning or tetrachloroethylene
exposure in humans 4-142
4-12. Summary of low-effect levels of exposure to tetrachloroethylene 4-189
4-13. Summary of potential modes of action for cancer 4-196
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 ix DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
4-14. Quantitative implications of different modes of action: candidate modeling
approaches 4-199
5-1. Summary of rationale for principal study selection 5-3
5-2. Inhalation studies considered in the development of an RfC 5-5
5-3. Oral studies considered in analysis of the oral RfD 5-25
5-4. Oral RfV: point of departure and uncertainty factors 5-26
5-5. Tumor incidence and estimated metabolized doses in mice exposed to
tetrachloroethylene 5-37
5-6. Historical control data of the Japan Bioassay Research Center, Crj/BDFl
mouse, 104-week studies 5-39
5-7. Incidence of mononuclear cell leukemia, kidney tumors, and brain gliomas
in rats exposed to tetrachloroethylene by inhalation 5-42
5-8. Historical control data of the Japan Bioassay Research Center, F344/DuCrj
(Fischer) rat, 104-week studies 5-43
5-9. Dose-response modeling summary for tumor sites using total
tetrachloroethylene metabolites as the dosimeter; tumor incidence data
from JISA (1993) and NTP (1986) 5-62
5-10. Human equivalent risk per unit concentration, in terms of continuous
environmental exposure, derived using total tetrachloroethylene
metabolites as the dosimeter; tumor incidence data from JISA (1993)
and NTP (1986) 5-63
5-11. Dose-response summary and cancer risk estimates using continuous
equivalent administered tetrachloroethylene levels as dosimeter, from
NTP (1986) and JISA (1993) 5-64
5-12. Summary of tetrachloroethylene oral slope factors, estimated from dose-
response modeling of inhalation-exposed animals and by extrapolation to
oral exposure using pharmacokinetic models 5-74
5-13. Summary of uncertainties in tetrachloroethylene cancer unit risk estimate 5-76
5-14. Summary of considerations for each rodent tumor type 5-82
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 x DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF TABLES (continued)
6-1. Summary of rationale for principal study selection 6-13
6-2. Summary of dose-specific extra risks (means and 95% confidence limits)
for four dose-response models fit to incidence of leukemias in male rats
exposed to tetrachloroethylene via inhalation 6-27
6-3. Summary of considerations for each rodent tumor type 6-30
6-4. Combined impact on tetrachloroethylene cancer risk estimates (per |ig/m3)
of statistical uncertainty, PBPK model and tumor site(s), using multistage
model in observed range and linear low-dose extrapolation 6-34
6-5. Considerations leading to the determination of a reasonable upper bound on
risk 6-37
6-6. Considerations that impact uncertainty in reasonable upper bound risk
estimates 6-38
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xi DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
3-1. Postulated scheme for the metabolism of tetrachloroethylene by the P450
oxidative pathway 3-7
3-2. Metabolism of tetrachloroethylene by the glutathione conjugation pathway 3-13
3-3. PBPK simulations of variations with age and gender in blood concentrations
of tetrachloroethylene and its main metabolite trichloroacetic acid (TCA) 3-32
3-4. Comparison of model predictions for blood concentration with inhalation
experiment 3-41
3-5. Comparison of model predictions for alveolar concentration of
tetrachloroethylene with experimental data on humans 3-42
3-6a. Comparison of model predictions for alveolar concentration as a fraction of
inhaled tetrachloroethylene concentration with experimental Opdam and
Smolders (1986) data on male human subjects 3-44
3-6b. Comparison of model predictions for alveolar concentration as a fraction of
inhaled tetrachloroethylene concentration with experimental data on female
human subjects 3-45
3-7. Comparison of tetrachloroethylene concentrations in blood in rats and humans.... 3-48
3-8. Comparison of various model predictions of tetrachloroethylene blood
concentration in humans and rats following steady state 3-49
3-9. Model predictions of total tetrachloroethylene metabolites produced
following a 6-hr inhalation exposure in rats, mice, and humans 3-50
3-10. Model predictions of rate of total metabolism in humans at steady state 3-51
3-1 la. Rate of metabolism in rat and human models: time course for low exposure 3-53
3-1 Ib. Rate of metabolism in rat and human models: time course for high exposure 3-54
3-12. Oral ingestion of tetrachloroethylene: blood concentration in humans
versus time 3-55
3-13. Rate of metabolism of tetrachloroethylene in humans: oral exposure 3-56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continuted)
4-1. Visual contrast sensitivity functions for control and exposed children (top),
adults that were identified as having impaired function (i.e., 5 of the total 11)
and their matched controls (middle), and the control and exposed individuals
over 60 years of age 4-85
4-2. Summary of the relationship between LOAEL concentrations (ppm) and
treatment duration (hours) 4-111
5-1. Array of PODs and reference values for a subset of neurotoxic effects of
studies in Table 5-2 5-16
5-2. Organ-specific reference values for inhalation exposure to tetrachloroethylene .... 5-19
5-3. Time course of venous blood concentration in humans as predicted by the
Bois et al. (1996), Rao and Brown (1993), and Reitz et al. (1996) PBPK
models for ingested tetrachloroethylene 5-23
5-4. Oral organ-specific reference values for exposure to tetrachloroethylene 5-28
5-5. Mouse liver tumor responses (hepatocellular adenomas and carcinomas) for
three chronic bioassays (Table 5-5), plotted against continuous equivalent
concentration (ppm) and total tetrachloroethylene metabolism
(mg-equivalents/kg-day), for male and female mice 5-40
5-6. Rat mononuclear cell leukemia responses (minus control) in two chronic
bioassays (Table 5-7), plotted against continuous equivalent exposure
(ppm) and total tetrachloroethylene metabolites, in mg-equivalents/kg-day,
for male and female rats 5-44
5-7. Sequence of steps for extrapolating from tetrachloroethylene bioassays in
animals to human-equivalent exposures expected to be associated with
comparable cancer risk 5-47
5-8a. Incidence of hepatocellular adenomas and carcinomas in male mice (JISA,
1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk 5-56
5-8b. Incidence of hepatocellular adenomas and carcinomas in male mice (JISA,
1993) corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at
10% extra risk 5-56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xiii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continuted)
5-9a. Incidence of hepatocellular adenomas and carcinomas in female mice
(JISA, 1993) corresponding to total tetrachloroethylene metabolism
(mg-eq/kg-day) and multistage model fit showing BMC and BMCL at
10% extra risk 5-57
5-9b. Incidence of hepatocellular adenomas and carcinomas in female mice
(JISA, 1993) corresponding to human equivalent continuous
tetrachloroethylene exposure (ppm) and multistage model fit showing
BMC and BMCL at 5% extra risk 5-57
5-10a. Incidence of malignant hemangiosarcomas in male mice (JISA, 1993)
corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk 5-58
5-10b. Incidence of malignant hemangiosarcomas in male mice (JISA, 1993)
corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at
5% extra risk 5-58
5-1 la. Incidence of mononuclear cell leukemia in male rats (JISA, 1993)
corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk 5-59
5-1 Ib. Incidence of mononuclear cell leukemia in male rats (JISA, 1993)
corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at
5% extra risk 5-59
5-12a. Incidence of mononuclear cell leukemia in female rats (JISA, 1993)
corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk 5-60
5-12b. Incidence of mononuclear cell leukemia in female rats (JISA, 1993)
corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at
5% extra risk 5-60
5-13a. Incidence of kidney adenomas and adenocarcinomas in male rats (NTP,
1986) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 5% extra risk 5-61
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xiv DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF FIGURES (continuted)
5-13b. Incidence of kidney adenomas and adenocarcinomas in male rats (NTP,
1986) corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at
5% extra risk 5-61
5-14. Comparison of inhalation risks per unit concentration for tetrachloroethylene
derived from rodent bioassays using four different dose metrics 5-72
5-15. Illustration of sensitivity to model selection for low-dose extrapolation 5-80
6-1. Array of PODs and reference values for a subset of neurotoxic effects in
inhalation studies 6-16
6-2. Organ-specific RfVs for inhalation exposure to tetrachloroethylene 6-17
6-3. Oral organ-specific reference values for exposure to tetrachloroethylene 6-18
6-4. Cancer risk estimates for tumor sites associated with tetrachloroethylene
exposure in rodent bioassays, using the multistage model 6-33
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xv DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS
8-OHdG 8-hydroxydeoxyguanosine
AAP alanine aminopeptidase
ALT alanine transferase
AST aspartase amino transaminase
ATSDR Agency for Toxic Substances and Disease Registry
AUC area-under-the-curve
BMC benchmark concentration
BMCL 95% lower bound benchmark dose
BMD benchmark dose
BMDS Benchmark Dose Software
BMDU 95% upper bound benchmark dose
BUN blood urea nitrogen
BW body weight
CARB California Air Resources Board
CASRN Chemical Abstracts Service Registry Number
CCI Color Confusion Index
CI confidence interval
CLL chronic lymphocytic leukemia
CNS central nervous system
CC>2 carbon dioxide
CT carbon tetrachloride
CYP cytochrome P450
CYP P450 cytochrome P450
DCA dichloroacetic acid
DEHP di(2-ethylhexyl)phthalate
EEGs electroencephalograms
EPA U.S. Environmental Protection Agency
FDA Food and Drug Administration
FMO3 flavin-containing monooxygenase 3
GOT gamma-glutamyltransferase
GSH glutathione
GST glutathione S-transferase
GSTx glutathione S-transferase isoform, where x denotes different isoforms (such as M,
T, P, S, Z)
FIEC human equivalent concentration
HSIA Halogenated Solvents Industry Alliance
i.p. intraperitoneal
IAP intestinal alkaline phosphatase
IARC International Agency for Research on Cancer
IOM Institute of Medicine
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xvi DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
IPCS International Programme on Chemical Safety
IRIS Integrated Risk Information System
IUGR intrauterine growth restriction
JISA Japan Industrial Safety Association
Km Michaelis-Menten constant
LECioS 95% lower confidence limits on the air concentrations associated with a 10%
extra risk of cancer incidence
LGL Large granular lymphocyte
LOAEL lowest-observed-adverse-effect level
MLE maximum likelihood estimate (please verify inserted correctly on pg 5-69, MLE
was there but not the exact definition)
MCA monochloroacetic acid
MCL-5 microsomal epoxide hydrolase
MCL mononuclear cell leukemia
MOA mode of action
MRL minimal risk level
NAG N-acetyl-p-D-glucosaminidase
NCI National Cancer Institute
NHL non-Hodgkin's lymphoma
NIOSH National Institutes of Occupational Safety and Health
NK natural killer
NOAEL no-observed-adverse-effect level
NRC National Research Council
NTP National Toxicology Program
NYS DOH New York State Department of Health
NYS O AG New York State Office of Attorney General
OR odds ratio
P450 cytochrome P450
PBPK physiologically based pharmacokinetic
PCO palmitoyl CoA oxidation
PHG public health goal
POD point of departure
PPAR peroxisome proliferater activated receptor
PPAR-a peroxisome proliferater activated receptor, alpha isoform
PPAR-5 peroxisome proliferater activated receptor, delta isoform
RBP retinol binding protein
REAL revised European-American Lymphoma
RfC reference concentration
RfD reference dose
RfV reference value
RR relative risk
SAP Scientific Advisory Panel
SCE sister chromatid exchange
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xvii DRAFT—DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
SES socio-economic status
SGA small for gestational age
SIR standardized incidence ratio
SMR standardized mortality ratio
SSB single-strand breaks
TCA trichloroacetic acid
TCE trichloroethylene
TCOH Trichloroethanol
TCVC S-(l,2,2,-trichlorovinyl)-L-cysteine
TCVCSO S-(l,2,2,-trichlorovinyl)-L-cysteine sulfoxide
TCVG S-(l,2,2-trichlorovinyl) glutathione
TNAP tissue non-specific alkaline phosphatase
TWA time-weighted average
U/L international units per liter
UDS unscheduled DNA synthesis
UF uncertainty factor
VCS visual contrast sensitivity
VE ventilation rate
VEP visually evoked potential
Vmax maximum velocity
WHO World Health Organization
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xviii DRAFT—DO NOT CITE OR QUOTE
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to
tetrachloroethylene. It is not intended to be a comprehensive treatise on the chemical or
toxicological nature of tetrachloroethylene.
In Chapter 6, Characterization of Hazard and Dose-Response, the United States
Environmental Protection Agency (EPA) has characterized its overall confidence in the
quantitative and qualitative aspects of hazard and dose-response by addressing knowledge gaps,
uncertainties, quality of data, and scientific controversies. The discussion is intended to convey
the limitations of the assessment and to aid and guide the risk assessor in the ensuing steps of the
risk assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xix DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER
KathrynZ. Guy ton1
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Karen A. Hogan1
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Robert E. McGaughy2
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
AUTHORS
Stanley Bar one
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Rebecca Brown
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Glinda Cooper
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Nagalakshmi Keshava
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
1 Chemical Manager since September 2007.
2 Retired prior to final revisions to document.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xxi DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
Leonid Kopylev
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Susan Makris
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Jean Parker2
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Cheryl Scott
National Center for Environmental Assessment
Office of Research and Development
Washington, DC
Ravi Subramaniam
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Larry Valcovic2
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
CONTRIBUTORS
Nancy Beck
AAAS Fellow
U.S. Environmental Protection Agency
Washington, DC
David Bussard
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xxi DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
Jane Caldwell
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Weihsueh Chiu
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Deborah Rice
Formerly with National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Marc Rigas
Formerly with Environmental Sciences Division
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Las Vegas, NV
Bob Sonawane
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Paul White
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
REVIEWERS
This document and the accompanying IRIS Summary have been peer reviewed by EPA
scientists and independent scientists external to the EPA. Comments from all peer reviewers
have been evaluated carefully and considered by the EPA during the preparation of this external
review draft. During the preparation of this draft, the IRIS Program Director achieved common
understanding of the assessment among the Office of Research and Development; the Office of
Air and Radiation; the Office of Prevention, Pesticides, and Toxic Substances; the Office of
Solid Waste and Emergency Response; the Office of Water; the Office of Policy, Economics,
and Innovation; the Office of Children's Health Protection; the Office of Environmental
Information, and the EPA's regional offices.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xxii DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
INTERNAL PEER REVIEWERS
Hugh Barton, U.S. Environmental Protection Agency, National Health and Environmental
Effects Research Laboratory, Research Triangle Park, NC
Robert Benson, U.S. Environmental Protection Agency, Office of Partnerships and Regulatory
Assistance, Region 8, Denver, CO
William Boyes, U.S. Environmental Protection Agency, National Health and Environmental
Effects Research Laboratory, Research Triangle Park, NC
Jane Caldwell, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Research Triangle Park, NC
Jim Cogliano, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Washington, DC
Herman Gibb, formerly with U.S. Environmental Protection Agency, National Center for
Environmental Assessment, Washington, DC
John Lipscomb, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Cincinnati, OH
Elizabeth Margosches, U.S. Environmental Protection Agency, Office of Pollution, Prevention,
and Toxics, Washington, DC
Dierdre Murphy, U.S. Environmental Protection Agency, Office of Air Quality and Planning and
Standards, Research Triangle Park, NC
Onyemaechi Nweke, U.S. Environmental Protection Agency, Office of Policy, Economics, and
Innovation, Washington, DC
Robert Park, National Institute for Occupational Safety and Health, Education and Information
Division, Cincinnati, OH
Brenda Perkovitch, U.S. Environmental Protection Agency, Office of the Administrator, Office
of Children=s Health Protection, Washington, DC
Bruce Rodan, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Washington, DC
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xxiv DRAFT—DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
Val Schaeffer, Occupational Safety and Health Administration, Directorate for Health Standards,
Washington, DC
Vanessa Vu, U.S. Environmental Protection Agency, Office of Prevention, Pesticides, and Toxic
Substances, Washington, DC
Diana M. Wong, U.S. Environmental Protection Agency, Office of Science and Technology
Office of Water, Washington, DC
Tracey Woodruff, formerly with U.S. Environmental Protection Agency, Office of Policy,
Economics, and Innovation, Washington, DC
CONSULTANTS
Anne Aschengrau, Department of Epidemiology, Boston University School of Public Health,
Boston, MA
Matt Bogdanffy, Lincoln University, PA
George Lucier, formerly with National Institute of Environmental Health Sciences, Research
Triangle Park, NC
Neurotoxicity Expert Review Panel Reviewers at public workshop on February 25, 2004
Kent Anger (Chair), Center for Research on Occupational and Environmental Toxicology,
Oregon Health and Science University, Portland, OR
Rosmarie Bowler, San Francisco State University, San Francisco, CA
Diana Echeverria, Battelle Center for Public Health Research and Evaluation, Seattle, WA
Fabriziomaria Gobba, Dipartimento di Scienze Igienistiche, Universita di Modena e Reggio
Emilia, Modena, Italy
William Merigan, Department of Ophthalmology and Center for Visual Science,
University of Rochester School of Medicine and Dentistry, Rochester, NY
ACKNOWLEDGMENTS
The authors would like to acknowledge the contributions of the following individuals:
Terri Konoza of NCEA who managed the document production activities; Cristopher Broyles of
Intellitech Systems, Inc. who provided editing support; Lana Wood of Intellitech Systems, Inc.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xxv DRAFT—DO NOT CITE OR QUOTE
-------
who provided word processing support; Patricia von Brook of KBM Group who provided editing
support for the previous draft; and Christine Chang of KBM Group who provided word
processing support for the previous draft.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 xxvi DRAFT—DO NOT CITE OR QUOTE
-------
1 1. INTRODUCTION
2
O
4 This document presents background information and justification for the Integrated Risk
5 Information System (IRIS) Summary of the hazard and dose-response assessment of
6 tetrachloroethylene. IRIS Summaries may include oral reference dose (RfD) and inhalation
7 reference concentration (RfC) values for chronic and other exposure durations, and a
8 carcinogenicity assessment.
9 The RfD and RfC, if derived, provide quantitative information for use in risk assessments
10 for health effects known or assumed to be produced through a nonlinear (presumed threshold)
11 mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
12 uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
13 population (including sensitive subgroups) that is likely to be without an appreciable risk of
14 deleterious effects during a lifetime. The RfC is defined as an estimate, with uncertainty
15 spanning perhaps an order of magnitude, of a continuous inhalation exposure to the human
16 population (including sensitive subgroups) that is likely to be without an appreciable risk of
17 deleterious non-cancer effects during a lifetime. The RfC considers toxic effects for both the
18 respiratory system (portal-of-entry) and for effects peripheral to the respiratory system
19 (extrarespiratory or systemic effects). Reference values are generally derived for chronic
20 exposures (up to a lifetime), but may also be derived for acute (#24 hrs), short-term (>24 hrs up
21 to 30 days), and subchronic (>30 days up to 10% of lifetime) exposure durations, all of which are
22 derived based on an assumption of continuous exposure throughout the duration specified.
23 Unless specified otherwise, the RfD and RfC are derived for chronic exposure duration.
24 The carcinogenicity assessment provides information on the carcinogenic hazard
25 potential of the substance in question and quantitative estimates of risk from oral and inhalation
26 exposure. The information includes a weight-of-evidence judgment of the likelihood that the
27 agent is a human carcinogen and the conditions under which the carcinogenic effects may be
28 expressed. Quantitative risk estimates may be derived from the application of a low-dose
29 extrapolation procedure. If derived, the oral slope factor is an upper bound on the estimate of
30 risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is an upper bound on the
31 estimate of risk per ug/m3 air breathed.
32 Development of these hazard identification and dose-response assessments for
33 tetrachloroethylene has followed the general guidelines for risk assessment set forth by the
34 National Research Council (NRC, 1983, 1994). U.S. Environmental Protection Agency (EPA)
35 Guidelines and Risk Assessment Forum Technical Panel Reports that may have been used in the
36 development of this assessment include the following: Guidelines for the Health Risk
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 1 -1 DRAFT—DO NOT CITE OR QUOTE
-------
1 Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines for Mutagenicity Risk
2 Assessment (U.S. EPA, 1986b), Recommendations for and Documentation of Biological Values
3 for Use in Risk Assessment (U.S. EPA, 1988), Guidelines for Developmental Toxicity Risk
4 Assessment (U.S. EPA, 1991), Interim Policy for Particle Size and Limit Concentration Issues in
5 Inhalation Toxicity (U.S. EPA, 1994a), Methods for Derivation of Inhalation Reference
6 Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b), Use of the
1 Benchmark Dose Approach in Health Risk Assessment (U. S. EPA, 1995), Guidelines for
8 Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for Neurotoxicity Risk
9 Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk Characterization (U.S.
10 EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b),
11 Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures (U.S.
12 EPA, 2000c), A Review of the Reference Dose and Reference Concentration Processes (U.S.
13 EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), Supplemental
14 Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens (U. S. EPA,
15 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A Framework
16 for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA, 2006b).
17 The literature search strategy employed for tetrachloroethylene was based on the
18 Chemical Abstracts Service Registry Number (CASRN) and at least one common name. Any
19 pertinent scientific information submitted by the public to the IRIS Submission Desk was also
20 considered in the development of this document. A comprehensive literature review was carried
21 out through July 2004. In addition, a number of relevant publications since that time have been
22 considered and incorporated in the document.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 1-2 DRAFT—DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR CHAPTER 1
2
O
4 NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
5 (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3F1 mice. 311, 1-190. U.S. Department of Health
6 and Human Services. Technical Report Series.
7
8 U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
9 mixtures. Fed. Register 51(185):34014-34025. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
10
11 U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk assessment. Federal
12 Register 51(185):34006-34012. Available online at http://www.epa.gov/ncea/raf.
13
14 U.S. EPA (Environmental Protection Agency). (1988) Recommendations for and documentation of biological values
15 for use in risk assessment. Environmental Criteria and Assessment Office, Office of Health and Environmental
16 Assessment, Cincinnati, OH; EPA 600/6-87/008. Available from: National Technical Information Service,
17 Springfield, VA; NTIS PB88-179874/AS.
18
19 U.S. EPA (Environmental Protection Agency). (1991) Guidelines for developmental toxicity risk assessment.
20 Federal Register 56(234):63798-63826.
21
22 U.S. EPA (Environmental Protection Agency). (1994a) Methods for derivation of inhalation reference
23 concentrations and application of inhalation dosimetry. Environmental Criteria and Assessment Office, Office of
24 Health and Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical
25 Information Service, Springfield, VA; NTIS PB88-179874/AS.
26
27 U.S. EPA (Environmental Protection Agency). (1994b) Peer review and peer involvement at U.S. Environmental
28 Protection Agency. Signed by the U.S. EPA Administrator Carol M. Browner, dated June 7, 1994.
29
30 U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose approach in health risk
31 assessment. Risk Assessment Forum, Washington, DC; EPA/630/R-94/007. Available from: National Technical
32 Information Service, Springfield, VA, PB95-213765, and online at http://www.epa.gov/raf.
33
34 U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive toxicity risk assessment. Federal
35 Register 61(212):56274-56322. Available online at http://www.epa.gov/ncea/raf.
36
37 U.S. EPA (Environmental Protection Agency). (1998a) Guidelines for neurotoxicity risk assessment. Federal
38 Register 63(93):26926-26954. Available online at http://www.epa.gov/ncea/raf.
39
40 U.S. EPA (Environmental Protection Agency). (1998b) Science policy council handbook: peer review. Prepared by
41 the Office of Science Policy, Office of Research and Development, Washington, DC; EPA 100-B-98-001.
42 Available from: National Technical Information Service, Springfield, VA, PB 98-140726, and online at
43 http://www.epa.gov/OSA/spc.
44
45 U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Federal Register
46 70(66)17765-18717. Available online at http://www.epa.gov/cancerguidelines.
47
48 U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing cancer susceptibility
49 from early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
50 online at http://www.epa.gov/cancerguidelines.
51
52 U.S. EPA (Environmental Protection Agency). (2006a) Peer review handbook, 3rd edition. Science Policy Council,
53 Washington, DC; EPA/100/B-06/002. Available online at
54 http://www.epa.gov/peerreview/pdfs/Peer%20Review%20HandbookMav06.pdf.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 1-3 DRAFT—DO NOT CITE OR QUOTE
-------
1 U.S. EPA (Environmental Protection Agency). (2006b) A framework for assessing health risks of environmental
2 exposures to children. National Center for Environmental Assessment, Washington, DC; EPA/600/R-05/093 A.
3 Available online at http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid= 158363.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 1-4 DRAFT—DO NOT CITE OR QUOTE
-------
1 2. BACKGROUND
2
O
4 2.1. USES AND PHYSICAL/CHEMICAL PROPERTIES
5 Tetrachloroethylene is a widely used solvent that is produced commercially for use in dry
6 cleaning, textile processing, and metal-cleaning operations. It has the following use pattern:
7 55% as a chemical intermediate, 25% for metal cleaning and vapor degreasing, 15% for dry
8 cleaning and textile processing, and 5% for other unspecified uses (ATSDR, 1997).
9 Table 2-1 lists the physical and chemical properties of tetrachloroethylene (ATSDR,
10 1997). The reference citations can be found in the Agency for Toxic Substances and Disease
11 Registry (ATSDR) document and are not included in the reference list for this document.
12
13 2.2. OCCURRENCE AND EXPOSURE
14 Tetrachloroethylene has been detected in ground water and surface water as well as in air,
15 soil, food, and breast milk. The primary exposure routes of concern are inhalation of vapor and
16 ingestion of contaminated water. Although dermal exposure is possible via contaminated tap
17 water during showering, bathing, or swimming, this is generally not considered a major route of
18 exposure.
19
20 2.2.1. Air
21 Because of its high volatility, there is considerable potential for release of
22 tetrachloroethylene into the atmosphere. Once in the air, it is not susceptible to wet deposition
23 because of its hydrophobicity. The primary method for removal is photooxidation to
24 trichloroacetyl chloride, trichloroacetic acid (TCA), carbon monoxide, ozone, and phosgene
25 (U.S. EPA, 1982). However, this reaction is very slow, so tetrachloroethylene is not implicated
26 in the buildup of any of the reaction products in the troposphere. Though the half-life of
27 perchloroethylene can vary based on season and environmental conditions, it has been estimated
28 at 96 days under typical conditions (ATSDR, 1997).
29 Ambient tetrachloroethylene concentrations vary from source to source and with
30 proximity to the source. It should be noted that outdoor concentrations can vary widely within a
31 period of a few hours as a function of wind velocity and direction, precipitation, humidity, and
32 sunlight. ATSDR (1997) reported mean tetrachloroethylene concentrations of 8.8 |ig/m3 in areas
33 close to points of release.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-1 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 2-1. Physical and chemical properties of tetrachloroethylene
Property
Molecular weight
Color
Physical state
Melting point
Boiling point
Density at 20EC
Density at 25EC
Odor
Odor threshold: water
Odor threshold: air
Solubility: water at 25EC
Solubility: organic solvent(s)
Partition coefficients:
Log KQW
Partition coefficients:
Log KOC
Vapor pressure at 25EC
Henry's law constant at 25EC
Autoignition temperature
Flashpoint
Flammability limits
Conversion factors, air
Explosive limits
Information
165.83
Colorless
Liquid (at room temperature)
-19EC
121EC
1.6227g/mL
No data
Ethereal
0.3 ppm
1 ppm
150mg/L
Miscible with alcohol, ether,
chloroform, benzene, solvent
hexane, and most of the fixed
and volatile oils
3.4
2.2B2.7
18.47mmHg
1.8H 10'2 atm-m3/mol
No data
None
Nonflammable
1 mg/L = 141.4 ppm
1 ppm = 6.78 mg/m3
No data
Reference
Lide (1990)
Sax and Lewis (1987)
Sax and Lewis (1987)
Lide (1990)
Lide (1990)
Lide (1990)
HSDB (1996)
U.S. EPA(1987b)
U.S. EPA(1987b)
HSDB (1996)
HSDB (1996)
HSDB (1996)
Seipetal. (1986)
Zytneretal. (1989a)
HSDB (1996)
Gossett(1987)
HSDB (1996)
HSDB (1996)
HSDB (1996)
Source: ATSDR(1997).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-2 DRAFT-DO NOT CITE OR QUOTE
-------
1 EPA has carried out modeling to characterize the geographic distribution of
2 tetrachloroethylene for its National-Scale Air Toxics Assessment database (U.S. EPA, 1996).
3 Median census tract-based tetrachloroethylene concentrations across the United States were
4 estimated at about 0.3 |ig/m3 for urban areas and 0.1 |ig/m3 for rural areas (75% upper percentiles
5 of 0.4 and 0.2 |ig/m3, respectively). The California Air Resources Board (CARB, 1998) reported
6 a statewide median air concentration of 0.3 |ig/m3 in 2001, which represents the lowest value in
7 what has been a decreasing trend since 1990. Note that these averages, which are based on
8 geographic areas, only characterize the likely exposure of individuals who spend an equal
9 amount of time in all parts of the defined area, and they may, therefore, significantly
10 underestimate the exposure of individuals who consistently spend time in subareas that have
11 higher tetrachloroethylene concentrations.
12 Near points of use, such as dry cleaners or industrial facilities, indoor exposure to
13 tetrachloroethylene is more significant than outdoor exposure (U.S. EPA, 2001). Indoor air
14 concentrations in an apartment above a dry cleaning shop have been measured at up to 4.9 mg/m3
15 (Verbek and Scheffers, 1980), whereas mean concentrations inside dry cleaning facilities have
16 been found to vary from 48 mg/m3 to 200 mg/m3, depending on type of facility (Solet et al.,
17 1990). Concentrations in facilities with post-1990 equipment are likely to be lower (U.S. EPA,
18 1998).
19 The off-gassing of garments that have recently been dry-cleaned may be of concern
20 (Tichenor et al., 1990). In the home, tetrachloroethylene vapors may off-gas from the clothes of
21 occupationally exposed individuals, or they may come directly from the exhaled breath of
22 exposed workers (ATSDR, 1997). Relatively high tetrachloroethylene air concentrations have
23 been measured in the proximity of freshly dry-cleaned clothing stored in small, close spaces. A
24 residential closet storing newly dry-cleaned clothing had an air concentration of 2.9 mg/m3 after
25 1 day, which rapidly declined to 0.5 mg/m3 and persisted for several days (Tichenor et al., 1990).
26 There is one documented mortality case: a 2-year-old boy was found dead after being put to
27 sleep in a room with curtains that had been incorrectly dry-cleaned (Gamier et al., 1996).
28 Dry-cleaned garments transported in an automobile may also lead to unexpectedly high
29 levels of exposure. Park et al. (1998) used simulated driving cycles to estimate the
30 concentrations of several contaminants emitted from in-vehicle sources. Using dry-cleaned
31 clothes as a source, tetrachloroethylene levels inside a stationary vehicle after 30 minutes reached
32 0.230 mg/m3. Approximating these exposures is not easy because specific exposure levels would
33 depend on many factors: car velocity, wind speed, ventilation, and time spent in the automobile.
34 Another study demonstrating exposure in a car found that transporting a freshly dry-cleaned
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-3 DRAFT-DO NOT CITE OR QUOTE
-------
1 down jacket in a car resulted in a cabin air concentration of 24.8 mg/m3 after 108 minutes (Chien,
2 1997).
3 Air exposure may also occur during showering or bathing as dissolved
4 tetrachloroethylene in the warm tap water is volatilized. Rao and Brown (1993) used an adult
5 physiologically based pharmacokinetic (PBPK) model combined with a microenvironmental
6 exposure model to estimate the dose received by inhalation exposure during showering and
7 bathing as well as by dermal exposure to the water. The tap water concentration of
8 tetrachloroethylene was 1 mg/L, which is probably a higher concentration than exists in most
9 water supplies. They also demonstrated that a majority of the tetrachloroethylene in the blood, as
10 a result of their bathing scenario, resulted from inhalation exposure, while about 15% resulted
11 from dermal absorption.
12
13 2.2.2. Water
14 Because of its relatively low aqueous solubility (see Table 2-1), it is not likely that
15 volatilized tetrachloroethylene will enter surface or rain water. However, it has been detected in
16 drinking water, ground water, and surface water (U.S. EPA, 2001; AT SDR, 1997). Most of this
17 contamination is probably due to release in water following industrial use or by public use of
18 consumer products.
19 Unless a surface water body is in the vicinity of a highly contaminated site, surface waters
20 are expected to have a lower concentration of tetrachloroethylene than ground water. In an
21 estimate of drinking water contamination in California, McKone and Bogen (1992) assumed that
22 surface water would have a negligible contribution to the concentration of tetrachloroethylene
23 measured in drinking water. Based on data from wells in California, they estimated an average
24 drinking water concentration of 0.3 |ig/L, with a standard deviation of 0.35 |ig/L.
25 In areas near sources of contamination, ground water, and surface water concentrations
26 can be considerably higher than average. Because the density of tetrachloroethylene is about
27 60% higher than that of water, tetrachloroethylene is expected to accumulate near the bottom of a
28 stagnant receiving water body after a large-volume point discharge. Water samples collected
29 near the bottom of the St. Clair River near Sarnia, Ontario, downstream from several petroleum-
30 based production facilities, contained tetrachloroethylene concentrations ranging from 0.002 to
31 34.6 |ig/L (EC, 1993). The concentrations in 17 samples of surface water from the lower Niagara
32 River in New York State in 1981 averaged 0.036 |ig/L (with a maximum of 0.134 |ig/L; EC,
33 1993).
34
35
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-4 DRAFT-DO NOT CITE OR QUOTE
-------
1 Exposure models have been developed to predict the fate and transport of organic
2 compounds such as tetrachloroethylene in environmental media, including air, water, and soil.
3 The outputs from two similar but independently developed environmental exposure models,
4 CalTOX and FugSONT, were compared for a scenario designed to reproduce a residential area
5 near an industrial contamination site (Maddalena et al., 1995), in which 75 moles/day are
6 released into the air and 0.7 moles/day are released into surface water. Although the soil
7 predictions differed, the predictions of tetrachloroethylene in air and ground water were similar,
8 with the concentration of air predicted by CalTOX approximately 6 |ig/m3 and the surface water
9 concentration 82 |ig/L. It should be noted that agreement of the models does not confirm the
10 validity of either one, but lends some support to the usefulness of the results.
11 The off-gassing of tetrachloroethylene from a drinking water supply can result in
12 exposure. In 1976, EPA measured tetrachloroethylene levels ranging from 800 to 2,000 |ig/L in
13 drinking water samples in Massachusetts (Paulu et al., 1999). Similar levels were reported
14 elsewhere in New England. These concentrations were attributed to the vinyl-lined asbestos-
15 cement pipes that were used to carry water in this area (Webler and Brown, 1993). Letkiewicz et
16 al. (1982) estimated that 53% of newborn infants are formula-fed from drinking water sources
17 and the other 47% receive all of their fluid from breast milk. Taking into account volatilization
18 during boiling of water, they indicate that the uptake of tetrachloroethylene in formula-fed infants
19 on a mg/kg-day basis is 10 times higher than in adults with the same level of drinking water
20 contamination.
21 Although dermal exposure is possible via contaminated tap water during showering,
22 bathing, or swimming, this is generally not considered a major route of exposure. Rao and
23 Brown (1993) demonstrated that only 15% of the tetrachloroethylene in the blood resulted from
24 dermal exposure as compared to inhalation of vapors.
25
26 2.2.3. Food
27 Certain foods have been found to be contaminated with tetrachloroethylene (U.S. EPA,
28 2001). Because of the lipophilic nature of tetrachloroethylene, it may bind to lipid molecules in
29 such foods as margarine, oils, meats, and other fatty foods stored in areas where there is
30 tetrachloroethylene in the air. In 1988, elevated tetrachloroethylene levels were seen in
31 margarine and butter samples obtained from grocery stores located near dry cleaning facilities
32 (Entz and Diachenko, 1988). Further studies confirmed that close proximity to a dry cleaning
33 facility was associated with elevated tetrachloroethylene levels in butter samples (Kacew and
34 Lambert, 1997). Nonetheless, food is not considered to be a major exposure pathway. Other
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-5 DRAFT-DO NOT CITE OR QUOTE
-------
1 sources of information about tetrachloroethylene in foods are the Food and Drug Administration
2 (FDA, 2003) and Fleming-Jones and Smith (2003).
3
4 2.2.4. Breast Milk
5 Due to its lipid solubility, tetrachloroethylene can concentrate in milk (NYS DOH, 2000;
6 Schreiber, 1993; Sheldon et al., 1985). Breast milk can contain high concentrations of
7 tetrachloroethylene and some of its toxic metabolites. Reported levels of tetrachloroethylene in
8 breast milk have ranged up to 43 |ig/L in the general population (U.S. EPA, 2001).
9 Schreiber (1993) used a PBPK model to estimate the dose a nursing infant might receive
10 from an exposed mother's breast milk. This study showed that it is possible for the dose an
11 infant receives through breast milk to approach levels that could result in adverse health effects
12 and exceed the 1988 EPA RfD of 0.01 mg/kg-day (U.S. EPA, 1988). Actual indoor air
13 concentrations (24-hr average), as measured in apartments in New York State, were used to
14 predict potential levels in breast milk in these modeling scenarios. The apartments included one
15 located above a dry cleaning facility that used an old dry-to-dry machine (average concentration,
16 45.8mg/m3), three located above facilities that used transfer machines (average concentration,
17 7.7mg/m3), and two located above facilities that used newer dry-to-dry machines (average
18 concentration, 0.25 mg/m3; Schreiber, 1993). The predicted breast milk concentrations in these
19 scenarios ranged from 16 to 3,000 |ig/L. Assuming that a 7.2 kg infant ingests 700 mL of breast
20 milk per day, Schreiber (1993) determined that the infant dose from milk could range from
21 0.0015 to 0.3 mg/kg-day.
22 Using the same exposure conditions as Schreiber (1993), Byczkowski et al. (1994)
23 predicted lower doses to the infant (0.0009-0.202 mg/kg-day), although these doses approached
24 levels that could result in adverse health effects. Exceedances of the RfD were seen only in those
25 apartments above old dry-to-dry machines (0.202 mg/kg-day) or above transfer machines (0.029
26 mg/kg-day). Ingestion through breast milk and infant exposures is discussed further in
27 Section 4.8. However, Schreiber (1997) has suggested that if infants live adjacent to or in close
28 proximity to dry cleaning facilitates, the dose received through breast milk ingestion will be
29 insignificant when compared with that from their inhalation exposure.
30 In one case study, the breast milk of a woman was found to contain 10 mg/L of
31 tetrachloroethylene 1 hr following a visit to her husband at his work in a dry cleaning
32 establishment. This concentration dropped to 3 mg/L after 24 hrs. Her child suffered from
33 obstructive jaundice and hepatomegaly, but these conditions improved when breastfeeding was
34 discontinued (Bagnell and Ellenberger, 1977).
35
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-6 DRAFT-DO NOT CITE OR QUOTE
-------
1 2.2.5. Direct Ingestion
2 In rare circumstances, direct ingestion of tetrachloroethylene has been documented. A
3 6-year-old boy who directly ingested 12-16 g tetrachloroethylene experienced drowsiness,
4 vertigo, agitation, and hallucinations. He then lost consciousness and went into a coma, and later
5 recovered (Koppel et al., 1985). Follow-up testing on the boy was not reported, so any potential
6 long-term effects of the exposure are unknown.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-7 DRAFT-DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR CHAPTER 2
2
3 ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
4 (update). Prepared by Sciences, International, under subcontract to Research Triangle Institute.
5
6 Bagnell, PC; Ellenberger, HA. (1977) Obstructive jaundice due to a chlorinated hydrocarbon in breast milk. Can
7 MedAssocJ 117:104761048.
8
9 Byczkowski, JZ; Kinkead, ER; Leahy, HF; et al. (1994) Computer simulation of the lactational transfer of
10 tetrachloroethylene in rats using a physiologically based model. Toxicol Appl Pharmacol 125:228B236.
11
12 CARD (California Air Resources Board). (1998) 1990B1996 Statewide perchloroethylene summary, ppb. California
13 Environmental Protection Agency. Available online at http://www.arb.ca/gov/aqd/perc/pcstate.htm.
14
15 Chien, Y-C. (1997) The influences of exposure pattern and duration on elimination kinetics and exposure assessment
16 of tetrachloroethylene in humans. PhD thesis, Rutgers University, New Brunswick, NJ. Available from: IRIS
17 Information Desk, U.S. Environmental Protection Agency, Washington, DC.
18
19 EC (Environment Canada). (1993) Priority substances list assessment report: tetrachloroethylene. En-40-215/28E.
20 Canadian Environmental Protection Agency, Ottawa, Canada.
21
22 Entz, RC; Diachenko, GW. (1988) Residues of volatile halocarbons in margarines. Food Addit Contam 5:267B276.
23
24 FDA (U.S. Food and Drug Administration). (2003) Food and Drug Administration total diet study: summary of
25 residues found, ordered by pesticide. 91-3-01-4. Center for Food Safety and Nutrition. Washington, DC. Available
26 online at http://www.cfsan.fda.gov/~acrobat/tdslbvps.pdf.
27
28 Fleming-Jones, ME; Smith, RE. (2003) Volatile organic compounds in foods: A five year study. J. Agric and Food
29 Chem 51:8120-8127
30
31 Gamier, R; Bedouin, J; Pepin, G; Gaillard, Y. (1996) Coin-operated dry cleaning machines may be responsible for
32 acute tetrachloroethylene poisoning: report of 26 cases including one death. J Toxicol Clin Toxicol 34:19lBl97.
33
34 Kacew, S; Lambert, GH. (1997) Environmental toxicology and pharmacology of human development. Washington,
3 5 DC: Taylor and Francis.
36
37 Koppel, C; Arndt, I; Arendt, U; et al. (1985) Acute tetrachloroethylene poisoning-blood elimination kinetics during
38 hyperventilation therapy. J Toxicol Clin Toxicol 23:103B115.
39
40 Letkiewicz, F; Johnston, P; Macaluso, C; et al. (1982) Occurrence of tetrachloroethylene in drinking water, food, and
41 air. Prepared by JRB Associates JRB Project No. 2-613-03-852-29, for EPA contract 86-01-6388, task 29.
42
43 Maddalena, RL; McKone, TE; Layton, DW; et al. (1995) Comparison of multi-media transport and transformation
44 models: regional fugacity model vs. CalTOX. Chemosphere 30:869B889.
45
46 McKone, TE; Bogen, KT. (1992) Uncertainties in health-risk assessment: an integrated case study based on
47 tetrachloroethylene in California groundwater. Regul Toxicol Pharmacol 15:86Bl03.
48
49 McKone, T; Bogen, K. (1993) CaLTOX: A multimedia total-exposure model for hazardous waste sites. Part II:
50 multimedia transport and transformation model. Prepared for the State of California, Department of Toxic
51 Substances Control. UCRLBCRB111456Ptl.
52
53 McKone, TE; Daniels, JI. (1991) Estimating human exposure through multiple pathways from air, water, and soil.
54 Regul Toxicol Pharmacol 13:36B61.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-8 DRAFT-DO NOT CITE OR QUOTE
-------
1 NYS DOH. (2000) Evaluation of residential exposure to tetrachloroethene using biomarkers of dose and neurological
2 tests (non peer-reviewed draft). Albany, NY.
O
4 Park, JH; Spengler, JD; Yoon, DW; et al. (1998) Measurement of air exchange rate of stationary vehicles and
5 estimation of in-vehicle exposure. J Expo Anal Environ Epidemiol 8:65-78.
6
7 Paulu, C; Aschengrau, A; Ozonoff, D. (1999) Tetrachloroethylene-contaminated drinking water in Massachusetts and
8 the risk of colon-rectum, lung, and other cancers. Environ Health Perspect 107:2656271.
9
10 Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
11 groundwater for a bathing and showering determination. Risk Anal 13:37B49.
12
13 Schreiber, JS. (1993) Predicted infant exposure to tetrachloroethene in human breastmilk. Risk Anal 13:515B524.
14
15 Schreiber. JS. (1997) Transport of organic chemicals to breast milk: Tetrachloroethene case study. In Kacew S,
16 Lambert G (eds): Environmental Toxicology and Pharmacology of Human Development. Washington, DC: Taylor
17 and Francis.
18
19 Sheldon, L; Handy, R; Hartwell, W; et al. (1985) Human exposure assessment to environmental chemicals: nursing
20 mothers study. Final report. Research Triangle Institute, Research Triangle Park, NC.
21
22 Solet, D; Robins, TG; Sampaio, C. (1990) Perchloroethylene exposure assessment among dry cleaning workers. Am
23 Ind Hyg Assoc J 51(10):566-574.
24
25 Tichenor, BA; Sparks, LE; Jackson, MD; et al. (1990) Emissions of perchloroethylene from dry cleaned fabrics.
26 Atmospheric Environment 24A: 1219B1229.
27
28 U.S. EPA (Environmental Protection Agency). (1982) An exposure and risk assessment for tetrachloroethylene.
29 Office of Water, Regulations, and Standards, Washington, DC; EPA-4404-85-015.
30
31 U.S. EPA (Environmental Protection Agency). (1988) IRIS summary of tetrachloroethylene RID. Available online at
32 http://www.epa. gov/iris/subst/0106.htm.
33
34 U.S. EPA (Environmental Protection Agency). (1996) Modeled ambient concentration for perchloroethylene
3 5 (CAS#127184). National Air Toxics Assessment, Technology Transfer Network, Office of Air and Radiation.
36 Available online at http://www.epa.gov/ttn/atw/nata/pdf/perc/conc.pdf.
37
38 U.S. EPA (Environmental Protection Agency). (2001) Sources, emission and exposure for trichloroethylene (TCE)
39 and related chemicals. National Center for Environmental Assessment, Washington, DC; EPA/600/R-00/099.
40 Available from: National Technical Information Service, Springfield, VA, and online at http://www.epa. gov/ncea.
41
42 Verberk, MM; Scheffers, TM. (1980) Tetrachloroethylene in exhaled air of residents near dry-cleaning shops.
43 Environ Res 21(2):432^37.
44
45 Webler, T; Brown, HS. (1993) Exposure to tetrachloroethylene via contaminated drinking water pipes in
46 Massachusetts: a predictive model. Arch Environ Health 48:293B297.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 2-9 DRAFT-DO NOT CITE OR QUOTE
-------
1 3. TOXICOKINETICS
2
O
4 3.1. ABSORPTION
5 Tetrachloroethylene is rapidly absorbed into the bloodstream following oral and
6 inhalation exposures. It can also be absorbed across the skin following dermal exposure to either
7 pure or diluted solvent or vapors (Stewart and Dodd, 1964; Nakai et al., 1999; Poet et al., 2002).
8
9 3.1.1. Inhalation
10 The major exposure route for tetrachloroethylene is considered to be inhalation (U.S.
11 EPA, 1985; IARC, 1995). Pulmonary uptake of tetrachloroethylene is rapid; however, complete
12 tissue equilibrium occurs only after several hours. Absorption into the systemic circulation
13 through pulmonary uptake is proportional to the ventilation rate, the duration of exposure, and, at
14 lower ambient concentrations to which humans are likely to be exposed, the concentration in the
15 inspired air (Hake and Stewart, 1977; Monster et al., 1979).
16 Chiu et al. (2007) reported that peak levels of tetrachloroethylene in venous blood and air
17 occurred near the end of a 6-hr inhalation exposure to 1 ppm and declined thereafter. In the
18 Monster et al. (1979) study, uptake after 4 hrs was 75% of its value at the onset of exposure.
19 Increased physical activity increases uptake but lowers the alveolar partial pressure, thus
20 removing more tetrachloroethylene from the alveoli, resulting in a longer time to reach tissue
21 equilibrium (Pezzagno et al., 1988).
22 The blood/gas partition coefficient for tetrachloroethylene describes how the chemical
23 will partition itself between the two phases. Specifically, it is the ratio of concentrations at
24 steady state; i.e., when all rates are constant after equilibrium has been reached. Reported values
25 for the coefficient in humans range from around 10 to 20 (e.g., Byczkowski and Fisher, 1994;
26 Reitz et al., 1996; Droz and Guillemin, 1986; Ward et al., 1988; Gearhart et al., 1993; Hattis et
27 al., 1990), meaning that if tetrachloroethylene is in equilibrium, the concentration in blood will
28 be 10 to 20 times higher than the concentration in the alveoli.
29 Opdam and Smolders (1986) determined concentrations of tetrachloroethylene in alveolar
30 air for 1- to 60-second residence times (the time interval from the beginning of an inhalation to
31 the end of the next inhalation) for six volunteers exposed to 0.5 to 9.8 ppm of chemical for 1 to
32 60 minutes. These investigators found the concentrations of tetrachloroethylene in alveolar air to
33 decrease with residence times for breaths during exposure periods but to increase during post-
34 exposure for residence times less than 10 seconds. Alveolar air tetrachloroethylene
35 concentration correlated with the concentrations in pulmonary artery mixed venous blood.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-1 DRAFT-DO NOT CITE OR QUOTE
-------
1 Like the studies in humans, inhalation studies in laboratory animals provide clear
2 evidence that tetrachloroethylene is readily absorbed via the lungs into the systemic circulation
3 (e.g., Pegg et al., 1979; Dallas et al., 1994a).
4
5 3.1.2. Oral
6 Gastric absorption of tetrachloroethylene occurs at a relatively rapid rate and is
7 essentially complete. Close to 100% of oral doses are absorbed from the gut, according to
8 reports of several studies conducted in mice, rats, and dogs (Dallas et al., 1994a, 1995; Frantz
9 and Watanabe, 1983; Pegg et al., 1979; Schumann et al., 1980). Absorption into the systemic
10 circulation was indicated by blood tetrachloroethylene levels of 21.5 |ig/mL following accidental
11 ingestion of the chemical by a 6-year-old boy (Koppel et al., 1985).
12
13 3.1.3. Dermal
14 Absorption of tetrachloroethylene by humans following dermal exposure to vapors of the
15 chemical has been reported to be relatively insignificant (only 1%) when compared with
16 absorption via inhalation of vapors (Riihimaki and Pfaffli, 1978; Nakai et al., 1999). The
17 amount of chemical absorbed during the immersion of one thumb in liquid tetrachloroethylene is
18 equivalent to the uptake during inhalation of 10 to 15 ppm of the compound for the same time
19 period (Stewart and Dodd, 1964).
20 Studies in animals confirm that dermal uptake of tetrachloroethylene following vapor
21 exposure is minimal when compared with pulmonary uptake (Tsuruta, 1989; McDougal et al.,
22 1990), whereas dermal uptake is greater following direct skin application (Jakobson et al., 1982).
23 Notably, the conclusions of Bogen et al. (1992), based on the results of their study in hairless
24 guinea pigs, indicate that dermal absorption of tetrachloroethylene from contaminated water
25 supplies could be an important route of exposure for humans. These investigators estimated that
26 a standard 70 kg man with 80% of his body immersed in water would completely absorb the
27 amount of tetrachloroethylene in 2 L of that water.
28
29 3.2. DISTRIBUTION AND BODY BURDEN
30 Once absorbed, tetrachloroethylene is distributed by first-order diffusion processes to all
31 tissues in the mammalian body. The highest concentrations of tetrachloroethylene are found in
32 adipose tissue due to the lipophilicity of the compound (U.S. EPA, 1985). Concentrations of
33 tetrachloroethylene reach higher levels in brain and liver than in many other tissues (Gamier et
34 al., 1996; Levine et al., 1981; Lukaszewski, 1979). Absolute tissue concentrations are directly
35 proportional to the body burden or exposure dose. Due to its lipid solubility, tetrachloroethylene
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-2 DRAFT-DO NOT CITE OR QUOTE
-------
1 is also concentrated in milk, and it has been measured in human breast milk (Schreiber, 1993,
2 1997; Schreiber et al., 2002; NYS DOH, 2000). Higher concentrations occur in milk having
3 higher fat content; e.g., a noticeable difference exists between the milk/blood partition
4 coefficients for rats (12) and for humans (2.8; Byczkowski and Fisher, 1994), reflecting the
5 higher fat content of rat milk. Tetrachloroethylene readily crosses both the blood-brain barrier
6 and the placenta. Partition coefficients for various tissues, relative to blood or air, have been
7 reported by several investigators (Ward et al., 1988; Dallas et al., 1994a, b; Gearhart et al., 1993;
8 Byczkowski and Fisher, 1994). Section 3.5 presents examples of these.
9 Repeated daily inhalation exposures of human volunteers to tetrachloroethylene indicate
10 accumulation of the compound in the body, which is thought to be due to its high lipid solubility.
11 Because of its long residence time in adipose tissue, repeated daily exposure results in an
12 accumulated concentration; tetrachloroethylene from new exposures adds to the residual
13 concentration from previous exposures until steady state is reached. Blood levels of
14 tetrachloroethylene increase over several days with continued daily exposures. Following
15 cessation of these exposures, it is still present in the blood. Exhalation of the compound
16 continues over a number of days due to its slow release from the adipose tissue (Stewart et al.,
17 1977; Altmann et al., 1990; Skender et al., 1991). For a given concentration in blood or air, the
18 half-time—the time necessary to equilibrate the adipose tissue to 50% of its final
19 concentration—is about 25 hrs (Monster, 1979; Fernandez et al., 1976). Therefore, during a
20 single 8-hr exposure, adipose tissue does not reach steady-state equilibrium.
21 Tetrachloroethylene uptake by fatty tissue during the working hours of the week is
22 countered by the elimination that occurs during nonexposure times of nights and weekends; thus,
23 for persons exposed to tetrachloroethylene on a five-day-a-week work schedule, an equilibrium
24 is eventually established, but it requires a time period of 3 to 4 weeks of exposure for adipose
25 tissue to reach plateau concentrations (U.S. EPA, 1985).
26 Animal studies provide clear evidence that tetrachloroethylene distributes widely to all
27 tissues of the body, readily crossing the blood-brain barrier and the placenta (Schumann et al.,
28 1980; Ghantous et al., 1986; Savolainen et al., 1977; Dallas et al., 1994b). Following exposure
29 of rats to tetrachloroethylene, the compound has been measured in blood, fat, brain, lungs, liver,
30 kidneys, heart, and skeletal muscle (Savolainen et al., 1977; Dallas et al., 1994b). Highest tissue
31 concentrations were found in adipose tissue (60 or more times blood level) and in brain and liver
32 (4 and 5 times blood level, respectively), as can be calculated from the rat tissue distribution data
33 of Savolainen et al. (1977). Dallas et al. (1994b) found the concentration in fat to be 9 to 18
34 times the concentrations found in nonfat tissues. Skeletal muscle contained the lowest
35 concentration. In one human fatality case, the concentration of tetrachloroethylene in the brain
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-3 DRAFT-DO NOT CITE OR QUOTE
-------
1 was 120 times higher than concentrations measured in the lung. In another case the
2 concentration in the liver was 8, 3.4, and 3.5 times higher, respectively, than concentrations
3 measured in the lung, kidney, and brain (Levine et al., 1981).
4
5 3.3. METABOLISM
6 This section describes the metabolism of tetrachloroethylene, identifying metabolites
7 thought to be causally associated with toxic responses as well as those used to evaluate the flux
8 of parent compound through the known metabolic pathways. Sex- and species-dependent
9 differences in the metabolism of tetrachloroethylene and potential contributors to interindividual
10 differences are identified. Factors that influence metabolism in humans are mentioned. See
11 Section 4.9 for further discussion of how these factors affect variability and susceptibility.
12
13 3.3.1. Introduction
14 The metabolism of tetrachloroethylene has been studied mostly in mice, rats, and humans
15 (for reviews, see Dekant et al., 1987, 1989; Anders et al., 1988; IARC, 1995; U.S. EPA, 1985,
16 1986, 1991; Lash and Parker, 2001). Tetrachloroethylene is metabolized in laboratory animals
17 and in humans through at least two distinct pathways: oxidative metabolism via the cytochrome
18 P450 (CYP [also abbreviated as P450 and CYP 450]) mixed-function oxidase system and
19 glutathione (GSH) conjugation followed by subsequent further biotransformation and processing,
20 either through the cysteine conjugate beta lyase pathway or by other enzymes including flavin-
21 containing monooxygenase 3 (FMO3) and CYP3A (Daniel, 1963; Filser and Bolt, 1979; Pegg et
22 al., 1979; Costa and Ivanetich, 1980; Dekant et al., 1987, 1989; Anders et al., 1988; U.S. EPA,
23 1985, 1991; IARC, 1995; Birner et al., 1996; Lash et al., 1998; Volkel et al., 1998; Lash and
24 Parker, 2001). The conjugative pathway, although the minor route quantitatively, is
25 lexicologically significant because it yields relatively potent toxic metabolites (Vamvakas et al.,
26 1987, 1989a, b, c; Dekant et al., 1986a, b, 1989; Werner et al., 1996; Anders et al., 1988; Lash
27 and Parker, 2001).
28
29 3.3.2. Extent of Metabolism
30 Studies in both animals and humans indicate that overall metabolism of
31 tetrachloroethylene is relatively limited (reviewed in U.S. EPA, 1985, 1991; Lash and Parker,
32 2001), as evidenced by the high percentage of absorbed dose excreted in the breath as the parent
33 molecule (Stewart et al., 1961, 1970; Monster et al., 1979, 1983; Boettner and Muranko, 1969;
34 Ikeda and Otsuji, 1972; Essing et al., 1973; Fernandez et al., 1976; May, 1976; Ohtsuki et al.,
35 1983; Yllner, 1961; Daniel, 1963; Filser and Bolt, 1979; Pegg et al., 1979; Frantz and Watanabe,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-4 DRAFT-DO NOT CITE OR QUOTE
-------
1 1983; Schumann et al., 1980; Buben and O'Flaherty, 1985; Volkel et al., 1998). Because of its
2 high lipid solubility, tetrachloroethylene can be sequestered in fat and, thus, not all metabolism is
3 evident in short sampling time periods.
4 The extent of metabolism after inhalation exposure in humans has been estimated by
5 measuring trichloro-compounds excreted in the urine and exhalation of tetrachloroethylene in
6 expired air (Bolanowska and Golacka, 1972; Fernandez et al., 1976; Monster et al., 1979, 1983;
7 Monster and Houtkooper, 1979; Ikeda et al., 1972; Boettner and Muranko, 1969; Essing et al.,
8 1973; May, 1976; Stewart et al., 1961, 1970). Several studies reported only about 1-3% of the
9 estimated amounts inhaled were metabolized to TCA and other chlorinated metabolites, although
10 additional tetrachloroethylene—as much as 20% or more of the dose—may be metabolized over
11 a longer period (Monster etal., 1979; U.S. EPA, 1985, 1991; Bois et al., 1996; Bogen et al.,
12 1992). For example, Chiu et al. (2007) noted that although an average of 0.4% of
13 tetrachloroethylene intake (1 ppm for 6 hrs) was recovered in urine as TCA, total recovery in
14 urine and exhaled air accounted for on average only 82% of intake. This would imply 18%
15 metabolized, but Chiu et al. (2007) noted substantial uncertainty and variability in these
16 calculations and concluded they were consistent with previous studies at higher exposures.
17 Interestingly, Chiu et al. (2007) also noted significant variability among the seven subjects and
18 among the four occasions, contributing to the uncertainty in measurements.
19 The extent of metabolism in animals has been estimated by conducting excretion-balance
20 studies using isotopically labeled tetrachloroethylene. In rodents, 2-88% of the dose was
21 metabolized, depending on dose level and species: the higher the dose the smaller the percent
22 metabolized. Rats metabolized a lower percent of a given tetrachloroethylene body burden than
23 did mice (Yllner, 1961; Daniel, 1963; Filser and Bolt, 1979; Pegg et al., 1979; Frantz and
24 Watanabe, 1983; Schumann et al., 1980). As an example, using data from the Pegg et al. (1979)
25 and Schumann et al. (1980) studies in rats, EPA calculated that the percent of body burdens
26 excreted were unchanged following exposure to 10 and 600 ppm for 6 hrs, were 68 and 99%,
27 respectively (U.S. EPA, 1985). For comparison, studies in mice exposed to 10 ppm for 6 hrs
28 found pulmonary excretion of only 12%, whereas 83% of the tetrachloroethylene was excreted
29 by the pulmonary route for a body burden of about 11 mg from oral administration (U.S. EPA,
30 1985). As body burden is increased, the proportion of tetrachloroethylene excreted unchanged
31 increases and the percent metabolized decreases.
32
33 3.3.3. Pathways of Metabolism
34 The two known biotransformation pathways for tetrachloroethylene metabolism are (1)
35 oxidation by cytochrome P450 (CYP) enzymes and (2) conjugation with GSH followed by
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-5 DRAFT-DO NOT CITE OR QUOTE
-------
1 further processing of the conjugate through various pathway bifurcation branches. The initial
2 step in the metabolism of tetrachloroethylene may be either epoxidation or chlorine migration for
3 the oxidative pathway or conjugation with GSH for the secondary pathway (Costa and Ivanetich,
4 1980; Miller and Guengerich, 1982, 1983; Dekant et al., 1986b, 1987, 1998; Lash et al., 1998;
5 Lash and Parker, 2001). It is possible that other as yet unrecognized pathways for
6 tetrachloroethylene exist in humans (Sakamoto, 1976; Monster et al., 1979; U.S. EPA, 1985,
7 1991; Boisetal., 1996).
8
9 3.3.3.1. Cytochrome P450-Dependent Oxidation
10 Oxidative metabolism by the cytochrome P450, or CYP-dependent, pathway is
11 quantitatively the major route of tetrachloroethylene biotransformation (U.S. EPA, 1991; IARC,
12 1995; Lash and Parker, 2001). This pathway was initially proposed by Powell (1945) for
13 trichloroethylene and was subsequently supported for tetrachloroethylene by the results of Yllner
14 (1961), Daniel (1963), Leibman and Ortiz (1970, 1977), Costa and Ivanetich (1980), and others.
15 The pathway is operative in humans and rodents and leads to several metabolites, some of which
16 are known toxins and carcinogens (U.S. EPA, 1991; IARC, 1995). Figure 3-1 depicts the overall
17 scheme of tetrachloroethylene P450 metabolism. Known metabolites presented in this figure are
18 identified by an asterisk.
19 The major excretory metabolite of the oxidative pathway, TCA, is excreted in the urine of
20 all species tested. Figure 3-1 identifies many common urinary metabolites, including
21 dichloroacetic acid (DCA), trichloroacetylethanolamide, oxalylethanalamide, and oxalic acid.
22 Trichloroethanol (TCOH) has been measured in some, but not all, studies (Bonse et al., 1975;
23 Bonse and Henschler, 1976; Yllner, 1961; Dmitrieva, 1967; Pegg et al., 1979; Ogata et al., 1962,
24 1971; Tanaka and Ikeda, 1968; Ikeda and Otsuji, 1972; Ikeda et al., 1972; Monster et al., 1983;
25 Weichardt and Lindner, 1975; Dekant et al., 1986b, 1987; Birner et al., 1996; U.S. EPA, 1985,
26 1986, 1991). Oxalic acid is a relatively major urinary metabolite in rats (Dmitrieva, 1967; Pegg
27 et al., 1979). Pulmonary excretion of carbon dioxide (CO2) has been identified in exhaled breath
28 from rodents exposed to 14C-labeled tetrachloroethylene (Pegg et al., 1979; Schumann et al.,
29 1980). Oxalic acid and formic acid plus CO2 are hypothesized to arise from action of
30 microsomal epoxide hydrase on the initial epoxide intermediate to yield tetrachloroethylene
31 glycol, which may then be further processed via two routes to these aforementioned end
32 products.
33 Oxidative metabolism of tetrachloroethylene, irrespective of the route of administration,
34 occurs predominantly in the liver but also occurs at other sites. For example, the kidneys exhibit
35 cytochrome P450 enzyme activities, mostly in the proximal tubules, although total activity is
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-6 DRAFT-DO NOT CITE OR QUOTE
-------
\
I
,o
n o ci
\/\/
c — c
/ z \
and
ITelrachloroclhylcnc -
P450 intermediate)
17
C1,C C
n
•s
on •«
o
*
"/"Ii
xnc!K-rn,-on
,
c ' OH on ( '
II
UN —C
H
— C — Oil
I
0
0
\
t)II
OH
J_6
1
2
3
4
5
6
7
8
9
10
11
12
13
Figure 3-1. Postulated scheme for the metabolism of tetrachloroethylene by
the P450 oxidative pathway. Tetrachloroethylene and identified (*) metabolites:
1 tetrachloroethylene, 2 tetrachloroethylene oxide, 1 trichloroacetyl chloride, 4
trichloroacetic acid, 5 dichloroacetic acid, 6 monochloroacetic acid, 7 chloral, 8
trichloroethanol, 9 trichloroethanol glucuronide, K) trichloroacetyl ethanolamide,
11 tetrachloroethylene glycol, J_2 ethandioyl dichloride, H oxalic acid, 14
glyoxylic acid chloride, 1_5 carbon dioxide, 16_ formic acid, j/7 P450 intermediate,
JJ5 chloral hydrate, 19 glyoxylic acid, and 20 glycine.
Sources: Adapted from Pegg et al. (1979), Costa and Ivanetich (1980), U.S. EPA
(1985), Dekant et al. (1986a), Lash and Parker (2001).
TTz/'s1 document /'s a draft for review purposes only and does not constitute Agency policy
06/06/08 3-7 DRAFT-DO NOT CITE OR QUOTE
-------
1 markedly less than in the liver (Lash and Parker, 2001; Lash et al., 2001). CYP enzymes
2 occurring in other extrahepatic tissues—brain and lungs, for example—may also contribute to
3 oxidative metabolism of tetrachloroethylene.
4
5 3.3.3.1.1. Formation of tetrachloroethylene oxide. The first step in the oxidation of
6 tetrachloroethylene is hypothesized to yield 1,1,2,2-tetrachloroethylene oxide, a relatively
7 unstable epoxide (Costa and Ivanetich, 1980; Miller and Guengerich, 1982, 1983). Although an
8 initial epoxide metabolite has not been unequivocally demonstrated for tetrachloroethylene,
9 evidence for this epoxide does exist. The epoxide has been chemically synthesized (Frankel et
10 al., 1957; Bonse et al., 1975; Kline et al., 1978). The several potential fates of
11 tetrachloroethylene epoxide include trichloroacetyl chloride, oxalate dichloride through
12 tetrachloroethylene glycol, trichloroacetyl aminoethanol, and possibly chloral hydrate (in
13 equilibrium with chloral; Bonse and Henschler, 1976; Henschler and Bonse, 1977; Pegg et al.,
14 1979; U.S. EPA, 1985, 1986). Formation of trichloroacetyl chloride directly from
15 tetrachloroethylene, without the formation of the epoxide intermediate, via the mechanism of
16 CYP-mediated olefin oxidation has also been postulated (Guengerich and Macdonald, 1984).
17
18 3.3.3.1.2. Metabolism to Trichloroacetic Acid (TCA) and possibly Trichloroethanol (TCOH).
19 Measurement of urinary TCA has been used as a biomarker for tetrachloroethylene exposure
20 (U.S. EPA, 1985; IARC, 1995), although TCA can be a by-product of metabolism of other
21 chemical compounds. TCA, a major tetrachloroethylene urinary metabolite in both humans and
22 laboratory rodents (Yllner, 1961; Daniel, 1963; Leibman and Ortiz, 1970, 1977; Birner et al.,
23 1996; Dekant et al., 1987; Ohtsuki et al., 1983; Volkel et al., 1998), is believed to result
24 primarily from the oxidation of tetrachloroethylene to trichloroacetyl chloride. This oxidation
25 may occur through the epoxide intermediate, with chloride migration leading to the reactive
26 trichloroacetyl chloride, which can then react with amino groups of cellular proteins or undergo
27 hydrolysis to produce the TCA. N-(di- and trichloroacetylated)-L-lysines, formed by interaction
28 of tetrachloroethylene reactive metabolites with protein, have been identified in liver and kidney
29 tissue of rats exposed to tetrachloroethylene (Birner et al., 1994; Pahler et al., 1999a).
30 The proposed chloral hydrate intermediate is another potential source of TCA, but chloral
31 hydrate can also be further metabolized to TCOH (Sellers et al., 1972; Birner et al., 1996). This
32 latter pathway to TCOH would be the favored reaction, and it is thought to be catalyzed by both
33 alcohol dehydrogenase (Larson and Bull, 1989) and CYP2E1 (Schultz and Weiner, 1979; Ni et
34 al., 1996). The resulting TCOH is then conjugated with glucuronide, a reversible reaction, and
35 both the alcohol and its glucuronide conjugate have been reportedly detected as urinary excretion
36 products following tetrachloroethylene exposures. TCOH has been detected in the urine of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-8 DRAFT-DO NOT CITE OR QUOTE
-------
1 subjects exposed to tetrachloroethylene in some studies (Birner et al., 1996; Ikeda and Otsuji,
2 1972; Ikeda et al., 1972; Ogata et al., 1962, 1971; Tanaka and Ikeda, 1968; Monster et al., 1983;
3 Weichardt and Lindner, 1975; Schreiber et al., 2002), but could not be identified by others
4 (Fernandez et al., 1976; Hake and Stewart, 1977; Monster et al., 1979; Volkel et al., 1998;
5 Yllner, 1961; Daniel, 1963; Buben and O'Flaherty, 1985; Costa and Ivanetich, 1980). TCOH is
6 thought to be an artifact of the methodology used or could arise due to unknown exposures to
7 other chemicals. Thus, because TCOH is clearly not a significant metabolite for
8 tetrachloroethylene, very little, if any, TCA produced from tetrachloroethylene metabolism is
9 likely to come through chloral, either directly or indirectly through TCOH (Lash and Parker,
10 2001).
11
12 3.3.3.1.3. Formation of dichloroacetic acid (DCA) and other products. TCA is the major
13 source of DCA from the tetrachloroethylene P450 oxidation pathway. Although DCA has been
14 identified as a tetrachloroethylene urinary metabolite (Yllner, 1961; Dekant et al., 1987; Volkel
15 et al., 1998), it is not clear whether the DCA is a product of further metabolism of TCA, of
16 another pathway originating with GSH conjugation, or both. The major organ site of DCA
17 production is likely to differ for each pathway, with DCA arising from oxidative metabolism
18 primarily in the liver and from GSH-dependent metabolism products mostly in the kidney. The
19 amount of DCA produced from tetrachloroethylene oxidative metabolism may vary across
20 species and is likely to be less than TCA. This is because DCA derived from P450 oxidation
21 comes only from dechlorination of TCA, which is not extensively metabolized, but rather, is
22 mostly excreted unchanged in urine.
23 The lack of a role for DCA in tetrachloroethylene liver toxicity is supported by the
24 limited findings of Maronpot and his coworkers (Anna et al., 1994; Maronpot et al., 1995),
25 which showed no similarities in mutation spectra between tetrachloroethylene-induced liver
26 tumors and DCA-induced liver tumors. It is interesting to note, however, that the kinetics of
27 metabolism and the sensitivity of target tissue to TCA and DCA and their precursors are likely of
28 key importance to understanding species differences in responsiveness to tetrachloroethylene.
29 Dechlorination of TCA to DCA is catalyzed by gut contents (ingested food and bacteria)
30 of the rat and mouse (Moghaddam et al., 1996); isolated mouse microflora have been shown to
31 convert TCA to DCA (Moghaddam et al., 1997). DCA can be rather quickly processed to other
32 chemical species, such as monochloroacetic acid (MCA), glycolic acid, glyoxylic acid, and
33 oxalic acid (Abbas and Fisher, 1997; Lash et al., 2000; Bull, 2000; Board et al., 1997; Tong et
34 al., 1998a, b; Lash and Parker, 2001). Conversion to glyoxylic acid is thought to occur by action
35 of the GST zeta (GSTZ in humans) isoform of glutathione S-transferase (GST; Lash et al., 2000).
36 DCA is a mechanism-based inactivator of GSTZ, of which five polymorphic variants exist
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-9 DRAFT-DO NOT CITE OR QUOTE
-------
1 (Tzeng et al., 2000). Potent and irreversible inhibition of GSTZ activity by DCA occurs, and the
2 substrate inhibition of the enzyme in vitro differs between rats and humans, with the enzyme
3 being relatively more sensitive to inhibition by DCA in rats. Further degradation of DCA in the
4 liver occurs primarily in hepatic cytosol (Lipscomb et al., 1995). Human liver cytosol is less
5 efficient than either rat or mouse liver cytosol in processing DCA (Lipscomb et al., 1995).
6 Trichloroacetyl ethanolamide also may be formed from the tetrachloroethylene oxide
7 intermediate or from the alternative chlorine migration in an oxygenated tetrachloroethylene
8 transition state. 14CO2 has been recovered from laboratory animals administered 14C-labeled
9 tetrachloroethylene (Frantz and Watanabe, 1983; Pegg et al., 1979; Schumann et al., 1980). A
10 measurable portion of tetrachloroethylene is completely metabolized in a dose-dependent manner
11 to CO2. The oxalate metabolite excretory product may be derived from DCA or MCA (long et
12 al., 1998a, b), although oxalic acid is also produced from the epoxide through tetrachlorodiacetyl
13 chloride and oxalic acid dichloride intermediates (Pegg et al., 1979; Costa and Ivanetich, 1980).
14 The occurrence of oxalic acid and of CO2 as major metabolites of tetrachloroethylene, at least in
15 rodents, indicates the existence of pathway(s) of metabolism other than the primary TCA
16 pathway.
17
18 3.3.3.1.4. Species-dependent differences. Although thought to be qualitatively similar, there are
19 clear differences among species in the quantitative aspects of tetrachloroethylene metabolism
20 (Schumann et al., 1980; Ikeda and Otsuji, 1972; Volkel et al., 1998; U.S. EPA, 1991; Lash and
21 Parker, 2001). These differences are in the relative yields and kinetic behavior of metabolites
22 (Volkel et al., 1998; Ohtsuki et al., 1983; Green et al., 1990; U.S. EPA, 1985, 1991). Rodents
23 and humans differ in relative rates of tetrachloroethylene metabolism in key target organs, in the
24 doses at which saturation of metabolism occurs, and in the half-times in the body.
25 The rate of metabolism of tetrachloroethylene is faster in rodents than in humans and
26 higher blood levels of metabolites are obtained in rodents as compared to humans. The higher
27 blood levels of metabolites in rodents are particularly noticeable at the higher tetrachloroethylene
28 exposure levels because saturation is approached at lower exposure levels in humans than in
29 rodents. The half-time in the body of these metabolites is, however, noticeably longer for
30 humans than for rodents (144 hrs in humans vs. approximately 10 hrs or less in rodents; see U.S.
31 EPA, 1985). It is for this reason that examinations of tetrachloroethylene concentration and
32 toxicity associations must reflect both blood concentration and time-integrated dose metrics such
33 as area-under-the-curve.
34 A study of species differences in tetrachloroethylene metabolism conducted by Dekant
35 and colleagues is presented in Volkel et al. (1998). These investigators compared both oxidative
36 and GSH-dependent metabolism in rats and humans exposed for 6 hrs to 10, 20, or 40 ppm
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-10 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene by inhalation. Rats were also exposed to 400 ppm concentrations. TCA was
2 the major urinary excretion product in both species; however, the elimination half-time was more
3 than four times slower in humans than in rats. Blood plasma concentrations of the metabolite
4 were higher (three- to eight-fold, depending on the dose) in rats than in humans exposed to
5 identical air concentration levels. These observations are in agreement with metabolic rates in
6 general, which are higher in mice than in rats; rats, in turn, have higher metabolic rates than do
7 larger animals, including humans. Dekant and his coworkers also reported urinary excretion of
8 DCA by rats but not humans. They concluded most of the DC A resulted from GSH-dependent
9 metabolism. DCA, however, is further metabolized by P450 enzymes, which, in turn, limits its
10 detectability in urine.
11
12 3.3.3.1.5. Cytochrome P450 (CYP) isoforms and genetic polymorphisms. Relatively few
13 studies provide information about which specific CYP isoforms play a role in tetrachloroethylene
14 oxidative metabolism. CYP2E1 is presumed to have an important role in tetrachloroethylene
15 metabolism (Lash and Parker, 2001); however, the chemical-specific related data are too sparse
16 to provide strong support for this assumption (Doherty et al., 1996). CYP2B1/2 may also be
17 important for the metabolism of tetrachloroethylene. CYP3A isoenzymes may contribute to the
18 generation of reactive sulfoxides from metabolites of the GSH pathway (see below). Costa and
19 Ivanetich (1980) showed increased hepatic metabolism following treatment with agents now
20 known to induce these isoenzymes specifically.
21 Genetic polymorphisms are DNA sequence variations that result in changes in protein
22 sequence of an enzyme that can alter the enzyme's ability to catalyze a reaction or alter the
23 expression of an allele. Polymorphisms are known for most of the CYP enzymes including
24 CYP2E1 (McCarver et al., 1998; Hu et al., 1999) and CYP3A4 (Sata et al., 2000; Westlind et al.,
25 1999).
26 Metabolism of tetrachloroethylene to its putative epoxide is likely affected by CYP
27 enzymes. The metabolism of the putative metabolite chloral hydrate to TCOH and TCA may be
28 catalyzed by both alcohol dehydrogenase and CYP2E1. Oxidation of TCOH is catalyzed by
29 P450 enzymes, with CYP2E1 the likely predominant isoform involved, although other
30 isoenzymes may also play a role, even substituting for CYP2E1 in processing
31 tetrachloroethylene. Rat kidney expresses CYP2E1 (Cummings et al., 1999; Speerschneider and
32 Dekant, 1995), although human kidney has not been shown to do so (Amet et al., 1997;
33 Cummings et al., 2000a). Therefore, renal CYP metabolism by this isoform in rat kidney would
34 be relevant only insofar as the involvement of other isoenzymes in metabolizing
3 5 tetrachl oroethy 1 ene vi a thi s route.
36
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-11 DRAFT-DO NOT CITE OR QUOTE
-------
1 3.3.3.2. Glutathione (GSH) Conjugation Pathway
2 Figure 3-2 shows the second metabolic pathway for tetrachloroethylene (Dekant et al.,
3 1987, 1988; Sausen and Elfarra, 1990; Volkel et al., 1998). This GSH pathway was
4 subsequently shown to exist in both rodents and humans (Volkel et al., 1998).
5 The GSH pathway is initiated by the conjugation of the parent tetrachloroethylene
6 molecule with GSH to form S-(l,2,2-trichlorovinyl)glutathione (trichlorovinyl glutathione, or
7 TCVG). This reaction, which is catalyzed by the GSH-S-transferase enzymes (GSTs), a group
8 of enzyme isoforms, was traditionally considered to be a detoxification reaction, leading to more
9 water-soluble compounds that are more readily excreted. In many cases, however, as with
10 certain halogenated alkanes and alkenes such as tetrachloroethylene, GSH conjugation can be
11 important for bioactivation. The critical step for the alkenes would occur after the enzymatic
12 removal of the glutamyl and glycine residues from the GSH conjugate to yield the corresponding
13 cysteine S-conjugate, which in the case of tetrachloroethylene would be S-(l,2,2-trichlorovinyl)
14 cysteine (trichlorovinyl cysteine, or TCVC; Dekant et al., 1987, 1989; Anders et al., 1988; Green
15 et al., 1990; Vamvakas et al., 1987, 1989a, b, c).
16 Tetrachloroethylene conjugation with GSH is thought to primarily occur through an
17 interorgan process. GSH conjugation occurs predominantly in the liver to form TCVG, which is
18 then further metabolized to the corresponding cysteine conjugate, TCVC, by the enzymes
19 gamma-glutamyltransferase (GGT) and cysteinylglycine dipeptidase. TCVC acts as a substrate
20 for several enzymes. Beta lyase cleaves TCVC to yield an unstable thiol, giving rise to cytotoxic
21 and mutagenic products, particularly the reactive thioketene. TCVC may also be activated by
22 cysteine conjugate S-oxidase activity, which also can rearrange to form the reactive thioketene.
23 Although Green et al. (1990) hypothesized that GSH conjugation and subsequent activation of
24 tetrachloroethylene did not occur in humans, the N-acetyl urinary metabolite has subsequently
25 been clearly identified in humans exposed to tetrachloroethylene in occupational settings, in
26 laboratory studies, and in residential buildings (Birner et al., 1996; Volkel et al., 1998; Schreiber
27 et al., 2002). Therefore, this pathway is now known to operate in humans as well as in rodents.
28 This GSH conjugation pathway was recognized much later than was the oxidative
29 pathway, probably because it is relatively minor quantitatively compared with the CYP pathway,
30 yet it may be lexicologically influential (U.S. EPA, 1991; IARC, 1995; Lash and Parker, 2001).
31 The evidence for this is based on in vitro kinetics and the relatively low recovery of urinary
32 mercapturates as compared with urinary TCA and other CYP-derived metabolites (Green et al.,
33 1990; Birner et al., 1996). Urinary mercapturates comprise from 1% to as little as 0.03% of total
34 recovered urinary metabolites, but this does not reflect the total flux through the GSH pathway
35 but rather only the portion that is excreted. In particular, the amount of the mercapturate product
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-12 DRAFT-DO NOT CITE OR QUOTE
-------
Cl Cl
GST
1
2
3
4
5
6
7
8
9
10
11
12
c c
Cl Cl
1
Cl O COO
C C CH2 C CH2
£T)
Cl S CH NH^NH,
GGT HN CH2 CH
V e
C CH2 COO
O
DP
P-Lyase
ci c:
c c
Cl 6 S
Cl ci V
C C CH, COO"
Cl S CH 3
tl7 K i l_j
ci
FM03
or
P450
Cl
CCNAT
Acylase
C C CH, COO"
CCS
Cl
C C R
Cl S
P-Lyase
or
Spontaneous
CljOHCCXT
9*
Cl S CH
C O CH,
C NH
e°
CYP3A1/2 (rat)
CYP3A4 (human)
Cl Cl V
C C CH, COO®
Cl . S CH O
5
O HN C
CH, 0°
Figure 3-2. Metabolism of tetrachloroethylene by the glutathione
conjugation pathway. Tetrachloroethylene and identified (*) metabolites: 1
tetrachloroethylene, 2 S-l,2,2-trichlorovinyl) glutathione (TCVG), 3
S-(l,2,2-trichlorovinyl)-L-cysteine (TCVC), 4 N-acetyl trichlorovinyl cysteine
(NAcTCVC), 5 NAcTCVC sulfoxide, 6 1,2,2-trichlorovinylthiol, 7 TCVCSO, 8
2,2-dichlorothioketene, and 9 dichloroacetate. Enzymes: glutathione-
S-transferase (GST), gamma-glutamyltransferase (GGT), dipeptidase (DP), beta
lyase, FMO3, CCNAT, acylase, CYP3A1/2, and CYP3A4. Unstable reactive
metabolites are shown in brackets.
Source: Lash and Parker (2001).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-13 DRAFT-DO NOT CITE OR QUOTE
-------
1 excreted in the urine also does not reflect the amount of the more important portion that is
2 converted to toxic by-products through further metabolism.
3 For tetrachloroethylene, the GSH pathway is associated with renal toxicity (Anders et al.,
4 1988; Dekant et al., 1989; U.S. EPA, 1991; IARC, 1995; Lash et al., 2000; Lash and Parker,
5 2001). The initial conjugation with GSH occurs mainly in the liver (Dekant et al., 1987; Green
6 et al., 1990; Vamvakas et al., 1987, 1989a), with transport of the conjugate and its cysteine
7 counterpart to the kidney target organ for further processing. This first step also occurs within
8 the kidney (Lash et al., 1998). As shown in Figure 3-2, tetrachloroethylene is initially
9 conjugated with GSH to form TCVG. This reaction is catalyzed by cytosolic and microsomal
10 GSTs. TCVG is then processed through the cysteinylglycine conjugate S-(l,2,2,trichlorovinyl)-
11 L-cysteinylglycine to TCVC by the enzymatic removal of glutamyl and glycine residues by GGT
12 and various membrane-bound dipeptidases known as cysteinylglycine dipeptidase (reviewed by
13 Anders et al., 1988; Dekant et al., 1989; U.S. EPA, 1991; Lash and Parker, 2001). These
14 enzymes reside in tissues other than the kidneys (e.g., the brain), indicating a potential for toxic
15 reactive metabolite formation in those tissues as well. Conversion of TCVG to TCVC by these
16 cleavage enzymes leads to a critical bifurcation point of the GSH pathway because the TCVC
17 may be processed by certain enzymes to yield reactive, toxic chemical species, although it may
18 be metabolized via a different route to yield an excretory product (Lash and Parker, 2001).
19 Importantly, the TCVC metabolite may also act as a substrate for renal beta lyases
20 (Dekant et al., 1988 reviewed by Anders et al., 1988; Dekant et al., 1989; U.S. EPA, 1991; Lash
21 et al., 2000; Lash and Parker, 2001). Renal beta lyases are known to cleave TCVC to yield an
22 unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a reactive chemical species that can
23 form covalent adducts with cellular nucleophiles, including DNA and proteins (Volkel et al.,
24 1999). Beta lyases are a family of pyridoxal phosphate-containing enzymes that are located in
25 several tissues besides the kidneys, including liver and brain, and in intestinal flora, although
26 their substrate specificities may vary. Hepatic beta lyase is distinct from renal beta lyase and has
27 not been found to have a role in TCVC metabolism. Beta lyase activity is higher in rat kidney
28 than in human kidney (Cooper, 1994; Lash et al., 1990), which is consistent with overall
29 metabolic rates being higher in smaller versus larger mammalian species.
30 In addition to activation by beta lyases, TCVC may be metabolized by a flavin-containing
31 monooxygenase, FMO3, or CYP enzymes to TCVC sulfoxide (TCVCSO), another reactive
32 metabolite (Ripp et al., 1997). TCVSO is a more potent nephrotoxicant than TCVC (Elfarra and
33 Krause 2007). These TCVC sulfoxide and beta lyase cleavage products rearrange, forming a
34 thioketene (Dekant et al., 1988; Ripp et al., 1997), which is a potent acylating agent capable of
35 binding to cellular macromolecules, including DNA (Birner et al., 1996; Pahler et al., 1999a, b;
36 Volkel et al., 1999). Interestingly, the thioketene can degrade to form DC A, potentially making
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-14 DRAFT-DO NOT CITE OR QUOTE
-------
1 this metabolite a product of both tetrachloroethylene metabolism pathways (Dekant et al., 1987;
2 Volkeletal., 1998).
3 In addition to beta lyase and FMO3/CYP activation of TCVC, reactive sulfoxides can
4 also be produced by further CYP3A metabolism of N-acetyl-S-(l,2,2 -trichlorovinyl)-L-cysteine
5 (NAcTCVC; Werner et al., 1996). This tetrachloroethylene-derived mercapturate metabolite
6 results from TCVC being acetylated via a reversible reaction (Bartels, 1994; Birner et al., 1996;
7 Duffel and Jakoby, 1982). N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine may be excreted in the
8 urine. However, in addition to its activation to sulfoxides via CYP3 A metabolism, it can also be
9 transported to other organs and deactylated intracellularly, regenerating the cysteine conjugate
10 TCVC and thus making it available to other enzymes for activation (Uttamsingh et al., 1998). It
11 should be noted that the N-acetylation reaction is catalyzed by an enzyme located in the
12 endoplasmic reticulum that is distinct from the cytosolic enzymes that are polymorphic in
13 humans (Lash and Parker, 2001).
14
15 3.3.3.2.1. Glutathione S-transferase (GST) isoenzymes/polymorphisms. GSTs are a family of
16 isoenzymes (Mannervik, 1985) found in cytoplasm. A distinct microsomal GST isoenzyme also
17 exists in most mammalian tissues (Otieno et al., 1997). Although GST activity occurs in most
18 cell types, the liver is by far the predominant site of GSH conjugation. GST alpha, designated as
19 GSTA in humans, is the predominate isoenzyme expressed in normal kidney from rodents and
20 humans (Campbell et al., 1991; Overby et al., 1994; Mitchell et al., 1997; Rodilla et al., 1998;
21 Cummings et al., 2000b). Available data thus far do not indicate that variability in activity of
22 this isoenzyme is important to differences in individual susceptibility to toxicity. GSTZ
23 catalyzes the oxidative metabolism of DCA to glyoxylate (Board et al., 1997; Tong et al., 1998a,
24 b), however, the tetrachloroethylene metabolite DCA has been shown to be a potent, irreversible
25 inhibitor of GSTZ activity (Tzeng et al., 2000).
26 There are five human polymorphic variants of this GSTZ isoenzyme (Tzeng et al., 2000;
27 U.S. EPA, 1998). These genetic polymorphisms may influence tetrachloroethylene metabolism
28 although human data regarding this hypothesis are lacking. There are some species differences
29 in the other three cytoplasmic GSTs relevant to liver and kidney. GSTP expression is the most
30 variable and appears to be polymorphic in humans (Rodilla et al., 1998). It has been found in rat
31 liver (Cummings et al., 1999), but only in biliary ducts in humans (Terrier et al., 1990; Campbell
32 et al., 1991). GSTP has been detected within the human kidney in various cell types (Terrier et
33 al., 1990) but has not been isolated from rat kidney cells (Cummings et al., 1999), although
34 GSTP has also been detected in rabbit kidney (Cummings et al., 1999).
35 Two homodimeric GST theta (GSTT) isoenzymes have been identified in human kidney
36 (Veitch et al., 1997; Cummings et al., 2000a). GSTT has been detected in rat and mouse liver
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-15 DRAFT-DO NOT CITE OR QUOTE
-------
1 and in mouse but not rat kidney (Cummings et al., 1999; Quondamatteo et al., 1998). GST mu
2 (GSTM) has been detected in rat kidney distal tubule cells (Cummings et al., 2000b) and in
3 mouse and rabbit liver and kidney (Overby et al., 1994; Mitchell et al., 1997), but it was not
4 detected in human kidney (Cummings et al., 2000a). It is not clear just how the differences in
5 these isoenzymes are related to species differences in tetrachloroethylene toxicity because the
6 isoenzyme specificity and reaction rates have not yet been studied with regard to
7 tetrachloroethylene (Lash and Parker, 2001).
8 Some controversy surrounds the importance of the GSH conjugation pathway with regard
9 to tetrachloroethylene metabolism in humans. As noted above, the GSH pathway for
10 tetrachloroethylene was originally demonstrated only in rodents, and interpretation of the then-
11 existing data led some scientists to conclude that the pathway was not operative in humans
12 (Green et al., 1990; U.S. EPA, 1991). More recent data clearly demonstrate the existence of the
13 pathway in humans (Birner et al., 1996; Volkel et al., 1998; Schreiber et al., 2002). There are
14 discrepancies regarding reported rates of tetrachloroethylene GSH metabolism, however (Green
15 et al., 1990; Dekant et al., 1987; Lash et al., 1998; Lash and Parker, 2001). These differences
16 may be due, in part, to different chemical assay methodology or to problems resulting from the
17 stability of the chemical product being measured or both (Lash and Parker, 2001). Some of the
18 published findings concerning TCVG production would not predict any less susceptibility for
19 humans than for rodents with regard to renal toxicity (Lash et al., 1998).
20
21 3.3.3.2.2. Gamma-glutamyltransferase (GGT). Species-dependent differences in GGT
22 (Hinchman and Ballatori, 1990) also are not thought to be limiting, because renal activity is
23 present at high enough levels even in humans so that GGT activity is not the rate-limiting step in
24 the metabolism. Species-dependent differences in this enzyme (described below) would have
25 only a very small quantitative effect on the overall metabolism of TCVG and other similar GSH
26 conjugates. Species differences in GGT activities, therefore, would not have a major role in
27 species differences in renal toxicity (Lash and Parker, 2001) in affecting transformation of
28 TCVG to TCVC, and thus, should not be important to differences in susceptibility to
29 tetrachloroethylene-induced renal toxicity.
30 GGT is the only enzyme that can split the gamma-glutamyl bond in the GSH conjugates
31 to form cysteine conjugates (Lash and Parker, 2001). It is this reaction that creates TCVC, the
32 substrate for the enzymes that generate the toxic metabolites. Therefore, the distribution of GGT
33 is important. Renal proximal tubular cells have the highest activities of GGT of all tissues,
34 although GGT activity also occurs in the liver, and the kidney-to-liver ratio of this enzyme varies
35 among species. In the rat, the specific activity ratio is 875 (Hinchman and Ballatori, 1990). The
36 ratio is lower in other species that have been studied. The tissue distribution and relative activity
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-16 DRAFT-DO NOT CITE OR QUOTE
-------
1 have not been fully studied in humans, but it is known that GOT activity is considerably higher
2 in human liver than in rodent liver (Lash and Parker, 2001). The kidney-to-liver ratio of GGT
3 for humans is thought to be closer to those of pigs (2) and Macaques (5) than to those of rats or
4 mice (423). For this reason, use of a rodent model for the processing of the tetrachloroethylene
5 GSH conjugate to the corresponding cysteine conjugate would overestimate the contribution of
6 the kidneys and underestimate the contribution of the liver in cleaving TCVG to TCVC. Even
7 so, the liver excretes most of the cysteine conjugates such as TCVC into the bile or plasma,
8 where it is cycled to the kidneys and taken up into renal epithelial cells. So, the TCVC will still
9 end up in the kidneys.
10
11 3.3.3.2.3. Betalyase. The beta lyase enzyme is among the most important activator of toxic
12 products in the conjugation pathway, a fact particularly well documented in the kidney. There
13 are some data, however, that indicate that renal beta lyase-dependent metabolism is greater in
14 rats than in mice or in humans and greater in male than in female rats (Green et al., 1990; Lash et
15 al., 1990; Volkel et al., 1998). This is not entirely in keeping with metabolic rates in general,
16 which are higher in mice than in rats, and rats, in turn have higher metabolic rates than do larger
17 animals, including humans. Studies that measured only cytoplasmic beta lyase activity did not
18 consider the importance of mitochondrial beta lyase activity, which may be key to
19 tetrachloroethylene metabolite toxicity (Lash et al., 2001).
20 In contrast, it must also be noted that species comparisons of tetrachloroethylene
21 metabolism in chronic exposures on a surface area- or metabolic-rate basis rather than on a direct
22 body-weight basis, particularly when including the total area-under-the-curve (AUC) for amount
23 metabolized, indicate that metabolite production in rats and humans may not differ significantly
24 (U.S. EPA, 1986; Rhomberg, 1992; Calabrese, 1983). The fact is that metabolic rates and the
25 amounts metabolized are not the same thing. Metabolic rates are always faster in smaller
26 species. Total AUC may or may not be similar among species. Even if AUC is the same, the
27 peak blood levels may differ greatly from species to species. In other words, the
28 pharmacokinetics are not the same.
29 The higher percentage of mercapturate found in rat versus human urine does not indicate
30 a higher level of production of toxic products in the rat, because excreted mercapturate allows no
31 estimate of the amount of TCVC or N-acetyl TCVC being processed through alternate routes
32 (Lash and Parker, 2001). The relatively higher percentage of DCA in the urine may, however,
33 indicate a relatively higher beta lyase enzyme activity and a higher thioketene production in rats
34 if the DCA is indeed largely the product of the GST pathway rather than the oxidative pathway
35 (Volkel et al., 1998). It is not known whether sex-dependent variation of beta lyase activity
36 exists in humans as it does in rats (Volkel et al., 1998).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-17 DRAFT-DO NOT CITE OR QUOTE
-------
1 And finally, it is important to note that because the enzymes involved in this activation
2 pathway are also present in other tissues (Dohn and Anders, 1982; Stevens, 1985; Stevens and
3 Jacoby, 1983; Tateishi et al., 1978; Tomisawa et al., 1984, 1986; Larsen, 1985; Larsen and
4 Stevens, 1986; Alberati-Giani et al., 1995; Malherbe et al., 1995), there exists a potential for
5 formation of the reactive metabolites at sites other than the kidney, e.g., in the brain. In one
6 carcinogenicity bioassay of tetrachloroethylene, a biologically significant elevation of gliomas in
7 the rat brain was reported (NTP, 1986). Whether or not toxic metabolites resulting from beta
8 lyase activity in the brain play a role in the development of the gliomas in the rat has not been
9 studied. The possibility that such tetrachloroethylene metabolites could be involved in the mode
10 of tumorigenic action producing gliomas is not unrealistic.
11
12 3.3.3.3. Relative Roles of the Cytochrome P450 (CYP) and Glutathione (GSH) Pathways
13 Although it is clear that the oxidative CYP pathway is quantitatively more important than
14 the GSH conjugation pathway, the interorgan patterns for some of the intermediate metabolites,
15 as well as the relative toxicity of certain key metabolites generated from these pathways,
16 influence the relative importance of the two pathways in determining toxicity. It is still not
17 certain which metabolites, alone or in combination, are explicitly responsible for specific
18 tetrachloroethylene toxicities, and it is likely that different metabolites contribute to toxicity at
19 different target sites. In general, CYP metabolism is associated with tetrachloroethylene-induced
20 liver toxicity, whereas GSH conjugation followed by further processing by beta lyase and other
21 enzymes is associated with tetrachloroethylene-induced renal toxicity. There is a possibility that
22 beta lyase products could contribute to toxicity in the brain, for example, and be a factor in the
23 gliomas observed in rats. The parent compound itself is also likely to be a contributing factor to
24 tetrachloroethylene neurotoxicity, particularly central nervous system (CNS) effects.
25 Data from experiments designed to assess the effects of enzyme modulation suggest
26 competition between the two pathways (Dekant et al., 1987; Lash et al., 1999; Volkel et al.,
27 1998; Lash et al., 2001). Other data show relatively low urinary excretion of mercapturates as
28 compared to CYP-derived products. On the basis of these findings, some scientists have
29 concluded that there is a lack of toxicological significance for the low-affinity, low-activity GSH
30 pathway except when the high-affinity CYP pathway approaches saturation (Green et al., 1990,
31 1997; Volkel et al., 1998). However, this conclusion does not consider the relative toxicological
32 potency or chemical reactivity of the metabolites from the two pathways or the fact that the
33 amount of mercapturate excreted is not a valid quantitative indicator of the extent of conjugative
34 pathway metabolism (Lash and Parker, 2001).
35 Specific tetrachloroethylene metabolites are known to be associated with certain
36 toxicities when they are administered directly. Exactly how these same compounds—as
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-18 DRAFT-DO NOT CITE OR QUOTE
-------
1 metabolites of tetrachloroethylene—contribute to the various toxicities associated with exposure
2 to the parent compound is not yet well understood.
O
4 3.3.4. Susceptibility
5 Differences in enzyme activity may lead to variations among individuals in their
6 sensitivity to tetrachloroethylene toxi cities. A 10-fold difference in CYP enzyme metabolic
7 capacity among humans is a generally accepted norm. Although individual variations in the
8 CYP2E1 enzymatic activity as high as 20- to 50-fold have been reported (Stephens et al., 1994;
9 Yoo et al., 1988; Lieber, 1997), these in vitro measurements would be taken out of physiological
10 context if used to estimate in vivo interindividual variations. Measurable and obvious
11 differences in CYP enzymatic activity are observed among various ethnic groups and age groups
12 (Goldstein et al., 1969; Raunio et al., 1995). No chemical-specific data regarding the manner in
13 which CYP enzyme isoforms might affect susceptibility to adverse effects are available for
14 tetrachloroethylene.
15 Diagnosis of polymorphisms in carcinogen-activating and -inactivating enzymes and
16 cancer susceptibility have been noted (Stephens et al., 1994; Yoo et al., 1988; Raucy, 1995).
17 Potential strain-dependent differences among rodents and human genetic polymorphisms in
18 metabolizing enzymes involved in biotransformation of tetrachloroethylene are now known to
19 exist. Whether CYP polymorphisms could account for interindividual variation in
20 tetrachloroethylene metabolism among humans-and thus differences in susceptibility to
21 tetrachloroethylene-induced toxi cities-is not known.
22 The GSTs involved in tetrachloroethylene metabolism are described in Section 3.3.2. A
23 potential exists for interindividual variation to occur in tetrachloroethylene metabolism as a
24 result of variability in GST enzyme expression. It is important to note that GST polymorphism
25 has been associated with increased risk of kidney cancer in people exposed to trichloroethylene.
26 This information is discussed in EPA's draft health assessment report on trichloroethylene (U.S.
27 EPA, 2001). There are no direct, chemical-distinctive data with regard to the specific isoenzyme
28 family responsible for TCVG formation in metabolism of tetrachloroethylene. There are
29 species-dependent differences as to which isozymes occur in liver and kidney, although it is
30 unknown how the various enzymes are related to differences in metabolism of
31 tetrachloroethylene. The compound is likely a good substrate for GSTA (Lash and Parker,
32 2001). GSTT and GSTP occur in human kidney, as does GSTA, the primary isozyme in human
33 kidney, meaning that there is a potential for differences in the ability to produce TCVG. GSTZ
34 transforms the tetrachloroethylene metabolite DCA. DCA has also been shown to have a potent
35 irreversible inhibitory effect on the GSTZ isoenzyme, which is known to have at least four
36 polymorphic variations.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-19 DRAFT-DO NOT CITE OR QUOTE
-------
1 Inhibition or induction of the enzymes responsible for tetrachloroethylene metabolism
2 can, and likely does, alter susceptibility to toxicity (U.S. EPA, 1985; IARC, 1995; Lash and
3 Parker, 2001). Numerous environmental pollutants and therapeutic agents alike have the
4 potential to induce or inhibit tetrachloroethylene-metabolizing enzymes. For example,
5 tetrachloroethylene metabolism is increased by inducers of cytochrome CYP enzymes such as
6 toluene, phenobarbital, and pregnenolone-16 alpha-carbonitrile, whereas CYP inhibitors such as
7 SKF 525A, metyrapone, and carbon monoxide decrease tetrachloroethylene metabolism (Moslen
8 et al., 1977; Ikeda and Imanura, 1973; Costa and Ivanetich, 1980). Chronic exposure to
9 tetrachloroethylene has been shown to cause self-induction of metabolism (Kaemmerer et al.,
10 1982; Savolainen et al., 1977; Vainio et al., 1976). Other factors, such as health status or disease
11 state, activity patterns, or concomitant exposure to other chemicals, can potentially influence
12 tetrachloroethylene metabolism and its resulting toxicity. Section 4.9 addresses issues
13 coexposures and cumulative risk in greater detail.
14
15 3.3.5. Comparison of Tetrachloroethylene Metabolism with Trichloroethylene Metabolism
16 Tetrachloroethylene is structurally related to trichloroethylene, and these two compounds
17 cause similar adverse health effects. The toxic effects, with the possible exception of
18 neurotoxicity, are attributed to metabolites. TCA, DCA, chloral, and TCOH are reported P450
19 biotransformation products of tetrachloroethylene and trichloroethylene; however, both the
20 relative amounts of these metabolites produced and the precursor intermediates in the oxidative
21 pathways are different for the two compounds. Interestingly, although tetrachloroethylene is not
22 as extensively biotransformed as trichloroethylene, it is slightly more toxic. Differences in
23 pharmacodynamics of precursors to P450 metabolic products as well as pharmacokinetic
24 differences between the two parent compounds may be related to their pharmacologic potencies
25 (Buben and O'Flaherty, 1985). Excretion of urinary mercapturates indicates that, relative to
26 P450 oxidation, tetrachloroethylene is more extensively metabolized via GSH conjugation than
27 is trichloroethylene. However, these urinary excretion products do not reflect the total flux
28 through the GSH pathway since the resulting glutathione and cysteine conjugates have been
29 shown to undergo further processing to products that are highly reactive. The Appendix for
30 Chapter 3 provides additional discussion of tetrachloroethylene/trichloroethylene comparative
31 metabolism.
32
33 3.4. EXCRETION
34 Tetrachloroethylene is eliminated from the body by two major processes: pulmonary
35 excretion and rate-limited metabolism. Tetrachloroethylene that is not metabolized is exhaled
36 unchanged, and this elimination process is the primary pathway of tetrachloroethylene excretion
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-20 DRAFT-DO NOT CITE OR QUOTE
-------
1 in humans for all routes of administration (Monster et al., 1979; Stewart et al., 1961, 1970, 1974,
2 1977; Guberan and Fernandez, 1974; Opdam and Smolders, 1986; Koppel et al., 1985; Stewart
3 and Dodd, 1964). Pulmonary elimination of (unchanged) parent compound is also important to
4 tetrachloroethylene excretion by animals (Pegg et al., 1979; Yllner, 1961; Frantz and Watanabe,
5 1983; Schumann et al., 1980; Bogen et al., 1992). A very small amount of tetrachloroethylene
6 has been shown to be eliminated through the skin (Bolanowska and Golacka, 1972); however, it
7 represents an insignificant percent of total tetrachloroethylene elimination.
8 Pulmonary elimination of unchanged tetrachloroethylene and other volatile compounds is
9 related to ventilation rate, cardiac output, and the solubility of the compound in blood and tissue.
10 The lung clearance of tetrachloroethylene in six adults exposed at rest to 72 ppm and 144 ppm of
11 tetrachloroethylene averaged 6.1 L/min initially and decreased to 3.8 L/min after 4 hrs (Monster
12 et al., 1979). Lung clearance represents the volume of air from which all tetrachloroethylene can
13 be removed per unit time. Normal ventilation rates in adults range from 5-8 L air/min at rest.
14 Pulmonary elimination of unchanged tetrachloroethylene at the end of exposure is a first-order
15 diffusion process across the lungs from blood into alveolar air, and it can be thought of as the
16 inverted equivalent of its uptake from the lungs. Pulmonary excretion occurs in three first-order
17 phases of desaturation of blood vessel-rich tissues, muscle tissue, and adipose tissues (Monster et
18 al., 1979; Guberan and Fernandez, 1974). For humans, the half-times of elimination from these
19 three tissue groups are 12 to 16 hrs, 30 to 40 hrs, and 55 to 65 hrs, respectively (Monster et al.,
20 1979).
21 The long half-time of tetrachloroethylene elimination from adipose tissue, due to the high
22 adipose tissue/blood partition coefficient and the low rate of blood perfusion of the fat tissue
23 (Eger, 1963), is independent of the body burden of tetrachloroethylene, indicated by parallel
24 blood and exhaled air concentration decay curves (U.S. EPA, 1985). However, the exhaled air or
25 end alveolar air concentrations and the blood concentrations after exposure and throughout
26 desaturation are proportional to the acquired body burden or exposure concentration and
27 duration, and they can serve as a means of estimating body burdens. The half-life of
28 tetrachloroethylene in the human body, measured as the inverse of the slope of the log-
29 concentration versus time curve of the exhaled chemical, varies from 5 to 20 minutes for the first
30 phase of elimination up to approximately 50 hrs during its extended phase (Chien, 1997; Monster
31 et al., 1979). The long half-time of tetrachloroethylene pulmonary excretion indicates that a
32 considerable time is necessary to completely eliminate the compound. This time is greater than
33 five times the half-life, or about 2 weeks for humans. For the rat, the half-time of pulmonary
34 elimination is about 7 hrs.
35 Metabolism of tetrachloroethylene provides another means of elimination of the parent
36 compound. Metabolism in humans is not considered to be as important to tetrachloroethylene
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-21 DRAFT-DO NOT CITE OR QUOTE
-------
1 excretion as is pulmonary excretion; however, at low ambient exposure concentrations, this may
2 not be the case. The rationale for assigning greater importance to elimination by metabolism in
3 humans is discussed later in this section. The biotransformation process is well accepted as
4 being important to elimination of tetrachloroethylene in rodents (see Metabolism, Section 3.3).
5 The mean half-time of elimination for total trichloro-compounds for 13 subjects exposed
6 to tetrachloroethylene was determined to be 144 hrs (Ikeda and Imamura, 1973). When TCA is
7 administered directly, however, the half-life is not that long. The longer half-life of TCA from
8 tetrachloroethylene metabolism is likely due to constant metabolic conversion of the parent
9 compound to TCA as tetrachloroethylene is cycled to the liver over the period of time it is
10 released from adipose tissue.
11 The urinary excretion of tetrachloroethylene biotransformation products, primarily TCA,
12 has been thought to represent only a small percent of the total absorbed dose of
13 tetrachloroethylene in humans (U.S. EPA, 1985; ATSDR, 1997). Urinary excretion of TCA (or
14 total trichloro-compounds) was estimated to be only 1 to 3% in balance studies conducted in
15 humans (Stewart et al., 1961, 1970; Monster et al., 1979, 1983; Monster and Houtkooper, 1979;
16 Boettner and Muranko, 1969; Ikeda et al., 1972; Essing et al., 1973; Fernandez et al., 1976; May,
17 1976). The shortcomings of the human balance studies include the lack of follow-up of the
18 subjects over a long time period. It is highly likely that a larger percent of the
19 tetrachloroethylene dose was eventually metabolized. Not all of the dose was accounted for in
20 these studies, indicating that more of the dose may be metabolized. Part of the dose may be
21 metabolized to biotransformation products, such as oxalic acid, that were not measured. It is
22 important to note that estimates of risk calculated directly from the data from such studies would
23 seriously underestimate risk of exposure, because the tetrachloroethylene dose in some of these
24 studies does not likely reflect low-dose exposure metabolism (U.S. EPA, 1985, 1991; Bois et al.,
25 1996).
26 A literature review published by Hattis et al. (1990) reported estimates of the fraction of
27 tetrachloroethylene metabolized at a low dose of 1 ppm to range from 2 to 86%. Based on data
28 from the 1979 Monster et al. study, Bois and his colleagues (Bois et al., 1996) determined that at
29 exposure levels above the current occupational standards, a median of approximately 1.5% of
30 inhaled tetrachloroethylene would be metabolized, whereas at ambient air levels (0.001 ppm) the
31 median estimate of the fraction of inhaled dose that would be metabolized is 36%, a considerably
32 higher fraction of the dose.
33 Metabolism of tetrachloroethylene has generally been reported to contribute more to its
34 elimination in rats and mice than in humans. The relative importance of metabolism elimination
35 of tetrachloroethylene in rodents depends on the species and the dose (Pegg et al., 1979;
36 Schumann et al., 1980; Dallas et al., 1994a; Bogen et al., 1992). As the body burden of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-22 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene is increased in the rat or mouse, the percentage excreted as unchanged parent
2 compound increases. Conversely, as metabolism is the other principal route of elimination of
3 tetrachloroethylene, when the body burden increases, the percentage of the burden metabolized
4 decreases, although the absolute amount metabolized increases (Pegg et al., 1979; Schumann et
5 al., 1980). These observations suggest that, in the rodent, metabolism of tetrachloroethylene and
6 urinary excretion of its metabolites are rate limited and dose dependent, whereas pulmonary
7 excretion is a first-order process and is dose independent, with half-time and rate constant being
8 independent of the dose. Data from studies by Filser and Bolt (1979) and Buben and O'Flaherty
9 (1985) suggest that elimination of tetrachloroethylene by metabolism is greater in mice than in
10 rats.
11
12 3.5. PHYSIOLOGICALLY BASED AND OTHER TOXICOKINETIC MODELING
13 Most of the understanding of the pharmacokinetics of tetrachloroethylene in humans is
14 based on a limited number of human data sets (Monster et al., 1979; Fernandez et al., 1976;
15 Volkel et al., 1998) and on extrapolations from animal data to humans using PBPK modeling.
16 PBPK models can provide estimates of tissue concentration as well as total metabolism of
17 tetrachloroethylene. Models that incorporate transfer into milk and subsequent infant exposure
18 have also been validated using measured human milk concentrations of tetrachloroethylene
19 (Byczkowski and Fisher, 1994). In addition, researchers have looked at the variability in the
20 measured and estimated PBPK model parameters and the implications for applying the models to
21 risk assessment. The critical difference among the various models is in their different
22 approaches to estimating the metabolic parameters.
23
24 3.5.1. Various Physiologically Based Pharmacokinetic (PBPK) Models
25 Chen and Blancato (1987) developed a PBPK model for rats, mice, and humans. The
26 metabolic parameters maximum velocity (Vmax) and Michaelis-Menten constant (Km) were
27 derived by fitting the model to the total amount of metabolized tetrachloroethylene.
28 Experimental data on total metabolite were available for rodents. However, for humans, it was
29 assumed that the urinary metabolite TCA, as measured by Monster et al. (1979), accounted for
30 30% of the total metabolite. This percentage was chosen because it resulted in a better fit. The
31 model consisted of five compartments: lung, fat tissue, richly perfused tissue, poorly perfused
32 tissue, and liver. The model was used to estimate cancer risk from inhalation and drinking water
33 exposures, based on total daily absorbed tetrachloroethylene.
34 Reitz et al. (1996) developed a PBPK model for rats, mice, and humans that describes the
35 total metabolism of tetrachloroethylene using Michaelis-Menten kinetics. The partition
36 coefficients for the five tissue compartments were measured independently and were similar to
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-23 DRAFT-DO NOT CITE OR QUOTE
-------
1 those used by Chen and Blancato (1987), giving confidence in the reasonableness of both sets of
2 numbers. For rats and mice, the metabolic parameters Vmax and Km, as well as the volume and
3 blood flow rates of the fat compartment, were obtained by simultaneously optimizing the fit to
4 three sets of in vivo data gathered in 6-hr inhalation radiolabeled tetrachloroethylene exposure
5 studies. These data were (a) concentration of tetrachloroethylene in exhaled breath, (b)
6 radioactive body burden present in animals at end of exposure, and (c) total post-exposure
7 radioactive metabolites recovered from all excreta and carcass homogenates.
8 The metabolic parameters for humans were estimated as follows using a "parallelogram
9 approach" (Reitz et al., 1989). First-order constants for the rate of metabolism were measured in
10 vitro using isolated liver microsomes of all three species. The ratio of these in vivo and in vitro
11 metabolic rates was assumed to be nearly constant across species, as was found to be the case for
12 rats and mice. Using this constant ratio, the human in vivo metabolic rate constant per gram of
13 liver could be determined from the human in vitro value. Km was assumed to be invariant across
14 species because it is derived solely from the reaction rate constants for the enzyme-catalyzed
15 metabolic reactions. In contrast, Chen and Blancato (1987) found very different values of Km for
16 each species because this was a fitted parameter in their model. Vmax, on the other hand, depends
17 on the concentration of the enzyme (substrate) present and is likely to exhibit large inter- and
18 intraspecies variability. As also noted by Reitz et al. (1996), there are inherent uncertainties in
19 estimates from in vitro studies.
20 Reitz et al. (1996) also used a second method for estimating Vmax, which was based on
21 extrapolation from in vivo animal studies of other chemicals metabolized by cytochrome P450
22 enzymes. Vmax, so estimated, was allometrically scaled to humans. The values obtained by Reitz
23 et al. (1996) through both these independent methods were comparable. The overall average
24 value of 32.9 mg/hr was then used in the PBPK model. This value compares with the value of
25 42.2 mg/hr used by Chen and Blancato (1987). The Chen and Blancato (1987) and Reitz et al.
26 (1996) models differ considerably in the values of Km for humans: 4.66 mg/L and 32.04 mg/L,
27 respectively. Chen and Blancato (1987) also demonstrated that because tetrachloroethylene is so
28 poorly metabolized, the levels of tetrachloroethylene in blood and tissues are not extremely
29 sensitive to the values of Vmax and Km.
30 A human PBPK model was developed for the purpose of investigating neurotoxicological
31 endpoints (Rao and Brown, 1993). In this case, tetrachloroethylene, not its metabolites, is of
32 toxicological interest. This model was similar to the others previously discussed except that it
33 included a skin compartment to allow for dermal absorption of tetrachloroethylene from shower
34 water and a brain compartment so that the researchers could evaluate tetrachloroethylene
35 concentrations in this organ. The model was coupled with an exposure model that predicted the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
amount of tetrachloroethylene a human would be exposed to from water during showering and
bathing.
The values for Vmax and Km in Rao and Brown (1993) were estimated using the method
of Reitz and Nolan (1986). The predictions of the model were fit to total metabolite levels
measured in rats and mice (Schumann et al., 1980; Pegg et al., 1979) to obtain the maximum rate
of metabolism, Vmax (which varied across species and was allometrically related to body weight
raised to 0.74 power), and Km (considered invariant across species). Other parameters for
tetrachloroethylene were derived from various experimental data reported in the literature. The
value of Vmax for humans was determined by fitting the predicted total metabolite level to that
estimated from urinary metabolite measurements in humans (Monster et al., 1979, and Fernandez
et al., 1976, combined), assuming that the ratio of urinary to total metabolites would be the same
in humans as that observed in rats (equal to 0.71). Although the value of Km for humans in the
Rao and Brown (1993) and Reitz et al. (1996) models were similar, their values for Vmax differed
significantly (see Table 3-1). Rao and Brown (1993) also provided parameter estimates for 6-
and 10-year-old children.
Table 3-1. Comparison of Vmax/Km ratios3
Model
Rao and Brown
Reitz
Bois et al.
Gearhart et al.
» max
(mg/hr)b
6.77
32.9
1.86
5.48
Km
(mg/L)
4.56
4.66
0.133C
7.7
» max' ^-m
(L/hr)
1.48
7.06
14
0.712
19
20
21
22
23
24
25
26
27
28
29
30
31
32
a Ratios are for the human models used for animal-to-human extrapolations in this assessment and those of Gearhart
et al. (1993), which accurately predicted production of major metabolite, TCA, in urine.
b 70 kg human.
0 The "posterior" value for Km in Bois et al. (1996) was multiplied by the liver volume and the liver tissue/blood
partition coefficient in order to conform to the format of the pharmacokinetic equations in this document.
Bois et al. (1996) used a Bayesian analysis in conjunction with a PBPK model that was
structurally similar to that used by Reitz et al. (1996). The analysis used "prior" empirically
determined distributions for the parameters in the model that were based on values in the
literature and other previously conducted PBPK analyses. The Markov Chain Monte Carlo
method was used with the PBPK model to compute updated "posterior" population distributions
of these parameters that provided optimal fits to the individual blood and exhaled
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-25 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene concentrations of subjects in Monster et al. (1979). Due to lack of good prior
2 knowledge about the population variability of physiological parameters, standard reference
3 values were assumed and standard deviations were selected by "reasonable guess" to generate a
4 relatively diffuse prior.
5 The Michaelis-Menten parameter, Vmax, was obtained first for rats and mice by fitting the
6 model to in vivo data on the rate of total metabolite formation (Bois et al., 1990). The
7 investigators then estimated human values by using the ratio of human and animal values of
8 Vmax, as determined by Reitz (1992). The geometric mean of these values was in agreement with
9 that obtained by extrapolation from the animal values, based on allometric scaling by body
10 weight. This value was used as the "prior" estimate of Vmax. Km was treated as invariant across
11 species, and the geometric mean of values determined for rats and mice was used as the "prior"
12 estimate. The Monte Carlo simulation was run for 10,000 iterations, after which time the
13 parameter distributions converged. Tetrachloroethylene concentrations in blood and exhaled air
14 were fit extremely well to the Monster et al. (1979) data. The shape of the prior distribution was
15 seen to have little impact on final results. Model predictions were compared against alveolar
16 concentrations of subjects in the Opdam and Smolders (1986) study, and all data points were
17 seen to fall within the 95th percentile envelope of predictions. The exposure concentrations in
18 this study were 5 to 100 times lower than those used in the Monster et al. (1979) study; thus, this
19 comparison provides further weight to the strength of the model.
20 The mean value for the posterior estimate of Vmax was 20 times lower than the prior
21 estimate. Thus, the Bois et al. (1990) results imply that the maximum rate of tetrachloroethylene
22 metabolism in humans is much lower than that extrapolated by body weight raised to 3/4 power
23 allometric scaling from rodents.
24 Other authors have developed models for tetrachloroethylene that specifically describe
25 the kinetics of its major metabolite, TCA. Gearhart et al. (1993) developed a model for
26 tetrachloroethylene that also included the kinetics of TCA, assuming that TCA comprised 60%
27 of the total tetrachloroethylene metabolized in the rodent and using similar parameters for TCA
28 as in a model for trichloroethylene. Tetrachloroethylene metabolism parameters for mice were
29 estimated by fitting the model to the time course of decrease in chamber concentration of
30 tetrachloroethylene in gas uptake studies. The model was independently validated at low oral
31 doses (acute oral gavage of tetrachloroethylene in corn oil) by comparing the time course of
32 blood concentrations of tetrachloroethylene and TCA in mice. Details pertaining to the
33 derivation of parameters for metabolism in humans are not provided in the original paper but are
34 available in the review by Clewell et al. (2005).
35 The parameters for describing tetrachloroethylene metabolism in humans were derived by
36 fitting the model to urinary excretion of TCA in two subjects in a study by Fernandez et al.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-26 DRAFT-DO NOT CITE OR QUOTE
-------
1 (1976), assuming the same ratio of TCA to total metabolite as in the rodent. This value was set
2 to 0.6 and attributed to Dekant et al. (1986b). The validity of using this value for humans has not
3 been evaluated. Reitz et al. (1996), in their radiolabeled tetrachloroethylene studies, determined
4 the fraction of urinary to total metabolites to range from 0.49 to 0.59 in rats and from 0.56 to
5 0.66 in mice for exposure concentrations that varied by two orders of magnitude.
6 Clewell et al. (2005) evaluated the Gearhart et al. (1993) model further, comparing its
7 predictions against the more recently available urinary and blood TCA data gathered by Volkel
8 et al. (1998) on human subjects exposed to tetrachloroethylene concentrations of 10 to 40 ppm
9 for 6 hrs. The predicted blood TCA concentrations were in general agreement with the
10 experimental data, but the rate of urinary excretion of TCA was overpredicted by roughly a
11 factor of 2. Clewell et al. (2005) extended the Gearhart et al. (1993) model to include
12 metabolism of tetrachloroethylene in the kidney, allowing for excretion directly into urine.
13 Assuming metabolism in this organ to be at 10% of the capacity of the liver, substantial
14 improvement was noted in the agreement with experimental data. An advantage in using the
15 Volkel et al. (1998) data is that they pertain to exposure concentrations that are lower than those
16 in other studies (e.g., 72 to 144 ppm in the Monster et al., 1979, study). In addition to
17 developing this refined model, the Clewell et al. (2005) work provides an extensive review and
18 evaluation of available PBPK models for tetrachloroethylene.
19 Loizou (2001) used a PBPK model that was structurally similar to that of Gearhart et al.
20 (1993). The model assumes a 15% stoichiometric yield for the total metabolite produced across
21 various dose levels (i.e., 15% of the parent compound in the liver is metabolized), but the basis
22 for these assumptions is not substantiated. The above yield is also assumed to hold for the
23 production of TCA because it is the major metabolite (E-mail dated June 26, 2002, from G.
24 Loizou, Health and Safety Laboratory, UK, to R. Subramaniam, U.S. EPA). Elimination rates of
25 TCA through blood and urine were chosen by calibrating the model to fit blood and urinary TCA
26 kinetics and exhaled tetrachloroethylene TCA concentration levels obtained from Monster et al.
27 (1979).
28 Other compartments have been added to human PBPK models to answer specific
29 questions. One model was developed for the purpose of predicting cancer risk in breastfed
30 infants (Byczkowski and Fisher, 1995). This model did not include a brain compartment but
31 instead included a milk compartment for the mother. Hence, milk concentrations were predicted
32 on a real-time basis, and the daily dose to a nursing infant could be computed. To assess cancer
33 risk, the authors used a standard method based on intake dose of the parent compound (U.S.
34 EPA, 1989).
35 This document has described three human PBPK models—those of Rao and Brown
36 (1993), Reitz et al. (1996), and Bois et al. (1996)—in order to extrapolate health risk from
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-27 DRAFT-DO NOT CITE OR QUOTE
-------
1 laboratory animals to human. The rationale for the selection of these models is discussed in
2 Section 3.5.3.
O
4 3.5.2. Variability and Uncertainty
5 A number of models and studies have been discussed in the preceding sections on animal
6 and human pharmacokinetics, and other models have appeared in older literature (see Hattis et
7 al., 1990; Clewell et al., 2005). These models can be shown to adequately predict human data on
8 concentrations of the parent tetrachloroethylene compound in blood and exhaled air and have
9 been used to varying degrees along with cancer risk models to attempt to predict the risk of
10 tetrachloroethylene exposure. There is likely to be considerable biological variability in many of
11 these parameters, and the uncertainties about the values and their interpretations are significant.
12 Variability in pharmacokinetic measurements can exist within an individual over repeated
13 measurements (intra-individual variability) and between different individuals in a population
14 (inter-individual variability). Although this variability can introduce uncertainty into risk
15 assessments that are based on single-point estimates, it is also a factor that can be explored using
16 physiologically based mathematical modeling (including PBPK models) along with statistical
17 techniques to estimate parameter distributions. Some studies have attempted to examine how
18 variability among individuals affects risk (Bois et al., 1990, 1996; Gearhart et al., 1993;
19 Isukapalli et al., 1998).
20 Bois et al. (1996) used Markov Chain Monte Carlo analysis to investigate the sensitivity
21 of model output to changes in parameters and to determine the parameter distributions needed to
22 explain the intersubject variability in humans. In simple Monte Carlo analysis, parameter values
23 are selected from predetermined empirical or experimental distributions to investigate variability
24 in model output. The Markov chain technique takes into account prior knowledge and collected
25 data to modify the parameter distributions. In this case, population distributions were developed
26 using knowledge of the six subjects in the Monster et al. (1979) data set for which a fit was
27 desired. The new (posterior) parameter distributions were then used to estimate the amount of
28 tetrachloroethylene metabolized during ambient exposures of approximately 1 ppb inhaled dose.
29 The investigators estimated that a median value of 36% was metabolized, with 95% confidence
30 bounds of 15% and 58% at low inhalation exposure concentrations of 0.001 ppm, in contrast to a
31 median value of 1.7% metabolized at a 50 ppm concentration.
32 In an earlier study, Bois et al. (1990) used the conventional Monte Carlo method in
33 conjunction with PBPK modeling to consider the effect of pharmacokinetic parameter
34 uncertainties on the precision in model predictions and, subsequently, on cancer risk estimates.
35 Empirically reasonable probability distributions were assigned to the scaling coefficients
36 associated with each parameter. It was acknowledged that certain parameters would co-vary (for
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-28 DRAFT-DO NOT CITE OR QUOTE
-------
1 example, body weight and liver weight), so these parameters would not be altered independently
2 in the Monte Carlo simulation. The metabolic parameters were seen to be the most important
3 determinants in the sensitivity of the results, as also observed by Reitz et al. (1996). The
4 investigators calculated a median rate of metabolism in humans of 58 ng/day/kg2/3, with 5th and
5 95th percentiles of 34 and 104 ng/day/kg2/3, upon continuous exposure to 1 ng/L of
6 tetrachloroethylene. The variability in the rate of metabolism was estimated to be much lower
7 than that expected from the variability in Vmax and Km due to the covariance of these two
8 parameters.
9 To assess variability in uptake and elimination within a single individual over multiple
10 exposures under different exposure patterns, Chi en (1997) collected exhaled breath
11 measurements on a single individual following four different exposure scenarios in a controlled
12 environmental facility (three replicates per scenario for a total of 12 exposures) and following
13 tetrachloroethylene exposure in 22 dry cleaning facilities, where ambient levels of
14 tetrachloroethylene were recorded and exposures were carefully timed. Hence, the variability in
15 exhaled breath as a biomarker measurement and surrogate for internal dose could be evaluated in
16 the same individual under clinical conditions in which exposure magnitude, duration, and pattern
17 were carefully controlled and in a field environment where only the duration of exposure was
18 controlled. The controlled exposures occurred for either 30 minutes or 90 minutes, with
19 exposure concentrations ranging from 0.5 to 3 ppm. The experiments were designed to result in
20 potential inhalation exposures of 297 |ig/L-min. Differences in percent uptake and elimination
21 half-life between exposure sessions at the same environmental concentration were statistically
22 insignificant. However, percent uptake was dependent on environmental concentration.
23 Gearhart et al. (1993) performed 600 runs of a PBPK model in Monte Carlo fashion to
24 produce a distribution of output results and attempted to look at the effect of the variation in the
25 values of partition coefficients on the prediction of different dose surrogates such as area under
26 the blood time curve for metabolites in the liver. For this dose surrogate, the investigators
27 determined that the coefficient of variation was 25% and that the maximum was less than twice
28 the mean. They concluded that parameter uncertainty in the models does not constitute a
29 significant source of variability in using PBPK models for risk assessment. It must be noted that
30 variation of the metabolic parameters was not included in their exercise.
31 The quantity of metabolite produced was observed to be very sensitive to Vmax and Km,
32 parameters that vary significantly across models. On the other hand, the concentration of
33 tetrachloroethylene in the blood is relatively insensitive to these parameters. Tables 3-1 and 3-2
34 indicate the range in the values of these parameters reported in the literature for humans and
35 laboratory animals.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-29 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 3-2. Variation in values of metabolic parameters for
tetrachloroethylene, as seen in the literature
Subject
Reference
'max
(mg/hr)
Km
(mg/L)
Comment
Animal
Sprague-
Dawley rat
B6C3F1
mouse
F344 rat
B6C3F1
mouse
Sprague-
Dawley rat
Chen and Blancato
(1987)
Chen and Blancato
(1987)
Reitzetal. (1996)
Reitzetal. (1996)
Dallas et al.
(1994a,b)
Byczkowski and
Fisher (1994)
0.35
0.18
0.325
0.355
0.009
0.012
2.94
1.47
5.62
3.69
0.019
0.32
Fit model to 10 and 600 ppm inhalation exposure
from Peggetal. (1979).
Least squares fit to total metabolized from a gavage
and inhalation study of Pegg et al. (1979).
Optimization to fit entire data set collected and
reported in this study for 6 hr inhalation exposures.
Metabolic parameters and size/perfusion rate of fat
compartment were fit.
Same as F344 rats in this paper (above).
Metabolic parameters and blood/air partition
coefficient estimated via nonlinear regression to an
intra-arterial injection of 10 mg/kg in the rats.
Fit model to exhaled tetrachloroethylene from
closed chamber inhalation studies.
Humans
10-year-old
child
3 -year-old
child
Chen and Blancato
(1987)
Reitzetal. (1996)
Rao and Brown
(1993)
Boisetal. (1996)
Rao and Brown
(1993)
Rao and Brown
(1993)
Byczkowski and
Fisher (1995)
42.2
32.9
6.77
1.86
4.25
2.64
2.94
32
4.66
4.56
0.1333
4.56
4.56
0.32
Least squares fit to urinary TCA data from Monster
et al. (1979) and Fernandez et al. (1976), assuming
that TCA represents 30% of overall metabolism.
Determined from in vitro studies with human liver
cells and in vitro and in vivo studies in animals.
Published literature.
Central estimate, Markov Chain Monte Carlo
posterior distribution fit to Monster et al. (1979)
exhaled and venous blood concentrations.
Published literature.
Published literature.
Allometric scaling from model optimizations based
on the exhaled breath of rats during closed chamber
inhalation exposure.
4
5
See text and footnote in Table 3-1.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-30 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Age and gender-specific differences in pharmacokinetics can have a significant impact on
tissue dosimetry. The immaturity of metabolic enzyme systems in the perinatal period may lead
to decreased clearance of toxic chemicals as well as decreased production of reactive
metabolites. Clewell et al. (2004) examined these differences for various stages in life using
PBPK modeling for tetrachloroethylene and five other chemicals that differed considerably in
their physicochemical (lipophilicity, solubility, and volatility) and metabolic characteristics.
Parameters describing growth of various tissues were taken from the literature, and blood flow
changes with age were assumed to change proportionally with tissue volumes. For
tetrachloroethylene, only oxidative metabolism—specifically the production of TCA—was
considered. Data on age-dependent development of CYP2E1 was used for this purpose (Vieira
et al., 1996). The parameters for tetrachloroethylene were taken from the Gearhart et al. (1993)
model, and the age-dependence of metabolism was based on the CYP2E1 data. The Gearhart et
al. (1993) model describes the amount of TCA produced as 60% of the total metabolized
tetrachloroethylene; this was fixed in the life-stage model.
The dose metrics examined were blood concentrations of the parent compound and the
metabolite TCA. Continuous lifetime oral exposure was simulated at a daily dose rate of 1
|ig/kg/day. Table 3-3 provides the average daily dose during different life stages of a male
expressed relative to that of a 25-year-old adult male. The gender and age differences in
tetrachloroethylene and TCA blood concentrations are detailed further in Figure 3-3.
Table 3-3. Ratio of average daily dose at various life stages to the average
daily dose for a 25-year-old adult: PBPK simulations
Dose metric
Tetrachloroethylene blood
concentration
TCA blood concentration
Life stage
0-6 months
0.33
0.057
0.5-5 years
0.42
0.16
5-25 years
0.76
0.59
25-75 years
1.2
1.4
24
25
26
27
Source: Clewell et al. (2004).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-31 DRAFT-DO NOT CITE OR QUOTE
-------
1.5E-4
2
3
4
5
6
1
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
o>
O
f£
UJ
Q-
"5
d
§
O
-Q
8
CO
1.0E-4--
5.0E-5-
O.OE+0
- mates
PERC - females
TCA - males
- females
5.0E-8
--4.0E-8
f 3.0E-8 P
4-2.0E-8 g
P
"D
O
1.0E-8 £
n | I 11 I jI11 I 111 11 I IiI IiIt I iI I I M I I (
O.OE+0
0 10 20 30 40 50 60 70 80
Age (years)
Figure 3-3. PBPK simulations of variations with age and gender in blood
concentrations of tetrachloroethylene and its main metabolite trichloroacetic
acid (TCA). Simulations are for continuous lifetime oral exposure at a constant
daily intake of 1 |ig/kg/day.
Source: Clewell et al. (2004).
Considerable gender differences in blood concentrations of TCA and tetrachloroethylene
were seen in these predictions. Internal dose during infancy differed most from the
corresponding dose in a 25-year-old. Tetrachloroethylene and TCA blood concentrations
increased with age, which the authors attributed to the lower metabolic and pulmonary clearance
of tetrachloroethylene when compared with other volatiles as well as its higher lipophilicity, both
resulting in storage of the compound in fat and other tissues. These age and gender differences
in pharmacokinetic sensitivity are significant, but they need to be considered together with
pharmacodynamic considerations in determining the contribution of exposure at a life stage to
lifetime risk.
The same group of authors (Gentry et al., 2003) developed a PBPK model for
tetrachloroethylene that compared maternal and fetal/neonatal blood and tissue dose metrics
during pregnancy and lactation. The manuscript contains the details on the structure of the
model. Oxidative metabolism (TCA) in the mother and lactating infant was modeled using data
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-32 DRAFT-DO NOT CITE OR QUOTE
-------
1 for CYP2E1 (Vieira et al., 1996); metabolism in the fetus was not included due to lack of
2 information pertaining to the development of this pathway during gestation. The dose metrics
3 were the fetal and infant blood concentrations of tetrachloroethylene and TCA. Changes in fetal
4 blood concentrations were not pronounced because changes in tissue composition occurred in
5 both mother and fetus during pregnancy (Gentry et al., 2003). A decrease of nearly three orders
6 of magnitude of blood concentrations in the lactating infant when compared with that of the fetus
7 was calculated. This decrease was attributed to the lower exposure rate during lactation as
8 compared with placental exposure. Concentrations in the lactating infant were considerably
9 lower, by more than two orders of magnitude, than in the mother. The largest variation in blood
10 concentration occurred in the early postnatal period.
11 As the authors indicated, validation of the results in the Clewell et al. (2004) and Gentry
12 et al. (2003) work and further refinement of the parameters in the models are necessary. It would
13 therefore be premature to consider the results of such analyses for use in risk assessment.
14 Further investigation of variability in the parameters used in the Clewell et al. (2004) analysis is
15 needed before the results from Table 3-3 can be used to weigh upon considerations of a
16 pharmacokinetic uncertainty factor for age and gender variability. Nonetheless, these models
17 will enable future studies to focus on the key factors that are likely to influence pharmacokinetic
18 susceptibility.
19
20 3.5.3. Animal-to-Human Extrapolation Using a Physiologically Based Pharmacokinetic
21 (PBPK) Model
22 3.5.3.1. Choice of Physiologically Based Pharmacokinetic (PBPK) Model
23 As explained above, the evidence suggests that by-products of tetrachloroethylene
24 metabolism are implicated in carcinogenesis in both rodent species. Inhaled concentration of the
25 parent compound is therefore not an appropriate dosimeter. The use of pharmacokinetic
26 modeling is expected to be useful in this regard. Various dose metrics are explored in detail in
27 Chapter 4. Because the choice of the most appropriate dose metric has bearing on our selection
28 of PBPK models, the issues are briefly summariuzed here.
29 Both the oxidative and GSH-dependent pathways of tetrachloroethylene metabolism are
30 known to be involved significantly in the various tumors. Tetrachloroethylene hepatotoxicity is
31 associated with cytochrome P450 metabolism occurring in the liver. TCA is considered to be the
32 predominant metabolite associated with this P450 oxidation pathway. However, TCA may not
33 be the sole contributory metabolite to tetrachloroethylene-induced hepatotoxicity and cancer, and
34 reactive intermediates such as tetrachloroethylene oxide and trichloroacetyl chloride may also be
35 involved. In the case of renal toxicity, GSH conjugates formed in the liver and transported to the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-33 DRAFT-DO NOT CITE OR QUOTE
-------
1 kidney are thought to be the primary agents. The GSH pathway is also implicated in the mode of
2 action for leukemia (see Chapter 4).
3 Tetrachloroethylene is a chemical that has generated prolific pharmacokinetic modeling
4 endeavors. A consideration in determining appropriate PBPK model structures for use in risk
5 assessment is the ability to use the same model to predict dose metrics for all the endpoints.
6 Although many models have been developed to predict concentrations of the parent compound
7 and total metabolite levels, only the model by Gearhart et al. (1993) and its variations, developed
8 by Clewell et al. (2005) and Loizou (2001), predict both tetrachloroethylene and TCA
9 concentrations. These models were reviewed in previous sections. As noted in Chapter 4, there
10 is no reliable quantitative data on GSH conjugates formed by tetrachloroethylene; accordingly,
11 there are no models that can specifically predict these metabolite levels.
12 Various uncertainties are associated with the use of PBPK models developed to predict
13 the kinetics of TCA produced as a result of tetrachloroethylene metabolism. One assumption
14 pertains to the fraction of tetrachloroethylene metabolized to TCA in humans. Loizou (2001)
15 made the assumption that 15% of tetrachloroethylene reaching the liver is metabolized. In
16 related models, Gearhart et al. (1993) and Clewell et al. (2005) estimated their parameters on this
17 assumption that urinary TCA in humans accounts for 60% of the total metabolism of
18 tetrachloroethylene. This percentage was assumed to be independent of dose and was based on a
19 range around the value seen in rodents (Reitz et al., 1996; Clewell et al., 2005); however, its
20 reliability for humans is not known.
21 At an exposure concentration of 72 ppm in humans, Monster et al. (1979) determined that
22 95% of inhaled tetrachloroethylene was exhaled unmetabolized, 2% was excreted as TCA in
23 urine, and 1% of TCA remained in systemic circulation. Thus, these data suggest that urinary
24 TCA may comprise roughly 40% of the total metabolite in humans. It may be noted that other
25 tetrachloroethylene metabolites are also known to be excreted in the urine of exposed humans
26 (Ikeda and Ohtsuji, 1972).
27 The measurement of urine levels of TCA using the photometric Fujiwara reaction
28 method, which was used in the Fernandez et al. (1976) and Monster et al. (1979) studies, is
29 hindered by analytical or methodological problems, such as providing information only on the
30 total trichloro content and blank levels being significantly high in unexposed subjects (Reitz et
31 al., 1996). The TCA measurements in the Fernandez et al. (1976) study were used in estimating
32 the metabolic parameters in the Gearhart et al. (1993) model. Other problems include the half-
33 life of TCA in humans being long—in the neighborhood of 100 hrs (Muller et al., 1974).
34 TCA alone may not be an adequate dose metric for liver tumors. Buben and O'Flaherty
35 (1985) compared various indices of liver toxicity for tetrachloroethylene and trichloroethylene,
36 finding tetrachloroethylene to be at least twice as potent as trichloroethylene on a molar basis for
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-34 DRAFT-DO NOT CITE OR QUOTE
-------
1 equivalent amounts of total metabolite generated. They concluded that the toxic metabolite in
2 tetrachloroethylene is considerably more toxic than that in trichloroethylene with regard to liver
3 toxicity. In Appendix 4A, we compare the potency of liver tumors in the TCA and
4 tetrachloroethylene bioassays and conclude that TCA (produced in the metabolism of
5 tetrachloroethylene) alone does not appear to be sufficient to account for the tumorigenicity of
6 tetrachloroethylene for the exposures that were examined. Clewell et al. (2005) had similar
7 conclusions and suggest a combination of metabolites as the responsible agents.
8 All the above factors combined led to the use of the rate of overall metabolism (the total
9 amount of tetrachloroethylene metabolized per day) as a surrogate for the toxic dose in the route-
10 to-route and animal-to-human extrapolations for liver tumors, leukemia, and kidney tumors.
11 This is more reasonable than using the parent chemical, even though total metabolized dose is
12 not a perfect dose metric in that it does not actually estimate the tissue concentration of toxic
13 metabolites. Inhaled concentration of the parent chemical was used as the basis for
14 extrapolations for other cancers.
15 In this assessment, three of the most recently developed human PBPK models that predict
16 total metabolite levels were considered: those of Rao and Brown (1993; the Rao and Brown
17 model), Reitz et al. (1996; the Reitz model), and Bois et al. (1996; the Bois model). These three
18 models were chosen to allow cancer risk estimates to reflect uncertainties that arise from using
19 different data and methods to calibrate human PBPK models. This enables provision of a range
20 of values for extrapolation from laboratory animals to human. In later sections, these models are
21 compared with each other and with experimental data. The three models were chosen on the
22 basis of their different approaches to estimating metabolic parameters, as summarized in the
23 previous section. Although the models describe the overall metabolism of the parent compound,
24 they do not describe the kinetics of the metabolites.
25 The Reitz model used in vivo rodent data on total metabolism and parent compound
26 concentrations in blood and exhaled breath. The development of a human model used a
27 "parallelogram" approach wherein in vivo metabolic rate constants were related to
28 experimentally determined in vitro values by assuming the relationship of in vivo to in vitro
29 metabolic rates to be invariant across species. The Bois et al. model, on the other hand, used
30 Bayesian inference methods to fit model predictions to laboratory data on exhaled air and blood
31 concentrations of tetrachloroethylene in human volunteers. The Rao and Brown model used the
32 same human study but assumed the ratio of urinary TCA to total metabolite levels to be equal to
33 0.71 in order to derive metabolic parameters. The Rao and Brown model was included to permit
34 examination of the range in risk values that would arise if metabolic parameters are derived from
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-35 DRAFT-DO NOT CITE OR QUOTE
-------
1 urinary TCA data extrapolated to total metabolites with this assumption. l These and other PBPK
2 models generally seem to predict parent concentrations well; however, they differ considerably
3 in their predictions of the amount metabolized. As shown in Tables 3-1 and 3-2, there are large
4 differences in the metabolic parameters used by various authors.
5 The human PBPK models chosen suffer from the limitation that their predictions of total
6 metabolite cannot be accurately evaluated because such data are not available. The models
7 predict only total metabolite levels, so it is not possible to validate them against specific
8 metabolites, such as TCA, that have been measured in experiments. These models have been
9 validated against concentrations of the parent compound in blood and exhaled breath. However,
10 because the metabolism of tetrachloroethylene is slow (particularly in humans, with roughly 95%
1 1 of tetrachloroethylene being exhaled unmetabolized), the concentrations of the parent compound
12 are not sensitive to the values of the metabolic parameters. A similar argument applies to our use
13 of the Bois model. The posterior distributions of parameters in this model were obtained by
14 fitting to the parent compound concentrations in the Monster et al. (1979) study. Further, as
15 explained above, these three models are limited in different ways.
16
17 3.5.3.2. Implementation of Physiologically Based Pharmacokinetic (PBPK) Models
18 Implementation of the Rao and Brown, Reitz, and Bois models follows the PBPK model
19 structure of Ramsey and Andersen (1984). The Reitz and Bois models are composed of four
20 compartments: poorly perfused tissues, well-perfused tissues, fat, and liver. The Rao and Brown
21 model contains, in addition, a separate compartment for the brain. In the implementation of the
22 Rao and Brown model herein, there is no separate skin compartment. The compartments are
23 assumed to be homogeneous, and distribution is limited by blood flow. The metabolism of
24 tetrachloroethylene is modeled by a Michaelis-Menten term in the differential equation for the
25 liver compartment. The simulation is represented by the following equations:
26
,
at
28 — = Q,(Cart -CJ -- ^— Cv/
dt ^l^ art vl) Km+Cvl vl
29
30
Another approach is to adjust the urinary TCA predicted by Clewell et al. (2005) by the inverse of this
factor to derive total metabolite levels. A preprint of this in-press manuscript was not received in time to be able to
exercise the Clewell et al. (2005) model for the purposes of this document.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-36 DRAFT-DO NOT CITE OR QUOTE
-------
1 where:
21= compartments other than liver
3 / = liver
4 M; = mass of tetrachloroethylene in the ith compartment
5 Cvi = venous concentration of tetrachloroethylene at the exit from compartment I
6 Cart = arterial concentration of tetrachloroethylene
7 Q; = blood flow rate into the ith compartment
8 Vmax, Km = Michaelis-Menten constants.
9
10 Pulmonary exchange is represented by:
11
12 Qalv (Cinh - Calv ) = Qtot (Cart - Cven )
13 Cart = "ba L-a/v
14 Cexh = 0.67Ca/v % 0.33C,nh
15
16 where:
17 Qaiv = alveolar ventilation rate (which is different from the inspiratory flow rate
18 because of the respiratory dead space). The alveolar ventilation rate and cardiac
19 output (the ratio that is referred to as the ventilation-to-perfusion ratio) increase
20 with activity but at different rates
21
22 Qnh = inhaled concentrations of the chemical
23
24 Caiv = alveolar concentrations of the chemical
25
26 Qtot = total blood flow rate (equal to the cardiac output)
27
28 Cart = arterial concentrations of the chemical
29
30 Cven = venous concentrations of the chemical
31
32 hba = blood/air partition coefficient
33
34 Cexh = exhaled concentrations of the chemical
35
36 For oral exposures, the gastric route was added by assuming "first-pass" metabolism—by
37 assuming that all tetrachloroethylene ingested is transported directly to the liver, the
38 metabolizing organ. A separate PBPK compartment for the stomach was, therefore, not
39 necessary. The absorption of tetrachloroethylene in the stomach was modeled as a first-order
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-37 DRAFT-DO NOT CITE OR QUOTE
-------
1 process with an absorption rate constant, ka. Then the mass balance equation for the liver may be
2 modified to have an additional source term as follows:
3
C +*
at v
"' Km+Lvl
5
6 MO is amount of tetrachloroethylene ingested and is itself a function of time. In these
7 simulations, tetrachloroethylene was administered via drinking water as a series of boluses.
8 These model equations were solved using the Simulink Module of the MATLAB
9 computational software package (The Mathworks, Natick, MA) and the single-point estimation
10 module in the software package MCSIM (Bois et al., 1996). It was verified that both packages
1 1 produced the same results when applied to the same set of equations and parameters.
12 A note is in order regarding this implementation of the Bois model. Bois's Bayesian
13 approach produces a (posterior) distribution of parameters and, therefore, a distribution rather
14 than a point estimate of dose. However, this assessment is not carried out within such a
15 statistical framework. The central estimate of the parameters in the Bois et al. posterior
16 distribution was used to provide point PBPK estimates of internal dose of tetrachloroethylene
17 and of its overall metabolic rate. The point estimates obtained in this manner reproduce
18 (coincide with) the median population estimated by Bois et al. for the amount of
19 tetrachloroethylene metabolized for a large range of exposure concentration, 0.001 to 50 ppm. It
20 is therefore reasonable to use the central estimates of Bois's posterior distribution of parameters
21 to provide point estimates of dose for extrapolation purposes.
22 Most human PBPK models have been implemented to investigate inhalation exposure
23 and do not incorporate gastric absorption rate constants. Values in the literature for the gastric
24 absorption rate vary widely. Ward et al. (1988) reported a gastric absorption rate constant in
25 mice of 0.5 L/hr. Dallas et al. (1995) reported oral absorption rate constants in rats and dogs as
26 1.5 and 20.4 L/hr, respectively, obtained by fitting blood concentrations following oral gavage.
27 For modeling purposes, a gastric absorption rate constant of 1 .6 L/hr was chosen. This predicts a
28 reasonably rapid gastric absorption consistent with the data. It was determined that the resulting
29 blood concentrations of tetrachloroethylene are not particularly sensitive to larger values of this
30 parameter. Simulations of gastric absorption of tetrachloroethylene were carried out for humans
3 1 for use in route-to-route extrapolation. Because these simulations were at low exposures, and
32 because of first-pass metabolism effects, the uncertainty in the gastric absorption rate constant is
33 not likely to significantly affect the results of the extrapolation. Increasing the gastric absorption
34 rate constant to 20 L/hr results in an approximately twofold increase in peak blood concentration.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-38 DRAFT-DO NOT CITE OR QUOTE
-------
1 Changing this parameter does not substantially impact the elimination profile. Table 3-4 shows
2 the parameter sets used in this modeling effort.
3 For inhalation exposures, ventilation rate is a key parameter. In rodents, ventilation rate
4 (VE) was calculated as a function of body weight using the following equations (U.S. EPA,
5 1994):
6 For mice: VE(L/min) = ea326+L05 ln^
1 For rats: VE(L/min) = e°-578+0-821 ln^
8
9 where w is body weight in kilograms and In represents the natural log operation. These
10 equations provide total ventilation rate. The alveolar ventilation rate is the total ventilation rate
11 less the volume of air that is inhaled through the physiological dead space (total effective volume
12 not involved in gas exchange) in a given time. For the rats and mice and for resting inspiratory
13 rates (7.5 L/min) in humans, Qaiv . 0.67 VE (Brown et al., 1997). For the exercising individual
14 (24 to 49 L/min), Qaiv increases up to 0.8 VE (Brown et al., 1997). For the ventilation rates
15 covered in this document, it was considered reasonable to use the relationship Qaiv . 0.67 VE
16 throughout. These values represent reasonable physiological values, recognizing that there is
17 substantial variation. The alveolar ventilation rate corresponding to the resting inhaled minute
18 volume is 5.5 L/min. However, the EPA typically assumes a total ventilation rate of 13.8 L/min
19 for a 70 kg human. Thus, unless otherwise stated, the calculations presented in this assessment
20 assume an alveolar ventilation rate of 9.3 L/min.
21 In order to extrapolate between equivalent metabolized doses in animals and humans, the
22 PBPK structure of the Reitz model was used for rats and mice, and all three PBPK models (Rao
23 and Brown, Reitz, and Bois) were used for humans. The animal PBPK models were run to
24 simulate the exposure conditions of the animal bioassay studies. During the human equivalent
25 exposures, the model was run to simulate continuous low-level chronic exposure at steady-state
26 conditions. Chapter 5 discusses the details and results of the extrapolation.
27
28 3.5.4. Comparison of Physiologically Based Pharmacokinetic (PBPK) Simulations With
29 Experimental Data
30 The models were run to simulate various experimental and clinical exposure scenarios
31 from the literature. Simulated concentration levels of tetrachloroethylene in the blood and in
32 exhaled air were compared with measured values. As discussed in a previous section, it was not
33 feasible within the constraints of these models to make credible quantitative comparisons with
34 data on urinary or blood levels of major tetrachloroethylene metabolites.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-39 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 3-4. Parameters for tetrachloroethylene PBPK modeling
Parameter
BW(kg)
Cardiac output (L/hr)b
Alveolar ventilation
(L/hr)b
Mouse
variable3
0.275 x 60 x
(BW)0-75
0.67 x VEC
Rat
variable3
0.275 x 60 x
(BW)0-75
0.67 x VEd
Human model
Rao and
Brown
70
430
558
Reitz
70
430
558
Bois
70
430
558
Tissue volumes6 (%)
Rapidly perfused
Slowly perfused
Fat
Brain
Liver
4.5
79.5
9.18
1
5.86
4.5
75.7
6.8
1
2.53
1.7
57
23.1
2
3.4
3.71
57
23.1
3.14
15.1
49.4
23.1
2.54
Blood flow (% cardiac output)
Rapidly perfused
Slowly perfused
Fat
Brain
Liver
44
19
1.98
10
25
39.2
19
6.84
10
25
41
19
5
11
24
52
19
5
24
63.7
12.9
4.88
17.9
Partition (tissue/blood)
Rapidly perfused
Slowly perfused
Fat
Brain
Liver
Blood/air
3
2.59
48.3
3
3
16.9
3.73
1.06
86.9
3.73
3.73
18.9
3.72
1.06
86.6
3.72
3.72
10.3
5.88
3.1
119.13
5.88
10.3
1.92
2.9
84.1
3.08
16
Metabolic parameters
Vmax (mg/hr)
Km (mg/L)
0.355
3.69
0.325
5.62
6.77
4.56
32.9
4.66
1.86
0.133f
Gastric absorption rate
Ka (1/hr)
1.6
1.6
1.6
1.6
1.6
4
5
6
7
8
9
10
11
The simulations in this document use 0.03 kg and 0.3 kg for the mouse and rat, respectively.
o 326+1 osinfBW)^ where VE is the minute ventiiation.
b VE(L/hr) = 60 x
C VE(L/hT) = 60 X
1 ln(BW).
d Values used in the animal-to-human extrapolation.
e A density of 0.92 and 1 g/cc was used for fat and for other compartments, respectively.
f The "posterior" value for Km in Bois et al. (1996) was multiplied by the liver volume and the liver tissue/blood
partition coefficient in order to conform to the format of the pharmacokinetic equations in this document.
BW = body weight.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-40 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Figure 3-4 shows a comparison of the blood concentration levels predicted by the three
models considered in this assessment with clinical data from the human inhalation study by
Monster et al. (1979). In that study, six male volunteers breathed 72 ppm or 144 ppm
tetrachloroethylene at rest. In a third session, they breathed 142 ppm tetrachloroethylene at rest
with two intermittent 30-minute exercise excursions. All exposures lasted a total of 4 hrs. The
researchers measured tetrachloroethylene in blood and exhaled air after exposure until almost no
tetrachloroethylene remained. They monitored TCA in blood and urine for up to 100 hrs after
exposure. Their data set is widely cited and is perhaps the most complete in terms of human
distribution and in vivo metabolism data. The concentration of tetrachloroethylene in exhaled air
and blood as measured in the Monster experiments have been used to validate several human
PBPK models. In the comparison presented here, only tetrachloroethylene exposure at 72 ppm
was considered, and simulations were carried out at two different ventilation-to-perfusion ratios
(ratio of alveolar ventilation rate to cardiac output), corresponding to occupational activity levels.
10 -,
O)
_§
c
O
CD
O
O
O
T3
O
O
CO.
O
CD
1 -
0.1 -
0.01
Bois model, VPR=1.2
Rao, Brown model, VPR=1.2
Reitz model, VPR=1.2
Bois model, VPR=1.6
Rao, Brown model, VPR=1.6
Reitz model, VPR=1.6
Monster Experiment
-20
20 40 60 80 100 120
Post Exposure Time (hrs)
140
160 180
Figure 3-4. Comparison of model predictions for blood concentration with
inhalation experiment. Tetrachloroethylene was inhaled at a concentration of 72
ppm. Simulations were performed at different ventilation-to-perfusion ratios
(VPR) and at an alveolar ventilation rate of 7 L/min (the geometric mean of
values in the Monster experiment). Standard deviations on the experimental data
were very small (e.g., 0.025 mg/L and 0.003 mg/L at 20 and 140 hrs,
respectively).
Source: Adapted from Monster et al. (1979).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-41 DRAFT-DO NOT CITE OR QUOTE
-------
1 For any particular model, the increase in ventilation-to-perfusion ratio from 1.2 to 1.6
2 does not appear to make much difference, as shown in Figure 3-4. The much closer
3 correspondence of the Bois model predictions to the Monster data is to be expected because the
4 model's posterior distribution of parameters was arrived at by fitting to the Monster data. The
5 Rao and Brown and Reitz model predictions are less than the experimental values, generally
6 within a factor of 2 and 3, respectively. These two models do not differ much in their predictions
7 of tetrachloroethylene blood concentrations.
8 Stewart et al. (1970) analyzed the expired breath of subjects repeatedly exposed to 100
9 ppm tetrachloroethylene (7 hrs/day for 5 days) using gas chromatography and infrared
10 spectrometry. Figure 3-5 compares the mean alveolar concentration of tetrachloroethylene in
11 these subjects with the results of model simulations. The subjects were assumed to be at rest
12 (alveolar ventilation rate of 5.02 L/min and a ventilation-to-perfusion ratio of 1). All three
13 models agree reasonably well with the experimental data; the Bois et al. model differs the least
14 from the experimental result.
o
o
O
JS
o
(D
10 -
1 -
O Bois model
V Rao & Brown model
D Reitz model
^ Stewart Expt
3*7 O
°v •
% v 0*
0
o
El
O
i
D
¥
50
100
150
200
250
300
350
Post-Exposure Time (hrs)
15 Figure 3-5. Comparison of model predictions for alveolar concentration of
16 tetrachloroethylene with experimental data on humans. Tetrachloroethylene was
17 inhaled at a concentration of 100 ppm, 7 hrs/day for 5 days. The experimental data
18 Stewart et al. (1970) show the mean alveolar concentration of tetrachloroethylene in these
19 subjects. Resting breathing conditions (alveolar ventilation rate of 5.02 L/min and a
20 ventilation-to-perfusion ratio of 1.0) were assumed. Some points early in the time course
21 were deleted because of difficulty in obtaining numerical values from the author's plot.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-42 DRAFT-DO NOT CITE OR QUOTE
-------
1 Opdam and Smolders (1986) exposed six human subjects to constant levels of
2 tetrachloroethylene ranging from 0.5 to 9 ppm and measured the concentration of
3 tetrachloroethylene in their exhaled breath (from which their alveolar concentrations could be
4 deduced) during exposure up to 50 to 60 minutes. Separate data were gathered on males and
5 females. Figures 3-6a and 3-6b compare their results for the ratio of alveolar to inhaled
6 concentrations of tetrachloroethylene with predictions from the three models. The experiments
7 were performed for different breathing scenarios that included normal breathing (with no breath
8 holding) as well as paused breathing with different durations of breath holding. However, the
9 simulations were carried out only for the normal breathing scenario at resting inspiratory rates
10 and with the ventilation-to-perfusion ratio set equal to 1. Body weights and lean body weights
11 were considered differently for males and females (as given in ICRP, 1975). While running the
12 Bois model, other parameters remained unchanged across gender. On the other hand, blood flow
13 distributions to various compartments differed across gender in the other two models (as given in
14 Brown etal., 1997).
15 Although simulations were carried out for a range of tetrachloroethylene exposure
16 concentrations from 0.5 to 9 ppm (only a range has been provided by the authors), the plots in
17 Figures 3-6a and 3-6b show only the simulations for 5 ppm, as there were no substantial
18 differences across this range of exposure concentrations. The agreement with this experimental
19 data is particularly noteworthy, considering that the Bois model was parameterized on the basis
20 of the Monster et al. (1979) data, in which exposures were 5-100 times higher than in the Opdam
21 and Smolders (1986) measurements.
22 The comparison shows the Reitz and Bois models' predictions to be generally closer to
23 the experimental results. The disagreement of the Rao and Brown model appears to be greatest
24 at the longer time durations: at 40 hrs, it overpredicts by roughly a factor of 1.5, whereas the
25 Reitz and Bois models are in close agreement. Alveolar concentrations in male subjects are
26 generally slightly less than those in females in both simulations and experiment.
27 Comparisons were also performed for the Altmann et al. (1990) study, in which subjects
28 were exposed to 10 ppm and 50 ppm tetrachloroethylene by inhalation for 4 hrs repeatedly on 4
29 days. Table 3-5 shows the values corresponding to measurements at the end of exposure on the
30 first day of exposure. Relative to the Reitz model, the Rao and Brown and Bois models appear
31 to compare well with the Altmann et al. (1990) experiment at lower exposure, but they predict
32 nearly twice the experimental levels at higher exposure. The comparison with the Altmann et al.
33 (1990) data may be less conclusive because the measurement was made immediately after
34 exposure, when variation in both experiment and simulation is large.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-43 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
0.45 -i
0.40 -
o
o 0.35 -
O
T3
-^ 0.30 H
CD
.c
_c
-Q °-25-
o
fe °'2CH
"o
j> 0.15 -
0.10 -
0.05
Bois model
Rao & Brown model
Reitz model
• Opdam Expt: male a
V Opdam Expt: male b
• Opdam Expt: male c
10
20
30 40
Time (hours)
50
60
70
Figure 3-6a. Comparison of model predictions for alveolar concentration as
a fraction of inhaled tetrachloroethylene concentration with experimental
Opdam and Smolders (1986) data on male human subjects. Some
physiological parameters specific for males were used (see text for details).
Experiment exposure concentrations ranged from 0.5 to 9 ppm; plots for
simulations depict only results for 5 ppm (no significant difference at other
exposures in this range). Breathing conditions at rest assumed for the
simulations: alveolar ventilation rate of 5.02 L/min and ventilation-to-perfusion
ratio of 1). Simulations and experimental data shown here are with no pause in
breathing.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-44 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
o
c
o
O
_CD
CD
O
O
O
_
o
CD
I
0.45 -
0.40 -
0.35 -
0.30 -
0.25 -
0.20 -
0.15 -
0.10 -
0.05
Bois model
Rao & Brown model
Reitz model
Opdam Expt: female a
Opdam Expt: female b
Opdam Expt: female c
.-' V
10
20
30 40
Time (hours)
50
60
70
Figure 3-6b. Comparison of model predictions for alveolar concentration as
a fraction of inhaled tetrachloroethylene concentration with experimental
data on female human subjects. Some physiological parameters specific for
females were used (see text for details). Experiment exposure concentrations
ranged from 0.5 to 9 ppm; plots for simulations depict only results for 5 ppm (no
significant difference at other exposures in this range). Resting breathing
conditions (alveolar ventilation rate of 5.02 L/min and a ventilation-to-perfusion
ratio of 1.0). Simulations and experimental data shown here are with no pause in
breathing.
Source: Opdam and Smolders (1986).
TTz/s document /'s a draft for review purposes only and does not constitute Agency policy
06/06/08 3-45 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 3-5. Comparison of venous blood tetrachloroethylene concentrations:
PBPK simulations and Altmann et al. (1990) studya
Inhaled exposure
concentration
(ppm)
10
50
Blood concentration (g/L)
Altmann
333
1,106
PBPK simulations
Bois et al.
350
1,855
Rao and Brown
385
1,940
Reitz
262
1,332
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
a Values correspond to measurements at the end of the first day of study in the Altmann et al. (1993) experiment.
It is concluded that these four comparisons provide no particular basis for preferring one
model over another. The comparisons in Figures 3-4, 3-5, 3-6a, and 3-6b indicate that, with
regard to alveolar and blood concentrations, all three models provide reasonably good
predictions and are not markedly different from each other.
The three models differ most in their values for the metabolic parameters (see Tables 3-1
and 3-2) and consequently, as shown later, in their predictions of the rate of total metabolite
production. The PBPK models presented in this document predict the rate of production of the
total amount of metabolites but do not describe their transformation and clearance. It is therefore
not possible to compare their predictions on metabolite produced with experimental data without
making major assumptions. Such a comparison was attempted with the experimental data of
Fernandez et al. (1976) on the amount of TCA excreted in urine.2 In this experiment, two
individuals were exposed to 150 ppm of tetrachloroethylene by inhalation for 8 hrs and followed
for 72 hrs. The data indicate that approximately 30 mg of TCA were eliminated through urine in
these subjects during the post-exposure period. In order to make a rough comparison, we
assumed that urinary excreted TCA constituted the bulk of overall metabolites formed.
Simulations using the Rao and Brown and Bois models predict approximately 60 mg and 320 mg
(tetrachloroethylene equivalent, determined by multiplying the amount of the metabolite by the
ratio of tetrachloroethylene and TCA molecular weights), respectively, of total metabolite
produced during the post-exposure period.
Other studies that have reported data on the concentrations of TCA in blood or urine include Stewart et al.
(1961, 1970), Monster et al. (1979, 1983), Boettner and Muranko (1969), Ikeda et al. (1972), Essing et al. (1973),
Guberan and Fernandez (1974), Volkel et al. (1998), and the New York State Department of Health (NYS DOH,
2000).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-46 DRAFT-DO NOT CITE OR QUOTE
-------
1 In contrast to this large difference, when the two models are applied to the exposure
2 scenario of the Monster et al. (1979) experiment, they predict approximately the same rate of
3 metabolite produced immediately following a short 4-hr exposure to 72 ppm tetrachloroethylene.
4 In addition to the problems in making the comparison, the discrepancies between the
5 models and with experimental data on TCA may point to large uncertainties in the parameters
6 used in these models. Because the accuracy of the models has been evaluated only against blood
7 and breath concentrations of the parent compound, their reliability for predicting the production
8 of overall metabolites is an unknown. The use of all three of these models to provide a range of
9 risk estimates is intended to capture some of this uncertainty.
10 Furthermore, there are many difficulties associated with estimates of the extent of
11 metabolism in humans based on TCA excretion reported in the experimental studies of
12 Fernandez et al. and Monster et al. Some of the problems encountered are the accurate
13 measurement of the retained dose of tetrachloroethylene from inhalation exposure, the
14 imprecision of the older methodologies using the Fujiwara reaction for metabolite quantification,
15 the possibility of an important contribution for metabolites other than TCA (e.g., oxalic acid,
16 CC>2, TCVC, or as yet unrecognized products) that may be excreted, and the relatively long half-
17 life of certain urinary metabolites, which necessitates extended collection of samples.
18 The fraction of TCA in urine relative to that in blood or that stored in body organs is not
19 known. Furthermore, TCA is only one component of metabolism. Although it is considered to
20 be the major metabolite (the conclusions from Monster et al. (1979) indicate that it may comprise
21 60% of the metabolites), some tetrachloroethylene is converted to other compounds, and not all
22 potential metabolites have been adequately evaluated. In addition, TCA, itself, might be further
23 metabolized, reducing the amount of TCA available for urinary excretion.
24 Gearhart et al. (1993) used a model similar to the one used by Rao and Brown to predict
25 the same urinary TCA data set and included a parameter for urinary excretion of TCA. A
26 comparison of the metabolic parameters is shown in Table 3-1. The ratio Vmax/Km is directly
27 proportional to the rate of metabolism at low doses of tetrachloroethylene. The values shown in
28 Table 3-1 indicate that the rate of metabolism predicted by the model of Gearhart et al. (1993),
29 which directly fit the urinary TCA data, is slightly lower than the one used by Rao and Brown.
30 The Reitz and Bois models use values of Vmax/Km that are greater than those of Gearhart et al. by
31 an order of magnitude or more.
32
33 3.5.5. Physiologically Based Pharmacokinetic (PBPK) Model Comparisons and
34 Interspecies Differences
35 For an example of tetrachloroethylene tissue concentrations in different species, blood
36 concentrations resulting from a 1 ppm inhalation exposure for a duration of 16 hrs were
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-47 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
simulated. Figure 3-7 indicates blood concentrations in both rats (using the Reitz model) and
humans (using the Rao and Brown model). The blood concentrations of the two species differ
by a factor of about 2. It will be shown later that the difference in metabolism is significantly
greater on a body-weight basis. Evidence of this exists in these plots; in particular, the shape of
the decay curve between the two species is different. The blood concentrations are higher in rats
up to approximately 20 hrs post exposure; the curves cross at this point in time. In Figure 3-7 the
decay appears faster in humans than in rats, which is likely a consequence of steady-state
behavior not having been attained.
10
20
30 40 50
Time (hours)
60
70
BO
Figure 3-7. Comparison of tetrachloroethylene concentrations in blood in
rats and humans. Blood concentrations in humans (solid line) and rats (dashed
line) from a 16-hr inhalation exposure to 1 ppm tetrachloroethylene. The PBPK
models used were those of Rao and Brown (1993) for humans and Reitz et al.
(1996) for rats.
In contrast to Figure 3-7, long-time exposure (12 days at 1 ppm) is simulated in
Figure 3-8 so as to ensure that steady state has been attained in both species. The blood
concentration shows at least two modes of decay. In the initial phase, the concentrations in the
Rao and Brown and Reitz human models decay faster than those in the model for the rat,
whereas those in the Bois et al. model decay more slowly. In the second phase, the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-48 DRAFT-DO NOT CITE OR QUOTE
-------
0.12 -i
0.10 -
£ 0.08 H
CD
0)
o
8
0.06 -
o 0.04 H
O
I
0.00 -
-0.02
Bois (Human)
Rao, Brown (Human)
Reitz (Human)
Reitz (Rat)
100
200 300
Time (hours)
400
500
2 Figure 3-8. Comparison of various model predictions of tetrachloroethylene
3 blood concentration in humans and rats following steady state. Blood
4 concentrations in humans and rats due to 12-day inhalation exposure to 1 ppm
5 tetrachloroethylene. Inspiratory rate is 13.9 L/min, ventilation-to-perfusion ratio
6 is 1.3.
7
8
9 concentrations in all three models for the human decay more slowly than those for the rat. The
10 decay as predicted by the Bois model is much slower than the decay for other models.
11 Figure 3-9 shows the daily rate of total amount of tetrachloroethylene metabolized
12 following a 6-hr exposure. It illustrates differences in metabolism among the three species
13 (mouse, rat, and human). The human model is that of Rao and Brown; the animal models are
14 those of Reitz. Steady state was not attained—at least not in the rat and human.
15 The human models differ most in the metabolic parameters Vmax and Km (see Table 3-2).
16 The effect of these differences is reflected graphically in Figure 3-10, which shows the rate of
17 metabolism after steady state has been attained as a function of inhaled exposure concentration in
18 units of milligrams per day per kilogram of body weight. At low exposures, the rate of
19 metabolism is nearly equal to the ratio Vmax/Km. The values of this ratio in the Rao and Brown
20 and Reitz models are one-tenth and one-half of the value in the Bois model. Km in the Rao and
21 Brown and Reitz models is greater than in the Bois model by a factor of 35. Therefore,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-49 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
10
03
5
"S>
.*:
O)
E
"o
.Q
03
O
0)
0.1
0.01
— Rat(BW= 0.3 kg)
.--- Human (BW = 70 kg)
. — Mouse (BW=0.025 kg)
50 100 150
Inhaled Concentration (ppm)
200
Figure 3-9. Model predictions of total tetrachloroethylene metabolites
produced following a 6-hr inhalation exposure in rats, mice, and humans.
The PBPK model parameters were from Reitz et al. (1996) for rodents and from
Rao and Brown (1993) for humans.
BW = body weight
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-50 DRAFT-DO NOT CITE OR QUOTE
-------
CD
o:
10e1
^
^
X
x
^x
s
^
•^
0.0001 0.001
/
^
F
1
^
3ois
?ao, Brown
Reitz
•'
^
^
—
^
_^»-
X
^
^*
^^
0.01 0.1 1
Inhaled Cone (ppm)
10
en
I5
13>
0.1
0.01
Bois
Rao, Brown
Reitz
10 100
Inhaled Cone (ppm)
1000
I
2
3
4
5
6
7
Figure 3-10. Model predictions of rate of total metabolism in humans at
steady state. Rate of metabolism in humans, normalized to body weight, with
continuous exposure after steady-state conditions have been reached, as predicted
by the Bois et al., Rao and Brown, and Reitz models. Inspiratory minute volume
is 13.9 L/min, ventilation-to-perfusion ratio is 1.3, and body weight is 70 kg.
77?/'s document /'s a draft for review purposes only and does not constitute Agency policy
06/06/08 3-51 DRAFT-DO NOT CITE OR QUOTE
-------
1 saturation occurs at correspondingly higher concentrations in these models (between 1,000 and
2 2,000 ppm) than in the Bois model (at about 50 ppm). At saturation, the rate of metabolism is
3 lowest in the Bois et al. model, as reflected in the relative values for Vmax. Ohtsuki et al. (1983)
4 monitored all the trichloro metabolite compounds excreted in the urine of 36 male and 25 female
5 workers exposed to tetrachloroethylene. Their line of regression of the urinary metabolite level
6 versus exposure concentration of tetrachloroethylene indicated saturation of metabolism
7 occurring at roughly 600 ppm of exposure concentration. On the other hand, in an earlier study
8 of 85 male workers, Ikeda et al. (1972) determined this saturation to occur at between 50 and
9 100 ppm.
10 In Figures 3-1 la and 3-1 Ib, the simulated rate of metabolism is shown as a function of
11 time for two extremes in exposures (0.001 ppm and 50 ppm) for the rat and humans. The
12 exposure time was considered to be 12 days (as in Figure 3-8). At a 50 ppm exposure
13 concentration, the rate of tetrachloroethylene metabolism decreases very slowly with time, post-
14 exposure.
15 Table 3-4 lists all the parameters used in the models. Vmax differs considerably between
16 the three models. Km of the Bois model is the least and is less than the value in the other human
17 models by a factor of 35. There are also other significant differences between the human
18 models. The volume of the rapidly perfused compartment with the associated blood flow and the
19 blood air partition coefficient in the Bois model are considerably different from those in the Rao
20 and Brown and the Reitz models. The perfusion per unit volume of rapidly perfused tissue is
21 much less in the Bois model (equal to 4.2) than in the Reitz (equal to 14) and Rao and Brown
22 (equal to 24) models. The partition coefficient for the slowly perfused compartment in the Rao
23 and Brown model is only about one-third that of the other two human models. The partition
24 coefficient for fat in the Reitz model is considerably higher.
25 A sensitivity analysis was carried out in order to determine the dominant parameters
26 underlying the differences between the results of the Bois and the Rao and Brown models. The
27 differences in the blood concentrations and in the amount metabolized were found to be largely
28 accounted for by the metabolic parameters (Vmax and Km), with the blood/air partition coefficient
29 playing a secondary role. Results were found to be insensitive to the other parameters in the
30 model.
31 The results of PBPK simulations of oral exposure to tetrachloroethylene are shown in
32 Figures 3-12 and 3-13. In these simulations tetrachloroethylene was orally delivered via
33 drinking water in nine bolus doses spaced 2 hrs apart during an 18-hr time period, followed by
34 6 hrs of no dosing. Because tetrachloroethylene concentrations and the rate of metabolism were
35 found to be negligible at the end of the 24-hr period, the simulation was terminated after 24 hrs.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-52 DRAFT-DO NOT CITE OR QUOTE
-------
5e-4 -i
ro 4e-4 -
T3
E 3e-4 -
*«<•
E
"o
2e-4 -
OT
0 1e-4 -
o>
Bois Human
Rao, Brown Human
Reitz Human
Reitz Rat
100 200 300 400
Time (hours)
500
600
700
1
2
3
4
5
6
7
Figure 3-lla. Rate of metabolism in rat and human models: time course for
low exposure. Rate normalized by body weight as predicted by the rat model
(Reitz model) and various human models (Bois, Rao and Brown, and Reitz) for
low-exposure concentration, 0.001 ppm. Human inspiratory minute volume is
13.9 L/min, ventilation-to-perfusion ratio is 1.3; for the rat, these parameters are
as given in Table 3-4.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-53 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
1
ro
CTi
^
E
tn
TO
4—1
QJ
E
4—
O
QJ
QL
14 -,
12 -
10 -
4 -
2 -
0 -
-2
/'
Bois Human
Rao, Brown Human
jitz Human
--------- Reitz Rat
0 100 200 300 400 500 600 700
Time (hours)
Figure 3-llb. Rate of metabolism in rat and human models: time course for
high exposure. Rate normalized by body weight as predicted by the rat model
(Reitz model) and various human models (Bois et al., Rao and Brown, and Reitz)
for high-exposure concentration, 50 ppm. Human inspiratory minute volume is
13.9 L/min, ventilation-to-perfusion ratio is 1.3; for the rat, these parameters are
as given in Table 3-2.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-54 DRAFT-DO NOT CITE OR QUOTE
-------
0.10 -i
1
2
3
4
5
6
7
8
9
10
11
12
13
10 15
Time (hours)
20
Figure 3-12. Oral ingestion of tetrachloroethylene: blood concentration in
humans versus time. Time course of venous blood concentration in humans as
predicted by the Rao and Brown model for ingested tetrachloroethylene. A total
of 76 mg tetrachloroethylene was delivered orally via drinking water in 9 bolus
doses spaced 2 hrs apart during an 18-hr time period, followed by 6 hrs of no
dosing. The Bois and Reitz models result in nearly the same blood concentrations
at this exposure concentration. The dashed line shows the corresponding steady-
state blood concentration due to inhaled tetrachloroethylene of 0.7 ppm exposure
concentration that results in the same area under the curve as the above curve
integrated over a 24-hr period. The inspiratory rate is 13.9 L/min and the
ventilation-to-perfusion ratio is 1.3.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-55 DRAFT-DO NOT CITE OR QUOTE
-------
^c
~3>
£,
E
"o
.a
CD
-i— »
0
M—
0
CD
-I-J
CD
0.0012 -
0
0
0
0
0
0
0010 -
0008 -
0006 -
0004 -
0002 -
0000 -
£ \
n
i *.
\
1 1
• i
i \
i \
\
V
! V
I v
\
\
\
\
i
? A
1 11
I M
1 '!
1 1
\ ! \
i i \
\ i \
\
\
\
\
\
\
\ \l \ \
\( \ •
ii 't t *
A A A |\ Raog
1 1 1 , l : ? '. Kao c
I |> |1 I Reitz
1 h M M
1 M M 1
1 M M
\ 1 h i \
i ' \ !
\ i \\ \\ \
\i \i \j \
* * • "X
^
'•>^
i
1
''•••^r— •— •_
10
15
20
25
Time (hours)
1 Figure 3-13. Rate of metabolism of tetrachloroethylene in humans: oral
2 exposure. Rate of metabolism (mg/min) versus time of tetrachloroethylene
3 ingested orally as predicted by two PBPK models (Rao and Brown; Reitz).
4 Tetrachloroethylene was delivered orally via drinking water in 9 bolus doses
5 spaced 2 hrs apart during an 18-hr time period, followed by 6 hrs of no dosing.
6 The oral exposures that resulted in production of 0.01 mg/kg/day of overall
7 metabolite were 21,5.1, and 2.25 mg of total ingested tetrachloroethylene for the
8 Rao and Brown, Reitz, and Bois models, respectively. The Bois model is not
9 shown here to avoid a congested draft.
10
11
12 3.5.6. Metabolic Interactions With Other Chemicals
13 Fisher et al. (2004) used PBPK modeling and complementary studies in mice to
14 investigate the effect of co-exposures of orally administered carbon tetrachloride (CT) and
15 tetrachloroethylene on metabolic interactions between the two chemicals. CT is known to inhibit
16 its own metabolism (referred to as suicide inhibition). TCA was used as a biomarker to assess
17 the inhibition of the cytochrome P450 system by CT. Oral bolus intubation in the dose range of
18 1-100 mg/kg of CT was followed by a dose of 100 mg/kg of tetrachloroethylene an hour later. It
19 was concluded that dose additivity could not be used to predict interactions between the
20 compounds in this dose range because the metabolic interactions were found to be highly
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-56 DRAFT-DO NOT CITE OR QUOTE
-------
1 nonlinear. The inhibition in metabolic capacity of tetrachloroethylene 2 hrs after administration
2 of CT and 1 hr after administration of tetrachloroethylene was found to be 5, 52, and 90% at CT
3 doses of 1.5, 10, and 19 mg/kg, respectively.
4 Dobrev et al. (2002) performed gas uptake studies in F344 rats and developed a mixture
5 PBPK model for humans to study interaction effects during co-exposure to mixtures of
6 trichloroethylene (TCE), tetrachloroethylene, and methylchloroform. Corresponding to a 10%
7 increase in TCE blood concentration, the production rates of toxic conjugative metabolites
8 exceeded 17%, pointing to a nonlinear interaction effect due to co-exposure to TCE.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-57 DRAFT-DO NOT CITE OR QUOTE
-------
1 APPENDIX FOR CHAPTER 3:
2 COMPARISONS OF TETRACHLOROETHYLENE METABOLISM
3 WITH TRICHLOROETHYLENE METABOLISM
4
5
6 3.A.I. EXTENT OF METABOLISM
7 The available data indicate that, overall, tetrachloroethylene is less extensively
8 metabolized than is the closely related chemical, trichloroethylene. The difference is due to the
9 fact that a lower fraction of a tetrachloroethylene dose is metabolized via the major oxidative
10 CYP pathway when compared with an equivalent dose of the trichloroethylene congener
11 (Ohtsuki et al., 1983; Volkel et al., 1998; Lash and Parker, 2001). For example, in balance
12 studies of humans, only about 1-3% of the estimated amounts of tetrachloroethylene inhaled
13 were shown to be metabolized to TCA and other chlorinated metabolites, although these studies
14 fail to account for total dose (see Section 3.3.2 for further discussion). These amounts can be
15 compared to the 40-75% of trichloroethylene shown to be metabolized in various human balance
16 studies similar to the ones conducted for tetrachloroethylene (U.S. EPA, 1985).
17 Because of its higher lipid solubility, tetrachloroethylene may appear to be less well
18 metabolized than trichloroethylene, at least to a certain degree, simply because it is more slowly
19 metabolized due to fat sequestration. However, the animal data from studies of the two
20 compounds provide results similar to those of the human studies regarding the relative extent of
21 metabolism. For example, using data from laboratory animal studies of tetrachloroethylene
22 (Pegg et al., 1979; Schumann et al., 1980), EPA reported the percent of tetrachloroethylene body
23 burdens excreted as unchanged parent compound following exposure to 10 and 600 ppm for
24 6 hrs to be 68 and 99%, respectively (U.S. EPA, 1985). By comparison, rats and mice exposed
25 to equivalent 10 and 600 ppm trichloroethylene doses (Stott et al., 1982) metabolized a higher
26 percentage of this compound, with mice metabolizing essentially all of the dose and rats
27 metabolizing 98 and 79% of the low and high doses, respectively.
28 Saturation of metabolism occurs at a higher dose for trichloroethylene than for
29 tetrachloroethylene; thus, at certain dose levels, the differences in the amounts of the two
30 compounds metabolized is relatively greater than at other dose levels. Tetrachloroethylene
31 appears to be a lower-affinity substrate for CYP enzymes than trichloroethylene (Ohtsuki et al.,
32 1983; Volkel et al., 1998). The Km value for tetrachloroethylene is certainly higher than the Km
33 value for trichloroethylene (Lipscomb et al., 1998).
34 Both tetrachloroethylene and trichloroethylene are liver toxicants and cause liver
35 hepatocellular carcinomas in mice. The liver toxicity, including carcinogenicity, of these
36 compounds is thought to be due to metabolites. It is interesting to note that although
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-58 DRAFT-DO NOT CITE OR QUOTE
-------
1 trichloroethylene appears to be more extensively metabolized—due to greater CYP metabolism
2 in the liver—the relative cancer potency for liver tumors is similar for the two compounds.
O
4 3.A.2. DIFFERENCES IN CYTOCHROME P450 (CYP) METABOLITES
5 TCA, DC A, chloral, and TCOH are reported biotransformation products of both
6 tetrachloroethylene and trichloroethylene; however, the relative amounts produced and the
7 precursor intermediates are different for the two compounds. TCA is the major urinary
8 metabolite for tetrachloroethylene and is also an excretion product of trichloroethylene, whereas
9 TCOH is the major trichloroethylene urinary excretion product. The formation of chloral in
10 metabolism of tetrachloroethylene has been called into question, and the measurements of TCOH
11 in urine following tetrachloroethylene exposures have also been challenged.
12 Regardless, TCOH clearly is not the significant metabolite for tetrachloroethylene that it
13 is for trichloroethylene (Lash and Parker, 2001). The fact that the major urinary metabolite for
14 tetrachloroethylene is TCA (with little, if any, TCOH being formed), whereas the major urinary
15 metabolite for trichloroethylene is TCOH in the form of its glucuronide, clearly indicates
16 qualitative and quantitative differences in precursor intermediates. Very little, if any, TCA
17 produced from tetrachloroethylene metabolism comes through chloral, either directly or
18 indirectly through TCOH (Lash and Parker, 2001). The TCA from tetrachloroethylene comes
19 through trichloroacetyl chloride, possibly via the epoxide. On the other hand, the TCA produced
20 from trichloroethylene metabolism is thought to come through chloral, both directly and through
21 TCOH enterohepatic circulation (Lash et al., 2000).
22 DCA is a biotransformation product of both tetrachloroethylene and trichloroethylene,
23 although it is believed that a greater portion of DCA coming from tetrachloroethylene
24 metabolism does not arise from CYP metabolism, but rather results from further processing of
25 TCVC, whereas the DCA coming from trichloroethylene metabolism results from CYP
26 oxidation. There are at least three potential routes to DCA in CYP metabolism of
27 trichloroethylene, yet only one likely route—dechlorination of TCA—in the CYP metabolism of
28 tetrachloroethylene. Furthermore, the amount of DCA produced from tetrachloroethylene
29 oxidative metabolism may vary across species.
30
31 3.A.3. CYTOCHROME P450 (CYP) ENZYMES
32 Quantitatively, the liver is by far the predominant site of tetrachloroethylene and
33 trichloroethylene oxidative metabolism; although most other tissues contain the CYPs that could
34 conceivably metabolize these compounds. CYP2E1 has been shown to be important in rodent
35 metabolism of trichloroethylene; however, the chemical-specific data are sparse with regard to
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-59 DRAFT-DO NOT CITE OR QUOTE
-------
1 its role in tetrachloroethylene metabolism (Doherty et al., 1996). Still, assuming that CYP2E1 is
2 important to tetrachloroethylene metabolism is not unreasonable.
3 CYP3A isoenzymes—and especially CYP2B1/2—may be important for
4 tetrachloroethylene. Costa and Ivanetich (1980) showed increased hepatic metabolism following
5 treatment with agents now known to induce these isoenzymes specifically. CYP2B1/2 is
6 probably the most important CYP isoenzyme involved in oxidative metabolism of
7 tetrachloroethylene, at least in the rat, although CYPSAs and CYP2E1 are also likely involved.
8
9 3.A.4. GLUTATHIONE-DEPENDENT METABOLISM
10 The GSH-dependent pathway for tetrachloroethylene exists in both rodents and humans,
11 and the pathway is also operative for trichloroethylene in these species (Birner et al., 1996;
12 Volkel et al., 1998). The flux through this pathway is thought to be quantitatively less than that
13 through the P450 pathway. Toxic metabolites can arise from several sources in the pathway;
14 however, for tetrachloroethylene, as well as for trichloroethylene, the GSH pathway is associated
15 with renal toxicity (Anders et al., 1988; Dekant et al., 1989; U.S. EPA, 1991; IARC, 1995; Lash
16 et al., 2000; Lash and Parker, 2001). For both compounds, recovery of urinary mercapturates,
17 the stable end-products of the GSH pathway, comprise 1% or less of the total dose (Lash and
18 Parker, 2001; Dekant et al 1986a), but this does not reflect the total flux through the GSH
19 pathway. In particular, the TCVC metabolite and the corresponding DCVC and their respective
20 N-acetylated forms derived from trichloroethylene might also act as substrates for renal beta
21 lyases and other enzymes: FMO3 and CYP3A (Dekant et al., 1988; reviewed by Anders et al.,
22 1988; Dekant et al., 1989; U.S. EPA, 1991; Lash et al., 2000; Lash and Parker, 2001; see
23 Section 3.2). It should be noted that a higher cysteine S-conjugate-to-mercapturate ratio exists
24 for tetrachloroethylene when compared to trichloroethylene, which could influence the relative
25 bioactivation and nephrotoxicity of these two compounds (Lash and Parker, 2001).
26
27 3.A.5. SUMMARY
28 Tetrachloroethylene is closely related structurally to trichloroethylene and the two
29 chemicals cause similar toxic effects, many of which are attributed to metabolic activation of the
30 parent compounds. Although tetrachloroethylene is not as extensively metabolized as
31 trichloroethylene, there is little difference in potency between the two chemicals. TCA, DCA,
32 chloral, and TCOH are reported P450 biotransformation products of both tetrachloroethylene and
33 trichloroethylene; however, the relative amounts of these metabolites produced, as well as the
34 precursor intermediates in the oxidative pathways, are different for the two compounds. The fact
35 that the two compounds produce different reactive intermediate P450 metabolites is important to
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-60 DRAFT-DO NOT CITE OR QUOTE
-------
1 consider. Excretion of urinary mercapturates indicates that, relative to P450 oxidation,
2 tetrachloroethylene is more extensively metabolized via GSH conjugation than is
3 trichloroethylene. However, these urinary excretion products do not reflect the total flux through
4 the GSH pathway since the glutathione and cysteine conjugates of both chemicals have been
5 shown to undergo further processing to products that are highly reactive. Thus, regardless of
6 similarities, both the qualitative and the quantitative differences between tetrachloroethylene and
7 trichloroethylene in metabolite production could have bearing on toxicity and tumor induction,
8 and the relative importance of various mechanisms and different modes of action contributing to
9 their toxic effects, including tumorgenesis, may vary between the two parent compounds.
10 Recognizing similarities and differences is important in attempting to understand how each of
11 these two compounds causes its toxic effects.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-61 DRAFT-DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR CHAPTER 3
2
O
4 Abbas, R; Fisher, JW. (1997) A physiologically based pharmacokinetic model for trichloroethylene and its
5 metabolites, chloral hydrate, trichloroacetate, dichloroacetate, trichloroethanol, and trichloroethanol glucuronide in
6 B6C3F1 mice. Toxicol Appl Pharmacol 147:15-30.
7
8 Alberati-Giani, D; Malherbe, P; Kohler, C; et al. (1995) Cloning and characterization of a soluble kynurenine
9 aminotransferase from rat brain: identity with kidney cysteine conjugate beta-lyase. J Neurochem 64:1448-1455.
10
11 Altmann, L; Bottger, A; Wiegand, H. (1990) Neurophysiological and psychophysical measurements reveal effects of
12 acute low-level organic solvent exposure in humans. Int Arch Occup Environ Health 62:493-499.
13
14 Amet, Y; Berthou, F; Fournier, G; et al. (1997) Cytochrome P450 4A and 2E1 expression in human kidney
15 microsomes [published erratum appears in Biochem Pharmacol 1997 Jun 15; 53(12): 1946]. Biochem Pharmacol
16 53:765-771.
17
18 Anders, MW; Lash, L; Dekant, W; et al. (1988) Biosynthesis and biotransformation of glutathione S-conjugates to
19 toxic metabolites. CritRev Toxicol 18:311-341.
20
21 Anna, CH; Maronpot, RR; Pereira, MA; et al. (1994) Ras proto-oncogene activation in dichloroacetic acid-,
22 trichloroethylene- and tetrachloroethylene-induced liver tumors in B6C3F1 mice. Carcinogenesis 15:2255-2261.
23
24 ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
25 (update). Prepared by Sciences, International, under subcontract to Research Triangle Institute, Research Triangle
26 Park, NC.
27
28 Bartels, MJ. (1994) Quantitation of the tetrachloroethylene metabolite N-acetyl-S- (trichlorovinyl)cysteine in rat
29 urine via negative ion chemical ionization gas chromatography/tandem mass spectrometry. Biol Mass Spectrom
30 23:689-694.
31
32 Birner, G; Richling, C; Henschler, D; et al. (1994) Metabolism of tetrachloroethene in rats: identification of N
33 epsilon- (dichloroacetyl)-L-lysine and N epsilon-(trichloroacetyl)-L-lysine as protein adducts. Chem Res Toxicol
34 7:724-732.
35
36 Birner, G; Rutkowska, A; Dekant, W. (1996) N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and 2,2,2-
37 trichloroethanol: two novel metabolites of tetrachloroethene in humans after occupational exposure. Drug Metab
38 Dispos 24:41-48.
39
40 Board, PG; Baker, RT; Chelvanayagam, G; et al. (1997) Zeta, a novel class of glutathione transferases in a range of
41 species from plants to humans. Biochem J 328(Pt 3):929-935.
42
43 Boettner, EA; Muranko, HJ. (1969) Animal breath data for estimating the exposure of humans to chlorinated
44 hydrocarbons. Am Ind Hyg Assoc 130:437-442.
45
46 Bogen, KT; Colston, BW, Jr.; Machicao, LK. (1992) Dermal absorption of dilute aqueous chloroform,
47 trichloroethylene, and tetrachloroethylene in hairless guinea pigs. Fundam Appl Toxicol 18:30-39.
48
49 Bois, FY; Zeise, L; Tozer, TN. (1990) Precision and sensitivity of pharmacokinetic models for cancer risk
50 assessment: tetrachloroethylene in mice, rats, and humans. Toxicol Appl Pharmacol 102:300-315.
51
52 Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol 70:347-
53 355.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-62 DRAFT-DO NOT CITE OR QUOTE
-------
1 Bolanowska, W; Golacka, J. (1972) Absorption and elimination of tetrachloroethylene in humans under
2 experimental conditions (English translation). Medycyna Pracy 23:10-119.
3
4 Bonse, G; Henschler, D. (1976) Chemical reactivity, biotransformation, and toxicity of polychlorinated aliphatic
5 compounds. CRC Crit Rev Toxicol 4:395-409.
6
7 Bonse, G; Urban, T; Reichert, D; et al. (1975) Chemical reactivity, metabolic oxirane formation and biological
8 reactivity of chlorinated ethylenes in the isolated perfused rat liver preparation. Biochem Pharmacol 24:1829-1834.
9
10 Brown, RP.; Delp, MD; Lindstedt, SL; etal. (1997) Physiological parameter values for physiologically based
11 pharmacokinetic models. Toxicol Industrial Health 13:407-484.
12
13 Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
14 and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
15
16 Bull, RJ. (2000) Mode of action of liver tumor induction by trichloroethylene and its metabolites, trichloroacetate
17 and dichloroacetate. Environ Health Perspect 108(suppl 2):241-259.
18
19 Byczkowski, JZ; Fisher, JW. (1994) Lactational transfer of tetrachloroethylene in rats. Risk Anal 14:339-349.
20
21 Byczkowski, JZ; Fisher, JW. (1995) A computer program linking physiologically based pharmacokinetic model
22 with cancer risk assessment for breast-fed infants. Comput Methods Programs Biomed 46:155-163.
23
24 Calabrese, E. (1983) Principles of animal extrapolation. New York: John Wiley and Sons.
25
26 Campbell, JA; Corrigall, AV; Guy, A; et al. (1991) Immunohistologic localization of alpha, mu, and pi class
27 glutathione S-transferases in human tissues. Cancer 67:1608-1613.
28
29 Chen, C; Blancato, J. (1987) Role of pharmacokinetic modeling in risk assessment: perchloroethylene as an
30 example. In: Safe Drinking Water Committee, SoPNRC, eds. Pharmacokinetics in risk assessment. Washington,
31 DC: National Academy Press; pp. 367-388.
32
33 Chien, Y-C. (1997) The influences of exposure pattern and duration on elimination kinetics and exposure
34 assessment of tetrachloroethylene in humans. Ph.D. Dissertation, Rutgers University, New Brunswick, NJ.
35
36 Chiu, WA; Micallef, S.; Monster, A.C.; and F.Y. Bois. (2007). Toxicokinetics of inhaled trichloroethylene and
37 tetrachloroethylene in humans at 1 ppm: empirical results and comparisons with previous studies. Toxicol Sci 95
38 (l):23-36.
39
40 Clewell, HS, III; Gentry, PR; Covington, TR; et al. (2000) Development of a physiologically based pharmacokinetic
41 model of trichloroethylene and its metabolites for use in risk assessment. Environ Health Perspect 108 (Suppl)
42 2:283-305.
43
44 Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
45 pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79(2):381-93.
46
47 Clewell, HJ; Gentry, PR; Kester, JE; et al. (2005) Evaluation of physiologically based pharmacokinetic models in
48 risk assessment: an example with perchloroethylene. Crit Rev Toxicol (accepted for publication).
49
50 Cooper, AJ. (1994) Enzymology of cysteine S-conjugate beta-lyases. Adv Pharmacol 27:71-113.
51
52 Costa, AK; Ivanetich, KM. (1980) Tetrachloroethylene metabolism by the hepatic microsomal cytochrome P- 450
53 system. Biochem Pharmacol 29:2863-2869.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-63 DRAFT-DO NOT CITE OR QUOTE
-------
1 Cummings, BS; Zangar, RC; Novak, RF; et al. (1999) Cellular distribution of cytochromes P-450 in the rat kidney.
2 Drug Metab Dispos 27:542-548.
3
4 Cummings, BS; Lasker, JM; Lash, LH. (2000a) Expression of glutathione-dependent enzymes and cytochrome
5 P450s in freshly isolated and primary cultures of proximal tubular cells from human kidney. J Pharmacol Exp Ther
6 293:677-685.
7
8 Cummings, BS; Parker, JC; Lash, LH. (2000b) Role of cytochrome P450 and glutathione S-transferase alpha in the
9 metabolism and cytotoxicity of trichloroethylene in rat kidney. Biochem Pharmacol 59:531-543.
10
11 Dallas, CE; Chen, XM; Muralidhara, S; et al. (1994a) Use of tissue disposition data from rats and dogs to determine
12 species differences in input parameters for a physiological model for perchloroethylene. Environ Res 67:54-67.
13
14 Dallas, CE; Chen, XM; O'Barr, K; et al. (1994b) Development of a physiologically based pharmacokinetic model
15 for perchloroethylene using tissue concentration-time data. Toxicol Appl Pharmacol 128:50-59.
16
17 Dallas, CE; Chen, XM; Muralidhara, S; et al. (1995) Physiologically based pharmacokinetic model useful in
18 prediction of the influence of species, dose, and exposure route on perchloroethylene pharmacokinetics. J Toxicol
19 Environ Health 44:301-317.
20
21 Daniel, J. (1963) The metabolism of 36-Cl-labeled trichloroethylene and tetrachloroethylene in the rat. Biochem
22 Pharmacol 12:795-802.
23
24 Davidson, IW; Beliles, RP. (1991) Consideration of the target organ toxicity of trichloroethylene in terms of
25 metabolite toxicity and pharmacokinetics. Drug Metab Rev 23:493-599.
26
27 Dekant, W; Vamvakas, S; Berthold, K; et al. (1986a) Bacterial beta-lyase mediated cleavage and mutagenicity of
28 cysteine conjugates derived from the nephrocarcinogenic alkenes trichloroethylene, tetrachloroethylene and
29 hexachlorobutadiene. Chem Biol Interact 60:31^4-5.
30
31 Dekant, W; Metzler, M; Henschler, D. (1986b) Identification of S-l,2-dichlorovinyl-N-acetyl-cysteine as a urinary
32 metabolite of trichloroethylene: a possible explanation for its nephrocarcinogenicity in male rats. Biochem
33 Pharmacol 35:2455-2458.
34
35 Dekant, W; Martens, G; Vamvakas, S; et al. (1987) Bioactivation of tetrachloroethylene. Role of glutathione S-
3 6 transferase-catalyzed conjugation versus cytochrome P-450-dependent phospholipid alkylation. Drug Metab Dispos
37 15:702-709.
38
39 Dekant, W; Berthold, K; Vamvakas, S; et al. (1988) Thioacylating intermediates as metabolites of S-(l,2-
40 dichlorovinyl)-L- cysteine and S-(l,2,2-trichlorovinyl)-L-cysteine formed by cysteine conjugate beta-lyase. Chem
41 Res Toxicol 1:175-178.
42
43 Dekant, W; Vamvakas, S; Anders, MW. (1989) Bioactivation of nephrotoxic haloalkenes by glutathione
44 conjugation: formation of toxic and mutagenic intermediates by cysteine conjugate beta-lyase. Drug Metab Rev
45 20:43-83.
46
47 Dekant, W; Birner, G; Werner, M; et al. (1998) Glutathione conjugation of perchloroethene in subcellular fractions
48 from rodent and human liver and kidney. ChemBiol Interact 116:31-43.
49
50 Dmitrieva, NV (1967) [On the metabolism of tetrachloroethylene]. Gig Tr Prof Zabol 11:54-56.
51
52 Dobrev, ID; Andersen, ME; Yang, RSH. (2002) In silico toxicology: simulating interaction thresholds for human
53 exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-trichloroethane. Environ Health Perspect
54 110(10):1031-1039.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-64 DRAFT-DO NOT CITE OR QUOTE
-------
1 Doherty, AT; Ellard, S; Parry, EM; et al. (1996) An investigation into the activation and deactivation of chlorinated
2 hydrocarbons to genotoxins in metabolically competent human cells. Mutagenesis 11:247-274.
3
4 Dohn, DR; Anders, MW. (1982) Assay of cysteine conjugate beta-lyase activity with S-(2-benzothiazolyl)cysteine
5 as the substrate. Anal Biochem 120:379-386.
6
7 Droz, PO; Guillemin, MP. (1986) Occupational exposure monitoring using breath analysis. JOccupMed 28:593-
8 602.
9
10 Duffel, MW; Jakoby, WB. (1982) Cysteine S-conjugate N-acetyltransferase from rat kidney microsomes. Mol
11 Pharmacol 21:444-448.
12
13 Eger, EL (1963) A mathematical model of uptake and distribution. In: Popper, EM; Ketz, RJ; eds. Uptake and
14 distribution of anesthetic agents. New York: McGraw-Hill; pp. 72-103.
15 Elfarra, A.A.; Krause R.J. (2007) S-(l,2,2-trichlorovinyl)-L-cysteine sulfoxide, a reactive metabolite of S-(l,2,2-
16 Trichlorovinyl)-L-cysteine formed in rat liver and kidney microsomes, is a potent nephrotoxicant. J Pharmacol Exp
17 Ther 321(3): 1095-101.
18
19 Essing, HG; Schacke, G; Schaller, KH; et al. (1973) [Occupational-medicine studies on the dynamics of
20 perchloroethylene in the organism]. Med Welt 24:242-244.
21
22 Fernandez, J; Guberan, E; Caperos, J. (1976) Experimental human exposures to tetrachloroethylene vapor and
23 elimination in breath after inhalation. Am Ind Hyg Assoc 137:143-150.
24
25 Filser, JG; Bolt, HM. (1979) Pharmacokinetics of halogenated ethylenes in rats. Arch Toxicol 42:123-136.
26
27 Fisher, J; Lumpkin, M; Boyd, J; et al. (2004) PBPK modeling of the metabolic interactions of carbon tetrachloride
28 and tetrachloroethylene in B6C3F1 mice. Environ Toxicol Phar 16:93-105.
29
30
31 Frankel, DM; Johnson, CE; Pitt, HM. (1957) Preparation and properties of tetrachloroethylene oxide. J Org Chem
32 22:1119-1124.
33
34 Frantz, SW; Watanabe, PG. (1983) Tetrachloroethylene: balance and tissue distribution in male Sprague-Dawley
3 5 rats by drinking-water administration. Toxicol Appl Pharmacol 69:66-72.
36
37 Gamier, R; Bedouin, J; Pepin, G; et al. (1996) Coin-operated dry cleaning machines may be responsible for acute
38 tetrachloroethylene poisoning: report of 26 cases including one death. J Toxicol Clin Toxicol 34:191-197.
39
40 Gearhart, JM; Mahle, DA; Greene, RJ; et al. (1993) Variability of physiologically based pharmacokinetic (PBPK)
41 model parameters and their effects on PBPK model predictions in a risk assessment for perchloroethylene (PCE).
42 Toxicol Lett 68:131-144.
43
44 Gentry, PR; Covington, TR; Clewell, HJ, III. (2003) Evaluation of the potential impact of pharmacokinetic
45 differences on tissue dosimetry in offspring during pregnancy and lactation. Regul Toxicol Pharmacol 38(1): 1-16.
46
47 Ghantous, H; Danielsson, BR; Dencker, L; et al. (1986) Trichloroacetic acid accumulates in murine amniotic fluid
48 after tri- and tetrachloroethylene inhalation. Acta Pharmacol Toxicol (Copenh) 58:105-114.
49
50 Goldstein, A; Aronow, L; Kalman, SM. (1969) Principles of drug action. New York: Harper and Row.
51
52 Green, T; Odum, J; Nash, JA; et al. (1990) Perchloroethylene-induced rat kidney tumors: an investigation of the
53 mechanisms involved and their relevance to humans. Toxicol Appl Pharmacol 103:77-89.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-65 DRAFT-DO NOT CITE OR QUOTE
-------
1 Green, T; Dow, J; Ellis, MK; et al. (1997) The role of glutathione conjugation in the development of kidney tumours
2 in rats exposed to trichloroethylene. ChemBiol Interact 105:99-117.
3
4 Guberan, E; Fernandez, J. (1974) Control of industrial exposure to tetrachloroethylene by measuring alveolar
5 concentrations: theoretical approach using a mathematical model. Br J Ind Med 31:159-167.
6
7 Guengerich, FP; Macdonald, TL. (1984) Chemical mechanisms of catalysis by cytochrome P450: a unified view.
8 Ace ChemRes 17:9-16.
9
10 Hake, CL; Stewart, RD. (1977) Human exposure to tetrachloroethylene: inhalation and skin contact. Environ Health
11 Perspect 21:231-238.
12
13 Hattis, D; White, P; Marmorstein, L; et al. (1990) Uncertainties in pharmacokinetic modeling for perchloroethylene.
14 I. Comparison of model structure, parameters, and predictions for low-dose metabolism rates for models derived by
15 different authors. Risk Anal 10:449-458.
16
17 Henschler, D; Bonse, G. (1977) Metabolic activation of chlorinated ethylenes: dependence of mutagenic effect on
18 electrophilic reactivity of the metabolically formed epoxides. Arch Toxicol 39:7-12.
19
20 Hinchman, CA; Ballatori, N. (1990) Glutathione-degrading capacities of liver and kidney in different species.
21 BiochemPharmacol40:1131-1135.
22
23 Hu, Y; Hakkola, J; Oscarson, M; et al. (1999) Structural and functional characterization of the 5'-flanking region of
24 the rat and human cytochrome P450 2E1 genes: identification of a polymorphic repeat in the human gene. Biochem
25 Biophys Res Commun 263:286-293.
26
27 IARC (International Agency for Research on Cancer). (1995) Tetrachloroethylene. In: IARC Monographs on the
28 Evaluation of Carcinogenic Risks to Humans, Vol. 63: dry cleaning, some chlorinated solvents and other industrial
29 chemicals. Lyon, France; pp. 159-221.
30
31 ICRP (International Commission on Radiological Protection). (1975) Report of the Task Group on Reference Man.
32 Publication 23. Annals of the ICRP.
33
34 Ikeda, M. (1977) Metabolism of trichloroethylene and tetrachloroethylene in human subjects. Environ Health
35 Perspect 21:239-245.
36
37 Ikeda, M; Imanura, T. (1973) Biological half-life of trichloroethylene and tetrachloroethylene in human subjects. Int
3 8 Arch Arbeitsmed 31:209-224.
39
40 Ikeda, M; Otsuji, H. (1972) A comparative study of the excretion of Fujiwara reaction-positive substances in urine
41 of humans and rodents given trichloro- or tetrachloro-derivatives of ethane and ethylene. Br J Ind Med 29:99-104.
42
43 Ikeda, M; Otsuji, H; Imamura, T; et al. (1972) Urinary excretion of total trichloro-compounds, trichloroethanol, and
44 trichloroacetic acid as a measure of exposure to trichloroethylene and tetrachloroethylene. Br J Ind Med 29:328-
45 333.
46
47 Isukapalli, SS; Roy, A; Georgopoulos, PG. (1998) Stochastic response surface methods (SRSMs) for uncertainty
48 propagation: application to environmental and biological systems. Risk Anal 18:351-363.
49
50 Jakobson, I; Wahlberg, JE; Holmberg, B; et al. (1982) Uptake via the blood and elimination of 10 organic solvents
51 following epicutaneous exposure of anesthetized guinea pigs. Toxicol Appl Pharmacol 63:181-187.
52
53 Kaemmerer, K; Fink, J; Keitzmann, M. (1982) Studies on the pharmacodynamics of perchloroethylene. Der
54 Prakitsch Tierarzt 2:171-182.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-66 DRAFT-DO NOT CITE OR QUOTE
-------
1 Kline, SA; Solomon, JJ; VanDuuren, BL. (1978) Synthesis and reactions of chloroalkene epoxides. J Org Chem
2 42:3596-3600.
3
4 Koppel, C; Arndt, I; Arendt, U; et al. (1985) Acute tetrachloroethylene poisoning—blood elimination kinetics
5 during hyperventilation therapy. JToxicol ClinToxicol23:103-115.
6
7 Kundig, S; Hogger, D. (1970) The importance of the determination of tri- and perchloroethylene metabolites in the
8 urine. Int Arch Arbeitsmed 26:3 06-315.
9
10 Jakobson, I; Wahlberg, JE; Holmberg, B; et al. (1982) Uptake via the blood and elimination of 10 organic solvents
11 following epicutaneous exposure of anesthetized guinea pigs. Toxicol Appl Pharmacol 63:181-187.
12
13 Larsen, GL. (1985) Distribution of cysteine conjugate beta-lyase in gastrointestinal bacteria and in the environment.
14 Xenobiotica 15:199-209.
15
16 Larsen, GL; Stevens, JL. (1986) Cysteine conjugate beta-lyase in the gastrointestinal bacterium Eubacterium
17 limosum. Mol Pharmacol 29:97-103.
18
19 Larson, JL; Bull, RJ. (1989) Effect of ethanol on the metabolism of trichloroethylene. J Toxicol Environ Health
20 28:395-406.
21
22 Lash, LH; Parker, JC. (2001) Hepatic and renal toxicities associated with perchloroethylene. Pharmacol Rev
23 53:177-208.
24
25 Lash, LH; Nelson, RM; Van Dyke, RA; et al. (1990) Purification and characterization of human kidney cytosolic
26 cysteine conjugate beta-lyase activity. Drug Metab Dispos 18:50-54.
27
28 Lash, LH; Qian, W; Putt, DA; et al. (1998) Glutathione conjugation of perchloroethylene in rats and mice in vitro:
29 sex-, species-, and tissue-dependent differences. Toxicol Appl Pharmacol 150:49-57.
30
31 Lash, LH; Lipscomb, JC; Putt, DA; et al. (1999) Glutathione conjugation of trichloroethylene in human liver and
32 kidney: kinetics and individual variation. Drug Metab Dispos 27:351-359.
33
34 Lash, LH; Fisher, JW; Lipscomb, JC; et al. (2000) Metabolism of trichloroethylene. Environ Health Perspect 108
35 Suppl 2:177-200.
36
37 Lash, LH; Qian, W; Putt, DA; et al. (2001) Renal and hepatic toxicity of trichloroethylene and its glutathione-
3 8 derived metabolites in rats and mice: sex-, species-, and tissue- dependent differences. J Pharmacol Exp Ther
39 297:155-164.
40
41 Leibman, KC; Ortiz, E. (1970) Epoxide intermediates in microsomal oxidation of olefins to glycols. J Pharmacol
42 Exp Ther 713:242-246.
43
44 Leibman, KC; Oritz, E. (1977) Metabolism of halogenated ethylenes. Environ Health Perspect 21:91-97.
45
46 Levine, B; Fierro, MF; Goza, SW; et al. (1981) A tetrachloroethylene fatality. J Forensic Sci 26:206-209.
47
48 Lieber, CS. (1997) Cytochrome P-4502E1: its physiological and pathological role. Physiol Rev 77:517-544.
49
50 Lipscomb, JC; Mahle, DA; Brashear, WT; et al. (1995) Dichloroacetic acid: metabolism in cytosol. Drug Metab
51 Dispos 23:1202-1205.
52
53 Lipscomb, JC; Mahle, DA; Brashear, WT; et al. (1996) A species comparison of chloral hydrate metabolism in
54 blood and liver. Biochem Biophys Res Commun 227:340-350.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-67 DRAFT-DO NOT CITE OR QUOTE
-------
1 Lipscomb, JC; Fisher, JW; Confer, PD; et al. (1998) In vitro to in vivo extrapolation for trichloroethylene
2 metabolism in humans. Toxicol Appl Pharmacol 152:376-387.
3
4 Loizou, GD. (2001) The application of physiologically based pharmacokinetic modelling in the analysis of
5 occupational exposure to perchloroethylene. Toxicol Lett 124:59-69.
6
7 Lukaszewski, T. (1979) Acute tetrachloroethylene fatality. Clin Toxicol 15:411-415.
8
9 Malherbe, P; Alberati-Giani, D; Kohler, C; et al. (1995) Identification of a mitochondria! form of kynurenine
10 aminotransferase/glutamine transaminase K from rat brain. FEES Lett 367:141-144.
11
12 Mannervik, B. (1985) The isoenzymes of glutathione transferase. Adv Enzymol Relat Areas Mol Biol 57:357-417.
13
14 Maronpot, RR; Fox, T; Malarkey, DE; et al. (1995) Mutations in the ras proto-oncogene: clues to etiology and
15 molecular pathogenesis of mouse liver tumors. Toxicology 101:125-156.
16
17 May, R. (1976) Metabolisierung und Ausscheidung von Tetrachlorathylen und Beeinflussung durch Gleichzeitige
18 Aufnahme von Athanol beim Menschen. Thesis, University of Wurzburg, Germany.
19
20 McCarver, DG; Byun, R; Hines, RN; et al. (1998) A genetic polymorphism in the regulatory sequences of human
21 CYP2E1: association with increased chlorzoxazone hydroxylation in the presence of obesity and ethanol intake.
22 Toxicol Appl Pharmacol 152:276-281.
23
24 McDougal, JN; Jepson, GW; Clewell, HJ, III; et al. (1990) Dermal absorption of organic chemical vapors in rats and
25 humans. Fundam Appl Toxicol 14:299-308.
26
27 Miller, RE; Guengerich, FP. (1982) Oxidation of trichloroethylene by liver microsomal cytochrome P-450: evidence
28 for chlorine migration in a transition state not involving trichloroethylene oxide. Biochemistry 21:1090-1097.
29
30 Miller, RE; Guengerich, FP. (1983) Metabolism of trichloroethylene in isolated hepatocytes, microsomes, and
31 reconstituted enzyme systems containing cytochrome P-450. Cancer Res 43:1145-1152.
32
33 Mitchell, AE; Morin, D; Lakritz, J; et al. (1997) Quantitative profiling of tissue- and gender-related expression of
34 glutathione S-transferase isoenzymes in the mouse. Biochem J 325(Pt 1):207-216.
35
36 Moghaddam, AP; Abbas, R; Fisher, JW; etal. (1996) Formation of dichloroacetic acid by rat and mouse gut
37 microflora, an in vitro study. Biochem Biophys Res Commun 228:639-645.
38
39 Moghaddam, AP; Abbas, R; Fisher, JW; et al. (1997) The role of mouse intestinal microflora in the metabolism of
40 trichloroethylene, an in vivo study. Hum Exp Toxicol 16:629-635.
41
42 Monster, AC; Houtkooper, JM. (1979) Estimation of individual uptake of trichloroethylene, 1,1,1- trichloroethane
43 and tetrachloroethylene from biological parameters. Int Arch Occup Environ Health 42:319-323.
44
45 Monster, AC; Smolders, JF. (1984) Tetrachloroethene in exhaled air of persons living near pollution sources. Int
46 Arch Occup Environ Health 53:331-336.
47
48 Monster, AC; Boersma, G; Steenweg, H. (1979) Kinetics of tetrachloroethylene in volunteers: influence of exposure
49 concentration and work load. Int Arch Occup Environ Health 42:303-309.
50
51 Monster, AC; Regouin-Peeters, W; van Schijndel, A; et al. (1983) Biological monitoring of occupational exposure
52 to tetrachloroethene. Scand J Work Environ Health 9:273-281.
53
54 Moslen, MT; Reynolds, ES; Szabo, S. (1977) Enhancement of the metabolism and hepatotoxicity of
55 trichloroethylene and perchloroethylene. Biochem Pharmacol 26:369-375.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-68 DRAFT-DO NOT CITE OR QUOTE
-------
1 Muller, G; Spassovski, M; Henschler, D. (1974) Metabolism of trichloroethylene in man: II. pharmacokinetics of
2 metabolites. Arch Toxicol 32:283-95.
3
4 Nakai, JS; Stathopulos, PB; Campbell, GL; et al. (1999) Penetration of chloroform, trichloroethylene, and
5 tetrachloroethylene through human skin. J Toxicol Environ Health A 58:157-170.
6
7 Ni, YC; Wong, TY; Lloyd, RV; et al. (1996) Mouse liver microsomal metabolism of chloral hydrate, trichloroacetic
8 acid, and trichloroethanol leading to induction of lipid peroxidation via a free radical mechanism. Drug Metab
9 Dispos 24:81-90.
10
11 NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
12 (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3F1 mice. NTP Technical Report 311.
13
14 NYS DOH (New York State Department of Health). (2000) Evaluation of residential exposure to tetrachloroethene
15 using biomarkers of dose and neurological tests (non peer-reviewed draft). Albany, NY.
16
17 Ogata, M; Sugiyama, K; Kuroda, YI. (1962) Investigation of a dry-cleaning shop using tetrachloroethylene, with
18 special reference to Fujiwara's substance in the urine of the employees. Okayama Igakkai Zasshi 74:247-253.
19
20 Ogata, M; Takatsuka, Y; Tomokuni, K. (1971) Excretion of organic chlorine compounds in the urine of persons
21 exposed to vapours of trichloroethylene and tetrachloroethylene. Br J Ind Med 28:386-391.
22
23 Ohtsuki, T; Sato, K; Koizumi, A; et al. (1983) Limited capacity of humans to metabolize tetrachloroethylene. Int
24 Arch Occup Environ Health 51:381-390.
25
26 Opdam, JJ; Smolders, JF. (1986) Alveolar sampling and fast kinetics of tetrachloroethene in man. I. Alveolar
27 sampling. BrJ Ind Med 43:814-824.
28
29 Otieno, MA; Baggs, RB; Hayes, JD; et al. (1997) Immunolocalization of microsomal glutathione S-transferase in rat
30 tissues. Drug Metab Dispos 25:12-20.
31
32 Overby, LH; Gardlik, S; Philpot, RM; et al. (1994) Unique distribution profiles of glutathione S-transferases in
33 regions of kidney, ureter, and bladder of rabbit. Lab Invest 70:468-478.
34
3 5 Pahler, A; Birner, G; Parker, J; et al. (1998) Generation of antibodies to di- and trichloroacetylated proteins and
3 6 immunochemical detection of protein adducts in rats treated with perchloroethene. Chem Res Toxicol 11:995-1004.
37
3 8 Pahler, A; Volkel, W; Dekant, W. (1999a) Quantitation of N epsilon-(dichloroacetyl)-L-lysine in proteins after
39 perchloroethene exposure by gas chromatography-mass spectrometry using chemical ionization and negative ion
40 detection following immunoaffinity chromatography. J Chromatogr A 847:25-34.
41
42 Pahler, A; Parker, J; Dekant, W. (1999b) Dose-dependent protein adduct formation in kidney, liver, and blood of
43 rats and in human blood after perchloroethene inhalation. Toxicol Sci 48:5-13.
44
45 Pastino, GM; Yap, WY; Carroquino, M. (2000) Human variability and susceptibility to trichloroethylene [In Process
46 Citation]. Environ Health Perspect 108(suppl2):201-214.
47
48 Pegg, DG; Zempel, JA; Braun, WH; et al. (1979) Disposition of tetrachloro(14C)ethylene following oral and
49 inhalation exposure in rats. Toxicol Appl Pharmacol 51:465-474.
50
51 Pereira, MA. (1996) Carcinogenic activity of dichloroacetic acid and trichloroacetic acid in the liver of female
52 B6C3F1 mice. Fundam Appl Toxicol 31:192-199.
53
54 Pezzagno, G; Imbriani, M; Ghittori, S; et al. (1988) Urinary concentration, environmental concentration, and
55 respiratory uptake of some solvents: effect of the work load. Am Ind Hyg Assoc J 49:546-552.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-69 DRAFT-DO NOT CITE OR QUOTE
-------
1 Poet, TS.; Weitz, KK; Gies, RA; et al. (2002) PBPK modeling of the percutaneous absorption of perchloroethylene
2 from a soil matrix in rats and humans. Toxicol Sci 67: 17-31.
3
4 Powell, JF. (1945) Trichloroethylene: absorption, elimination and metabolism. Br J Ind Med 2:142-145.
5
6 Quondamatteo, F; Schulz, TG; Bunzel, N; et al. (1998) Immunohistochemical localization of glutathione S-
7 transferase-Tl in murine kidney, liver, and lung. Histochem Cell Biol 110:417-423.
8
9 Ramsey, JC; Andersen ME. (1984). A physiologically based description of the inhalation pharmacokinetics of
10 styrene in rats and humans. Toxicol Appl Pharmacol 73:159-175.
11
12 Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
13 groundwater for a bathing and showering determination. Risk Anal 13:37^19.
14
15 Raucy, JL. (1995) Risk assessment: toxicity from chemical exposure resulting from enhanced expression of
16 CYP2E1. Toxicology 105:217-224.
17
18 Raunio, H; HusgafVel-Pursiainen, K; Anttila, S; et al. (1995) Diagnosis of polymorphisms in carcinogen-activating
19 and inactivating enzymes and cancer susceptibility-a review. Gene 159:113-121.
20
21 Reitz, RH. (1992) New data on tetrachloroethylene metabolism. Office of Environmental Health Hazard
22 Assessment Advisory Panel Meeting. California Environmental Protection Agency. Berkeley.
23
24 Reitz, RH; Nolan, RJ. (1986) Physiological pharmacokinetic modeling for perchloroethylene dose adjustment. Draft
25 Report to EPA's Science Advisory Board, pp. 1-23. Available from: IRIS desk, U.S. Environmental Protection
26 Agency, Washington, DC.
27
28 Reitz, RH; Mendrala, AL; Guengerich, FP. (1989) In vitro metabolism of methylene chloride in human and animal
29 tissues: use in physiologically based pharmacokinetic models. Toxicol Appl Pharmacol 97:230^16.
30
31 Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
32 for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
33 306.
34
3 5 Rhomberg, L. (1992) A cross-species scaling factor for carcinogen risk assessment based on equivalence of
36 mg/kg3/4/day. Federal Register 57:4152-24173.
37
38 Riihimaki, V; Pfaffli, P. (1978) Percutaneous absorption of solvent vapors in man. Scand J Work Environ Health
39 4:73-85.
40
41 Ripp, SL; Overby, LH; Philpot, RM; et al. (1997) Oxidation of cysteine S-conjugates by rabbit liver microsomes and
42 cDNA- expressed flavin-containing mono-oxygenases: studies with S-(l,2- dichlorovinyl)-L-cysteine, S-(l,2,2-
43 trichlorovinyl)-L-cysteine, S-allyl- L-cysteine, and S-benzyl-L-cysteine. Mol Pharmacol 51:507-515.
44
45 Rodilla, V; Benzie, AA; Veitch, JM; et al. (1998) Glutathione S-transferases in human renal cortex and neoplastic
46 tissue: enzymatic activity, isoenzyme profile and immunohistochemical localization. Xenobiotica 28:443-456.
47
48 Sakamoto, N. (1976) [Metabolism of tetrachloroethylene in guinea pigs (author'stransl)]. Sangyo Igaku 18:11-16.
49
50 Sata, F; Sapone, A; Elizondo, G; et al. (2000) CYP3A4 allelic variants with amino acid substitutions in exons 7 and
51 12: evidence for an allelic variant with altered catalytic activity. Clin Pharmacol Ther 67:48-56.
52
53 Sausen, PJ; Elfarra, AA. (1990) Cysteine conjugate S-oxidase. Characterization of a novel enzymatic activity in rat
54 hepatic and renal microsomes. J Biol Chem 265:6139-6145.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-70 DRAFT-DO NOT CITE OR QUOTE
-------
1 Savolainen, H; Pfaffli, P; Tengen, M; et al. (1977) Biochemical and behavioural effects of inhalation exposure to
2 tetrachloroethylene and dichloromethane. J Neuropathol Exp Neurol 36:941-949.
3
4 Schreiber, JS. (1993) Predicted infant exposure to tetrachloroethylene in human breast milk. Risk Anal 13:515-
5 524.Schreiber, JS; House, S; Prohonic, E; et al. (1993) An investigation of indoor air contamination in residences
6 above dry cleaners. Risk Anal 13:355-344.
7
8 Schreiber, J. (1997) Transport of organic chemicals to breast milk: tetrachloroethene case study. In: Kacew, S;
9 Lambert, GH; eds. Environmental toxicology and pharmacology of human development. Washington, DC: Taylor
10 and Francis.
11
12 Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
13 tetrachloroethylene (perc) and deficits in visual contrast sensitivity. Environ Health Perspect 110:655-664.
14
15 Schultz, J; Weiner, H. (1979) Alteration of the enzymology of chloral hydrate reduction in the presence of ethanol.
16 Biochem Pharmacol 28:3379-3384.
17
18 Schumann, AM; Quast, JF; Watanabe, PG. (1980) The pharmacokinetics and macromolecular interactions of
19 perchloroethylene in mice and rats as related to oncogenicity. Toxicol Appl Pharmacol 55:207-219.
20
21 Sellers, EM; Lang, M; Koch-Weser, J; et al. (1972) Interaction of chloral hydrate and ethanol in man. I.
22 Metabolism. Clin Pharmacol Ther 13:37-49.
23
24 Sheldon, L; Handy, R; Hartwell, W; et al. (1985) Human exposure assessment to environmental chemicals: nursing
25 mothers study. Final report. Research Triangle Institute, Research Triangle Park, NC.
26
27 Skender, LJ; Karacic, V; Prpic-Majic, D. (1991) A comparative study of human levels of trichloroethylene and
28 tetrachloroethylene after occupational exposure. Arch Environ Health 46:174-178.
29
30 Speerschneider, P; Dekant, W. (1995) Renal tumorigenicity of 1,1-dichloroethene in mice: the role of male- specific
31 expression of cytochrome P450 2E1 in the renal bioactivation of 1,1-dichloroethene. Toxicol Appl Pharmacol
32 130:48-56.
33
34 Stephens, EA; Taylor, JA; Kaplan, N; et al. (1994) Ethnic variation in the CYP2E1 gene: polymorphism analysis of
35 695 African-Americans, European-Americans and Taiwanese. Pharmacogenetics 4:185-192.
36
3 7 Stevens, JL. (1985) Isolation and characterization of a rat liver enzyme with both cysteine conjugate beta-lyase and
38 kynureninase activity. J Biol Chem 260:7945-7950.
39
40 Stevens, J; Jakoby, WB. (1983) Cysteine conjugate beta-lyase. Mol Pharmacol 23:761-765.
41
42 Stewart, RD; Dodd, HC. (1964) Absorption of carbon tetrachloride, trichloroethylene, tetrachloroethylene,
43 methylene chloride and 1,1,1-trichloroethane through the human skin. J Ind Hyg 25:429-446.
44
45 Stewart, RD; Erley, DS; Schaeffer, AW; et al. (1961) Accidental vapor exposure to anesthetic concentrations of a
46 solvent containing tetrachloroethylene. Indus Med Sug 30:327-330.
47
48 Stewart, RD; Baretta, ED; Dodd, HC; et al. (1970) Experimental human exposure to tetrachloroethylene. Arch
49 Environ Health 20:225-229.
50
51 Stewart, RD; Hake, CL; Forster AJ; et al. (1974) Tetrachloroethylene: development of a biologic standard for the
52 industrial worker by breath analysis. National Institute of Occupational Safety and Health. Cincinnati, OH.
53 NIOSH-MCOW-ENVM-PCE-74-6.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-71 DRAFT-DO NOT CITE OR QUOTE
-------
1 Stewart, RD; Hake, CL; Wu, A. (1977) Effects of perchloroethylene drag interaction on behavior and neurological
2 function. DHEW (NIOSH) Publ. No. 77-191. Department of Health, Education, And Welfare, Washington, DC.
3
4 Stott, WT; Quast, JF; Watanabe, PG. (1982) The pharmacokinetics and macromolecular interactions of
5 trichloroethylene in mice and rats. Toxicol Appl Pharmacol 62:137-151.
6
7 Tanaka, S; Ikeda, M. (1968) A method for determination of trichloroethanol and trichloroacetic acid in urine. Br J
8 IndMed 25:214-219.
9
10 Tateishi, M; Suzuki, S; Shimizu, H. (1978) Cysteine conjugate beta-lyase in rat liver. A novel enzyme catalyzing
11 formation of thiol-containing metabolites of drags. J Biol Chem 253:8854-8859.
12
13 Terrier, P; Townsend, AJ; Coindre, JM; et al. (1990) An immunohistochemical study of pi class glutathione S-
14 transferase expression in normal human tissue. Am JPathol 137:845-853.
15
16 Tomisawa, H; Suzuki, S; Ichihara, S; et al. (1984) Purification and characterization of C-S lyase from
17 Fusobacterium varium. A C-S cleavage enzyme of cysteine conjugates and some S- containing amino acids. J Biol
18 Chem 259:2588-2593.
19
20 Tomisawa, H; Ichihara, S; Fukazawa, H; et al. (1986) Purification and characterization of human hepatic cysteine-
21 conjugate beta-lyase. Biochem 1235:569-575.
22
23 Tong, Z; Board, PG; Anders, MW. (1998a) Glutathione transferase zeta-catalyzed biotransformation of
24 dichloroacetic acid and other alpha-haloacids. Chem Res Toxicol 11:1332-1338.
25
26 Tong, Z; Board, PG; Anders, MW. (1998b) Glutathione transferase zeta catalyses the oxygenation of the carcinogen
27 dichloroacetic acid to glyoxylic acid. Biochem J 33 l(Pt 2):371-374.
28
29 Tsurata, H. (1989) Skin absorption of organic solvent vapors in nude mice in vivo. Ind Health 27:37-47.
30
31 Tzeng, HF; Blackburn, AC; Board, PG; et al. (2000) Polymorphism- and species-dependent inactivation of
32 glutathione transferase zeta by dichloroacetate. Chem Res Toxicol 13:231-236.
33
34 U.S. EPA (Environmental Protection Agency). (1985) Health assessment document for tetrachloroethylene
3 5 (perchloroethylene). Office of Health and Environmental Assessment, Office of Research and Development,
36 Washington, DC; EPA/600/8-82/005F. Available from: National Technical Information Service, Springfield, VA;
37 PB-86-174489/AS.
38
39 U.S. EPA (Environmental Protection Agency). (1986) Addendum to the health assessment document for
40 tetrachloroethylene (perchloroethylene) [review draft]. Office of Health and Environmental Assessment, Office of
41 Research and Development, Washington, DC; EPA/600/8-82/005FA. Available from: National Technical
42 Information Service, Springfield, VA; PB-86-174489/AS.
43
44 U.S. EPA (Environmental Protection Agency). (1989) Risk assessment guidance for Superfund (RAGS), vol. I:
45 human health evaluation manual, part A (1989). Office of Emergency and Remedial Response, Washington, DC;
46 EPA/540/1-89/002. Available online at http://www.epa.gov/superfund/programs/risk/ragsa/index.htm.
47
48 U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
49 carcinogenicity of tetrachloroethylene (perchloroethylene). Office of Health and Environmental Assessment,
50 Washington, DC; EPA/600/6-91/002F. Available from: National Technical Information Service, Springfield, VA.
51
52 U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
53 and application of inhalation dosimetry. Office of Health and Environmental Assessment, Environmental Criteria
54 and Assessment Office, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical Information
55 Service, Springfield, VA, PB2000-500023, and online at http://www.epa.gov/ncea.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-72 DRAFT-DO NOT CITE OR QUOTE
-------
1 U.S. EPA (Environmental Protection Agency). (1998) Dichloroacetic acid: carcinogenicity identification
2 characterization summary. National Center for Environmental Assessment, Office of Research and Development,
3 Washington, DC; EPA NCEA-W-0372.
4
5 U.S. EPA (Environmental Protection Agency). (2001) Trichloroethylene health risk assessment: synthesis and
6 characterization. National Center for Environmental Assessment, Office of Research and Development,
7 Washington, DC; EPA/600/P-01/002A. Available online at
8 http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=4580.
9
10 Uttamsingh, V; Keller, DA; Anders, MW. (1998) Acylase I-catalyzed deacetylation of N-acetyl-L-cysteine and S-
11 alkyl— acetyl-L-cysteines. Chem Res Toxicol 11:800-809.
12
13 Vainio, H; Parkki, MG; Mamiemi, J. (1976) Effects of aliphatic chlorohydrocarbons on drug-metabolizing enzymes
14 in rat liver in vivo. Xenobiotica 6:599-604.
15
16 Vamvakas, S; Dekant, W; Berthold, K; et al. (1987) Enzymatic transformation of mercapturic acids derived from
17 halogenated alkenes to reactive and mutagenic intermediates. Biochem Pharmacol 36:2741-2748.
18
19 Vamvakas, S; Dekant, W; Henschler, D. (1989a) Genotoxicity of haloalkene and haloalkane glutathione S-
20 conjugates in porcine kidney cells. Toxicol In Vitro 3:151-156.
21
22 Vamvakas, S; Herkenhoff, M; Dekant, W; et al. (1989b) Mutagenicity of tetrachloroethene in the Ames test-
23 metabolic activation by conjugation with glutathione. J Biochem Toxicol 4:21-27.
24
25 Vamvakas, S; Dekant, W; Henschler, D. (1989c) Assessment of unscheduled DNA synthesis in a cultured line of
26 renal epithelial cells exposed to cysteine S-conjugates of haloalkenes and haloalkanes. Mutat Res 222:329-335.
27
28 Veitch, JM; Murray, GI; Juronen, E; et al. (1997) Theta-class glutathione S-transferases in human kidney and renal
29 tumours. Biochem Soc Trans 25:8605.
30
31 Vieira, I; Sonnier, M; Cresteil, T. (1996) Developmental expression of CYP2E1 in the human liver:
32 hypermethylation control of gene expression during the neonatal period. Eur J Biochem 238(2):476-83.
33
34 Volkel, W; Friedewald, M; Lederer, E; et al. (1998) Biotransformation of perchloroethene: dose-dependent
3 5 excretion of trichloroacetic acid, dichloroacetic acid, and N-acetyl-S- (trichlorovinyl)-L-cysteine in rats and humans
36 after inhalation. Toxicol Appl Pharmacol 153:20-27.
37
38 Volkel, W; Pahler, A; Dekant, W. (1999) Gas chromatography-negative ion chemical ionization mass spectrometry
39 as a powerful tool for the detection of mercapturic acids and DNA and protein adducts as biomarkers of exposure to
40 halogenated olefins. JChromatogr A 847:35-46.
41
42 Ward, RC; Travis, CC; Hetrick, DM; et al. (1988) Pharmacokinetics of tetrachloroethylene. Toxicol Appl
43 Pharmacol 93:108-117.
44
45 Waters, EM; Gerstner, HB; Huff, JE. (1977) Trichloroethylene. I. an overview. J Toxicol Environ Health 2:671-
46 707.
47
48 Weichardt, H; Lindner, J. (1975) Gesundheitsgefahren durch Perchlorathlen in Chemisch-Reinigungsbetrieben aus
49 arbiet medizinischtoxikologischer Sicht. Staub Reinhalt Luft 35:416^4-20.
50
51 Weiss, G. (1969) [Observation of the course of trichloroacetic acid excretion in occupational tetrachloroethylene
52 poisoning]. Zentralbl Arbeitsmed 19:143-146.
53
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-73 DRAFT-DO NOT CITE OR QUOTE
-------
1 Werner, M; Birner, G; Dekant, W. (1996) Sulfoxidation of mercapturic acids derived from tri- and tetrachloroethene
2 by cytochromes P450 3 A: a bioactivation reaction in addition to deacetylation and cysteine conjugate beta-lyase
3 mediated cleavage. Chem Res Toxicol 9:41-49.
4
5 Westlind, A; Lofberg, L; Tindberg, N; et al. (1999) Interindividual differences in hepatic expression of CYP3 A4:
6 relationship to genetic polymorphism in the 5'-upstream regulatory region. Biochem Biophys Res Commun
7 259:201-205.
8
9 Yllner, S. (1961) Urinary metabolites of 14-C-tetrachloroethylene in mice. Nature 191:820-821.
10
11 Yoo, JS; Guengerich, FP; Yang, CS. (1988) Metabolism of N-nitrosodialkylamines by human liver microsomes.
12 Cancer Res 48:1499-1504.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 3-74 DRAFT-DO NOT CITE OR QUOTE
-------
1 4. HAZARD IDENTIFICATION
2
O
4 4.1. OVERALL APPROACH
5 This chapter discusses tetrachloroethylene toxicity on an organ-specific basis, with liver,
6 kidney, neurotoxicity, and developmental/reproductive effects as the major emphasis in separate
7 sections. For each of the major organ systems, human effects are presented first, followed by
8 effects in animals and in in vitro systems. Cancer and noncancer toxicity and mode of action
9 (MOA) are also included in the discussions. Of note, site concordance of effect between animals
10 and humans is generally not assumed.
11 Evidence for each organ system is summarized, but an in-depth discussion or data
12 evaluation is not provided for any individual studies, especially those evaluated and discussed in
13 previous EPA documents and other Agency reports. Several existing publications provide more
14 detailed study descriptions: the World Health Organization-International Programme on
15 Chemical Safety (WHO-IPCS, 2006), the ATSDR' s Toxicologic Profile for Tetrachloroethylene
16 (ATSDR, 1997) and the International Agency for Research on Cancer (IARC) review the health
17 effect evidence on tetrachloroethylene, trichloroethylene, and their common metabolites; other
18 studies review this evidence on dry cleaner as an occupational title (IARC, 1995);
19 Tetrachloroethene Ambient Air Criteria Document (NYS DOH, 1997); and Public Health Goal
20 for Tetrachloroethylene in Drinking Water (Cal EPA, 2001). The details for earlier toxicity and
21 carcinogenicity studies may also be found in previous EPA assessments (e.g., U.S. EPA, 1980,
22 1985a, b, 1986a, 1991a).
23
24 4.2. OVERVIEW OF TETRACHLOROETHYLENE METABOLISM
25 Most tetrachloroethylene toxicity and cancer-causing activity, other than neurotoxicity, is
26 generally attributed to its metabolites. For example, historically, a direct relationship has been
27 demonstrated between the level of hepatic microsomal cytochrome P450, the extent of
28 metabolism of tetrachloroethylene in vivo, and cellular damage (Bonse et al., 1975; Bonse and
29 Henschler, 1976; Moslen et al., 1977; Pegg et al., 1979; Schumann et al., 1980; Buben and
30 O'Flaherty, 1985; U.S. EPA, 1985a, 1991a). In addition, several oxidative (P450) and GSH-
31 derived tetrachloroethylene metabolites have been shown to induce toxic and carcinogenic
32 effects in similar targets when they are administered directly (IARC, 1995; Herren-Freund et al.,
33 1987; Bull et al., 1990; Bull, 2000; Pereira, 1996; DeAngelo et al., 1991, 1999; Daniel, 1963;
34 Carter et al., 2003; Elfarra and Krause, 2007). This metabolism prerequisite for certain toxic
35 effects is true for related halogenated ethylenes and ethanes as well (IARC, 1995; U.S. EPA,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-1 DRAFT-DO NOT CITE OR QUOTE
-------
1 1991a, 2001a; Dekant, 2001). CNS effects are a notable exception in that they are largely
2 attributable to the parent compound.
3 As detailed in Chapter 3, tetrachloroethylene is metabolized through at least two major
4 pathways. The oxidative P450, or CYP, pathway is quantitatively most important, and it
5 accounts for the greatest amount of observed metabolite in all species at all doses tested (see
6 Section 3.3.3.1). The other is the GSH conjugation pathway, which is associated with renal
7 toxicity and renal carcinogenicity (see Section 3.3.3.2). A significant portion of absorbed dose in
8 human studies (as much as 20-40%) cannot be tracked either as parent compound or metabolites
9 of known pathways, which introduces uncertainty about the identity and the amounts of the
10 metabolites formed in humans (U.S. EPA, 1991 a; Bogen and McKone, 1988). TCA, a product
11 of the oxidative pathway, is the major urinary metabolite derived from tetrachloroethylene
12 metabolism. TCA is a mouse liver carcinogen. TCA may contribute in part to findings of liver
13 toxicity and cancer observed in tetrachloroethylene-exposed animals; however, according to
14 some investigators (Clewell et al., 2004, 2005), the amount of TCA produced from
15 tetrachloroethylene in rodent bioassays is insufficient to account in total for observed
16 hepatocarcinogenicity (see Appendix 4A). DC A is another known tetrachloroethylene urinary
17 metabolite that is formed in both the oxidative pathway by dechlorination of TCA and, in organs
18 other than the liver, in the GSH pathway. DC A is known to cause liver cancer in both rats and
19 mice. Whether DC A contributes to tetrachloroethylene-induced toxicity or carcinogenicity in the
20 liver is not known.
21 Tetrachloroethylene oxide, trichloroacetyl chloride, and chloral/chloral hydrate are
22 proposed reactive intermediates in tetrachloroethylene P450 oxidation. Tetrachloroethylene
23 oxide and trichloroacetyl chloride have the potential to contribute to tetrachloroethylene toxicity
24 and carcinogenicity, particularly in the liver. Detection of TCOH in the urine of
25 tetrachloroethylene-exposed humans and animals would provide evidence for the existence of
26 the chloral hydrate intermediate. However, TCOH—and therefore the evidence that its
27 chloral/chloral hydrate precursor is formed from tetrachloroethylene—is not consistently
28 detected and it might be an artifact of the methodology used in some but not all studies. Chloral
29 hydrate is a liver carcinogen in mice.
30 The glutathione pathway entails initial glutathione-S-transferase catalyzed conjugation of
31 tetrachloroethylene with GSH to form TCVG. The cellular damage and genotoxic effects of
32 these conjugation products are thought to be from their further metabolism via beta-lyase, FMO3,
33 and/or P450 metabolism to highly reactive toxic products.
34 All of these metabolites have effects that may contribute to the toxicity and
35 carcinogenicity of tetrachloroethylene, although the role of specific intermediates has not been
36 elucidated. Genotoxicity of the oxidative metabolites TCA, DC A, chloral hydrate,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-2 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene oxide and the GSH-derived intermediates TCVC, TCVG, and NAcTCVC is
2 discussed in Section 4.3; for TCA and DCA, see Section 4.4.4.3 for peroxisome proliferator-
3 activated receptor alpha (PPAR-a) form activation and Section 4.4.4.4 for hypomethylation.
4 Section 4.10.3 provides a summary of the cancer MOA conclusions for tetrachloroethylene.
5
6 4.3. GENOTOXICITY
7 Tetrachloroethylene has been extensively studied for genotoxic activity in a variety of in
8 vitro assay systems such as bacteria, yeast, and mammalian cells (see reviews by U.S. EPA,
9 1985c, 1991a; IARC, 1995; ATSDR, 1997). Also, a review of the mutagenicity of
10 trichloroethylene (Moore and Harrington-Brock, 2000) contains a discussion of several of known
11 (TCA, DCA) and proposed (chloral hydrate) tetrachloroethylene metabolites.
12 The application of mutagenicity data to the question of potential carcinogenicity is based
13 on the premise that genetic alterations are found in all cancers. Mutagenesis is the ability of
14 chemicals to alter the genetic material in a manner that permits changes to be transmitted during
15 cell division. Although most tests for mutagenicity detect changes in DNA or chromosomes,
16 modifications of the epigenome, including proteins associated with DNA or RNA, can also cause
17 transmissible changes. Genetic alterations can occur via a variety of mechanisms including gene
18 mutations, deletions, translocations or amplification; evidence of mutagenesis provides
19 mechanistic support for the inference of potential for carcinogenicity in humans.
20 The following discussion focuses on the conclusions of the earlier studies and includes
21 details of recent studies that may provide some insight into the potential genotoxicity of
22 tetrachloroethylene. Positive findings were reported in some experiments using technical-grade
23 tetrachloroethylene that contained impurities or used epichlorohydrin or epoxybutane as
24 stabilizers, both of which are clearly mutagenic in a number of biological systems. Purified
25 tetrachloroethylene was negative in the same systems tested without or with mixed-function
26 oxidation activity provided by either rat or hamster liver S9 (Haworth et al., 1983). The results
27 of a large number of in vitro genotoxicity tests in which tetrachloroethylene was the test agent do
28 not clearly support the conclusion that tetrachloroethylene exhibits direct mutagenic activity,
29 although the few studies of conditions that would generate the GSH conjugate were positive (U. S.
30 EPA, 1991a; IARC, 1995; ATSDR, 1997).
31 An increased level of DNA single-strand breaks (SSB) was seen in liver and kidney
32 tissues but not in the lung tissue of mice 1 hr after single intraperitoneal (i.p.) injections of 4-8
33 mmol/kg (663-1326 mg/kg) of tetrachloroethylene (Walles, 1986). Potter et al. (1996) found no
34 increases in DNA strand breaks in kidneys of male F344 rats after a single gavage treatment with
35 1,000 mg/kg tetrachloroethylene. However, differences in species and/or route of exposure
36 preclude direct comparisons of these studies. Muzzullo (1987) found DNA binding of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-3 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene in mouse liver and rat kidney. Cytosols from several organs were more
2 effective than liver microsomes in enhancing in vitro DNA or protein binding of
3 tetrachloroethylene, and enrichment with GSH enhanced the activity of liver microsomes
4 (Muzzullo, 1987).
5 Toraason et al. (1999) found no increase in 8-hydroxydeoxyguanosine (8-OHdG) in the
6 urine, the liver, or the kidneys of male F344 rats after a single i.p. injection of
7 tetrachloroethylene at 100, 500, or 1,000 mg/kg (8-OHdG in peripheral lymphocytes was
8 measured only in the 500 mg/kg group). In a subsequent paper, Toraason et al. (2003) reported
9 no increase in 8-OHdG in urine of 18 dry cleaner workers sampled pre- and post-shift work
10 (time-weighted average [TWA] concentration of tetrachloroethylene was 3.8 + 5.3 ppm).
11 Tetrachloroethylene induced damage was observed in the sister chromatid exchange
12 (SCE) assay and in the single-cell gel test in human blood culture treated with up to 5 mM (-830
13 rng/L) tetrachloroethylene, at which viability was reduced by 40% (Hartmann and Speit, 1995).
14 Tetrachloroethylene exposure increased the frequency of micronuclei in peripheral blood
15 reticulocytes or hepatocytes of ddY mice given single i.p. injections at 1,000 or 2,000 mg/kg
16 tetrachloroethylene given after, but not prior to partial hepatectomy (Murakami and Horikawa,
17 1995). Tetrachloroethylene-induced micronuclei have also been reported in cultured Chinese
18 hamster kidney cells (Wang et al., 2001) and in human cells (Doherty et al., 1996; White et al.,
19 2001). Micronucleus induction was enhanced by tetrachloroethylene exposure in human
20 lymphoblastoid cells by stable expression of cDNAs encoding either CYP2E1 (hEl cells) or
21 human CYP1A2, 2A6, 3A4, 2E1 and microsomal epoxide hydrolase (Doherty et al., 1996). In
22 contrast to these findings, neither chromosome aberrations nor SCE were induced in Chinese
23 hamster ovary cells following in vitro exposure to tetrachloroethylene (Galloway et al., 1987).
24 Tetrachloroethylene when incubated with rat liver GST, GSH, and a rat kidney fraction,
25 exhibited a clear dose-response in the Ames test (Vamvakas et al., 1989b). In addition, it was
26 demonstrated that TCVG was produced from tetrachloroethylene in isolated perfused rat liver
27 and excreted into bile; in the presence of a rat kidney fraction, the collected bile was mutagenic
28 in Salmonella, as was purified TCVG (Vamvakas et al., 1989b). Dreesen (2003) also
29 demonstrated, for TCVG, an unequivocal dose-dependent mutagenic response in the TA 100
30 strain in the presence of the rat kidney S9-protein fraction; TCVC was mutagenic without
31 metabolic activation in this strain. In a separate study, the tetrachloroethylene metabolite TCVC
32 was also positive in Salmonella (strains TA 98 and TA 100) and inhibition of beta lyase activity
33 blocked the effect (Dekant et al., 1986). A subsequent study indicated that Salmonella also were
34 capable of deacetylating the urinary metabolite NAcTCVC when TA 100 showed a clear positive
35 response without exogenous activation (Vamvakas et al., 1987). Vamvakas et al. (1989a) also
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-4 DRAFT-DO NOT CITE OR QUOTE
-------
1 reported concentration-related increases in unscheduled DNA synthesis (UDS) in LLC-PK1 (a
2 porcine kidney cell line) exposed to TCVC, with the effect abolished by a beta lyase inhibitor.
3 Several identified or putative P450 metabolites of tetrachloroethylene are mutagenic.
4 Tetrachloroethylene-epoxide, a hypothesized intermediate in tetrachloroethylene P450 oxidative
5 metabolism (Henschler et al., 1977a, b), is mutagenic in bacteria (Kline et al., 1982). As
6 reviewed by Moore and Harrington-Brock (2000), the oxidative metabolite TCA, the major
7 urinary excretion product, exhibits little, if any, genotoxic activity. However, in vitro
8 experiments with TCA should be interpreted with caution if steps have not been taken to
9 neutralize pH changes caused by the compound. TCA was positive in genotoxicity studies
10 conducted by Bhunya and Behera (1987), Bhunya and Jena (1996), and Birner et al. (1994) in in
11 vivo mouse and chick test systems. TCA has also been reported to induce DNA SSB in hepatic
12 DNA of mice. A single dose of TCA was administered to Sprague-Dawley rats and B6C3F1
13 mice by gavage (Nelson and Bull, 1988). The animals were sacrificed 4 hrs later and SSB in
14 liver DNA were analyzed by alkaline unwinding assay. SSB were observed in a dose-dependent
15 manner. The lowest dose of TCA that produced significant SSB in the rats was 0.6 mmol/kg (98
16 mg/kg). For mice, the lowest dose of TCA that produced significant increases was 0.006
17 mmol/kg (0.98 mg/kg). Further, in another study by the same authors (Nelson et al., 1989), the
18 incidence of SSB was elevated at 1 hr after a single i.p. dose TCA exposure of 500 mg/kg; the
19 level returned to control levels by 8 hrs. In a second experiment, no increase in SSB in hepatic
20 DNA was observed 24 hrs after 10 days of daily gavage of 500 mg/kg TCA. A later study by
21 Styles et al. (1991), using essentially the same procedures, failed to detect any increase in SSB.
22 Chang et al. (1992) observed a marginally significant increase in SSB in hepatocyte DNA
23 of mice but not rats at 4 hrs after a single TCA dose of 10 mmol/kg (1,633.9 mg/kg)
24 administered orally. However, the authors considered this finding to be not biologically
25 significant, because SSB were not increased at 1 hr and there were no detectable SSB in isolated
26 hepatocytes exposed to concentrations of TCA as high as 10 mM (-1,650 mg/L). Storer et al.
27 (1996), after evaluating 81 chemicals (carcinogens, noncarcinogens, mutagens, and
28 nonmutagens) for SSB using the alkaline unwinding assay, demonstrated that increased DNA
29 SSB at high doses can be the result of cytotoxicity involving endonucleocytic degradation of
30 DNA.
31 As reviewed elsewhere (see Salmon et al., 1995; Moore and Harrington-Brock, 2000),
32 chloral hydrate is mutagenic in the standard battery of screening assays. Effects include positive
33 results in bacterial mutation tests for point mutations and in the mouse lymphoma assay for
34 mutagenicity at the Tk locus (e.g., Haworth et al., 1983). In vitro tests showed that chloral
35 hydrate also induced micronuclei and aneuploidy in human peripheral blood lymphocytes or
36 Chinese hamster pulmonary cell lines. Micronuclei were induced in Chinese hamster embryonic
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-5 DRAFT-DO NOT CITE OR QUOTE
-------
1 fibroblasts. Several studies demonstrate that chloral hydrate induces aneuploidy (loss or gain of
2 whole chormosomes) in both mitotic and meiotic cells, including yeast (Singh and Sinha, 1976,
3 1979; Kafer, 1985; Gualandi, 1987; Sora and Agostini-Carbone, 1987), cultured mammalian
4 somatic cells (Degrassi and Tanzarella, 1988), and spermatocytes of mice (Russo et al., 1984;
5 Liang and Pacchierotti, 1988). Chloral hydrate has also been shown to block spindle elongation
6 in insect spermatocytes (Ris, 1949). Chloral hydrate was negative for sex-linked recessive lethal
7 mutations in drosophila (Yoon et al., 1985). It induces SSB in hepatic DNA of mice and rats
8 (Nelson and Bull, 1988) and mitotic gene conversion in yeast (Bronzetti et al., 1984). Schatten
9 and Chakrabarti (1998) showed that chloral hydrate affects centrosome structure, which results
10 in the inability to reform normal microtubule formations and causes abnormal fertilization and
11 mitosis of sea urchin embryos.
12 The chloroacid metabolite, DCA, is also mutagenic in the standard battery of screening
13 tests (reviewed by Moore and Harrington-Brock, 2000). DCA was positive in bacterial mutation
14 tests, in the in vitro mouse lymphoma assay, the micronucleus induction test, the Big Blue mouse
15 system and other tests (DeMarini et al., 1994; Fuscoe et al., 1996; Nelson and Bull, 1988;
16 Harrington-Brock et al., 1998; Leavitt et al., 1997; Chang et al., 1989; Bignami et al., 1980).
17 Anna et al. (1994) compared mutations in the ras gene in liver tumors in mice treated orally with
18 tetrachloroethylene, DCA and trichloroethylene with those in untreated mice. The frequency of
19 mutations at codon 61 ofH-ras was significantly lower in liver tumors of tetrachloroethylene-
20 exposed mice but not in DCA or trichloroethylene tumors. Thus, the phenotype of
21 tetrachloroethylene-induced mouse tumors appeared to differ from trichloroethylene, DCA or
22 spontaneous occurring tumors. While not sufficient to indicate the MO A, tumor phenotype data
23 regarding H-ras codon 61 suggests that tetrachloroethylene-induced liver tumors differ from
24 those induced by DCA, TCA, or trichloroethylene and those arising spontaneously in the mouse.
25 In summary, tetrachloroethylene has been shown to induce some genotoxic effects
26 (micronuclei and SCEs following in vitro exposure, DNA binding and SSBs in tumor tissue).
27 Results of in vitro mutagenicity (Ames) or DNA binding assays of tetrachloroethylene have
28 largely been negative except in the few tests of conditions where metabolites of the GSH
29 pathway are generated. The GSH metabolites are clearly mutagenic. TCVC is the most potent
30 bacterial mutagen of the tetrachloroethylene metabolites and induces UDS in a porcine kidney
31 cell line; TCVG and NAcTCVC are also mutagenic in bacteria. The known (DCA) or putative
32 (tetrachloroethylene oxide, chloral hydrate) P450 metabolites also exhibit mutagenicity.
33 Uncertainties with regard to the genotoxicity characterization include that not all
34 tetrachloroethylene metabolites have been identified, nor have all the known or postulated
35 metabolites been sufficiently tested in the standard genotoxicity screening battery. Of note,
36 bacterial mutation testing protocols typically specify the inclusion of cytotoxic concentrations of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-6 DRAFT-DO NOT CITE OR QUOTE
-------
1 the test article, and the relative potency of the metabolites in vitro may not necessarily inform
2 their relative contribution to the overall mechanistic effects of the parent chemical. This may be
3 especially relevant when evaluating in vitro testing results for tetrachloroethylene, which can
4 undergo inter-organ metabolic processing involving multiple enzyme systems to yield highly
5 reactive species. In addition, such tests do not provide data for all effects that are relevant for
6 carcinogenesis. Thus, other data gaps include incomplete characterization of the metabolites in
7 tests beyond the standard battery of genotoxicity tests including on important genetic and
8 epigenetic endpoints.
9 Section 4.10.3 addresses the contribution of mutagenicity of tetrachloroethylene and its
10 oxidative and GSH-derived metabolites to the MO A of carcinogeniticy for tetrachloroethylene.
11 Overall, the finding is that the MOA for tetrachloroethylene-induced carcinogenesis is not yet
12 fully characterized, completely tested, or understood. The database for hepatocarcinogenesis is
13 especially limited with regard to chemical-specific studies. It is concluded that the role of
14 genotoxicity in hepatocarcinogenicity, an effect that is thought to be related to products of CYP
15 metabolism, is uncertain (see Section 4.4.4.5). While the complete mechanisms are not yet
16 understood, the weight of evidence, including the known mutagenicity of GSH-derived
17 metabolites produced in the kidney, suggests a mutagenic MOA cannot be ruled out for
18 tetrachloroethylene-induced renal carcinogenesis (see Section 4.5.4.3.3).
19
20 4.4. LIVER TOXICITY
21 4.4.1. Human Effects
22 A number of hepatotoxic effects, including hepatomegaly, hepatocellular damage, and
23 elevations of several hepatic enzymes and bilirubin degradation byproducts, have been observed
24 after acute high-level exposure to tetrachloroethylene (levels not identified; Meckler and Phelps,
25 1966; Coler and Rossmiller, 1953; Hake and Stewart, 1977; Saland, 1967; Stewart et al., 1961,
26 as reported in ATSDR, 1997). One case report noted obstructive jaundice and hepatomegaly in
27 an infant exposed orally to tetrachloroethylene (1 mg/dL; Bagnell and Ellenberger, 1977, as
28 reported in ATSDR, 1997).
29
30 4.4.1.1. Liver Damage
31 Four cross-sectional studies were available that evaluated the prevalence of liver damage
32 among dry cleaner populations (Lauwerys et al., 1983; Cai et al., 1991; Gennari et al.; 1992;
33 Brodkin et al., 1995). These studies assessed serum concentration of a number of hepatic
34 enzymes in dry cleaner and control populations. Additionally, sonographic changes to hepatic
35 parenchymal tissue were examined in one study (Brodkin et al., 1995). An elevated
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-7 DRAFT-DO NOT CITE OR QUOTE
-------
1 concentration of the serum enzyme GOT and mild hepatic changes were notable observations in
2 two studies (Gennari et al., 1992; Brodkin et al., 1995).
3 Gennari et al. (1992) measured the electrophoretic fractionation patterns of serum GGT
4 isozymes among 141 tetrachloroethylene-exposed dry cleaners and 130 nonexposed controls
5 selected from staff and students from the academic institution of the principal investigators.
6 Both the exposed subjects and the controls had similar lifestyle (smoking, alcohol consumption)
7 and clinical medical histories. The TWA tetrachloroethylene concentration in the dry cleaning
8 facilities was 11.3 ppm. Total GGT was higher in exposed workers (exposed: mean of 12.4
9 international units per liter [U/L; standard deviation, 6.9 U/L]; controls: 8.8 U/L [4.9 U/L],/? <
10 0.01). The GGT-2 isoenzyme component was higher in exposed workers (6.8 U/L [5.7 U/L] in
11 exposed vs. 3.5 U/L [3.3 U/L] in controls,/* < 0.01) and the GGT-4 component was detectable in
12 exposed workers but not measurable in controls. The authors regarded a GGT-2/GGT-3 ratio of
13 greater than 1 as a sensitive index of the reciprocal behavior of the two isoenzymes. GGT-2 is
14 generally associated with activation of liver microsomal enzymes. GGT-4 is common in liver
15 diseases and indicates hepato-biliary impairment.
16 This study excluded individuals who presented values for GGT, or other liver enzymes
17 above a normal range, and individuals who had past or current liver disease. None of the
18 workers showed any clinical symptoms of liver disease, and their enzymatic profiles, including
19 GGT, aspartase amino transaminase (AST), alanine amino transaminase, 5'-nucleotidase, and
20 alkaline phosphotase, were within the clinically normal reference limits. Given the study's
21 exclusion criteria, it is not surprising that liver enzyme concentrations were within a normal
22 range. The authors stated that more research is required to develop this GGT fractionation assay
23 into a clinically useful method of measuring liver function. Nevertheless, the study showed that
24 these dry cleaners had markers of tetrachloroethylene oxidative metabolism (GGT-2) and liver
25 impairment (GGT-4).
26 The study by Brodkin et al. (1995) examined liver function and carried out sonography
27 measurements in a population of 27 dry cleaners and 26 nonexposed laundry workers. Dry
28 cleaners were older and had a longer duration of employment than did laundry workers. The
29 noninvasive imaged penetration of ultrasound into liver tissue can reveal the presence of fat
30 accumulation and fibrous structures. The mean TWA exposure (8 hrs) among all dry cleaners
31 was 15.8 ppm (range: 0.4-83 ppm). The investigators found a higher prevalence of abnormal
32 hepatic sonograms among the dry cleaners (67%) than among laundry workers (38%;p < 0.05),
33 the control group. Hepatic parenchymal changes, as assessed by sonography, were graded as
34 mild, moderate, or severe. The prevalence of hepatic parenchymal changes increased both with
35 increasing current concentration and with cumulative exposure (p < 0.05). Subjects with
36 serological evidence of active hepatitis infection were excluded from these analyses.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-8 DRAFT-DO NOT CITE OR QUOTE
-------
1 Brodkin et al. (1995) fit logistic regression models to examine possible associations
2 between mild or greater parenchymal changes and tetrachloroethylene exposure. These analyses
3 included adjustment for the effects of ethanol consumption within the past six months, sex, body
4 mass index, age, and serological evidence of active and past hepatitis infection. Subjects with
5 serological evidence of active hepatitis infection were included in the logistic regression analysis
6 due to the ability of the statistical method to account for the effects associated with this factor.
7 These analyses showed subjects exposed during older wet or dry-to-dry transfer processes
8 (average concentration: 19.8 ppm; range: 1.8-83 ppm) was strongly—but imprecisely—
9 associated with mild or greater sonographic changes (odds ratio [OR] = 4.2, 95% confidence
10 interval [CI] = 0.9-20.4) as compared with controls. No association was shown with subacute
11 exposure in new dry-to-dry operations (OR = 0.7, 95% CI = 0.1-5.9). An inverse dose-response
12 association was found with cumulative exposure after adjustment for age due to a strong but
13 imprecise association between tetrachloroethylene exposure and hepatic sonographic changes in
14 younger workers (workers less than 35 years of age, OR =15; 95% CI = 1.33-170).
15 Only 21% of the exposed study subjects who had changes graded as mild or greater had
16 increases in any hepatic enzyme (Brodkin et al., 1995). Mean concentrations of GGT, AST, and
17 alanine transferase (ALT) tended to be higher among the dry cleaners as compared with laundry
18 workers; however, the differences were not statistically significant and all mean values were
19 within the normal range of reference values. However, all of the subjects who had elevated ALT
20 concentrations had moderate or severe sonographic changes. Hence, sonographic imaging of the
21 liver appeared to be a more sensitive indicator of toxicity than was measurement of serum
22 hepatic enzymes.
23 Lauwerys et al. (1983) performed behavioral, renal, hepatic, and pulmonary tests on 22
24 subjects exposed to tetrachloroethylene in six dry cleaning shops and compared the results with
25 those obtained for 33 subjects nonoccupationally exposed to organic solvents. The mean TWA
26 concentration was 21 ppm. The investigators found no statistically significant differences in
27 mean serum hepatic enzyme concentration between exposed subjects and controls, but they did
28 not describe the statistical methods used to test for differences between the exposed and control
29 groups.
30 Cai et al. (1991) investigated subjective symptoms, hematology, serum biochemistry, and
31 other clinical signs in 56 dry cleaners exposed to tetrachloroethylene at 20 ppm (as a geometric
32 mean of 8 hr TWA) and compared the results with findings for 69 nonexposed controls from the
33 same factories. Exposure-related increases were observed in the prevalence of subjective
34 symptoms during the workday as well as in the past 3-month period, whereas no significant
35 changes in hematology were seen. There was no effect on liver and kidney function, as
36 measured by enzyme activities, blood urea nitrogen (BUN), and creatinine in the serum.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-9 DRAFT-DO NOT CITE OR QUOTE
-------
2
O
4
5
6
7
8
9
10
11
Table 4-1 presents a summary of the human liver toxicity studies in dry cleaners. Two of
the four studies (Brodkin et al., 1995; Gennari et al., 1992) showed clinical signs of liver toxicity,
namely, sonographic changes in the liver and higher serum concentrations of liver enzymes
indicative of liver injury in the absence of frank toxicity. Subjects in these two studies were
exposed to tetrachloroethylene for a longer duration than were subjects in Cai et al. (1991) or
Lauwerys et al. (1983), and for this reason these two studies carry greater weight in this analysis.
Moreover, the studies by Brodkin et al. (1995) and Gennari et al. (1992) assessed potential liver
damage using a different set of markers than those of Cai et al. (1991) or Lauwerys et al. (1983).
Table 4-1. Summary of studies of human liver toxicity
Subjects
27 tetrachloroethylene -
exposed dry cleaners
26 nonexposed laundry
workers
141 tetrachloroethylene-
exposed dry cleaners
130 controls
24 tetrachloroethylene -
exposed dry cleaners
33 controls non-
occupationally exposed to
organic solvents
56 tetrachloroethylene -
exposed dry cleaners
69 nonexposed factory
controls
Effects
Sonographic scattering of
fat in liver (in vivo)
Severity greater with
higher cumulative exposure
No liver toxicity
Elevation of total GGT due
to GGT-2
GGT-4 detected in exposed
but not in control workers
No effect on serum hepatic
enzymes
Increased subjective
symptoms
No effects on serum
indicators of liver and
kidney toxicity
Exposure
Group mean TWA =
15.8 ppm
Mean duration of exposure =
12 years
Mean TWA =11.3 ppm
Mean duration of
exposure = 20 years
Mean TWA = 21 ppm
Mean duration of
exposure = 6 years
Geometric mean TWA
= 20 ppm
Mean duration of
exposure = 3 years
Author
Brodkin et
al. (1995)
Gennari et
al. (1992)
Lauwerys
etal.
(1983)
Cai et al.
(1991)
12
13
14
15
16
17
Biological markers of liver effects permit the early identification of adverse effects of
xenobiotic exposure. They are an important link between biological markers of exposure and
frank liver toxicity, and they offer the most potential for clinical intervention before irreversible
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-10 DRAFT-DO NOT CITE OR QUOTE
-------
1 effects have occurred (NRC, 1995). The observations of Brodkin et al. (1995) and Gennari et al.
2 (1992) support the indication that tetrachloroethylene exposure affects liver function; hence, the
3 lowest-observed-adverse-effect level (LOAEL) for liver effects in humans can be established as
4 a range from 12 to 16 ppm (TWA).
5
6 4.4.1.2. Liver Cancer
1 Cohort and case-control studies assessing possible association between liver cancer and
8 dry cleaner and laundry workers, or tetrachloroethylene specifically, are identified in Tables
9 4B-la, 4B-lb, and 4B-3 (Appendix 4B). An incidence study by Andersen et al. (1999) of dry
10 cleaning and laundry workers in Denmark, Finland, Norway, and Sweden reported the following
11 liver cancer risks: males, standardized indicence ratio (SIR) =1.3 (95% CI = 0.6-2.3); females,
12 SIR= 1.3 (95% CI = 0.9-1.9); combined, SIR= 1.3 (95% CI = 0.9-1.8). This study included
13 some of the same subjects as the studies by Lynge and Thygesen, who also reported an elevation
14 in liver cancer incidence among Danish female dry cleaners and laundry workers (Lynge and
15 Thygesen, 1990; Lynge, 1994), and the study by Travier et al. (2002) of Swedish dry cleaners,
16 launderers, and pressers.
17 Risk for primary liver cancer in these analyses was larger than the risk for the liver and
18 biliary tract cancer, which indicates the potential for bias due to disease misclassification in
19 studies that examine liver cancer as a broad category. A nested case-control study (Lynge et al.,
20 1995) suggests that the excess primary liver cancer risk among females observed in Lynge
21 (1994) is attributable to laundry workers rather than to dry cleaners (Table 4B-3 in Appendix 4B).
22 This type of information is not available for primary liver cancer cases in Andersen et al. (1999).
23 Mortality studies are biased due to misclassification of liver cancer on death certificates; and
24 these studies do not report statistically significant elevated risks for liver and biliary tract cancer.
25 Primary liver cancer mortality was not elevated, and observations from case-control studies that
26 assessed generic organic solvents or dry cleaning fluid mixtures did not show a consistent liver
27 response (Wartenberg et al., 2000).
28
29 4.4.2. Animal Studies
30 4.4.2.1. Liver Toxicity
31 Hepatic ffects observed after subchronic or chronic inhalation exposure to
32 tetrachloroethylene include increased liver weight (Kjellstrand et al., 1984; Kyrklund et al.,
33 1990); hypertrophy (Odum et al., 1988); fatty changes (Kylin et al., 1965; Odum et al., 1988);
34 peroxisome proliferation, an increase in the size and numbers of peroxisome organelles (Odum et
35 al., 1988; Goldsworthy and Popp, 1987; Bergamaschi et al., 1992); other histological lesions
36 (Kjellstrand et al., 1984; NTP, 1986a); and necrosis and tumors (NTP, 1986a; JISA, 1993). Liver
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-11 DRAFT-DO NOT CITE OR QUOTE
-------
1 toxicity observed in animal studies has been reviewed (see U.S. EPA, 1980, 1985a, b, 1986a,
2 1991a; IARC, 1995; ATSDR, 1997; NYS DOH, 1997; Cal EPA, 2001).
3 Species differ in their susceptibility to tetrachloroethylene-induced hepatic toxicity. For
4 example, mice appear to be more sensitve than rats to the adverse liver effects caused by
5 tetrachloroethylene exposure (U.S. EPA, 1985a; NTP, 1986a; Lash and Parker, 2001). In Rowe
6 et al. (1952), guinea pigs exposed to 100 to 2,500 ppm proved to be more susceptible than rabbits,
7 monkeys, and rats to liver toxicity. The lowest reported level for liver effects in laboratory
8 animals is in tetrachloroethylene-exposed NMRI mice at 9 ppm (61 mg/m3; Kjellstrand et al.,
9 1984). These investigators exposed male and female mice to 9 ppm and higher concentrations of
10 tetrachloroethylene for 30 days and observed changes indicative of adverse health effects
11 including statistically significant increases in liver weight as well as changes in liver morphology.
12 Increases in levels of blood plasma enzyme butyrylcholinesterase (BuChE) were reported at all
13 tetrachloroethylene concentration levels at or above 9 ppm. A recovery period reversed the
14 effects on BuChE, although liver weight was still slightly elevated at 120 days after cessation of
15 tetrachloroethylene exposure for 30 days at 150 ppm.
16 Chronic lifetime inhalation bioassays of tetrachloroethylene in mice have been conducted
17 by the National Toxicology Program (NTP, 1986a), the Japan Industrial Safety Association
18 (JISA, 1993), and Nagano et al. (1998). In the NTP study, B6C3F1 mice were exposed to 0, 100,
19 and 200 ppm tetrachloroethylene for 104 weeks. In addition to liver tumors in mice of both
20 sexes, the authors reported liver degeneration in 2/49, 8/49, and 14/50 males and in 1/49, 2/50,
21 and 13/50 females. Liver necrosis was seen in some of the mice (1/49, 6/49, and 15/50 males;
22 3/48, 5/50 and 9/50 females). The authors also observed nuclear inclusions in male mice (2/49,
23 5/49, and 9/50). No dose-related liver effects were reported in the rats.
24 In the Japan Industrial Safety Association (JISA, 1993) study (some results reported in
25 Nagano et al., 1998), male and female Crj/BDFl mice were exposed to 0, 10, 50, and 250 ppm
26 tetrachloroethylene for 104 weeks and sacrificed at 110 weeks. In addition to hepatocellular
27 carcinomas and adenomas in the mice, telangiectasis (vascular lesions formed by dilation of a
28 group of small blood vessels) and focal necrosis occurred in males at 50 ppm and above. Liver
29 degeneration was observed at 250 ppm in both sexes. Liver hemangiosarcomas were also
30 reported in the male mice. The authors described effects in F344/DuCrj rats exposed to 0, 50,
31 200, 600 ppm for 104 weeks and sacrificed at 110 weeks. Male, but not female, rats had excess
32 incidence of spongiosis hepatitis at 200 ppm and above and hyperplasia at 600 ppm. Liver
33 tumors were not observed in either male or female rats.
34 Tetrachloroethylene was found to cause liver toxicity in laboratory animals by the oral
35 route in several studies (e.g., Buben and O'Flaherty, 1985; Story et al., 1986; Hayes et al., 1986).
36 The observed effects included increased liver weights, biochemical changes, histological lesions,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-12 DRAFT-DO NOT CITE OR QUOTE
-------
1 necrosis, and polyploidy. Buben and O'Flaherty treated male Swiss-Cox mice with
2 tetrachloroethylene doses of 0, 20, 100, 200, 500, 1,000, 1,500, or 2,000 mg/kg-day for 5
3 days/week for 6 weeks. These investigators demonstrated that indices of tetrachloroethylene
4 hepatotoxicity (increased liver weight, liver triglyceride accumulation, glucose-6-phosphotase
5 activity, and serum glutamic pyruvic transaminase activity) were highly correlated with the
6 amount of tetrachloroethylene metabolized by the mice. The degree of liver response, as
7 measured by each toxicity parameter when plotted against total urinary metabolites, was linear in
8 all cases. The several dose-related liver effects were reported at doses above the lowest dose of
9 20 mg/kg-day. Increased liver triglycerides and increased liver-to-body weight ratios were seen
10 in mice receiving 100 mg/kg-day and higher doses. At doses of 500 mg/kg-day and higher,
11 effects in the treated mice also included reduction of DNA content, increased serum levels of
12 liver enzymes, liver degeneration, necrosis, and polyploidy. The LOAEL was 100 mg/kg-day.
13 Ebrahim et al. (1996) administered 3 g/kg/day tetrachloroethylene in sesame oil to mice
14 for 15 days and observed a significant increase in liver weight and degeneration and necrosis of
15 hepatocytes. These changes occurred simultaneously with a decrease in blood glucose; elevated
16 activities of enzymes hexokinase, aldolase, and phosphoglucoisomerase; and decreased activities
17 of gluconeogenic enzymes. Table 4-2 presents a summary of liver toxicity studies in animals.
18
19 4.4.2.2. Liver Cancer
20 In carcinogenicity bioassays, tetrachloroethylene has been shown to cause a statistically
21 significant increase in the incidence of hepatocellular carcinomas in both sexes of B6C3F1 mice
22 following either oral gavage administration or inhalation exposure (NCI, 1977; NTP, 1986a).
23 Both sexes of Crj :BDF 1 mice have also been shown to develop an increased incidence of
24 hepatocellular carcinomas when exposed to tetrachloroethylene by inhalation (Nagano et al.,
25 1998; JISA, 1993). The National Cancer Institute (NCI) and NTP bioassays were reviewed
26 previously by EPA (U.S. EPA, 1985a, 1986a, 1991a) and are briefly summarized here.
27 Observations regarding liver cancer from the more recent study (Nagano et al., 1998; JISA,
28 1993), which confirms the earlier findings of liver tumors in B6C3F1 mice, is also briefly
29 summarized. The tumor incidence data from these studies, with the accompanying tables and
30 figures, are presented in Section 5.3.2.
31 Several metabolites of tetrachloroethylene have been found to be carcinogenic in mice,
32 and it is thought that the hepatocarcinogenicity of the parent compound is mediated through the
33 action of one or more of its metabolites. Metabolites of tetrachloroethylene, including TCA,
34 DCA, and the putative metabolite chloral hydrate, have been observed to cause liver cancer in
35 mice (Daniel et al., 1992; Rijhsinghani et al., 1986; Herren-Freund et al., 1987; Bull et al., 1990;
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-2. Summary of rodent liver toxicity studies
Subjects
NMRI mice (both
sexes), inhalation
Swiss-Cox mice
(male), gavage
B6C3F1 mice
Crj/BDFl mice
(both sexes)
F344/DuCrj rats
(both sexes)
Effects
Increase in liver weight
Morphological changes
Increased plasma butylcholinesterase
Increased liver/body weight ratio at 100
mg/kg-day
Increased triglycerides at 100 mg/kg-day
No change at 20 mg/kg-day
Liver degeneration and necrosis at 100
ppm and higher in males and at 200 in
females
Focal necrosis in males at 50 ppm and
higher
Liver degeneration in males and females
at 250 ppm
Spongiosis hepatitis in males at 200 ppm
and higher
Hyperplasia in males at 600 ppm
Exposure
9 ppm and above for
4 weeks, inhalation
0, 20, 100, 200, 500,
1,000, 1,500, 2,000
mg/kg-day for 6
weeks, gavage
0, 100, 200 ppm for
104 weeks
0, 10, 50, 250 ppm
for 110 weeks
0, 50, 200, 600 ppm
for 110 weeks
Authors
Kjellstrand et
al. (1984)
Buben and
O' Flaherty
(1985)
NTP, (1986a)
JISA(1993)
JISA(1993)
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Richmond et al., 1995; Pereira, 1996; DeAngelo et al., 1991, 1996, 1999; NTP, 2000a, b). In
addition, DCA causes liver cancer in rats (DeAngelo et al., 1996; Richmond et al., 1995).
In the mouse gavage study (NCI, 1977), groups of 50 male mice received TWA doses of
536 or 1,072 mg/kg tetrachloroethylene in corn oil by intragastric gavage for 78 weeks (450 or
900 mg/kg for 11 weeks, then 550 or 1,100 mg/kg for 67 weeks). Groups of 50 female mice
received TWA doses of 386 or 772 mg/kg of tetrachloroethylene in corn oil by gavage (300 or
600 mg/kg for 11 weeks, then 400 or 800 mg/kg for 67 weeks). Mice were dosed 5 days/week.
The tetrachloroethylene used in the study was greater than 99% pure, but impurities were not
identified (NCI, 1977; U.S. EPA, 1985a). The test sample was estimated to contain
epichlorohydrin concentrations of less than 500 ppm (U.S. EPA, 1985a). It was considered
unlikely, however, that the tumor response resulted from this low concentration of
epichlorohydrin. Tetrachloroethylene caused statistically significant increases (p < 0.001) in the
incidences of hepatocellular carcinoma in both sexes of mice in both treatment groups when
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-14 DRAFT-DO NOT CITE OR QUOTE
-------
1 compared with untreated controls or vehicle controls. The time to tumor was significantly
2 decreased in treated mice.
3 The inhalation study (NTP, 1986a) confirmed the finding of hepatocellular carcinoma in
4 B6C3F1 mice. Groups of 50 mice of each sex were exposed to (epichlorohydrin free)
5 tetrachloroethylene concentrations of 0, 100, or 200 ppm, 6 hrs/day, 5 days/week, for 103 weeks.
6 Tetrachloroethylene caused statistically significant dose-related increases in the incidences of
7 hepatocellular carcinoma and in combined hepatocellular adenoma and carcinoma in both sexes.
8 More recent tetrachloroethylene inhalation studies conducted in Japan using Crj :BDF1
9 mice resulted in the observation of hepatocellular carcinomas in both sexes (JISA, 1993 [results
10 reported in Nagano et al., 1998]). Groups of 50 male and 50 female mice were exposed to 0, 10,
11 50, and 250 ppm tetrachloroethylene, 6 hrs/day, 5 days/week, for 104 weeks, and the terminal
12 sacrifice was performed at 110 weeks. Both males and females showed dose-related increased
13 incidences of liver carcinomas and combined liver adenomas and carcinomas. Malignant liver
14 hemangioendotheliomas were also increased in males. Both malignant and combined benign and
15 malignant hemangioendotheliomas in the spleen were increased in males. The investigators also
16 observed Harderian gland adenomas and enlargement of the nucleus in the kidney proximal
17 tubular cells in male mice at the highest dose.
18
19 4.4.3. Summary of Liver Effects in Humans and Animals
20 Two of four studies of occupationally exposed dry cleaners showed indications of liver
21 toxicity, namely sonographic changes of the liver and altered serum concentrations of liver
22 enzymes indicative of liver injury. Frank liver disease was not seen among these workers for a
23 number of possible reasons: individuals with frank liver disease may not have been included in
24 cross-sectional studies because they had left the workforce due to their conditions, the healthy
25 worker effect, and other selection biases. LOAELs in these human studies were between 12 and
26 16 ppm (TWA).
27 Primary liver cancer incidence was not consistently elevated across incidence studies and
28 appeared to be associated with laundry work (Andersen et al., 1999; Travier et al., 2002; Lynge
29 and Thygesen, 1990; Lynge et al., 1995). Additionally, elevated risks were also seen for
30 incidence in the combined category of primary liver cancer and cancer of the biliary passages.
31 Primary liver cancer is often misclassified on death certificates; hence, mortality studies that
32 examine mortality from liver and biliary passage cancer are less informative than are studies of
33 incidence. For this reason, greatest weight is placed on the observations of incident studies of
34 primary liver disease. Observations from case-control studies that assess organic solvent
35 generically or dry cleaning fluid mixtures do not show a consistent carcinogenic effect on the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-15 DRAFT-DO NOT CITE OR QUOTE
-------
1 liver (Wartenberg et al., 2000). There are no human studies of drinking water or other oral
2 exposure.
3 In animals, liver toxicity, manifested by fatty changes, liver enlargement, and enzyme
4 changes in blood, has been observed in rats and mice in several studies. The LOAEL for the
5 inhalation studies, 9 ppm, is from a 30-day-exposure mouse study. A chronic mouse inhalation
6 bioassay showed liver necrotic foci at 50 ppm and higher. In two lifetime inhalation cancer
7 bioassays, increases in liver cancer occurred at 100 ppm and above, and there was a significant
8 dose-response trend in both studies. With oral administration, liver effects have been observed at
9 100 mg/kg-day, although these were not considered to be irreversible effects. The lowest dose at
10 which liver tumors have appeared is 386 mg/kg-day, administered long term.
11
12 4.4.4. Mode of Action for Liver Toxicity
13 This section summarizes scientific data regarding the MOA for tetrachloroethyene-
14 induced hepatic toxicity and carcinogenicity in mice and its relevance to humans. The MOA for
15 tetrachloroethylene-induced mouse liver cancer is not well understood, and it is highly likely that
16 more than one MOA is operative. The following topics are relevant to the MOA for liver
17 toxicity.
18 (1) Tetrachloroethylene metabolites and liver toxicity. Metabolic activation of
19 tetrachloroethylene is required to produce adverse effects in the liver. TCA is the major urinary
20 excretion product, and it is also a hepatocarcinogen in mice; however, insufficient amounts of
21 TCA are produced from tetrachloroethylene metabolism to quantitatively account for the mouse
22 liver tumor incidences observed in cancer bioassays. In addition, the liver tumor phenotypes
23 with regard to H-ras codon 61 mutation do not appear to be similar between TCA, DC A, and
24 tetrachloroethylene. Therefore, it is likely that other tetrachloroethylene metabolites, such as the
25 potentially reactive trichloroacetyl chloride, are contributing to the production of liver tumors.
26 The potential role of GST conjugates of tetrachloroethylene in liver toxicity, although unknown,
27 is presumed to be less than that in the kidney.
28 (2) Role of receptor activation. Data exist to advance the hypothesis that peroxisome
29 proliferators can contribute to liver tumorigenesis in rodents; however, the causal role of PPAR-
30 mediated events in tumorigenesis, and human sensitivity to these effects needs further scientific
31 examination and analysis. Data suggest that tetrachloroethylene is a very weak peroxisome
32 proliferator. The strongest evidence supporting the PPAR MOA for tetrachloroethylene is the
33 data for TCA; however, TCA also has other MO As, TCA alone does not account quantitatively
34 for tetrachloroethylene induced tumors, and tetrachloroethylene- and TCA-induced tumors are
35 phenotypically distinct (Bull et al., 2002).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-16 DRAFT-DO NOT CITE OR QUOTE
-------
1 (3) Genotoxic effects. Tetrachoroethylene has been shown to induce some genotoxic
2 effects (micronuclei and SCEs following in vitro exposure, DNA binding and SSBs in liver).
3 Results of in vitro mutagenicity (Ames) or DNA binding assays of tetrachloroethylene have
4 largely been negative except in the few studies of conditions where metabolites of the GSH
5 pathway are generated. The GSH metabolites are clearly mutagenic. In addition, several known
6 (DCA) and putative (tetrachloroethylene oxide) P450 metabolites exhibit mutagenicity. The
7 mutagenic potential of reactive metabolites of tetrachloroethylene has not been adequately
8 studied. Moreover, the identity of all metabolites is not known.
9 (4) Nongenotoxic effects. Existing data suggest the involvement of events related to
10 tumor induction that are nonspecific to activation of PPAR-a, being common to other
11 nongenotoxic MO As. Hypomethylation is a common early molecular event in most tumors, and
12 alterations in DNA methylation following exposure to chemicals, both hypomethylation and
13 hypermethylation, may be factors in tetrachloroethylene-induced tumorigenesis. Although
14 tetrachloroethylene-specific data are lacking, its metabolites DCA and TCA are known to induce
15 hypomethylation of DNA and protooncogenes in mouse liver.
16
17 4.4.4.1. Background
18 Although hepatocellular tumors are common endpoints in mouse carcinogenicity studies,
19 their biological significance with respect to identifying human hazard has long been a subject of
20 intense controversy and debate (Tomatis et al., 1989; Ward et al., 1979; Nutrition Foundation,
21 1983; U.S. EPA, 1985c; U.S. EPA, 1986b, 1991a; Popp, 1984; Stevenson et al., 1990). The
22 current controversy in the case of tetrachloroethylene-induced hepatocellular carcinoma in mice
23 involves identifying the operative MO As and their relevance to human situations.
24 Hemangiosarcomas, unlike hepatocellular carcinomas, are not a common finding in
25 mouse bioassays (U.S. EPA, 2002, 2000b); in fact, they are considered relatively rare, and their
26 relevance to human health hazard, therefore, is generally accepted. Findings of a positive trend
27 for liver and spleen hemangiosarcomas in the most recent mouse carcinogenicity bioassay of
28 tetrachloroethylene (JISA, 1993; Nagano et al., 1998) in a strain of mice not known to have any
29 type of high background tumor incidence constitute important information about risk of exposure
30 to humans.
31 The focus of the human relevance of the hepatocellular carcinomas observed in mice has
32 turned to the emerging information on modes of carcinogenic action. Peroxisome proliferation,
33 which is associated with certain rodent liver carcinogens, has gained increasing attention due to
34 its possible relationship to a hypothesized MOA. Studies of tetrachloroethylene and its
35 chloroacid metabolites suggest that the compound is a peroxisome proliferator chemical, albeit a
36 very weak one.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-17 DRAFT-DO NOT CITE OR QUOTE
-------
1 A lack of human relevance for the MOA associated with peroxisome proliferator
2 carcinogens has been proposed (Klaunig et al., 2003; Meek et al., 2003). However, agreement is
3 lacking on the extent to which the MOA hypothesis has been validated or whether the MOA or
4 quantitative differences among species are sufficiently understood to rule out a potential risk of
5 carcinogenicity to humans (Melnick, 2001; U.S. EPA, 2005a). EPA's Federal Insecticide,
6 Fungicide, and Rodenticide Act Scientific Advisory Panel (SAP) reviewed a draft EPA proposed
7 science policy report (U.S. EPA, 2003d) and considered the role of PPAR-a agonists in
8 activation of PPAR-a leading to an increase in cell proliferation, a decrease in apoptosis, and
9 eventual clonal expansion of preneoplastic cells leading to liver cancer, and the human relevance
10 of PPAR-a agonist-induced hepatocarcinogenesis. While the majority of the panel concluded
11 that there was sufficient evidence in support of the proposed MOA for PPAR-a agonist-induced
12 rodent hepatocarcinogenesis, some panel members complete disagreed. The majority of the
13 panel agreed that there are relevant data indicating that humans are less sensitive than rodents to
14 the hepatic effects of PPAR-a agonists. However, it was noted that humans are not refractory to
15 the effects of PPAR agonism and many questions remain regarding the specific events PPAR
16 activation entails.
17 This assessment attempts to evaluate scientific information and review the current issues
18 on MOA hypotheses pertinent to tetrachloroethylene.
19
20 4.4.4.2. Relationship of Metabolism to Potential Mode of Action and Organ Toxicity
21 Metabolic activation of tetrachloroethylene is required for adverse effects to occur in the
22 liver. The cancer-causing activity of tetrachloroethylene and other chlorinated ethylenes is
23 generally considered to reside in metabolites rather than in the parent compounds (U.S. EPA,
24 1991a; Davidson and Bellies, 1991; Lash et al., 2000a; Lash and Parker, 2001; see Section 3.3).
25 Certain metabolites of tetrachloroethylene—specifically, TCA, DC A, and chloral hydrate—have
26 been shown to cause liver tumors in mice (IARC, 1995; Daniel et al., 1992; Rijhsinghani et al.,
27 1986; Herren-Freund et al., 1987; Bull et al., 1990; Pereira, 1996; Odum et al., 1988; DeAngelo
28 et al., 1991, 1999; DeAngelo, 2000; NTP, 2000a, b; Carter et al., 2003), and they may be
29 involved in hepatocarcinogenicity following exposure to the parent compound.
30 TCA is the metabolite that has received the most attention as being potentially associated
31 with tetrachloroethylene-induced liver tumorigenesis. Although it may play a role in
32 tetrachloroethylene hepatocarcinogenicity, not enough TCA is produced from metabolism to
33 account for all tetrachloroethylene-induced mouse liver tumors (see Appendix 4A) observed in
34 bioassays. Although TCA can be further metabolized to DCA, most DCA originating from
35 tetrachloroethylene is proposed by some investigators to be derived predominantly from the renal
36 beta lyase-mediated cleavage of the TCVC conjugate, where DCA production ultimately results
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-18 DRAFT-DO NOT CITE OR QUOTE
-------
1 from dichlorothioketene in the kidney (Volkel et al., 1998). If this is the case, most
2 tetrachloroethylene-derived DCA would not occur in the liver target organ, implying that DCA
3 would not likely be critically involved in tetrachloroethylene-induced liver tumorigenesis.
4 The potential significance of precursor metabolites—e.g., trichloroacetyl chloride, a
5 major and potentially reactive P450 intermediate—should not be underestimated. The possible
6 roles of key reactive precursor intermediates in causing hepatotoxicity need to be better
7 elucidated, as they may be important to understanding MOA for tetrachloroethylene.
8
9 4.4.4.3. Description of a Hypothesized Mode of Action (MOA): Peroxisome Proliferator-
10 Activated Receptor (PPAR) Mediated Hepatocarcinogenesis
11 4.4.4.3.1. Background summary. Some peroxisome proliferators cause increased incidence of
12 rodent liver tumors. Several investigators (Reddy et al., 1980; Reddy and Lalwai, 1983; Moody
13 et al., 1991; Ashby et al., 1994) have hypothesized a causal relationship between proliferation of
14 peroxisomes and hepatocellular carcinogenicity because peroxisome proliferation in rodent
15 hepatocytes frequently occurs alongside hepatocyte hypertrophy and liver hyperplasia and
16 disproportionate transcriptional increases of peroxisomal enzymes involved in B-oxidation of
17 fatty acids (reviewed by Cattley et al., 1998).
18 Insight into the possible MOA by which chemicals induce peroxisome proliferation—and
19 possibly cancer—was revealed by the discovery of the PPAR receptor, which was shown to be
20 activated by peroxisome proliferators (Issemann and Green, 1990). Activation of this receptor
21 regulates transcription of the genes that encode the enzymes responsible for biochemical changes,
22 including peroxisomal enzymes responsible for beta-oxidation, liver fatty acid-binding protein
23 (Issemann et al., 1993), certain microsomal P450 (CYP4A, CYP2B, and CYP2C) family
24 enzymes (Heuvel, 1999; Gorton et al., 1998; Fan et al., 2003; Simpson et al., 1995, 1996) and
25 other enzymes (Barbier et al., 2003).
26 The evidence regarding whether peroxisome proliferation induced by tetrachloroethylene
27 or its metabolite TCA is the primary or sole mode of action for carcinogenesis is equivocal at
28 best (Ashby et al., 1994; IARC, 1995a; Goldsworthy and Popp, 1987; Odum et al., 1988;
29 Elcombe, 1985; Elcombe et al., 1985; Goldsworthy and Popp, 1987; DeAngelo et al., 1989;
30 Laughter et al., 2004).
31
32 4.4.4.3.2. Summary description of postulated mode of action (MOA)—peroxisome
33 proliferation via modification of cell signal pathways through the peroxisome proliferator-
34 activated receptor (PPAR) receptor. A recent, in-depth review by Klaunig et al. (2003)
35 summarized the PPAR MOA and supporting data; see also the OPP draft science policy paper
36 and the SAP review (U.S. EPA, 2003d). Klaunig et al. (2003) proposed three events to be
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-19 DRAFT-DO NOT CITE OR QUOTE
-------
1 causally related to tumorigenesis: activation of PPAR-a, perturbation of cell proliferation and
2 apoptosis, and selective clonal expansion. The causal role is largely based on evidence that the
3 induction of these events is attenuated in PPAR-a -null mice (or in hepatocytes isolated from
4 such mice) in response to the prototypical agonist WY 14,643 (Lee et al., 1995; Peters et al.,
5 1997). A number of intermediary events are considered associative including: expression of
6 peroxisomal and nonperoxisome genes, peroxisome proliferation, inhibition of gap junction
7 intracellular communication, hepatocyte oxidative stress, as well as Kupffer cell-mediated events.
8 Historically, the increase in peroxisomal organelles, peroxisomal fatty acid beta-
9 oxidation, and alteration in ratios and production of marker enzymes such as increased acyl-CoA
10 oxidase, observed in rodents treated with the peroxisome proliferator chemicals, was thought to
11 induce oxidative stress in hepatocytes and potentially result in oxidative damage to proteins and
12 DNA, leading to carcinogenesis. Two key factors—oxidative injury and enhanced cell
13 proliferation—were implicated in the rodent hepatocarcinogenicity of peroxisome proliferating
14 agents (Cattley et al., 1998; Klaunig et al., 2003). PPAR-a has been identified as the specific
15 PPAR receptor associated with cell proliferation and hepatocarcinogenesis in mouse liver (Lee et
16 al., 1995; Peters et al., 1997a; Corton et al., 2000). PPAR-a activation has been shown to trigger
17 multiple events, and events other than the proliferation of peroxisomes—or at least
18 manifestations not limited to this phenomenon only—are thought relevant to tumorigenesis.
19 Peroxisome proliferators have been shown to alter hepatocyte growth and survival by
20 induction of DNA synthesis and suppression followed by enhancement and then depression of
21 apoptosis (cell death; Cattley and Popp, 1989; Roberts et al., 1995; Bursch et al., 1984; Marsman
22 et al., 1992). Cell proliferation is thought to play an important role through specifically
23 enhanced proliferation of normal hepatocytes, resulting in an increase in the frequency of
24 initiated cells or in the selective growth of pretransformed hepatocytes, with subsequent
25 tumorigenesis (Cattley and Popp, 1989; Kraupp-Grasl et al., 1990, 1991; Grasl-Kraupp et al.,
26 1993; Cattley et al., 1991; Marsman et al., 1988; Eacho et al., 1991; Marsman, 1991; Marsman
27 and Popp, 1994). The currently hypothesized MOA for liver carcinogenesis assumes that events
28 such as the increased cell proliferation, inhibition of apoptosis, and clonal expansion of
29 preneoplastic lesions are linked directly to PPAR-a activation.
30
31 4.4.4.3.3. Identification of potential key events in mode of action (MOA) for liver. Certain
32 biochemical and cellular events have been associated with hepatocarcinogenic effects of
33 peroxisome proliferating chemicals. Whether these key events are causally related to liver
34 tumorigenesis remains to be determined. Potential key events include (a) peroxisome
35 proliferation—an increase in the number of peroxisomes and also an increase in their volume
36 density (Meijer and Afzelius, 1989; Ganning et al., 1983; Thangada et al., 1989); (b) certain
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-20 DRAFT-DO NOT CITE OR QUOTE
-------
1 disproportionate alterations in levels of peroxisomal enzymes, especially increases in fatty acyl-
2 CoA oxidase levels of 20- to 30-fold, whereas catalase and urate oxidase are increased only 2- to
3 3-fold, with resulting excess production of hydrogen peroxide, which may affect hepatocytes by
4 oxidative injury; (c) increases in members of the cytochrome P450 CYP4A subfamily; (d)
5 increases in rates of cell proliferation due to increased DNA synthesis and suppression of
6 apoptosis, particularly in basophilic preneoplastic lesions; (e) hepatomegaly; and (f) expression
7 and activation of the a subtype of the PPAR (PPAR-a; Hardwick et al., 1987; Dreyer et al, 1992;
8 Marcus et al., 1993; Tugwood et al., 1992; Zhang et al., 1992; Cattley et al., 1998; Chevalier and
9 Roberts, 1998).
10
11 4.4.4.3.3.1. Peroxisome proliferator-activated receptor alpha (PPAR-a) activation. The
12 hypothesized causal event most associated—and best supported by existing data—with the
13 proposed peroxisome proliferation MOA is the one having true specificity for PPAR-a MOA.
14 This is the activation of the PPAR-a receptor. PPAR-a expression and activation, inducing
15 transcription of selected genes, possibly in concert with altered cell signal transduction initiated
16 by release of cytokines by hepatic macrophages (Kupffer cells), is probably the important
17 process to evaluate as a potential key event to the development of liver hyperplasia and
18 hepatocyte carcinogenesis. The strongest support for a causal relationship between PPAR-a
19 activation and hepatocellular tumorigenesis is found in studies in null mice, i.e., mice lacking
20 PPAR-a, particularly the eleven-month study of WY-14643 (Peters et al., 1997a, b). Such
21 "knockout" mice do not respond to this prototype peroxisome proliferator with increased cell
22 proliferation and decreased apoptosis or with development of other events potentially associated
23 with PPAR-a activation leading to liver cancer. Although a short-term study in null mice has
24 been performed for TCA (Laughter et al., 2004), such a study has not been conducted for
25 tetrachloroethylene. This short-term study of TCA provides data consistent with a relationship
26 between PPAR-a activation and peroxisome proliferation, but provides minimal, if any, support
27 for PPAR-a activation and liver cancer. Although null mouse studies have flaws, some due to
28 deficiencies in the altered mice such as physiological and biological differences in response to
29 stress when compared to wild mice, a well-designed, lifetime tetrachloroethylene carcinogenicity
30 study in null mice could provide valuable information.
31
32 4.4.4.3.3.2. Alterations in cell replication and death rate. Other events could be claimed to be
33 causal for rodent liver tumor induction. One such event is selective clonal expansion. Stauber et
34 al. (1998) and Bull et al. (2004) reported data indicating that TCA acts to induce liver tumors by
35 increasing clonal expansion of a specific group of initiated cells in mouse liver. TCA stimulates
36 growth of colonies of hepatocytes expressing c-Jun phenotype, which is representative of tumors
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-21 DRAFT-DO NOT CITE OR QUOTE
-------
1 caused by TC A. TCA also lowers replication rates of normal liver cells. TCA treatment,
2 therefore, results in negative selection, or growth advantage being given to specific characteristic
3 tumor cells over normal cells. Although selective clonal expansion has been associated with
4 PPAR-a MO A, it is not clear that this is the case with the tetrachloroethylene metabolite. In fact,
5 these investigators recently stated that they believe TCA causes cancer independently of
6 peroxisome proliferation (Bull, 2004). The DCA metabolite also selectively stimulates the
7 growth of clones of cells. Clonal expansion is thought to occur with all cancer-causing agents,
8 so it is not limited to PPAR-a MOA. The available data indicate that the MO As for the two
9 chloroacid metabolites are different.
10 Another possible causal key event is cell proliferation and apoptosis, although it is not
11 unique to peroxisome proliferator chemicals.
12
13 4.4.4.3.3.3. Potential key events specific for peroxisome proliferator-activated receptor alpha
14 (PPAR-a). Several other events have potential for being key events in the PPAR-a MOA for
15 liver tumors observed in rodents exposed to peroxisome proliferator chemicals. Two events
16 considered specific to PPAR-a activation are the actual proliferation of peroxisome organelles
17 and the expression of peroxisomal genes. These events can be considered biomarkers for
18 peroxisome proliferator chemicals, but a cause-and-effect relationship to liver tumor induction
19 cannot be made. These events are historically linked to the PPAR-a MOA.
20
21 4.4.4.3.3.4. Other potential key events not limited to prottferator-activated receptor alpha
22 (PPAR-a) mode of action (MOA). Yet another key event to consider is an alteration in the
23 expression and activities of nonperoxisomal lipid-metabolizing enzymes that mediate
24 hypolipidemia. PPAR-a agonists shown to cause liver tumors in rodents also induce genes that
25 encode lipid metabolizing enzymes, although these same genes can be altered by other agents.
26 Alterations in DNA methylation, both hypomethylation and hypermethylation, may be
27 factors in tumorigenesis and occur following exposure to chemicals that also cause peroxisome
28 proliferation (Pereira et al., 2004a, b). Hypomethylation is a common early molecular event in
29 most tumors (Pereira et al., 2004b). DCA and TCA are known to induce hypomethylation of
30 DNA and protooncogenes in mouse liver (Tao et al., 1998, 2000; Pereira et al., 2004a).
31 Inhibition of gap junction cellular communication has been attributed to certain
32 peroxisome proliferator chemicals (Klaunig et al., 1988; Dybing et al., 1995). This event is thus
33 correlated with the rodent tumorigenesis caused by such chemicals, although similar inhibition of
34 the gap junction cellular communication process also occurs with other nongenotoxic liver
35 carcinogens (Klaunig et al., 2003) and, therefore, cannot be considered specific to peroxisome
36 proliferators and PPAR-a MOA.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-22 DRAFT-DO NOT CITE OR QUOTE
-------
1 Oxidative stress and resulting DNA damage to hepatocytes secondary to that stress have
2 been attributed to peroxisome proliferator chemicals causing liver cancer in rodents, but the
3 exact role is not clear and remains a controversial issue. As discussed above, alterations in the
4 ratios of peroxisomal enzymes by induction of beta-oxidation enzymes results in an imbalance
5 that causes overall increases in hydrogen peroxide leading to oxidative damage.
6 The nonparenchymal Kupffer cell macrophages may be involved in peroxisome
7 proliferator chemical tumor induction (Klaunig et al., 2003; Rusyn et al., 2000a, 2001). Kupffer
8 cells do not express PPAR-a (Peters et al., 2000). Peroxisome proliferator chemicals activate
9 Kupffer cells directly (Rose et al., 1999; Peters et al., 2000) and independently of PPAR-a
10 activation (Peters et al., 2000). Klaunig et al. (2003) suggest that Kupffer cell mediated events
11 are associated with (i.e., are not causally related to) hepatic tumors induced by the PPAR-a
12 MO A; it is noted that responses of these cells are independent of PPAR-a and are not restricted
13 only to peroxisome proliferator chemicals.
14
15 4.4.4.3.4. Correlation between proliferator-activated receptor alpha (PPAR-a)
16 activation/peroxisomeproliferation and tumor induction. Historically, chemicals have been
17 characterized as peroxisome proliferators on the basis of either observations of increases in
18 volume density of peroxisomes or increases in peroxisomal fatty acid beta-oxidation enzyme
19 activity, with characterization by both of these parameters being preferable. Demonstration of
20 induction of the cyanide-insensitive palmitoyl CoA enzyme is viewed as a key biochemical
21 marker acceptable for the detection and quantitation of peroxisome proliferation. Usually,
22 palmitoyl CoA oxidation (PCO) is measured, although palmitoyl CoA oxidase activity can be
23 determined directly where hydrogen peroxide production is measured.
24 The potential key events described above in Section 4.4.4.3.3 have been correlated with
25 tumorigenesis, although some of these events are not restricted to PPAR-a MO A, and a cause-
26 and-effect relationship is questionable for others. Certain key events are associated with
27 PPAR-a activation and can also be associated with tumorigenesis; however, evidence supporting
28 the link between the receptor activation and tumorigenesis through these key events lacks
29 compelling persuasiveness.
30 The strongest case currently available for a cause-and-effect link is the results of cancer
31 studies using PPAR-a null mice (see Peters et al., 1997, and Ito et al, 2007). When exposed to
32 the peroxisome proliferator WY-14,643, the null mice do not show increased cell proliferation or
33 decreased apoptosis or evidence of developing hepatocellular carcinogenesis (Peters et al., 1997).
34 The occurrence of events known to be associated with tumorigenesis following exposure to
35 peroxisome proliferator chemicals allows an association to be made between PPAR-a activation
36 and these other events. The events clearly specific to PPAR-a receptor activation—actual
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-23 DRAFT-DO NOT CITE OR QUOTE
-------
1 peroxisome proliferation and expression of peroxisomal genes and enzymes—are likely markers
2 for the receptor activation and not cause-and-effect events for carcinogenesis. The other events
3 may be related to tumorigenesis but are not restricted to PPAR-a activation MOA. The TCA
4 metabolite of tetrachloroethylene has been studied in null mice, and although no tumors were
5 observed, this was only a short-term study and findings are only minimally supportive of
6 PPAR-a being related to liver tumorigenesis. Section 4.4.4.3.5 describes the study in greater
7 detail. Tetrachloroethylene has not been studied in such mice. Although some investigators
8 attributed tetrachloroethylene-induced hepatocarcinogenesis to TCA, analysis of the amount of
9 tetrachloroethylene metabolism at the doses administered in the mouse carcinogenicity bioassays
10 demonstrates that not enough TCA is produced to account for the tumor response (see Appendix
11 4A).
12 The evidence for peroxisome proliferation by tetrachloroethylene and the chloroacid
13 metabolites is published in Zhou and Waxman (1998), Zanelli et al. (1996), Bruschi and Bull
14 (1993), Nelson et al. (1989), Elcombe (1985), Elcombe et al. (1985), Goldsworthy and Popp
15 (1987), Odum et al. (1988), Channel et al. (1998), DeAngelo et al. (1989), and Daniel et al.
16 (1993).
17 Studies by Goldsworthy and Popp (1987) indicate that tetrachloroethylene and its
18 metabolite TCA elevate cyanide-insensitive PCO activity in mouse liver, yet only TCA caused
19 increased PCO activity in rat liver. The elevation in PCO activity in mouse liver caused by
20 tetrachloroethylene was not great. In a study conducted by Zanelli et al. (1996), TCA was shown
21 to increase PCO activity in the liver of treated rats of a different strain, so the metabolite does
22 cause the effect, and it may be responsible for the response of the parent compound in mice.
23 Different rat strains were used in the two studies (F344 and Wistar, respectively), and the
24 increase reported by Goldsworthy and Popp was a relatively weak response.
25 Tetrachloroethylene increased both PCO enzyme activity and peroxisome volume density in
26 exposed mice in the study by Goldsworthy and Popp. In the tetrachloroethylene study conducted
27 by Odum et al. (1988), peroxisome proliferation was increased in the livers of mice but not rats.
28 Elcombe (1985) found TCA to cause peroxisome proliferation—as measured by an increase in
29 peroxisomal enzyme activity—in hepatocytes of both rats and mice, both in vivo and in vitro,
30 after short-term exposure. Interestingly, Elcombe reported that the Wistar rat showed a greater
31 peroxisome proliferation response than did mice, as measured by increases in cyanide-insensitive
32 acyl-CoA oxidase activity induction. Clearly, strain and species differences exist.
33 DeAngelo et al. (1989) demonstrated peroxisome proliferation induction by TCA and by
34 DCA exposures in mice and rat livers, as indicated by increased PCO activity and peroxisomal
35 volume and possibly the observed increased carnitine acetyl transferase activity as well. The
36 investigators examined peroxisome proliferation activity in three strains of rats (Sprague-Dawley,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-24 DRAFT-DO NOT CITE OR QUOTE
-------
1 F344, and Osborne-Mendel) and in four strains of mice (Swiss-Webster, C57BL/6, C3H, and
2 B6C3F1). The conclusion from the DeAngelo et al. (1989) study is that mice are more sensitive
3 than rats with respect to the enhancement of liver peroxisome proliferation by TCA. More recent
4 studies conducted by Waxman and colleagues (Maloney and Waxman, 1999; Zhou and Waxman,
5 1998) showed induction of peroxisome proliferation in rodents by the tetrachloroethylene
6 metabolites TCA and DCA. DCA and TCA activated the PPAR-a receptor in both mouse and
7 human cells in these studies. Walgren and colleagues (Walgren et al., 2000a, b, 2004) reported
8 expression of PPAR-a in human hepatocytes, activation by TCA and DCA, and peroxisome
9 proliferation by a series of acetates. High concentrations of TCA were used in the studies.
10
11 4.4.4.3.5. Strength, consistency, specificity of association of the hepatocellular tumor response
12 with key events. Whether or not any cause-and-effect relationship exists between peroxisome
13 proliferation per se and cancer-causing activity leading to rodent liver cancer is not clear
14 (Capone, 1994; Cattley et al., 1998; Cattley and Roberts, 2000; Youssef and Badr, 1999). The
15 current majority opinion regards activation of PPAR-a as the important causal key event—the
16 obligatory step—for the MOA of rodent liver carcinogenesis. Even so, not all scientists agree,
17 and not all data support this hypothesis (see Section 4.4.4.1).
18 The strongest support for a causal relationship between the PPAR-a activation MOA and
19 hepatocellular tumorigenesis using the compound WY-14643 is from studies in null mice by
20 Peters et al. (1997a, b). The existing data show null mice to be refractive to other possible key
21 events—suppression of apoptosis and cell proliferation—and also refractive to tumor formation
22 following 11 months of exposure to this prototype peroxisome proliferator. The null mouse,
23 when challenged by 11 months of exposure to WY-14643, did not respond to a dose that causes
24 100% tumor response in wild-type mice.
25 Results from null mouse studies would be convincing data in support of the PPAR-a
26 MOA hypothesis, except for serious shortcomings. For example, the studies are less-than-
27 lifetime studies, or they are in vitro studies conducted in tissues from animals exposed to test
28 compounds. Such studies clearly cannot be considered equal to the standard rodent lifetime
29 bioassays conducted to detect carcinogenic activity, and because of this deficiency, they are not
30 considered adequate for assessing lifetime cancer risk. Also, they cannot be used to demonstrate
31 conclusively that PPAR-a activation is an obligatory step in rodent hepatocellular tumorigenesis
32 simply because some of the key events that could be associated with tumorigenesis are not
33 observed. The complexity of the multitude of effects and the lack of understanding about which
34 of the myriad downstream events result at particular dose levels—mechanisms and steps linking
35 those events to PPAR-a activation—also render the null mouse data somewhat less than
36 adequate for understanding the PPAR-a activation relationship to tumorigenesis. Additionally,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-25 DRAFT-DO NOT CITE OR QUOTE
-------
1 the null mice are innately dissimilar from wild mice—for example, they respond to stress
2 differently (Watanabe et al., 2000; Huss and Kelly, 2005). Their differences include both
3 physiological and biochemical aspects. Some are likely related to PPAR-a-dependent changes in
4 gene expression, but others are not PPAR-a dependent (Valles et al., 2003; Jalouli et al., 2003;
5 Hasmall et al., 2002; Meyer et al, 2003). Because of such inherent differences between null mice
6 and wild mice prior to exposure to test chemicals, the data from studies using these mice should
7 be interpreted with caution.
8 The most relevant data pertinent to tetrachloroethylene comes from studying TCA.
9 Laughter et al. (2004) attempted to determine whether effects of TCA in the liver associated with
10 carcinogenesis were mediated by PPAR-a. Male wild-type and PPAR-a-null mice were given
11 TCA at 0.25, 0.5, 1, or 2g/L in the drinking water for 7 days. TCA increased liver-to-body
12 weight ratios, but the increases were not significant. The livers from wild-type but not PPAR-a-
13 null mice exposed to 2 g/L TCA exhibited centrilobular hepatocyte hypertrophy. Further, global
14 gene expression was assessed using the mouse Atlas cancer 1.2 array. The induction of CYP4a
15 and acyl-CoA oxidase was examined in the livers of mice after exposure to TCA. In wild-type
16 mice, but not PPAR-a-null mice, CYP4a was induced at Ig TCA/L and above, Aacyl-CoA was
17 induced by TCA at 2g/L, and palmitoyl-CoA oxidase activity was induced at 2 g/L. These data
18 suggest that peroxisome proliferation induced by such compounds could be potentially mediated
19 by PPAR-a (Nakajima et al., 2000). They do not indicate a cause-and-effect relationship
20 between PPAR-a and liver tumorigenesis, however. These studies of TCA were not designed to
21 examine tumor development or show any evidence for a cause-and-effect relationship between
22 receptor activation and tumor development. No comparable study exists for tetrachloroethylene.
23 If peroxisome proliferation is causally related to the induction of liver cancer, then a
24 detectable quantitative relationship between the two events could be expected. That is, potent
25 peroxisome proliferators should also be potent hepatocarcinogens. However, this does not
26 appear to be the case (Elcombe and Mitchell, 1986; Marsman et al., 1988, 1992; Eacho et al.,
27 1991; U.S. EPA, 1991a). A comparison of the tetrachloroethylene chloroacid metabolites
28 indicates that DCA is a more potent hepatocarcinogen than TCA. For example, in a chronic
29 65-week study of DCA and TCA in male B6 mice, Herren-Freund et al. (1987) found that equal
30 concentrations in drinking water resulted in a nearly threefold higher incidence of liver cancer in
31 DCA-dosed animals than in TCA-dosed animals. DeAngelo et al. (1989), however, reported that
32 TCA was more potent than DCA as a peroxisome proliferator in male B6 mice. Nelson et al.
33 (1989) also reported that TCA produced greater peroxisome proliferation than did DCA in B6
34 mice dosed for only 10 days. Additionally, evaluation of TCA and DCA indicates that these two
35 metabolites act through different MO As because they exhibit clearly unparallel dose-response
36 curves.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-26 DRAFT-DO NOT CITE OR QUOTE
-------
1 Compared with oxidative damage, hepatomegaly caused by cell proliferation appears to
2 be better correlated with hepatocarcinogenesis in rodents (Marsman et al., 1988, 1992; Barrass et
3 al., 1993; Cattley et al., 1998). A reversible increase is seen in cell death along with the
4 increased cell proliferation (Bursch et al., 1984; Roberts et al., 1995; Marsman et al., 1992;
5 Cattley et al., 1998). Both TCA and DCA can produce hepatomegaly at carcinogenic doses,
6 although, at least in the case of DCA, the increase is more likely due to cytomegaly (i.e., an
7 increase in cell size), whereas in the case of TCA it is more likely due to an increase in the
8 number of liver cells (Bull, 2000).
9 Tetrachloroethylene and its major oxidative metabolite, TCA, cause liver tumors in mice,
10 yet do not induce liver tumors in rats (NCI, 1977; NTP, 1986a; DeAngelo et al., 1997). TCA has
11 been demonstrated, however, to produce hepatic peroxisome proliferation in rats as well as in
12 mice (Elcombe, 1985; Zanelli et al., 1996; DeAngelo et al., 1989).
13 Tetrachloroethylene is not as potent a peroxisome proliferator as are its metabolites. This
14 information is meaningful because some investigators have postulated that although
15 tetrachloroethylene may possess an intrinsic ability to induce peroxisomes, it may be less
16 effective as a peroxisome proliferator and carcinogen in the rat due to a metabolic inability of
17 that species to form sufficient amounts of peroxisome proliferator metabolites, whereas TCA is
18 formed in mice in sufficient amounts from the parent compound bioassay doses to result in a
19 sustained level of peroxisome proliferation. Carcinogenicity bioassay studies in rats disprove
20 that theory because TCA does not cause liver tumors in rats.
21 The hepatocellular cancer-causing activity of tetrachloroethylene has not been heavily
22 associated with DCA. DCA is thought not to be produced in the liver in sufficient quantity
23 because its only source from tetrachloroethylene oxidative metabolism is the further
24 biotransformation of TCA. TCA, on the other hand, is generally considered to contribute to the
25 mode of carcinogenic action for tetrachloroethylene. Because DCA may be somewhat rapidly
26 further metabolized to other compounds, such a conclusion may be in error.
27 It is important to note that the peroxisome proliferation effects observed in rodents
28 exposed to tetrachloroethylene are equivocal. In a summary plot of tumor incidence versus
29 peroxisome proliferation from the two studies reporting tetrachloroethylene data, the effects are
30 significant in male mice but not in females (NCI, 1977; NTP, 1986a; Ashby et al., 1994;
31 Goldsworthy and Popp, 1987; Odum et al., 1988). Compared with the effects of potent
32 peroxisome proliferator chemicals, the effects caused by tetrachloroethylene are relatively weak
33 (Ashby etal., 1994).
34 PPAR-a isolated from mouse liver can be activated by certain tetrachloroethylene
35 chloroacid metabolites. Both the TCA and the DCA metabolites have been shown to activate
36 PPAR-a (Maloney and Waxman, 1999).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-27 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.4.4.3.6. Dose-response relationship. The data for a clear dose-response for increased
2 peroxisome proliferation in mouse liver resulting from tetrachloroethylene exposure are
3 equivocal, especially in female mice (Section 4.4.4.3.5). Some evidence exists for increased
4 tumor response with increased tetrachloroethylene dose in both gavage and inhalation
5 carcinogenicity bioassays, although the dose/concentration levels in those studies were relatively
6 high. Approaching saturation of metabolism blurs the dose-response at these levels, most
7 noticeably in the oral gavage study. There is also some evidence, from studies conducted
8 separately from the cancer bioassays, to support increase in tetrachloroethylene peroxisome
9 proliferation with increase in dose, although these data are not particularly convincing (Ashby et
10 al., 1994).
11 Likewise, positive dose-responses for cancer-causing activity in lifetime carcinogenicity
12 bioassays and for peroxisome proliferation in short-term studies have been observed in mice
13 following exposure to relatively high doses of TCA (Ashby et al., 1994; Daniel et al., 1993;
14 Goldsworthy and Popp, 1987; Elcombe, 1985; Herren-Freund et al., 1987).
15 Other potentially adverse effects associated with DCA exposure (e.g., changes in
16 carbohydrate metabolism, as well as other alterations in cell signaling) are expected to occur, in
17 some cases, at lower doses than are required for peroxisome proliferation (Bull, 2000), indicating
18 occurrence of events possibly associated with tumorigenesis at doses below those causing the
19 peroxisome proliferation response.
20
21 4.4.4.3.7. Temporal association. Increases in peroxisome volume density as well as in marker
22 cyanide-insensitive (PCO) oxidation indicative of peroxisome proliferation have been shown to
23 occur following a few days of treatment with tetrachloroethylene, DCA, or TCA (Goldsworthy
24 and Popp, 1987; Elcombe, 1985; Odum et al., 1988; Daniel et al., 1993). On the other hand,
25 following DCA treatment, DNA SSB has been observed prior to peroxisome proliferation
26 (Nelson and Bull, 1988; Nelson et al., 1989), although investigators using a different
27 methodology did not observe the DCA-induced SSB (Chang et al., 1992) after DCA treatment.
28
29 4.4.4.3.8. Species similarities and differences: human evidence. The relevance to humans of
30 rodent hepatocellular carcinomas thought to be induced specifically by peroxisome proliferator
31 chemicals has been questioned. Humans do have a functional PPAR receptor (Sher et al., 1993),
32 which is comparable to PPAR receptors of mice and rats in its affinity for PPAR-a ligands
33 (Klaunig et al., 2003) and it is capable of activating many of the genes regulated in the mouse by
34 PPAR-a (Yu et al., 2001).
35 PPAR-a, the PPAR subtype considered to be the causal factor for peroxisome
36 proliferation in rodent hepatocytes, has been found in tissue from several species, including mice,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-28 DRAFT-DO NOT CITE OR QUOTE
-------
1 rats, and humans as well as dogs, guinea pigs, hamsters, and nonhuman primates (Yousef and
2 Badr, 1999; Schultz et al., 1999; Lake et al., 1993; Roberts et al., 2000; Reddy et al., 1984;
3 Graham et al., 1994; Kurata et al., 1998). Humans express PPAR-a in liver (Auboeuf et al.,
4 1997), although reportedly to a lesser extent than do rats and mice (Klaunig et al., 2003).
5 PPAR-a mRNA in human liver samples have been reported by some investigators to be one
6 order of magnitude lower than those observed in mice (Palmer et al., 1998; Tugwood et al.,
7 1996).
8 Only a few human liver samples have been examined for quantification of PPAR-a
9 transcription factors (Klaunig et al., 2003), and one study by Walgren et al. (2000a) reported that
10 one of six human samples was equivalent to mice in expression of PPAR-a protein. Although
11 the number of human liver samples examined is limited, evidence exists for mutations in
12 PPAR-a (Flavell et al., 2000; Sapone et al., 2000; Vohl et al., 2000; Yamakawa-Kobayashi et al.,
13 2002), which could contribute to the large variations in PPAR-a levels (Walgren et al., 2000a).
14 Intel-individual variability (Tugwood et al., 1996), along with the inducibility of PPAR-a
15 expression by chemicals and other factors (Sterchele et al., 1996), indicate the likelihood of a
16 susceptible subpopulation (Heuvel, 1999).
17 Although some studies have shown no increase in DNA synthesis in primary human
18 hepatocytes following treatment with several peroxisome proliferator chemicals, other evidence
19 indicates that humans may indeed be responsive to adverse effects of peroxisome proliferators.
20 For example, investigations of human hepatocytes following treatment with certain fibrate
21 chemotherapeutic agents found dose-dependent induction of acyl-CoA oxidase activity and, in
22 one case, increased peroxisome density (Cimini et al., 2000; Perrone et al., 1998). Increases of
23 liver peroxisomes have been reported in human patients taking the hypolipidemic therapeutic
24 agents clofibrate and ciprofibrate (Hanefeld et al., 1983; Bentley et al., 1993; Hinton et al., 1986).
25 The increases in volume density of peroxisomes (23-30%) are comparable to or greater than
26 those observed in rodents exposed to tetrachloroethylene. Also, both humans and rodents
27 respond to peroxisome proliferators with reduction of serum lipids, indicating similar capabilities
28 for modification of gene expression. Epidemiologic evidence of cancer from exposure to
29 peroxisome proliferator chemicals is limited to only a few studies in patients taking fibrate drugs,
30 and is inconclusive (Newman and Hulley, 1996; Melnick, 2001).
31 Taking into account kinetic and dynamic factors, the proposed animal MOA is plausible
32 in humans. If peroxisome proliferation is involved in a mode of carcinogenic action for
33 tetrachloroethylene, the cancer-causing activity cannot be dismissed for humans, especially since
34 PPAR-a has been identified in human liver, and both TCA and DCA metabolites similarly
35 activate human as well as mouse PPAR-a, even if to different degrees. Chemical-specific data
36 regarding the ability of the major tetrachloroethylene metabolite TCA to activate PPAR-a in
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-29 DRAFT-DO NOT CITE OR QUOTE
-------
1 humans as well as in mice indicate cross-species relevance (Maloney and Waxman, 1999),
2 although there exists a quantitative difference between these two species in hepatic PPAR-a
3 activation (Walgren et al., 2000a, b). Site concordance is not a requirement for extrapolation of
4 tumorigenesis in animal models to the human situation, however, and PPAR-a is found in
5 several different organs. It is highly expressed in cells having active fatty acid capacity, such as
6 hepatocytes in liver, and also in renal proximal tubule cells, cardiomyocytes, and enterocytes.
7 Some PPAR-a agonists cause tumors in rodents at sites other than the liver.
8 PPAR-a is proposed to be involved in causing Leydig cell tumors. The human relevance
9 of PPAR-a induction of Leydig cell tumors is not so controversial. The understanding of the
10 science is not good enough to explain why humans have a functional PPAR-a capable of gene
11 expression modulation in a manner similar to that of rodents but does not respond similarly.
12 PPAR-a in humans may be capable of modulating lipid homeostasis through alteration of
13 expression of other, different, genes in the liver and genes in other target organs that express
14 higher levels of PPAR-a. Similar events occurring in various tissues and organs could lead to
15 carcinogenesis in those tissues and organs; therefore, target organs could differ among species.
16 The liver carcinogenesis in mice could be a red flag for tumorigenesis at some other site in
17 humans. This is the case for other chemicals, such as arsenic.
18 A different PPAR receptor also involved in lipid homeostasis may be involved in human
19 liver cancer, because cross-talk is known to occur among these receptors. Glinghammar et al.
20 (2003) found that PPAR delta (PPAR-5) receptor activation in human hepatocellular carcinoma
21 cells induced COX2 expression, a factor associated with carcinogenesis, and increased cellular
22 proliferation. Such results suggest a potential role for PPAR-5 in human hepatocellular
23 carcinoma induction. PPAR-a also can mediate mRNA alterations associated with prostaglandin
24 synthesis in rodents, i.e., COX2 expression (Peters and Vanden Heuvel, 2002). PPAR gamma
25 has been shown to induce COX2 expression as well. Because cross-talk is known to occur
26 among the PPAR receptors, the potential exists for cross-talk to be involved in receptor
27 activation related to carcinogenicity.
28
29 4.4.4.3.9. Biological plausibility and coherence of the database. Tetrachloroethylene induces
30 liver cancer in treated mice, but it has not been shown to cause liver cancer in treated rats; thus,
31 an inconsistency exists between rodent species. Epidemiologic evidence is insufficient for
32 determining whether tetrachloroethylene causes liver cancer in humans (see Section 4.3.1.2).
33 Inadequate amounts of TCA are produced from tetrachloroethylene metabolism to
34 account totally for the liver tumor induction observed in bioassays (Appendix 4A). DCA is less
35 likely to contribute to tetrachloroethylene hepatocarcinogenesis because relatively small amounts
36 would be produced in the liver target organ from metabolism of tetrachloroethylene. Limited
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-30 DRAFT-DO NOT CITE OR QUOTE
-------
1 data on tumor phenotype indicate that tetrachloroethylene tumors differ phenotypically from
2 TCA tumors or DCA tumors (Maronpot et al., 1995; Anna et al., 1994).
3 An important distinction needs to be drawn between biomarkers that are causally related
4 to carcinogenic activity and those that are merely correlative. There is much confusion and
5 debate over the reliability of hepatic peroxisome proliferation as a marker for
6 hepatocarcinogenesis. As reviewed in Section 4.4.4.1, some scientists maintain that there is a
7 causal relationship between peroxisome proliferation and hepatocarcinogenesis, whereas others
8 have questioned the validity of this relationship and suggest that the degree of peroxisome
9 proliferation correlates poorly with relative hepatocarcinogenic effectiveness and potency.
10 Although many of the responses generally observed in the overall peroxisome proliferation
11 phenomena are often manifested in tumorigenesis, causality is uncertain, and in the specific case
12 of tetrachloroethylene, the data are especially limited.
13 Several recent studies have expanded the scientific understanding of the PPAR-a mode of
14 action (see Caldwell et al., 2008). First, Yang et al. (2007) demonstrated that PPAR-a activation
15 in hepatocytes induces peroxisome proliferation but not liver tumors. The approach entailed
16 targeting expression of PPAR-a to hepatocytes by placing the VP16 PPAR-a transgene gene
17 under control of the liver enriched activator protein (LAP) promoter. LAP-VP16 PPAR-a
18 transgenic mice showed a number of PPAR-a-mediated effects: decreased serum triglycerides
19 and free fatty acids, peroxisome proliferation, enhanced hepatocyte proliferation, and induction
20 of cell-cycle and PPAR-a target genes. However, compared with wild-type mice exposed to
21 Wy-14,643, the extent of hepatomegaly was reduced and no hypertrophy or eosinophilic
22 cytoplasms was seen in LAP-VP16 PPAR-a mice. Also in contrast with wild-type mice exposed
23 to Wy-14,643, no evidence of non-parenchymal cell proliferation was observed in the LAP-
24 VP16 PPAR-a transgenic mice. Moreover, at one year of age no evidence of preneoplastic
25 hepatic lesions or hepatocellular neoplasia was observed in LAP-VP16 PPAR-a transgenic mice.
26 As noted by the authors, PPAR-a activation only in mouse hepatocytes is sufficient to induce
27 peroxisome proliferation and hepatocyte proliferation but"... is not sufficient to induce liver
28 tumors."
29 Secondly, Ito et al. (2007) found that di (2-ethylhexyl)phthalate (DEHP), a proposed
30 robust example of PPAR-a agonism-induced hepatocarcinogenesis, yields liver tumors in a
31 2-year study in PPAR-a knock-out mice. This study demonstrates the limitations, cited by the
32 FIFRA SAP, of drawing conclusions from the one-year bioassay of high doses of WY-14,643
33 referenced above (e.g., Peters, 1997). It supports the view that knock-out mouse bioassays
34 should be carefully characterized and conducted for 2 years to assess whether PPAR-a activation
35 is indeed necessary for induction of liver cancer.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-31 DRAFT-DO NOT CITE OR QUOTE
-------
1 In summary, limited evidence supports the hypothesis that tetrachloroethylene tumor
2 induction could be related to PPAR-a activation, but critical review of the scientific literature
3 reveals significant data gaps regarding the relationship between the PPAR-a activation and
4 neoplasia induced by tetrachloroethylene. If PPAR-a does play a role in tetrachloroethylene-
5 induced tumorigenesis, available information suggests relevance to humans cannot be ruled out.
6
7 4.4.4.4. Effects That Could Be Related to Other Potential Modes of Action
8 4.4.4.4.1. Mutagenicity and genotoxic effects. The available evidence is inconclusive regarding
9 mutagenicity of tetrachloroethylene or its metabolites and hepatocarcinogenesis (see Section 4.3).
10 Tetrachloroethylene induces SSB and DNA binding in liver tissue, and existing data implicate a
11 potential role for genotoxic effects of certain metabolites, such as DCA and the proposed
12 intermediate chloral hydrate; the epoxide tetrachloroethylene oxide is a bacterial mutagen.
13 Interestingly, and the phenotype and frequency of tumors produced by DCA and
14 tetrachloroethylene tumors differ (Bull, 2000; Anna et al., 1994; Maronpot et al., 1995; Moore
15 and Harrington-Brock, 2000; also see Section 4.4.4.2). The mutagenic potential of several
16 metabolites has not been studied. Not all of the P450 metabolites, including the unstable,
17 potentially reactive intermediates, such as trichloroacetyl chloride, for example, have been
18 sufficiently tested in the standard genotoxicity screening battery.
19
20 4.4.4.4.2. Immunosuppressive effects. Although tetrachloroethylene-specific data are lacking,
21 it is possible that inhibition of the natural immune surveillance could be related to
22 hepatocarcinogenic properties of tetrachloroethylene (see also Section 4.8.3). Immune
23 suppression could play a role in the induction of cancer as many immunosuppressive agents are
24 human carcinogens (Tomatis et al., 1989). Exposure to organic solvents has been generally
25 associated with autoimmune diseases such as scleroderma (Nietert et al., 1998). A strong
26 association has been reported between exposure to solvents structurally similar to
27 tetrachloroethylene and systemic sclerosis in patients who have autoantibodies (Nietert et al.,
28 1998).
29 Binding of reactive compounds to cellular macromolecules has been proposed as an
30 important step in the pathogenesis of several diseases, both for cancer (Hinson and Roberts,
31 1992) and for chemically induced autoimmune disease (Uetrecht et al., 1988). Reactive
32 metabolites of tetrachloroethylene have been shown to bind irreversibly to cellular
33 macromolecules in vitro (e.g., Costa and Ivanetich, 1980) and in vivo (Pegg et al., 1979;
34 Schumann et al., 1980). Binding occurs proportionally to the amount metabolized, and
35 metabolism is proportional to toxicity (e.g., Buben and O'Flaherty, 1985). Several published
36 studies have demonstrated formation of trichloroacylated protein adducts, for example, in liver
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-32 DRAFT-DO NOT CITE OR QUOTE
-------
1 and kidney of rats (Birner et al., 1994) and in plasma of rats and humans (Pahler et al., 1999)
2 following exposures to tetrachloroethylene. Another example is the detection of
3 trichloroacetylated protein adducts formed in mice treated with tetrachloroethylene (Green et al.,
4 2001). Further studies designed to identify the adducted proteins may help to elucidate an MOA
5 for tetrachloroethylene-induced autoimmune response, which, in turn, may be related to cancer-
6 causing activity.
7
8 4.4.4.4.3. Effects on the insulin receptor/glucose metabolism, Tumorigenesis is associated
9 with changes in enzymes involved in carbohydrate metabolism (Ahn et al., 1992) and tumor cells
10 depend uniquely on glucose as an energy source. Transformation of many cell types is
11 associated with an increase in glucose metabolism including increases in important glycolytic
12 enzyme activities (Baggetto, 1992), and often these enzyme activities correlate with malignancy
13 (Weber and Lea, 1966; Harap, 1975). Ebrahim et al. (1996) observed tetrachloroethylene-
14 induced alterations in glycolytic and gluconeogenic enzymes in liver and kidney of mice treated
15 with the chemical. Administration of 2-deoxy-D-glucose and vitamin E controlled the changes
16 in glycolytic and gluconeogenic enzymes induced by tetrachloroethylene.
17
18 4.4.4.4.4. Alteration in DNA methylation. No tetrachloroethylene-specific data are available
19 regarding a role of alteration in DNA methylation in tumorigenesis. Such changes are reported
20 to be a common early molecular event in most tumors (Zingg and Jones, 1997; Baylin et al.,
21 1998). Alterations in DNA methylation following exposure to chemicals, both hypomethylation
22 and hypermethylation, may be factors in tetrachloroethylene-induced tumorigenesis. Although
23 tetrachloroethylene-specific data are lacking, its metabolites DCA and TCA are known to induce
24 hypomethylation of DNA and protooncogenes in mouse liver.
25
26 4.4.4.4.5. Alterations in cell replication and death rate. Although modification of cell
27 replication and death rates may be important to tetrachloroethylene liver tumorigenesis, no
28 tetrachloroethylene-specific data are available.
29
30 4.4.4.4.6. Cytotoxicity and compensatory hyperplasia. Cytotoxicity and reparative hyperplasia
31 are not marked findings resulting from tetrachloroethylene exposures capable of causing liver
32 cancer in mice.
33
34 4.4.4.4.7. Hepatomegaly/cytomegaly. Increase in liver size is highly correlated with liver
35 tumorigenesis in mice. Treatment with tetrachloroethylene can lead to increased liver weight. In
36 carcinogenicity studies, hepatomegaly occurred following exposures in the dose ranges that
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-33 DRAFT-DO NOT CITE OR QUOTE
-------
1 cause liver tumors and at experimental exposures well below the carcinogenicity bioassay dose
2 levels. It is not clear exactly how the phenomenon is related to tumorigenesis in the case of
3 tetrachloroethylene.
4
5 4.4.4.5. Summary and Conclusions
6 At the present time, the specific mechanisms or MOA for tetrachloroethylene-induced
7 hepatocarcinogenesis in mice are not known. The cancer-causing activity of tetrachloroethylene
8 is thought to be metabolism-dependent, however the specific contribution of particular
9 metabolites has not been elucidated. TCA induces liver tumors with different phenotype and is
10 not produced in sufficient amounts to account quantitatively for the liver tumor response
11 observed with tetrachloroethylene. Not all metabolites have been identified or characterized, but
12 several known metabolites including those derived from P450 as well as GSH pathways are
13 clearly mutagenic in the standard battery of tests. Tetrachloroethylene is mutagenic in bacterial
14 assays in the presence of GST and GSH whereas the standard S9 fraction has typically yielded
15 negative results. Tetrachloroethylene at higher concentrations also induces SSBs and DNA
16 binding in liver tissue. The metabolite DC A is the most potent mutagen of the P450-derived
17 metabolites, exhibiting mutagenic activity in a number of assays. A putative P450 derived
18 metabolite, 1,1,2,2-tetrachloroethylene oxide, is also mutagenic; the mutagenicity of this epoxide
19 would be predicted from structure-activity relationships. Given the demonstrated mutagenicity
20 of several tetrachloroethylene metabolites, it is expected that mutagenicity contributes to the
21 MOA for tetrachloroethylene carcinogenesis, although the specific metabolic species or
22 mechanistic effects are not known.
23 Chemical-specific data for PPAR-a activation are limited but suggest that this is not the
24 primary MOA for hepatocarcinogenesis. The existing data for tetrachloroethylene show minimal
25 peroxisome proliferator activity, and no chemical-specific data correlate peroxisome
26 proliferation with tumor induction for tetrachloroethylene. As noted above, TCA produces
27 tumors of a different phenotype than tetrachloroethylene and TCA is not produced in sufficient
28 amounts to account quantitatively for the tetrachloroethylene liver tumor response. Moreover,
29 several recent studies have expanded the scientific understanding of the PPAR-a mode of action
30 (see Caldwell et al., 2008) including the demonstration that PPAR-a activation in hepatocytes
31 induces peroxisome proliferation but not liver tumors (Yang et al., 2007). In particular,
32 peroxisome proliferation and hepatocyte proliferation, but not liver tumors, were observed with
33 PPAR-a activation in mouse hepatocytes. Furthermore, Ito et al. (2007) found that DEHP, a
34 proposed robust example of PPAR-a agonism-induced hepatocarcinogenesis, yields liver tumors
35 in a 2-year study in PPAR-a knock-out mice.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-34 DRAFT-DO NOT CITE OR QUOTE
-------
1 In summary, the MOA for tetrachloroethylene-induced liver toxicity and tumorigenesis is
2 not understood. Data are lacking particularly for tetrachloroethylene P450 intermediates that
3 could be involved in mutagenicity and carcinogenicity of the parent compound. Among the data
4 gaps is the incomplete characterization of the metabolites in tests beyond the standard battery of
5 genotoxicity tests, including on important genetic and epigenetic endpoints.
6
7 4.5. KIDNEY TOXICITY
8 4.5.1. Human Studies
9 4.5.1.1. Kidney Toxicity in Humans
10 High concentrations of inhaled tetrachloroethylene given acutely as an anaesthetic are
11 associated with symptoms of renal dysfunction, including proteinuria and hematuria (Hake and
12 Stewart, 1977, AT SDR, 1997). Controlled inhalation exposure to tetrachloroethylene at levels of
13 0, 20, 100, or 150 ppm for up to 11 weeks did not affect a number of urine parameters or BUN (a
14 measure of kidney function) in 12 healthy individuals (Stewart et al., 1977, as reported in
15 ATSDR, 1997). Whether renal effects would occur from these acute exposure levels in a larger,
16 more diverse population than the one studied by Stewart et al. (1977) is not known.
17 The evidence for kidney effects from chronic inhalation of tetrachloroethylene is limited
18 because many of the available reports do not include information on even a minimal core battery
19 of tests for kidney function. The ATSDR (Amler et al., 1998; Lybarger et al., 1999)
20 recommends a core battery of kidney function tests that includes serum creatinine, urinalysis
21 with microscopic examination of urine sediment, albumin, retinol binding protein (RBP),
22 N-acetyl-p-D-glucosaminidase (NAG), alanine aminopeptidase (AAP), osmolality, and urine
23 creatinine (Lybarger et al., 1999). These indicators evaluate a range of toxicity, from effects on
24 general kidney function to effects on specific segments of the nephron. For example, the overall
25 integrity of the nephron can be evaluated from the urinalysis, and albumin is an indicator of the
26 integrity of the glomerulus; three indicators—RBP, NAG, and AAP—assess damage to the
27 proximal tubules (Lybarger et al., 1999). The proximal tubules house beta lyase enzymes and
28 are hypothesized to be a target of tetrachloroethylene toxicity due to the bioactivation of reactive
29 metabolites produced from the further metabolism of TCVC (see Section 4.2). For this reason,
30 this analysis places greater weight on urinary indicators of proximal tubule function.
31 The epidemiologic studies are suggestive of subtle damage to the renal tubules. Five
32 studies (Trevisan et al., 2000; Verplanke et al., 1999; Mutti et al., 1992; Solet and Robins, 1991;
33 Lauwerys et al., 1983) have examined the three core indicators of tubule function—RBP, NAG,
34 or AAP—in urine of dry cleaners. Three studies measured RBP, with two of the studies
35 reporting a statistically significant elevated prevalence of abnormal values among study
36 participants (Mutti et al., 1992) or a statistically significant elevated geometric mean
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-35 DRAFT-DO NOT CITE OR QUOTE
-------
1 concentration of RBP (Verplanke et al., 1999) for tetrachloroethylene-exposed workers as
2 compared with controls. The mean concentration of RBP for exposed subjects (75.4 ug/g
3 creatinine) in the Verplanke et al. (1999) study is within a normal range,1 indicating the absence
4 of concurrent tubul e toxi city.
5 As a comparison, Nomiyama et al. (1992) suggest a critical level of RBG of 200 ug/g
6 creatinine as indicative of cadmium-induced kidney toxicity. Exposure levels were to a median
7 of 15 ppm (range: limit of detection to 85 ppm) in Mutti et al. (1992) and 1.2 ppm (range:
8 0.3-6.5 ppm) in Verplanke et al. (1999). Lauwerys et al. (1983), the only other study to assess
9 RBP, did not observe any differences in the geometric mean concentration of RBP between dry
10 cleaners with mean tetrachloroethylene exposure of 21 ppm and their controls; however, this
11 study contained fewer exposed subjects with a shorter duration of exposure than did that of Mutti
12 etal. (1992).
13 The four studies that measured urinary excretion of NAG (Solet and Robins, 1991; Mutti
14 et al., 1992; Verplanke et al., 1999; Trevisan et al., 2000) and the one study that measured AAP
15 (Verplanke et al., 1999) did not observe any differences between exposed subjects and controls.
16 These findings are not surprising; NAG is not a sensitive and specific marker of tubular
17 dysfunction (Lybarger et al., 1999). Mean exposures were 14 ppm in Solet and Robins (1991)
18 and 9 ppm in Trevisan et al. (2000); both studies assessed exposure from personal monitoring of
19 exhaled breath.
20 The above findings are further supported by the observation of elevated urinary excretion
21 of other proteins that are also indicators of damage to the proximal tubules: beta2-microglobulin,
22 intestinal alkaline phosphatase (LAP), tissue non-specific alkaline phosphatase (TNAP),
23 lysozyme, beta2-glucuronidase, and glutamine synthetase. Both IAP and TNAP are indicators of
24 proximal tubule brush border integrity (Price et al., 1996), whereas lysozyme and
25 beta2-microglobulin indicate a failure of the tubule to reabsorb protein (Lybarger et al., 1999;
26 Bernard and Lauwerys, 1995; Kok et al., 1998). Glutamine synthetase is a mitochondrial
27 enzyme located in the proximal tubules and has been recently suggested as a marker of tubular
28 damage in rats exposed to 1,3-hexachlorobutadiene (Trevisan et al., 1999).
29 Mutti et al. (1992) observed an elevated prevalence of abnormal values for
30 beta2-microglobulin and brush border antigens, a higher geometric mean concentration of brush
31 border antigens in urine, and a higher concentration of TNAP in urine among 50 exposed dry
32 cleaners as compared with 50 blood donors matched by sex and age with the exposed subjects.
33 Furthermore, markers of renal damage were highly predictive of exposure status in discriminant
1 Lapsley et al. (1998) found a median and an upper 98% confidence limit of 67 and 143 ug/g creatinine,
respectively, in a survey of 70 adults, and this range closely matches the findings of Topping et al. (1986), who
observed a mean and a 98% upper limit of 64 and 185 ug/g creatine in 118 subjects.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-36 DRAFT-DO NOT CITE OR QUOTE
-------
1 analysis. Beta2-microglobulin, however, was not elevated among exposed subjects as compared
2 with controls in the other two studies that examined this protein (Lauwerys et al., 1983; Vyskocil
3 et al., 1990), although the mean concentration of beta2-microglobulin appeared higher in subjects
4 studied by Vyskocil et al. than the mean concentration in controls. Both these studies contained
5 fewer numbers of exposed subjects than did the study by Mutti et al. (1992), and reduced power
6 as a consequence of fewer subjects may be a reason for the null observations. Further,
7 tetrachloroethylene exposure appears to affect reabsorption in the renal tubules. Two studies that
8 assessed lysozyme or beta-glucuronidase observed a statistically significant elevated mean
9 concentration of these proteins among dry cleaners as compared with controls (Franchini et al.,
10 1983; Vyskocil et al., 1990).
11 It is not clear whether tetrachloroethylene exposure engenders an effect on other parts of
12 the kidney. The study by Mutti et al. (1992) is suggestive of damage to the glomerulus; however,
13 the lack of an elevated excretion of albumin, an indicator of glomerular function (Lybarger et al.,
14 1999), in the study by Verplanke et al. (1999) argues for further assessment of possible
15 glomerular effects.
16 Taken together, the epidemiologic studies support an inference of subtle effects on the
17 renal proximal tubules. Effects are seen in populations of both males and females, and potential
18 differences in susceptibility due to sex-related differences in rates of metabolism (see Section
19 4.2) cannot be determined from the available evidence. Median exposure levels in the studies
20 that observed alterations in renal enzymes were 9 ppm (Trevisan et al., 2000), 10 ppm (Franchini
21 et al., 1983), and 15 ppm (Mutti et al., 1992), representing LOAELs for these studies. Only the
22 study by Trevisan et al. (2000) observed an exposure-response relationship, a correlation
23 between urinary tetrachloroethylene and the concentration of glutatmine synthatase (p < 0.001).
24 None of the other studies reported exposure-response relationships, which is a limitation on the
25 inference of an association between tetrachloroethylene and renal damage. However, as pointed
26 out by Mutti et al. (1992), this is a common finding among solvent-exposed populations, and
27 inadequate definition of the dose metric most likely contributes to the null finding. Table 4-3
28 summarizes the human kidney toxicity studies.
29
30 4.5.1.2. Kidney Cancer
31 The evidence supporting a hypothesis of an association between tetrachloroethylene
32 exposure and kidney cancer consists of the observation of elevated risks in the larger case-
33 control studies (Asal et al., 1988; Aschengrau et al., 1993; Dosemeci et al., 1999; Mandel et al.,
34 1995; McCredie and Stewart, 1993; Mellemgaard et al., 1994; Schlehofer et al., 1995; Partanen
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-37 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-3. Summary of human kidney toxicity marker studies in dry cleaners
Author
Mutti et
al. (1992)
Verplanke
etal.
(1999)
Solet and
Robins
(1991)
Lauwreys
etal.
(1983)
Vyskocil
etal.
(1990)
Franchini
etal.
(1983)
Trevisan
etal.
(1999)
Number
of subjects
50 dry cleaners and 50
matched controls
82 exposed and 19
nonexposed dry cleaners
192 exposed dry cleaners,
no control group
24 exposed dry cleaners
and 33 unexposed controls
16 exposed dry cleaners
and 13 nonoccupationally
unexposed controls
57 exposed dry cleaners
and (a)
81 unexposed and (b) 80
unexposed factory workers
40 dry cleaners
and 45 laundry workers
Estimated mean
exposure
concentration"
(ppm)/duration
(yrs)
15/10
PM
1.2 (TWA)/
5.1
PM
14/11.6
PM
21/6.4
PM
23 (TWA)/9
U
10 (TWA)2/14
PM
8.8/15
PM, B, U
RBP
D
D
NS
NAG
NS
NS
NS
AAP
NS
Beta2-
micro-
globulin
D
NS
NS
Albumin
D
NS
NS
NS
Total
protein
NS
NS
NS
NS
Lyso-
zyme
D
D
TNAP
NS
Beta-
glucuro-
nidase
D
Gluta-
mine
synthe-tase
Dose-
response
observed
a TWA tetrachloroethylene concentration developed using the relationship of between excretion of TCA in urine and breathing-zone, TWA concentration of
Ikeda and Otsuji (1972).
B = Biological monitoring of blood
D = Detected with statistically significant elevation with respect to controls
IA = Indoor air monitoring
NS = Not statistically significant
PM = Personal monitoring of breath
U = Biological monitoring of urine for trichloroacetic acid (U-TCA)
-------
1 et al., 1991; Pesch et al., 2000a). The studies by Aschengrau et al. (1993), Partanen et al. (1991),
2 and Pesch et al. (2000a) are of high quality because they had good exposure information,
3 controlled adequately for confounding, and used histologic confirmation of outcomes. For these
4 reasons, observations in these two case-control studies carry greater weight than observations in
5 the other case-control studies identified in Table 4B-4 (Appendix 4B). The remaining studies
6 included large numbers of cases self-reported to determine exposure. These types of reports are
7 more subject to misclassification errors.
8 In many of the case-control studies there are concerns about selection bias, blinding of
9 investigators or interviewers, and, particularly, exposure characterization (Wartenberg et al.,
10 2000). Three studies (Pesch et al., 2000a; Dosemeci et al., 1999; Schlehofer et al., 1995) present
11 risks for tetrachloroethylene exposure explicitly. The studies by Pesch et al. (2000a) and
12 Dosemeci et al. (1999) both suggest that there may be gender differences in renal cell carcinoma
13 risk with occupational exposure to tetrachloroethylene; in both studies the risks were higher in
14 males than in females. Exposure-response gradients were not observed in any of the three
15 studies.
16 Cohort studies of kidney cancer incidence among dry cleaners and laundry workers in
17 Sweden and Denmark (Andersen et al., 1999; Travier et al., 2002; Lynge and Thygesen, 1990;
18 McLaughlin et al., 1987) did not observe excess risks of kidney cancer (see Table 4B-la and
19 Appendix 4B)—an inconsistency with the case-control studies. Few kidney cancer deaths were
20 observed in cohort studies assessing mortality among dry cleaners (Ruder et al., 2001; Blair et al.,
21 2003; see Table 4B-lb and Appendix 4B). The highest risks (not statistically significant) were
22 reported for tetrachloroethylene-exposed subjects (Ruder et al., 2001) and for subjects identified
23 with higher levels of exposure as compared with subjects with little or no exposure (Blair et al.,
24 2003). There are too few cases of kidney cancer in the tetrachloroethylene subcohorts (degreaser
25 studies) to assess any relationship with tetrachloroethylene (see Table 4B-2 and Appendix 4B).
26
27 4.5.2. Animal Studies
28 4.5.2.1. Kidney Toxicity in Animals
29 Tetrachloroethylene causes renal toxicity across several species, including rats, mice,
30 rabbits, dogs, guinea pigs, and humans (for reviews, see U.S. EPA, 1985a, ATSDR, 1997; NYS
31 DOH, 1997; Cal EPA, 2001).
32 Adverse effects on the kidney have been observed in studies of animals exposed to high
33 concentrations of tetrachloroethylene by inhalation, oral intake, and i.p. injection. These effects
34 include hyperplasia and increased kidney-to-body weight ratios, hyaline droplet formation,
35 glomerular "nephrosis," karyomegaly (enlarged nuclei), cast formation, and other lesions or
36 indicators of renal toxicity. The effects occurred following very high doses or chronic, relatively
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-39 DRAFT-DO NOT CITE OR QUOTE
-------
1 high-doses of tetrachloroethylene exposures. The LOAEL for renal toxicity reported in the
2 scientific literature is 100 ppm (678 mg/m3) for inhalation exposure in mice (NTP, 1986a).
3 Oral administration of tetrachloroethylene in sesame oil (3 g/kg/day for 15 days) to mice
4 caused an increase in kidney weight as well as increases in glomerular nephrosis and
5 degeneration (Ebrahim et al., 1996). A lifetime animal carcinogenicity study in which
6 tetrachloroethylene was administered to rats and mice by oral gavage in corn oil for 78 weeks
7 resulted in clear evidence of kidney toxicity in both species (NCI, 1977). The TWA doses
8 (mg/kg-day) used in the bioassay were 471 and 941 for male rats, 474 and 949 for female rats,
9 536 and 1,072 for male mice, and 386 and 772 for female mice. Nephropathy was observed in
10 almost all of the test animals.
11 Hayes et al. (1986) reported renal effects in rats exposed to 400 mg/kg-day
12 tetrachloroethylene in drinking water for 90 days. In a study by Jonker et al. (1996),
13 tetrachloroethylene nephrotoxicity was observed in female Wistar rats administered the chemical
14 in corn oil by oral gavage for 32 days. Nephrotoxic effects were noted at 2,400 mg/kg.
15
16 4.5.2.2. Kidney Cancer in Animals
17 In the studies conducted by NTP (1986a), groups of 50 male and 50 female F344/N rats
18 were exposed for 6 hrs/day, 5 days/week, for 103 weeks by inhalation to atmospheres containing
19 0, 200, or 400 ppm tetrachloroethylene. Tubule cell hyperplasia was observed in male rats
20 (control, 0/49; low dose, 3/49; high dose, 5/50) and in one high-dose female rat. Renal tubule
21 adenomas and adenocarcinomas were observed in male rats (control, 1/49; low dose, 3/49; high
22 dose, 4/50).
23 The spontaneous incidence rate for renal tubule tumors in F344/N rats, the strain used in
24 the NTP bioassay, as well as for other rat strains reported by NTP was less than 1%, making the
25 appearance of tubule neoplasms in 8% of the treated animals in the NTP study (low-dose and
26 high-dose groups combined) convincing evidence of a treatment-related effect (Goodman et al.,
27 1979; Solleveld et al., 1984; U.S. EPA, 1986a, 1991a). Also notable is the fact that no malignant
28 renal tubule neoplasms had ever been observed in any control rats examined by NTP—including
29 chamber controls from the performing laboratory and the untreated controls and vehicle controls
30 from gavage studies—whereas two of the tumors observed in high-dose animals in the NTP
31 study were carcinomas. The probability of two rare carcinomas appearing by chance in a group
32 of 50 animals has been calculated to be less than 0.001 (U.S. EPA, 1987a, 1991a; NTP, 1986a).
33 In addition, when statistically compared with historical control incidences of renal tubule
34 tumors, a significant dose-related positive trend exists, and tumor incidences in both low-dose
35 and high-dose groups are significantly elevated. Standard statistical analyses of tumor incidence
36 data did not reveal a significant increase in kidney tumors, and the tumor incidence is not
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-40 DRAFT-DO NOT CITE OR QUOTE
-------
1 statistically significant when compared with concurrent controls; however, when the incidences
2 of tubule cell hyperplasia and neoplasms and tumor severity are all considered, a dose-response
3 relationship is apparent.
4 A slight increase in renal tumors was observed in other studies in male Sprague-Dawley
5 (SD) rats receiving tetrachloroethylene by gavage or by inhalation (Maltoni and Cotti, 1986;
6 Rampy et al., 1978; cited in U.S. EPA, 1991a), consistent with the findings reported in the NTP
7 studies. However, in the rat chronic bioassay reported by JISA (1993), there was no increase in
8 the incidence of kidney tubular cell adenoma or carcinoma in excess of that in the concurrent or
9 historical control animals (see Tables 5-7 and 5-8) at administered concentrations of 50, 200, and
10 600 ppm.
11 The findings of rare kidney tumors in some cancer bioassays constitute suggestive
12 evidence that tetrachloroethylene can induce kidney cancer in humans. The findings of rare
13 kidney tumors in tetrachloroethylene bioassays in laboratory animals have been reviewed in EPA
14 assessment documents on tetrachloroethylene (U.S. EPA, 1985a, 1986a, 1991a) and by an EPA
15 risk assessment forum technical panel (U.S. EPA, 1991b) reporting on the association of alpha-
16 2u-globulin with renal lesions in the male rat. The closely related tetrachloroethylene congener
17 trichloroethylene also induces low increased incidences of rare renal tumors in rats (U.S. EPA,
18 200la) and NTP has found low incidences of tubule neoplasms in rats dosed with other
19 chlorinated ethanes and ethylenes (NTP, 1983, 1988, 1986a, b, 1987, 1989, 1990b).
20
21 4.5.3. Summary of Kidney Effects in Humans and Animals
22 Taken together, the epidemiologic studies support an inference of subtle effects on the
23 renal proximal tubules from inhalation exposure in tetrachloroethylene. The elevated urinary
24 RBP levels seen in two studies (Mutti et al., 1992; Verplanke et al., 1999) provide some evidence
25 for effects to the proximal tubules from tetrachloroethylene exposure. Exposures in the two
26 studies that observed renal toxicity were 1.2 ppm and 15 ppm (means), representing an
27 observational LOAEL for human kidney effects. None of the reviewed studies reported
28 exposure-response relationships, and this is an important limitation of the available data.
29 However, as pointed out by Mutti et al. (1992), this is a common finding among solvent-exposed
30 populations, and inadequate definition of the dose metric most likely contributes to the absence
31 of exposure-response relationships. No human studies investigating drinking water or other oral
32 exposures on kidney toxicity have been published.
33 Positive associations between kidney cancer (renal cell carcinoma) and exposure to dry
34 cleaning and laundry workers or to tetrachloroethylene specifically were observed in several
35 well-conducted studies (Mandel et al., 1995; McCredie and Stewart, 1993; Pesch et al., 2000a;
36 Schlehofer et al., 1995).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-41 DRAFT-DO NOT CITE OR QUOTE
-------
1 Adverse effects on the kidney have been observed in studies of animals exposed to high
2 concentrations of tetrachloroethylene by inhalation, oral gavage, and i.p. injection. These effects
3 include hyperplasia and increased kidney-to-body weight ratios, hyaline droplet formation,
4 glomerular "nephrosis," karyomegaly, enlarged nuclei, cast formation, and other lesions or
5 indicators of renal toxicity. Increased incidences of relatively rare renal tumors have been
6 observed in multiple studies of male rats. The renal effects occurred following very high (or
7 chronic, relatively high) doses of tetrachloroethylene exposures. The LOAEL for renal toxicity
8 reported in the scientific literature is 100 ppm (678 mg/m3) for inhalation exposure in mice (NTP,
9 1986a).
10
11 4.5.4. Mode of Action for Kidney Toxicity and Carcinogenicity
12 4.5.4.1. Background
13 The data support the conclusion that the chronic administration of tetrachloroethylene
14 produces nephrotoxicity in both sexes of mice and rats and an increased incidence of
15 proliferative lesions of the kidney tubules in male rats. The renal tumors observed in male rats
16 exposed to tetrachloroethylene are of a rare type and include carcinomas. However, the use of
17 these data to infer risk of carcinogenesis to humans has been a focus of scientific debate. Of
18 particular consequence in this debate are two issues: the possibility that quantitative species
19 differences in conjugative metabolism of tetrachloroethylene may greatly reduce the potential
20 risk of human hazard and the possibility that the induction of renal tubule tumors by
21 tetrachloroethylene may be unique to male rats and, therefore, is inappropriate for deducing
22 potential human health hazard.
23 There are multiple hypothesized MO As for kidney toxicity induced with
24 tetrachloroethylene exposure, including mutagenicity, alpha-2u-globulin accumulation, and
25 cytotoxicity unrelated to alpha-2u-globulin. When clearly demonstrated to develop from the
26 sequence of events induced by alpha-2u-globulin accumulation, kidney tumors in male rats
27 caused by exposure to a test chemical, are generally considered to be species and sex specific
28 and not relevant for assessing human hazard. Limited data from studies of tetrachloroethylene
29 indicating hyaline droplet formation provide some evidence for the alpha-2u-globulin MOA.
30 The phenomenon occurs only at very high doses of tetrachloroethylene, however, above the
31 doses used in cancer bioassays in which tumors were observed. There is also data supporting
32 other MO As for tetrachloroethylene-induced renal tumors, particularly findings of mutagenicity,
33 and also cytotoxicity not associated with alpha-2u-globulin accumulation. This mutagenicity
34 and cytotoxicity are attributed to the further biotransformation of glutathione and cysteine
35 conjugates of tetrachloroethylene to reactive chemical intermediates. Humans are known to
36 conjugate tetrachloroethylene with glutathione and excrete the mercapturate end product,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-42 DRAFT-DO NOT CITE OR QUOTE
-------
1 therefore, exhibiting evidence for operation of the same metabolic pathway. Quantitative
2 measurements of urinary excretionmetabolites from this pathway do not provide data for
3 estimations of the amount of chemical coverted to toxic intermediates; therefore, relative
4 amounts of tetrachloroethylene processed by enzymes activating the conjugates to toxic products
5 in rats versus humans are not known.
6
7 4.5.4.2. Summary Description of a Postulated Mode of Action—alpha-2u-globulin
8 Accumulation
9 A variety of organic compounds have been shown to cause sex- and species-specific
10 lesions in the renal tubules of male rats in the form of what is known as "hyaline droplet
11 nephropathy" (NTP, 1983, 1986a, b, 1987, 1990a; U.S. EPA 1991b; Alden, 1985; MacNaughton
12 and Uddin, 1984; Alden and Repta, 1984; Phillips et al., 1987). These chemicals have been
13 associated with interference of normal renal proximal tubule reabsorption of protein from the
14 glomerular filtrate, resulting in accumulation of alpha-2u-globulin in phagolysosomes of renal
15 proximal tubule cells (U.S. EPA, 1991a, b). This accumulation is believed to be the reason for
16 an excessive number of hyaline droplets (Stonard et al., 1986; Olson et al., 1987) and associated
17 nephropathy. The sequence of functional changes in the epithelial cells of proximal tubules, with
18 subsequent tubule necrosis and compensatory cell proliferation, is hypothesized to culminate in
19 the renal tubule tumors observed in the male rats exposed to these compounds in bioassays
20 (UAREP, 1983; Alden et al., 1984; Haider et al., 1984; Swenberg et al., 1989; U.S. EPA, 1991b).
21 Alpha-2u-globulin is considered unique to the male rat and is the major component of its
22 urinary protein load. Alpha-2u-globulin is synthesized in the liver under hormonal control, but it
23 has not been detected in the liver of female rats or in other species (Roy et al., 1975; U.S. EPA,
24 1991b), although homologous proteins do exist in other species, including humans (Flower et al.,
25 1993).
26 The renal tubule tumors associated with alpha-2[j, nephropathy appear to be the end
27 product in the following histopathological sequence of functional changes in the epithelial cells
28 of proximal tubules:
QQ
m
31 1. Excessive accumulation due to increased number and size of hyaline droplets in the P2
32 segment of renal proximal tubule cells, representing lysosomal overload, leads to tubule
33 cell degeneration, cell loss, and regenerative cellular proliferation. The excessive hyaline
34 droplet accumulation occurs in male rats only, and the accumulating protein is identified
35 as alpha-2u-globulin.
36
37 2. Cell debris in the form of granular casts accumulates at the corticomedullary junction,
38 with associated dilation of the affected tubule segment and, more distally, mineralization
39 of tubules within the renal medulla.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-43 DRAFT-DO NOT CITE OR QUOTE
-------
1 3. The chronic progressive nephropathy characteristically found in aging rats is exacerbated
2 as a consequence of the induced nephrotoxicity.
O
4 4. Renal tubule hyperplasia and neoplasia develop subsequently. The increased cellular
5 proliferation is thought to cause development of renal cell tumors due to increases in
§ DNA damage in replicating cells.
8
9 This proposed MOA for kidney tumorigenesis seems plausible to many scientists and
10 may provide an adequate explanation of the specific susceptibility of the male rat to renal tubule
11 tumor induction by certain chemicals. However, data gaps still exist, and the mechanism of
12 cellular damage in alpha-2u-globulin nephropathy is not known (Melnick et al., 1997). Several
13 chemical compounds have been shown to cause the acute renal nephropathy associated with
14 alpha-2u-globulin accumulation, but not to cause tumors (Melnick et al., 1997; Swenberg and
15 Lehman-McKeeman, 1999) even though a critical level of regenerative cellular proliferation may
16 have to be attained for renal tumorigenesis to occur (Swenberg and Lehman-McKeeman, 1999).
17 EPA has developed specific criteria for use in evaluating the likelihood of a chemical's
18 inducing renal tumors through the hypothesized alpha-2u-globulin MOA (U.S. EPA, 1991b).
19 Although EPA downgrades the finding of kidney cancer in male rats as being unimportant to the
20 human situation if it can be shown that the criteria for alpha-2u-globulin are clearly met, the
21 proposed MOA, although reasonable, is still hypothetical, and other reasonable alternative
22 hypotheses have been proposed. As described and discussed below, in the case of
23 tetrachloroethylene, evidence for alpha-2u-globulin accumulation exists only at doses above
24 cancer-causing doses, and other alternative MO As are well supported.
25
26 4.5.4.2.1. Human relevance of alpha-2u-globulin nephropathy. The U.S. EPA has specific
27 guidance (U.S. EPA, 1991b) for evaluating chemically induced male rat renal tumors to
28 determine the use of the data for human risk assessment. It is interesting to note, however, that
29 controversy still exists within the scientific community regarding this mode of carcinogenic
30 action and its relevance to human health risk assessment (Lash et al., 2000b; Ashby, 1996; de la
31 Iglesia et al., 1997; Dietrich, 1997; Huff, 1995, 1996; Melnick et al., 1997; Melnick, 2001, 2002).
32
33 4.5.4.2.2. Identification of tetrachloroethylene-specific key events in support of the alpha-2u
34 hypothesis. Goldsworthy et al. (1988) observed increases in alpha-2u-hyaline droplets in
35 exposed male but not female F344 rats following 10 days of gavage with 1,000 mg/kg
36 tetrachloroethylene. This finding was correlated with both protein droplet nephropathy
37 (crystalloid accumulation) and increases in cellular proliferation. The cell replication was
38 enhanced in the male rats specifically in damaged P2 segments, suggesting a link between the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-44 DRAFT-DO NOT CITE OR QUOTE
-------
1 alpha-2u-globulin accumulation and kidney tumors. These investigators reported similar
2 findings for pentachloroethane in the same study, but at a dose of 150 mg/kg for 10 days.
3 Trichloroethylene has a similar structure but did not cause any alpha-2[j, accumulation or increase
4 in protein droplets, nor did it stimulate cellular proliferation in either male or female rats in this
5 study when a dose of 1,000 mg/kg was administered for 10 days. Bergamaschi et al. (1992) also
6 demonstrated alpha-2u-accumulation in P2 segments of rat proximal tubule cells resulting from a
7 daily exposure of rats to 500 mg/kg tetrachloroethylene in corn oil for 4 weeks.
8 In short-term, high-dose studies, Green et al. (1990) found that the oral administration of
9 from 1,000 to 1,500 mg/kg of tetrachloroethylene daily for up to 42 days caused an accumulation
10 of alpha-2u-globulin in the proximal tubules of male rats. The animals were sacrificed within
11 24 hrs of the last dose of tetrachloroethylene. The effect was accompanied by evidence of
12 nephrotoxicity, with the formation of granular tubular casts and evidence of tubular cell
13 regeneration. These effects were not observed in female rats or in mice. Inhalation exposure to
14 1,000 ppm tetrachloroethylene for 10 days resulted in the formation of hyaline droplets in the
15 kidneys of male rats, but granular casts and tubule cell regeneration were not observed, although
16 the time period may have been too short to allow progression to this stage. These results show
17 that very high doses of tetrachloroethylene are capable of precipitating hyaline droplet
18 nephropathy in male rats. The results also show that male rats are more sensitive to the effect
19 than are female rats or mice of either sex. It is possible, therefore, that alpha-2u-globulin
20 accumulation may indeed play a role in the tumorigenesis observed in male rats exposed to
21 tetrachloroethylene. EPA has listed tetrachloroethylene as an alpha-2u-accumulator (U.S. EPA,
22 1991b) and in the same report specifically identified trichloroethylene as not being an alpha-2u-
23 accumulator.
24 It is interesting to note that tetrachloroethylene-induced alpha-2u-globulin accumulation
25 is probably more likely to be caused by the parent molecule rather than by its metabolites,
26 because its occurrence is related more to the charge and lipid solubility of the inducing agent
27 than to specific interactions with reactive chemical species (Lash and Parker, 2001).
28
29 4.5.4.2.3. Points relevant to biological plausibility and coherence of this mode of action
30 (MOA) for tetrachloroethylene-induced kidney tumors. The following points show that factors
31 other than the specific protein droplet nephropathy may have as much—or more—of a
32 significant role in explaining renal tumor formation resulting from tetrachloroethylene exposure,
33 although some contributions of alpha-2u-globulin accumulation cannot be entirely ruled out.
34 The alpha-2u-globulin response reported following exposure to tetrachloroethylene is
35 relatively mild, and the fact that renal tumors have been observed at doses lower than the ones
36 shown to cause the alpha-2u-globulin response is inconsistent with this phenomenon being
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-45 DRAFT-DO NOT CITE OR QUOTE
-------
1 responsible for tumorigenesis. Although the alpha-2u-globulin response occurs in male rats
2 exposed to tetrachloroethylene, it has been observed following only high doses. Green et al.
3 (1990) tested lower inhaled tetrachloroethylene doses in rats—up to 400 ppm for 6 hrs/day for 28
4 days, with the animals being sacrificed within 18 hrs of termination of the final exposure—but
5 found no evidence of hyaline droplet formation; however, there may have been time for recovery
6 prior to sacrifice.
7 It is noteworthy that the 400 ppm concentration was the same exposure level used for the
8 high-dose rats in the NTP inhalation carcinogenicity bioassay (NTP, 1986a). In the NTP study,
9 the 400 ppm concentration caused a high incidence of nontumor nephropathy and resulted in the
10 formation of kidney tubule adenomas and adenocarcinomas. The renal pathology of rats in the
11 NTP study was reported to be different from the specific alpha-2u-globulin nephropathy, but the
12 age of the rats as well as the length of time that elapsed between final exposure and sacrifice may
13 explain some of the differences. However, mineralization in the inner medulla and papilla of the
14 kidney—a characteristic trait of alpha-2u-globulin nephropathy—was not seen.
15 Green et al. (1990) proposed the possibility that longer-term exposure to the 400 ppm
16 concentration of tetrachloroethylene is required for the hyaline droplet accumulation in the
17 kidney of rats. Alpha-2u-globulin accumulation can be demonstrated, however, after only short-
18 term exposures (even a single administration) to several agents, such as d-limonene, decalin,
19 unleaded gasoline, and trimethylpentane (Charbonneau et al., 1987; NTP, 1988). Lack of
20 hyaline droplet formation, increase in alpha-2u-globulin, or signs of the characteristic renal
21 nephropathy at the high dose level of the NTP inhalation study (NTP, 1986a) may indicate a
22 threshold effect and thus diminish the likelihood that the renal tumors associated with exposure
23 to tetrachloroethylene are induced through this mechanism (Green et al., 1990).
24 Pharmacokinetic differences between oral and inhalation exposure may contribute to the
25 observed discrepancies in some of the results.
26 NTP did not report the presence of hyaline droplets in rats that had been exposed to either
27 200 or 400 ppm tetrachloroethylene for up to 2 years. These doses were associated with the
28 production of renal tubule neoplasms in male rats. However, the fact that NTP did not report the
29 presence of hyaline droplets in the 14-day, 90-day, or 2-year studies is not definitive, because the
30 NTP protocol at that time was not designed specifically to detect hyaline droplets or alpha-2u-
31 globulin accumulation in the kidney (NTP, 1990a). Thus, the procedures followed at the time of
32 the study were not necessarily conducive to detecting hyaline droplets. For example, in the
33 chronic study of tetrachloroethylene, at least 1 week elapsed between the final
34 tetrachloroethylene exposure and the scheduled sacrifice of the surviving animals. It is possible
35 that had hyaline droplets been present, they could have regressed. Also, the nephropathy
36 observed at the end of a 2-year bioassay could be difficult to distinguish from the old-age
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-46 DRAFT-DO NOT CITE OR QUOTE
-------
1 nephropathy that occurs in these rats. Other investigators (Goldsworthy et al., 1988; Green et al.,
2 1990) observed hyaline droplets containing alpha-2u-globulin following very high doses of
3 tetrachloroethylene administered to male rats.
4 On the other hand, the renal pathology reported in the NTP bioassay is not entirely
5 consistent with the results generally found for chemicals where there is alpha-2u-globulin
6 accumulation (NTP, 1986a; letter from Scot Eustis, National Toxicology Program, to William
7 Farland, Director, Office of Health and Environmental Assessment, EPA, 1988). For example,
8 there was no mineralization in the inner medulla and papilla of the kidney, a frequent finding in
9 bioassays of chemicals that induce alpha-2u-globulin accumulation (e.g., for pentachloroethane,
10 the incidence of renal papillar mineralization was 8% in controls, 59% in the low-dose group,
11 and 58% in the high-dose group). In addition, it is important to note that some aspects of toxic
12 tubular nephropathy were also observed in female rats and male mice exposed to
13 tetrachloroethylene, clearly contrary to sex and species specificity.
14 Thus, chronically induced tetrachloroethylene nonneoplastic kidney lesions exhibit
15 neither species nor sex specificity. Unlike with other chemicals that induce alpha-2u-globulin
16 accumulation and have been tested by NTP in chronic carcinogenicity bioassays, renal lesions
17 occurring in animals exposed to tetrachloroethylene were not limited to the male rat. Although
18 the female rat did not develop any renal tubule tumors, the incidence of karyomegaly was
19 significantly elevated in the female rat as well as in the male rat; 1 of 50 female rats exposed at
20 the high dose developed tubule cell hyperplasia.
21 In the mouse, "nephrosis" was observed at increased incidences in dosed females, casts
22 were observed at increased incidences in dosed males and high-dose females, and karyomegaly
23 of the tubular cells was observed at increased incidences in both sexes of treated mice. The
24 severity of the renal lesions was dose related, and one low-dose male had a renal tubular cell
25 adenocarcinoma.
26 In the NCI gavage study of tetrachloroethylene (NCI, 1977), toxic nephropathy, which
27 was not detected in the control animals, occurred in both male and female Osborne-Mendel rats
28 administered tetrachloroethylene. Tetrachloroethylene also clearly caused nephropathy in both
29 sexes of mice in the study. Unfortunately, animal survival in the rat study was not adequate to
30 support any conclusions about tetrachloroethylene carcinogenicity.
31 Other chlorinated ethanes and ethylenes also produce nephrotoxicity and renal tubule
32 tumors in laboratory animals. Hexachloroethane causes accumulation of hyaline droplets and
33 renal tubule tumors in male rats (NTP, 1989). On the other hand, trichloroethylene, which was
34 also tested by NTP, induces kidney tumors in male rats and also possibly in female rats, but it
35 does not cause an accumulation of hyaline droplets or an increase in levels of alpha-2u-globulin
36 (Goldsworthy et al., 1988). Consequently, kidney tumors induced by this compound are not
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-47 DRAFT-DO NOT CITE OR QUOTE
-------
1 considered to be associated with alpha-2u-globulin accumulation (U.S. EPA, 1991b, 2001a).
2 Tetrachloroethylene is related in structure to trichloroethylene, and both chemicals have been
3 shown to be metabolized in the kidney to cytosolic and mutagenic compounds.
4
5 4.5.4.3. Other Modes of Action for Tetrachloroethylene-Induced Renal Tumors in Rats
6 4.5.4.3.1. Genotoxicity. The glutathione conjugation of tetrachloroethylene in the kidney,
7 discussed in Chapter 3, leads sequentially to S(l,2,2-trichlorovinyl)glutathione and
8 S(l,2,2-trichlorovinyl)cysteine—TCVG and TCVC. TCVC can be further processed by beta-
9 lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a highly reactive
10 thioketene, a chemical species that can form covalent adducts with cellular nucleophiles
11 including DNA. Additionally, sulfoxidation of both TCVC and its N-acetylated product via
12 FMO3 or P450s occurs, resulting in reactive metabolites (Ripp et al, 1997, 1999; Werner et al.,
13 1996). While most of these intermediates have not been characterized for mutagenic potential,
14 TCVG, TCVC, and NAcTCVC are clearly mutagenic in Salmonella tests. In addition,
15 tetrachloroethylene exhibited mutagenicity in Salmonella in the few studies of conditions that
16 could generate GSH-derived metabolites and, following in vivo exposures, induces SSB and
17 DNA binding in kidney. See Section 4.3 for more details of genotoxicity.
18
19 4.5.4.3.2. Peroxisomeproliferation. Peroxisome proliferation has been linked to tumorigenesis
20 in rodents; however, the mechanisms involved have not been clearly elucidated (see Section 4.3).
21 The PPARs, a class of nuclear receptors, are believed to be transcriptionally activated to mediate
22 the effects of peroxisome proliferators (Issemann et al., 1993; Desvergne and Wahli, 1995).
23 Although most of the focus on peroxisome proliferation and PPAR receptor activation, and their
24 relationship to tumor development, has been on the liver (see Section 4.3.4 for a more detailed
25 discussion), the phenomenon can also occur in other tissues. In fact, peroxisomes were first
26 noted in rodent renal tubule epithelial cells, and peroxisome proliferation in these cells in
27 response to peroxisome proliferating agents is not unusual (reviewed by Stott, 1988; Klaunig et
28 al., 2003). Data exist to support increased peroxisome proliferation in rodent kidney following
29 exposure to tetrachloroethylene (Goldsworthy and Popp, 1987; Zanelli et al., 1996).
30 Goldsworthy and Popp (1987) investigated the ability of tetrachloroethylene to induce
31 peroxisome proliferation in both liver and kidney of rats and mice using increases in cyanide-
32 insensitive PCO activity as a marker enzyme. Tetrachloroethylene caused elevations in enzyme
33 activity in mouse kidney as well as liver but not in rat kidney. It seems somewhat unlikely that
34 any peroxisome proliferation observed following tetrachloroethylene exposure would be
35 associated with renal tumors if one considers the magnitude of the measurable peroxisome
36 proliferation effect across species. Renal tumors occur in rats, but greater peroxisome
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-48 DRAFT-DO NOT CITE OR QUOTE
-------
1 proliferation is observed in the kidneys of mice (Goldsworthy and Popp, 1987). Modification of
2 cell signaling pathways that control rates of cell division and apoptosis, for example, may occur
3 through activation of PPAR receptors. The occurrence of peroxisome proliferation per se may
4 be only a marker for PPAR receptor activation, the key event having the most support for being
5 causally related to other tumor types. The dissimilarity in the peroxisome response observed in
6 different tissues may be related to variability in the levels of PPAR receptors in these tissues.
7 It is important to note that, relatively speaking, chlorinated ethylenes, particularly
8 tetrachloroethylene, and their chloroacid metabolites are not very potent peroxisome proliferators
9 compared to many other compounds that are known to cause the phenomenon. PPAR receptors
10 belong to the superfamily of proteins that control almost every metabolic and developmental
11 event in mammals, and PPAR receptor activation is known to result in numerous biochemical,
12 physiological, and molecular events. Therefore, the possibility of PPAR receptor activation
13 being related to tetrachloroethylene-induced kidney tumorigenesis in rats is plausible, especially
14 since peroxisome proliferation, the phenomenon that could be considered a biomarker for PPAR
15 activation, has been shown to occur in the kidneys of mice following exposure to
16 tetrachloroethylene. Moreover, a causal link between PPAR activation and kidney tumorigenesis
17 has yet to be established for tetrachloroethylene or other PPAR agonists.
18
19 4.5.4.3.3. Cytotoxicity/sustained chronic nephrotoxicity not associated with alpha-2fi-globulin
20 nephropathy. The kidney is a major target organ for tetrachloroethylene-induced toxicity
21 through the reactive metabolites of TCVC. Tetrachloroethylene has been reported to produce
22 nephrotoxicity across species, although its relative potency is not extremely high. Other
23 chlorinated ethanes and ethylenes also induce nephrotoxicity, although the toxicity manifests
24 itself differently with specific chemicals. The observed effects vary across species and between
25 sexes and may include tubular cell cytomegaly, karyomegaly and pleomorphism, tubular cell
26 dilation, or the formation of granular casts. There may be a link between renal toxicity and
27 tumorigenesis, and sustained kidney damage may be a risk factor for tumorigenesis. It is
28 reasonable, therefore, to suspect that renal tubule neoplasia observed in tetrachloroethylene-
29 exposed male rats may be influenced by cytotoxicity and subsequent cellular regeneration.
30 It has been suggested that renal neoplasms induced by tetrachloroethylene may be
31 secondary to renal cytotoxicity and subsequent cellular proliferation without regard to alpha-2u-
32 accumulation. Thus, sustained chronic nephrotoxicity, independent of alpha-2u-globulin
33 accumulation and its resulting neuropathic cascade of events, may be a possible MO A for
34 tetrachloroethylene carcinogenesis. If this is the case, renal tubule neoplasia observed to occur in
35 male rats would not be expected to be a species- or sex-specific response because the nontumor
36 lesions appear in both sexes of both rodent species tested. In support of this expectation, the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-49 DRAFT-DO NOT CITE OR QUOTE
-------
1 renal lesions occurring in animals exposed to tetrachloroethylene are not limited to the male rat.
2 Signs of tetrachloroethylene-induced kidney damage appeared in both rats and mice during the
3 early phases of the NTP inhalation study, for example, indicating that animals of both species
4 surviving to the scheduled termination of the study had long-standing nephrotoxicity. Although
5 the female rats did not develop any renal tubule tumors, the incidence of karyomegaly was
6 significantly elevated in females as well as in males, and 1 of 50 female rats exposed at the high
7 dose developed tubule cell hyperplasia.
8 In the NTP study of the mouse, "nephrosis" was observed at increased incidences in
9 dosed females, casts were observed at increased incidences in dosed males and high-dose
10 females, and karyomegaly of the tubule cells was observed at increased incidences in both sexes
11 of treated mice. The severity of the renal lesions was dose related, and one low-dose male had a
12 renal tubule cell adenocarcinoma. In the NCI gavage study of B6C3F1 mice and Osborne-
13 Mendel rats exposed to tetrachloroethylene, toxic nephropathy was not detected in control
14 animals but did occur in both male and female rats as well as in mice. On the other hand,
15 findings using in vitro models studied by Lash et al. (2002) suggest a marked sex difference
16 between male and female rats in the severity of acute renal toxicity caused by both
17 tetrachloroethylene and its TCVG metabolite. Tetrachloroethylene and TCVG also produced
18 signs of toxicity in mitochondria; i.e., mitochondrial dysfunction, such as inhibition of state 3
19 respiration by specific inhibition of several sulfhydryl-containing enzymes in both sexes of mice
20 (Lash et al., 2000, 2001,2002).
21 Mechanistic studies of tetrachloroethylene nephrotoxicity are relatively sparse. More and
22 better data are available for trichloroethylene. Most studies performed to elucidate information
23 related to understanding tetrachloroethylene renal toxicity have concentrated on the GSH
24 pathway metabolites rather than on the parent chemical; this is because much available data for
25 both tetrachloroethylene and trichloroethylene suggest that it is flux through this pathway that
26 generates reactive chemical species responsible for nephrotoxicity. Vamvakas et al. (1989c, d)
27 have shown the tetrachloroethylene conjugate metabolites TCVG and TCVC to cause dose-
28 related cytotoxicity in renal cell preparations and prevention of this toxicity by beta lyase
29 enzyme inhibitor. Renal beta lyases are known to cleave TCVC to yield an unstable thiol,
30 1,2,2-trichlorovinylthiol, that may give rise to a highly reactive thioketene, a chemical species
31 that can form covalent adducts with cellular nucleophiles. Additionally, sulfoxidation of both
32 TCVC and its N-acetylated product occurs, resulting in toxic metabolites (Ripp et al, 1997, 1999;
33 Werner et al., 1996). Contribution to overall toxicity is unknown, however, and it is interesting
34 to note that human CYP3A4 catalyzed sulfoxidation of the N-acetyl metabolite of a structurally
35 related chemical, HCBD, at rates comparable to those of rat CYP3A1 (Werner et al., 1995),
36 indicating a relative value of the human-to-rat rate constant of 1 (Lash and Parker, 2001).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-50 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.5.4.3.4. Immunotoxicity/immunosuppression. Although specific data about
2 tetrachloroethylene are lacking, immune suppression could contribute to the induction of kidney
3 tumors caused by tetrachloroethylene exposure. Many immunosuppressive therapeutic agents
4 are human carcinogens (see Tomatis et al., 1989), although they are usually associated almost
5 exclusively with lymphoma. Organic solvent exposure in general is associated with autoimmune
6 disease such as schleroderma and other autoimmune responses (Nietert et al., 1998). Several
7 published studies have demonstrated formation of trichloroacylated protein adducts, for example,
8 in liver and kidney of rats (Birner et al., 1994) and in plasma of rats and humans following
9 exposures to tetrachloroethylene (Pahler et al., 1999). Another recent example is the detection of
10 trichloroacetylated protein adducts formed in mice treated with tetrachloroethylene (Green et al.,
11 2001).
12
13 4.5.4.4. Summary
14 The kidney is a target organ in mammalian species for tetrachloroethylene and other
15 related chlorinated ethanes and ethylenes, and tetrachloroethylene causes kidney cancer in male
16 rats. It is likely that several mechanisms contribute to tetrachloroethylene-induced kidney
17 toxicity, including cancer, and that the relative importance of specific MO As varies from high-
18 dose to low-dose exposure. Peroxisome proliferation, alpha-2u-globulin nephropathy,
19 mutagenicity, and cytotoxicity not associated with alpha-2u-globulin accumulation are MO As
20 that have been investigated. The pathogenesis of immunosuppression is another potential MOA
21 that may also be related to tumorigenesis.
22 Tetrachloroethylene-induced renal toxicity is likely associated with its metabolites rather
23 than with the parent compound, except for toxicity associated with alpha-2u-globulin
24 accumulation, which is more likely due to tetrachloroethylene itself (Lash and Parker, 2001).
25 The GSH conjugate and reactive metabolites generated from further processing of TCVC
26 or its acetylated metabolite NAcTCVC by beta lyase, FMO3/P450 and/or CYP3 A, are the most
27 likely tetrachloroethylene metabolites to induce renal toxicity and tumorigenicity, as opposed to
28 the metabolites resulting from oxidative CYP processing. These conjugate metabolites are
29 known to be mutagenic, and they are known to occur in rats, mice, and humans.
30 Due to tetrachloroethylene's nephrotoxic effects, it has been suggested that the low-level
31 renal tumor production observed in exposed rats is secondary to sustained cytotoxicity and
32 necrosis leading to activation of repair processes and cellular regeneration. However,
33 "nephrotoxicity" occurs in both sexes of rats and mice whereas cell replication and
34 tumorigenesis occurs in male but not in female rats In addition, tetrachloroethylene induces
35 kidney tumors at lower doses than those required to cause alpha-2u-globulin accumulation,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-51 DRAFT-DO NOT CITE OR QUOTE
-------
1 raising serious doubt that alpha-2u-globulin plays a key role—especially any major role—in the
2 rat kidney tumor formation.
3 Because tetrachloroethylene has been shown to induce peroxisome proliferation, an
4 indicator of PPAR activation, the possibility exists that certain responses resulting from
5 activation of PPAR receptors might be involved in cancer-causing activity leading to
6 tetrachloroethylene-induced renal tumors. However, there is no evidence causally linking
7 PPAR-a activation to kidney tumorigenesis for tetrachloroethylene or other compounds.
8 The weight of evidence suggests that for tetrachloroethylene the further processing of
9 conjugative metabolites by beta lyase, FMO3 and/or CYP3 A leads to reactive and mutagenic
10 metabolites responsible for nephrotoxicity and carcinogenicity. The glutathione conjugation of
11 tetrachloroethylene in the kidney, discussed in Chapter 3, leads sequentially to
12 S(l,2,2-trichlorovinyl)glutathione and S(l,2,2-trichlorovinyl)cysteine—TCVG and TCVC.
13 TCVC can be further processed by beta-lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol,
14 that may give rise to a highly reactive thioketene, a chemical species that can form covalent
15 adducts with cellular nucleophiles including DNA. TCVC can also undergo FMO3 or P450
16 oxidation to reactive intermediates; additionally, sulfoxidation of both TCVC and its N-
17 acetylated product occurs, resulting in reactive metabolites (Ripp et al, 1997, 1999; Werner et al.,
18 1996). While most of these intermediates have not been characterized for mutagenic potential,
19 TCVG, TCVC, and NAcTCVC are clearly mutagenic in Salmonella tests. In addition,
20 tetrachloroethylene exhibited mutagenicity in Salmonella in the few studies of conditions that
21 could generate GSH-derived metabolites. Tetrachloroethylene, following in vivo exposures, also
22 binds to kidney DNA and induces SSB in kidney.
23 In summary, the complete mechanisms of tetrachloroethylene-induced renal
24 carcinogenesis are not yet understood. Given the known mutagenicity of the GSH-derived
25 tetrachloroethylene metabolites that are formed in the kidney, and the observed in vitro
26 mutagenicity of tetrachloroethylene under conditions that would generate these metabolites, a
27 mutagenic MOA contributing to the development of the kidney tumors clearly cannot be ruled
28 out.
29
30 4.6. NEUROTOXICITY
31 4.6.1. Human Studies
32 A wide range of effects on neurologic function are well-documented for both acute and
33 chronic exposure to tetrachloroethylene. Acute controlled inhalation exposures of 100 ppm and
34 higher have induced symptoms consistent with depression of the CNS and included dizziness
35 and drowsiness (ATSDR, 1997). Changes in electroencephalograms (EEGs) have also been
36 noted with controlled inhalation exposures at this level (Stewart et al., 1977). Acute exposure to
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-52 DRAFT-DO NOT CITE OR QUOTE
-------
1 lower levels of tetrachloroethylene has induced alterations in neurobehavioral function. For
2 example, Altmann et al. (1990, 1992) reported increases in the latency of visually evoked
3 potentials (VEPs) and significant performance deficit for vigilance and eye-hand coordination in
4 subjects with inhalation exposure to 50 ppm for 4 hrs/day for 4 days as compared with that seen
5 among subjects exposed to 10 ppm (the control group in these reports). These observations
6 indicate visual system dysfunction, including delayed neuronal processing time, is related to
7 tetrachloroethylene exposure. ATSDR (1997) developed an acute exposure minimal risk level
8 from this study and considered 10 ppm as the no-observed-adverse-effect level (NOAEL).
9 Studies by Stewart et al. (1997) and Hake and Stewart (1977), funded primarily by the
10 National Institutes of Occupational Safely and Health (NIOSH), are third-party studies and are
11 considered by EPA to be protective of human subjects. EPA and other federal agencies
12 subscribe fully to principles articulated in EPA's Protection of Human Subjects Rule ("the
13 Common Rule"), 40 CFR Part 26. EPA recently issued an interim policy on accepting human
14 test data stating that the Agency will continue to accept third-party test data on a case-by-case
15 basis (U.S. EPA, 2005d). A description of the studies by Altmann et al. (1990, 1992) is also
16 included because ATSDR used these studies to develop an acute minimal risk level (MRL;
17 ATSDR, 1977). EPA also considers these studies to be third-party. No information is provided
18 in the published papers regarding the procedures the study investigators adopted for informed
19 consent or protection of human subjects, and staff of the National Center for Environmental
20 Assessment (NCEA) contacted study investigators requesting this information (e-mail
21 communication from Robert McGaughy, U.S. EPA, to L. Altmann, Heinrich-Heine University,
22 Dusseldorf, Germany, October 8, 2003).
23 Epidemiologic studies of workers or residents with chronic exposure to
24 tetrachloroethylene show that the nervous system is a target, with decrements reported in several
25 nervous system domains (Altmann et al., 1995; Schreiber et al., 2002; Seeber, 1989; Ferroni et
26 al., 1992; Cavalleri et al., 1994; Gobba et al., 1998; Spinatonda et al., 1997; Echeverria et al.
27 1994, 1995). Table 4-4 presents select details of available chronic studies evaluating
28 neurological function. Furthermore, neurotoxic effects from chronic-duration exposure to
29 tetrachloroethylene are similar to neurotoxic effects reported for other solvents (Arlien-Sorborg,
30 1982). Case reports and case studies also show the nervous system as a target of organic solvent
31 exposures such as tetrachloroethylene (Seppalainen and Antti-Poika, 1983; Antti-Poika,
32 1982a, b; Onofirj et al., 1998). Electrophysiological abnormalities in tetrachloroethylene- and
33 other organic solvent-exposed patients, with diagnosed chronic solvents intoxication, persisted
34 post-exposure (Seppalainen and Antti-Poika, 1983; Antti-Poika, 1982b).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-53 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-4. Summary of human neurotoxicology studies
Subjects
24 dry cleaners exposed to
tetrachloroethylene, 33
subjects nonoccupationally
exposed to other solvents
101 dry cleaners (both
sexes), 84 nonexposed
controls
56 tetrachloroethylene-
exposed dry cleaning
workers, 69 non-exposed
factory controls
64 dry cleaners exposed to
tetrachloroethylene,
120 controls
Dry cleaners, 60 females
and 30 nonsolvent-exposed
controls
22 dry cleaners, 13 ironers,
35 controls
Exposure
Mean TWA = 21 ppm, PM
Mean duration of exposure
= 6 years
Low-exposure group: 12
ppm TWA, IA, PM, 11.8
years
High-exposure group: 53
ppm TWA IA, PM, 10.6
years
Geometric mean TWA =
20 ppm, PM
Mean duration of exposure
= 3 years
Mean duration of
employment not reported
Geometric mean =15 ppm
(males), 1 1 ppm; (females),
PM
15 ppm IA, 10.1 years
Mean TWA = 6 ppm
(7 ppm, dry cleaning
workers; 5 ppm, ironers),
PM, 8.8 years (exposed
subjects)
Effects
Statistically significant
differences for simple
reaction time and critical
flicker fusion
Decrease in information
processing speed
(perceptual threshold,
choice reaction time),
visual scanning
(cancellation dZ test),
visuo-spatial function
No fine motor function
deficits
Neurological signs
Statistically significant
increase in the prevalence
of several CNS symptoms
such as nasal irritation and
dizziness
No statistically significant
differences between
exposed and controls in
color vision loss
Impaired performance on
three tests (simple reaction
time, vigilance, stress)
No fine motor function
deficit
No effects on digit symbol
test
CCI statistically
significantly elevated
among dry cleaners with a
statistically significant
exposure (TWA) -response
relationship.
No effect on CCI in ironers
Author(s)
Lauwerys et al.
(1983)
Seeber(1989)
Caietal. (1991)
Nakatsuka et al.
(1992)
Ferroni et al.
(1992)
Cavalleri et al.
(1994)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-54 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-4. Summary of human neurotoxicology studies (continuted)
Subjects
45 dry cleaners matched to
69 laundry workers, 59
pressers or counter clerks
65 dry cleaners, no
unexposed group
Residents near dry cleaning
facilities, 14 exposed and 23
age- and gender-matched
nonexposed controls
24 solvent-exposed workers
and 24 controls
35 dry cleaners, 39 age- and
education-matched controls
Exposure
<0.2 ppm, 3 ppm, 9 ppm,
PM, 2. 6 to 11 years
1 1 ppm, 23 ppm, 41 ppm,
PM, 1.2 to 14.6 years
0.7 ppm (mean), IA (7-day
monitoring period), B, 10.6
years
Past use of solvent mixture,
including tetrachloro-
ethylene (10%) for at least
3 years, solvent exposure
for 6 years average
8 ppm TWA, IA, grab
sample
Mean duration of
employment not reported
Effects
Statistically significant
association between chronic
lifetime tetrachloroethylene
exposure and reduced test
performance on three
cognitive tests: switching,
pattern memory, and
pattern recognition.
Statistically significant
differences between "high"
and "low" lifetime
exposure groups in three
tests of visuo-spatial
function
No effect on digit span test
Increase in simple reaction
time
Decrements in continuous
performance and visuo-
spatial function
No fine motor function
deficits
Color vision impairment
(p < 0.05) among exposed
subjects as compared with
controls
Increase in vocal reaction
time to visual stimuli
Concentration-response
relationship
Author(s)
Echeverria et al.
(1994)
Echeverria et al.
(1995)
Altmann, et al.
(1995)
Muttray et al.
(1997)
Spinatonda et al.
(1997)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-55 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-4. Summary of human neurotoxicology studies (continuted)
Subjects
Exposure
Effects
Author(s)
33 dry cleaners and ironers,
self control (follow-up of
Cavalleri et al, 1994,
subjects)
Group A, 4 ppm, PM
Group B, 0.7 ppm, PM
Worse CCI in subjects who
experienced a higher mean
TWA exposure than that
reported in Cavalleri et al.
No impairment in CCI in
subjects who experienced a
lower mean TWA exposure
than that reported in
Cavalleri et al.
Gobba et al.
(1998)
30 patients with solvent-
induced encephalopathy
Several solvents, including
tetrachloroethylene, 20
years
No peripheral neuropathy
Albers et al.
(1999)
Apartment residents living
above dry cleaning
facilities: 17 exposed and
17 age- and gender-matched
unexposed controls
0.4 ppm (mean,
monitoring taken before
closure of dry cleaners) IA,
PM, B, U
Duration of residence, 6
years (mean)
No effects on visual acuity
or color discrimination
(comparison of group
means)
Lower (worse) scores on
tests of visual contrast
sensitivity
Schreiberetal.
(2002)
Employees of a day care
facility located in the same
building as a dry-cleaning
business, 9 exposed and 9
age- and gender-matched
unexposed controls
0.32 ppm (mean,
monitoring taken before
closure of dry cleaners) IA
No information on duration
of employment
No effects on visual acuity
color discrimination
(comparison of group
means)
Lower (worse) scores on
tests of visual contrast
sensitivity
Schreiberetal.
(2002)
14 dry cleaners exposed to
tetrachloroethylene, 27 non-
exposed (Control 1) and 27
support staff of Universiti
Kebangsaan Malaysia
(Control 2)
No exposure information
presented in paper other
than tetrachloroethylene
was used for dry cleaning
No effect on color vision
using Ishihara plates. 43%
and 93% of dry cleaners
compared to no controls
had errors on the color
vision D-15 test and FM
100 Hue test, respectively.
Dry cleaners had more
errors on FM 100 Hue than
control group 2 (p < 0.05)
but not control group 1.
Sharanj eet-Kaur
et al. (2004)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-56 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-4. Summary of human neurotoxicology studies (continuted)
Subjects
Exposure
Effects
Author(s)
65 households (67 adults
and 68 children) in
residential buildings with
co-located dry cleaners, 61
households (61 adults and
71 children) in residential
buildings without dry
cleaners
Geometric mean = 5 ppb
(0.005 ppm), IA
Duration of residence, 10
years (mean)
Association (p < 0.05)
between tetrachloroethylene
concentrations in indoor air
and in blood and
performance on test of
visual contract sensitivity in
children. No association
observed in adults.
Color vision impairment
(p < 0.05) among children
but not adult exposed
subjects as compared with
controls
NYSDOH
(2005a),
McDermott et al.
(2005)
4-year follow-up of 13
former students, 13 children
matched to exposed
children on age, gender, and
daycare experience
Exposure had ceased 4
years earlier.
No difference in visual
function (VCS, color
vision) or neurobehavioral
function between exposed
children and controls.
NYSDOH
(2005c)
88,820 births from
1964-1976 identified from
the Jerusalem Perinatal
Study and linked to Israel's
national Psychiatric
Registry for hospitalization
with a schizophrnia-related
diagnosis as of 1-1-98.
Occupation of mother and
father on birth certificate as
dry cleaner.
Four cases were identified
in 144 offspring of dry
cleaners. Unadjusted
relative risk (RR) of 3.4
(95%CI=1.3-9.2)for
schizophrenia in the
offspring of dry cleaners
using proportional hazard
modeling. Control for a
number of potentially
confounding variables did
not lead to appreciable
different RR estimates.
Perrin et al.
(2007)
2 B = Biological monitoring of blood
3 IA = Indoor air
4 PM = Personal monitoring of breath
5 U = Biological monitoring of urine for trichloroacetic acid
6 VCS = visual contrast sensitivity
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-57 DRAFT-DO NOT CITE OR QUOTE
-------
1 Additionally, other reports (Laslo-Baker et al., 2004; Till et al. 200la, b, 2005) suggest a
2 vulnerability of the fetus to organic solvent exposures, including tetrachloroethylene exposure.
3 Deficits in neurobehavioral parameters and in visual system functioning in young children of
4 mothers exposed during pregnancy were observed in these reports. These reports are not fully
5 discussed in this section. Case series help identify target organ toxicity and can support
6 generating hypotheses; however, the lack of information in an unexposed population in these
7 types of studies limits the ability to infer observations reported among cases to other populations.
8 Most human data on tetrachloroethylene exposure and nervous system effects—from
9 cross-sectional or prevalence studies—is of dry cleaner and laundry workers and assessment of
10 neurobehavioral effects; one report on neurobehavioral effects is of residents living in close
11 proximity to a dry cleaning establishment. Three studies assessed the visual system; two reports
12 of the same population are of dry cleaner and laundry workers, and one report is of two
13 populations living or working in a building co-located with a dry cleaning establishment. Few
14 studies are available on neurologic diseases such as Parkinson's disease, amyotrophic lateral
15 sclerosis, and Alzheimer's disease and organic solvents (IOM, 2002), and none of these reports
16 uniquely assess tetrachloroethylene.
17 Tetrachloroethylene concentrations reported in the dry cleaning and laundry worker
18 studies ranged from an 8-hr TWA mean of 6 ppm for dry cleaner and laundry workers in
19 Cavalleri et al. (1994) to an 8-hr TWA of 41 ppm for operators of wet-transfer dry cleaning
20 machine in Echeverria et al. (1995). Tetrachloroethylene concentrations reported for exposed
21 residents were 0.4 ppm (mean) to residents living in an apartment building containing an
22 operating dry cleaning business (Schreiber et al., 2002) and 0.7 ppm (mean) in indoor air of
23 residents in close proximity to a dry cleaning business (Altmann et al., 1995).
24
25 4.6.1.1. Environmental Chamber Studies
26 4.6.1.1.1. Stewart, R.O., E.D. Baretta, H.C. Doddand T.R. Torkelson. 1970. Experimental
27 human exposure to tetrachloroethylene. Arch. Environ. Health. 20:225—229. In a study by
28 Stewart et al. (1970), 12 healthy adults were exposed to 100 ppm for 7 hrs; eye and nose
29 irritation was reported by 60% of the subjects, a slight frontal headache by 26%, slight light-
30 headedness by 26%, feeling slightly sleepy by 40%, and difficulty in speaking by 25%. Some of
31 these complaints were made during the first 2 hrs. Of five healthy men exposed to 100 ppm for
32 7 hrs/day on 5 consecutive days, one reported a mild frontal headache during each exposure and
33 two consistently reported mild eye and throat irritation. Other symptoms were not reported, and
34 individual responses during exposures to 0 ppm were not assessed. Subjects reported that their
35 ability to detect the odor of tetrachloroethylene decreased during the course of daily and weekly
36 exposure.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-58 DRAFT-DO NOT CITE OR QUOTE
-------
1 Three tests of equilibrium (a modified Romberg test,2 where an individual stands on one
2 foot with eyes closed and arms at side; a heel-to-toe test; and a fmger-to-nose test) were
3 performed every 60 minutes during each day of exposure. After 6 hrs, neurobehavioral tests of
4 motor function (the Crawford manual dexterity and Flanagan coordination tests), cognitive
5 function (arithmetic test), and motor/cognitive function (inspection test) were also performed.
6 Three of the subjects had increased difficulty in maintaining their equilibrium when tested within
7 the first 3 hrs of exposure (i.e., performance on the Romberg equilibrium test was impaired).
8 The three subjects were able to perform the test normally when given a second chance.
9 Performance on the other tests was not impaired. An additional subject, exposed during the third
10 day of testing, showed a slight deterioration in his Romberg test and complained of slight
11 dizziness and slight impairment of his intellectual faculties after 1 hr of exposure. No
12 improvement in his Romberg test occurred during the next hour, and he was removed from the
13 test chamber. The subject performed the test normally when retested 30 minutes later.
14 The investigators concluded that there were CNS effects in some subjects exposed to 100
15 ppm and that there exists a large range of individual susceptibility to tetrachloroethylene. The
16 latter conclusion was based on the observations that only 3 of 17 subjects reported light-
17 headedness and 4 of 17 subjects had an abnormal Romberg test.
18
19 4.6.1.1.2. Hake, C.L., andR.D. Stewart. 1977. Human exposure to tetrachloroethylene:
20 inhalation and skin contact Environ. Health Perspect. 21:231-238. As part of a 6-week study,
21 four healthy men were exposed 7.5 hrs/day to 0 ppm (2 days in week one, 1 day in week three,
22 and 2 days in week six), 21 ppm (4 consecutive days in week three), 100 ppm (5 consecutive
23 days in week two), and a TWA of 100 ppm (5 consecutive days in week four) when exposure
24 levels were more than 53, 100, or 155 ppm (5 consecutive days during week five). In addition,
25 four healthy women were exposed to 100 ppm for 7.5 hrs/day on 5 consecutive days and to 0
26 ppm on 2 days. A fifth woman became sick with influenza during the study and was exposed to
27 100 ppm on only 2 days.
28 All subjects were cautioned to either abstain from the use of alcohol and drugs (or to limit
29 their use to very moderate amounts) and they were asked not to drink coffee until 1 hr after the
30 end of each exposure period. The subjects were told that they would be exposed to various
31 concentration of tetrachloroethylene, but they were not told their sequence of exposures (a
32 single-blind protocol). All subjects were sedentary during exposure except that men exercised
33 on a bicycle ergometer for 6 minutes at 1 and 5 hrs of exposure.
2 The Romberg test measures CNS depression.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-59 DRAFT-DO NOT CITE OR QUOTE
-------
1 Reports of symptoms (e.g., headache) varied among individuals but, overall, complaints
2 during exposures were similar to those during exposures to 0 ppm of tetrachloroethylene. All
3 subjects were able to detect the odor of tetrachloroethylene at all levels of exposure immediately
4 upon entering the chamber; thereafter, they varied in their ability to detect odors. Some subjects
5 retained the ability to detect odors during the entire experimental period, particularly at 155 ppm.
6 A few other subjects detected no odor after the first hour of the first day.
7 The evaluation of EEG recordings made during exposure suggested altered patterns
8 indicative of cortical depression in three of four men and four of five women exposed to 100
9 ppm (constant or TWA). In five subjects, altered EEG recordings occurred during hours 4 to 7
10 of exposure; another subject had altered recordings within 10 minutes of exposure, which
11 gradually returned to normal during continued exposure, and the seventh subject showed changes
12 between 30 minutes and 6-7 hrs of exposure. Recordings of VEPs in response to bright flashes
13 of light (i.e., neurophysiological measurements of the electrical signals generated by the visual
14 system in response to visual stimuli) and equilibrium tests (Romberg and heel-to-toe) were
15 normal in men and women.
16 The performance of men on neurobehavioral tests of cognitive function (arithmetic),
17 motor function (alertness), motor/cognitive function (inspection), and time estimation were not
18 significantly affected by any exposure. The performances of men on a second test of motor
19 function (Flanagan coordination) were significantly decreased (p < 0.05) on 1 of 3 days during
20 each of 2 weeks of exposure to 100 ppm and on 2 of 3 days during the week of exposure to 155
21 ppm, but the investigators concluded that only the results at 155 ppm were related to
22 tetrachloroethylene. In women, alertness (the only neurobehavioral endpoint evaluated) was not
23 affected by exposure to tetrachloroethylene.
24 The study authors concluded that (1) there is considerable interindividual variation in
25 response to tetrachloroethylene vapors, (2) EEG analysis indicates preliminary signs at narcosis
26 in most subjects exposed to 100 ppm for 7.5 hrs, (3) impairment of coordination may occur in
27 subjects exposed to 155 ppm for 7.5 hrs, and (4) the effects are likely due to tetrachloroethylene
28 itself, given its slow metabolism in humans. They also reported that their data suggested that a
29 threshold limit value of 100 ppm contains no margin of safety for susceptible subjects—both
30 subjectively and neurologically—to the vapors of tetrachloroethylene, a surprising finding, given
31 the study's sample size.
32
33 4.6.1.1.3. Hake, C.L., R.D. Stewart, A. Wu, J. Kalbfleisch, RE. Newton, S. K. Marlow andM.
34 Vucicevic-Salama. 1977. Effects of perchloroethylene/dmg interaction on behavior and
35 neurological function. DHEW (NIOSH) Publ No. 77-191. Stewart et al. (1977) conducted a
36 complex study with 12 healthy adults (6 men and 6 women) on the behavioral effects of inhaled
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-60 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene and oral doses of alcohol or Valium. Individuals were typically exposed for
2 6.5 hrs to 0 ppm on Monday or Tuesday, 100 ppm on Wednesday and Friday, and 25 ppm on
3 Thursday during each of the 11 weeks of exposure and were given a placebo capsule, alcohol,
4 Valium, or nothing during each period. Numerous neurological tests were performed throughout
5 each exposure, and all testing was done in a double-blind mode (neither the testers nor the
6 subjects were told the exposure level).
7 Exposure to 25 or 100 ppm of tetrachloroethylene for 6.5 hrs did not increase the overall
8 prevalence of reported symptoms (e.g., headache) or alter the subjects' mood. There were
9 exposure-related increases in the strength and persistence of the tetrachloroethylene odor
10 perceived by the subjects. Exposure did not significantly reduce performance on two
11 equilibrium tests (Romberg and heel-to-toe) and two neurobehavioral tests of motor function
12 (Michigan eye-hand coordination test and rotary pursuit test). At 100 ppm (but not 25 ppm)
13 there was a significant decrease (p < 0.05) in scores on a third test of motor function (Flanagan
14 coordination test) on some days of exposure. Statistical analyses performed by the investigators
15 revealed no effect of tetrachloroethylene exposure alone on EEGs and no significant interactive
16 effects between tetrachloroethylene and either alcohol or Valium.
17 The study authors reported that exposure to 100 ppm tetrachloroethylene did not have a
18 significant consistent effect on performance, although it did have a significant but inconsistent
19 detrimental effect on the performance of the Flanagan coordination test (given during the 3rd and
20 4th hrs of exposure).
21
22 4.6.1.1.4. Altmann, L., A. Bottgor andH. Wiegand. 1990. Neurophysiological and
23 psychophysical measurements reveal effects of acute low-level organic solvent exposure in
24 humans. Int. Arch. Occup. Environ. 3:493-499. Altmann, L., H. Wiegand, A. Bottger, F.
25 Eistwmelor and G. Winneke. 1992. Neurobehavioral and neurophysiological outcomes of
26 acute repeatedperchloroethylene exposure. Appl. Psych. 41:269-279. This study, conducted in
27 Germany, reports intentional inhalation exposure of human subjects to tetrachloroethylene for
28 the purpose of measuring potentially adverse health outcomes. No information is provided about
29 ethical principles espoused by the U.S. government for exposure to human subjects. Therefore,
30 the principal author was contacted by EPA staff (e-mail communication from R. McGaughy,
31 EPA to L. Altmann, October 8, 2003). Information was requested regarding procedures that
32 were used to select the subjects and inform them about the nature of the exposure, institutional
33 procedures that were taken to review the design of the study, and ethical standards and guidelines
34 that the institution was operating under at the time of the study. No response had been received
35 by EPA staff as of October 19, 2004. Although the report is not of crucial importance in the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-61 DRAFT-DO NOT CITE OR QUOTE
-------
1 evaluation of chronic neurotoxic effects of tetrachloroethylene, the scientific and ethical issues
2 associated with intentional dosing of human subjects is of importance to EPA.
3 Altmann et al. (1990, 1992) used neurophysiological and neurobehavioral techniques to
4 evaluate the neurological effects of tetrachloroethylene on healthy adults exposed to 10 ppm or
5 50 ppm for 4 hrs on 4 consecutive days. All subjects denied prior occupational exposure to
6 solvents and drug use at the time of the study. Visual acuity of all subjects was normal or
7 corrected to normal. The study was a single-blind study (subjects were not told their level of
8 exposure), and subjects were randomly assigned to either group. Sixteen subjects were exposed
9 to 10 ppm and 12 subjects were exposed to 50 ppm, but neurophysiological measurements were
10 made on only 22 subjects (12 at the low level and 10 at the high level). There was no unexposed
11 control group.
12 Three neurophysiological measurements were taken on the day before exposure started
13 and on each of the four exposure days: (1) VEPs in response to black-and-white checkerboard
14 patterns were measured; the VEPs of some subjects (exact number not reported) were also
15 measured on the day after exposure ceased; (2) a visual contrast sensitivity (VCS) test (a test of
16 the central spatial vision that determines the minimum contrast necessary for an individual to see
17 patterns of various sizes) was given to five subjects (three from the low-exposure group and two
18 from the high-exposure group); (3) recordings of brainstem auditory-evoked potentials
19 (neurophysiological measurements of the electrical signals generated by the hearing system in
20 response to auditory stimuli) were made to evaluate peripheral hearing loss. All measurements
21 were started 2 hrs after a subject entered the chamber and were completed within 1 hr.
22 A German version of the Neurobehavioral Evaluation System was used to assess motor,
23 motor/cognitive, and cognitive function of subjects. The battery included nine tests (finger
24 tapping, eye-hand coordination, simple reaction time, continuous performance, symbol digit,
25 visual retention, pattern recognition, digit span, and paired associates). A vocabulary test and a
26 test of emotional state (moods) were also given. Each subject was assessed with a complete
27 battery of tests during the pre-exposure baseline assessment and at the end of the study. Subsets
28 of the battery covering motor function and mood were given repeatedly at the beginning and end
29 of each 4-hr exposure period.
30 Tetrachloroethylene was not detected in blood samples collected before the start of the
31 first exposure period. The detection limit was less than 0.0005 mg/L. Mean tetrachloroethylene
32 blood levels increased slightly over the 4-day period. Among subjects exposed to 10 ppm, mean
33 blood levels were 0.33, 0.36, 0.4, and 0.38 mg/L at the end of days one, two, three, and four of
34 exposure, respectively. Among subjects exposed to 50 ppm, mean blood levels were 1.1, 1.2, 1.4,
35 and 1.5 mg/L at the end of days one, two, three, and four of exposure, respectively.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-62 DRAFT-DO NOT CITE OR QUOTE
-------
1 On the first day of testing, a faint solvent odor was reported by 33% and 29% of the
2 subjects exposed to 10 ppm and 50 ppm, respectively. On the fourth day, these incidences
3 changed to 17% and 36%, respectively. The VEP latencies of subjects during the 3rd hr of
4 exposure to 50 ppm on days one, two, three, and four of exposure were significantly longer
5 (p< 0.05) from those measured on the control day, and the differences became progressively
6 longer on successive exposure days. One set of VEP latencies on the day after the end of the
7 exposure period remained longer than the control day values (statistical significance not
8 reported). VEP latencies in subjects with exposure to 10 ppm were not statistically significantly
9 longer than those recorded on the control day. There were significant differences (p < 0.05)
10 between the VEP latencies of subjects exposed to 10 ppm and those exposed to 50 ppm.
11 Data on contrast sensitivity indicated greater effects at 50 ppm than at 10 ppm; effects
12 were most pronounced on the last day of exposure. However, statistical analysis was not
13 reported, and the data are limited by the small number of subjects. There were no indications of
14 peripheral hearing loss at either exposure level.
15 Neurobehavioral tests results were reported for only those tests given repeatedly on 4
16 consecutive days (finger tapping, eye-hand coordination test, simple reaction time, continuous
17 performance, and moods). There were significant post-exposure performance deficits (p 0.05)
18 among subjects exposed to 50 ppm when compared with the group exposed to 10 ppm in tests of
19 motor/cognitive function (continuous performance test for vigilance) and motor function (eye-
20 hand coordination), and a near-significant difference (p = 0.09) on a test of motor function
21 (simple reaction time). In all cases, the degree of improvement shown by the subjects exposed to
22 50 ppm was less than that shown by the subjects exposed to 10 ppm. There were no exposure-
23 related effects on the finger-tapping or moods test.
24 The study authors concluded that visual function in healthy, young, adult males is mildly
25 affected by tetrachloroethylene exposures to 50 ppm maintained for 4 hrs on each of 4 days
26 (Altmann et al., 1990), and they stated that the impaired performance on tests of motor/cognitive
27 and motor function suggests that 50 ppm cannot be considered a NOAEL for neurobehavioral
28 end-points indicative of CNS depression (Altmann et al., 1992).
29
30 4.6.1.2. Chronic Exposure Studies
31 4.6.1.2.1. Lauwerys, R., J. Herbrand, J.P. Bucket, A. Bernard and J. Gaussin. 1983. Health
32 surveillance of workers exposed to tetrachloroethylene in dry-cleaning shops. Int. Arch. Occup.
33 Environ. Health. 52:69-77. Lauwerys et al. (1983) studied 263 workers (24 women and 2 men)
34 occupationally exposed to tetrachloroethylene in six dry cleaning shops in Belgium for a mean of
3
Abstract of paper reports 22 subjects were exposed to perchloroethylene.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-63 DRAFT-DO NOT CITE OR QUOTE
-------
1 6.4 years. The authors evaluated potential effects on the CNS, the kidney, the liver, and the
2 lungs in these workers and in controls (31 women and 2 men) working in a chocolate factory
3 (20) or an occupational health service (13) who did not report occupational exposure to organic
4 solvents. No information is provided in the paper on the methods used to identify subjects or
5 their reasons for participating in the study. Several characteristics of the two groups were similar
6 (sex ratio, mean age [32.9 vs. 34.5 years], and level of education). However, 13 of the 26 dry
7 cleaning workers—but only 9 of the 33 controls—were smokers. Neurobehavioral tests of motor
8 function (simple and choice reaction time), sensory function (critical flicker fusion), and
9 cognitive function (sustained attention test) were given twice to each worker, once before work
10 and once after work. Both groups were tested in the middle of the work week. Individuals also
11 were questioned about chronic symptoms related to nervous system disturbances. Blood samples
12 were collected both before and after work. Urine samples for kidney function tests were
13 collected after work.
14 The mean tetrachloroethylene air concentration (8-hr TWA) was 21 ppm and the range of
15 TWA values was 9 to 38 ppm, using results from active sampling of personal air. The mean
16 tetrachloroethylene blood level (30 minutes after the end of work) was 1.2 mg/L (range of means
17 from the shops was 0.6 to 2.4 mg/L). There was no significant connection between air
18 concentrations and blood levels. Trichloroacetic acid, a metabolite of tetrachloroethylene, was
19 not detected (LOD not identified in published paper) in urine specimens from exposed subjects.
20 Seventeen of 22 symptoms related to nervous system disturbances were reported by study
21 investigators as being more prevalent among the workers than among unexposed controls.
22 Although statistical testing was conducted, the study investigators did not describe statistical
23 methods or tests in the published paper. The investigators reported no statistically significant
24 differences in prevalences for most symptoms and no relationships with duration of exposure.
25 Lauwerys et al. (1983) reported more complaints in the exposed group than in control workers,
26 particularly memory loss (7/26 exposed vs. 3/33controls) and difficulty falling asleep (11/26
27 exposed vs. 6/33 controls). EPA analysis of the data found the latter complaint to be statistically
28 significant using Fisher's exact test (p = 0.04). The mean cycles/second, the score of the critical
29 flicker fusion test (a test of sensory function) was significantly increased (better performance) in
30 the exposed group than in controls when given both before work and after work. The dry
31 cleaning workers did not differ from controls on the other three neurobehavioral tests. The
32 prevalence of abnormal scores (those beyond the 5th or 95th percentile of the control group) did
33 not vary significantly between the two groups.
34
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-64 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.6.1.2.2. Seeber, A. 1989. Neurobehavioral toxicity of long-term exposure to
2 tetrachloroethylene. Neurotoxicol. Teratol. 11:579—583.4 Seeber (1989) evaluated the
3 neurobehavioral effects of tetrachloroethylene on 101 German dry cleaning workers (machine
4 operators, ironers, touch-up workers, counter attendants, and other employees) who were
5 employed in coin-operated or while-you-wait shops, all affiliated with one organization. The
6 workers were separated into a low-exposure group (50 women, 7 men) and a high-exposure
7 group (39 women, 5 men). A third group of 84 sales personnel (64 women, 20 men) from
8 several department stores and receptionists from large hotels served as unexposed controls. No
9 information was provided on the methods used to identify subjects or their reasons for
10 participating in the study, although the authors reported that 29 service technicians were
11 excluded from the study because of either discontinuous exposure conditions with peak
12 concentrations or long periods of no exposure. Predominant characteristics of both groups
13 included primarily standing work, contact with customers, and moderate physical exercise. The
14 mean ages of the low-exposure, high-exposure, and control groups were 38.2, 38.4, and 31.8
15 years, respectively.
16 Details on air monitoring methods were sparsely reported, but mean tetrachloroethylene
17 concentrations (8-hr TWA) for the low- and high-exposure groups were 12 (+8) ppm and 53
18 (±17) ppm, respectively, using results from active sampling of room air and passive sampling of
19 personal air. The mean duration of occupational exposure for the low- and high-exposure groups
20 was 11.8 and 10.6 years, respectively.
21 A number of tests of neuropsychological functioning were administered, including
22 standardized tests of symptoms and personality; tests of sensorimotor function, including finger
23 tapping and aiming; and the Mira and Santa Ana dexterity tests, which are published
24 standardized tests. Threshold of perceptual speed was assessed by recognition of stimuli flashed
25 briefly on a screen; whether this procedure used a standardized instrument was not noted.
26 Choice reaction time was also determined using "nine light and tone stimuli." It is not clear
27 whether the auditory and visual stimuli occurred together or whether some trials consisted of an
28 auditory stimulus and others a visual stimulus. Details of the timing of the stimulus presentation
29 were not provided. One of the response variables, "delayed reactions," was not defined. The
30 typical dependent variable measured in this task—response reaction time—apparently was not
31 measured; only the number of correct reactions was reported. Subtests of the Wechsler
32 Intelligence Test (digit span, digit symbol, and cancellations) were used, as was recognition of
33 words, faces, and digits. The instrument used and the scoring of the last three tests were not
4 Dr. Seeber provided additional information on this study in written correspondence to the New York State
Department of Health dated January 19 and May 20, 1996. This information appears in NYS DOH (1997).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-65 DRAFT-DO NOT CITE OR QUOTE
-------
1 described. Intelligence was assessed using the logical thinking subtest of the German
2 Performance Test System.
3 Each subject was examined during a 1.5-day stay at a clinic located at a large institute for
4 occupational medicine. Each subject came to the clinic in the evening hours, stayed overnight,
5 and started the examination and testing process the next morning. The clinic examined
6 numerous people daily, and the dry cleaners and the control group took up only a small part of
7 the daily routine of the clinic staff. Neurobehavioral tests were given by two specialized clinic
8 staff who did not question the subjects about exposure status. However, clinic psychologists (six
9 at the time of the study) did inquire about the exposure and living conditions of the subjects.
10 Because the dry cleaner groups and the control group differed in gender ratios, age, and scores
11 on the intelligence test, stratified analysis was used to statistically control the influence of these
12 confounding factors on test scores. As discussed in the section that describes Altmann et al.
13 (1995), the use of dichotomous or categorical variables may not fully control for confounding
14 effects of these factors on the endpoint. The groups also differed in alcohol consumption, so a
15 separate analysis was used to examine the role of alcohol on effects associated with
16 tetrachl oroethy 1 ene.
17 Performance of both the low-exposure and high-exposure groups differed significantly
18 (p< 0.01) from that of the unexposed control group on the threshold of perceptual speed and
19 "delayed responses" on a choice reaction time task, both of which are measures of information
20 processing speed (p = 0.08 and 0.03 for low exposure and high exposure, respectively). Both
21 exposed groups also had worse scores (p < 0.01) on two tests of attention (digit reproduction and
22 digit symbol) and on visual scanning (cancellations). Group mean scores for digit reproduction
23 and digit symbol did not appear to increase from the low-exposure to the high-exposure group.
24 The low-exposure group also showed significantly higher scores than did the control group on
25 questionnaires, on neurological signs (p < 0.01), and emotional liability (p < 0.05). Scores of the
26 high-exposure group for these measures appeared higher than those for the control group;
27 however, the scores did not show a statistically significant difference. There were no differences
28 between groups on the other tests. Controlling for group differences in alcohol consumption did
29 not alter any test results.
30
31 4.6.1.2.3. Cm, S.X., M. Y. Huang, Z. Chen, Y. T. Liu, C. Jin, T. Watanabe, H. Nakatsuka, K.
32 Seiji, O. Inoue andM. Ikeda. 1991. Subjective symptom increase among dry-cleaning workers
33 exposed to tetrachloroethylene vapor. Ind. Health. 29:111-121. Cai et al. (1991) evaluated the
34 CNS effects of tetrachloroethylene exposure among 56 dry cleaning workers (27 women and 29
35 men) from three shops in China. The control group (37 women and 32 men) were of similar
36 mean age (34 years vs. 35 years for dry cleaning workers), but the male dry cleaning workers
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-66 DRAFT-DO NOT CITE OR QUOTE
-------
1 were 4 years younger than the male controls and the women were 4.9 years older than the female
2 controls. The controls were recruited from the same factories as the dry cleaning workers but
3 from workshops without known solvent exposures. No information is provided in the paper on
4 the methods used to identify subjects or their reasons for participating in the study. Further, no
5 information was provided on test procedures or the questionnaire used to assess subjective
6 symptoms. The geometric mean tetrachloroethylene air concentration (8-hr TWA) was 20 ppm
7 and the range of TWA values was 4 to 97 ppm, using results from passive sampling of personal
8 air. The mean duration of occupational (tetrachloroethylene) exposure was 3 years.
9 The prevalence of symptoms of tetrachloroethylene exposure was significantly higher
10 among the dry cleaning workers (men, women, and men and women combined; p < 0.001) than
11 among the unexposed controls. Five symptoms (dizziness, drunken feeling, floating sensation, a
12 heavy feeling in the head, and facial flushes) in men and women (combined) were significantly
13 more prevalent in the dry cleaning workers than in the controls (p < 0.001). Nasal irritation and
14 unusual smell were also reported significantly more often by the dry cleaning workers than by
15 controls (p < 0.05). Similar findings were reported when the workers were asked about the
16 symptoms they had noticed during the 3 months before the study. The investigators found
17 exposure-related increases in the prevalence of subjective symptoms among dry cleaning
18 workers exposed to 21 ppm (8-hr TWA).
19
20 4.6.1.2.4. Nakatsuka, H., T. Watanabe, Y. Takeuchi, N. Hisanaga, E. Shibata, H. Suzuki, M. Y.
21 Huang, Z. Chen, Q.S. Qu andM. Ikeda. 1992. Absence of blue-yellow color vision loss among
22 workers exposed to toluene or tetrachloroethylene, mostly at levels below occupational
23 exposure limits. Int. Arch. Occup. Environ. Health. 64:113-117. Nakatsuka et al. (1992)
24 evaluated the effects of tetrachloroethylene exposure on the color vision of 64 dry cleaning
25 workers (34 women and 30 men) in China. The workers were from the same shops studied by
26 Cai et al. (1991). Control workers (72 women and 48 men) were recruited from the clerical
27 sections of dry cleaning shops and from other factories (paint production plants or plants
28 producing tetrachloroethylene from trichloroethylene). No information is provided in the paper
29 on the methods used to identify subjects or their reasons for participating in the study. The mean
30 ages of the dry cleaning workers (34 years for men, 35 years for women) were lower than those
31 of the controls (34 years for men, 33 years for women). A screening color test (Lanthony's new
32 color test) and a test used for confirmation of red-green vision loss were carried out by
33 ophthalmologists or occupational health doctors in charge of the factories under one of two
34 lighting conditions (natural sunlight or a daylight fluorescent light). The published report does
35 not identify what procedure was used on which test; illumination is a critical component in
36 administering color vision tests to subjects (Geller and Hudnell, 1997).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-67 DRAFT-DO NOT CITE OR QUOTE
-------
1 The geometric mean air concentrations of tetrachloroethylene (averaging time not
2 reported) were 16 and 11 ppm for the men and women, respectively, using results from passive
3 sampling of personal air. The overall geometric mean was 13 ppm.
4 There was no significant difference in the performance of the dry cleaning workers and
5 unexposed controls on Lanthony's new color test. The study authors reported that the
6 percentages of dry cleaning workers who correctly separated colored caps from monochromatic
7 caps were not significantly different from the percentages in the corresponding control group. A
8 statistical analysis of these data reported in public comments of the Halogenated Solvents
9 Industry Alliance to EPA (HSIA, 2004) on the Neurotoxicity of Tetrachloroethylene Discussion
10 Paper (U.S. EPA, 2003b) showed—using a chi-square test for differences in proportions—that
11 tetrachloroethylene-exposed women were more likely to have normal color vision as compared
12 with unexposed women. An EPA analysis of male workers did not show any differences, either
13 better color vision or worse color vision, in exposed males compared with unexposed male
14 controls. Nakatsuka et al. concluded, overall, that they found no distinct case of color vision loss
15 among the dry cleaning workers.
16
17 4.6.1.2.5. Ferroni, C., L. Selis, A. Mutti, D. Folli, E. Bergamaschi and I. Franchini. 1992.
18 Neurobehavioral and neuroendocrine effects of occupational exposure to perchloroethylene.
19 Neurotoxicol. 13:243—247.5 Ferroni et al. (1992) evaluated neuroendocrine and neurobehavioral
20 effects of tetrachloroethylene exposure among 60 female dry cleaners and 30 unexposed female
21 controls who were comparable in age (mean ages 39.7 and 37.6 years, respectively) and
22 vocabulary level. Each dry cleaning shop in a small town outside of Parma, Italy was visited.
23 The workers were invited to participate in the study, which was part of a preventive health
24 program implemented by the local health office and professional associations of small businesses.
25 There were no refusals. Controls were selected from the workers at a hospital who cleaned
26 clothes using a water-based process. Their jobs were essentially the same as those of the dry
27 cleaners, but they were not exposed to any organic solvents. Both groups filled out a
28 questionnaire on their health status, medication (including oral contraceptives), lifestyle, and
29 current and past jobs. Both groups met the following criteria: no history of metabolic disorders,
30 no history of psychiatric disorders, and low level of daily alcohol intake. The two groups were
31 similar in height, weight, body mass index, smoking habits, and use of medication, but alcohol
32 intake was about 5% higher (p < 0.03) in the control group than in the dry cleaner group.
5 Dr. Mutti provided details on the selection process of exposed and control subjects and also clarified
reported results to Dr. Ken Bodgen, New York State Department of Health, in written correspondence dated July 29
and September 5, 1995 (see NYS DOH, 1997).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-68 DRAFT-DO NOT CITE OR QUOTE
-------
1 Workers and controls were given five neurobehavioral tests (part of the Swedish
2 Performance Evaluation System, "adapted" Italian version: finger tapping with both dominant
3 hand and nondominant hand, simple reaction time, digit symbol test, shape comparison-vigilance,
4 and shape comparison-response to stress). All subjects were examined in the morning before
5 their work shift in the same room by the same examiners (NYS DOH, 1997). The tests were part
6 of a computer-based battery, and the same machines and software were used to administer the
7 tests and score the results. The same sequence of tests and protocols were used for all subjects.
8 Although the examiners were not blind to the status of the subjects (dry cleaner or control), they
9 were blind to the worker's exposure level (NYS DOH, 1997). Serum prolactin levels were
10 measured in all subjects; blood samples were collected during the working day during summer
11 and winter. Prolactin secretion by the pituitary is under control by hypothalamic dopamine;
12 dopamine is also important to neurotransmitter systems.
13 One proposed alternative for assessment of nervous system toxicity is the study of
14 biochemical signals in peripheral tissues as biomarkers of nervous system function (Manzo et al.,
15 1996). Samples from dry cleaners and controls were alternated and analyzed in the same
16 experimental runs. For women, only those samples obtained during the proliferative phase of the
17 menstrual cycle were used for comparison between groups (41 dry cleaners and 23 controls).
18 Workplace air samples were randomly collected throughout the work week during
19 summer and winter to account for variability related to either the work cycle or seasonal
20 environmental fluctuations. The median tetrachloroethylene air concentration (4-hr TWA) was
21 15 ppm and the range of TWA values was 1 to 67 ppm. The subjects' range of
22 tetrachloroethylene blood levels was 0.012 to 0.864 mg/L (median = 0.145 mg/L; incorrectly
23 expressed in Ferroni et al., 1992, as 12,864 and 145 mg/L [NYS DOH, 1997]). The mean
24 duration of occupational exposure was 10 years.
25 The dry cleaners showed significantly reduced performance when compared with the
26 unexposed matched controls in three tests (simple reaction time, p < 0.0001; vigilance,
27 p< 0.005; and stress,/? < 0.005), as reported by Ferroni et al. (1992). Performance on the fmger-
28 tapping test (both hands) and digit symbol test was not affected (NYS DOH, 1997). Additionally,
29 the mean serum level of prolactin was significantly higher in the workers than in the matched
30 controls (p < 0.001). None of the three measures of exposure (duration of exposure and air or
31 blood concentration of tetrachloroethylene) was significantly associated with decreased test
32 scores or increased serum prolactin levels among the dry cleaners.
33 The study authors concluded that tetrachloroethylene exposure in dry cleaning shops may
34 impair performance and affect pituitary function but that the cross-sectional design prevented
35 distinguishing acute effects from chronic effects. Ferroni et al. (1992) also reported that the most
36 likely bias of cross-sectional studies is a spontaneous selection of the sample (i.e., workers who
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-69 DRAFT-DO NOT CITE OR QUOTE
-------
1 believe exposure is making them sick or workers who actually become sick may quit work
2 prematurely and not be included in the study) and, as a result, the actual risk was likely to be
3 underestimated rather than overestimated, although no data are presented in the paper with which
4 to evaluate this statement. On the other hand, the exposed and unexposed study population of
5 women was tested during the proliferation phase of menstruation, which may better capture
6 changes in prolactin secretion, but also may potentially confound findings if there are individual
7 differences in severity of menstruation and in the timing of test session relative to the day of
8 menstruation (U.S. EPA, 2004).
9
10 4.6.1.2.6. Cavalleri, A., F. Gobba, M. Paltrinieri, G. Fantuzzi, E. Righi and C.L. Aggazzoti.
11 1994. Perchloroethylene exposure can induce colour vision loss. Neuroscience Lett. 179:162-
12 166.6 Gobba, F., E. Righi, G. Fantuzzi, G. Predieri, L. Cavazzuti and G. Aggozzotti. 1998.
13 Two-year evaluation of perchloroethylene-induced color-vision loss. Arch. Environ. Health
14 53:196-198. Cavalleri et al. (1994) reported on a control-matched, cross-sectional,
15 observational, occupational study that evaluated the effects of tetrachloroethylene exposure on
16 the color vision of dry cleaners. The investigators compiled a list of all the dry cleaning shops in
17 the municipality of Modena, Italy, (110 shops employing 189 workers) and randomly selected 60
18 dry cleaners from 28 premises for recruitment into the study (Aggazzotti et al., 1994a). Only
19 full-time workers (n = 52) were asked to participate, and two declined. No information is
20 provided in the paper on a subject's motivation for participating or not participating in the study.
21 All 50 workers provided, via questionnaires, information on work history, health status,
22 occupational and hobby use of solvents, drinking and smoking habits, and drug use. Thirty-five
23 of the 50 dry cleaners (33 women, 2 men) met the inclusion criteria; others were excluded for
24 hypertension, smoking more than 30 cigarettes a day, alcohol consumption exceeding 50 g of
25 alcohol a day, oculo-visual pathology, or working less than 1 year. Another worker was
26 excluded because a matched control could not be found.
27 The controls were factory workers who were not occupationally exposed to solvents or
28 other neurotoxic chemicals; they were selected and recruited into the study using the same
29 methods that were used for dry cleaners. The controls (n = 35) were from factories in the
30 Modena area and met the same inclusion criteria as the dry cleaners. They were matched to dry
31 cleaners by gender, age (+3 years), alcohol consumption (+10 g/day), and cigarette use (+5
32 cigarettes a day). The mean age of both groups (35 years) and the percentages of each group that
33 were smokers (43%) or alcohol drinkers (71%) were comparable.
6 Dr. Cavalleri provided additional information on this study in written correspondence to the New York
State Department of Health dated October 8, 1996 (see NYS DOH, 1997).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-70 DRAFT-DO NOT CITE OR QUOTE
-------
1 All subjects appeared healthy and met minimal status of visual acuity. None of the
2 subjects reported hobby exposure to solvents or other substances toxic to the eye. There were no
3 known systematic differences between exposed and control groups or between machine
4 operators and ironers.
5 Color vision was assessed using Lanthony's D-15 desaturated panel test, in which
6 subjects are asked to put a series of small round "caps" in order by color. The types of errors
7 made can distinguish specific types of color vision deficiency; e.g., red-green color confusion
8 errors (blindness) is a common condition in males, mostly but not entirely of congenital origin,
9 whereas blue-yellow color confusion errors are very rarely due to congenital conditions and
10 therefore are considered as a hallmark of an acquired condition. Impairments in color vision,
11 beginning as blue-yellow confusion errors, have been reported in numerous populations exposed
12 to organic solvents (Mergler and Blain, 1987; Mergler et al., 1987, 1988a, b, 1991; Campagna et
13 al., 1995, 1996). Test scores are based on the ability of each subject to arrange a set of 15 caps
14 colored with desaturated colors according to a definite chromatic sequence, with each mistake
15 increasing the score above a perfect score of 1.00. A formula (the Color Confusion Index [CCI])
16 is used to calculate total errors.
17 Exposed and control subjects were tested in a random order (NYS DOH, 1997). All
18 subjects were tested at the same time of day (in the morning, before work) under the same
19 lighting conditions by the same investigator. With respect to exposed subjects, the investigator
20 was unaware of both the exposure levels and the job (operator or ironer) of each dry cleaner.
21 For all dry cleaners, the mean tetrachloroethylene air concentration (8-hr TWA) was 6
22 ppm and the range of TWA values was 0.4-31 ppm, using results from passive sampling of
23 personal air. For operators (n = 22), the mean air concentration 8-hr TWA was 7 ppm and the
24 range of TWA values was 0.4-31 ppm. For ironers (n = 13), mean air concentration (8-hr TWA)
25 was 5 ppm and the range of TWA values was 0.5-11 ppm. The mean duration of occupational
26 exposure was 8.8 years. Tetrachloroethylene concentrations were also measured in alveolar air
27 for a subset of these dry cleaners, with a high correlation observed between tetrachloroethylene
28 concentration in alveolar air and the 8-hr TWA levels in ambient air (r = 0.8, p < 0.001;
29 Aggazzotti et al., 1994a).
30 Only three dry cleaning workers, as opposed to 13 controls, scored a perfect test score
31 (p < 0.01). Mistakes were made mainly in the blue-yellow range. Overall, the workers showed
32 poorer performance on the test as compared to controls, and they had a significantly higher mean
33 CCI using a Student's t-test (p = 0.03). The effect was statistically significant among operators
34 but not among ironers. Study investigators also evaluated whether CCI values were normally
35 distributed, which is important if using a Student's t-test, but they did not present any
36 information about the result of the Kolmogorov-Smirnov test. The observation for ironers may
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-71 DRAFT-DO NOT CITE OR QUOTE
-------
1 reflect a lower statistical power in this group due to fewer subjects (13 ironers vs. 22 operators).
2 There also was a statistically significant positive correlation (p < 0.01) between TWA air
3 concentrations and the CCI (r = 0.52), which remained after multivariate analysis considered
4 previous tetrachloroethylene exposure duration, age, number of cigarettes a day, and daily intake
5 of alcohol as covariates.
6 The effect on color vision may not be rapidly reversible; preliminary data showed that the
7 scores of some workers did not improve when retested after 4 weeks of vacation (NYS DOH,
8 1997). Moreover, some of these workers showed poorer performance on this test in the follow-
9 up study by Gobba et al. (1998), described below, suggesting color vision impairment as a
10 chronic effect. The CCI values were not associated with two other measures of
11 tetrachloroethylene exposure (mean duration and an integrated index of exposure, yearly TWA
12 level). The study authors suggested that this may reflect the difficulty in controlling for the
13 interactive effects of age and exposure and accurately evaluating exposure.
14 Gobba et al. (1998) reexamined color vision after a period of 2 years in 33 of the 35 dry
15 cleaners and ironers examined by Cavalleri et al. (1994). Two subjects had retired during the
16 2-year period between examinations. These investigators used the Lanthony D-15 test, the test
17 used by Cavalleri et al. (1994) to assess color vision, and performance was compared with the
18 subject's score from the initial survey (self-control). Tetrachloroethylene concentration in the
19 occupational setting was determined in the breathing zone using personal passive samplers.
20 Monitoring was carried out during the afternoon shift, as Cavalleri et al. (1994) did not show any
21 differences between morning and afternoon samples. Gobba et al. (1998) found that
22 tetrachloroethylene concentration had increased during the 2-year period for 19 subjects,
23 identified as Group A, (geometric mean, from 1.67 ppm at the first survey to 4.35 ppm at the
24 second survey) and had decreased for 14 subjects, identified as Group B (geometric mean, from
25 2.95 ppm to 0.66 ppm). For the 33 workers overall, tetrachloroethylene concentration did not
26 change over the 2-year period (geometric mean, from 2.4 ppm to 1.94 ppm at the second survey,
27 /?>0.05).
28 Color vision deteriorated between the two surveys for the entire group, a reflection of the
29 color vision loss among Group A subjects, whose exposure had increased by the second survey.
30 As found in the first survey, color vision was impaired primarily in the blue-yellow range of
31 color, with few subjects presenting a red-green deficit. Color vision performance for the entire
32 group was related significantly to age (r = 0.45) and tetrachloroethylene concentration (r = 0.39).
33 The mean CCI score for Group A subjects showed a statistically significant difference between
34 the two surveys (p < 0.05). Analysis of variance methods that controlled for an effect of age
35 further supported the finding of color vision deterioration among these subjects. For Group B
36 subjects, who experienced lower exposure concentrations by the second survey, the CCI score
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-72 DRAFT-DO NOT CITE OR QUOTE
-------
1 did not change from that of the initial survey. The findings in Groups A and B were also
2 supported using analysis of variance methods that adjusted for age, alcohol consumption, or
3 cigarette smoking between the subgroups.
4
5 4.6.1.2.7. Echeverria, D., N. Hyer, J.A. Bitner, H. Checkoway, G. Toutonghi andN.
6 Ronhovde. 1994. Behavioral effects of low level exposure to perchloroethylene (PCE) among
1 dry cleaners. In: Battelle Centers for Public Health Research and Evaluation. A behavioral
8 investigation of occupational exposure to solvents: perchloroethylene among dry cleaners and
9 styrene among reinforced fiberglass laminations. Final Report SSRC-10OM4/040. Seattle,
10 WA: pp. 6.1—6.37. Echeverria et al. (1994) reported a study designed to evaluate a hypothesis in
11 a previous study (Echeverria et al., 1995)7 of frontal/limbic system effects8 as the site underlying
12 tetrachloroethylene pathology, where degradation in behavior may be the earliest indicator of
13 acute, subchronic, or chronic neurotoxicity. The study was conducted in the Seattle/Tacoma,
14 Washington area from 1989 through 1993, when the area's dry cleaning industry was switching
15 from wet-transfer to dry-to-dry machines. Initially, 320 dry cleaning shops and laundries were
16 sent introductory letters requesting permission to allow their employees to participate in the
17 study. Of the 181 owners who responded, 39 agreed to participate. Of the owners who did not
18 agree to participate, 22% expressed no specific reason, 19% cited time constraints, 17% feared
19 legislative reprisal from federal agencies, 17% did not speak English, 15% were unavailable or
20 never contacted, and 10% cited various other reasons.
21 Recruitment ended when a total of 45 operators were enrolled in the study (total n = 173).
22 Each operator was matched with a less-exposed person from the same shop. The subjects
23 included laundry workers (n = 69), pressers or counter clerks (n = 59), and operators or former
24 operators (n = 45). The mean ages of the groups were 42.5, 34.2, and 46.2 years, respectively.
25 Women comprised 63% of the study population (109/173). The subjects, who were paid
26 volunteers, were eligible if they spoke English, had no history of diabetes or CNS disorders, and
27 had worked for more than one year in the trade. The final sample excluded three subjects for
28 limited knowledge of English and reading skills and six subjects for not wearing glasses or
29 missing covariate information such as vocabulary performance on the test.
7 Although published a year after this study, the study by Echeverria et al. (1995), discussed in Section
4.6.1.2.8, was conducted in 1986, 3 years before this study.
8 Echeverria et al. (1994) hypothesized that exposure to solvents particularly affected attention, executive
function, visuospatial skill, short-term memory, and mood, leaving motor, language-based skills, and long-term
memory intact. The frontal system has been associated with executive function, such as that measured by Switching
Attention, Trailmaking A and B, and the Wisconsin Card Sort. Tests associated with the limbic system include
mood, as measured by the Profile of Mood States, short-term memory as measured by digit span, visual
reproductions, and pattern memory.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-73 DRAFT-DO NOT CITE OR QUOTE
-------
1 An index of chronic exposure and measures of subchronic and acute exposure were
2 developed for each subject. The chronic exposure index was based on a detailed work history,
3 including consideration of the type of dry cleaning machine, job title, percentage of time at each
4 job title, estimated air levels associated with each job title, and employment duration. The
5 measures of subchronic and acute current exposure were based on mean 8-hr TWA air
6 concentrations measured on the day of neurobehavioral testing. Mean chronic indices were zero
7 for the never-exposed group of laundry workers, 68 for the dry cleaning workers with low
8 exposure (pressers/clerks), and 1,150 for the dry cleaning workers with high exposure (operators).
9 Mean exposures (8-hr TWA, using results from passive sampling of personal air) for workers
10 placed in these chronic exposure categories were <0.2 ppm (laundry workers), 3 ppm
11 (pressers/clerks), and 9 ppm (operators). Dry cleaning workers placed in the chronic exposure
12 categories of low and high had been employed in their current job for 2.6 and 11 years,
13 respectively.
14 The subjects also were placed in acute and subchronic exposure categories of <1 ppm
15 (laundry workers and some dry cleaning workers, e.g., clerks), low (mainly pressers), and high
16 (operators), with corresponding current tetrachloroethylene 8-hr mean concentrations of 0.5, 3,
17 and 20 ppm. Dry cleaning workers placed in the low and high categories had been employed in
18 their current job for 5 and 9 years, respectively. Because of the changes in dry cleaning practices
19 over the course of the study, many subjects who were placed in the high chronic-exposure
20 category, which was based on detailed work history, were frequently found in the low acute- or
21 subchronic-exposure group, which was based on air concentrations on the day of testing.
22 The test battery included tests of cognitive function, including visuospatial ability, motor
23 skills, mood, CNS symptoms, and basic verbal and arithmetic skills. The chronic and subchronic
24 assessment was based on tests given during the morning of each subject's day off and on pre-
25 shift scores. Additional tests considered to have an acute component were given 1 hr before
26 work on the first day of the work week and again at the end of the day, allowing acute effects to
27 be examined by using pre-shift performance to predict post-shift performance and then
28 comparing predicted with observed performances. On their day off, the subjects were tested at
29 home. At the work site, subjects were tested in a minivan. Each subject signed a consent form,
30 provided a breath sample at each test session, and completed a questionnaire covering transient
31 factors that could affect performance (e.g., headache). This was followed by questionnaires on
32 medical history, medication, drug and alcohol use, occupational and nonoccupational exposure to
33 chemicals, symptoms, and mood. The subject was then administered the neurobehavioral tests.
34 Multivariate analysis was used to evaluate the relationship between exposure indices and
35 levels and performance on neurobehavioral tests after adjusting for the variables of age, gender,
36 race, vocabulary level (surrogate for education and test-taking), and alcohol consumption. After
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-74 DRAFT-DO NOT CITE OR QUOTE
-------
1 adjustment for those variables that were significant confounders, associations between increased
2 indices of chronic (lifetime) exposure and reduced test performance were found in three tests of
3 cognitive function: switching (p = 0.1), pattern memory (p = 0.03), and pattern recognition
4 (p = 0.09). The magnitude of change attributable to tetrachloroethylene was a 3% loss in
5 function for the latency of pattern memory and an 11% loss in function for the correct number in
6 visual reproductions; losses in function that are well within pre-clinical values. Subjective
7 measures of mood and symptoms were not significantly associated with exposure. Dry cleaning
8 workers scored lower (but not significantly) on all but one of the remaining tests (the digit span
9 test).
10 Analysis of the association between test scores and measures of subchronic exposure
11 (8-hr TWA tetrachloroethylene concentrations on the day of testing) confirmed the findings of
12 the chronic analysis: reduced scores on tests of switching (p = 0.1) and pattern recognition
13 (p = 0.04) as exposure increased. Analysis of effects of acute exposures showed no relationship
14 between workday exposure at any level and post-work performance on nine neurobehavioral
15 tests.
16 Echeverria et al. (1994) detected deficits in visuospatial function (reduced performance in
17 tests of pattern memory and pattern recognition) in the dry cleaning workers categorized as
18 having high lifetime chronic exposure and whose current exposure level was 9 ppm, 8-hr TWA.
19 However, the exposure level of 9 ppm was not considered representative of past chronic
20 exposure levels because the industry in the study area was switching from wet-transfer to dry-to-
21 dry machines during the study. The investigators attributed the reduced performance to prior
22 exposures that were about two to four times higher 3 to 5 years previously, and they
23 hypothesized that a few years of reduced exposure may not be long enough to eliminate the
24 residual effects on visuospatial skills caused by the exposures associated with wet-transfer
25 machines.
26
27 4.6.1.2.8. Echeverria, D., R.F. White and C. Sampaio. 1995. A behavioral evaluation ofPCE
28 exposure in patients and dry cleaners: a possible relationship between clinical andpreclinical
29 effects. J. Occup. Environ. Med. 37:667—680. Echeverria et al. (1995) assessed neurobehavioral
30 effects and mood disturbances in four patients diagnosed with tetrachloroethylene
31 encephalopathy. Subject 1 was exposed chronically over a 1-year period when the interior
32 woodwork of her home was mistakenly treated with tetrachloroethylene. The three other cases
33 were occupationally exposed. Subject 2 was exposed during two separate periods: first, for 3
34 years in a dry cleaning establishment and, second, for 7 years cleaning parts. Subject 3 was
35 exposed for 16 years as a dry cleaning worker. Subject 4 was also exposed as a dry cleaning
36 worker, but her duration of employment was not reported. Subjects 2, 3, and 4 were working
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-75 DRAFT-DO NOT CITE OR QUOTE
-------
1 with tetrachloroethylene when first tested. Air monitoring data were not available; however,
2 occupational health physicians diagnosed each case with tetrachloroethylene encephalopathy on
3 the basis of symptoms, neurophysiological assessment, and their own examinations.
4 A large battery of standard neurobehavioral tests was given to each subject. For most
5 tests, impairment was inferred clinically when a subject's score was greater than one standard
6 error of measurement below expectation, which is less restrictive than the criterion (more than
7 two standard deviations below mean) commonly used in neurobehavioral testing to separate
8 normal from abnormal scores (Lezak, 1995). Test results for the four subjects most consistently
9 indicated complaints of fatigue and confusion, accompanied by cognitive deficits on tests
10 assessing memory and motor, visuospatial, and executive function. Repeated testing of subjects
11 3 and 4 indicated post-exposure improvement on neurobehavioral tests of all the affected
12 functional domains, although performance on some of the more difficult tests in each domain
13 remained impaired. These results suggest an association between CNS effects and
14 tetrachloroethylene exposure, but a conclusion of a causal relationship is precluded by the lack of
15 data on the duration and severity of the tetrachloroethylene exposure.
16 The investigators also assessed the performance of 66 dry cleaning workers on
17 neurobehavioral tests designed to detect the same impairments noted in the clinical cases. The
18 testing was conducted in 1986. The owners of 125 shops in Detroit, Michigan were contacted,
19 and 23 agreed to allow their workers to participate in the study. Within each shop, operators
20 were matched on education and age (+5 years) with a lower-exposure subject.
21 The subjects (35 men and 30 women) were grouped into three categories of chronic
22 tetrachloroethylene exposure (low, moderate, and high), based on type of shop (wet-transfer or
23 dry-to-dry), job title (counter clerk, presser, or operator), and years of employment. All the
24 operators were placed in the high-exposure category. There was no unexposed control group.
25 Dry cleaning workers placed in the chronic exposure categories of low, moderate, and high had
26 been employed at their main job for 2.1, 3.9, and 14.6 years, respectively. Their mean age was
27 40.9, 40.6, and 43 years. The three groups were also characterized by estimates of current
28 exposure (low, medium, and high), which corresponded to mean tetrachloroethylene air
29 concentrations (8-hr TWA) of 11, 23, and 41 ppm, respectively, for counter clerks, pressers, and
30 operators in the more common wet-transfer shops (17 of 23 shops). Estimated air concentrations
31 for counter clerks, pressers, and operators in the dry-to-dry shops were 0.5, 10, and 11 ppm. The
32 estimates were based on a relationship between breath and air concentrations derived from a
33 larger independent study (Solet et al., 1990). The study authors noted that the estimates were
34 comparable to those found in other surveys of dry cleaning facilities in the United States.
35 All subjects were tested in groups of two in the afternoon after work on the first or
36 second day of their work week. The tests were conducted in a minivan. Each subject provided a
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-76 DRAFT-DO NOT CITE OR QUOTE
-------
1 breath sample and completed a medical, symptom, work history, and hobby questionnaire. The
2 subjects were administered six neurobehavioral tests, a test of verbal skills, and questionnaires
3 on emotional states (moods) and CNS symptoms. The neurobehavioral test battery consisted of
4 one test of motor/cognitive function (symbol digit) and five tests of cognitive function (digit span,
5 trailmaking A and B, visual reproduction, pattern memory, and pattern recognition), including
6 three tests of an individual's ability to process and remember visuospatial stimuli (the latter three
7 tests).
8 Multivariate analysis was used to evaluate the relationship between a chronic index of
9 lifetime exposure and performance on neurobehavioral tests, after adjusting for the confounding
10 variables of current exposure, a 3-year index of exposure, age, education, verbal skill, alcohol
11 consumption, hours of sleep, fatigue, mood, symptoms, medication, and secondary exposures to
12 neurotoxicants. After adjustment for factors affecting performance, the scores of the dry
13 cleaning workers with high chronic exposure were statistically significantly lower (p < 0.01) than
14 those of the workers with low chronic exposure in three tests of visual function: visual
15 reproduction, pattern memory, and pattern recognition. Adjusted scores were reduced from 6 to
16 15%; the two most sensitive tests were those that measured short-term memory of visual designs.
17 These impairments of visually mediated function were consistent with the impairment of
18 visuospatial functions observed in the four patients previously studied by Echeverria et al. who
19 were diagnosed with tetrachloroethylene encephalopathy. Other effects seen in the patients
20 (mood changes and decreased cognitive function in nonvisual tests) were not found in the dry
21 cleaning workers with high lifetime exposures. Among complaints by the dry cleaning workers,
22 only the number of complaints of dizziness from standing up rapidly and "solvent-induced
23 dizziness" over the previous 3 months was significantly elevated (p<0.04) in the high-exposure
24 group.
25 The study authors concluded that effects on visuospatial function were consistently found
26 in subjects employed as operators for an average of 14.6 years and exposed to an estimated
27 tetrachloroethylene 8-hr TWA air concentration of 41 ppm, suggesting a vulnerability of visually
28 mediated functions with tetrachloroethylene exposure. This conclusion was based on the
29 impaired performance of the high-exposure group when compared with a group of dry cleaning
30 workers with low lifetime exposure, including 16/22 workers who were probably clerks in wet-
31 transfer shops where the mean current exposure level was 12 ppm. This exposure level is
32 substantially above background ambient levels, and whether the performance of the low-
33 exposure group was impaired when compared with that of a group without occupational
34 exposure (i.e., an unexposed control group) is not known. The lack of an unexposed control
35 group limits the ability of the study to fully characterize the magnitude of the effects on
36 visuospatial ability and to detect exposure-related symptoms or effects on tests of nonvisual
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-77 DRAFT-DO NOT CITE OR QUOTE
-------
1 cognitive ability. It also limits the extrapolation of the results to other populations exposed to
2 tetrachloroethylene.
O
4 4.6.1.2.9. Altmann, L., H.F. Neuhann, U. Kramer, J. Witten andE. Jermann. 1995.
5 Neurobehavioral and neurophysiological outcomes of chronic low-level tetrachloroethylene
6 exposure measured in neighborhoods of dry cleaning shops. Environ. Res. 69:83-89.
1 Altmann et al. (1995) used neurophysiological and neurobehavioral techniques to assess the
8 effects of long-term exposures to tetrachloroethylene. A total of 19 tetrachloroethylene-exposed
9 subjects (residents of Mulheim, Germany) were chosen from a population of 92 subjects living in
10 neighborhoods close to dry cleaning facilities. Three criteria were used to select subjects: a
11 tetrachloroethylene blood level above 0.002 mg/L, a period of living above or next to a dry
12 cleaning facility for at least 1 year, and no occupational exposure to organic solvents. The mean
13 age of the exposed subjects was 39.2 years (range: 27-58 years) and the mean duration of living
14 near a dry cleaning facility was 10.6 years (range: 1-30 years). The daily activity pattern of the
15 exposed subjects was not reported. A total of 30 controls were selected from volunteers; their
16 mean age was 37.2 years (range: 24-63 years). One or two controls, matched for age (+1 year,
17 but 3 years in one case and 6 years in another case) and gender, were chosen for each exposed
18 subject. The control subjects were recruited mainly from the staff of a public health office or an
19 institute for environmental hygiene, and none reported a history of solvent exposure. No
20 information is provided in the paper on the motivation for exposed and control subject to
21 participate in the study. Voluntary consent was obtained from all subjects prior to the initiation
22 of testing.
23 All subjects were given medical examinations. Five exposed (26%) and seven control
24 subjects (23%) were excluded for various medical reasons, including impaired vision, diseases
25 with potential neuropathy, hypertension, and joint impairment. The reasons for exclusion were
26 similar in both groups. All subjects met standards for visual acuity and vibration perception.
27 The final exposed group was composed of 5 men and 9 women and the control group was
28 composed of 9 men and 14 women. The two groups did not differ with regard to consumption of
29 alcoholic beverages, regular medication, smoking, or body mass index, but they did differ in
30 degree of education, which was used as an indicator of social status.
31 The effect of tetrachloroethylene exposure on the neurophysiological and
32 neurobehavioral measurements was evaluated using both univariate and covariate linear
33 regression. The multivariate regression analysis accounted for age, gender, and education as
34 covariates. Degree of education was defined as "low," "medium," or "high" level.
35 VEPs in response to black-and-white checkerboard patterns were recorded for all
36 individuals. Vibration perception using a tuning fork—a crude measure of peripheral
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-78 DRAFT-DO NOT CITE OR QUOTE
-------
1 neuropathy—was assessed at the ankle. Five tests included in the Neurobehavioral Evaluation
2 System developed in the United States and adapted for testing on a German population were
3 used: (1) finger-tapping speed with the index finger of both the dominant and the nondominant
4 hand; (2) hand-eye coordination using a joy stick to follow a sine wave on a computer screen; (3)
5 a continuous performance test for assessment of vigilance, which requires a response to a
6 specific stimulus appearing on the computer screen and failure to respond to other stimuli; (4)
7 simple reaction time, which requires the fastest possible response to a simple visual stimulus
8 (measured twice); and (5) visual memory on the Benton visual retention test, which requires a
9 match of a previously displayed stimulus out of several choices after a short delay interval. All
10 of these tests are commonly used to assess occupationally exposed adults, and the software for
11 testing and analysis is available for purchase. All testing was completed in a single 3-hr session;
12 testing times were selected randomly for both exposed or control subjects.
13 Blood samples were taken once in the examination room immediately before testing (all
14 subjects) and, if possible, once when the exposed subjects were at home. The mean blood level
15 for exposed subjects, based on samples collected in the examination room, was 0.0178 mg/L
16 (standard deviation, 0.469 mg/L). For seven of the nine exposed subjects, blood concentrations
17 in samples collected at home were higher than those in samples collected in the examination
18 room. None of the blood concentrations in the control group exceeded the detection limit of
19 0.0005 mg/L. For the exposed subjects (data from 13 apartments), indoor air sampling indicated
20 that the mean (7-day TWA) air concentration was 0.7 ppm (standard deviation, 1 ppm) and the
21 median was 0.2 ppm. For the control group, the mean and median values were 0.0005 ppm
22 (standard deviation, 0.0005 ppm) and 0.0003 ppm. There was a good correlation between home
23 indoor air concentrations and blood levels of tetrachloroethylene in the exposed subjects
24 (r = 0.81). The correlation was much lower when the examination room blood samples were
25 used (r = 0.24).
26 After adjusting for covariates and possible confounders of age, gender, and education,
27 there were statistically significant group differences between the adjusted mean scores of
28 exposed and control subjects on three neurobehavioral tests (simple reaction time,/? < 0.05 for
29 the first test andp < 0.01 for the second test; continuous performance, p < 0.05; and visual
30 memory,/' < 0.05). In all cases, the exposed subjects had slower response times or more errors
31 than did the unexposed controls. No statistically significant differences were observed between
32 the performance of the exposed and control groups on the finger-tapping or hand-eye
33 coordination tests, which are measures of fine motor function; on VEP, which may be less
34 sensitive than direct measurement of visual function; or on vibration perception at the ankle
35 using a tuning fork.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-79 DRAFT-DO NOT CITE OR QUOTE
-------
1 The relationship between indoor tetrachloroethylene concentration and individual
2 performance was not reported, so it was not possible to evaluate concentration-response
3 relationships. The small numbers of study subjects in this study compared to the occupational
4 studies is a limitation; however, this study appeared to have sufficient power to detect
5 associations with tetrachloroethylene exposure. Additionally, exposed and control groups did
6 not differ with regard to consumption of alcoholic beverages, regular medication, smoking, or
7 body mass index, but they did differ in degree of education. Statistical analysis of the data took
8 into account effects from important covariates such as age, gender, and education. Univariate
9 analyses showed that age, gender, and education were not predictors of deficits in tests of
10 continuous performance, visual memory, and the second reaction time, although education and
11 gender were predictors of deficits in the first reaction time.
12 The use of three categories for education in the multivariate regression analyses may not
13 fully account for all effects from these covariates (U.S. EPA, 2004), although it is not possible to
14 evaluate whether residual confounding from these covariates may explain observations on the
15 neurobehavioral tests. Furthermore, because the responses in the exposed group for the tests
16 highlighted above (simple reaction time, continuous performance, visual retention) were
17 statistically significantly different from those of the control group, whether or not the covariates
18 were considered, an approximate estimate of the impact of the tetrachloroethylene exposures can
19 be derived by comparing the reported response levels for the two groups. The degree of change
20 from control was approximately 15-20% for this subset of tests.
21
22 4.6.1.2.10. Spinatonda, G., R. Colombo, E.M. Capodaglio, M. Imbriani, C. Pasetti, G. Minuco
23 and P. Pinelli. 1997. [Study on speech production processes: application for a group of
24 subjects chronically exposed to organic solvents (part II).J Med. Lav. 19:85—88. Spinatonda et
25 al. (1997) assessed the effect of tetrachloroethylene exposure on vocal reaction times among 35
26 dry cleaners and 39 unexposed controls. Controls were matched to exposed individuals for age
27 (mean age of 35 years for both groups) and education. The published paper did not identify the
28 population from which exposed and controls were drawn, the inclusion criteria for exposed
29 subjects and controls—and hence, whether potential study subjects may have been excluded—
30 and duration of exposure in a tetrachloroethylene-exposed j ob.
31 Exposure was assessed by a "grab sample" and not as a weighted average (as often
32 reported in other occupational studies reviewed in this section). Exposure monitoring indicated a
33 median concentration of tetrachloroethylene of 8 ppm (range: 2-136 ppm). An index of
34 cumulative exposure to tetrachloroethylene was also developed for each exposed subject by
35 multiplying the tetrachloroethylene concentration by the number of years worked.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-80 DRAFT-DO NOT CITE OR QUOTE
-------
1 Latency to and duration of vocal response to the stimulus (reading) were measured in
2 each subject after the presentation of a sequence of words on a computer screen. For each
3 condition, subjects were asked to say the word immediately or following delays of 0.1 or 0.5
4 seconds. The test was performed using a random sequence of concrete or meaningless disyllabic
5 words. These tests were carried out at the place of employment for dry cleaners and in a clinical
6 setting for controls, indicating that the investigators were not blinded as to a subject's exposure
7 status. Testing conditions may have differed between exposed group and controls.
8 Compared with the control group, the exposed group had statistically significant longer
9 mean reaction times and/or vocalization durations under all response conditions (immediate or
10 delayed response) with either real or meaningless words. Furthermore, statistically significant
11 positive correlations were observed between cumulative tetrachloroethylene exposure and
12 immediate reading and delayed reading tasks r = 0.69 and r = 0.73, respectively). No
13 information on alcohol consumption or other potential differences between exposed subjects and
14 controls was reported, precluding an analysis of how these factors may have affected the
15 observed association between tetrachloroethylene and reaction time.
16
17 4.6.1.2.11. Schreiber, J.S., H.K. Hudnell, A.M. Getter, D.E. House, KM. Aldous, M.E. Force,
18 K. W. Langguth, E.J. Prohonic and J. C. Parker. 2002. Apartment residents' and day care
19 workers' exposure to tetrachloroethylene (perc) and deficits in visual contrast sensitivity.
20 Environ, Health Perspect. 110:655-664. Schreiber et al. (2002) reported the findings from
21 investigations using visual tests to assess neurologic function in two populations: apartment
22 residents and day care employees who had potential environmental tetrachloroethylene exposure
23 due to close proximity to dry cleaning facilities.9 Residential exposure to tetrachloroethylene
24 can result in nearly continuous exposure (NYS OAG, 2004a) and is distinct from the pattern of
25 tetrachloroethylene exposure experienced by the occupational populations described in the
26 preceding paragraphs. Objectives of the residential and day care investigations were to
27 characterize tetrachloroethylene exposure and to screen for subclinical neurological effects using
28 a battery of visual function tests. All participants—or their guardians in the case of the
29 residential study—signed voluntary consent forms prior to study commencement.
30 For the residential study, the exposed group consisted of 17 tetrachloroethylene-exposed
31 subjects (11 adults between the ages of 20 and 50, 2 adults over the age of 60, and 4 children)
9 The apartment residents lived in two separate buildings in New York City that each contained a dry
cleaning business. The residential study served as a pilot for a larger study that is investigating visual effects among
tetrachloroethylene-exposed residents. The day care study was part of an investigation of staff and children carried
out by the NYS DOH and the Centers for Disease Control and Prevention. The day care facility, located in Albany,
New York, was in a building that also housed a business that did dry cleaning. Visual testing for both studies was
carried out by the same investigator using the same testing apparatus.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-81 DRAFT-DO NOT CITE OR QUOTE
-------
1 from six families residing for an average of 5.8 years (6 years median) in two apartment
2 buildings in New York City. Preliminary monitoring of these buildings indicated
3 tetrachloroethylene concentrations were elevated compared to eight other buildings also
4 monitored by the NYS DOH. These eight buildings were identified by NYS DOH from
5 discussions with the New York City Department of Health and from a review of files on dry
6 cleaning facilities (NYS OAG, 2004b).
7 Study subjects were identified through several methods: (1) both families in the first
8 building (Buildling A) had been referred to the NYS DOH for information about participating in
9 the study by Consumer Union/Hunter College researchers, (2) one family in the second building
10 (Buildling B) had previously contacted NYS DOH about exposure concerns and desired to
11 participate in a study, and (3) three other families in Building B were recruited by a participating
12 family (NYS OAG, 2004b). Exposed residents were an affluent, English-speaking, Caucasan
13 population living near New York City's Central Park (telephone communication from K.
14 Hudnell, EPA, to D. Rice, EPA, February 2003). Exposed participants were generally unaware
15 of the tetrachloroethylene exposure, although some study participants did observe
16 tetrachloroethylene-like odors prior to the study period.
17 Control subjects were recruited from among NYS DOH Albany, New York employees
18 and their families. They were considered representative of the general population not living near
19 dry cleaning facilities. All controls were Caucasan, except for one Asian individual, and were
20 age- and sex-matched to exposed apartment residents. In some cases, more than one control
21 participant was matched to an exposed subject, and an average of the multiple control visual
22 function test scores was used for comparison to that of an exposed subject. Mean age was 34.5
23 years for exposed apartment residents and 33.2 years for control subjects.
24 Nine adult staff (all females) of a day care facility agreed to participate in the day care
25 study. Controls were age- and gender-matched acquaintances of the exposed participants, local
26 retail shop employees, NYS DOH employees, or staff from other local day care centers with no
27 known tetrachloroethylene exposure. All subjects in the exposed and control groups were
28 Caucasan (telephone communication from K. Hudnell, EPA, to D. Rice, EPA, February 2003).
29 Mean age was 27.7 years for control participants and 27.2 years for day care staff; mean duration
30 of employment for exposed subjects was 4 years at the center.
31 Information on sociodemographics; lifestyle factors such as exposure to direct or passive
32 smoke, alcohol consumption, and exercise; medical history; and neurotoxicant exposure in
33 addition to the visual tests was obtained by questionnaire from both study populations and their
34 controls. Exposed participants had no known exposure to other neurotoxicants, ongoing illness,
35 current use of neuroactive drugs, or a medical history indicative of neurologic dysfunction, and
36 both exposed participants and controls reported low or moderate alcohol consumption that did
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-82 DRAFT-DO NOT CITE OR QUOTE
-------
1 not differ between either exposed group and their controls. Moreover, the profile of moods test
2 scores of all residential exposed subjects were within normal limits. The investigators also
3 administered visual tests of acuity, contrast sensitivity, and color discrimination to exposed
4 subjects and their referents. The investigators were not blinded as to a subject's status as either
5 exposed or nonexposed.
6 The assessment of tetrachloroethylene exposure of residents consisted of
7 tetrachloroethylene concentrations in indoor air and personal air samples, exhaled breath, and
8 blood, which were collected at the time of visual testing. Testing was performed during a period
9 of active dry cleaning for four of the families and one month after closure of the facility for the
10 remaining two families in the residential study. Additionally, adult residents provided urine
11 samples, which were analyzed for tetrachloroethylene as well as for three products of its
12 metabolism: TCA, trichloroethanol, and the urinary acetyl metabolite. Breast milk samples
13 were provided from two exposed breastfeeding mothers.
14 Ambient concentrations of tetrachloroethylene were available for all study participants
15 for an earlier time frame (from 1 to 3 months before the date of visual testing), when active dry
16 cleaning was occurring in both apartment buildings. These measurements were used by NYS
17 DOH to identify study sites. Concentrations of airborne tetrachloroethylene levels in apartment
18 rooms were higher in these samples than in the monitoring data obtained at the time of the visual
19 testing. Median concentrations in these samples, which were taken during the day during active
20 periods of dry cleaning, were 0.21 ppm (mean = 0.36 ppm; range: 0.1-0.9 ppm). Airborne
21 tetrachloroethylene concentrations had decreased in samples collected at the time of visual
22 testing; median tetrachloroethylene concentration was 0.09 ppm (mean = 0.18 ppm; range:
23 0.01-0.78 ppm). Tetrachloroethylene levels in blood correlated well with levels in room air,
24 personal air, and breath.
25 Atmospheric monitoring of the day care facility before closure of the dry cleaning
26 business showed airborne concentrations of tetrachloroethylene ranging from 0.27 to 0.35 ppm,
27 with median and mean concentrations of 0.32 ppm. Samples obtained at the time of visual
28 testing, five weeks after removal of the dry cleaning machines, approached background
29 concentrations (range: 0.0012-0.0081 ppm).
30 Visual function testing consisted of near visual acuity, near VCS, and color vision. All
31 study participants who wore corrective lenses for reading wore their lenses when taking the
32 vision tests. The visual acuity test measured the ability to discriminate high- frequency (i.e.,
33 small) images at high contrast; e.g., reading successively smaller black-on-white letters as part of
34 an examination for corrective lenses. This measure typically is dependent on the optics of the
35 eye (and corrective lenses when needed) and is insensitive to subclinical deficits in neurologic
36 function. In neither assessment did the groups differ in visual acuity.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-83 DRAFT-DO NOT CITE OR QUOTE
-------
1 The contrast sensitivity test is sensitive to subclinical deficits in neurologic function in
2 the visual pathways. The test measured the least amount of luminance difference between dark
3 and light bars that was needed to detect the bar pattern. Luminance varied between the bars in
4 sine-wave fashion, and each test pattern represented one size of bars or spatial frequency. The
5 bar patterns were presented at five different spatial frequencies, thereby breaking spatial visual
6 function into its essential components. The least amount of luminance contrast needed to detect
7 each bar size was measured. The contrast sensitivity data are presented in Figure 4-1. A
8 strength of this study is that the test of contrast sensitivity employed a forced-choice procedure,
9 providing better reliability and consistency than other approaches.
10 Multivariate analysis of variance was used to analyze the VCS data. Group mean scores
11 for VCS across spatial frequencies were statistically significantly lower in exposed residents than
12 in controls and in day care employees as compared with controls, indicating poorer visual
13 function in the exposed groups. An exposure-response analysis did not show an association
14 between poorer performance and increasing tetrachloroethylene concentration. Among
15 apartment residents, mean scores of VCS in all four children and in both older adults (60 years of
16 age) were lower than the 12th percentile score of all control subjects. (The 12th percentile
17 represents the two control subjects with the poorest performance out of the 17 total data points.)
18 In contrast, 5 of the 11 adults aged less than 60 years scored below the 12th percentile. It is
19 unknown whether the difference between groups would have been statistically significant on the
20 basis of the adults under 60 years alone. However, there was a statistically significant lower
21 group mean VCS score across all spatial frequencies when day care employees were compared
22 with the control group (data not shown).
23 In the residential study, exposed subjects were retested twice after the initial assessment,
24 6 to 10 months and 17 to 21 months after closure of the dry cleaning facility. Performance
25 appeared to worsen over successive evaluations, although statistical comparisons were not
26 performed (NYS DOH, 2000). Control subjects from the initial testing were not retested,
27 preventing a comparison with observations from exposed subjects.
28 Color vision was also assessed in both the residential and the day care groups. Subjects
29 were asked to put a series of small round "caps" in order by color. The types of errors made
30 could distinguish specific types of color vision deficiency: e.g., red-green color blindness, which
31 is common in males, or blue-yellow color blindness, which is associated with solvent exposure
32 (Mergler and Blain, 1987; Mergler et al., 1987, 1988a, b, 1991; Campagna et al., 1995, 1996).
33 Group differences in the CCI were assessed using two-tailed Student's t-tests for matched-pair
34 analyses. CCI scores of exposed groups did not show statistically significant impairment as
35 compared with referents, although the performance of the exposed groups, particularly the
36 residential group, appeared worse than that of control.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-84 DRAFT-DO NOT CITE OR QUOTE
-------
140
>, 120
a 100
W
ra
o
15
3
Ifl
140
>, 120
S 100
P 80
Ifl
ra
o
O
60 -
O
T5
Control, <18 years old
Exposed, <18 years old
1.5 3 6 12 18
Spatial Frequency (Cycles/Degree)
Spatial Frequency (Cycles / Degree)
Control, >60 years old
Exposed, >60 years old
Spatial Frequency (Cycles/Degree)
4
5
6
7
8
9
10
11
Figure 4-1. Visual contrast sensitivity functions for control and exposed
children (top), adults that were identified as having impaired function (i.e.,
5 of the total 11) and their matched controls (middle), and the control and
exposed individuals over 60 years of age. The x-axis represents the frequency
of the stimulus bars, with finer bars toward the right. The y-axis represents the
inverse of the contrast at which the subject could no longer distinguish the
orientation of the bars (threshold). For any frequency, a higher contrast
sensitivity threshold represents better visual function. It is apparent that the group
of children is relatively more impaired than the impaired group adults.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-85 DRAFT-DO NOT CITE OR QUOTE
-------
1 Observations in the study paper have been questioned, particularly on selection bias in
2 the residential investigation as an explanation of observed VCS deficits (HSIA, 2004). Although
3 motivation for study participation is not known, the New York State Office of Attorney General
4 (NYS OAG, 2004b) noted that test results were provided to individual study subjects, which
5 probably encouraged participation; however, the principal investigator does not believe selection
6 bias was a factor for study participation. Some information in the study paper may also be used
7 to judge the potential for selection bias. The study authors noted that the profile of moods test
8 scores of all exposed residential subjects were within normal limits, with no cases of clinical
9 depression or other neuropsychiatric conditions. Hence, it does not appear that exposed residents
10 had major psychological impairments. Additionally, bias may be introduced through the use of
11 controls living in Albany for comparison with exposed residents living in New York City.
12 Information on covariates is lacking, and the impact of these covariates on VCS function cannot
13 be adequately assessed.
14 Some general information is available to evaluate potential confounding due to education,
15 occupation, and residential location. Factors such as education, socioeconomic status, and
16 smoking do not affect the VCS test (NYS OAG, 2004b; Hudnell et al., 2001; Mergler et al.,
17 1991; Frenette et al., 1991; U.S. EPA, 2004). Occupation is highly correlated with
18 socioeconomic status (Deonandan et al., 2000) and is not likely to confound the VCS test.
19 Moreover, urban-rural differences between exposed and control subjects are not thought to
20 strongly bias findings. For example, Kaufman et al. (1988) did not show that urban or rural
21 residence was related to performance on specific subtests of the Wechsler Adult Intelligence
22 Scale, although associations were seen with other variables such as sex, age, and education,
23 variables that are similar or matched for exposed and referent subjects in the current study.
24 Public comments to EPA (NYS DOH, 2004) discuss that two of the four children in the
25 residential study had medically verified diagnoses of learning disabilities or developmental
26 delays; however, no information was provided in these public comments about these conditions
27 in referent children. Without comparable information on control children, it is difficult to draw
28 any conclusions about whether these conditions may or may not have also contributed to the
29 VCS deficits observed in residential subjects.
30 Finally, as with all other studies discussed in this section, unmeasured differences or
31 residual confounding between exposed and referent groups may possibly explain observations;
32 however, in the absence of information, it is not possible to evaluate the unmeasured variables.
33
34 4.6.1.2.12. Sharanjeet-Kaur, MursyidA, Kamamddin A, Ariffln A. 2004. Effect of petroleum
35 derivatives and solvents on colour perception. Clin Exp Optom 87:339-343. Fourteen healthy
36 subjects of ages 24-53 years working in 3 dry cleaning facilities using tetrachloroethylene are
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-86 DRAFT-DO NOT CITE OR QUOTE
-------
1 included in a study assessing color vision. This study was part of a larger study assessing color
2 vision in two other occupationally-exposed populations, 39 workers in a factory producing
3 polyethylene resins plastic storage containers and 40 workers manufacturing polystyrene plastic
4 bags. The published study is poorly reported, lacks many details, and adopts post-hoc statistical
5 testing. The paper reports neither how facilities were identified nor recruitment methods for
6 study subjects. Furthermore, the paper does not present any information on tetrachloroethylene
7 concentrations or on tetrachloroethylene biomarkers, making it difficult to judge the degree of
8 exposure to tetrachloroethylene. Control selection criteria are not identified in the published
9 paper other than 27 healthy subjects (mean age 27 + 4 years) composed Control Group 1 and 2
10 healthy subjects (mean age 33 + 4 years) who were support staff of Universiti Kebangsaan
11 Malaysia. Dry cleaning workers differed from controls on several variables: work duration,
12 hours worker per day, cigarette smoking, mean age (compared to Control Group 1), and race.
13 Also, no information is presented on possible difference between dry cleaners and controls on
14 socio-economic status (SES). Voluntary consent was provided by all subjects.
15 Visual testing was carried out at the factory or dry cleaner, for exposed subjects, and at
16 the Optometry Clinic in the Universiti Kebangsaan Malaysia for control subjects. Given these
17 testing conditions, a subject's exposure status was known, i.e., no blinding. Visual acuity was
18 measured at distance using the Snellen chart and at near using a reading chart. Subjects were
19 excluded with poor visual acuity or with systemic, ocular, or neurological diseases; the number
20 of excluded subjects is not identified in the published paper. Color vision was assessed
21 binocularly using Ishihara plate, D-15 test, and Farnsworth Munsell 100 Hue test under a light
22 box at illumination of 1,000 lux. Subjects wore the best corrective lens and testing was carried
23 out at a distance of 35 to 40 cm.
24 The number of subjects with abnormal scores, using criteria of Vingrys and King-Smith
25 (1988), is presented but not group-mean color confusion index scores. None of the controls or
26 dry cleaners had color vision errors with the Ishihara plates. In contrast, 6 dry cleaners (43%)
27 and 13 subjects (93%) compared to no controls were identified with errors on the D-15 test and
28 FM 100 Hue test, respectively. Statistical testing of differences is lacking. Total error scores for
29 the FM 100 Hue test differed between dry cleaners and control group 2 (p < 0.05) but not with
30 control group 1. It is difficult to interpret these findings due to the lack of exposure information
31 on potential tetrachloroethylene exposure other than job title, and differences between dry
32 cleaners and controls regarding test conditions, SES, and smoking.
33
34 4.6.1.2.13. New York Department of Health (NYS DOH). 2005a. Improving human risk
35 assessment for tetrachloroethylene by using biomarkers and neurobehavioral testing. Final
3 6 Technical Report to US EPA Star Grant #R82 7446. Accessed 5 December 2006,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-87 DRAFT-DO NOT CITE OR QUOTE
-------
1 http://cfpub.epa.gov/ncer_abstracts/index.cfm/fuseaction/display.abstractDetail/abstract/97
2 7/report/O. NYS DOH (2005a) examines the effect of tetrachloroethylene exposure on visual
3 function in two populations, residents living in a building co-located with a dry cleaning
4 establishment and among employees and former students of a day care establishment exposed 4
5 years previously during a period when the day care was co-located near a dry cleaner. The first
6 study, the New York City Perc Project, did not include subjects studied by Schreiber et al. (2002)
7 and employed different methods for testing visual contrast sensitivity and color vision. NYS
8 DOH (2005a) enlisted 65 households in 24 residential buildings with dry cleaners using
9 tetrachloroethylene on-site and 61 households in 36 buildings without dry cleaners located in the
10 study area in Manhattan, New York City. Health outcome and tetrachloroethylene
11 concentrations as measured from indoor air monitoring and in exposed subject's breath and
12 blood were obtained over the period from 2001 to 2003. The full report of the residential study
13 has not received public peer review nor has it been published as a literature paper although
14 McDermott et al. (2005) presents exposure monitoring findings from the dry cleaner households.
15 The second project, the Pumpkin Patch Day Care Center (PPDCC) Follow-up Evaluation,
16 is a 5-year follow-up of visual function among some employees and neurobehavioral function
17 among children in a day care center that had been previously co-located in a building with a dry
18 cleaning establishment. The PPDCC Follow-up Study also included first-time visual function
19 tests of former students. NYS DOH together with the U.S. Centers for Disease Control and
20 Prevention carried out the initial evaluation of PPDCC staff and students in 1998 (NYS DOH,
21 2005b). Funding to NYS DOH for the residential study and the PPDCC Follow-up study was
22 provided though U.S. EPA STAR Grant #827446010 (NYS DOH, 2005a, 2005c).
23
24 4.6.1.2.13.1. New York City Perc Project. The objectives of the New York City Perc Project are
25 as follows: 1) to document tetrachloroethylene exposures in buildings where dry cleaners were
26 present; 2) to evaluate whether living in a building with a dry cleaner was associated with CNS
27 effects; 3) to evaluate the relationship(s) between measures of tetrachloroethylene exposure and
28 CNS effects; and, 4) to assess whether children were disproportionately exposed to and/or
29 affected by tetrachloroethylene compared to adults. Study design and protocols were approved
30 by Institutional Review Boards at the NYS DOH and other collaborating institutes (Mr. Sinai
31 Medical Center and U.S. CDC).
32 Subjects were identified in buildings from eleven zip code areas surrounding Central Park,
33 New York City, contiguous with one another, but different in demographic and socioeconomic
34 characteristics. Eligible households for participating in this study include at least one adult
35 (20-55 years old) and one child (5-14 years old), so as to assess whether residential
36 tetrachloroethylene exposure would disproportionately affect children. Initial monitoring
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-88 DRAFT-DO NOT CITE OR QUOTE
-------
1 indicated few residences in dry cleaner buildings with elevated indoor air concentration of
2 tetrachloroethylene above the current NYS DOH residential air guideline of 0.015 ppm
3 (0.1 mg/m3). The study area was broadened to include buildings subject of a resident complaint
4 and to include buildings in additional zip codes, primarily characterized by lower SES or higher
5 percentage of minority residents. This decision was made after the finding of elevated
6 tetrachloroethylene levels in households in dry cleaner buildings in one zip code area, a low
7 income, minority area, compared to other zip codes area. Mail and telephone contact were the
8 primary methods of recruiting subjects, with door-to-door recruitment by a not-for-profit child
9 advocacy organization (Northern Manhattan Perinatal Partnership, Inc.) providing assistance
10 with recruiting bilingual subjects primarily living in 3 of the 11 zip code areas. Of the 1,261 dry
11 cleaner and 1,252 reference households contacted, 132 dry cleaner households and 175 reference
12 households included age-eligible adult-child pairs and a total of 65 dry cleaner (67 adults, 68
13 children) and 61 reference households (61 adults, 71 children) participated in the study. All
14 participants or their guardians signed voluntary consent forms prior to study commencement.
15 Tetrachloroethylene indoor air concentrations in dry cleaner buildings had decreased
16 since 1997, the period of the pilot study (Schreiber et al., 2002), and ranged up to around 0.77
17 ppm (5 mg/m3) with a geometric mean of 0.005 ppm (0.035 mg/m3). Monitoring was carried out
18 using passive monitoring badges. In comparison, tetrachloroethylene concentrations in building
19 without dry cleaners ranged up to 0.014 ppm (0.09 mg/m3) with a geometric mean of 0.0004
20 ppm (0.003 mg/m3). Both breath and blood tetrachloroethylene levels were significantly
21 (p< 0.05) correlated with indoor air concentration for adult and for child subjects of dry cleaner
22 buildings. Levels of detection (LODS) were 5 |ig/m3 air and 0.048 mg/ml blood. Air, breath,
23 and blood tetrachloroethylene concentrations were inversely correlated with income and were
24 higher among minority compared to non-minority subjects.
25 NYS DOH staff visited participants in their residences to collect 24-hr indoor air samples,
26 breath samples, and to give adult participants a questionnaire seeking information on residential,
27 occupational, and medical history for themselves and their children. Opthalmologic
28 examinations were scheduled at the same time for the Mt Sinai School of Medicine Department
29 of Opthalmology research clinic. Participants received financial compensation after completing
30 the home visit ($50.00) and ophthalmology clinic visit ($50.00).
31 No differences between exposure groups were observed for participants recruited using
32 the mail and telephone method, although this was not so for participants recruited using door-to-
33 door methods. For these individuals, language and adult age differed significantly between
34 exposed and non-exposed groups with more English speaking households participating in the
35 non-exposed group and non-exposed adults were slightly older than exposed adults. Overall,
36 differences between adult residents of reference buildings or buildings with dry cleaners in SES
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-89 DRAFT-DO NOT CITE OR QUOTE
-------
1 characteristics, residence duration, education level, age smoking or alcohol use are not apparent.
2 Differences between child residents in gender or residence duration are not apparent, but the
3 highest exposure group is about a year younger and has about one less year of education than
4 children in the other exposure groups.
5 Ophthalmologic examinations and visual function tests were given to study participants
6 at the Mt. Sinai Medical School of Medicine. The final report does not describe whether
7 examiners were or were not blinded as to a subject's exposure status (NYS DOH, 2005a). The
8 examination included determination of past ocular and medical history; measurement of visual
9 acuity, pupil size, extrocular motility, and intraocular pressure; and anterior and posterior
10 segment exams. Subjects with abnormalities or taking medications that could influence VCS
11 and/or color vision were excluded from further testing. Furthermore, visual functional tests for
12 some children were excluded from the statistical analysis because of their young age or because
13 they were identified by their parents as learning disabled or having attention deficit hyperactivity
14 disorder. VCS was determined using the Functional Acuity Contrast Test (FACT) distance chart
15 placed 10 feet from the participant under light conditions of 68-240 cd/m2. These testing
16 conditions differ from those employed by Schreiber et al. (2002) in their residential study where
17 visual test was carried out assessing near contrast sensitivity.
18 Adults and children demonstrated a ceiling effect with VCS performance, i.e., a
19 maximum score at 1.5, 3, 6, 12, and 18 cycles per degree (cpd) is achieved by some study
20 participants. VCS scores among adults were not correlated with any SES factor or personal
21 characteristics (smoking, alcohol use, education level, duration of residence). Among all
22 children, poorer VCS at 1.5, 3, and 6 cpd were significantly correlated with speaking primarily
23 Spanish at home.
24 NYS DOH examined possible association between VCS and tetrachloroethylene
25 exposure by, (1) comparing the percent of exposed subjects with maximum VCS score (no
26 errors) to referents, (2) comparing mean differences in VCS scores between adult and child
27 subjects living in the same residence across exposure categories, and (3) using logistic regression
28 to assess the effect of tetrachloroethylene in indoor air, blood or breathe on the achievment of
29 maximum VCS score at 6 and 12 cpd. Analyses examining relationships between
30 tetrachloroethylene and visual function were conducted with the referent exposure group
31 (background exposure, living in a building without a dry cleaner), <0.015 ppm (<100 |ig/m3),
32 and >0.015 ppm (>100 |ig/m3).
33 Several analyses suggest a susceptibility of exposed subjects, particularly among children,
34 to tetrachloroethylene on VCS performance at higher spatial frequencies. A decreasing trend
35 (p < 0.05) was observed between increasing residential tetrachloroethylene exposure and the
36 proportion of adults achieving the maximum contrast sensitivity score at 6 cpd and in the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-90 DRAFT-DO NOT CITE OR QUOTE
-------
1 proportion of children achieving the maximum contrast sensitivity score at 6 and 12 cpd; i.e., a
2 lower proportion of participants with a maximum VCS score in the highest exposure category
3 compared to referents living in a building without a dry cleaner. Stratified analyses suggested a
4 lower percentage of low income and minority children with maximum VCS scores at a given cpd
5 than higher income and non-minority children, but sample sizes in the highest exposure group,
6 especially in higher income, non-minority groups, limit reliability of this observation. Race did
7 not appear to confound the association in adults between VCS at 6 cpd and tetrachloroethylene.
8 VCS scores in children at a given cpd were generally higher (better contrast vision) than
9 the VCS score of an adult living in the same apartment. Using differences between adult-child
10 pairs in each exposure grouping to assess possible tetrachloroethylene effects, the advantage of
11 children over adults appeared much smaller in the >0.015 ppm (>100 |ig/m3) category at 12 cpd
12 (mean difference of 10.9) compared to referents (mean difference of 21.6) but was not
13 statistically significant from the mean difference in the referent population (p = 0.16).
14 Results from logistic regression analyses further support susceptibility of children but not
15 adults. Whereas adult VCS at 6 or 12 cpd was not significantly influenced by any measure of
16 tetrachloroethylene exposure, VCS performance at 12 cpd among children was significantly
17 influenced (p < 0.05) by tetrachloroethylene concentrations in either indoor air or in blood; that
18 is, a lower percentage of children achieved a maximum VCS score with higher
19 tetrachloroethylene exposure. Analyses of tetrachloroethylene breath concentrations and VCS
20 performance at 12 cpd in children appeared to support the findings with indoor air and blood, but
21 were of borderline statistical significance. Logistic regression models examining VCS findings
22 in either children or adults are not adjusted for potentially confounding factors such as SES,
23 education, smoking, alcohol use, age (for children) and gender (for children); these variables
24 were correlated with one another as well as with tetrachloroethylene, but not with VCS
25 performance.
26 Color vision was assessed biocularly using both the Farnsworth D15 and Lanthony's
27 Desaturated 15 Hue Test. Both tests were administered under light conditions specified by the
28 manufacturer. The number of errors for each eye was recorded by noting instances of inversions
29 involving a single cap (minor error) and instances of inversions involving two or more caps
30 (major errors). Total Color Distance Scores (TCDS) were determined and a CCI was calculated
31 for each participant according to Geller (2001) and Bowman (1982).
32 Analyses were carried out using the proportion of subjects with no errors, comparing
33 quantitative differences in CCI, and logistic regression modeling to assess associations between
34 tetrachloroethylene exposure measures and occurrence of any major errors. A comparison of
35 differences in CCI and major error between children and adults residing in the same household
36 was used to assess the possible vulnerability of children. A high proportion of adult and child
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-91 DRAFT-DO NOT CITE OR QUOTE
-------
1 participants scored perfectly on both the Farnsworth and Lanthony color vision tests. Lower
2 annual household income, being a member of a minotiry group, speaking primarily Spanish at
3 home, and fewer years of education were all significantly associated with increased CCI on both
4 color vision tests.
5 Tetrachloroethylene measures of exposure were unrelated to color vision performance
6 among adults; however, similar to VCS performance, children appear as a susceptible population.
7 There were no differences between exposure groups for either adults or children in the percent of
8 subject with major effort on both color vision tests. A comparison of mean CCI between
9 exposure groups showed that children in the highest exposure category performed worse (mean
10 CCI of 1.3, range 1.0-1.9) than children in the low exposure category (mean CCI of 1.1, range
11 1.0-1.7) and to referent children (mean CCI of 1.2, range 1.0-2.0) on the Lanthony test; the test
12 for trend for the three exposure groups was statistically significant (p < 0.05). Performance
13 (mean CCI) on the less sensitive Farnsworth test was not associated with tetrachloroethylene
14 exposure in either adults or children. Moreover, for children, tetrachloroethylene in breath was
15 significantly associated (p < 0.05) with making one or more major errors on the Lanthony color
16 vision test in logistic regression analyses that adjusted for the effects of age and gender. Logistic
17 regression analyses examining color vision and other tetrachloroethylene measures such as
18 indoor tetrachloroethylene concentration or breath concentration were not discussed in NYS
19 DOH (2005a.). Last, the higher mean difference in CCI between children and adults in the
20 highest exposure category, >0.015 ppm (>100 |ig/m3), and referents was statistically significant.
21 Children in the high exposure group were a year younger than in other exposure groups; age was
22 correlated with CCI and with tetrachloroethylene exposure in this study. The highly correlated
23 variables and the few numbers of children in the high exposure group limits analysis of age
24 effects on the association between breath tetrachloroethylene concentration and CCI.
25 In summary, this study adopts a different approach than Schreiber et al. (2002) to assess
26 vision, using far vision methods as opposed to the near vision methods of Schreiber et al. (2002).
27 For both contrast vision and color vision, a number of analyses in NYS DOH (2005a) are
28 suggestive of vulnerability among children. The association with vision effects in children and
29 exposure to >0.015 ppm (>100 |ig/m3) tetrachloroethylene support findings from the earlier pilot
30 study (Schreiber et al., 2002). Exposure to >0.015 ppm (>100 |ig/m3) tetrachloroethylene was
31 highly correlated with race and children's age, and the sample sizes in the highest exposure
32 group, especially in higher income, non-minority groups, makes it difficult to fully examine
33 possible effects of income, race, and age on vision. However, association of tetrachloroethylene
34 exposure >0.015 ppm (>100 |ig/m3) with visual deficits suggests a susceptibility of the
35 population studied.
36
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-92 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.6.1.2.13.2. Pumpkin Patch Day Care Center follow-up evaluation. The objective of the
2 PPDCC Follow-up Evaluation was to assess neurobehavioral function in former students of
3 PPDCC after a 5 year post exposure period and to carry out first-time testing of visual function
4 of the former students. Additionally, visual function testing was carried out on five staff exposed
5 to tetrachloroethylene 5 years previously. The NYS DOH final report to EPA (NYS DOH,
6 2005c) provides a full description of testing in children but not adults. The discussion of visual
7 tests on former PPDCC is contained in NYS DOH (2005b). The initial testing in 1998 of vision
8 in PPDCC staff and of neurobehavior in children is contained in NYS DOH (2005b).
9 Children eligible for testing in the current evaluation were enrolled in the New York State
10 Volatile Organic Chemical (VOC) Registry and had attended PPDCC. There were 115 children
11 who met this criteria. Of this group, 27 children with the highest number of hours spent at
12 PPDCC were asked through letters or by phone to participate; 17 children completed vision
13 testing and 13 children completed some or all of the neurobehavioral assessment. Referents
14 were children who attended other day care centers and who were about the same age as PPDCC
15 participants. No information is provided on methods employed for referent participation.
16 Exposed and referent subjects were matched on day care experience, age, and gender. Overall,
17 17 PPDCC and 13 comparison children completed vision testing and 13 PPDCC and 13
18 comparison children completed neuropsychological testing. Of these subjects, 13 matched pairs
19 completed vision test; but only 8 matched pairs completed the neurobehavioral test. No
20 information is provided in the NYS DOH final report on how many of the 13 former PPDCC
21 students were part of the PPDCC student group who underwent neurological testing 5 years
22 previously one month after the close of the dry cleaner facility.
23 Neurobehavioral evaluations consisted of a battery of tests that assess general intellectual
24 function, attention/information processing speed, visuospatial ability, reasoning and logical
25 analysis, memory, motor functions, and sensory-perceptual functions. Tests were administered
26 in fixed order on two different days. All children completed the same tests with the exception of
27 the Halstead-Reitan Neuropsychological Batteries. Children age eight or younger were
28 administered the Reitan-Indiana Neuropsychological Test Battery and the Halstead-Reitan
29 Neuropsychological Test Battery for Old Children was administered to children age nine or older.
30 Children also performed portions of the computerized Neurobehavioral Evaluation System-2
31 (NES-2) which assessed perceptual-motor skills, attention, visual memory, and mood. A parent
32 or guardian completed the Child Behavioral Checklist and a background history questionnaire.
33 All neurobehavioral evaluations were conducted at the office of Albany Psychological
34 Associated, P.C., in Albany, NY.
35 Independent samples t-tests were performed on the scores from the Wechsler Intelligence
36 Scale for Children, Children's Memory Scale, the Halstead-Reitan Neuropsychological Test
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-93 DRAFT-DO NOT CITE OR QUOTE
-------
1 Battery for Old Children, and Reitan-Indiana Neuropsychological Test Battery. Age was
2 significantly correlated with performance on the Purdue Pegboard and many subtests in the NES-
3 2 and analysis of covariance was completed on subtests from the NES-2 and Purdue Pegboard
4 with age as a covariate. Each child's performance level on the neurobehavioral tests was
5 determined by comparing his/her test score to normative information for the specific test or
6 battery. For NES-2, performance of the referent children, children who attended other day care
7 centers in Albany and who were about the same age as PPDCC participants, was used as the
8 normative basis, with scores 2 S.D. below the mean of the same age and gender from the
9 normative data being classified as impaired.
10 Neurobehavioral function of the 13 PPDCC children evaluated in this follow-up study did
11 not differ from that of the 13 referent children. PPDCC children performed better than referent
12 children on several tests but performance was within normative ranges. These results are not
13 surprising. Neuropsychological or behavioral testing was conducted in October 1998 by the
14 auspices of the U.S. Centers for Disease Control (U.S. CDC) on children then of ages four and
15 five. No consistent differences in neurological function were found between 18 children who
16 then attended the day care center and 18 age- and gender-matched control children who did not
17 attend the day care center, although a statistically significant association was found between
18 duration of attendance at PPDCC and poorer performance on the Purdue pegboard test with the
19 dominant hand (NYS DOH, 2005b).
20 Visual function testing consisted of visual acuity, far VCS, and color vision. Visual
21 contrast sensitivity was determined monocularly using the Functional Acuity Contrast Test
22 distance chart placed 10 feet from the participant under light conditions specified by the
23 manufacturer. Scores for each eye were recorded on a graph showing a normal range (90%
24 confidence interval) of VCS at each spatial frequency. Color vision was assessed using both the
25 Farnsworth D15 and Lanthony's Desaturated 15 Hue Test under light conditions specified by the
26 manufacturer. For both tests, for each eye, participants were shown a rectangular box containing
27 16 color caps arranged in chromatic order. The test administrator removed 15 caps, leaving the
28 first as a stand, and randomized them in front of the participant. Participants were asked to place
29 the cap which most closely matched the stand in hue in the box next to the stand, and to continue
30 the process until all colored caps were in the box. When the participant was done, the order of
31 cap placement was recorded and diagramed on templates accompanying the tests. Both color
32 vision and contrast sensitivity tests were performed monocularly. Ophthalmologic examinations
33 and visual function testing was performed by Cornea Consultants of Albany, NY. Examiners
34 were not blinded but were not told whether participants were associated with the PPDCC.
35 VCS results for all 13 matched pairs of children were analyzed using the Wilcoxon
36 matched-pairs signed-ranks test. Two pairs of PPDCC and comparison children were six years
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-94 DRAFT-DO NOT CITE OR QUOTE
-------
1 old during vision testing; all other children were aged seven or more years. So as to examine the
2 effect of age on visual function, analyses were conducted using the 13 child pairs and excluding
3 two pairs who were <6 years old. PPDCC children performed better on the VCS test compared
4 to referent children.
5 Color vision results were evaluated in several ways. Statistical analyses were performed
6 for matched pairs (n= 13) only. Proportions of pairs of children with discordant clinical
7 judgements and with discordant numbers of major errors were assessed using McNemar's Exact
8 Test for Correlated Proportions. Furthermore, difference in CCI between matched PPDCC and
9 referent children were assessed using Wilcoxon Matched-Pairs signed-rank test. As for VCS
10 results, analyses were conducted using the all child pairs and excluding pairs who were <6 years
11 old. No significant difference in proportions of children with abnormal color vision or with
12 children making major errors between PPDCC and comparison children for either of the color
13 vision tests were found. Similarly, PPDCC and referent children were not significantly different
14 on CCI for either color vision test.
15
16 4.6.1.2.14. Perrin, MC; Opler, MG; Harlap, S; Harkavy-Friedman, J; Kleinhau,s K; Nation,
17 D; Fennig, S; Susser, ES; Malaspina, D. 2007. Tetrachloroethylene exposure and risk of
18 schizophrenia: Offspring of dry cleaners in apopulaton birth cohort, preliminary findings.
19 Schizophr Res. 90(1-3):251-254. Perrin et al. (2007) evaluates the time to a diagnosis of
20 schizophrenia among a cohort of 88,829 births born between 1964-1976 in the Jerusalum
21 Perintal Project, a population-based cohort. Births in this cohort were linked to the database of
22 Israel's Psychiatric Case Registry (PCR), with cases identified using a broad definition of
23 schizophrenia-related disorders as recorded as hospital discharge codes. Diagnoses for
24 individuals with psychosis were validated and the date of onset was identified as the date of first
25 psychiatric admission. Of the 88,829 births, 136 offspring were born to parents identified as
26 having a job title of dry cleaner on the birth certificate; 120 offspring whose fathers but not
27 mothers were dry cleaners, 20 whose mothers but not fathers were dry cleaners; and 4 with both
28 parents as dry cleaners; 4 of the 136 births had a later diagnosis of schizophrenia. The relative
29 risk (crude) between schizophrenia and parental employment in dry cleaning was 3.9 (95% CI =
30 1.3-9.2) using proportional hazard methods. The investigators noted risk estimates did not
31 greatly change when fitting proportional hazard models that adjusted for a number of potentially
32 confounding variables; although adjusted relative risk (RR) estimates are not reported in the
33 paper. Variables considered as possible confounders were parents' age, father's social class,
34 duration of marriage, rural residence, religion, ethnic origin, parental immigration status,
35 offspring's birth order, sex, birth weight and month of birth. Family history of mental illness
36 was not included as a covariate; rates of schizophrenia are higher among relatives of patients
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-95 DRAFT-DO NOT CITE OR QUOTE
-------
1 than in the general population (Mueser and McGurk, 2004). The magnitude of this possible bias
2 on the association between parental occupational employment as a dry cleaner and schizophrenia
3 in offspring can not be judged given the information provided in the paper.
4
5 4.6.1.3. Summary of Neuropsychological Effects in Low- and Moderate-Exposure Studies
6 It is important to compare outcomes across studies in order to determine whether it is
7 possible to identify a pattern of neuropsychological deficits produced by tetrachloroethylene.
8 Table 4-5 is presented as an aid for this comparison. Primarily these studies have assessed
9 neurobehavioral and, to a limited extent, neurophysiological effects of tetrachloroethylene
10 exposure using a number of statistical methods of varying sensitivity, from simple methods that
11 are more susceptible to multiple comparison errors to regression analyses that control for
12 potentially confounding effects. A clinical neurological examination that includes the Romberg
13 test, tests of body balance, and neuroradiological examination has not been widely incorporated
14 into the tetrachloroethylene epidemiologic studies. Neurophysiological tests such as EEGs,
15 nerve conduction tests, and evoked potentials (EPs) have seen limited use for assessing
16 neurotoxicologic effects in tetrachloroethylene-exposed populations. Although statistically
17 significant alterations in VEPs were reported by Altmann et al. (1990, 1992) with 4-hr acute
18 exposure at 10 ppm, they were not altered in residents exposed chronically to a median of around
19 1 ppm tetrachloroethylene (Altmann et al., 1995).
20 Acute and chronic exposures are of different patterns, short-term peak exposure versus
21 longer duration exposure, and, therefore, may result in a different pattern of toxicity.
22 Furthermore, studies assessing peripheral neuropathy and tetrachloroethylene uniquely were not
23 found, and studies reporting tetrachloroethylene exposure as one of a number of solvent
24 exposures (Albers et al., 1999; Antti-Poika, 1982a, b) are not informative, as discussed in
25 Section 4.6.1.
26 Several occupational studies of dry cleaner and laundry workers and the residential study
27 by Altmann et al. (1995) share a common set of tests from a neurobehavioral battery. Tests in
28 this battery have been widely administered to occupational populations in different settings with
29 a reasonably high degree of reliability (Anger et al., 2000). Moreover, these tests have been used
30 in clinical or experimental research to assess normal nervous system functioning, and they
31 measure a range of sensory and cognitive function. Studies that used a test battery include
32 Ferroni et al. (1992), Seeber (1989), Echeverria et al. (1994, 1995), and Altmann et al. (1995).
33 Both the Seeber and the Echeverria et al. studies involved more subjects than did the studies by
34 Ferroni et al. and Altmann et al. and statistical analyses, such as in Altmann et al., controlled for
35 a number of potentially confounding factors. The Ferroni et al. study was not well-reported and
36 was methodologically poorer than the other studies.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-96 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-5. Summary of neuropsychological effects of tetrachloroethylene in humans
Neurological
effects
Contrast
sensitivity
(spatial vision)
CCI
(color vision)
VEP
(vision)
Fine motor
function
Simple RT
(attention,
motor)
Continuous
performance
(vigilance)
Visuo spatial
function
Study, number of subjects, exposure level (ppm)
Cavalleri
etal.
(1994)
70
7
+/-b
Echeverria
etal.
(1994, 1995)a
173
<0.2
3
9
+,+,+
pattern
recognition,
reproduc-tion,
memory
65
11
23
41
+,+,+
pattern
recogni-tion,
reproduc-tion,
memory
Ferroni et
al. (1992)
90
15
-
+
+
Seeber
(1989)
185
12
53
-
+
digit
reproduc-
tion
Spinatonda
et al. (1997)
74
8
Altmann et
al. (1995)
37
0.7
-
-
+
+
+
Benton
Schreiber et al. (2002)
Residents
34
0.4
+
trend +
Daycare
18
0.3
+
Lauwerys
et al. (1983)
59
21
—
Nakatsuka
et al. (1992)
184
15
C
-------
Table 4-5. Summary of neuropsychological effects of tetrachloroethylene in humans (continued)
Neurological
Information
processing
speed
Digit span,
digit symbol
Cancellation
(visual
scanning)
Trailmaking
(executive
function)
Study, number of subjects, exposure level (ppm)
Cavalleri
et al.
(1994)
70
7
Echeverria
et al.
(1994, 1995)a
173
<0.2
3
9
65
11
23
41
_
-, - (both)
Ferroni et
al. (1992)
90
15
_
-
Seeber
(1989)
185
12
53
+
perceptual
threshold
"delayed
reaction" on
choice
reaction time
+/-,+
+
Spinatonda
et al. (1997)
74
8
+
vocal
reproduction
Altmann et
al. (1995)
37
0.7
Schreiber et al. (2002)
Residents
34
0.4
Daycare
18
0.3
Lauwerys
et al. (1983)
59
21
Nakatsuka
et al. (1992)
184
15
a Field study, no unexposed controls.
b Positive in dry cleaners, negative in ironing workers with lower exposure.
0 Using less sensitive test instrument and data analysis procedure than the other studies.
RT = Reaction time
+ = Impaired performance in exposed group
- = No effect of tetrachloroethylene
-------
1 Cognitive domains affected by tetrachloroethylene include visuospatial function,
2 attention, vigilance, and speed of information processing (choice reaction time; Table 4-5).
3 Effects on visuospatial function are of particular interest, given the finding in the four studies
4 that examined this domain and similar reports for other solvents (Morrow et al., 1990; Daniel et
5 al., 1999). Echeverria et al. (1995) found effects on tests of pattern memory, visual reproduction,
6 and pattern recognition in the absence of effects on attention (digit symbol and digit span) or
7 executive function (Trailmaking A and B). Further, Echeverria and colleagues (1994) confirmed
8 these findings in an independent sample of dry cleaners in their follow-up study (U.S. EPA,
9 2004).
10 Seeber (1989) also reported impaired visuospatial recognition in both exposure groups,
11 and Altmann et al. (1995) observed deficits on a test of visuospatial function in residents with
12 much lower exposure concentrations than those of the two occupational studies. These studies
13 are considered to provide strong weight, given the numbers of subjects and their use of
14 appropriate statistical methods, including adjustment for potentially confounding factors.
15 Additionally, they considered potential bias and confounding more carefully than did other
16 studies in this review.
17 Altmann et al. (1995) and Ferroni et al. (1992) assessed vigilance using a continuous
18 performance procedure in which the subject faces a screen that presents one of several different
19 stimuli at random intervals. The subject must make a response to a specified stimulus and not to
20 the others. This test measures sustained attention and is correlated with performance on tests of
21 executive function. Both studies found deficits as a result of tetrachloroethylene exposure on
22 this task. Seeber (1989) found effects on two tests of attention (cancellation d2 and digit
23 symbol) that are subsets of the Weschler IQ tests and were designed to be sensitive to
24 performance within the normal range. These investigators also found positive effects on a visual
25 scanning test that is usually used to assess laterality of brain damage but has also proved
26 sensitive to toxicant (lead) exposure (Bellinger et al., 1994). In contrast, Echeverria et al. (1995)
27 and Ferroni et al. (1992, as described in NYS DOH, 1997) did not find effects on digit span,
28 which is given as a test of attention and memory, or digit symbol, despite higher levels of
29 exposure than in Seeber (1989).
30 Two of these studies—an occupational study with relatively higher exposure (Ferroni et
31 al., 1992) and the Altmann et al. (1995) residential study—also assessed simple reaction time, a
32 task that uses a motor response and demands a relatively modest amount of attention; results
33 were positive in both studies. Speed of information processing was assessed in two studies,
34 Seeber (1989) and Spinatonda et al. (1997). Seeber used two tasks: recognition and choice
35 reaction time. Effects were observed in both groups on a task requiring recognition of briefly
36 presented stimuli. In a choice reaction time task, effects were borderline in the lower-exposure
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-99 DRAFT-DO NOT CITE OR QUOTE
-------
1 group and negative in the higher-dose group, with no exposure-response relationship.
2 Spinatonda et al. (1997) found effects on response to vocal and visual stimuli. A third study,
3 Lauwerys et al. (1983), reported better performance on simple and choice reaction times.
4 Of the occupational studies, greatest weight is placed on the Seeber (1989) observations
5 due to the larger number of study subjects and to their consideration in the statistical analysis of
6 potentially confounding factors. Ferroni et al. (1992), Spinatonda et al. (1997), and Lauwerys et
7 al. (1983) all reported limited information in their published papers, particularly regarding
8 potential confounding and bias, and because of this, they have greater inherent uncertainties than
9 does Seeber (1989).
10 Tetrachloroethylene exposure has not been reported to affect fine motor tests. Seeber
11 (1989), Ferroni et al. (1992), and Altmann et al. (1995) each assessed fine motor control using
12 various instruments and all three found no significant decrements in fine motor performance.
13 Deficits in blue-yellow color vision, a well established effect of solvents, were observed
14 in the high-exposure group (mean tetrachloroethylene concentration of 7 ppm) but not the low-
15 exposure group (mean tetrachloroethylene concentration of 5 ppm) in Cavalleri et al. (1994) and
16 in Muttray et al. (1997)—a study carrying lesser weight than that of Cavalleri et al.—of workers
17 previously exposed to a mixture of solvents that contained tetrachloroethylene. Overall, the
18 findings of the Cavalleri et al. study and its follow-up study (Gobba et al., 1998) are in
19 agreement with previous reports on other solvents (Geller and Hudnell, 1997; Mergler et al.,
20 1996; Mergler and Blain, 1987): the blue-yellow range of color vision was primarily affected in
21 the dry cleaners, with only a few workers showing an effect on red-green perception.
22 The absence of a color vision effect in Nakatsuka et al. (1992), who used confirmatory
23 methods to augment their screening method of Lanthony's new color test, may not be
24 inconsistent with the findings of Cavalleri et al. (1994) and Gobba et al. (1998). There are
25 uncertainties regarding testing lighting conditions in Nakatsuka et al. (1992)—an important
26 determinant of a subject's response (Geller and Hudnell, 1997)—and the fewer subjects in this
27 study than in Cavalleri et al. (1994). A pilot study of residents living above dry cleaners with
28 mean tetrachloroethylene exposure during active dry cleaning of 0.4 ppm (Schreiber et al., 2002)
29 also reported a trend of decreasing color vision, although this finding was not statistically
30 significant. The follow-up study of NYS DOH (2005a), reported to U.S. EPA as a final grant
31 report, is further suggestive of tetrachloroethylene effects on color vision, particularly in children
32 compared to their parents. Tetrachloroethylene exposure concentrations had decreased since
33 Schreiber et al. (2002), making it it difficult to find higher-exposed subjects. Higher
34 tetrachloroethylene exposure, that is, exposure at or over 0.1 ppm, was highly correlated with
35 SES and belonging to a minor population. This study is not able to adjust for these possible
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-100 DRAFT-DO NOT CITE OR QUOTE
-------
1 confounders given the sample size. Studies of a larger number of residents with similar exposure
2 concentrations are needed to draw more definitive conclusions.
3 Only Schreiber et al. (2002) and NYS DOH (2005a) assessed spatial vision, an effect
4 reported for exposure to other solvents (Bowler et al., 1991; Broadwell et al., 1995; Campagna et
5 al., 1995; Donoghue et al., 1995; Frenette et al., 1991; Hudnell et al., 1996a, b; Mergler et al.,
6 1991). VCS deficits in subjects with normal visual acuity were observed at low exposure
7 concentrations in residential populations that were subject to very different dose-rates than those
8 incurred by occupational workers (U.S. EPA, 2004). This finding is based on few subjects in
9 this study and is noteworthy for this reason, even in light of questions regarding potential biases.
10 Potential bias and confounding could be introduced, in part, from a lack of blinding of testers,
11 differences in motivation between exposed and referent subjects for participating in the study,
12 and individual differences in exposed and control populations (U.S. EPA, 2004). As discussed in
13 Section 4.6.1.2.11, occupation has not been found to strongly affect contrast sensitivity (Hudnell
14 et al., 2001), nor does motivation (U.S. EPA, 2004).
15 Peer consultation comments on EPA's earlier draft "Neurotoxicity of Tetrachloroethylene
16 (Perchloroethylene) Discussion Paper" (U.S. EPA, 2003b) noted that the deficit in contrast
17 sensitivity could reflect a sensitivity of the visual system to tetrachloroethylene, or it may be that
18 this test was simply carried out by a superior test method (U.S. EPA, 2004). Furthermore, the
19 peer consultants also suggested that contrast sensitivity loss may reflect impaired function
20 throughout the brain, because contrast sensitivity is affected by retinal, optic nerve, or central
21 brain dysfunction (U.S. EPA, 2004). Nonetheless, drawing strong conclusions from a single
22 study is difficult, particularly in light of the paucity of data on this test in occupational
23 populations with higher exposure concentrations and in animal studies. The finding of poorer
24 performance among children with exposure to >0.1 ppm tetrachloroethylene compared to adults
25 in NYS DOH, (2005a) adds some support for observations in Schreiber et al. (2002).
26 Mental disease of neurologic origin has not been well studied with respect to
27 environmental factors. Perrin et al. (2007), who reports an association between schizophrenia
28 and parental exposure in dry cleaning, is the only such study. Other studies are needed to
29 understand the role of parental tetrachloroethylene exposure in the development of mental
30 disease in children.
31 The epidemiologic studies all have limitations. The body of evidence is characterized
32 generally by a lack of studies that adopt sensitive tests of important functions affected by
33 solvents. The exceptions are the studies employing vision assessment and the Echeverria et al.
34 (1994, 1995) studies. In most studies, investigators were not blinded to subject status, a potential
35 source of bias, particularly in situations in which the investigator interacted directly with the
36 subject during testing. All studies used a cross-sectional design, which is weaker than a
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-101 DRAFT-DO NOT CITE OR QUOTE
-------
1 longitudinal design for a number of reasons, including a greater potential for selection bias and
2 exposure misclassification (the latter of which would bias the results toward the null). One
3 possible source of selection bias is motivational differences between exposed and control
4 populations. The designs of the Schreiber et al. (2002) and Ferroni et al. (1992) studies may
5 have introduced such unwanted bias, although motivation has been found to more strongly
6 influence performance on color vision but not contrast sensitivity tests (U.S. EPA, 2004).
7 Several studies provided insufficient details on the population from which controls were selected
8 (Seeber, 1989; Spinatonda et al., 1997; Ferroni et al., 1992), or the details provided raise
9 concerns regarding the appropriateness of the control group (Seeber, 1989; Spinatonda et al.,
10 1997; Schreiber et al., 2002 [residents only]).
11 For some of the occupational studies, the descriptions of behavioral testing procedures or
12 results were insufficient or ambiguous (Ferroni et al., 1992; Seeber, 1989; Nakatsuka et al.,
13 1992; Spinatonda et al., 1997). A number of studies had either insufficient control for possible
14 influences of education (Cavalleri et al., 1994; Gobba et al., 1998; Schreiber et al. (2002) [day
15 care study]) or provided insufficient detail on the study populations (Schreiber et al., 2002
16 [residents]; Nakatsuka et al., 1992; Ferroni et al., 1992; Spinatonda et al., 1997). Further,
17 adjustment for education in the statistical analysis may not have been fully adequate due to the
18 use of categorical variables (Seeber, 1989; Altmann et al., 1995). Additionally, Spinatonda et al.
19 (1997) did not provide detailed information on variables other than age of subjects, precluding a
20 determination of whether subjects may or may not have been comparable.
21 Alcohol by itself cannot explain the observed deficits in neurobehavioral functions,
22 because either study designs excluded subjects who were moderate to heavy drinkers, or
23 statistical analyses of the epidemiologic observations controlled for this covariate. However,
24 effects from the interaction between tetrachloroethylene exposure and alcohol consumption were
25 not well investigated in these studies. Valic et al. (1997) showed greater decrements in color
26 vision among subjects with exposures to both tetrachloroethylene and ethanol when compared
27 with individuals with solvent exposure only to the solvents or to neither substance.
28 Many studies did not include exposure monitoring of individual subjects, and the
29 statistical analyses compare groups using t-tests or chi-square tests, with the result of a greater
30 dependency on the performance in the control group. Dose response analyses are statistically
31 more powerful. However, despite the use of t-tests or chi-square tests, deficits in
32 neurobehavioral or neurophysiological functions were reported in these studies. A number of
33 statistical comparisons were made in these studies, increasing the possibility of a type I, or false
34 positive, error. The issue of multiple comparisons is always present in risk assessment and
35 evaluation activities; however, unfortunately, reducing the type I error increases the type II, or
36 false negative, error for those associations that are not null. As mentioned above, inferences
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-102 DRAFT-DO NOT CITE OR QUOTE
-------
1 about associations between exposure and effects are best drawn from a body of evidence where
2 consistency across studies of different designs, populations, and statistical methods may be
3 obtained; value and richness can be found when consistency emerges from the diversity and
4 despite the flaws.
5 These studies do have important strengths. They describe susceptibility to
6 tetrachloroethylene toxicity in humans, providing evidence to augment findings from animal
7 toxicity testing. Further, the majority of studies, with the exception of Schreiber et al. (2002),
8 exceeded a smallest cell size of 40 exposed subjects that is generally considered sufficient to
9 detect preclinical effects in a group that range from 3 to 18, or 20%, from normal function.
10 Several studies (Altmann et al., 1995; Schreiber et al., 2002; Echeverria et al., 1995) employed
11 multiple measures of exposure (indoor air monitoring, personal monitoring, and in some cases,
12 biological monitoring), with a high degree of correlation between tetrachloroethylene
13 concentration as assessed from indoor air monitoring or personal monitoring and biological
14 metrics such as blood tetrachloroethylene concentration, suggesting indoor air concentration as a
15 reasonable exposure metric.
16 Several independent lines of evidence can be found in the occupational and residential
17 studies to support an inference of a broad range of cognitive and behavioral deficits following
18 tetrachloroethylene exposure (U.S. EPA, 2004). First, adverse effects on visuospatial function
19 are reported in three studies (Seeber, 1989; Altmann et al., 1995; Echeverria et al., 1995), with
20 Echeverria et al. (1994) as a confirmatory study of Echeverria et al. (1995). The results across
21 these three studies appear reasonably consistent, despite substantial differences in study design.
22 A second line of evidence can be found in both the occupational (Nakatsuka et al., 1992;
23 Cavalleri et al., 1994) and residential studies (Schreiber et al., 2002), both of which evaluated
24 performance on the Lanthony color vision test. Cavalleri et al. (1994) reported a decrement in
25 color vision in the high exposure group, but not the low exposure group, and a significant dose-
26 response relationship between CCI value and tetrachloroethylene concentration. The lack of an
27 association between color vision and tetrachloroethylene exposure in Nakatsuka et al. (1992)
28 may not be inconsistent, given significant weaknesses in this study (U.S. EPA, 2004).
29 Performance of the residents and day care workers who worked in buildings with a co-located
30 dry cleaner appeared worse (particularly that of residents) than the performance of controls,
31 although CCI scores were not statistically significantly different from referents (Schreiber et al.,
32 2002). Last, VCS deficits were observed in these residents and day care workers. These
33 subjects received exposures of lower dose rates, but a different and, for residents, a more
34 prolonged daily exposure duration than typical occupational exposures occurred. Overall, the
35 evidence reveals a high degree of consistency in visually mediated function.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-103 DRAFT-DO NOT CITE OR QUOTE
-------
1 Effects on spatial vision are well-known consequences of solvent exposure in industrial
2 workers (Bowler et al., 1991; Broadwell et al., 1995; Campagna et al., 1995; Donoghue et al.,
3 1995; Frenette et al., 1991; Hudnell et al., 1996a; Mergler et al., 1991). Other organic solvents,
4 as well as alcohol, induce effects on memory and color vision (Altmann et al., 1995; Mergler et
5 al., 1991; Hudnell et al., 1996a, b). By analogy, the observations on other solvents also support
6 an inference of neurobehavioral deficits following exposure to tetrachloroethylene.
7 In conclusion, the weight of evidence across the available studies of humans exposed to
8 tetrachloroethylene—and by analogy to other organic solvents—indicates that chronic exposure
9 to tetrachloroethylene may be associated with adverse decrements in nervous system function.
10
11 4.6.2. Animal Studies
12 4.6.2.1. Inhalation Studies
13 Mattsson et al. (1998) studied the effects of acute exposure to tetrachloroethylene for 13
14 weeks observing flash-evoked potentials (FEPs), somatosensory-evoked potentials (SEPs), EEGs,
15 and rectal temperature in F344 rats. During the acute (pilot) study, male rats were exposed to 0
16 or 800 ppm tetrachloroethylene for 6 hrs/day for 4 days and tested before and after exposure on
17 the 4th day. Changes in FEP, SEP, and EEG components were observed after acute exposure. In
18 the subchronic study, the above evoked potentials and caudal nerve conduction velocity were
19 determined in male and female rats exposed to 0, 50, 200, or 800 ppm for 6 hrs/day for 13 weeks.
20 Testing was performed during the week following cessation of exposure. Changes in FEP were
21 observed at the highest dose (800 ppm). Several measures of the evoked potential were affected,
22 at 50 ppm but not at higher doses. Other measures were not affected, and no dose response was
23 observed. The finding of an overall greater effect following short-term (4-day) exposure as
24 compared with longer-term exposure is similar to the findings of Moser et al. (1995) on a
25 number of measures of a neurotoxicity battery.
26 The effects of exposure to 90-3,600 ppm tetrachloroethylene for 1 hr on motor activity
27 were examined in male MRI mice (Kjellstrand et al., 1985). A strong odor (cologne) was used
28 as the control condition. Total activity was monitored during the dark period during exposure
29 and for several hours thereafter. All doses produced increased activity during exposure; activity
30 decreased over several hours after cessation of exposure. Although apparently no statistical
31 analyses were performed, it is clear from the figures that the lowest dose produced an average
32 performance that was well outside the boundary of the 95% CIs of the cologne-treated controls
33 and was dose-dependent. Tetrachloroethylene induced motor activity at concentrations lower
34 than those of any of the other organic solvents tested (methylene chloride, toluene,
35 trichloroethylene, 1,1,1-trichloromethane).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-104 DRAFT-DO NOT CITE OR QUOTE
-------
1 De Ceaurriz et al. (1983) exposed male Swiss OF1 mice to 596, 649, 684, or 820 ppm
2 tetrachloroethylene for 4 hrs. Immediately following exposure, subjects were immersed in a
3 cylinder filled with water and the duration of immobility was observed for 3 minutes. The term
4 "behavioral despair" has been coined for this initial immobility, and the length of immobility is
5 shortened by antidepressant administration. Tetrachloroethylene exposure also shortened the
6 period of immobility, with a no-observed-effect level (NOEL) of 596 ppm.
7 Nelson et al. (1980) of NIOSH, investigated developmental neurotoxicity in SD rats by
8 exposing pregnant dams to tetrachloroethylene at concentrations of 100 ppm or 900 ppm during
9 both early pregnancy (gestation days 7 to 13) or late pregnancy (gestation days 14 to 20). The
10 investigators made morphological examinations of the fetuses and performed behavioral testing
11 and neurochemical analysis of the offspring. There were no alterations in any of the measured
12 parameters in the 100 ppm groups. At 900 ppm there were no skeletal abnormalities, but the
13 weight gain of the offspring as compared with controls was depressed about 20% at weeks 3-5.
14 Developmental delay was observed in both the early and late pregnancy groups. Offspring of the
15 early pregnancy-exposed group performed poorly on an ascent test and on a rotorod test, whereas
16 those in the late pregnancy group underperformed on the ascent test only at postnatal day 14.
17 However, later in development (days 21 and 25), their performance was higher than that of the
18 controls on the rotorod test. These pups were markedly more active in the open field test at days
19 3 land 32.
20 There were no effects on running in an activity wheel on days 32 or 33 or avoidance
21 conditioning on day 34 and operant conditioning on days 40 to 46. Neurochemical analyses of
22 whole brain (minus cerebellum) tissue in 21-day-old offspring revealed significant reductions in
23 acetylcholine levels at both exposure periods, whereas dopamine levels were reduced among
24 those exposed on gestation days 7-13. Unfortunately, none of the statistics for the 100 ppm
25 treatments was presented. The authors observed that more behavioral changes occurred in
26 offspring exposed during late pregnancy than in those exposed during early pregnancy.
27 Szakmary et al. (1997) exposed CFY rats to tetrachloroethylene via inhalation throughout
28 gestation (i.e., gestation days 1—20) for 8 hrs/day at concentrations of 0, 1,500, or 4,500 mg/m3
29 tetrachloroethylene. The primary focus of the study was prenatal developmental evaluations (see
30 Section 4.7.2). However a cohort of rats (15 litters/group) was allowed to deliver, and the
31 offspring (standardized to 8 pups/litter) were maintained on study until postnatal day 100 and
32 evaluated for growth, development and neurotoxic effects. The report did not specify whether
33 the animals were exposed to tetrachlorotehylene after birth. Pre-weaning observations included
34 weekly body weights, developmental landmarks (pinna detachment, incisor eruption, and eye
35 opening), and functional assessments (forward movement, surface righting reflex, grasping
36 ability, swimming ontogeny, rotating activity, auditory startle reflex, and examination of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-105 DRAFT-DO NOT CITE OR QUOTE
-------
1 stereoscopic vision). After weaning, exploratory activity in an open field, motor activity in an
2 activity wheel, and development of muscle strength were assessed. The study authors reported
3 that adverse findings included a decreased survival index (details were not provided), a minimal
4 decrease of exploratory activity and muscular strength in treated offspring (presumably at both
5 exposure levels) which normalized by postnatal day 51, and significantly increased motor
6 activity on postnatal day 100 of females exposed to 4,500 mg/m3. Litter was evaluated as the
7 statistical unit of measure for all outcomes. There is no clear indication of group means for
8 postnatal measures reported. The lack of experimental detail in the postnatal evaluation part of
9 this study reduces the overall confidence in the findings. There was no evaluation of postnatal
10 histopathology of the nervous system reported or cognitive testing during the post weaning
11 period or during adulthood.
12 Wang et al. (1993) exposed male SD rats to 300 ppm tetrachloroethylene continuously
13 for 4 weeks or 600 ppm for 4 or 12 weeks. Exposure to 600 ppm at either duration resulted in
14 reduced brain weight gain, decreased regional brain weight, and decreased DNA in frontal cortex
15 and brain stem but not hippocampus. Four specific proteins (S-100 [an astroglial protein], glial
16 fibrallary acidic protein, neurone specific enolase, and neurofilament 68 kD polypeptide) were
17 decreased at 4 and/or 12 weeks exposure to 600 ppm; 300 ppm had no effect on any endpoint.
18 The effects of exposure to 200 ppm tetrachloroethylene 6 hrs/day for 4 days in male SD
19 rats were examined on a number of endpoints (Savolainen et al., 1977a, b). Rats were killed on
20 the 5th day following a further 0-6 hrs of exposure. Tetrachloroethylene levels were highest in
21 fat, followed by liver, cerebrum, cerebellum, lung, and blood. Tissue levels increased in all
22 tissues over the 6 hrs of exposure. Brain RNA content decreased, and brain nonspecific
23 cholinesterase was increased on the 5th day, although no statistical comparisons were performed.
24 Locomotion in an open field was increased immediately following the end of exposure on the 4th
25 day, with no difference 17 hrs after exposure, although no statistical comparisons were made.
26 Brain protein, GSH, and acid proteinase were unaffected.
27 A series of experiments were performed on the effects of tetrachloroethylene on brain
28 lipid patterns. Exposure to 320 ppm for 90 days (Kyrklund et al., 1990) or 30 days (Kyrklund et
29 al., 1988) in male SD rats resulted in changes in the fatty acid composition of cerebral cortex,
30 which persisted after a 30-day recovery period (Kyrklund et al., 1990). Similar results were
31 observed in cerebral cortex and hippocampus after exposure to 320 ppm in the Mongolian gerbil
32 (sex unspecified) in the presence of reduced brain weight (Kyrklund et al., 1987). Exposure of
33 male Mongolian gerbils to 120 ppm for 12 months also resulted in decreases in long-chain,
34 linolenic acid-derived fatty acids in cerebral cortex and hippocampus (Kyrklund et al., 1984).
35 The effect of tetrachloroethylene on neurotransmitter levels in the brain was explored in
36 male SD rats exposed continuously to 200, 400, or 800 ppm for a month (Honma et al., 1980a, b).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-106 DRAFT-DO NOT CITE OR QUOTE
-------
1 The 800 ppm dose produced a decrease in ACh in striatum, and there was a dose-related increase
2 in a peak containing glutamine, threonine, and serine in whole brain preparations. GAB A, NE,
3 5-HT, and other amino acids were not affected.
4 In a study from the same laboratory (Rosengren et al., 1986), Mongolian gerbils of both
5 sexes were exposed to 60 or 300 ppm tetrachloroethylene for 3 months, followed by a 4-month
6 solvent-free period. Changes in both S-100 and DNA concentrations in various brain regions
7 were observed at the higher concentration, and decreased DNA in frontal cortex was observed
8 after exposure to 60 ppm. The higher concentration also produced decreased brain but not body
9 weight. The results at 60 ppm were replicated in a follow-up study (Karlsson et al., 1987).
10 In a related study (Briving et al., 1986), Mongolian gerbils were exposed for 12 months
11 to tetrachloroethylene at 120 ppm. At the end of exposure, out of a total of 8 amino acids
12 assayed, taurine was significantly decreased in the two brain regions assessed (hippocampus and
13 cerebellum), and glutamine was elevated in hippocampus. y-Aminobutyric acid (GABA) levels
14 were unaffected, as was uptake of GAB A and glutamate.
15 Kyrklund and Haglid (1991) exposed pregnant guinea pigs to airborne
16 tetrachloroethylene continuously from day 33 through day 65 of gestation. The exposure was
17 continuous at 160 ppm except for 4 days at the beginning and end of the exposure period, when
18 it was reduced to 80 ppm. In the control group there were three dams with litter sizes of four,
19 three and two pups, and in the exposed group there were three dams with litter sizes of two each.
20 The pup body weights differed between litters. In the data analysis, three pups in the control
21 group were eliminated and the six pups in the treatment and control groups were assumed to be
22 independent, which is an invalid assumption. According to the authors' analysis, the offspring
23 had a slightly altered brain fatty acid composition, with a statistically significant reduced stearic
24 acid content in the tetrachloroethylene treatment group, which is consistent with the authors'
25 earlier findings in rats. This conclusion might have been different if the investigators had
26 grouped litters rather than pups as independent groups. The results suggest that
27 tetrachloroethylene could have reduced the litter size, but a much larger study would be
28 necessary to establish reduced litter size as an effect of tetrachloroethylene.
29 Caucasan male and female NMRI mice were exposed to 9, 37, 75, or 150 ppm
30 continuously for 30 days, to 150 ppm for one of several exposure periods ranging from 5 to 30
31 days, or to 150 ppm tetrachloroethylene for 30 days with various recovery periods (Kjellstrand et
32 al., 1984). Other groups were exposed intermittently on several dosing and exposure regimens
33 that resulted in a TWA of 150 ppm for 30 days. Plasma BuChE levels, organ weights, liver
34 morphology, and motor activity were assessed. BuChE was elevated after continuous exposure
35 to 37 ppm or greater. Liver weight was increased at all doses following continuous exposure,
36 and body weight decreased at 37 ppm or above. Motor activity results following continuous
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-107 DRAFT-DO NOT CITE OR QUOTE
-------
1 exposure were not reported. BuChE and liver weight were both elevated at a TWA of 150 ppm
2 for 30 days, regardless of the length of the exposure pulse. This was true even for an hour's
3 exposure (at 3,600 ppm) as well as at the lowest concentration (225 ppm). All concentrations of
4 intermittent exposure increased motor activity. A recovery period reversed the effects on BuChE,
5 whereas liver weight was still slightly elevated at 150 days after cessation of exposure. Changes
6 in liver morphology were detected following exposure to 9 ppm for 30 days and reversed after
7 cessation of exposure.
8 Tinston (1994) performed a multi-generation study of the effects on rats exposed to
9 airborne concentrations of tetrachloroethylene. The details of the study are discussed in Section
10 4.7.2. The investigators observed several developmental effects. Of interest here were the signs
11 of CNS depression (decreased activity and reduced response to sound) observed for the first 2
12 weeks in both adult generations and when the exposure was resumed on day 6 postpartum in the
13 Fl generation (adults and pups). These effects disappeared about 2 hrs after cessation of the
14 daily exposure. Other overt signs of tetrachloroethylene poisoning among the adults included
15 irregular breathing and piloerection at both 300 and 1,000 ppm. These changes stopped
16 concurrently with cessation of exposure or shortly thereafter.
17
18 4.6.2.1.1. Summary of animal inhalation neurotoxicity studies. In order to compare the animal
19 inhalation neurotoxicity studies with each other and to evaluate whether there is any relationship
20 across studies between the LOAEL of the administered dose and the duration of treatment, the
21 data were summarized (Table 4-6). In order to estimate the lowest concentration at which a
22 given effect occurs, the experiment showing that effect must have both a LOAEL (the lowest
23 concentration at which the effect occurred) and a NOAEL (the next lower concentration where
24 the effect did not occur). The experiments that meet this criteria are Mattsson et al. (1998), De
25 Ceaurriz et al. (1983), Wang et al. (1993), Honma et al. (1980 a, b), and Kjellstrand et al. (1984).
26 The total duration of exposure in these experiments is plotted in Figure 4-2 as a function
27 of the LOAEL concentration in order to discover whether there is a systematic trend in this
28 relationship. The plot shows that there is no systematic trend. It also shows that the LOAEL
29 varies over a 22-fold range: from 37 ppm for 30 days for increased brain butyl cholinesterase in
30 mice observed by Kjellstrand et al. (1984) to 800 ppm for 13 weeks for alteration in the flash-
31 evoked potential in rats observed by Mattsson et al. (1998). Table 4-6 shows other observations
32 at comparatively low concentrations: decreased DNA in gerbils by Rosengren et al. (1986) and
33 Karlsson et al. (1987) at 60 ppm and increased motor activity in mice at 90 ppm, observed by
34 Kjellstrand et al. (1985). The LOAEL for these studies as a group is therefore in the range of 37
35 to 90 ppm, and the effects at these levels are changes in neurotransmitter levels and increased
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-108 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-6. Summary of animal inhalation neurotoxicology studies
Subjects
F344 rats
Pilot study: male
10/dose
Follow-up study: males
and females
12/sex dose
NMRI mice, males
Swiss OF1 mice, males
10/dose
SD rats
pregnant females
13-21 litters/dose
males and female
Offspring assessed
CFY rats
pregnant females
15 litters/dose
male and female offspring
assessed
SD rats, males
8/dose
SD rats, males
10/dose
SD rats, males
5-6/dose
SD rats, males
5-6/dose
Mongolian gerbils
males and females
6/sex/dose
Effect
Changes in FEP, SEP, EEC
Increased amplitude and
latency in late component
of FEP
Increase in motor activity
Decrease in duration of
immobility
Decreased weight gain
Behavioral changes, more
extensive for late
pregnancy exposure
Decreased brain
acetylcholine
Transient decreases in
muscular strength and
exploratory behavior.
Latent increases in motor
activity in females at 100
days postnatally
Reduced brain weight,
DNA, protein
Decrease in brain RNA,
increase in brain
cholinesterase and increase
motor activity
Change in fatty acid
composition of cerebral
cortex
Neurotransmitter changes,
brain regions
Decrease in DNA, frontal
cortex
Decrease in brain weight
NOAEL/ LOAEL"
(ppm)
0, 800. 4 days, 6 hr/d
50, 200. 800.
13 wks, 6 hr/d, 5 d/wk
2Q, 3,600 1 hr
596. 649, 684, 820
4hrs
0, 100. 900 on
GD7-13oronGD 14-20
7 hr/d
0, 1,500 or 4.500 me/m3
GD 1-20 for 8 hrs/day
300. 600.
4 or 12 wks continuous (24
hr/d)
200, 4 days
320. 12 wks
continuous (24 hr/d),
30-day washout period
320. 4 wks continuous (24
hr/d)
200, 400. 800,
4 wks continuous
(24 hr/d)
6fl, 300, 12 wks,
continuous (24 hr/d)
16-week washout period
Authors
Mattssonetal. (1998)
Kjellstrandetal. (1985)
DeCeaurrizetal. (1983)
Nelson etal. (1980)
Szakmary et al. (1997)
Wang etal. (1993)
Savolainen et al. (1977a, b)
Kyrklund etal. (1990,
1988)
Honmaetal. (1980a,b)
Rosengren et al. (1986)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-109 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-6. Summary of animal inhalation neurotoxicology studies (continued)
Subjects
Mongolian gerbils
gender unspecified
Mongolian gerbils
males and females
8/sex/dose
Mongolian gerbils
gender unspecified
6/dose
Mongolian gerbils
males 6/dose
Guinea pigs
pregnant females
3/litters/dose
males and female
Offspring assessed
NMRI mice,
males and females
3-8/sex dose
Males and females
10/sex dose
SD rats, multigeneration
study
28 litters/dose
Effect
Decrease in DNA, frontal
cortex
Decrease in brain weight
Taurine, glutamine changes
in brain regions
Decrease in brain weight,
change in fatty acids
Decreased brain long-chain
fatty acids
Decrease in brain stearic
acid in offspring after in
utero exposure13
Increase in butyl
cholinesterase
Increased motor activity
CNS depression in first 2
wksofFlandF2
generations, which ceased
2 hrs after daily exposures
NOAEL/LQAEL
(ppm)
60, 12 wks, continuous (24
hr/d)
120. 12 mos continuous (24
hr/d)
320, 12 wks continuous (24
hr/d)
120, 52 wks continuous (24
hr/d)
Maximum exposure 160.
GD 33 to 65 continuous
(24 hr/d)
9C, 32, 75, 150
4 wks continuous
(24 hr/d)
150. 4 wks intermittent-
(1, 2, 4, 8, or 16 hr/d)
0, 100. 300. 1,000
6 hr/d, 5 d/wk, except
during mating, 6 hr/d-7
d/wk
Authors
Karlssonetal. (1987)
Brivingetal. (1986)
Kyrklundetal. (1987)
Kyrklundetal. (1984)
Kyrklund and Hagid (199 1)
Kjellstrandetal. (1984)
Kjellstrandetal. (1984)
Tinston (1994)
4
5
6
7
8
9
10
a Experimental/observational NOAEL is underlined, LOAEL is double-underlined.
b Questionable findings because litter was not used as the unit of measure in analysis.
0 LOAEL for changes in liver weight.
FEP = Flash-evoked potential
GD = Gestational day
SEP = Somatosensory-evoked potential
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-110 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Animal Neurotoxicity Inhalation Studies
Duration vs. LOAEL
mnnn
IUUUU
i/r
^ mnn
3 IUUU
.c
cf
o ^nn
^= IUU
ro
3
Q
"~^ m
0) 1U
o
_i
-i
•
1
•
i
10 100 1000
Log (LOAEL, ppm)
Figure 4-2. Summary of the relationship between LOAEL concentrations
(ppm) and treatment duration (hours).
LOAEL = Lowest-observed-adverse-effect level
motor activity. Changes in fatty acid composition were observed at somewhat higher
concentrations (320 ppm).
4.6.2.2. Oral and Intraperitoneal Studies
A study in male SD rats assessed the acute or short-term effects of tetrachloroethylene by
gavage on several screening tests (Chen et al., 2002). A single dose of 500 mg/kg to adult rats
produced changes on three different tests of pain threshold, locomotor activity, and seizure
susceptibility threshold following pentylenetetrazol infusion, whereas 50 mg/kg had statistically
significant effects on seizure threshold only. In the short-term study, young 45-50 gram rats
were dosed 5 days/week for 8 weeks with 5 or 50 mg/kg. Behavioral testing began 3 days after
the last dose. Locomotion was affected only at the high dose, whereas both doses produced
effects on the other four endpoints. The 8-week exposure resulted in retarded weight gain in
both treated groups, which was about 10% at the end of the dosing period.
The interpretation of these results is problematic. The tests were all observational in
nature, requiring scoring by the observer. The study by Chen et al. (2002) does not state whether
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-111 DRAFT-DO NOT CITE OR QUOTE
-------
1 the observer(s) was blind to the treatment group of the animals, a condition that is essential for
2 such tests to be valid. In fact, because there were differences in weight between control and
3 treated rats, it would probably be easy to distinguish treated from control animals simply by
4 looking at them. Further, the paper does not state whether all animals were tested by the same
5 person for each task or, if not, whether there was any indication of inter-observer correlation.
6 The potential effect of the difference in weight between the control and the treated groups on
7 these measures is also unknown. Given that the difference between the control and the treated
8 groups in response latency to painful stimuli is tenths or hundredths of a second with no dose
9 response, these issues are of serious concern.
10 Various behavioral endpoints were assessed in 8-week-old ICR male mice at the
11 beginning of an experiment by Umezu et al. (1997). Righting reflex was affected after single-
12 dose i.p. administration of tetrachloroethylene at 4,000 but not at 2,000 mg/kg or less, and ability
13 to balance on a wooden rod was decreased at 2,000 but not at 1,000 mg/kg or less. Response rate
14 on a fixed-ratio 20 (FR20) schedule—which requires 20 responses for each reinforcement—was
15 affected at 2,000 but not at 1,000 mg/kg or less 30 minutes after administration. In a procedure
16 in which a thirsty mouse was shocked every 20th lick of a water spout, mice dosed with 500
17 mg/kg but not with higher or lower doses received an increased number of shocks. In an
18 FR20-FR20 punishment schedule, responding in the punishment condition was increased at
19 1,000 but not at 500 mg/kg or less. A puzzling aspect of the study is the mention in the methods
20 section of "breeding animals," with no further explanation. If the investigators bred their own
21 mice, there is no indication of how pups were assigned to treatment groups.
22 Moser et al. (1995) examined the effects of a number of potentially neurotoxic agents,
23 including tetrachloroethylene, on a neurotoxicity screening battery in adult female F344 rats
24 following either a single gavage dose (acute exposure) or repeated gavage doses over 14 days
25 (subacute exposure). For the acute study, subjects were tested 4 and 24 hrs following exposure.
26 After acute exposure, a LOAEL of 150 mg/kg was identified for increased reactivity to being
27 handled 4 hrs after dosing, with increased lacrimation, decreased motor activity, abnormal gate,
28 decreased response to an auditory stimulus,
29 decreased righting ability, and increased landing foot-splay at higher doses at 4 and/or 24 hrs
30 post-dosing. A NOAEL was not identified. In the subacute study, no endpoint was significantly
31 different from those of controls at doses of 50-1,500 mg/kg. This presumably represents
32 behavioral adaptation following repeated exposure to tetrachloroethylene.
33 Locomotor activity was monitored in NMRI mice gavaged with 5 or 320 mg/kg
34 tetrachloroethylene for 7 days beginning at 10 days of age (Fredriksson et al., 1993). Twelve
35 male pups from three or four litters were assigned to each treatment group. This study design
36 does not conform to traditional developmental toxicity testing guidelines. Locomotion, rearing,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-112 DRAFT-DO NOT CITE OR QUOTE
-------
1 and total activity (vibration of the cage) were measured for 60 minutes at 17 and 60 days of age.
2 A stastically significant increase in locomotor activity of treated mice in both dose groups was
3 observed, and rearing behavior decreased as compared with controls for all three measures at 60
4 days of age but not at 17 days of age in which testing followed shortly after the last dose.
5 The persistent effects of subacute developmental exposures in this study raises some
6 concerns. Some caution in interpreting the results of the effect of tetrachloroethylene exposure is
7 warranted for two reasons: (1) the results at 320 mg/kg were no different than at 5 mg/kg,
8 indicating no clear dose-response relationship between exposure and this effect, and (2) litter
9 mates were used as independent observations in the statistical analysis. This procedure can
10 increase the apparent a and result in an erroneous statistical result. For example, Holson and
11 Pearce (1992) demonstrated that for body weight, using three or four littermates as independent
12 observations, as in the above study, resulted in the nominal a increasing from 0.05 to a range of
13 0.23 to 0.38. Similar litter effects have been demonstrated for behavioral data (Buelke-Sam et al.,
14 1985).
15 Fredriksson et al. completed a study that parametrically compared the effects of postnatal
16 dosing and resulting alternations in motor activity using both litter as the unit of measure and
17 their own within-litter randomization (Ericksson et al., 2005). Their results were similar in both
18 the magnitude of effect across dose groups and in the variability within each dose group for both
19 experimental designs. The authors' key assertion for using this randomization within a small
20 number of litters rather than the traditional litter as the unit of measure is that it reduces the
21 overall number of animals needed to be generated to statistically determine an effect of chemical
22 exposure.
23 Locomotor activity was assessed in 6-week-old male Wister rats following i.p. doses of
24 100, 500, or 1,000 mg/kg tetrachloroethylene for 3 consecutive days, with activity being
25 monitored for at least 1 week following cessation of administration (Motohashi et al., 1993).
26 Animals were monitored 24 hrs/day, and locomotor activity (measured as change in electrical
27 capacitance of a circuit beneath the floor of the cage) was analyzed by time-series analysis and
28 spectral analysis. All doses of tetrachloroethylene changed circadian rhythm in a dose-
29 dependent manner, with the increased activity at the start of the dark period delayed by
30 tetrachloroethylene exposure. Recovery took 3-5 days after cessation of exposure.
31 Operant performance on a fixed-ratio 40 schedule of reinforcement was assessed in adult
32 male SD rats gavaged with 160 or 480 mg/kg tetrachloroethylene immediately before testing
33 (Warren et al., 1996). The lower dose produced no effect on response rate over the 90-minute
34 session, whereas the higher dose produced a transient rate decrease in three of six animals (with
35 recovery after 20 to 40 minutes) and induced a complete cessation of response in two of the six
36 animals. Tetrachloroethylene concentrations increased rapidly after administration in blood,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-113 DRAFT-DO NOT CITE OR QUOTE
-------
1 brain, fat, liver, and muscle. For the duration of the 90-minute period of testing, blood
2 tetrachloroethylene levels were approximately linearly related to the administered dose, but brain
3 tetrachloroethylene levels were similar for both dose groups. This study did not evaluate the
4 persistent effects of exposure to tetrachloroethylene on cognitive performance.
5 Table 4-7 presents a summary of the oral neurotoxicity animal studies. For the six oral
6 neurotoxicity studies in rodents reviewed here, only one (Fredriksson et al., 1993) describes
7 effects lasting more than 1 week. In that study the effect (increased motor activity) was the same
8 at 5 and 320 mg/kg, and the results do not represent a clear dose-response relationship across two
9 orders of magnitude of administered doses. The lowest LOAEL occurring in the four remaining
10 studies is 100 mg/kg for delayed onset of circadian activity in rats (Motohashi et al., 1993). This
11 LOAEL is based on an i.p.-administered dose describing transient neurological effects and is not
12 comparable to inhalation or ingestion LOAELs without pharmacokinetic modeling of an
13 appropriate dose metric. No information is available for irreversible neurological effects via the
14 oral route because no studies have evaluated the potential for neurotoxicity following chronic
15 oral exposure.
16
17 4.6.3. Summary of Neurotoxic Effects in Humans and Animals
18 Taken together, the animal and epidemiologic evidence is supportive of an association
19 between neurobehavioral deficits and tetrachloroethylene exposure. The pattern of effects on the
20 visual system in humans may be consistent with decrements in visually mediated dysfunction, as
21 suggested by Echeverria et al. (1995). The test for VCS in humans is sensitive to neurological
22 dysfunction associated with many diseases affecting the nervous system (NYS DOH, 2000).
23 Moreover, VCS deficits as well as color discrimination deficits are commonly present prior to
24 detectable pathology in the retina or optic nerve head, making this one of the earliest signs of
25 disease (Regan, 1989). Additionally, other organic solvents, as well as alcohol, induce effects on
26 memory and color vision (Altmann et al., 1995; Mergler et al., 1991; Hudnell et al., 1996a, b).
27 The consistency of these observations suggests construct validity for organic solvents as a class
28 because of their effects on visually mediated function. Hence, these observations, by analogy,
29 add support to an inference of tetrachloroethylene-induced neurobehavioral effects.
30 Studies of occupational (Seeber, 1989; Echeverria 1994, 1995) and residential (Altmann
31 et al., 1995) exposures indicate that cognitive performance in humans exposed to
32 tetrachloroethylene is affected with effects on choice reaction times, visual-spatial information
33 processing, and other measures of cognitive performance.
34 The three epidemiological studies on dry cleaners chronically exposed to
35 tetrachloroethylene showed decrements in color vision at 7 ppm (Cavalleri et al., 1994, with a
36 follow-up of these workers [Gobba et al., 1998] showing greater loss in color discrimination in
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-114 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-7. Summary of oral neurotoxicity animal studies
Subjects
SD rats,
male
SD rats,
male
ICR mice,
male
8-10/dose
F344 rats,
female
NMRI male
mice,
post-natal
exposure
12 pups/
dose
(derived
from 3
litters)
Wistar rats,
male
n/dose?
Effect
Pain threshold, pain susceptibility,
weight gain decrement
Interpretation is unclear
Operant responses stopped
immediately after 480 mg/kg dose,
then 2/3 of animals recovered by 40
minutes.
Brain tetrachloroethylene
concentrations were the same at
both doses
NOAEL/LOAEL:
Righting reflex, 2,000/4,000
Balance, 1,000/2,000
Operant responses, 1,000/2,000
Punishment, 500/1,000
Increased reactivity, decreased
motor activity, decreased righting
ability, increased landing foot
splay, abnormal gait after one dose
No effect after repeated doses
Increased locomotion and
decreased rearing at day 60 in both
dose groups
No effect immediately after
treatment
Transient delay in circadian
activity, dose-related
NQAEL/LQAEL3 (mg/kg)
5, 50 mg/kg daily for 8 weeks
Gavage single dose at 0, 160^ 480
mg/kg
Single intraperitoneal doses at 0,
500. 1.000. 2,000, 4,000 mg/kg
Single doses: 150 mg/kg is
LOAEL
Repeated dosing for 14 days: 1.500
mg/kg is NOAEL
Gavage treatment 1, 320 mg/kg
daily for postnatal days 10-16
Intraperitoneal doses 0. 100, 500,
1,000 mg/kg-day for 3 days
Author
Chen et al.
(2002)
Warren et al.
(1996)
Umezu et al.
(1997)
Moser et al.
(1995)
Fredriksson et
al. (1993)
Motohashi et al.
(1993)
4
5
6
a Experimental/observational NOAEL is underlined, LOAEL is double-underlined.
n/dose = Number of animals per dose not clearly defined
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-115 DRAFT-DO NOT CITE OR QUOTE
-------
1 those who were subsequently exposed to a higher concentration and smaller loss in those
2 exposed for lower concentrations), longer reaction times to visual stimuli at 8 ppm (Spinatonda
3 et al., 1997) and decrements in cognitive function at 12 ppm (Seeber, 1989).
4 Two studies of tetrachloroethylene exposure in residences near a dry cleaning facility
5 (Altmann et al., 1995) and a day care facility (Schreiber et al., 2002) found decrements in several
6 neurological parameters at lower exposures than did the studies of occupational exposures cited
7 above. This indicates that CNS effects can occur at a lower concentration than inferred from
8 occupational studies of dry cleaners, which have been of exposures several-fold higher than
9 those in these residential studies. LOAELs in the human studies of CNS effects ranged from 0.7
10 ppm to 41 ppm.
11 The lack of exposure-response relationships is a limitation; however, this lack may be a
12 reflection of poorer characterization of exposure in these studies. Exposure-response
13 relationships are based only on an estimate of current exposure; historical exposure to
14 tetrachloroethylene is lacking and may be more important to an analysis of exposure response.
15 Moreover, most analyses use a concentration x time [C x t]) dose metric. Another metric, such
16 as peak concentration, may be more relevant of an exposure-response relationship.
17 Alcohol by itself cannot account for the observed deficits in neurobehavioral functions,
18 because statistical analyses of the epidemiologic observations accounted for this covariant.
19 However, effects from the interaction between tetrachloroethylene exposure and alcohol
20 consumption was not well investigated in these studies. Valic et al. (1997) showed greater
21 decrements in color vision among subject with both exposures as compared with individuals with
22 solvent exposure only or with neither exposure.
23 No epidemiological studies investigating drinking water or other oral exposures to
24 tetrachloroethylene have explored the potential for neurotoxicity.
25 The research in animal models (rodents) on the effects of tetrachloroethylene on
26 functional endpoints consists almost exclusively of screening studies (functional observation
27 battery, motor activity) or effects on sensory system function, as assessed by evoked potentials.
28 Effects on motor activity and motor function have been observed with some consistency
29 following either adult or developmental exposure. Changes in VEPs were also reported
30 following acute (4-day) and subchronic (13-week) exposure. In addition, changes in brain DNA,
31 RNA, or protein levels and lipid composition were altered following inhalation, with changes
32 observed in cerebellum, hippocampus, and frontal cortex. The replication of these changes in
33 biochemical parameters and effects in brain weight in both rats and gerbils is pathognomonic.
34 Changes in neurotransmitters systems (Honma et al., 1980 a, b, Driving et al., 1986) and
35 circadian rhythm (Motohashi et al., 1993) in animal studies are consistent with neuroendocrine
36 alterations observed in humans (Ferroni et al., 1992). Operant tasks that test cognitive
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-116 DRAFT-DO NOT CITE OR QUOTE
-------
1 performance have demonstrated performance deficits in rats and mice following acute
2 tetrachloroethylene oral (Warren et al., 1996) and i.p. (Umezu et al., 1997) exposures. These
3 findings in animal studies are consistent with observed effects on cognition and memory in
4 humans. However, no studies to date have evaluated the persistent effects of tetrachloroethylene
5 exposure on cognitive performance deficits in animal models. This is a clear data need that
6 could help resolve the dose-response relationship in cognitive performance observed in both
7 human occupational and residential studies. The neurophysiological findings in animal studies,
8 albeit at high doses (800 ppm), are consistent with the physiological dysfunction observed in
9 visually mediated functions in humans. In addition, the persistent changes in neurotransmitter
10 levels, regional DNA content, and brain weight in animal studies is consistent with neurological
11 effects in humans. Therefore, effects observed in human and rodent models exhibit a reasonable
12 degree of congruence.
13 The inhalation LOAEL for neurotoxic effects in humans is 0.2-41 ppm. For animals it is
14 37-90 ppm, with no apparent correlation between the LOAEL of administered concentration and
15 duration of treatment (see Section 4.6.2.1). Information for oral effects in humans is missing,
16 and the only animal data applicable to an oral exposure are from gavage administration of
17 tetrachloroethylene. No information on long-term neurological effects in animals via the oral
18 route is available.
19
20 4.6.4. Mode of Action for Neurotoxic Effects
21 The MOA for the neurotoxic effects of tetrachloroethylene is unknown; however, at
22 present, the best surrogate for the dose metric for neurotoxicity is blood tetrachloroethylene.
23 There may be multiple mechanisms or MO As, which may differ for adult and developmental
24 exposure. The acute effects of tetrachloroethylene share much in common with those of other
25 solvents such as toluene, volatile anesthetics, and alcohols. There is emerging evidence that such
26 agents act on the ligand-gated ion channel superfamily in vitro (Shafer et al., 2005), particularly
27 on the inhibitory amino acids NMD A, nicotinic, and GABA receptors in vivo (Bale et al., 2005).
28 Volatile anesthetics and alcohol both interact with the glycine receptor (Yamakura et al., 1999;
29 Wick et al., 1998; Mihic, 1999). Affinities depend on specific subunits of the receptor and are
30 correlated with behavioral effects on tests such as loss of righting ability. Similarly, ethanol and
31 volatile anesthetics enhance GAB AA receptor function (Mihic, 1999). Chronic effects of these
32 agents may also be dependent on the GABAA system (Grobin et al., 1998).
33 Other receptors, such as the dopaminergic/N-methyl-D-aspartate (NMDA) receptor, may
34 also be involved in the mediation of the effects of these agents, as may the glutamate kainate or
35 a-amino-3-hydroxy-5-methyl-4-isoxazole-4-propionic acid (AMPA) receptors (Harris et al.,
36 1995; Cruz et al., 1998). The solvents 1,1,1-trichloromethane and trichloroethylene enhanced
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-117 DRAFT-DO NOT CITE OR QUOTE
-------
1 neurotransmitter-activated currents at a ipl GABAA and a 1 glycine receptors (Beckstead et al.,
2 2000). It seems reasonable to speculate that tetrachloroethylene would also have as part of its
3 MO A modulation of these systems, although specificity for different receptor subunits would
4 undoubtedly differ from those for other nervous system depressants. Consistent with this
5 hypothesis, glutamine levels were elevated in hippocampus following tetrachloroethylene
6 administration (Driving et al., 1986), although GABA levels and uptake were unchanged.
7 Tetrachloroethylene also affects the fatty acid composition of the brain following 30- or
8 90-day exposure and persists for at least 30 days after cessation of exposure (Kyrklund et al.,
9 1984, 1987, 1988, 1990). It has long been known that the potency of an anesthetic is
10 proportional to its lipid solubility, from which it was inferred that anesthetics act on the lipid
11 bilayer of the plasma membrane. However, this observation has not led to elucidation of the
12 mechanism of action of anesthetic agents or solvents. It is nonetheless interesting that
13 tetrachloroethylene produces changes in fatty acid composition of the brain.
14 Tetrachloroethylene effects on the nervous system are not explained simply by any direct
15 dosing studies with metabolites. In particular, the pattern of neurotoxicity observed with DCA is
16 qualitatively different. DCA has been found to produce hind-limb paralysis, altered gait, muscle
17 weakness, and pathology in the spinal cord (Moser et al., 1999). In the one study that examined
18 the effects of chronic exposures to tetrachloroethylene on the histology and function of the
19 nervous system (Mattsson et al., 1998) there was no observed neuropathology or functional
20 deficits (i.e., hind-limb paralysis, altered gait, or muscle weakness). This discrepancy between
21 tetrachloroethylene and DCA could be explained by differences in target tissue dose available
22 from metabolized tetrachloroethylene and more direct exposure to DCA.
23
24 4.7. DEVELOPMENTAL/REPRODUCTIVE STUDIES
25 4.7.1. Human Studies
26 Adverse effects on reproduction and development assessed by epidemiologic
27 investigation include effects on fecundity (defined as reproductive potential and measured by
28 time to pregnancy); effects on sperm; the risk of adverse pregnancy outcomes such as
29 spontaneous abortion, stillbirth, congenital malformation, or low birth weight; and effects on
30 postnatal development (which in this evaluation includes the occurrence of childhood cancer).
31 Several of the adverse pregnancy outcome studies evaluated exposure during a critical window,
32 the first trimester of the pregnancy. In general, the epidemiologic studies evaluating effects on
33 reproduction and the developing fetus do not present quantitative information on level of
34 exposure to tetrachloroethylene. When information is available (identified in the discussion
35 below), an assertion of exposure to tetrachloroethylene in most cases was derived from self-
36 reported information provided by study subjects in mailed questionnaires or interviews. More
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-118 DRAFT-DO NOT CITE OR QUOTE
-------
1 rarely, biological measures of exposure, such as tetrachloroethylene in blood or in urine, were
2 available for a subset of these subjects.
3 A number of studies found elevated risks of spontaneous abortions among women
4 employed as dry cleaners (Doyle et al., 1997; Windham et al., 1991; Olsen et al., 1990;
5 Lindbohm et al., 1990; Kyyronen et al., 1989; Bosco et al., 1987). Two reports of the same
6 study population do not note associations between spontaneous abortions and dry cleaning and
7 laundry employment (McDonald et al., 1986, 1987). These reports are not inconsistent with the
8 remaining body of literature due to possible bias. Approximately 25% of the women
9 hospitalized for a spontaneous abortion were not interviewed, and this may have introduced a
10 bias into the study if a decision to participate was related to solvent exposure. The studies
11 assessing spontaneous abortions among the wives of men exposed to tetrachloroethylene
12 (Taskinen et al., 1989; Eskenazi et al., 1991a) were not remarkable due to the few numbers of
13 exposed cases.
14 The study by Doyle et al. (1997) is the largest: 3,517 pregnant women who were
15 currently or previously employed in dry cleaning or laundry shops. The authors analyzed the
16 data by applying several different approaches and taking into account a number of important
17 covariates, which is a strength of this study. The findings were all suggestive of an increased
18 risk of spontaneous abortions among pregnancies reported by women who were employed as dry
19 cleaners at any time during pregnancy or three months before conception as compared with
20 unexposed pregnancies. In fact, lower 95% CIs for many of these approaches were above a
21 relative risk of 1. Adding support was the observation that risk for pregnancies in dry cleaning
22 operators was larger than risks observed for pregnancies reported by women working in jobs in
23 laundry or nonoperator dry cleaning, suggesting the presence in this study of an exposure-
24 response association.
25 Doyle et al. (1997) is considered to carry greater weight than the other studies discussed
26 below due to its use of a pregnancy as the unit of analysis. Analyses that do not adjust for
27 previous pregnancy loss, such as those presented in McDonald et al. (1987), could lead to biased
28 estimates because a previous spontaneous abortion is a risk factor for a spontaneous abortion
29 with the current pregnancy. The study design of Doyle et al. (1997) minimizes the potential for
30 this type of bias because a woman with repeated pregnancy losses may be counted in both
31 exposed and unexposed categories, depending on her exposure status at the time of the
32 pregnancy.
33 Three other studies (Olsen et al., 1990; Bosco et al., 1987; Windham et al., 1991)
34 examined the association between spontaneous abortions and occupational exposure to
35 tetrachloroethylene. Olsen et al. (1990) presented findings from a four-country Nordic study of
36 spontaneous abortion, low birth weight, and congenital anomalies and observed a relative risk of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-119 DRAFT-DO NOT CITE OR QUOTE
-------
1 2.9 (95% CI = 1.0-8.4; eight exposed cases) for all data sets between spontaneous abortion and
2 "high exposure" to tetrachloroethylene during the first trimester of pregnancy, primarily due to
3 the large risk seen among subjects from Finland (OR = 4.5, 95% CI = 1.1-18.5, six exposed
4 cases). These analyses were based on 3,279 pregnancies among women dry cleaners and laundry
5 workers linked to national registers of birth and reproductive failures. Biological monitoring
6 data were available for some of the subjects from Finland; blood tetrachloroethylene
7 concentration ranged from 0.1 |imol/L to 2.6 |imol/L for cases (n = 4) and from 0.3 |imol/L to
8 3.6 |imol/L for controls (n = 3; Kyyronen et al., 1989). Unfortunately, the number of subjects in
9 Kyyronen et al. (1989) who were also included in Olsen et al. (1990) is not known.
10 The result for Finnish workers reported by Olsen et al. (1990) is of a similar magnitude as
11 that reported for essentially the same study subjects by two other Finnish investigators
12 (Kyyronen et al., 1989; Lindbohm et al., 1990). Overall, greater weight is placed on the Olsen et
13 al. (1990) findings due to the investigators' more systematic approach for evaluating an
14 exposure-effect association.
15 Bosco et al. (1987) observed a 4-fold higher history of prior spontaneous abortions
16 among women working in dry cleaning shops than among these same women when they were
17 not employed outside their homes. These findings were based on a small number of subjects and
18 were not statistically significant. Mean urinary TCA levels among women employed in dry
19 cleaning shops was 5 |ig/L, compared to 1.4 jig/L for women employed in shops that operated
20 only as an ironing service. Windham et al. (1991) reported a statistically significant elevated risk
21 of spontaneous abortions (OR = 4.7, 95% CI = 1.1-21.1) in tetrachloroethylene-exposed women
22 in analyses that adjusted for age, race, education, prior fetal loss, smoking, and number of hours
23 worked. This analysis was based on seven women identified with tetrachloroethylene exposure,
24 of which four were identified as having exposure to trichloroethylene, for which the odds ratio
25 was also elevated (OR = 3.1, 95% CI = 0.9-10.4). Both trichloroethylene and
26 tetrachloroethylene share a number of common or like metabolites, although human metabolism
27 via the P450 oxidative pathway is more extensive for trichloroethylene than for
28 tetrachloroethylene.
29 There is more limited evidence for reduced fecundity and effects on sperm with exposure
30 to tetrachloroethylene, but it is suggested in several studies (Rachootin and Olsen, 1983;
31 Eskenazi et al., 1991a, b; Sallmen et al., 1995). Sallmen et al. (1995) observed a lower
32 probability of achieving a clinically recognized pregnancy among women employed in dry
33 cleaning shops (incidence density ratio [IDR] = 0.44, 95% CI = 0.22-0.86, 11 women) in
34 analyses that adjusted for a number of other covariates. Furthermore, tetrachloroethylene
35 exposure was associated with a decreased probability of pregnancy, although an exposure-
36 response pattern was not apparent (low exposure to tetrachloroethylene, IDR = 0.63, 95% CI =
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-120 DRAFT-DO NOT CITE OR QUOTE
-------
1 0.34-1.17, based on 13 women; high exposure to tetrachloroethylene, IDR = 0.69, 95% CI =
2 0.31-1.52, based on 7 women). Exposure was defined as frequency of tetrachloroethylene use,
3 with no attention paid to level of exposure. Hence, exposure misclassification may partially
4 explain the lack of an exposure-response relationship. Sallmen et al. (1995) examined these
5 exposures as part of a larger evaluation of general organic solvent exposure for which a
6 statistically significant association was noted (high level exposure, IDR = 0.41, 95% CI =
7 0.27-0.62).
8 Eskenazi et al. (1991a, b) reported a statistically significant reduced probability of
9 pregnancy among highly exposed individuals. Eskenazi et al. (1991a) noted a lower per-cycle
10 pregnancy rate among wives of men who received higher-level exposure to tetrachloroethylene
11 (RR = 0.94, 95% CI = 0.85-1.04) as compared with wives of men who received lower-level
12 exposure. The potential contribution of tetrachloroethylene exposure on time to conception was
13 small compared to the contribution observed from Hispanic ethnicity and smoking, which were
14 found to be stronger and statistically significant predictors of time to conception.
15 Eskenazi et al. (1991b) also found subtle spermatogenic effects among dry cleaners when
16 compared with laundry workers. These effects were characterized as a greater proportion of
17 round sperm and a lower proportion of narrow sperm. Furthermore, tetrachloroethylene level
18 was a statistically significant predictor of decreased number of narrow sperm and of increased
19 numbers of spermatids with amplitude of lateral head displacement, a measure of the unsteady
20 movement of the sperm head about its average path of motion. Mean tetrachloroethylene
21 exposure was 1.2 ppm for dry cleaners and 0.01 ppm for laundry workers. More traditional
22 measures of semen quality, such as number of sperm, concentration, volume, average percentage
23 of motile sperm or average percentage of abnormally shaped sperm, as well as the prevalence of
24 study subjects identified as azospermic or oligospermic, did not differ between these two job
25 titles. Moreover, it did not appear that the effects on semen parameters or the per-cycle
26 pregnancy rate had a large impact on fertility rates. Partners of these male dry cleaners did not
27 have fewer pregnancies as compared with a national standard (Eskenazi et al., 199la).
28 The findings from these studies are consistent with those observed by Rachootin and
29 Olsen (1983), whose study subjects were couples seeking treatment for infertility. These
30 investigators noted that employment as a dry cleaner was associated with hormonal disturbances
31 and delayed conception in analyses that took into account the woman's age, education, residence,
32 and parity. Exposure to tetrachloroethylene was inferred but not documented.
33 Few epidemiologic studies exist that evaluate other developmental toxicity endpoints
34 such as decreased birth weight, intrauterine growth restriction (IUGR; also known as small for
35 gestation age [SGA]), and congenital anomalies. Many of the analyses were included in the
36 reports of spontaneous abortions discussed above. Overall, no associations were noted in several
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-121 DRAFT-DO NOT CITE OR QUOTE
-------
1 studies that assessed maternal or paternal exposure to tetrachloroethylene and increased
2 incidences of stillbirths, congenital anomalies, or decreased birth weight (Olsen et al., 1990;
3 Kyyronen et al., 1989; Taskinen et al., 1989; Windham et al., 1991). These findings may be a
4 reflection of the small numbers of exposed cases, or they may be attributable to exposure
5 misclassification (biological marker data were available for only a few study subjects) or disease
6 misclassification (which could be introduced from the grouping of several different outcomes
7 into one category).
8 The case-control study by Windham et al. (1991) observed a strong but imprecise
9 association between IUGR and exposure to tetrachloroethylene (OR = 12.5, 95% CI not given in
10 the published paper and too few data for NCEA staff to calculate). This observation was based
11 on only one exposed case who also had exposure to trichloroethylene; an RR greater than 1 was
12 also observed for trichloroethylene exposure (OR = 4.2).
13 Studies of populations serviced by drinking water containing several contaminants,
14 including tetrachloroethylene and trichloroethylene, report elevated risks such as adverse
15 pregnancy or postnatal outcomes attributed to living in a residence receiving contaminated water
16 effects (Lagakos et al., 1986; Bove et al., 1995; ATSDR, 1998). Lagakos et al. examined the
17 relationship between several birth outcomes that were identified from questionnaires given to a
18 sample of residents from Woburn, MA. This study was part of a larger study evaluating the
19 association between childhood leukemia among residents of this town and living in a residence
20 receiving drinking water from two wells contaminated with trichloroethylene,
21 tetrachloroethylene, and chloroform. The levels of these contaminants in the wells at the time
22 they were closed were 267 ppb (|ig/L) trichloroethylene, 21 ppb (|ig/L) tetrachloroethylene, and
23 12 ppb (|ig/L) chloroform. The investigators observed statistically significant associations
24 between a residence receiving contaminated water and three outcomes: perinatal deaths since
25 1970, eye and ear anomalies, and CNS/oral cleft anomalies.
26 An analysis by Shawn and Robins (1986) of events among residents of East Woburn, the
27 location of the contaminated wells, noted a statistically significant exposure-response trend only
28 for perinatal deaths. This analysis was presented in comments by the study authors to the study
29 by Lagakos et al. (1986) and was carried out to evaluate recall bias in that study, which these
30 authors concluded did not exist.
31 A case-control study of leukemia cases among children in Woburn, MA (MA DPH,
32 1997) noted a large risk between maternal exposure (e.g., living during the first trimester of
33 pregnancy in a residence that received contaminated water) and leukemia (ORadj = 8.3, 90% CI =
34 0.7-94.7, 10 exposed case). Risks increased significantly (p < 0.05) with increasing exposure
35 (never exposed, least exposed, most exposed). This study is more fully discussed in the section
36 reviewing the epidemiologic evidence on cancer effects.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-122 DRAFT-DO NOT CITE OR QUOTE
-------
1 In a prevalence study, Bove et al. (1995) assessed the relationship between a number of
2 birth outcomes, as identified from the birth certificate or from the New Jersey Birth Defects
3 Registry, and residence in 75 towns for which monitoring data were available. Concentrations of
4 trihalomethanes, trichloroethylene, and tetrachloroethylene, along with a wide range of other
5 solvents, were identified from the monitoring data. Bove et al. (1992, 1995) observed risks
6 above 1.5 between residence in a town with >10 ppb tetrachloroethylene detected in drinking
7 water and oral cleft defects (OR = 3.5, 90% CI = 1.3-8.78, four exposed cases). No associations
8 were reported for other birth outcomes such as CNS defects, neural tube defects, low birth
9 weight, and small for gestational age and tetrachloroethylene exposure.
10 This analysis lacks information on risk factors for an individual. To address this
11 limitation, the investigators conducted a case-control study of oral cleft defects (Bove et al.,
12 1992), where findings did not support the observations in the ecological study. The association
13 between tetrachloroethylene in water (>5 ppb) and oral clefts was not elevated (ORadj = 0.4, 95%
14 CI = 0-4.3, four exposed cases) in the case-control analyses. Given the better design of the case-
15 control study and its ability to include information on individual study participants, this study
16 carries a greater weight than does the 1995 ecological study in the overall evaluation of the
17 relationship between tetrachloroethylene exposure and developmental effects.
18 An analysis by ATSDR (1998; results published by Sonnenfeld et al., 2001) examined
19 birth weight and gestational age among births of residents living in base family housing at Camp
20 Lejeune, NC. The residences received drinking water contaminated by solvents, including
21 trichloroethylene, tetrachloroethylene, and/or benzene. A large number of births (n = 6,117)
22 between 1968 and 1985 were identified from birth records and were classified as exposed to
23 tetrachloroethylene, i.e., the mother had resided at any time of pregnancy in the base housing,
24 specifically Tarawa Terrace, which had received tetrachloroethylene-contaminated drinking
25 water. Although quantitative information on exposure is limited in this study (water samples
26 were collected on only three different occasions between 1982 and 1985), it is thought that the
27 well providing water to Tarawa Terrace was contaminated with tetrachloroethylene for as long as
28 30 years (ATSDR, 1998; Sonnenfeld et al., 2001). The highest concentrations of contaminants
29 measured in tap water from Tarawa Terrace were 215 ppb (|ig/L) tetrachloroethylene, 8 ppb
30 (ng/L) trichloroethylene (single sample), and 12 ppb (|ig/L) 1,2,-dichloroethylene (single
31 sample). No information on level of tetrachloroethylene exposure was available prior to 1982.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-123 DRAFT-DO NOT CITE OR QUOTE
-------
1 The frequency of a birth of an SGA10 infant was slightly increased among women who
2 were identified as tetrachloroethylene exposed (10.2%) as compared with those who were not
3 (9%; OR = 1.2, 90% CI =1.0-1.3, 622 exposed births). The investigators also observed a
4 smaller mean birth weight of exposed infants as compared with infants of mothers who lived in
5 unexposed housing—a difference of 24 g, which was not considered to be of biological
6 significance. Two susceptible groups were identified from this analysis: mothers 35 years or
7 older and mothers with previous fetal deaths. For older mothers, the adjusted difference in mean
8 birth weight between tetrachloroethylene-exposed and unexposed births was 205 g (90% CI =
9 78-333), with a risk (OR) of 4 (95% CI = 1.6-10.2, 11 exposed births) between exposure and
10 birth of an SGA infant. For mothers with prior fetal losses, exposure to tetrachloroethylene in
11 drinking water was associated with a 60% higher risk for an infant that was identified as SGA
12 (95% CI = 1.2-2.1, 147 exposed births).
13 Inferences regarding developmental and reproductive effects from tetrachloroethylene are
14 limited due to small risks of low precision, the lack of a direct measure of tetrachloroethylene in
15 many studies, the small numbers of exposed cases, and possible biases such as recall or
16 misclassification bias. The epidemiologic evidence is strongest for spontaneous abortions among
17 exposed women. A number of studies have reported an elevated risk of spontaneous abortions
18 and maternal exposure to tetrachloroethylene, primarily exposure received through employment
19 as a dry cleaner (Doyle et al., 1997; Windham et al., 1991; Olsen et al., 1990; Lindbohm et al.,
20 1990; Kyyronen et al., 1989; Bosco et al., 1987). The epidemiologic evidence for infertility is
21 further suggestive of an association with tetrachloroethylene exposure (Rachootin and Olsen,
22 1983; Eskenazi et al., 1991a, b; Sallmen et al., 1995). Any conclusions of effects on birth weight,
23 IUGR (SGA), or congenital anomalies and tetrachloroethylene exposure cannot be drawn from
24 the available occupational studies, although drinking water studies of exposures to multiple
25 chemicals, including tetrachloroethylene, provide some limited evidence. There is very little
26 information about what exposure pattern (concentration and duration) is associated with these
27 effects. Table 4-8 summarizes these studies.
28
29 4.7.2. Animal Studies
30 Evaluation of the developmental and reproductive effects of tetrachloroethylene exposure
31 in animal models is based on several studies of in utero exposures to maternal animals during
32 specific periods of pregnancy. These studies include embryo explant in rats, multigeneration
10 SGA is measured by comparing birth weight at specific gestational ages with a gestational-age-specific-
birth-weight distribution. Live births of infants weights less than the 10th percentile are classified as SGA. Three
standards were examined by Sonnenfeld et al. (2001), and the standard of Williams et al. (1982) provided the best fit
to the data.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-124 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-8. Developmental/reproductive studies in humans
Subjects
3,517 pregnancies in women
employees in dry cleaners
and laundries
3,279 pregnancies among
dry cleaners in four Nordic
countries. Occupational
records were linked to
national registries of birth
and reproductive failures.
53 women dry cleaners and
ironers whose pregnancy
outcomes were compared
with outcomes during time
as a housewife.
Case-control study of
spontaneous abortions
(1,361 cases, controls with
live births)
20 women in dry cleaning
shops, 90 unexposed control
women
Wives of 17 men exposed to
tetrachloroethylene dry
cleaners, wives of 32 male
laundry workers
34 male dry cleaners, 48
unexposed male laundry
workers
Case-control study of
infertility and delayed
conception: 1,069 infertile
case couples; 4,305 fertile
control couples
Effect
Spontaneous abortions
Spontaneous abortions
elevated among "high-
exposure" workers in all
countries, primarily
attributed to excesses in
Finland
Spontaneous abortions
fourfold higher than among
nonexposed housewives but
not statistically significantly
different
Spontaneous abortion excess
in 7 women, 4 of whom
were also exposed to
trichloroethylene
Lower probability of
clinically recognized
pregnancy
Lower per-cycle pregnancy
rate but no decrease in
fertility rate
Subtle sperm changes but no
change in clinical sperm
quality criteria
Delayed conception and
hormonal disturbances:
OR = 3 for exposure to dry-
cleaning chemicals (95% CI
= 1.2-7.4)
Exposure
Elevated risks by several
definitions of "exposed."
No tetrachloroethylene
measurements
Blood levels of TCA not
elevated in some of these
workers
TCA in dry cleaners was
5 (ig/L; TCA in ironers was
1.4jig/L
No dose-response with
tetrachloroethylene exposure
High vs. low
tetrachloroethylene exposure
categories
Mean exposure was 1 .2 ppm
Tetrachloroethylene
exposure inferred
Authors
Doyle et al.
(1997)
Olsen et al.
(1990)
Bosco et al.
(1987)
Windham et
al. (1991)
Sallmen et al.
(1995)
Eskenazi et
al. (1991a)
Eskenazi et
al. (1991b)
Rachootin,
and Olsen
(1983)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-125 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-8. Developmental/reproductive studies in humans (continued)
Subjects
4,396 pregnancies among
residents of Woburn, MA.
80,938 live births and 594
fetal deaths among residents
in 75 New Jersey towns
Case-control study of
selected birth outcomes in
New Jersey; 49 cases of oral
cleft defects, 138 controls
1 1,798 births among women
living in United States
Marine Corp base housing
Effect
Statistically significant
positive association between
access to contaminated
water and
(a) perinatal deaths since
1970 and
(b) eye/ear birth anomalies
No association between
water access and
(a) incidence of
spontaneous abortion
(b) low birth rate
(c) perinatal deaths before
1970
(d) musculoskeletal birth
anomalies
(e) cardiovascular birth
anomalies
Oral cleft defects, OR = 3.5,
95% CI= 1.3-8.8) based on
four exposed cases
The association between
tetrachloroethylene in water
(>5 ppb) and oral cleft
defects was not elevated.
Excess of age (SGA) births
in women >35 years of age
among mothers with prior
fetal losses
Exposure
Tetrachloroethylene: 21
Mg/L
Trichloroethylene: 267 (ig/L
Chloroform: 12 (ig/L
>10 (ig/L in drinking water
Tetrachloroethylene: <215
Mg/L
Authors
Lagakos et al.
(1986)
Bove et al.
(1995)
Bove et al.
(1992)
ATSDR,
(1998), Son-
nenfeld et al.
(2001)
3
4
5
6
7
8
9
10
11
12
13
reproduction in rats, and an in vitro oocyte fertilization assay following in vivo exposure of adult
female rats.
In an inhalation developmental toxicity study (Schwetz et al., 1975), Sprague-Dawley
rats and Swiss-Webster mice were exposed to airborne tetrachloroethylene at 300 ppm 7 hrs/day
on days 6-15 of gestation. Following laparohysterectomy on gestation days 21 or 18 (for rats
and mice, respectively), fetuses were weighed and measured, examined for external
abnormalities, and processed for the evaluation of either soft tissue or skeletal abnormalities.
Three other organic solvents were also tested with the same protocol; the concentration of all
agents was chosen to be approximately twice their threshold limit values. Although the study
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-126 DRAFT-DO NOT CITE OR QUOTE
-------
1 authors concluded that there was no significant maternal, fetal, or embryo toxicity for any of the
2 solvents tested, the maternal and fetal data demonstrated a number of statistically significant
3 differences from control values following gestational exposures to tetrachloroethylene in rats and
4 mice. In the rats, exposures to tetrachloroethylene produced slight but statistically significant
5 maternal toxicity (4-5% reductions in mean maternal body weight gains) and embryotoxicity
6 (increased resorptions; 9% in treated vs. 4% in controls). In the mice, maternal toxicity consisted
7 of a significant 21% increase in mean relative liver weight as compared with controls. The mean
8 fetal weight in mice was significantly (9%) less than in the concurrent control, and the percent of
9 litters with delayed ossification of the skull bones, delayed ossification of the sternebra, and
10 subcutaneous edema were significantly increased.
11 Szakmary et al. (1997) exposed CFY rats to tetrachloroethylene via inhalation throughout
12 gestation (i.e., gestation days 1-20) for 8 hrs/day at concentrations of 1,500, 4,500, or
13 8,500 mg/m3. In the same study, the study authors exposed C57B1 mice via inhalation on
14 gestation days 7-15 (i.e., during the period of organogenesis) to a concentration of 1,500 mg/m3
15 and New Zealand white rabbits during organogenesis (gestation days 7-20) to a concentration of
16 4,500 mg/m3. Maternal animals were killed approximately 1 day prior to expected delivery; a
17 gross necropsy was conducted, organ weights were recorded, blood was taken by aorta puncture
18 for hematology and clinical chemistry evaluations, ovarian corpora lutea were counted, and
19 uterine contents were examined (number and position of living, dead, or resorbed fetuses; and
20 fetal and placental observations and weights). The numbers of litters available for evaluation
21 were as follows: 20 control and 21 or 22 per treated group in the rat, 77 control and 10 treated in
22 the mice, and 10 control and 16 treated in the rabbit. One-half of the fetuses from each litter
23 were evaluated for visceral abnormalities, and the other half were evaluated for skeletal
24 development. The study authors reported that the organs of five dams and five embryos from
25 each group were also evaluated by routine histological methods. To evaluate the concentration
26 of tetrachloroethylene in maternal and fetal blood and in amniotic fluid, another subset of rats
27 (number not specified) was studied. (For the 1,500 and 8,500 mg/m3 exposure levels, maternal
28 blood concentrations of tetrachloroethylene were 17.8+8.9 and 86.2+13.0 uL/mL, respectively.
29 Concentrations in the fetal blood were 66% and 30% of maternal blood concentrations, and
30 amniotic fluid concentrations were 33% and 20% of maternal blood concentrations.) In the rat,
31 at 4,500 and 8,500 mg/m3, maternal body weight gain during gestation was significantly
32 decreased (37 and 40%, respectively), relative maternal liver mass was significantly increased
33 (10 and 6%, respectively), and serum aspartate amino transferase activity was increased (data not
34 provided) as compared to controls. Percent pre-implantation loss was significantly increased
35 from controls by 133 and 117% at these exposure levels, while percent post-implantation loss
36 was increased non-significantly from controls by 80% in each group. Also, at 4,500 and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-127 DRAFT-DO NOT CITE OR QUOTE
-------
1 8,500 mg/m3, fetal weight was significantly decreased in 98.5 and 100% of all fetuses, the
2 number of fetuses with skeletal retardation was significantly increased in 98.5% and 100% of
3 fetuses, and the percent of fetuses with malformations was both significantly increased to 6.4%
4 and 15.7% as compared to the control incidence of 2.0%. Although the study authors judged the
5 1,500 mg/m3 exposure level to be the NOAEL for the rat study, it is noted that there were
6 concentration-dependent non-significant decreases in maternal body weight gain (13% lower
7 than control), and increases in pre- and post-implantation loss (49% and 38% greater than control,
8 respectively). The percent of weight-retarded fetuses increased to 3.4 times the control incidence,
9 and the incidences of fetuses with skeletal retardation (48% increased) or total malformations
10 increased by 2.3 times the control incidence observed at the low-exposure level of 1,500 mg/m3.
11 Therefore, these findings are judged to be adverse consequences of treatment. The attribution of
12 these findings to treatment, and the designation of 1,500 mg/m3 as the study LOAEL is
13 consistent with the adverse developmental findings of Schwetz et al. (1975). In mice
14 (1,500 mg/m3) and rabbits (4,500 mg/m3), relative liver mass was significantly increased;
15 decreased maternal body weight gain was also observed in the rabbits. In the mice, a
16 significantly increased number of fetuses with visceral malformations (details not specified) was
17 observed, while in the rabbits, two (of 16) does aborted, total resorption of four litters was
18 reported, and the percent of post-implantation loss was significantly increased. The percent of
19 rabbit fetuses with malformations (details not provided in the report) was also increased,
20 although not significantly.
21 An additional cohort of rats from the Szakmary et al. study (15 litters/group at exposure
22 levels of 1,500 or 4,500 mg/m3 tetrachloroethylene) was allowed to deliver, and the offspring
23 (standardized to 8 pups/litter) were maintained on study to postnatal day 100. It was not clearly
24 specified in the report whether the daily inhalation exposures continued throughout the postnatal
25 period. Pre-weaning observations included weekly body weights, developmental landmarks
26 (pinna detachment, incisor eruption, and eye opening), and functional assessments (forward
27 movement, surface righting reflex, grasping ability, swimming ontogeny, rotating activity,
28 auditory startle reflex, and examination of stereoscopic vision). After weaning, exploratory
29 activity in an open field, motor activity in an activity wheel, and development of muscle strength
30 were assessed. The study authors reported that adverse findings included a decreased survival
31 index (details not provided), a minimal decrease of exploratory activity and muscular strength in
32 treated offspring (presumably at both exposure levels) that normalized by postnatal day 51, and
33 significantly increased motor activity on postnatal day 100 of females exposed to 4,500 mg/m3 of
34 tetrachloroethylene.
35 Nelson et al. (1980) investigated developmental neurotoxicity in Sprague-Dawley rats by
36 exposing pregnant dams to tetrachloroethylene at concentrations of 100 ppm and 900 ppm during
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-128 DRAFT-DO NOT CITE OR QUOTE
-------
1 either early pregnancy (gestation days 7 to 13) or late pregnancy (gestation days 14 to 20). They
2 performed morphological examinations of the fetuses (gross, visceral, and skeletal) and
3 behavioral testing and neurochemical analyses of the offspring.
4 There were no alterations in any of the measured parameters in the 100 ppm groups. At
5 900 ppm there were no skeletal abnormalities, but the weight gain of the offspring as compared
6 with controls was depressed about 20% at postnatal weeks 3-5. Developmental delay was
7 observed in both the groups exposed in early and in late pregnancy. Offspring of the early
8 pregnancy-exposed group performed poorly on an ascent test and on a rotorod test, whereas
9 those in the late pregnancy group underperformed on the ascent test at only postnatal day 14.
10 However, later in development (days 21 and 25) their performance was higher than that of the
11 controls on the rotorod test. These pups were markedly more active in the open field test at days
12 31 and 32. Activity wheel testing on days 32 and 33 did not reveal statistically significant
13 changes. Avoidance conditioning on day 34 and operant conditioning on days 40-46 failed to
14 suggest effects. Neurochemical analyses of whole brain (minus cerebellum) tissue in 21-day-old
15 offspring revealed significant reductions in acetylcholine levels at both exposure periods,
16 whereas dopamine levels were reduced among those exposed on gestation days 7-13.
17 All of the described effects in the 900 ppm group were statistically significant as
18 compared with controls. Unfortunately, none of the statistics for the 100 ppm treatments were
19 presented. The authors observed that more behavioral changes occurred in offspring exposed
20 during late pregnancy than in those exposed during early pregnancy.
21 Beliles et al. (1980) described an experiment in which male rats and mice were exposed
22 via inhalation to tetrachloroethylene concentrations of 100 and 500 ppm for 7 hrs/day for 5 days.
23 Sperm head abnormalities and abnormal sperm were evaluated at 1, 4, and 10 weeks after the last
24 dose. Rats were unaffected. At 4 weeks but not at 1 or 10 weeks after exposure there was a
25 significant increase (p < 0.05) in the percentage of mice with abnormal sperm heads (19.7%) for
26 animals inhaling 500 ppm. For the 100 ppm and control groups the percentages were 10.3% and
27 6% (not statistically significant at the/? < 0.05 level), respectively. A positive control group
28 administered triethylene melanime was adversely affected (11.1%). The authors suggested that
29 the temporal appearance of the abnormal sperm heads indicated that the spermatocyte and/or
30 spermatogonia were the stages most sensitive to the effects of inhaled tetrachloroethylene. In
31 this study the NOAEL was 100 ppm and the LOAEL was 500 ppm.
32 Hardin et al. (1981; see also Beliles et al., 1980) found no developmental toxicity among
33 the fetuses from Sprague-Dawley rats or New Zealand White rabbits inhaling 500 ppm of
34 tetrachloroethylene for 7 hrs/day, 5 days/week. Tetrachloroethylene was administered with and
35 without three-week pregestation exposures and with both full-term and terminal two-thirds-term
36 exposure.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-129 DRAFT-DO NOT CITE OR QUOTE
-------
1 In a developmental toxicity study, Carney et al. (2006) investigated the effects of whole-
2 body inhalation exposures to pregnant Sprague-Dawley rats at nominal concentrations of 0, 75,
3 250, or 600 ppm (actual chamber concentrations of 0, 65, 249, or 600 ppm) tetrachloroethylene
4 for 6 hrs/day, 7 days/week on gestation days (GD) 6-19. This study was conducted under Good
5 Laboratory Practice (GLP) regulations according to current EPA and OECD regulatory testing
6 guidelines. Maternal toxicity consisted of slight but statistically significant decreases in body
7 weight gain during the first 3 days of exposure to 600 ppm, establishing a no-adverse-effect
8 concentration of 249 ppm for dams. A slight, statistically significant decrease in gravid uterine
9 weight at 600 ppm correlated with significant reductions in mean fetal body weight (9.4%) and
10 placental weight (15.8%) at GD 20 cesarean section. At >249 ppm, mean fetal and placental
11 weights were significantly decreased by 4.3% and 12.3% from control, respectively. A
12 significant increase in the incidence of incomplete ossification of the thoracic vertebral centra at
13 this exposure level was consistent with fetal growth retardation. No treatment-related alterations
14 in fetal growth or development were noted at 65 ppm. Therefore the LOAEL for this study is
15 249 ppm.
16 Saillenfait et al. (1995), using a rat whole embryo (day 10) culture system, found
17 tetrachloroethylene-induced embryo toxicity, including mortality, malformations, and delayed
18 growth and differentiation. No adverse effect was produced at the 2.5 mM concentration, but
19 concentration-related trends of increasing toxicity occurred from 3.5 mM through 15 mM.
20 Statistical tests for a concentration-related trend were not reported. The investigators found that
21 trichloroethylene produced similar effects, with potency somewhat less than that of
22 tetrachloroethylene. They also found that TCA and DCA caused a variety of abnormalities in
23 this culture system.
24 In a developmental toxicity screening study, timed-pregnant F344 rats were treated by
25 gavage with tetrachloroethylene doses of 900 or 1,200 mg/kg-day in corn oil vehicle on gestation
26 days 6-19 (Narotsky and Kavlock, 1995). There were 17 dams in each of the
27 tetrachloroethylene-treated groups and 21 in the control groups. The dams were allowed to
28 deliver, and their litters were examined on postnatal days 1, 3, and 6. At 1,200 mg/kg no live
29 pups were delivered on day 22 of gestation. At 900 mg/kg-day there was maternal ataxia, and
30 weight gain was markedly less than in the controls. The number of pups per litter was reduced
31 (p < 0.01) as compared with the controls at day 22 of gestation. On postnatal day 6 the number
32 of pups per litter was reduced (p < 0.001) as compared with the controls. The investigators noted
33 that full-litter resorptions were not observed with other chemicals they tested in the presence of
34 maternal toxicity. An increase in micro/anophthalmia was found in the offspring. There was no
35 evaluation for skeletal changes, and not all available pups were examined for soft tissue changes.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-130 DRAFT-DO NOT CITE OR QUOTE
-------
1 Because of the high dose levels and limited evaluation of the soft tissue changes, the
2 malformations described are of limited impact.
3 A multigeneration study of the effects on rats of exposure to airborne concentrations of
4 tetrachloroethylene was performed by Tinston (1994). Although this study has not been
5 published, it was submitted to EPA (Office of Prevention, Pesticides, and Toxic Substances and
6 to the IRIS Office as a result of the data call-in for the IRIS update). It was conducted under
7 good laboratory practice standards and received frequent quality assurance audits. In this study,
8 weanling male and female (Alpk:APfSD) rats (FO) were exposed to airborne tetrachloroethylene
9 concentrations of 0, 100, 300, or 1,000 ppm 6 hrs/day, 5 days/week, for 11 weeks prior to mating
10 and then for 6 hrs/day during mating and through day 20 of gestation. There were no exposures
11 from day 21 of gestation through day 5 postpartum. One litter was produced in the first
12 generation (F1A). The first-generation dams and their litters were exposed to
13 tetrachloroethylene from postnatal day 6 through 29, at which time parental animals for the
14 second generation were selected. The second-generation parents (Fl) were then exposed 5
15 days/week during the 11-week pre-mating period. In the second generation, three litters were
16 produced: F2A, F2B, and F2C. The F2A dams and litters were exposed from days 6 to 29
17 (control and 100 ppm) or days 7 to 29 (300 ppm). The 1,000 ppm exposure for the Fl dams
18 stopped after the F2A littering.
19 F2B litters were generated by mating the Fl parental males and females in the control,
20 300, and 1,000 ppm groups; the dams and F2B litters were not exposed to tetrachloroethylene
21 during lactation. An F2C litter was produced by mating Fl males exposed to 1,000 ppm with
22 unexposed females. These females and the F2C litters were killed on postnatal day 5 and
23 discarded without further examination. Overall, the FO males were exposed for 19 weeks and the
24 Fl males were exposed up to 35 weeks. Postmortem evaluation in adults and selected weanlings
25 included organ weight and histopathology examination of liver, kidney, and reproductive organs;
26 sperm measures were not assessed.
27 Table 4-9 summarizes the results of the Tinston study. Signs of CNS depression
28 (decreased activity and reduced response to sound) were observed at 1,000 ppm for the first 2
29 weeks in both adult generations and again when the exposure was resumed on day 6 postpartum
30 in the F1 generation (adults and pups). Other signs of overt tetrachloroethylene toxicity in the
31 adults included irregular breathing and piloerection at both 1,000 and 300 ppm and salivation
32 and tip-toe gait (in one Fl female) at 1,000 ppm. These changes stopped with the cessation of
33 exposure or within approximately 30 minutes thereafter.
34 There were a number of changes relative to controls that were of minor biological
35 significance. One change was transient statistically significant reductions of mean body weights
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-131 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-9. Exposure concentrations (ppm) at which effects occurred in a
two-generation study
Parameter
Clinical signs
(piloerection, irregular
breathing)
Behavioral effects
(decreased activity;
reduced response to sound)
Transient decreased body
weight gains
Decreased mean testes
weight
Increased liver and kidney
weights
Renal histopathology
Decreased pups born alive
(%)
Decreased mean % pup
survival days 1-5
Decreased mean % pup
survival days 5-22
Decreased mean male pup
weight day 1
Decreased mean female
pup weight day 1
Decreased mean male pup
weight day 29
Decreased mean female
pup weight day 29
Generation
FO
1,000, 300
1,000
1,000, 300
1,000
1,000
F1A
1,000
1,000
l,000b
1,000
l,000b
1,000C
1,000C
l,000b,
300b, 100
b,d
l,000b,
300b, 100d
Fl
1,000, 300
1,000
1,000, 300
1,000
1,000
1,000
F2A
1,000C
1,000C
l,000b
1,000C
l,000b
F2B
1,000C
1,000C
1,000C
F2Ca
1,000C
NA
NA
NA
4
5
6
7
8
9
10
11
12
a Not exposed after delivery.
V<0.05.
d trend;? < 0.05.
NA = Not applicable (pups terminated on day 5 postnatal)
Source: Adapted from Tinston (1994).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-132 DRAFT-DO NOT CITE OR QUOTE
-------
1 (originating from treated males and nontreated females) suggests the absence of male-mediated
2 effects on reproductive outcome. Nevertheless, the alterations in testes weight cannot be
3 discounted as a possible effect of treatment.
4 In females, dystocia was noted in one FO dam at 100 ppm, two Fl dams at 300 ppm, and
5 a total of four dams (two each FO and Fl) at 1,000 ppm; these dams were terminated without
6 completion of delivery. From the data for surviving dams and litters, it can be assumed that the
7 difficulties in parturition were not associated with or attributable to alterations in mean gestation
8 length or increased mean pup or litter weights. In fact, mean pup body weights showed a
9 statistically significant decrease throughout the lactation period at 300 and 1,000 ppm for Fl A
10 litters and in early lactation for F2A and F2B litters. Additionally, mean Fl A male pup body
11 weight was significantly decreased (5% less than controls; p < 0.05) at 100 ppm on postnatal day
12 29. These postnatal day 29 mean body weight deficits in all treated groups were observed in the
13 animals selected as parents of the second generation, but by the second week of the Fl pre-
14 mating period, mean body weights were similar to those of controls for both 100 and 300 ppm
15 animals.
16 Mean litter size was decreased at 1,000 ppm for F2A and F2B litters. Statistically
17 significant decreases in the number of live pups on postnatal day 1 (25% and 37% lower than
18 controls for F2A and F2B, respectively) are suggestive of either an adverse effect on fertilization
19 or on in utero survival. Early postnatal survival (i.e., on postnatal day 1 and between postnatal
20 days 1 and 5) was also compromised in F2A and F2B pups at 1,000 ppm, with mean litter sizes
21 decreasing to 48% and 53% of those of controls, respectively. The number of dead pups and
22 litters with dead pups was also increased, although not significantly, at 300 ppm for F2A litters.
23 Clinical observations data for 1,000 ppm litters reported an increased incidence of F2A and F2B
24 pups that were found dead, were killed in extremis or were missing and presumed dead. The
25 apparent increase in adverse survival findings at 300 and 1,000 ppm in the second generation as
26 compared with the first generation could not be definitively attributed to any particular aspect of
27 study design or conduct (e.g., differences in the duration of treatment), although it is noted that,
28 unlike the second generation (Fl) parental animals, the first generation (FO) rats were not
29 exposed to tetrachloroethylene during preconception and in utero development.
30 A deficiency of the Tinston study is that the pregnant rats were not exposed from
31 gestation day 21 through lactation day 6 or 7, and the exposure at the 1,000 ppm treatment level
32 stopped for the Fl dams at the littering of the F2B pups. The F2B pups were not exposed
33 postnatally. A summary of the doses at which the effects were observed in the study is presented
34 in Table 4-9.
35 In a study designed to examine the fertilizability of rat oocytes, female rats were exposed
36 to inhaled tetrachloroethylene at 12,000 mg/m3 (2 hrs/day, 5 days/week) for 2 weeks (Berger and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-133 DRAFT-DO NOT CITE OR QUOTE
-------
1 Horner, 2003). The percentage of extracted oocytes that were fertilized in vitro was reduced for
2 tetrachloroethylene-treated females as compared with controls.
O
4 4.7.2.1. Summary of Animal Studies
5 Table 4-10 summarizes the findings of the animal studies described in this section. The
6 data show that inhalation of tetrachloroethylene by pregnant mice and rats during various periods
7 of gestation resulted in fetal growth retardation and mortality in several studies and in delayed
8 behavioral changes in the three studies that measured these effects (SzakmacZy et al 1997;
9 Nelson et al. 1980; Tinston, 1994). Single studies have shown changes in brain acetyl choline
10 and dopamine, altered brain fatty acid composition, and altered sperm morphology. These
11 effects occurred at doses higher than 300 to 1,000 ppm in various studies.
12 The overall NOAEL for the animal developmental/reproductive inhalation studies is 100
13 ppm, based on Tinston (1994). The overall LOAEL is 300 ppm, based on Tinston (1994) and
14 Schwetz et al. (1975), in which increased mortality and decreased body weight of the offspring
15 were observed. All of these studies used the inhalation route of exposure.
16
17 4.7.3. Summary of Human and Animal Developmental/Reproductive Studies
18 Inferences regarding developmental and reproductive effects from tetrachloroethylene
19 exposure in humans are limited due to small risks of low precision, the lack of a direct measure
20 of tetrachloroethylene in many studies, the small numbers of exposed cases, and possible biases,
21 particularly in studies where the information on birth outcome or exposure is obtained by
22 questionnaire or inferred by residence. The epidemiologic evidence is strongest for spontaneous
23 abortions among exposed women. A number of studies have reported an elevated risk of
24 spontaneous abortion and maternal exposure to tetrachloroethylene, primarily exposure received
25 through employment as a dry cleaner (Doyle et al., 1997; Windham et al., 1991; Olsen et al.,
26 1990; Lindbohm et al., 1990; Kyyronen et al., 1989; Bosco et al., 1987).
27 The epidemiologic evidence for infertility is further suggestive of an association with
28 tetrachloroethylene exposure (Rachootin and Olsen, 1983; Eskenazi et al., 1991a, b; Sallmen et
29 al., 1995). Strong conclusions about effects on birth weight—IUGR (SGA)—or congenital
30 anomalies and tetrachloroethylene exposure cannot be drawn from the available occupational
31 studies, although drinking water studies of exposures to multiple chemicals, including
32 tetrachloroethylene, provide some limited evidence. There is very little information about what
33 exposure pattern (concentration and duration) is associated with these effects.
34 Inhalation of tetrachloroethylene by pregnant mice and rats during various fractions of
35 the gestation period has resulted in fetal growth retardation and mortality in several studies and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-134 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-10. Summary of animal developmental/reproductive studies
fortetrachloroethylene, in chronological order
Subjects
Effects
Concentration
Authors
SW Mice
Maternal toxicity, decreased fetal
weight, delayed ossification, 9%
decrease in birth weight
300 ppm on gestation days 6-15
Schwetz et al.
(1975)
SD Rats
Maternal toxicity, increased
resorptions (fetal death)
300 ppm on days 6-15
Schwetz et al.
(1975)
CFY Rats
Maternal toxicity (decreased
body weight gain, increased liver
weight and serum enzymes);
increased pre- and post-
implantation loss, skeletal
retardation, and total
malformations; decreased fetal
weight
1,500, 4,500, 8,500 mg/m3 on
gestation days 1-20
LOAEL= 1,500 mg/m3
Szakmary et al.
(1997)
CFY Rats
Decreased postnatal survival,
minimal transient decreases in
exploratory activity and muscular
strength, and increased motor
activity in females on postnatal
day 100
1,500, 4,500 mg/m3 on gestation
days 1-20 (and perhaps
postnatally to PND 100)
LOAEL= 1,500 mg/m3
Szakmary et al.
(1997)
C57B1 Mice
Maternal toxicity (increased liver
weight); visceral malformations
1,500 mg/m3 on gestation days
7-15
LOAEL= 1,500 mg/m3
Szakmary et al.
(1997)
NZW Rabbits
Maternal toxicity (decreased
body weight gain, increased liver
weight); abortions, total litter
resorptions, increased
postimplantation loss,
malformations
4,500 mg/m3 on gestation days
7-20
LOAEL = 4,500 mg/m3
Szakmary et al.
(1997)
SDRats
Decreased weight gain,
behavioral changes (more
extensive for late pregnancy
exposure), decreased brain
acetylcholine
0, 100, 900 ppm on days
7-13 or on days 14-20
NOAEL = 100 ppm
LOAEL = 900 ppm
Nelson et al.
(1980)
Mice
Abnormal sperm heads at 500
ppm but not at 100 ppm,
spermatogonia or spermatocyte
stage affected
0, 100, 500 ppm for 5 days
LOAEL = 500 ppm
Beliles et al.
(1980)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-135 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4-10. Summary of animal developmental/reproductive studies
fortetrachloroethylene, in chronological order (continued)
Subjects
Rats, rabbits
SD rats
Rat embryo in
culture
F344 rats
SD rats, two-
generation
study
Rats
Effects
No developmental toxicity
Maternal toxicity (decreased
body weight gain; decreased
gravid uterine weight); fetal body
weight and placental weight
decrements, increased delays in
thoracic vertebral ossification
Mortality, malformations,
delayed growth and
differentiation
100% mortality at 1,200 mg/kg-
day, increased mortality and
micro/anophthalmia at 900
mg/kg-day; soft tissues not
examined
Increased death of F1A and F2A
and F2B pups, decreased body
weight
Reduced fertilizability of
extracted oocytes
Concentration
Exposures throughout gestation
NOAEL = 500 ppm
0, 75, 250 or 600 ppm (actual
concentrations: 0, 66, 249, 600
ppm), 6 hr/day, 7 days/week, on
gestation days 0-19
Maternal LOAEL = 600 ppm
Fetal LOAEL = 250 ppm
No effect at 2.5 mM, effects at
3.5 mM and higher
Gavage, 900, 1,200 mg/kg-day on
gestation days 6-19
0, 100, 300,
1,000 ppm
NOAEL = 100 ppm for
body weight reduction
12,000 mg/m3, 2 hrs/day, 5
days/week for 2 weeks
Authors
Hardin et al.
(1981)
Carney et al.
(2006)
Saillenfait et al.
(1995)
Narotsky and
Kavlock(1995)
Tinston(1994)
Berger and
Horner (2003)
4
5
6
7
8
9
10
11
12
13
14
in delayed behavioral changes in the two studies that measured these effects. Single studies have
shown changes in brain acetyl choline and dopamine, altered brain fatty acid composition, and
altered sperm morphology. These effects occurred at doses higher than 300 to 1,000 ppm in
various studies.
The overall NOAEL for the animal developmental/reproductive inhalation studies is 100
ppm, based on Tinston (1994). The overall LOAEL is 300 ppm, based on Tinston (1994) and
Schwetz et al. (1975), in which increased mortality and decreased body weight of the offspring
were observed. All studies used the inhalation route of exposure except for one gavage study
(Fredriksson et al., 1993), which showed behavioral toxicity.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-136 DRAFT-DO NOT CITE OR QUOTE
-------
1 The finding of spontaneous abortions in several human studies of dry cleaners is
2 supported by the occurrence of reduced birth weight and mortality in several animal studies. The
3 finding of low birth weight in the Camp Lejeune studies by ATSDR is supported by reduced
4 birth weight in five animal studies (Schwetz et al., 1975; SzakmacZy et al., 1997; Nelson et al.,
5 1980; Carney et al., 2006; and in the Fl generation but not the F2 generation of Tinston (1994).
6 There are no human observations of behavioral changes to compare with the animal evidence of
7 CNS effects. The subtle nonadverse effects on sperm seen in humans correspond to one report
8 of abnormal sperm in mice. The LOAEL for developmental/reproductive effects in animals is
9 300 ppm, but no quantitative measures are available for human effects, although other dry
10 cleaner studies cited in previous sections have 8-hr TWA exposures of 10-20 ppm.
11 For oral exposures in humans there are only suggestive effects associated with
12 tetrachloroethylene exposure in drinking water, with no reliable data on exposures; therefore, this
13 information does not contribute to the determination of a LOAEL. There is little information on
14 developmental or reproductive effects in animals by the oral route of exposure.
15
16 4.7.4. Mode of Action for Developmental Effects
17 Because of its lipid solubility, tetrachloroethylene can cross both the blood-brain barrier
18 and the placental barrier and, therefore, it can be present in all tissues, including the brain, during
19 development.
20 Peroxidation of the lipids of the cell membranes (Cojocel et al., 1989), alteration of
21 regulation of fatty acid composition of the membrane (Kyrklund and Haglid, 1991), disturbances
22 in the properties of the nerve membrane (Juntunen, 1986), and progressively increased activity in
23 one or more of the phosphoinositide-linked neurotransmitters (Subramoniam et al., 1989) have
24 all been suggested as MO As for neurotoxic effects. These mechanisms could be involved during
25 development phases, as well as in adults.
26 The metabolite TCA may be the causative agent for developmental toxicity expressed as
27 morphological changes, lethality, or growth. Evidence in support of this speculative position is
28 presented in the following discussion. TCA is a weak organic acid, as are many developmental
29 toxicants, such as ethylhexanoic acid and valproic acid. These materials accumulate to a greater
30 extent in the embryo/fetal compartment than in the mother, based on the pKa of the acid and the
31 pH gradient between the maternal plasma and the embryo compartments (O'Flaherty et al.,
32 1992). TCA could induce developmental toxicity by changing the intracellular pH or through
33 peroxisome proliferation. Ghantous et al. (1986) detected TCA in the amniotic fluid of pregnant
34 mice exposed to tetrachloroethylene via inhalation.
35 Smith et al. (1989) found that oral gavage doses of TCA (330, 800, 1,200, and 1,800
36 mg/kg-day) delivered on gestation days 6-15 to pregnant Long-Evans rats produced soft tissue
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-137 DRAFT-DO NOT CITE OR QUOTE
-------
1 malformations, principally in the cardiovascular system. Johnson et al. (1998) found cardiac
2 defects in rat fetuses whose mothers received 2,730 ppm TCA in drinking water during the
3 period of cardiac development. Saillenfait et al. (1995), using the rat whole embryo (day 10)
4 culture system, found that both tetrachloroethylene and TCA induced embryo toxicity, including
5 mortality, malformations, and delayed growth and differentiation. TCA produced a reduction in
6 the first branchial arch as well as other morphological changes at a lower concentration (2.5 mM)
7 than that at which tetrachloroethylene induced no adverse effect (3.5 mM). TCA also induced a
8 reduction of the yolk sac diameter at 1 mM.
9
10 4.8. TOXIC EFFECTS IN OTHER ORGAN SYSTEMS
11 This section discusses effects in organ systems not covered previously. It does not
12 include effects on the liver, kidney, or nervous system; nor does it include developmental or
13 reproductive effects. To be consistent with other sections of the document, effects in humans are
14 presented separately from those in animals. Immune effects and lymphoid cancer are the most
15 studied, and these are the predominant topics of the noncancer and cancer sections, respectively.
16
17 4.8.1. Human Studies
18 4.8.1.1. Noncancer Effects
19 4.8.1.1.1. Immune-related effects in humans. Adverse effects on the immune system resulting
20 from chemical exposure fall within the following principal domains: immunosuppression (host
21 resistance), immunostimulation, autoimmunity, and allergy-hypersensitivity. Various
22 immunologic measurements (e.g., T-cell counts, immunoglobulin (Ig) E levels, specific
23 autoantibodies) may provide evidence of altered immune response that may subsequently be
24 related to risk of clinically expressed diseases such as infections, asthma, or systemic lupus
25 erythematosus. Tetrachloroethylene exposure via air or water may result in immune-mediated
26 organ-specific or systemic effects, as described in a case report of hypersensitivity pneumonitis
27 in a 42-year-old female dry cleaner worker (Tanois et al., 2004). Another case report described
28 severe fatigue, weight loss, myalgia, arthralgia, cardiac arrhythmia, decreased T-cell count, high-
29 titer (1:160) antinuclear antibodies, and neurological symptoms that were linked to an unusual
30 chemical sensitivity to tetrachloroethylene in a municipal water supply (Rea et al., 1991).
31
32 4.8.1.1.1.1. Tetrachloroethylene and immunologic parameters. Byers et al. (1988) provide
33 data pertaining to immune function from 23 family members of leukemia patients in Woburn,
34 Massachusetts. In 1979, testing of the wells in this town revealed that the water in two of the
35 wells was contaminated with a number of solvents, including tetrachloroethylene (21 ppb) and
36 trichloroethylene (267 ppb; as cited in Lagakos et al., 1986). These wells had been in operation
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-138 DRAFT-DO NOT CITE OR QUOTE
-------
1 from 1964 to 1979. Byers et al. collected serum samples in May and June of 1984 and in
2 November of 1985. They determined the total lymphocyte counts and lymphocyte
3 subpopulations (CDS, CD4, CDS), and the CD4/CD8 ratio in these samples, and in samples from
4 a combined control group of 30 laboratory workers and 40 residents of Boston selected through a
5 randomized probability area sampling process. The study authors also assessed the presence of
6 autoantibodies (antismooth muscle, antiovarian, antinuclear, antithyroglobulin, and
7 antimicrosomal antibodies) in the family member samples and compared the results with
8 laboratory reference values. The lymphocyte subpopulations were higher and the CD4/CD8
9 ratio was lower in the Woburn family members compared to the controls in both of the samples
10 taken in 1984. In the 1985 samples, however, the subpopulation levels had decreased and the
11 CD4/CD8 ratio had increased; the values were no longer statistically different from the controls.
12 None of the family member serum samples had antithyroglobulin or antimicrosomal antibodies,
13 but 10 family member serum samples (43%) had antinuclear antibodies (compared to <5%
14 expected based on the reference value). Because the initial blood sample was taken in 1984, and
15 because of the considerable mixture of exposures that occurred in this setting, it is not possible to
16 determine the patterns at a time nearer to the time of the exposure, or to infer the exact role of
17 tetrachloroethylene in alterations of the immunologic parameters.
18 Andrys et al. (1997) examined immunologic parameters in 21 dry cleaning workers (20
19 women) and 16 office workers in the dry cleaning plant (14 women) and compared them to
20 reference values based on samples from blood donors and "healthy persons in the same region"
21 (n = 14-311, depending on the test). The mean age of the exposed workers and office controls
22 was 45.7 years and 31.9 years, respectively; no information was provided on the age or sex
23 distribution of the reference controls. The tests included measures of Ig, A, G, M, and E levels,
24 complement (C3 and C4) levels, phagocyte activity, C-reactive protein, a-macroglobulin, T-
25 lymphocytes, and a blast transformation test. Several differences were observed between the
26 exposed workers and the office workers (e.g., higher levels of serum complement C3 and C4,
27 and of salivary IgA in the exposed), and between the exposed workers and the reference controls
28 (reduced T-lymphocytes, higher phagocytic activity, higher C3 levels in exposed). However,
29 there were also many differences noted between the office workers and reference group
30 (including reduced T-lymphocytes in office workers). The lack of information about the
31 reference group adds to the difficulty in interpreting these results.
32
33 4.8.1.1.1.2. Tetrachloroethylene andimmunosuppression. In 1982, Lagakos et al. (1986)
34 conducted a telephone survey of residents of Woburn, Massachusetts, collecting information on
35 residential history and history of 14 types of medically diagnosed conditions. The survey
36 included 4,978 children born since 1960 who lived in Woburn before age 19. Completed
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-139 DRAFT-DO NOT CITE OR QUOTE
-------
1 surveys were obtained from approximately 57% of the town residences with listed phone
2 numbers. Lagakos et al. used information from a study by the Massachusetts Department of
3 Environmental Quality and Engineering to estimate the contribution of water from the two
4 contaminated wells to the residence of each participant, based on zones within the town
5 receiving different mixtures of water from various wells, for the period in which the
6 contaminated wells were operating. This exposure information was used to estimate a
7 cumulative exposure based on each child's length of residence in Woburn. A higher cumulative
8 exposure measure was associated with history of kidney and urinary tract disorders (primarily
9 kidney or urinary tract infections) and with lung and respiratory disorders (asthma, chronic
10 bronchitis, or pneumonia). There are no other human data that characterize the effects of
11 tetrachloroethylene-only exposure on immunosuppression, as measured by increased
12 susceptibility to infections.
13
14 4.8.1.1.1.3. Tetrachloroethylene and autoimmune disease. In the 1970s, recognition of a
15 scleroderma-like disease characterized by skin thickening, Raynaud's phenomenon, and
16 acroosteolysis and pulmonary involvement in workers exposed to vinyl chloride (Gama et al.,
17 1978) prompted research pertaining to the role of organic solvents in autoimmune diseases.
18 Exposure to the broad categories of solvents, organic solvents, or chlorinated solvents has been
19 associated with a 2- to 3-fold increased risk of systemic sclerosis (scleroderma) in epidemiologic
20 studies summarized in a recent meta-analysis (Aryal et al., 2001) and in subsequent studies
21 (Garabrant et al., 2003; Maitre et al., 2004). Similar results were seen in studies of other
22 systemic autoimmune diseases including undifferentiated connective tissue disease (Lacey et al.,
23 1999), rheumatoid arthritis (Lundberg et al., 1994; Sverdrup et al., 2005), and antineutrophil-
24 cytoplasmic antibody (ANCA)-related vasculitis (Lane et al., 2003; Beaudreuil et al., 2005). In
25 contrast, there was little evidence of an association between solvent exposure and systemic lupus
26 erythematosus in two recent case-control studies (Cooper et al., 2004; Finckh et al., 2007).
27 As described in the preceding paragraph, the epidemiologic data in relation to the role of
28 solvents, as a broad category, in systemic autoimmune diseases, varies among these conditions.
29 Much more limited data is available pertaining to specific solvents, including tetrachloroethylene,
30 and risk of autoimmune diseases. Case reports have been published describing a condition
31 similar to vinyl-chloride induced scleroderma in a man who worked as a presser in a dry cleaning
32 plant, and who also helped clean the tetrachloroethylene-containing drums on a weekly basis
33 (Sparrow, 1977), and a localized scleroderma in a man who had worked with tetrachloroethylene
34 as a metal degreaser, with workplace exposures reported to be between 10-25 ppm (Hinnen et al.,
35 1995; in German). Among 279 cases with connective tissue disease, Goldman (1996) observed a
36 higher frequency of individuals who reported employment as a dry cleaner among systemic
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-140 DRAFT-DO NOT CITE OR QUOTE
-------
1 sclerosis patients (4 of 33) compared with patients with other connective tissue diseases (1 of
2 246; p < 0.01). Similar patterns were seen with self-reported history of tetrachloroethylene
3 exposure (3 of 33 systemic sclerosis patients compared with 2 of 246 other patients,/* < 0.01),
4 but the author noted the difficulty in obtaining this type of information.
5 One registry-linkage study from Sweden of rheumatoid arthritis (Lundberg et al., 1994)
6 and three case-control studies of undifferentiated connective tissue disease (Lacey et al., 1999),
7 scleroderma (Garabrant et al., 2003), and antineutrophil-cytoplasmic antibody (ANCA) related
8 diseases (Beaudreuil et al., 2005) provide data concerning dry cleaning work or
9 tetrachloroethylene exposure (Table 4-11). As expected in population-based studies, the
10 exposure prevalence is low, with approximately 4% of controls reporting work in dry cleaning
11 and 1% reporting exposure to tetrachloroethylene. The observed associations are generally weak
12 (odds ratios for dry cleaning around 1.5 for the 3 large studies of women) and none of the
13 individual studies are statistically significant. The results seen for the exposure to
14 tetrachloroethylene in the three studies that attempted this kind of assessment were more varied
15 (Lacey et al., 1999; Garabrant et al., 2003; Beaudreuil et al., 2005). Only the study of ANCA-
16 related diseases resulted in an elevated odds ratio, but again this estimate was somewhat
17 imprecise (OR = 2.0, 95% CI = 0.6, 6.9; Beaudreuil et al., 2005). These studies are clearly
18 limited by the low prevalence of and difficulty in accurately characterizing occupational
19 exposure to tetrachloroethylene in population-based or clinical settings.
20
21 4.8.1.1.1.4. Tetrachloroethylene and allergy and hypersensitivity. Allergy and hypersensitivity,
22 as assessed with measures of immune system parameters or immune function tests (e.g., asthma,
23 atopy) in humans, have not been extensively studied with respect to the effects of
24 tetrachloroethylene.
25 Th2 cytokines (e.g., interleukin-4) stimulate production of IgE and Thl cytokines (e.g.,
26 interferon-y) act to inhibit IgE production. Lehmann et al. (2001) examined IgE levels and
27 cytokine producing cells (interferon-y, tumor necrosis factor-a, and interleukin-4) in relation to
28 indoor levels of volatile organic compounds among children (age 36 months) selected from a
29 birth cohort study in Leipzip, Germany. The hypothesis underlying this work is that a shift in
30 Thl to Th2 cytokine profile is a risk factor for IgE-mediated allergic disease in children (Tang et
31 al., 1994; Warner et al., 1994). Enrollment into the birth cohort occurred between 1995 and
32 1996. The children in this allergy study represent a higher-risk group for development of allergic
33 disease, with eligibility criteria that were based on low birth weight (between 1,500 and 2,500 g),
34 or cord blood IgE greater than 0.9 kU/1 with double positive family history of atopy. These
35 eligibility criteria were met by 429 children; 200 of these children participated in the allergy
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-141 DRAFT-DO NOT CITE OR QUOTE
-------
o
Oi
Table 4-11. Immune-related conditions in studies of dry cleaning or tetrachloroethylene exposure in humans"
oo S3'
I
I
§
***.
£3'
1
TO'
il
I
TO
Co
Condition, location, diagnosis period,
sample size, demographics, source of data
Results
Authors
Autommune diseases
Rheumatoid arthritis, Sweden (13 counties),
hospitalized 1981-1983, 896 male cases,
629 female cases; population comparison
(total 370,035 men, 140,139 women), ages
35-74. Registry linkage to 1960 and 1970
census occupation data
launderers and dry cleaning
men: 1 exposed cases; OR = 0.8 (95% Cl = 0.1-5.0)
women: 7 exposed cases; OR = 1.5 (95% Cl = 0.7-3.2)
Lundberg et al. (1994)
Undifferentiated connective tissue disease,
Michigan and Ohio, diagnosed 1980-1991
(Michigan) 1980-1992 (Ohio). 205 cases,
2095 population controls. Women, ages 18
and older. Structured interview (specific jobs
and materials; jobs held 3 or more months)
dry cleaning
cases: 4.3%, controls 3.8% OR = 1.4 (95% Cl = 0.68, 2.8)
tetrachl oroethy 1 ene:
cases: 0%, controls 1% OR = 0.00
Lacey etal. (1999)
H I
O >
HH Oq
H TO
O
Scleroderma, Michigan and Ohio. Diagnosed
1980-1991 (Michigan), 1980-1992 (Ohio).
660 cases, 2227 population controls.
Women, ages 18 and older. Structured
interview (specific jobs and materials; jobs
held 3 or more months)
dry cleaning
cases: 4.7%, controls 3.7% OR = 1.4 (95% Cl = 0.9, 2.2)
tetrachl oroethy 1 ene:
self report cases: 1.1%, controls 1.0% OR = 1.4 (95% Cl =
0.6, 3.4)
expert review cases: 0.8%, controls 0.8% OR = 1.1 (95%
CI = 0.4, 2.9)
Garabrant et al. (2003)
ANCA-related diseases,13 France. Diagnosed
1999-2000. 60 patients, 120 hospital
controls, men and women (50% each), mean
age 61 years
tetrachl oroethy 1 ene
cases: 8.3%, controls 4.1% OR = 2.0 (0.6-6.9)
Beudreuil et al. (2005)
H
W
-------
Table 4-11. Immune-related conditions in studies of dry cleaning or tetrachloroethylene exposure in humans*
(continued)
Condition, location, diagnosis period,
sample size, demographics, source of data
Results
Authors
Allergy and hypersensitivity
IgE levels, Germany, 1995-1996. 121
children (ages 36 months), selected based on
high risk profile for allergic diseases, blood
sample and indoor air sampling (child's
bedroom) of 26 volatile organic chemicals (4
weeks around age 36 months)
no association between tetrachloroethylene measures and
total IgE or IgE-specific allergen antibodies
Lehmannetal. (2001)
CD3 T-cell subpopulations from cord blood,
Germany 1997-1999. 85 newborns, cord
blood and indoor air sampling (child's
bedroom) of 28 volatile organic chemicals (4
weeks immediately after birth)
Tetrachloroethylene exposure associated with decreased
interferon-y cells, but no association with interleukin-4,
interleukin-2, or tumor necrosis factor-a
Lehmann et al. (2000)
Exacerbation of asthmatic symptoms, Los
Angeles, 1999-2000. 21 children (ages
10-16 years), 3 month diaries, ambient
levels and exhaled breath measures of 8
volatile organic compounds and 8 criteria
pollutant gases
Little evidence of an association between ambient
tetrachloroethylene exposure or exhaled tetrachloroethylene
measures and asthma symptoms
Delfino et al.
(2003 a, b)
a Includes case-control studies and cross-sectional studies, but does not include case reports.
b ANCA = antineutrophil-cytoplasmic antibody. Diseases included Wegener glomerulonephritis (n = 20), microscopic polyangiitis (n = 8), pauci-immune
glomerulonephritis (n = 10), uveitis (n = 6), Churg-Strauss syndrome (n = 4), stroke (n = 4) and other diseases (no more than 2 each)
-------
1 study described below, but complete data (IgE and volatile organic compound measurements)
2 were available for only 121 of the study participants.
3 Lehmann et al. measured 26 volatile organic compounds via passive indoor sampling in
4 the child's bedroom for a period of 4 weeks around the age of 36 months. The highest exposures
5 were seen for limonene (median 19.1 ug/m3), a-pinene (median 16.3 ug/m3), and toluene
6 (median 13.3 ug/m3). The median exposure of tetrachloroethylene was 2.5 ug/m3 (0.87 ug/m3
7 and 5.1 ug/m3 for the 25th and 75th percentiles, respectively). The only strong correlation
8 (r > 0.3) between tetrachloroethylene and the other volatile organic compounds measured was a
9 correlation of 0.72 with trichloroethylene. Blood samples were taken at the 36-month-study
10 examination and were used to measure the total IgE and specific IgE antibodies directed to egg
11 white, milk, indoor allergens (house dust mites, cat, molds), and outdoor allergens (timothy -
12 perenial grass, birch- tree). There was no association between tetrachloroethylene exposure and
13 any of the allergens tested in this study, although some of the other volatile organic compounds
14 (e.g., toluene, 4-ethyltoluene) were associated with elevated total IgE levels and with
15 sensitization to milk or eggs.
16 Another study by Lehmann et al. (2002) examined the relationship between indoor
17 exposures to volatile organic compounds and T-cell subpopulations measured in cord blood of
18 newborns. The study authors randomly selected 85 newborns (43 boys and 42 girls) from a
19 larger cohort study of 997 healthy, full-term babies, recruited between 1997 and 1999 in
20 Germany. Exclusion criteria included a history in the mother of an autoimmune disease or
21 infectious disease during the pregnancy. Twenty-eight volatile organic compounds were
22 measured via passive indoor sampling in the child's bedroom for a period of 4 weeks after the
23 birth (a period which is likely to reflect the exposures during the prenatal period close to the time
24 of delivery). The levels were generally similar or slightly higher than the levels seen in the
25 previous study using samples from the bedrooms of the 36-month-old children. The highest
26 levels of exposure were seen for limonene (median 24.3 ug/m3), a-pinene (median 19.3 ug/m3)
27 and toluene (median 18.3 ug/m3), and the median exposure of tetrachloroethylene was 3.4 ug/m3
28 (1.8 ug/m3 and 7.3 ug/m3 for the 25th and 75th percentiles, respectively). Flow cytometry was
29 used to measure the presence of CD3 T-cells obtained from the cord blood labeled with
30 antibodies against interferon-y, tumor necrosis factor-a, interleukin-2 and interleukin-4.
31 Tetrachloroethylene was the only one of the measured volatile organic compounds that was
32 associated with a reduced level of interferon-y. In the univariate analysis, the median percentage
33 of interferon-y cells was 3.6% and 2.6% in the groups that were below the 75th percentile and
34 above the 75th percentile of tetrachloroethylene exposure, respectively. The odds ratio between
35 high (above the 75th percentile) tetrachloroethylene exposure and reduced (less than the 25th
36 percentile) levels of interferon-y cells was 2.9 (95% CI = 1.0-8.6), adjusting for family history of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-144 DRAFT-DO NOT CITE OR QUOTE
-------
1 atopy, gender and smoking history of the mother during pregnancy. There was no association
2 between tetrachloroethylene exposure and interleukin-4 cells, but naphthalene and
3 methylcyclopentane were associated with elevated levels of interleukin-4 cells.
4 Delfino et al. (2003a, b) examined the exacerbation of asthmatic symptoms following
5 exposure to volatile organic compounds that occurred due to variation in air quality over a 3
6 month period in 1999-2000 in Los Angeles. This study included daily repeated exposures to
7 ambient air pollutants and peak expiratory flow rates over a 3-month period in 21 children (17
8 males and 4 females) of Hispanic origin ages 10-16 years; an additional child participated in the
9 ambient air but not in the exhaled air portion of the study. Daily diaries were used to record
10 severity of symptoms and asthmatic episodes. Exposure metrics included exhaled breath
11 measures and ambient levels of eight volatile organic compounds (benzene, methylene chloride,
12 styrene, toluene, w,/?-xylene, o-xylene, />-dichlorobenzene, and tetrachloroethylene) and eight
13 criteria pollutant gases. An association between criteria air pollutants and subsequent symptoms
14 of asthma in children in the Los Angeles area suggest an increased risk of adverse health
15 outcomes with exposure to SO2 and NO2 (Delfino et al., 2003a). Although ambient levels of
16 tetrachloroethylene were associated with bothersome asthma symptoms (OR = 1.37, [95% CI =
17 1.09, 1.71]) per an interquartile range change), this association was reduced with the adjustment
18 for SO2 or NO2 (Delfino et al., 2003a). In the 21 children who partcipated in the peak expiratory
19 flow measurements, the mean breath level of tetrachloroethylene was 4.40 ng/L (sd 10.77 ng/L),
20 the mean ambient level was 3.52 (sd 2.17) ng/L, and the correlation between the same-day
21 measures was 0.31 (p < 0.01; Delfino et al., 2003b). There was little relation between asthma
22 symptoms and exhaled breath levels of tetrachloroethylene. The mean exhalation of
23 tetrachloroethylene was 2.50 and 2.69 ng/L, respectively, in the two groups of asthma symptoms
24 (none or not bothersome; bothersome and more severe). Stronger associations were reported
25 between asthma symptoms and some of the other volatile organic chemicals, specifically for
26 benzene, toluene, m,p-xy\Qne.
27 The limited available data from these studies (Lehmann et al., 2001; Lehmann et al.,
28 2002; Delfino et al., 2003a, b), provide weak evidence of an effect of tetrachloroethylene
29 exposure during childhood on allergic sensitization or exacerbation of asthma symptomology.
30 However, the observation of the association between increased tetrachloroethylene exposure and
31 reduced interferon-y in cord blood samples may reflect a sensitive period of development, and
32 points to the current lack of understanding of the potential immunotoxic effects of prenatal
33 exposures.
34
35 4.8.1.1.2. Endocrine system effects. Only one study of endocrine system effects was found in
36 the published literature. Ferroni et al. (1992) observed an increased serum concentration of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-145 DRAFT-DO NOT CITE OR QUOTE
-------
1 prolactin among tetrachloroethylene-exposed dry cleaners as compared with controls (12.1 + 6.7
2 Hg/L vs. 7.4 + 3.1 |ig/L). The median tetrachloroethylene concentration to which these workers
3 were exposed was 15 ppm. The variance of serum prolactin concentration was wider for
4 exposed subjects than for unexposed controls, and 4 of 41 subjects had serum prolactin
5 concentrations above the upper normal limit (defined by Ferroni et al. as 25 |ig/L), compared
6 with none of 23 controls. The prevalence of abnormal concentrations, however, was not
7 statistically significantly elevated. Positive correlations between the response and
8 tetrachloroethylene were not observed with either exposure duration or biomarker measures.
9 The evaluation of prolactin was part of an overall assessment of neurobehavioral functioning
10 (see the discussion of this study in the section on neurobehavioral effects) for which these
11 investigators hypothesized a relationship between dopamine and serum prolactin concentration
12 (Ferroni et al., 1992; Mutti and Smargiassi, 1998).
13 Epidemiologic studies of other parameters of endocrine function are lacking. The
14 endocrine system can be considered a potential target for tetrachloroethylene because another
15 like solvent, trichloroethylene, has been shown to induce endocrine system changes in both
16 humans (Goh et al., 1998; Chia et al., 1996, 1997) and experimental animals (Kumar et al., 2000).
17
18 4.8.1.2. Cancer
19 The body of literature reporting carcinogenic effects in humans associated with exposure
20 to tetrachloroethylene consists of cohort, proportional mortality, and case-control studies. A
21 small number of studies, including studies of cohorts involved in metal degreasing or in aircraft
22 manufacturing/maintenance (Boice et al., 1999; Anttila et al.; 1995, Spirtas et al., 1991), have
23 assessed tetrachloroethylene exposure explicitly. These cohort studies present risks associated
24 with site-specific cancer mortality (Boice et al., 1999; Spirtas et al., 1991) or incidence (Anttila
25 et al., 1995) for a subcohort of the larger study population who were exposed to
26 tetrachloroethylene. Additionally, a few case-control studies were able to examine the
27 relationship between cancer at specific sites and tetrachloroethylene exposure (Vaughan et al.,
28 1997; Schlehofer et al., 1995; Pesch et al., 2000a; Heineman et al., 1994).
29 A larger body of evidence on cancer exists for workers employed in dry cleaning. Dry
30 cleaners have potential exposures to a number of solvents, including tetrachloroethylene, which
31 has been in widespread use since the early 1960s (IARC, 1995). Information on the potential for
32 tetrachloroethylene exposure and concentration measurements is lacking for individual study
33 subjects in these studies: however, the cohort studies of Ruder et al. (1994, 2001) and Blair et al.
34 (1990, 2003) are of individuals primarily exposed to tetrachloroethylene (Lynge et al., 1997).
35 The exposure assessment approach of Lynge et al. (2006), a case-control study nested within a
36 cohort of dry-cleaning and laundry workers, relies on job title to increase sensitivity for
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-146 DRAFT-DO NOT CITE OR QUOTE
-------
1 tetrachloroethylene exposure identification, with dry cleaners identified as having greater
2 potential for tetrachloroethylene exposure. This study lacks information on tetrachloroethylene
3 concentration for individual study subjects, is unable to classify job title for 20% of study
4 subjects, and has a large percentage of cases from Sweden and Norway with job title provided by
5 next-of-kin.
6 Although several community-based drinking water studies are available (Aschengrau et
7 al., 1993, 1998, 2003; Paulu et al., 1999; Fagliano et al., 1990; Cohn et al., 1994; MA DPH,
8 1997; Vartiainen et al., 1993; Lagakos et al., 1986), exposure in most of these studies was to a
9 mixture of solvents, including tetrachloroethylene and trichloroethylene except for the studies by
10 Aschengrau et al. (1993, 1998, 2003) and Paulu et al. (1999) that examined tetrachloroethylene
11 specifically. Summary tables of these analyses are presented in Appendix 4B.
12
13 4.8.1.2.1. Lymphoid cancer. A number of epidemiologic studies including degreaser cohort
14 (tetrachloroethylene subcohorts), dry cleaner and laundry worker cohort, case-control, and
15 community studies have examined tetrachloroethylene exposure and lymphoid cancer. Elevated
16 risks of lymphoid cancer incidence, specifically non-Hodgkin's lymphoma (NHL), were
17 observed in studies of degreasers exposed to tetrachloroethylene; a total of 15 cases were
18 observed in three available studies (Table 4B-2, Appendix 4B) versus 6.8 expected cases or
19 deaths (95% CI = 1.2-3.7). Spirtas et al. (1991) observed a statistically significant elevated risk
20 for NHL for males and females combined (standardized mortality ratio [SMR] = 4.0, 95% CI =
21 1.1-10.2, four deaths). Two of the four deaths occurred among females, with females having the
22 highest risk. The tetrachloroethylene cohorts of Anttila et al. (1994) and of Boice et al. (1999)
23 support the findings in Spirtas et al. (1991). NHL risk is elevated—but not statistically
24 significantly—in both studies (routine exposed, SMR = 1.7, 95% CI = 0.7-3.3, eight deaths,
25 Boice et al., 1999; SIR = 3.8, 95% CI = 0.8-11.0, three cases, Anttila et al., 1994). Boice etal.
26 (1999) also present an analysis of duration of exposure-response gradient for NHL mortality;
27 however, the inclusion of deaths with either routine or intermittent exposure in this analysis (a
28 total of 16 deaths), with nonsolvents-exposed factory workers as referents prevents a comparison
29 to the excess NHL risk reported for the eight deaths with routine exposure to tetrachloroethylene
30 (Table 8 in Boice et al., 1999). NHL relative risk for subjects with >5 years duration of exposure
31 (intermittent or routine exposure) to tetrachloroethylene was 1.6 (95% CI = 0.8-3.2) in Poisson
32 regression analysis using internal controls (factory workers not exposed to any solvent) and no
33 indication of a linear trend of increasing RR with increasing duration of exposure (p < 0.20).
34 The inclusion of subjects with differing exposure patterns in an analysis of exposure duration
35 likely introduces misclassification bias and does not diminish observations of routinely exposed
36 subjects.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-147 DRAFT-DO NOT CITE OR QUOTE
-------
1 The three degreaser cohort studies provide only limited information on lymphoid tumors
2 other than NHL due to few total numbers of observed deaths or incident cases for lymphoid
3 neoplasms and, in general, a high proportion of cancers attributable to NHL. Furthermore, none
4 of these cohort studies provide information on leukemia subtype. For example, Anttila et al.
5 (1994) observed three cases of lymphoid tumors, and all three were attributable to NHL.
6 Noteworthy, however, is the observation in one study of aircraft maintenance workers (Spirtas et
7 al., 1991) of a large but imprecise risk for multiple myeloma and exposure to tetrachloroethylene
8 (SMR = 17.0, 95% CI = 2.1-61.6) among females, which was not seen in another cohort of
9 mostly male aircraft manufacturing workers (Boice et al., 1999).
10 Eight studies of incidence in dry cleaners and laundry workers in Scandanavian countries,
11 one study of leukemia incidence in Portland, Oregon, and two studies of mortality of dry
12 cleaners in the United States were available for review: Morton and Marjanovic (1984); Lynge
13 and Thygesen (1990); Andersen et al. (1999); Ruder et al. (2001); Cano and Pollan, (2001);
14 Travier et al. (2002); Blair et al. (2003) and Ji and Hemminki, (2005b, 2006). Observations in
15 several of these studies are summarized in Table 4B-la and Table 4B-lb, Appendix 4B, while
16 others are discussed below. Anderson et al. (1999) examines cancer incidence by occupational
17 title in the 1970 census among citizens of Denmark, Finland, Norway, and Sweden, between
18 1971 and 1987-1991, depending on country. Site-specific lymphoma incidence, but not
19 leukemia subtype, is reported for launderers and dry cleaners. Further analysis of lymphoma
20 incidence in Swedish subjects, many of whom overlap with the larger cohort of Anderson et al.
21 (1999), was presented by Travier et al. (2002) who examined leukemia subtype and by Cano and
22 Pollan (2001) who examined non-Hodgkin's lymphoma incidence. Ji and Hemminki (2005b,
23 2006) expanded the Swedish cohort, following launderers and dry cleaners identified on 1960,
24 1970, 1980, and 1990 censuses another 10 years to 2002. Lynge and Thygesen (1990) examine
25 lymphoma incidence among Danish dry cleaners and launderers who were identified with this
26 job title in the 1970 census and followed for 10 years to 1980.
27 Overall, association with Hodgkin's disease, NHL, or chronic lymphocytic leukemia are
28 suggested, although site-specific risks are not always elevated or statistically significant in all
29 studies nor were they elevated in both sexes (see Table 4B-la). Two studies examined
30 Hodgkin's disease (Anderson et al., 1999; Travier et al., 2002) and both reported statistically
31 significant associations in females but not males: SIR =1.9 (95% CI = 1.1-2.9) in Andersen et
32 al. (1999), RR = 3.6 (95% CI = 1.2-11.1) in Travier et al. (2002). The elevated risk was
33 observed particularly among the subjects who were below 40 years of age in 1960 (Travier et al.,
34 2002). These are subjects who used mainly tetrachloroethylene with possibly some
35 trichloroethylene.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-148 DRAFT-DO NOT CITE OR QUOTE
-------
1 NHL risks in the four-country study of Andersen et al. (1999) were 1.5 (95% CI =
2 1.0-2.1) for males and 1.0 (95% CI = 0.7-1.2) for females. Separate analyses of Swedish and
3 Danish workers supported these observations: males, RR =1.4 (95% CI = 0.6-3.4); females,
4 RR = 0.5 (95% CI = 0.2-1.6; Travier et al., 2002); males, RR = 1.9 (95% CI = 0.8-4.1; Cano and
5 Pollan , 2001); males, SIR = 2.8 (95% CI = 0.9-6.5); females, SIR = 0.5 (95% CI = 0.1-1.5;
6 Lynge and Thygesen, 1990).
7 Three studies examine association with leukemia subtype (Ji and Hemminki, 2005b,
8 2006; Travier et al., 2002; Morton and Marjanovic, 1984) and all report statistically significant
9 association with chronic lymphocytic leukemia in females but not males. A noteworthy finding
10 in Travier et al. (2002) was a statistically significant elevated risk for leukemia among dry
11 cleaners or launderers in both the 1960 and 1970 Swedish census (RR = 1.8, 95% CI = 1.1-3.1),
12 mostly due to chronic lymphocytic leukemia (CLL); 6 of the 15 total leukemia cases, RR =1.9,
13 95% CI = 0.8-4.1; of which 5 of the 6 CLL cases were in females, RR = 2.9, 95% CI = 1.2-7.0.
14 Ji and Hemminki (2005b) also examined associations with leukemia subtypes and reported a
15 similar finding with chronic lymphocytic leukemia in females but not males (females: SIR =1.5,
16 95% CI= 1.1-2.1; males: SIR = 0.9, 95% CI = 0.5-1.3). Similarly Morton and Marjanovic
17 (1984) reported leukemia incidence, particularly lymphocytic leukemia incidence, was
18 statistically significantly higher in women laundry and dry cleaners in the Portland, Oregon area.
19 Age-standardized incidence rates (per 100,000) for women dry cleaners and laundry workers
20 compared to all women were: all leukemias, 23.7 compared to 6.7; lymphatic leukemia, 20.9
21 compared to 2.6. Many chronic lymphocytic leukemias and NHLs may arise from a common
22 cell type (Beers and Berkow, 1999, in Bukowski et al., 2003).
23 The update of the cohort mortality study by Ruder et al. (2001) found only six deaths
24 attributable to cancer of the lymphatic and hematopoietic system. The few deaths due to
25 lymphatic cancer greatly impact the statistical power in this study. Blair et al. (2003), in an
26 updated mortality analysis of a cohort that is predominately female, present observed and
27 expected number of deaths for categories of lymphomas. Although based on a small number of
28 deaths in several categories, these authors discuss this study as supporting an excess risk of
29 Hodgkin's disease (SMR = 2.0, 95% CI = 0.6-4.6, 5 cases) and this observation is consistent
30 with observations in the Scandinavian studies discussed above. Neither of the mortality studies
31 of United States dry cleaners examined leukemia subtype.
32 Ten case-control studies examine site-specific lymphomas and occupational exposure to
33 tetrachloroethylene or job title of dry cleaner or launderer (Lynge et al., 2006; Mester et al.,
34 2006; Miligi et al, 2006; Kato et al., 2005; Fabbro-Peray et al., 2001; Seniori Costantini et al.,
35 2001; Clevel et al. 1998; Blair et al., 1993; Scherr et al., 1992; and Hardell et al., 1981; Table
36 4B-5, Appendix 4B). Several studies examine cell type (Mester et al., 2006, Miligi et al., 2006,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-149 DRAFT-DO NOT CITE OR QUOTE
-------
1 Clevel et al., 1998). The only study available on Hodgkin's disease observed a statistically
2 significant elevated risk for male with a job title of dry cleaner or laundry worker (Seniori
3 Costantini et al., 2001). Risks above 1.0 were observed in several studies of NHL and CLL
4 although risks were not statistically significantly elevated, likely due to several factors discussed
5 below (Mester et al., 2006 Miligi et al., 1999; 2006; Kato et al., 2005; Clevel et al., 1998; Blair et
6 al., 1993; Scherr et al., 1992; Hardell et al., 1981).
7 Overall, case-control studies examining occupational exposure are quite limited for
8 evaluating lymphoma and tetrachloroethylene for a number of reasons. Reviewed case-control
9 studies are typically population-based studies. More recent studies are of large number of cases
10 and controls compared to older studies. These recent studies adopt procedures to blind
11 interviewers and apply more refined exposure assessment methods, examining
12 tetrachloroethylene specifically as opposed to a grouping of dry cleaners and laundry workers
13 (Lynge et al., 2006; Mester et al., 2006; and Miligi et al., 2006). However, the prevalence of
14 exposure to tetrachloroethylene or as a dry cleaner, or launderer, particularly long-duration
15 exposure, exposure, is low in these studies. A consequence of this is few exposed cases and
16 imprecise risks (wide confidence intervals) that reflect lower statistical power to examine
17 lymphoma and tetrachloroethylene exposure (Miligi et al., 2006; Mester et al., 2006; Seniori
18 Costantini et al., 2005). Four other aspects of population case-control studies are important to
19 consider in their interpretation. Risk is difficult to determine for low-prevalence jobs when
20 studying a regional population; associations may be nonlinear because categorical assignment of
21 duration and intensity are arbitrary and do not necessarily represent a linear dose relationship,
22 and duration and cumulative exposure variables do not address age at first exposure, which also
23 affects cancer risk (NRC, 2005). Additionally, missing information either as a result of lower
24 participation rates or missing job history data, can introduce bias—particularly misclassification
25 bias. For example, the lack of information to classify job title for roughly 20% of NHL cases
26 and controls in Lynge et al. (2006), in addition to a large number of next-of-kin interviews, limits
27 this study due to an increased potential for exposure misclassification and, as assessed using two
28 exposure groups (yes/no) in this study, results in risks close to 1.0 (no risk). Lynge et al. (2005)
29 is considered a null or uninformative study for this reason rather than a study supporting no
30 association between tetrachloroethylene and NHL. Furthermore, quantitative exposure
31 information is missing from all studies and leads to substantial misclassification of exposure.
32 Four case-control studies are available on childhood leukemia (acute lymphocytic
33 leukemia, ALL) and parental occupational exposure to tetrachloroethylene or to drinking water
34 contaminated with trichloroethylene, tetrachloroethylene, and other chlorinated solvents (Infante-
35 Rivard et al., 2005; Costas et al., 2002; Shu et al., 1999; Lowengart et al. 1987; Table 4B-5,
36 Appendix 4B). Many aspects discussed above for case-control studies examining occupational
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-150 DRAFT-DO NOT CITE OR QUOTE
-------
1 exposure are found in ALL studies. While some studies appear consistent (Costas et al.,2002;
2 Shu et al., 1999; Lowengart et al., 1987), these studies are insensitive for assessing association,
3 or lack thereof, between ALL and tetrachloroethylene exposure because observations are based
4 on few exposed cases. Other studies are needed to clarify the role of tetrachloroethylene and
5 ALL.
6 Four studies examine drinking water exposure and lymphoma: a case-control study by
7 Aschengrau et al. (1993), and the ecological analyses by Cohn et al. (1994); Fagliano et al.,
8 1990; and Vartiainen, 1985 (see Table 4B-13 and Appendix 4B). In a study by Aschengrau et al.
9 (1993), where tetrachloroethylene was identified as the putative exposure, an elevated risk of
10 leukemia was observed for those most exposed (90th percentile of exposure; with no latency,
11 OR =8.3; 95%CI= 1.5-45.3; considering a latency period of 5 years, OR = 5.9; 95% CI =
12 1.4-24.9). An exposure-response relationship is suggested; the crude unadjusted risk for high-
13 level exposure (exposure at the 90th percentile; OR = 6.0, 95% CI = 0.6-32.0) is larger than the
14 unadjusted risk for any exposure (OR =1.8, 95% CI = 0.6-4.3).
15 Moreover, the case-control study of Costas et al. (2002) and the ecologic studies by
16 Fagliano et al. (1990) and Cohn et al. (1994) provide some evidence of an association between
17 NHL or leukemia and drinking water that includes trichloroethylene and tetrachloroethylene.
18 The actual level of exposure to tetrachloroethylene and other solvents in these studies is not
19 known, and in the case of Costas et al. (2002), trichloroethylene was measured in the well water
20 at concentrations an order of magnitude higher than tetrachloroethylene. Each of these solvents
21 is hypothesized to be metabolized and bioactivated to TCA (see metabolism discussion). Thus,
22 exposure to tetrachloroethylene can be considered to add to the level of these metabolites
23 generated through trichloroethylene exposure.
24 The classification of lymphoid neoplasms, specifically lymphomas, has recently
25 undergone a revision, primarily on the basis of new findings from molecular biology, genetics,
26 and immunology, which have changed older concepts of lymphoid cancer, making them obsolete
27 (Herrinton, 1998). Although lymphomas have been classified in the past into distinct categories
28 (e.g., leukemia, reticulosarcoma/lymphosarcoma), lymphomas can share common biological
29 properties (Weisenburger, 1992) and differentiation pathways. For example, advances in
30 molecular biology have blurred the distinction between lymphoid leukemia and lymphoma
31 (Herrinton, 1998). Few studies assessing tetrachloroethylene exposure have included analyses of
32 diagnostic subcategories and no studies have examined cellular or molecular markers.
33 In 1994, the International Lymphoma Study Group published the revised European-American
34 Lymphoma (REAL) classification, and this system has been adopted by WHO (Harris et al.,
35 2000a). Modifications in the REAL include the grouping of lymphatic leukemia with NHL
36 (Herrinton, 1998; Miligi et al., 1999); the REAL/WHO Classification considers lymphomas and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-151 DRAFT-DO NOT CITE OR QUOTE
-------
1 lymphoid leukemias of the same cell type as one disease, with different clinical presentation or
2 stages (Harris et al., 2000b). This implies that the classification system used in many of the
3 epidemiologic studies is imprecise for both diseases. The resulting bias related to disease
4 misclassification affects observed risk estimates by masking underlying associations, biasing
5 risks towards the null or RR close to 1.0. Hence, the elevated risks observed for different
6 categories of lymphoid tumors in individual studies are noteworthy because of study
7 insensitivities and, additionally, these risks may not be inconsistent with the pathogenesis of
8 disease and with an etiology associated with tetrachloroethylene.
9 Rats exposed chronically to tetrachloroethylene for 2 years developed increased
10 incidences of mononuclear cell leukemia (MCL) or large granular lymphocyte leukemia (JISA,
11 1993; NTP, 1986a). Section 4.8.2.2.1 describes these observations and their interpretation as a
12 human cancer hazard. Large granular lymphocyte (LGL) cells exist in humans that are
13 morphologically, biochemically, and functionally similar to the cells involved in MCL in the
14 F344 rat (Stromberg, 1985). In humans, clonal disorders of LGLs represent a biologically
15 heterogeneous spectrum of lymphoid malignancies thought as originating either from mature
16 T-cell or natural killer (NK) cells (Sokol and Loughran, 2006). LGL disorders can clinically
17 present as an indolent (chronic) or aggressive diseases (Sokol and Loughran, 2006). The
18 indolent form of LGL leukemia is a disease of the elderly, with a median age at diagnosis of 60
19 years. A number of clinical conditions have been seen in patients with LGL leukemia. These
20 include the following: red cell aplasia and aplastic anemia; other lymphoproliferative disorders
21 such as NHL, Hodgkin's lymphoma, multiple myeloma, hairy cell leukemia, and B-cell
22 lymphoprliferative disorders; and autoimmune diseases such as rheumatoid arthritis and systemic
23 lupus erythematousus (Rose and Berliner, 2004). The etiology of LGL disorders is not known
24 (Rose and Berliner, 2004; Sokol and Loughran, 2006). Several possible etiologies have been
25 proposed including chronic activation of T-cell by a viral antigen or autoantigen in which case
26 LGL leukemia could be considered as an autoimmune disorder (Sokol and Loughran, 2006).
27 Existing epidemiologic studies of tetrachloroethylene exposure are simply not able to
28 inform human relevance examinations of rat MCL. Lymphoid tumor pathobiology in rats and
29 humans, its historical and current classification, and epidemiology, including observations in
30 tetrachloroehylene-exposed populations, have bearing on examination of the human relevance of
31 rat mononuclear cell leukemia. Important to any examination are the changes in diagnostic and
32 classification criteria of human lymphoid tumors and lack of data on molecular markers in the
33 tetrachloroethylene epidemiologic studies, as discussed above. Diagnostic and classification
34 criteria may not be uniform across studies and hinders comparison of consistency within
35 epidemiologic studies of lymphoid cancers and tetrachloroethylene exposure and, also, between
36 human and rat lymphoid tumor observations. Furthermore, adoption of consensus nomenclatures
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-152 DRAFT-DO NOT CITE OR QUOTE
-------
1 of human lymphoid tumors, i.e., the WHO scheme, for rats will facilitate cross-species
2 comparisons, as was recently conducted by the hematopathology subcommittee of the Mouse
3 Models for Human Cancers Consortium (Morse et al., 2002).
4
5 4.8.1.2.2. Esophageal cancer. Both cohort and case-control studies support an association
6 between tetrachloroethylene and excess risk of esophageal cancer. An overall excess in the
7 number of observed deaths was seen in the recent updates of the dry cleaner mortality studies
8 (Ruder et al., 2001; Blair et al., 2003): a total of 31 deaths as compared with 13.7 expected
9 deaths (RR = 2.3, 95% CI = 1.5-3.2; Table 4B-lb, Appendix 4B). This finding is the same as
10 that of Wartenberg et al. (2000), who examined a slightly different set of studies for esophageal
11 cancer. The recent PMR study by Walker et al. (1977) provides further support (aged <65 years
12 at time of death, PMR= 1.7, 95% CI= 1.1-2.5). Both Ruder et al. (2001) and Blair et al. (2003)
13 reported a similar magnitude of risk among workers employed before 1960 and after 1960. No
14 clear picture of increasing risk was seen in either study between duration of exposure and
15 esophageal cancer risk. Except for Boice et al. (1999), studies of the degreasers do not present
16 risks for esophageal cancer (Table 4B-2, Appendix 4B). Boice reported a risk of 1.5 (95% CI =
17 0.5-3.2), based on six deaths.
18 The cohort and PMR studies cannot directly address possible effects due to smoking or
19 alcohol, which are risk factors for the squamous cell histologic type of esophageal cancer. It is
20 not known whether elevated risk may reflect smoking and alcohol effects. Data from the
21 National Health Interview Survey (Nelson et al., 1994) suggest that the prevalence of smoking
22 among dry cleaners and laundry operators as equal to that of "blue collar" workers. Moreover,
23 Ruder et al. (2001) and Blair et al. (2003) note that the magnitude of the risks for several
24 smoking-related cancers was greater than could be explained by smoking alone, suggesting a
25 further contribution from another risk factor, such as occupational exposure.
26 The incidence of esophageal cancer is generally higher for nonCaucasan males than for
27 Caucasan males (Blot and McLaughlin, 1999; Brown et al., 2001). In contrast, Ruder et al.
28 (2001) observed similar SMRs for esophageal cancer across all race-sex groupings
29 (supplementary table at http://www.cdc.gov/niosh/dc-mort.html), providing further support for
30 occupational exposure as a risk factor. For these reasons, the observations in Ruder et al. (2001)
31 and Blair et al. (2003) together suggest that the excess esophageal cancer risks seen in the dry
32 cleaner studies cannot be entirely due to smoking, alcohol, or some other factor that may be
33 as soci ated with race.
34 Additionally, in the case-control study by Vaughan et al. (1997), the OR for cumulative
35 (ppm-yr) tetrachloroethylene exposure and squamous cell esophageal cancer, which were
36 adjusted for effects of both smoking and alcohol consumption, although imprecise (large
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-153 DRAFT-DO NOT CITE OR QUOTE
-------
1 confidence intervals), were significantly elevated (Table 4B-6, Appendix 4B). In fact, the
2 magnitude of risk for tetrachloroethylene exposure after adjustment in the statistical analysis for
3 smoking and alcohol consumption was larger for the effect of tetrachloroethylene exposure when
4 compared with the crude or unadjusted OR, suggesting that the association between occupational
5 exposure and esophageal cancer may be underestimated in those studies that could not control
6 for these factors.
7 Lacking information to classify job title for 25% and 19% of cases and controls,
8 respectively, Lynge et al. (2006) provides little weight for informing an examination of the
9 presence or absence of association between tetrachloroethylene and esophageal cancer.
10
11 4.8.1.2.3. Cervical cancer. There is some evidence for an excess in risk for cervical cancer
12 mortality. The total number of observed and expected numbers of cancers in the
13 tetrachloroethylene cohort mortality studies was 31 observed deaths versus 19 expected (RR =
14 1.6, 95% CI = 1.1-2.3; Table 4B-lb, Appendix 4B). An association with dry cleaning is
15 supported by an exposure-response trend in Ruder et al. (2001), although RRs for cervical cancer
16 mortality in Blair et al. (2003) did not appear to differ between subjects with medium/high
17 exposure and those with little or no exposure to tetrachloroethylene. The number of female
18 subjects in studies of workers exposed to tetrachloroethylene as a degreasing agent is few, with a
19 consequence of limited statistical power (Table 4B-2, Appendix 4B). Data availability on
20 socioeconomic and lifestyle factors in the dry cleaner studies precludes an evaluation of these
21 factors.
22
23 4.8.1.2.4. Suggestive evidence of cancer at other sites. More limited are the findings of excess
24 risks from cancers of the bladder, lung, pancreas, and small bowel.
25
26 4.8.1.2.4.1. Bladder cancer. The recent updates of two cohort mortality studies of dry cleaners
27 with tetrachloroethylene as the primary exposure (Blair et al., 2003; Ruder et al., 2001) provide
28 some evidence for an excess risk for bladder cancer mortality (Table 4B-lb, Appendix 4B).
29 Ruder et al. (2001) observed statistically significant differences in bladder cancer risk among the
30 entire cohort (SMR = 2.2, 95% CI = 1.1-4.1, 10 observed deaths). No deaths were observed
31 among subjects employed after 1960, a date corresponding with greater usage of
32 tetrachloroethylene. The magnitude of bladder cancer risk in Blair et al. (2003), on the other
33 hand, was similar regardless of level of exposure (little or no exposure vs. medium/high
34 exposure). The Nordic incidence studies are consistent with the mortality studies (Table 4B-la,
35 Appendix 4B). The study of Lynge et al. (2006), who examine bladder cancer incidence and
36 tetrachloroethylene exposure using a nested case-control approach in a cohort of Nordic dry
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-154 DRAFT-DO NOT CITE OR QUOTE
-------
1 cleaners and laundry workers, contributes little weight to causal evidence. The limitations of this
2 study preclude drawing conclusions on whether this study is suggestive of, or lack of, an
3 association between tetrachloroethylene and site-specific cancers such as bladder cancer. This
4 study is uninformative due to a high percentage of study subjects whose job title could not be
5 classified (for example, 16% of bladder cancer cases and controls) and potential bias resulting
6 from the large number of next-of-kin interviews.
7 Of the studies of workers exposed to tetrachloroethylene as a degreasing agent, only the
8 study by Boice et al. (1999) reports data for bladder cancer, based on two deaths (Table 4B-2,
9 Appendix 4B). Several case-control studies of bladder cancer (Table 4B-7, Appendix 4B) have
10 examined a job title as a dry cleaner or laundry worker and present risks adjusted for a number of
11 factors, including cigarette smoking, a known risk factor for this cancer. These studies also
12 provide weak support for an association: RRs ranged from 1.3 to 2.8, although increased risks
13 were generally not statistically significant.
14 Two population case-control studies examined tetrachloroethylene exposures specifically,
15 Pesch et al. (2000b) and Aschengrau et al. (1993). Pesch et al. (2000b) examined occupational
16 exposure to tetrachloroethylene using a job exposure and job task exposure matrix. Urothelial
17 cancer cases (a category that includes cancer of the urinary bladder, ureter, and renal pelvis)
18 were histologically confirmed. A statistically significant excess risk (OR) was reported in both
19 exposure assessment methods for males with substantial exposure to tetrachloroethylene in
20 analyses that adjusted for age, study center, and smoking.
21 Aschengrau et al. (1993) reported an adjusted OR of 4.9 (95% CI = 0.7-25.1) with high
22 exposure to tetrachloroethylene in drinking water (Table 4B-13, Appendix 4B). Strengths of this
23 study are the use of exposure modeling to reconstruct tetrachloroethylene delivery to a home and
24 adjustment of ORs for sex, age at diagnosis, vital status, educational level, and smoking.
25
26 4.8.1.2.4.2. Lung cancer. Lung cancer risk was elevated in the mortality studies by Blair et al.
27 (2003) and Ruder et al. (2001), where the total number of observed deaths was 144, with 106.6
28 deaths expected (RR = 1.4, 95% CI = 1.1-1.6; Table 4B-lb, Appendix 4B); in the incidence
29 study by Lynge and Thygesen (1990; RR = 1.2, 95% CI = 0.9-1.6) 60 cases among males and
30 females, 49.4 expected cases (Table 4B-la, Appendix 4B); and in the degreaser studies, a total of
31 51 cases observed with 43 expected; RR = 1.2, 95% CI = 0.9-1.6; Table 4B-2, Appendix 4B).
32 The possible effect of smoking cannot be examined in the cohort studies.
33 Two case-control studies (Pohlabeln et al. 2000; Brownson et al., 1993) examined the
34 association between occupational risk factors and lung cancer among nonsmokers (Table 4B-8,
35 Appendix 4B). Both studies reported an association between lung cancer (among nonsmokers)
36 and dry cleaning work. Additionally, the case-control study by Paulu et al. (1999) provides
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-155 DRAFT-DO NOT CITE OR QUOTE
-------
1 further support for an association between tetrachloroethylene exposure and lung cancer (Table
2 4B-13, Appendix 4B). This study examined oral (drinking water) exposure to
3 tetrachloroethylene, and the statistical analysis adjusted for both active and passive smoking.
4
5 4.8.2. Animal Studies
6 In addition to the toxic effects in animals already mentioned (in liver, kidney, nervous
7 system, and developmental/reproductive system), effects have also been reported on cardiac
8 function, immunosuppression, and cancer at other sites.
9
10 4.8.2.1. Noncancer Effects
11 4.8.2.1.1 Whole animal toxicity. Hayes et al. (1986) administered tetrachloroethylene to SD
12 rats in drinking water at doses of 0, 14, 400, and 1,400 mg/kg-day for 90 days and observed the
13 body weights of all animals weekly throughout the experiment; and the liver, kidney, and brain
14 weights of all animals at necropsy. They also performed measurements of hematological and
15 serum chemistry parameters on 10 animals per group at the end of the period. They found
16 significant (p < 0.05) decrements in body weight gain in males at 1,400 mg/kg-day and in
17 females at 400 and 1,400 mg/kg-day. No effects that could be attributed to administered dosing
18 were observed in hematology, serum chemistry, urinalysis, mortality, or organ weights. The
19 body-weight-gain decrements in this experiment signify a LOAEL of 400 mg/kg-day and a
20 NOAEL of 14 mg/kg-day.
21
22 4.8.2.1.2. Cardiac toxicity. Kobayashi et al. (1982) treated animals using intravenous injections
23 of tetrachloroethylene. In the animals examined (rabbits, cats, and dogs), tetrachloroethylene
24 enhanced the vulnerability of the ventricles to epinepherine-induced arrhythmias. The threshold
25 doses were 10, 24, and 13 mg/kg in rabbits, cats and dogs, respectively.
26 Cardiac effects of some tetrachloroethylene metabolites have been examined in animals.
27 As mentioned in Section 4.6.4, Smith et al. (1989) and Johnson et al. (1998) observed cardiac
28 anomalies in rat fetuses after exposure of pregnant rats to TCA. Epstein et al. (1992) saw cardiac
29 defects in rat fetuses after exposure to DC A. This work indicated a developmental LOAEL of
30 1,900 mg/kg-day DCA. DCA has also been shown to concentrate in rat myocardial
31 mitochondria (Kerbey et al., 1976). More research into cardiac toxicity resulting from exposures
32 to tetrachloroethylene and its metabolites is needed to fully characterize possible adverse cardiac
33 effects.
34
35 4.8.2.1.3. Immunotoxicity. The animal evidence for immunotoxicity following exposure to
36 tetrachloroethylene is also very limited. These studies consist of a mixed solvent exposures and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-156 DRAFT-DO NOT CITE OR QUOTE
-------
1 some inhalation and oral studies in which experimental animals were dosed with
2 tetrachloroethylene alone.
3 Immune systems parameters were altered in a mouse study (female B6C3F1)
4 administered tetracholroethylene (maximum concentration 6.8 ppm) along with a mixture 24
5 contaminants frequently found in ground water near superfund sites. Exposure lasted 14 or 90
6 days and mice were sacrificed to assess immune system parameters. Evidence of
7 immunosuppression was seen, with a dose related decrease in antibody response to sheep red
8 blood cells and decreased host resistance to following challenge to Plasmodium yoelli. There
9 were no changes in lymphocyte number, T-cell subpopulations, NK cell activity, or in challenge
10 listeria monocytgens or PYB6 tumor cells. While these findings may be attributed to
11 B-cell/humoral immunity these effects cannot be attributed to tetrachloroethylene alone
12 (Germolecetal., 1989).
13 In another inhalation study, mice were given a single exposure of 170 mg/m3
14 tetrachloroethylene (50 ppm) for 3 hrs and then challenged with Klebsiella pneumoniae; an
15 increase in streptococcal pneumonia was observed (Aranyi et al., 1986). Interpretation of the
16 significance of these findings is confounded with a high degree of mortality in the control group.
17 In a study by Hanioka et al. (1995), atrophy of the spleen and thymus was observed in
18 rats receiving 2,000 mg/kg/d tetrachloroethylene via corn oil gavage for 5 days. No effect was
19 seen in the 1,000 mg/kg/d group. In a 14-day corn oil gavage (1,000 mg/kg/d) study of
20 tetrachloroethylene, no effects were observed on thymus and spleen weights of adult rats at dose
21 that produced liver toxicity (Berman et al., 1995). Another study employed 3 daily ip doses of
22 tetrachloroethylene to mice (Schlichting et al., 1992). No effects were observed on ex vivo
23 natural killer cell activity or humoral responses of T-cells to exogenous mitogens.
24 A series of experiments in the lupus-prone MRL +/+ mice examined the effect of
25 trichloroethylene on the expression of features of lupus (Griffin et al., 2000a; Gilbert et al., 2004).
26 Activation and expansion of CD4+ T-cells has been demonstrated, through a mechanism that
27 appears to be mediated through the CYP2E1 metabolism of trichloroethylene and inhibition of
28 FasL expression on the surface of the CD4+ T-cells (Griffin et al., 2000b; Blossom et al., 2004;
29 Blossom and Gilbert, 2006). Trichloroethylene exposure via drinking water was also shown to
30 induce an autoimmune hepatitis, characterized by mononuclear infiltration around the portal vein,
31 in the MRL +/+ mice (Griffin et al., 2000c). This evidence of immunological alterations
32 following trichloroethylene exposure provides suggestive evidence for perchloroethylene, a
33 halogenated solvent which shares some common metabolites with trichloroethylene. To date,
34 similar studies have not been conducted with other solvents, so the extent to which these findings
35 pertain to tetrachloroethylene is not known.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-157 DRAFT-DO NOT CITE OR QUOTE
-------
1 Additional data from inhalation, oral, and dermal exposures of different durations are
2 needed to assess the potential immunotoxicity of tetrachloroethylene along multiple dimensions,
3 including immunosuppression, autoimmunity, and allergic sensitization. This lack of data taken
4 together with the concern that other structurally related solvents have been associated with
5 immunotoxicity contributes to uncertainty in the database for tetrachloroethylene.
6
7 4.8.2.2. Cancer Effects
8 4.8.2.2.1. Mononuclear cell leukemia in rats. NTP (1986a) reported that the chronic inhalation
9 administration of tetrachloroethylene to male and female F344/N rats at concentration levels of 0,
10 200, and 400 ppm caused positive trends in the incidence of MCL in both sexes. The incidence
1 1 data are shown in Table 5-8 (Chapter 5). For males, there was a statistically significant trend (p
12 = 0.004), and for females the trend was marginally significant (p = 0.053). Pairwise comparisons
13 of tumor incidences in dosed and control groups of males (life table analysis) disclosed
14 statistically significant increases in both the low- and high-dose groups. Analysis of the data for
15 female rats revealed a significant increase in the low-dose group and a marginally significant
16 increase in the high-dose group. Interpretation of these data is somewhat clouded by the fact that
17 overall incidences of MCL in the concurrent chamber control groups were high relative to
18 historical chamber control groups at the performing laboratory (males 28/50 [56%] vs. 1 17/250
19 [47%]; females 18/50 [36%] vs. 73/249 [29%]). The concurrent control group rates were also
20 higher than the NTP program historical rate for untreated control groups (males 583/1,977
21 [29%]; females 375/2,021 [18%]).
22 Because of these factors, NTP conducted supplemental analyses of the progression of the
23 disease, the effect of tetrachloroethylene on the time of onset of advanced MCL, and the
24 contribution of MCL to early deaths in control and dosed animals. The results of these
25 supplemental analyses showed the following:
28 • In both males and females, tetrachloroethylene produced a dose-related increase in the
29 severity of MCL.
30
31 • Tetrachloroethylene exposure significantly shortened the time to onset of MCL in female
32 rats.
33
34 • Although there was no remarkable effect of tetrachloroethylene exposure on survival of
35 female rats, there was an increased incidence of advanced MCL in female rats that died
36 before the scheduled termination of the study. Thus, a more appropriate statistical
37 analysis was conducted in which only the incidences of advanced MCL in rats were
38 considered. Significantly positive trends and significant increases in the incidences of
39 advanced MCL were observed in both male and female rats in the high-dose groups.
40
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-158 DRAFT-DO NOT CITE OR QUOTE
-------
1 In 1987, the U.S. EPA's Science Advisory Board took exception to the use of these
2 special analyses because they did not represent generally accepted approaches to evaluating
3 increased incidences of MCL. According to the NTP report, however, the interpretation of MCL
4 incidences in the tetrachloroethylene study was based on standard methods of data evaluation
5 (NTP, 1986a). The special analyses were conducted to support rather than to establish the
6 interpretation.
7 The Japan bioassay (JISA, 1993) in F344/DuCrj rats exposed for 104 weeks at
8 concentrations of 0, 50, 200, and 600 ppm also reported clearly significant trends in MCL in
9 males. In females, MCL showed a marginally significant trend with dose. These data are also
10 shown in Table 5-8 (Chapter 5). In this assay, the control incidences for both males and females
11 were also higher than those for historical controls (for males, study control incidence was 11/50,
12 [22%], whereas for the historical controls it was 147/1,149 [13%]). This higher incidence in
13 concurrent controls versus historical controls also occurred in the NTP assay. The historical
14 control incidence data for the Japan rat studies are shown in Table 5-10 (Chapter 5). The Japan
15 bioassay report did not include an analysis of the tumor latency.
16
17 4.8.2.2.1.1. Discussion of issues associated with rat mononuclear cell leukemia (MCL).
18 Under the conditions of the bioassay s, a carcinogenic effect of tetrachloroethylene in male and
19 female rats was evidenced by significant increases of MCL in both sexes. However, the
20 reliability of MCL in the rat in predicting human carcinogenic risk associated with
21 tetrachloroethylene exposure has been questioned for several reasons, such as high spontaneous
22 background incidences, use of special supplemental analyses to aid in data interpretation, and the
23 relevance of MCL in F344/N rats to human hazard. Some of the issues have been reviewed by
24 Caldwell (1999) and others.
25
26 4.8.2.2.1.2. Background. Lymphohematopoietic neoplasms, leukemias, and lymphomas
27 represent uncontrolled proliferation or clonal expansion of bone marrow or lymphoid cells that
28 can no longer differentiate to mature blood cells (U.S. EPA, 1997a; Sawyers et al., 1991; Nowell,
29 1991). Like other cancers, they are thought to develop via a multi-step process in the
30 transformation of a normal cell to a malignant cell. Although rodent models—especially mouse
31 models—are considered to be highly relevant for understanding most aspects of hematopoiesis
32 (Bagby, 1994), several differences exist between humans and rodents. Differences in cell
33 composition, number, and proliferation rates and organization of the stem-cell compartment
34 likely influence the hematotoxic and carcinogenic effects observed in human and rodent systems
35 following exposure to carcinogenic chemicals (U.S. EPA, 1997b).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-159 DRAFT-DO NOT CITE OR QUOTE
-------
1 In adult humans, hematopoiesis occurs in the bone medullary spaces, with extramedullary
2 hematopoiesis occurring in the spleen, liver, and lymph nodes only under stress conditions. In
3 rodents, however, hematopoietic cells are commonly found in the spleen and, to some extent, the
4 liver. The primary type of lymphohematopoietic cancer induced by chemicals in humans is
5 myeloid leukemia; however, the induction of lymphoma has been associated with
6 immunosuppressive agents. In contrast, lymphohematopoietic tumors in rats and mice originate
7 primarily in lymphoid tissue.
8
9 4.8.2.2.1.3. Issues. The usefulness of increased incidences of MCL in predicting human
10 carcinogenic risk associated with exposure to tetrachloroethylene has been questioned on several
11 grounds, and these issues are discussed below.
12 MCL is a recognized as a common, spontaneously occurring neoplasm in F344 rats, and
13 its rate of appearance in historical control groups is highly variable. High-incidence MCL occurs
14 only in the F344 rat strain and not in mice. For this reason, Caldwell (1999), for example, has
15 stated that marginal increases in incidences are of questionable biological significance. High and
16 variable control incidence is also an issue with liver tumors in mice, and it has always been a
17 source of uncertainty in risk assessments.
18 Although the occurrence in a single strain indicates a genetic susceptibility of these rats
19 to spontaneous MCL, the occurrence in humans of a similar genetic susceptibility is by no means
20 ruled out. Humans are more genetically heterogeneous than are inbred rat strains, and some
21 individuals could potentially possess the same inherited susceptibility that is exhibited in F344
22 rats.
23 Another issue is the pathobiology of MCL. Some scientists believe it is too poorly
24 understood to allow the tumors to be used to determine human health risk. However, MCL is a
25 relatively well defined and well understood rodent neoplasm that is characterized by infiltration
26 of pleomorphic blastlike mononuclear cells in numerous organs. The disease per se, which is
27 splenic in origin but later infiltrates the liver, lung, bone marrow, lymph nodes, and other organs,
28 is readily and unequivocally diagnosed by standard histopathological techniques. MCL has also
29 been described as large, granular, lymphocyte leukemia and is known to be a rapidly
30 progressing and fatal neoplasm whose incidence is age related. The tumor is transplantable; its
31 eti ol ogi cal factor i s unknown.
32 The similarities and differences in the tissue of origin, precursor cell line, and pathologic
33 characteristics of rat MCL and human lymphoid cancers have been reviewed by Caldwell (1999)
34 Ishmael and Dugard (2006) and EPA (U.S. EPA, 1997b). Both diseases result from abnormal
3 5 development and maturation of lymphocytes. It is known that the tissue of origin in rats is the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-160 DRAFT-DO NOT CITE OR QUOTE
-------
1 spleen and that lymphomas and leukemias in humans originate in the bone marrow. The
2 precursor cell line is not known for either rat MCL or the human lymphoid cancers.
3 Large granular lymphocyte cells exist in humans that are morphologically, biochemically,
4 and functionally similar to the cells involved in MCL in the F344 rat (Stromberg, 1985). The
5 pathological characteristics of rat MCL are similar in some respects to one of the human T-cell
6 leukemias (Caldwell, 1999), and some investigators believe MCL serves as a model for T-cell
7 leukemia (Stromberg, 1985). However, discounting a rodent neoplasm simply because it has no
8 exact human counterpart is not a scientifically defensible position. Strict site concordance is not
9 a requirement for relevancy in extrapolation of hazard potential. For example, many aromatic
10 amines are probable bladder carcinogens in humans but are likely to produce Zymbal gland
11 tumors in rats, for which there is no analogous organ in humans.
12 The specific mechanism of leukemogenesis in rats is not understood, but neither is it well
13 understood in humans. A possible link to MO A for tetrachloroethylene-induced MCL in rats
14 comes from early reports of toxicity of cysteine S-conjugates, where S-(l,2,-dichlorovinyl)-L-
15 cysteine, the trichloroethylene metabolite, was implicated in induction of aplastic anemia and
16 marked biochemical alteration of DNA in bone marrow, lymph nodes, and thymus in calves
17 (Bhattacharya and Schultze, 1971, 1972).
18 As discussed elsewhere, the GSH conjugate of tetrachloroethylene is hydrolyzed in the
19 kidney to the comparable cysteine S-conjugate, a compound that can be cleaved to form a
20 mutagen. Humans as well as rodents activate the conjugate via FMO3, CYP3 A and/or the beta
21 lyase pathway. Thus, the possibility exists that the tetrachloroethylene S-conjugate
22 S-(l,2,2-trichlorovinyl)-L-cysteine may be involved in inducing leukemia in rats and may have
23 the potential to produce blood dyscrasias in humans as well. However, a recent report of a study
24 in which TCVC was given to two calves did not find that it produced bone marrow injury in
25 these animals at dose levels comparable to those of DCVC that caused bone marrow toxicity in
26 calves in the same study (Lock et al., 1996).
27
28 4.8.2.2.1.4. Summary and conclusions regarding the leukemia finding in rats. Leukemia
29 incidences were significantly increased in both male and female rats, in spite of high
30 spontaneous background incidences. Although caution is recommended with regard to
31 interpreting results for species that have high spontaneous incidences for any specific tumor, it is
32 also important to remember that high spontaneous incidences of lymphohematopoietic cancer are
33 not unique to rodent strains—e.g., high incidences of leukemia occur in certain genetically
34 susceptible humans as well (U.S. EPA, 1997a). In addition to causing increased tumor
35 incidences, tetrachloroethylene caused a dose-related increase in severity of MCL in both sexes
36 and shortened the time to tumor in female rats in one of the chronic bioassays.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-161 DRAFT-DO NOT CITE OR QUOTE
-------
1 The principal type of chemically induced lymphohematopoietic cancer in humans is
2 myeloid leukemia, with the exception being the lymphohematopoietic cancers induced by
3 immunosuppressive agents, which are usually associated with development of lymphomas.
4 There is some epidemiologic evidence that occupational exposure to tetrachloroethylene is
5 associated with NHL; thus, lymphohematopoietic cancers are observed in both rats and humans.
6 Although the specific etiology of NHL is unknown at the present time, it is thought likely to be
7 related to imbalances or disturbances in the immune system. Two potent immunosuppressive
8 chemotherapeutic agents, cyclosporin and azathioprine, are associated with NHL.
9 If a chemical produces a significant increase in MCL in the F344 rat, the finding cannot
10 be ignored. The observation of a significant increase of MCL in rats signals that the chemical
11 may possibly cause similar or different types of tumors in humans.
12
13 4.8.2.2.2. Tumors at other sites in animal bioassays. In the NTP inhalation study, an elevated
14 incidence of rare brain gliomas in rats was observed. In males in the control and the mid- and
15 high-tetrachloroethylene concentration groups, the incidences were 1/50, 0/50, 4/50, respectively,
16 and there was a significantly positive dose-related trend by the life table test but not by the
17 incidental tumor trend test. In females the incidence was 1/50, 0/50, 2/50. These are rare tumors
18 in NTP rat bioassays; the historical control incidence for males and females combined in the
19 laboratory was 2/247 (0.8%), and in the overall NTP program it was 6/1,971 (0.3%). Because
20 these tumors had not been observed in the previous NTP studies of trichloroethylene (NTP,
21 1990b) or pentachloroethane (NTP, 1983), and because they appeared in the untreated groups,
22 NTP concluded that they were not related to tetrachloroethylene exposure.
23 On the other hand, the tumors in the high-dose males occurred slightly earlier (88, 96,
24 102, and 103 weeks) than in the control group (99 weeks); in the high-dose females they
25 occurred more definitively earlier (75 and 78 weeks in the high-dose group vs. 104 weeks in the
26 control group). In addition, a Fisher's exact test of the significance of combined male and
27 female incidences in the tetrachloroethylene-treated animals shows significance with respect to
28 both lab and NTP program historical data, whereas the control incidence is not significant with
29 respect to either of the historical data sets. Therefore, although the data showing that
30 tetrachloroethylene is causing brain gliomas in rats is not strong, it is suggestive.
31 The incidence of interstitial testicular tumors in male F344 rats treated with
32 tetrachloroethylene was significantly higher than in controls in the study. However, it is
33 common in control rats (the NTP program historical control incidence was 1,729/1,949 [89%]),
34 and the incidence in treated animals was not higher than in historical laboratory or historical
35 program controls. Also, the combined incidence of hyperplasia and tumors was not significant
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-162 DRAFT-DO NOT CITE OR QUOTE
-------
1 with respect to that of the study controls. For these reasons, the NTP concluded that the
2 marginally higher incidence was not related to tetrachloroethylene exposure.
3 Hemangioendotheliomas in the liver and spleen in male mice were observed in the Japan
4 bioassay. These were mentioned in Section 4.4.2.2 in connection with liver cancer, and the data
5 are given in Section 5.2.2. These tumors were not observed in the NTP studies.
6
7 4.8.3. Summary of Immunotoxicologic Effects in Humans and Animals and Potential Mode
8 of Action
9 The epidemiologic evidence pertaining to tetrachloroethylene exposure in relation to
10 immune-related conditions, is limited (Table 4-11). Estimated associations based on population-
11 based case-control studies of various systemic autoimmune diseases have low statistical power
12 (and thus are highly imprecise) because of the relatively low prevalence of occupational
13 exposure to this chemical in the general population. Exposure misclassification is likely in these
14 studies, and in general would be expected to result in an attenuation of the observed effect. To
15 date, no relevant occupational cohorts have examined the risk of developing these diseases. The
16 few studies that have been conducted related to allergy and hypersensitivity have used direct
17 measures of tetrachloroethylene and other compounds in ambient or exhaled air samples, but
18 these studies do not provide much evidence of an adverse effect relating to tetrachloroethylene
19 exposures. However, the only study of asthma severity was quite small (n = 21), and our
20 understanding of the impact of changes in specific T-cell subsets (i.e., interferon-y) is currently
21 limited.
22 The immune system is clearly crucial to the prevention of disease caused by infectious
23 agents. Alterations in immune function may also contribute to the development of non-
24 infectious diseases including cancer, autoimmune diseases, and hypersensitivity disorders. Many
25 immunosuppressive agents are human carcinogens (Tomatis et al., 1989), and as described in the
26 previous section, inhibition of the natural immune surveillance could play a role in the
27 hepatocarcinogenic properties of tetrachloroethylene. The numerous immune-mediated activities
28 of relevance to the pathogenesis of a variety of disease include the binding and processing of
29 antigens by B-cells and T-cells, alteration and loss of tolerance to self-antigens, defects in
30 apoptosis which may effect the clearance of cells, and the secretion of pro- and anti-
31 inflammatory cytokines (Seliger, 2005; Ayensu et al., 2004).
32 Binding of reactive compounds to cellular macromolecules has been proposed as an
33 important step in the pathogenesis of several diseases, both for cancer (Hinson and Roberts,
34 1992) and for chemically induced autoimmune disease (Uetrecht et al., 1988). The modification
35 of proteins may lead to more immunoreactive products, and these may lead to the development
36 of autoantibodies and the cellular damage seen in alcoholic liver disease and in autoimmune
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-163 DRAFT-DO NOT CITE OR QUOTE
-------
1 diseases (e.g., systemic lupus erythematosus, rheumatoid arthritis; Kurien et al., 2006). Reactive
2 metabolites of tetrachloroethylene have been shown to bind irreversibly to cellular
3 macromolecules in vitro (e.g., Costa and Ivanetich, 1980) and in vivo (Pegg et al., 1979;
4 Schumann et al., 1980). Binding occurs proportionally to the amount metabolized, and
5 metabolism is proportional to toxicity (e.g., Buben and O'Flaherty, 1985). Several published
6 studies have demonstrated formation of trichloroacylated protein adducts, for example, in liver
7 and kidney of rats (Birner et al., 1994) and in plasma of rats and humans (Pahler et al., 1999)
8 following exposures to tetrachloroethylene. Another example is the detection of
9 trichloroacetylated protein adducts formed in the liver of MRL-lpr/lpr and MRL +/+ mice treated
10 with tetrachloroethylene (Green et al., 2001). These strains are "lupus-prone" mice, because of
11 their genetic susceptibility to the development of lupus-like disease. Further studies designed to
12 identify the adducted proteins may help to elucidate an MO A for tetrachloroethylene-induced
13 autoimmune response, which, in turn, may be related to cancer-causing activity (not clear how
14 this autoimmune response is related to cancer).
15 Apoptosis, a form of natural cell death differing from necrosis, is essential for proper
16 functioning of the immune system and clearance of tumor cells. Apoptosis also plays an
17 important role in the pathogenesis of autoimmune diseases. Genetic or environmental exposures
18 leading to increased apoptosis or to decreased clearance of apoptotic debris may stimulate the
19 production of autoantibodies directed against intracellular antigens (Gaipl et al., 2006). The
20 production of free radicals, increased lipid peroxidation, and increased apoptosis was
21 demonstrated in a recent study using human lung adenocarcinoma cells treated with
22 tetrachloroethylene (Chen et al., 2002b). Alternatively, the inhibition of apoptosis of CD4+
23 T-cells may also effect the development of autoimmune disorders, as demonstrated by the recent
24 studies of trichloroethylene metabolites (Blossom and Gilbert, 2006). Thus, with respect to
25 autoimmune diseases, as well as neurodegenerative and other diseases, the strict regulation of
26 apoptosis signalling is crucial (Ethell and Buhler, 2003; Schattenberg et al., 2006).
27
28 4.9. SUSCEPTIBLE POPULATIONS
29 Variation in response among segments of the population may be due to age, genetics, and
30 ethnicity, as well as to differences in lifestyle, nutrition, and disease status. These could be
31 potential risk factors that play an important role in determining an individual's susceptibility and
32 sensitivity to chemical exposures. Studies on tetrachloroethylene toxicity and MOA in relation
33 to some of these risk factors are discussed below.
34
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-164 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.9.1. Life Stages
2 Individuals of different life stages are physiologically, anatomically, and biochemically
3 different. Early and later life stages differ greatly from mid-life stages in body composition,
4 organ function, and many other physiological parameters that can influence the absorption,
5 distribution, metabolism, and elimination of chemicals and their metabolites from the body (ILSI,
6 1992). The limited data on tetrachloroethylene exposure suggest that these subpopulations—
7 particularly individuals in early life stages—may have greater susceptibility than does the
8 general population. This section presents and evaluates the pertinent published literature
9 available to assess how individuals of differing life stages may respond differently to
10 tetrachl oroethy 1 ene.
11
12 4.9.1.1. Life Stage-Specific Exposures
13 Section 2.2 describes the various exposure routes of concern for tetrachloroethylene. For
14 all postnatal life stages, the primary exposure routes of concern include inhalation (see Section
15 2.2.1) and contaminated water (see Section 2.2.2). In addition, certain exposure pathways to
16 tetrachloroethylene are unique to early life stages, such as through placental transfer or via breast
17 milk ingestion, or may be increased during early or later life stages. In utero, there is biological
18 plausibility of transfer of tetrachloroethylene across the human placental barrier as seen in
19 rodents (Ghantous et al., 1986; Szakmary et al., 1997). Fetal blood concentrations have been
20 modeled for human exposure (Gentry et al., 2003).
21 For infants, a unique exposure route of concern is ingestion of breast milk (see Section
22 2.2.4). The breast milk of one woman was found to contain 10 mg/L tetrachloroethylene 1 hr
23 following a visit to her spouse working at a dry cleaning establishment, dropping to 3 mg/L after
24 24 hrs (Bagnell and Ellenberger, 1977). Tetrachloroethylene has also been measured in the
25 breast milk of a woman living in an apartment located in a building housing a dry cleaning
26 facility (Schreiber et al., 2002; NYS DOH, 2005b). PBPK models have been used to estimate
27 the dose a nursing infant might receive from an exposed mother's breast milk (Gentry et al.,
28 2003; Schreiber, 1993). Using different exposure scenarios, Schrieber predicted that breast milk
29 concentrations could range from 1.5 ug/L for a typical residential scenario, 16-3,000 ug/L for a
30 residential scenario near a dry cleaner, to 857-8,440 ug/L for an occupational scenario.
31 Assuming that a 7.2 kg infant ingests 700 mL of breast milk per day, Schreiber estimated dose to
32 the infant could range from 0.0001 to 0.82 mg/kg/day (Schreiber, 1993). Therefore, the dose an
33 infant could receive through breast milk may exceed the previous EPA RfD level (0.01 mg/kg-
34 day). Byczkowski and Fisher (1995) refined the approach used by Schrieber (1993) and found
35 that with the same exposure conditions, the results predicted lower doses to the infant (0.0009-
36 0.202 mg/kg/day).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-165 DRAFT-DO NOT CITE OR QUOTE
-------
1 For infants on formula, ingestion of contaminated water may be of concern. Taking into
2 account tetrachloroethylene volatilization in boiling water, Letkiewicz et al. (1982) estimated
3 that 22% of formula-fed infants received fluids contaminated with tetrachloroethylene levels
4 found in the water supply. Data showed that about 11% (0.5 x 22%) of formula-fed infants
5 could receive an increased exposure as compared with adults on a mg/kg basis through drinking
6 contaminated water.
7 Dairy products have been found to have elevated concentrations of tetrachloroethylene
8 (see Section 2.2.3), and children ingest larger quantities of dairy products compared to adults
9 (NRC, 1993). Therefore, there may be concern for ingestion of contaminated dairy products in
10 early life stages, although this exposure route for tetrachloroethylene has not been well
11 characterized for any life stage.
12 Inhalation exposures may be increased for both early and later life stages compared to
13 adults, since children and the elderly have increased ventilation rates per kg body weight
14 compared to adults (NRC, 1993; U.S. EPA, 2006) and since they spend the majority of their time
15 indoors (NRC, 1993; U.S. EPA, 2002), where increased concentrations of tetrachloroethylene
16 have been found (U.S. EPA, 2001b). Section 2.2.1 describes increased indoor air concentrations
17 measured inside apartments containing dry cleaned clothing (Thomas et al., 1991), in apartments
18 above or adjacent to dry cleaners (Altmann et al., 1995; Chien, 1997; Garetano and Gochfeld,
19 2000; McDermott et al., 2005; Schreiber et al., 1993, 2002), and in daycare centers adjacent to
20 dry cleaners (NYS DOH, 2005b, c). In addition, inhalation may also occur during showering or
21 bathing as dissolved tetrachloroethylene in the warm tap water is volatilized (Rao and Brown,
22 1993).
23 Dermal exposures may be increased for both early and later life stages compared to adults,
24 since infants have increased skin area per kg body weight (NRC, 1993) and the elderly
25 experience changes in permeability (U.S. EPA, 2006). Dermal exposure may occur in an
26 occupational setting from direct handling of tetrachloroethylene or in a residential setting from
27 showering, bathing, or swimming in contaminated water (Rao and Brown, 1993; U.S. EPA 2001).
28 (see Section 2.2.2) While dermal exposure is generally not considered a major route of exposure,
29 this route of exposure is not well characterized for early life stages (prenatal or postnatal), or
30 later life stages.
31
32 4.9.1.2. Early Life Stage Effects
33 Although limited data exist on tetrachloroethylene toxicity as it relates to early life stages,
34 there is enough information to discuss the qualitative differences. In addition to the evidence
35 described below, Section 4.7 contains information on both human and animal evidence for
36 reproductive and developmental outcomes such as spontaneous abortion/fetal loss, low birth
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-166 DRAFT-DO NOT CITE OR QUOTE
-------
1 weight, IUGR, SGA, congenital abnormalities, sperm quality, developmental delays, and
2 behavioral changes. Together, Sections 4.4 on liver toxicity, 4.5 on kidney toxicity, 4.6 on
3 neurotoxicity, and 4.8 on toxic effects in other organ systems characterize a wide array of
4 postnatal developmental effects.
5
6 4.9.1.2.1. Differential effects in early life stages. Preconception exposure has been associated
7 with altered semen quality in occupationally exposed humans (Eskenazi et al., 1991b; Rachootin
8 and Olsen, 1983; Sallmen et al., 1995), as well as in mice, but not in rats (Beliles et al., 1980).
9 Additionally, reduced testes weight was seen in rats after inhalation exposure (Tinston, 1994).
10 A number of human studies have shown spontaneous abortion among women employed
11 as laundry workers (Hemminki et al., 1980) or dry cleaners (Ahlborg, 1990; Bosco et al., 1987;
12 Doyle et al., 1997; Olsen et al., 1990; Kyyronen et al., 1989), married to men employed as dry
13 cleaners (Taskinen et al., 1989; Eskenazi et al., 1991a), exposed to other occupational solvents
14 (Windham et al., 1991; Lindbohm et al., 1990), or living in residences receiving contaminated
15 water (Lagakos et al., 1986; Bove et al., 1995; ATSDR, 1998). However, another study
16 population of women working as laundry or dry cleaning workers did not experience
17 spontaneous abortion (McDonald et al., 1986, 1987). Reduced fertility has been seen in women
18 and men exposed occupationally (Eskenazi et al, 1991a, b; Sallmen et al., 1995; Rachootin and
19 Olsen, 1983).
20 In the literature concerning animal studies, there is evidence of increased pre- and post-
21 implantation loss (Szakmary et al., 1997) and increased resorption of rodent pups after maternal
22 inhalation (Schwetz et al., 1975; Szakmary et al., 1997), reduction in litter size after maternal
23 gavage (Narotsky and Kavlock, 1995), and litters with dead pups (Tinston, 1994). However,
24 fetal loss was not seen in other animal studies (Carney et al., 2006; Hardin et al., 1981). In vitro
25 studies show decreased fertilized oocytes (Berger and Horner, 2003), and increased mortality,
26 malformations, and delayed growth and differentiation of embryos (Saillenfait et al., 1995) when
27 exposed to tetrachloroethylene.
28 After residential exposure to contaminated water, birth outcomes related to in utero
29 exposure in humans include perinatal death, birth defects (eye and ear anomalies, and CNS/oral
30 cleft anomalies; Lagakos et al., 1986), and IUGR (Windham et al., 1991) or SGA (Sonnenfeld et
31 al., 2001). Also, childhood leukemia has been associated with in utero exposure to
32 tetrachloroethylene due to maternal ingestion of contaminated water (MA DPH, 1997; see
33 Section 4.9.1.2.4). The study population reported in Sonnenfeld et al., (2001) is currently being
34 further examined to determine any association between maternal ingestion of contaminated water
35 and the incidences of birth defects (e.g., neural tube defects and oral clefts; ATSDR, 2003).
36 However, other human studies have not shown effects after occupational exposure for other birth
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-167 DRAFT-DO NOT CITE OR QUOTE
-------
1 outcomes, such as stillbirths, congenital anomalies, or decreased birth weight (Bove et al., 1992;
2 Olsen et al., 1990, Kyyronen et al., 1989, Taskinen et al., 1989, Windham et al., 1991). In
3 addition, preconception or prenatal exposure may lead to other latent outcomes such as an
4 increased risk for schizophrenia as seen in a large prospective study of parental occupational
5 exposure to tetrachloroethylene (Perrin et al., 2007).
6 Birth outcomes in animals exposed to tetrachloroethylene in utero include skeletal
7 retardation and malformations, decreased body weight and weight gain, altered brain fatty acid
8 composition, and developmental delay. Skeletal retardation and malformations were increased
9 in rodent pups after maternal inhalation exposure 7 days per week (Carney et al., 2006;
10 Szakmary et al., 1997), but no birth defects were seen in other animal studies using a similar
11 dose but for 5 days per week (Hardin et al., 1981). Exposure resulted in decreased fetal or pup
12 body weights (Carney et al., 2006; Szakmary et al., 1997; Tinston, 1994), along with reduction in
13 weight gain (Nelson et al., 1980). Also, Kyrklund and Haglid (1991) noticed slightly altered
14 brain fatty acid composition after maternal exposure during pregnancy was also noted (Kyrklund
15 and Haglid, 1991). Developmental delay was seen in rat offspring after maternal exposure
16 during pregnancy (Nelson et al., 1980). In addition, cardiac anomalies have been seen in rats
17 exposed to the metabolites TCA (Smith et al., 1989; Johnson et al., 1998) and DC A (Epstein et
18 al., 1992; see Sections 4.6.2, 4.7.2, and 4.8.2).
19 Neurotoxicological effects in children have been reported after low exposure levels to
20 tetrachloroethylene (see Section 4.6 and Table 4-4). While other neurotoxic effects are seen in
21 adults (see Table 4-5), decreased VCS has been the main observation in children, including in
22 those who resided in an apartment building with a dry cleaning establishment (Schreiber et al.,
23 2002; NYS DOH, 2005a). Children who attended a day care center adjacent to a dry cleaner
24 were too young to take a visual exam given to the adult workers that demonstrated decreased
25 VCS. Other neuropsychological tests conducted on the children attending this day care center 5
26 weeks after exposure ceased did not consistently find any effect on cognition or behavior (NYS
27 DOH, 2005b). A follow-up evaluation of a different set of children attending the same day care
28 center 4 to 5 years after exposure showed no residual changes in VCS or color vision, although
29 these children were not tested immediately after exposure (NYS DOH, 2005c). A case study
30 reported reduced VCS in a 2V2 year old boy after prenatal exposure to tetrachloroethylene (Till et
31 al., 2003). Sections 4.6.2 and 4.7.2 discuss studies of postnatal neurological effects in animals
32 after prenatal exposure. Altered brain biochemistry was seen in the offspring of exposed rodents
33 (Kyrklund & Haglid, 1991; Nelson et al., 1980), and the offspring showed signs of
34 developmental delay (Nelson et al, 1980), altered motor activity (Szakmary et al., 1997; Tinston,
35 1994), decreased muscular strength (Szakmary et al., 1997), and short-term reduced response to
36 sound in pups (Tinston, 1994). In addition, young animals have been directly exposed
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-168 DRAFT-DO NOT CITE OR QUOTE
-------
1 postnatally to tetrachloroethylene. One gavage study on young 45-50 gram rats showed
2 behavioral and locomotor effects (Chen et al., 2002a), and another gavage study on 10-day old
3 mice showed increased locomotor activity and decreased rearing behavior (Fredriksson et al.,
4 1993). Following i.p. dosing, 8-week-old male mice showed effects on the righting reflex and
5 balancing (Umezu et al., 1997), and 6-week-old rats showed effects on locomotor activity
6 (Motohashi et al., 1993). Both human and animal evidence supports an association between
7 neurodevelopmental effects and tetrachloroethylene exposure.
8 Section 4.8.1.1.1 and Table 4-10 describe studies relating tetrachloroethylene to immune
9 response in children. Lehmann et al. (2002) examined cord blood samples for T-cell
10 subpopulations and associated them with indoor exposure to VOCs measured 4 weeks after birth
11 (likely to reflect late-prenatal exposures); however, another study examining indoor exposure to
12 VOCs and allergic sensitization and cytokine secretion in 3-year-old children at high risk for
13 development of allergic disease (low birth weight, high cord blood IgE, family history of atopy)
14 found no association between tetrachloroethylene exposure and any of the allergens tested in this
15 study (Lehmann et al. 2001). In a study of inhalation exposure, Delfmo et al. (2003a, b)
16 measured the concentration of ambient air pollutants, including tetrachloroethylene, and
17 correlated it with subsequent symptoms of asthma in children in the Los Angeles area. The
18 results suggest an increased risk with exposure to tetrachloroethylene (Delfmo et al., 2003a).
19 However, another analysis of the data examined the amount of tetrachloroethylene and other
20 volatile organic compounds in exhaled breath of asthmatic children (Delfmo et al., 2003b).
21 Although there was a significant correlation between ambient and exhaled concentrations, the
22 investigators did not find any association with exhalation concentrations and asthma symptoms
23 or ambient air concentrations and asthma symptoms, although the OR for exhaled breath was
24 larger than for ambient air exposure (OR = 1.94, 95% CI = 0.8-4.7; Delfmo et al., 2003b). An
25 18-year-old without personal or family history of bronchial asthma developed respiratory
26 symptoms (cough, dyspnea, altered forced expiratory volume) after maintaining dry cleaning
27 machines (Boulet, 1988). The limited, available data from these studies provide weak evidence
28 of an effect of tetrachloroethylene exposure during childhood on allergic sensitization or
29 exacerbation of asthma symptomology. However, the observation of the association between
30 increased tetrachloroethylene exposure and reduced interferon-y in cord blood samples may
31 reflect a sensitive period of development, and points to our current lack of understanding of the
32 potential immunotoxic effects of prenatal exposures.
33 Other postnatal health effects after tetrachloroethylene exposure have been seen in
34 children. In one case study with inexact exposure information, tetrachloroethylene vapors off-
35 gassing from dry-cleaned fabrics were implicated in causing the death of a 2-year-old boy after
36 sleeping in a room with curtains that had been incorrectly dry cleaned (Gamier et al., 1996).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-169 DRAFT-DO NOT CITE OR QUOTE
-------
1 Bagnell and Ellenberger (1977) reported that a child suffered from obstructive jaundice and
2 hepatomegaly after consuming contaminated breast milk, with conditions improving when
3 breastfeeding was discontinued. In the one case of a child's direct ingestion of
4 tetrachloroethylene, a 6-year-old boy who swallowed 12-16 g tetrachloroethylene lost
5 consciousness and lapsed into a coma (Koppel et al., 1985). This 6-year-old also experienced
6 drowsiness, vertigo, agitation, and hallucinations, but he later recovered. Follow-up testing on
7 the boy was not reported; therefore, any potential long-term effects of the exposure are unknown
8 (see Section 2.2.5).
9
10 4.9.1.2.2. Toxicokinetics and tetrachloroethylene in early life stages. Chapter 3 describes the
11 toxicokinetics of tetrachloroethylene. Early life stage-specific information regarding absorption,
12 distribution, metabolism, and excretion needs to be considered for a child-specific and chemical-
13 specific PBPK model. To adequately address the risk to infants and children, age-specific
14 parameters for these values should be used in PBPK models that can approximate the internal
15 dose a infant or child receives based on a specific exposure level (Byczkowski and Fisher, 1994;
16 Clewell et al., 2004; Gentry et al., 2003; Rao and Brown, 1993; see Section 3.5).
17 As discussed in Section 3.1, exposure may occur via inhalation, ingestion, and skin
18 absorption. In addition, prenatal exposure may result in absorption via the transplacental route.
19 Exposure via inhalation is proportional to the ventilation rate, duration of exposure, and
20 concentration of expired air, and children have increased ventilation rates per kg body weight
21 compared to adults, with an increased alveolar surface area per kg body weight for the first two
22 years (NRC, 1993). For lipophilic compounds such as tetrachloroethylene, percent adipose
23 tissue, which varies with age (NRC, 1993), will affect absorption and retention of the absorbed
24 dose. It is not clear to what extent dermal absorption may be different for pregnant women and
25 children compared to adults, given their increased surface areas and thinner outer skin layers.
26 The distribution of tetrachloroethylene to specific organs will depend on organ blood
27 flow and the lipid and water content of the organ (NRC, 1993), which may vary between life
28 stages. Rodent studies demonstrate that tetrachloroethylene crosses the placental barrier when
29 pregnant dams are exposed (Ghantous et al., 1986; Szakmary et al., 1997), and in humans it has
30 been shown that during lactation, tetrachloroethylene distributes to breast milk (NYS DOH,
31 2005b; Schreiber, 1993; Sheldon et al., 1985). However, a noticeable difference exists between
32 the milk/blood partition coefficients for rats (12) and for humans (2.8; Byczkowski and Fisher,
33 1994), reflecting the higher fat content of rat milk.
34 Tetrachloroethylene can also cross the blood-brain barrier during both prenatal and
35 postnatal development; this may occur to a greater extent in younger children. Based on the
36 modeled dose of tetrachloroethylene to the brain after a showering/bathing scenario, a study by
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-170 DRAFT-DO NOT CITE OR QUOTE
-------
1 Rao and Brown (1993) showed that for a given set of exposures, the younger a person is, the
2 greater the estimated concentration of tetrachloroethylene in the brain. Modeling showed that
3 after a 30-minute bathing scenario, a 3-year-old child accumulated higher brain tissue
4 concentrations of tetrachloroethylene as compared with a 10-year-old and an adult. An autopsy
5 conducted on the previously mentioned 2-year-old boy found dead after exposure to dry-cleaned
6 curtains revealed the highest levels of tetrachloroethylene in the brain, 77 mg/kg. Levels in his
7 blood, heart, and lungs were 66 mg/L, 31 mg/kg, and 46 mg/kg, respectively (Gaillard et al.,
8 1995; Gamier et al., 1996).
9 Animal studies provide clear evidence that tetrachloroethylene distributes widely to all
10 tissues of the body (see Section 3.2) but it is not clear whether distribution may vary
11 differentially with lifestage.
12 Section 3.3.3 describes the production of CYP enzymes involved in the metabolism of
13 tetrachloroethylene. Expression of these enzymes changes during various stages of fetal
14 development (Hakkola et al., 1996a, 1996b, 1998) and during postnatal development (Tateishi et
15 al., 1997). One study modeled the role of the age-dependent development of CYP2E1 in
16 oxidative metabolism (TCA) in the mother and lactating infant (Vieira et al., 1996). A number
17 of other human studies suggest that CYP2B6 may also play a role in the metabolism of
18 tetrachloroethylene (White et al., 2001), although this enzyme was not detected in placental or
19 fetal liver samples (Hakkola et al., 1996a, b), and differences between a group of 10 perinatal
20 and infant patients showed significantly lower CYP2B6 protein expression in placental hepatic
21 microsomes as compared with an adult group (Tateishi et al., 1997).
22 The major processes of excretion of tetrachloroethylene and its metabolites are discussed
23 in Section 3.3 and 3.4. Tetrachloroethylene or its metabolites have been measured in blood
24 (NYS DOH, 2005a; Popp et al., 1992), exhaled breath (Delfmo et al, 2003b; Schreiber et al.,
25 2002; NYS DOH, 2005a), and urine of children (NYS DOH, 2005b; Schreiber et al., 2002; Popp
26 etal., 1992).
27
28 4.9.1.2.3. Toxicodynamics and tetrachloroethylene in early life stages. Toxicodynamic
29 responses to chemical exposures can change throughout different life stages. TCA, a metabolite
30 of tetrachloroethylene, is hypothesized to be the causative agent for developmental toxicity
31 expressed as morphological changes, lethality, and/or growth. TCA could accumulate to a
32 greater extent in the embryo/fetal compartment than in the mother, based on the pKa of the acid
33 and the pH gradient between the maternal plasma and the embryo compartments (O'Flaherty et
34 al., 1992). TCA could induce developmental toxicity by changing the intracellular pH or
35 through peroxisome proliferation. Ghantous et al. (1986) detected TCA in the amniotic fluid of
36 pregnant mice exposed to tetrachloroethylene via inhalation (see Section 4.7.4).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-171 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.9.1.2.4. Susceptibility to cancer in early life stages. The epidemiologic and experimental
2 animal evidence is limited regarding susceptibility to cancer from exposure to
3 tetrachloroethylene during early life stages. The human epidemiological evidence is summarized
4 above for cancer in the liver (see Section 4.4.1.2), kidney (see Section 4.5.1.2), and other organ
5 systems (see Section 4.8.1.2). The animal research is summarized above for cancer in the liver
6 (see Section 4.4.2.2), kidney (Section 4.5.2.2), and other organ systems (see Section 4.8.2).
7 Few studies have examined cancer in children after exposure to tetrachloroethylene;
8 however, those few have found evidence for concern for leukemia (see Section 4.8.1.2.1). One
9 case-control study of children residing in Woburn, Massachusetts diagnosed with leukemia
10 examined exposure to drinking water contaminated with multiple solvents, including
11 tetrachloroethylene (Costas et al., 2002; MA DPH, 1997). This study reported a large but
12 imprecise association and a dose-response relationship between maternal exposure during
13 pregnancy and childhood leukemia in the offspring when compared with exposure prior to
14 pregnancy or postnatal exposure to the infant via lactation (Costas et al., 2002; MA DPH, 1997),
15 and altered immune response was found in family members of the cases (Byers et al., 1988; see
16 Section 4.8.1.1.1). However, it is difficult to uniquely identify tetrachloroethylene as the
17 causative agent given the higher concentrations of trichloroethylene reported in these studies.
18 Similarly, in another case-control study of childhood leukemia, paternal exposure to chlorinated
19 solvents have been associated with increased risk (Lowengart et al., 1987), although this and
20 other case-control studies do not show an increased risk from paternal (Lowengart et al., 1987;
21 Shu et al., 1999) or maternal (Infante-Rivardet al., 2005; Shu et al., 1999) occupational exposure
22 to tetrachloroethylene, possibly due to the relatively small sample size. Another study
23 population is currently being further examined to determine any association between maternal
24 ingestion of contaminated water and the incidence of childhood cancers (ATSDR, 2003). One in
25 vitro study of human mononuclear cord blood cells exposed to tetrachloroethylene found that
26 pathways involved in cancer induction were affected through altered gene expression of
27 inflammatory responses, tumor and metastatis progression, and the apoptotic process (Diodovich
28 et al., 2005). Leukemia has also presented in adult humans after tetrachloroethylene exposure
29 (see Section 4.8.1.2.1, Appendix 4B, and Tables 4B-5 and 4B-13), and MCL has been seen in
30 exposed adult rats in spite of high spontaneous background incidences (see Section 4.8.2.4).
31 While interspecies extrapolation of rat MCL to humans has been questioned (see Section
32 4.8.2.4.1), the findings in humans suggest relevance for the leukemia findings in animals.
33 However, no data are available on leukemia risk in young animals exposed to tetrachloroethylene.
34
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-172 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.9.1.3. Later Life Stages
2 Few studies examine the effects of tetrachloroethylene exposure in elderly adults. One
3 study found elevated blood tetrachloroethylene levels (310-1770 ug/L) and urine trichloroacetic
4 acid levels (22-1650 ug/L) in an elderly couple living above a dry cleaning facility (Popp et al.,
5 1992). Another residential study examined two individuals over the age of 60 years and found
6 that the mean scores of VCS were lower than the 12th percentile of all control subjects (Schreiber
7 et al., 2002). These studies suggest that older adults may experience increased exposure to
8 tetrachloroethylene and resulting increased VCS deficits than younger adults. However, there is
9 no further evidence for elderly individuals exposed to tetrachloroethylene beyond these two
10 studies.
11
12 4.9.2. Other Susceptibility Factors
13 Aside from age, many other factors may affect susceptibility to tetrachloroethylene
14 toxicity. A partial list of these factors includes gender, genetic polymorphisms, pre-existing
15 disease status, nutritional status, diet, and previous or concurrent exposures to other chemicals.
16 The toxicity that results due to changes in multiple factors may be quite variable, depending on
17 the exposed population and the type of exposure. Qualitatively, the presence of multiple
18 susceptibility factors will increase the variability that is seen in a population response to
19 tetrachl oroethy 1 ene toxi city.
20
21 4.9.2.1. Health and Nutritional Status
22 It is known that kidney diseases can affect the clearance of chemicals from the body, and
23 therefore poor kidney health may lead to increased half-lives for tetrachloroethylene and its
24 metabolites. Similarly, liver disease may change the metabolic profiles in the liver, thus
25 potentially altering tetrachloroethylene metabolism.
26 Co-exposure to a-tocopherol (vitamin E) along with tetrachloroethylene resulted in
27 decreased rat (Costa et al., 2004) and mouse (Ebrahim et al., 1996, 2001) liver cell toxicity. A
28 similar protective effect was also seen with co-exposure to 2-deoxy-D-glucose in mice (Ebrahim
29 et al., 1996, 2001) and taurine in mice (Ebrahim et al., 2001). However, no associations were
30 found for blood levels of vitamin E and beta-carotene in rats (Toraason et al., 2003; see Sections
31 4.3 and 4.4.4.4.3).
32
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-173 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.9.2.2. Gender
2 In humans, it has not been determined whether there is a gender difference in response to
3 exposure to tetrachloroethylene. However, because gender also affects metabolic capabilities,
4 which vary throughout development, it is important to consider sex-specific changes.
5 In the case of tetrachloroethylene, there is some indication that tetrachloroethylene
6 metabolism is different between males and females. One PBPK model found gender-specific
7 differences that were small (although significant) in tetrachloroethylene blood concentrations but
8 considerable (2-fold at age 40) with regard to TCA blood concentration levels (Clewell et al.,
9 2004; see Section 3.5.2 and Figure 3-3.). Opdam and Smolders (1986) exposed six human
10 subjects to concentrations ranging from 0.5-9 ppm and found alveolar concentrations in male
11 subjects to be only slightly less than those in females (see Figures 3-6a, b). It is not known
12 whether gender variation of beta lyase activity (see Section 3.3.3.2.3), the most important
13 activator of toxic products in the conjugation pathway, exists in humans as it does in rats, with
14 metabolism in males being faster than in females (Volkel et al., 1998), although there seems to
15 be little gender difference in the concentrations of metabolites in blood, regardless of age
16 (Sarangapani et al., 2003).
17 Ferroni et al. (1992) evaluated neurological effects of tetrachloroethylene exposure
18 among female dry cleaners and concluded that tetrachloroethylene exposure in dry cleaning
19 shops may impair neurobehavioral performance and affect pituitary function. The pituitary is
20 controlled in part by hypothalamic dopamine, which is important to neurotransmission. Study
21 participants were tested during the proliferation phase of menstruation which may better capture
22 changes in prolactin secretion but also may potentially confound findings if there are individual
23 differences in severity of menstruation and in the timing of test session relative to the day of
24 menstruation (U.S. EPA, 2004; see Section 4.6.1.2.5).
25 In a study of aircraft maintenance employees, Spirtas et al. (1991) observed an increased
26 risk for NHL in females compared to males (see Section 4.8.1.2.1). Although quantitative
27 exposure information on tetrachloroethylene was not obtained in this study, differences in
28 exposure potential and level of exposure may explain the difference in risk between women and
29 men. Differences in physiological parameters may also explain the observed gender difference
30 in risk.
31 The studies by Pesch et al. (2000a) and Dosemeci et al. (1999) suggest that there may be
32 gender differences in risk to renal cell carcinoma with occupational exposure to
33 tetrachloroethylene; in both studies the risks were higher in males than in females (see Section
34 4.5.1.2). In a rat inhalation study, tubule cell hyperplasia was observed in eight males at various
35 doses, but in only one female at high dose. Also, renal tubule adenomas and adenocarcinomas
36 were observed only in males; however, chronically induced tetrachloroethylene neoplastic
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-174 DRAFT-DO NOT CITE OR QUOTE
-------
1 kidney lesions do not exhibit sex specificity (NTP, 1986a). In a rat gavage study, there was no
2 gender difference for toxic nephropathy (NCI, 1977). A marked gender difference was seen
3 between male and female rats in the severity of acute renal toxicity with male rats being more
4 affected than female rats (Lash et al., 2002), but otherwise no gender variation was observed for
5 chronic nephrotoxicity not associated with alpha-2u-globulin nephropathy (see Sections 4.5.2.2
6 and 4.5.4.3.3).
7 In the liver, male rats showed an increased incidence of spongiosis hepatitis as compared
8 with females, but there was no gender difference in hepatocellular adenomas and carcinomas;
9 however, the spleen showed increased effects in males versus females (JISA, 1993; see Sections
10 4.4.2.1 and 4.4.2.2).
11
12 4.9.2.3. Race/Ethnicity
13 One residential study found that buildings with >100 |ig/m3 tetrachloroethylene were
14 more likely in minority neighborhoods (OR = 6.7; 95% CI = 1.5-30.5; NYS DOH, 2005a). In
15 addition to possible increased exposure, different racial or ethnic groups may express metabolic
16 enzymes in different ratios and proportions due to genetic variability.
17 In a follow-up study on the mortality of a cohort of dry cleaners, bladder cancer was
18 elevated among Caucasian men and women, and kidney cancer was elevated among black men
19 and women; however, these associations were not strongly related to duration or estimated level
20 of exposure to tetrachloroethylene (Blair et al., 2003). One study found that following
21 tetrachloroethylene exposure, TCA concentration in the urine of six Asian subjects was no
22 different from the levels found in six Caucasians; however, this study was confounded by
23 significant differences in alcohol consumption between the Caucasian and Asian populations
24 (Jang and Droz, 1997).
25 Eskenazi et al. (1991a) noted a slightly lower per-cycle pregnancy rate among wives of
26 men who received higher level exposure to tetrachloroethylene, but the potential contribution of
27 tetrachloroethylene exposure to time to conception was small when compared with the
28 contribution observed from Hispanic ethnicity and smoking, which were found to be stronger
29 and statistically significant predictors of time to conception.
30
31 4.9.2.4. Genetics
32 Human variation in response to tetrachloroethylene exposure may be associated with
33 genetic variation. For example, in a study of six adults, Monster et al. (1979) found that the
34 mean coefficient of interindividual variation for tetrachloroethylene uptake was 17%. Human
35 genetic polymorphisms in metabolizing enzymes involved in biotransformation of
36 tetrachloroethylene are now known to exist (U.S. EPA, 1991; IARC, 1995; Lash and Parker,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-175 DRAFT-DO NOT CITE OR QUOTE
-------
1 2001). Section 3.3.3.1.5 discusses CYP isoforms and genetic polymorphisms, Section 3.3.3.2.1
2 covers GST isoenzymes and polymorphisms, and Section 3.3.4 describes differences in
3 enzymatic activity.
4 Reitz et al. (1996) examined tetrachloroethylene metabolism in seven adult human liver
5 samples and found a fivefold difference in the rate of tetrachloroethylene metabolism between
6 the 50th and 99th percentiles. Opdam (1989) found a 2-fold spread in tetrachloroethylene blood
7 concentrations in a study population of nine adult human subjects. In this study, the amount of
8 fat and the blood concentrations seemed to be positively correlated but could not be confirmed;
9 the author suggested that if the subjects had a wider range of body fat levels (range in this study
10 was only 7-22 kg), a larger amount of interindividual variation would be expected.
11 Computer modeling was used to examine the toxicokinetic variability of tetrachloroethylene
12 (Bois et al., 1996; Chiu and Bois, 2006). However, whether CYP or GSH polymorphisms
13 account for interindividual variation in tetrachloroethylene metabolism among humans, and thus
14 differences in susceptibility to tetrachloroethylene-induced toxicities, is not known.
15
16 4.9.3. Multiple Exposures and Cumulative Risks
17 When considering health risks, it is important to consider the cumulative impact of
18 effects that may be due to multiple routes of exposure. EPA published Framework for
19 Cumulative Risk Assessment (U.S. EPA, 2003c) to address these issues. A human aggregate
20 exposure model developed by McKone and Daniels (1991) incorporated likely exposures from
21 air, water, and soil media through inhalation, ingestion, and dermal contact. They asserted that
22 the aggregate exposure may be age dependent, but did not present any data for persons of
23 differing life stages.
24 The limited data summarized by the ATSDR in its draft interaction profile on
25 tetrachloroethylene, trichloroethylene, 1,1-dichloroethane, and 1,1,1-trichloroethane suggest that
26 additive joint action is plausible (ATSDR, 2001). Co-exposure to other pollutants, including
27 trichloroethylene and methylchloroform which produce some of the same metabolites and
28 similar health effects as tetrachloroethylene, is likely to occur in occupational settings as well as
29 in non-occupational sources such as in ground water contamination (e.g., Bove et al., 1995;
30 Lagakos et al., 1996; MA DPH, 1997; ATSDR, 1998; Sonnenfeld et al., 2001). However, no
31 evidence was among available studies indicates greater-than-additive effects for liver and kidney
32 toxicity.
33 Due to the effects that many chemicals have on inducing and/or repressing metabolic
34 enzymes as well as on organ systems, co-exposures may alter the way in which
35 tetrachloroethylene is metabolized and cleared from the body. Inhibition or induction of the
36 enzymes responsible for tetrachloroethylene metabolism can—and likely does—alter
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-176 DRAFT-DO NOT CITE OR QUOTE
-------
1 susceptibility to toxicity (U.S. EPA, 1985a; IARC, 1995; Lash and Parker, 2001). Numerous
2 environmental pollutants and therapeutic agents have the potential to induce or inhibit
3 tetrachloroethylene-metabolizing enzymes. For example, tetrachloroethylene metabolism is
4 increased by inducers of CYP enzymes such as toluene, phenobarbital, and pregnenolone-
5 16-a-carbonitrile, whereas CYP inhibitors such as SKF 525A, metyrapone, and carbon monoxide
6 decrease tetrachloroethylene metabolism (Moslen et al., 1977; Ikeda and Imanura, 1973; Costa
7 and Ivanetich, 1980). Likewise, tetrachloroethylene exposure may increase the effects of
8 exposures to other chemicals or stressors. For instance, adverse effects due to exposure to
9 chlorinated solvents and alcohol may be increased because tetrachloroethylene may induce
10 shared metabolic enzymes (see Section 3.3.4).
11 The acute effects of tetrachloroethylene share much in common functionally with those
12 of other solvents (e.g., toluene, volatile anesthetics, and alcohols) such as changes in reaction
13 time, nerve conduction velocity, and sensory deficits. There is emerging evidence that such
14 agents act on the ligand-gated ion channel superfamily in vitro (Shafer et al., 2005), particularly
15 on the inhibitory amino acids NMD A, nicotinic, and GAB A receptors in vivo (Bale et al., 2005).
16 Other organic solvents induce effects on memory and color vision (Altmann et al., 1995; Mergler
17 et al., 1991; Hudnell et al., 1996a, b). The consistency of these observations suggests a common
18 MO A of organic solvents to altered vision pattern. Hence, a concern exists for neurobehavioral
19 effects from interaction or competitive inhibition between tetrachloroethylene and exposures
20 with similarly hypothesized MOAs.
21 The interaction between tetrachloroethylene, trichloroethylene, and 1,1,1-trichloroethane
22 (methylchloroform) was modeled in rats (Dobrev et al., 2001) and in computer models for
23 humans (Dobrev et al., 2002) and were shown to compete for metabolic capacity. The
24 interaction between tetrachloroethylene and trichloroethylene showed a less-than-additive effect
25 on the liver and kidney through inhibition of TCA formation (Pohl et al., 2003). Similarly, when
26 exposed to tetrachloroethylene, rat liver cells had increased toxicity when co-exposed to
27 peroxidation drugs such as cyclosporine A, valproic acid, and amiodarone (Costa et al., 2004),
28 and n-hexane and ethylbenzene inhibited the metabolism of tetrachloroethylene in rats (Skowron
29 etal., 2001).
30 Alcohol and smoking are generally regarded as confounders, although the additive or
31 interactive effects of these exposures along with tetrachloroethylene are not well characterized.
32 Exposure to alcohol changes the metabolic profiles in the liver, thus increasing metabolism to
33 TCA through the CYP2E1 pathway (Pastino et al., 2000). Alcohol by itself cannot account for
34 the observed deficits in neurobehavioral functions, because statistical analyses of the
35 epidemiologic observations accounted for this covariate. Meskar et al. (2001) also contended
36 that alcohol induces CYP2E1, which in turn has the ability to activate other compounds such as
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-177 DRAFT-DO NOT CITE OR QUOTE
-------
1 halogenated solvents. This could potentially cause higher toxicity of tetrachloroethylene for
2 those who use alcohol, if the oxidative metabolism leads to a proto-oxidant. Valic et al. (1997)
3 showed greater decrements in color vision among subjects with both exposures as compared with
4 individuals with solvent exposure only or with neither exposure (see Sections 4.6.1.3 and 4.6.3).
5 Regarding esophageal cancer, occupational observations suggest that the magnitude of
6 the risks for several smoking-related cancers among dry cleaners was greater than could be
7 explained by smoking alone, suggesting a further contribution from another risk factor, such as
8 occupational exposure (Blair et al., 2003; Ruder et al., 2001; see Section 4.8.1.2.2).
9
10 4.9.4. Uncertainty of Database for Susceptible Populations
11 There is a need to better characterize the implications of tetrachloroethylene exposures to
12 susceptible populations. A number of areas where the data base is currently insufficient are
13 identified below.
14
15 4.9.4.1. Uncertainties of Exposure
16 A number of uncertainties exist regarding exposure to subpopulations. Further evaluation
17 of the effects of multiple routes and pathways of exposures and aggregate risks is needed.
18 Similarly, the effects due to co-exposures to other compounds with similar or different MO As
19 need to be evaluated. An estimate of multiple exposures is needed to know where along the
20 dose-response curve to place an incremental exposure to tetrachloroethylene. The size of any
21 increased risk will be different in dissimilar regions of the dose-response curve. This means that
22 a dose that is safe for an unexposed population will not necessarily be a safe dose if background
23 and other exposures are considered. Until quantitative conclusions can be made for each
24 susceptibility factor, it will be very hard to consider the impacts of changes in multiple
25 susceptibility factors.
26 Although there is more information on early life exposure to tetrachloroethylene than on
27 other potentially susceptible populations, there remain a number of uncertainties regarding
28 children's susceptibility. For example, it is not clear to what extent tetrachloroethylene may pass
29 through the placenta in humans, as shown in a rodent study (Ghantous et al., 1986). Also, there
30 is limited information that evaluates nonoccupational exposures to tetrachloroethylene (e.g., in
31 homes and automobiles) for all susceptible populations or on additional exposures that may
32 modify an individual's exposure to tetrachloroethylene. Improved PBPK modeling and
33 validation of these models will aid in determining how variations in metabolic enzymes affect
34 tetrachloroethylene metabolism.
35 Although inhalation is believed to be of most concern for tetrachloroethylene, the
36 pathways of exposure as well as the MO A for children are not well characterized. Inhalation
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-178 DRAFT-DO NOT CITE OR QUOTE
-------
1 exposures may occur when tetrachloroethylene vapors are released from treated clothing or the
2 clothing worn by occupationally exposed individuals, as well as when vapors are exhaled in the
3 breath of exposed workers (ATSDR, 1997; Aggazzotti et al., 1994a, b). Dry-cleaned garments
4 transported in an automobile may also lead to unexpectedly high levels of exposure to children
5 who sit in the rear seats of cars, nearest to where most items are stored (Park et al., 1998; Chien,
6 1997). Inhalation exposure may also occur during showering or bathing as dissolved
7 tetrachloroethylene in the warm tap water becomes volatilized (Rao and Brown, 1993; see
8 Section 2.2.1).
9 Although there is more information on early life exposure to tetrachloroethylene than on
10 other potentially susceptible populations, a number of uncertainties remain regarding children's
11 susceptibility. For example, though demonstrated in a study of placental transport by Ghantous
12 et al. (1986), it is not clear to what extent tetrachloroethylene may pass through the human
13 placenta. Also, there is limited information that evaluates nonoccupational exposures to
14 tetrachloroethylene (e.g., in homes and automobiles) for all susceptible populations or on
15 additional exposures that may modify an individual's exposure to tetrachloroethylene. Improved
16 PBPK modeling and validation of these models will aid in determining how variations in
17 metabolic enzymes affect tetrachloroethylene metabolism.
18 Although ingestion of tetrachloroethylene through breast milk may be a significant
19 pathway of exposure for some infants (see Sections 2.2.4 and 3.2), it has been suggested that if
20 these infants live adjacent to or in close proximity of dry cleaning facilities, the dose received
21 through ingestion of breast milk will become insignificant when compared with the inhalation
22 exposure and subsequent dose (Schreiber, 1997).
23 Certain foods have been found to be contaminated with tetrachloroethylene (see Section
24 2.2.3). Because children consume a high level of dairy due to their need for calcium for bone
25 growth, the lipophilicity of tetrachloroethylene may pose a higher concern for children than for
26 adults. Dairy intake is generally highest during infancy and decreases throughout life (NRC,
27 1993).
28 It is not clear to what extent dermal absorption is possible for children. Although an
29 infant's skin has similar permeability to adults, a premature infant may have increased
30 permeability (Guzelian et al., 1992). Also, an infant has approximately double the ratio of
31 surface area to body weight compared to adults (NRC, 1993), which could imply increased
32 exposure during bathing and swimming, which has already been modeled for adults by Rao and
33 Brown (1993).
34 It is not known to what extent tetrachloroethylene is absorbed by a child and to which
35 organs the chemical and its metabolites may be distributed. A validated PBPK model is needed
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-179 DRAFT-DO NOT CITE OR QUOTE
-------
1 that contains physiologic parameter information for infants and children, including the effects of
2 maternal inhalation exposure and the resulting concentration in breast milk.
O
4 4.9.4.2. Uncertainties of Effects
5 More studies specifically designed to evaluate effects in early and later life stages are
6 needed in order to more fully characterize potential life stage-related tetrachloroethylene toxicity.
7 Because the neurological effects of tetrachloroethylene constitute the most sensitive endpoints of
8 concern for noncancer effects, it is quite likely that the early life stages may be more susceptible
9 to these outcomes than are adults. Life stage-specific neurotoxic effects, particularly in the
10 developing fetus, need further evaluation. It is important to consider the use of age-appropriate
11 testing for assessment of these and other outcomes, both for cancer and noncancer outcomes.
12 The reduction in fertility seen in some studies (Eskenazi et al., 1991a, b; Rachootin and
13 Olsen, 1983; Sallmen et al., 1995) occurs by an unknown mechanism. Altered sperm quality is
14 one possibility (Beliles et al., 1980; Eskenazi et al., 1991b), as is spontaneous abortion/fetal loss
15 occurring early in gestation without maternal knowledge of the pregnancy, thereby being
16 misclassified as infertility (see Section 4.7.1).
17 Data specific to the carcinogenic effects of tetrachloroethylene exposure during the
18 critical periods of development of experimental animals and humans also do not exist. The
19 perinatal period, which encompasses the end of pregnancy and the early postnatal period, may be
20 the most susceptible window for exposure for tetrachloroethylene across species (Beliles, 2002).
21 Several of the adverse pregnancy outcome studies evaluated exposure during a critical window,
22 the first trimester of the pregnancy, a critical window for exposure (MA DPH, 1997; Kyyronen
23 et al., 1989; Ahlborg, 1990; Taskinen et al., 1989). Exposure during another developmental
24 period may not result in certain outcomes that occur from exposure during this critical window.
25 Alternately, another window of exposure may result in a different outcome than occurs during
26 the first trimester.
27
28 4.9.5. Conclusions on Susceptibility
29 There is some evidence that certain subpopulations may be more susceptible to exposure
30 to tetrachloroethylene. These subpopulations include early and later life stages, health and
31 nutrition status, gender, race/ethnicity, genetics, and multiple exposures and cumulative risk.
32 Cancer outcomes of concern for perinatal exposure are not well characterized in either
33 the human epidemiological or the experimental animal literature. Data-derived noncancer
34 outcomes of concern in early life stages are spontaneous abortion/fetal loss, mortality, and
35 neurological impairment. As described above, the evidence for spontaneous abortion following
36 prenatal exposures to tetrachloroethylene is well characterized in humans, and fetal loss is well
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-180 DRAFT-DO NOT CITE OR QUOTE
-------
1 characterized in experimental animals. However, the human epidemiological data that support
2 this conclusion do not provide information on the maternal dose to tetrachloroethylene that may
3 have resulted in spontaneous abortion. Further, data from the experimental animal studies
4 suggest that this finding may be a high-dose effect. Together, this evidence suggests that
5 reference values that are established on the basis of more sensitive neurological endpoints should
6 mitigate potential risk of fetal lethality.
7 Regarding postnatal hazard, the correlation of tetrachloroethylene exposure to childhood
8 mortality is based on a single report in which a 2-year-old toddler died following exposure to
9 dry-cleaned curtains (Gamier et al., 1996). Ambient dose was not well characterized, although
10 tissue levels of tetrachloroethylene indicated the possibility of bioaccumulation in the brain. Yet,
11 it is noted that early postnatal mortality was not observed in animal studies. The overall
12 confidence in this endpoint, observed in a single individual, is minimal relative to predicting risk
13 for the broader population.
14 Likewise, evidence of neurobehavioral impairment in children is based on a minimal data
15 set, consisting of only four children who resided in an apartment building with a dry cleaning
16 establishment and who demonstrated visual system impairment (Schreiber et al., 2002).
17 Confidence in the results of this study for the assessment of risk to the broader population of
18 children is minimal due to the size of the studied population (i.e., only four individuals). Given
19 the lack of information on these factors in referent children, it is difficult to evaluate the possible
20 contribution of other factors that may have contributed to the observed visual system impairment
21 in children (Storm and Mazor, 2004). The lack of associations reported in Storm and Mazor
22 does not contradict the findings of Schreiber et al., given the differences in study protocol and
23 procedures (Hudnell and Schreiber, 2004). Decreased VCS has been observed in children who
24 resided in an apartment building with a dry cleaning establishment (Schreiber et al., 2002; NYS
25 DOH, 2005a). Children who attended a day care center adjacent to a dry cleaner did not
26 consistently show any effect on cognition or behavior when assessed 5 weeks after exposure
27 ceased (NYS DOH, 2005b), and no effect on VCS was seen 5 years after exposure (NYS DOH,
28 2005c). Moreover, the findings of Till et al. (2001a, b, 2005) and Laslo-Baker et al. (2004),
29 although not definitive, further suggest that the developing fetus is susceptible to maternal
30 organic solvent exposures (see Section 4.6.1).
31 Other subpopulations with potential for susceptibility to tetrachloroethylene include the
32 elderly, diminished health status, gender, race/ethnicity, and multiple/cumulative exposures.
33 There is suggestive evidence that there may be greater susceptibility for exposures to the elderly.
34 Diminished health status (e.g., impaired kidney liver or kidney) will likely affect an individual's
35 ability to metabolize tetrachloroethylene, whereas certain nutrients may have a protective effect
36 on exposure. Gender and race/ethnic differences in susceptibility are likely due to variation in
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-181 DRAFT-DO NOT CITE OR QUOTE
-------
1 physiology and exposure, and genetic variation likely has an effect on the toxicokinetics of
2 tetrachloroethylene. Multiple and cumulative exposures are likely to cause competition in
3 metabolic capacity. Future research should better characterize possible susceptibility for certain
4 life stages or subpopulations.
5
6 4.10. SUMMARY OF HAZARD IDENTIFICATION
7 4.10.1. Description of Effects and Exposure Levels at Which They Occur
8 In the previous sections of this document the effects in each organ system were discussed
9 pairwise in three categories: humans/animals, noncancer/cancer, and inhalation/oral. The
10 summaries in Sections 4.4.3, 4.5.3, 4.6.3, and 4.7.3 pertain to each of the organ systems
11 individually. In this section, the same effects are integrated across organ systems, with the
12 primary subdivision being humans/animals, the secondary subdivision being noncancer/cancer,
13 and later subdivisions being the exposure route and organ system. Section 4.10.2 summarizes
14 the potential modes of action. The dose levels, effect levels, and concentrations discussed here
15 are those observed in the studies and are not corrected for continuous exposure or for human
16 equivalency. In Chapter 5, these corrections are made before deriving the RfCs and RfDs.
17
18 4.10.1.1. Summary of Effects in Humans
19 4.10.1.1.1. Human noncancer effects. The epidemiologic evidence indicates that the primary
20 targets of tetrachloroethylene noncancer toxicity are the CNS, kidneys, liver, and developing
21 fetus. The epidemiologic evidence supporting these inferences is derived primarily from studies
22 of tetrachloroethylene-exposed dry cleaners—with two studies reporting neurobehavioral effects
23 in residents living in housing located in close proximity to a dry cleaning facility—and from
24 studies reporting effects to the developing fetus in populations exposed to drinking water
25 contaminated with tetrachloroethylene and other solvents. In the drinking water studies, several
26 of the contaminants are congeners (e.g., tetrachloroethylene and trichloroethylene), which are
27 metabolized in the body to TCA and DCA.
28 The epidemiologic database is primarily composed of studies of a prevalence or cross-
29 sectional design. Although a cohort study, by definition, is able to identify that exposure has
30 indeed occurred before disease, the available epidemiologic studies, including those of a cross-
31 sectional design, support a causal role of tetrachloroethylene.
32 In most cases, the number of study subjects was not large; however, the issue of sample
33 size affects the power of the study to detect underlying risk. Hence, observed effects are
34 considered noteworthy if chance and bias are minimized. Furthermore, the number of studied
35 individuals in the epidemiologic studies of tetrachloroethylene is not any smaller than that of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-182 DRAFT-DO NOT CITE OR QUOTE
-------
1 epidemiologic studies for many chemicals identified in the U.S. EPA's IRIS. In fact, it is a rare
2 case when hazard inferences are based on a large human population.
3 Studies have adopted a number of methods to infer exposure to tetrachloroethylene.
4 Exposure was ascertained in many cases indirectly by questionnaire, by job title, or by the
5 subject living in a residence receiving drinking water containing tetrachloroethylene. There is
6 higher confidence of exposure potential to tetrachloroethylene in those studies using the
7 occupational title of dry cleaner because tetrachloroethylene is the solvent of choice.
8 Atmospheric montoring of dry cleaning facilities show 8-hr TWA concentrations in the range of
9 10-20 ppm, with short-term exposures many times this value. However, the analyses that
10 combined dry cleaners with laundry workers carry more uncertainty than do studies whose
11 analyses included only dry cleaners, because laundry workers have a lower probability for
12 exposure to tetrachloroethylene. More rarely, studies incorporated biological measures such as
13 tetrachloroethylene excretion in breath or urinary TCA.
14 When looked for, exposure-response gradients have not been observed, generally, and
15 this is another uncertainty associated with the inferences regarding a causal association.
16 However, exposure misclassification may partially explain the lack of exposure-response
17 associations, hence, the lack of an exposure-response gradient does not diminish the observed
18 associations between tetrachloroethylene exposure and adverse effects. Moreover, observed
19 effects cannot be considered to arise from confounding; investigators have taken great effort to
20 take into account the effects of smoking, age, and other factors through matching exposed
21 subjects with like controls or through statistical analysis of the data. This synthesis of the
22 epidemiologic evidence places greatest weight on those studies where confounding has been
23 adequately controlled and identifies those studies where confounding may be a possible
24 explanation for observed results. It is not possible to examine residual confounding or effects
25 not explained by variables adjusted for in the study's design or statistical analysis. Residual
26 confounding is an issue for both epidemiologic and toxicologic studies and may explain
27 observed study findings.
28 In general, observations from the epidemiologic studies are consistent with the biology of
29 tetrachloroethylene. Tetrachloroethylene is lipophilic and would distribute to organs rich in
30 lipids, e.g., the CNS. The liver and kidney are considered target organs due to their ability to
31 metabolize tetrachloroethylene. Moreover, systemic effects, specifically to the kidney, liver,
32 CNS, and the developing organism, have been observed in experimental animals. Thus, humans
33 do not appear as an exception to the systemic toxicity elicited by tetrachloroethylene. The
34 pattern of effects seen with tetrachloroethylene exposure is similar to that seen with other
35 solvents, such as trichloroethylene. These findings together support a causal role of
36 tetrachloroethylene in the development a number of systemic effects in humans.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-183 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 4.10.1.1.2. Human cancer effects. Overall, the epidemiologic evidence considered as a whole
3 has associated tetrachloroethylene exposure with excess risks for a number of site-specific
4 cancers. Studies of tetrachloroethylene and cancer showed positive associations between
5 exposure and cancer of the lymphoid system, esophagus, and cervix, with more limited evidence
6 for cancer of the bladder, kidney, and lung. For both lymphoid and esophageal cancer, excess
7 risk was observed in studies of dry cleaners and in degreasers, populations who have exposure to
8 tetrachloroethylene and other solvents. In both cases, average risks were doubled as compared
9 with those of referents. Studies of drinking water exposure also support an association between
10 lymphoid cancer and tetrachloroethylene and other solvents, as do case-control studies that
11 assessed employment as a dry cleaner or laundry worker. Chance and confounding by smoking
12 are unlikely explanations for the observed excesses is risks. Furthermore, the finding of elevated
13 risk for lymphohematopoietic system cancer incidence in a Swedish cohort of subjects who
14 developed suspected solvent-related disorders from organic solvent exposures supports the
15 findings of the tetrachloroethylene studies (Berlin etal., 1995). EPA judged that these data,
16 though limited and not consistently observed across all studies, suggested an association between
17 lymphoma and tetrachloroethylene.
18 For esophageal cancer, indirect evidence suggests that esophageal risk in these studies is
19 larger than that expected due to smoking. Blair et al. (2003) stated that if the magnitude of the
20 difference in smoking for dry cleaners and the general population is in the range of 10% or less,
21 confounding from smoking in their study is unlikely to result in an RR greater than 1.2, a finding
22 similar to that of Kriebel et al. (2004). Hence, the finding of a doubling in risk strongly suggests
23 occupational exposure as a contributing (etiologic) factor. Observations from one case-control
24 study that was able to adjust for the effects of smoking support the cohort study findings.
25 The epidemiologic evidence also is suggestive of excess risks for cervical cancer, based
26 on observations in dry cleaner and laundry worker cohorts, with few cases in the degreaser
27 studies. Unfortunately, information is not available on possibly confounding factors such as
28 socioeconomic and lifestyle factors. Associations with kidney, bladder, and lung cancers and dry
29 cleaning employment or, more specifically, with tetrachloroethylene, were reported in recent
30 updates of the American and Nordic cohorts and in case-control studies. Conclusions are more
31 uncertain for these sites, because they are based either on heterogenous observations between
32 differing study designs or on a small number of available studies. Overall, EPA judged these
33 findings as suggestive of an association.
34 Other reviews of the tetrachloroethylene epidemiologic evidence have concluded that
35 "little consistent evidence existed for an association with a specific cancer such as kidney"
36 (McLaughlin and Blot, 1997) or that there is "limited evidence" for cancers of the cervix,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-184 DRAFT-DO NOT CITE OR QUOTE
-------
1 esophagus, bladder, or kidney or for NHL (IARC 1995; Weiss 1995; Lynge et al., 1997; Ulm et
2 al., 1996). The Institute of Medicine (IOM) reviewed a similar—but not the same—body of
3 epidemiologic literature as EPA (IOM, 2002). For example, the large four-country Nordic
4 cohort of dry cleaners and laundry workers in Andersen et al. (1999) was not considered, nor was
5 the most recent update of the cohort of American dry cleaner and laundry workers in Blair et al.
6 (2003). The IOM committee concluded limited or suggestive evidence of an association between
7 bladder and kidney cancers and tetrachloroethylene and dry cleaning solvents. No conclusions
8 were presented on the epidemiologic evidence on esophageal and lung cancers and
9 tetrachloroethylene, given that the committee could not reach a consensus opinion. Some
10 committee members believed that the overall evidence was limited by potential confounding
11 from smoking in cohort studies, whereas other committee members considered, for esophageal
12 cancer, the lack of other smoking-related cancers in cohort studies or, for lung cancer, the
13 presence of exposure-response relationships as supportive of an conclusion of limited/suggestive
14 evidence. For other cause-specific cancers, the committee concluded that there was inadequate
15 or insufficient evidence to determine whether an association existed.
16 U.S. EPA's analysis is similar to those of the IOM committee's on kidney and bladder
17 cancer, and, like some committee members, EPA considered the evidence on esophageal and
18 lung cancers as suggestive of an association. U.S. EPA's conclusions on lymphoma are
19 supported, in part, by observations from studies not considered by the IOM committee.
20 Mundt et al. (2003) reviewed a body of epidemiologic studies similar to U.S. EPA's and
21 presented conclusions as to whether an association was "likely" or "not likely." The authors
22 reported that little support existed on which to base a conclusion that tetrachloroethylene was a
23 strong occupational risk factor, but that "because of a number of positive findings suggested
24 from some of these epidemiological studies, one cannot definitely rule out the possibility that
25 associations between PCE [tetrachloroethylene] and some cancers exist in humans." This
26 conclusion is consistent with conclusions in this assessment, although it is expressed differently.
27 Although epidemiologists acknowledge that using guidelines to assess causation is an
28 imperfect process, some find that the aspects developed by A.B. Hill (1965) are helpful in
29 making these difficult judgments. Making this determination may precede an understanding of
30 the underlying mechanisms and involves consideration of several aspects that would be
31 characteristic of a cause-and-effect relationship (Hill, 1965; Rothman and Greenland, 1998).
32
33 1. Strength of the observed association. The finding of large and precise risks increases
34 confidence that the association is likely not due to chance, bias, or other factors. For
35 tetrachloroethylene, observed risks are generally modest, 2-fold or less. The observed
36 risks for esophageal cancer are not thought to be attributable to smoking or alcohol,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-185 DRAFT-DO NOT CITE OR QUOTE
-------
1 although insufficient data exist on socioeconomic factors important to cervical cancer to
2 assess their impact on observed elevated risks for these site-specific cancers.
3 2. Consistency of the observed association. An inference of causality is strengthened when
4 a pattern of elevated risks is observed across several independent studies. Excess risks
5 for lymphoid cancers are seen in studies of dry cleaners, degreasers, and populations
6 exposed to drinking water containing tetrachloroethylene—and in some cases
7 trichloroethylene—and for esophageal cancer in studies of dry cleaners and degreasers.
8 Excess risks for the other site-specific cancers are less consistently observed across these
9 populations.
10
11 3. Specificity of the observed association. Traditionally, specificity has been described in
12 terms of one cause, one disease (Hill, 1965). This implies that one factor is associated
13 with the observed effect and no other effects are associated with the putative factor.
14 Tetrachloroethylene causes cancer at several sites in rats and mice; hence, there is no
15 expectation that tetrachloroethylene would be associated with only one human cancer.
16 Furthermore, many agents cause cancer at multiple sites, and many cancers have multiple
17 causes. Specificity has little meaning in this case, and therefore the lack of specificity
18 does not detract from the weight of the overall epidemiologic evidence.
19
20 4. Temporal relationship of the observed association. Causal relationships have temporality,
21 i.e., the cause precedes the effect. Associations between tetrachloroethylene exposure
22 and several forms of cancer are established primarily by cohort and case-control studies,
23 in which the temporal relationship is well described. Many drinking water studies are
24 ecologic or prevalence studies, in which knowledge of the temporal relationship is
25 lacking. The exceptions are those studies assessing exposure to residents of Cape Cod,
26 MA, Woburn, MA, and Camp Lejeune, NC. For this reason, the conclusions place
27 greater weight on the cohort and case-control studies.
28
29 5. Biological gradient (exposure-response relationship). A clear exposure-response
30 relationship often suggests cause and effect. For tetrachloroethylene, biological gradients
31 are only sporadically observed, though most studies identify exposure only as a
32 dichotomous variable (yes/no), or the number of site-specific cancers is often too small to
33 identify biological gradients. For esophageal cancer, the dry cleaner studies of
34 tetrachloroethylene exposure (Blair et al., 2003; Ruder et al., 2001) showed no clear
35 picture of exposure response relationships. Exposure response analyses in the drinking
36 water studies collectively suggest that greater exposure to drinking water contaminated
37 with tetrachloroethylene—and in a smaller number of studies, with chlorinated solvents
38 both tetrachloroethylene and trichloroethylene—is associated with lymphoid cancer,
39 particularly leukemia and NHL (Aschengrau et al., 1993; MA DPH 1997; Fagliano et al.,
40 1990; Cohnetal., 1994).
41
42 6. Biological plausibility. The mechanistic studies (discussed in another section in this
43 assessment) investigating tetrachloroethylene carcinogenic effects in rats or mice and
44 their relevance to humans indicate that carcinogenesis is complex and likely involves
45 multiple mechanisms. Overall, the MOA for site-specific cancer is not know at this time.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-186 DRAFT-DO NOT CITE OR QUOTE
-------
1 7. Coherence. Coherence means that the causal interpretation of the data should not
2 seriously conflict with generally known facts about the natural history and biology of the
3 disease. The strongest associations between tetrachloroethylene and human cancer are
4 for the lymphopoietic system and esophagus, with more limited evidence for cervix,
5 bladder, kidney, and lung. Several of these organ systems are also targets for noncancer
6 toxicity. The associations between cervical and esophageal cancer have no suitable
7 animal counterparts.
8
9 8. Experimental evidence (from human populations). Experimental evidence (e.g., a
10 "natural experiment" that measures effects with exposure and in the absence of exposure)
11 is seldom available from human populations and exists only when conditions of exposure
12 are altered to create a kind of quasi-experiment. There are few data to evaluate this
13 criterion. The only study that does present information notes that childhood leukemia
14 cases appeared to be more evenly distributed throughout Woburn, MA, after closure of
15 the two wells contaminated with trichloroethylene and tetrachloroethylene (MA DPH,
16 1997).
17
18 9. Analogy. The pattern of effects associated with tetrachloroethylene, particularly cancers
19 of the lymphoid system, cervix, kidney, and pancreas, has similarities to that of several
20 other chlorinated solvents and to mixed-solvent exposures.
21
22 Together, the evidence on tetrachloroethylene partially fulfills several of these criteria
23 and is suggestive of a cause-and-effect relationship between tetrachloroethylene and human
24 cancer. The body of human evidence is not sufficient to regard tetrachloroethylene as a known
25 human carcinogen.
26
27 4.10.1.1.3. Susceptibility. Many of tetrachloroethylene's metabolites are formed through the
28 enzyme system that also metabolizes ethanol and other drugs and environmental pollutants.
29 Exposures to these chemicals can alter tetrachloroethylene's toxicity, not only altering the
30 pharmacokinetics of tetrachloroethylene but also the pharmacodynamics for toxicity. For
31 tetrachloroethylene's effects on the nervous system, kidney, and liver, the available limited data
32 suggest that a joint effect that reflects the addition of all exposures is plausible. In addition,
33 susceptibility to tetrachloroethylene's toxicity may vary among individuals because of both
34 intrinsic factors, (age, sex, and genetic factors, including metabolic polymorphisms) and
35 acquired factors (disease status, nutritional status).
36
37 4.10.1.2. Summary of Effects in Animals
38 4.10.1.2.1. Animal noncancer effects. Tetrachloroethylene exposure in animals results in
39 toxicity to the liver, kidney, and nervous system and also causes developmental and reproductive
40 effects. These are all sites of high metabolic activity, and the CNS is also a lipid accumulation
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-187 DRAFT-DO NOT CITE OR QUOTE
-------
1 site. The immune system is potentially affected, but there are very few studies of these effects,
2 and none of them are in intact animals. No information is available on the effects of
3 tetrachloroethylene on the endocrine system in animals. The effects have been discussed in
4 several other review documents and are described in Sections 4.4.2.1, 4.5.2.1, 4.6.2.1, 4.6.2.2,
5 and 4.7.2.1.
6 In the liver, several measures of toxicity have been observed, such as increased liver
7 weight, infiltration of fat, necrosis, peroxisome proliferation, polyploidy of hepatocytes, and
8 increased triglycerides. In kidney, increased weight, hyperplasia, hyaline droplets, and protein
9 cast formation in tubules have been observed. In the CNS, alteration of brain neurotransmitter
10 levels, increased motor activity, and delayed reaction times to visual stimuli have been observed.
11 Fetal growth retardation, increased fetal mortality, and behavioral changes occurring after birth
12 to animals exposed in utero have been observed. Section 4.10.1.3 describes the doses at which
13 these effects occurred.
14
15 4.10.1.2.2. Animal cancer effects. Carcinogen bioassays in rats and mice have shown benign
16 and malignant tumors at various sites, as summarized in Sections 4.4.2.2, 4.5.2.2, and 4.8.2.
17 Data on the incidence and dose levels at which these effects occur are presented in Section 5.3.2
18 and Tables 5-6, 5-8, and 5-9 (Chapter 5). One study that used the oral route found liver
19 adenomas and carcinomas in mice. Two inhalation bioassays, both of which used both mice and
20 rats, found tetrachloroethylene-induced excess incidence of hepatocellular adenomas and
21 carcinomas (mice) and mononuclear cell leukemia (rats). One of these studies found
22 hemangioendothelioma (mice) and the other found brain glioma, kidney tubular cell tumors, and
23 testicular interstitial cell tumors, all in male rats. Brain and kidney tumors are rare in unexposed
24 animals, but they were found to be slightly elevated above control levels in only one of the two
25 inhalation studies.
26 Testicular tumors are extremely common in control animals, and the statistically
27 significant elevation in one of the bioassays was not considered by the investigators as related to
28 tetrachloroethylene exposure; however, the testicular tumors and kidney tumors are consistent
29 with exposure of rats to trichloroethylene, a structural analogue of tetrachloroethylene.
30 Mononuclear cell leukemias in rats were elevated in both inhalation bioassays. As discussed in
31 Section 4.8.2.4.1, this is a relatively common tumor in nonexposed animals, with a possible site
32 concordance with human lymphoid cancer, but because the mechanism of formation for both
33 human and animal hematopoietic neoplasms is not understood in all of its complexity, they are of
34 possible relevance to humans.
35
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-188 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
4.10.1.3. Summary of Effect Levels
Table 4-12 summarizes the lower ranges of air concentrations and oral doses at which
effects occur in each of the organ systems discussed in this document. The lowest of these air
concentrations is 0.7 ppm (mean), which is associated with neurological effects observed in
residents living above dry cleaning facilities (Altmann et al., 1995) in Germany. This is close to
the mean concentration measured by Schreiber et al. (2002) for a similar exposure situation in
the United States (0.4 ppm). The lowest concentration showing effects in animals is 9 ppm,
where liver toxicity was observed in mice which are more sensitive than the rat test strains.
Table 4-12. Summary of low-effect levels of exposure to tetrachloroethylene
Organ System
Liver
Kidney
Neurological
Developmental,
reproductive
Other organs
Humans
Inhalation
(ppm)
12-16 (Table 4-1)
1.2 and 8. 8
(Table 4-3)
0.3 (Table 4-5)
1.2 (Table 4-8)
Exposure uncertain0
Oral
(ug/kg/day)
-
-
6, uncertain
(Table 4-8)
-
Animals
Inhalation
(ppm)
9-50, mice
(Table 4-2)
100 for cancer in
micea
100 ppm, miceb
37-90, mice and
gerbils (Table 4-6)
100, rats
(Table 4-9)
No conclusiond
Oral
(mg/kg/day)
100 (Table 4-2)
386 in mice for
cancer3
No studiesb
No chronic study
(Table 4-7)
-
-
12
13
14
15
16
17
18
19
20
21
22
23
24
25
a See Section 4.4.2.2.
b See Section 4.5.2.1.
c See Section 4.8.1.
d See Section 4.8.2.
- = No studies available
For subchronic oral exposures, the lowest dose for which adverse effects occurred in
animals is 100 mg/kg-day. These data come from the Buben and O'Flaherty (1985) gavage
study, where the exposure duration was 6 weeks. There are no reliable data for humans exposed
orally.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-189 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.10.2. Characterization of Cancer Hazard
2 Tetrachloroethylene is "Likely to be a human carcinogen by all routes of exposure"
3 within the framework of the 2005 carcinogen risk assessment guidelines (U.S. EPA, 2005b).
4 This conclusion is based on reported associations in epidemiologic studies between
5 tetrachloroethylene exposure and site-specific cancers and by the induction of site-specific
6 tumors in rodents given tetrachloroethylene by oral gavage and inhalation. Several metabolites
7 of tetrachloroethylene also are considered rodent carcinogens. Metabolites from the oxidative
8 pathway, TCA and DCA, produce liver tumors in mice, and DCA also induces liver tumors in
9 rats. Metabolites from the GST pathway have not been tested in a standardized 2-year bioassay.
10 This hazard characterization is discussed in more detail in Section 4.10.2.2. The context for this
11 statement is described in the following section.
12
13 4.10.2.1. Background
14 As specified in the guidelines, the descriptor "Likely to be carcinogenic to humans"
15 expresses the conclusion regarding the weight of evidence for carcinogenic hazard potential, and
16 it is presented only in the context of a weight of evidence narrative. Although the term "likely"
17 can have a probabilistic connotation in other contexts, its use as a weight of evidence descriptor
18 does not correspond to a quantifiable probability of whether the chemical is carcinogenic. The
19 five recommended standard hazard descriptors are as follows:
20 "Carcinogenic to humans"
21 "Likely to be carcinogenic to humans"
22 "Suggestive evidence of carcinogenic potential"
23 "Inadequate information to assess carcinogenic potential"
24 "Not likely to be carcinogenic to humans"
25
26 These descriptors are not unlike those used by the IARC, NTP, and other health agencies
27 that weigh carcinogenicity evidence. If there are no or insufficient pertinent data, then the
28 descriptors "Inadequate information to assess carcinogenic potential" or "Suggestive evidence of
29 carcinogenic potential" are used. If the evidence is stronger, as is the case with
30 tetrachloroethylene, the descriptor "Likely to be carcinogenic to humans" is used; convincing
31 evidence, usually conclusive demonstration of causality in epidemiological studies, would
32 support "Carcinogenic to humans." On the other hand, if the conclusion is negative (i.e., strong,
33 consistent and compelling information indicating the absence of human health hazard), the agent
34 would be described as "Not likely to be carcinogenic to humans." Thus, going down the list of
35 descriptors from "Carcinogenic to humans" to "Inadequate information to assess carcinogenic
36 potential" indicates a decrease in the level of evidence or of a human health hazard. In summary,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-190 DRAFT-DO NOT CITE OR QUOTE
-------
1 use of the weight of evidence descriptor "Likely to be carcinogenic to humans" for
2 tetrachloroethylene is intended to communicate that the available information indicates the
3 presence of a human health hazard.
4 The weight-of-evidence conclusion represented by the top three levels of evidence is
5 related to but distinct from the quantitative dose-response assessment/conclusions in that the
6 judgment that an agent is a human carcinogen does not guarantee adequate data to quantitatively
7 estimate human risk. Notably, evaluation of an agent that is judged a likely human carcinogen
8 may offer data conducive to estimating human risk. Indeed, dose-response assessments are
9 generally completed for agents considered "Carcinogenic to humans" and "Likely to be
10 carcinogenic to humans." Section 5.4 provides the dose-response analyses for
11 tetrachloroethylene.
12
13 4.10.2.2. Hazard Characterization for Tetrachloroethylene
14 Overall, the epidemiologic evidence considered as a whole has associated
15 tetrachloroethylene exposure with excess risks for a number of site-specific cancers. Lymphoid
16 cancer is now recognized as a combination of NHL, Hodgkin's disease, lymphosarcoma,
17 multiple myeloma, and lymphatic leukemia. Cohort studies of dry cleaner and laundry workers
18 and of degreasers suggest excess risks of lymphoid cancers, as do case-control studies of
19 drinking water exposure and occupational exposure. Exposure to a number of solvents is likely
20 in most of the case-control studies; however, these solvents have a qualitatively similar profile of
21 metabolites, although quantitative differences are expected. One study of exposure only to
22 tetrachloroethylene in drinking water reported a statistically significant association, based on a
23 small number of exposed cases, between leukemia and a residence receiving tetrachloroethylene-
24 contaminated water (Aschengrau et al., 1993).
25 Both cohort and case-control studies of dry cleaning workers support an association
26 between tetrachloroethylene and excess risk of esophageal cancer. Recent updates of dry
27 cleaners and laundry worker cohorts (Ruder et al., 2001; Blair et al., 2003) carry great weight in
28 this evaluation because dry cleaners are predominately exposed to tetrachloroethylene and a
29 statistically significant elevated mortality from this cancer continued to be observed. Little
30 weight is given to the Lynge et al. (2006) study due to potential biases that likely dampen their
31 observations and because of these biases it is considered a null study. No clear patterns are seen
32 in the tetrachloroethylene studies for either level or duration of exposure and response.
33 The possibility that other exposures such as smoking and alcohol consumption may
34 potentially confound the associations observed in Blair et al. (2003) and Ruder et al. (2001)
35 cannot be directly addressed. Indirect evidence suggests that the esophageal risk in these studies
36 is larger than that expected due to smoking. Moreover, the case-control study by Vaughan et al.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-191 DRAFT-DO NOT CITE OR QUOTE
-------
1 (1997) provides support for an association with tetrachloroethylene; a statistically significant
2 association was observed between tetrachloroethylene exposure and esophageal cancer after
3 adjustment for smoking, alcohol, and socioeconomic status. Support by analogy is derived from
4 the finding of excess esophageal cancer incidence in a cohort that was occupationally exposed to
5 trichloroethylene (Hansen et al., 2001).
6 More deaths from cervical cancer were observed among American and Nordic female dry
7 cleaners or laundry workers than were expected. The observation of exposure-response trends in
8 the studies that presented this information (Blair et al., 1990; Ruder et al., 1994, 2001) support an
9 association with dry cleaning. Lack of data on socioeconomic status—a proxy for exposure to
10 the human papilloma virus, a known risk factor for cervical cancer—indicates great uncertainty
11 for asserting this association with tetrachloroethylene exposure.
12 There is also some support, albeit less than for the sites above, for an association between
13 dry cleaning occupations and other cancers, specifically, cancers of the kidney, bladder, and lung.
14 These findings are based on heterogenous observations of differing study designs, on a small
15 number of available studies, or on small numbers of study subjects.
16 An open question in the dry cleaner studies is the specificity of exposure to
17 tetrachloroethylene. Elevated mortality for cancer of the esophagus and cervix were observed in
18 two cohorts that were considered to have primarily tetrachloroethylene exposures. However,
19 individuals who may have had exposures to other dry cleaning solvents were also included in
20 these studies. There are only three studies of cancer incidence or mortality among degreasers
21 exposed to tetrachloroethylene, and they are of a small number of subjects with
22 tetrachloroethylene exposure and, consequently, of few site-specific cases. These studies are
23 only now collectively beginning to provide insight on associations between tetrachloroethylene
24 exposures and site-specific cancers.
25 In rodents, hepatocellular carcinomas in both male and female B6C3F1 mice have been
26 observed following inhalation and oral gavage exposure, and the same tumor response was
27 observed in male and female Crj :BDF1 mice after inhalation exposure. MCL, a common tumor
28 site in treated and untreated F344 rats, was significantly increased in both males and females in
29 inhalation bioassays carried out in both Japan and the United States. Malignant liver
30 hemangiosarcomas and splenic hemangioendotheliomas were also observed in male mice in the
31 Japan bioassay. In the U.S. inhalation study in F344 rats, a small excess incidence of rare renal
32 tubule cell carcinoma and adenoma was observed in males. Testicular interstitial cell tumors, a
33 common tumor in treated and untreated F344 rats, were significantly elevated in the U.S.
34 bioassay, and an elevation of rare brain glioma incidence was also observed in these rats.
35 The major metabolite of tetrachloroethylene in humans and rodents, TCA, is carcinogenic
36 by gavage in male mice, and another metabolite in rodents, DCA, is also carcinogenic by gavage
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-192 DRAFT-DO NOT CITE OR QUOTE
-------
1 in male mice. The MO A of tetrachloroethylene or its metabolites in the likely causation of
2 cancer is not known. Extensive testing of tetrachloroethylene showed that it does not damage
3 DNA except in a few studies of conditions where the GSH metabolites would be generated, and
4 it induces chromosome aberrations in some studies. Several of the known or putative oxidative
5 metabolites are mutagenic. Metabolism through kidney GSH conjugation produces
6 trichlorovinyl GSH and trichlorovinyl cysteine, which were mutagenic in the Salmonella test but
7 which have not been tested in mammalian genotoxicity assays. The latter metabolite reacts with
8 beta lyase in the kidney to produce reactive thiol compounds. Other metabolites, including
9 reactive sulfoxides, can be also be produced by FMO3 or CYP3 A metabolism of TCVC. This is
10 a plausible MOA for the rare rat kidney tumors observed in one bioassay. However, the MO As
11 for human tumors and the mice liver tumors is still unknown. Therefore, there is little
12 mechanistic basis for choosing a low-dose extrapolation model.
13
14 4.10.3. Mode-of-Action Summary
15 The MOA for tetrachloroethylene-induced carcinogenesis is not yet fully characterized,
16 completely tested, or understood. The database for hepatocarcinogenesis is especially limited
17 with regard to chemical-specific studies. The available evidence points to multiple MO As being
18 involved. Furthermore, although there is some evidence for common MOAs, there is also
19 evidence indicating differences in the potential MOA across organ systems.
20 Tetrachloroethylene exposure has been associated with peroxisome proliferation in
21 rodent liver and kidney. Compelling insight into the hypothesized MOA by which certain
22 chemicals induce proliferation of peroxisomal organelles and possibly cancer—specifically, the
23 focus has been on liver cancer—was disclosed by the discovery of the PPAR receptors, a class of
24 nuclear receptors closely related to the thyroid hormone and retinoid receptors that were first
25 shown to be activated by peroxisome proliferators by Issemann and Green (1990). To date, three
26 known subtypes of PPAR have been described in mammals: PPAR gamma, PPAR-5, and
27 PPAR-a.
28 Evidence exists to support PPAR-a as being the specific receptor that is necessary for
29 transient cell proliferation and its role in hepatocarcinogenesis has been the subject of several
30 investigations, although most studies have explored the potent agonist Wy-14,643 (Lee et al.,
31 1995; Peters et al., 1997a; Corton et al., 2000). Activation of the steroid-like PPAR receptor
32 regulates transcription of the genes. The PPAR target genes encode enzymes involved in
33 peroxisomal and mitochondrial beta-oxidation and ketone body synthesis as well as certain P450
34 4A enzymes, fatty-acid binding proteins, apolipoproteins, lipoprotein lipase, malic enzyme, and
35 phosphoenolpyruvate carboxykinase (Issemann et al., 1993; Desvergne and Wahli, 1995; Reddy
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-193 DRAFT-DO NOT CITE OR QUOTE
-------
1 et al., 1986). The PPAR genes are expressed in a wide range of tissues, and PPAR occurs across
2 species.
3 Several recent studies have expanded the scientific understanding of the PPAR-a mode of
4 action proposed by Klaunig et al (2003; see Caldwell et al., 2008). First, Yang et al (2007)
5 demonstrated that PPAR-a activation in hepatocytes induces peroxisome proliferation but not
6 liver tumors. The approach entailed targeting expression of PPAR-a to hepatocytes by placing
7 the VP16 PPAR-a transgene gene under control of the liver enriched activator protein (LAP)
8 promoter. LAP-VP16 PPAR-a transgenic mice showed a number of PPAR-a -mediated effects:
9 decreased serum triglycerides and free fatty acids, peroxisome proliferation, enhanced
10 hepatocyte proliferation, and induction of cell-cycle and PPAR-a target genes. However,
11 compared with wild-type mice exposed to Wy-14,643, the extent of hepatomegaly was reduced
12 and no hypertrophy or eosinophilic cytoplasms was seen in LAP-VP16 PPAR-a mice. Also in
13 contrast with wild-type mice exposed to Wy-14,643, no evidence of non-parenchymal cell
14 proliferation was observed in the LAP-VP16 PPAR-a transgenic mice. Moreover, at one year of
15 age no evidence of preneoplastic hepatic lesions or hepatocellular neoplasia was observed in
16 LAP-VP16 PPAR-a transgenic mice. As noted by the authors, PPAR-a activation only in mouse
17 hepatocytes is sufficient to induce peroxisome proliferation and hepatocyte proliferation but
18 "... is not sufficient to induce liver tumors."
19 Secondly, Ito et al. (2007) found that DEHP, a proposed robust example of PPAR-a
20 agonism-induced hepatocarcinogenesis, yields liver tumors in a 2-year study in PPAR-a knock-
21 out mice. This study demonstrates the limitations, cited by the FIFRA SAP, of drawing
22 conclusions from the one-year bioassays of high doses of Wy-14,643 referenced above (e.g.,
23 Peters 1997). It supports the view that knock-out mouse bioassays should be carefully
24 characterized and conducted for 2 years to assess whether PPAR-a activation is indeed necessary
25 for induction of liver cancer. Thus, although a weak peroxisome proliferator, chemical-specific
26 data supporting the hypothesis that PPAR-a activation plays a prominent or essential role in
27 tetrachloroethylene tumor induction are lacking. Critical review of the scientific literature
28 reveals significant data gaps regarding the relationship between the PPAR-a activation and
29 neoplasia induced by peroxisome proliferators as a group and tetrachloroethylene specifically. If
30 PPAR-a does play a role in tetrachloroethylene-induced tumorigenesis, available information
31 suggests relevance to humans cannot be ruled out.
32 Although accumulation of alpha-2u-globulin has been suggested as an MOA leading to
33 nephropathy that culminates in the formation of renal tumors, the available data do not support
34 this MOA for tetrachloroethylene. Indeed, the available data suggest that alpha-2u-globulin
35 accumulation following tetrachloroethylene exposure occurs only at doses higher than those used
36 in the carcinogenicity bioassays. In addition, tetrachloroethylene does not meet all the criteria to
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-194 DRAFT-DO NOT CITE OR QUOTE
-------
1 suggest that alpha-2u-globulin accumulation is the MOA. Therefore, an important role for
2 alpha-2u-globulin accumulation in tetrachloroethylene-induced renal tumors is highly unlikely.
3 The role of genotoxicity in tetrachloroethylene liver cancer, an effect that is thought to be
4 related to products of CYP metabolism, is uncertain. The available data suggest that several of
5 the chloroacid metabolites are mutagenic. In particular, tetrachloroethylene oxide, the primary
6 metabolite hypothesized to be formed during CYP metabolism, is a known bacterial mutagen; it
7 has not been tested in mammalian systems, although genotoxicity in such tests could be
8 anticipated based on the expected DNA reactivity of the epoxide moiety. GSH-derived
9 intermediates also exhibit genotoxicity. The glutathione conjugation of tetrachloroethylene in
10 the kidney leads sequentially to S(l,2,2-trichlorovinyl)glutathione and
11 S(l,2,2-trichlorovinyl)cysteine—TCVG and TCVC. TCVC can be further processed by beta-
12 lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a highly reactive
13 thioketene which can form covalent adducts with cellular nucleophiles including DNA. TCVC
14 can also undergo FMO3 or P450 oxidation to reactive intermediates; additionally, sulfoxidation
15 of both TCVC and its N-acetylated product occurs, resulting in reactive metabolites (Ripp et al,
16 1997, 1999; Werner et al., 1996). While most of these intermediates have not been characterized
17 for mutagenic potential, TCVG, TCVC and NAcTCVC are clearly mutagenic in Salmonella tests.
18 In addition, tetrachloroethylene exhibited mutagenicity in Salmonella in the few studies of
19 conditions that could generate GSH-derived metabolites and, following in vivo exposures,
20 induces SSB and DNA binding in kidney. A mutagenic MOA is therefore likely to play a role in
21 the development of tetrachloroethylene-induced renal cancer. A mutagenic MOA could also
22 play a role in the formation of other tumors such as brain gliomas and MCL in rats, if sufficient
23 concentration of the potentially genotoxic metabolites arising from the initiation of GSH
24 conjugation occurs at target sites. In the kidney, the conjugates are concentrated after being
25 transported from the liver, in addition to being generated on site in that target tissue. Further
26 processing by beta lyase, FMO3 or CYPs yields mutagenic products in what could be
27 sufficiently high target tissue concentrations. Beta lyase is also found in the brain and other
28 tissues.
29 In summary, cancers resulting from tetrachloroethylene exposures are likely due to
30 multiple MO As that vary from target tissue to target tissue. The MO As for tetrachloroethylene-
31 induced cancers are not yet well understood. The potential MO As for cancer discussed in this
32 section are summarized in Table 4-13 below. The implications of these potential MO As to risk
33 extrapolation to concentrations lower than those producing effects in animal bioassays are
34 explored in Table 4-14. These considerations underlie the discussion in Chapter 5 of the
35 uncertainties in modeling risk at low concentrations.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-195 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-13. Summary of potential modes of action for cancer
Organ, tumor type
Kidney
adenocarcinoma in
male rats.
Potential MOA
Mutagenicity. Via the GSH pathway in
liver and kidney, glutathione and
cysteine conjugates are produced.
Identification of the urinary
mercapturate metabolite in humans and
rodents shows that this pathway is
operative in both species.
Tubular cell necrosis and
nephrotoxicity followed by hyperplasia
and neoplastic transformation.
Accumulation of arpha-2u-globulin, a
male rat specific protein, in hyaline
droplets of tubule cells leading to a
specific pattern of nephropathy
(involving tubule degeneration,
compensatory hyperplasia, neoplastic
transformation).
PPAR-a receptor activation in kidney.
Evidence for MOA
Reactive thiol and
sulfoxide compounds
are formed in kidney
from TCVG/TCVC
that are mutagenic.
In vitro kidney damage
from processing of
GSH metabolites to
cytotoxic species.
Hyaline droplets
observed in short-term
(42-day) experiments
BUT ONLY at high
dose (1,000 mg/kg-
day).
Peroxisome
proliferation occurs in
kidney, but no
evidence supports
causal association with
tumorigenesis
Limitations/evidence against MOA
Several metabolites (TCVG/TCVC,
NAcTCVC) are mutagenic in bacteria
but are not characterized in other
systems; other GSH metabolites have
not been identified and characterized.
Mutagenicity is commonly assumed to
contribute to cancer.
Relationship between non-genotoxic
kidney necrosis and carcinogenicity has
not been studied.
Tetrachloroethylene-induced renal
nephropathy also occurs in female rats
and mice of both sexes. Hyaline
droplets are not observed in 28-day
experiments at lower concentrations
(400 ppm) that induce kidney tumors.
alpha-2u-globulin protein not
identified. Two features characteristic
of this MO A do not occur: (1)
mineralization of tubules in the chronic
bioassay; (2) male-rat-only
nephrotoxicity.
Peroxisome proliferation is greater in
mouse kidney than rat kidney; however
male rats develop kidney tumors.
Weight of evidence
Experimental evidence
for this pathway and
identification of its
urinary mercapturate
metabolites supports
the MOA.
Inadequate
tetrachloroethylene-
specific data exist to
support this potential
MOA.
Kidney tumors are not
specific to the male rat.
This MOA only occurs
at concentrations
greater than those
necessary for tumor
induction.
Little evidence to
support this MOA.
-------
o
Oi
Table 4-13. Summary of potential modes of action for cancer (continued)
This document is a draft for review purposes only and does not constitute Agency policy
06/08 4-197 DRAFT: DO NOT CITE OR QUOTE
Organ, tumor type
Liver hepatocellular
carcinoma in male and
female mice.
Liver and spleen
hemangiosarcoma in
male mice.
Blood mononuclear
cell leukemia in male
and female rats.
Brain glioma in male
rats. (Rare)
Potential MOA
PPAR-a receptor activation
Mutagenicity produced by P450
pathway in liver.
Necrosis of liver cells leads to
compensatory hyperplasia, excess
mutation rate and neoplastic
transformation.
Mutagenicity of P450 and/or GSH-
derived reactive intermediates.
Mutagenicity. There is a potential for
circulating genotoxic metabolites, or
their precursors produced by liver
metabolism, to be further processed to
reactive products in other tissues.
Perc incorporation into brain cell
membranes, disruption of membrane
ion channels, interference with
neurotransmitters.
Evidence for MOA
PPAR-a is found in
human liver, testis,
pancreas.
Tetrachloroethylene
oxide and DCA are
mutagenic and other
metabolites
(trichloroacetyl
chloride) are capable of
binding to DNA.
No evidence, but
plausible hypothesis.
Tetrachloroethylene
oxide and DCA are
mutagenic and other
metabolites
(trichloroacetyl
chloride) are capable of
binding to DNA.
DCVC, a metabolite of
trichloroethylene,
induced blood
dyscrasias (aplastic
anemia, DNA alteration
in bone marrow, lymph
nodes and thymus in
calves.
Perc induces persistent
changes in fatty acid
composition of brain in
rodents.
Limitations/evidence against MOA
Perc1 only weakly activates PPAR-a.
There are no studies in PPAR-a null
mice evaluating PERC-induced tumors.
Genotoxicity is commonly assumed to
contribute to carcinogenesis.
Cytotoxicity and compensatory
hyperplasia are not observed at
bioassay doses.
Genotoxicity is commonly assumed to
contribute to carcinogenesis.
TCVC, the analogous perc metabolite,
did not produce bone marrow injury in
calves in a single study.
No information about cell membrane
changes at low doses or longer than
acute exposures. Connection with
carcinogenesis is speculative.
Weight of evidence
Some, but not
conclusive, evidence
for this MOA
Some, but not
conclusive, evidence
for this MOA
Some evidence against
this MOA.
No studies have
explored this MOA for
tetrachloroethylene-
induced
hemangiosarcomas.
No further studies have
explored this MOA.
No studies have
explored this MOA.
-------
Table 4-13. Summary of potential modes of action for cancer (continued)
Organ, tumor type
Brain glioma in male
rats. (Rare)
(continued)
Potential MOA
Mutagenicity of metabolites generated
in situ.
Evidence for MOA
Beta lyase has been
observed in rat brain
tissue. Therefore a
potential exists for
metabolizing
TCVG/TCVC to
reactive compounds.
Limitations/evidence against MOA
Pathway from genotoxic metabolites to
brain cancer is not defined.
Weight of evidence
Little evidence for this
MOA.
1 Perc = Tetrachloroethylene
-------
o
Oi
Table 4-14. Quantitative Implications of different modes of action: candidate modeling approaches
Organ
Kidney
adenocarcimoma in
male rats.
Liver hepatocellular
carcinoma in male and
female mice.
Liver and spleen
hemangiosarcoma in
male mice.
Potential MOA
Mutagenicity of GSH-
derived metabolites.
Necrosis, hyperplasia,
neoplastic
transformation.
Alpha-2ji in hyaline
droplets in male rats.
PPAR-a activation.
PPAR-a receptor
activation, peroxisome
proliferation,
hyperplasia, neoplastic
transformation.
Mutagenicity of P450
metabolites.
Necrosis, hyperplasia,
neoplastic
transformation.
Mutagenicicity (e.g.,
of GSH metabolites).
Weight of evidence
Evidence for MOA.
No direct evidence.
Not likely to occur at
bioassay doses.
Little evidence.
Some, but not
conclusive, evidence
for this MOA.
Some, but not
conclusive, evidence
for this MOA.
Evidence against this
MOA. Probably not
occurring.
No studies have
explored this MOA.
Inference about low-
dose shape
Linear no threshold.
Non-linear model
resulting from
distribution of
thresholds for
individuals in a
population.
MOA does not occur in
humans.
Too little evidence to
inform modeling
approach.
Receptor binding shape
could be linear, sub-
linear or supra-linear,
depending on binding
constants.
Linear no threshold.
Too little evidence to
inform modeling
approach .
Linear no threshold.
Candidate
models
POD, linear
extrapolation.
Log probit or
Log logistic.
Model for
receptor binding
induced by gene
activation.
(Kohnetal.,
1993).
POD, linear
extrapolation.
POD, linear
extrapolation.
Limitations of extrapolation
procedure
Model is commonly used. As
implemented in the current BMDL
software, upper confidence limits on
dose are not calculated.
Variability of thresholds is not the
same in humans as in animals. BMDS
implementation of models is not
believed to be accurate.
Binding data to inform modeling
approach not available.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
oo S3'
I
I
§
***.
£3'
1
TO'
l
2
•••2
I
TO
H I
O >
HH Oq
H TO
O
H
W
-------
Table 4-14. Quantitative Implications of different modes of action: candidate modeling approaches (continued)
Organ
Blood mononuclear cell
leukemia in male and
female rats.
Brain glioma in male
rats (rare tumor).
Potential MOA
Mutagenic GSH
metabolites.
Perc1 incorporation
into cell membranes.
Mutagenicity of
metabolites generated
in situ.
Weight of evidence
Only two
conflicting,
inconclusive reports.
No evidence.
Little evidence but
plausible.
Inference about low-
dose shape
Linear no threshold.
Linear no threshold.
Linear no threshold.
Candidate
models
POD, linear
extrapolation.
POD, linear
extrapolation.
POD, linear
extrapolation.
Limitations of extrapolation
procedure
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
1 Perc = Tetrachloroethylene
-------
1 4.10.4. Rationale for Selection of Dose Metric
2 4.10.4.1. Liver
3 There are several possible choices to consider as the dose metric for tetrachloroethylene-
4 induced liver toxicity and carcinogenicity. First is administered dose or exposure concentration.
5 Tetrachloroethylene hepatotoxicity is associated, however, with cytochrome P450 metabolism
6 occurring in the liver. Several investigators have reported hepatotoxicity in rodent studies to be
7 directly related to metabolism. Because liver toxicity, including carcinogenicity, is generally
8 considered to be caused by metabolites rather than by the parent compound, choosing the
9 specific chemical species responsible for adverse effects in the liver would be the preferred
10 choice over administered dose/exposure concentration, particularly since tetrachloroethylene
11 metabolism is nonlinear with dose of parent compound, with the percent metabolized decreasing
12 with increasing dose.
13 TCA is considered a key product of this P450 oxidation pathway. TCA is the major
14 urinary metabolite from tetrachloroethylene biotransformation, and it is the principal metabolite
15 in the systemic circulation. TCA, like the parent compound, also causes liver toxicity and
16 carcinogenicity in mice. Therefore, the second plausible dose metric for use with the liver target
17 tissue is the concentration and AUC for TCA.
18 The MOA for tetrachloroethylene-induced liver toxicity and carcinogenicity is not clear,
19 however, and whether TCA is the sole contributory metabolite to tetrachloroethylene-induced
20 hepatotoxicity and cancer is unknown. Other possible P450 oxidation products, such as DCA,
21 are also associated with liver toxicity when administered directly. In addition, it is not known
22 whether reactive intermediates such as tetrachloroethylene oxide and trichloroacetyl chloride are
23 involved in tetrachloroethylene-induced liver toxicity.
24 Hepatic toxicity correlates better with metabolism than with administered dose. In other
25 words, a better linear relationship exists between metabolism and hepatotoxicity than between
26 administered dose and hepatotoxicity. Because of the uncertainty about which metabolite
27 species are involved in causing liver toxicity and the degree to which they are involved, the most
28 appropriate dose metric is considered to be total metabolism. Production of the putative
29 metabolites is then considered to be directly proportional to the total amount of
30 tetrachloroethylene metabolized, a reasonable assumption.
31
32 4.10.4.2. Kidney
33 More than one choice was considered for the kidney target organ dose metric. The most
34 simplistic dose metric is administered dose or exposure concentration. However, renal toxicity,
35 including kidney cancer, is associated with metabolism. It is specifically associated with GSH-
36 dependent metabolism, although P450 metabolism could potentially contribute to renal toxicity.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-201 DRAFT-DO NOT CITE OR QUOTE
-------
1 It is generally accepted that the interorgan GSH-dependent pathway, which also occurs
2 completely in the kidney target organ, results in the production and accumulation of mutagenic
3 metabolites. Unfortunately, the measurements of GSH-dependent metabolism are from in vitro
4 studies or are of urinary excretion products, and are not representative of the toxic species in
5 vivo.
6 The total production of the thioketene reactive intermediate divided by the volume of the
7 kidney has been proposed as the dose metric for use in the PBPK model for kidney target organ.
8 In order to use this dose metric, however, several assumptions must be made. One assumption
9 might be that all GSH conjugate formed in the liver is transported to the kidney. Excretion of N-
10 acetyl TCVC is the measurement used to represent flux through the pathway. Clearance of
11 TCVC would be modeled, and production of toxic metabolites would be assumed to be
12 proportional to overall flux. Unfortunately, the flux through the beta lyase and FMO3/CYP3A—
13 or sulfoxide-producing branches of the pathway—has not been measured in vivo.
14 Better methods are needed to quantitate the reactive species that are generated during
15 tetrachloroethylene metabolism, particularly in the beta lyase, FMO3 and CYP3A sections of the
16 pathway, to improve the usefulness of data in development and validation of PBPK models. The
17 amounts of N-acetyl TCVC excreted in urine represent only a portion of the flux through the
18 overall GSH-dependent pathway. This excretion of mercapturate does not represent the
19 processing of TCVC and also N-acetyl TCVC to the reactive and toxic products important to
20 toxicity. The fraction of overall flux represented by the excretory product is simply unknown,
21 and how it is related to the fraction processed through the beta lyase branch of the path is also
22 unknown. Several products formed in the pathway are unstable and reactive, and, therefore, they
23 are difficult to quantitate. Because the quantitative information about toxic metabolites from the
24 GSH-dependent pathway is not available, and there is no way of knowing whether the
25 measurement of excretory mercapturate is proportional to production of the toxic species
26 produced in this pathway, the administered exposure and total metabolites are both considered as
27 dose metrics (see Section 5.3.3.2). Production of the putative metabolites, then, is considered to
28 be directly proportional to the total amount of tetrachloroethylene metabolized.
29
30 4.10.4.3. Hematopoietic Target Organ
31 Tetrachloroethylene causes mononuclear cell leukemia in rats. Although the specific
32 mechanism of leukemogenesis in rats is not understood, neither is it well understood in humans.
33 Whether the parent compound, a metabolite, or several metabolites are involved in the
34 tetrachloroethylene induction of the leukemia is not known. In the case of the
35 tetrachloroethylene congener trichloroethylene, the comparable DCVC conjugate metabolite has
36 been associated with causing adverse effects to the hematopoietic system. A possible link to
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-202 DRAFT-DO NOT CITE OR QUOTE
-------
1 MO A for tetrachloroethylene-induced MCL in rats comes from early reports of toxicity of
2 cysteine S-conjugates where DCVC, the trichloroethylene metabolite, was implicated in
3 induction of aplastic anemia and marked biochemical alteration of DNA in bone marrow, lymph
4 nodes, and thymus in calves (McKinney et al., 1957; Schultze et al., 1959; Bhattacharya and
5 Schultze, 1971, 1972). To the contrary, however, the single study of TCVC, the
6 tetrachloroethylene conjugate, in calves, did not result in the adverse effects observed in studies
7 of exposures to DCVC. Therefore, because considerable uncertainty surrounds the identification
8 of the causative chemical species, the administered exposure and total metabolites are both
9 considered as dose metrics (see Section 5.3.3.2). Production of the putative metabolites is
10 considered to be directly proportional to the total amount of tetrachloroethylene metabolized.
11
12 4.10.4.4. Central Nervous System
13 As discussed in Section 4.6.4, the best surrogate for internal dose is blood
14 tetrachloroethylene concentration.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-203 DRAFT-DO NOT CITE OR QUOTE
-------
1 APPENDIX 4A:
2 CONSISTENCY OF TETRACHLOROETHYLENE AND TRICHLOROACETIC
3 ACID HEPATOCARCINOGENICITY
4
5
6 TCA, a metabolite of tetrachloroethylene, is associated with hepatocarcinogenicity in
7 male and female mice (Bull et al., 1990, 2002; Daniel et al., 1993; DeAngelo et al., 2008;
8 Herren-Freund et al., 1995; Ferreira-Gonzalez, 1987; Pereira, 1996), as is tetrachloroethylene
9 (NCI, 1977; NTP, 1986; JISA, 1992).
10 There has been some suggestion that TCA does not account for all of the toxicity
11 observed with tetrachloroethylene exposure (Buben and O'Flaherty, 1985; Clewell et al., 2005).
12 The purpose of this investigation was to compare the incidence of hepatocarcinogenicity
13 observed with tetrachloroethylene exposure to that observed with TCA exposure, in order to
14 examine whether the TCA that is expected to be generated by tetrachloroethylene can account
15 for tetrachloroethylene's hepatocarcinogenicity. This was carried out by pooling the separate
16 TCA studies, fitting a time-to-tumor model to the TCA data, and comparing the incidence of
17 hepatocellular tumors expected based on the TCA studies with that observed in the
18 tetrachloroethylene bioassays.
19
20 4A.1. METHODS
21 Table 4A-1 summarizes data from the available TCA studies considered for carrying out
22 dose-response modeling. As detailed below, a number of these TCA studies lack information for
23 a complete comparison of hepatocarcinogenicity between tetrachloroethylene and TCA.
24
25 4A.1.1. Response Data
26 EPA generally emphasizes combining hepatocellular adenomas and carcinomas in
27 developing cancer risk values, for three reasons: (1) Hepatocellular adenomas develop from the
28 same cell lines as carcinomas and can progress to carcinomas; (2) Adenomas are often
29 distinguished from carcinomas only on the basis of size; and (3) histopathologic decision criteria
30 may vary between laboratories or over time.
31 However, most of the TCA studies either did not consider adenomas or did not report
32 combined incidence of adenomas and carcinomas. Lacking data on adenomas, the studies that
33 only provided carcinoma incidence may under-represent hepatocellular tumor incidence. For
34 studies not reporting combined incidence of adenomas and carcinomas, there could be some
35 double-counting of animals when the separate totals of adenomas and carcinomas are added
36 together.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-204 DRAFT-DO NOT CITE OR QUOTE
-------
1 Therefore, for the purposes of this analysis, only the chronic data of DeAngelo et al.
2 (2008) and the chronic data of Pereira (1996) were considered further for comparing with the
3 tetrachloroethylene bioassays in male and female mice, respectively. The DeAngelo et al. study
4 was conducted at the lowest TCA levels of all the available studies, with exposures spanning
5 about 6 to 60 mg/kg-day; these exposure levels span the range of TCA equivalents in the
6 tetrachloroethylene bioassays. Comparison of the Pereira data with the responses of the female
7 mice in the tetrachloroethylene bioassays is limited by the availability of carcinoma data only,
8 and by the study being conducted only through Week 82, not through Week 104.
9 Table 4A-3 provides the hepatocellular adenoma or carcinoma incidence data from the
10 two tetrachloroethylene bioassays considered in this assessment, NTP (1986) and JISA (1993)
11 (for convenience, the studies will be referred to in the remainder of this appendix as the NTP and
12 JISA studies). For comparison across data sets, all incidences were normalized by converting
13 each to extra risk, [P(d)-P(0)]/[l-P(0)].
14
15 4A.1.2. Exposure Level Conversions
16 TCA bioassay exposures were generally reported in terms of water concentration, in
17 mg/L or mmol/L. Table 4A-1 provides the exposure levels as reported by each set of authors.
18 Some reports provided mg/kg-day equivalents. TCA exposures in mg/kg-day for the Pereira
19 (1996) study were interpolated from the other TCA studies which reported exposures in mg/kg-
20 day (see Table 4A-2).
21 The Reitz et al. (1996) PBPK model was used to estimate total metabolites corresponding
22 to the bioassay exposures in the NTP and JISA studies. Then it was assumed that 60% of the
23 total metabolites were TCA, as assumed in the model of Gearhart et al. (1993). Although it is
24 possible that the extent of metabolism to TCA may be dose dependent, as for trichloroethylene-
25 induced TCA, there were insufficient data to characterize a dose dependency of TCA formation
26 for tetrachloroethylene.
27 The estimates of TCA induced by tetrachloroethylene exposure are internal doses, while
28 the exposures in the TCA bioassays were administered doses. Because orally administered TCA
29 has been estimated to be 95% absorbed in mice, the tetrachloroethylene-induced TCA estimates
30 were adjusted by dividing by 0.95, in order to approximate administered TCA exposures that
31 would be compatible with the dose-response modeling of the TCA drinking water studies.
32
33 4A.1.3. Dose-Response Model
34 The TCA data sets for male and female mice were fit separately. The male mice TCA
35 were modeled using the multistage model (BMDS 1.4.1; U.S.EPA, 2007), given by:
36
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-205 DRAFT-DO NOT CITE OR QUOTE
-------
1 P(d) = 1 - exp[(-q0 - qi x d - q2 x d2 x... q6 x d6 ),
2 where d = exposure level.
3 The TCA data set for female mice was fit using a multistage-Weibull model because the
4 only available data were limited to two time points less than the 104-week length of the
5 tetrachloroethylene bioassays; this model provided a means of including both time points in the
6 same analysis and facilitated extrapolation to 104 weeks. The multistage-Weibull model is
7 given by:
8
9 P(d,t) = 1- exp[(-q0 - qi x d - q2 x d2 x... q6 x d6 ) x tz,
10
11 where:
12 d = exposure level
13 t = time to observation of the tumor
14 q;, z = parameters estimated in fitting the model
15
16 Time of scheduled sacrifice was input as the time to observation of each tumor. All tumors were
17 taken to be incidental to the death of affected animals. The software used was Tox_Risk (see
18 Section 5.4.4.1).
19 For comparison with the observed tetrachloroethylene data, model predictions were also
20 adjusted to estimate extra risk.
21
22 4A.2. RESULTS
23 4A.2.1. Trichloroacetic Acid (TCA), Male Mice
24 Figure 4A-2 provides the result of fitting a multistage model to the DeAngelo et al. data.
25 The responses at the control and low dose levels did not follow a monotonically increasing
26 pattern (the low-dose response was lower than the control), but a nearly linear one-stage model
27 provided an adequate fit (p = 0.15; model output included with Figure 4A-2).
28
29 4A.2.2. Trichloroactic Acid (TCA) Data, Female Mice
30 The evaluation of the model fit of the female mouse TCA data (Pereira, 1996) followed
31 the same steps as for the male mice. The hepatocellular tumor data for female mice exposed to
32 TCA in drinking water are shown in Table 4A-2 and Figure 4A-2. These data include groups of
33 animals evaluated at three exposure levels plus control at two time points, for a total of eight
34 groups. The response at the high dose (463 mg/kg-day) was very similar for both time points, at
35 25 - 28%. A one-stage model also provided the best fit to these data, with the two highest doses
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-206 DRAFT-DO NOT CITE OR QUOTE
-------
1 at week 82 fitting least well. Because the fit at the lower doses was relatively good, no other
2 attempts were made to refine the dose-response model for the TCA female mouse data.
O
4 4A.2.3. Comparison of Tetrachloroethylene Hepatocellular Tumor Data With Predictions
5 Based on Trichloroacetic Acid Data
6 For the male mice, the extra risk of adenomas or carcinomas observed following 104
7 weeks of inhalation exposure to tetrachloroethylene in the two bioassays is provided in Table
8 4A-4, for comparison with the predicted extra risk of adenomas or carcinomas from the TCA-
9 based dose-response modeling for male mice (based on the data of DeAngelo et al., 2008). For
10 each male exposure group in the tetrachloroethylene bioassays, the observed proportion
11 responding is higher than that predicted using the TCA drinking water study, by 2- to 12-fold.
12 Mitigating factors to investigate further include possible differences in histopathology protocols
13 between laboratories and adequacy of the assumptions used to derive the TCA-equivalents
14 corresponding to the tetrachloroethylene exposure levels. Comparison between the
15 tetrachloroethylene and TCA studies for the male mice at Week 104 suggests concordance, but
16 "inconclusive" appears to be a plausible conclusion as well.
17 For the female mice, the extra risk of carcinomas observed following 104 weeks of
18 inhalation exposure to tetrachloroethylene in the two bioassays is provided in Table 4A-5, for
19 comparison with the predicted extra risk of carcinomas from the TCA-based dose-response
20 modeling for female mice (based on the data of Pereira, 1996). The female mouse TCA model
21 appears to agree with the lack of carcinomas in female mice at the lower two exposures in the
22 JISA study, at approximately 3 and 11 mg/kg-day, but underestimates the observed incidence at
23 30 mg/kg-day of 29% by 90-fold. In contrast, the TCA model underpredicts the observed
24 carcinomas in both exposed groups of female mice in the NTP study by more than 200-fold.
25 In addition to the mitigating factors mentioned above, note that the tetrachloroethylene bioassays
26 were conducted at exposures associated with lower TCA levels than were used in the female
27 mouse TCA study. That is, for JISA female mice, the highest bioassay exposures were
28 associated with 28 mg/kg-day TCA (NTP) and 30 mg/kg-day, and the lowest exposure level in
29 the TCA study was approximately 47 mg/kg-day. Consequently, there is a degree of
30 extrapolation beyond the TCA data set that may impact the predictions. Also note that the
31 tetrachloroethylene bioassays do not have sufficient resolution (let alone statistical power) to
32 detect response levels as low as those predicted by the TCA model in this range of exposures
33 (bioassays with 50 animals/group cannot provide estimates below 1/50 or 2%).
34
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-207 DRAFT-DO NOT CITE OR QUOTE
-------
1 4A.3. DISCUSSION AND CONCLUSIONS
2 This analysis suggests that TCA may not explain the incidence of carcinomas observed in
3 the available tetrachloroethylene bioassays, at least at TCA levels near 2 mg/kg-day in male mice
4 and 20 mg/kg-day in female mice. Otherwise, these data are inconclusive.
5 As mentioned earlier in this appendix, some of the assumptions made in quantifying the
6 dose-response relationships may have contributed to an overestimate of TCA's carcinogenicity.
7 The Pereira study involved planned sacrifices at the reported time points, while the
8 tetrachloroethylene studies did not. This would have led to earlier detection of tumors in the
9 TCA studies relative to the tetrachloroethylene bioassays due to the detection of some tumors
10 before they may have become fatal, and, therefore, a slightly higher estimate of carcinoma
11 incidence in the time-to-tumor model. Therefore, dose-response estimates based on the female
12 TCA study may contribute to overestimating risk, all else being equal.
13 And as mentioned earlier, another uncertainty is the use of PBPK estimates of TCA
14 levels resulting from inhalation or oral exposure to tetrachloroethylene. Another interpretation
15 of tetrachloroethylene-induced TCA levels has been provided by Clewell et al. (2005), who
16 provided TCA levels corresponding to the bioassay levels in the NTP bioassay, but not the JISA
17 bioassay. These levels were 16.3 mg/kg-day for the low-dose males, 30.6 mg/kg-day for the
18 high-dose males, 16.9 mg/kg-day for the low-dose females, and 31.6 mg/kg-day for the high-
19 dose females. These levels differ from those estimated here by no more than 15%, which does
20 not explain the differences in response levels compared in this analysis. Given the current state
21 of the science, the impact of this source of uncertainty is not well understood.
22 The differing results from the other TCA studies underscore the need to consider the joint
23 incidence of adenoma and carcinomas, which could have a substantial impact on this analysis.
24 The relative time courses of and correlation between adenomas and carcinomas in the TCA
25 bioassays are less clear, because relevant data were not included in the TCA reports. This is
26 perhaps the most uncertain part of this analysis. Additional information should be obtained from
27 the original investigators for further evaluation if possible, perhaps in a meta-analysis.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-208 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4A-1. Trichloroacetic Acid (TCA) drinking water studies in male mice: incidence of hepatocellular
adenomas and carcinomas
Source
Bull et al.
(1990)a
Bull et al.
(2002)
Herren-Freund
etal. (1987)
Ferreira-
Gonzalez et al.
(1995)
DeAngelo et al.
(2008)
Weeks of
exposure
37
52
52
61
104
104
TCA
exposure,
g/L
2
0
1
2
0
0.5
2
0
5
0
4.5
0
0.05
0.5
Equivalent TCA
exposure
(mg/kg-day)
330
0
170
330
0
NR
NR
0
NR
0
NR
0
8
68
N
11
35
11
24
20
20
20
22
22
16C
11
56
48
51
Incidence of
adenomas
0
0
2
1
0
5
6
2
8
NR
NR
10
10
20
Incidence of
carcinomas
3
0
2
4
0
3
3
0
7
3C
8
26
14
32
Incidence of
adenomas or
carcinoma
3
0
NR
NR
0
6
8
2
NR
NR
NR
31
21
36
Proportion
responding with
carcinomas
0.27
0.0
0.18
0.17
0.0
0.15
0.15
0.0
0.32
0.19
0.73
0.55
0.44
0.71
a Cumulative TCA exposures were provided in g/kg for the mice evaluated at 52 weeks. Those exposures were converted to mg/kg-day by multiplying by
(1,000 mg/g)/(7 days/week * 52 weeks).
0 Estimated from the reported proportion responding by selecting the smallest group size and incidence value consistent with the precision of the reported
proportion.
NR = not reported
-------
Table 4A-2. Trichloroactic acid (TCA) drinking water study in female
mice—incidence of hepatocellular adenomas and carcinomas
Weeks of
exposure
52
82
TCA
exposure,
g/La
0.0
0.3
1.1
3.3
0.0
0.3
1.1
3.3
Equivalent
TCA
exposure1"
(mg/kg-day)
0
47
189
463
0
47
189
463
N
40
40
19
20
90
53
27
18
Incidence of
adenomas
1
3
3
2
2
4
3
7
Incidence of
carcinomas
0
0
0
5
2
0
5
5
Incidence of
adenomas or
carcinomas
1
3
3
NR
NR
4
NR
NR
Proportion
responding
with
carcinomas
0.0
0.0
0.0
0.25
0.022
0.0
0.19
0.28
a Exposure concentration was reported in mmol/L.
b Estimated by interpolating exposures in Table 4A-1.
NR = not reported
Source: Adapted from Pereira (1996).
1
2
3
4
Table 4A-3. Incidence of hepatocellular adenomas and carcinomas in
B6C3F1 mice exposed to tetrachloroethylene in two inhalation bioassays
Sex
Male
Female
Bioassay
NTP
(1986)
JISA
(1993)
NTP
(1986)
JISA
(1993)
Administered
exposures,
(ppm)
0
100
200
0
10
50
250
0
100
200
0
10
50
250
Cumulative liver tumor incidence at week 104
Adenomas
12
8
19
7
13
8
26
3
6
2
3
3
7
26
Carcinomas
7
25
26
7
8
12
25
1
13
36
0
0
0
14
Adenomas
or
carcinomas
17
31
41
13
21
19
40
4
17
38
3
3
7
33
Total at risk8
49
47
50
46
49
48
49
45
42
48
50
47
48
49
5
6
7
a Animals dying before the first appearance of a hepatocellular tumor, but no later than week 52, were omitted from
the totals because these animals were presumed not to have adequate time on study to develop tumors.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-210 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
Table 4A-4. Comparison of cumulative hepatocellular tumor incidence in
male mice exposed for 104 weeks to tetrachloroethylene in chronic inhalation
bioassays, to predictions based on trichloroacetic acid (TCA) exposure via
drinking water
Study
NTP
(1986)
JISA
(1993)
N
49
47
50
46
49
48
49
Tetrachloro-
ethylene
exposure
(ppm)
0
100
200
0
10
50
250
Total
metabolites
(mg/kg-day)
0
27
41
0
3
14
36
Tetrachloro-
ethylene-induced
TCAa
(mg/kg-day)
0
17
26
0
2
9
23
Observed
proportion
responding1"'0
0.0
0.479
0.724
0.0
0.203
0.158
0.744
Predicted
proportion
responding1"
0.0
0.139
0.204
0.0
0.017
0.076
0.182
Observed/
predicted
-
o
J
4
.
12
2
4
6
7
8
9
10
11
12
13
14
15
16
17
Estimated using PBPK model of Reitz et al. (1996) and adjusted for use with the TCA dose-response model by
dividing by 0.95 to approximate a drinking water exposure to TCA (see Section 4A. 1.2).
Extra risk.
Calculated from Table 4A-1.
Table 4A-5. Comparison of cumulative hepatocellular carcinoma incidence
in female mice exposed for 104 weeks to tetrachloroethylene in chronic
inhalation bioassays, to predictions based on trichloroacetic acid (TCA)
exposure via drinking water
Study
NTP
(1986)
JISA
(1993)
N
45
42
48
50
47
48
49
Tetrachloro-
ethylene
exposure
(ppm)
0
100
200
0
10
50
250
Total
metabolites
(mg/kg-day)
0
31
45
0
4
18
47
Tetrachloro-
ethylene-
induced
TCAa
(mg/kg-day)
0
20
28
0
3
11
30
Observed
proportion
responding1"'0
0.0
0.287
0.728
0.0
0.0
0.0
0.286
Predicted
proportion
responding1"
0.0
0.0014
0.0029
0.0
0.00002
0.00041
0.0031
Observed/
predicted
.
210
260
.
-
-
90
18
19
20
21
22
23
a Estimated using PBPK model of Reitz et al. (1996) and adjusted for use with the TCA dose-response model by
dividing by 0.95 to approximate a drinking water exposure to TCA (see Section 4A. 1.2).
b Extra risk.
c Calculated from Table 4A-2.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-211 DRAFT-DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
1
2
5
6
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
0.8
0.7
0.6
0.5
0.4
0.3
Multistage
BMDL
MD
0 10
16:4806/092008
20
30
dose
40
50
60
Figure 4A-1. Multistage dose-response fit of male mouse hepatocellular
tumor incidence associated with exposure to trichloroacetic acid in drinking
water; data from DeAngelo et al. (2008).
The form of the probability function is:
ound + (1-background)*[1-EXP(
-betal*dose^l)]
Dependent variable = hep a c
Independent variable = mg kg d
Total number of observations = 3
Total number of records with missing value
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 DRAFT-DO NOT CITE OR QUOTE
4-212
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
"_/
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
Default Initial Parameter Values
Background = 0.489396
Asymptotic Co
Background
Background
Beta (1)
Interval
Variable
Conf. Limit
Background
Beta (1)
1
-0.54
0.
* - Indicates that this
Model
Full model
Fitted model
Reduced model
AIC:
Dose Est
0 . 0000 0 .
5 . 6000 0 .
58.0000 0.
Chi^2 = 2.07
Benchmark Dose
Specified effect
Risk Type
Confidence level
BMD
BMDL
BMDU
Taken together, (
interval for the
Beta(l) = 0.00926311
rrelation Matrix of Parameter Estimates
Beta(l)
-0.54
1
Parameter Estimates
95.0% Wald Confidence
Estimate Std. Err. Lower Conf. Limit Upper
0.493963 * *
00878738 * *
value is not calculated.
Analysis of Deviance Table
Log (likelihood) # Param's Deviance Test d.f. P-value
_
_
_
. Prob.
4940
5183
6960
d.f.
102.285 3
103.323 2 2.07524 1 0.1497
106.011 1 7.4518 2 0.02409
210. 645
Goodness of Fit
Scaled
Expected Observed Size Residual
27.662 31 56 0.892
24.877 21 48 -1.120
35.497 36 51 0.153
= 1 P-value = 0.1499
Computation
=
=
_
=
=
=
6.53618
BMD
0.1
Extra risk
0. 95
11. 99
6.53618
43.6117
, 43.6117) is a 90 % two-sided confidence
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 DRAFT-DO NOT CITE OR QUOTE
4-213
-------
15:1906/03/2005
1
0.8
0.6
0.4
0.2
0
1
2
3
4
5
6
Incidental Graph
tca_fem_carc.ttd - TCA, females, hep. care.
Model: Multistage Weib
Dose (mg/kg/day)=47
Dose (mg/kg/day)=189
Dose (mg/kg/day)=463
HoelWalburg(189)
Hoel Walburg (463)
H
H
20
40
60
Time (wks)
80
100
120
Figure 4A-2. Multistage-Weibull dose-response fit of female mouse
hepatocellular carcinoma incidence associated with exposure to
trichloroacetic acid in drinking water; data in Table 4A-2. Multistage-
-5
Weibull model parameters: q0 = 3.39 x 10 , q2 = 6.10 x 10°, z = 1.9.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 DRAFT-DO NOT CITE OR QUOTE
4-214
-------
1 APPENDIX 4B: HUMAN STUDIES OF CANCER
2
3 The body of literature assessing carcinogenic effects associated with exposure to
4 tetrachloroethylene is composed of cohort, proportionate mortality, and case-control studies. A
5 small number of studies, including studies of cohorts involved in metal degreasing or in aircraft
6 manufacturing/maintenance (Boice et al., 1999; Anttila et al., 1995; Spirtas et al., 1991), have
7 assessed tetrachloroethylene exposure explicitly. These cohort studies present risks associated
8 with site-specific cancer mortality (Boice et al., 1999; Spirtas et al., 1991) or incidence (Anttila
9 et al., 1995) for a subcohort of the larger study population who were exposed to
10 tetrachloroethylene. Additionally, a few case-control studies were able to examine the
11 relationship between cancer at specific sites and tetrachloroethylene exposure (Vaughan et al.,
12 1997; Schlehofer et al., 1995; Pesch et al., 2000a; Heineman et al., 1994).
13 A larger body of evidence on cancer exists for workers employed as dry cleaners. Dry
14 cleaners have potential exposures to a number of solvents, including tetrachloroethylene, which
15 began to be widely used in the early 1960s (IARC, 1995). Two cohorts, Ruder et al. (1994,
16 2001) and Blair et al. (1990, 2003), are of individuals primarily exposed to tetrachloroethylene
17 (Lyngeetal., 1997).
18 Last, several community-based drinking water studies are available (Aschengrau et al.,
19 1993, 1998; Paulu et al., 1999; Fagliano et al., 1990; Cohn et al., 1994; MA DPH, 1997;
20 Vartiainen et al., 1993; Lagakos et al., 1986). Exposure in most of these studies was to a mixture
21 of solvents, including tetrachloroethylene and trichloroethylene, although the studies by
22 Aschengrau et al. and Paulu et al. examined tetrachloroethylene specifically.
23 Observations from these epidemiologic studies are summarized in Tables 4B-la through
24 4B-13. Table 4B-la presents results of incidence studies of dry cleaners and laundry workers
25 and Table 4B-lb) those of mortality studies of dry cleaners. Three studies of aircraft
26 maintenance/manufacturing workers or metal degreasing workers identified workers (a
27 subcohort) exposed to tetrachloroethylene. Tables 4B-3 through 4B-13 present results of case-
28 control studies of specific organ sites (e.g., lymphoma, liver, kidney, esophagus). A few studies
29 assessed exposure to tetrachloroethylene, with the majority of studies evaluating the relationship
30 between site-specific cancer and employment in dry cleaning.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 DRAFT-DO NOT CITE OR QUOTE
4-215
-------
Table 4B-la. Standardized incidence ratios (95% confidence intervals) in cohort studies of dry cleaners and
laundry workers
Site
Bladder
Breast
Cervix
Esophagus
Hodgkin's
disease
Kidney
Leukemia
Ji and Hemminki
(2005 a, b; 2006);
Ji et al. (2005)
Male
0.9b
(0.7-1.1)
61 obs/
0.8C
(0.6-1.1)
47 obs/
Ji and Hemminki
(2005 a, b; 2006); Ji
et al. (2005)
Female
1.4b
(1.1-1.7)
92 obs/
1.3C
(1.0-1.6)
80 obs/
Andersen et al.
(1999) a
Male
1.1
(0.9-1.5)
62 obs/
54.4 exp
0.8
(0.3-1.7)
7 obs/8.5 exp
0 obs/4.0 exp
1.0
(0.7-1.5)
24 obs/
23.3 exp
0.7
(0.03-1.2)
12 obs/
17.8 exp
Andersen et al.
(1999) a
Female
0.9
(0.7-1.2)
57 obs/
64.0 exp
0.9
(0.8-0.97)
634 obs/
7 12 exp
1.2
(1.01-1.4)
155 exp/
131.4 exp
1.0
(0.5-1.6)
14 obs/ 14. 4 exp
1.9
(1.1-2.9)
19obs/10.1exp
0.9
(0.7-1.2)
57 obs/64.8 exp
0.9
(0.7-1.2)
46 exp/50.9 exp
Lynge and
Thygesen
(1990)
Male
0.6
(0.2-1.3)
6 obs/9.7 exp
1.5
(0.6-3.3)
6 obs/4 exp
0.7
(0.1-2.6)
2 obs/2.8 exp
Lynge and
Thygesen (1990)
Female
0.9
(0.4-1.7)
8obs/9.1 exp
0.8
(0.7-1.1)
94obs/110.7exp
0.8
(0.6-1.2)
34 obs/40.3 exp
0.6
(0.2-1.4)
5 obs/8.6 exp
0.8
(0.2-1.7)
5 obs/6.7 exp
-------
o
Oi
oo 3s
I
I
§
***.
£3'
1
TO'
Table 4B-la. Standardized incidence ratios (95% confidence intervals) in cohort studies of dry cleaners and
laundry workers (continued)
Site
Primary liver
Lung
Multiple
myeloma
Non-
Hodgkin's
Pancreas
Prostate
Rectum
Ji and Hemminki
(2005 a, b; 2006);
Ji et al. (2005)
Male
1.4
(1.2-1.5)
247 obs/
1.0
(0.7-1.4)
29 obs/
1.0
(0.8-1.3)
59 obs/
Ji and Hemminki
(2005 a, b; 2006); Ji
et al. (2005)
Female
1.3
(1.1-1.5)
156 obs/
0.9
(0.6-1.2)
3 lobs/
0.9
(0.7-1.1)
64 obs/
Andersen et al.
(1999) a
Male
1.3
(0.6-2.3)
1 1 obs/
8.7 exp
1.2
(1.1-1.5)
141 obs/
113.7 exp
1.4
(0.8-2.3)
14 obs/
10.1 exp
1.5
(0.96-2.1)
27 obs/
18.5 exp
1.4
(0.98-2.0)
35 exp/
24.8 exp
1.1
(0.9-1.3)
118 obs/
11.3 exp
0.9
(0.6-1.3)
33 obs/
37.1 exp
Andersen et al.
(1999) a
Female
1.3
(0.9-1.9)
28obs/21.2exp
1.2
(1.08-1.4)
172 obs/ 148. 3 exp
0.9
(0.6-1.3)
3 lobs/ 34.8 exp
1.0
(0.7-1.2)
55 obs/57.9 exp
1.0
(0.8-1.3)
85 obs/83.3 exp
1.0
(0.8-1.2)
106 obs/
110.4 exp
Lynge and
Thygesen
(1990)
Male
1.2b
(0.3-3.5)
3 obs/2.5 exp
1.1
(0.8-1.7)
28 obs/24.5 exp
3.3
(0.9-8.5)
4 obs/ 1.2 exp
2.8
(0.9-6.5)
5 obs/ 1.8 exp
2.4
(1.1-4.5)
9obs/3.8exp
1.5
(0.7-2.6)
11 obs/7.6 exp
1.4
(0.6-2.6)
9 obs/6.5 exp
Lynge and
Thygesen (1990)
Female
2.7b
(1.5-4.5)
14 obs/5.2 exp
1.3
(0.9-1.8)
32 exp/24.9 obs
1.1
(0.2-3.1)
3 obs/2.8 exp
0.5
(0.1-1.5)
3 obs/6 exp
1.4
(0.7-2.4)
13 obs/9.3 exp
0.7
(0.4-1.3)
11 obs/15.5 exp
to
§
I
TO
H I
O >
HH Oq
H TO
O
H
W
-------
o
Oi
oo S3'
I
I
§
^.
Co'
1
TO'
to
00 2
I
TO
Table 4B-la. Standardized incidence ratios (95% confidence intervals) in cohort studies of dry cleaners and
laundry workers (continued)
Site
Skin
Stomach
Ji and Hemminki
(2005 a, b; 2006);
Ji et al. (2005)
Male
Ji and Hemminki
(2005 a, b; 2006); Ji
et al. (2005)
Female
Andersen et al.
(1999) a
Male
1.0
(0.8-1.3)
60 obs/
60.7 exp
1.3
(0.9-1.7)
47 obs/
37.6 exp
Andersen et al.
(1999) a
Female
0.9
(0.8-1.04)
174 obs/
196.1 exp
1.0
(0.8-1.2)
89obs/93.7exp
Lynge and
Thygesen
(1990)
Male
1.1
(0.6-1.9)
14 obs/12.7 exp
1.3
(0.5-2.7)
7 obs/5.3 exp
Lynge and
Thygesen (1990)
Female
0.7
(0.5-1.1)
23obs/32.1exp
1.3
(0.6-2.3)
11 obs/8.6 exp
a Anderson et al. (1999), a study of cancer incidence and occupation in Nordic populations (Norway, Sweeden, finland, and Denmark), includes
some subjects in these previously-published and earlier studies: McLaughlin et al. (1987), Lynge and Thygesen (1990) and Chow et al. (1995).
Cano and Pollan (2001) and Travier et al. (2002) present observations from Sweedish subjects included in Anderson et al. (1999).
b Ji et al. (2005) updates the analysis of McLaughlin et al.(1987) who also presented kidney cancer incidence rate for dry cleaners and laundry
workers.
0 Ji and Hemminki (2005b) present standardized incidence ratios for these leukemia subtypes. For males, chronic lymphatic leukemia, 0.9 (0.5-
1.3), 19 obs.; acute myelogeneous leukemia, 0.6 (0.2-1.2), 7 obs.; chronic myelogenous leukemia, 0.9 (0.3-1.9), 5 obs.; polycythemia vera, 1.0
(0.4-2.0), 7 obs. For females, chronic lymphocytic leukemia 1.5 (1.1-2.1), 32 obs.; acute myelogenous leukemia, 1.4 (0.8-2.0), 20 obs.; chronic
myelogenous leukemia, 0.3 (0.03-0.9), 2 obs.: polycythemia vera, 1.7 (0.9-2.7), 14 obs.
H I
O ^
HH Oq
H TO
O
obs = Observed
exp = Expected
H
W
-------
1
2
Table 4B-lb. Standardized mortality ratios in cohort studies of dry cleaners
Site
Total mortality
Bladder
Breast
Buccal
Cervix
Esophagus
Hodgkin's disease
Kidney
Larynx
Leukemia
Liver/biliary passages
Lung
Lymphatic-hematopoietic
Multiple myeloma
Non-Hodgkin's
Ruder et al. (2001)a
1.0
(0.9-1.1)
302 obs/299 exp
Oobs
0.8
(0.3-1.7)
6 obs/7.7 exp
1.9
(0.5-4.8)
4 obs/2. 1 exp
2.7
(0.9-6.2)
5 obs/1.9 exp
1.7
(0.2-6.3)
2 obs/1.2 exp
0.2b
(0.0-1.3)
1 obs/5 exp
1.2
(0.7-1.8)
19 obs/17.3 exp
1.1
(0.4-2.4)
6 obs/5. 6 exp
1.4b
(0.6-2.9)
7 obs/5 exp
Blair et al. (2003)
1.0
(1.0-1.1)
23 5 lobs/23 51 exp
1.3
(0.7-2.4)
12 obs/9.2 exp
1.0
(0.8-1.3)
68 obs/68 exp
1.1
(0.5-2.0)
10obs/9.1exp
1.6
(1-2.3)
27 obs/16.9 exp
2.2
(1.5-3.3)
26obs/11.8exp
2.0
(0.6-4.6)
5 obs/2. 5 exp
1.0
(0.4-2.0)
8 obs/8 exp
1.7
(0.6-3.7)
6 obs/3.5 exp
0.8
(0.4-1.4)
12 obs/15 exp
0.8
(0.4-1.5)
10 obs/12.5 exp
1.4
(1.1-1.6)
125 obs/89.3 exp
1.0
(0.7-.3)
39 obs/39 exp
0.8
(0.3-1.6)
7 obs/16 exp
0.9
(0.5-1.6)
12 obs/13.2 exp
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-219 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-lb. Standardized mortality ratios in cohort studies of dry cleaners
(continued)
Site
Pancreas
Prostate
Rectum
Skin
Stomach
Ruder et al. (2001)a
0.8
(0.2-2.4)
3 obs/3.8 exp
0.7C
(0.1-2.4)
2 obs/3.1 exp
Oobs
Blair et al. (2003)
1.1
(0.7-1.5)
28 obs/25.5 exp
1.0
(0.6-1.6)
17 obs/17 exp
1.3
(0.2-2.1)
15 obs/11.5 exp
0.8
(0.2-2.1)
4 obs/5 exp
0.9
(0.6-1.4)
20 obs/22.2 exp
4
5
6
7
a Standard mortality ratios (SMRs) for subjects whose employment begun after 1960, e.g., the
tetrachloroethylene subcohort.
b SMR for the PCE subcohort for liver cancer and for non-Hodgkin's lymphoma are not presented in
Ruder et al. (2001). The SMRs in this table are for the entire cohort.
0 Ruder reports SMR for the larger category of "male genital organ cancer."
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-220 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-2. Standardized incidence and mortality ratios in cohort studies of
tetrachloroethylene-exposed degreasers
Site
Total mortality
Bladder
Breast
Buccal
Cervix
Esophagus
Hodgkin's disease
Kidney
Larynx
Leukemia
Liver/biliary
passages
Lung
Lymphatic-
hematopoietic
Multiple myeloma
Non-Hodgkin's
Boice et al.
(1999)a
(mortality)
0.9
(0.9-1.0)
476 obs/528.9 exp
0.7
(0.1-2.5)
2 obs/2.9 exp
1.2
(0.3-3.0)
4 obs/3.5 exp
0.6
(0.1-2.0)
2 obs/3.6 exp
(0-7.8)
0 obs/0.5 exp
1.5
(0.5-3.2)
6 obs/4.1 exp
0 obs/0.6 exp
0.7
(0.1-2.5)
2 obs/2.9 exp
0.7
(0-3.6)
1 obs/1.6 exp
1.1
(0.4-2.6)
5 obs/4.6 exp
2.1
(0.8-4.2)
7 obs/3.4 exp
1.1
(0.8-1.4)
46 obs/42.6 exp
1.1
(0.6-1.9)
14 obs/12.4 exp
0.4
(0-2.3)
1 obs/2.5 exp
1.7
(0.7-3.3)
8 obs/4.7 exp
Anttila et al.
(1995)
(incidence)
3.2
(0.4-11.6)
2 obs/1.6 exp
1.8
(0.2-6.6)
2 obs/1.1 exp
1.9
(0.6-4.5)
5 obs/2.6 exp
1.4
(0.3-4.0)
3 obs/2.2 exp
(0-9.8)
0 obs/0.4 exp
3.8
(0.8-11.0)
3 obs/0.8 exp
Spirtas et al. (1991)
Males
(mortality)
Oobs
1.9
(0.2-6.9)
2 obs/1.1 exp
Females
(mortality)
17.1
(2.1-61.6)
2 obs/0. 1 exp
9.7
(1.2-35.0)
2 obs/0. 2 exp
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-221 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-2. Standardized incidence and mortality ratios in cohort studies of
tetrachloroethylene-exposed degreasers (continued)
Site
Pancreas
Prostate
Rectum
Skin
Stomach
Boice et al.
(1999)a
(mortality)
1.5
(0.7-2.8)
10 obs/6.7 exp
1.1
(0.6-2.0)
12 obs/10.5 exp
1.2
(0.3-3.6)
3 obs/2.4 exp
lb
(0.1-3.4)
2 obs/2.1 exp
1.4
(0.6-2.9)
7 obs/4.9 exp
Anttila et al.
(1995)
(incidence)
3.1
(0.6-9.0)
3 obs/1 exp
Spirtas et al. (1991)
Males
(mortality)
Females
(mortality)
4
5
a Standard mortality ratios (SMRs) are for subjects routinely exposed to tetrachloroethylene.
b SMR is for melanoma of the skin.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-222 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-3. Case-control studies: liver cancer
Reference
Lynge et al.
(2006)
Lynge et al.
(1995)
Suarez et al.
(1989)
Hernberg et al.
(1988)
Austin et al.
(1987)
Hardell et al.
(1984)
Hernberg et al.
(1984)
Stemhagen et
al. (1983)
Houten and
Sonnesso
(1980)
Exposure
assessment
Occupation
Occupation
Death
certificate
Mail survey
andlH
Interview
Mail survey
Mail survey
andlH
Interview
Interview
Exposure
classification
DC, other in
dry cleaning,
or
unclassifiable
DC
DC services
DC operators
Solvent
exposure
(tetrachloro-
ethylene, TCE,
CC14)
DC
Solvent
exposure
Solvent
exposure
DC, laundry,
and other
garment
services
DC
Number of
subjects
(cases/controls)
DC, 11/95
Other, 2/22
Unclassifiable,
23/121
17/85
1,742/1,742
377/385
80/146
102/200
126/175
265/530
102/
Odds ratio
(95% CI)
DC. 0.8 (0.4-
1.5)
Other in DC,
0.4(0.1-1.9)
Unclassifiable,
1.1(0.6-2.1)
No exposed
cases
1.0 (0.4-2.2)
PLC
0.6 (0.2-1.8)
0.6(0.3-1.2)
male, PLC
3.4(1.3-8.6)
female, PLC
no exposed
cases
1.8(1.0-3.4)
PLC
2.1(1.1-4.0)
HCC
2.3 (0.8-7.0)
PLC
2.5(1.0-6.1)
PLC
2.3(0.9-6.1)
HCC
Too small to
assess; 2 cases
in DC
Exposure
response
-
-
-
3
4
5
DC = Dry cleaner
PLC = Primary liver cancer
HCC = Hepatocellular cancer
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-223 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-4. Case-control studies: kidney cancer
Reference
Lynge et al.
(2006)
Parent et al.
(2000)
Pesch et al.
(2000a)
Dosemeci et
al. (1999)
Delahunt et
al. (1995)
Lynge et al.
(1995)
Mandel et al.
(1995)
Schlehofer et
al. (1995)
Auperin et al.
(1994)
Exposure
assessment
Occupation
Questionnaire,
interview
Questionnaire,
interview
Occupation
from
Interview to
JEM
Survey
Occupation
Interview
Interview
Occupation
Exposure
classification
DC, other in dry
cleaning, or
unclassifiable
DC/laundry worker
Tetrachloroethylene
JEM
males
females
JTEM
males
females
Tetrachloroethylene
(male)
Tetrachloroethylene
(female)
Tetrachloroethylene
(both)
DC
DC
DC industry (male)
DC solvents (male)
DC solvents
(female)
Iron/steel industry
(male)
Tetrachloroethylene
Metal industry
DC
Number of
subjects
(cases/controls)
DC, 29/196
Other, 9/34
Unclassifiable,
43/241
142/533
population and
1,900 cancer
935/4,298
273/462
165/225
438/687
914/12,756a
16/80
1,732/2,309
277/286
196/347
Odds ratio
(95% CI)
DC. 0.7 (0.4-
1.1
Other in DC,
1.2(0.5-2.5)
Unclassifiable,
0.8(0.5-1.2)
1.7(0.6-4.7)
1.4(1.0-2.0)
0.7 (0.3-2.2)
1.3 (0.7-2.3)
2.0 (0.5-7.8)
1.1 (0.7-1.7)
0.8(0.3-2.1)
1.1 (0.7-1.6)
1.9(0.3-13.9)
0.7 (0.2-3.6)
0.9 (0.3-2.4)
1.4(1.1-1.7)
1.6(1.0-2.7)
1.6(1.2-2.2)
2.5 (1.2-5.2)
1.6(1.1-2.5)
Too few
exposed
Exposure
response
None with
exposure
duration
None
-
-
None
-
-
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-224 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-4. Case-control studies: kidney cancer
Reference
Mellemgaard
etal. (1994)
McCredie
and Stewart
(1993)
Partanen et
al. (1991)
Harrington et
al. (1989)
Sharpe et al.
(1989)
Asal et al.
(1988)
Exposure
assessment
Interview
Interview
Mail survey
Questionnaire
Mail survey
Occupation
Exposure
classification
DC (male)
DC (female)
Solvents (male)
Solvents (female)
Iron/steel industry
(male)
DC industry
Solvent exposure
Iron/steel industry
DC, solvents
Iron/metalwork
Solvent exposure
DC
Organic solvents
Degreasing solvents
DC (male)
DC(female)
Metal degreasing
Number of
subjects
(cases/controls)
368/396
489/523
338/338
54/54
164/161
315/313 + 336
(hospital +
population)
Odds ratio
(95% CI)
2.3 (0.2-27.0)
2.9(0.3-33.0)
1.5(0.9-2.4)
6.4(1.2-23.0)
1.4(0.8-2.4)
2.7(1.1-6.7)
1.5(1.1-2.1)
2.1 (1.0-4.4)
Too few
exposed
1.9(0.9-3.8)
1.0(0.2-4.9)
No exposed
cases
1.7(0.9-3.2)
3.4 (0.9-12.7)
0.7 (0.2-2.3)
2.8 (0.8-9.8)
1.7(0.7-3.8)
Exposure
response
Trend in
years
worked
for
iron/steel
-
3
4
5
6
7
a Controls were subjects with primary tumors from sites other than the urinary tract.
DC = Dry cleaner
JEM = Job exposure matrix
JTEM = Job task exposure matrix
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-225 DRAFT-DO NOT CITE OR QUOTE
-------
o
Oi
Table 4B-5. Case control studies: lymphoid cancer
Reference
Exposure
assessment
Exposure
classification
Number of
subjects
(cases/controls)
Odds ratio
(95% CI)
Exposure
response
Hodgkin's Disease
Senior!
Costantini et al. (2001);
Miligietal. (1999)
Interview
Occupation: DC,
laundry worker,
presser
1,250/1,779
2.5 (0.3-24.6)a [Males]
3.5 (1.5-8.2)a [Females]
Non-Hodgkin's lymphoma
Lynge et al. (2006)
Mester et al. (2006)
(Hodgkin's and non-
Hodgkin's)
Miligi et al. (2006),
Seniori
Costantini et al. (2001),
Miligietal. (1999)
(non-Hodgkin's and
chronic lymphatic
leukemia)
Country-wide pension
database (for Danish and
Finnish subjects) or from
blinded interview (for Norse
and Swedish subjects)
Interview
Interview
DC, other in dry
cleaning, or
unclassifiable
Job title: DC,
launderer and
presser
Tetrachloroethylene
using JEM;
Occupation: DC,
launderer and
presser
DC, 42/219
Other, 8/48
Unclassifiable
52/255
710/710
1,428 /1,520
(Miligi et al.,
2006):
1,250/1,779
(Seniori
Costantini et
al. (2001)
DC, 1.0 (0.7-1. 4)b
Other in DC, 0.7 (0.3-1. 6)b
Unclassifiable, 0.9 (0.6-1. 4)b
Duration of employment:
1-10 yr, 0.8 (0.3-2.5)c
>10yr, 3.4(0.6-18.5)c
Allyrs, 1. 3(0.5-3 .2)c
Tetrachloroethylene, exposure
intensity:
Very low/low, 0.6 (0.3-l-2)d
Medium/high, 1.2 (0.6-2.5)d
Duration of medium/high
exposure:
<15yr, 1.3 (0.5-3.3)d
>15 yr, 3 cases-ORnot
presented
Tetrachloroethylene :
Diffuse NHL 1.9 (0.7-5. 7)d
DC, 1.6 (0.3-9. l)c [Males]
DC, 0.7 (0.3-1. 5)c [Females]
Linear trend,
p'0.72
oo S3'
I
I
§
***.
£3'
1
TO'
to
I
TO
H I
O >
HH Oq
H TO
O
H
W
-------
o
Oi
Table 4B-5. Case control studies: lymphoid cancer (continued)
Reference
Kato et al. (2005)
Fabbro-Peray et al.
(2001)
Blair etal. (1993)
Scherretal. (1992)
Hardell etal. (1981)
(Hodgkin's and non-
Hodgkin's)
Exposure
assessment
Interview
Interview
Interview
Interview
Survey
Exposure
classification
Occupation/Dry
Cleaning fluids
Dry cleaning
solvents
Occupation/JEM
Laundry/Garment
Services
Occupation: DC,
launderer, leather
product fabrication
Trichloroethylene,
Tetrachloroethy lene ,
styrene, benzene
Number of
subjects
(cases/controls)
376/463
445/1025
622/1,245
303/303
169/338
Odds ratio
(95% CI)
1.6 (0.5-5. l)e
1.0(0.6-1.6)f
2.0
(0.97-4.3)
1.6(0.6-4.0)
4.6
(1.9-11.4)
Exposure
response
-
—
<
Leukemia / Lymphoma
Senior!
Costantini et al. (2001)
Clavel etal. (1998)
Malone etal. (1989)
Interview
Self-administered
questionnaire
Interview
Occupation: DC,
launderer and
presser
Occupation: DC,
launderer
Occupation: DC,
launderer
383/269
226/425
427/1,683
All leukemia,
3. 3 (0.3-32.4) a [Males}
Hairy Cell Leukemia,
3.0(0.2-49.2)8
Chronic lymphocytic leukemia,
1.1 (0.6-2.0)h [All respondents]
0.9 (0.4-1. 8)h [Self-respondents]
—
oo S3'
I
I
§
^.
Co'
1
TO'
to
I
TO
H I
O
H
W
-------
o
Oi
Table 4B-5. Case control studies: lymphoid cancer (continued)
Reference
Exposure
assessment
Exposure
classification
Number of
subjects
(cases/controls)
Odds ratio
(95% CI)
Exposure
response
Leukemia, Childhood (Acute lymphocytic leukemia)
Infante-Rivard et al.
(2005)
Costas et al. (2002)
Interview; questionnaire on
maternal occupation. JEM
used to assign chemical-
specific exposures.
Interview; questionnaire to
parents which included
information on use of public
drinking water in the home.
Hydraulic mixing model used
to infer drinking water
containing TCE and other
solvents delivery to residence
Tetrachloroethylene
Residence receiving
water from Wells G
andH
790/790
19/37
Exposure period:
2 years before pregnancy up to
birth, 1.0(0.4-2.3)'
During pregnancy, 0.8 (0.3-2.3)1
Any exposure,
0.9 (0.4-2.2)1
Lowest exposure rank, 1.0 (0.4-
2.6)1
Highest exposure rank, 0.6 (0.5-
12.8)1
Exposure period of mother:
From 2 years before conception
to case diagnosis,
Never exposed, 1.0
Least exposed, 5.0 (0.8-33.5)1
Most exposed, 3.6 (0.5, 24.8)1
During 2 years before
conception,
Never exposed, 1.0
Least exposed, 2.5 (0.4-15.2)1
Most exposed, 2.8 (0.3-26.4)1
From birth to diagnosis,
Never exposed, 1.0
Least exposed, 1.8 (0.3-1 0.8)1
Most exposed, 0.9 (0.2-4.6)1
-
oo S3'
I
I
§
***.
£3'
1
TO'
to
I
TO
H I
O
H
W
-------
o
Oi
Table 4B-5. Case control studies: lymphoid cancer (continued)
Reference
Shu etal. (1999)
Lowengart et al. (1987)
Exposure
assessment
Interview: questionnaire to
assess parental occupation
using job-industry title and
self-reported exposure
history.
Interview; questionnaire
sought information on
occupation and occupational
exposures of both parents.
Exposure
classification
Tetrachloroethylene
Tetrachloroethylene
Number of
subjects
(cases/controls)
1842/1986
123/123
Odds ratio
(95% CI)
During pregnancy
Never exposed, 1.0
Least exposed, 3.5 (0.2, 58. 1)1
Most exposed, 14.30 (0.9-224)1
Exposure period of mother:
Anytime, 0.4 (0.1-1. 4)k
Preconception, 1.4 (0.2-8.6)k
During pregnancy, 1.3 (0.2-8. 4)k
Postnatal, 0.4 (0.1-1. 5)k
Exposure period of father:
One year before pregnancy,
00(^ = 0.39)'
During pregnancy,
00(^ = 0.39)'
After delivery,
00(0.19-00)'
Exposure
response
Linear trend,
/K0.05
oo S3'
I
I
§
***.
£3'
1
TO'
to
I
TO
Co
H I
O >
HH Oq
H TO
O
H
W
a Adjusted for age
b RR from logistic regression models adjusting for matching criteria (country, sex, five-year age group and five-year calendar period at the time of
diagnosis of the case).
0 Adjusted for smoking and alcohol consumption.
d Adjusted for sex, age, education, and area.
e Adjusted for age at index date, family history of hematologic cancer, college education, surrogate status, interview year, body mass index 10
years before interview, pain-relieving drugs use, antibiotic use, household pesticide products use, and work duration involving pesticide exposure.
f Adjusted for age, gender, area, and education level.
B Conditional logistic regression with the baseline category of unexposed workers, adjusted for smoking and farming as an occupation.
h Adjusted for race, age, education, sex, and study site
1 Adjusted for maternal age and level of schooling.
J Adjusted using a composite covariate to control for socio-economic status, maternal age at birth of child and breast-feeding.
k Adjusted for maternal education, race, and family income.
-------
1 Discordant pair analysis
DC = Dry cleaner
JEM = Job exposure matrix
RR = Relative risk
-------
1
2
Table 4B-6. Case-control studies: esophagus
Reference
Lynge et
al. (2006)
Parent et
al. (2000)
Vaughan et
al. (1997)
Exposure
assessment
Occupation
Occupation-
interview
Occupation-
interview
Exposure
classification
DC, other in dry
cleaning, or
unclassifiable
Solvents
Dry cleaner
Tetrachloroethy lene ,
probable exposure
Cumulative
exposure:
-1-29 ppm-y
-30+ ppm-y
Number of
subjects
(cases/controls)
DC, 8/86
Other, 5/31
Unclassifiable
18/108
99/533
315/313+336
(hospital +
population)
Odds ratio
(95% CI)
DC, 0.8 (0.3-
1.7)
Other in DC, 1.2
(0.4-3.6)
Unclassifiable,
2.0 (0.9-4.6)
1.6" (0.8-2.5)
Too few
exposed
subjects
precluded an
analysis of this
occupational
group.
6.4a (0.6-68.9)
11.9s (1.1-
124.0)
0 exposed cases
Exposure
response
None
a Odds ratio for squamous cell carcinoma, adjusted for a number of covariates including smoking and
alcohol consumption.
4
5
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-231 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-7. Case-control studies: bladder cancer
Reference
Lynge et
al. (2006)
Pesche et
al. (2000b)a
Teschke et
al. (1997)
Silverman
etal.
(1989a, b)
Smith et al.
(1985)
Exposure
assessment
Occupation
Questionnaire-
interview
Occupation-
interview
Occupation-
interview
Occupation
Exposure
classification
DC, other in dry
cleaning, or
unclassifiable
Tetrachloroethylene
- substantial exp
JEM
males
females
ITEM
males
DC/laundry worker
DCb, ironer, presser
- male, white
- male,
non-white
DC/laundry worker
Number of
subjects
(cases/controls)
CD, 93/292
Other, 12/52
Unclassifiable,
57/234
1,035/4,298
105/139
(population
controls)
2,100/3,874
126/383
103/1,869
Odds ratio
(95% CI)
DC, 1.4(1.1-
1.9)
Other in DC,
1.1(0.6-2.1)
Unclassifiable,
1.2(0.8-1.8)
1.4(1.0-1.9)
0.7 (0.2-2.5)
1.8(1.1-3.7)
2.3d
(0.4-13.9)
2.8(l.l-7.4)c
1.3 (0.9-2.0),
nonsmoker;
3.0 (1.8-5.0),
former smoker;
2.9 (2.4-6.5),
current smoker
Exposure
response
Increased risk
with
>10 years of
exposure
Duration of
employment
showed no
effect
Duration of
exposure
showed no
effect
3
4
5
6
7
a Cases are histological confirmed urothelial cancers, a category comprised of urinary bladder, ureter, and renal
pelvis. Odds ratio and 95% CI adjusted for age, study center, and smoking.
b A prior susceptible occupation.
0 Odds ratio adjusted for smoking and employment in other high-risk occupations.
d Odds ratio adjusted for sex, age, and cigarette smoking.
DC = Dry cleaner
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-232 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-8. Case-control studies: lung
Reference
Pohlabeln
etal.
(2000)
Brownson
etal.
(1993)
Exposure
assessment
Occupation-
interview
Occupation-
interview
Exposure
classification
DC/laundry
worker
DC
Number of
subjects
(cases/controls)
650a/l,542
429b/ 1,021
Odds ratio
(95% CI)
Females, 1.83
(0.98-3.4)
Ever, 1.8
(1.1-3.0)
> 1.1 25 years,
2.9(1.5-5.4)
Exposure
response
"
Trend with
duration of
employment
4
5
6
a Cases are nonsmokers.
b Cases are either nonsmokers or ex-smokers.
DC = Dry cleaner
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-233 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-9. Case-control studies: pancreas
Reference
Lin and
Kessler
(1981)
Partanen et
al. (1991)
Kernan et
al. (1999)
Exposure
assessment
Occupational
history
obtained by
interview
Job title and
industrial
branch
Occupation
as coded on
death
certificate
Exposure
classificatio
n
DC/gasoline
DC,
launderer,
presser
Tetrachloro-
ethylene
Number of
subjects
(cases/controls)
61I6T
625/1,700
63,0977
252,386
Odds ratio
(95% CI)
2.8(1.1-7.1)
2.4(0.3-17.2)b
High intensity of
exposure:
BF,0.8C (0.6-1.2)
BM, 1.2C (0.9-1.5)
WF, lc (0.9-1. 2)
WM, 0.9C (0.8-1.0)
High probability
of exposure:
BF,0.7d (0.5-1.1)
BM, ld (0.7-1. 4)
WF, 0.8C (0.6-1.1)
WM, 0.9C (0.8-1.0)
Exposure
response
Risk increased
with increasing
duration of
exposure.
>10 years
exposure,
OR = 5.1
(95% CI:
1.2B20.0)
-
-
3
4
5
6
7
8
9
10
11
12
13
14
15
a Cases and controls selected from hospital admissions. Control subjects were free of cancer and were
selected at random from contemporaneous admissions.
b Odds ratio (OR) adjusted for age, gender, smoking, alcohol consumption, and diabetes.
0 OR adjusted for age, metropolitan status, region of residence, and marital status.
BF = Black female
BM = Black male
DC = Dry cleaner
WF = White female
WM = White male
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-234 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-10. Case-control studies: intestineCsmall bowel or colon
Reference
Fredriksson et al.
(1989)
Colon
Kaerlev et al.
(2000)
Adeno-carcinoma
of the small bowel
Exposure
assessment
Occupational
survey-
Questionnaire
Questionnaire-
interview
Exposure
classification
DC
Trichloroethylene
among DC
DC industry
Number of
subjects
(cases/controls)
312/623
79/2,649
(population+
hospital
controls)
Odds ratio
(95% CI)
2.0(0.5-7.1),
females
7.4
(1.1-47.0)
6.5b(1.6-
26.9)
Exposure
response
Positive trend
with duration
of employment
(industry and
occupation)
4
5
6
7
8
9
10
11
12
13
a Odd ratio adjusted for sex, country, and year of birth.
BF = Black female
BM =Black male
DC = Dry cleaner
WF =White female
WM =White male
Table 4B-11. Case-control studies: brain
Reference
Heineman
etal. (1994)
Exposure
assessment
Occupation
from
interview to
JEM
Exposure classification
Organic solvents
Tetrachloroethylene
Trichloroethylene
MeCl2
Number of
subjects
(cases/controls)
300/320
Odds ratio
(95% CI)
1.3 (0.9-1.8)
1.2(0.8-1.6)
1.1 (0.8-1.6)
1.3 (0.9-1.8)
Exposure
response
Positive
trend for
methylene
chloride
14
15
16
17
18
19
JEM= Job exposure matrix
Table 4B-12. Case-control studies: breast
Reference
Band et al.
(2000)
Exposure
assessment
Occupational
survey-
questionnaire
Exposure
classification
Dry cleaner/
Laundry worker
Number of
subjects
(cases/controls)
1,018/1,025
Odds ratio
(95% CI)
1.8a (0.4-7.7),
pre-menopausal
4.9b (1.3-18.7), post-
menopausal
Exposure
response
20
21
22
a Ever employed as dry cleaner or laundry worker.
b Usually employed as dry cleaner or laundry worker.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-235 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-13. Community-based studies
Outcome
Hodgkin's
disease
Leukemia
Multiple
myeloma
Non-
Hodgkin's
lymphoma
(NHL)
Bladder
Kidney
Exposure
assessment
Residence in town
VOCs other than
THMs in town water
Tetrachloroethylene
in town water
Residence in town;
VOCs in water
Residence in town
Distribution of
drinking water with
tetrachloroethylene
Residence in town
Tetrachloroethylene
in town water
Residence in town
Distribution of
drinking water with
tetrachloroethylene
Solvents in town
water (including
tetrachloroethylene,
TCE)
Distribution of
drinking water with
tetrachloroethylene
Reference
Vartiainen
(1985)
Fagliano et
al.
(1990)
Cohn et al.
(1994)
Lagakos et
al. (1986)
Vartiainen
(1985)
Aschengrau
etal. (1993)
Vartiainen
(1985)
Cohn et al.
(1994)
Vartiainen
(1985)
Aschengrau
etal. (1993)
Mallin
(1990)
Aschengrau
etal. (1993)
Relative risk
0.8 (0.3-1.7)
Hausjavi, Finland
1.4 (0.7-2.5)
Hattula, Finland
1.5 (1-2.2) f
1.0 (0.7-1. 5) m
1.4(1. 1-1.9) f
1.1 (0.8-1.4) m
2.2 (1.5-2.9)
1.2 (0.8-1.7)
Hansjavi, Finland
0.7(0.4-1.1)
Hattula, Finland
2.1(0.9-5.2)a'b
8.3(1. 5-45. 3)b'c
0.7 (0.3-1.3)
Hausjavi, Finland
0.6 (0.2-1.3)
Hattula, Finland
1.1 (0.9-1.3) f
1.1 (0.9-1.3) nf
0.6(0.3-1.1)
Hausjavi, Finland
1.4 (1.0-2.0)
Hattula, Finland
1.4 (0.7-2.9)3
4.0 (0.7-25. l)b'c
1.7m
2.6 f
1.1 (OA-2.^b
Exposed
cases
6
11
28
25
56
63
20
33
19
7
2
7
6
87
78
14
31
13
4
Not in
paper
6
Comments
Exposure response for
females
Exposure response for
females
Exposure response not
consistent
Risk for 90th percentile
higher than risk for
ever exposed
High-grade NHL;
2.7 (1.3-5.6) f
0.4 (0.1-1. 8) m
Risk for 90th percentile
higher than risk for
ever exposed
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-236 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 4B-13. Community-based studies (continued)
Outcome
Liver
Colon-
rectum
Lung
Breast
Exposure
assessment
Residence in village
Distribution of
drinking water with
tetrachloroethylene
Distribution of
drinking water with
tetrachloroethylene
Distribution of
drinking water with
tetrachloroethylene
Reference
Vartiainen et
al. (1993)
Paulu et al.
(1999)
Paulu et al.
(1999)
Aschengrau
etal. (1998)
Paulu et al.
(2002)
Aschengrau
et al. (2003)
Vieira et al.
(2005)
Relative risk
0.7 (0.3-1.4)
Hausjavi, Finland
0.6 (0.2-1.3)
Hattula, Finland
1.1(0.7-1.5)"' b
l(0.3-2.8)b'c
l.l(Q.l-l.lfb
3.7(1.0-11.7)b'c
l.l(Q.l-l.lfb
1.1 (0.3-3.8)b'c
2.2(l.l-4.3)b'c'd
1.3(0.7-2.6)b'c'e
1.0 (0.4-2.5/
Exposed
cases
7
6
44
5
33
5
36
4
28
18
10
Comments
-
Risk for 90th percentile
higher than risk for
ever exposed
Exposure response
with tetrachloro-
ethylene exposure
level
Risk for 90th percentile
higher than risk for
ever exposed
None
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Resided in home receiving tetrachloroethylene-contaminated drinking water.
Odds ratio adjusted for sex, age at diagnosis, vital status at interview, educational level, and occupational
exposure to solvents. The usual number of cigarettes smoked and history of cigar and pipe use, living with a
smoker, and occupational history associated with lung cancer were additionally controlled in the lung cancer
analyses. All multivariate models of breast cancer included adjustment for age of diagnosis, vital status at
interview, family history of breast cancer, age at first live birth or stillbirth, prior history of breast cancer/disease,
and occupational exposure to solvents.
Resided in home receiving water at a high relative delivered dose of tetrachloroethylene, defined by Aschengrau
et al. (1993, 1998, 2004) and Paulu et al. (1999) as a level above the 90th percentile.
Analyses were conducted among the present study population.
Analyses were conducted among the present study population and the study population reported in Aschengrau et
al. (1998).
Odds ratio adjusted for age at diagnosis or index year, family history of breast cancer, personal history of breast
cancer before current diagnosis or index year, age at first live-birth or still birth, vital status at interview, and
occupational exposure to solvents. The tetrachloroethylene exposure variable, personal delivered dose or FDD,
was inferred from a model that extended the relative delivered dose (ROD) model to include information on tap
water consumption and bathing habits.
f = Female
m = Male
TCE = Trichloroethylene
THMs = Trihalomethanes
VOCs = Volatile organic compounds
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-237 DRAFT-DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR CHAPTER 4
2
3
4 Aggazzotti, G; Fantuzzi, G; Predieri, G; et al. (1994a) Indoor exposure to perchloroethylene (PCE) in individuals
5 living with dry-cleaning workers. Sci Total Environ 156:133-137.
6
7 Aggazzotti, G; Fantuzzi, G; Righi, E; et al. (1994b) Occupational and environmental exposure to perchloroethylene
8 (PCE) in dry cleaners and their family members. Arch Environ Health 49(6):487-93.
9
10 Ahlborg, G, Jr. (1990) Pregnancy outcome among women working in laundries and dry cleaning shops using
11 tetrachloroethylene. Am J Ind Med 17:567-575.
12
13 Ahn, YS; Zerban, H; Grobholz, R; et al. (1992) Sequential changes in glycogen content, expression of glucose
14 transporters and enzymic patterns during development of clear/acidophilic cell tumors in rat kidney. Carcinogenesis
15 13:2329-2334.
16
17 Albers, JW; Wald, JJ; Werner, RA; et al. (1999) Absence of polyneuropathy among workers previously diagnosed
18 with solvent-induced toxic encephalopathy. J Occup Environ Med 41:500-509.
19
20 Alden, CL. (1985) Species, sex, and tissue specificity in toxicologic and proliferative responses. Toxicol Pathol
21 13:135-140.
22
23 Alden, WW; Repta, AJ. (1984) Exacerbation of cisplatin-induced nephrotoxicity by methionine. Chem Biol Interact
24 48:121-124.
25
26 Altmann, L; Bottger, A; Wiegand, H (1990) Neurophysiological and psychophysical measurements reveal effects of
27 acute low-level organic solvent exposure in humans. Int Arch Occup Environ Health 62:493-499.
28
29 Altmann, L; Weigand, H; Bottger, A; et al. (1992) Neurobehavioral and neurophysiological outcomes of acute
30 repeated perchloroethylene exposure. Applied Psychology: An International Review 41:269-279.
31
32 Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
33 low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
34
3 5 Amler, RW; Mueller, PW; Schuyltz, MG. (1998) Biomarker of kidney function for environmental health field
36 studies.
37
3 8 Andersen, A; Barlow, L; Engeland, A; et al. (1999) Work-related cancer in the Nordic countries. Scand J Work
39 Environ Health 25(suppl 2): 116.
40
41 Andrys, C; Hanovcova, I; Chylkova, V; et al. (1997) Immunological monitoring of dry-cleaning shop workers--
42 exposure to tetrachloroethylene. Cent EurJ Public Health 5:136-142.
43
44 Anger, WK; Liang, YX; Nell, V; et al. (2000) Lessons learned: 15 years of the WHO-NCTB: a review.
45 Neurotoxicology 21:837-846.
46
47 Anna, CH; Maronpot, RR; Pereira, MA; et al. (1994) Ras proto-oncogene activation in dichloroacetic acid-,
48 trichloroethylene-and tetrachloroethylene-induced liver tumors in B6C3F1 mice. Carcinogenesis 15:2255-2261.
49
50 Antti-Poika, M. (1982a) Prognosis of symptoms in patients with diagnosed chronic organic solvent intoxication. Int
51 Arch Occup Environ Health 51:81-89.
52
53 Antti-Poika, M. (1982b) Overall prognosis of patients with diagnosed chronic organic solvent intoxication. Int Arch
54 Occup Environ Health 51:127-138.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-238 DRAFT-DO NOT CITE OR QUOTE
-------
1 Anttila, S; Hirvonen, A; Husgafvel-Pursiainen, K; et al. (1994) Combined effect of CYP1 Al inducibility and
2 GSTM1 polymorphism on histological type of lung cancer. Carcinogenesis 15:1133-1135.
3
4 Anttila, A; Pukkala, E; Sallmen, M; et al. (1995) Cancer incidence among Finnish workers exposed to halogenated
5 hydrocarbons. JOccup Environ Med 37:797-806.
6
7 Aranyi, C; O'Shea, W J; Graham, J A; Miller, F J (1986) The effects of inhalation of organic chemical air
8 contaminants on murine lung host defenses. Fundam Appl Toxicol 6:713-720.
9
10 Arlien-Soborg, P. (1982) Solvent neurotoxicity. Boca Raton, FL: CRC Press.
11
12 Aryal, B K; Khuder, S A; Schaub, E A (2001) Meta-analysis of systemic sclerosis and exposure to solvents. Am J
13 Ind Med 40:271-274.
14
15 Asal, NR; Geyer, JR; Risser, DR; et al. (1988) Risk factors in renal cell carcinoma. II. medical history, occupation,
16 multivariate analysis, and conclusions. Cancer Detect Prev 13:263-279.
17
18 Aschengrau, A; Seage, GR. (2003) Essentials of epidemiology for public health. Boston: Jones and Bartlett.
19
20 Aschengrau, A; Ozonoff, D; Paulu, C; et al. (1993) Cancer risk and tetrachloroethylene-contaminated drinking water
21 in Massachusetts [see comments]. Arch Environ Health 48:284-292.
22
23 Aschengrau, A; Paulu, C; Ozonoff, D. (1998) Tetrachloroethylene-contaminated drinking water and the risk of
24 breast cancer. Environ Health Perspect 106 Suppl 4:947-953.
25
26 Aschengrau, A; Rogers, S; Ozonoff, D. (2003) Perchloroethylene-contaminated drinking water and the risk of breast
27 cancer: additional results from Cape Cod, Massachusetts, USA. Environ Health Perspect 111:167-173.
28
29 Ashby, J. (1996) Alpha 2 mu-globulin nephropathy in white ravens. Environ Health Perspect 104:1264-1267.
30
31 Ashby, J. Brady, A; Elcombe, CR; et al. (1994) Mechanically-based human hazard assessment of peroxisome
32 proliferator-inducedhepatocarinogenesis. Hum Exp Toxicol 13(suppl2):Sl-117.
33
34 ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
3 5 (Update). Prepared by Sciences, International, under subcontract to Research Triangle Institute. ATSDR. Atlanta,
36 GA.
37
3 8 ATSDR (Agency for Toxic Substances and Disease Registry). (1998) Volatile organic compounds in drinking water
39 and adverse pregnancy outcomes. Atlanta, GA.
40
41 ATSDR (Agency for Toxic Substances and Disease Registry). (2001) Interaction Profile: 1,1,1-Trichloroethane, 1,1-
42 Dichloroethane, Trichloroethylene, and Tetrachloroethylene (Draft). Available online at
43 http://www.atsdr.cdc.gov/interactionprofiles/ip02.html. Accessed November 12, 2004.
44
45 ATSDR (Agency for toxic substances and disease registry) (2003) Survey of specific childhood cancers and birth
46 defects among children whose mothers were pregnant while living at U.S. Marine Corps Base Camp Lejeune, North
47 Carolina, 1968-1985.
48
49 Auboeuf, D; Rieusset, J; Fajas, L; et al. (1997) Tissue distribution and quantification of the expression of mRNAs of
50 peroxisome proliferator-activated receptors and liver X receptor-alpha in humans: no alteration in adipose tissue of
51 obese and NIDDM patients. Diabetes 46:1319-1327.
52
53 Auperin, A; Benhamou, S; Ory-Paoletti, C; et al. (1994) Occupational risk factors for renal cell carcinoma: a case-
54 control study. Occup Environ Med 51:426-428.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-239 DRAFT-DO NOT CITE OR QUOTE
-------
1 Austin, H; Delzell, E; Grufferman, S; et al. (1987) Case-control study of hepatocellular carcinoma, occupation, and
2 chemical exposures. JOccupMed 29:665-669.
3
4 Ayensu, W K; Tchounwou, P B; McMurray, R W (2004) Molecular and cellular mechanisms associated with
5 autoimmune diseases. Int J Environ Res Public Health 1:39-73.
6
7 Bagby, GCJ. (1994) Hematopoiesis. In: Stamatoyannopoulos, G; Nienhuis, AW; Majerus, PW; et al., eds. The
8 molecular basis of blood diseases. Philadelphia: W.B. Saunders; pp. 7-103.
9
10 Baggetto, LG. (1992) Deviant energetic metabolism of glycolytic cancer cells. Biochimie 74:959-974.
11
12 Bagnell, PC; Ellenberger, HA. (1977) Obstructive jaundice due to a chlorinated hydrocarbon in breast milk. Can
13 MedAssoc 1117:1047-1048.
14
15 Bale, A S; Adams, T L; Bushnell, P J; Shafer, T J; Boyes, W K (2005) Role of NMD A, nicotinic, and GAB A
16 receptors in the steady-state visual-evoked potential in rats. Pharmacol Biochem Behav 82:635-645.
17
18 Band, PR; Le, ND; Fang, R; et al. (2000) Identification of occupational cancer risks in British Columbia: a
19 population-based case-control study of 995 incident breast cancer cases by menopausal status, controlling for
20 confounding factors. JOccup Environ Med 42:284-310.
21
22 Barbier, O; Duran-Sandoval, D; Pineda-Torra, I; et al. (2003) Peroxisome proliferator-activated receptor-alpha
23 induces hepatic expression of the human bile acid glucuronidating UDP-gllucuronosyltransferase 2B4 enzyme. J
24 Biol Chem 278(35):32852-32860.
25
26 Barrass, N; Stewart, M; Warburton, S; et al. (1993) Cell proliferation in the liver and thyroid of C57B1/10J mice
27 after dietary administration of chlordane. Environ Health Perspect 101 Suppl 5:219-223.
28
79
Baylin, SB; Herman, JG; Graff, JR; et al. (1998) Alterations in DNA methylation: a fundamental aspect of neoplasia
30 Adv Cancer Res 72:141-196.
31
32 Beaudreuil, S; Lasfargues, G; Laueriere, L; El, G Z; Fourquet, F; Longuet, C; Halimi, J M; Nivet, H; Buchler, M
33 (2005) Occupational exposure in ANCA-positive patients: a case-control study. Kidney Int 67:1961-1966.
34
3 5 Beckstead, MJ; Weiner, JL; Eger, El; et al. (2000) Glycine and gamma-aminobutyric acid(A) receptor function is
36 enhanced by inhaled drugs of abuse. Mol Pharmacol 57:1199-1205.
37
3 8 Beliles, RP. (2002) Concordance across species in the reproductive and developmental toxicity of
3 9 tetrachloroethylene. Toxicol Ind Health 18:91-106.
40
41 Beliles, RP; Brusick, DJ; Mecler, FJ. (1980) Teratogenic-mutagenic risk of workplace contaminants:
42 trichloroethylene, perchloroethylene and carbon disulfide. Prepared by Litton Bionetics, Inc. NTIS Publication No.
43 PB-82 185-075, NIOSH Contract Report No. 201-77-0047. Available from: National Technical Information Service,
44 Springfield, VA.
45
46 Bellinger, D; Hu, H; Titlebaum, L; et al. (1994) Attentional correlates of dentin and bone lead levels in adolescents.
47 Arch Environ Health 49:98-105.
48
49 Bentley, P; Calder, I; Elcombe, C; et al. (1993) Hepatic peroxisome proliferation in rodents and its significance for
50 humans. Food Chem Toxicol 31:857-907.
51
52 Bergamaschi, E; Mutti, A; Bocchi, MC; et al. (1992) Rat model of perchloroethylene-induced renal dysfunctions.
53 Environ Res 59:427-439.
54
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-240 DRAFT-DO NOT CITE OR QUOTE
-------
1 Berger, T and Horner, CM (2003) In vivo exposure of female rats to toxicants may affect oocyte quality. Reprod
2 Toxicol 17:273-281.
3
4 Berlin, K; Edling, C; Persson, B; Ahlborg, et al. (1995) Cancer incidence and mortality of patients with suspected
5 solvent-related disorders. Scand J Work Environ Health 21:362-367.
6
7 Berman, E; Schlicht, M; Moser, VC; MacPhail, RC. (1995) A multidisciplinary approach to toxicological
8 screening: I. Systemic toxicity. J Toxicol Environ Health 45:127-143.
9
10 Bernard, A; Lauwerys, R. (1995) Low-molecular-weight proteins as markers of organ toxicity with special reference
11 to Clara cell protein. Toxicol Lett 77:145-151.
12
13 Bhattacharya, RK; Schultze, MO. (1971) Properties of DNA isolated from tissues of calves treated with S-(l,2-
14 dichlorovinyl)-L-cysteine. II. Primer-template activity for bacterial DNA polymerases. Arch Biochem Biophys
15 145:575-582.
16
17 Bhattacharya, RK; Schultze, MO. (1972) Properties of DNA treated with S-(l,2-dichlorovinyl)-L-cysteine and a
18 lyase. Arch Biochem Biophys 153:105-115.
19
20 Bhunya SP, Behera BC. (1987). Relative genotoxicity of trichloroacetic acid (TCA) as revealed by different
21 cytogenetic assays: bone marrow chromosome aberration micronucleus and sperm-head abnormality in the mouse.
22 Mutat Res 188: 215-221.
23
24 Bhunya, SP; Jena GB. (1996) The evaluation of clastogenic potential of trichloroacetic acid (TCA) in chick in vivo
25 test system. Mutat Res 367: 254-259.
26
27 Birner, G; Richling, C; Henschler, D; et al. (1994) Metabolism of tetrachloroethene in rats: identification of N
28 epsilon- (dichloroacetyl)-L-lysine and N epsilon-(trichloroacetyl)-L-lysine as protein adducts. Chem Res Toxicol
29 7:724-732.
30
31 Birner, G; Rutkowska, A; Dekant, W. (1996) N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and 2,2,2-
32 trichloroethanol: two novel metabolites of tetrachloroethene in humans after occupational exposure. Drug Metab
33 Dispos 24:41-48.
34
3 5 Blair, A; Stewart, PA; Tolbert, PE; et al. (1990) Cancer and other causes of death among a cohort of dry cleaners.
36 BrJIndMed 47:162-168.
37
3 8 Blair, A; Linos, A; Stewart, PA; et al. (1993) Evaluation of risks for non-Hodgkin's lymphoma by occupation and
39 industry exposures from a case-control study. Am JIndMed 23:301-312.
40
41 Blair, A; Petralia, SA; Stewart. PA. (2003) Extended mortality follow-up of a cohort of dry cleaners. Ann Epidemiol
42 13(l):50-56.
43
44 Blossom, SJ; Gilbert, KM. (2006) Exposure to a metabolite of the environmental toxicant, trichloroethylene,
45 attenuates CD4+ T cell activation-induced cell death by metalloproteinase-dependent FasL shedding. Toxicol Sci
46 92:103-114.
47
48 Blossom, SJ; Pumford, NR; Gilbert, KM. (2004) Activation and attenuation of apoptosis of CD4+ T cells following
49 in vivo exposure to two common environmental toxicants, trichloroacetaldehyde hydrate and trichloroacetic acid. J
50 Autoimmun 23:211-220.
51
52 Blot, WJ; McLaughlin, JK. (1999) The changing epidemiology of esophageal cancer. Semin Oncol 26:2-8.
53
54 Bogen, KT; McKone, TE. (1988) Linking indoor air and pharmacokinetic models to assess tetrachloroethylene risk.
55 Risk Anal 8:509-520.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-241 DRAFT-DO NOT CITE OR QUOTE
-------
1 Boice, JD; Marano, DE; Fryzek, JP; et al. (1999) Mortality among aircraft manufacturing workers. Occup Environ
2 Med 56:581-597.
3
4 Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol
5 70(6):347-55.
6
7 Bonse, G; Henschler, D. (1976) Chemical reactivity, biotransformation, and toxicity of polychlorinated aliphatic
8 compounds. CRC Crit Rev Toxicol 4:395-409.
9
10 Bonse, G; Urban, T; Reichert, D; et al. (1975) Chemical reactivity, metabolic oxirane formation and biological
11 reactivity of chlorinated ethylenes in the isolated perfused rat liver preparation. Biochem Pharmacol 24:1829-1834.
12
13 Bosco, MG; Figa-Talamanca, I; Salerno, S. (1987) Health and reproductive status of female workers in dry cleaning
14 shops. Int Arch Occup Environ Health 59:295-301.
15
16 Boulet, LP (1988) Increases in airway responsiveness following acute exposure to respiratory irritants. Reactive
17 airway dysfunction syndrome or occupational asthma?. Chest 94:476-481.
18
19 Bove, FJ; Fulcomer, MC; Klotz, JB; et al. (1992) Public drinking water contamination and birthweight, and selected
20 birth defects: a case-control study. New Jersey Department of Health, New Brunswick, NJ.
21
22 Bove, FJ; Fulcomer, MC; Klotz, JB; et al. (1995) Public drinking water contamination and birth outcomes [see
23 comments]. AmJEpidemiol 141:850-862.
24
25 Bowler, RM; Mergler, D; Huel, G; et al. (1991) Neuropsychological impairment among former microelectronics
26 workers. Neurotoxicology 12:87-103.
27
28 Bowman, KJ. (1982) A method for quantitative scoring of the Farnsworth Panel D-15. Acta Opthalmol 60:907-916.
29
30 Breslow, NE; Day, NE. (1994) Statistical method in cancer research. Vol. 2: The design and analysis of cohort
31 studies. New York: Oxford University Press.
32
33 Briving, C; Jacobson, I; Hamberger, A; et al. (1986) Chronic effects of perchloroethylene and trichloroethylene on
34 the gerbil brain amino acids and glutathione. Neurotoxicology 7:101-108.
35
36 Broadwell, DK; Darcey, DJ; Hudnell, HK; et al. (1995) Work-site clinical and neurobehavioral assessment of
37 solvent-exposed microelectronics workers. Am J Ind Med 27:677-698.
38
39 Brodkin, CA; Daniell, W; Checkoway, H; et al. (1995) Hepatic ultrasonic changes in workers exposed to
40 perchloroethylene. Occup Environ Med 52:679-685.
41
42 Bronzetti, G; Galli, A; Corsi, C; et al. (1984) Genetic and biochemical investigation on chloral hydrate in vitro and
43 in vivo. Mutat Res 141: 19-22.
44
45 Brown, RP; Delp, MD; Lindstedt, SL; et al. (1997) Physiological parameter values for physiologically based
46 pharmacokinetic models. Toxicol Industrial Health 13:407-484.
47
48 Brown, LM; Hoover, R; Silverman, D; et al. (2001) Excess incidence of squamous cell esophageal cancer among
49 US Black men: role of social class and other risk factors. Am J Epidemiol 153:114-122.
50
51 Brownson, RC; Alavanja, MC; Chang, JC. (1993) Occupational risk factors for lung cancer among nonsmoking
52 women: a case-control study in Missouri (United States). Cancer Causes Control 4:449-454.
53
54 Bruschi, SA; Bull, RJ. (1993) In vitro cytotoxicity of mono-, di-, and trichloroacetate and its modulation by hepatic
55 peroxisome proliferation. FundamAppl Toxicol 21:366-375.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-242 DRAFT-DO NOT CITE OR QUOTE
-------
1 Buben, JA; O=Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
2 and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
O
4 Buelke-Sam, J; Kimmel, CA; Adams, J; et al. (1985) Collaborative behavioral teratology study: results.
5 Neurobehav Toxicol Teratol 7:591-624.
6
7 Bukowski, JA; Huebner, WW; Schnatter, AR; et al. (2003) An analysis of the risk of B-lymphocyte malignancies in
8 industrial cohorts. J Toxicol Environ Health Part A 66:5 81 -5 97.
9
10 Bull, RJ. (2000) Mode of action of liver tumor induction by trichloroethylene and its metabolites, trichloroacetate
11 and dichloroacetate. Environ Health Perspect 108 Suppl 2:241-259.
12
13 Bull, RJ. (2004) Trichloroethylene and liver tumors in mice. Symposium on new scientific research related to the
14 health effects of TCE. Washington, DC.
15
16 Bull, RJ; Sanchez, IM; Nelson, MA; etal. (1990) Liver tumor induction in B6C3F1 mice by dichloroacetate and
17 trichloroacetate. Toxicology 63:341-359.
18
19 Bull, RJ; Orner, GA; Cheng, RS; et al. (2002) Contribution of dichloroacetate and trchloroacetate to liver tumor
20 induction in mice by trichloroethylene. Toxicol App Pharmacol 182(l):55-65.
21
22 Bull, RJ; Sasser, LB; Lei, XC. (2004) Interactions in the tumor-promoting activity of carbon tetrachloride,
23 trichloroacetate, and dichloroacetate in the liver of male B6C3F1 mice. Toxicol 199:169-83.
24
25 Bursch, W; Lauer, B; Timmermann-Trosiener, I; et al. (1984) Controlled death (apoptosis) of normal and putative
26 preneoplastic cells in rat liver following withdrawal of tumor promoters. Carcinogenesis 5:453-458.
27
28 Byczkowski, JZ; Fisher, JW. (1994) Lactational transfer of tetrachloroethylene in rats. Risk Anal 14:339-349.
29
30 Byczkowski, JZ; Fisher, JW. (1995) A computer program linking physiologically based pharmacokinetic model
31 with cancer risk assessment for breast-fed infants. Comput Methods Programs Biomed 46:155-163.
32
33 Byers, VS; Levin, AS; Ozonoff, DM; et al. (1988) Association between clinical symptoms and lymphocyte
34 abnormalities in a population with chronic domestic exposure to industrial solvent- contaminated domestic water
35 supply and a high incidence of leukaemia. Cancer Immunol Immunother 27:77-81.
36
37 Cai, SX; Huang, MY; Chen, Z; et al. (1991) Subjective symptom increase among dry-cleaning workers exposed to
3 8 tetrachloroethylene vapor. Ind Health 29:111-121.
39
40 Cal EPA (California Environmental Protection Agency). (2001) Public health goal for tetrachloroethylene in
41 drinking water. Prepared by Office of Environmental Health Hazard Assessment, California EPA. Available online
42 at http://www.oehha.org/water/phg/pdf/PCEAug2001.pdf.
43
44 Caldwell, DJ. (1999) Review of mononuclear cell leukemia in F-344 rat bioassays and its significance to human
45 cancer risk: A case study using alkyl phthalates. Regul Toxicol Pharmacol 30:45-53.
46
47 Caldwell, JC; Keshava, N; Evans, MV. (2008). Difficulty of mode of action determination for trichloroethylene: An
48 example of complex interactions of metabolites and other chemical exposures. Environ Mol Mutagen 49(2): 142-154.
49
50 Campagna, D; Mergler, D; Huel, G; et al. (1995) Visual dysfunction among styrene-exposed workers. Scand J Work
51 Environ Health 21:382-390.
52
53 Campagna, D; Gobba, F; Mergler, D; et al. (1996) Color vision loss among styrene-exposed workers
54 neurotoxicological threshold assessment. Neurotoxicology 17:367-373.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-243 DRAFT-DO NOT CITE OR QUOTE
-------
1 Cano, MI; Pollan, M. (2001) Non-Hodgkin=s lymphomas and occupation in Sweden. Int Arch Occup Environ
2 Health 74:443-449.
O
4 Capone, JP. (1994) Mechanisms of peroxisome proliferator-induced carcinogenesis: historical perspectives and
5 current status. Prepared by Eastern Research Group, Inc., Lexington, MA EPA Contract No. 69-D9-0133.
6 Available from: National Center for Environmental Assessment, Office of Research and Development, U.S. EPA,
7 Washington, DC.
8
9 Carney, EW; Thorsrud, BA; Dugard, PH; et al. (2006) Developmental toxicity studies in Crl:CD (SD) rats following
10 inhalation exposure to trichloroethylene and perchloroethylene. Birth Defects Research (Part B) 77:405-412
11
12 Carter, JH; Carter, HW; Deddens, JA; et al. (2003) A 2-year dose-response study of lesion sequences during
13 hepatocellular carcinogenesis in the male B6C3F1 mouse given the drinking water chemical dichloroacetic acid.
14 Environ Health Perspect 111 (1): 5 3 -64.
15
16 Cattley, RC; Glover, SE. (1993) Elevated 8-hydroxydeoxyguanosine in hepatic DNA of rats following exposure to
17 peroxisome proliferators: relationship to carcinogenesis and nuclear localization. Carcinogenesis 14:2495-2499.
18
19 Cattley, RC; Popp, JA. (1989) Differences between the promoting activities of the peroxisome proliferator WY-
20 14,643 and phenobarbital in rat liver. Cancer Res 49:3246-3251.
21
22 Cattley, RC; Roberts, RA. (2000) Peroxisome proliferators and carcinogenesis: editorial perspectives. Mutat Res
23 448:117-119.
24
25 Cattley, RC; Marsman, DS; Popp, JA. (1991) Age-related susceptibility to the carcinogenic effect of the peroxisome
26 proliferator WY-14,643 in rat liver. Carcinogenesis 12:469-473.
27
28 Cattley, RC; DeLuca, J; Elcombe, C; et al. (1998) Do peroxisome proliferating compounds pose a
29 hepatocarcinogenic hazard to humans? Regul Toxicol Pharmacol 27:47-60.
30
31 Cavalleri, A; Gobba, F; Paltrinieri, M; et al. (1994) Perchloroethylene exposure can induce colour vision loss.
32 Neurosci Lett 179:162-166.
33
3 4 Chang, LW; Daniel, FB; DeAngelo, AB. (1992) Analysis of DNA strand breaks induced in rodent liver in vivo,
3 5 hepatocytes in primary culture, and a human cell line by chlorinated acetic acids and chlorinated acetaldehydes.
36 Environ MolMutagen 20:277-288.
37
38 Channel, SR; Latendresse, JR; Kidney, JK; et al. (1998) A subchronic exposure to trichloroethylene causes lipid
39 peroxidation and hepatocellular proliferation in male B6C3F1 mouse liver. Toxicol Sci 43:145-154.
40
41 Charbonneau, M; Short, BB; Lock, EA; et al. (1987) Mechanism of petroleum-induced sex-specific protein droplet
42 nephropathy and renal cell proliferation in Fischer 344 rats: Relevance to humans. In: Hemphill, DD, ed. Trace
43 substances in environmental health Vol. 21. Columbia, MO: University of Missouri; pp. 263-273.
44
45 Chawla, A; Schwarz, EJ; Dimaculangan, DD; et al. (1994) Peroxisome proliferator-activated receptor (PPAR)
46 gamma: adipose- predominant expression and induction early in adipocyte differentiation. Endocrinology 135:798-
47 800.
48
49 Chen, HH; Chan, MH; Fu, SH. (2002a) Behavioural effects of tetrachloroethylene exposure in rats: acute and
50 subchronic studies. Toxicology 170:201-209.
51
52 Chen, SJ; Wang, JL; Chen, JH; et al. (2002b) Possible involvement of glutathione and p53 in trichloroethylene- and
53 perchloroethylene-induced lipid peroxidation and apoptosis in human lung cancer cells. Free Radic Biol Med
54 33:464-472.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-244 DRAFT-DO NOT CITE OR QUOTE
-------
1 Chevalier, S; Roberts, RA. (1998) Perturbation of rodent hepatocyte growth control by nongenotoxic
2 hepatocarcinogens: mechanisms and lack of relevance for human health (review). Oncol Rep 5:1319-1327.
3
4 Chia, SE; Ong, CN; Tsakok, MF; et al. (1996) Semen parameters in workers exposed to trichloroethylene. Reprod
5 Toxicol 10:295-299.
6
7 Chia, SE; Goh, VH; Ong, CN. (1997) Endocrine profiles of male workers with exposure to trichloroethylene. Am J
8 IndMed 32:217-222.
9
10 Chien, Y-C. (1997) The influences of exposure pattern and duration on elimination kinetics and exposure
11 assessment of tetrachloroethylene in humans. PhD thesis, Rutgers University, New Brunswick, NJ.
12
13 Chiu, WA; Bois, FY. (2006) Revisiting the population toxicokinetics of tetrachloroethylene 30. Arch Toxicol
14 80:382-385.
15
16 Chow, WH; McLaughlin, JK; Malker, HS; et al. (1995) Esophageal cancer and occupation in a cohort of Swedish
17 men. Am J IndMed 27:749-757.
18
19 Chu, S; Huang, Q; Alvares, K; et al. (1995) Transformation of mammalian cells by overexpressing H2O2-generating
20 peroxisomal fatty acyl-CoA oxidase. Proc Natl Acad Sci U S A 92:7080-7084.
21
22 Clevel, J; Manderau, L; Conso; F; et al. (1998) Occupational exposure to solvents and hairy cell leukaemia. Occup
23 EnvironMed 55:59-64.
24
25 Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
26 pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79(2):381-393.
27
28 Clewell, HJ; Gentry, PR; Kester, JE; et al. (2005) Evaluation of physiologically based pharmacokinetic models in
29 risk assessment: an example with perchloroethylene. Crit Rev Toxicol (accepted for publication).
30
31 Cimini, A; Cristiano, L; Bernardo, A; et al. (2000) Presence and inducibility of peroxisomes in a human
32 glioblastoma cell line. Biochim Biophys Acta 1474:397-409.
33
34 Cohn, P; Klotz, J; Bove, F; et al. (1994) Drinking water contamination and the incidence of leukemia and non-
35 Hodgkin's Lymphoma. Environ Health Perspect 102:556-561.
36
37 Cojocel, C; Beuter, W; Muller, W; et al. (1989) Lipid peroxidation: a possible mechanism of trichloroethylene-
3 8 induced nephrotoxicity. Toxicology 5 5:131 -141.
39
40 Coler HR; Rossmiller, HR. (1953) Tetrachloroethylene exposure in a small industry. AMA Arch Ind Hyg Occup
41 Med 8:227-233.
42
43 Cooper, GS; Parks, CG; Treadwell, EL; et al. (2004) Occupational risk factors for the development of systemic
44 lupus erythematosus. J Rheumatol 31:1928-1933.
45
46 Gorton, JC; Fan, LQ; Brown, S; et al. (1998) Down-regulation of cytochrome P450 2C family members and positive
47 acute-phase response gene expression by peroxisome proliferator chemicals. Mol Pharmacol 54:463-473.
48
49 Gorton, JC; Anderson, SP; Stauber, A. (2000) Central role of peroxisome proliferator-activated receptors in the
50 actions of peroxisome proliferaters. Annu Rev Pharmacol Toxicol 40:491-518.
51
52 Costa, AK; Ivanetich, KM. (1980). Tetrachloroethylene metabolism by the hepatic microsomal cytochrome P-450
53 system. Biochem Pharmacol 29(20):2863-2869.
54
55 Costa, C; Barbara, M; Catania, et al. (2004) Cytotoxicity evaluation after coexposure to perchloroethylene and
56 selected peroxidant drugs in rat hepatocytes. Toxicol In Vitro 18:37-44.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-245 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 Costantini, AS; Miligi, L; Kriebel, D; et al. (2001) A multicenter case-control study in Italy on
3 hematolymphopoietic neoplasms and occupation. Epidemiology 12:78-87.
4
5 Costas K; Knorr RS; Condon SK. (2002) A case-control study of childhood leukemia in Woburn, Massachusetts: the
6 relationship between leukemia incidence and exposure to public drinking water. Sci Tot Environ 300:23-35.
7
8 Cruz, SL; Mirshahi, T; Thomas, B; et al. (1998) Effects of the abused solvent toluene on recombinant N-methyl-D-
9 aspartate and non-N-methyl-D-aspartate receptors expressed inXenopus oocytes. J Pharmacol Exp Ther 286:334-
10 340.
11
12 Czirjak, L; Pocs, E; Szegedi, G. (1994) Localized scleroderma after exposure to organic solvents. Dermatology
13 189:399-401.
14
15 Czirjak, L; Kumanovics, G. (2002) Exposure to solvents in female patients with scleroderma. Clin Rheumatol
16 21:114-118.
17
18 Dallas, CE; Chen, XM; O=Barr, K; et al. (1994) Development of a physiologically based pharmacokinetic model for
19 perchloroethylene using tissue concentration-time data. Toxicol Appl Pharmacol 128-159.
20
21 Daniel, J. (1963) The metabolism of 36-Cl-labeled trichloroethylene and tetrachloroethylene in the rat. Biochem
22 Pharmacol 12:795-802.
23
24 Daniel, FB; DeAngelo, AB; Stober, JA; et al. (1992) Hepatocarcinogenicity of chloral hydrate, 2-
25 chloroacetaldehyde, and dichloroacetic acid in the male -6C3F1 mouse. Fundam Appl Toxicol 19:159-168.
26
27 Daniel, FB; Meier, JR; DeAngelo, AB. (1993) Advances in research on carcinogenic and genotoxic by-products of
28 chlorine disinfection: chlorinated hydroxyfuranones and chlorinated acetic acids. Ann 1st Super Sanita 29:279-291.
29
30 Daniell, WE; Claypoole, KH; Checkoway, H; et al. (1999) Neuropsychological function in retired workers with
31 previous long-term occupational exposure to solvents. Occup Environ Med 56:93-105.
32
33 Davidson, IW; Beliles, RP. (1991) Consideration of the target organ toxicity of trichloroethylene in terms of
34 metabolite toxicity and pharmacokinetics. Drug Metab Rev 23:493-599.
35
36 DeAngelo, AB. (2000) Response to AEpigenetic mechanisms of chemical carcinogenesis@ by James E. Klaunig,
37 Lisa M. Kamendulis, and Xu Yong. Hum Exp Toxicol 19:561-562.
38
39 DeAngelo, AB; Daniel, FB; McMillan, L; et al. (1989) Species and strain sensitivity to the induction of peroxisome
40 proliferation by chloroacetic acids. Toxicol Appl Pharmacol 101:285-298.
41
42 DeAngelo, AB; Daniel, FB; Stober, JA; et al. (1991) The carcinogenicity of dichloroacetic acid in the male B6C3F1
43 mouse. Fund Appl Toxicol 16:337-347.
44
45 DeAngelo, AB; Daniel, FB; Most, BM; et al. (1996) The carcinogenicity of dichloroacetic acid in the male Fischer
46 344 rat. Toxicology 114:207-221.
47
48 DeAngelo, AB; Daniel, FB; Most, BM; et al. (1997) Failure of monochloroacetic acid and trichloroacetic acid
49 administered in the drinking water to produce liver cancer in male F344/N rats. J Toxicol Environ Health 52:425-
50 445.
51
52 DeAngelo, AB; George, MH; House, DE. (1999) Hepatocarcinogenicity in the male B6C3F1 mouse following a
53 lifetime exposure to dichloroacetic acid in the drinking water: dose-response determination and modes of action. J
54 Toxicol Environ Health 58:485-507.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-246 DRAFT-DO NOT CITE OR QUOTE
-------
1 DeAngelo, AB; Daniel, FB; Wong, DM; et al. (2008). The induction of hepatocellular neoplasiaby trichloroacetic
2 acid administered in the drinking water of the male -6C3F1 mouse. J Tox and Env Health 71:1053-1065.
3
4 De Ceaurriz, J; Desiles, JP; Bonnet, P; et al. (1983) Concentration-dependent behavioral changes in mice following
5 short-term inhalation exposure to various industrial solvents. Toxicol Appl Pharmacol 67:383-389.
6
7 Degrassi, F and Tanzarella C. (1988). Immunofluorescent staining of kinetochores in micronuclei: A new assay for
8 the detection of aneuploidy. Mutat. Res. 203: 339-345.
9
10 Dekant, W. (2001) Chemical-induced nephrotoxicity mediated by glutathione S-conjugate formation. Toxicol Lett
11 124:21-36.
12
13 Dekant, W; Vamvakas, S; Berthold, K; et al. (1986) Bacterial beta-lyase mediated cleavage and mutagenicity of
14 cysteine conjugates derived from the nephrocarcinogenic alkenes trichloroethylene, tetrachloroethylene and
15 hexachlorobutadiene. Chem Biol Interact 60:31-45.
16
17 Delahunt, B; Bethwaite, PB; Nacey, JN. (1995) Occupational risk for renal cell carcinoma. A case-control study
18 based on the New Zealand Cancer Registry. Br J Urol 75:578-582.
19
20 de la Iglesia, FA; Gough, AW; Sigler, RE; et al. (1997) Alpha 2u-globulin nephropathy and ravens: do ravens of a
21 different feather flock together? Environ Health Perspect 105:903-904.
22
23 Delfino, RJ; Gong, H, Jr.; Linn, WS; et al. (2003a) Asthma symptoms in Hispanic children and daily ambient
24 exposures to toxic and criteria air pollutants. Env Health Perspect 111(4):647-656.
25
26 Delfino, RJ; Gong, H; Linn, WS; et al. (2003b) Respiratory symptoms and peak expiratory flow in children with
27 asthma in relation to volatile organic compounds in exhaled breath and ambient air. Journal of Exposure Analysis
28 and Environ Epidemiol 13:348-363.
29
30 Deonandan, R; Campbell, K; Ostbye, T; et al. (2000) A comparison of methods for measuring socio-economic
31 status by occupation or postal area. Chronic Dis Can21:114-118.
32
33 Desvergne, B; Wahli, W. (1995) PPAR: A key factor in nutrient/gene interactions. In: Baeujerle, P., ed. Inducible
34 transcription. Boston: Birkhauser; pp. 142-176.
35
36 Dietrich, DR. (1997) Doubting nongenotoxic mechanisms of renal cancer: comparing apples and oranges in the
3 7 alpha2u-globulin hypothesis. Environ Health Perspect 105:898-902.
38
39 Diodovich, C; Ferrario, D; Casati, B; Malerba, I; Marafante, E; Parent-Massin, D; Gribaldo, L (2005) Sensitivity of
40 human cord blood cells to tetrachloroethylene: cellular and molecular endpoints. Arch Toxicol 79:508-514.
41
42 Dobrev, ID; Andersen, ME; Yang, RS. (2001) Assessing interaction thresholds for trichloroethylene in combination
43 with tetrachloroethylene and 1,1,1-trichloroethane using gas uptake studies and PBPK modeling. Arch Toxicol
44 75(3):134-144.
45
46 Dobrev, ID; Andersen, ME; Tang RS. (2002) In silica toxicology: simulating interaction thresholds for human
47 exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-trichloroethane. Env Health Perspect
48 110(10):1031-1039.
49
50 Doherty, AT; Ellard, S; Parry, EM; et al. (1996) An investigation into the activation and deactivation of chlorinated
51 hydrocarbons to genotoxins in metabolically competent human cells. Mutagenesis 11:247-274.
52
53 Donoghue, AM; Dryson, EW; Wynn-Williams, G. (1995) Contrast sensitivity in organic-solvent-induced chronic
54 toxic encephalopathy. J Occup Environ Med 37:1357-1363.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-247 DRAFT-DO NOT CITE OR QUOTE
-------
1 Dosemeci, M; Cocco, P; Chow, WH. (1999) Gender differences in risk of renal cell carcinoma and occupational
2 exposures to chlorinated aliphatic hydrocarbons. Am JIndMed 36:54-59.
3
4 Doyle, P; Roman, E; Beral, V; et al. (1997) Spontaneous abortion in dry cleaning workers potentially exposed to
5 perchloroethylene. Occup EnvironMed 54:848-853.
6
7 Dreessen, B.; Westphal, G.; Hunger, I; et al. (2003). "Mutagenicity of the glutathione and cysteine S-conjugates of
8 the haloalkenes l,l,2-trichloro-3,3,3-trifluoro-l-propene and trichlorofluoroethene in the Ames test in comparison
9 with the tetrachloroethene-analogues." Mutat Res 539(1-2): 157-66.
10
11 Dreyer, C; Krey, G; Keller, H; et al. (1992) Control of the peroxisomal beta-oxidation pathway by a novel family of
12 nuclear hormone receptors. Cell 68:879-887.
13
14 Dybing, E; Mikalsen, S-O; Huttunen, J; et al. (1995) Peroxisome proliferation, genotoxicity and carcinogenicity.
15 IARC Technical Report 24:55-85.
16
17 Eacho, PI; Larder, TL; Brodhecker, CA. (1991) Hepatocellular DNA synthesis in rats given peroxisome
18 proliferating agents: comparison of WY-14,643 to clofibric acid, nafenopinandLY171883. Carcinogenesis
19 12:1557-1561.
20
21 Ebrahim, AS; Babakrishnan, K; Sakthisekaran, D. (1996) Perchloroethylene-induced alterations in glucose
22 metabolism and their prevention by 2-deoxy-D-glucose and vitamin E in mice. J Appl Toxicol 16:339-348.
23
24 Ebrahim, AS; Babu, E; Thirunavukkarasu, C; et al. (2001) Protective role of vitamin E, 2-deoxy-D-glucose, and
25 marine on perchloroethylene induced alterations in ATPases. Drug Chem Toxicol 24(4):429-437.
26
27 Echeverria, D; Heyer, N; Checkoway, H; et al. (1994) Behavioral investigation of occupational exposure to solvents:
28 Perchloroethylene among dry cleaners, and styrene among reinforced fiberglass laminators. Final Report. Report
29 prepared for the Centers for Disease Control and Prevention under Grant No. 5 R01 OHo2719-03. Battelle Centers
3 0 for Public Health Research and Evaluation.
31
32 Echeverria, D; White, RF; Sampaio, C. (1995) A behavioral evaluation of PCE exposure in patients and dry
33 cleaners: a possible relationship between clinical and preclinical effects. J Occup Environ Med 37:667-680.
34
35 Elcombe, CR. (1985) Species differences in carcinogenicity and peroxisome proliferation due to trichloroethylene: a
36 biochemical human hazard assessment. Arch Toxicol Suppl 8:6-17.
37
38 Elcombe, CR; Mitchell, AM. (1986) Peroxisome proliferation due to di(2-ethylhexyl) phthalate (DEHP): species
39 differences and possible mechanisms. Environ Health Perspect 70:211-219.
40
41 Elcombe, CR; Rose, MS; Pratt, IS. (1985) Biochemical, histological, and ultrastructural changes in rat and mouse
42 liver following the administration of trichloroethylene: possible relevance to species differences in
43 hepatocarcinogenicity. Toxicol ApplPharmacol79:365-376.
44
45 Elfarra, AA.; Krause RJ. (2007) S-(l,2,2-trichlorovinyl)-L-cysteine sulfoxide, a reactive metabolite of S-(l,2,2-
46 Trichlorovinyl)-L-cysteine formed in rat liver and kidney microsomes, is a potent nephrotoxicant. J Pharmacol Exp
47 Ther 321(3): 1095-101.
48
49 Epstein, DL; Nolen, GA; Randall, JL; et al. (1992) Cardiopathic effects of dichloroacetate in the fetal Long-Evans
50 rat. Teratology 46:225-235.
51
52 Ericksson, P; von Rosen, D; Viberg, H; et al. (2005) Developmental toxicology in the neonatal mouse: the use of
53 randomly selected individuals as statistical unit compared to the litter in mice neonatally exposed to PBDE 99.
54 Toxicologist84(S-l):219-220.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-248 DRAFT-DO NOT CITE OR QUOTE
-------
1 Eskenazi, B; Fenster, L; Hudes, M; et al. (1991a) A study of the effect of perchloroethylene exposure on the
2 reproductive outcomes of wives of dry-cleaning workers. Am J Ind Med 20:593-600.
3
4 Eskenazi, B; Wyrobek, AJ; Fenster, L; et al. (1991b) A study of the effect of perchloroethylene exposure on semen
5 quality in dry cleaning workers. Am J Ind Med 20:575-591.
6
7 Ethell, DW; Buhler, LA. (2003) Fas ligand-mediated apoptosis in degenerative disorders of the brain. J Clin
8 Immunol 23:439-446.
9
10 Fabbro-Peray, P; Daures, JP; Rossi, JF. (2001) Environmental risk factors for non-Hodgkin=s lymphoma: a
11 population-based case-control study in Languedoc-Roussillon, France. Cancer Causes Control 12:201-212.
12
13 Fagliano, J; Berry, M; Bove, F; et al. (1990) Drinking water contamination and the incidence of leukemia: an
14 ecologic study. Am J Public Health 80:1209-1212.
15
16 Fan, LQ; Coley, J; Miller, RT; et al. (2003) Opposing mechanisms of NADPH-cytochrome P450 oxidoreductase
17 regulation by peroxisome proliferators. Biochem Pharmacol. 65:949-959.
18
19 Ferreira-Gonzalez, A; DeAngelo, AB; Nasim, S; et al. (1995) Ras oncogene activation during hepatocarcinogenesis
20 inB6C3Fl male mice by dichloroacetic and trichloroacetic acids. Carcinogenesis 16:495-500.
21
22 Ferroni, C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
23 perchloroethylene. Neurotoxicology 13:243-247.
24
25 Finckh, A; Cooper, GS; Chibnik; et al. (2007) Occupational exposures and risk of systemic lupus erythematosus.
26 Arthritis Rheum 54:3648-3654.
27
28 Flavell, DM; Pineda, TI; Jamshidi, Y; et al. (2000) Variation in the PPAR-a gene is associated with altered function
29 in vitro and plasma lipid concentrations in Type II diabetic subjects. Diabetologia 43:673-680.
30
31 Flower, DR; North, AC; Attwood, TK. (1993). Structure and sequence relationships in the lipcalins and related
32 proteins. Protein Sci 2:753-761.
33
34 Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
3 5 solvents. Int Arch Occup Environ Health 52:1-9.
36
37 Fredriksson, M; Bengtsson, NO; Hardell, L; et al. (1989) Colon cancer, physical activity, and occupational
38 exposures. A case-control study. Cancer 63:1838-1842.
39
40 Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
41 tetrachloroethylene as neonates. Toxicol Lett 66:13-19.
42
43 Frenette, B; Mergler, D; Bowler, R. (1991) Contrast-sensitivity loss in a group of former microelectronics workers
44 with normal visual acuity. Optom Vis Sci 68:556-560.
45
46 Gaillard, Y; Billault, F; Pepin, G. (1995) Tetrachloroethylene fatality: case report and simple gas chromatographic
47 determination in blood and tissues. Forensic Sci Int 76:161-168.
48
49 Gaipl, US; Kuhn, A; Sheriff, A; et al. (2006) Clearance of apoptotic cells in human SLE. Curr Dir Autoimmun
50 9:173-187.
51
52 Galloway, SM; Armstrong, MJ; Reuben, C; et al. (1987) Chromosome aberrations and sister chromatid exchanges in
53 Chinese hamster ovary cells: evaluations of 108 chemicals. Environ Mol Mutagen 10 Suppl 10:1-175.
54
55 Gama, C; Meira, JB. (1978) Occupational acro-osteolysis. J Bone Joint Surg Am 60:86-90.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-249 DRAFT-DO NOT CITE OR QUOTE
-------
1 Canning, AE; Brunk, U; Dallner, G. (1983) Effects of dietary di(2-ethylhexyl)phthalate on the structure and function
2 of rat hepatocytes. Biochim Biophys Acta 763:72-82.
3
4 Garabrant, DH; Lacey, J; Laing, TJ; et al. (2003) Scleroderma and solvent exposure among women. Am J Epidemiol
5 157/6:493-500.
6
7 Garetano, G; Gochfeld, M. (2000) Factors influencing tetrachloroethylene concentrations in residences above dry-
8 cleaning establishments. Arch Environ Health 55:59-68.
9
10 Gamier, R; Bedouin, J; Pepin, et al. (1996) Coin-operated dry cleaning machines may be responsible for acute
11 tetrachloroethylene poisoning: report of 26 cases including one death. JToxicol ClinToxicol 34:191-197.
12
13 Geller, AM; Hudnell, HK. (1997) Critical issues in the use and analysis of the Lanthony Desaturate Color Vision
14 test. Neurotoxicol Teratol 19:455-465.
15
16 Geller, AM (2001) A table of color distance scores for quantitative scoring of the Lanthony
17 Desaturate color vision test. Neurotoxicol Teratol 23:265-267.
18
19 Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
20 tetrachloroethylene. Am JIndMed 21:661-671.
21
22 Gentry, PR; Covington, TR; Clewell, HJ. (2003) Evaluation of the potential impact of pharmacokinetic differences
23 on tissue dosimetry in offspring during pregnancy and lactation. Regul Toxicol Pharmacol 38:1-16.
24
25 George, MH; Moore, T; Kilburn, S; et al. (2000) Carcinogenicity of chloral hydrate administered in drinking water
26 to the male F344/N rat and male B6C3F(1) mouse. Toxicol Pathol 28:610-618.
27
28 Germolec, DR; Yang, RS; Ackermann, MF; et al. (1989) Toxicology studies of a chemical mixture of 25
29 groundwater contaminants. II. ImmunosuppressioninB6C3Fl mice. Fundam Appl Toxicol 13:377-387.
30
31 Ghantous, H; Danielsson, BR; Dencker, L; et al. (1986) Trichloroacetic acid accumulates in murine amniotic fluid
32 after tri- and tetrachloroethylene inhalation. Acta Pharmacol Toxicol (Copenh) 58:105-114.
33
34 Gilbert, KM; Whitlow, AB; Pumford, NR. (2004) Environmental contaminant and disinfection by-product
3 5 trichloroacetaldehyde stimulates T cells in vitro. Int Immunopharmacol 4:25-36.
36
37 Oilman, AG; Rail, TW; Nies, AS; et al. (1990) The pharmacological basis of therapeutics. New York: McGraw-
38 Hill; pp. 244-573.
39
40 Glinghammar, B; Skogsberg, J; Hamsten, A; et al. (2003) PPARdelta activation induces COX-2 gene expression and
41 cell proliferation in human hepatocellular carcinoma cells. Biochem Biophys Res Commun 308(2):361-368.
42
43 Gobba, F; Righi, E; Fantuzzi, G; et al. (1998) Two-year evolution of perchloroethylene-induced color-vision loss.
44 Arch Environ Health 53:196-198.
45
46 Goh, VH; Chia, SE; Ong, CN. (1998) Effects of chronic exposure to low doses of trichloroethylene on steroid
47 hormone and insulin levels in normal men. Environ Health Perspect 106:41-44.
48
49 Goldman, JA. (1996) Connective tissue disease in people exposed to organic chemical solvents: systemic sclerosis
50 (scleroderma) in dry cleaning plant and aircraft industry workers. JCR: J Clin Rheumatol 2:185-190.
51
52 Goldsworthy, TL; Popp, JA. (1987) Chlorinated hydrocarbon-induced peroxisomal enzyme activity in relation to
53 species and organ carcinogenicity. Toxicol Appl Pharmacol 88:225-233.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-250 DRAFT-DO NOT CITE OR QUOTE
-------
1 Goldsworthy, TL; Lyght, O; Burnett, VL; et al. (1988) Potential role of alpha-2 mu-globulin, protein droplet
2 accumulation, and cell replication in the renal carcinogenicity of rats exposed to trichloroethylene,
3 perchloroethylene, and pentachloroethane. Toxicol Appl Pharmacol 96:367-379.
4
5 Goodman, DG; Ward, JM; Squire, RA; et al. (1979) Neoplastic and non-neoplastic lesions in aging F344 rats.
6 Toxicol Appl Pharmacol 48:237-248.
7
8 Graham, MJ; Wilson, SA; Winham, MA; et al. (1994) Lack of peroxisome proliferation in marmoset liver following
9 treatment with ciprofibrate for 3 years. Fundam Appl Toxicol 22:58-64.
10
11 Grasl-Kraupp, B; Huber, W; Schulte-Hermann, R. (1993) Are peroxisome proliferators tumour promoters in rat
12 liver? In: Gibson, G and Lake, B , eds. Perixosomes: biology and importance in toxicology and medicine. London:
13 Taylor and Francis; pp. 667-693.
14
15 Green, T; Odum, J. Nash, JA; et al. (1990) Perchloroethylene-induced rate kidney tumors: an investigation of the
16 mechanisms involved and their relevance to human. Toxicol Appl Pharmacol 103:77-89.
17
18 Green, SM; Khan, MF; Kaphalia, BS; et al. (2001) Immunohistochemical localization of trichloroacylated protein
19 adducts in tetrachloroethene-treated mice. J Toxicol Environ Health A 63:145-157.
20
21 Griffin, J M; Blossom, S J; Jackson, S K; Gilbert, K M; Pumford, N R (2000a) Trichloroethylene accelerates an
22 autoimmune response by Thl T cell activation in MRL +/+ mice. Immunopharmacology 46:123-137.
23
24 Griffin, JM; Gilbert, KM; Pumford, NR. (2000b) Inhibition of CYP2E1 reverses CD4+ T-cell alterations in
25 trichloroethylene-treated MRL+/+ mice. Toxicol Sci 54:384-389.
26
27 Griffin, JM; Gilbert, KM; Lamps, LW; et al. (2000c) CD4(+) T-cell activation and induction of autoimmune
28 hepatitis following trichloroethylene treatment in MRL+/+ mice. Toxicol Sci 57:345-352.
29
30 Grobin, AC; Matthews, DB; Devaud, LL; et al. (1998) The role of GABA(A) receptors in the acute and chronic
31 effects of ethanol. Psychopharmacology (Berl) 139:2-19.
32
33 Gualandi, G. (1987) Use of alpha- and beta-tubulin mutants for the study of spontaneous and induced chromosomal
34 mis-distribution inAspergillus nidulans. Mutat. Res. 178: 33-41.
35
36 Hake, CL; Stewart, RD. (1977) Human exposure to tetrachloroethylene: inhalation and skin contact. Environ Health
37 Perspect 21:231-238.
38
39 Hakkola, J; Pasanen, M; Hukkanen, J; et al. (1996a) Expression of xenobiotic-metabolizing cytochrome P450 forms
40 in human full-term placenta. Biochem Pharmacol 51:403 -411.
41
42 Hakkola, J; Raunio, H; Purkunen, R; et al. (1996b) Detection of cytochrome P450 gene expression in human
43 placenta in first trimester of pregnancy. Biochem Pharmacol 52:379-83.
44
45 Hakkola, J; Pelkonen, O; Pasanen, M; et al. (1998) Xenobiotic-metabolizing cytochrome P450 enzymes in the
46 human feto- placenta! unit: role in intrauterine toxicity. Crit Rev Toxicol 28:35-72.
47
48 Haider, CA; Warne, TM; Hatoum, NS. (1984) Renal toxicity of gasoline and related petroleum naphthas in male rats.
49 In: Mehlman, MA; Hemstreet, GP; Thorpe, JJ; et al., eds. Advances in modern environmental toxicology, vol. VII.
50 Renal effects of petroleum hydrocarbons. Princeton, NJ: Princeton Scientific.
51
52 Hanefeld, M; Kemmer, C; Kadner, E. (1983) Relationship between morphological changes and lipid-lowering action
53 of p-chlorphenoxyisobutyric acid (CPIB) on hepatic mitochondria and peroxisomes in man. Atherosclerosis 46:239-
54 246.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-251 DRAFT-DO NOT CITE OR QUOTE
-------
1 Hanioka, N; Jinno, H; Toyo'oka, T; et al. (1995) Induction of rat liver drug-metabolizing enzymes by
2 tetrachloroethylene. Arch Environ Contam Toxicol 28:273-280.
3
4 Hansen, J; Raaschou-Nielsen, O; Christensen, JM; et al. (2001) Cancer incidence among Danish workers exposed to
5 trichloroethylene. JOccup Environ Med 43:133-139.
6
7 Harap, KR. (1975) Deviant metabolic patterns in malignant disease. In: Ambrose, EJ; Roe, FJC; eds. Biology of
8 cancer, 2nd ed. London: Cox and Wyman.
9
10 Hardell, L; Eriksson, M; Lenner, P; et al. (1981) Malignant lymphoma and exposure to chemicals, especially organic
11 solvents, chlorophenols and phenoxy acids: a case-control study. Br J Cancer 43:169-176.
12
13 Hardell, L; Bengtsson, NO; Jonsson, U; et al. (1984) Aetiological aspects on primary liver cancer with special
14 regard to alcohol, organic solvents and acute intermittent porphyriaBan epidemiological investigation. Br J Cancer
15 50:3896397.
16
17 Hardin, BD; Bond, GP; Sikov, MR; et al. (1981) Testing of selected workplace chemicals for teratogenic potential.
18 Scand J Work Environ Health 7 Suppl 4:66-75.
19
20 Hardwick, JP; Song, BJ; Huberman, E; et al. (1987) Isolation, complementary DNA sequence, and regulation of rat
21 hepatic lauric acid omega-hydroxylase (cytochrome P-450 LA omega): Identification of a new cytochrome P-450
22 gene family. J Biol Chem 262:801-810.
23
24 Harrington, JM; Whitby, H; Gray, CN; et al. (1989) Renal disease and occupational exposure to organic solvents: a
25 case referent approach. Br J Ind Med 46:643-650.
26
27 Harrington-Brock, K; Doerr, CL; Moore, MM. (1998) Mutagenicity of three disinfection by-products: di- and
28 trichloroacetic acid and chloral hydrate in L5178Y/TK +/- (-)3.7.2C mouse lymphoma cells. Mutat Res 413:265-
29 276.
30
31 Harris, RA; Mihic, SJ; Dildy-Mayfield, JE; et al. (1995) Actions of anesthetics on ligand-gated ion channels: role of
32 receptor subunit composition. FASEB J 9:1454-1462.
33
34 Harris, NL; Jaffe, ES; Diebold, J; et al. (2000a) The World Health Organization classification of hematological
35 malignancies report of the Clinical Advisory Committee Meeting, Airlie House, VA, November 1997. Mod Pathol
36 13:193-207.
37
38 Harris, NL; Jaffe, ES; Diebold, J; et al. (2000b) Lymphoma classification-from controversy to consensus: the
39 R.E.A.L. and WHO Classification of lymphoid neoplasms. AnnOncol 11 Suppl 1:3-10.
40
41 Hartmann, A; Speit, G. (1995) Genotoxic effects of chemicals in the single cell gel (SCO) test with human blood
42 cells in relation to the induction of sister-chromatid exchanges (SCE). Mutat Res 346:49-56.
43
44 Hasmall, S; Orphanides, G; James, N; et al. (2002) Down regulation of lactoferrin by PPAR-a ligands: role in
45 perturbabion of hepatocyte proliferation and apoptosis. Toxicol Sci 68:304-313.
46
47 Haworth, S; Lawlor, T; Mortelmans, K; et al. (1983) Salmonella mutagenicity test results for 250 chemicals.
4 8 Environ Mutagen 5 Suppl 1:1-142.
49
50 Hayes, JR; Condie, LW, Jr.; Borzelleca, JF. (1986) The subchronic toxicity of tetrachloroethylene
51 (perchloroethylene) administered in the drinking water of rats. Fundam Appl Toxicol 7:119-125.
52
53 Heineman, EF; Cocco, P; Gomez, MR; et al. (1994) Occupational exposure to chlorinated aliphatic hydrocarbons
54 and risk of astrocytic brain cancer. Am J Ind Med 26:155-169.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-252 DRAFT-DO NOT CITE OR QUOTE
-------
1 Hemminki, K; Franssila, E; Vainio, H (1980) Spontaneous abortions among female chemical workers in Finland.
2 Int Arch Occup Environ Health 45:123-126.
3
4 Hernberg, S; Korkala, ML; Asikainen, U; et al. (1984) Primary liver cancer and exposure to solvents. Int Arch
5 Occup Environ Health 54:147-153.
6
7 Hernberg, S; Kauppinen, T; Riala, R; et al. (1988) Increased risk for primary liver cancer among women exposed to
8 solvents. ScandJ Work Environ Health 14:356-365.
9
10 Herren-Freund, SL; Pereira, MA; Khoury, MD; et al. (1987) The carcinogenicity of trichloroethylene and its
11 metabolites, trichloroacetic acid and dichloroacetic acid, in mouse liver. Toxicol Appl Pharmacol 90:183-189.
12
13 Herrinton, LJ. (1998) Epidemiology of the revised European-American lymphoma classification subtypes.
14 Epidemiol Rev 20:187-203.
15
16 Heuvel, JPV. (1999) Peroxisome proliferator-activated receptors (PPARS) and carcinogenesis. Toxicol Sci 47:1-8.
17
18 Hill, AB. (1965) The environment and disease: association or causation? Proc R Soc Med 58:295-300.
19
20 Hinnen, U; Schmid-Grendelmeier, P; Muller, E; et al. (1995) Exposure to solvents in scleroderma: disseminated
21 circumscribed scleroderma (morphea) in a painter exposed to perchloroethylene. Schweiz Med Wochenschr
22 125:2433-2437.
23
24 Hinson, JA; Roberts, DW. (1992) Role of covalent and noncovalent interactions in cell toxicity: effects on proteins.
2 5 Annu Rev Pharmacol Toxicol 3 2:471 -510.
26
27 Hinton, RH; Mitchell, DE; Mann, A; et al. (1986) Effect of phthalic acid esters on the liver and thyroid Environ
28 Health Perspect 70:195-210.
29
30 Holson, RR; Pearce, B. (1992) Principles and pitfalls in the analysis of prenatal treatment effects in multiparous
31 species. Neurotoxicol Teratol 14:221-228.
32
33 Honma, T; Hasegawa, H; Sato, M; et al. (1980a) Changes of free amino acid content in rat brain after exposure to
3 4 trichloroethylene and tetrachloroethylene. Ind Health 18:1-7.
35
36 Honma, T; Sudo, A; Miyagawa, M; et al. (1980b) Effects of exposure to trichloroethylene and tetrachloroethylene
37 on the contents of acetylcholine, dopamine, norepinephrine and serotonin in rat brain. IndHealth 18:171-178.
38
39 Houten, L, Sonnesso, G. (1980) Occupational exposure and cancer of the liver. Arch Environ Health 35:51-53.
40
41 HSIA (Halogenated Solvents Industry Alliance). (2004) Comment entitled AEvaluation of EPA'sneurotoxicity of
42 tetrachloroethylene.@ ORD-2003-0014-0005. Available online at
43 http://docket.epa.gov/edkpub/do/EDKStaffItemDetailView?objectId=090007d48025686c.
44
45 Hudnell, HK; Schreiber, JS. (2004) Residential tetrachloroethylene exposure: response. Environ Health Perspect
46 112(15):A864.
47
48 Hudnell, HK; Boyes, WK; Otto, DA; et al. (1996a) Battery of neurobehavioral tests recommended to ATSDR:
49 solvent-induced deficits in microelectronic workers. Toxicol Ind Health 12:235-243.
50
51 Hudnell, HK; Otto, DA; House, DE. (1996b) The influence of vision on computerized neurobehavioral test scores: a
52 proposal for improving test protocols. Neurotoxicol Teratol 18:391-400.
53
54 Hudnell, HK; House, D; Schmid, J; et al. (2001) Human visual function in the North Carolina clinical study on
55 possible estuary-associated syndrome. J Toxicol Environ Health A 62:575-594.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-253 DRAFT-DO NOT CITE OR QUOTE
-------
1 Huff, J. (1995) Mechanisms, chemical carcinogenesis, and risk assessment: cell proliferation and cancer. Am J Ind
2 Med 27:293-300.
3
4 Huff, J. (1996) Response: alpha-2-mu-Globulin nephropathy, posed mechanisms, and white ravens. Environ Health
5 Perspect 104:1264-1267.
6
7 Huss, JM; Kelly, DP. (2005) Mitochondria! energy metabolism in heart failure: a question of balance. J Clin Invest
8 115:547-555.
9
10 IARC (International Agency for Research on Cancer). (1995) Tetrachloroethylene. In: IARC monographs on the
11 evaluation of carcinogenic risks to humans; vol. 63: dry cleaning, some chlorinated solvents and other. Lyon,
12 France.
13
14 Ikeda, M; Imanura, T. (1973) Biological half-life of trichloroethylene and tetrachloroethylene in human subjects. Int
15 Arch Arbeitsmed. 31(3):209-224.
16
17 Ikeda, M; Otsuji, H. (1972) A comparative study of the excretion of Fujiwara reaction-positive substances in urine
18 of humans and rodents given trichloro- or tetrachloro-derivatives of ethane and ethylene. Br J Ind Med 29:99-104
19
20 ILSI (International Life Sciences Institute). (1992) Similarities and differences between children and adults:
21 implications for risk assessment. New York: ILSI Press.
22
23 Infante-Rivard, C; Siemiatycki, J; Lakhani, R; et al. (2005) Maternal exposure to occupational solvents and
24 childhood leukemia. Environ Health Perspect 113(6):787-92.
25
26 IOM (Institute of Medicine). (2002) Gulf war and health. Vol. 2: Insecticides and solvents. Washington, DC:
27 National Academies Press.
28
29 Ishmael, J and Dugard, P H (2006) A review of perchloroethylene and rat mononuclear cell leukemia. Regul
30 Toxicol Pharmacol 45:178-184.
31
32 Issemann, I; Green, S. (1990) Activation of a member of the steroid hormone receptor superfamily by peroxisome
33 proliferators. Nature 347:645-650.
34
35 Issemann, I; Prince, RA; Tugwood, JD; et al. (1993) The peroxisome proliferator-activated receptor:retinoid X
3 6 receptor heterodimer is activated by fatty acids and fibrate hypolipidaemic drugs. J Mol Endocrinol 11:37-47.
37
3 8 Ito, YO ; Yamanoshita, N ; Asaeda, Y; et al. (2007). Di(2-ethylhexyl)phthalate induces hepatic tumorigenesis
39 through a peroxisome proliferator-activated receptor -a-independent pathway. J Occup Health 49(3): 172-82.
40
41 Jalouli, M; Carlsson, L; Ameen, C; et al. (2003) Sex difference in hepatic peroxisome proliferator-activated receptor
42 -a expression: influence of pituitary and gonadal hormones. Endocrinology 144:101-109.
43
44 Jang, JY; Droz, PO. (1997) Ethnic differences in biological monitoring of several organic solvents. II. A simulation
45 study with a physiologically based pharmacokinetic model. Int Arch Occup Environ Health 70:41-50.
46
47 Ji, J; Hemminki, K. (2005a) Occupation and upper aerodigestive tract cancers: a follow-up study in Sweden. J
48 Occup Environ Med 47:785-795.
49
50 Ji, J; Hemminki, K. (2005b) Occurrences of leukemia subtypes by socioeconomic and occupational group. J Occup
51 Environ Med 47:1131-1140.
52
53 Ji, J; Hemminki, K. (2006) Socioeconomic/occupational risk factors for lymphproliferative diseases in Sweden.
54 Ann Epidemiol 16:370-376.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-254 DRAFT-DO NOT CITE OR QUOTE
-------
1 Jirtle, RL; Meyer, SA. (1991) Liver tumor promotion: effect of phenobarbital on EOF and protein kinase C signal
2 transduction and transforming growth factor-beta 1 expression. Dig Dis Sci 36:659-668.
3
4 JISA (Japan Industrial Safety Association). (1993) Carcinogenicity study of tetrachloroethylene by inhalation in rats
5 and mice. Data No. 3-1. Kanagawa, Japan. Available from: IRIS Information Desk, U.S. Environmental Protection
6 Agency, Washington, DC.
7
8 Johnson, PD; Dawson, BV; Goldberg, SJ. (1998) Cardiac teratogenicity of trichloroethylene metabolites. J Am Coll
9 Cardiol 32:540-545.
10
11 Jonker, D; Woutersen, RA; Feron, VJ. (1996) Toxicity of mixtures of nephrotoxicants with similar or dissimilar
12 mode of action. Food ChemToxicol 34:1075-1082.
13
14 Juntunen, J. (1986) Occupational toxicology of trichloroethylene with special reference to neurotoxicity. Dev
15 Toxicol Environ Sci 12:189-200.
16
17 Kaerlev, L; Teglbjaerg, PS; Sabroe, S; et al. (2000) Occupation and small bowel adenocarcinoma: a European case-
18 control study. Occup Environ Med 57:760-766.
19
20 Kafer, E. (1985) Tests which distinguish induced aneuploidy and crossing-over from secondary segregation in
21 Aspergillus treated with chloral hydrate or x-rays. Mutat. Res. 167: 9-34.
22
23 Karlsson, JE; Rosengren, LE; Kjellstrand, P; et al. (1987) Effects of low-dose inhalation of three chlorinated
24 aliphatic organic solvents on deoxyribonucleic acid in gerbil brain. Scand J Work Environ Health 13:453-458.
25
26 Kato, I; Koenig, KL; Watanabe-Meserve, H; et al. (2005) Personal and occupational exposure to organic solvents
27 and risk of non-Hodgkin's lymphoma (NHL) in women (United States). Cancer Causes and Control 16:1215-1224.
28
29 Kaufman AS; McLean JE; Reynolds CR. (1988) Sex, race, residence, region, and education differences on the 11
30 WAIS-R subtests. J ClinPsychol 44:231-248.
31
32 Kerbey, AL; Randle, PJ; Cooper, RH; et al. (1976) Regulation of pyruvate dehydrogenase in rat heart: mechanism of
3 3 regulation of proportions of dephosphorylated and phosphorylated enzyme by oxidation of fatty acids and ketone
34 bodies and of effects of diabetes: role of coenzyme A, acetyl-coenzyme A and reduced and oxidized nicotinamide-
35 adenine dinucleotide. BiochemJ 154:327-348.
36
37 Kernan, GJ; Ji, BT; Dosemeci, M; et al. (1999) Occupational risk factors for pancreatic cancer: a case-control study
3 8 based on death certificate from 24 U.S. states. Am J Ind Med. 36(2):260-70.
39
40 Kjellstrand, P; Holmquist, B; Kanje, M; et al. (1984) Perchloroethylene: effects on body and organ weights and
41 plasma butyrylcholinesterase activity in mice. Acta Pharmacol Toxicol (Copenh) 54:414-424.
42
43 Kjellstrand, P; Holmquist, B; Jonsson, I; et al. (1985) Effects of organic solvents on motor activity in mice.
44 Toxicology 35:35-46.
45
46 Klaunig, J; Ruch, R; DeAngelo, A; et al. (1988) Inhibition of mouse hepatocyte intercellular communication by
47 phthalate esters. Cancer Lett 43:65-71.
48
49 Klaunig, JE; Babich, MA; Baetcke, KP; et al. (2003) PPAR-a agonist-induced rodent tumors: modes of action and
50 human relevance. Crit Rev Toxicol 33(6):655-780.
51
52 Kobayashi, S; Hutcheon, DE; Regan, J. (1982) Cardiopulmonary toxicity of tetrachloroethylene. J Toxicol Environ
53 Health 10:23-30.
54
55 Kok, RJ; Haas, M; Moolenaar, F; et al. (1998) Drug delivery to the kidneys and the bladder with the low molecular
56 weight protein lysozyme. RenFail 20:211-217.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-255 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 Koppel, C; Arndt, I; Arendt, U; et al. (1985) Acute tetrachloroethylene poisoning-blood elimination kinetics during
3 hyperventilation therapy. JToxicol ClinToxicol23:103-115.
4
5 Kraupp-Grasl, B; Huber, W; Putz, B; et al. (1990) Tumor promotion by the peroxisome proliferator nafenopin
6 involving a specific subtype of altered foci in rat liver. Cancer Res 50:3701-3708.
7
8 Kraupp-Grasl, B; Huber, W; Taper, H; et al. (1991) Increased susceptibility of aged rats to hepatocarcinogenesis by
9 the peroxisome proliferator nafenopin and the possible involvement of altered liver foci occurring spontaneously.
10 Cancer Res 51:666-671.
11
12 Kriebel, D; Zeja, A; Eisen, EA; et al. (2004) Quantitative evaluation of the effects of uncontrolled confounding by
13 alcohol and tobacco in occupational cancer studies. Int J Epidemiol 33:1040-1045.
14
15 Kumar, P; Prasad, AK; Dutta, KK. (2000) Steroidogenic alterations in testes and sera of rats exposed to
16 trichloroethylene (TCE)by inhalation. HumExp Toxicol 19:117-121.
17
18 Kurata, Y; Kidachi, F; Yokoyama, M; et al. (1998) Subchronic toxicity of di(2-ethylhexyl)phthalate in common
19 marmosets: Lack of hepatic peroxisome proliferation, testicular atrophy or pancreatic acinar cell hyperplasia.
20 Toxicol Sci 42:49-56.
21
22 Kurien, B T; Hensley, K; Bachmann, M; Scofield, R H (2006) Oxidatively modified autoantigens in autoimmune
23 diseases. Free Radic Biol Med 41:549-556.
24
25 Kylin, B; Sumegi, I; Yllner, S. (1965) Hepatotoxicity of inhaled trichloroethylene and tetrachloroethylene: long-
26 term exposure. Acta Pharmacol Toxicol (Kbh) 22:379-385.
27
28 Kyrklund, T; Haglid, K. (1991) Brain lipid composition in guinea pigs after intrauterine exposure to
29 perchloroethylene. Pharmacol Toxicol 68:146-148.
30
31 Kyrklund, T; Ailing, C; Kjellstrand, P; et al. (1984) Chronic effects of perchloroethylene on the composition of lipid
32 and acyl groups in cerebral cortex and hippocampus of the gerbil. Toxicol Lett 22:343-349.
33
34 Kyrklund, T; Kjellstrand, P; Haglid, KG. (1987) Lipid composition and fatty acid pattern of the gerbil brain after
35 exposure to perchloroethylene. Arch Toxicol 60:397-400.
36
37 Kyrklund, T; Kjellstrand, P; Haglid, KG. (1988) Effects of exposure to Freon 11, 1,1,1-trichloroethane or
3 8 perchloroethylene on the lipid and fatty-acid composition of rat cerebral cortex. Scand J Work Environ Health
39 14:91-94.
40
41 Kyrklund, T; Kjellstrand, P; Haglid, KG. (1990) Long-term exposure of rats to perchloroethylene, with and without
42 a post-exposure solvent-free recovery period: effects on brain lipids. Toxicol Lett 52:279-285.
43
44 Kyyronen, P; Taskinen, H; Lindbohm, ML; et al. (1989) Spontaneous abortions and congenital malformations
45 among women exposed to tetrachloroethylene in dry cleaning. J Epidemiol Community Health 43:346-351.
46
47 Lacey, JV, Jr.; Garabrant, DH; Laing, TJ; et al. (1999) Petroleum distillate solvents as risk factors for
48 undifferentiated connective tissue disease (UCTD). Am J Epidemiol 149:761-770.
49
50 Lagakos, SW; Wessen, B; Zelen, M. (1986) Analysis of contaminated well water and health effect in Woburn,
51 Massachusetts. Journal of the American Statistical Association 81:583-596.
52
53 Liang, JC and Pacchierotti, F. (1988). Cytogenetic investigation of chemically-induced aneuploidy in mouse
54 spermatocytes. Mutat. Res. 201: 325-335.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-256 DRAFT-DO NOT CITE OR QUOTE
-------
1 Lake, BG; Evans, JG; Cunninghame, ME; et al. (1993) Comparison of the hepatic effects of nafenopin and
2 WY14643 on peroxisome proliferation and cell replication in the rat and Syrian hamster. Environ Health Perspect
3 5:241-247.
4
5 Lane, S E; Watts, R A; Bentham, G; Innes, N J; Scott, D G (2003) Are environmental factors important in primary
6 systemic vasculitis? A case-control study. Arthritis Rheum 48:814-823.
7
8 Lapinskas, PJ; JC, Gorton. (1997) Peroxisome proliferator-activated receptor alpha: central mediator of peroxisome
9 proliferatortoxicity. CUT Activities. Vol 17 No 1. 1-9. Research Triangle Park, NC.
10
11 Lapsley, M; Akers, K; Norden, AG. (1998) Sensitive assays for urinary retinol-binding protein and beta-2-
12 glycoprotein-1 based on commercially available standards. AnnCrinBiochem35 (Pt 1):115-119.
13
14 Lash, LH; Fisher, JW; Lipscomb, JC; et al. (2000a) Metabolism of Trichloroethylene. Environ Health Perspect 108:
15 (2):177B200.
16
17 Lash, LH; Parker, JC; Siegel-Scott, C. (2000b) Modes of Action of trichloroethylene for kidney tumorigenesis.
18 Environ Health Perspect 108:(2):225-240.
19
20 Lash, LH; Parker, JC. (2001) Hepatic and renal toxicities associated with perchloroethylene. Pharmacol Rev
21 53:177-208.
22
23 Lash, LH; Qian, W; Putt, DA; et al. (2002) Renal toxicity of perchloroethylene and s-(l,2,2-
24 trichlorovinyl)glutathione in rats and mice: sex- and species-dependent differences. Toxicol App Pharmacol
25 179:163-171.
26
27 Laslo-Baker, D; Barrera, M; Knittel-Keren, D; et al. (2004) Child neurodevelopmental outcome and maternal
28 occupational exposure to solvents. Arch Pediatr Adolesc Med 158(10):956-61.
29
30 Laughter, AR; Dunn, CS, Swanson CL; et al. (2004) Role of the peroxisome proliferator-activated receptor a
31 (PPARa) in responses to trichloroethylene and metabolite, trichloroacetate and dichloroacetate in mouse liver.
32 Toxicology 203:83-98.
33
34 Lauwerys, R; Herbrand, J; Buchet, J P; et al. (1983) Health surveillance of workers exposed to tetrachloroethylene
35 in dry- cleaning shops. Int Arch Occup Environ Health 52:69-77.
36
37 Lee, SS; Pineau, T; Drago, J; et al. (1995) Targeted disruption of the alpha isoform of the peroxisome proliferator-
3 8 activated receptor gene in mice results in abolishment of the pleiotropic effects of peroxisome proliferators. Mol
39 Cell Biol 15:3012-3022.
40
41 Lehmann, I; Rehwagen, M; Diez, U; etal. (2001) Enhanced in vivo IgE production and T cell polarization toward
42 the type 2 phenotype in association with indoor exposure to VOC: results of the LARS study. Int J Hyg Environ
43 Health 204:211-221.
44
45 Lehmann, I; Thoelke, A; Rehwagen, M; et al. (2002) The influence of maternal exposure to volatile organic
46 compounds on the cytokine secretion profile of neonatal T cells. Environ Toxicol 17:203-210.
47
48 Letkiewicz, F; Johnston P; Macaluso C; et al. (1982) Occurrence of tetrachloroethylene in drinking water, food, and
49 air. Prepared by JRB Associates JRB Project No. 2-613-03-852-29, for U.S. Environmental Protection Agency
50 contract 86-01-6388, task 29.
51
52 Lezak, MD. (1995) Neuropsychological assessment, 3rd ed. New York: Oxford University Press.
53 Lin, RS; Kessler, II. (1981) A multifactorial model for pancreatic cancer in man. Epidemiologic evidence. JAMA
54 245:147-152.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-257 DRAFT-DO NOT CITE OR QUOTE
-------
1 Lindbohm, ML; Taskinen, H; Sallmen, M; et al. (1990) Spontaneous abortions among women exposed to organic
2 solvents. Am JIndMed 17:449-463.
3
4 Lock, EA; Sani, Y; Moore, RB; et al. (1996) Bone marrow and renal injury associated with haloalkene cysteine
5 conjugates in calves. Arch Toxicol 70:607-619.
6
7 Loizou, GD. (2001) The application of physiologically based pharmacokinetic modeling in the analysis of
8 occupational exposure to perchloroethylene. Toxicol Lett 124:59-69.
9
10 Lowengart, RA; Peters, JM; Cicioni, C; et al. (1987) Childhood leukemia and parents' occupational and home
11 exposures. J Natl Cancer Inst 79:39-46.
12
13 Lumpkin, MH; Bruckner, JV; Cambell, JL; et al. (2003) Plasma binding of trichloroacetic acid in mice, rats, and
14 humans under cancer bioassay and environmental exposure conditions. Drug Metab Disp 31 (10):1203-1207.
15
16 Lundberg, I; Alfredsson, L; Plato, N; Sverdrup, B; Klareskog, L; Kleinau, S (1994) Occupation, occupational
17 exposure to chemicals and rheumatological disease. A register based cohort study. Scand J Rheumatol 23:305-310.
18
19 Lybarger, JA; Lichtveld, MY; Amler, RW. (1999) Biomedical testing of the kidney forpersons exposed to
20 hazardous substances in the environment. RenFail 21:263-274.
21
22 Lynge,E. (1994) Danish Cancer Registry as a resource for occupational research. JOccupMed 36:1169-1173.
23
24 Lynge, E; Thygesen, L. (1990) Primary liver cancer among women in laundry and dry-cleaning work in Denmark.
25 Scand J Work Environ Health 16:108-112.
26
27 Lynge, E; Carstensen, B; Andersen, O. (1995) Primary liver cancer and renal cell carcinoma in laundry and dry-
28 cleaning workers in Denmark. Scand J Work Environ Health 21:293-295.
29
30 Lynge, E; Anttila, A; Hemminki, K. (1997) Organic solvents and cancer. Cancer Causes Control 8:406-419.
31
32 MA DPH (Massachusetts Department of Public Health) (1997) Woburn childhood leukemia follow-up study. Final
33 eport. Available online at
34 http://www.mass.gov/Eeohhs2/docs/dph/environmental/investigations/woburn childhood leukemia follow.pdf.
35
36 MacNaughton, MG; Uddin, DE. (1984) Toxicology of mixed distillate and high-energy synthetic fuels. Mehlman,
37 MA; Hemstreet, GP; Thorpe, JJ; et al., eds. Advances in modern environmental toxicology. Vol.7. Renal effects of
38 petroleum hydrocarbons. Princeton, NJ: Princeton Scientific Publishers, Inc.; pp. 121-132.
39
40 Maitre, A; Hours, M; Bonneterre, V; et al. (2004) Systemic sclerosis and occupational risk factors: role of solvents
41 and cleaning products. JRheumatol 31:2395-2401.
42
43 Mallin, K. (1990) Investigation of a bladder cancer cluster in northwestern Illinois. Am J Epidemiol 132:896-106.
44
45 Malone, KE; Koepsell, TD; Dating, JR; et al. (1989) Chronic lymphcytic leukemia in relation to chemical exposures.
46 Am J Epidemiol 130:1152-1158.
47
48 Maloney, EK; Waxman, DJ. (1999) Trans-activation of PPAR-alpha and PPARgamma by structurally diverse
49 environmental chemicals. Toxicol Appl Pharmacol 161:209-218.
50
51 Maltoni, C; Cotti, G. (1986) Results of long-term carcinogenicity bioassay s of tetrachloroethylene on Sprague-
52 Dawley rats administered by ingestion. Acta Oncologica 7(1): 11-26.
53
54 Mandel, JS; McLaughlin, JK; Schlehofer, B; et al. (1995) International renal-cell cancer study. IV. Occupation. Int
55 J Cancer 61:601-605.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-258 DRAFT-DO NOT CITE OR QUOTE
-------
1 Manzo, L; Artigas, F; Martinez, E; et al. (1996) Biochemical markers of neurtoxicity: a review of mechanistic
2 studies and applications. Hum Exper Toxicol 15(suppl l):S20-235.
3
4 Marcus, SL; Miyata, KS; Zhang, B; et al. (1993) Diverse peroxisome proliferator-activated receptors bind to the
5 peroxisome proliferator-responsive elements of the rat hydratase/dehydrogenase and fatty acyl-CoA oxidase genes
6 but differentially induce expression. ProcNatlAcad Sci USA 90:5723-5727.
7
8 Maronpot, RR; Fox, T; Malarkey, DE; et al. (1995) Mutations in the ras proto-oncogene: clues to etiology and
9 molecular pathogenesis of mouse liver tumors. Toxicology 101:125-156.
10
11 Marsman, DS. (1991) Peroxisome proliferatorhepatocarcinogenesis in rats. Thesis. University of North Carolina,
12 Chapel Hill, NC.
13
14 Marsman, DS; Popp, JA. (1994) Biological potential of basophilic hepatocellular foci and hepatic adenoma induced
15 by the peroxisome proliferator, Wy-14,643. Carcinogenesis 15:111-117.
16
17 Marsman, DS; Cattley, RC; Conway, JG; et al. (1988) Relationship of hepatic peroxisome proliferation and
18 replicative DNA synthesis to the hepatocarcinogenicity of the peroxisome proliferators di(2-ethylhexyl)phthalate
19 and [4-chloro-6-(2,3-xylidino)-2- pyrimidinylthio]acetic acid (Wy-14,643) in rats. Cancer Res 48:6739-6744.
20
21 Marsman, DS; Goldsworthy, TL; Popp, JA. (1992) Contrasting hepatocytic peroxisome proliferation, lipofuscin
22 accumulation and cell turnover for the hepatocarcinogens Wy-14,643 and clofibric acid. Carcinogenesis 13:1011-
23 1017.
24
25 Mattsson, JL; Albee, RR; Yano, BL; et al. (1998) Neurotoxicologic examination of rats exposed to 1,1,2,2-
26 tetrachloroethylene (perchloroethylene) vapor for 13 weeks. Neurotoxicol Teratol 20:83-98.
27
28 Mayes, MD. (1999) Epidemiologic Studies of Environmental Agents and Systemic Autoimmune Diseases. Environ
29 Health Perspect 107:743-748.
30
31 McCredie, M; Stewart, JH. (1993) Risk factors for kidney cancer in New South Wales. IV. Occupation. Br J Ind
32 Med 50:349-354.
33
34 McDermott, M J; Mazor, K A; Shost, S J; Narang, R S; Aldous, K M; Storm, J E (2005) Tetrachloroethylene (PCE,
3 5 Perc) levels in residential dry cleaner buildings in diverse communities in New York City. Environ Health Perspect
36 113:1336-1343.
37
38 McDonald, AD; Armstrong, B; Cherry, NM; et al. (1986) Spontaneous abortion and occupation. J Occup Med
39 28:1232-1238.
40
41 McDonald, AD; McDonald, JC; Armstrong, B; et al. (1987) Occupation and pregnancy outcome. Br J Ind Med
42 44:521-526.
43
44 McKinney, LL; Weakley, FB; Eldridge, AC; et al. (1957) S-(dichlorovinyl) L-cysteine: an agent causing fatal
45 aplastic anemia in calves. J AM Chem Soc 51:16946.
46
47 McKone, TE; Daniels, JL (1991) Estimating human exposure through multiple pathways from air, water, and soil.
48 Regul Toxicol Pharmacol 13:36-61.
49
50 McLaughlin, JK; Blot, WJ. (1997) A critical review of epidemiology studies of trichloroethylene and
51 perchloroethylene and risk of renal-cell cancer. Int Arch Occup Environ Health 70:222-231.
52
53 McLaughlin, JK; Malker, HS; Stone, BJ; et al. (1987) Occupational risks for renal cancer in Sweden. Br J Ind Med
54 44:119-123.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-259 DRAFT-DO NOT CITE OR QUOTE
-------
1 Meckler, LC; Phelps, DK. (1966) Liver disease secondary to tetrachloroethylene exposure. A case report. JAMA
2 197:662-663.
3
4 Meek, ME; Bucher, JR; Cohen, SM; et al. (2003) A framework for human relevance analysis of information on
5 carcinogenic modes of action. Crit Rev Toxicol 33(6):591-653.
6
7 Meijer, J; Afzelius, BA. (1989) Effects of clofibrate treatment and of starvation on peroxisomes, mitochondria, and
8 lipid droplets in mouse hepatocytes: a morphometric study. J Ultrastruct Mol Struct Res 102:87-94.
9
10 Mellemgaard, A; Engholm, G; McLaughlin, JK; et al. (1994) Occupational risk factors for renal-cell carcinoma in
11 Denmark. ScandJ Work Environ Health 20:160-165.
12
13 Melnick, RL. (2001) Is peroxisome proliferation an obligatory precursor step in the carcinogenicity of di(2-
14 ethylhexyl)phthalate (DEHP)? Environ Health Perspect 109:437-442.
15
16 Melnick, RL. (2002) The IARC evaluation of di02-ethylhexl)phthalate (DEHP): a flawed decision based on an
17 untested hypothesis. Int J Occup Environ Health 8:284-286.
18
19 Melnick, RL; Kohn, MC; Huff, J. (1997) Weight of evidence versus weight of speculation to evaluate the alpha2u-
20 globulin hypothesis. Environ Health Perspect 105:904-906.
21
22 Mergler, D; Blain, L. (1987) Assessing color vision loss among solvent-exposed workers. Am J Ind Med 12:195-
23 203.
24
25 Mergler, D; Blain, L; Lagace, JP. (1987) Solvent related colour vision loss: an indicator of neural damage? Int Arch
26 Occup Environ Health 59:313-321
27
28 Mergler, D; Blain, L; Lemaire, J; et al. (1988a) Colour vision impairment and alcohol consumption. Neurotoxicol
29 Teratol 10:255-260.
30
31 Mergler, D; Belanger, S; De Grosbois, S; et al. (1988b) Chroma! focus of acquired chromatic discrimination loss
32 and solvent exposure among printshop workers. Toxicology 49:341-348.
33
34 Mergler, D; Huel, G; Bowler, R; et al. (1991) Visual dysfunction among former microelectronics assembly workers.
35 Arch Environ Health 46:326-334.
36
37 Mergler, D; Huel, G; Belanger, S; et al. (1996) Surveillance of early neurotoxic dysfunction. Neurotoxicol 17:803-
38 812.
39
40 Meskar, A; Plee-Gautier, E; Amet, Y; et al. (2001) Alcohol-xenobiotic interactions: role of cytochrome P4502E1.
41 Pathol Biol (Paris) 49(9):696-702.
42
43 Mester, B; Neiters, A; Deeg, E; et al. (2006) Occupation and malignant lymphoma: a population based case control
44 study in Germany. Occup Environ Med 63:17-26.
45
46 Meyer, K; Lee, JS; Dyck, PA; et al. (2003) Molecular profiling of hepatocellular carcinomas developing
47 spontaneouwly in acyl-CoA oxidase deficient mice: comparison with liver tumors induced in wild-type mice by a
48 peroxisome proliferator and a genotoxic carcinogen. Carcinogenesis 24:975-984.
49
50 Mihic, SJ. (1999) Acute effects of ethanol on GAB AA and glycine receptor function. Neurochemlnt 35:115-123.
51
52 Miligi, L; Seniori Constantini, A; Crosignani, P; et al. (1999) Occupational, environmental, and life-style factors
53 associated with the risk of hematolymphopoietic malignancies in women. Am J Ind Med 36:60-69.
54
55 Miligi, L; Seniori Costantini, A; Benvenuti, A; et al. (2006) Occupational exposure to solvents and the risk of
56 lymphomas. Epidemiology 17:552-561.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-260 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 Monster, AC; Boersma, G; Steenweg, H. (1979) Kinetics of tetrachloroethylene in volunteers; influence of exposure
3 concentration and work load. Int Arch Occup Environ Health 42:303-309.
4
5 Moody, DE; Reddy, JK; Lake, BG; et al. (1991) Peroxisome proliferation and nongenotoxic carcinogenesis:
6 commentary on a symposium. Fundam Appl Toxicol 16:233-248.
7
8 Moore, MM; Harrington-Brock, K. (2000) Mutagenicity of trichloroethylene and Its metabolites: implications for
9 the risk assessment of trichloroethylene. Environ Health Perspect 108 Suppl 2:215-223.
10
11 Morrow, LA; Ryan, CM; Hodgson, MJ; et al. (1990) Alterations in cognitive and psychological functioning after
12 organicsolvent exposure. J Occup Med 32:444- 450.
13
14 Morton, W; Marjanovic, D. (1984) Leukemia incidence by occupation in the Portland-Vancouver metropolitan area.
15 Am JInd Med 6:185-205.
16
17 Morton, W and Marjanovic, D (1984) Leukemia incidence by occupation in the Portland-Vancouver metropolitan
18 area. Am J Ind Med 6:185-205.
19
20 Moser, VC; Cheek, BM; MacPhail, RC. (1995) A multidisciplinary approach to lexicological screening: III.
21 Neurobehavioral toxicity. J Toxicol Environ Health 45:173-210.
22
23 Moser, VC; Phillips, PM; McDaniel, KL; et al. (1999) Behavioral evaluation of the neurotoxicity produced by
24 dichloroacetic acid in rats. Neurotoxicol Teratol 21:719-731.
25
26 Moslen, MT; Reynolds, ES; Szabo, S. (1977).Enhancement of the metabolism and hepatotoxicity of
27 trichloroethylene and perchloroethylene. Biochem Pharmacol 26(5):369-75.
28
29 Motohashi, Y; Miyazaki, Y; Takano, T. (1993) Assessment of behavioral effects of tetrachloroethylene using a set
30 of time-series analyses. Neurotoxicol Teratol 15:3-10.
31
32 Mundt, KA; Birk, T; Burch, MT. (2003) Critical review of the epidemiological literature on occupational exposure
33 to perchloroethylene and cancer. Int Arch Occup Environ Health 76:473-491.
34
35 Mueser, KT; McGurk, SR. (2004) Schizophrenia. Lancet 363:2063-2072.
36
37 Murakami, K; Horikawa, K. (1995) The induction of micronuclei in mice hepatocytes and reticulocytes by
38 tetrachloroethylene. Chemosphere 31:3733-3739.
39
40 Mutti, A; Smargiassi, A. (1998) Selective vulnerability of dopaminergic systems to industrial chemicals: risk
41 assessment of related neuroendocrine changes. Toxicol Ind Health 14:311-23.
42
43 Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-cleaners.
44 Lancet 340:189-193.
45
46 Muttray, A; Wolff, U; Jung, D; et al. (1997) Blue-yellow deficiency in workers exposed to low concentrations of
47 organic solvents. Int Arch Occup Environ Health 70:407-412.
48
49 Mazzullo, MS; Grilli, G; Lattanzi, G; et al. (1987). Evidence of DNA binding activity of perchloroethylene. Res
50 Commun Chem Pathol Pharmacol 58(2): 215-35.
51
52 Nagano, K; Nishizawa, T; Yamamoto, S; et al. (1998) Inhalation carcinogenesis studies of six halogenated
53 hydrocarbons in rats and mice. In: Chiyotani, K; Hosoda, Y; Aizawa, Y, eds. Advances in the prevention of
54 occupational respiratory diseases: Proceedings of the 9th international conference on occupational respiratory
55 diseases, Kyoto, Japan, 13-16, October 1997. Amsterdam: Elsevier; pp. 741-746.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-261 DRAFT-DO NOT CITE OR QUOTE
-------
1 Nakai, JS; Stathopulos, PB; Campbell, GL; et al. (1999) Penetration of chloroform, trichloroethylene and
2 tetrachloroethylene through human skin. J Toxicol Environ Health A 58:157-170.
3
4 Nakajima, T; Kamijo, Y; Usuda, N; et al. (2000) Sex-dependent regulation of hepatic peroxisome proliferation in
5 mice by trichloroethylene via peroxisome proliferator-activated receptor alpha (PPAR-alpha). Carcinogenesis
6 21:677-682.
7
8 Nakatsuka, H; Watanabe, T; Takeuchi, Y; et al. (1992) Absence of blue-yellow color vision loss among workers
9 exposed to toluene or tetrachloroethylene, mostly at levels below occupational exposure limits. Int Arch Occup
10 Environ Health 64:113-117.
11
12 Narotsky, MG; Kavlock, RJ. (1995) A multidisciplinary approach to toxicological screening: II. Developmental
13 toxicity. J Toxicol Environ Health 45:145-171.
14
15 NCI (National Cancer Institute). (1977) Bioassay of tetrachloroethylene for possible carcinogenicity. DHEW Pub.
16 (NIH)77-813.
17
18 Nelson, BK; Taylor, BJ; Setzer, JV; et al. (1980) Behavioral teratology of perchloroethylene in rats. J Environ
19 Pathol Toxicol 3:233-250.
20
21 Nelson, MA; Bull, PJ. (1988) Induction of strand breaks in DNA by trichloroethylene and metabolites in rat and
22 mouse liver in vivo. Toxicol Appl Pharmacol 94:45-54.
23
24 Nelson, MA; Lansing, AJ; Sanchez, IM; et al. (1989) Dichloroacetic acid and trichloroacetic acid-induced DNA
25 strand breaks are independent of peroxisome proliferation. Toxicology 58:239-248.
26
27 Nelson, DE; Emont, SL; Brackbill, RM; et al. (1994) Cigarette smoking prevalence by occupation in the United
28 States. A comparison between 1978 to 1980 and 1987 to 1990. J Occup Med 36:516-525.
29
3 0 Nemali, MR; Usuda, N; Reddy, MK; et al. (1988) Comparison of constitutive and inducible levels of expression of
31 peroxisomal beta-oxidation and catalase genes in liver and extrahepatic tissues of rat. Cancer Res 48:5316-5324.
32
33 Newman, TB; Hulley, SB. (1996) Carcinogenicity of lipid-lowering drugs. JAMA 275:55-660.
34
3 5 Nietert, PJ; Sutherland, SE; Silver, RM; et al. (1998) Is occupational organic solvent exposure a risk factor for
36 scleroderma? [published erratum appears in Arthritis Rheum 1998 Aug;41(8):1512]. Arthritis Rheum
37 41:111161118.
38
39 Nomiyama, K; Liu, SJ; Nomiyama, H. (1992) Critical levels of blood and urinary cadmium, urinary beta 2-
40 microglobulin and retinol-binding protein for monitoring cadmium health effects. IARC Sci Publ 325-340.
41
42 Nowell, PC. (1991) Genetic instability and tumor development. Basic Life Sci 57:221-228.
43
44 NRC (National Research Council). (1993) Pesticides in the diets of infants and children. Washington, DC: National
45 Academy Press.
46
47 NRC (National Research Council). (1995) Biologic markers in urinary toxicology. Washington, DC: National
48 Academy Press.
49
50 NRC (National Research Council). (2005) Assessing the human health risks of trichloroethylene: Key scientific
51 issues.
52
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-262 DRAFT-DO NOT CITE OR QUOTE
-------
1 NTP (National Toxicology Program). (1983) Toxicology bioassay of pentachloroethane (CAS No.76-01-7) in
2 F344/N rats and B6C3F1 mice (gavage study). NIH Publication No. 83-1788, NTP TR 232.
3
4 NTP (National Toxicology Program). (1986a) Toxicology and carcinogenesis studies of tetrachloroethylene
5 (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3F1 mice. 311. Available from: National
6 Toxicology Program, National Institutes of Health, Public Health Service, U.S. Department of Health and Human
7 Services.
8
9 NTP (National Toxicology Program). (1986b) Toxicology and carcinogenesis studies of isophorone (CAS No. 78-
10 59-1) in F344/N rats and B6C3F1 mice (gavage studies). NIH Publication No. 86-2547. NTPTR291.
11
12 NTP (National Toxicology Program). (1987) Toxicology and carcinogenesis studies of dimethyl methylphosphonate
13 (CAS No. 756-79-6) in F344/N rats and B6C3F1 mice (gavage studies). Nffl Publication No. 87-2579, NTP TR
14 323.
15
16 NTP (National Toxicology Program). (1988) Toxicology and carcinogenesis studies of d-limonene (CAS No.
17 598927-5) in F344/N rats and B6C3F1 mice (gavage studies). NIH Publication No. 88-2902, NTP TR 347.
18
19 NTP (National Toxicology Program). (1989) Toxicology and carcinogenesis studies of hexachloroethane (CAS No.
20 67-72-1) in F344/N rats (gavage studies). NTP TR No. 361, NIH Publication No. 89-2816. Available from:
21 National Toxicology Program, National Institutes of Health, Public Health Service, U.S. Department of Health and
22 Human Services.
23
24 NTP (National Toxicology Program). (1990a) Toxicology and carcinogenesis studies of d-limonene in F344/N rats
25 and B6C3F1 mice (gavage studies). NTP TR 347. Washington, DC, U.S. Department of Health and Human
26 Services.
27
28 NTP (National Toxicology Program). (1990b) Carcinogenesis studies of trichloroethylene (without
29 epiochlorohydrin) in F344/N rats and B6C3F1 mice. NIH Publication No. 90-1799, NTP TR 243.
30
31 NTP (National Toxicology Program). (2000a) Toxicology and carcinogenesis studies of chloral hydrate (CAS NO.
32 302-17-0) in B6C3F1 mice (gavage study). NTP TR 502, Nffl Publication No. 00-4436. Available from: National
33 Toxicology Program, National Institutes of Health, Public Health Service, U.S. Department of Health and Human
34 Services.
35
36 NTP (National Toxicology Program). (2000b) Toxicology and carcinogenesis studies of chloral hydrate (ad libitum
37 and feed restricted) (CAS NO. 302-17-0) in male B6C3F1 mice (gavage study). NTP TR 503, Nffl Publication No.
38 00-4437. Available from: National Toxicology Program, National Institutes of Health, Public Health Service, U.S.
3 9 Department of Health and Human Services.
40
41 Nutrition Foundation. (1983) The relevance of mouse liver hepatoma to human carcinogenic risk: a report of the
42 International Expert Advisory Committee to the Nutrition Foundation. Washington, DC.
43
44 NYS DOH (New York State Department of Health). (1997) Tetrachloroethylene ambient air criteria document.
45 Final Report. Albany, NY.
46
47 NYS DOH (New York State Department of Health) (2000) Evaluation of residential exposure to tetrachloromethane
48 biomarkers of dose and neurological tests. Final deliverable from NYS DOH to U.S. Environmental Protection
49 Agency under Cooperative Agreement #CR 824400-01.
50
51 NYS DOH (New York State Department of Health) (2004) Comments on ANeurotoxicity of Tetrachloroethylene
52 (perchloroethylene)-Discussion Paper. @ EPA/600/P-03/005A.
53
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-263 DRAFT-DO NOT CITE OR QUOTE
-------
1 NYS DOH (New York State Department of Health) (2005a) Improving human risk assessment for
2 tetrachloroethylene by using biomarkers and neurobehavioral testing. U.S. EPA Star Grant #R827445. Grant
3 #R827446. Available online at
4 http://cfpub.epa.gov/ncer abstracts/index.cfm/fuseaction/displav.abstractDetail/abstract/977/reprort/O.
5
6 NYS DOH (New York State Department of Health) (2005b) Pumpkin patch day care center investigation. Final
7 Report. New York State Department of Health, Bureau of Toxic Substance Assessment, Division of Environmental
8 Health Assessment, Center for Environmental Health
9
10 NYS DOH (New York State Department of Health) (2005c) Pumpkin patch day care center follow-up evaluation.
11 Final Report. New York State Department of Health, Bureau of Toxic Substance Assessment, Division of
12 Environmental Health Assessment, Center for Environmental Health.
13
14 NYS OAG (New York State Office of Attorney General). (2004a) Comments from Judy Schreiber, Ph.D. for EPA
15 Docket ORD-2003-0014. March 1, 2004.
16
17 NYS OAG (New York State Office of Attorney General). (2004b) Comment focusing on APeer review workshop of
18 the neurotoxicity of tetrachloroethylene discussion paper: submitted by Judith S. Schreiber, Ph.D. Chief Scientist,
19 Albany Environmental Protection Bureau. ORD-2003-0014-0007. Available online at
20 http://docket.epa.gov/edkpub/do/EDKStaffItemDetailView?objectId=090007d48038f97c.
21
22 Odum, J; Green, T; Foster, JR; et al. (1988) The role of trichloracetic acid and peroxisome proliferation in the
23 differences in carcinogenicity of perchloroethylene in the mouse and rat. Toxicol Appl Pharmacol 92:103-112.
24
25 O=Flaherty, EJ; Scott, W; Schreiner, C; et al. (1992) A physiologically based kinetic model of rat and mouse
26 gestation: disposition of a weak acid. Toxicol Appl Pharmacol 112:245-256.
27
28 Olsen, J; Hemminki, K; Ahlborg, G; et al. (1990) Low birthweight, congenital malformations, and spontaneous
29 abortions among dry-cleaning workers in Scandinavia. Scand J Work Environ Health 16:163-168.
30
31 Olson, MJ; Garg, BD; Murty, CV; et al. (1987) Accumulation of alpha 2u-globulin in the renal proximal tubules of
32 male rats exposed to unleaded gasoline. Toxicol Appl Pharmacol 90:43-51.
33
34 Onofrj, M; Thomas, A; Paci, C; et al. (1998) Optic neuritis with residual tunnel vision in perchloroethylene toxicity.
35 J Toxicol Clin Toxicol 36:603-607.
36
37 Opdam, JJ. (1989) Intra and interindividual variability in the kinetics of a poorly and highly metabolising solvent.
38 BrJIndMed46:831-845.
39
40 Opdam, JJ; Smolders, JF. (1986) Alveolar sampling and fast kinetics of tetrachloroethylene in man. I. Alveolar
41 sampling. Br J Ind Med 43:814-824.
42
43 Pahler, A; Parker, J; Dekant, W. (1999) Dose-dependent protein adduct formation in kidney, liver, and blood of rats
44 and in human blood after perchloroethene inhalation. Toxicol Sci 48:5-13.
45
46 Palmer, CN; Hsu, MH; Griffin, KJ; et al. (1998) Peroxisome proliferator activated receptor-alpha expression in
47 human liver. Mol Pharmacol 53:14-22.
48
49 Parent, ME; Siemiatycki, J; Fritschi, L. (2000) Workplace exposures and oesophageal cancer. Occup Environ Med
50 57:3256334.
51
52 Park, JH; Spengler, JD; Yoon, DW; et al. (1998) Measurement of air exchange rate of stationary vehicles and
53 estimation of in-vehicle exposure. J Expo Anal Environ Epidemiol 8:65-78.
54
55 Partanen, T; Heikkila, P; Hernberg, S; et al. (1991) Renal cell cancer and occupational exposure to chemical agents.
56 Scand J Work Environ Health 17:231-239.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-264 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 Pastino, GM; Yap, WY; Carroquino, M. (2000) Human variability and susceptibility to trichloroethylene [In Process
3 Citation]. Environ Health Perspect 108 Suppl 2:201-214.
4
5 Paulu, C; Aschengrau, A; Ozonoff, D. (1999) Tetrachloroethylene-contaminated drinking water in Massachusetts
6 and the risk of colon-rectum, lung, and other cancers. Environ Health Perspect 107:265-271.
7
8 Paulu, C; Aschengrau, A; Ozonoff D. (2002) Exploring associations between residential location and breast cancer
9 incidence in a case-control study. Environ Health Perspect 110:471-478.
10
11 Pegg, DG; Zempel, JA; Braun, WH; et al. (1979) Disposition of tetrachloro(14C)ethylene following oral and
12 inhalation exposure in rats. Toxicol Appl Pharmacol 51:465-474.
13
14 Pereira, MA. (1996) Carcinogenic activity of dichloroacetic acid and trichloroacetic acid in the liver of female
15 B6C3F1 mice. Fundam Appl Toxicol 31:192-199.
16
17 Pereira, MA; Wang, W; Kramer, PM; et al. (2004a) Prevention by methionine of dichloroacetic acid-induced liver
18 cancer and DNA hypomethylation in mice. Toxicol Sci 77:243-248.
19
20 Pereira, MA; Tao, L; Wang, W; et al. (2004b) Modulation by celecoxib and difluoromethylornithine of the
21 methylation of DNA and the estrogen receptor-{-a} gene in rat colon tumors. Carcinogenesis 25:1917-1923.
22
23 Perrin, MC; Opler, MG; Harlap, S, et al. (2007) Tetrachloroethylene exposure and risk of schizophrenia: Offspring
24 of dry cleaners in a populaton birth cohort, preliminary findings. Schizophr Res. 90(l-3):251-4.
25
26 Perrone, CE; Shao, L; Williams, GM. (1998) Effect of rodent hepatocarcinogenic peroxisome proliferators on fatty
27 acyl-CoA oxidase, DNA synthesis, and apoptosis in cultured human and rat hepatocytes. Toxicol. Appl Pharmacol.
28 150:277-286.
29
30 Pesch, B; Haerting, J; Ranft, U; et al. (2000a) Occupational risk factors for renal cell carcinoma: agent-specific
31 results from a case-control study in Germany. MURC Study Group. Multicenter urothelial and renal cancer study.
32 IntJEpidemiol 29:1014-1024.
33
34 Pesch, B; Haerting, J; Ranft, U; et al. (2000b) Occupational risk factors for urothelial carcinoma: agent-specific
3 5 results from a case-control study in Germany. MURC Study Group. Multicenter Urothelial and Renal Cancer. Int J
36 Epidemiol 29:238-247.
37
3 8 Peters, JM; Vanden Heuvel, JP. (2002) Peroxisome, peroxisome proliferators and peroxisome-proliferator activated
39 receptors (PPARs). In: Vanden Heuvel, JP; Perdew, GH; Mattes, WB; et al.; eds. Cellular and molecular toxicology.
40 Amsterdam: Elsevier Science, pp 133-158.
41
42 Peters, JM; Cattley, RC; Gonzalez, FJ. (1997a) Role of PPARalpha in the mechanism of action of the nongenotoxic
43 carcinogen and peroxisome proliferator Wy-14,643. Carcinogenesis 18:2029-2033.
44
45 Peters, JM; Hennuyer, N; Staels, B; et al. (1997b). Alterations in lipoprotein metabolism in peroxisome proliferator-
46 activated receptor-alpha-deficient mice. J Biol Chem 272:27303-27312.
47
48 Peters, J; Rusyn, I; Rose, M; et al. (2000) Peroxisome proliferator-activated receptor alpha is restricted to hepatic
49 parenchymal cells, not Kupffer cells: implications for the mechanism of action of peroxisome proliferators in
50 hepatocarcinogenesis Carcinogenesis 21:823-826.
51
52 Phillips, RD; Moran, EJ; Dodd, DE, et al. (1987) A 14-week vapor inhalation toxicity study of methyl isobutyl
53 ketone. Fundam Appl Toxicol 9:380-388.
54
55 Poet, TS; Weitz, KK; Gies, RA; et al. (2002) PBPK modeling of the percutaneous absorption of perchloroethylene
56 from a soil matrix in rats and humans. Toxicol Sci 67:17-31.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-265 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 Pohl, HR; Roney, N; Wilbur, S; et al. (2003) Six interaction profiles for simple mixtures. Chemosphere 53:183-197.
3
4 Pohlabeln, H; Boffetta, P; Ahrens, W; et al. (2000) Occupational risks for lung cancer among nonsmokers.
5 Epidemiology 11:532-538.
6
7 Popp, JA. (1984) Mouse liver neoplasia. Washington, DC: Hemisphere.
8
9 Popp, W; Muller, G; Baltes-Schmitz, B; et al. (1992) Concentrations of tetrachloroethylene in blood and
10 trichloroacetic acid in urine in workers and neighbors of drycleaning shops. Int Arch Occup Environ Health
11 63(6):393-395.
12
13 Potter, CL; Chang, LW; DeAngelo, AB; et al. (1996) Effects of four trihalomethanes on DNA strand breaks, renal
14 hyaline droplet formation and serum testosterone in male F-344 rats. Cancer Lett 106:235-242.
15
16 Price, RG; Taylor, SA; Olivers, I; et al. (1996) Development and validation of new screening tests for nephrotoxic
17 effects. HumExpToxicol 15 Suppl 1:S10-S19.
18
19 Rachootin, P; Olsen, J. (1983) The risk of infertility and delayed conception associated with exposures in the Danish
20 workplace. J Occup Med 25:394-402.
21
22 Rampy, LW; Quast, JF; Balmer, MF; et al. (1978) Results of a long-term inhalation toxicity study:
23 perchloroethylene in rats. Toxicology Research Laboratory, Health and Environmental Research, Dow Chemical
24 Company, Midland, MI.
25
26 Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
27 groundwater for a bathing and showering determination. Risk Anal 13:37-49.
28
29 Rao, MS; Reddy, JK. (1991) An overview of peroxisome proliferator-induced hepatocarcinogenesis. Environ
30 Health Perspect 93:205-209.
31
32 Rea, WJ; Ross, GH; Johnson, AR; et al. (1991) Chemical sensitivity in physicians. Bol Asoc Med P R 83:383-388.
33
34 Reddy, JK; Lalwani, ND. (1983) Carcinogenesis by hepatic peroxisome proliferators: evaluation of the risk of
3 5 hypolipidemic drugs and industrial plasticizers to humans. Crit Rev Toxicol 12:1-58.
36
37 Reddy, JK; Rao, MS. (1986) Peroxisome proliferators and cancer: mechanisms and implications. Trends in
38 Pharmacol Sci 7:438-443.
39
40 Reddy, JK; Azarnoff, DL; Hignite, CE. (1980) Hypolipidaemic hepatic peroxisome proliferators form a novel class
41 of chemical carcinogens. Nature 283:397-398.
42
43 Reddy, JK; Lalwani, ND; Qureshi, SA; et al. (1984) Induction of hepatic peroxisome proliferation in nonrodent
44 species, including primates. Am JPathol 114:171-183.
45
46 Reddy, JK; Reddy, MK; Usman, MI; et al. (1986) Comparison of hepatic peroxisome proliferative effect and its
47 implication for hepatocarcinogenicity of phthalate esters, di(2- ethylhexyl) phthalate, and di(2-ethylhexyl) adipate
48 with a hypolipidemic drug. Environ Health Perspect 65:317-327.
49
50 Regan, D. (1989) Human brain electrophysiology. New York: Elsevier Science.
51
52 Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
53 for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
54 306.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-266 DRAFT-DO NOT CITE OR QUOTE
-------
1 Renwick, AG. (1998) Toxicokinetics in infants and children in relation to the ADI and TDI. Food Additives and
2 Contaminants 15 Suppl: 17-35.
3
4 Richmond, RE; Carter, JH; Carter, HW; et al. (1995) Immunohistochemical analysis of dichloroacetic acid (DCA)-
5 induced hepatocarcinogenesis in male Fischer (F344) rats. Cancer Lett 92:67-76.
6
7 Rijhsinghani, KS; Abrahams, C; Swerdlow, MA; et al. (1986) Induction of neoplastic lesions in the livers of C57BL
8 x C3HF1 mice by chloral hydrate. Cancer Detect Prev 9:279-288.
9
10 Ris, H. 1949. The anaphase movement of chromosomes in the spermatocytes of the grasshopper. Biol Bull. 96:90-
11 106.
12
13 Roberts, RA; Soames, AR; Gill, JH; et al. (1995) Non-genotoxic hepatocarcinogens stimulate DNA synthesis and
14 their withdrawal induces apoptosis, but in different hepatocyte populations. Carcinogenesis 16:1693-1698.
15
16 Roberts, RA; James, NH; Hasmall, SC; et al. (2000) Apoptosis and proliferation in nongenotoxic carcinogenesis:
17 species differences and role of PPAR alpha. Toxicol Lett 112-113:49-57.
18
19 Rose, ML; Rivera, CA; Bradford, BU; et al. (1999) Kupffer cell oxidant production is central to the mechanism of
20 peroxisome proliferators. Carcinogenesis 20:27-33.
21
22 Rosengren, LE; Kjellstrand, P; Haglid, KG. (1986) Tetrachloroethylene: levels of DNA and S-100 in the gerbil CNS
23 after chronic exposure. Neurobehav Toxicol Teratol 8:201-206.
24
25 Rothman, KJ; Greenland, S. (1998) Modern epidemiology. New York: Lippincott, Williams, and Wilkins.
26
27 Rowe, VK; McCollister, DD; Spencer, HC; et al. (1952) Vapor toxicity of tetrachloroethylene for laboratory animals
28 and human subjects. Arch Ind Hyg Occup Med 5:556-579.
29
30 Roy, AK; McMinn, DM; Biswas, NM. (1975). Estrogenic inhibition of the hepatic synthesis of a2u-globurininthe
31 rat. Endocrinology 97:1501-1508.
32
33 Ruder, AM; Ward, EM; Brown, DP. (1994) Cancer mortality in female and male dry-cleaning workers. J Occup
34 Med 36:867-874.
35
36 Ruder, AM; Ward, EM; Brown, DP. (2001) Mortality in dry-cleaning workers: an update. Am J Ind Med 39:121-
ST 132.
38
39 Russo, A; Pacchierotti, F; Metalli, P. (1984) Nondisjunction induced in mouse spermatogenesis by chloral hydrate, a
40 metabolite of trichloroethylene. Environ. Mutagen. 6: 695-703.
41
42 Rusyn, I; Rose, ML; Bojes, HK; et al. (2000a) Novel role of oxidants in the molecular mechanism of action of
43 peroxisome proliferators. Antioxid Redox Signal 2:607-621.
44
45 Rusyn, I; Denissenko, MF; Wong, V; et al. (2000b) Expression of base excision repair enzymes in rat and mouse
46 liver is induced by peroxisome proliferators and is dependent upon carcinogenic potency. Carcinogenesis 21:2141-
47 2145.
48
49 Rusyn, I; Kadiiska, MB; Dikalova, A; et al. (2001) Phthalates rapidly increase production of reactive oxygen species
50 in vivo: Role of Kupffer cells. Mol. Pharmacol 59:744-750.
51
52 Saillenfait, AM; Langonne, I; Sabate, JP. (1995) Developmental toxicity of trichloroethylene, tetrachloroethylene
53 and four of their metabolites in rat whole embryo culture. Arch Toxicol 70:71-82.
54
55 Saland, G. (1967) Accidental exposure to perchloroethylene. NY State J Med 67:2359-2361.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-267 DRAFT-DO NOT CITE OR QUOTE
-------
1 Sallmen, M; Lindbohm, ML; Kyyronen, P; et al. (1995) Reduced fertility among women exposed to organic
2 solvents. Am JIndMed 27:699-713.
3
4 Salmon, AG; Kizer, KW; Zeise, L; et al. (1995) Potential carcinogenicity of chloral hydrateCa review. J Toxicol
5 Clin Toxicol 33:115-121.
6
7 Sapone, A; Peters, JM; Sakai, S; et al. (2000) The human peroxisome proliferator-activated receptor alpha gene:
8 Identification and functional characterization of two natural allelic variants. Pharmacogenetics 10:321-333.
9
10 Sarangapani, R; Gentry, PR; Covington, TR; et al. (2003) Evaluation of the potential impact of age- and gender-
11 specific lung morphology and ventilation rate on the dosimetry of vapors. Inhal Toxicol 15:987-1016.
12
13 Savolainen, H; Pfaffli, P; Tengen, M; Vainio, H. (1977a) Biochemical and behavioural effects of inhalation
14 exposure to tetrachloroethylene and dichloromethane. J Neuropathol Exp Neurol 36:941-949.
15
16 Savolainen, H; Pfaffli, P; Tengen, M; Vainio, H. (1977b) Trichloroethylene and 1,1,1-trichloroethane: effects on
17 brain and liver after five days intermittent inhalation. Arch Toxicol 38:229-237.
18 Sawyers, CL; Denny, CT; Witte, ON. (1991) Leukemia and the disruption of normal hematopoiesis. Cell 64:337-
19 350.
20
21 Schatten, H; Chakrabarti, A. (1998) Centrosome structure and function is altered by chloral hydrate and diazepam
22 during the first reproductive cell cycles in sea urchin eggs. Eur J Cell Biol 75:9-20.
23
24 Schattenberg, JM; Galle, PR; Schuchmann, M. (2006) Apoptosis in liver disease. Liver Int 26:904-911.
25
26 Scherr, PA; Hutchison, GB; Neiman, RS. (1992) Non-Hodgkin's lymphoma and occupational exposure. Cancer Res
27 52:5503s-5509s.
28
29 Schlehofer, B; Heuer, C; Blettner, M; et al. (1995) Occupation, smoking and demographic factors, and renal cell
30 carcinoma in Germany. Int JEpidemiol 24:51-57.
31
32 Schlesselman, J. (1982) Case-control studies. Design, conduct, analysis. New York: Oxford University Press.
33
34 Schlichting, LM; Wright, PF; Stacey, NH. (1992) Effects of tetrachloroethylene on hepatic and splenic
3 5 lymphocytotoxic activities in rodents. Toxicol Ind Health 8:255-266.
36
37 Schreiber, JS; House, S; Prohonic, E; et al. (1993) An investigation of indoor air contamination in residences above
38 dry cleaners. Risk Anal 13:335-344.
39
40 Schreiber, JS. (1993) Predicted infant exposure to tetrachloroethene in human breastmilk. Risk Anal 13:515-524.
41 Schreiber. JS. (1997) Transport of organic chemicals to breast milk: Tetrachloroethene case study. In Kacew S,
42 Lambert G (eds): Environmental Toxicology and Pharmacology of Human Development. Washington, DC, Taylor
43 and Francis.
44
45 Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
46 tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect 110:655-664.
47
48 Schultz, R; Yan, W; Toppari, J; et al. (1999) Expression of peroxisome proliferator-activated receptor alpha
49 messenger ribonucleic acid and protein in human and rat testis. Endocrinology 140:2968-2975.
50
51 Schultze, MO; Klubes, P; Perman, V; et al. (1959) Blood dyscrasia in calves induced by S-(dichlorovinyl)-L-
52 cysteine. Blood 14:1015-1025.
53
54 Schumann, AM; Quast, JF; Wantanabe, PG. (1980) The pharmacokinetics and macromolecular interactions of
55 perchloroethylene in mice and rats as related to oncogenicity. Toxicol Appl Pharmacol 55:207-219.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-268 DRAFT-DO NOT CITE OR QUOTE
-------
1 Schwetz, BA; Leong, KJ; Gehring, PJ. (1975) The effect of maternally inhaled trichloroethylene, perchloroethylene,
2 methyl chloroform, and methylene chloride on embryonal and fetal development in mice and rats. Toxicol Appl
3 Pharmacol 32:84-96.
4
5 Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
6 11:579-583.
7
8 Seliger, B. (2005) Strategies of tumor immune evasion. BioDrugs 19:347-354.
9
10 Seniori Costantini, A; Miligi, L; Kreibel, D; et al. (2001) A multicenter case-control study in Italy on
11 Hematolymphopoietic neoplasms and occupation. Epidemiology 12:78-87.
12
13 Seppalainen, AM; Antti-Poika, M. (1983) Time course of electrophysiological findings for patients with solvent
14 poisoning: adescriptive study. ScandJ Work Environ Health 9:15-24.
15
16 Shafer, TJ; Bushnell, PJ; Benignus, VA; et al. (2005) Perturbation of voltage-sensitive Ca2+ channel function by
17 volatile organic solvents. 129. J Pharmacol Exp Ther 315:1109-1118.
18
19 Sharanjeet-Kaur; Mursyid, A; Kamaruddin, A; et al. (2004) Effect of petroleum derivatives and solvents on colour
20 perception. ClinExp Optom 87:339-343.
21
22 Sharpe, CR; Rochon, JE; Adam, JM; et al. (1989) Case-control study of hydrocarbon exposures in patients with
23 renal cell carcinoma. CMAJ 140:1309-1318.
24
25 Shawn, SH; Robins, JM. (1986) Comment. J Am StatAssoc 84:604-609.
26
27 Sheldon, L; Handy, R; Hartwell, W; et al. (1985) Human exposure assessment to environmental chemicals: nursing
28 mothers study. Final report. Available from: Research Triangle Institute, Research Triangle Park, NC.
29
30 Sher, T; Yi, HF; McBride, OW; et al. (1993) CDNA cloning, chromosomal mapping, and functional characterization
31 of the human peroxisome proliferator activated receptor. Biochemistry 32:5598-5604.
32
33 Shu, XO; Steward, P; Wen, WO; et al. (1999) Parental occupational exposure to hydrocarbons and risk of acute
34 lymphocytic leukemia in offspring. Cancer Epidemiol Biomarkers Prev 8(9):783-791.
35
36 Silverman, DT; Levin, LI; Hoover, RN; et al. (1989a) Occupational risks of bladder cancer in the United States: I.
37 Whitemen. JNatl Cancer Inst 81:1472-1480.
38
39 Silverman, DT; Levin, LI; Hoover, RN. (1989b) Occupational risks of bladder cancer in the United States: II
40 Nonwhitemen. JNatl Cancer Inst 81:1480-1483.
41
42 Simpson ,AE; Brammar, WJ; Pratten, MK; et al. (1995) Translactational induction of CYP4A expression in 10.5-
43 day neonatal rats by the hypolipidemic drug clofibrate. Biochem Pharmacol. 50:2021-2032.
44
45 Simpson, AE; Brammar, WJ; Pratten, MK; et al. (1996) Placenta! transfer of the hypolipidemic drug, clofibrate,
46 induces CYP4A expression in 18.5-day fetal rats. Drug Metab Disp 24:547-554.
47
48 Singh, M; Sinha, U. (1976). Chloral hydrate induced haploidization mAspergillus nidulans. Experientia
49 32:114-115.
50
51 Singh, M; Sinha, U. (1979). Mitotic haploidization and growth ofAspergillus nidulans on media containing chloral
52 hydrate. JCytol Genet 14:1-4.
53
54 Skowron, J; Miranowicz-Dzierzawska, K; Zapor, L; et al. (2001) Interactions of some organic solvents:
55 hydrocarbons and chloroalkene. Int J Occup Saf Ergon 7(l):35-47.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-269 DRAFT-DO NOT CITE OR QUOTE
-------
1 Smith, EM; Miller, ER; Woolson, RF; et al. (1985) Bladder cancer risk among laundry workers, dry cleaners, and
2 others in chemically-related occupations. JOccupMed 27:295-297.
3
4 Smith, MK; Randall, JL; Read, EJ; et al. (1989) Teratogenic activity of trichloroacetic acid in the rat. Teratology
5 40:445-451.
6
7 Sokol, L; Loughran, TP. (2006) Large granular lymphocyte leukemia. The Oncologist 11:263-273.
8
9 Solet, D; Robins, TG. (1991) Renal function in dry cleaning workers exposed to perchloroethylene. Am J Ind Med
10 20:601-614.
11
12 Solet, D; Robins, TG; Sampaio, C. (1990) Perchloroethylene exposure assessment among dry cleaning workers. Am
13 Ind HygAssoc 151:566-574.
14
15 Solleveld, HA; Haseman, JK; McConnell, EE. (1984) National history of body weight gain, survival, and neoplasia
16 in the F344 rat. J Natl Cancer Inst 72:929-940.
17
18 Sonnenfeld, N; Hertz-Picciotto, I; Kaye, WE. (2001) Tetrachloroethylene in drinking water and birth outcomes at
19 the US Marine Corps base at Camp Lejeune, NC. Am J Epidemiol 154:902-908.
20
21 Sora, S; Agostini-Carbone, ML. (1987) Chloral hydrate, methylmercury hydroxide and ethidium bromide affect
22 chromosomal segregation during meiosis of Saccharomyces cervisiae. Mutat Res. 190: 13-17.
23
24 Sparrow, GP. (1977) A connective tissue disorder similar to vinyl chloride disease in a patient exposed to
25 perchlorethylene. ClinExpDermatol2:17-22.
26
27 Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
28 of subjects chronically exposed to organic solvents (II)]. G Ital Med Lav Ergon 19:85-88.
29
30 Spirtas, R; Stewart, PA; Lee, JS; et al. (1991) Retrospective cohort mortality study of workers at an aircraft
31 maintenance facility. I. Epidemiological results. Br J Ind Med 48:515-530.
32
3 3 Stauber, AJ; Bull, RJ. (1997) Differences in phenotype and cell replicative behavior of hepatic tumors induced by
34 dichloroacetate (DCA) and trichloroacetate (TCA). Toxicol Appl Pharmacol 144:235-246.
35
36 Stauber, A; Bull, RJ; Thrall, BD. (1998) Dichloroacetate and trichloroacetate promote clonal expansion of
37 anchorage-independent hepatocytes in vivo and in vitro. Toxicol Appl Pharmacol 260:287-294.
38
39 Stemhagen, A; Slade, J; Altaian, R; et al. (1983) Occupational risk factors and liver cancer. A retrospective case-
40 control study of primary liver cancer in New Jersey. Am J Epidemiol 117:443-454.
41
42 Sterchele, PF; Sun, H; Peterson, RE; et al. (1996) Regulation of peroxisome proliferator-activated receptor-alpha
43 mRNA in rat liver. Arch Biochem Biophys 326:281-289.
44
45 Stevenson, DE; McClain, RM; Popp, JA; et al. (1990) Mouse liver carcinogenesis: mechanisms and species
46 comparisons. New York: Wiley-Liss.
47
48 Stewart, RD; Dodd, HC. (1964) Absorption of carbon tetrachloride, trichloroethylene, tetrachloroethylene,
49 methylene chloride and 1,1,1-trichloroethane through the human skin. J Ind Hyg 25:429-446.
50
51 Stewart, RD; Erley, DS; Schaeffer, AW; et al. (1961) Accidental vapor exposure to anesthetic concentrations of a
52 solvent containing tetrachloroethylene. Indus Med Sug 30:327-330.
53
54 Stewart, RD; Baretta, ED; Dodd, HC; et al. (1970) Experimental human exposure to tetrachloroethylene. Arch
55 Environ Health 20:225-229.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-270 DRAFT-DO NOT CITE OR QUOTE
-------
1 Stewart, RO; Hake, CL; Wu, A; et al. (1977) Effects of perchloroethylene drag interactions on behavior and
2 neurological function. DHEW (NIOSH) Publ. No. 77-191. Department of Health, Education, and Welfare.
3 Washington, DC.
4
5 Stonard, MD; Phillips, PG; Foster, JR; et al. (1986a) Alpha 2u-globulin: measurement in rat kidney and relationship
6 to hyaline droplets. Clin Chim Acta 160:197-203.
7
8 Storer, RD; McKelvey, TW; Kraynak, AR; et al. (1996) Revalidation of the in vitro alkaline elution/rat hepatocyte
9 assay for DNA damage: improved criteria for assessment of cytotoxicity and genotoxicity and results for 81
10 compounds. Mutat Res 368:59-101.
11
12 Storm, JE; Mazor, KA. (2004) Update of residential tetrachloroethylene exposure and decreases in visual contrast
13 sensitivity. Environ Health Perspect 112(15):A862-863.
14
15 Story, DL; Meierhenry, EF; Tyson, CA; et al. (1986) Differences in rat liver enzyme-altered foci produced by
16 chlorinated aliphatics and phenobarbital. Toxicol Ind Health 2:351-362.
17
18 Stott, WT. (1988). Chemically induced proliferation of peroxisomes: Implications to risk assessment. Regul Toxicol
19 Pharmacol 8:125-159.
20
21 Stromberg, PC. (1985) Large granular lymphocyte leukemia in F344 rats: model for human T gamma lymphoma,
22 malignant histiocytosis, and T-cell chronic lymphocytic leukemia. Am J Pathol 119:517-519.
23
24 Styles, JA; Wyatt, I; Coutts, C. (1991) Trichloroacetic acid: studies on uptake and effects on hepatic DNA and liver
25 growth in mouse. Carcinogenesis 12:1715-1719.
26
27 Suarez, L; Weiss, NS; Martin, J. (1989) Primary liver cancer death and occupation in Texas. Am J Ind Med 15:167-
28 175.
29
30 Subramoniam, A; Goel, SK; Pandya, KP; et al. (1989) Influence of trichloroethylene treatment on phosphoinositides
31 in rat brain. Toxicol Lett 49:55-60.
32
33 Sverdrap, B; Kallberg, H; Bengtsson, C; et al. (2005) Association between occupational exposure to mineral oil and
34 rheumatoid arthritis: results from the Swedish EIRA case-control study. Arthritis Res Ther 7:R1296-R1303.
35
36 Swenberg, JA; Lehman-McKeeman, LD. (1999) A 2 Urinary-globulin-associated nephropathy as a mechanism of
37 renal tubule cell carcinogenesis in male rats. In: Capen, C; Dybing, E; Rice, J; et al.; eds. Species differences in
3 8 thyroid, kidney and urinary bladder carcinogenesis. Lyon: IARC Scientific Publications No 147; pp 95-118.
39
40 Swenberg, JA; Short, B; Borghoff, S; et al. (1989) The comparative pathobiology of alpha 2u-globulin nephropathy.
41 Toxicol Appl Pharmacol 97:35-46.
42
43 Szakmary, E; Ungvary, G; Tatrai, E. (1997) The offspring-damaging effect of tetrachloroethylene in rats, mice, and
44 rabbits. Central Eur J Occup Environ Med 3 (1): 31 -3 9.
45
46 Tang, ML; Kemp, AS; Thorburn, J; et al. (1994) Reduced interferon-gamma secretion in neonates and subsequent
47 atopy. Lancet 344:983-985.
48
49 Tanios, MA; El, GH; Rosenberg, B J; et al. (2004) Can we still miss tetrachloroethylene-induced lung disease? The
5 0 emperor returns in new clothes. Respiration 71:642-645.
51
52 Tao, L; Kramer, PM; Ge, R; et al. (1998) Effect of dichloroacetic acid and trichloroacetic acid on DNA methylation
53 in liver and tumors of female B6C3F1 mice. Toxicol Sci 43:139-144.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-271 DRAFT-DO NOT CITE OR QUOTE
-------
1 Tao, L; Yang, S; Xie, M; et al. (2000) Hypomethylation and overexpression of c-jun and c-myc protooncogenes and
2 increased DNA methyltransferase activity in dichloroacetic and trichloroacetic acid-promoted mouse liver tumors.
3 Cancer Lett 158:185-193.
4
5 Taskinen, H; Anttila, A; Lindbohm, ML; et al. (1989) Spontaneous abortions and congenital malformations among
6 the wives of men occupationally exposed to organic solvents. Scand J Work Environ Health 15:345-352.
7
8 Tateishi, T; Nakura, H; Asoh, M; et al. (1997) A comparison of hepatic cytochrome P450 protein expression
9 between infancy and postinfancy. Life Sciences 61(21):2567-74.
10
11 Teschke, K; Morgan, MS; Checkoway, H; et al. (1997) Surveillance of nasal and bladder cancer to locate sources of
12 exposure to occupational carcinogens. Occup Environ Med 54:443 -451.
13
14 Thangada, S; Alvares, K; Mangino, M; et al. (1989) An in vitro demonstration of peroxisome proliferation and
15 increase in peroxisomal beta-oxidation system mRNAs in cultured rat hepatocytes treated with ciprofibrate. FEES
16 Lett 250:205-210.
17
18 Thomas, KW; Pellizzari, ED; Perritt, RL; et al. (1991) Effect of dry-cleaned clothes on tetrachloroethylene levels in
19 indoor air, personal air, and breath for residents of several New Jersey homes. J Expo Anal Environ Epidemiol
20 1:475-490.
21
22 Till, C; Koren, G; Rovet, JF. (2001a) Prenatal exposure to organic solvents and child neurobehavioral performance.
23 Neurotoxicol Teratol 23(3):235-245.
24
25 Till, C; Westall, CA; Rovet, JF; et al. (200 Ib) Effects of maternal occupational exposure to organic solvents on
26 offspring visual functioning: a prospective controlled study. Teratology 64(3):134-141.
27
28 Till C; Rovet JF; Koren G; et al. (2003) Assessment of visual functions following prenatal exposure to organic
29 solvents. Neurotoxicology 24(4-5):725-31.
30
31 Till, C; Westall, CA; Koren, G; et al. (2005) Vision abnormalities in young children exposed prenatally to organic
32 solvents. Neurotoxicology 26:599-613.
33
34 Tinston, DJ. (1994) Perchloroethylene: A multigeneration inhalation study in the rat. CTL/P/4097. Available from:
35 IRIS Information Desk, U.S. Environmental Protection Agency.
36
37 Tomatis, L; Aitio, A; Wilbourn, J; et al. (1989) Human carcinogens so far identified. Jpn J Cancer Res 80:795-807.
38
39 Tontonoz, P; Hu, E; Spiegelman, DM. (1994) Stimulation of adipogenesis in fibroblasts by PPARgamma2, alipid-
40 activated transcription factor. Cell 79:1147-1156.
41
42 Topping, MD; Forster, HW; Dolman, C; et al. (1986) Measurement of urinary retinol-binding protein by enzyme-
43 linked immunosorbent assay, and its application to detection of tubular proteinuria. ClinChem 32:1863-1866.
44
45 Toraason, M; Clark, J; Dankovic, D; et al. (1999) Oxidative stress and DNA damage in Fischer rats following acute
46 exposure to trichloroethylene or perchloroethylene [in process citation]. Toxicology 138:43-53.
47
48 Toraason, M; Butler, MA; Ruder, A; et al. (2003) Effect of perchloroethylene, smoking, and race on oxidative DNA
49 damage in female dry cleaners. Mutat Res 539:9-18.
50
51 Travier N; Gridley, G; De Roos, AJ; et al. (2002) Cancer incidence of dry cleaning, laundry, and ironing workers in
52 Sweden. Scand J Work Environ Health 28:341-348.
53
54 Trevisan, A; Cristofori, P; Fanelli, G. (1999) Glutamine synthetase activity in rat urine as sensitive marker to detect
55 S3 segment-specific injury of proximal tubule induced by xenobiotics. Arch Toxicol 73:255-262.
56
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-272 DRAFT-DO NOT CITE OR QUOTE
-------
1 Trevisan, A; Macca, I; Rui, F; et al. (2000) Kidney and liver biomarkers in female dry-cleaning workers exposed to
2 perchloroethylene. Biomarkers 5:399-409.
3
4 Tugwood, JD; Issemann, I; Anderson, RG; et al. (1992) The mouse peroxisome proliferator activated receptor
5 recognizes a response element in the 5* flanking region of the rat acyl CoA oxidase gene. EMBO 11:433-439.
6
7 Tugwood, JD; Aldridge, TC; Lambe, KG; et al. (1996) Peroxisome proliferator-activated receptors: structures and
8 function. Ann NY Acad Sci 804:252-265.
9
10 U.S. EPA (Environmental Protection Agency). (1980) Ambient water quality criteria for tetrachloroethylene. Office
11 of Water, Washington, DC; EPA 440/5-80-073. Available from: National Technical Information Service,
12 Springfield, VA.
13
14 U.S. EPA (Environmental Protection Agency). (1985a) Health assessment document for tetrachloroethylene
15 (perchloroethylene). Office of Health and Environmental Assessment, Office of Research and Development,
16 Washington, DC; EPA/600/8-82/005F. Available from: National Technical Information Service, Springfield, VA.
17 PB-85-249696/AS.
18
19 U.S. EPA (Environmental Protection Agency). (1985b) Drinking water criteria document for tetrachloroethylene.
20 Available from: National Technical Information Service, Springfield, VA; PB86-118114.
21
22 U.S. EPA (Environmental Protection Agency), (1985c) Chemical carcinogens: review of the science and its
23 associated principles. Office of Science and Technology Policy. Federal Register 50:10372-10442.
24
25 U.S. EPA (Environmental Protection Agency). (1986a) Addendum to the health assessment document for
26 tetrachloroethylene (perchloroethylene). Prepared by the Office of Health and Environmental Assessment,
27 Washington, DC; EPA/600/8-82/005FA. Available from: National Technical Informtion Service, VA; PB-86-
28 174489/AS.
29
30 U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for carcinogen risk assessment. Federal Register
31 51(185):34014-34025.
32
33 U.S. EPA (Environmental Protection Agency). (1987a) Evaluation of the carcinogenicity of unleaded gasoline.
34 Office of Health and Environmental Assessment, Office of Research and Development, Washington, DC;
35 EPA/600/8-87/001. Available from: National Technical Information Service, Springfield, VA; PB-87-186151/AS.
36
37 U.S. EPA (Environmental Protection Agency). (1987b) Health assessment document for beryllium. Office of
3 8 Health and Environmental Assessment, Office of Research and Development, Washington, DC; EP A/600/8-84/026F.
39 (Pages 7-84.) Available from: National Technical Information Service, Springfield, VA.
40
41 U.S. EPA (Environmental Protection Agency). (1991a) Response to issues and the data submissions on the
42 carcinogenicity of tetrachloroethylene (perchloroethylene). Office of Health and Environmental Assessment,
43 Washington, DC; EP A/600/6-91/002F. Available from: National Technical Information Service, Springfield, VA.
44
45 U.S. EPA (Environmental Protection Agency). (1991b) Alpha-2u-globulin: association with chemically induced
46 renal toxicity and neoplasia in the male rat. Risk Assessment Forum, Office of Research and Development,
47 Washington, DC; EPA/625/3-91/0IF. Available from: National Technical Information Service, Springfield, VA.
48
49 U.S. EPA (Environmental Protection Agency). (1997a) Chemical and radiation leukemogenesis in humans and
5 0 rodents and the value of rodent models for assessing risks of lymphohematopoietic cancers. Office of Research and
51 Development, Washington, DC; EPA/600/R-97/090. Available from: National Technical Information Service,
52 Springfield, VA; PB97-208185.
53
54 U.S. EPA (Environmental Protection Agency). (1997) Exposure factors handbook. Vol 1. National Center for
55 Environmental Assessment, Office of Research and Development, Washingotn, DC; EPA/600/P-95/002Fa.
56 Available online at http://www.epa.gov/ncea.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-273 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 U.S. EPA (Environmental Protection Agency). (2000a) Toxicological review of chloral hydrate. National Center for
3 Environmental Assessment, Washington, DC; EPA/635/R-00/006. Available online at http://www.epa.gov/iris.
4
5 U.S. EPA (Environmental Protection Agency). (2000b) Toxicological review of vinyl chloride. Integrated Risk
6 Information System (IRIS). National Center for Environmental Assessment, Washington, DC; EPA/635/R-00/004.
7 Available online at http://www.epa.gov/iris.
8
9 U.S. EPA (Environmental Protection Agency). (200la) Trichloroethylene health risk assessment: synthesis and
10 characterization [external review draft]. National Center for Environmental Assessment, Office of Research and
11 Development, Washington, DC; EPA/600/P-01/002A. Available online at
12 http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=4580.
13
14 U.S. EPA (Environmental Protection Agency). (200Ib) Sources, emission and exposure for trichloroethylene (TCE)
15 and related chemicals. National Center for Environmental Assessment, Washington, DC; EPA/600/R-00/099.
16
17 U.S. EPA (Environmental Protection Agency). (2002) Toxicological review of vinyl chloride. National Center for
18 Environmental Assessment, Washington, DC; EPA/635/R-00/004. Available online at http://www.epa.gov/iris.
19
20 U.S. EPA (Environmental Protection Agency). (2002) Child-specific exposure factors handbook. Interim report.
21 EPA-600-P-00-002B. U.S. Environmental Protection Agency, Washington, DC.
22
23 U.S. EPA (Environmental Protection Agency). (2003a) IRIS summary of toxicological effects of dichloroacetic acid.
24 National Center for Environmental Assessment, Washington, DC. Available online at http://www.epa.gov/iris.
25
26 U.S. EPA. (Environmental Protection Agency). (2003b) Neurotoxicity of tetrachloroethylene (perchloroethylene):
27 discussion paper. National Center for Environmental Assessment, Washington, DC; EPA/600/P-03/005 A.
28 Available online at http://www.epa.gov/ncea.
29
30 U.S. EPA (Environmental Protection Agency). (2003c) Framework for cumulative risk assessment. Risk
31 Assessment Forum, Washington, DC; EPA/630/P-02/001F. Available online at http://www.epa.gov/ncea/raf.
32
33 U.S. EPA (Environmental Protection Agency). (2003d) Proposed science policy: PPAR-aagonist-mediated
34 hepatocarcinogenesis in rodents and relevance to human health risk assessment (available online at
35 http://www.epa.gov/scipolv/sap/meetings/2003/december9/peroxisomeproliferatorsciencepolicvpaper.pdf) and
36 minutes of FIFRA Scientific Advisory Panel Meeting, Arlington, VA (available online at
37 http://www.epa.gov/oscpmont/sap/2003/december9/meetingminutes.pdf).
38
39 U.S. EPA (Environmental Protection Agency). (2004) Summary report of the peer review workshop on the
40 neurotoxicity of tetrachloroethylene (perchloroethylene) discussion paper. National Center for Environmental
41 Assessment, Washington, DC; EPA/600/R-04/041. Available online at http://www.epa.gove/ncea.
42
43 U.S. EPA (Environmental Protection Agency). (2005a) Trichloroethylene issue paper 3: role of peroxisome
44 proliferator-activated receptor agonism and cell signaling in trichloroethylene toxicity. National Center for
45 Environmental Assessment, Washington, DC; EPA/600/R-05/024. Available online at
46 http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=438646.
47
48 U.S. EPA (Environmental Protection Agency). (2005b) Guidelines for carcinogen risk assessment. Federal Register
49 70(66)17765-18717. Available online at http://www.epa.gov/cancerguidelines.
50
51 U.S. EPA (Environmental Protection Agency). (2005c) Supplemental guidance for assessing cancer susceptibility
52 from early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
53 online at http://www.epa.gov/cancerguidelines.
54
55 U.S. EPA (Environmental Protection Agency). (2005d) Human testing; proposed plan and description of review
56 process. Federal Register 70(25):6661-6667.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-274 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 U.S. EPA (Environmental Protection Agency). (2006) Aging and toxic response: issues relevant to risk assessment.
3 EPA/600/P-03/004A. Available online at http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=156648.
4
5 UAREP (Universities Associated for Research and Education in Pathology). (1983) Hydrocarbon toxicity; acute,
6 subchronic and chronic effects in relation to unleaded gasoline exposure of rodents with comments on the
7 significance to human health. Contract No. PS-6 UAREP (504-2) with the American Petroleum Institute. Bethesda,
8 MD.
9
10 Uetrecht, J; Zahid, N; Rubin, R. (1988) Metabolism of procainamide to a hydroxylamine by human neutrophils and
11 mo no nuclear leukocytes. Chem Res Toxicol 1:74-78.
12
13 Ulm, K; Henschler, D; Vamvakas, S. (1996) Occupational exposure to perchloroethylene. Cancer Causes Control
14 7:284-286.
15
16 Umezu, T; Yonemoto, J; Soma, Y; et al. (1997) Behavioral effects of trichloroethylene and tetrachloroethylene in
17 mice. Pharmacol Biochem Behav 58:665-671.
18
19 Valles, EG; Laughter, AR; Dunn, CS; et al. (2003) Role of the peroxisome proliferator-activated receptor alpha in
20 responses to diisononyl phthalate. Toxicology 191:211-225.
21
22 Valic, E; Waldhor, T; Konnaris, C; et al. (1997) Acquired dyschromatopsia in combined exposure to solvents and
23 alcohol. Int Arch Occup Environ Health 70:403-406.
24
25 Vamvakas, S; Dekant, W; Berthold, K; et al. (1987) Enzymatic transformation of mercapturic acids derived from
26 halogenated alkenes to reactive and mutagenic intermediates. Biochem Pharmacol 36:2741-2748.
27
28 Vamvakas, S; Dekant, W; Henscher, D. (1989a) Genotoxicity of haloalkene and haloalkane glutathione S-
29 conjugates in porcine kidney cells. Toxicol In Vitro 3:151-156.
30
31 Vamvakas, S; Herkenhoff, M; Dekant, W; et al. (1989b) Mutagenicity of tetrachloroethene in the Ames test--
32 metabolic activation by conjugation with glutathione. J Biochem Toxicol 4:21-27.
33
34 Vamvakas, S; Dekant, W; Henschler, D. (1989c) Assessment of unscheduled DNA synthesis in a cultured line of
3 5 renal epithelial cells exposed to cysteine S-conjugates of haloalkenes and haloalkanes. Mutat Res 222:329-335.
36
37 Vamvakas, S; Kochling, A; Berthold, K; et al. (1989d) Cytotoxicity of cysteine S-conjugates: structure-activity
38 relationships. Chem Biol Interact 71:79-90.
39
40 Vartiainen, T; Pukkala, E; Strandman, T; et al. (1993) Population exposure to tri- and tetrachloroethylene and cancer
41 risk: two cases of drinking water pollution. Chemosphere 27:1171-1181.
42
43 Vaughan, TL; Stewart, PA; Davis, S; et al. (1997) Work in dry cleaning and the incidence of cancer of the oral
44 cavity, larynx, and oesophagus. Occup Environ Med 54:692-695.
45
46 Verplanke, AJ; Leummens, MH; Herber, RF. (1999) Occupational exposure to tetrachloroethene and its effects on
47 the kidneys. J Occup Environ Med 41:11-16.
48
49 Vieira, I; Sonnier, M; Cresteil, T. (1996) Developmental expression of CYP2E1 in the human liver:
50 hypermethylation control of gene expression during the neonatal period. Eur J Biochem 238(2):476-83.
51
52 Vieira, V; Aschengrau, A; Ozonoff, D. (2005) Impact of tetrachloroethylene-contaminated drinking water on the
53 risk of breast cancer: Using a dose model to assess exposure in a case-control study. Environ Health 4:3.
54
55 Vingrys, AJ; King-Smith, PE. (1988) A quantitative scoring technique for panel tests of color vision. Invest
56 Ophthalmol Vis Sci 29:50-53.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-275 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 Vohl, MC; Lepage, P; Gaudet, D; et al. (2000) Molecular scanning of the human PPAR-a gene. Association of the
3 1162v mutation with hyperapobetalipoproteinemia. J Lipid Res 41:945-952.
4
5 Volkel, W; Friedewald, M; Lederer, E; et al. (1998) Biotransformation of perchloroethene: dose-dependent
6 excretion of trichloroacetic acid, dichloroacetic acid, and N-acetyl-S- (trichlorovinyl)-L-cysteine in rats and humans
7 after inhalation. Toxicol Appl Pharmacol 153:20-27.
8
9 Vyskocil, A; Emminger, S; Tejral, J; et al. (1990) Study on kidney function in female workers exposed to
10 perchloroethylene. Hum Exp Toxicol 9:377-380.
11
12 WHO, IPCS Tetrachloroethene (Cicads 68, 2006). World Health Organization, International Programme on
13 Chemical Saftey. Available online at http://www.inchem.org/documents/cicads/cicads/cicad68.htm.
14
15 Walgren, JE; Kurtz, DT; McMillan, JM. (2000a) The effect of the trichloroethylene metabolites trichloroacetate and
16 dichloroacetate on peroxisome proliferation and DNA synthesis in cultured human hepatocytes. Cell Biol Toxicol
17 16:257-273.
18
19 Walgren, JE; Kurtz, DT; McMillan, JM. (2000b) Expression of PPAR(-a) in human hepatocytes and activation by
20 trichloroacetate and dichloroacetate. Res Commun Mol Pathol Pharmacol 108:116-132.
21
22 Walgren, JL; Jollow, DJ; McMillan, JM. (2004) Induction of peroxisome proliferation in cultured hepatocytes by a
23 series of halogenated acetates. Toxicol 197(3)189-197.
24
25 Walker, JT; Burnett, CA; Lalich, NR; et al. (1997) Cancer mortality among laundry and dry cleaning workers. Am J
26 IndMed 32:614-619.
27
28 Walles, SA. (1986) Induction of single-strand breaks in DNA of mice by trichloroethylene and tetrachloroethylene.
29 Toxicol Lett 31:31-35.
30
31 Wang, S; Karlsson, JE; Kyrklund, T; et al. (1993) Perchloroethylene-induced reduction in glial and neuronal cell
32 marker proteins in rat brain. Pharmacol Toxicol 72:273-278.
33
34 Wang, JL; Chen, WL; Tsai, SY; et al. (2001) An in vitro model for evaluation of vaporous toxicity of
3 5 trichloroethylene and tetrachloroethylene to CHO-K1 cells. Chem Biol Interact 137:139-154.
36
37 Ward, JM; Griesemer, RA; Weisburger, EK. (1979) The mouse liver tumor as an endpoint in carcinogenesis tests.
38 Toxicol Appl Pharmacol 51:389-397.
39
40 Warner, JA; Miles, EA; Jones, AC; et al. (1994) Is deficiency of interferon gamma production by allergen triggered
41 cord blood cells apredictor of atopic eczema? 103. ClinExp Allergy 24:423-430.
42
43 Warren, DA; Reigle, TG; Muralidhara, S; et al. (1996) Schedule-controlled operant behavior of rats following oral
44 administration of perchloroethylene: time course and relationship to blood and brain solvent levels. J Toxicol
45 Environ Health 47:345-362.
46
47 Wartenberg, D; Reyner, D; Scott, CS. (2000) Trichloroethylene and cancer: epidemiologic evidence. Environ
48 Health Perspect 108 Suppl 2:161-176.
49
50 Watanabe, K; Fuji, H; Takahashi, T; et al. (2000) Constitutive regulation of cardiac fatty acid metabolism through
51 peroxisome proliferator-activated receptor -a associated with age-dependent cardiac toxicity. J Biol Chem
52 275:22293-22299.
53
54 Weber, G; Lea, MA. (1966) The molecular correlation concept of neoplasia. Adv Enzyme Regul 4:115-145.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-276 DRAFT-DO NOT CITE OR QUOTE
-------
1 Weisenburger, DD. (1992) Pathological classification of non-Hodgkin's lymphoma for epidemiological studies.
2 Cancer Res 52:5456s-5462s.
3
4 Weiss, NS. (1995) Cancer in relation to occupational exposure to perchloroethylene. Cancer Causes Control
5 6:257B66.
6
7 White, IN; Razvi, N; Gibbs, AH; et al. (2001) Neoantigen formation and clastogenic action of HCFC-123 and
8 perchloroethylene in human MCL-5 cells. Toxicol Lett 124:129-138.
9
10 Wick, MJ; Mihic, SJ; Ueno, S; et al. (1998) Mutations of gamma-aminobutyric acid and glycine receptors change
11 alcohol cutoff: evidence for an alcohol receptor? Proc Natl Acad Sci U S A 95:6504-6509.
12
13 Williams, RL; Creasy, RK; Cunningham, GC; et al. (1982) Fetal growth and perinatal viability in California. Obstet
14 Gynecol 59:624-632.
15
16 Windham, GC; Shusterman, D; Swan, S; et al. (1991) Exposure to organic solvents and adverse pregnancy outcome.
17 Am J of IndMed 20:241-259.
18
19 Yamakawa-Kobayashi, K; Ishiguro, H; Arinami, T; et al. (2002) A Val227Ala polymorphism in the peroxisome
20 proliferator activated receptor -a (PPAR-a) gene is associated with variations in serum lipid levels. J Med Genet
21 39:189-191.
22
23 Yamakura, T; Mihic, SJ; Harris, RA. (1999) Amino acid volume and hydropathy of a transmembrane site determine
24 glycine and anesthetic sensitivity of glycine receptors. J Biol Chem 274:23006-23012.
25
26 Yang, Q.; Ito S; Gonzalez, FJ. (2007). Hepatocyte-restricted constitutive activation of PPAR-a induces
27 hepatoproliferationbut not hepatocarcinogenesis. Carcinogenesis 28(6): 1171-7.
28
29 Yoon, JS; Mason JM; Valencia R; et al. (1985) Chemical mutagenesis testing in Drosophila. IV. Results of 45
30 coded compounds tested for the National Toxicology Program. Environ Mutagen 7: 349-367.
31
32 Yount, EA; Felten, SY; O=Connor, BL; et al. (1982) Comparison of the metabolic and toxic effects of 2-
33 chloropropionate and dichloroacetate. J Pharmacol Exp Ther 222:501-508.
34
3 5 Youssef, J; Badr, M. (1999) Biology of senescent liver peroxisomes: role in hepatocellular aging and disease.
36 Environ Health Perspect 107:791-797.
37
38 Yu, S; Cao, WQ; Kashireddy, P; et al. (2001) Human peroxisome proliferator-activated receptor a (PPARa) supports
39 the induction of peroxisome proliferation in PPARa deficient mouse liver. J Biol Chem 276:42485-42491.
40
41 Zanelli, U; Puccini, P; Acerbi, D; et al. (1996) Induction of peroxisomal beta-oxidation and P-450 4A-dependent
42 activities by pivalic and trichloroacetic acid in rat liver and kidney. Arch Toxicol 70:145-149.
43
44 Zhang, B; Marcus, SL; Sajjadi, FG; et al. (1992) Identification of a peroxisome proliferator-responsive element
45 upstream of the gene encoding rat peroxisomal enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase.
46 Proceedings of the National Academy of Sciences of the US A 89:7541 -7545.
47
48 Zhang, B; Marcus, S; Miyata, K; et al. (1993) Characterization of protein-DNA interactions within the peroxisome
49 proliferator-responsive element of the rat hydratase-dehydrogenase gene. J Biol Chem 268:12939-12945.
50
51 Zhou, YC; Waxman, DJ. (1998) Activation of peroxisome proliferator-activated receptors by chlorinated
52 hydrocarbons and endogenous steroids. Environ Health Perspect 106(suppl 4):983-988.
53
54 Zingg, JM: Jones, PA. (1997) Genetic and epigenetic aspects of DNA methylation on genome expression, evolution,
55 mutation and carcinogenesis. Carcinogenesis 18:869-882.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 4-277 DRAFT-DO NOT CITE OR QUOTE
-------
1 5. DOSE-RESPONSE EVALUATION
2
O
4 5.1. INHALATION REFERENCE CONCENTRATION (RfC)
5 Although the RfD is commonly presented first in the IRIS toxicological reviews, the RfC
6 is presented in Section 5.1 and the RfD in Section 5.2 because the available data were primarily
7 from inhalation exposure and pharmacokinetic modeling was available to carry out route-to-
8 route extrapolation of the RfC to the oral route of exposure.
9 The RfCl for tetrachloroethylene is derived through a process of (1) considering all
10 studies and selecting the adverse health effects that occur at the lowest exposure concentration,
11 (2) selecting the point of departure (POD)2 at which the adverse health effect either is not
12 observed or would occur at a relatively low prevalence (e.g., 10%), (3) deriving the POD in
13 terms of the human equivalent concentration (HEC), and (4) reducing this exposure
14 concentration by uncertainty factors (UFs) to account for uncertainties in the extrapolation from
15 the study conditions to an estimate of human environmental exposure. This is EPA's first
16 attempt to define a tetrachloroethylene RfC for IRIS. Health assessments from other agencies,
17 more fully described in Appendix A, have included a criterion for noncarcinogenic effects
18 associated with inhalation exposure based on neurotoxic effects observed in human
19 epidemiologic studies (NYS DOH, 1997; ATSDR, 1997).
20
21 5.1.1. Choice of Principal Study and Critical Effect
22 The database of human and animal studies on inhalation toxicity of tetrachloroethylene is
23 adequate to support derivation of inhalation reference values. A number of targets of toxicity
24 from chronic exposure to tetrachloroethylene, include the nervous system, liver, kidney,
25 reproductive system, and developing fetus, with published reports in both animals and humans.
26 Greatest consideration is given to human data, if adequate, to develop an RfC.
27 Neurological effects were judged to be associated with lower tetrachloroethylene
28 concentrations. This finding is in agreement with Rao and Brown (1993), who, using categorical
29 analysis methods, identified neurological effects as the most sensitive noncancer toxicity
The RfC is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to
the human population (including sensitive subgroups) that is likely to be without appreciable risk of deleterious
effects during a lifetime. It can be derived from a NOAEL, a LOAEL, or a benchmark concentration, with
uncertainty factors generally applied to reflect limitations of the data used. It is generally used in EPA's noncancer
health assessments. The RfC, like the RfD for oral exposure, is based on the assumption that thresholds exist for
certain toxic effects, such as liver pathology, but may not exist for other toxic effects, such as carcinogenicity.
2
The POD denotes a dose at the lower end of the observed dose-response curve where extrapolation to
lower doses begins. For effects other than cancer, the POD is either a NOAEL, a LOAEL if no NOAEL can be
identified, or a modeled point (e.g., a BMCL10 or an LED10) if the data are suitable for dose-response modeling.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-1 DRAFT-DO NOT CITE OR QUOTE
-------
1 endpoint. A number of studies assessing neurobehavioral effects in both humans and rodents are
2 available for RfC analysis. The epidemiologic body of evidence is characterized by studies that
3 used standardized neurobehavioral batteries. In addition, some studies employed assessment of
4 visual function (Cavalleri et al., 1994; Schreiber et al., 2002; Echeverria et al., 1994, 1995), a
5 neurological outcome known to be sensitive to volatile organic compounds. Most
6 epidemiological studies have examined occupational exposure to tetrachloroethylene. Two
7 epidemiological studies examined residential exposure to tetrachloroethylene (Altmann et al.,
8 1995; Schreiber et al., 2002). Together, the epidemiologic evidence supports an inference of a
9 broad range of cognitive, behavioral, and visual functional deficits following tetrachloroethylene
10 exposure (U.S. EPA, 2004).
11 The research in animal models on the effects of tetrachloroethylene on functional
12 neurological endpoints consists of screening studies (functional observation battery, motor
13 activity) or effects on sensory system function as assessed by evoked potential. Some
14 consistency is seen in the animal models, with effects on motor activity and motor function
15 following exposure to tetrachloroethylene in either the adult or the developmental period,
16 changes in evoked potentials following acute and subchronic exposures, and replication of
17 observed alterations in brain DNA, RNA, or protein levels and brain weight changes.
18 Of the studies discussed in Chapter 4, a number of epidemiologic studies of neurological
19 effects in either occupational workers or residential subjects with tetrachloroethylene exposure
20 are considered for the principal study with which to identify the POD. No single epidemiologic
21 study stands out as a superior candidate for identifying the POD, as all of the available studies
22 have limitations. However, some studies are considered more desirable as a principal or critical
23 study than other studies for the reasons below. The epidemiologic studies by Ferroni et al.
24 (1992) and Spinatonda et al. (1997) have associated uncertainties related to incomplete reporting
25 and are methodologically weaker than other epidemiologic studies, such as those of Seeber
26 (1989), Altmann et al. (1995), or Echeverria et al. (1994, 1995), that assessed neurobehavioral
27 functions. The studies by Echeverria et al. (1994, 1995) are not informative for supporting a
28 POD because the authors attribute observed visual function effects to past higher
29 tetrachloroethylene concentrations and historical exposure data that are not available. Table 5-1
30 identifies study characteristics and the rationale for the principal study considered in the RfC
31 analysis.
32 Seeber (1989) reports effects on visuo-spatial function, as does the residential study by
33 Altmann et al. (1995). Both studies are considered adequate for quantitative analysis, given the
34 numbers of study subjects (with Seeber et al., 1989, having the larger number of study subjects),
35 and their use of appropriate statistical methods, including methods to adjust for potentially
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-1. Summary of rationale for principal study selection
Consideration/
approach
Type of data
Decision
Quality of study
Animal
neurotoxicity
studies
Animal neurotoxicity studies are considered as supporting studies. An
RfC/RfD from human data, if available and of adequate quality, is
preferred for reducing interspecies extrapolation uncertainties.
Quality of study
Human
neurotoxicity
studies
Both occupational and residential studies on tetrachloroethylene exposure
contain uncertainties regarding their use for quantitative analysis. Some of
these epidemiology studies carry greater weight for quantitative analysis
than other studies. The occupational studies of Seeber (1989) and Cavelleri
et al. (1994), and the residential study of Altmann et al. (1995) are
considered to carry greater weight for quantitative analysis than Ferroni et
al. (1992), Echeverria et al. (1994, 1995), and Spinatonda et al. (1997).
Measurement
tool
Standardized
neurobehavioral
battery
Both occupational and residential epidemiology studies assessed
neurobehavioral function using a standardized neurobehavioral battery.
The battery has been widely administered to occupational population in
different setting with a reasonably high degree of validity. WHO and
ATSDR recommend these test methods to evaluate nervous system deficits
in adults and children.
Endpoint
Deficits in
neurological
domains such as
attention, motor
function,
vigilance,or
visuo-spatial
function.
There is congruence of neurological effects observed in studies of both
residential and occupational populations. These domains are also sensitive
to acute tetrachloroethylene exposure in controlled human studies. The
consistency of observed effects between occupational and residential
populations and their persistence with lower tetrachloroethylene
concentration, as experienced by residential populations, provide a strong
rationale for a study of lower-level exposures as the basis for the RfC.
Relevance of
exposure
scenario
Epidmiology
studies of
residential
populations
Tetrachloroethylene exposure to residential populations is of lower
concentration and of chronic duration compared to acute duration and
higher concentration exposure to occupational populations. Additionally,
potential to tetrachloroethylene peak or intensity concentrations is more
common with occupational exposures. A study of residential exposure, if
adequate and of similar quality as an occupational epidemiology study, is
preferred for supporting the RfC because it better represents exposure
scenarios of interest to EPA.
3
4
5
6
7
8
9
10
11
12
confounding factors. However, in both studies, statistical analyses that adjusted for potential
confounders may have not have been fully complete due to the use of categorical variables.
The report by Cavalleri et al. (1994) is consistent with the growing body of literature
indicating that chronic exposure to a variety of volatile organic solvents, including
tetrachloroethylene, toluene, styrene, and carbon disulfide, is associated with deficits in visual
perception measured either as deficits in color vision or deficits in VCS (see Section 4.6.1);
visual perception is a sensitive test of neurological impairment. The study authors reported
poorer performance on a test of color vision among dry cleaning operators. A statistically
significant lower prevalence of tetrachloroethylene-exposed dry cleaners had perfect scores on a
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-3 DRAFT-DO NOT CITE OR QUOTE
-------
1 test of color perception, and their mean score was higher as compared with those of controls,
2 indicating impaired color vision among dry cleaners. No effects were observed in laundry
3 workers with exposure to a lower mean TWA concentration than that of dry cleaning operators.
4 These effects of tetrachloroethylene-induced deficits in color vision are supported by other
5 studies (Muttray et al., 1997; Gobba et al., 1998).
6 Deficits in visual function were also reported in Schreiber et al. (2002), a study originally
7 designed as a pilot for a study composed of a larger number of subjects, assessed
8 tetrachloroethylene with air monitoring and markers in biological samples (biomarkers). This
9 pilot study was expanded after its inception to include control subjects and tests of VCS (NYS
10 OAG, 2004). The findings in Schreiber et al. (2002), a first report of VCS deficits, need
11 replication. For these reasons, although this study contributes to the weight of evidence for
12 hazard identification, it is less desirable than the residential study of Altmann et al. (1995) as a
13 critical study for developing an RfC.
14 The study by Altmann et al. (1995) is chosen as the principal study for a number of
15 reasons. First, it adopted a standardized neurobehavioral battery to evaluate neurological effects
16 in residents. Tests in this battery have been widely administered to occupational populations in
17 different settings with a reasonably high degree of validity (Anger et al., 2000). Additionally,
18 several public health organizations, such as the World Health Organization and the ATSDR in
19 the United States, recommend these test methods to evaluate nervous system deficits in adults
20 and children (Anger et al., 1994, 2000, 2003; ATSDR, 1996; Amler et al., 1994). Second, there
21 is congruence of neurological effects observed in studies of both residential and occupational
22 populations. As shown in Table 4-1 (Chapter 4), decrements in a number of neurological
23 domains such as attention, motor function, and vigilance reported by Altmann et al. (1994) are
24 also reported for occupationally exposed populations. The consistency of these effects between
25 the two populations and their persistence with lower tetrachloroethylene concentration, as
26 experienced by residential populations, provide a strong rationale for a study of lower-level
27 exposures as the basis for the RfC. Last, a study of residential exposures is preferred for
28 quantitative analysis because it better represents exposure scenarios of interest to EPA.
29 Table 5-2 identifies studies and outcomes considered for quantitative analysis.
30 Table 5-1 summarizes chronic, subchronic or longer-term, and developmental toxicity
31 studies considered for derivation of the inhalation RfC and are a subset of the body of evidence
32 on tetrachloroethylene more fully described in Chapter 4. These studies are considered
33 supportive of a POD and an RfC because they report effects associated with lower exposure
34 concentrations or are studies with multiple experimental exposures, allowing exploration of
35 benchmark dose (BMD) approaches. For each study, Table 5-1 identifies the species; the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-4 DRAFT-DO NOT CITE OR QUOTE
-------
o
Oi
Table 5-2. Inhalation studies considered in the development of an RfC
Organ/ system
Liver
Kidney
CNS
Study
Brodkin et al.
(1995)
Gennari et al.
(1984)
Kjellstrand et al.
(1984)
NTP(1986)
JISA(1993)
Verplanke et al.
(1999)
Trevisan et al.
(2000)
Franchini et al.
(1983)
Mutti et al.
(1992)
NTP(1986)
Altmann et al.
(1995)
Schreiber et al.
(2002)
Species
Human
Human
Mouse
Mouse
Mouse
Human
Human
Human
Human
Rat
Human
Human
Duration/ exposure
route
20 years mean duration
12 years mean duration
Subchronic (4 weeks)
continuous
Chronic bioassay
(104 weeks)
Chronic
(104 weeks)
4 years duration
(geometric mean)
15 years duration
(geometric mean)
14 years mean duration
10 years duration
Chronic bioassay
(104 weeks)
10.6 years median
duration
5.8 years mean duration
NOAEL/LOAEI/
(ppm)
4Ji, 19.8
(overall mean, 16)
11.3 (mean)
0,2,37,75,150
0, 100, 200
0, 10, SO, 250
L2 (mean)
M
(mean)
Ifl
15 (median)
0, 200. 400
0.7 (mean)
0.2 (median)
0.1 (residents, median
and mean), maybe as
high as 0.4 (mean) and
0.3 (median
Effect
Hepatic
parenchymal
changes
Gamma glutamyl
transpeptidase
Increased liver
weight
Increased liver
degeneration,
necrosis
Increased
angiectasis
Retinol binding
protein
Glutamine
sythetase
Lysozyme
Urine and serum
markers of
nephrotoxicity
Increased
karyomegaly,
megalonuclear-
cytosis
Verbal function ,
cognitive
function,
vigilance, vision
Visual contrast
sensitivity
Human equivalent continuous
concentrations'" (ppm)
NOAEL/LOAEL
LOAEL: 2
LOAEL: 4
LOAEL: 9
LOAEL: 18
NOAEL: 2
LOAEL: 0.3
LOAEL: 3
LOAEL: 4
LOAEL: 5
LOAEL: 36
LOAEL: 0.7
LOAEL: 0.4
BMCLC
BMCL10: 0.5
NA
BMCLS: 0.6
BMCLS/P: 1.4-10
NA
BMCL10: 2
BMCL10/P: 4.3-23
NA
NA
NA
BMCL10: 2.2
NA
NA
oo S3'
I
I
§
***.
£3'
1
TO'
^ %
I Co
"" §
I
H I
O
H
W
-------
Table 5-2. Inhalation studies considered in the development of an RfC (continued)
Organ/ system
CNS (cont.)
Study
Schreiber et al.
(2002)
Cavalleri et al.
(1994)
Spinatonda et al.
(1997)
Seeber(1989)
Ferroni et al.
(1992)
Echeverria et al.
(1995)
Kjellstrand et al.
(1984)
Rosengren et al.
(1986)
Mattsson et al.
(1998)
Wang et al.
(1993)
Species
Human
Human
Human
Human
Human
Human
Mouse
Gerbil
Rat
Rat
Duration/ exposure
route
4 years mean duration
8.8 years mean duration
Inhalation (no
information on
duration)
>10 years mean
duration
10.6 years mean
duration
15 (high-exposure
group) years mean
duration
Subchronic (4 weeks)
continuous
Subchronic (12 weeks,
with 16- week follow-
up) continuous
Subchronic (13 weeks)
6hrs/d, 5d/w
Subchronic (12 weeks)
continuous
NOAEL/LOAEI/
(ppm)
0.3 (daycare workers,
mean and median)
2
8 (median)
12,53
JJ
41 (operators)
0,9,22,75,150
0, 6Q, 300
0, 50, 200, 800
0, 300. 600
Effect
Visual contrast
sensitivity
Dyschromatopsia
Reaction time
Visuospatial
function,
information
processing speed
Reaction time,
continuous
performance
Endocrine:
prolactin
Visuospatial
function
Butyryl
cholinesterase
Brain: protein,
DNA
concentration
Flash-evoked
potential
Reduced brain
weight, DNA,
protein
Human equivalent continuous
concentrations'" (ppm)
NOAEL/LOAEL
LOAEL: 0.1
LOAEL: 3
LOAEL: 3
LOAEL: 4
LOAEL: 5
LOAEL: 4
LOAEL: 15
NOAEL: 9
LOAEL: 60
NOAEL: 36
NOAEL: 300
BMCLC
NA
NA
NA
NA
NA
NA
NA
NAe
NA
NA
NA
-------
Table 5-2. Inhalation studies considered in the development of an RfC (continued)
Organ/ system
Reproductive/
developmental
Study
Eskenazi et al.
(1991)
Nelson et al.
(1980)
Beliles et al.
(1980)
Tinston(1994)
Carney et al.
(2006)
Species
Human
Rat
Mouse
Rat
Rat
Duration/ exposure
route
9.5 years mean
duration, occupational
7 hrs/day on gestation
day 7-1 3 or 14-20
5 days exposure;
1, 4, and 10 week
follow-up
Developmental —
multigeneration; 6
hrs/day, 5 days/wk
Developmental — 6
hrs/day on GD 6-19
NOAEL/LQAELa
(ppm)
12
0,100,900
0, 100. 500
0,100,300,1,000
0, 65, 250, 600
Effect
Sperm quality
Decreased
weight gain in
offspring;
CNS: behavior,
brain
acetylcholine
Sperm quality
F2A pup deaths
by day 29;
FlandF2
generations:
CNS depression
Decreased fetal
and placental
weight and
incomplete
ossification of
thoracic
veretebral
centra
Human equivalent continuous
concentrations'" (ppm)
NOAEL/LOAEL
LOAEL: 0.4
NOAEL: 29
NOAEL: 21
NOAEL: 18
NOAEL: 16
BMCLC
NA
NA
NA
BMCLoi: 1.8
NA
Experimental/observational NOAEL is underlined, LOAEL is double-underlined.
Calculated using RfC methodology for a Category 3 gas, extrathoracic effects, and adjusted to equivalent continuous exposure; occupational exposures were
multiplied by 5/7(days) x 10/20 (nrVday, breathing rate).
BMCL10 is the lower bound on concentration associated with a 10% response over background. BMCLS is the lower bound on dose associated with a one
standard deviation change in the mean over the control response. This corresponds to an excess risk of approximately 10% for developing a level of the
endpoint above the 98th percentile (or below the 2nd percentile) of the control distribution for normally distributed effects. BMCLx/p is the estimation, based on
total metabolism, for either type of benchmark response (percent change or standard deviation).
-------
Table 5-2. Inhalation studies considered in the development of an RfC (continued)
d Atmospheric monitoring indicated slightly higher exposure levels were experienced by subjects. Schreiber et al. (2002) found mean tetrachloroethylene
concentrations of 0.2 ppm (0.09 ppm, median) of four families living in apartments above active dry cleaning and two families living in an apartment building
where dry cleaning had ceased 1 month earlier. Ambient monitoring of these six apartments during a period of active dry cleaning indicated exposure to
higher concentrations, mean = 0.4 ppm (median 0.2 ppm).
e Benchmark modeling not feasible; exposure-response relationship showed very little gradation among responses aside from apparently maximal response in
mid- and high-dose groups relative to control and low-dose groups.
-------
1 exposure duration; the ambient (experimental) concentration or, for epidemiologic studies, the
2 mean concentration; the observed effect; and the exposure concentration, identified as the
3 NOAEL or the LOAEL. Additionally, HECs for LOAELs or NOAELs are presented so as to
4 better allow examination of effect levels across studies and species. The HECs in Table 5-1 are
5 calculated using the RfC methodology for a category 3 gas, extrathoracic effects, and adjusted to
6 equivalent continuous exposure (U.S. EPA, 1994).
7
8 NOAEL * [HEcj = NOAEL * [ADJ] (ppm) x (Hb/g)A/Hb/g)H
9 where:
10 NOAEL* [REG] = the NOAEL or analogous effect level such as the benchmark
11 concentration (BMC)
12
13 NOAEL*[ADJ] = the NOAEL or analogous effect level adjusted for duration of
14 experimental regimen; experimental exposure times duration (number of
15 hrs exposed/24 hrs) times week (number of days of exposure/7 days)
16
17 (Hb/g)A/Hb/g)n = the ratio of the blood/gas (air) partition coefficient of the chemical for the
18 laboratory animal species to the human value. The value of 1 is used for
19 the ratio if (Hb/g)A > Hb/g)H
20
21 Response levels in Table 5-1 are presented as either a NOAEL, a LOAEL if the study did not
22 identify a NOAEL, or a modeled BMCL3 if the study results were suitable for modeling, using
23 BMD methodology (U.S. EPA, 2000). Five studies (Brodkin et al., 1985; Kjellstrand et al.,
24 1984; JISA, 1993; NTP, 1986; Tinston, 1994) reported toxicity in other organs besides the
25 nervous system and are presented for the comparative purposes. These studies had experimental
26 designs with multiple exposure concentrations and quantitative information that were sufficient
27 for quantitative analysis using BMD methodologies.
28 Ideally, an examination using pharmacokinetically derived dose metrics is preferred
29 when well validated models are available. An alternative procedure for deriving a POD using
30 pharmacokinetic modeling to estimate metabolite production for liver toxicity data sets is also
31 examined in this document.
3 A BMCL10 is the lower 95% bound on the dose or concentration associated with a 10% extra risk
compared to background. In general, the benchmark approach is a superior methodology to the LOAEL/NOAEL
approach because it makes more complete use of exposure/ response data rather than being limited by the sample
size of the study group that happened to be exposed at the LOAEL or NOAEL. The BMD approach identifies doses
that are not restricted to being one of the study exposure levels, another improvement over the LOAEL/NOAEL
approach, particularly when doses are widely spaced. BMDs correspond to specific response levels, such as 10%
extra risk, facilitating comparisons across studies and endpoints. Because the BMD corresponds to an adverse effect
level, it should be treated conceptually as a LOAEL.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-9 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.1.2. Method of Analysis
2 The present analysis defines a POD using the traditional NOAEL/LOAEL approach in
3 addition to using BMD modeling where feasible. Further, PBPK modeling was used with
4 suitable studies in animals in order to inform the process of extrapolating to human equivalent
5 exposures. The use of these alternative approaches has the potential to add information to the
6 NOAEL/LOAEL approach.
7 Altmann et al. (1995) reports a mean 8-hr TWA of 0.7 ppm (4.8 mg/m3) for residents
8 exposed to tetrachloroethylene from living in close proximity to a dry cleaning establishment.
9 This mean concentration is used as the POD for the RfC derivation. The POD is not adjusted for
10 exposure duration as is the general practice when using an occupational study. Instead, an
11 assumption was made that residents were continuously exposed. In other words, no further
12 adjustments to the estimated exposure level to approximate continuous exposure levels were
13 considered necessary due to the lack of information concerning the duration and length of
14 exposure of the study population.
15 The application of BMD methodologies offer advantages over traditional
16 LOAEL/NOAEL approaches, however, exposure in Altmann et al. (1995) is reported as a group
17 mean concentration and does not allow fitting of BMD models. Data sufficient for BMD
18 modeling generally came from animal experiments, with the exception of one human study. This
19 was the Brodkin et al. (1985) study, which found an increasing incidence of hepatic parenchymal
20 changes in laundry workers with increasing exposure to perchloroethylene. There was no
21 concurrent control group, so substitution of a background level was necessary (Hartwell et al.,
22 1985). BMD modeling of the two reported groups plus the substituted control group yielded a
23 BMCLio of 0.5 ppm (see Table 5-1). In addition to the lack of a control group, another
24 limitation of the result modeled from the Brodkin et al. (1985) study is that hepatic parenchymal
25 changes appear to be a less severe endpoint, as all of the participants had normal liver function
26 measurements, and its relationship to frank liver disease is not known. Despite these
27 uncertainties, the result provides support for the POD derived from the Altmann et al. (1995)
28 study. Furthermore, Eskenazi et al. (1991), who observed effect on sperm quality at a similar
29 mean exposure concentration as that of Altmann et al. (1995) and Schreiber et al. (2002), support
30 the POD of the critical study. Table 5A-1 in the appendix provides details of the BMD
31 modeling.
32 The animal studies suitable for BMD modeling addressed liver and kidney toxicity and
33 pup death. For liver toxicity in mice (increased liver weights [Kjellstrand et al., 1984]), a human
34 equivalent BMCLS4 of 0.6 ppm was estimated using administered concentration, and a BMCLs/p
BMCLS = Lower bound on dose (concentration) associated with a one standard deviation change in the
mean over the control response. This corresponds to an excess risk of approximately 10% for the proportion of
individuals above the 98th percentile (or below the 2nd percentile) of the control distribution for normally distributed
effects (U.S. EPA, 2000).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-10 DRAFT-DO NOT CITE OR QUOTE
-------
1 of 1.4-10 ppm (9.6-69 mg/m3) was derived using pharmacokinetic models, assuming that total
2 metabolism was the relevant dosimeter for animals and humans (see Tables 5A-2 and 5A-3 for
3 modeling details). For liver toxicity in rats (increased angiectasis [JISA, 1993]), BMD modeling
4 yielded a BMCLio of 3.7 ppm (25 mg/m3) using administered concentration and a BMCLio/p of
5 4.3-23 ppm (29-156 mg/m3) using pharmacokinetic modeling (see Table 5A-6). PBPK models
6 are described in Section 3.5 and are more fully considered in the cancer dose-response analysis
7 (see Section 5.4.4.2). Karyomegaly (kidney) was observed in male rats in the chronic study by
8 the NTP (1986). BMD modeling yielded a BMCL75 of 29 ppm (near the lowest exposure tested)
9 and a BMCLio of 2.2 ppm (15 mg/m3). Because the lowest exposure was associated with a
10 relatively high response, estimation of the BMCLio is somewhat tenuous, and modeling using
11 total metabolism was not pursued. Last, BMD modeling of pup deaths through Day 29 in the
12 F2A generation of a multigeneration study (Tinston, 1994) yielded a BMCLoi of 1.8 ppm (12
13 mg/m3). These BMD analyses are more fully presented in Tables 5A-2 through 5A-9 and are
14 provided in support of the choice of the Altmann et al. (1995) study as the most relevant data
15 source for developing the RfC.
16
17 5.1.3. Reference Concentration (RfC) Derivation, Including Application of Uncertainty
18 Factors
3
19 The NOAEL of 0.7 ppm (4.8 mg/nr) from Altmann et al. (1995) is the POD, as described
20 above. The POD is reduced by the following UFs.5
BMCLx/p = Lower bound on dose (concentration), where X denotes the benchmark response (either in
percent or one standard deviation), based on dose metric of total metabolism as estimated by a pharmacokinetic
model. This subscript distinguishes these BMCLs from those based on administered exposure.
RfCs apply to lifetime human environmental exposure, including exposures of sensitive subgroups.
Differences between study conditions and conditions of human environmental exposure may make a dose that
appears safe in an experiment not safe in the environment. UFs account for differences between study conditions
and conditions of human environmental exposure. These include the following:
1. Variation from average humans to sensitive humans: RfCs apply to the human population, including sensitive
subgroups, but studies rarely target sensitive humans. Sensitive humans could be adversely affected at doses
lower than those that affect a general study population; consequently, general population NOAELs are reduced
to cover sensitive humans.
2. Uncertainty in extrapolation from animals to humans: If an RfC is developed from animal studies, the animal
NOAEL is reduced to reflect pharmacokinetic and pharmacodynamic factors that may make humans more
sensitive than animals.
3. Uncertainty in extrapolating from subchronic NOAELs to chronic NOAELs: RfCs apply to lifetime exposure,
but sometimes the best data come from shorter studies. Lifetime exposure can have effects that do not appear
in a shorter study; consequently, a safe dose for lifetime exposure can be less than the safe dose for a shorter
period. If an RfC is developed from less-than-lifetime studies, the less-than-lifetime NOAEL is adjusted to
estimate a lifetime NOAEL.
4. Uncertainty in extrapolation from LOAELs to NOAELs: RfCs estimate a dose with appreciable risks, but
sometimes adverse effects are observed at all study doses. If an RfC is developed from a dose where there are
adverse effects, that dose is adjusted to estimate a NOAEL.
5. Other factors to reflect professional assessment of scientific uncertainties not explicitly treated above,
including completeness of the overall database, minimal sample size, or poor exposure characterization.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-11 DRAFT-DO NOT CITE OR QUOTE
-------
1 1. Human variation: The UF of 10 is applied for human variation. Although human
2 residential data were used as the basis for the POD (Altmann et al., 1995), the overall
3 database does not support the use of a value other than the default 10-fold UFn. The
4 rationale for this determination is based on several considerations. First, Altmann et al.
5 (1995) excluded subjects with disorders such as hypertension, neurological or
6 endocrinological diseases (e.g., diabetes), impaired vision, or impairment of joints.
7 Hence, subjects in Altmann et al. (1995) are considered to be a select population,
8 analogous to an occupational population and subject to selection bias known as the
9 "healthy worker effect." The use of these exclusion criteria and the small numbers of
10 subjects in Altmann et al. (1995; « = 37, 14 exposed and 23 control subjects) suggest that
11 the range of human variation in a larger and more diverse population is not likely
12 represented by this study. Second, no information is presented in Altmann et al. (1995)
13 with which to examine variation between subjects. Third, the sparse data available on
14 tetrachloroethylene indicate the presence of pharmacokinetic variation in the human
15 population. One report described variation in tetrachloroethylene blood concentrations
16 among nine subjects exposed acutely to tetrachloroethylene and observed a twofold
17 spread in the ratio of alveolar air concentrations to atmospheric concentrations (Opdam,
18 1989).
19
20 Gentry et al. (2003), Clewell et al. (2004), and Pelekis et al. (2001) present
21 pharmacokinetic modeling simulations of pharmacokinetic variation between adults and
22 children in tetrachloroethylene parent and its metabolites. As the authors themselves
23 indicated, validation of these results for various life stages and further refinement of the
24 parameters in the model are necessary before the results of such an analysis can be
25 considered for use in risk assessment. Further investigation of variability in the
26 parameters used in the Clewell et al. (2004) analysis is also needed before their results
27 can be used to address pharmacokinetic uncertainty for age and gender variability.
28
29 Given an adequate database, or after adjustment in the assessment for deficiencies in the
30 database, a reference value incorporating the default 10-fold factor for human variation is
31 believed to adequately address likely susceptibilities in children. A thorough evaluation
32 of the animal and human hazard data for tetrachloroethylene identified the developing
33 fetus and the young child as susceptible life stages (populations). As described in Section
34 4.9, data-derived noncancer outcomes of concern in children for perinatal exposure are
35 (1) spontaneous abortions, (2) childhood mortality, and (3) neurological impairment.
36 This assessment contains a database uncertainty factor addressing, in part, limitations in
37 life stage-related data in human and rodent studies.
38
39 Section 4.9 also describes differential opportunities for exposure to children. However,
40 susceptibilities of this nature are not addressed in the determination of the UF for human
41 variation, but rather are used to establish a context for the hazard and dose-response
42 evaluation and are further addressed in the exposure assessment and subsequent risk
43 characterization.
44
45 2. Animal-to-human uncertainty. This UF is used when the POD is supported by an animal
46 study. When the POD is supported by a human study, this UF is not needed.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-12 DRAFT-DO NOT CITE OR QUOTE
-------
1 3. Subchronic-to-chronic uncertainty. A factor to address the potential for more severe
2 toxicity from chronic or lifetime exposure to tetrachloroethylene is not used in this
3 assessment. The epidemiologic studies, except for Schreiber et al. (2002), are all of
4 median duration of exposures of more than 15% of a 70-year lifespan. There are no data
5 to suggest that continuing exposure to tetrachloroethylene can increase the severity of
6 effects; duration-response trends are not generally evident in the human studies.
7
8 4. LOAEL-to-NOAEL uncertainty. The default value of 10 is applied for use of a LOAEL
9 because of the lack of a NOAEL in Altmann et al. (1995).
10
11 5. Database uncertainty. A threefold database UF has been applied to address the lack of
12 data to adequately characterize the hazard and dose-response in the human population.
13 The rationale for this database UF is based on several considerations. There is human
14 evidence of neurotoxicity following tetrachloroethylene exposure, with both visual
15 system dysfunction and cognitive performance deficits. However, these studies have
16 limitations, and in particular lack adequate data to address childhood or other life stage
17 susceptibility. There is also a lack of animal studies (including in developing animals)
18 designed to clearly investigate these neurotoxicity findings and define and characterize
19 the exposure-response relationship.
20 A broad range of animal toxicology data are available for the hazard assessment of
21 tetrachloroethylene, as described throughout this document. Included in these studies are
22 short-term and long-term bioassays in rats and mice (see Chapter 4 and Table 4-2);
23 neurotoxicology studies in rats, mice, and gerbils (see Tables 4-6 and 4-7); prenatal
24 developmental toxicity studies in rats, mice, and rabbits and a two-generation
25 reproduction study in rats (see Table 4-10); and numerous supporting genotoxicity and
26 metabolism studies. Nevertheless, critical data gaps have been identified. Data from
27 acute studies in animals (Warren et al., 1996; Umeza et al., 1997) suggest that cognitive
28 function is affected by exposure to tetrachloroethylene. These studies do not address the
29 exposure-response relationship for subchronic and chronic tetrachloroethylene exposures
30 on cognitive functional deficits observed in humans (e.g., Seeber, 1989; Echeverria et al.,
31 1994; and Altmann et al., 1995).
32
33 Even more importantly, there is a lack of cognitive testing in both developmentally
34 exposed animals and adult animals following exposures to tetrachloroethylene that are
35 longer than acute durations of exposure. For another critical outcome, visual function,
36 there has been a limited evaluation of visual function in rodents, with the exception of the
37 evoked potential studies by Mattsson et al. (1998). Visual system dysfunction and
38 processing of visual spatial information are sensitive endpoints in human studies. The
39 exposure-response relationship of these functional deficits could be evaluated more
40 definitively with studies using homologous methods that examine retinal and visual
41 function in experimental animals. These types of studies could help elucidate whether
42 there are both peripheral and central effects of tetrachloroethylene exposure on visual
43 perception, and they could be used as an animal model to better define the exposure-
44 response relationships.
45
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-13 DRAFT-DO NOT CITE OR QUOTE
-------
1 Additionally, the database of human epidemiological studies includes studies of liver and
2 kidney toxicity (see Tables 4-1 and 4-3), neurotoxicity (see Tables 4-4 and 4-5), and
3 developmental/reproductive effects (see Table 4-8). The reference value is established on
4 the basis of a sensitive neurological effect in adult humans (Altmann et al., 1995), and
5 some characterization of the response of children to tetrachloroethylene exposure was
6 found in limited data for a similar neurological (visual system) parameter (Schreiber et
7 al., 2002).
8
9 Although the toxicological database is considered adequate for establishing a reference
10 value, some uncertainties remain. In both the Altmann et al. (1995) and the Schreiber et
11 al. (2002) studies, there was a lack of robust sample size and an inadequate dose-response
12 characterization for potentially susceptible human populations following
13 tetrachloroethylene exposures. Although the Altmann et al. (1995) study (with a LOAEL
14 of 0.7 ppm for healthy adult subjects) was used in setting the reference value (based on a
15 number of considerations that are summarized in Section 5.2.1), the Schreiber et al.
16 (2002) study, using an alternative visually based testing paradigm, identified adverse
17 visual effects at 0.4 ppm (see Section 4.6.1.2.11). Additionally, in a postnatal
18 neurotoxicity study in mice (Fredriksson et al., 1993), persistent neurological effects (i.e.,
19 increased locomotion and decreased rearing behavior at 60 days of age, measured 43 days
20 after exposure ceased) were observed at an oral dose of 5 ppm, with no NOAEL,
21 although this study did not conform to traditional toxicity testing guidelines (see Section
22 4.6.2.2). The possibility exists that if adequate, robust, dose-response data based on the
23 most appropriate neurophysiological and cognitive tests were available, the exposure
24 eliciting an adverse response (and hence the POD for the reference value) could be lower
25 than that established on the basis of deficits in visuo-spatial and cognitive function
26 following tetrachloroethylene exposure in healthy adults (Altmann et al., 1995).
27
28 A total UF of 300 was applied to this effect level: 10 for human variation, 10 for
29 consideration of LOAEL to NOAEL, and 3 for database uncertainties.
30
31 RfC = NOAEL I UF
32 =0.7ppm(4.8mg/m3)/300
33 = 0.002 ppm (0.02 mg/m3, rounded to 1 significant digit}
34
35 where:
36 0.7 ppm = LOAEL for neurological effects in residents exposed to tetrachloroethylene
37 (Altmann etal., 1995)
38
39 300 = composite UF chosen to account for extrapolation from a LOAEL to a
40 NOAEL in humans, intra-individual variability in humans, and uncertainties
41 in the database
42
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-14 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.1.4. Supporting Studies
2 PODs and reference values (RfVs) that could be derived from supporting neurotoxicity
3 studies identified in Table 5-2 (see Section 5.1.1) are presented below in Figure 5-1 to allow a
4 comparison with the critical study. Not all studies of neurotoxic effects identified in Table 5-2
5 are presented in Figure 5-1; however, these studies are a sample of human and animal data sets
6 for some of the more sensitive measures of neurotoxic endpoints. Vision or visual function
7 effects are observed in the human and rodent studies and one study in gerbils reports changes in
8 brain chemistry (Schreiber et al., 2002; Altmann et al., 1995; Cavalerri et al., 1994; Seeber,
9 1989; Mattsson et al., 1998; Rosengren et al., 1986). Effect magnitudes could be identified for
10 three studies and ranged from a 5% change in color vision index to roughly a 15% decrement in
11 several tests on a neurobehavioral evaluation battery (summarized in Figure 5-1). In the absence
12 of tetrachloroethylene data to inform uncertainty factors, the analysis uses the default values as
13 discussed in Section 5.1.3: a factor of 10 to extrapolate from a LOAEL to a NOAEL; a factor of
14 10 for human variation; and a factor of 3 for database deficiencies. For the rodent studies of
15 Mattson et al. (1998) and Rosengren et al. (1986), PODs represent human equivalent
16 concentrations for a category 3 gas adopting EPA's RfC methodology (U.S. EPA, 1994) and an
17 uncertainty factor of 3 addresses uncertainties associated with extrapolating from animal data to
18 an average human. Three studies are of subchronic exposure duration, and extrapolation to
19 chronic exposure duration is achieved using a factor of 10 for the studies of Mattsson et al.
20 (1998) and Schreiber et al. (2002; daycare employees). A subchronic to chronic factor of 3,
21 rather than 10, was applied for Rosengren et al. (1986) in light of the large overall uncertainty for
22 this study associated with extrapolating from a LOAEL to NOAEL, from animal to humans, for
23 human variation, and for database deficiencies; the total uncertainty factor was 3,000.
24 One study in Figure 5-1 used a neurological test for visual contrast sensitivity and yields
25 LOAELs (PODs) of 0.1 and 0.4 ppm for day care workers or apartment dwellers adjacent to dry
26 cleaners, respectively (Schreiber et al., 2002). An RfV developed from this study would be
27 lower than that of Altmann et al. (1995) but, as more fully discussed in Section 5.1.1, this study
28 was not considered as the critical study due to its design as a pilot for a larger study conducted
29 by the NYS DOH (see Sections 5.2.1 and 5.2.2). LOAELs are higher than that of Altmann et al.
30 (1995) for the occupational studies of Cavalleri et al. (1994) and Seeber (1989) and reflect higher
31 tetrachloroethylene concentrations in the occupational setting when compared to residential
32 concentrations.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-15 DRAFT-DO NOT CITE OR QUOTE
-------
Inhalation Neurotoxicity RfVs
Human Human Human Human Human
Gerbil
-- 100
Q.
c
o
O
0.001
0.0001
0.00001
0.01
c
01
u
c
o
O
-- 0.01
-- 0.001
KEY
+ Point of departure
(POD)
I I UF, extrapolation from a
lowest-observed-
adverse-effect level
(LOAEL) to a no-
observed-adverse-effect
level (NOAEL)
UF, animal-to-human
extrapolation
UF, human variation
UF, extrapolation from
subchronic exposure
duration to chronic
exposure duration
H UF, database
deficiencies
O Reference value (RfV)
Visual Visual Vision, cognitive Vision, 5% Visual function, Visual Brain
contrast contrast function, change in ~10% change function chemistry
(day care (apartment vigilance," ~15% color in neuro- (Mattsson (Rosengren
employees) residents) decrement confusion behavioral etal., 1998) etal.,1986)
(Schreiber (Schreiber (Altmann index tests
etal.,2002) etal.,2002) etal., 1995) (Cavalleri (Seeber, 1989)
etal., 1994)
Principal study.
Figure 5-1. Array of PODs and reference values for a subset of neurotoxic effects of studies in Table 5-2.
-------
1 PODs and inhalation organ-specific RfVs are presented for selected studies in Figure 5-2
2 to give perspective on the RfC derived from the adverse neurotoxic effects in Altmann et al.
3 (1995) and to provide information on other systemic effects associated with tetrachloroethylene
4 exposure. Toxicity to the liver, kidney, developing fetus, and reproductive organs are observed
5 at higher mean or median tetrachloroethylene concentrations than Altmann et al. (1995). The
6 POD for a given study is the tetrachloroethylene concentration associated with the LOAEL,
7 NOAEL, or lower bound on a benchmark concentration (BMCL). Benchmark concentration
8 models are fit to JISA (1993) and Tinston (1994; see Appendix 5, Tables 5A-5 and 5A-7).
9 Furthermore, a pharmacokinetic model of Bois et al. (1996) and scaling of the body weight to the
10 3/4th power is adopted for liver weight changes in JISA (1993) to obtain human equivalent
11 concentrations, a treatment consistent with the cancer dose-response analysis of liver tumors and
12 more fully described in Section 5.4.3.1. PODs developed from pharmacokinetic models of Rao
13 and Brown (1993) and Reitz et al. (1996), reflecting uncertainties associated with
14 tetrachloroethylene metabolism, are up to an order of magnitude higher than that of Bois et al.
15 (1996). The POD for JISA (1993) using the Bois et al. (1996) pharmacokinetic model is shown
16 on Figure 5-2.
17 Organ-specific inhalation RfVs for endpoints besides neurotoxicity are developed by a
18 procedure similar to that for neurotoxicity and was carried out for Altmann et al. (1995). Default
19 values in the absence of data-informed adjustment factors are adopted to account for
20 uncertainties in the analysis. With human studies, the default values as discussed in
21 Section 5.1.3 are a factor of 10 to extrapolate from a LOAEL in Altmann et al. (1995) and Mutti
22 et al. (1992) to a NOAEL; a factor of 10 for human variation; and a factor of 3 for database
23 deficiencies. Another uncertainty factor is adopted to account for uncertainty in extrapolating
24 laboratory animal data to the case of average healthy humans (U.S. EPA, 1994). Typically, this
25 factor addresses residual uncertainties not associated with default dosimetric adjustment like the
26 human equivalent concentration. The POD for Tinston (1994) is a human equivalent
27 concentration for a category 3 gas and an uncertainty factor of 3 addresses uncertainty associated
28 with extrapolating animal data to the average healthy human case. In the case of liver
29 angiectasis, cross-species scaling of total rate of metabolism using body weight to the 3/4th
30 power was used for describing toxicological equivalence because of the extensive rationale
31 supporting it (U.S. EPA, 1992). The methodology achieves a human equivalent concentration
32 that is expected to approximate AUC or ppm equivalence across species for a category 3 gas. An
33 animal to human uncertainty factor of 3 addresses non-pharmacokinetic uncertainties such as
34 pharmacodynamics as suggested in the RfC framework (U.S. EPA, 1994).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-17 DRAFT-DO NOT CITE OR QUOTE
-------
Inhalation Organ-Specific RfVs
100
Human
Human
Mouse
Q.
a.
c
o
«
I
o
c
o
o
0.001
0.0001
0.01
KEY
Point of departure
(POD)
UF, extrapolation from a
lowest-observed-
adverse-effect level
(LOAEL) to a no-
observed-adverse-effect
level (NOAEL)
UF, animal-to-human
extrapolation
UF, human variation
UF, database
deficiencies
Reference value (RfV)
CMS, -15%
decrement in verbal
function, cognitive
function, vigilance,
and vision3
(Altmann et al.,
1995)
Kidney, increase in Developmental/
mean urinary Reproductive, 1%
concentration of
several renal
enzymes
(Multti
etal., 1992)
increase in fetal
death, BMCLoi
as POD
(Tinston
etal., 1994)
Liver, 10%
increase in liver
angiectasis,
BMCLio/p as POD
(JISA, 1993)
aPrincipal study.
Figure 5-2. Organ-specific reference values for inhalation exposure to tetrachloroethylene.
-------
1 5.1.5. Previous Inhalation Assessment
2 There is no previous EPA RfC assessment for tetrachloroethylene with which to compare
3 and contrast the RfC developed in this assessment. Other assessments identified in Appendix A
4 have derived a noncancer reference value from the human evidence. California's drinking water
5 assessment on tetrachloroethylene (Cal EPA, 2001) derived a public health goal (PHG)6 for
6 noncancer effects from a geometric mean in Altmann et al. (1995) and two occupational studies:
7 Spinatonda et al. (1997) and Ferroni et al. (1992). The most recent assessment of
8 tetrachloroethylene by the NYS DOH (1997) used a slightly different set of neurotoxicity studies
9 than did California and presented reference criteria for neurotoxicity derived from Cavalleri et al.
10 (1994), Altmann et al. (1995), and Seeber (1989). NYS DOH (1997) considered the reference
11 criterion of Seeber (1989) as best providing a sufficient margin of exposure over the air levels of
12 tetrachloroethylene associated with CNS effects.
13 ATSDR (1997), on the other hand, based its chronic MRL on Ferroni et al. (1992).
14 ATSDR (1997) considered Altmann et al. (1995) to provide a NOAEL, a conclusion inconsistent
15 with the assessments by New York State and California. ATSDR, however, noted that the
16 Altmann et al. (1995) study suggested a need to characterize neurotoxic effects in populations
17 exposed to very low levels of tetrachloroethylene. A second report (Schreiber et al., 2002) of
18 visual functional deficits in two populations exposed to tetrachloroethylene at lower ambient
19 concentrations that were similar to those of Altmann et al. (1995) has become available since the
20 publication of the ATSDR toxicological profile.
21 A difference between this and previous assessments is in the previous assessments'
22 treatment of human variation, particularly the residential study by Altmann et al. (1995). A
23 choice other than the UF of 10 has been adopted in the assessments by California and New York
24 State. A presumption underpinning this choice is that the residential population studied by
25 Altmann et al. (1995) is more reflective of the general adult population than of an occupational
26 population and any accompanying selection bias that is often associated with a healthier worker
27 population. Although there is some merit in this opinion, observations in Altmann et al. (1995)
28 are of a German population: 14 adults with exposure to tetrachloroethylene. These individuals
29 likely do not represent the full range of human variation found in a large and ethnically diverse
30 population such as the United States population. Furthermore, as noted, the Altmann et al.
31 (1995) study excluded subjects with disorders such as hypertension, neurological or
32 endocrinological diseases (e.g., diabetes), impaired vision, or impairment of joints; hence, these
33 subjects can be considered as having an overall good health status, and individuals whose
PHG is conceptually similar to an RfD.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-19 DRAFT-DO NOT CITE OR QUOTE
-------
1 diseases may have increased their susceptibility to tetrachloroethylene effects were not included
2 in this study.
O
4 5.2. ORAL REFERENCE DOSE (RfD)
5 Ideally, the studies of greatest duration of exposure and conducted via the oral route of
6 exposure have the most confidence for derivation of an RfD.7 An earlier assessment of
7 tetrachloroethylene oral noncancer toxicity by EPA, for example, identified liver toxicity in
8 Buben and O'Flaherty (1985) as the critical effect for developing an RfD (U.S. EPA, 1988).
9 However, the application of pharmacokinetic models for a route-to-route extrapolation of the
10 inhalation studies expands the oral database. Cal EPA (2001), for example, carried out a route-
11 to-route extrapolation of the human inhalation studies of neurotoxic effects to develop a PHG for
12 oral tetrachloroethylene exposure, based on a route-to-route extrapolation of inhalation
13 neurotoxicity studies.
14
15 5.2.1. Choice of Principal Study and Critical Effects
16 Toxicity to several targets, including the liver, kidney, nervous system, and reproductive
17 system and to the developing fetus is seen in rodents with oral tetrachloroethylene exposure.
18 Effects have been observed at these targets in acute studies (28 days or less), longer
19 term/sub chronic studies (90 days), or chronic studies (1 year or more). At higher doses (above
20 approximately 1,000 mg/kg-day), targets of oral tetrachloroethylene toxicity include the liver,
21 kidney, nervous system, lymphatic system, reproductive system, and developing fetus (ATSDR,
22 1997). There are few studies at lower doses as compared to the number of studies of inhalation
23 exposure. Several targets of toxicity from oral exposure are similar to targets observed with
24 inhalation exposures, i.e., liver and kidney.
25 No epidemiologic studies of oral exposure were suitable for quantitative analysis,
26 although these studies did provide information for hazard identification. Four studies of
27 subchronic oral exposure in mice or rats (Buben and O'Flaherty, 1985; Jonker et al., 1996;
28 Berman et al., 1995; Hayes et al., 1986) are available, as is a developmental study in mice of oral
29 exposure to tetrachloroethylene (Fredriksson et al., 1993). As discussed above, EPA previously
30 developed an RfD from Buben and O'Flaherty (1985). A significant effect on liver weight was
31 seen in this study.
The RfD is expressed in units of milligrams per kilogram body weight per day (mg/kg-day). In general,
the RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious noncancer
effects during a lifetime.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-20 DRAFT-DO NOT CITE OR QUOTE
-------
1 The CNS is a sensitive target for tetrachloroethylene inhalation toxicity, as discussed in
2 Section 5.1.1. This assessment has attempted to expand the database for derivation of an RfD
3 using relevant inhalation data and route-to-route extrapolation with the aid of a PBPK model (see
4 Section 3.5) to the POD of Altmann et al. (1995). The nervous system is an expected target with
5 lower oral tetrachloroethylene exposures, in view of the fact that other organ systems such as the
6 liver the kidney are also common targets associated with both inhalation and either oral routes of
7 subchronic or chronic exposure. The similarity of effects in these organ systems with either oral
8 or inhalation exposure to tetrachloroethylene supports the use of route extrapolation to compare
9 PODs for oral and inhalation exposure. For these reasons, the inhalation study in humans by
10 Altmann et al. (1995) is chosen as the principal study for supporting the RfD.
11
12 5.2.2. Methods of Analysis, Including Models
13 The present analysis defines a POD using the traditional NOAEL/LOAEL approach in
14 addition to using BMD modeling where feasible. This assessment has attempted to expand the
15 database for derivation of an RfD using relevant inhalation data and route-to-route extrapolation
16 with the aid of a PBPK model (see Section 3.5). Several factors support the use of route-to-route
17 extrapolation for tetrachloroethylene. Tetrachloroethylene has been shown to be rapidly and
18 well absorbed by both the oral and inhalation routes of exposure (ATSDR, 1997). Additionally,
19 the metabolic pathways and kinetics of excretion with oral exposure are similar to those of
20 inhalation exposure (ATSDR, 1997). Furthermore, the data for oral administration indicate a
21 pattern of effects similar to that of inhalation exposure, including effects on the liver and kidney.
22 PBPK modeling was also used with suitable studies in animals in order to inform the process of
23 extrapolating to HECs. The use of these alternative approaches has the potential to add
24 information to the NOAEL/LOAEL approach.
25 PBPK modeling was used to derive the oral dose that would result in the same
26 tetrachloroethylene in blood AUC as that following a continuous inhalation exposure of 0.7 ppm,
27 the LOAEL from the inhalation study by Altmann et al. (1995). A hypothetical drinking water
28 scenario of 9 equal drinking water incidents, spaced 2 hrs apart allowing for an 8-hr sleep period,
29 was judged to be a reasonable baseline. Use of this scenario generated an estimated total oral
30 ingestion of 1.1 mg/kg-day of tetrachloroethylene, leading to the same steady-state blood
31 tetrachloroethylene AUC as a continuous inhalation exposure of 0.7 ppm using the PBPK model
32 of Rao and Brown (1993).
33 A route-to-route extrapolation based on a venous blood dose metric is more robust than
34 one based on another dose metric such as the amount of metabolized tetrachloroethylene, and it
35 provides a strong rationale for using blood AUC as a dose metric for extrapolating between
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-21 DRAFT-DO NOT CITE OR QUOTE
-------
1 exposure routes. Venous blood concentration is well-validated in the Rao and Brown (1993)
2 model; hence, little model uncertainty is associated with its estimation (see the pharmacokinetic
3 discussion in Chapter 3). Furthermore, as noted in Section 5.1.1, the use of blood
4 tetrachloroethylene provides some attempt to account for breathing rates and to adjust for kinetic
5 nonlinearities related to tetrachloroethylene absorption, and it is assumed to better reflect
6 tetrachloroethylene pharmacokinetics than use of default methodologies. The chemical species
7 responsible for tetrachloroethylene-induced neurotoxic effects has not been demonstrated, but
8 blood tetrachloroethylene is presumed to be one step in the MOA pathway and is used as a
9 marker for the dose metric associated with neurologic effects.
10 The route-to-route extrapolation starts with the estimation of the average venous blood
11 tetrachloroethylene AUC resulting from continuous inhalation exposure at the LOAEL of 0.7
12 ppm. The venous blood tetrachloroethylene AUC at steady state resulting from continuous
13 exposure to 0.7 ppm tetrachloroethylene is estimated to be 68.3 mg/min/L, according to the Rao
14 and Brown (1993) model. This model does not address pharmacokinetic variation in the human
15 population. An analogous curve corresponding to a drinking water scenario is provided in
16 Figures 5-3 and 3-10. The drinking water scenario models a subject consuming water every
17 2 hrs except during sleep, which was assumed to be for 8 hrs. An assumption of the amount of
18 water consumed is also not necessary, because blood concentrations of tetrachloroethylene are
19 solely dependent on the amount of compound ingested during each drinking episode. The curve
20 depicted in Figures 5-3 and 3-10 yields the same AUC as does continuous inhalation exposure to
21 0.7 ppm and corresponds to ingestion of 76 mg/day. Therefore, the extrapolation of an
22 inhalation exposure of 0.7 ppm (4.8 mg/m3) using the PBPK model of Rao and Brown (1993)
23 yields the same blood concentration of tetrachloroethylene as does ingestion of 1.1 mg/kg-day.
24 Table 5-3 summarizes the results of animal studies of oral exposures that represent the
25 lower end of the dose-response curve. The doses shown in Table 5-3 are expressed in human
26 equivalent terms—using mg/kg3/4-day scaling—to enable interspecies comparisons (U.S. EPA,
27 1992). For liver effects, an alternative procedure of using the pharmacokinetic model of total
28 metabolism as the dose metric for extrapolating between species was carried out. Potential
29 PODs are presented as either a NOAEL or a modeled LEDX when the study results were suitable
30 for modeling. Among the four studies identified in Table 5-2, a significant effect on liver weight
31 is seen in the study by Buben and O'Flaherty (1985), with a NOAEL at a duration-adjusted
32 human equivalent dose of 2 mg/kg-day. The modeled human equivalent BMDLs was 5 mg/kg-
33 day in terms of administered exposure and 3.4-32 mg/kg/day using the available
34 pharmacokinetic models. As discussed above, EPA previously developed an RfD from Buben
35 and O'Flaherty (1985).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-22 DRAFT-DO NOT CITE OR QUOTE
-------
0.10 n
1
2
3
4
5
6
7
8
9
10
11
12
13
10 15
Time (hours)
20
Figure 5-3. Time course of venous blood concentration in humans as
predicted by the Bois et al. (1996), Rao and Brown (1993), and Reitz et al.
(1996) PBPK models for ingested tetrachloroethylene. A total of 76 mg of
tetrachloroethylene was orally delivered via drinking water in nine bolus doses
spaced 2 hrs apart for a duration of 18 hrs, followed by 8 hrs of no dosing. The
dashed line indicates the steady-state blood concentration level due to inhaled
tetrachloroethylene of 0.7 ppm exposure concentration that results in the same
area under the curve as above the curve, integrated over a 24-hr period. The
alveolar ventilation rate was 9.3 L/min (total inspiratory rate 13.9 L/min) and the
ventilation-to-perfusion ratio was equal to 1.3. The three models result in nearly
equal concentrations at this exposure concentration.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-23 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.2.3. Reference Dose (RfD) Derivation, Including Application of Uncertainty Factors
2 The POD of 1.1 mg/kg-day obtained from route-to-route extrapolation from Altmann et
3 al. (1995) is similar to the POD of 2 mg/kg-day observed for liver toxicity in mice orally exposed
4 to tetrachloroethylene (see Table 5-4). Several studies have identified similar PODs for adverse
5 liver effects, indicating some degree of confidence in Altmann et al. (1995) as the basis for the
6 POD and for calculating an RfD. Studies of inhalation exposure show that humans respond to
7 tetrachloroethylene's neurotoxic effects at lower concentrations than do rodents, indicating
8 humans as a potentially more sensitive species for these effects. Humans are also expected to be
9 sensitive to neurotoxicity from oral tetrachloroethylene exposure, and there are no data to
10 indicate otherwise. Hence, the LOAEL of 1.1 mg/kg-day derived from Altmann et al. (1995)
11 represents the POD for the RfD for oral exposure to tetrachloroethylene.
12 To address differences between study conditions and conditions of lifetime human
13 environmental exposure, the POD is reduced by UFs that consider specific areas of uncertainty.
14 The following areas of uncertainty were evaluated for this RfD:
15 1. Human variation. Because the critical study is the same for this RfD derivation as for
16 the RfC, the 10-fold factor applied to the POD for the RfC is used as a default in this
17 assessment for human variation.
18
19 2. Animal-to-human uncertainty. Because the critical study is in humans, this factor is
20 not needed. For animal data, a factor of 3 would be used for extrapolation of
21 pharmacodynamics, because body weight scaling was used for calculating the human
22 equivalent doses (see Table 5-4).
23
24 3. Subchronic-to-chronic uncertainty. As with the RfC derivation, described in Section
25 5.1.3, the POD is based on a study of 10.6-year median duration of exposure, or 15%
26 of a 70-year lifespan. There are no data to suggest that continuing exposure to
27 tetrachloroethylene can increase the severity of effects; exposure-response trends are
28 not evident in the human studies, and there is no systematic trend in the relation
29 between the LOAEL and treatment duration for neurotoxicity in animals. A factor to
30 address the potential for more severe toxicity from lifetime exposure to
31 tetrachloroethylene is not used in this assessment.
32
4. LOAEL-to-NOAEL uncertainty. LOAELs are observed in this human study, and, as
in the RfC derivation, a 10-fold factor is adopted in this assessment to approach the
range where a negligible response could be expected.
33
34 5. Database uncertainty. A factor of 3 is used for the database uncertainties for the
35 same reasons as those for the derivation of the inhalation RfC.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-24 DRAFT-DO NOT CITE OR QUOTE
-------
Table 5-3. Oral studies considered in analysis of the oral RfD
Organ/
system
Liver
Liver,
kidney
Liver
Central
nervous
system
Whole
animal
Study
Buben and
O'Flaherty (1985)
Jonker et al.
(1996)
Herman et al.
(1995)
Fredriksson et al.
(1993)
Hayes et al.
(1986)
Species
Mouse
(40 g)
Rat
Rat
Mouse
Rat
Duration/exposure
route
Subchronic (6
weeks)/oral gavage
Subchronic (4
weeks)/oral gavage
14 days/oral gavage
Post-natal days 10-
16/oral, gavage
Subchronic (91
days)/oral drinking
water
Dose/exposure
(NOAEL/LOAEU)"
(mg/kg-day)
0, 20, 100. 200, 500,
1,000, 1,500, 2,000
0, 600. 2,400
0, 50, 150, 500, 1.500.
5,000
0, 5, 320
0, 0 (vehicle control),
14, 400, 1,400
Effect
Liver weight,
triglycerides
Liver weight, enzyme
levels; kidney weight,
kidney enzyme levels
Liver weight, ALT
Day 60: increased
locomotion, decreased
rearing
Body weightd
Human equivalent doses b
(mg/kg-day)
NOAEL/
LOAEL
NOAEL: 2
LOAEL: 133
NOAEL: 77
LOAEL: 0.5C
NOAEL: 4
BMDLC
BMDLS: 5.0
BMDLS/P: 3.4-32
NA
NA
NA
NA
a NOAELs are underlined once, LOAELs are double-underlined.
b Human equivalent doses calculated using RfD methodology (Barnes and Dourson, 1988) and scaled to the ratio of body weight to the 0.75 power, i.e.,
multiplied by [(animal body weight in kg)/human body weight (70kg)]025. Also adjusted for daily exposure by multiplying by (5 days)/(7 days) where
relevant.
0 BMDL is the lower bound on dose associated with a one standard deviation change in the mean over the control response. This corresponds to an excess risk
of approximately 10% for developing a level of the endpoint above the 98th percentile (or below the 2nd percentile) of the control distribution for normally
distributed effects. BMDLS/P is the BMDLS using pharmacokinetic modeling for relating metabolites in the experimental animals to the responses, and for
reflecting extrapolation to humans (see Appendix 5A for details).
0 For Fredriksson et al. (1993), a regression was fit to the data with body weight estimated as 8g, the body weight at day 13, the midpoint of 10-16 days. The
human equivalent NOAEL is 5 x (0.008/70)°25 = 0.5 mg/kg-day.
d Female rats (LOAEL was 400 mg/kg-day). Authors describe lower body weight gain as significant; however, statistical testing is not presented in the
published paper (only that there was a statistical difference \p < 0.05] between treatment and control groups).
-------
Table 5-4. Oral RfV: point of departure and uncertainty factors
Study
Altmann et al.
(1995): humans
Buben and
O'Flaherty (1985):
mice, increased
liver/body weight
Fredriksson et al.
(1993)
Hayes etal. (1986)
Human
equivalent
concentration"
0.7 ppm
(LOAEL)
NA
NA
NA
Oral
human
equivalent
dose
(mg/kg-day)
l.lb
3.4-32c
0.5
4
Human
variation
10
10
10
10
Animal
to
human
1
3C
3
3
Subchronic
to
chronic
1
10
1
10
LOAEL
to
NOAEL
10
1
10
1
Database
o
J
o
J
3
o
J
Composite
uncertainty
factor
300
1000
1000
1000
RfV
(mg/kg-day)
4 x 10'3
3 x 1(T3 to
3 x 10'2
5 x 1Q-4
4 x lO'3
a Ambient concentration is assumed to represent continuous exposure in the residential studies.
b Equivalent oral exposure from application of PBPK model of Rao and Brown (1993), on the basis of equivalent AUC of blood tetrachloroethylene for humans.
0 See Table 5A-9 for the dose-response modeling summary and extrapolation to human equivalent exposure. These human equivalent doses lie in the nonlinear
portion of the exposure versus rate of metabolism relationship. However, for the purpose of comparing alternate reference doses, the calculation presented
here is equivalent to applying the uncertainty factors prior to conversion from human equivalent metabolized dose to human equivalent administered dose.
-------
1 A total UF of 300 was applied to this effect level: 10 for human variation, 10 for
2 consideration of LOAEL to NOAEL, and 3 for uncertainties in the database.
O
4 RfD =NOAELIUF
5 = 1.1 mg/kg-day / 300
6 = 0.0037 mg/kg-day (rounded to 0.004 mg/kg-day)
7
8 where:
9 1.1 mg/kg-day = the oral exposure POD equivalent of 0.7 ppm (4.8 mg/m3) continuous
10 inhalation exposure LOAEL for neurologic effects in residents
11 exposed tetrachloroethylene (Altmann et al., 1995) estimated via
12 PBPK modeling
13
14 300 = composite UF chosen to account for extrapolation from a LOAEL to a
15 NOAEL in humans, intra-individual variability in humans, and
16 uncertainties in the database
17
18 In summary, an RfD for tetrachloroethylene was developed through a route-to-route
19 extrapolation from the POD in Altmann et al. (1995), which reports neurological toxicity in
20 residents exposed to tetrachloroethylene. The oral exposure POD equivalent to the 0.7 ppm (4.8
21 mg/m3) continuous inhalation exposure LOAEL was estimated via PBPK modeling to be 1.1
22 mg/kg-day. A composite UF of 300 was obtained by multiplying factors of 10 for average to
23 sensitive human variation, 10 for starting from an effect level instead of a NOAEL, and 3 for
24 database uncertainties. Dividing the POD by a composite UF of 300 yields an RfD of 4 xlO"3
25 mg/kg-day. This RfD is equivalent to a drinking water concentration of 0.12 mg/L, assuming a
26 body weight of 70 kg and a daily water consumption of 2 L.
27
28 5.2.4. Supporting Studies
29 PODs and oral RfVs from studies in Table 5-3 are arrayed in Figure 5-4 to give a
30 perspective on the oral RfD supported by Altmann et al. (1995). Effects include liver weight
31 changes (POD [BMDLs/p] of 3.4 mg/kg-day from Buben and O'Flaherty [1985]), body weight
32 changes (POD [NOAEL] of 4 mg/kg-day from Hayes et al. [1986]), and developmental
33 neurotoxicity (POD [LOAEL] of 0.5 mg/kg-day from Fredriksson et al. [1993]). These PODs
34 span a range which includes the oral-equivalent POD (LOAEL) of 1.1 mg/kg-day of Altmann et
35 al. (1995), converted using a pharmacokinetic model for route-to-route extrapolation (Rao and
36 Brown [1993]). The POD from Buben and O'Flaherty (1985) is a standard deviation change in
37 mean liver-to-body weight over control using benchmark dose models and using
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-27 DRAFT-DO NOT CITE OR QUOTE
-------
Oral Organ-Specific RfVs
100
Human
Mouse
Mouse
Rat
KEY
+ Point of departure
(POD)
I I UF, extrapolation from a
lowest-observed-
adverse-effect level
(LOAEL) to a no-
observed-adverse-effect
level (NOAEL)
UF, animal-to-human
extrapolation
UF, human variation
I I UF, extrapolation from
subchronic exposure
duration to chronic
exposure duration
^H UF, database
deficiencies
O Reference value (RfV)
0.0001
aPrincipal study.
CMS, -15% decrement Neuro-
in verbal function, developmental
cognitive function, (Fredriksson
vigilance, and vision3 et al., 1993)
(Altmann et al., 1995)
Liver, one S.D.
change in the mean
liver weight over the
control response
(Buben and
O'Flaherty, 1985)
Systemic
toxicity
(Hayes et al.,
1986)
Figure 5-4. Oral organ-specific reference values for exposure to tetrachloroethylene.
-------
1 pharmacokinetic models for relating metabolites in the experimental animals to the responses,
2 and for reflecting extrapolation to humans (see Appendix 5A and Table 5A-8 for details).
3 Default uncertainty factor values in the absence of data-informed adjustment factors are adopted
4 to account for uncertainties in the analysis (see Table 5-3). A factor of 10 is used to extrapolate
5 from a LOAEL in Altmann et al. (1995), and Fredriksson et al. (1993) to a NOAEL because an
6 effect occurred at the lowest dose studied; this factor is 1.0 for the Buben and O'Flaherty (1985)
7 and Hayes et al. (1986) studies because there was no effect observed at the lowest dose studied.
8 In all studies a factor of 10 is used for human variation and a factor of 3 for database
9 deficiencies. For rodent studies, an uncertainty factor of 3 is adopted to account for uncertainty
10 in extrapolating laboratory animal data to the case of average healthy humans (U.S. EPA, 1994).
11 Typically, this factor addresses residual uncertainties not associated with default dosimetric
12 adjustment like the human equivalent concentration. Scaling experimental doses in the rodent
13 studies to the 3/4th power achieves a human equivalent dose that is considered lexicologically
14 equivalent. For Buben and O'Flaherty (1985), specifically, body weight scaling of the
15 pharmacokinetically derived total rate of metabolism similarly produces a human equivalent
16 dose that is assumed to be lexicologically equivalent. The uncertainty factor of 3 is used to
17 account for residual uncertainty such as pharmacodynamic processes. For studies of subchronic
18 exposure duration, a factor of 10 accounts for uncertainty associated with extrapolating to
19 chronic exposure duration.
20 RfVs in the oral rodent studies of weight changes in Buben and O'Flaherty (1985) and
21 Hayes et al. (1986) were 3 x 10"3 and 4 x 10"3 mg/kg-day, respectively. The RfD of Altmann et
22 al. (1995) at the higher end of the range and with a total uncertainty factor (300) less than total
23 uncertainty factor of 1000 was applied to the two rodent studies. A developmental neurotoxicity
24 study in animals with persistent effects on motor activity yielded a NOAEL of 5 mg/kg-day and
25 a RfV of 5 x 10"4 mg/kg-day (Fredriksson et al., 1993). However, this study was not considered
26 as the principal study for chronic exposure due to uncertainties associated with study design and
27 its level of confidence (see Section 4.6.2.2).
28
29 5.2.5. Previous Oral Assessment
30 EPA previously suggested an RfD of 1 x 10'2 mg/kg-day (U.S. EPA, 1988), which was
31 supported by an adjusted NOAEL of 14 mg/kg-day in Buben and O'Flaherty (1985), and a
32 composite UF of 1,000 (10 for extrapolation from the rat to humans, 10 for human variation, and
33 10 for extrapolating to chronic exposure conditions). A human study is now available, using
34 route-to-route extrapolation, and is preferred to animal data. The composite UF of 300 in the
35 current analysis is smaller than that used in the previous analysis, reflecting fewer uncertainties
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-29 DRAFT-DO NOT CITE OR QUOTE
-------
1 associated with using human data. More recently, Cal EPA (2001) developed a PHG for oral
2 exposure to tetrachloroethylene from the studies by Altmann et al. (1995), Spinatonda et al.
3 (1997), and Ferroni et al. (1992) and conversion factors for breathing and absorption rates for an
4 extrapolation from the inhalation to the oral exposure route. The PHG, calculated by taking the
5 geometric mean of these three studies, of 0.032 mg/kg-day is higher than the current RfD (U.S.
6 EPA, 1987) and the RfD developed in this assessment. On the other hand, ATSDR (1997) did
7 not develop subchronic- and chronic-duration oral MRLs, although an acute MRL was
8 developed from the study by Fredriksson et al. (1993). ATSDR (1997) noted neurological
9 effects as the principal effect of tetrachloroethylene in humans and the scarcity of data in animals
10 from subchronic and chronic studies on this endpoint.
11 One difference between this assessment and the health assessment from California (Cal
12 EPA, 2001) are choices of studies and UFs. The Schreiber et al. (2002) and NYSDOH (2005)
13 studies, which support the findings of Altmann et al. (1995) of neurotoxic effects in residentially
14 exposed populations, were not available at the time of the California assessment. Another
15 difference between this assessment and previous assessments is their treatment of human
16 variation. A choice other than the default of 10 was adopted in the California assessment, based
17 on a presumption that the residential population studied by Altmann et al. (1995) is more
18 reflective of the general adult population than of an occupational population and any
19 accompanying selection bias that is often associated with a healthier worker population.
20 Although there is some merit in this opinion, the observations in Altmann et al. (1995)
21 are of a German population of 14 adults with exposure to tetrachloroethylene. These individuals
22 do not likely represent the full range of human variation found in a large and ethnically diverse
23 population such as the UNITED STATES population. The study excluded subjects with
24 disorders such as hypertension, neurological or endocrinological diseases (e.g., diabetes),
25 impaired vision, or impairment of joints; therefore, the subjects in Altmann et al. (1995) likely
26 had overall good health status. Individuals with diseases that may have increased their
27 susceptibility to tetrachloroethylene effects were not included in the study.
28
29 5.3. UNCERTAINTIES IN INHALATION REFERENCE CONCENTRATION (RFC)
30 AND ORAL REFERENCE DOSE (RfD)
31 Risk assessments need to portray associated uncertainty. The following discussion
32 identifies uncertainties associated with the RfC or RfD for tetrachloroethylene. As presented
33 earlier in this chapter (see Sections 5.1.2, 5.1.3, 5.2.2, and 5.2.3), the uncertainty factor approach,
34 following EPA practices and RfC and RfD guidance (U.S. EPA, 1993, 1994), was applied to a
35 POD, a LOAEL from an epidemiologic study of neurobehavioral effects. Factors accounting for
36 uncertainties associated with a number of steps in the analyses were adopted to account for
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-30 DRAFT-DO NOT CITE OR QUOTE
-------
1 extrapolating the POD, the starting point in the analysis, to a no-adverse-effect concentration or
2 dose (LOAEL to NOAEL) given insufficient data in the principal study for benchmark dose
3 modeling, to a diverse population of varying susceptibilities, and to account for database
4 deficiencies. These extrapolations are carried out with default approaches instead of factors
5 derived from data on tetrachloroethylene given the paucity of experimental tetrachloroethylene
6 data to inform individual steps. As further explained below, limited information is available on
7 human variation in blood tetrachloroethylene concentration and can provide qualitative
8 information on uncertainties associated with human variation. Evaluation of a
9 tetrachloroethylene exposure dose or concentration likely to be without an appreciable risk of
10 chronic adverse health effects over a lifetime and associated uncertainties relies on chemical -
11 specific data to describe dose-response curves, on the breadth of the database for evaluating
12 toxicity in a number of organs, and on characteristics of these data.
13 A broad range of animal toxicology and human epidemiologic data is available for the
14 hazard assessment of tetrachloroethylene, as described throughout the previous section
15 (Chapter 4). Included in these studies are short-term and long-term bioassays in rats and mice
16 (see Table 4-2, Chapter 4); neurotoxicology studies in humans, rats, mice, and gerbils (see
17 Tables 4-4, 4-5, 4-6, and 4-7); prenatal developmental toxicity studies in rats, mice, and rabbits
18 and a two-generation reproduction study in rats (see Table 4-10); and numerous supporting
19 genotoxicity and metabolism studies. Toxicity associated with inhalation exposure to
20 tetrachloroethylene is observed in the liver, kidney, central nervous system, reproductive organs,
21 and the developing fetus (see Chapter 4, Table 5-1, and Figure 5-2). Liver, kidney, and
22 neurodevelopmental effects are observed with oral exposure (see Chapter 4, Table 5-2, and
23 Figure 5-4). Nevertheless, critical data gaps have been identified and uncertainties associated
24 with data deficiencies are more fully discussed below.
25 Neurotoxicity appears to be a sensitive organ system as previously identified in more
26 limited analyses of Rao and Brown (1993) and Guth et al. (1997). The neurotoxic effects
27 observed in a residential population (Altmann et al., 1995) are similar to those observed in
28 occupational populations exposed at higher mean tetrachloroethylene concentration (Seeber,
29 1989, Echeverria et al., 1994). Schreiber et al. (2002) observed visual effects (visual contrast
30 sensitivity) among residents co-located near dry cleaning establishments; however, this study
31 was a pilot for a larger study. The larger study (NYS DOH, 2005) has become available as a
32 final report and appears supportive of this pilot study
33
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-31 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.3.1. Point of Departure
2 A POD based on a LOAEL or NOAEL is, in part, a reflection of the particular exposure
3 concentration or dose at which a study was conducted. It lacks characterization of the dose-
4 response curve and for this reason is less informative than a POD defined as a BMC or a BMD
5 obtained from benchmark dose-response modeling. With respect to neurotoxicity of
6 tetrachloroethylene, benchmark dose-response models are fit to five data sets (Buben and
7 O'Flaherty, 1985; JISA, 1993; NTP, 1986; Brodkin et al., 1995; Tinston, 1994) with sufficient
8 information. The choice of benchmark dose model does not lead to significant uncertainty in
9 estimating the POD since benchmark effect levels were within the range of experimental data.
10 Parameter uncertainty can be assessed through confidence intervals and probabilistic analysis.
11 Each description of parameter uncertainty assumes that the underlying model and associated
12 assumptions are valid. Uncertainty in the animal dose-response data can be assessed through the
13 ratio of BMCs to their BMCLs. These generally do not exceed a factor of two at the POD
14 identified in Tables 5-1 and 5-2.
15 Effects in the CNS and in other organ systems (liver, kidney, reproductive, and
16 developmental) in occupational populations and in animals are observed at higher average
17 tetrachloroethylene concentrations than the Altmann et al. (1995) residential study. As more
18 fully discussed in Section 5.1, uncertainties in other studies of neurotoxicity and of other organ
19 systems differ from those of Altmann et al. (1995). For both occupational and residential
20 populations, studies do not describe a NOAEL and human variation is not well characterized in
21 study subjects. Uncertainties associated with the occupational studies include the following: (1)
22 potential for neurobehavioral effects at lower exposures and (2) exposure pattern differences
23 between occupational and residential studies with peaks characterizing occupational exposures.
24 For animal studies, uncertainties are associated with extrapolating high concentration exposure
25 typically of subchronic duration to genetically inbred rodents to infer a concentration of
26 tetrachloroethylene that is likely to be without an appreciable risk of adverse health effects over a
27 lifetime to a diverse human population.
28
29 5.3.2. Extrapolation from Laboratory Animal Studies to Humans
30 Extrapolating from animals to humans embodies further issues and uncertainties. First,
31 the effect and its magnitude associated with the concentration at the POD in rodents is
32 extrapolated to human response. Pharmacokinetic models are useful to examine species
33 differences in pharmacokinetic processing. This was possible for liver toxicity where limited
34 MO A information suggests metabolism as important to toxicity. The ranges of BMCLs
35 presented for liver effects (a 10-fold range of estimates of tetrachloroethylene metabolism)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-32 DRAFT-DO NOT CITE OR QUOTE
-------
1 demonstrates the uncertainty in tetrachloroethylene pharmacokinetic models. The discrepancies
2 between the models and with experimental data may point to large uncertainties in the
3 parameters used in these models. Because the accuracy of the models has been evaluated only
4 against blood and breath concentrations of the parent compound, their reliability for predicting
5 total metabolites is an unknown. The use of all three of these pharmacokinetic models to provide
6 a range of risk estimates is intended to capture some of this uncertainty.
7
8 5.3.3. Human Variation
9 Heterogeneity among humans is another uncertainty associated with extrapolating doses
10 from animals to humans. Uncertainty related to human variation needs consideration, also, in
11 extrapolating dose from a subset or smaller sized population, say of one sex or a narrow range of
12 life stages typical of occupational epidemiologic studies, to a larger, more diverse population.
13 In the absence of tetrachloroethylene-specific data on human variation, a factor of 10 was
14 used to account for uncertainty associated with human variation. Human variation may be larger
15 or smaller; however, tetrachloroethylene-specific data to examine the potential magnitude of
16 over- or under-estimation are few. The pharmacokinetic model of Clewell et al. (2004) of mean
17 physiological parameters to explore age-dependent pharmacokinetic differences is suggestive of
18 a 2-fold variation in blood tetrachloroethylene levels (see Chapters 3 and 5). Bois et al. (1996),
19 revised by Chiu and Bois (2006), have examined uncertainty and variation in a
20 tetrachloroethylene pharmacokinetic model describing the amount of tetrachloroethylene
21 metabolism. This analysis suggests large uncertainty is associated with estimating the quantity
22 of tetrachloroethylene metabolism in humans.
23
24 5.3.4. Database Uncertainties
25 Critical data gaps have been identified: uncertainties associated with database
26 deficiencies on developmental, immunological, and neurotoxic effects. Most notably, data
27 characterizing dose-response relationships and chronic visual-spatial functional deficits and the
28 cognitive effects of tetrachloroethylene exposure under controlled laboratory conditions are
29 lacking. Several halogenated organic solvents have been linked with altered immune system
30 function in both animals and humans (e.g., toluene, TCE). Additional data from inhalation, oral,
31 and dermal exposures, at different durations, are needed to assess the potential immunotoxicity
32 of tetrachloroethylene. This lack of data combined with the concern that other structurally
33 related solvents have been associated with immunotoxicity contributes to uncertainty in the
34 database for tetrachloroethylene.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-33 DRAFT-DO NOT CITE OR QUOTE
-------
1 Data from acute studies in animals (Warren et al., 1996; Umeza et al., 1997) suggest that
2 cognitive function is affected by exposure to tetrachloroethylene. These studies do not address
3 the exposure-response relationship for subchronic and chronic tetrachloroethylene exposures on
4 cognitive functional deficits observed in humans (e.g., Seeber, 1989; Echeverria et al., 1994; and
5 Altmann et al., 1995). Even more importantly, there is a lack of cognitive testing in both
6 developmentally exposed animals and adult animals following exposures to tetrachloroethylene
7 that are longer than acute durations of exposure. Visual system dysfunction and processing of
8 visual spatial information are sensitive endpoints in human studies. The exposure-response
9 relationship of these functional deficits could be evaluated more definitively with studies using
10 homologous methods that examine retinal and visual function in experimental animals.
11 However, there has been a limited evaluation of visual function in rodents, with the exception of
12 the evoked potential studies by Mattsson et al. (1998). These types of studies could help
13 determine whether there are both peripheral and central effects of tetrachloroethylene exposure
14 on visual perception, and they could be used as an animal model to better define the exposure-
15 response relationships.
16 Subjects in the epidemiologic studies comprise adults, and some characterization of the
17 response of children to tetrachloroethylene exposure was found in limited data for a similar
18 neurological (visual system) parameter (Schreiber et al., 2002) and in a larger number of subjects
19 (NYS DOH, 2005) using other visually based testing paradigms. Additionally, in a postnatal
20 neurotoxicity study in mice (Fredriksson et al., 1993), persistent neurological effects (i.e.,
21 increased locomotion and decreased rearing behavior at 60 days of age, measured 43 days after
22 exposure ceased) were observed at an oral dose of 5 ppm, with no NOAEL, although this study
23 did not conform to traditional toxicity testing guidelines (see Section 4.6.2.2). There is
24 uncertainty that if adequate, robust, dose-response data based on the most appropriate
25 neurophysiological and cognitive tests were available, the exposure eliciting an adverse response
26 (and hence the POD for the reference value) could be lower than that established on the basis of
27 deficits in visuo-spatial and cognitive function following tetrachloroethylene exposure in healthy
28 adults (Altmann et al., 1995).
29
30 5.4. CANCER DOSE-RESPONSE ASSESSMENT
31 The following dose-response assessment was developed following the guidelines in the
32 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a). As discussed in Section 4.10.1,
33 there is some indication from epidemiologic investigations that a human cancer risk is associated
34 with exposure to tetrachloroethylene. Sufficient human data linked with exposure
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-34 DRAFT-DO NOT CITE OR QUOTE
-------
1 characterizations from these studies have not been available, but estimating cancer risk from
2 these studies might be feasible in the future.
3 Further, as detailed in Section 4.10.4., the available body of MO A information is not
4 sufficient to derive quantitative, biologically based, or toxicodynamic models for low-dose
5 extrapolation from animal data. Moreover, current literature does not identify a nonlinear MOA
6 for tetrachloroethylene carcinogenicity. Therefore, consistent with the 2005 cancer guidelines, a
7 default low-dose linear model is indicated for use with the animal data to estimate human cancer
8 risk.
9 There is evidence that one or more tetrachloroethylene metabolites may be involved in
10 some of the carcinogenicity associated with tetrachloroethylene exposure (see Section 4.10.4).
11 PBPK models are available to estimate total metabolism in laboratory rodents and humans from
12 inhalation and oral exposure to tetrachloroethylene. The dose-response discussion below
13 describes where the PBPK models have been used to estimate human carcinogenic risk arising
14 from tetrachloroethylene exposure through their impact on high-dose to low-dose extrapolation
15 in animals, interspecies extrapolation, and route-to-route extrapolation.
16
17 5.4.1. Choice of Study/Data with Rationale and Justification
18 As discussed in Chapter 4, there are several chronic studies in rats and mice: an oral
19 gavage study in mice and female rats by NCI (1977) and two inhalation studies in mice and rats
20 (NTP, 1986; JISA, 1993). These studiesestablished that the administration of
21 tetrachloroethylene, either by ingestion or by inhalation to sexually mature rats and mice, results
22 in increased incidence of tumors. In at least two studies, several tumor sites showed statistically
23 significantly increased rates with increasing tetrachloroethylene administration: MCL in male
24 and female rats and hepatocellular adenomas and carcinomas in male and female mice. Other
25 cancer dose response assessments of tetrachloroethylene have relied on the tumor data from the
26 NCI and NTP studies (see Appendix A).
27 This analysis considers all three bioassays but focuses primarily on the JISA (1993) study
28 results. First, the JISA (1993) study included lower exposures than did the two earlier bioassays
29 for both species tested, which makes it a stronger study for deriving dose-response relationships
30 for risk assessment purposes, insofar as all other aspects of these studies can be considered
31 comparable. For mice, the lowest exposure concentration of 10 ppm was 10-fold lower and the
32 mid-dose of 50 ppm was 2-fold lower than the lower exposure concentration in the NTP (1986)
33 inhalation study (100 ppm). For rats, the low-exposure concentration of 50 ppm was fourfold
34 lower than in the NTP study (200 ppm). Second, no other dose response modeling appears to be
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-35 DRAFT-DO NOT CITE OR QUOTE
-------
1 available for the JISA (1993) study, whereas the incidence of hepatocellular tumors and MCL in
2 the NTP (1986) study have been extensively analyzed for previous assessments.
O
4 5.4.2. Dose-Response Data
5 5.4.2.1. Liver Tumors in Mice
6 All three bioassays showed increases in hepatocellular tumors in male and female mice.
7 Table 5-5 summarizes these incidence patterns. Because hepatic adenomas and carcinomas are
8 considered part of the same continuum of tumor development, and adenomas may be
9 differentiated from carcinomas only on the basis of size, this analysis emphasizes the combined
10 incidence of these two tumor types. Historical data from the Japan Bioassay Research Center
11 (JBRC), where the JISA (1993) study was conducted, indicate that the control liver tumor
12 incidences in this study were fairly typical for this laboratory (see Table 5-6). Specifically, the
13 incidence in controls was 28% for males and 6% for females; the averages for the laboratory
14 were 23% and 2% and the upper bounds were 42% and 8%, respectively, for carcinomas.8
15 The results of the inhalation studies are reasonably consistent when adjusted for
16 background tumor incidence (see Figures 5-5a and 5-5b). Liver tumor incidence among male
17 mice in the JISA (1993) study did not follow a clearly monotonic pattern, with a higher response
18 in the low-dose group than seems consistent with the pattern in responses in the other dose
19 groups. Taking into account this variability in the responses, however, the dose-response
20 patterns for the male and female mice in the NTP (1986) and JISA (1993) studies appear
21 reasonably concordant.
22 Several issues complicate comparisons of the NCI (1977) gavage study results with those
23 of the other chronic bioassays. First, dosing lasted 78 weeks rather than 104 weeks as in the
24 inhalation studies, so in making direct comparisons it might be expected that the observed tumor
25 incidence in the NCI (1977) study would underestimate the incidence associated with 104 weeks
26 of exposure. Second, this oral gavage study had a variable dosing schedule, with doses that were
27 increased by 100 mg/kg-day in the low-dose group and by 200 mg/kg-day in the high-dose group
28 after 11 weeks of study. Consequently, association of a constant level of exposure with the
29 observed effects must be inferred rather than measured. In addition, surviving animals were
30 maintained without further exposure until final sacrifice in week 90.
31 The NCI (1977) exposures were recalculated on a basis consistent with other EPA
32 estimates of chronic toxicity, in which the cumulative exposure received over the full period of
o
Combined historical incidence of adenomas and carcinomas was not available. Presumably the incidence
of carcinomas slightly underestimates the combined incidence of adenomas and carcinomas.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-36 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-5. Tumor incidence and estimated metabolized doses in mice
exposed to tetrachloroethylene
Bioassay
Doses/Exposures
Administered
Continuous
Equivalent
Sex
Body
Weight3
(kg)
Total
Metabolism1"
(mg/kg-day)
Survival-
Adjusted
Tumor
Incidence0
(%)
Hepatocellular adenomas and carcinomas
NCI (1977)d
B6C3FJ mice
Gavage:
5 days/wk,
78wks
NTP (1986)
B6C3FJ mice
Inhalation:
6 hrs/day,
5 days/wk,
104 wks
JISA (1993)
Crj:BDFl mice
Inhalation:
6 hrs/day,
5 days/wk,
104 wks
Vehicle control
450 mg/kg-day
900 mg/kg-day
Vehicle control
300 mg/kg-day e
600 mg/kg-day
Oppm
100 ppm
200 ppm
0 ppm
100 ppm
200 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oe
332 mg/kg-day
663 mg/kg-day
Oe
239 mg/kg-day
478 mg/kg-day
0
18 ppm
36 ppm
0
18 ppm
36 ppm
0
1.8 ppm
9.0 ppm
45 ppm
0
1.8 ppm
9.0 ppm
45 ppm
Male
Female
Male
Female
Male
Female
0.030
0.025
0.037
0.032
0.048
0.035
0
35
47
0
34
46
0
27
41
0
31
45
0
3.4
14
36
0
4
18
47
2/20 (10)
32/48 (67)
27/45 (60)
0/20 (0)
19/48 (40)
19/45 (42)
17/49 (35)
31/47 (70)
41/50 (82)
4/45 (9)
17/42 (40)
38/48 (79)
13/46 (28)
21/49 (43)
19/48 (40)
40/49 (82)
3/50 (6)
3/47 (6)
7/48 (15)
33/49 (67)
Malignant hemangiosarcomasf, liver or spleen
JISA (1993)
Crj:BDFl mice
6 hrs/day,
5 days/wk,
104 wks
Oppm
10 ppm
50 ppm
250 ppm
0
1.8 ppm
9.0 ppm
45 ppm
Male
0.048
0
3.4
14
36
2/46 (4)
1/49 (2)
6/48 (13)
9/49 (18)
4
5
6
7
8
9
10
11
12
13
14
15
16
a Average body weight reached during adulthood.
b As calculated using the Reitz et al. (1996) pharmacokinetic model for mice, using alveolar ventilation rate, at
67% of total ventilation (see Section 3.5). Total metabolism was estimated from the simulated bioassay exposure
pattern, i.e., the amount estimated to be metabolized following an increment of exposure (gavage dose or 6-hr
inhalation exposure). Adjustment for continuous exposure followed by multiplying the exposure by (5 days/7
days). Figure 3-9 illustrates the correspondence of total metabolism with administered exposure estimated by this
model for mice weighing 0.025 kg: at 100 ppm, approximately 47 mg-equivalent (eq)/kg-day of metabolites are
estimated to be produced. For the purposes of this assessment, this is assumed to be equivalent to 47 mg-eq/kg-
day x 5/7 = 34 mg-eq/kg-day of metabolites resulting from continuous exposure. Note that this level is higher
than the 31 mg-eq/kg-day estimated for 0.032 kg female mice and the 27 mg-eq/kg-day for 0.037 kg male mice in
the NTP (1986) study in this table, illustrating the dependence of the PBPK model on body size. This dependence
is not tabulated or graphed in this document.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-37 DRAFT-DO NOT CITE OR QUOTE
-------
1 Table 5-5. Tumor incidence and estimated metabolized doses in mice
2 exposed to tetrachloroethylene (continued)
O
4 ° Animals dying before the first appearance of the tumor of interest but no later than week 52 were omitted from
5 the totals because these animals were presumed not to have adequate time on study to develop tumors.
6 d No adenomas were reported in this study. Because hepatic adenomas and carcinomas are considered part of the
7 same continuum of tumor development, and adenomas have been distinguished from carcinomas only on the basis
8 of size, the correspondence of this observation to the other studies is not clear.
9 e Gavage doses listed were increased after 11 weeks by 100 mg/kg-day in each low-dose group or by 200 mg/kg-
10 day in each high-dose group. Animals surviving the 78-week exposure period were observed until the week 90
11 study termination. Lifetime average daily (administered) doses (LADDs) were calculated as follows:
12
13 LADD (mg/kg-day) = Cumulative administered dose (mg/kg)/(total days on study)
14 = {[(initial dose level x 11 weeks) + (increased dose level x 67 weeks)]/90 weeks}
15 x (5 days/7 days)
16
17 Male mice received LADDs of 332 and 663 mg/kg-day, and female mice received 239 and 478 mg/kg-day.
18 f These tumors were reported as hemangioendotheliomas in the JISA (1993) report. The term has been updated to
19 hemangiosarcoma. Note that these incidences do not match those tabulated in Table 12 of the JISA report
20 summary. The incidences reported here represent a tabulation of malignant hemangioendotheliomas from the
21 individual animal data provided in the JISA report.
22
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-38 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-6. Historical control data of the Japan Bioassay Research Center,
Crj/BDFl mouse, 104-week studies
Tumor types
Inhalation, feeding, and drinking
studies (19 studies)
Total incidence
(%)
Range
(%)
Inhalation studies only (9 studies)
Total incidence
(%)
Range
(%)
Male mice
Liver
hepatocellular adenoma
hepatocellular carcinoma
Spleen
hemangioma3
hemangiosarcoma3
165/947 (17.4)
215/947 (22.7)
17/946 (1.8)
30/946 (3.2)
4.0-34.0
2.0-42.0
0-10.0
0-8.0
92/448 (20.5)
105/448 (23.4)
8/448 (1.8)
12/448 (2.7)
10.0-30.6
10.0-36.7
0-8.0
0-6.0
Female mice
Liver
hepatocellular adenoma
hepatocellular carcinoma
Spleen
hemangioma3
hemangiosarcoma3
50/949 (5.3)
22/949 (2.3)
8/949 (0.9)
3/949 (0.3)
2.0-10.0
0-8.0
0-6.0
0-2.0
18/449 (4.0)
14/449 (3.1)
5/449 (1.1)
3/449 (0.7)
2.0-6.0
0-8.0
0-6.0
0-2.0
4
5
6
7
3 The terms "hemangioendothelioma: benign" and "hemangioendothelioma" in the original study have been
changed to "hemangioma" and "hemangiosarcoma," respectively.
Source: Attachment to letter dated September 5, 2001, from K. Nagano, Japan Bioassay Research Center, Japan
Industrial Safety and Health Association, to R. McGaughy, U.S. EPA. Available from hotline.iris@epa.gov.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-39 DRAFT-DO NOT CITE OR QUOTE
-------
Male Mice
Female Mice
t°
0)0'
to
§-
Q.
W 02 -
£
Administered concentration (ppm)
o
iO.8
§.,
2J0.6
£
_§°5
0)0.4
W>
§«
Q.
0)0.2
0)
Administered concentration (ppm)
Male Mice
Female Mice
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
s,
§».H
3 '
C
|0.5
0)0.4
to
£o.3j
Total metabolism (mg-equiv/kg/d)
C
00.7
(A
§0,
Q.
»0.2
£..
d.
Total metabolism (mg-equiv/kg/d)
Figure 5-5. Mouse liver tumor responses (hepatocellular adenomas and
carcinomas) for three chronic bioassays (Table 5-5), plotted against
continuous equivalent concentration (ppm) and total tetrachloroethylene
metabolism (mg-equivalents/kg-day), for male and female mice.
observation is prorated to obtain an average lifetime daily exposure. Table 5-5 provides the
recalculated exposures. It is not clear in this case, however, that a simple TWA over the period
of observation is the most suitable representation of tetrachloroethylene exposure in the NCI
(1977) study, due to the substantial changes in the dosing pattern, as noted above. In addition,
mortality was significantly increased in both treated groups over that of controls, suggesting that
the maximum tolerated dose had been exceeded. Note that although no adenomas were reported
in the NCI (1977) study, some of the reported carcinomas may have been adenomas. This factor
should be taken into account when comparing the incidence of carcinomas in the NCI (1977)
study with the combination of adenomas and carcinomas in the inhalation studies (see
Table 5-5). Consequently, it was not feasible to compare this dose-response with that from the
inhalation studies on an administered mg/kg-day basis.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-40 DRAFT-DO NOT CITE OR QUOTE
-------
1 In addition to hepatocellular adenomas and carcinomas, the JISA (1993) study
2 demonstrated increased hemangiomas and hemangiosarcomas in the liver and spleen of mid- and
3 high-dose male mice (see Table 5-5; Cochran-Armitage trend test, p = 0.004). The incidence in
4 control and low-dose male mice was similar to the JBRC historical control incidence for spleen
5 only (3.2%, range 0-8%; see Table 5-6). This finding was not replicated in the NCI (1977) or
6 NTP (1977) studies (tumors noted in the NTP male mice livers: controls, 3/49; low dose, 2/49;
7 high dose, 2/50; tumors noted in the NTP historical controls for all sites: 4.4%, range 2-8%
8 (http://ntp.niehs.nih.gov/ntp/research/database searches/historical controls/path/m inhar.txt).
9 Both the JBRC and NTP historical controls showed similar background levels of hemangiomas
10 and hemangiosarcomas, although the JBRC data included only the spleen, whereas the NTP data
11 included all sites.
12 Because tetrachloroethylene's metabolites have been implicated in its liver toxicity (see
13 Section 4.10.4.1), and because a pharmacokinetic model was available to estimate metabolism
14 levels in mice (based on the work of Reitz et al., 1996; see Section 3.5), the hepatocellular tumor
15 responses in the three chronic bioassays were compared in terms of total metabolism of
16 tetrachloroethylene (see Figures 5-5c and 5-5d). Here it can be seen that the hepatocellular
17 tumor dose-response in the gavage study appears to be quite similar to that of the inhalation
18 studies. Note further that, from an empirical point of view, the dose-response patterns for the
19 inhalation studies collectively appear to follow an approximately linear relationship, whether the
20 exposure measure is the administered concentration or total metabolism. In other words, these
21 data do not clearly suggest one dose metric over the other as being more closely associated with
22 the liver tumors or that some other dose metric would be preferred for characterizing cancer
23 incidence in this range of exposure.
24
25 5.4.2.2. Mononuclear Cell Leukemia in Rats
26 The NTP (1986) and JISA (1993) studies demonstrated increased MCL incidences for
27 male and female rats (see Table 5-7). The NCI study did not demonstrate any MCL increases in
28 rats. However, the investigators considered this study inconclusive because of low survival, so
29 the NCI study neither confirms nor refutes the findings of the NTP and JISA studies.
30 The responses in the NTP (1986) study were approximately twofold higher than for the
31 corresponding groups in the JISA (1993) study, including the control groups. Control groups for
32 both laboratories were consistent with their respective historical controls (see Table 5-8 for the
33 JISA historical controls). Like the hepatocellular tumor results in mice (see Section 5.4.2.1), the
34 MCL results from the NTP and JISA studies were plotted in terms of additional risk versus
35 administered concentration (see Figure 5-6). Note that MCL risk has been considered previously
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-41 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
Table 5-7. Incidence of mononuclear cell leukemia, kidney tumors, and
brain gliomas in rats exposed to tetrachloroethylene by inhalation
Bioassay
Exposure concentration
(ppm)
Administered
Continuous
equivalent
Sex
Body
weight (kg)
Total
metabolism"
(mg/kg-day)
Su rvival-ad justed
tumor incidence1"
(%)
Mononuclear cell leukemia
NTP (1986)
F344/N rats
Inhalation
6 hrs/day,
5 days/wk,
104 wks
JISA (1993)
F344/DuCrj rats
Inhalation
6 hrs/day,
5 days/wk,
104 wks
0
200
400
0
200
400
0
50
200
600
0
50
200
600
0
36
72
0
36
72
0
9
36
108
0
9
36
108
Male
Female
Male
Female
0.44
0.32
0.45
0.3
0.0
3.6
5.0
0.0
5.1
6.9
0.0
1.4
3.6
6.1
0.0
2.0
5.1
8.7
28/50 (56)
37/48 (77)
37/50 (74)
18/50 (36)
30/50 (60)
29/50 (58)
11/50(22)
14/50 (28)
22/50 (44)
27/50 (54)
10/50 (20)
17/50 (34)
16/50 (32)
19/50 (38)
Kidney: tubular cell adenoma or adenocarcinoma
NTP (1986)
JISA (1993)
0
200
400
0
50
200
600
0
36
71
0
9
36
110
Male
Male
0.44
0.45
0.0
3.6
5.0
0.0
1.4
3.6
6.1
1/49 (2)
3/47 (6)
4/50 (8)
1/50 (2)
2/50 (4)
1/50 (2)
2/50 (4)
Brain gliomas
NTP (1986)
JISA (1993)
0
200
400
0
50
200
600
0
36
71
0
9
36
110
Male
Male
0.44
0.45
0.0
3.6
5.0
0.0
1.4
3.6
6.1
1/50 (2)
0/48 (0)
4/50 (8)
2/50 (4)
0/50 (0)
0/50 (0)
0/50 (0)
4
5
6
7
8
9
10
11
12
13
14
15
16
a As calculated by the Reitz et al. (1996) pharmacokinetic model for rats using alveolar ventilation rate at 67% of total
ventilation (see Section 3.5). Total metabolism was estimated from the simulated bioassay exposure pattern, that is,
the amount estimated to be metabolized following an increment of exposure (6-hr inhalation exposure). Adjustment
for continuous exposure followed by multiplying the exposure by (5 days/7 days). Figure 3-9 illustrates the
correspondence of total metabolism with administered exposure estimated by this model for rats weighing 0.3 kg: at
200 ppm, approximately 7.1 mg-equivalent (eq)/kg-day of metabolites are estimated to be produced. For the
purposes of this assessment, this is assumed to be equivalent to 7.1 mg-eq/kg-day x 5/7 = 5.1 mg-eq/kg-day of
metabolites resulting from continuous exposure (see metabolite levels above for the JISA female rats). Note that
this level is higher than the 3.6 mg-eq/kg-day estimated for 0.45 kg rats in the JISA study, illustrating the
dependence of the PBPK model on body size. This dependence is not tabulated or graphed in this document.
b Animals dying before the first appearance of the tumor of interest but no later than week 52 were omitted from the
totals because these animals were presumed to have had inadequate time on study to develop these tumors.
Sources: NTP (1986) and JISA (1993).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-42 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-8. Historical control data of the Japan Bioassay Research Center,
F344/DuCrj (Fischer) rat, 104-week studies
Tumor types
Inhalation, feeding, and drinking
studies
(23 studies)
Total incidence (%)
Range
(%)
Inhalation studies only
(11 studies)
Total incidence
(%)
Range
(%)
Male rats
Mononuclear cell leukemia
Kidney
Renal cell adenoma
Renal cell carcinoma
147/1149(12.8)
2/1149(0.2)
2/1149(0.2)
6.0-22.0
0-2.0
0-2.0
76/549 (13.8)
1/549 (0.2)
2/549 (0.4)
6.0-22.0
0-2.0
0-2.0
Female rats
Mononuclear cell leukemia
Kidney
Renal cell adenoma
Renal cell carcinoma
147/1048 (14.0)
1/1048(0.1)
0/1048 (0.0)
2.0-26.0
0-2.0
NA
68/448 (15.2)
1/448 (0.2)
0/448 (0.0)
8.0-20.0
0-2.0
NA
4
5
Source: Attachment to letter from K. Nagano to R. McGaughy 9/5/01. Available from IRIS Information Desk.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-43 DRAFT-DO NOT CITE OR QUOTE
-------
Male Rats
Female Rats
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
0.4 -i
^*
O
-go.3
o
o
to
3
c
I02
0)
to
c
Q.0.1
to
a:
n n s
-A- NTP, 1986
-e-JISA, 1993
^__^°
^^_^-~-"~~~~~^
^^_^-—-~~"'^
%C^^
/y~~~~~~~~-t.
//
//
//
//
r
/ a.
«,,,,,,,,,,,
0.4 -i
^*
O
"C 0.3
O
o
to
3
C
I02
0)
to
c
o
Q0.1
to
-------
1 For female rats (see Figures 5-6b and d), the responses adjusted for control rates do not
2 show as much concordance between studies as those for the male rats, with the JISA study still
3 showing lower adjusted responses than the NTP study at comparable exposures. The low-dose
4 females in the JISA study had a higher response relative to the pattern of responses in the other
5 three groups and a higher response than would be expected from the dose-response pattern in the
6 NTP study. The dose-response relationship for the JISA study appears to be slightly more linear
7 for total metabolism than for administered concentration; however, both studies suggest some
8 degree of saturation of effects in the available range of the dose metrics considered. Although
9 F344 rats were used in both studies, it is possible that there could be some differences
10 attributable to the specific lines of animals used at each laboratory and laboratory-specific
11 procedures.
12
13 5.4.2.3. Other Tumor Sites in Male Rats
14 Other elevated tumor incidences—brain gliomas and kidney tubule adenomas and
15 adenocarcinomas—were observed in male F344/N rats in the NTP study but not in the JISA
16 study (again, there were no corresponding data available for the NCI male rats). Table 5-7
17 summarizes the incidence data from both laboratories for these sites. Brain gliomas in rats in the
18 NTP inhalation study were elevated. In males, the incidences were 1/50, 0/48, and 4/50 in the
19 control, 200 ppm, and 400 ppm tetrachloroethylene groups, respectively. This was a statistically
20 significant dose-related trend by the life table test (p = 0.039) but not by the incidental tumor
21 trend test. A similar trend was seen in female rats (1/50, 0/50, 2/50), but this trend was not
22 statistically significant. Brain gliomas are rare tumors in NTP rat bioassays; in male rats the
23 historical control incidence is 2/247 (0.8%) in the laboratory where this study was conducted,
24 and 4/1971 (0.2%) in the overall program. Because these tumors had not been observed in
25 previous NTP studies of tetrachloroethylene, trichloroethylene, or pentachloroethane, and
26 because they appeared in the untreated groups, the NTP investigators concluded that they were
27 not related to tetrachloroethylene exposure.
28 On the other hand, the previous study of tetrachloroethylene was the NCI (1977) gavage
29 study, not an inhalation study; route-to-route differences are not implausible. Although the other
30 solvents mentioned have been associated with similar adverse effects, there are other differences
31 among them. Also, the brain tumors in the high-dose males started occurring earlier (weeks 88,
32 96, 102, and 103) than in the control group (99 weeks), and in the high-dose females they
33 occurred even earlier (75 and 78 weeks in the high-dose group vs. 104 weeks in the control
34 group). Therefore, although the association between tetrachloroethylene exposure and brain
35 gliomas in rats in this data set is not strong, it is still suggestive, especially considering that the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-45 DRAFT-DO NOT CITE OR QUOTE
-------
1 nervous system is a target of tetrachloroethylene exposure in humans and animals (see Sections
2 4.5.3 and 5.1.1).
3 In the JISA study, brain gliomas were observed only in male control rats. Historical
4 brain tumor incidence data for this laboratory were not available, however. Although F344 rats
5 were used in both studies, it is possible that there could be some differences between laboratories
6 attributable to the specific lines of animals used and laboratory-specific procedures. Given the
7 low overall incidences relative to the other tumor sites, these data were not modeled.
8 Kidney tubule cell adenomas and adenocarcinomas (see Table 5-7) were elevated in the
9 exposed male rats, but they were not statistically significantly elevated. Statistical significance is
10 a secondary consideration in determining the biological significance of these tumors because
11 they are considered to be uncommon in NTP studies of rats. The investigators noted that these
12 tumors were observed among historical controls at about 0.2% in 1968 untreated control rats.
13 Further support for considering the relevance of this site comes from the evidence relating
14 trichloroethylene and rat kidney tumors (U.S. EPA, 2001).
15 There was no apparent trend in the incidence of kidney tubule cell adenomas and
16 adenocarcinomas among JISA male rats. The incidence in all groups was consistent with JISA
17 historical control data (see Table 5-8).
18
19 5.4.3. Estimation of Dose Metrics for Dose-Response Modeling
20 The sequence of steps in estimating human equivalent risks is illustrated in Figure 5-7,
21 with estimation of the dose inputs to the dose-response modeling being the first step.
22 Considerations for estimating continuous exposure levels equivalent to the intermittent animal
23 bioassay exposures differ according to whether administered exposures or metabolized doses
24 were used as the measure of dose, and they are discussed following the selection of each dose
25 metric.
26
27 5.4.3.1. Dose Metric for Hepatocellular Carcinogenicity
28 There are several possible rationales to consider for the appropriate dose metric for
29 tetrachloroethylene-induced liver toxicity and carcinogenicity. The specific chemical species
30 responsible for adverse effects on the liver would be the preferred choice. As discussed in
31 Chapter 3 and Section 4.10.4.1, several metabolites associated with P450 metabolism occurring
32 in the liver have been identified. TCA, which is associated with liver toxicity when administered
33 directly, is considered a key product of this P450 oxidation pathway. However, the MOA for
34 tetrachloroethylene-induced liver toxicity and carcinogenicity is not clear. Further, TCA does
35 not appear to explain the liver carcinogenicity observed with tetrachloroethylene, because a
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-46 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Administered dose/
concentration in bio assay
Animal PBPK model
Human equivalent
unit risk (per ppm) or
slope factor
(p er m g/kg B W- day)
Total metabolism
in anim als, m g- eq/kg B W- day
BMR - POD
Continuous exposure
adjustment for 5/7 days
Equivalent continuous total
metabolism
in anim als, m g- eq/kg B W- day
Human equivalent point
of departure, in
continuous environmental
exp osure level, in ppm
(inhalation) or
m g/k g B W- day (oral)
Dose-response model,
with, response data
Human PBPK model
Point of departure, in
m g- eq/kg B W- day of total
metabolism
Bw scaling
Human equivalent point
of departure,
in m g- e q/k g B W-day of
total m etabolism
Figure 5-7. Sequence of steps for extrapolating from tetrachloroethylene
bioassays in animals to human-equivalent exposures expected to be
associated with comparable cancer risk.
comparison of hepatocellular tumor incidence associated with direct TCA exposure appears to
underpredict the hepatocellular carcinoma incidence in the NTP and JISA studies when
characterized in terms of equivalent TCA exposures (see Appendix 4A in Chapter 4).
Additional metabolites could play a role. For instance, reactive intermediates such as
tetrachloroethylene oxide or trichloroacetyl chloride are hypothesized to be precursors to TCA,
and as reactive compounds they would be likely candidates. However, their involvement in liver
toxicity remains unknown, and they have not been confirmed in the tetrachloroethylene
metabolic pathways. Consequently, although it appears plausible that at least another compound
besides TCA contributes to tetrachloroethylene-induced hepatocarcinogenicity, none has been
identified nor can the amounts be estimated.
Because of the uncertainty over which metabolite species are involved in causing liver
toxicity—and to what degree they are involved—total metabolism was considered the most
appropriate dose metric. Use of this dose metric does not require assuming that all of the
metabolites are responsible for tetrachloroethylene's liver carcinogenicity, however, only that the
rate of total metabolite production is proportional to the actual target dose in the target tissue, at
least at very low exposures. That is, if there is a constant relationship between the surrogate dose
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-47 DRAFT-DO NOT CITE OR QUOTE
-------
1 measure and the relevant measure of the actual carcinogen, at least at very low exposures, then
2 dose-response modeling of each measure (if both were available) would yield the same cancer
3 risk value in terms of environmental exposure. For instance, the concentration of the actual liver
4 carcinogen(s) could be in equilibrium with the circulating blood, or it could be primarily
5 generated and active in the liver; as long as proportionality of the surrogate dose measure is a
6 reasonable assumption, cancer risk values can be estimated and interpreted accordingly.
7 Consequently, the daily production rate of all metabolites of tetrachloroethylene corresponding
8 to the bioassay exposure patterns, as estimated using the PBPK model of Reitz et al. (1996), was
9 used for the dose-response modeling of mouse liver tumors.
10 The second adjustment made prior to dose-response modeling was to characterize the
11 intermittent bioassay exposures in terms of equivalent continuous exposure. Because the
12 pharmacokinetic model cannot generate an AUC for total metabolism, due to the lack of
13 clearances for the individual metabolites, the daily production rate of all metabolites for five
14 days was averaged over seven days by multiplying it by 5/7 (0.71), under the assumption that
15 concentration multiplied by time maintains a constant effect (C x t = k\ is likely to hold for very
16 low tetrachloroethylene exposures. The metabolism rates reported in Tables 5-5 and 5-7 reflect
17 this adjustment.
18
19 5.4.3.2. Dose Metric for Rat Leukemias and Kidney Tumors
20 Experimental evidence suggests that a GST-pathway metabolite (TCVC) is more likely to
21 be associated with the kidney tumors and possibly the leukemias than the P450 pathway (see
22 Sections 4.9.4.2 and 4.9.4.3). However, the measurements of glutathione-dependent metabolism
23 are from in vitro studies or they are measures of urinary excretion products and are, therefore,
24 not representative of the toxic species in vivo. Consequently, insufficient data exist to
25 incorporate the GST-derived metabolites explicitly in the PBPK models.
26 Given the approximately linear dose-response relationship observed between leukemias
27 and total metabolism for male rats (see Figure 5-6c), it appears plausible that the carcinogen
28 responsible for the leukemias may be approximately proportional to total metabolism. The
29 situation is somewhat less clear for female rats due to the nonmonotonic dose-response patterns,
30 although the degree of saturation was less pronounced when the dose-response relationship was
31 considered in terms of total metabolism (Figure 5-6d). Accordingly, total metabolism was
32 considered a better surrogate than administered concentration for the proximate carcinogen.
33 Because kidney tumors are associated with the same GST-metabolite as the leukemias—
34 somewhat more definitively than the leukemias—total metabolite production was also
35 considered as a dose metric for estimating the male rat kidney dose-response relationship.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-48 DRAFT-DO NOT CITE OR QUOTE
-------
1 Adjustment for continuous exposure was the same as for the liver tumors (see Section
2 5.4.3.1). That is, the only continuous exposure adjustment to the total metabolite dose metric
3 that was needed was to average the metabolic rate for five days over seven days (by multiplying
4 each by 5/7).
5
6 5.4.3.3. Dose Metric for Sites Not Addressed by Physiologically Based Pharmacokinetic
1 (PBPK) Modeling
8 For tumor sites for which the MOA is not clear, such as for male rat brain tumors (see
9 Section 4.10.4), the administered concentration of tetrachloroethylene was used as the dose
10 metric. Use of this dose metric should provide plausible results as long as the concentration of
11 the proximate carcinogen(s) is proportional to administered concentration—at least at low
12 concentrations. Because there is uncertainty in the identification of the carcinogenic agent for all
13 the sites, the dose-response relationships for all tumor sites were also estimated using this default
14 dose metric, for comparison purposes.
15 Allowance for extrapolation to continuous exposures was made before dose-response
16 modeling. In all cases, administered inhalation concentrations (in ppm) were adjusted for
17 continuous exposure by averaging the five 6-hr daily exposures over the full week. That is,
18 administered concentrations were multiplied by 6 hrs/24 hrs x 5 days/7 days (0.179) to yield
19 equivalent continuous concentrations. Tables 5-5 and 5-7 provide these adjusted concentrations.
20
21 5.4.4. Extrapolation Methods
22 Extrapolation of tetrachloroethylene cancer risks observed in animal bioassays to humans
23 with continuous environmental exposure involved a number of methods, including dose-response
24 modeling in the range of observation, interspecies extrapolation, extrapolation to low exposures,
25 and route-to-route extrapolation. Section 5.4.4.1 and Figure 5-7 summarize the methods used to
26 extrapolate from the experimental data to humans.
27
28 5.4.4.1. Dose-Response Models and Extrapolation to Low Doses
29 As discussed in Section 4.10.3, the available body of MOA information is not sufficient
30 to derive biologically based quantitative models for low-dose extrapolation. No key events in
31 the tumor development process for tetrachloroethylene have been identified that would
32 determine the overall dynamics of such a model, nor are there experimental data specific to
33 tetrachloroethylene describing any of the underlying toxicodynamic processes, such as cell
34 replication rates.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-49 DRAFT-DO NOT CITE OR QUOTE
-------
1 The multistage model (and the multistage-Weibull) has been used by EPA in the vast
2 majority of quantitative cancer assessments because it has some parallelism to the multistage
3 carcinogenic process and it fits a broad array of dose-response patterns. Occasionally the
4 multistage model does not fit the available data, in which case an alternate model should be
5 considered. The related multistage-Weibull model has been the preferred model when individual
6 data are available for time-to-tumor modeling, which considers more of the observed response
7 than does the simpler dichotomous response model. Use of this decision scheme has contributed
8 to greater consistency among cancer risk assessments.
9 Consequently, the multistage model was the primary tool considered for fitting the dose-
10 response data and is given by:
11
12 P(d) =1 - exp(- q0-qixd-q2x£f-...-q6xd6)
13
14 where:
15 d = exposure level and
16 q; = parameters estimated in fitting the model
17
18 The multistage model in HMDS (Benchmark Dose Software, version 1.3.2; U.S. EPA, 2000) was
19 used for all multistage model fits.
20 Two tumor sites with statistically significantly decreased time to tumor were noted: brain
21 gliomas in NTP male rats and MCL in the NTP female rats, especially for the most severe stage
22 of leukemia observed (Stage 3). The multistage-Weibull model, given by the following
23 equation, was also used to evaluate the importance of decreased time to tumor and intercurrent
24 mortality in interpreting these responses.
25
26 P(d,t) = 1- exp[(- q0- qi x d - q2 x d2 -. . . - q6 x d5) x f]
27
28 where:
29 d = exposure level
30 t = time to observation of the tumor
31 q;, z = parameters estimated in fitting the model
32
33 Note that when the time to observation of the tumor is not a significant contributor to the dose-
34 response relationship, the parameter z is estimated to have a value of 1, and the model reduces to
35 the simpler multistage model described just before the multistage-Weibull. Tox_Risk (K.S.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-50 DRAFT-DO NOT CITE OR QUOTE
-------
1 Crump Group, Inc., ICF Kaiser International, Ruston, LA) was used for all multistage-Weibull
2 model fits.
3 Following dose-response modeling in the range of observation, the cancer risk values for
4 extrapolation to low doses were derived from the lower bound on the dose/concentration
5 associated with a level of risk from the low end of the observed range, usually 10% extra risk,
6 consistent with the 2005 cancer assessment guidelines. Extra risk has been used consistently
7 throughout EPA risk assessments and is given by:
8
9 Extra risk = [P(d) -P(0)] / [1 -P(0)]
10
11 where:
12 P(d) = estimated response at dose d and
13 P(0) = estimated response in the control group
1 A
14
15 The slope factor (risk per mg/kg-day for oral exposure) and risk per unit concentration (risk per
16 mg/L for drinking water exposure, or per |ig/m3 for inhalation exposure) are estimated by
17 dividing the risk level by its associated POD:
18
19 Risk/(unit of exposure) = Extra risk/(lower confidence bound on associated exposure)
20
21 5.4.4.2. Extrapolation to Human Equivalent Environmental Exposure
22 For extrapolation of risk to humans, this assessment used two approaches that were
23 dependent on the relevant dose metric: the EPA RfC methodology (U.S. EPA, 1994), which
24 applies when chemical-specific pharmacokinetic data are lacking, and EPA's cross-species
25 scaling methodology (U.S. EPA, 1992), which applies to exposures characterized in mg/kg-day,
26 whether parent chemical or metabolite.
27
28 5.4.4.2.1. Metabolized dose. Because of the availability of PBPK models to estimate a plausible
29 dose metric that addresses the differential metabolism of tetrachloroethylene between laboratory
30 rodents and humans, extrapolation to human equivalent environmental exposure entailed two
31 steps. First, consistent with the 2005 cancer guidelines (U.S. EPA, 2005a), EPA's methodology
32 for cross-species scaling (U.S. EPA, 1992) was considered when toxicological equivalence for
33 the relevant tumor sites was addressed. Then, the human equivalent exposure in terms of
34 metabolized dose was estimated via the human PBPK models. These considerations are further
35 described below.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-51 DRAFT-DO NOT CITE OR QUOTE
-------
1 EPA's cross-species scaling methodology was used for describing toxicological
2 equivalence because of the extensive rationale supporting it (U.S. EPA, 1992). Briefly, the
3 methodology maintains that, in the absence of adequate information to the contrary, toxicological
4 equivalence across species is determined through equal average lifetime concentrations or AUCs
5 of the carcinogen. The most typical application of this methodology is to oral exposures in
6 mg/kg-day, with no pharmacokinetic or pharmacodynamic information. In this circumstance, the
7 correspondence of equal AUCs is equivalent to considering the exposures in terms of
8 mg/kg3/4-day, and is achieved by multiplying animal exposures by (BWanimai/BWhuman)1/4. Note
9 that this equivalence across species entails the cross-species correspondence of internal doses in
10 terms of AUCs or mg/kg3/4-day, which is implicit in the frequent default case, i.e., oral
11 carcinogens without chemical-specific pharmacokinetic data. In other words, each time a
12 carcinogen is scaled from animals to humans on the basis of mg/kg3/4-day, an implicit
13 assumption is that internal doses are equipotent in terms of mg/kg3/4-day ("cross-species
14 scaling"), not mg/kg-day ("body-weight scaling"). Accordingly, when pharmacokinetic data are
15 available that relate administered concentration to the overall metabolized dose of the
16 carcinogen, this methodology is still applicable; internal doses, as a fraction of administered
17 dose, should still tend to produce equivalent effects when considered in terms of AUCs or
18 mg/kg3/4-day because metabolites are also subject to scale-affected clearance processes. In other
19 words, the scaling may be thought of as applied to the administered dose adjusted by the fraction
20 metabolized. There is a wide body of empirical evidence that overall metabolic rates associated
21 with enzymatic processes scale with body weight to the 3/4 power (U.S. EPA, 1992).
22 Furthermore, because in this assessment the scaling is applied to an internal dose (namely, the
23 overall metabolic rate), it is applicable regardless of the route of exposure.
24 EPA has experience applying cross-species scaling methodology in a number of
25 carcinogen assessments that have relied on pharmacokinetic modeling to characterize risks from
26 inhalation exposure. Further, the vast majority of EPA carcinogen assessments have relied on
27 this method—all oral slope factor estimates developed from animal bioassay data and all cancer
28 risk values developed from bioassay data and relying on PBPK models. Specific assessments
29 relying on PBPK models include the previous tetrachloroethylene assessment, dichloromethane,
30 vinyl chloride, and trichloroethylene. In all cases, a scientific rationale was provided for the
31 cross-species scaling approach taken.
32 The previous tetrachloroethylene assessment (U.S. EPA, 1986) also used cross-species
33 scaling of total rate of metabolism for the liver tumors and leukemias. The dichloromethane
34 assessment used cross-species scaling (BW2/3) of the daily amount of inhaled dichloromethane
35 metabolized by a GST pathway (U.S. EPA, 1987, 1995). The vinyl chloride inhalation risk per
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-52 DRAFT-DO NOT CITE OR QUOTE
-------
1 unit concentration involved a reactive metabolite whose AUC was judged to be proportional to
2 the metabolite's tissue concentration (U.S. EPA, 2000); that is, AUCs (and responses) for this
3 metabolite would tend to be equal for doses in terms of mg/kg-day rather than mg/kg3/4-day.9
4 Most recently, EPA's trichloroethylene assessment (external draft) used AUCs of metabolites
5 produced in the liver following inhalation exposure as being predictive of human risk using the
6 liver tumors observed in mice; for kidney tumors, the human equivalent risk was estimated using
7 BW3/4 scaling of daily production of thiol in the kidney (U.S. EPA, 2001).
8 Following the cross-species scaling methodology, metabolized tetrachloroethylene was
9 scaled using mg/kg3/4-day in order to estimate equivalent toxic effects in humans (U.S. EPA,
10 1992). This determination followed consideration of the reactivity of the dose metric and the
11 ability to estimate AUCs for the dose metric. The involvement of reactive metabolites through
12 which all other metabolites may follow has been hypothesized; however, body-weight scaling
13 was not considered pertinent for tetrachloroethylene because the possible reactive metabolites
14 have not been confirmed and because the majority of the metabolites formed is accounted for by
15 TCA, a stable metabolite. Concerning estimation of AUCs, the PBPK models for
16 tetrachloroethylene provide the rate of overall metabolism in units of mg-equivalents/kg-day,
17 which is a rate or flux. The models do not describe the kinetics of the overall metabolism and
18 therefore cannot provide an AUC. As discussed in Section 3, this is because the clearances of all
19 but one of tetrachloroethylene's metabolites are unknown, and many of the metabolites
20 themselves have not been identified. The metabolite whose clearance has been estimated is
21 TCA. While TCA is the predominant metabolite, it is not clear that TCA is responsible for all
22 the observed toxicity (see Appendix 4A). For animals, the study-specific body weights were
23 used (see Tables 5-5 and 5-7), and for humans the default of 70 kg was used.
24 It might appear that the use of such a procedure constitutes a "double counting" of
25 allometric scaling. This is not the case as is evident from the following explanation. The AUC
26 of the circulating stable metabolite (if available) leads to an equivalent average tissue
27 concentration of the metabolite X, Cx, for both species. This average concentration, when
28 applied over the lifetime of a species, leads to equivalent risk across species. For simplicity,
29 consider a one-compartment model. At steady-state, the production of X will be equal to the
30 clearance of X, so that
Available human incidence data were judged to be concordant with this interpretation of the animal data
for vinyl chloride; consequently, no cross-species scaling factor was considered necessary. The cross-species
scaling methodology (U.S. EPA, 1992) points out that, in general, body-weight scaling for reactive metabolites
entails assuming that the metabolite is removed from its target by spontaneous action, never leaves the tissue in
which it is formed, does not form lexicologically active macromolecular adducts, and that there are no species
differences in persistence. That is, body-weight scaling of a reactive metabolite would not be expected to result in
cross-species lexicological equivalence in all cases.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-53 DRAFT-DO NOT CITE OR QUOTE
-------
1 Rmet=VdxBWxCxxkcl,
2
3 where:
4 Rmet = rate of production of X
5 Vd = fractional volume of distribution
6 BW = body weight (converted to liters)
7 Cx = concentration of X and
8 kci = clearance of X in units of I/time
9
10 Then, for the concentration Cx equivalent in both species:
11
12 CX = [Rmet/BW X kcl xVd]H = [Rmet/BW X kcl X Vd\A
13
14 where H and A refer to human and animal. It is safe to assume that Vd is the same across
15 species. Then, [RmetIBW x kci\H = [RmetIBW x &C/]A. Now, kd (with units of I/time) is known to
16 scale according to BW1'4 (U.S. EPA, 2005a). Thus, the AUC approach leads to
17
18 Rmet (H/8 WH = Rmet (A)IB WA
19
20 This is the scaling approach used in this assessment due to lack of data to pursue an AUC
21 approach explicitly.
22 In the last step of the extrapolation to human equivalent PODs, the PODs in terms of
23 metabolized dose were extrapolated to environmental inhalation and oral exposures using
24 pharmacokinetic modeling. As discussed in Section 3.5, three human PBPK models were
25 considered, owing to insufficient data to distinguish between these models at low environmental
26 concentrations, especially concerning validation of total metabolite levels. These models
27 represent the work of Reitz et al. (1996), Rao and Brown (1993), and Bois et al. (1996), as
28 adapted by EPA (see Section 3.5 for more details). Because use of the human PBPK models
29 indicated that the correspondence between total metabolism and administered concentration was
30 linear below 0.1-1 ppm (see Figure 3-10), conversion factors (slopes) derived from Figure 3-10
31 were applied to estimate the human equivalent PODs in terms of administered concentration;
32 e.g., human equivalent POD (ppm) = human equivalent POD (mg-eq/kg-day) x conversion
33 factor ([mg-eq/kg-day]/ppm). See footnotes d-f in Table 5-9 and footnotes b-d in Table 5-11 for
34 the inhalation and oral conversion factors.
35
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-54 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.4.4.2.2. Administered concentration as dose metric. For those sites for which
2 pharmacokinetic-adjusted doses were not available or not otherwise relevant, EPA's default RfC
3 methodology was used (U.S. EPA, 1994). Tetrachloroethylene is considered a Category 3 gas
4 because it is water soluble and perfusion limited, and it has systemic (extra-respiratory) effects.
5 Because the ratio of blood/air partition coefficients for the experimental animal species relative
6 to humans is greater than or equal to 1 (for F344 rats, 18.9/10.3 = 1.8; for B6C3Fi mice,
7 17.5/10.3 = 1.7), a default value of 1 was used for this ratio (U.S. EPA, 1994). Consequently,
8 when administered inhalation concentrations were used as the dose metric, the concentrations
9 were considered equipotent across species for extrapolating risk to humans. Therefore, no
10 further extrapolation was necessary with the resulting PODs in the units of human equivalent
11 environmental exposure levels.
12
13 5.4.5. Cancer Risk Values
14 Human cancer risk was assessed using six different sex-species animal tumor data sets
15 and three different human PBPK models of total metabolism. The results of the dose-response
16 modeling using the data from the inhalation animal studies are discussed below, followed by
17 route-to-route extrapolation for estimating human cancer risk via oral exposure to
18 tetrachloroethylene. Finally, a discussion of quantitative and qualitative uncertainties underlying
19 the risk estimation process is provided.
20
21 5.4.5.1. Dose-Response Modeling Results
22 The dose-response modeling relying on total metabolism as the dosimeter is illustrated in
23 Figures 5-8a through 5-13a, and it is summarized in Table 5-9. The estimation of risk per unit
24 concentration associated with each tumor site is summarized in Tables 5-9 (identification of
25 PODs) and 5-10 (conversions of PODs to risk per unit concentration). The dose-response
26 modeling relying on administered concentration is illustrated in Figures 5-8b through 5-13b and
27 summarized in Table 5-11. Site-specific modeling results and conversions to human equivalent
28 risk values are discussed below. In all cases, linear extrapolation from the PODs was carried out
29 because of the lack of information supporting another extrapolation approach (U.S. EPA, 2005a).
30
31 5.4.5.1.1. Mouse tumors.
32 5.4.5.1.1.1. Hepatocellular tumors, male mice. The dose-response modeling results from the
33 hepatocellular adenomas or carcinomas in male mice of the JISA bioassay using total
34 metabolism (via PBPK modeling) led to human equivalent PODs (BMCLios) ranging from 1.8
35 ppm (Bois model) to 18 ppm (Rao and Brown model) tetrachloroethylene in air (see Table 5-9
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-55 DRAFT-DO NOT CITE OR QUOTE
-------
0.9
O.E!
0.7
T3
'If
ts
I
•I 0.5
fO
£ 0.4
0.3
0.2
Multistage Model with 0.95 Confidence Level
Multistage
BMDL
BMD
10 15 20
dose
25
30
35
17:0705/222002
2
3
4
5
6
7
8
9
10
Figure 5-8a. Incidence of hepatocellular adenomas and carcinomas in male
mice (JISA, 1993) corresponding to total tetrachloroethylene metabolism
(mg-eq/kg-day) and multistage model fit showing BMC and BMCL at 10%
extra risk. Data from Table 5-5.
17" 10 05^22072
Figure 5-8b. Incidence of hepatocellular adenomas and carcinomas in male
mice (JISA, 1993) corresponding to human equivalent continuous
tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
and BMCL at 10% extra risk. Data from Table 5-5
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-56 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
0 10
144:06272002
30
40
51
dose
Figure 5-9a. Incidence of hepatocellular adenomas and carcinomas in female
mice (JISA, 1993) corresponding to total tetrachloroethylene metabolism
(mg-eq/kg-day) and multistage model fit showing BMC and BMCL at 10%
extra risk. Data from Table 5-5.
Multistage Model with 0.95 Confidence Level
6
1
8
9
10
40
dose
14:5206/272002
Figure 5-9b. Incidence of hepatocellular adenomas and carcinomas in female
mice (JISA, 1993) corresponding to human equivalent continuous
tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
and BMCL at 5% extra risk. Data from Table 5-5.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-57 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
dose
17:2005/222002
Figure 5-10a. Incidence of malignant hemangiosarcomas in male mice (JISA,
1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk. Data
from Table 5-5.
Multistage Model with 0.95 Confidence Level
0.3
0.25
S 0.2
u
Qj
I
O
~o
¥
0.15
0.1
0.05
Multistage
BMDL
BMD
10
20
30
40
dose
13:4305/252005
Figure 5-10b. Incidence of malignant hemangiosarcomas in male mice
(JISA, 1993) corresponding to human equivalent continuous
tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
and BMCL at 5% extra risk. Data from Table 5-5.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-58 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
•c
Qj
g
o
O.G
0.5
0.4
0.3
0.2
0.1
Multistage
BMDL BMD
1
3
dose
17:2705/222002
Figure 5-lla. Incidence of mononuclear cell leukemia in male rats (JISA,
1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk. Data
from Table 5-7.
Multistage Model with 0.95 Confidence Level
0.6
0.5
0.4
0.3
0.2
0.1
Multistage
.BMDL BMP
20
40
60
SO
100
120
dose
17:3805/222002
Figure 5-1 Ib. Incidence of mononuclear cell leukemia in male rats (JISA,
1993) corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at 5%
extra risk. Data from Table 5-7.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-59 DRAFT-DO NOT CITE OR QUOTE
-------
0.5
0.45
, 0.4
m
I 0,35
0.3
0.25
0.2
0.15
0.1
Multistage
BMDL
dose
1
2
3
4
5
6
7
8
9
10
17:4205/222002
Figure 5-12a. Incidence of mononuclear cell leukemia in female rats (JISA,
1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
and multistage model fit showing BMC and BMCL at 10% extra risk. Data
from Table 5-7.
Multistage Model with 0.95 Confidence Level
-a
Oi
T>
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
Multistage
BMDL
BMD
0
20
40
GO
80
100
120
dose
17:4405/222002
Figure 5-12b. Incidence of mononuclear cell leukemia in female rats (JISA,
1993) corresponding to human equivalent continuous tetrachloroethylene
exposure (ppm) and multistage model fit showing BMC and BMCL at 5%
extra risk. Data from Table 5-7.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-60 DRAFT-DO NOT CITE OR QUOTE
-------
Multistage Model with 0.95 Confidence Level
I
2
3
4
5
0.15 -
III
t:
CD
o
iLJ
£
0.05 -
09:47 03/09 2005
Figure 5-13a. Incidence of kidney adenomas and adenocarcinomas in male
rats (NTP, 1986) corresponding to total tetrachloroethylene metabolism (mg-
eq/kg-day) and multistage model fit showing BMC and BMCL at 5% extra
risk. Data from Table 5-7.
m
Tj
IZ
o
tj
£5
0.15
0.1
0.05
10
20
30
40
50
60
70
dose
6 Figure 5-13b. Incidence of kidney adenomas and adenocarcinomas in male
7 rats (NTP, 1986) corresponding to human equivalent continuous
8 tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
9 and BMCL at 5% extra risk. Data from Table 5-7.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-61 DRAFT-DO NOT CITE OR QUOTE
-------
Table 5-9. Dose-response modeling summary for tumor sites using total tetrachloroethylene metabolites as the
dosimeter; tumor incidence data from JISA (1993) and NTP (1986)
Tumor type
Hepatocellular
adenomas or
carcinomas
(JISA, 1993)
Hemangiosarcomas,
spleen or liver
(JISA, 1993)
Mononuclear cell
leukemia, rats
(JISA, 1993)
Kidney tumors
(NTP, 1986)
Group
Male
mice
Female
mice
Male
mice
Male rats
Female
rats
Male rats
Modeling summary"
MLE coefficients
qo = 0.34
qj = 0.0049
q3 = 2.3 x ID'5
qo = 0.055
q2 = 4.6 x ID'4
q0 = 0.032
qj= 0.0051
qo = 0.21
Q! = 0.078
q2 = 0.0020
qo = 0.24
qj = 0.026
q0 = 0.020
Q!= 0.013
/7-value
0.15
0.75
0.48
0.68
0.48
1
POD in terms of metabolized dose (mg-
eq/kg-day)
Bioassay
estimate1"
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC05
BMCL05
12
3.6
15
9.8
21
12
1.3
0.81
4.1
2.0
4.1
1.9
Human
equivalent0
1.9
0.58
2.2
1.5
3.4
1.9
0.37
0.23
1.1
0.51
1.2
0.53
Human equivalent POD, in terms of
environmental exposure (ppm)
Rao and
Brown
(1993)d
59
18
68
44
100
59
11
7
32
16
35
16
Reitz et al.
(1996)e
14
4.2
16
11
24
14
2.6
1.6
7.5
3.7
8.2
3.8
Bois et al.
(1996)f
5.9
1.8
6.8
4.4
10
5.9
1.1
0.7
3.2
1.6
3.5
1.6
aModel: multistage model, extra risk. Coefficients estimated in terms of mg-equivalents/kg-day, as estimated for the experimental animals, and adjusted to
estimate equivalent continuous exposure.
bBioassay estimates illustrated in Figures 5-8a through 5-13a.
°Human equivalent points of departure were derived by dividing the bioassay estimate by [70 kg/ animal body weight (kg)]0 25. Animal body weights provided
in Tables 5-5 and 5-7.
dAt exposures below 1 ppm, approximately 0.033 (mg-eq/kg-day)/ppm inhaled tetrachloroethylene was estimated to be metabolized at steady state (see
Figure 3-10).
eAt exposures below 1 ppm, approximately 0.14 (mg-eq/kg-day)/ppm inhaled tetrachloroethylene was estimated to be metabolized at steady state (see
Figure 3-10).
fAt exposures below 0.1 ppm, approximately 0.33 (mg-eq/kg-day)/ppm inhaled tetrachloroethylene was estimated to be metabolized at steady state (see
Figure 3-10).
-------
1
2
3
4
5
Table 5-10. Human equivalent risk per unit concentration, in terms of
continuous environmental exposure, derived using total tetrachloroethylene
metabolites as the dosimeter; tumor incidence data from JISA (1993) and
NTP (1986)
Tumor type
Hepatocellular
adenomas or
carcinomas
(JISA, 1993)
Hemangiosarcomas,
spleen or liver
(JISA, 1993)
Mononuclear cell
leukemia, rats
(JISA, 1993)
Kidney tumors
(NTP, 1986)
Group
Male mice
Female mice
Male mice
Male rats
Female rats
Male rats
Concentrations above which these risks per
unit concentration should not be used due
to nonlinearity of metabolism and dose-
response
Human equivalent risk per unit concentration", continuous
environmental exposure, (ppm) *
Rao and Brown
(1993)
5.7 x ID'3
2.3 x lO'3
1.7 x 10'3
1.4 x ID'2
6.4 x ID'3
3.1 x 10'3
1 ppm
Reitz et al.
(1996)
2.4 x ID'2
9.6 x lO'3
7.2 x 10'3
6.1 x ID'2
2.7 x ID'2
1.3 x lO'3
0.1 ppm
Bois et al.
(1996)
5.7 x ID'2
2.3 x lO'2
1.7 x lO'2
1.4 x lO'1
6.4 x ID'2
3.1 x 10'2
1 ppm
aRisk per unit concentration calculated by dividing the risk level by the lower bound on its risk-specific
environmental concentration. See Table 5-9 for the risk levels and risk-specific concentrations.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-63 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 5-11. Dose-response summary and cancer risk estimates using
continuous equivalent administered tetrachloroethylene levels as dosimeter,
from NTP (1986) and JISA (1993)
Tumor type
Hepatocellular
adenomas or
carcinomas, male
mice
Hepatocellular
adenomas or
carcinomas, female
mice
Hemangiosarcomas,
spleen or liver,
male mice
Mononuclear cell
leukemia, male rats
Mononuclear cell
leukemia, female rats
Kidney tubular cell
adenoma or
adenocarcinoma,
male rats
Source
JISA
(1993)
JISA
(1993)
JISA
(1993)
JISA
(1993)
JISA
(1993)
NTP
(1986)
Modeling summary"
MLE dose
coefficients
q0 = 0.34
qj= 0.013
q3 = 7.8 x ID'6
qo = 0.056
q1 = 0.0076
q2 = 3.6 x lO'4
qo = 0.041
qj = 0.0041
q0 = 0.25
qj= 0.0051
qo = 0.26
qj= 0.0017
qo = 0.022
q1 = 9.6 x ID'4
/7-value
0.17
0.83
0.27
0.51
0.34
0.75
MLE PODb
(ppm)
BMC10 = 8.1
BMC05 = 5.4
BMC05=12
BMC05 = 10
BMC05 = 29
BMC05 = 53
Lower bound
on POD
(ppm)
BMCL10 = 2.8
BMCL05 = 2.1
BMCLoi = 6.9
BMCL05 = 6.4
BMCL05 = 13
BMCL05 = 24
Risk per unit
concentration
(ppm)la'c'd
3.6 x ID'2
2.4 x ID'2
7.2 x 10'3
7.8 x ID'3
3.8 x ID'3
2.1 x ID'3
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
aUsing dose coefficients in terms of administered ppm of tetrachloroethylene adjusted to equivalent continuous
exposure, consistent with RfC methodology (U.S. EPA, 1994), and the multistage model, extra risk:
P(d) = 1 - exp( - q0 - qi x d - q2 x d2 x . . . q6 x d6). See Tables 5-5 and 5-7 for input data.
kpOD results (MLEs and lower bounds) illustrated in Figures 5-8b through 5-13b.
Consistent with 2005 cancer guidelines; risk per unit concentration calculated by dividing the appropriate risk level
by its risk-specific total metabolite level.
dRisks per unit concentration, which are approximations for low-dose extrapolation, should not be used with
exposures greater than the POD from which they were derived without considering the curvature of the dose-
response function (at left).
Dose-response modeling of the male mouse liver tumor data using administered exposure fit the
data points as well as when using total metabolism, with the control and lowest exposure groups
again having the poorest fit. This dose-response modeling led to a human equivalent POD
(BMCLio) of 2.8 ppm tetrachloroethylene in air (see Table 5-11 and Figure 5-8b). The
corresponding central tendency estimate was approximately threefold higher, at 8.1 ppm. Linear
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-64 DRAFT-DO NOT CITE OR QUOTE
-------
_2
1 extrapolation from this POD led to a human equivalent risk per unit concentration of 3.6 x 10
2 per ppm, about twofold lower than the upper end of the range obtained using total metabolism.
3 Hepatocellular tumors, female mice. The dose-response modeling of the hepatocellular
4 adenomas or carcinomas in female mice from the JISA bioassay and the consideration of total
5 metabolism (via the PBPK models) led to human equivalent PODs (BMCLios) ranging from 4.4
6 ppm (Bois et al. [1996] model) to 44 ppm (Rao and Brown [1993] model) tetrachloroethylene in
7 air (see Table 5-9; Figure 5-9a). The corresponding central tendency estimates are
8 approximately 1.5-fold higher, at 6.8-68 ppm.
9 Linear extrapolation from the PODs above for hepatocellular tumors in female mice was
10 carried out because of the lack of information supporting another extrapolation approach. This
11 led to risks per unit concentration that were approximately 2.5-fold lower than those for the male
12 mice, at 2.3 x 10"3 per ppm (Rao and Brown [1993] model 0.1/44 = 0.0023) to 2.3 x 10"2 per ppm
13 (Bois et al. [1996] model; 0.1/0.4.4 = 0.023; see Table 5-9 and Figure 5-9a).
14 The dose-response modeling results from these same tumor data—but using administered
15 inhalation exposure as the dose metric (without PBPK modeling)—led to a human equivalent
16 POD (BMCL05) of 2.1 ppm (see Table 5-11 and Figure 5-9b). Note that, because the range of
17 experimental data extended below 10% extra risk, the risk per unit concentration was based on
18 5% extra risk. The corresponding central tendency estimate is approximately 2.5-fold higher, at
19 5.4 ppm. Linear extrapolation from this POD led to a risk per unit concentration of 2.4 x 10"2 per
20 ppm (0.05/2.1 = 0.024), which is virtually identical to the upper end of the risk per unit
21 concentration range obtained using total metabolism as the dose metric.
22 The dose-response relationship in terms of administered exposure (Figure 5-9b) appears
23 somewhat more linear than when expressed in terms of metabolized dose (Figure 5-9a), but the
24 PODs relying on the PBPK models have relatively narrower confidence intervals. However, the
25 confidence intervals associated with both dose metrics are fairly typical of adequate dose-
26 response fits, and neither dose metric is clearly better on a purely empirical basis.
27
28 5.4.5.1.1.2. Hemangiosarcomas. Hemangiosarcomas of the liver and spleen were also observed
29 in the JISA male mice. Because these tumors differ etiologically from the hepatocellular
30 adenomas and carcinomas, they were modeled separately. Dose-response modeling using total
31 metabolism led to human equivalent PODs (BMCLios) ranging from 5.9 ppm (Bois et al. [1996]
32 model) to 59 ppm (Rao and Brown [1993] model) tetrachloroethylene in air (see Table 5-9;
33 Figure 5-10a). The corresponding central tendency estimates are approximately 1.7-fold higher,
34 at 10-100 ppm.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-65 DRAFT-DO NOT CITE OR QUOTE
-------
1 Linear extrapolation from the PODs above for hemangiosarcomas in male mice led to
-3
2 human equivalent risk per unit concentration ranging from 1.7 x 10" per ppm (Rao and Brown
-v-2
3 [1993] model; 0.1/59 = 0.0017) to 1.7 x 10'z per ppm (Bois et al. [1996] model; 0.1/5.9 = 0.017)
4 tetrachloroethylene in air (see Table 5-10), approximately 3.5-fold less than the corresponding
5 risks per unit concentration for the other male mouse liver tumors. These results raise some
6 concern that total cancer risk based on the male mice data may be underestimated slightly by
7 considering only the hepatocellular adenomas and carcinomas. An analysis combining the risks
8 from these two sites indicates an overall risk from the male mice data of 3.4 x 10"2 per
9 mg-equiv/kg-day, about 20% higher than the risk per unit concentration estimated for
10 hepatocellular adenomas and carcinomas alone.10
11 Dose-response modeling using human equivalent continuous administered concentration
12 led to a human equivalent POD (BMCLos) of 6.9 ppm tetrachloroethylene in air (see Table 5-11
13 and Figure 5-10b). The corresponding central tendency estimate is approximately 1.7-fold
14 higher, at 12 ppm. However, although this fit was technically adequate (p > 0.1), the model did
15 not fit as well as the model using total metabolism in the region of the low and middle exposures;
16 the dose-response relationship is essentially a straight line between the high-dose group and the
17 control group.
10In order to gain some understanding of the total risk from multiple tumor sites in male mice, a sum of
risks across tumor sites was considered. This combined risk does not constitute double-counting if it can be
assumed that the hepatic adenomas and carcinomas were mechanistically independent from the hemangiosarcomas.
If there is some dependence between the tumor types, then the combined risk would tend to be an overestimate of
the total risk.
A statistically appropriate approach was used to sum the maximum likelihood estimates (MLE) of unit
potency across these tumor sites for male mice in the JISA study, assuming independence of the tumor sites.
Specifically, an estimate of the 95% upper bound on the summed unit risk, corresponding to the region of 10~4 extra
risk in the two dose-response curves, where the slopes were reasonably constant and stable. Assuming a normal
distribution for the individual risk estimates, the variance of the risk estimate for each tumor site can be derived
from its 95% upper confidence limit (UCL) according to the formula
95% UCL = MLE + 1.645 x standard error (MLE) (1)
after solving for the standard error
standard error (MLE) = (95% UCL - MLE)/1.645 (2)
where 1.645 is the z-statistic corresponding to a one-sided 95% confidence interval. Then the result is squared the
result to obtain the variance of each MLE.
The variances of the MLEs for the two tumor sites were summed to obtain the variance of the sum of the
MLEs. Then the standard error of the summed risk was obtained by taking the square root of the variance. The
95% UCL on the sum of the MLEs was then calculated using equation (1) above.
The resulting 95% UCL on the summed unit risk was 3.4 x 10~2 per mg-equiv/kg-day, about 20% higher
than the unit risk estimated at the POD at 10% for hepatocellular adenomas and carcinomas alone. That is, at 3.6
mg-equiv/kg-day (see Table 5-8), the extra risk for hepatocellular adenomas and carcinomas in male mice is 2.8
x 10"2 per mg-equiv/kg-day (0.1/3.6 mg-equiv/kg-day).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-66 DRAFT-DO NOT CITE OR QUOTE
-------
1 Linear extrapolation from the POD based on administered concentration led to a human
2 equivalent risk per unit concentration of 7.2 x 10"3 per ppm (0.01/1.4 = 0.0074)
3 tetrachloroethylene in air (see Table 5-11), approximately twofold less than the risk per unit
4 concentration obtained for male mouse hepatocellular tumors and administered concentration.
5
6 5.4.5.1.2. Rat leukemias.
1 5.4.5.1.2.1. Male rats. Table 5-9 provides the dose-response model coefficients for the curve fit
8 of the male rat MCL data shown in Figure 5-1 la. The dose-response modeling using total
9 metabolism (via the PBPK models) led to human equivalent PODs (BMCLios) ranging from 0.7
10 ppm (Bois et al. [1996] model) to 7 ppm (Rao and Brown [1993] model) tetrachloroethylene in
11 air. The corresponding central tendency estimates are approximately 1.5-fold higher, at 1.1-11
12 ppm.
13 Linear extrapolation from the PODs above for MCL in male rats led to human equivalent
14 risks per unit concentration ranging from 1.4 x 10"2 per ppm (Rao and Brown [1993] model;
15 0.1/7 = 0.014) to 1.4 x 10'1 per ppm (Bois et al. [1996] model; 0.1/0.7 = 0.14)
16 tetrachloroethylene in air (see Table 5-10).
17 The dose-response modeling results from these same tumor data but using administered
18 inhalation exposure as the dose metric (without PBPK modeling) led to a human equivalent POD
19 (BMCLos) of 6.4 ppm (see Table 5-11 and Figure 5-1 Ib). The corresponding central tendency
20 estimate is approximately 1.5-fold higher, at 10 ppm. Linear extrapolation from this POD led to
21 a risk per unit concentration of 7.8 x 10"3 per ppm (0.05/6.4 = 0.0073). Although the model fit in
22 terms of administered concentration was technically adequate (p = 0.51), the model
23 overestimated the control response and underestimated the mid-dose group response, leading to a
24 risk per unit concentration 2- to 20-fold lower than those obtained using total metabolism as the
25 dose metric.
26
27 5.4.5.1.2.2. Female rats. It was noted earlier (see Section 5.4.2.2) that the dose-response pattern
28 of MCL for female rats in the JISA study also was not monotonic. In this case, the "best" model
29 fit in terms of both dose metrics (see Figures 5-12a and 5-12b) provided adequate fits overall
30 (p = 0.48 and 0.27), but fit the control and low-exposure group responses least well. The model
31 fit in terms of metabolized dose provided a better fit of the control response, although both dose
32 metrics lead to approximately the same estimated response for the low dose, at -27%, compared
33 with the 34% observed. Although both models would appear to underestimate extra risk in this
34 region of the dose-response for female rat leukemias, it is not clear in this particular set of
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-67 DRAFT-DO NOT CITE OR QUOTE
-------
1 responses that the fit to the low dose should be emphasized over fitting as many of the responses
2 as possible.
3 The dose-response modeling of MCL in female rats using total metabolism (via the
4 PBPK models) led to human equivalent PODs (BMCLi0s) ranging from 1.6 ppm (Bois et al.
5 [1996] model) to 16 ppm (Rao and Brown [1993] model) tetrachloroethylene in air. The
6 corresponding central tendency estimates are twofold higher, at 3.2-32 ppm.
7 Linear extrapolation from the PODs above for MCL in female rats led to human
8 equivalent risks per unit concentration ranging from 6.4 x 10"3 per ppm (Rao and Brown [1993]
9 model; 0.1/7 = 0.014) to 6.4 x 10"2 per ppm (Bois et al. [1996] model; 0.1/0.7 = 0.14)
10 tetrachloroethylene in air (see Table 5-10). The dose-response modeling using administered
11 inhalation exposure as the dose metric led to a human equivalent POD (BMCLio) of 13 ppm (see
12 Table 5-11 and Figure 5-12b). The corresponding central tendency estimate is approximately
13 twofold higher, at 29 ppm. Linear extrapolation from this POD led to a risk per unit
14 concentration of 3.8 x 10"3 per ppm (0.1/13 = 0.0038), about 1.6- to 16-fold lower than those
15 obtained using total metabolism as the dose metric. Although the risks per unit concentration for
16 female rats were about twofold lower than those for the male rats, this relationship among the
17 dose metrics is very similar to that seen with male rat MCL.
18 There was an indication of accelerated occurrence of leukemias in female rats in the NTP
19 study, but the addition of time-to-tumor in the multistage model did not significantly affect the
20 estimate from that study. There was no similar observation of earlier leukemia incidence with
21 increasing exposure in the JISA study.
22
23 5.4.5.1.3. Rat kidney tumors. Table 5-9 provides the dose-response model coefficients for the
24 curve fit of the male rat kidney adenocarcinomas and carcinomas seen in the NTP study (see
25 Figure 5-13a). The dose-response modeling using total metabolism (via the PBPK models) led
26 to human equivalent PODs (BMCL0ss) ranging from 1.6 ppm (Bois et al. [1996] model) to 16
27 ppm (Rao and Brown [1993] model) tetrachloroethylene in air. Note that, because the 10% extra
28 risk response level fell above the range of experimental data, the POD was based on 5% extra
29 risk which the available data did span. The corresponding central tendency estimates are
30 approximately twofold higher than their lower bounds, at 3.5-35 ppm.
31 Linear extrapolation from the PODs above for kidney tumors in male rats led to human
32 equivalent risks per unit concentration ranging from 3.1 x 10"3 per ppm (Rao and Brown [1993]
33 model; 0.05/1.6 = 0.0031) to 3.1 x 10'2 per ppm (Bois et al. [1996] model; 0.05/16 = 0.031)
34 tetrachloroethylene in air (see Table 5-10). These risks per unit concentration were the lowest of
35 those estimated for all sites.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-68 DRAFT-DO NOT CITE OR QUOTE
-------
1 The dose-response modeling results from these same tumor data but using administered
2 inhalation exposure as the dose metric (without PBPK modeling) led to a human equivalent POD
3 (BMCLos) of 24 ppm (see Table 5-11, Figure 5-13b). The corresponding central tendency
4 estimate is approximately twofold higher, at 53 ppm. Linear extrapolation from this POD led to
5 a risk per unit concentration of 2.1 x 10"3 per ppm (0.05/24 = 0.0021).
6
7 5.4.5.1.4. Summary and discussion of site-specific dose response modeling. Dose-response
8 modeling of the candidate data sets presented no particular difficulties. As noted in the
9 preceding descriptions of modeling results, lower bounds on the central tendency estimates
10 (maximum likelihood estimate [MLEs]) of the PODs tended to be within twofold of the central
11 estimates. The only exception was for male mice, with a threefold difference between the MLEs
12 and their lower bounds.
13 The slopes of the dose-response curves at the PODs were estimated and found to be
14 within 1.6-fold of the corresponding risks per unit concentration in all cases, reflecting the
15 mostly low-dose linear dose-response relationships estimated within the lower region of the
16 observed data ranges. Because of the similarity of the slopes to the risks per unit concentration
17 and the apparent lack of potential for sublinear dose response behavior in the range of exposure
18 below the experimental data, these slopes are not shown.
19 Figure 5-14 shows the relative magnitudes of the risks per unit concentration associated
20 with each tumor site. It is interesting to note that the risks per unit concentration estimated using
21 administered concentration are not consistently the lowest or highest risk values among the
22 different estimates for each tumor site.
23 For example, the risk per unit concentration estimated from female mouse hepatocellular
24 tumors using administered concentration (2.4 x 10"2 per ppm) is approximately equal to the
25 upper end of the range estimated using metabolized tetrachloroethylene (2.3 x 10"2 per ppm).
26 Similarly for the male mice, the risk per unit concentration using administered concentration is
27 about twofold lower than the upper end of the range using metabolized dose (3.6 x 10"2 per ppm
28 vs. 5.7 x 10"2 per ppm, respectively). In contrast, the risks per unit concentration for MCL
29 estimated using administered concentration (7.8 x 10"3 per ppm, males; 3.8 x 10"3 per ppm,
30 females) are about twofold lower than the lower end of the range estimated using metabolized
31 tetrachloroethylene (1.4 x 10"2 per ppm, males; 6.4 x 10"3 per ppm, females). Some of this
32 variation is attributable to the differing shapes of the dose-response curves for the two different
33 dose metrics for each site and the variability in the bioassay responses. Overall, the interleaving
34 of the results from the two types of dosimetric, administered concentration and PBPK-estimated
35 metabolism, underscores some uncertainty in identifying the appropriate dosimetric(s).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-69 DRAFT-DO NOT CITE OR QUOTE
-------
1
2 5.4.5.1.5. Concordance of animal and human risk estimates. Although sufficient human data
3 linked with exposure characterizations are not available to derive cancer risk values, an analysis
4 by van Wijngaarden and Hertz-Picciotto (2004) provides a limited perspective on the human
5 cancer risk values estimated from animal bioassays. Van Wijngaarden and Hertz-Picciotto
6 (2004) demonstrated a simple methodology using epidemiologic data for four chemical
7 exposures including tetrachloroethylene. For tetrachloroethylene specifically, a linear dose-
8 response model was fit to laryngeal cancer observations in the upper airway cancer case-control
9 study of Vaughan et al. (1997). Van Wijngaarden and Hertz-Picciotto (2004) presented both an
10 EDoi and LEDoi (effective dose for a 1% additional lifetime risk over background and the lower
11 confidence interval on this dose, called the TD1 and LCL1 in their paper) for humans exposed
12 for 45 years, 240 days/year, a standard occupational exposure scenario. The ED0i was 228.40
13 mg/day and LEDoi was 60.16 mg/day. In order to compare these results with those derived from
14 the JISA (1993) study, we assumed a continuous lifetime exposure (70 years, 365 days/year, and
15 20 mVday breathing rate), resulting in an equivalent ED0i of 4.8 mg/m3 and LED0i of 1.3 mg/m3.
16 Using the continuous lifetime equivalent LEDoi as the POD and a low-dose linear approach, a
17 risk per unit concentration based upon Vaughan et al. (1997) is 8 x 10"6 per |ig/m3 (or, 0.01/1.3
18 x io3 |ig/m3). This estimate falls in the lower end of the range of cancer risk estimates from
19 male and female rat MCL tumors in JISA (1993). A cancer risk estimate from human data using
20 the EDoi as the POD is 2 x 10'6 per |ig/m3 (or, 0.01/4.8 x IO3 |ig/m3).
21 While the analysis of van Wijngaarden and Hertz-Picciotto (2004) can provide some
22 insight on the rodent-based tetrachloroethylene cancer risk estimate, it is still quite limited due to
23 the possible biases in Vaughan et al. (1997) and other factors. While individual bias in Vaughan
24 et al. (1997) may influence observed risk estimates from this study in either a positive
25 (overestimate) or null (underestimate) direction, the overall direction of all bias is likely toward
26 the null. First, Vaughan et al. (1997) do not have exposure information on individual cases and
27 controls and make an assumption that case and controls are exposed to tetrachloroethylene
28 concentrations as described by industrial hygiene surveys in dry cleaning establishments. For
29 this reason, bias related to exposure misclassification is likely great in this study. Second, as is
30 common to many population case-control studies, exposure prevalence to tetrachloroethylene is
31 low. Only 5 of 235 laryngeal cancer cases were identified as having exposure to
32 tetrachloroethylene, and 4 of these 5 cases as more likely than not as being exposed. Low
33 exposure prevalence may lead to reduced study power lead and imprecise estimates of the
34 relative risk (OR) that are not statistically significant. Last, epidemiologic evidence is available
35 to suggest an association between esophageal cancer and tetrachloroethylene, a site also
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-70 DRAFT-DO NOT CITE OR QUOTE
-------
1 examined by Vaughan et al. (1997). The OR between esophageal cancer and tetrachloroethylene
2 exposure was larger than that for laryngeal cancer, OR =11.9 (95% CI = 1.1-124). Cancer risk
3 estimates based on esophageal cancer observations would lead to a higher estimate than that
4 identified by van Wijngaarden and Hertz-Picciotto (2004). Also, Vaughan et al. (1997) only
5 considered respiratory cancers; given the epidemiological results discussed earlier these results
6 may represent an underestimate of total risk. An examination of site concordance with animal
7 observations, additionally, is not possible because the rat is a poor model for laryngeal cancer.
8 Odds ratios in Vaughan et al. (1997) are adjusted for a number of possible confounders such as
9 age, sex, education, study period, alcohol consumption, and cigarette smoking, and the use of
10 adjusted odds ratios is a strength of the van Wijngaarden and Hertz-Pi cci otto (2004) analysis.
11
12 5.4.5.2. Recommended Inhalation Unit Risk
13 Human inhalation cancer risk has been assessed using several different gender-species
14 animal tumor data sets and three different human PBPK models of total metabolism rate. These
15 results have been discussed above and are summarized in Figure 5-14.
16 In choosing which species-sex combination is most relevant for extrapolating to humans,
17 the MOA information does not provide a clear rationale. Although target organ concordance is
18 not a prerequisite for evaluating the implications of animal study results for humans (U. S. EPA,
19 2005a), it is notable that the leukemias (in both sexes of rats) support the observation of
20 lymphopoietic cancers in individuals employed as dry cleaners and degreasers, and the liver
21 tumors (in both sexes of mice) support the observation of liver tumors in dry cleaners (see
22 Section 4.10.1.1.2).
23 The male rat leukemia data provide the most sensitive response of the four
24 species-sex combinations in the JISA study for deriving a unit risk, defined as the
25 plausible upper-bound excess lifetime cancer risk estimated to result from continuous
26 exposure to tetrachloroethylene per unit of concentration. From Table 5-10, the
27 recommended unit risk value range is 1.4 x 10~2 to 1.4 x 10"1 per ppm, or 2 x 10~6 to
28 2 x 10~5 per ug/m3. This range reflects uncertainty in the choice of pharmacokinetic
29 model.
30 Comparison with previous EPA assessment: EPA (U.S. EPA, 1986, 1991) reported an
31 overall unit risk of 5.8 x 10"7 per |ig/m3 (3.9 x 10"3 per ppm), which was a geometric mean of six
32 risks per unit concentration from the 1986 NTP study: male and female rat leukemias, male and
33 female mouse liver carcinomas, and male and female mouse liver adenomas and carcinomas.
34 The highest risk per unit concentration in that range was 9.5 x 10"7 per |ig/m3, corresponding to
35 the leukemias in male rats from the NTP study (using total metabolism as the dosimeter).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-71 DRAFT-DO NOT CITE OR QUOTE
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
O-I---A----
O---A O-
..-.{V} [7] MCL, male rats
-Q---O---B
A----
A--O
•O H>
-Q
a
MCL, female rats
Hepatocellular tumors, male mice
Hepatocellular tumors, female mice
Hemangiosarcomas, male mice
I
Kidney tumors,(male rats
0.001 0.01 0.1 1
Human equivalent tetrachloroethylene unit risks (per ppm)
10
Figure 5-14. Comparison of inhalation risks per unit concentration for
tetrachloroethylene derived from rodent bioassays using four different dose
metrics—continuous equivalent inhalation concentration (0), Bois et al. (1996)
PBPK model (n), Reitz et al. (1996) PBPK model (o), and Rao and Brown (1993)
PBPK model (A). See Table 5-9 for PBPK model-derived estimates and Table 5-
10 for estimates relying on administered tetrachloroethylene.
This analysis supports a unit risk 14-fold higher than in EPA's 1991 assessment. This
difference is attributable to number of considerations. A comparison of the results from the two
bioassays, using the Reitz et al. (1996) model to characterize internal dose for both data sets but
not extrapolating to humans, indicates that the JISA study leads to risks per unit concentration
that are approximately twofold lower than those from the NTP study (not shown), if all else can
be considered equal. The remaining differences between the human equivalent inhalation risks
per unit concentration are attributable to differences in the particular PBPK models used, the
change in cross-species scaling factor from BW2/3 to BW3/4 (U.S. EPA, 1992), and use of the
most sensitive response rather than a (geometric) mean of the significant tumor responses.
Concerning the latter decision, use of a mean response treats the observations as if all are equal
likely alternatives (in the case of geometric means, the highest responses are disproportionately
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-72 DRAFT-DO NOT CITE OR QUOTE
-------
1 discounted relative to the lower responses). Use of the most sensitive response acknowledges
2 the weight contributed by all of the observed responses as independent indicators of human risk,
3 and provides a plausible upper bound on potential human risk.
4
5 5.4.5.3. Recommended Oral Slope Factor
6 The oral slope factor was developed from inhalation data because the only available oral
7 bioassay was less relevant for extrapolating to lifetime risk in humans, for several reasons. First,
8 the study was conducted by gavage at relatively high doses. Human exposures are more likely
9 not to occur in boluses, and high doses are associated at least with saturable metabolism
10 processes which may involve a different profile of toxicological processes than those prevalent at
11 more likely environmental exposure levels. Also, the animals were dosed for only
12 approximately 75% of the more usual 2-year period (NCI, 1977), making the oral study less
13 useful for estimating lifetime risk. Route-to-route extrapolation from the inhalation PODs
14 developed from the JISA study (see Table 5-9) was carried out using the human pharmacokinetic
15 models described in Section 3.5. Table 5-12 summarizes the resulting slope factors. Because the
16 oral slope factors are linear conversions of the inhalation risks per unit concentration, no figure
17 analogous to Figure 5-14 is provided; such a figure would be identical to Figure 5-14 with the
18 exception that the x-axis would reflect mg/kg-day units rather than ppm units.
19 The same arguments that led to selecting the range based on male rat leukemias
20 for the inhalation unit risk apply to the oral slope factor. In order to account for the
21 uncertainty contributed by the human PBPK models, the oral slope factor is given by the
22 range 1 x 10"2 to 1 x 10"1 per mg/kg-day. This range is equivalent to drinking water risks
23 per unit concentration of 4 x 10"7 to 4 x 10"6 per |ig/L of tetrachloroethylene in water
24 (assuming 70 kg body weight and a daily water consumption of 2 L/day). The
25 recommended slope factor range is 1 x 10~2 to 1 x 10"1 per mg/kg-day. This range
26 reflects uncertainty in the choice of pharmacokinetic model.
27 Comparison with previous EPA assessment: EPA (U.S. EPA, 1985) reported a slope
28 factor of 5.1 x 10"2 per mg/kg-day, based on the liver tumor incidence in female mice in the NCI
29 (1977) oral gavage study, total metabolized dose, and BW2/3 cross-species scaling. This value
30 falls near the center of the range developed in the current assessment.
31
32 5.4.5.4. Quantitative Adjustment for Sensitive Populations
33 Although a mutagenic MO A would indicate increased early-life susceptibility, there are
34 no data exploring whether there is differential sensitivity to tetrachloroethylene carcinogenicity
35 across life stages. This lack of understanding about potential differences in metabolism and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-73 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 5-12. Summary of tetrachloroethylene oral slope factors, estimated from
dose-response modeling of inhalation-exposed animals and by extrapolation to oral
exposure using pharmacokinetic models
Tumor type
Hepatocellular adenomas
or carcinomas, mice
Hemangiosarcomas, male
mice
Mononuclear cell
leukemia, rats
Kidney tumors
Group
Male
mice
Female
Mice
Male
mice
Male rats
Female
rats
Male rats
(NTP)
POD (and extra
risk), in terms of
human equivalent
metabolized dose
(mg-eq/kg-day) a
0.58
1.5
1.9
0.23
0.51
0.53
(10%)
(10%)
(10%)
(10%)
(10%)
(5%)
Concentrations above which these risks per unit concentration
should not be used due to nonlinearity of metabolism and dose-
response.
Oral slope factor (mg/kg-day) * b
Rao and
Brown
(1993)
model c
5.6 x 10'3
2.2 x lO'3
1.7 x 10'3
1.4 x 1Q-2
6.3 x 1Q-3
1.5 x lO'3
1 mg/kg-day
Reitz et al.
(1996)
model d
2.4 x 10'2
9.4 x lO'3
7.0 x 10'3
5.9 x 1Q-2
2.7 x 1Q-2
6.2 x lO'3
1 mg/kg-day
Bois et al.
(1996) model6
5.3 x 10'2
9.4 x lO'3
1.6 x 10'2
1.4 x 1Q-1
6.3 x 1Q-2
1.5 x lO'2
1 mg/kg-day
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
See Table 5-9 for derivation of human equivalent metabolite estimates.
Points of departure in the previous column were converted to human equivalent oral doses using the
pharmacokinetic models detailed below (intermediate calculation not shown), then converted to risks per unit
concentration by dividing extra risk by the corresponding risk-specific oral doses.
At exposures below about 1 mg/kg-day (divided between nine equally spaced doses during waking hours),
approximately 0.033 (mg-eq/kg-day)/(mg/kg-day) of ingested tetrachloroethylene was estimated to be
metabolized at steady-state, using the Rao and Brown (1993) model modified for oral exposure (see Section 3.5
and Figure 3-13) The conversion factor was calculated by dividing 0.01 mg-eq/kg-day (the total metabolite
production in 24 hrs) by the total oral dose of tetrachloroethylene estimated to produce that level of metabolites,
21mg/70kg.
At exposures below about 1 mg/kg-day (divided between nine equally spaced doses during waking hours),
approximately 0.14 (mg-eq/kg-day)/(mg/kg-day) of ingested tetrachloroethylene was estimated to be metabolized
at steady-state, assuming that the proportional relationship observed between the Rao and Brown (1993) model
and the Reitz et al. (1996) model for the inhalation route holds for oral exposure (see Section 3.5 and Figure 3-
13). The conversion factor was calculated by dividing 0.01 mg-eq/kg-day (the total metabolite production in 24
hrs) by the total oral dose of tetrachloroethylene estimated to produce that level of metabolites, 5.1 mg/70 kg.
At exposures below about 0.1 mg/kg-day (divided between nine equally spaced doses during waking hours),
approximately 0.31 (mg-eq/kg-day)/(mg/kg-day) of ingested tetrachloroethylene was estimated to be metabolized
at steady-state, assuming that the proportional relationship observed between the Rao and Brown (1993) model
and the Bois et al. (1996) model for the inhalation route holds for oral exposure (see Section 3.5 and Figure 3-13).
The conversion factor was calculated by dividing 0.01 mg-eq/kg-day (the total metabolite production in 24 hrs) by
the total oral dose of tetrachloroethylene estimated to produce that level of metabolites, 2.25 mg/70 kg.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-74 DRAFT-DO NOT CITE OR QUOTE
-------
1 susceptibility across exposed human populations thus represents a source of uncertainty.
2 Nevertheless, the existing data do support the possibility of a heterogenous response that may
3 function additively to ongoing or background exposures, diseases, and biological processes. As
4 noted in Section 4.9.5, there is some evidence that certain subpopulations may be more
5 susceptible to exposure to tetrachloroethylene. These subpopulations include early and later life
6 stages and groups defined by health and nutrition status, gender, race/ethnicity, genetics, and
7 multiple exposures and cumulative risk. As discussed below, these considerations strengthen the
8 scientific support for the choice of a linear non-threshold extrapolation approach. However,
9 because the MOA for tetrachloroethylene has not been established, it is not appropriate to derive
10 age-adjustment factors for early life exposures, as discussed in Supplemental Guidance for
11 Assessing Susceptibility from Early-Life Exposure to Carcinogens (U.S. EPA, 2005b).
12
13 5.4.6. Discussion of Uncertainties in Cancer Risk Values
14 A number of uncertainties underlie the cancer unit risk for tetrachloroethylene. These are
15 discussed in the following paragraphs. Specifically addressed is the impact on the assessment of
16 issues such as the use of models and extrapolation approaches, the reasonable alternatives and
17 the choices made and the data gaps identified. In addition, the use of assumptions, particularly
18 those underlying the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) is explained
19 and the decision concerning the preferred approach is given and justified. Several of the
20 uncertainties with the largest impact cannot be considered quantitatively. Thus an overall
21 integrated quantitative uncertainty analysis is not presented. Section 5.4.6.1 and Table 5-13
22 summarize principal uncertainties.
23
24 5.4.6.1. Sources of Uncertainty
25 5.4.6.1.1. Human population variability. The extent of inter-individual variability in
26 tetrachloroethylene metabolism has not been characterized. As noted above, several enzymes of
27 the oxidative and GSH metabolism, notably CYP2E1, CYP3A4, GSTZ, GSTA, GSTM, and
28 GSTT, show genetic polymorphisms with the potential for variation in production of specific
29 metabolites. Tetrachloroethylene metabolism has been shown to increase by inducers of
30 CYP450 enzymes such as toluene, phenobarbital, and pregnenolone-16 alpha-carbonitrile,
31 whereas CYP enzyme inhibitors such as SKF 525A, metyrapone, and carbon monoxide have
32 been shown to decrease tetrachloroethylene metabolism. Additionally, chronic exposure to
33 tetrachloroethylene has been shown to cause self-induction of metabolism. Human population
34 variability has also been discussed in Chapter 3.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-75 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-13. Summary of uncertainties in tetrachloroethylene cancer unit
risk estimate
Consideration/
Approach
Human
population
variability in
metabolism and
response/
sensitive
subpopulations
Low-dose
extrapolation
procedure
Dose metric
Species /gender
combination
(see Table 5-14)
PBPK model
Cross-species
scaling
Impact on unit risk
Low-dose risk f to
an unknown extent
Departure from
EPA's Guidelines for
Carcinogen Risk
Assessment POD
paradigm, if justified,
could | or t risk per
unit concentration an
unknown extent
Alternatives could t
or J, risk per unit
concentration by an
unknown extent
Human risk could J,
or t, depending on
relative sensitivity
10-fold range in risk
per unit
concentration among
the three available
models
Alternatives could J,
or t risk per unit
concentration (e.g.,
3.5-fold| [scaling by
BW) or t 2-fold
[scaling by BW2/3])
Decision
Considered
qualitatively
Multistage
model to
determine POD,
linear low-dose
extrapolation
from POD
(default
approach)
Considered
total
metabolism and
administered
concentration
Male rat MCL
All are
considered
BW3/4 (default
approach)
Justification
No data to support range of human
variability/sensitivity, including whether children are
more sensitive. Mutagenic MOA (if established)
would indicate increased early-life susceptibility.
Available MOA data do not inform selection of
dose-response model but do not support non-
linearity (mutagenicity is plausible contributor and
cannot be ruled out); male rat MCL data are linear in
observed range; linear approach in absence of clear
support for an alternative is generally supported by
scientific deliberations supporting EPA's Guidelines
for Carcinogen Risk Assessment.
Experimental evidence supports a role for
metabolism in toxicity, but actual responsible
metabolites are not clearly identified.
MCL is the largest response and is reproducible
across studies, despite high background response
rate. There are no MOA data to guide extrapolation
approach for any choice. It was assumed that
humans are as sensitive as the most sensitive rodent
gender/species tested; true correspondence is
unknown. The carcinogenic response occurs across
species. Generally, direct site concordance is not
assumed; consistent with this view, some human
tumor types are not found in rodents (i.e., cervical,
esophageal cancer deaths) and rat and mouse tumor
types also differ.
There is no scientific basis for choosing among
pharmacokinetic results for estimating total
metabolism of tetrachloroethylene given limitations
in available data. The highest value provides a
reasonable upper estimate of potential human cancer
risk.
There are no data to support alternatives. Because
the dose metric was not an AUC, BW3/4 scaling was
used to calculate equivalent cumulative exposures
for estimating equivalent human risks.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-76 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-13. Summary of uncertainties in tetrachloroethylene cancer unit
risk estimate (continued)
Consideration/
Approach
Bioassay
Statistical
uncertainty at
POD
Impact on unit risk
t risk per unit
concentration 2-fold
if NTP study used
J, risk per unit
concentration 1.6-
foldif ECio used
rather than LECio
Decision
JISA study
LEC (default
approach for
calculating
plausible upper
bound)
Justification
JISA study used the lowest experimental exposures
(reduces extrapolation uncertainty)
Limited size of bioassay results in sampling
variability; lower bound is 95% confidence interval
on concentration.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
LECIO = 95% lower confidence limits on the air concentrations associated with a 10% extra risk of cancer incidence
A separate issue is that the human variability in response to tetrachloroethylene is also
poorly understood. The effect of metabolic variation, including potential implications for
differential toxicity, has not been well studied.
Although a mutagenic MOA would indicate increased early-life susceptibility, there are
no data exploring whether there is differential sensitivity to tetrachloroethylene carcinogenicity
across life stages. This lack of understanding about potential differences in metabolism and
susceptibility across exposed human populations thus represents a source of uncertainty.
Nevertheless, the existing data do support the possibility of a heterogenous response that may
function additively to ongoing or background exposures, diseases, and biological processes. As
noted in Chapter 4 (see Section 4.9.5), there is some evidence that certain subpopulations may be
more susceptible to exposure to tetrachloroethylene. These subpopulations include early and
later life stages and groups defined by health and nutrition status, gender, race/ethnicity,
genetics, and multiple exposures and cumulative risk. As discussed in the section on low-dose
extrapolation below, these considerations strengthen the scientific support for the choice of a
linear non-threshold extrapolation approach.
5.4.6.1.2. Choice of low-dose extrapolation approach. The MOA is a key consideration in
clarifying how risks should be estimated for low-dose exposure. MOA data are lacking or
limited for all candidate cancer endpoints for tetrachloroethylene (i.e., rat MCL and kidney
tumors, mouse hepatocellular tumors and hemangiosarcomas). When the MOA cannot be
clearly defined, EPA uses a linear approach to estimate low-exposure risk, based on the
following broad and long-term scientific assumptions, which supported the development of the
EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-77 DRAFT-DO NOT CITE OR QUOTE
-------
1 • A chemical's carcinogenic effects may act additively to ongoing biological processes,
2 given that diverse human populations are already exposed to other agents and have
3 substantial background incidence of various tumors.
4
5 • A broadening of the dose-response curve in the human population (less rapid fall-off with
6 dose) and, accordingly, a greater potential for risks from low-dose exposures (see Zeise et
7 al., 1987; Lutz et al., 2005) would result for two reasons. First, even if there is a
8 threshold concentration at the cellular level, that threshold is likely to be different among
9 different individuals. Secondly, greater variability in response to exposures in the
10 heterogeneous human population would be anticipated than in controlled laboratory
11 species and conditions (due to, e.g., genetic variability, disease states, age).
12
13 • The use of linear extrapolation provides consistency across assessments as well as
14 plausible upper-bound risk estimates that are believed to be health-protective (U.S. EPA,
15 2005a).
16
17 The extent to which the overall uncertainty in low-dose risk estimation could be reduced
18 if the MOA for tetrachloroethylene were known with a high degree of confidence is of interest,
19 but clear data on the MOA of tetrachloroethylene is not available, and even if it were,
20 incorporation of MOA into dose-response modeling might not be straightforward and might not
21 significantly reduce the uncertainty about low-dose extrapolation. This is because the MOA as
22 well as other factors, especially human response variability, are determinants of the dose-
23 response function in humans.
24 This chemical assessment also evaluates the extent to which a collection of mathematical
25 functions, fit to one of the tetrachloroethylene bioassay data sets and extrapolated down to low
26 doses, could inform uncertainty. There is not sufficient information regarding the MOA to
27 support a chemical-specific inference about dose-response behavior at low dose for
28 tetrachloroethylene. Thus, it is of interest to observe how different functions fit to the tumor data
29 may diverge when extrapolated downward. Much previous experience has supported a general
30 mathematical property that different curves, though fitting observed experimental data well,
31 often diverge widely when extrapolated to doses well outside the observed range. Indeed, the
32 inability of curve-fitting procedures to provide useful compound-specific information about low
33 dose risks has been a principal motivation for the "model free" approach of straight line
34 extrapolation from a POD within the observed range of the data (Krewski and van Ryzin, 1981;
35 NRC, 1983).
36 Calculations here considered four alternative functional forms frequently used for
37 noncancer dose-response assessment in the observable range of the experimental data
38 (multistage, Weibull, log-logistic, and log-probit). These can accommodate a wide variety of
39 dose-response shapes, including threshold-like behavior. These models were fit to the MCL data
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-78 DRAFT-DO NOT CITE OR QUOTE
-------
1 in male rats using the EPA BMDS program without restricting the shape parameters of the latter
2 three models (see Appendix 5B for more details). Parameters describing the risk distribution
3 (mean, median and 95% upper and lower confidence bounds) were estimated by a bootstrap
4 procedure because these parameters are not all readily available in the current BMDS software.
5 In comparing these distributions from the bootstrap procedure, the mean (as a measure of central
6 tendency) and the 95% upper and lower bounds calculated from the bootstrap procedure are of
7 interest. The resulting risk distributions were compared at two exposure levels—at a generalized
8 POD and at an environmental level approximately 105-fold lower than the POD.
9 At POD C, corresponding to a risk of approximately 0.1 using the mean estimate from
10 the multistage model, the bootstrap procedure yielded similar risk distributions for the four
11 models (see Figure 5-15a). The means and corresponding confidence bounds agree within an
12 order of magnitude and the spreads in the distributions (the distance between the upper and lower
13 confidence bounds) are within two orders of magnitude. Note that the probability calculations
14 are in terms of metabolized dose in the male rat and do not directly characterize human risks.
15 EPA also examined the bootstrap results from those same models at a dose that is lower
16 than the POD C by a factor of 105 (although EPA's actual low dose risk estimates are developed
17 using a linear extrapolation from a POD to the origin rather than using low-dose estimates from a
18 model). Figure 5-15 illustrates these results. In the region of extrapolated concentrations
19 (C x 10"5), the mean risks of the latter three models (Weibull, log-logistic and log-probit) are
20 about one to three orders of magnitude higher than the mean of the multistage model risks. The
21 spreads of all the models are quite broad, with a six order of magnitude 95% confidence interval
22 for the multistage and much greater spreads for the other three models. The upper bounds of risk
23 for the other three models are higher than that for multistage model, within about three orders of
24 magnitude, and their lower bounds of risk are much lower than that of multistage, by nine or
25 more orders of magnitude. With such large spreads in confidence intervals, the extrapolated
26 models in effect provide little information about low-dose risks. The extrapolation of the
27 multistage model does result in estimates reasonably close to the low-dose estimates from the
28 model-independent straight line extrapolation from the POD, in that the mean and upper-bound
29 risks at the lower concentration are both within 10% of the estimates resulting from applying
30 linear extrapolation to the results at the higher concentration.
31 This comparison of risk distributions has several limitations. First, the selected models
32 do not represent all possible models one might fit, serving primarily to illustrate a range of
33 possibilities. That is, other models could be selected to yield more extreme results, both higher
34 and lower than those shown here. Further, the results apply only to the prediction of MCL in
35 male rats. For reasons discussed above concerning expected additivity to background processes
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-79 DRAFT-DO NOT CITE OR QUOTE
-------
•c
a. Predicted risks of MCL in male rats, corresponding
to 1.5 mg-eq/kg-day total metabolism
Multistage
Weibull Log-logistic
Dose-response model
Log-probit
•
§, 1.E-03 -
§ 1.E-04
1
O1 1.E-05 -
Q.
£ 1.E-06 -
b. Predicted risks of MCL for male rats, corresponding
to 1.5E-5 mg-eq/kg-day total metabolism
LB=1E-12
LB=0
LB=1E-21
Me iian=8E-26
LB: 1E-152
Multistage
Weibull Log-logistic
Dose-response model
Log-probit
1
2 Figure 5-15. Illustration of sensitivity to model selection for low-dose extrapolation. The risk
3 distributions associated with four dose-response models which adequately fit the
4 tetrachloroethylene dose-response data for MCL in male rats (JISA, 1993) were compared. The
5 mean (0) and median (n) risks for each model are indicated with symbols, and the 5th and 95th
6 percentiles are indicated by bars. Risks are in terms of metabolized tetrachloroethylene in male
7 rats. Figure a shows the comparison at a generalized POD, selected as the mean exposure estimate
8 from the multistage model corresponding to a risk of approximately 0.1—that is, 1.5 mg-eq/kg-
9 day, equivalent to about 50 ppm as administered in the bioassay. Figure b compares risk
10 distributions at an exposure corresponding to an environmental concentration of
11 tetrachloroethylene-approximately 105-fold lower than the POD, or 1.5 x 10~5 mg-eq/kg-day,
12 which is equivalent to about 50 x 10"5 ppm if administered as in the bioassay. Note that three
13 lower bounds (Weibull, log-logistic, and log-probit) and one median (log-probit) could not be
14 plotted on the graph. See Appendix 5B for more details.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-80 DRAFT-DO NOT CITE OR QUOTE
-------
1 in humans and greater heterogeneity of human populations, more linear fits are expected to better
2 capture the anticipated response in a human population. The low-dose extrapolation to humans
3 from threshold-like models (i.e., log-logistic and log-probit) carries a relatively greater degree of
4 uncertainty than extrapolation from the multistage and Weibull fits. These calculations illustrate
5 the expected finding that alternative functional forms fit to the tetrachloroethylene tumor data
6 yield a wide range of numerical values for probability of response when extrapolated down to
7 low dose and are uninformative of the actual risk.
8 Given the current state of scientific knowledge about tetrachloroethylene carcinogenicity,
9 the straight line based risk estimates presented above form the preferred recommendation for
10 estimating a plausible upper-bound estimate of potential human risks from tetrachloroethylene.
11 This approach is supported by both general scientific considerations, including those supporting
12 the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), as well as chemical-specific
13 findings. The former include the scientific principles articulated above (the expectation that a
14 chemical functions additively to background exposures, diseases, and processes, that variability
15 within the human population would broaden the dose-response curve and eliminate individual
16 thresholds if present, and that the approach provides consistency across assessments facilitating
17 direct comparison of the derived risk values). The latter includes evidence that, within the dose
18 range of the cancer bioassays, the observable tumor response data are consistent with a linear
19 model and do not suggest occurrence of a threshold, and that variability in the human response
20 across the population is expected (see Human population variability, above).
21
22 5.4.6.1.3. Dose metric. Tetrachloroethylene is metabolized to several intermediates with
23 carcinogenic potential. Although much data exist for TCA, several analyses indicate that TCA
24 alone is not able to explain the toxicity associated with tetrachloroethylene exposure; therefore,
25 at least one other toxic agent appears to be involved. Whether total metabolism, either as a
26 measure of a precursor or intermediate or as a surrogate directly proportional to the toxic
27 agent(s), is an adequate indicator of potential risk is unclear. Use of administered dose (without
28 use of a PBPK model) yields risk estimates intermediate between those based on the higher and
29 lower PBPK models. Consequently, a role for the parent compound has not been ruled out, nor
30 is it clear that the toxic agent(s) are not proportional to administered concentration.
31
32 5.4.6.1.4. Choice of species/gender. The factors influencing the choice of rodent tumor data set
33 for human risk characterization are summarized in Table 5-14. The carcinogenic response
34 occurs in rodents as well as in humans. There is no information on tetrachloroethylene to
35 indicate that the observed rodent tumors are not relevant to humans, and there are no non-rodent
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-81 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5-14. Summary of considerations for each rodent tumor type
Magnitude of response
Specificity of response
Replication of findings
in multiple bioassays
Other considerations
contributing to
uncertainty in rodent
data
Relevance to humans:
a) qualitative
(biologic) site
concordance
b) occurrence in
human studies
c) confidence in
MOA
Overall considerations
for choice of tumor
type
Mononuclear cell
leukemia
+++
Rats, both genders
4/4 study datasets
High background
rate
Yes
Yes, but exact
match of tumor
classification is not
found/may not be
possible
No data
Rodent response of
highest magnitude,
reproducible; no
MOA data
Liver
++
Mice, both
genders
6/6 study datasets
High background
rate in male mice
Yes
Yes, but
association is
weak
PPAR-a
activation may
contribute, but is
not sole MOA
Rodent response
of considerable
magnitude,
reproducible,
some MOA data
Kidney
++
Male rats only
1/2 study datasets
Rare tumor
(unlikely to be due
to chance, but low
incidence)
Yes
Yes, but
association is weak
Multiple MO As
may play a role
Rare tumor in
rodents and
humans; MOA
data are strongest
Hemangiosarcoma
+
Male mice only
1/3 study datasets
Rare tumor (unlikely
to be due to chance,
but low incidence)
Yes
No, but tumor type is
rare
No data
Rare tumor in rodents
and humans; no
MOA data
4
5
6
7
8
9
10
11
12
13
14
cancer bioassay data. Further, no tetrachloroethylene data exist to guide quantitative adjustment
for differences in sensitivity among rodents and humans. Human-rodent site concordance
generally is not assumed, e.g., due to potential differences in pharmacokinetics, DNA repair,
other protective systems across species and tissues (U.S. EPA, 2005a). In keeping with this
view, certain tumors associated with tetrachloroethylene exposure in human mortality studies
(e.g., cervix and esophagus) were not observed in rodents; cancer of the lymphoid system was
associated with tetrachloroethylene exposure in humans, with some evidence for an association
with bladder, kidney, and lung cancer. In addition, rat and mouse tumor types also differ from
each other. Finally, conclusive MOA data are lacking for the observed rodent and human
tumors.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-82 DRAFT-DO NOT CITE OR QUOTE
-------
1 MCL is the cancer response of highest magnitude, and it is reproducible in two bioassays
2 and both genders. Although MCL has a high and variable incidence in unexposed F344 rats, a
3 biologically and statistically significant increase over background was observed. The qualitative
4 similarities among MCL to certain lymphoid cancers, and the implications regarding human
5 relevance, are addressed in Section 4.8.2.4.1.2; also addressed is that elevated lymphoma
6 mortality has been associated with tetrachloroethylene exposure in humans. The MOA for MCL
7 remains unexplored.
8 Occasionally, if the multistage model does not adequately fit a data set, an alternate
9 model can be used to determine the POD. In the case of female rat MCL data, the best-fitting
10 model (Weibull) allowed for a plateau and yielded estimates of risk per unit concentration 10-
11 fold higher than those from the multistage model fit of the male rat MCL data. While the female
12 rat MCL data suggest a plateau (also apparent for female rat MCL data from the NTP bioassay),
13 the multistage model fit was technically adequate (p = 0.48).
14 The mouse liver tumor is a robust finding in several studies, including in both sexes. As
15 is the case with MCL, the background for this tumor type is high especially in males. A
16 biologically and statistically significant increase over background was observed in males and
17 females. There is evidence that activation of the PPAR-a receptor by the tetrachloroethylene
18 metabolite TCA contributes in part to the induction of mouse liver tumors. However, it is not the
19 only operative MOA involved in hepatocellular tumorigenesis. Thus, the MOA remains
20 unresolved.
21 Two tumor types were observed in only one bioassay. Kidney tumors rarely occur in
22 unexposed rodents and were significantly elevated with tetrachloroethylene exposure in the male
23 rat NTP bioassay. The MOA is better understood for kidney tumors than for the other sites.
24 Hemangiosarcoma is another rare tumor associated with tetrachloroethylene exposure in the
25 male mouse JISA study. There are no MOA data for hemangiosarcomas.
26
27 5.4.6.1.5. Physiologically basedpharmacokinetic (PBPK) model. Toxicokinetic models are
28 used in this assessment for deriving dose metrics to support dose-response analyses. The
29 evidence suggests that by-products of tetrachloroethylene metabolism are responsible for liver
30 and kidney toxicity and for carcinogenicity. Inhaled concentration of the parent compound is,
31 therefore, not an appropriate dosimeter for these effects, and pharmacokinetic modeling of daily
32 overall metabolized dose is expected to be an improvement in spite of the many attendant
33 uncertainties in the modeling. Of the available toxicokinetic models on tetrachloroethylene, the
34 assessment considers three recently developed models that describe parent tetrachloroethylene
35 and overall metabolism of the parent compound in humans. These models do not describe the
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-83 DRAFT-DO NOT CITE OR QUOTE
-------
1 kinetics and transformation of total metabolic products or any individual metabolite. All three
2 models provide reasonably good predictions of exhaled breath and blood tetrachloroethylene
3 concentrations, so there is no particular basis for preferring one model over another. A 10-fold
4 difference is shown in model predictions of the rate of metabolism in humans, a reflection of
5 model differences in the values for the metabolic parameters. Because the accuracy of the
6 models has been evaluated only against blood and breath concentrations of the parent
7 compound—quantities that are insensitive to these parameters—the reliability of these models
8 for predicting the rate of total metabolism in humans is unknown. Data on total metabolite levels
9 are not available in humans, and the use of available urinary and blood TCA data is problematic.
10 The overall difference in risk estimates using these three models is approximately 10-fold.
11
12 5.4.6.1.6. Cross-species scaling. An adjustment for cross-species scaling (BW3/4) was applied
13 to address toxicological equivalence of internal doses between each rodent species and humans,
14 consistent with the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a); the
15 approach is detailed in Section 5.4.4.2.1. It is assumed that, without data to the contrary, equal
16 risks result from equivalent constant exposures. While the true correspondence of equipotent
17 tetrachloroethylene exposures across species is unknown, the use of BW3/4 scaling is expected
18 neither to over- or underestimate human risk (U.S. EPA, 1992).
19
20 5.4.6.1.7. Choice ofbioassay. The JISA inhalation bioassay provides data on the lowest
21 experimental exposures, and its use, therefore, reduces extrapolation uncertainty slightly. For
22 mice, the lowest-exposure concentration of 10 ppm was 10-fold lower than the lowest-exposure
23 concentration in the NTP inhalation study (NTP, 1986). For rats, the low-exposure concentration
24 of 50 ppm was fourfold lower than in the NTP study. Although the JISA and NTP inhalation
25 bioassays used similar rodent strains, it is possible that differences in the animals used (in
26 addition to other unidentified factors) may have contributed to the twofold higher incidence of
27 hepatocellular tumors and MCL in the NTP study. The estimated risks for these sites are
28 consequently twofold lower than in previous EPA assessments which relied on the NTP bioassay
29 (U.S. EPA, 1991).
30
31 5.4.6.1.8. Statistical uncertainty at the Point of Departure (POD). Parameter uncertainty
32 within the chosen model reflects the limited sample size of the cancer bioassay. For the
33 multistage model applied to this data set, there is a reasonably small degree of uncertainty at the
34 10% extra risk level (the POD for linear low-dose extrapolation).
35
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-84 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.4.6.2. Summary and Conclusions
2 The uncertainties presented in Table 5-13 have a varied impact on risk estimates. Some
3 suggest risks could be higher than was estimated, while others would decrease risk estimates or
4 have an impact of an uncertain direction. Several uncertainties are quantitatively
5 characterized—the range of uncertainty in the PBPK models considered, together with the
6 statistical uncertainty in the multistage modeling estimate, for the significantly increased rodent
7 tumors. Sensitivity to model selection is quantitatively explored in Figure 5-15, with a focus on
8 thresholded, non-linear alternatives, illustrating the expected finding that such alternatives yield
9 a wide range of estimates that are uninformative of the actual risk. Alternatives that would yield
10 higher risk estimates (e.g., supralinear models), which are equally scientifically valid, are not
11 presented. In addition, the results apply only to the prediction of MCL in male rats, not in
12 humans. Due to limitations in the data, particularly regarding the MO A and relative human
13 sensitivity and variability, the quantitative impact of other uncertainties of potentially equal or
14 greater impact has not been explored. As a result, an integrated quantitative analysis that
15 considers all of these factors independently was not undertaken.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-85 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
APPENDIX 5A:
BENCHMARK DOSE MODEL RESULTS
ABBREVIATIONS
AIC = Akaike Information Criterion (see, e.g., U.S. EPA, 2000)
BMCXX = Effective concentration at a specified level of extra risk; e.g., BMCio is the
concentration corresponding to 10% extra risk, BMC0s corresponds to 5% extra risk
BMCs = Effective concentration corresponding to a one standard deviation difference in the
mean response from the control mean response (for continuous data). This is
approximately equivalent to 10% of the responses at the effective concentration
being more extreme than 98% of the controls if the adverse response is an increase
relative to the controls, or 10% of the responses being more extreme than 2% of the
controls if the adverse response is a decrease relative to controls.
BMCLxx = Lower 95% confidence bound on the estimated BMCXX
BMCLs = Lower 95% confidence bound on the estimated BMCs
NA = Not applicable
POD = Point of departure
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-86 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 5A-1. Benchmark modeling summary: hepatic parenchymal changes
in humans with occupational exposure to tetrachloroethylene, data from
Brodkin et al. (1995)
8-hr TWA exposure
(ppm)
0.0007
4.6
19.8
Quantal models
Multistage13
Gamma, Weibull,
Logistic
Probit
Probit, log-transformed dose
Equivalent continuous
exposure"
(ppm)
0.0003
1.6
7.1
Goodness of fit
/7-value
0.71
0.71
0.63
0.62
0.53
Number examined
26
9
18
AIC
72.4
72.4
72.5
72.5
72.7
Maximum
X2 residual
near POD
0.2
0.3
0.4
0.4
0.6
Number (%) with
parenchymal changes in
hepatic ultrasonograph
10 (38)
5(55)
13 (72)
BMC10
(ppm)
0.9
0.9
1.2
3.5
1.7
BMCL10
(ppm)
0.5
0.5
0.7
2.2
0.8
5
6
7
8
9
10
11
12
13
14
15
16
a Exposures adjusted by 10/20 (m3/day) x 5/7 (days) to estimate equivalent continuous exposure levels.
Measurements were taken from personal samplers for a subset of the individuals in the two higher-exposure
groups. The background level of 0.0007 ppm is the high end of a range from Hartwell et al. (1985).
b Multistage model selected as best fitting-model. Models had similar fits, multistage had lowest AIC and closest fit
near the ECi0. Multistage model given by:
P(d) = 1 - exp(- q0 -
where: d = continuous exposure level (ppm)
q0 = 0.40
i = o.n
Multistage Model with 0.95 Confidence Level
O.B
0.7
0,
°£
0.4
0.3
Multistage
BMDL
BMP
0 1
13:4402/262002
3 4
dose
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-87 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 5A-2. Benchmark modeling summary: increased liver weight in
female mice exposed to tetrachloroethylene, data from Kjellstrand et al.
(1984)
Administered exposure,
24 hrs/day, 30 days
(ppm)
0
9
37
75
150
Continuous models"
Hillb
Power
Linear, Polynomial
Human equivalent
continuous exposure
(ppm)
0
9
37
75
150
Goodness of fit
p-value
0.56
0.001
0.003
N
42
11
10
10
10
AIC
510
524
522
Maximum
X2 residual
near POD
0.3
0.6
0.6
Female Mice
Liver weight (g)
mean V s.d.
108+13.5
142 + 25.6
210 + 24.2
241+41.4
230 + 31.1
BMCS
(ppm)
3.2
6.6
6.6
BMCLS
(ppm)
0.6
5.1
5.1
5
6
7
8
9
10
11
12
13
14
1 Nonconstant variance models fitted.
3Hill model was the only adequately fitting model:
P(d) = intercept + v * dosen/(kn + dose"),
where: intercept = 108.0
v= 193.6
n= 1.08
k=35.3
Hill Model with 0.95 Confidence Level
150
100
1E:5402/D52002
70
Fin
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-88 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
Table 5A-3. Benchmark modeling summary: increased liver weight in
female mice exposed to tetrachloroethylene, using response data from
Kjellstrand et al. (1984) and human equivalent metabolized dose as dose
metric
Administered exposure,
24 hr/day, 30 days
(ppm)
0
9
37
75
150
Continuous models'"
Hiir
Power0
Polynomial (linear)
Human equivalent
metabolized dose"
(mg-eq/kg-day)
0.0
3.2
12
19
28
Goodness of fit
p-value
0.58
0.78
0.03
N
42
11
10
10
10
AIC
510
515
513
Maximum
X2 residual
near POD
0.4
0.3
0.4
Female Mice
Liver weights (g)
mean V s.d.
108+13.5
142 + 25.6
210 + 24.2
241+41.4
230 + 31.1
BMCS
(mg-eq/kg-day)
1.1
0.7
1.8
BMCLs
(mg-eq/kg-day)
0.3
0.2
1.4
6
7
8
9
10
11
12
13
a Metabolite levels were estimated using the Reitz et al. (1996) PBPK model, and adjusted to equivalent human
doses using surface area scaling by multiplying by [0.03 kg/70 kg]025 = 0.144.
b Nonconstant variance models fitted. Highest-dose group omitted due to poor fits for all models.
0 Among the models with adequate fits (p > 0.1), the Hill and Power models had very similar BMCLss. The
average of these BMCLss was 0.3 mg-eq/kg-day. The Hill model fit is shown as a representative of the two fits.
Using the three human pharmacokinetic models, the human equivalent inhalation exposures ranged from 1.4 to 10
ppm (see Figure 3-9 and Chapter 3).
Hill Model with 0.95 Confidence Level
150 -
100
09:43 02/DB 2002
20
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-89 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5A-4. Benchmark modeling summary: hepatic angiectasis in male
mice exposed to tetrachloroethylene, data from JISA (1993)
Administered exposure,
6 hrs/day, 5 days/wk,
2 years (ppm)
0
10
50
250
Quantal models
Gamma b
Log-logistic b
Multistage b
Weibullb
Log-Probitb
Probit
Logistic
Human equivalent
continuous exposure"
(ppm)
0.0
1.8
8.9
45.0
Goodness of fit
p-value
0.57
0.78
0.66
0.59
1
0.02
0.02
N
50
50
50
50
AIC
161
161
160
161
161
166
167
Maximum
X2 residual
near POD
0.4
0.2
0.3
0.4
0
2.2
1.5
Incidence (%) of hepatic
angiectasis in male mice
1
3
12
30
BMC10
(ppm)
3.6
3.8
4.9
3.6
3.8
12.2
13.3
(2)
(6)
(24)
(60)
BMCL10
(ppm)
1.4
1.7
3.7
1.5
1.8
10
11
4
5
6
7
a Exposure adjusted by 6/24 (hrs/day) x 5/7 (days/wk) to estimate equivalent continuous exposure levels.
b All models except logistic and probit, had acceptable fits (p > 0.1), similar AICs, and BMCL10s within a factor of
3 of each other. Average of these 5 BMCL10s is 2 ppm. The log-probit model fit is shown as a representative fit.
0.8
0.7
0.6
"5 0.5
U
O)
g 0.4
o
'•3 0-3
CO
^ 0.2
0.1
0
Probit Model with 0.95 Confidence Level
Probit
BMDL
BMD
0
10
40
dose
13:2502/082002
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-90 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 5A-5. Benchmark modeling summary: hepatic angiectasis in male
mice exposed to tetrachloroethylene, using data from JISA (1993) and
human equivalent metabolized dose as dose metric
Administered exposure,
6 hr/day, 5 days/wk,
2 years
(ppm)
0
10
50
250
Quantal models
Gammab
Logisticb
Log-logistic13
Multistage13
Probitb
Log-Probitb
Weibullb
Human equivalent
metabolized dose"
(mg-eq/kg-day)
0.0
0.6
2.3
5.8
Goodness of fit
p-value
0.87
0.23
0.73
0.85
0.35
0.51
0.95
N
50
50
50
50
AIC
161
162
161
161
161
161
161
Maximum
X2 residual
near POD
0.1
1.3
0.2
0.1
1.1
0.4
0.04
Incidence (%) of hepatic
angiectasis in male mice
1
3
12
30
BMC10
(mg-eq/kg-
day)
1.2
2
1.2
1.1
1.8
1.2
1.2
(2)
(6)
(24)
(60)
BMCL10
(mg-eq/kg-
day)
0.6
1.6
0.6
0.6
1.5
0.6
0.6
5
6
7
8
9
10
11
12
Metabolite levels were estimated using the Reitz et al. (1996) PBPK model, were estimated adjusted to equivalent
human doses using surface area scaling by multiplying by [0.048 kg/70 kg]0'25 = 0.162.
All models achieved satisfactory fits (p > 0.1, similar AICs), except the logistic and probit models were least
consistent with the data, having acceptable but relatively large %2 residuals near the BMCL10. The BMCL10s from
the remaining models were all 0.6 mg-eq/kg/day. The multistage model fit is shown as a representative fit. Using
the three human pharmacokinetic models with the POD of 0.6 mg-eq/kg-day, the human equivalent inhalation
exposures ranged from 4.3-23 ppm.
Multistage Model with 0.35 Confidence Level
0 1
12:4702/082002
3
1 ?
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-91 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5A-6. Benchmark modeling summary: incidence of karyomegaly in
male rats exposed to tetrachloroethylene, data from NTP (1986)
Administered exposure,
6 hrs/day, 5 days/wk,
2 years (ppm)
0
200
400
Quantal models
Gamma, Log-logistic, Log-
Probit, Weibull
Logistic
Multistage13
Probit
Human equivalent
administered dose
(ppm)a
0
36
71
Goodness of fit
p-value
N
49
49
50
AIC
Maximum
X2 residual
near POD
Incidence (and %) of
Karyomegaly in male rats
1
37
47
BMC10
(ppm)
(2)
(76)
(96)
BMCL10
(ppm)
NA
<0.01
0.98
<0.01
98.3
91
101
1.3
0.01
1.9
10.6
2.7
9.8
7.8
2.2
7.5
4
5
6
7
8
9
10
11
12
13
a Exposure adjusted by 6/24 (hr/day) x 5/7 (days/wk) to estimate equivalent continuous exposure levels.
b With only two nonzero exposure groups, options for fitting these data were limited. Among the models with two
or less parameters to estimate, the multistage model was the only one to fit adequately (p>0. 1).
P(d) = 1 - exp(-q0 -
where: d = continuous exposure level (ppm)
qo = 0.02
qi = 0.04
Multistage Model with 0.95 Confidence Level
0.8
O.B
0.4
0.2
0
Multistage
BMDLBMD
0 10
16:36 10/10 2001
20 30 40 50
dose
60
70
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-92 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5A-7. Benchmark modeling summary: deaths by Day 29 in offspring
of rats exposed to tetrachloroethylene, data from Tinston (1994)
Administered exposure,
6 hrs/day, 5 days/wk, 11
weeks; every day during
gestation
(ppm)
0
100
300
1000
Nested models'"
Nested logistic
NCTRd
Rai and vanRyzind
Human equivalent
administered
exposure"
(ppm)
0
18
54
180
Goodness of fit
p-value
0.072
0.12
0.12
Number of F2A
Litters
23
22
21
19
AIC
656.9
656.9
656.9
Average per-litter percent
of deaths by Day 29
8.4
9.5
11.4
33.5
BMC0ic
(ppm)
24.4
23
23
BMCIV
(ppm)
2.1
1.8
1.8
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
a Exposures adjusted to equivalent continuous exposures by multiplying by 6/24 (hrs/day) x 5/7 (days/wk).
b All nested models fit best when including intralitter correlations. Litter size was not used as a litter-specific
covariate.
0 BMCLoi selected as relevant response level because of the severity of the response (pup death) and because the
response level occurred within the range of the data set.
d The NCTR and the Rai and van Ryzin model fits were identical for these data, whereas the nested Logistic did not
provide an adequate fit.
P(d)= l-exp[-(a + (3 xdp)],
where: d = continuous exposure level (ppm),
a =0.089
P = 4.38 x 10'5
p=1.73
0.4
0.35
-a 0.3
0
I 0.25
<
I 0.2
ro
"• 0.15
0.1
0.05
NCTR
NCTR Model with 0.95 Confidence Level
BMDL |BMD
50
100
150
dose
15:1911/022004
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-93 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
Table 5A-8. Benchmark modeling summary: liver to body weight ratio in
rats exposed to tetrachloroethylene, data from Buben and O'Flaherty (1985)
Administered doses (5 days/wk,
6 weeks) (mg/kg-day)
0
20
100
200
500
1000
1500
2000
Continuous models'"
Hillc
Power
Polynomial, Linear
Goodness of fit
p-value
0.11
O.001
O.001
N
26
13
13
15
15
19
6
6
AIC
-24.1
28
24
Maximum
X2 residual
near POD
0.8
1
1
Liver weight/ body weight, mean
% + s.d.
5.21+0.46
5.51+0.40
5.97+0.40
6.45+0.46
7.35+0.62
7.89+0.70
8.10+0.66
9.00+0.27
BMDS
(mg/kg-day)
8.3
38
38
BMDLs
(mg/kg-day)
6.4
33
33
4
5
6
7
8
9
10
11
12
13
14
15
a Gavage doses were adjusted for daily exposure (x 5/7) and were adjusted to equivalent human doses using surface
area scaling by multiplying by [0.048 kg/70 kg]025 = 0.162.
b Constant variance models used.
0 Hill model was the only model to fit adequately.
P(d) = intercept + v * dose7(kn + dose"),
where: intercept = 5.27
v = 4.18
n=l
k=58.2
Hill Model with 0.95 Confidence Level
17:2201/302002
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-94 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 5A-9. Benchmark modeling summary: liver to body weight ratio in
rats exposed to tetrachloroethylene, using data from Buben and O'Flaherty
(1985) and human equivalent metabolized dose as dose metric
Administered doses
(5 days/wk, 6 wks)
(mg/kg-day)
0
20
100
200
500
1000
1500
2000
Continuous models'"
Hiir
Power"
Linear, Polynomial °
Continuous equivalent
metabolized dose"
(mg-eq/kg-day)
0.0
0.58
2.52
3.51
5.83
8.15
9.94
11.6
Goodness of
fit
/7-value
0.51
0.35
0.35
AIC
-30.9
-30.5
-32.5
N
26
13
13
15
15
19
6
6
Maximum
X2 residual
near POD
0.2
0.2
0.6
Liver weight/ body
weight, mean % + s.d.
5.21+0.46
5.51+0.40
5.97+0.40
6.45+0.46
7.35+0.62
7.89+0.70
8.10+0.66
9.00+0.27
BMDS
(mg-eq/kg-day)
8.8
7.9
7.9
BMDLS
(mg-eq/kg-day)
NA
6.5
6.5
5
6
7
8
9
10
11
12
13
' Metabolites for mice were estimated using the Reitz et al. (1996) model and were adjusted for continuous daily
exposure (x 5/7).
' Nonhomogeneous variance models.
! All continuous models fit reasonably well, except the Hill model could not provide a BMDL. The power model
fit is shown as a representative of the other two fits. The BMDLS was converted to an equivalent human dose
using surface area scaling by multiplying by [0.048 kg/70 kg]0'25 = 0.162, or 1.1 mg-eq/kg-day. Using the three
human pharmacokinetic models, the human equivalent oral exposures ranged from 3.4 to 32 mg/kg-day (see Table
5-10 for conversion factors).
Power Model with 0.95 Confidence Level
12
15:0307/11 2002
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-95 DRAFT-DO NOT CITE OR QUOTE
-------
1 APPENDIX 5B:
2 PROBABILITY DISTRIBUTIONS OF CANCER RISK ESTIMATES
O
4
5 Given the importance of characterizing central tendency estimates of risk when feasible
6 (U.S. EPA, 2005a), and the observation that MLEs resulting from typical dose-response analyses
7 can be unstable, an analysis of the distributions underlying the range of site-specific
8 tetrachloroethylene estimates of risk per unit concentration was undertaken. In addition, the
9 distributions underlying estimates of risk per unit concentration based only mononuclear cell
10 leukemias was explored for three dose-response models frequently used for noncancer dose-
11 response assessment in the observable range of the experimental data: the Weibull, log-logistic,
12 and log-probit models.
13 The bootstrap analysis (Efron and Tibshirani, 1993) was used to characterize the
14 distributions of risk per unit concentration for the six tumor/sex types identified for
15 tetrachloroethylene. For each of the six data sets in Figure 5-14 (see Tables 5-4 and 5-6 for
16 group data), for each exposure group a simulated incidence level was generated using binomial
17 distribution with probability of success equal to the observed incidence. This was repeated until
18 there were 10,000 simulated experiments for each tumor type. Then each simulated data set was
19 used to obtain estimates of BMDs using BMDS (U.S. EPA, 2000) in the same manner as for the
20 tetrachloroethylene assessment, including using the multistage model. The BMDs were
21 estimated at a benchmark response (BMR) of 10% extra risk for all sites except kidney tumors,
22 which were evaluated at 5% extra risk because 10% fell above the observed data. Distributions
23 of cancer slope values were obtained by calculating the distributions of the ratios BMR/BMDs.
th th
24 Upper and lower bounds on the linear extrapolation were determined by the 95 and 5
25 percentiles of the resulting distributions.
26 In the same manner as in the preceding paragraph, the bootstrap analysis was used to
27 characterize the distributions of risks per unit concentration resulting from fitting the male rat
28 leukemia data with the multistage, Weibull, log-logistic, and log-probit models. These models
29 were fit to the mononuclear cell leukemia data in male rats using the EPA BMDS program
30 without restricting the shape parameters of the latter three models. Parameters describing the
31 risk distribution (mean, median and 95% upper and lower confidence bounds) were estimated by
32 a bootstrap procedure, because these parameters are not all readily available in the current
33 BMDS software. The resulting risk distributions were compared at two exposure levels—at a
34 generalized POD selected near the 10% response level estimated by the multistage model fit, at
35 1.5 mg-eq/kg-day, and at an environmental level 105-fold lower than the generalized POD, at 1.5
36 x 10"5 mg-eq/kg-day. In comparing these distributions from the bootstrap procedure, the mean
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-96 DRAFT-DO NOT CITE OR QUOTE
-------
1 (as a measure of central tendency) and the 95% upper and lower bounds calculated from the
2 bootstrap procedure are of interest.
3 5.B.I. MULTISTAGE MODEL RESULTS
4 Table 5B-1 compares cancer risk values calculated based on BMDS output (left half of
5 table) with those calculated based on the bootstrap distribution of BMDs (right half of table).
6 BMDS estimates the BMD as an MLE, and derives the 95% lower bound (BMDL) and the 95%
7 upper bound (BMDU) on the BMD using the asymptotic distribution of the profile likelihood.
8 Dividing the BMR by these values, one obtains estimates of the slopes of linear extrapolation to
9 background responses from the BMDLs and BMDUs.
10 One can observe that there is generally a very good correspondence between asymptotic
11 (BMDS) results and re-sampling (bootstrap- based) results. This is in agreement with analysis of
12 other models in BMDS (Moerbeek et al., 2004), but differs from the conclusions of Bailer and
13 Smith (1994). However, the latter paper's conclusions were based on 1,000 runs, and Moerbeek
14 et al. (2004) demonstrated that at least 2,000 runs are needed to stabilize confidence interval
15 estimates. Additionally, risk estimates corresponding to the BMD were derived using the
16 average of the inverse distribution of BMDs. While agreement with risk estimates calculated
17 using BMDS is generally good, for one data set (male mice liver tumors) the discrepancy is
18 noticeable, with the MLE (BMD) and bootstrap estimates differing by about 50% (8.07 x io"3 vs.
19 1.16x10"). The difference is due to asymmetry of the distribution of BMDs, so that the MLE
20 may be different from the average of the distribution in such situations. The estimate based on
21 the bootstrap average is therefore a preferred estimate of central tendency in such a case.
22
23 5.B.2. RESULTS USING ALTERNATE MODELS
24 At 1.5 mg-eq/kg-day, corresponding to a risk of approximately 0.1 using the mean
25 estimate of risk from the multistage model, the bootstrap procedure yielded similar risk
26 distributions for the four models (see Table 5B-2 and Figure 5B-la). The means and
27 corresponding confidence bounds agree within an order of magnitude and the spreads in the
28 distributions (the distance between the upper and lower confidence bounds) are within two
29 orders of magnitude. Note that the probability calculations are in terms of metabolized dose in
30 the male rat, and do not directly characterize human risks.
31 EPA also examined the bootstrap results from those same models at a dose IO5 -fold
32 lower (although EPA's actual low dose risk estimates are developed using a linear extrapolation
33 from a POD to the origin rather than using low-dose estimates from a model). These results are
34 shown in Table 5B-2 and Figure 5B-lb. In the region of extrapolated concentrations (1.5 x 10
-5
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-97 DRAFT-DO NOT CITE OR QUOTE
-------
1 mg-eq/kg-day), the mean risks of the latter three models (Weibull, log-logistic and log-probit)
2 are about one to three orders of magnitude higher than the mean of the multistage model risks.
3 The spreads of all the models are quite broad, with a six order of magnitude 95% confidence
4 interval for the multistage and much greater spreads for the other three models. The upper
5 bounds of risk for the other three models are higher than that for multistage model, within about
6 three orders of magnitude, and their lower bounds of risk are much lower than that of multistage,
7 by nine or more orders of magnitude. With such large spreads in confidence intervals, the
8 extrapolated models in effect provide little information about low-dose risks. The extrapolation
9 of the multistage model does result in estimates reasonably close to the low-dose estimates from
10 the model-independent straight line extrapolation from the POD, in that the mean and upper-
11 bound risks at the lower concentration are both within 10% of the estimates resulting from
12 applying linear extrapolation to the results at the higher concentration.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-98 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
Table 5B-1. Comparison of BMDS and bootstrap-based cancer risk values
derived using the multistage model to fit rodent bioassay data for tumor
types associated with tetrachloroethylene exposure, and using total
metabolites as the dose metric
Tumor
type
Male rat
kidney
tumorsb
Male mice
liver
hemangio-
sarcomas
Male mice
liver
tumors
Female
mice liver
tumors
Male rat
leukemias
Female rat
leukemias
Model
ordera
2
1
o
3
i
3
3
o
5
2
3
1
BMR
0.05
0.05
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Slopes, (mg-eq/kg-day) , of linear
extrapolation using BMDS
estimates of:
BMD
1.22 x 10"2
1.22 x 10"2
4.82 x 10"3
4.82 x 10"3
8.07 x 10"3
5.33 x 10"3
7.64 x 10"2
7.64 x 10"2
2.46 x 10"2
2.46 x 10"2
BMDL
2.63 x 10"2
2.63 x 10"2
8.04 x 10"3
8.04 x 10"3
2.77 x 10"2
1.04 x 10"2
1.24 x 10"1
1.24 x 10"1
4.93 x 10"2
4.93 x 10"2
BMDU
NDC
ND
2.06 x 10"3
2.06 x 10"3
5.46 x 10"3
4.14 x 10"3
2.67 x 10"2
2.93 x 10"2
ND
ND
Bootstrap-based slopes,
-i
(mg-eq/kg-day)
Mean
1.20 x 10"2
1.24 x 10"2
4.61 x 10"3
4.88 x 10"3
1.16 x 10"2
6.02 x 10"3
6.75 x 10"2
6.89 x 10"2
2.36 x 10"2
2.49 x 10"2
Upper
Bound
2.32 x 10"2
2.52x 10"2
7.66 x 10"3
8.01 x 10"3
2.77 x 10"2
9.69 x 10"3
1.16 x 10"1
1.16 x 10"1
4.81 x 10"2
5.03 x 10"2
Lower
Bound
ND
ND
2.26 x 10"3
1.97 x 10"3
5.76 x 10"3
4.30 x 10"3
2.92 x 10"2
3.30 x 10"2
ND
ND
6
7
8
9
10
a Order of polynomial in multistage model. Different order models were compared where various order models fit
equally well.
b Data set had 2 non-zero dose groups, all others had three. See Tables 5-4 and 5-6 for data and study details.
0 ND = could not be determined.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-99 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
Table 5B-2. Comparison of BMDS and bootstrap-based cancer risk values
derived using several dose-response models to fit rodent bioassay data for
mononuclear cell leukemia incidence in male rats associated with
tetrachloroethylene exposure, and using total metabolites as the dose metric
Model
Multistage
Log-probit
Log-logistic
Weibull
Dose,
mg-eq/kg-day
1.5
1.5x 10'5
1.5
1.5 x ID'5
1.5
1.5 x 10'5
1.5
1.5 x 10'5
MLE of risk
1.143 x ID'1
1.168 x ID'6
9.717 x ID'2
4 x ID'25
1.018 x 10'1
2.553 x 10'8
1.064 x 10'1
2.41 x 10'7
Bootstrap distribution of extra risk
Mean
9.442 x 1Q-2
9.172 x ID'7
1.037 x 1Q-1
8.172xlO-4
1.079 x 10'1
1.078 x 10'3
1.124 x 10'1
1.339 x 10'3
Median
9.853 x 1Q-2
9.933 x 1Q-7
9.544 x 1Q-2
1 x 1Q-26
9.937 x 10'2
1.401 x 10'8
1.052 x 10'1
1.643 x 10'7
S^percentile
1.163 x 1Q-2
4.546 x 1Q-13
8.366 x 1Q-4
2 x 10-170
4.320 x 10'3
2 x 10'21
5.594 x 10'3
0
95th percentile
1.672 x 1Q-1
1.819 x 1Q-6
2.426 x 1Q-1
1.008 x 1Q-5
2.429 x 10'1
7.776 x 10'4
2.452 x 10'1
1.893 x 10'3
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-100 DRAFT-DO NOT CITE OR QUOTE
-------
_>.
'55
0}
a
Risks for 4 models for dose=1.5; log scale
IO
O
I
• Multistage (M)
• Weibull (W)
n Logprobif (P)
• Loglogistic (L)
L 5th percentile 1 1
U 95th percentile M
— ~&r. . j\
LiVI UIVI
LW UW
LL LIL
l l l l
-15 -10 -5 0
LOG of Risk
Risks for 4 models for dose=1.5E-5; log scale
0)
Q
1
2
3
4
5
6
7
8
9
10
11
to
o
o
I
• Multistage (M)
• Weibull (W)
n Logprobit (P)
• Loglogistic (L)
/
f\ L 5th percentile
1 1 U 95th percentile
I
I I
/' i
\ _^___^. I - -
—ttfr uw
LW* UW
LL* UL
1 1 1 1
-15 -10 -5 0
LOG of Risk
* points are outside of the range of the graph
Figure 5B-1. Illustration of sensitivity to model selection for low-dose
extrapolation. The risk distributions associated with four dose-response models
which adequately fit the tetrachloroethylene dose-response data for MCL in male
rats (NTP, 1986) were compared. Risks are in terms of metabolized
tetrachloroethylene in male rats. Figure 5-2 shows the comparison at a
generalized POD, selected as the mean exposure estimate, from the multistage
model corresponding to a risk of approximately 0.1. Figure 5-3 compares risk
distributions at an exposure corresponding to environmental concentrations of
tetrachloroethylene, approximately 105-fold lower than the POD.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-101 DRAFT-DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR CHAPTER 5
2
O
4 Anger, WK; Letz, R; Chrislip, DW; et al. (1994) Neurobehavioral test methods for environmental health studies of
5 adults. Neurotoxicol Teratol 16:489-497.
6
7 Anger, WK; Liang, YX; Nell, V; et al. (2000) Lessons learned-15 years of the WHO-NCTB: a review.
8 Neurotoxicology 21:837-846.
9
10 Anger, WK. (2003) Neurobehavioural tests and systems to assess neurotoxic exposure in the workplace and
11 community. Occup Environ Med 60:531-538.
12
13 Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
14 low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
15
16 Amler, RW; Lybarger, JA: Anger, WK; et al. (2004) Adoption of an adult environmental neurobehavioral test
17 battery. Neurotoxicol Teratol 16:525-530.
18
19 ATSDR (Agency for Toxic Substances and Disease Registry). (1996) Pediatric Environmental Neurobehavioral Test
20 Battery. Amler, RW; Gibertini, M; eds. Atlanta, GA: U.S. Department of Health and Human Services, Public
21 Health Service, Agency for Toxic Substances and Disease Registry.
22
23 ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene.
24 Springfield, VA.
25
26 Bailer, AJ: Smith, RJ. (1994) Estimating Upper Confidence Limits for Extra Risk in Quantal Multistage Models.
27 Risk Analysis 14(6): 1001-1010.
28
29 Barnes, DG; Dourson, M. (1988) Reference dose (RfD): description and use in health risk assessments. Regul
30 Toxicol Pharmacol 8:471-486.
31
32 Bellies, RP; Brusick, DJ; Mecler, FJ. (1980) Teratogenic-mutagenic risk of workplace contaminants:
33 trichloroethylene, perchloroethylene and carbon disulfide. Prepared by Litton Bionetics, Inc. NTIS Publication No.
34 PB-82 185-075, NIOSH Contract Report No. 201-77-0047. Available from: National Technical Information
3 5 Service, Springfield, VA.
36
37 Berman, E; Schlicht, M; Moser, VC; et al. (1995) A multidisciplinary approach to lexicological screening: I.
38 Systemic toxicity. J Toxicol Environ Health 45:127-143.
39
40 Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol
41 70:347-355.
42
43 Brodkin, CA; Daniell, W; Checkoway, H; et al. (1995) Hepatic ultrasonic changes in workers exposed to
44 perchloroethylene. Occup Environ Med 52:679-685.
45
46 Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
47 and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
48
49 Cal EPA (California Environmental Protection Agency). (2001) Public health goal for tetrachloroethylene in
50 drinking water. Prepared by the Office of Environmental Health Hazard Assessment. Available online at
51 http://www.oehha.org/water/phg/pdf/PCEAug2001 .pdf.
52
53 CARD (California Air Resources Board). (1991) Proposed identification of perchloroethylene as a toxic air
54 contaminant. Oakland, CA.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-102 DRAFT-DO NOT CITE OR QUOTE
-------
1 CARD (California Air Resources Board). (1998) 1990-1996 Statewide perchloroethylene summary, ppb. Available
2 online at http://www.arb.ca.gov/aqd/perc/pcstate.htm.
3
4 Cavalleri, A; Gobba, F; Paltrinieri, M; et al. (1994) Perchloroethylene exposure can induce colour vision loss.
5 Neurosci Lett 179:162-166.
6
7 Chiu, WE; Bois, FY. (2006) Revisiting the population toxicokinetics of tetrachlorotethylene. Arch Toxicol 80:382-
8 385.
9
10 Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
11 pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79:381-393.
12
13 Echeverria, D; Heyer, N; Checkoway, H; et al. (1994) Behavioral investigation of occupational exposure to solvents:
14 Perchloroethylene among dry cleaners, and styrene among reinforced fiberglass laminators. Final Report. Prepared
15 for the Centers for Disease Control and Prevention under Grant No. 5 R01 OHo2719-03. Battelle Centers for Public
16 Health Research and Evaluation.
17
18 Echeverria, D; White, RF; Sampaio, C. (1995) A behavioral evaluation of PCE exposure in patients and dry
19 cleaners: a possible relationship between clinical and preclinical effects. J Occup Environ Med 37:667-680.
20
21 Efron, B; Tibshirani, PJ. (1993) An Introduction to the Bootstrap. Chapman and Hall, San Francisco.
22
23 Eskenazi, B; Wyrobek, AJ; Fenster, L; et al. (1991) A study of the effect of perchloroethylene exposure on semen
24 quality in dry cleaning workers. Am J Ind Med 20:575-591.
25
26 Ferroni, C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
27 perchloroethylene. Neurotoxicology 13:243-247.
28
29 Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
30 solvents. Int Arch Occup Environ Health 52:1-9.
31
32 Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
33 tetrachloroethylene as neonates. Toxicol Lett 66:13-19.
34
3 5 Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
36 tetrachloroethylene. Am J Ind Med 21:661-671.
37
38 Gentry, PR; Covington, TR; Clewell, HJ, III. (2003) Evaluation of the potential impact of pharmacokinetic
39 differences on tissue dosimetry in offspring during preganancy and lactation. Regul Toxicol Pharmacol 38:1-16.
40
41 Gobba, F; Righi, E; Fantuzzi, G; et al. (1998) Two-year evolution of perchloroethylene-induced color-vision loss.
42 Arch Environ Health 53:196-198.
43
44 Guth, DJ; Carroll, RJ; Simpson, DG et al. (1997) Categorical regression analysis of acute exposure to
45 tetrachloroethylene. Risk Analysis 17:321-332.
46
47 Hartwell, TD; Crowder, JH; Sheldon, LA; et al. (1985) Levels of volatile organics in indoor air. In: Proceedings of
48 the Air Control Pollution Association, 78th annual meeting, Vol. 3, 85-30.B., 2-12.
49
50 Hayes, JR; Condie, LW, Jr.; Borzelleca, JF. (1986) The subchronic toxicity of tetrachloroethylene
51 (perchloroethylene) administered in the drinking water of rats. Fundam Appl Toxicol 7:119-125.
52
53 JISA (Japan Industrial Safety Association). (1993) Carcinogenicity study of tetrachloroethylene by inhalation in rats
54 and mice. Data No. 3-1. Available from: EPA-IRIS Information Desk.
55
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-103 DRAFT-DO NOT CITE OR QUOTE
-------
1 Jonker, D; Woutersen, RA; Feron, VJ. (1996) Toxicity of mixtures of nephrotoxicants with similar or dissimilar
2 mode of action. Food Chem Toxicol 34:1075-1082.
3
4 Kjellstrand, P; Holmquist, B; Kanje, M; et al. (1984) Perchloroethylene: effects on body and organ weights and
5 plasma butyrylcholinesterase activity in mice. Acta Pharmacol Toxicol (Copenh) 54:414-424.
6
7 Krewski, D; vanRyzin, J. (1981) Dose response models for quantal response toxicity data, in statistics and related
8 topics. In: Csorgo, M; Dawson, DA; Rao, INK; Saleh, ADMdE; eds. North-Holland Publishing Company.
9
10 Lutz, WK; Gay lor, DW; Conolly, RB; et al. (2005) Nonlinearity and thresholds in dose-response relationships for
11 carcinogenicity due to sampling variation, logarithmic dose scaling, or small differences in individual susceptibility.
12 Toxicol Appl Pharmacol 207:565-569.
13
14 Mattsson, JL; Albee, RR; Yano, BL; et al. (1998) Neurotoxicologic examination of rats exposed to 1,1,2,2-
15 tetrachloroethylene (perchloroethylene) vapor for 13 weeks. Neurotoxicol Teratol 20:83-98.
16
17 Moerbeek, M; Piersma, AH; Slob, W. (2004) A Comparison of Three Methods for Calculating Confidence Intervals
18 for the Benchmark Dose. Risk Analysis 24(1):31-40.
19
20 Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-
21 cleaners. Lancet 340:189-193.
22
23 Muttray, A; Wolff, U; Jung, D; et al. (1997) Blue-yellow deficiency in workers exposed to low concentrations of
24 organic solvents. Int Arch Occup Environ Health 70:407-412.
25
26 NCI (National Cancer Institute). (1977) Bioassay of tetrachloroethylene for possible carcinogenicity. DHEW Pub.
27 (NIH):77-813.
28
29 Nelson, BK; Taylor, BJ; Setzer, JV; et al. (1980) Behavioral teratology of perchloroethylene in rats. J Environ
30 Pathol Toxicol 3:233-250.
31
32 NRC (National Research Council). (1983) Risk assessment in the Federal government: managing the process.
3 3 Committee on the Institutional Means for Assessment of Risks to Public Health, Commission on Life Sciences,
34 NRC. Washington, DC; National Academy Press
35
36 NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies fo tetrachloroethylene
37 (perchloroethylene) (CAS No. 127-18-4) inF344/N rats and B6C3F1 mice. NTP Technical Report 311. National
38 Institutes of Health, Public Health Service, U.S. Department of Health and Human Services, Washington, DC.
39
40 NYS DOH (New York State Department of Public Health). (1997) Tetrachloroethene ambient air criteria document.
41 Final Report. Albany, NY.
42
43 NYS DOH (New York State Department of Health) (2005) Improving human risk assessment for
44 tetrachloroethylene by using biomarkers and neurobehavioral testing. U.S. EPA Star Grant #R827445. Grant
45 #R827446. Available online at
46 http://cfpub.epa.gov/ncer abstracts/index.cfm/fuseaction/displav.abstractDetail/abstract/977/report/O.
47
48 NYS OAG (New York State Office of Attorney General). (2004) Letter from Judy Schreiber, Ph.D., for EPA
49 Docket ORD-2003-0014. July 30, 2004.
50
51 Opdam, JJ. (1989) Intra and interindividual variability in the kinetics of a poorly and highly metabolising solvent. Br
52 JIndMed46:831-845.
53
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-104 DRAFT-DO NOT CITE OR QUOTE
-------
1 Pelekis, M; Gephart, LA; Lerman, SE. (2001) Physiological-model-based derivation of the adult and child
2 pharmacokinetic intraspecies uncertainty factors for volatile organic compounds. Regul Toxicol Pharmacol 33:12-
3 20.
4
5 Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
6 groundwater for a bathing and showering determination. Risk Anal 13:37^9.
7
8 Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
9 for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
10 306.
11
12 Rhomberg, LR. (2000) Dose-response analyses of the carcinogenic effects of trichloroethylene in experimental
13 animals. Environ Health Perspect 108(S2):343-358.
14
15 Rosengren, LE; Kjellstrand, P; Haglid, KG. (1986) Tetrachloroethylene: levels of DNA and S-100 in the gerbil CNS
16 after chronic exposure. Neurobehav Toxicol Teratol 8:201-206.
17
18 Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
19 tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect 110:655-664.
20
21 Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
22 11:579-583.
23
24 Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
25 of subjects chronically exposed to organic solvents (II)]. G Ital Med Lav Ergon 19:85-88.
26
27 Tinston, DJ. (1994) Perchloroethylene: A multigeneration inhalation study in the rat. CTL/P/4097. Available from:
28 EPA IRIS Information Desk.
29
3 0 Trevisan, A; Macca, I; Rui, F; et al. (2000) Kidney and liver biomarkers in female dry-cleaning workers exposed to
31 perchloroethylene. Biomarkers 5:399-409.
32
33 Umezu, T; Yonemoto, J; Soma, Y; Miura, T. (1997) Behavioral effects of trichloroethylene and tetrachloroethylene
34 in mice. Pharmacol Biochem Behav 58:665-671.
35
36 U.S. EPA (Environmental Protection Agency). (1985) Health assessment document for tetrachloroethylene
3 7 (perchloroethylene). Office of Health and Environmental Assessment, Office of Research and Development,
38 Washington, DC; EPA/600/8-82/005F. Available from: National Technical Information Service, Springfield, VA;
39 PB-86-174489/AS.
40
41 U.S. EPA (Environmental Protection Agency). (1986) Addendum to the health assessment document for
42 tetrachloroethylene (perchloroethylene) [review draft]. Office of Health and Environmental Assessment, Office of
43 Research and Development, Washington, DC; EPA/600/8-82/005FA. Available from: National Technical
44 Information Service, Springfield, VA; PB-86-174489/AS.
45
46 U.S. EPA (Environmental Protection Agency). (1987) Technical analysis of new methods and data regarding
47 dichloromethane hazard assessments [review draft]. Office of Health and Environmental Assessment, Office of
48 Research and Development, Washington, DC; EPA/600/8-87/029A. Available from: National Technical
49 Information Service, Springfield, VA.
50
51 U.S. EPA (Environmental Protection Agency). (1988) IRIS summary of tetrachloroethylene RID. Available online
52 at http://www.epa.gov/iris/subst/0106.htm.
53
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-105 DRAFT-DO NOT CITE OR QUOTE
-------
1 U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
2 carcinogenicity of tetrachloroethylene (perchloroethylene). Office of Health and Environmental Assessment,
3 Washington, DC; EPA/600/6-91/002F. Available from: National Technical Information Service, Springfield, VA.
4
5 U.S. EPA (Environmental Protection Agency). (1992) Draft report: a cross-species scaling factor for carcinogen risk
6 assessment based on equivalence of mg/kg3/4/day. Federal Register 24152-24173.
7
8 U.S.EPA (1993) Reference Dose (RfD): Description and Use in Health Risk Assessments.
9
10 U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
11 and application of inhalation dosimetry. Office of Health and Environmental Assessment, Environmental Criteria
12 and Assessment Office, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical Information
13 Service, Springfield, VA; PB2000-500023, and online at http://www.epa.gov/ncea.
14
15 U.S. EPA (Environmental Protection Agency). (1995) IRIS summary of dichloromethane. Available online at
16 http://www.epa.gov/iris/subst/0070.htm.
17
18 U.S. EPA (Environmental Protection Agency). (2000) Benchmark dose technical guidance document [external
19 review draft]. Risk Assessment Forum. Washington, DC; EPA/630/R-00/001. Available online at
20 http://www.epa.gov/ncea/raf.
21
22 U.S. EPA (Environmental Protection Agency). (2001) Trichloroethylene health risk assessment: synthesis and
23 characterization. National Center for Environmental Assessment, Office of Research and Development,
24 Washington, DC; EPA/600/P-01/002A. Available online at
25 http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=4580.
26
27 U.S. EPA (Environmental Protection Agency). (2004) Summary report of the peer review workshop on the
28 neurotoxicity of tetrachloroethylene (perchloroethylene) discussion paper. National Center for Environmental
29 Assessment, Washington, DC; EPA/600/R-04/041. Available online at http://www.epa.gov/ncea.
30
31 U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Federal Register
32 70(66)17765-17817. Available online at http://www.epa.gov/cancerguidelines.
33
34 U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
35 early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
36 online at http://www.epa.gov/cancerguidelines.
37
38 Verplanke, AJ; Leummens, MH; Herber, RF. (1999) Occupational exposure to tetrachloroethylene and its effects on
39 the kidneys. J Occup Environ Med 41:11-16.
40
41 Wang, S; Karlsson, JE; Kyrklund, T; et al. (1993) Perchloroethylene-induced reduction in glial and neuronal cell
42 marker proteins in rat brain. Pharmacol Toxicol 72:273-278.
43
44 Warren, DA; Reigle, TG; Muralidhara, S; et al. (1996) Schedule-controlled operant behavior of rats following oral
45 administration of perchloroethylene: time course and relationship to blood and brain solvent levels. J Toxicol
46 Environ Health 47:345-362.
47
48 Zeise, L; Wilson, R; Crouch, E A. (1987) Dose-response relationships for carcinogens: a review. Environ Health
49 Perspect 73:259-306.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 5-106 DRAFT-DO NOT CITE OR QUOTE
-------
1 6. CHARACTERIZATION OF HAZARD AND DOSE-RESPONSE
2
3
4 6.1. SUMMARY OF HUMAN HAZARD POTENTIAL
5 6.1.1. Exposure
6 Tetrachloroethylene (CASRN 127-18-4) is a solvent used for cleaning clothes and for
7 metal cleaning and degreasing. It is a volatile liquid at room temperature. The largest human
8 exposure occurs indoors to workers in dry cleaning, laundry, and metal finishing facilities.
9 Release of tetrachloroethylene into the air from these facilities also results in measurable outdoor
10 ambient air concentrations. Indoor residential exposure can also occur when dry cleaning
11 facilities are located within residential areas. It has been detected in breast milk of women
12 exposed to tetrachloroethylene in ambient air in or near these facilities. Tetrachloroethylene can
13 enter water supplies, and it has been detected in drinking water. Exposure to airborne
14 tetrachloroethylene can occur in homes via volatilization from tap water during showering as
15 well as from water ingestion in homes with contaminated drinking water (see Chapter 2 for more
16 information).
17
18 6.1.2. Absorption, Metabolism, Distribution, and Excretion
19 Tetrachloroethylene rapidly enters body tissues after inhalation, ingestion, and dermal
20 exposure. Tetrachloroethylene metabolism is considered to be well characterized in rodents but
21 not in humans. A significant portion of tetrachloroethylene inhaled by humans at ambient
22 concentrations is not metabolized (about 64% according to the pharmacokinetic model of Bois,
23 et al. 1996). The recovered metabolites in the urine represent only a fraction of what is actually
24 metabolized (Bogen and McKone, 1988). Possible explanations for metabolites not reaching the
25 urine are (a) binding to plasma proteins, (b) biliary excretion, (c) enterohepatic circulation of
26 metabolites, (d) further metabolism of fat-sequestered parent compounds after the completion of
27 the studies, and (e) metabolism to currently unidentified metabolites. However, data to support
28 these hypotheses are sparse. The fraction of tetrachloroethylene metabolized appears to have a
29 strong dependence on the exposure concentration. The PBPK model by Bois et al. (1996)
30 predicts this fraction to be about 36% in humans at low environmental concentrations, whereas
31 the human data indicate a very small fraction would be metabolized at higher concentrations
32 (such as those corresponding to the laboratory animal bioassays; see Section 3.5 for more
33 details.)
34 There are two major routes of metabolism: (1) the predominant oxidative pathway,
35 which results in TCA and other urinary metabolites, as well as reactive intermediates and carbon
36 dioxide; and (2) the GSH conjugation pathway, which results in TCVG and TCVC that are
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 further processed to chemically reactive products that can bind to tissue macromolecules.
2 Metabolism occurs predominately in the liver. Further, metabolism of the GSH metabolites,
3 including activation by beta lyase, occurs in the kidney. In addition, the CYP enzymes of the
4 oxidative pathway as well as enzymes important to the GSH pathway are present in other tissues.
5 Therefore, a potential exists for extrahepatic metabolism and formation of reactive metabolites at
6 sites other than the liver and kidney (see Section 3.3.3.2).
7 Many steps in the oxidative metabolism of tetrachloroethylene are well characterized in
8 both animals and humans; however, not all proposed intermediates have been identified or
9 detected. Although an initial epoxide metabolite has not been unequivocally demonstrated for
10 tetrachloroethylene, the epoxide intermediate is a reasonable1 proposal. It has been chemically
11 synthesized, and it is metabolized to TCA when injected into rodents. The tetrachloroethylene
12 epoxide intermediate is considered to be unstable and short-lived in vivo and is thought to
13 spontaneously rearrange and convert to other intermediate metabolites. Formation of
14 trichloroacetyl chloride directly from tetrachloroethylene via the mechanism of CYP-mediated
15 olefin oxidation without the obligatory formation of the epoxide intermediate has also been
16 postulated. The formation of trichloroacylated protein adducts in liver and kidney of rats, liver
17 of mice, and plasma of rats and humans following exposures to tetrachloroethylene provides
18 evidence of a metabolic intermediate that can react with tissue proteins. TCA, a stable
19 metabolite, is believed to result primarily from the oxidation of tetrachloroethylene to the
20 potentially reactive trichloroacetyl choride. TCA has been detected in the blood and urine of
21 humans and laboratory rodents, and excretion in urine is used as a biomarker for
22 tetrachloroethylene exposure (see Section 3.3.3 for more details).
23 Other steps in tetrachloroethylene oxidative metabolism are not as well characterized.
24 Both TCOH and TCOH-glucuronide have been detected in the urine of mice and humans
25 following tetrachloroethylene exposure; however, there is uncertainty about formation of TCOH
26 and its chloral hydrate precursor from tetrachloroethylene because not all studies have detected
27 TCOH as a urinary excretion product. These different findings could be explained in several
28 ways: (a) TCOH could be an artifact of the analytical methodology, (b) differences could exist
29 in analytical methodologies, (c) contamination could have occurred by unknown exposures to
30 another chemical, and (d) excretion of TCOH and its glucuronide might be dependent on dose.
31 If the TCOH pathway exists in humans, the overall contribution to TCA from TCOH is expected
32 to be relatively small when compared with the amount of TCA resulting from trichloroacetyl
33 chloride.
1 "Reasonable," as used is this chapter, is intended to imply the use of reasoned scientific judgment in data
evaluation and decision-making, consistent with U.S. EPA practices and guidance (e.g., Guidelines for Carcinogen
Risk Assessment, 2005a)
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 DC A is an intermediate metabolite that has been identified in the urine of rats, but not
2 humans, exposed by inhalation to tetrachloroethylene. It is not clear whether DCA is a product
3 of further metabolism of TC A or a product of another pathway originating with GSH
4 conjugation, or both. The amount of DCA produced from tetrachloroethylene oxidative
5 metabolism could vary across species and is likely to be relatively small when compared with the
6 amount of TCA produced. Data on mutational changes in tetrachloroethylene-induced liver
7 tumors in mice also support a limited role for DCA in tetrachloroethylene toxicity.
8 Quantitatively, the GSH conjugation pathway is relatively minor when compared with the
9 P450 conjugation pathway. Urinary mercapturates comprise from 1% to as little as 0.03% of
10 total recovered urinary metabolites (Green et al., 1990; Birner et al., 1996); however, this urinary
11 excretion product does not reflect the amount of compound going through the GSH pathway, but,
12 rather, it reflects only the portion that is excreted. The amount of the mercapturate product
13 excreted in the urine also does not reflect the amount of the more important portion that is
14 converted to toxic by-products through further metabolism. TCVG is the conjugation product of
15 tetrachloroethylene, and its cleavage product, TCVC, reacts with a kidney enzyme, beta lyase, to
16 produce metabolites that are mutagenic in bacteria and are also cytotoxic. These metabolites are
17 believed to contribute to tetrachloroethylene-induced kidney toxicity. Uncertainty exists as to
18 the relative contribution of GSH metabolism to toxicity in humans as compared with the rat due
19 to study differences in reported amounts of human tetrachloroethylene GSH metabolism as
20 measured by excreted mercapturate. These differences may be due, in part, to different chemical
21 assay methodology or to problems resulting from the stability of the chemical product being
22 measured, or both. In spite of these uncertainties, some of the published findings concerning
23 TCVG production would not predict any less susceptibility for humans than for rodents with
24 regard to renal toxicity. The higher percentage of mercapturate found in rat versus human urine
25 does not indicate a higher level of production of toxic products in the rat, because excreted
26 mercapturate allows no estimate of the amount of TCVC or N-acetyl TCVC being processed
27 through alternate routes. Furthermore, it is not known whether sex-dependent variation of beta
28 lyase activity exists in humans as it does in rats. Human variation might also explain study
29 differences in reported excretion rates (see Section 3.3.3.2).
30 Several enzymes of the oxidative and GSH metabolism, notably CYP2E1, CYP3A4,
31 GSTZ, GSTA, GSTM, and GSTT, show genetic polymorphisms with the potential for variation
32 in production of specific metabolites. The effect of metabolic variation, including potential
33 implications for differential toxicity, has not been well studied. The limited data available on
34 tetrachloroethylene metabolites show DCA to be a potent, irreversible inhibitor of GSTZ
35 activity, with greater inhibition of this enzyme in mice than in humans. Studies show that
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 inducers of CYP enzymes, such as toluene, phenobarbital, and pregnenolone-16 alpha-
2 carbonitrile, increase tetrachloroethylene metabolism, whereas CYP enzyme inhibitors such as
3 SKF 525A, metyrapone, and carbon monoxide decrease tetrachloroethylene metabolism.
4 Additionally, chronic exposure to tetrachloroethylene has been shown to cause self-induction of
5 metabolism (see Section 3.3.4).
6
7 6.1.3. Noncancer Toxicity in Humans and Laboratory Animals
8 Targets of toxicity observed in human and animal studies include the liver, kidney, CNS,
9 reproductive system, and developing fetus. Both occupational and residential epidemiologic
10 studies have examined the effects of tetrachloroethylene exposure via inhalation. Humans were
11 found to be particularly sensitive for neurological effects, including decrements in vision or
12 visuo-spatial function, and other neurobehavioral (cognitive) effects following inhalation
13 exposure. These findings are supported by the consistency of the observations in a number of
14 epidemiologic studies of different designs, populations, and statistical analyses, despite study
15 flaws. Altmann et al. (1995) identified a pattern of neurobehavioral deficits in a study of
16 residents living in buildings co-located with a dry cleaning establishment that is similar to the
17 pattern observed in occupational populations with tetrachloroethylene exposures, thus providing
18 evidence of an association with nonoccupational exposure. A second residential study
19 (Schreiber et al., 2002) also suggests that children may be uniquely susceptible to visuo-spatial
20 effects, but larger studies in humans and studies using animal models are needed to confirm this
21 observation as well as reports of color vision discrimination and contrast sensitivity (black-white
22 discrimination) changes. The large body of evidence assessing neurobehavioral effects and
23 tetrachloroethylene does not permit a distinction between acute effects and effects of repeated
24 exposure. Furthermore, no studies are available to evaluate chronic disabling neurological
25 disease. Occupational studies have examined the effects of tetrachloroethylene on other
26 endpoints as well, with the strongest evidence being for markers of liver and kidney damage and
27 for reproductive/developmental effects such as spontaneous abortion. The few studies
28 examining inhalation exposure to tetrachloroethylene and immune or endocrine system effects
29 are inadequate for fully evaluating potential associations.
30 The measure of the extent of exposure in many of the epidemiologic studies is imprecise,
31 and, in occupational situations, there are potential exposures to other solvents, although to a
32 lesser extent than with tetrachloroethylene. Relationships between exposure to
33 tetrachloroethylene and responses are not generally observed. Possible explanations for this are
34 exposure misclassification due to use of current exposure measurements, an exposure or
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 response function that is above the increasing portion of the exposure-response curve, or, more
2 unlikely, a response that does not increase with increasing tetrachloroethylene exposure.
3 Epidemiologic studies of oral exposures to tetrachloroethylene have only examined
4 adverse pregnancy outcomes or postnatal effects (see Section 4.7.1). There is some evidence for
5 growth retardation in infants born to mothers residing in housing with drinking water
6 contaminated with tetrachloroethylene.
7 Tetrachloroethylene exposure to animals by the inhalation or oral route results in toxicity
8 to the liver, kidney, and nervous system; by inhalation, it also causes developmental and
9 reproductive effects. Specifically, several measures of toxicity have been observed in the liver,
10 such as, increased liver weight, infiltration of fat, necrosis, peroxisome proliferation, polyploidy
11 of hepatocytes, and increased triglycerides. In the kidney, increased weight, hyperplasia, hyaline
12 droplets, and protein cast formation in tubules have been observed. In the CNS, alteration of
13 brain neurotransmitter levels, increased motor activity, and delayed reaction times to visual
14 stimuli have been observed. Animals exposed in utero to tetrachloroethylene by inhalation
15 showed signs of fetal growth retardation, increased fetal mortality, and behavioral changes
16 occurring after birth. There is little information on developmental or reproductive effects in
17 animals by the oral route of exposure. There are very few studies of immune system toxicity,
18 and none of these studies are in intact animals. No information is available on the effects of
19 tetrachloroethylene on the endocrine system in animals.
20 Targets of toxicity are the same in animal and human studies (i.e., liver, kidney, CNS,
21 reproductive system, and developing fetus). The effect domain in animals and humans indicates
22 that both cognition and visual function are affected by tetrachloroethylene. Affected organs are
23 all sites of high metabolic activity, and the CNS is also a lipid accumulation site, consistent with
24 the absorption, distribution, metabolism, and elimination profile of tetrachloroethylene.
25
26 6.1.4. Carcinogenicity in Humans and Laboratory Animals
27 Overall, the epidemiologic evidence has associated tetrachloroethylene exposure with
28 excess risks for a number of cancers, although a causal association has yet to be definitively
29 established. Studies of tetrachloroethylene and cancer showed positive associations between
30 exposure and cancer of the lymphoid system, esophagus, and cervix, with more limited evidence
31 for cancer of the bladder, kidney, and lung. For both lymphoid and esophageal cancer, excess
32 risk was observed in studies of human populations exposed to tetrachloroethylene and other
33 solvents, including studies of exposures to dry cleaners or workers involved with degreasing
34 metal parts. In these cases, average risks were doubled as compared with those of referents.
35 Furthermore, studies of drinking water exposure also support an association between lymphoid
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 cancer and tetrachloroethylene and other solvents, as do case-control studies that assessed
2 employment as dry cleaners or laundry workers. Chance and confounding by smoking are
3 unlikely the sole explanations for the observed excesses in risks. Information is lacking on life
4 style and socioeconomic factors, which are indirect surrogates for human papilloma virus
5 infection, a known risk factor for cervical cancer (see Section 4.8.1.2 for more details).
6 The laboratory animal database includes 10 lifetime rodent bioassay data sets that
7 demonstrate increased cancer incidence. (Two additional study data sets, in male and female rats
8 exposed orally, were inconclusive due to excessive mortality caused by pneumonia or
9 tetrachloroethylene-related toxic nephropathy). Hepatocellular adenomas and carcinomas in
10 mice and MCL in rats occurred in multiple lifetime rodent bioassays, and
11 hemangioendotheliomas in male mice (JISA, 1993) and cancers of the kidney and brain (glioma)
12 in male rats (NTP, 1986) occurred in single lifetime bioassays. Also known as
13 hemangiosarcomas, hemangioendotheliomas are rare tumors of the epithelial lining of blood
14 vessels. These tumors have been observed in a limited number of bioassays, including vinyl
15 chloride and 1,3-butadiene. Although the dose-response relationships for kidney and brain
16 tumors observed in male rats were not as strong as for the preceding cancers, and the increasing
17 dose-response trend for kidney tumors was not statistically significantly, both tumor types were
18 considered tetrachloroethylene-related and biologically relevant (see Section 5.4.2.3).
19 The statistically significantly elevated incidences of hepatocellular carcinomas and
20 adenomas in male and female mice and MCL in male and female rats are considered to be
21 indicators of potential human health hazard, despite questions regarding high background
22 incidences of these tumors in controls and MOA hypotheses (see Section 6.1.5.1). The finding
23 of an increased incidence of hepatocellular carcinomas and adenomas in female mice with a low
24 background incidence of these tumors suggests tetrachloroethylene is the etiological agent and
25 supports an inference of tetrachloroethylene as a risk factor for liver tumors in male mice that
26 have a higher background incidence. Moreover, kidney cancer and MCL in rats as indicators of
27 a potential human cancer hazard appear reasonable, given the observations in the epidemiologic
28 studies.
29 Although there are segments of the population who may be especially susceptible to the
30 toxic effects of tetrachloroethylene, there are too few studies specifically on tetrachloroethylene
31 to examine this hypothesis directly. A potential exists for infant exposures from several
32 pathways, including breast or other milk containing tetrachloroethylene. Infants younger than 6
33 months of age have slower renal clearance and less active liver metabolizing enzymes. The
34 nervous system in the developing fetus and in infants matures later than other systems. Elderly
35 persons and those with liver and kidney diseases also have slower clearance of toxic
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-6 DRAFT - DO NOT CITE OR QUOTE
-------
1 substances—especially lipophilic chemicals. Existing PBPK studies are not yet reliable for
2 quantitative use for estimating pharmacokinetic susceptibility in infants or the elderly (see
3 Section 4.9.1 for more details).
4
5 6.1.5. Mode-of-Action Information
6 Although a wealth of new data related to understanding the toxic effects caused by
7 tetrachloroethylene exposure has been published over the past decade, the MOA is not yet
8 sufficiently characterized, tested, or understood for any one of these adverse effects. A number
9 of alternative hypotheses are identified and examined as possible MO As for liver and kidney
10 toxicity. Hypothesized MO As for mononuclear cell leukemia, neurotoxicity, and
11 developmental/reproductive effects are indirect and are based on experimental observations of
12 exposures to agents other than tetrachloroethylene. The available evidence points to multiple
13 hypothesized MO As as being involved, and, in each case, no one MOA can be uniquely
14 identified (see Section 4.10.3 for more details). The sections following immediately below
15 summarize the MOA information available for liver, kidney, and other targets of
16 tetrachl oroethy 1 ene toxi city.
17
18 6.1.5.1. Liver Mode-of-A ction Information
19 The MOA for tetrachloroethylene-induced liver toxicity, including tumor induction, is
20 not known. Tetrachloroethylene-induced liver tumors in mice are believed to result from
21 chloroacetic acid metabolites and other intermediate products of the oxidative pathway, with
22 MOA hypotheses focused on the role of the major urinary metabolite TCA. Because both
23 tetrachloroethylene and TCA have been shown to activate the PPAR-a, as evidenced by
24 peroxisome proliferation, the ability of PPAR receptors to trigger a number of cellular events
25 suggests a possible relationship with tumor induction. However, metabolism to TCA does not
26 obviously explain tetrachloroethylene-induced liver tumors, suggesting that other metabolites or
27 intermediates contribute to tetrachloroethylene liver toxicity. Key steps in one MOA hypothesis,
28 namely that TCA alters cell signaling processes through activation of PPAR-a, have yet to be
29 fully identified both in mice and in humans.
30 Experimental evidence does not support peroxisome proliferation, per se, as a proposed
31 MOA. Specifically, peroxisome proliferation does not correlate well with tumor incidence.
32 Peroxisomes are seen at exposure concentrations higher than those that induce liver tumors, and
33 peroxisome proliferation is also seen in rat liver and mouse kidney, sites that do not show
34 carcinogenicity (see Section 4.10.3). The ability of PPAR receptors to trigger nonperoxisomal
35 events suggests that toxicity and tumor induction may not be causally related to peroxisome
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-7 DRAFT - DO NOT CITE OR QUOTE
-------
1 proliferation, but that tumorigenesis may be only a concurrent happening with many other
2 events. The relationship between these events and tetrachloroethylene tumor induction is not
3 understood. At the current time, sufficient evidence does not exist to suggest that
4 tetrachloroethylene or its oxidative metabolites could initiate hepatocarcinogenesis via a
5 mutagenic MOA.
6
7 6.1.5.2. Kidney Mode-of-Action Information
8 Several MO As for kidney toxicity are possible, although the supporting evidence is
9 limited. Induction of alpha-2|i-globulin occurs only at doses higher than the doses that induce
10 kidney cancer in male rat bioassays, and it is not likely to have an important role in toxicity or
11 tumor induction. Peroxisome proliferation has been weakly detected in rat kidneys—which do
12 show carcinogenicity—but peroxisome proliferation is more extensive in mouse kidneys, which,
13 too, has not demonstrated cancer. Scientific evidence is more supportive of the possibility that
14 reactive metabolites from the GSH conjugation pathway are in some way responsible for kidney
15 toxicity. These metabolites are associated with cytotoxicity and are mutagenic in Salmonella.
16
17 6.1.5.3. Mode-of-Action Information for Other Targets of Toxicity
18 The MOA of tetrachloroethylene-induced leukemogenesis in rats is not well understood;
19 specifically whether the parent compound, a metabolite, or several metabolites are involved.
20 Metabolites from GSH metabolism may contribute to toxicity, as supported by the finding of
21 aplastic anemia and DNA changes in lymphatic tissues in calves exposed to
22 S-(l,2-dichlorovinyl)-L-cysteine DCVC, which is structurally similar to the TCVC that is
23 produced through tetrachloroethylene GSH metabolism, although the study of TCVC in calves
24 was negative.
25 For neurotoxicity, the parent compound, rather than the metabolites, might be exerting an
26 anesthetic-like effect on the lipid membranes in the nervous system or interacting with several
27 neurotransmitter receptors. However, this hypothesis is not supported by specific studies on
28 tetrachloroethylene.
29 The MO As hypothesized for developmental toxicity differ according to effect. The
30 neurobehavioral effects during development may be mediated by the same MOA as the
31 neurotoxic effects discussed above. For fetal toxicity, TCA, an organic acid, lowers the pH of
32 the fetal compartment (see Section 4.7.4); this may be a contributing factor, given the finding of
33 developmental toxicity with TCA exposure. These proposed hypotheses, however reasonable,
34 lack experimental support.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-8 DRAFT - DO NOT CITE OR QUOTE
-------
1 The binding of reactive metabolites of tetrachloroethylene to proteins in liver, kidney and
2 serum, has the potential to contribute to the pathogenesis of several diseases, including cancer
3 and autoimmune disease.
4
5 6.1.5.4. Mode-of-Action Conclusions and Implications for Dose-Response Analyses
6 In summary, there is no obvious common MOA for the different toxicological effects of
7 tetrachloroethylene, nor has a sequence of key events been identified for any of the individual
8 adverse effects. MOA information does not indicate in any instance that toxicity endpoints in
9 animals are not relevant to humans, nor does it provide a basis for non-default procedures for
10 estimating risk or establishing reference values. Specifically, hypothesized rodent-only MO As
11 are not sufficiently established, and it is reasonable to use animal tumors as an indicator of a
12 potential human cancer hazard. Rodent tumors, leukemia, and cancer of the liver and kidney
13 have human analogues. For example, mononuclear cell leukemia in rats is also known as large
14 granular lymphocytic leukemia; large granular lymphocyte leukemia represents a well-
15 recognized group of lymphoid neoplasms in humans (Stromberg, 1985).
16 In the absence of a well characterized MOA that could explain dose-response
17 relationships at doses lower than those leading to observed effects, the cancer dose-response
18 modeling is carried out using a linear extrapolation performed in accordance with default
19 recommendations in the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).
20 The available data on noncancer toxicity of tetrachloroethylene support using EPA's RfC/RfD
21 methodologies to derive noncancer toxicity values. These approaches are detailed in Section 6.2.
22
23 6.1.6. Weight-of-Evidence Descriptor for Cancer Hazard
24 Tetrachloroethylene is "Likely to be carcinogenic to humans" by all routes of exposure,
25 within the framework of the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA,
26 2005a). As specified in the guidelines, the descriptor "Likely to be carcinogenic to humans"
27 expresses the conclusion regarding the weight of evidence for carcinogenic hazard potential and
28 is presented only in the context of a weight-of-evidence narrative. Although the term "likely"
29 can have a probabilistic connotation in other contexts, its use as a weight-of-evidence descriptor
30 does not correspond to a quantifiable probability of whether the chemical is carcinogenic. The
31 five recommended standard hazard descriptors are as follows:
32 • "Carcinogenic to humans"
33 • "Likely to be carcinogenic to humans"
34 • "Suggestive evidence of carcinogenic potential"
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 • "Inadequate information to assess carcinogenic potential"
2 • "Not likely to be carcinogenic to humans"
3
4 These descriptors are not unlike those used by the IARC, NTP, and other health agencies
5 that weigh carcinogenicity evidence. If there are no or insufficient pertinent data, then the
6 descriptors "Inadequate information to assess carcinogenic potential" or "Suggestive evidence of
7 carcinogenic potential" are used. If the evidence is stronger, as is the case with
8 tetrachloroethylene, the descriptor "Likely to be carcinogenic to humans" is used; convincing
9 evidence, usually conclusive demonstration of causality in epidemiologic studies, would support
10 "Carcinogenic to humans." On the other hand, if the conclusion is negative (i.e., strong,
11 consistent and compelling information indicating the absence of human health hazard), the agent
12 would be described as "Not likely to be carcinogenic to humans." Thus, going down the list of
13 descriptors from "Carcinogenic to humans" to "Inadequate information to assess carcinogenic
14 potential" indicates a decrease in the level of evidence of a human health hazard. In summary,
15 use of the weight-of-evidence descriptor "Likely to be carcinogenic to humans" for
16 tetrachloroethylene is intended to communicate that the available information indicates the
17 presence of a human health hazard.
18 The weight-of-evidence conclusion represented by the top three levels of evidence is
19 related to, but distinct from, the quantitative dose-response assessment/conclusions in that the
20 judgment that an agent is a human carcinogen does not guarantee adequate data to quantitatively
21 estimate human risk. Notably, evaluation of an agent that is judged a likely human carcinogen
22 may offer data conducive to estimating human risk. Indeed, dose-response assessments are
23 generally completed for agents considered "Carcinogenic to humans" and "Likely to be
24 carcinogenic to humans." Section 6.2 provides the dose-response analyses for
25 tetrachloroethylene.
26 Three lines of evidence in the hazard database support the weight-of-evidence descriptor
27 for the cancer hazard for tetrachloroethylene: (1) tetrachloroethylene exposure is associated with
28 excess risks for a number of cancers in human epidemiologic studies, although a causal
29 association has yet to be sufficiently established; (2) tetrachloroethylene is a rodent carcinogen in
30 10 of 10 lifetime bioassay data sets, including by oral and inhalation routes; and (3) the available
31 information indicates that the cancer bioassay data are relevant to use as indicators of potential
32 human cancer hazard. Briefly, the epidemiologic evidence has associated tetrachloroethylene
33 exposure with excess risks for a number of cancers including cancer of the lymphoid system,
34 esophagus, and cervix, with more limited evidence for cancer of the bladder, kidney, and lung.
35 For both lymphoid and esophageal cancer, excess risk was observed in studies of people who
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 work as dry cleaners or degreasers, populations who experience inhalation exposure to
2 tetrachloroethylene and other solvents. In both cases, average risks were doubled as compared
3 with those of referents. Furthermore, studies of drinking water exposure also support an
4 association between lymphoid cancer and tetrachloroethylene and other solvents, as do case-
5 control studies that assessed employment as a dry cleaner or laundry worker. Chance and
6 confounding by smoking are unlikely explanations for the observed excesses in risks.
7 As summarized in Section 6.1.4, the laboratory animal database includes 10 lifetime
8 rodent bioassay data sets demonstrating increased cancer incidence. The findings include liver
9 cancer in both sexes of mice and mononuclear cell leukemia in both sexes of rats following
10 either oral or inhalation exposures and, in single bioassays, male rat kidney and brain tumors
11 (gliomas) and mouse hemangiosarcomas of the liver or spleen. In addition, although not all
12 tetrachloroethylene metabolites have been tested for carcinogen!city in rodents, the oxidative
13 metabolites TCA and DC A are hepatocarcinogens in one or more species. Taken together, these
14 data support a weight-of-evidence descriptor of "Likely to be carcinogenic to humans" by all
15 routes of exposure.
16
17 6.2. DOSE-RESPONSE CHARACTERIZATION
18 Quantitative estimates of risk to humans are derived separately for noncancer and cancer
19 effects. RfD and RfC values are derived from epidemiologic studies of residential populations
20 exposed to tetrachloroethylene from nearby dry cleaning facilities. Residents in these studies
21 have shown an impaired ability to detect and respond to visual stimuli compared to responses of
22 controls (see Section 5.1.1). Inhalation cancer risk has been estimated from animal data on
23 malignant tumors induced in tests involving lifetime exposure to tetrachloroethylene at known
24 concentrations.
25
26 6.2.1. Noncancer Toxicity (Reference Concentration [RfC]/Reference Dose [RfD])
27 A broad range of animal toxicology and human epidemiologic data are available for the
28 hazard assessment of tetrachloroethylene. The nervous system appears to be a sensitive organ
29 system, particularly in human studies (see Section 4.6.1). Nevertheless, critical data gaps have
30 been identified and uncertainties associated with data deficiencies are more fully discussed in
31 Chapter 5 and in the remainder of this section. Even with these uncertainties, the database of
32 human and animal studies on inhalation and oral toxicity of tetrachloroethylene can support
33 derivation of inhalation and oral reference values. A number of epidemiologic studies of
34 neurological effects in either occupational workers or residential subjects with
35 tetrachloroethylene exposure or toxicological studies in rodents are considered for developing an
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 RfC and RfD. No single epidemiologic study is considered to be without flaws and
2 uncertainties, although these are different among studies and studied populations. Among the
3 epidemiologic studies, seven studies were considered for supporting an inhalation reference
4 value, and a study of neurobehavioral deficits in people residing near dry cleaning facilities
5 (Altmann et al., 1995) was identified as the principal study using a weight-of-evidence approach
6 (see Sections 5.1. and 5.2. for more details). The small number of subjects (14 exposed of 37
7 subjects studied) can introduce uncertainties particularly regarding stability of statistical
8 inferences. However, statistically significant group differences between the adjusted mean
9 scores of exposed and control subjects on three neurobehavioral tests (simple reaction time,
10 p< 0.05 for the first test and/? < 0.01 for the second test; continuous performance,/' < 0.05; and
11 visual memory, p < 0.05) were observed after adjusting for covariates and possible confounders
12 of age, gender, and education. In all cases, the exposed subjects had slower response times or
13 more errors than did the unexposed controls. Other factors were also considered in the overall
14 weight-of-evidence analysis. Table 6-1 summarizes the rationale for selection of the principal
15 study (the rationale is also addressed in Sections 5.1 and 5.2).
16
17 6.2.1.1. Assessment Approach Employed
18 Noncancer toxicity RfC and RfD are developed using EPA's RfC and RfD
19 methodologies (U.S. EPA, 1993, 1994). The RfC for tetrachloroethylene is derived through a
20 process of (1) considering all studies and selecting the critical effects that occur at the lowest
21 exposure concentration, (2) selecting the point of departure (POD) at which critical effects either
22 are not observed or would occur at a relatively low prevalence (e.g., 10%), (3) deriving the POD
23 in terms of the HEC, and (4) reducing this exposure concentration by UFs to account for
24 uncertainties in the extrapolation from the study conditions to an estimate of human
25 environmental exposure.
26 The RfC is developed from the point of departure (POD) of 4.8 mg/m3 (0.7 ppm),
27 which was associated with impaired cognitive function and visual information processing in a
28 study of people residing near dry cleaning facilities (Altmann et al., 1995). The assumption that
29 the residents were continuously exposed to tetrachloroethylene eliminated the need for a duration
30 adjustment to the POD. There is sufficient evidence from occupational studies of higher
31 tetrachloroethylene concentrations to confirm that the nervous system is the primary target for
32 the effects of tetrachloroethylene, with several studies showing a similar pattern of effects in the
33 residential study (Seeber, 1989; Ferroni et al., 1992; Cavalleri et al., 1994, Echeverria et al.,
34 1994, 1995). The median concentration in Altmann et al. (1995) is similar to the concentration
35 in a pilot residential study reporting deficits in visual contrast sensitivity (Schreiber et al., 2002),
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-12 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
Table 6-1. Summary of rationale for principal study selection
Consideration/
approach
Type of data
Decision
Quality of study
Animal neurotoxicity
studies.
Animal neurotoxicity studies are considered as supporting
studies. An RfC/RfD from human data, if available and of
adequate quality, is from the species of interest to EPA,
reduces interspecies extrapolation uncertainties and is
preferred.
Quality of study
Human neurotoxicity
studies.
Both occupational and residential studies on
tetrachloroethylene exposure contain uncertainties regarding
their use for quantitative analysis. Some of these
epidemiology studies carry greater weight for quantitative
analysis than the less informative studies of Ferroni et al.
(1992), Echeverria et al. (1994, 1995), and Spinatonda et al.
(1997) with reporting or exposure assessment deficiencies.
Measurement
tool
Standardized
neurobehavioral
battery.
Both occupational and residential epidemiology studies
assessed neurobehavioral function using a standardized
neurobehavioral battery. The battery has been widely
administered to occupational populations in different
settings with a reasonably high degree of validity. WHO
and ATSDR recommend these test methods to evaluate
nervous system deficits in adults and children.
Endpoint
Deficits in
neurological domains
such as attention,
motor function,
vigilance, or visuo-
spatial function.
There is congruence of neurological effects observed in
studies of both residential and occupational populations.
These domains are also sensitive to acute tetrachloroethylene
exposure in controlled human studies. The consistency of
observed effects between occupational and residential
populations and their persistence with lower
tetrachloroethylene concentration, as experienced by
residential populations, provide a strong rationale for a study
of lower-level residential exposures as the basis for the RfC.
Relevance of
Exposure
Scenario
Epidemiology studies
of residential
populations.
Tetrachloroethylene exposure to residential populations is of
lower concentration and of chronic duration compared to
acute duration and higher concentration exposure to
occupational populations. Additionally, potential to
tetrachloroethylene peak or intensity concentrations is more
common with occupational exposures. A study of
residential exposure, if adequate and of similar quality as an
occupational epidemiology study, is preferred for supporting
the RfC because it represents exposure scenarios of interest
to EPA.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 and it is lower than the concentration associated with markers of kidney and liver damage and
2 reproductive/developmental effects (spontaneous abortion and changes in sperm quality) in
3 several epidemiologic studies of occupational exposure (Franchini et al., 1983; Gennari et al.,
4 1984; Olsen et al., 1990; Eskenazi et al., 1991a, b; Mutti et al., 1992; Brodkin et al., 1995; Doyle
5 et al., 1997; Verplanke et al., 1999; Trevisan et al, 2000). Effect levels in animal studies are,
6 generally, similar to those in the occupational epidemiologic studies, with BMC modeling of
7 liver toxicity showing effect concentrations only slightly higher than the POD from residents
8 (Kjellstrand et al., 1984; USA, 1993).
9 A composite UF of 300 is adopted to account for human variation (factor of 10),
10 extrapolation from an observed-effect level to a no-effect level (factor of 10), and uncertainties
11 in the database (factor of 3). Limited data on tetrachloroethylene blood concentration among
12 human subjects indicated that a choice of 3 for the pharmacokinetic portion of the 10-fold human
13 variation UF is reasonable. The rationale for a 3-fold database UF is based on critical data gaps
14 and takes into account a lack of animal studies designed to clearly investigate the human findings
15 in cognition and visual system dysfunction and a lack of cognitive testing in both
16 developmentally exposed animals and adult animals exposed to tetrachloroethylene for longer
17 than acute durations (see Sections 5.1, 5.2, and 5.3 for further discussion of these issues). These
18 data are needed to allow for a fuller characterization of the exposure-response relationship. The
19 RfC was calculated by dividing the POD by the composite UF = 4.8 mg/m3/300 = 1.6 x 10'2
20 mg/m3.
21 The database for oral exposure to tetrachloroethylene is limited to four subchronic gavage
22 studies, one subchronic drinking water study, and no human studies. In addition to using the
23 animal data on oral exposure, the assessment attempted to expand the database for derivation of
24 an RfD using relevant inhalation data and route-to-route extrapolation with the aid of a
25 pharmacokinetic model. Route extrapolation of human inhalation data is considered a
26 reasonable alternative to using the limited oral data in animals because tetrachloroethylene has
27 been shown to be rapidly and well absorbed by the oral and inhalation routes of exposure, and
28 the metabolic pathways and kinetics of excretion with oral exposure are similar to those of
29 inhalation exposure. Furthermore, human data, when adequate, are preferred for supporting the
30 RfD, and human data of inhalation exposure are available.
31 The residential inhalation study of Altmann et al. (1995) of neurobehavioral deficits and
32 three acute and subchronic toxicological studies were examined for supporting an RfD. The RfD
33 was derived from Altmann et al. (1995) with the aid of an extrapolation from the inhalation to
34 the oral route using pharmacokinetic modeling. The daily oral ingestion dose that results in the
35 same tetrachloroethylene blood concentration associated with the POD for inhalation, 4.8 mg/m3,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-14 DRAFT - DO NOT CITE OR QUOTE
-------
1 is 1.1 mg/kg-day. This value is equivalent to an oral LOAEL. Blood tetrachloroethylene
2 concentration is a well-validated dose metric, and the estimate varies little between models. The
3 human LOAEL falls within the PODs from oral studies in animals. The UFs, as used for the
4 inhalation RfC, are adopted for the RfD for oral exposure, namely, a composite factor of 300 (10
5 for human variation, 10 for extrapolation from a LOAEL to a NOAEL, and 3 for database
6 uncertainties; see also Sections 5.1, 5.2, and 5.3). The oral RfD is, therefore, 1.1 mg/kg-day/300
7 =4 xlQ'3 mg/kg-day.
8 To show the range of tetrachloroethylene concentrations at which different neurotoxic
9 effects and toxic effects in other organ systems have been observed, the points of departure and
10 reference values that could have been derived from these other studies were compared with that
11 of the principal study. These graphs allow a direct visualization of how the values compare to
12 the data from which the principal conclusions have been derived. This has been done for both
13 inhalation reference concentrations and oral reference doses in Figures 6-1, 6-2, and 6-3.
14
15 6.2.1.2. Impact of Assumptions, Uncertainties and Alternatives on Reference Concentration
16 and Reference Dose
17 A number of uncertainties underlie the RfC and RfD for tetrachloroethylene and are
18 discussed below in this section. A quantitative characterization of the uncertainty in the RfC and
19 RfD for tetrachloroethylene is not feasible because of the varied nature of the available database
20 and the limited data available for many of the studies. Most significantly, the available chronic
21 toxicity studies of tetrachloroethylene exposure demonstrated varying degrees of support for a
22 POD for the RfC and RfD. A weight-of-evidence approach was adopted to identify principal or
23 critical studies, with the additional studies supporting the principal studies (see also Sections 5.1
24 and 5.2).
25
26 6.2.1.2.1. Point of departure. Most of the available studies did not provide enough data to
27 support benchmark dose modeling; they only supported PODs based on LOAELs, especially for
28 the human studies, or LOAELs and NOAELs. Such a POD has a number of shortcomings
29 relative to a POD obtained from benchmark dose-response modeling (i.e., a benchmark
30 concentration). First, LOAELs and NOAELs are a reflection of the particular exposure levels at
31 which a study was conducted, contributing some inaccuracy to the POD determination. Second,
32 LOAELs and NOAELs reflect the number of study subjects or test animals and typically are
33 dissimilar in detection ability and statistical power, with smaller studies tending to identify
34 higher exposure levels as PODs relative to larger but otherwise similarly designed studies. This
35 is an important consideration for studies with multiple exposure groups and studies that did not
36 identify LOAELs but has much less impact for the single-group studies that identified a LOAEL.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-15 DRAFT - DO NOT CITE OR QUOTE
-------
o
Oi
Inhalation Neurotoxicity RfVs
O
oo
I
Co
H £
Human Human Human Human
Human
Rat
Gerbil
Q.
Q.
o
5
•s
o
o
0.001
0.0001
0.00001
0.01
--100
-- 10
--1
- 0.1
-- 0.01
o
c
o
o
- 0.001
Visual
contrast
(day care
employees)
(Schreiber
etal.,2002)
Vision, cognitive Vision, 5% Visual function, Visual
function, change in ~10% change function
Visual
contrast
(apartment vigilance,3 ~15%
residents) decrement
(Schreiber (Altmann
etal.,2002) etal.,1995)
color
confusion
index
(Cavalleri
etal., 1994)
in neuro-
behavioral
tests
(Seeber, 1989)
(Mattsson
etal., 1998)
Brain
chemistry
(Rosengren
etal., 1986)
Principal study.
KEY
4 Point of departure
(POD)
I I UF, extrapolation from a
lowest-observed-
adverse-effect level
(LOAEL) to a no-
observed-adverse-effect
level (NOAEL)
I I UF, animal-to-human
extrapolation
UF, human variation
UF, extrapolation from
subchronic exposure
duration to chronic
exposure duration
• UF, database
deficiencies
O Reference value (RfV)
Figure 6-1. Array of PODs and reference values for a subset of neurotoxic effects in inhalation studies.
-------
H £
Inhalation Organ-Specific RfVs
100
10
Human
Human
Rat
Mouse
o
•a
ra
is
c
8
O
o
0.1
0.01
0.001
0.0001
n i
II 1
^ 1 UUUUUU1
I I 1
• 111.
1
CMS, ~'\5% Kidney, increase in Developmental/ Liver, 10%
decrement in verbal mean urinary Reproductive, 1% increase in liver
function, cognitive concentration of increase in fetal angiectasis,
function, vigilance, several renal death, BMCLoi BMCUo/p as POD
and vision3 enzymes as POD (J ISA, 1993)
(Altmann et al., (Multti (Tinston
1995) etal.,1992) et al., 1994)
- 100
- 10 j~*
E
-1 I
p
:oncentra1
-0.01
KEY
A Point of departure
(POD)
I I UF, extrapolation
from a lowest-
observed-adverse-
effect level (LOAEL)
to a no-observed-
adverse-effect level
| 1 (NOAEL)
UF, animal-to-human
cmi extrapolation
^_ UF, human variation
^^ UF, database
deficiencies
Reference value
(RfV)
Principal study.
TO
Figure 6-2. Organ-specific RfVs for inhalation exposure to tetrachloroethylene.
-------
o
Oi
O
oo
I
Co
00
I
Co
§
H I
O >
HH Oq
H TO
si
100
Oral Organ-Specific RfVs
Human Mouse Mouse Rat
01
1/1
0.01
0.001
0.0001
CMS, -15% decrement
in verbal function,
cognitive function,
vigilance, and vision8
(Altmann et al., 1995)
Neuro-
developm ental
(Fredriksson
et al., 1993)
Liver, one S.D.
change in the m ean
liver weight over the
control response
(Buben and
O'Flaherty, 1985)
System ic
toxicity
(Hayes et al.,
1986)
aPrincipal study.
KEY
4 Point of departure
(POD)
\ \ UF, extrapolation from a
lowest-observed-
adverse-effect level
(LOAEL) to a no-
observed-adverse-effect
level (NOAEL)
^^ UF, animal-to-human
extrapolation
""" UF, human variation
I I UF, extrapolation from
subchronic exposure
duration to chronic
exposure duration
^H UF, database
deficiencies
O Reference value (RfV)
Figure 6-3 Oral organ-specific reference values for exposure to tetrachloroethylene.
H
W
-------
1 Third, LOAELs and NOAELs represent different response rates, as noted on Figures 6-1, 6-2,
2 and 6-3, and qualitative and quantitative comparisons are not possible lacking characterization of
3 the underlying dose-response curve.
4 PODs identified from fitting benchmark dose models overcome some of the deficiencies
5 associated with LOAELs and NOAELs. Benchmark dose models were fit to five data sets
6 (Buben and O'Flaherty, 1985; JISA, 1993; NTP, 1986; Brodkin et al., 1995; Tinston, 1994) with
7 sufficient information. The choice of benchmark dose model did not generally lead to significant
8 uncertainty in estimating the POD since benchmark effect levels were within the range of
9 experimental data. While this examination of a subset of chronic toxicity studies on
10 tetrachloroethylene exposure provides some insight on study and endpoint differences in PODs,
11 lacking characterization of dose-response curves for all studies, especially the more critical
12 studies, uncertainty associated with the PODs cannot be adequately quantified in this database.
13 Effects in the CNS and in other organ systems in occupational populations and in animals
14 are observed at higher average tetrachloroethylene concentrations than the Altmann et al. (1995)
15 residential study. Uncertainties in other studies of neurotoxicity and of other organ systems
16 differ from those of Altmann et al. (1995). For both occupational and residential populations,
17 studies do not describe a NOAEL and human variation is not well characterized in study
18 subjects. Uncertainties associated with the occupational studies include (1) potential for
19 neurobehavioral effects at lower exposures and (2) exposure pattern differences between
20 occupational and residential studies with peaks characterizing occupational exposures. Using an
21 occupational study to support the RfC may not be fully protective of neurological effects as has
22 been observed in populations co-located near dry cleaners (Altmann et al., 1995; Schreiber et al.,
23 2002; and NYSDOH, 2005a, c). For animal studies, uncertainties are associated with
24 extrapolating high concentration exposure, typically of subchronic duration to genetically inbred
25 rodents, to infer a concentration of tetrachloroethylene that is likely to be without an appreciable
26 risk of adverse health effects over a lifetime to a diverse human population.
27
28 6.2.1.2.2. Extrapolation from laboratory animal studies to humans. Extrapolating from
29 animals to humans embodies further issues and uncertainties. First, the effect and its magnitude
30 associated with the concentration at the point of departure in rodents is extrapolated to human
31 response. Pharmacokinetic models are useful to examine species differences in pharmacokinetic
32 processing. This was possible for liver toxicity where limited MOA information suggests
33 metabolism as important to toxicity. The ranges of BMCLs presented for liver effects (a 10-fold
34 range of estimates of tetrachloroethylene metabolism) demonstrate the uncertainty in
35 tetrachloroethylene pharmacokinetic models. The discrepancies among the models and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-19 DRAFT - DO NOT CITE OR QUOTE
-------
1 experimental data may point to large uncertainties in the parameters used in these models.
2 Because the accuracy of the models has been evaluated only against blood and breath
3 concentrations of the parent compound, their reliability for predicting total metabolites is
4 unknown.
5
6 6.2.1.2.3. Human variation. Heterogeneity among humans is another uncertainty associated
7 with extrapolating doses from animals to humans. Uncertainty related to human variation needs
8 consideration, also, in extrapolating dose from a subset or smaller sized population, say of one
9 sex or a narrow range of life stages typical of occupational epidemiologic studies, to a larger,
10 more diverse population.
11 In the absence of tetrachloroethylene-specific data on human variation, a factor of 10 was
12 used to account for uncertainty associated with human variation. Human variation may be larger
13 or smaller; however, tetrachloroethylene-specific data for examining the potential magnitude of
14 over- or under-estimation are few. The pharmacokinetic model of Clewell et al. (2004) of mean
15 physiological parameters used to explore age-dependent pharmacokinetic differences suggests a
16 2-fold variation in blood tetrachloroethylene levels (Chapters 3 and 5). Bois et al. (1996) and
17 Chiu and Bois (2006) have examined uncertainty and variation in a tetrachloroethylene
18 pharmacokinetic model describing the amount of tetrachloroethylene metabolism. This analysis
19 suggests large uncertainty is associated with estimating the quantity of tetrachloroethylene
20 metabolism in humans.
21
22 6.2.1.2.4. Database uncertainties. Critical data gaps have been identified with uncertainties
23 associated with database deficiencies on developmental, immunologic, and neurotoxic effects,
24 particularly data to characterize dose-response relationships and chronic visuo-spatial functional
25 deficits and cognitive effects of tetrachloroethylene exposure under controlled laboratory
26 conditions. Several halogenated organic solvents have been linked with altered immune system
27 function in both animals and humans (e.g., toluene, trichloroethylene). Additional data from
28 inhalation, oral, and dermal exposures at different durations are needed to assess the potential
29 immunotoxicity of tetrachloroethylene. This lack of data, combined with the concern that other
30 structurally related solvents, has been associated with immunotoxicity and contributes to
31 uncertainty in the database for tetrachloroethylene.
32 Data from acute studies in animals (Warren et al., 1996; Umezu et al., 1997) suggest that
33 cognitive function is affected by exposure to tetrachloroethylene. These studies do not address
34 the exposure-response relationship for subchronic and chronic tetrachloroethylene exposures on
35 cognitive functional deficits observed in humans (e.g., Seeber, 1989; Echeverria et al., 1994; and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 Altmann et al., 1995). Even more importantly, cognitive testing is lacking in both
2 developmentally exposed animals and adult animals following exposures to tetrachloroethylene
3 that are longer than acute durations of exposure. Visual system dysfunction and processing of
4 visuo-spatial information are sensitive endpoints in human studies. The exposure-response
5 relationship of these functional deficits could be evaluated more definitively with studies using
6 homologous methods that examine retinal and visual function in experimental animals.
7 However, there has been a limited evaluation of visual function in rodents, with the exception of
8 the evoked potential studies by Mattsson et al. (1998). These types of studies could help
9 determine whether there are both peripheral and central nervous system effects of
10 tetrachloroethylene exposure on visual perception, and they could be used as an animal model to
11 better define the exposure-response relationships.
12 Subjects in the epidemiologic studies comprise adults, and some characterization of the
13 response of children to tetrachloroethylene exposure was found in limited data for a similar
14 neurological (visual system) parameter (Schreiber et al., 2002) and in a larger number of subjects
15 (NYS DOH, 2005 a,c) using other visually based testing paradigms. Additionally, in a postnatal
16 neurotoxicity study in mice (Fredriksson et al., 1993), persistent neurological effects (i.e.,
17 increased locomotion and decreased rearing behavior at 60 days of age, measured 43 days after
18 exposure ceased) were observed at an oral dose of 5 ppm, with no NOAEL, although this study
19 did not conform to traditional toxicity testing guidelines (see Section 4.6.2.2). These results
20 suggest that if adequate, robust, dose-response data using the most appropriate
21 neurophysiological and cognitive tests were available, the exposure eliciting an adverse response
22 (and hence the POD for the reference value) could be lower than that established based on
23 deficits in visuo-spatial and cognitive function following tetrachloroethylene exposure in healthy
24 adults (Altmann et al., 1995).
25
26 6.2.2. Cancer Risk Estimates
27 Following the scientific principles and procedures outlined in EPA's Guidelines for
28 Carcinogen Risk Assessment (U.S. EPA, 2005a), the cancer risk values are based on the 95%
29 lower confidence limits on the air concentrations associated with a 10% extra risk of cancer
30 incidence (LECios). The LECio values were calculated from data on MCL in male rats, the most
31 sensitive species/gender in the rodent cancer bioassay conducted at the lowest concentration
32 range, using the multistage dose-response model. A linear low-dose extrapolation was then used,
33 in accordance with default recommendations in EPA's Guidelines for Carcinogen Risk
34 Assessment (U.S. EPA, 2005a). The approach and associated choices and assumptions are
35 described in Sections 6.2.2.1.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-21 DRAFT - DO NOT CITE OR QUOTE
-------
1 A broad range of animal toxicology and human epidemiologic data are available for the
2 identification of a carcinogenic hazard from exposure to tetrachloroethylene. Nevertheless,
3 critical data gaps have been identified, and uncertainties associated with data deficiencies are
4 more fully discussed in Chapter 5 and Section 6.2.2.2. Given the choices of tumor type, point of
5 departure, and low-dose extrapolation approach necessary to provide a plausible upper bound
6 risk estimate, there are additional considerations that contribute to uncertainty in the cancer risk
7 values. These uncertainties have a varied impact on risk estimates. Some (i.e., the bioassay or
8 cross-species scaling approach) suggest risks could be higher than estimated while others would
9 decrease estimates or have an impact of uncertain direction (i.e., the human population
10 variability, dose metric, and model-based uncertainty at the POD). While some uncertainties
11 could be quantitatively characterized, it is likely that the residual uncertainties remain the largest,
12 yet can only be qualitatively expressed: i.e., low dose extrapolation, MOA, and human
13 sensitivity and variability. Even if experimental data could further elucidate these uncertainties,
14 extrapolation of animal bioassay data to human (done here using allometric scaling) will remain
15 a substantial and unknown uncertainty. The tetrachloroethylene unit risk estimate, calculated
16 using three PBPK models, ranges from 2 x 10"6 to 2 x 10"5 per |ig/m3. From this range, the upper
17 end unit risk of 2 x 10"5 per |ig/m3 is the most public health protective value for the upper bound
18 risk estimate.
19 The tetrachloroethylene oral slope factor, using the three PBPK models for route-to-route
20 extrapolation from the experimental data to humans, ranges from 1 x 10"2 to 1 x 10"1 per mg/kg-
21 day. From this range, the upper end slope factor of 1 x 10"1 per mg/kg-day is the most public
22 health protective value for the upper bound risk estimate. With the exception of the route-to-
23 route extrapolation step, the uncertainties associated with the slope factor estimation are the same
24 as for the unit risk estimation.
25 Section 6.2.2.2 describes the uncertainties outlined above, their impact on cancer risk
26 estimation, the choices made and justification for each, and the associated data gaps. Section
27 6.2.2.3 provides a quantitative analysis of the potential numeric impact of three of these sources
28 of uncertainty on the unit risk estimate (the statistical uncertainty, PBPK model, and tumor site)
29 using the multistage model in the observed range and linear low-dose extrapolation. Section
30 6.2.2.4 and the table therein provide a summary of the cancer risk estimate.
31
32 6.2.2.1. Assessment Approach Employed
33 Animal bioassay data are used to derive quantitative cancer risk estimates for humans due
34 to the lack of quantitative exposure information in the occupational epidemiology studies. The
35 cancer dose-response analysis considers three bioassays but relies on the JISA (1993) study
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 results. This is primarily because the JISA (1993) study included lower exposures than did the
2 two earlier bioassays for both species tested, and, therefore, its use reduces extrapolation
3 uncertainty slightly. For mice, the lowest exposure concentration of 10 ppm was 10-fold lower
4 than the lowest exposure concentration in the NTP inhalation study (NTP, 1986). For rats, the
5 low-exposure concentration of 50 ppm was fourfold lower than in the NTP (1986) study.
6 Dose-response analyses for hepatocellular tumors in male and female mice,
7 hemangiosarcomas in male mice, and mononuclear cell leukemia and kidney tumors in male and
8 female rats are carried out using the rate of total metabolite production as estimated from the
9 three recently developed toxicokinetic models as the dose metric. Dose-response analyses are
10 also carried out using administered dose as the dose metric to allow comparison to the
11 pharmacokinetic model-based risk estimates. EPA's methodology for cross-species scaling was
12 applied for relevant tumor sites to address toxicological equivalence across species (U.S. EPA,
13 1992). This methodology is based on the observation that equal average lifetime concentrations
14 or AUC of the toxic moiety has been associated with toxicological equivalence across species.
15 This cross-species relationship has been shown to accommodate the general species variation in
16 pharmacokinetics and the carcinogenic response to internal doses. Although the available
17 pharmacokinetic data for tetrachloroethylene do not allow estimates of AUC, the use of
18 metabolized tetrachloroethylene scaled to mg/kg3/4-day in order to estimate the dose resulting in
19 the same lifetime risk in animals and humans is consistent with the EPA methodology and
20 further substantiated in the present document (see Section 5.4.4.2.1). This consideration of
21 cross-species scaling and toxicological equivalence is consistent with EPA's other carcinogen
22 assessments and its treatment of pharmacokinetic dose metrics.
23 The steps involved in generating the unit risk from the dose-response data are illustrated
24 using the male rat MCL data, as follows:
25 (1) A fit of the tumor incidence versus total metabolite curve using a multistage model
26 (BMDS, version 1.3.2) gave an LECio, or 95% lower confidence bound on the exposure
27 associated with 10% extra risk, of 0.81 per mg-eq/kg-day (Figure 5-8a);
28 (2) The point of departure (LECio) was then transformed to a human equivalent value by
29 dividing the animal value by (human body weight /animal body weight)0'25 = (70/0.45)°'25
30 = 3.53 to give a human equivalent value of 0.23 mg-eq/kg-day of metabolite formation;
31 (3) Three different models (see Section 3.5) of total human metabolite formation from
32 tetrachloroethylene exposure were used to estimate the environmental exposure that
33 would correspond to the human equivalent LECio (Bois et al., 1996; Rao and Brown,
34 1993; Reitz et al., 1996). The lowest human equivalent LECio resulting from the three
35 models (Rao and Brown) is 47,000 |ig/m3. The highest human equivalent LECio
36 resulting from the models is 4,700 |ig/m3 (Bois et al.);
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 (4) The unit risk calculated using three PBPK models ranges from 2 x 10"6 to 2 x 10"5 per
2 Hg/m3. From this range, the upper-end unit risk of 2 x 10"5 per |ig/m3 is the most public
3 health protective value for the upper bound risk estimate.
4
5 Age adjustment factors for early life exposures as discussed in the Supplemental
6 Guidance for Assessing Susceptibility for Early-Life Exposure to Carcinogens (U. S. EPA,
7 2005b) are not recommended because little evidence exists to indicate that tetrachloroethylene or
8 its oxidative metabolites directly damage DNA, information about genotoxicity of GSH
9 metabolites in cell assays other than Salmonella or in in vitro experiments are lacking, and the
10 MO A for tetrachloroethylene has not been established.
11
12 6.2.2.2. Impact of Assumptions, Uncertainties and Alternatives on Unit Risk Estimates
13 A number of uncertainties underlie the cancer unit risk for tetrachloroethylene. These are
14 discussed in the following paragraphs. Specifically addressed is the impact on the assessment of
15 issues such as the use of models and extrapolation approaches, the reasonable alternatives, the
16 choices made, and the data gaps identified. In addition, the use of assumptions, particularly
17 those underlying the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), is explained
18 and the decision concerning the preferred approach is given and justified. Several of the
19 uncertainties with the largest impact cannot be considered quantitatively, such as human
20 population variability and the most relevant dose metric. Thus, an overall integrated quantitative
21 uncertainty analysis is not presented.
22
23 6.2.2.2.1. Human population variability. The extent of inter-individual variability in
24 tetrachloroethylene metabolism has not been characterized. As noted in Section 6.1.2, several
25 enzymes of the oxidative and GSH metabolism, notably CYP2E1, CYP3A4, GSTZ, GSTA,
26 GSTM, and GSTT show genetic polymorphisms with the potential for variation in metabolite
27 production. The limited data available on tetrachloroethylene metabolites show DCA to be a
28 potent, irreversible inhibitor of GSTZ activity, with greater inhibition of this enzyme in mice
29 than in humans. Tetrachloroethylene metabolism has been shown to increase by inducers of
30 CYP enzymes such as toluene, phenobarbital, and pregnenolone-16 alpha-carbonitrile, whereas
31 CYP enzyme inhibitors such as SKF 525A, metyrapone, and carbon monoxide have been shown
32 to decrease tetrachloroethylene metabolism. Additionally, chronic exposure to
33 tetrachloroethylene has been shown to cause self-induction of metabolism. Human population
34 variability is summarized above (see Section 6.2.1.2.3) and in covered in more detail in
35 Chapters.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 A separate issue is that the human variability in response to tetrachloroethylene is also
2 poorly understood. The effect of metabolic variation, including potential implications for
3 differential toxicity, has not been well studied. Although a mutagenic MOA would indicate
4 increased early-life susceptibility, there are no data exploring whether there is differential
5 sensitivity to tetrachloroethylene carcinogenicity across life stages. Thus, this lack of
6 understanding about potential differences in metabolism and susceptibility across exposed
7 human populations represents a source of uncertainty. Nevertheless, the existing data support
8 the possibility of a heterogeneous response that may function additively to ongoing or
9 background exposures, diseases, and biological processes. As noted in Section 4.9.5., some
10 evidence shows certain subpopulations may be more susceptible to tetrachloroethylene exposure.
11 As discussed under (2) below, these considerations strengthen the scientific support for the
12 choice of a linear non-threshold extrapolation approach. In summary, the human equivalent risk
13 estimates for tetrachloroethylene, therefore, do not reflect this source of uncertainty.
14
15 6.2.2.2.2. Choice of low-dose extrapolation approach. A key consideration in clarifying how
16 risks should be estimated for low-dose exposure is the MOA. As noted above in Section 6.1.5,
17 MOA data are lacking or limited for all of the candidate cancer endpoints for tetrachloroethylene
18 (i.e., rat MCL and kidney tumors, mouse hepatocellular tumors and hemangiosarcomas). When
19 the MOA cannot be clearly defined, EPA uses a linear approach to estimate low-dose exposure
20 risk, based on the following broad and long-held scientific assumptions, which supported
21 development of EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a):
22 • A chemical's carcinogenic effects may act additively to ongoing biological processes,
23 given that diverse human populations are already exposed to other agents and have
24 substantial background incidence of various tumors. Under these conditions, a nonzero
25 slope of the response as a function of chemical exposure is expected.
26 • A broadening of the dose-response curve in the human population (less rapid fall-off with
27 dose) and, accordingly, a greater potential for risks from low-dose exposures (see Zeise et
28 al., 1987; Lutz et al., 2005) would result for two reasons. First, even if there is a
29 threshold concentration at the cellular level, that threshold is likely to be different among
30 different individuals. Second, greater variability is anticipated in response to exposures
31 in the heterogeneous human population than in controlled laboratory species and
32 conditions (due to, e.g., genetic variability, disease states, age).
33 • The general use of linear extrapolation provides plausible upper-bound risk estimates that
34 are believed to be health-protective (U.S. EPA, 2005a) and also provides consistency
35 across assessments.
36
37 The extent to which the overall uncertainty in low-dose risk estimation could be reduced
38 if the MOA for tetrachloroethylene were known with a high degree of confidence is of interest,
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-25 DRAFT - DO NOT CITE OR QUOTE
-------
1 but clear data on the MOA of tetrachloroethylene is not available, and even if it were,
2 incorporation of MOA into dose-response modeling might not be straightforward and might not
3 significantly reduce the uncertainty about low-dose extrapolation. This is because the MOA, as
4 well as other factors, especially human response variability, are determinants of the dose-
5 response function in humans.
6 This chemical assessment also evaluates the extent to which a collection of mathematical
7 functions fit to one of the tetrachloroethylene bioassay data sets and extrapolated down to low
8 doses, could inform uncertainty. There is not sufficient information regarding the MOA to
9 support a chemical-specific inference about dose-response behavior at low dose for
10 tetrachloroethylene. Thus, it is of interest to observe how different functions fit to the tumor data
11 may diverge when extrapolated downward. Much previous experience has supported a general
12 mathematical property that different curves, though fitting observed experimental data well,
13 often diverge widely when extrapolated to doses well outside the observed range. Indeed, the
14 inability of curve-fitting procedures to provide useful compound-specific information about low-
15 dose risks has been a principal motivation for the "model free" approach of straight line
16 extrapolation from a point of departure within the observed range of the data (Krewski and van
17 Ryzin, 1981; NRC, 1983).
18 Calculations here encompassed four alternative functional forms frequently used for
19 noncancer dose-response assessment in the observable range of experimental data (multistage,
20 Weibull, log-logistic, and log-probit) that can accommodate a wide variety of dose-response
21 shapes, including threshold-like behavior. These models were fit to the mononuclear cell
22 leukemia data in male rats using the EPA BMDS program, and distributions of model results
23 were evaluated (see Appendix 5B for more details). These calculations confirm the expected
24 finding that alternative functional forms fit to this tetrachloroethylene tumor data set are
25 consistent with a wide range of numerical values for probability of response when extrapolated
26 down to low dose, as illustrated in Table 6-2.
27 With such large spreads in confidence intervals, the extrapolated models, in effect,
28 provide little information about actual low-dose risks. These results are not presented as the
29 basis for alternative estimates of human risk, because they do not provide sound or useable
30 scientific estimates for the compound-specific risks from tetrachloroethylene. As noted
31 previously, such results serve to underscore the EPA Cancer Guidelines' rationale for the use of
32 a consistent model-independent approach.
33 A number of different biological motivations have been put forward to support functional
34 forms that might be used to estimate risks from low-dose exposure to carcinogens or other toxic
35 substances. For cancer, the most prominent class of models treats tumorigenesis as a multi-event
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-26 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 6-2. Summary of dose-specific extra risks (means and 95% confidence
limits) for four dose-response models fit to incidence of leukemias in male
rats exposed to tetrachloroethylene via inhalation (JISA, 1993)
Model
Log-probit
Multistage
Log-logistic
Weibull
Estimated extra risk, (mg-eq/kg-day)"1,
corresponding to 1.5 x 10"5 mg-eq/kg-day internal dose in ratsa
5th percentile
2 x 10-170
4.546 x 10'13
2 x 10'21
0
Mean
8.172xlO'4
9.172 x 10'7
1.078 x 10'3
1.339 x 10'3
95th percentile
1.008 x 10'5
1.819 x 10'6
7.776 x 10'4
1.893 x 10'3
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
From Appendix 5B.
process and characterizes the probability of accumulation of a series of changes (conceptualized
as mutations or other events) that, together, will result in formation of a malignant tumor. In
particular, EPA utilized the multistage model for low dose extrapolation of cancer risks in many
assessments. Risk estimates utilizing EPA's application of the multistage model have been
shown to be similar to the linear (straight line) risk estimation procedure now used by EPA
(Subramaniam et al., 2006). More complex multi-event models allow for the modeling of
formation and growth of populations of initiated and transformed cells and are still well
recognized tools for investigating biologically based dose response modeling for carcinogenesis.
The concept of a distribution of individual thresholds is a second approach used to
motivate functional forms for dose-response modeling. Such models assume that there is an
"individual threshold" for each member of the human population, and interindividual variation in
these thresholds determines the dose-response curve for a population. A recent National
Research Council report on risk assessment issues for TCE (NRC, 2006) included a discussion of
models based on distributions of thresholds. That report noted that if one assumes a normal or
logit distribution for individual thresholds this leads to a probit or logistic dose-response function
for the population and suggests that a variety of other distributions for thresholds would also lead
to sigmoidal shaped dose-response functions. The NRC report expressed the view that,
"Although linear extrapolation has been advocated as an intentionally conservative approach to
protect public health, there are some theoretical reasons to think that sublinear nonthreshold
dose-response models may be more relevant for human exposure to toxicants, regardless of the
mode of action" (p. 319). On the other hand, the same report also noted that a very broad class
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-27 DRAFT - DO NOT CITE OR QUOTE
-------
1 of dose-response functions can be obtained using distributions of thresholds models: "In fact any
2 monotonic dose-response model, including the linearized multistage model, can be defined
3 solely in terms of a tolerance distribution without resorting to mechanistic arguments. These
4 considerations suggest that one must consider both the role of mode of action and the role of
5 response variability among humans in determining the likely shape of the dose-response
6 function" (p. 323).
7 The discussion from the NRC TCE document emphasizes some key points in risk
8 assessment. Variability in the human population will have an important influence on the shapes
9 of the dose-response relationships for that population. This is distinct from the amount of
10 variability that may be observed in inbred animal strains. As noted in the NRC report, "One
11 might expect these individual tolerances to vary extensively in humans depending on genetics,
12 coincident exposures, nutritional status, and various other susceptibility factors..." (p. 320).
13 Thus, if a distribution-of-thresholds approach is considered for a carcinogen risk assessment,
14 application would depend on the ability of modeling to reflect the degree of variability in
15 response in human populations. By design, most cancer bioassays are conducted in inbred
16 rodent strains; accordingly, the parameters provided by curve fits of distribution-of-thresholds
17 models to bioassay data would not be predicted to reflect the dose-response patterns in diverse
18 human populations. It is important to note that the NRC text has no recommendation for an
19 approach where a tolerance distribution model for humans is estimated by a statistical fit to
20 rodent bioassay data.
21 The question of whether a tolerance distribution model is indeed an appropriate basis for
22 a risk assessment also warrants consideration. Low-dose linearity can arise in other contexts
23 distinct from effects of population variability and may be directly appropriate to a MOA. Low-
24 dose linearity can also arise due to additivity of a chemical's effect on top of background
25 chemical exposures and biological processes. In the case of chemicals such as
26 tetrachloroethylene, basic biological data do not exist to support the appropriateness of an
27 individual threshold model above models having inherent low-dose linearity. However, if
28 distribution of thresholds modeling were supported, it would need to be developed based on an
29 examination of predicted variability within in human population.
30 Given the current state of scientific knowledge about tetrachloroethylene carcinogen!city,
31 the straight-line-based risk estimates presented above form the preferred recommendation for
32 estimating a plausible upper-bound estimate of potential human risks from tetrachloroethylene.
33 This approach is supported by both general scientific considerations, including those supporting
34 the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), as well as chemical-specific
35 findings. The former include the scientific principles articulated previously (ie, the expectation
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-28 DRAFT - DO NOT CITE OR QUOTE
-------
1 that a chemical functions additively to background exposures, diseases, and processes; that
2 variability within the human population would broaden the dose-response curve and eliminate
3 individual thresholds if present; and that the approach provides consistency across assessments
4 facilitating direct comparison of the derived risk values). The latter include evidence that, within
5 the dose range of the cancer bioassays, the observable tumor response data are consistent with a
6 linear model and do not suggest occurrence of a threshold, and that variability in the human
7 response across the population is expected (eg, Bois et al., 1996; Clewell et al., 2004).
8
9 6.2.2.2.3. Dose metric. Tetrachloroethylene is metabolized to several intermediates with
10 carcinogenic potential. Although much data exist for the metabolite TCA, several analyses
11 indicate that TCA alone is not able to explain the toxicity associated with tetrachloroethylene
12 exposure; therefore, at least one other toxic agent appears to be involved. It is unclear whether
13 total metabolism—either as a measure of a precursor or intermediate, or as a surrogate directly
14 proportional to the toxic agent(s)—is an adequate indicator of potential risk. Since the
15 experimental evidence supports a role for metabolism in tetrachloroethylene's toxicity, use of
16 total metabolism (the only measure of metabolism available) to estimate cancer risk is germane
17 to this assessment. Use of administered dose (without use of a PBPK model) yields risk
18 estimates intermediate between those based on the higher and lower PBPK models.
19
20 6.2.2.2.4. Choice of species/gender. Table 6-3 summarizes the factors influencing the choice of
21 rodent tumor data set for human risk characterization. It is assumed that the observed rodent
22 tumors are relevant to humans, an assumption supported by a number of factors. Primary among
23 these factors is that a carcinogenic response is also observed in humans. Human-rodent site
24 concordance is not generally assumed (e.g., due to potential differences in pharmacokinetics,
25 DNA repair, other protective systems across species and tissues [U.S. EPA, 2005a]). In keeping
26 with this view, certain tumors associated with tetrachloroethylene exposure in human mortality
27 studies (e.g., cervix and esophagus) were not observed in rodents; cancer of the lymphoid system
28 was associated with tetrachloroethylene exposure in humans, with some evidence for an
29 association with bladder, kidney, and lung cancer. In addition, rat and mouse tumor types also
30 differ from each other. Finally, conclusive MOA data are lacking for the observed rodent and
31 human tumors.
32 MCL is the cancer response of highest magnitude and is reproducible in two bioassays
33 and in both genders. Although MCL has a high and variable incidence in unexposed F344 rats, a
34 biologically and statistically significant increase over background was observed (see
35 Section 5.4.1). Section 4.8.2.2.1.4 addresses the qualitative similarities among MCL to certain
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-29 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
Table 6-3. Summary of considerations for each rodent tumor type
Magnitude of
response
Specificity of
response
Replication of
findings in multiple
bioassays
Other considerations
contributing to
uncertainty in rodent
data
Relevance to humans:
a) qualitative
(biologic) site
concordance
b) occurrence in
human studies
c) confidence in
MOA
Overall
considerations for
choice of tumor type
Mononuclear cell
leukemia
+++
Rats, both genders
4/4 study datasets
High background
rate
Yes
Yes, but exact
match of tumor
classification is not
found/may not be
possible
No data
Rodent response of
highest magnitude,
reproducible; no
MOA data
Liver
++
Mice, both
genders
6/6 study datasets
High background
rate in male mice
Yes
Yes, but
association is
weak
PPAR-a
activation may
contribute, but is
not sole MOA
Rodent response
of considerable
magnitude,
reproducible,
some MOA data
Kidney
++
Male rats only
1/2 study datasets
Rare tumor
(unlikely to be due
to chance, but low
incidence)
Yes
Yes, but
association is weak
Multiple MO As
may play a role
Rare tumor in
rodents and
humans; MOA
data are strongest
Hemangiosarcoma
+
Male mice only
1/3 study datasets
Rare tumor (unlikely
to be due to chance,
but low incidence)
Yes
No, but tumor type is
rare
No data
Rare tumor in rodents
and humans; no
MOA data
4
5
6
7
8
9
10
11
12
13
lymphoid cancers, and the implications regarding human relevance. This section also addresses
the association of elevated lymphoma mortality with tetrachloroethylene exposure in humans.
The MOA for MCL remains unexplored.
Male rats had the higher response level of MCL as estimated using the multistage model.
Occasionally, if the multistage model does not adequately fit a data set, an alternate model can be
used to determine the POD. In the case of female rat MCL data, the best-fitting model (Weibull)
yielded central tendency risk estimates 10-fold higher than those from the multistage model fit of
the male rat MCL data. Consequently, there is some uncertainty in characterizing the magnitude
of MCL response, with the use of the male rat MCL data possibly underestimating risk. While
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-30 DRAFT - DO NOT CITE OR QUOTE
-------
1 the female rat MCL data suggest a supralinear fit extending into lower exposure levels (also
2 apparent for female rat MCL data from the NTP bioassay), the multistage model fit was
3 technically adequate (p = 0.48). In keeping with EPA's past practice of preferring the multistage
4 model in order to provide some measure of consistency across different carcinogen assessments,
5 the more linear multistage fit to the female rat MCL data supports the risk estimate derived from
6 the male rat MCL data.
7 The mouse liver tumor is a robust (i.e., of significant magnitude) finding in several
8 studies, including in both sexes. As is the case with MCL, the background for this tumor type is
9 high—especially in males. A biologically and statistically significant increase over background
10 was observed in males and females. There is evidence that activation of the PPAR-a receptor by
11 the tetrachloroethylene metabolite TCA contributes to the induction of mouse liver tumors.
12 However, it is not the only operative MOA involved in hepatocellular tumorigenesis. Thus, the
13 MOA remains unresolved.
14 Two tumor types were observed in only one bioassay. Kidney tumors rarely occur in
15 unexposed rodents but were significantly elevated with tetrachloroethylene exposure in the male
16 rat NTP bioassay. The MOA is better understood for kidney tumors than for the other sites.
17 Hemangiosarcoma is another rare tumor associated with tetrachloroethylene exposure in the
18 male mouse JISA (1993) study. There are no MOA data for hemangiosarcomas.
19
20 6.2.2.2.5. PBPK model. Toxicokinetic models are used in this assessment for deriving dose
21 metrics to support dose-response analyses. The evidence suggests that the by-products of
22 tetrachloroethylene metabolism are responsible for liver and kidney toxicity and for
23 carcinogenicity. Inhaled concentration of the parent compound is, therefore, not an appropriate
24 dosimeter for these effects, and pharmacokinetic modeling of daily overall metabolized dose is
25 expected to be an improvement in spite of the many attendant uncertainties in the modeling. Of
26 the available toxicokinetic models on tetrachloroethylene, the assessment considers three
27 recently developed models that describe parent tetrachloroethylene and overall metabolism of the
28 parent compound in humans. These models do not describe the kinetics and transformation of
29 total metabolic products or any individual metabolite. All three models provide reasonably good
30 predictions of exhaled breath and blood tetrachloroethylene concentrations, so there is no
31 particular basis for preferring one model over another. A 10-fold difference is shown in model
32 predictions of the rate of metabolism in humans, a reflection of model differences in the values
33 for the metabolic parameters. Because the accuracy of the models has been evaluated only
34 against blood and breath concentrations of the parent compound—quantities that are insensitive
35 to these parameters—the reliability of these models for predicting the rate of total metabolism in
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-31 DRAFT - DO NOT CITE OR QUOTE
-------
1 humans is unknown. Data on total metabolite levels are not available in humans, and the use of
2 available urinary and blood TCA data is problematic. The overall difference in risk estimates
3 using these three models is approximately 10-fold.
4
5 6.2.2.2.6. Cross-species scaling. An adjustment for cross-species scaling (BW3/4) was applied
6 to address toxicological equivalence of internal doses between each rodent species and humans,
7 consistent with the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a); the
8 approach is detailed in Section 5.4.4.2.1. It is assumed that, without data to the contrary, equal
9 risks result from equivalent constant exposures. While the true correspondence of equipotent
10 tetrachloroethylene exposures across species is unknown, the use of BW3/4 scaling is expected
11 neither to over- or underestimate human risk (U.S. EPA, 1992).
12
13 6.2.2.2.7. Choice ofbioassay. The JISA (1993) inhalation bioassay provides data on the lowest
14 experimental exposures, and its use, therefore, reduces extrapolation uncertainty slightly. For
15 mice, the lowest exposure concentration of 10 ppm was 10-fold lower than the lowest exposure
16 concentration in the NTP inhalation study (NTP, 1986). For rats, the low-exposure concentration
17 of 50 ppm was fourfold lower than in the NTP study. Although the JIS A and NTP inhalation
18 bioassays used similar rodent strains, differences in the animals used (in addition to other
19 unidentified factors) may have contributed to the twofold higher incidence of hepatocellular
20 tumors and MCL in the NTP study. Consequently, the estimated risks are twofold lower than
21 previous EPA assessments which relied on the NTP bioassay (U.S. EPA, 1991).
22
23 6.2.2.2.8. Statistical uncertainty at the point of departure. Parameter uncertainty within the
24 chosen model reflects the limited sample size of the cancer bioassay. For the multistage model
25 applied to this data set, there is a relatively small degree of uncertainty at the 10% extra risk level
26 (the point of departure for linear low-dose extrapolation).
27
28 6.2.2.3. Quantitative Analysis of Multiple Uncertainties on Cancer Unit Risk
29 Figure 6-4 and Table 6-4 show the central estimates and upper and lower confidence
30 limits of the inhalation risk per unit concentration for the rodent data sets under consideration, as
31 determined using BMDS (version 1.3.2). The upper bound inhalation risk per unit concentration
32 has been calculated as the ratio of the benchmark response (10% extra risk for all data sets
33 except the kidney tumors, which was 5%) to the 95% lower confidence limit of the benchmark
34 dose (the LECio). These results show that the lower bound of risks ranges from 20% to 40% of
35 the upper bound of risks. The values at the right end of each bar represent the unit risk estimates
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-32 DRAFT - DO NOT CITE OR QUOTE
-------
* ta—i
* o
Male rat, MCL
Male mouse, hepatocellular tumors
and hemangiosarcomas
Female rat, MCL
Female mouse, hepatocellular tumors
Kidney tumors, male rat (NTP)
Hemangiosarcomas, male mouse
IE-OS 1E-07 0.000001 0.00001 0.0001 0.001 0.01 0.1
Human equivalent tetrachloroethylene risk estimates (per ug/cu.m)
1
2
3
4
5
6
7
8
9
10
11
12
13
Figure 6-4. Cancer risk estimates for tumor sites associated with
tetrachloroethylene exposure in rodent bioassays, using the multistage model.
The four gender/species data sets are provided in the upper section of the graph
while two tumor types observed in single bioassays are provided in the lower
section. The symbols denote the slopes (to background risk) from the mean
estimate of exposure corresponding to 10% extra risk, using the Rao and Brown
(1993) PBPK model (n) and the Bois et al. (1996) model (o) to extrapolate to
human equivalent exposures. The bars indicate the slopes from the lower and
upper bounds on the mean estimates. * indicates lower bounds that could not be
estimated.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-33 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 6-4. Combined impact on tetrachloroethylene cancer risk estimates
(per ug/m3) of statistical uncertainty,3 PBPK model and tumor site(s), using
multistage model in observed range and linear low-dose extrapolation
Gender/species (tumor type)
Male rat (MCL)
Male mouse (hepatocellular tumor and
hemangiosarcoma)
Female rat (MCL)
Female mouse (hepatocellular tumor)
LB
C
UB
LB
C
UB
LB
C
UB
LB
C
UB
PBPK model
Rao and Brown
(1993)
5 xlO'7
1 x 10'6
2 x 10'6
*
3 x 10'7
1 x 10'6
*
5 x 10'7
9 x 10'7
1 x 10'7
2 x 10'7
3 x 10'7
Bois et al.
(1996)
5 xlO'6
1 x 10'5
2 x 10 5
*
3 x ID'6
1 x 10'5
*
5 x ID'6
9 x 10'6
1 x 10'6
2 x 10'6
3 x 10'6
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
a In some cases, the lower bounds on risk could not be estimated.
LB = Lower risk estimate derived from upper statistical confidence limit on the POD concentration (UECio).
C = Central risk estimate derived from the MLE estimate of EQ 0, and from the mean of the bootstrap distribution
of BMR/ECio values (equal to each other in this case, see Appendix 5B).
UB = Upper bound risk estimate derived from the lower bound statistical confidence limit POD concentration
(LECio).
Bolded value is used to derive assessment's unit risk estimate.
supported by each data set. Figure 6-4 and Table 6-4 also show the range of upper bound
inhalation risks due to the highest and lowest metabolic rate pharmacokinetic models used to
describe the rate of metabolism of tetrachloroethylene, as described in Sections 6.2.2.2 and 3.5
and Tables 5-8 and 5-9. A third model (Reitz et al., 1996), not shown in Figure 6-4, yields
results between the other two. For each PBPK model, unit risk estimates based on the male
mouse and female rat are similar, each about twofold lower than the male rat MCL unit risk
estimate. The unit risk estimate based on the least sensitive species/gender (female mouse,
hepatocellular tumor) is about eightfold less than that given by the male rat MCL estimate. Two
tumor types, each seen in only one bioassay, would respectively give unit risk estimates eightfold
lower (male mouse hemangiosarcoma in the JISA [1993] bioassay) and fivefold lower (male rat
kidney tumor in the NTP bioassay) than the JISA male rat MCL unit risk estimate. Unit risk
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-34 DRAFT - DO NOT CITE OR QUOTE
-------
1 estimates based on administered dose (without use of a PBPK model) do not correlate with those
2 from any particular PBPK model, but they generally fall between the higher and lower unit risk
3 estimates derived from the PBPK models.
4 Because rodent carcinogenicity is consistently evident in all data sets, with multiple types
5 of tumors occurring, the concern for human carcinogenic risks is increased. This supports
6 selection of the most sensitive observation as a basis for risk estimation. For tetrachloroethylene,
7 MCL in male rats is the basis for risk estimation. While this tumor type tends to have a high
8 background response in rats, it was not as high in the JISA (1993) rats, at about 20%, compared
9 with the male NTP rats at about 56%. Consistent with EPA's 2005 Guidelines for Carcinogen
10 Risk Assessment, it is still important to communicate the potential for increased cancer incidence
11 over background. In addition, while there is no exact analogue of MCL in humans, as noted
12 earlier, the EPA 2005 Guidelines for Carcinogen Risk Assessment notes that site concordance is
13 not necessary for assessing potential carcinogenic risk to humans.
14
15 6.2.2A. Conclusions
16 Tetrachloroethylene is "Likely to be carcinogenic to humans" by all routes of exposure,
17 using the framework specified in the Guidelines for Carcinogen Risk Assessment (U.S. EPA,
18 2005a). Three lines of evidence in the hazard database support this weight-of-evidence
19 descriptor for the cancer hazard for tetrachloroethylene:
20 (1) Tetrachloroethylene is a carcinogen in rodents in 10 of 10 lifetime bioassay data sets—
21 including by oral and inhalation routes
22 (2) It is reasonable to use these animal tumors as indicators of potential human cancer hazard
23 (3) Tetrachloroethylene exposure is consistently associated with excess risks for a number of
24 cancers in human epidemiologic studies, although a causal association has yet to be
25 definitively established.
26
27 The laboratory animal database includes 10 lifetime rodent bioassay data sets
28 demonstrating increased cancer incidence (two more study data sets were inconclusive due to
29 excessive mortality from pneumonia or tetrachloroethylene-related toxic nephropathy). The
30 findings include liver cancers in both sexes of mice following either oral or inhalation exposures,
31 and following inhalation exposures, mononuclear cell leukemias in both sexes of rats (multiple
32 bioassays), as well as male rat kidney and brain tumors (gliomas) and male mouse
33 hemangioendotheliomas of the liver or spleen (single bioassays). In addition, although not all
34 tetrachloroethylene metabolites have been tested for carcinogenicity in rodents, the oxidative
35 metabolites TCA and DC A are hepatocarcinogens in one or more species. Although insufficient
36 to establish causality, the epidemiologic evidence has consistently shown a positive association
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-3 5 DRAFT - DO NOT CITE OR QUOTE
-------
1 of inhalation and oral tetrachloroethylene exposure with excess risks for a number of neoplasms.
2 These include cancer of the lymphoid system, esophagus, and cervix, with more limited evidence
3 for cancer of the bladder, kidney, and lung. Taken together, these data support a weight-of-
4 evidence descriptor of "Likely to be carcinogenic to humans" by all routes of exposure for
5 tetrachloroethylene. Use of the weight-of-evidence descriptor "Likely to be carcinogenic to
6 humans" for tetrachloroethylene is intended to communicate that the available information
7 indicates the presence of a human health hazard.
8 Consistent with this view, dose-response assessments are generally completed for agents
9 considered "Likely to be carcinogenic to humans." The unit risk is intended to be a plausible
10 upper bound estimate of risk and, accordingly, all such estimates described below are based on
11 the following: (1) the most sensitive tumor type in rodents, with regard to species, gender, and
12 type of malignancy; (2) the POD based on the upper confidence bound on risk derived from
13 statistical modeling of the observed dose-response data; and (3) a linear low-dose extrapolation
14 approach. A linear extrapolation was performed in accordance with default recommendations in
15 the Guidelines of Carcinogen Risk Assessment (U.S. EPA, 2005a) because of the lack of
16 substantial biological basis for doing otherwise (particularly, the lack of knowledge about the
17 MO A for any of the observed tumors), and other approaches to estimate upper bounds on risk
18 were not considered informative for risk estimation. Table 6-5 gives a summary of the impact
19 and justification of these choices. On the other hand, alternative choices for these approaches,
20 while providing a perspective as to the overall uncertainty in human cancer risk, would not
21 provide upper bounds on risk.
22 Given the choices of tumor type, point of departure, and low-dose extrapolation approach
23 described in Table 6-5, there are additional considerations that contribute to uncertainty in the
24 plausible upper bound unit risk, which are summarized in Table 6-6. These uncertainties have a
25 varied impact on risk estimates. Some (i.e., the bioassay or cross-species scaling approach)
26 suggest risks could be higher than estimated, while others would decrease estimates or have an
27 impact of uncertain direction (i.e., the human population variability, dose metric, and model-
28 based uncertainty at the POD). While some uncertainties could be quantitatively characterized,
29 it is likely that the residual uncertainties represent the largest and can only be qualitatively
30 expressed. Such uncertainties pertain to MOA and human sensitivity and variability. Even if
31 these could be further elucidated by additional data, extrapolation of animal bioassay data to
32 humans (done here using allometric scaling) will remain a substantial and unknown uncertainty.
33 The PBPK model uncertainty is the only one for which there is no basis for preferring one
34 alternative to another, so the tetrachloroethylene unit risk estimate, calculated using three PBPK
35 models, ranges from 2 x 10"6 to 2 x 10"5 per |ig/m3.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-36 DRAFT - DO NOT CITE OR QUOTE
-------
o
Oi
Table 6-5. Considerations leading to the determination of a reasonable upper bound on risk
This document is a draft for review purposes only and does not constitute Agency policy
06/08 6-37 DRAFT: DO NOT CITE OR QUOTE
Consideration/
approach
Most sensitive
tumor response
Statistical sampling
uncertainty in
observable range
Dose-response
relationship at low-
dose
Impact on estimated risk
8-fold range from most to least sensitive rodent
tumor type. For example, Bois PBPK model
gives the following risk estimates using the
LECio and linear low-dose extrapolation in
(ug/m3)-1:
Rats, MCL
Males
Female
Mice, liver tumors
Males
Female
2 x 10'5
9 x 10'6
1 x 1Q-5
3 x IQ-6
J, 1.6-fold if ECio used rather than LECio, e.g.
Bois PBPK model gives the following risk
estimates in (ug/m3)-1 using linear low-dose
extrapolation with male rat MCL data:
Using UECi0: 5x 10'6
Using ECi0: 1 x 10'5
Using LECi0: 2x 10'5
Could J, to a negligible
unknown extent, and is
uncertainties.
value or t to an
among the largest
Decision
Male rats, MCL.
Lower bound on 10% risk
concentration (LECio)-the
EPA's default approach for
calculating a plausible upper
bound.
Linear low-dose extrapolation
from POD (default approach).
Justification
MCL had the greatest response and is reproducible
across studies.
Limited size of bioassay results in sampling
variability; lower bound (LECio) is the lower 95%
confidence interval on the concentration yielding a
10% risk and thus makes it less likely the estimate
underestimates risk due to small sample size.
Low-dose linear approach is supported by general
considerations (additivity to background,
population heterogeneity) and tetrachloroethylene-
specific data (male rat MCL data are linear in
observed range).
-------
Table 6-6. Considerations that impact uncertainty in reasonable upper bound risk estimates
Consideration/
approach
Human population
variability in
metabolism and
response/ sensitive
subpopulations
Dose metric
(uncertainty about
active moiety)
PBPK model
Cross-species
scaling
Bioassay
Model-based
uncertainty in POD
Impact on estimated risk
There may be subpopulations at higher risk and
there may be individuals for whom risk is
negligible.
Alternatives could t or J, risk estimate by an
unknown extent.
10-fold range among the three available models,
e.g. see below estimates of risk per unit
concentration (ug/m3)"1 using linear
extrapolation from LECio
Male Rats, MCL Bois: 2 x 10"5
Reitz: 9 x 10"6
Rao: 2 x 10'6
Alternative generic scaling approaches could J,
or t risk estimate (e.g., 3.5-fold J, [scaling by
BW] or t 2-fold [scaling by BW2/3 ]). Residual
uncertainty in scaling may t or J, risk estimate
by an unknown extent.
t 2-fold if NTP study used.
1.4-fold range in models that were explored.
Decision
Considered qualitatively.
Considered total metabolism
and administered
concentration.
The highest value provides a
reasonable upper bound
estimate of potential human
cancer risk; a range of
estimates is provided.
BW3/4 (default approach).
JISA study.
Multistage model used.
Justification
No data to support alternative estimates.
Experimental evidence supports a role for
metabolism in toxicity, but actual responsible
metabolites are not clearly identified.
There is no scientific basis for choosing among
pharmacokinetic results for estimating total
metabolism of tetrachloroethylene given
limitations in available data.
BW3/4 limits bias in estimate.
JISA study used the lowest experimental
exposures (reduces extrapolation uncertainty).
Flexible model; limited additional uncertainty
based on comparison with three other models.
-------
1 As addressed in the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
2 derivation of the central and lower bound risk estimate can be of value in some settings, such as
3 screening analyses and for ranking agents as to their carcinogenic hazard. For such purposes, the
4 cancer risk values based on the ECio represent the central and lower bound estimates of risk,
5 respectively, for a particular data set. For tetrachloroethylene, a range of central estimates (based
6 on the ECio using male rat MCL data from the JISA study and linear low-dose extrapolation and
7 based on three PBPK model choices, as addressed in Table 6-4) is from 1 x 10"6 to 1 * 10"5 per
8 ng/m3. The corresponding range of lower bound estimates (derived from the UECio and based
9 on the same choices of tumor type, low-dose extrapolation approach and using the three
10 available PBPK model choices, as addressed above) is from 5 x 10"7 to 5 x 10"6 per |ig/m3.
11 To summarize, tetrachloroethylene is "Likely to be carcinogenic to humans" by all routes
12 of exposure. A lack of human carcinogenicity, while not ruled out, is considered unlikely.
13 Existing data indicate that (1) tetrachloroethylene is a rodent carcinogen in 10 of 10 lifetime
14 bioassay datasets, including by oral and inhalation routes (2) the observed animal effects are
15 relevant to use as indicators of human carcinogenic risk; and (3) tetrachloroethylene exposure is
16 associated with excess risks for several cancers in human epidemiological studies, although a
17 causal relationship has yet to be established. In addition, the carcinogenicity of
18 tetrachloroethylene is also supported by other lines of evidence, including data on its metabolism
19 and pharmacokinetics and the demonstrated hepatocarcinogenicity of the oxidative metabolites
20 TCA and DC A in one or more species. In view of the likely carcinogenicity, a dose-response
21 assessment was undertaken with the purpose of identifying a plausible upper bound estimate of
22 risk. A range of unit risk estimates for tetrachloroethylene is from 2 x 10"6 to 2 x 10"5 per |ig/m3,
23 with the upper-end unit risk of 2 x 10"5 per |ig/m3 being the most public health protective value
24 for the upper bound risk estimate.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-39 DRAFT - DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR CHAPTER 6
2
O
4 Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
5 low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
6
7 Birner, G; Rutkowska, A; Dekant, W. (1996) N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and 2,2,2-
8 trichloroethanol: two novel metabolites of tetrachloroethene in humans after occupational exposure. Drug Metab
9 Disp 24:41-48.
10
11 Bogen, KT; McKone, TE. (1988) Linking indoor air and pharmacokinetic models to assess tetrachloroethylene risk.
12 Risl Anal 8:509-520.
13
14 Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol
15 70(6):347-55.
16
17 Brodkin, CA; Daniell, W; Checkoway, H; et al. (1995) Hepatic ultrasonic changes in workers exposed to
18 perchloroethylene. Occup Environ Med 52:6796685.
19
20 Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
21 and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
22
23 Cavalleri, A; Gobba, F; Paltrinieri, M; et al. (1994) Perchloroethylene exposure can induce colour vision loss.
24 Neurosci Lett 179:162-166.
25
26 Chiu, WE; Bois, FY. (2006) Revisiting the population toxicokinetics of tetrachlorotethylene. Arch Toxicol
27 80:382-385.
28
29 Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
30 pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79:381-393.
31
32 Doyle, P; Roman, E; Beral, V; et al. (1997) Spontaneous abortion in dry cleaning workers potentially exposed to
33 perchloroethylene. Occup Environ Med 54:8486853.
34
35 Echeverria, D; Heyer, N; Checkoway, H; et al. (1994) Behavioral investigation of occupational exposure to solvents:
36 perchloroethylene among dry cleaners, and styrene among reinforced fiberglass laminators. Final Report. Report
37 prepared for the Centers for Disease Control and Prevention under Grant No. 5 R01 OHo2719-03. Battelle Centers
3 8 for Public Health Research and Evaluation.
39
40 Echeverria, D; White, RF; Sampaio, C. (1995) A behavioral evaluation of PCE exposure in patients and dry
41 cleaners: a possible relationship between clinical and preclinical effects. J Occup Environ Med 37:667-680.
42
43 Eskenazi, B; Fenster, L; Hudes, M; et al. (1991a) A study of the effect of perchloroethylene exposure on the
44 reproductive outcomes of wives of dry-cleaning workers. Am J Ind Med 20:5936600.
45
46 Eskenazi, B; Wyrobek, AJ; Fenster, L; et al. (1991b) A study of the effect of perchloroethylene exposure on semen
47 quality in dry cleaning workers. Am J Ind Med 20:5756591.
48
49 Ferroni, C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
50 perchloroethylene. Neurotoxicology 13:243-247.
51
52 Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
53 solvents. Int Arch Occup Environ Health 52:1-9.
54
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-40 DRAFT - DO NOT CITE OR QUOTE
-------
1 Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
2 tetrachloroethylene as neonates. Toxicol Lett 66:13Bl9.
O
4 Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
5 tetrachloroethylene. Am J Ind Med 21:661-671.
6
7 Green, T; Odum, J. Nash, JA; et al. (1990) Perchloroethylene-induced rate kidney tumors: an investigation of the
8 mechanisms involved and their relevance to human. Toxicol Appl Pharmacol 103:77B89.
9
10 Hayes, JR; Condie, LW, Jr.; Borzelleca, JF. (1986) The subchronic toxicity of tetrachloroethylene
11 (perchloroethylene) administered in the drinking water of rats. Fundam Appl Toxicol 7:119-125.
12
13 JISA (Japan Industrial Safety Association). (1993) Carcinogenicity study of tetrachloroethylene by inhalation in rats
14 and mice. Data No. 3-1. Available from: EPA-IRIS Information Desk.
15
16 Kjellstrand, P; Holmquist, B; Kanje, M; et al. (1984) Perchloroethylene: effects on body and organ weights and
17 plasma butyrylcholinesterase activity in mice. Acta Pharmacol Toxicol (Copenh) 54:414-424.
18
19 Krewski, D; vanRyzin, J. (1981) Dose response models for quantal response toxicity data, in statistics and related
20 topics. In: Csorgo, M; Dawson, DA; Rao, JNK; Saleh, ADMdE; eds. North-Holland Publishing Company.
21
22 Lutz, WK; Gay lor, DW; Conolly, RB; et al. (2005) Nonlinearity and thresholds in dose-response relationships for
23 carcinogenicity due to sampling variation, logarithmic dose scaling, or small differences in individual susceptibility.
24 Toxicol Appl Pharmacol 207:565-569.
25
26 Mattsson, JL; Albee, RR; Yano, BL; et al. (1998) Neurotoxicologic examination of rats exposed to 1,1,2,2-
27 tetrachloroethylene (perchloroethylene) vapor for 13 weeks. Neurotoxicol Teratol 20:83B98.
28
29 Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-
30 cleaners. Lancet 340:1896193.
31
32 Nagano, K; Nishizawa, T; Yamamoto, S; et al. (1998) Inhalation carcinogenesis studies of six halogenated
33 hydrocarbons in rats and mice. In: Chiyotani, K; Hosoda, Y; Aizawa, Y, eds. Advances in the prevention of
34 occupational respiratory diseases: proceedings of the 9th international conference on occupational respiratory
35 diseases, Kyoto, Japan, October 13-16, 1997. Amsterdam: Elsevier; pp. 741-746.
36
37 National Toxicology Program (NTP). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
38 (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3Flmice. National Institutes of Health, Public
39 Health Service, U.S. Department of Health and Human Services. Available online at http://ntp.niehs.nih.gov.
40
41 NRC (National Research Council). (1983) Risk assessment in the Federal government: managing the process.
42 Committee on the Institutional Means for Assessment of Risks to Public Health, Commission on Life Sciences,
43 NRC. Washington, DC; National Academy Press.
44
45 NYS DOH (New York State Department of Health). (1997) Tetrachloroethylene ambient air criteria document.
46 Final Report. Albany, NY.
47
48 NYS DOH (New York State Department of Health). (2005a) Improving human risk assessment for
49 tetrachloroethylene by using biomarkers and neurobehavioral testing. U.S. EPA Star Grant #R827445. Grant
50 #R827446. Available online at
51 http://cfpub.epa.gov/ncer abstracts/index.cfm/fuseaction/displav.abstractDetail/abstract/977/reprort/O.
52
53 NYS DOH (New York State Department of Health). (2005c) Pumpkin patch day care center follow-up evaluation.
54 Final Report. New York State Department of Health, Bureau of Toxic Substance Assessment, Division of
5 5 Environmental Health Assessment, Center for Environmental Health.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-41 DRAFT - DO NOT CITE OR QUOTE
-------
2 Olsen, J; Hemminki, K; Ahlborg, G; et al. (1990) Low birthweight, congenital malformations, and spontaneous
3 abortions among dry-cleaning workers in Scandinavia. Scand J Work Environ Health 16:163B168.
4
5 Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
6 groundwater for a bathing and showering determination. Risk Anal 13:37^9.
7
8 Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
9 for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
10 306.
11
12 Rosengren, LE; Kjellstrand, P; Haglid, KG. (1986) Tetrachloroethylene: levels of DNA and S-100 in the gerbil CNS
13 after chronic exposure. Neurobehav Toxicol Teratol 8:201-206.
14
15 Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
16 tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect 110:655-664.
17
18 Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
19 11:579-583.
20
21 Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
22 of subjects chronically exposed to organic solvents (II)]. G Ital Med Lav Ergon 19:85-88.
23
24 Stromberg, PC. (1985) Large granular lymphocyte leukemia in F344 rats. Model for human T gamma lymphoma,
25 malignant histiocytosis, and T-cell chronic lymphocytic leukemia. Am J. Pathol 119:517-519.
26
27 Tinston, DJ. (1994) Perchloroethylene: A multigeneration inhalation study in the rat. CTL/P/4097. Available from:
28 EPA IRIS Information Desk.
29
3 0 Trevisan, A; Macca, I; Rui, F; et al. (2000) Kidney and liver biomarkers in female dry-cleaning workers exposed to
31 perchloroethylene. Biomarkers 5:399B409.
32
33 Umezu, T; Yonemoto, J; Soma, Y; Miura, T. (1997) Behavioral effects of trichloroethylene and tetrachloroethylene
34 in mice. Pharmacol Biochem Behav 58:665-671.
35
36 U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
37 carcinogenicity of tetrachloroethylene (perchloroethylene). Office of Health and Environmental Assessment,
38 Washington, DC; EPA/600/6-91/002F. Available from: National Technical Information Service, Springfield, VA.
39
40 U.S. EPA (Environmental Protection Agency). (1992) Draft report: a cross-species scaling factor for carcinogen risk
41 assessment based on equivalence of mg/kg3/4/day. Federal Register 24152-24173.
42
43 U.S. EPA (Environmental Protection Agency). (1993) Reference Dose (RfD): Description and Use in Health Risk
44 Assessments Background Document 1A, March 15, 1993.
45
46 U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
47 and application of inhalation dosimetry. Office of Health and Environmental Assessment, Environmental Criteria
48 and Assessment Office, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical Information
49 Service, Springfield, VA; PB2000-500023, and online at http://www.epa.gov/ncea.
50
51 U.S. EPA (Environmental Protection Agency). (2005a). Guidelines for carcinogen risk assessment. Federal Register
52 70(66)17765-17817. Available online at http://www.epa.gov/cancerguidelines.
53
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-42 DRAFT - DO NOT CITE OR QUOTE
-------
1 U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
2 early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
3 online at http://www.epa.gov/cancerguidelines.
4
5 Verplanke, AJ; Leummens, MH; Herber, RF. (1999) Occupational exposure to tetrachloroethene and its effects on
6 the kidneys. J Occup Environ Med 41:11B16.
7
8 Warren, DA; Reigle, TG; Muralidhara, S; et al. (1996) Schedule-controlled operant behavior of rats following oral
9 administration of perchloroethylene: time course and relationship to blood and brain solvent levels. J Toxicol
10 Environ Health 47:345-362.
11
12 Zeise, L; Wilson, R; Crouch, EA. (1987) Dose-response relationships for carcinogens: a review. Environ Health
13 Perspect 73:259-306.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 6-43 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14 APPENDIX A:
15
16 SUMMARY OF EARLIER ASSESSMENTS
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-l DRAFT—DO NOT CITE OR QUOTE
-------
1 APPENDIX A: SUMMARY OF EARLIER ASSESSMENTS
2
3 A.I. ORAL INGESTION ASSESSMENTS A-4
4 A.I.I. U.S. EPA Oral Ingestion Assessments A-4
5 A.I.1.1. IRIS Database, U.S. EPA, 2005 A-4
6 A. 1.1.2. Health Assessment Document for Tetrachloroethylene,
7 U.S. EPA, 1985 A-4
8 A.I.2. Oral Ingestion Assessments Conducted by Non-U.S. EPA Agencies A-4
9 A. 1.2.1. California Environmental Protection Agency, 2000, Draft
10 Public Health Goal for Tetrachloroethylene in Drinking Water A-4
11 A. 1.2.2. Agency for Toxic Substances and Disease Registry (ATSDR)
12 Toxicological Profile for Tetrachloroethylene (Update),
13 September 1997 A-5
14 A. 1.3. Summary of Ingestion Risk Estimates A-6
15
16 A.2. INHALATION ASSESSMENTS A-7
17 A.2.1. U.S. EPA Inhalation Assessments A-7
18 A.2.1.1. Cleaner Technologies Substitutes Assessment: Professional
19 Fabricare Processes, Office of Pollution Prevention and Toxics,
20 U.S. EPA, 1998 A-7
21 A.2.1.2. Addendum to the Health Assessment Document for
22 Tetrachloroethylene, U.S. EPA, 1986 A-7
23 A.2.1.3. Health Assessment Document for Tetrachloroethylene,
24 U.S. EPA, 1985 A-8
25 A.2.2. Inhalation Assessments Conducted by Non-U.S. EPA Agencies A-8
26 A.2.2.1. California Environmental Protection Agency, 2002 A-8
27 A.2.2.2. Massachusetts'Derivation of Inhalation Unit Risk for
28 Tetrachloroethylene, Massachusetts Department of
29 Environmental Protection, 1998 A-9
30 A.2.2.3. ATSDR Toxicological Profile for Tetrachloroethylene (Update),
31 September 1997 A-9
32 A.2.2.4. Tetrachloroethane—Ambient Air Criteria Document, New
33 York State Department of Health, 1997 A-10
34 A.2.2.5. Priority Substances List Assessment Report:
3 5 Tetrachloroethylene, Canada Health and Welfare, 1993 A-11
36 A.2.3. Summary of Inhalation Risk Estimates A-ll
37 A.3. QUALITATIVE RISK ASSESSMENTS A-ll
38 A.3.1. U.S. EPA Qualitative Risk Assessments A-13
39 A.3.1.1. Response to Issues and Data Submission on the Carcinogenicity
40 of Tetrachloroethylene, U.S. EPA, 1991 A-13
41 A.3.2. Qualitative Risk Assessments Conducted by Non-U.S. EPA Agencies A-13
42 A.3.2.1. Organization for Economic Cooperation and Development
43 (OECD), Screening Information Data Set (SIDS) Initial
44 Assessment Report: Comprehensive Risk Assessment Report
45 for Tetrachloroethylene, 1996 A-13
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-2 DRAFT—DO NOT CITE OR QUOTE
-------
1 CONTENTS (continued)
2
O
4 A.3.2.2. IARC Monographs on the Evaluation of Carcinogenic Risks to
5 Humans, Volume 63, 1995 A-13
6 A.3.2.3. Report on Carcinogens, Eleventh Edition, NTP, 2005 A-14
7
8 References A-15
9
10
11
12
13
14
15
16 LIST OF TABLES
17
18 Table A-l. Estimates of ingestion unit risk using different methods A-7
19
20 Table A-2. Ingestion, noncancer endpoints A-7
21
22 Table A-3. Estimates of cancer inhalation unit risk using different methods A-14
23
24 Table A-4. Inhalation, noncancer endpoints A-14
25
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-3 DRAFT—DO NOT CITE OR QUOTE
-------
1 A.l. ORAL INGESTION ASSESSMENTS
2 A.1.1. U.S. EPA Oral Ingestion Assessments
3 A.l.1.1. IRIS Database, U.S. EPA, 1988
4 In 1988, the U.S. Environmental Protection Agency (EPA) established a reference dose
5 (RfD) for the ingestion of tetrachloroethylene (U.S. EPA, 2005). An RfD is an estimate (with
6 uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
7 population (including sensitive subgroups) that is likely to be without an appreciable risk of
8 deleterious noncarcinogenic effects during a lifetime. For the oral RfD, EPA used the Buben and
9 O'Flaherty (1985) gavage study. A no-observed-adverse-effect level (NOAEL) of 20mg/kg-day
10 was determined, based on hepatotoxicity in mice. This value was duration adjusted and an
11 uncertainty factor of 1,000 was applied (10 for intraspecies variability, 10 for interspecies
12 variability, 10 for extrapolation from a subchronic study). EPA places medium confidence in the
13 RfD derivation because the data set lacks information on reproductive and teratology endpoints.
14 An RfD of 0.01 mg/kg-day was derived.
15
16 A.I.1.2. Health Assessment Document for Tetrachloroethylene, U.S. EPA, 1985
17 Dose response data for hepatocellular carcinomas observed in female mice in the
18 National Cancer Institute gavage study (NCI, 1977) were used to derive the unit risk. The
19 potency estimate for tetrachloroethylene was calculated using the linearized multistage model
20 and the dose metabolized and eliminated in urine. The unit risk was derived by multiplying the
21 assumed daily intake of 2 L of water contaminated with 1 [ig/L tetrachloroethylene for a person
22 (2.9 x 10"5 mg/kg-day) by the potency estimate for tetrachloroethylene (5.1 x 10"2) to derive the
23 unit risk. The upper-bound estimate of the incremental lifetime risk due to consuming water
24 contaminated with 1 |J,g/L of tetrachloroethylene was calculated to be 1.5 x 10"6.
25
26 A.1.2. Oral Ingestion Assessments Conducted by Non-EPA Agencies
27 A.l.2.1. World Health Organization, Concise International Chemical Assessment
28 Document 68, 2006
29 "...The available information on oral exposure was inadequate for derivation of a TDI by
30 the oral route. However, as tetrachloroethene is well absorbed after inhalation or ingestion and
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-4 DRAFT—DO NOT CITE OR QUOTE
-------
1 there is little evidence of first-pass metabolism, a PBPK model was used to derive a TDI. The
2 model predicted that tetrachloroethene consumed in drinking-water at a dose level of 0.047
3 mg/kg body weight per day would yield an AUC in plasma similar to that from continuous
4 exposure to tetrachloroethene at 0.2 mg/m3 in inhaled air. This oral figure was rounded to give a
5 TDI of 50 jig/kg body weight."
6 An oral cancer risk value was not derived.
7
8 A. 1.2.2. California Environmental Protection Agency, 2001, Draft Public Health Goal for
9 Tetrachloroethylene in Drinking Water
10 The California Environmental Protection Agency (Cal EPA) developed a Public Health
11 Goal (PHG) for tetrachloroethylene in drinking water on the basis of hepatocellular carcinomas
12 observed in male and female mice orally exposed to tetrachloroethylene (Cal EPA, 2001). PHGs
13 are based solely on health effects impacts and are set at levels that do not pose any significant
14 health risk, as determined by the California Office of Environmental Health Hazard Assessment.
15 For water-derived inhalation exposures, estimates were derived from studies showing
16 hepatocellular adenoma or carcinoma in male mice and mononuclear cell leukemia in both male
17 and female rats exposed by inhalation to tetrachloroethylene (NTP, 1986). The pharmacokinetic
18 model described by Bogen et al. (1987) was used to estimate the "effective" dose for use in
19 quantitative calculations. The Bogen et al. study was chosen over other studies (Bois et al.,
20 1990; Chen and Blancato, 1987) because it provided dose estimates for mice and rats exposed
21 orally and by inhalation.
22 Tetrachloroethylene was treated as a directly acting genotoxic carcinogen, and a linear
23 low-dose extrapolation model was used. The PHG established by Cal EPA is 0.056 ng/L. This
24 value corresponds to a unit risk estimate of 1.3 x 10"5 (fig/L)"1. This health-protective
25 concentration includes an estimate of inhalation exposure from showering in tetrachloroethylene-
26 contaminated water, flushing toilets, and other household activities involving tap water.
27 Chronic toxicity, excluding cancer, was evaluated on the basis of neurobehavioral
28 endpoints (delayed reaction time) observed in epidemiological studies of exposed humans.
29 These studies evaluated persons who were exposed to inhaled tetrachloroethylene. Cal EPA
30 concluded that no single study was sufficiently reliable to be used as the primary basis for a
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-5 DRAFT—DO NOT CITE OR QUOTE
-------
1 health-protective standard; both the Altmann et al. (1995) and the Spinatonda et al. (1997)
2 studies were quite small (14 and 35 subjects, respectively) and the Ferroni et al. (1992) study
3 lacked details. Therefore, the geometric mean from the three studies was used to derive an
4 estimated health-protective concentration in drinking water.
5 In calculating the mean, each study provided a lowest-observed-adverse-effect level
6 (LOAEL) value, and a study-specific uncertainty factor was used (10 to account for the use of a
7 LOAEL and 10 or 3 to account for potentially sensitive human subpopulations). A factor of 3%
8 was applied for the relative source contribution because drinking water supplies only 3% of the
9 total tetrachloroethylene exposure, and a water intake of 6.31 L/day was the calculated
10 equivalent drinking water ingestion rate that would supply the total tetrachloroethylene dose
11 from inhalation via showering and direct ingestion. The geometric mean of these safe
12 concentrations calculated from the three studies is 1.1 * 10 "2 mg/L (11 ng/L). The investigators
13 concluded that this is the health-protective drinking water concentration for noncarcinogenic
14 effects. With an assumption of 100% absorption from drinking water and an intake of 6.31
15 L/day, the equivalent dose corresponding to 11 |J,g/L is 1 |j,g/kg-day. This can be used to
16 compare the California safe limits to the RfD of other organizations.
17
18 A. 1.2.3. Agency for Toxic Substances and Disease Registry (ATSDR) Toxicological Profile
19 for Tetrachloroethylene (Update), September 1997
20 ATSDR has established a minimal risk level (MRL) for the acute ingestion of
21 tetrachloroethylene. MRLs are estimates of the daily human exposure to a hazardous substance
22 that is likely to be without appreciable risk of adverse noncancer health effects over a specified
23 duration of exposure. These values are based only on noncancer effects and are generally based
24 on the most sensitive endpoint considered to be of relevance to humans. The acute oral MRL
25 was derived from studies that showed hyperactivity in 60-day-old male mice that were treated
26 with tetrachloroethylene for 7 days beginning at 10 days of age (Fredriksson et al., 1993). The
27 MRL is based on a LOAEL (5 mg/kg-day) that was adjusted by an uncertainty factor of 100 to
28 account for the use of the LOAEL (10) and extrapolation from animals to humans (10). For
29 tetrachloroethylene, the acute oral MRL is 0.05 mg/kg-day.
30
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-6 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
A.1.3. Summary of Ingestion Risk Estimates
The following tables summarize the quantitative risk estimates that have been developed
by EPA and other agencies. Table A-l shows the cancer risk values. Table A-2 depicts the risk
estimates developed for noncarcinogenic endpoints.
Table A-l. Estimates of ingestion unit risk using different methods
Agency guideline
Public Health Goal (Cal EPA,
2000), based on metabolized
dose
Dose metabolized (U.S. EPA,
1985)
Unit risk value
(Hg/L)1
1.3 x 1Q-5
1.5 x 1Q-6
Studies used
NCI (1977),
NTP (1986)
NCI (1977)
Critical target effect(s)
Liver adenomas and carcinomas
in male and female mice,
mononuclear cell leukemia in
male and female rats
Liver carcinomas in female mice
9
10
11
12
13
14
Table A-2. Ingestion, noncancer endpoints
Agency guideline
Cal EPA (1999)
Minimal risk level
(ATSDR, 1997)
Reference dose (U.S.
EPA, 1988)
Standard
1 1 |o,g/L (water
concentration)
0.001 mg/kg-day
equivalent dose
0.05 mg/kg-day
0.01 mg/kg-day
Studies used
Altmannetal. (1995),
Spinatondaetal. (1997),
and Ferroni et al. (1992)
Fredrikssonetal. (1993)
Buben and O' Flaherty
(1985)
Critical target effect(s)
Delayed reaction times in
humans.
Hyperactivity in male
mice
Liver toxicity in mice
15
16
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-7 DRAFT—DO NOT CITE OR QUOTE
-------
1 A.2. INHALATION ASSESSMENTS
2 A.2.1. U.S. EPA Inhalation Assessments
3 A.2.1.1. Cleaner Technologies Substitutes Assessment: Professional Fabricare Processes,
4 Office of Pollution Prevention and Toxics, U.S. EPA, 1998
5 The Office of Pollution Prevention and Toxics (OPPT) developed the Cleaner
6 Technologies Substitutes Assessment (CTSA) to provide comparative cost, risk, and
7 performance information on professional fabricare processes. As part of this assessment, the
8 health risks associated with the use of tetrachloroethylene in dry cleaning establishments was
9 evaluated. For carcinogenic effects, the CTSA used human-equivalent metabolized doses using
10 mouse and rat tumor data from the National Toxicology Program (NTP) (1986) study. The
11 approach used was similar to that used by EPA (U.S. EPA, 1986), but the mouse carcinoma-only
12 data set was omitted from the assessment to avoid double-counting of animals with adenomas
13 and carcinomas. The analyses are based on taking the geometric mean of the unit risk from four
14 data sets that evaluated the incidence of male and female mouse liver adenomas and/or
15 carcinomas and male and female rat mononuclear cell leukemia. Using a linear-at-low-doses
16 approach, the unit risk was estimated to be 7.1 x 10"7 per |J,g/m3 of tetrachloroethylene in air.
17 The CTSA report states that the unit risk should not be used for lifetime average daily exposures
18 greater than 1.4 x 104 |j,g/m3.
19 Noncarcinogenic effects were also evaluated in the CTSA report. A provisional RfC was
20 derived on the basis of mild renal tubule damage seen in a cross-sectional occupational study
21 (Franchini et al., 1983). The average level of tetrachloroethylene exposure was equivalent to 10
22 mg/m3, and this value was used as the LOAEL. The LOAEL was adjusted to account for
23 duration of exposure, and an uncertainty factor of 10 was applied to account for the use of a
24 LOAEL. An uncertainty factor to account for sensitive individuals was not applied, because the
25 derived RfC was to be used in the CTSA screening to evaluate occupational populations. The
26 provisional RfC was established at 0.17 mg/m3.
27
28 A.2.1.2. Addendum to the Health Assessment Document for Tetrachloroethylene, U.S. EPA,
29 1986
30 This assessment was conducted to reevaluate tetrachloroethylene carcinogenicity on the
31 basis of the released NTP (1986) inhalation animal bioassay. On the basis of the evidence of
32 carcinogenicity in rats and mice, together with the inconclusive epidemiologic evidence,
33 tetrachloroethylene was recategorized as a Group B2 probable human carcinogen.
34 A new inhalation unit risk value was derived using the NTP (1986) inhalation study. The
35 NTP bioassay doses for rats and mice were converted to metabolized doses using the previously
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-8 DRAFT—DO NOT CITE OR QUOTE
-------
1 established dose-metabolism relationship (U.S. EPA, 1985). A linearized multistage model was
2 used for the low-dose extrapolation. Six different data sets from the NTP study were used to
3 derive unit risk estimates. These data included endpoints on leukemia in male and female rats,
4 liver carcinoma in male and female mice, and liver carcinomas and adenomas in male and female
5 mice. The revised upper-bound estimate of the incremental cancer risk due to lifetime exposure
6 to 1 |J,g/m3 of tetrachloroethylene in air was determined to range from 2.9 x 10"7 to 9.5 x 10"7.
7 This range includes the value determined using the NCI gavage study (NCI, 1977). The unit risk
8 was stated to be applicable only for low-level exposures where the relationship between ambient
9 air concentrations and metabolized dose is linear.
10
11 A.2.1.3. Health Assessment Document for Tetrachloroethylene, U.S. EPA, 1985
12 Tetrachloroethylene was categorized as a Group C possible human carcinogen on the
13 basis of limited evidence of carcinogenicity in animals and inconclusive epidemiologic data.
14 Dose response data for hepatocellular carcinomas observed in female mice in the NCI gavage
15 study (NCI, 1977) were used to derive the unit risk. The potency estimate for
16 tetrachloroethylene was calculated using the linearized multistage model and the dose
17 metabolized and eliminated in the urine. Urinary metabolites were considered to account for
18 80% of total metabolites, as in Buben and O'Flaherty (1985). Unit risk was then calculated
19 using human body burden data from Bolanowska and Golacka (1972). This study provided
20 information on the relationship between the air concentration and the amount metabolized in
21 urinary excretion in human subjects. The amount metabolized was assumed to be proportional to
22 the air concentration and the duration of exposure. The upper-bound estimate of the incremental
23 cancer risk due to 1 ng/m3 of tetrachloroethylene in air was determined to be 4.8 x 10"7.
24
25 A.2.2. Inhalation Assessments Conducted by Non-U.S. EPA Agencies
26 A.2.2.1. World Health Organization, Concise International Chemical Assessment
27 Document 68, 2006
28 "In occupationally exposed cohorts, the most consistent adverse finding was
29 neurotoxicity; therefore, the most informative study on neurotoxic effects in exposed workers
30 was used to derive a TC. The mean exposure level (83 mg/m3) was taken as a LOAEC. This
31 was converted to an equivalent concentration for continuous exposure (20 mg/m3), and two
32 uncertainty factors of 10 were applied (one to account for interindividual differences, the other
33 because the selected concentration was a LOAEC rather than a NOAEC), to derive a TC of 0.2
34 mg/m3. For comparative purposes, a similar approach was used for studies reporting
35 nephrotoxicity. The most informative study yielded a mean occupational exposure of 100
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-9 DRAFT—DO NOT CITE OR QUOTE
-------
1 mg/m3, which generated a TC of 0.24 mg/m3, a value in good agreement with the TC protective
2 against neurotoxic effects. Available data indicate that liver toxicity would occur only at
3 exposures higher than those that affect the CNS and kidney. A TC for spontaneous abortions
4 was not derived. However, the TC of 0.2 mg/m3 is more than 3 orders of magnitude lower than
5 the exposure concentration that induced mild adverse effects in laboratory animals, and so it was
6 considered to be protective against reproductive toxicity in humans."
7 "Tetrachloroethene has induced several types of tumour in rats and mice. Currently,
8 there is no convincing evidence that these tumours arise via modes of action that operate only in
9 rodents, and hence their relevance to humans cannot be dismissed. Therefore, a BMC approach
10 was used, and a BMC and its lower confidence limit (BMCL) were calculated for each animal
11 tumour. Of the tumours observed in experimental animals, hepatocellular adenomas and
12 carcinomas in male mice yield highest predicted risks. The TC derived above, 0.2 mg/m3,
13 corresponds to a cumulative lifetime risk of 0.4 x 10~3 when a linear extrapolation is applied to
14 the BMCio as the point of departure."
15
16 A.2.2.2. California Environmental Protection Agency, 2002
17 The California EPA Air Toxics Hot Spots program has derived an inhalation unit risk
18 value for tetrachloroethylene (Cal EPA, 2002). The value was determined from data on
19 hepatocellular adenomas and carcinomas in male mice reported in the NTP (1986) bioassay
20 study. Two pharmacokinetic models were used to estimate the human inhaled concentrations
21 equivalent to the bioassay concentrations. These two models were described only as (1) a
22 steady-state model and (2) a physiologically based pharmacokinetic (PBPK) model. An
23 assumption that 18.5% of the applied dose is metabolized in humans was incorporated. The
24 cancer potency values expressed in terms of human dose rates and derived using the two
25 different models and the rat and mouse studies ranged from 0.0025 to 0.093 per mg/kg-day.
26 Considering the quality of the cancer bioassay s and the uncertainty in human metabolism, Cal
27 EPA decided that the best value for the inhalation unit risk was 5.9 x 10"6 per |ig/m3.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-10 DRAFT—DO NOT CITE OR QUOTE
-------
1 A.2.2.2. Massachusetts' Derivation of Inhalation Unit Risk for Tetrachloroethylene,
2 Massachusetts Department of Environmental Protection, 1998
3 The Massachusetts Department of Environmental Protection (MA DEP) classifies
4 tetrachloroethylene as a Group B2 carcinogen with suggestive evidence for mutagenicity. The
5 unit risk was calculated on the basis of male and female liver tumors found in mice in the NCI
6 gavage study (NCI, 1977). The dose calculations used were similar to those used by EPA (U.S.
7 EPA, 1985), with two differences. MA DEP adjusted the lifetime average dose, which is based
8 on urinary metabolites, to a dose of total metabolites. This adjustment was made on the
9 assumption that the urinary metabolites are 80% of the total metabolism. To convert the
10 carcinogenic potency value to an inhalation exposure, MA DEP assumed that the metabolized
11 dose is equal to 70% of the inhaled dose. This differs from the 1985 EPA assessment, where
12 0.66% of the total inhaled dose in humans is assumed to be metabolized to urinary metabolites.
13 MA DEP also calculated a unit risk using the NTP inhalation study, but did not consider this to
14 provide a reasonable quantitative estimate due to uncertainty in the calculations of the
15 metabolized dose. Using the NCI study, 5.5 x 10"5 per ng/m3 is recommended as the unit risk.
16
17 A.2.2.3. ATSDR Toxicological Profile for Tetrachloroethylene (Update), September 1997
18 ATSDR has promulgated both acute and chronic MRLs for the inhalation of
19 tetrachloroethylene. The acute inhalation MRL was derived from studies where male volunteers
20 were exposed to 50 ppm tetrachloroethylene for 4 hrs/day for 4 days. The volunteers showed
21 increased pattern reversal visually evoked potential (VEP) latencies and deficits for vigilance and
22 eye-hand coordination (Altmann et al., 1992). Deficits were not seen at 10 ppm, and this value
23 was used as the NOAEL. This value was duration adjusted to extrapolate from intermittent
24 exposure, and an uncertainty factor of 10 was used to account for human variability. The acute
25 inhalation MRL was established at 0.2 ppm (1.36 mg/m3).
26 The chronic duration MRL for the inhalation of tetrachloroethylene was based on a study
27 that showed increased reaction times in neurobehavioral tests given to female workers exposed
28 to tetrachloroethylene in dry cleaning shops (Ferroni et al., 1992). Air exposures averaged 15
29 ppm tetrachloroethylene for an average of 10.1 years. The LOAEL in this study was 15 ppm.
30 This value was adjusted from an occupational exposure to a continuous exposure; an uncertainty
31 factor of 10 was used to account for the use of a LOAEL, and an additional factor of 10 was used
32 to account for human variability. The chronic inhalation MRL was established at 0.04 ppm (0.27
33 mg/m3).
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-l 1 DRAFT—DO NOT CITE OR QUOTE
-------
1 A.2.2.4. Tetrachloroethane-Ambient Air Criteria Document, New York State Department of
2 Health, 1997
3 The New York State Department of Health agrees with the International Agency for
4 Research on Cancer (IARC) classification of tetrachloroethylene as a Group 2A (known animal)
5 carcinogen. A linearized multistage model was applied to metabolized dose data for liver tumors
6 in mice and mononuclear cell leukemia in rats (NTP, 1986). Estimates of the metabolized dose
7 were based on predictions of a physiologically based pharmacokinetic (PBPK) model (mice) and
8 on experimental data on the production of urinary metabolites (mice and rats). For mononuclear
9 cell leukemia in rats, estimates based on air concentrations were also derived. ED 10 procedures
10 were also used to calculate unit risk values. Due to uncertainty regarding which method provides
11 a better estimate, the central tendency of all these estimates (linearized multistage model, ED 10,
12 metabolized dose, and air concentration) was used to derive an upper-bound unit risk value. The
13 central tendency estimate for liver tumors in mice established an upper bound unit risk value of
14 0.88 x 10"6 (i^g/m3)"1; for mononuclear cell leukemia in rats, the central tendency estimate was
15 1.3 x 10"6 (i^g/m3)"1. An upper-bound risk estimate of 1 x 10"6 (jig/m3)"1 is the central tendency
16 of the mouse- and rat-based unit risk estimates and is the recommended criteria for evaluating
17 the excess human carcinogenic risk associated with chronic exposure to l|j,g/m3
18 tetrachloroethylene in ambient air.
19 New York State has determined that the strength of human evidence on the
20 noncarcinogenic effects of tetrachloroethylene exposure support the use of human data for
21 determining an ambient air criteria for noncarcinogenic effects. A weight-of-evidence approach
22 was used, and multiple endpoints and epidemiologic studies were evaluated. Endpoints used in
23 the derivation of the ambient air criterion included evidence of central nervous system (motor
24 and cognitive effects) (Seeber, 1989), kidney (Mutti et al., 1992), and liver dysfunction (Gennari
25 et al., 1992). The lack of reproductive and developmental studies was identified as a significant
26 data gap because epidemiologic studies did not provide sufficient exposure data for criteria
27 evaluation.
28 Lowest-observed-effect level (LOEL) data were provided in the epidemiologic studies
29 listed above. LOELs were duration adjusted to account for continuous exposure using EPA
30 inhalation guidelines (U.S. EPA, 1994). For adult criteria, an uncertainty factor of 100 was
31 applied to each duration-adjusted LOEL (10 for variation in sensitivity among humans and 10 for
32 the use of a LOEL from a subchronic study). For criteria protective of children, the appropriate
33 scaling and uncertainty factors were used. Child-adjusted LOELs were derived using physical
34 and physiological data for children. An uncertainty factor of 100 was applied to the child-
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-12 DRAFT—DO NOT CITE OR QUOTE
-------
1 adjusted LOELs (10 for variation in sensitivity among humans, 3 for the use of a LOEL, and 3
2 for concerns about the increased sensitivity of children to tetrachloroethylene toxicity).
3 Listed below are the results of the safe ambient air level derivations:
4
5 Effects Adults Children
6 Central nervous system (Seeber, 1989) 0.30 mg/m3 0.12mg/m3
7 Kidney (Mutti et al., 1992) 0.36 mg/m3 0.14 mg/m3
8 Liver (Gennari et al., 1992) 0.28 mg/m3 0.10 mg/m3
9
10 New York State estimated that an ambient air criterion of 0.1 mg/m3 would provide the
11 general population, including sensitive subpopulations of infants, children, the infirm and
12 elderly, a sufficient margin of exposure over the air levels of tetrachloroethylene associated with
13 noncarcinogenic effects in humans and animals. The ambient air criterion was established at 0.1
14 mg/m3.
15
16 A.2.2.5. Priority Substances List Assessment Report: Tetrachloroethylene, Canada Health
17 and Welfare Agency, 1993
18 Tetrachloroethylene has been classified in Group 3 (possibly carcinogenic to humans) of
19 the classification scheme developed for use in the derivation for the guidelines for Canadian
20 drinking water quality. A tolerable daily intake (TDI) was derived using data from the NTP
21 (1986) study. It was assumed that 100% of the inhaled tetrachloroethylene was retained in the
22 mice. A LOAEL of 100 ppm for reduced survival and hepatotoxic effects in male mice and lung
23 congestion and nephrotoxic effects in male and female mice was used. An uncertainty factor of
24 5,000 was applied to account for intraspecies variation (10), use of a LOAEL (10), interspecies
25 variation (10), and limited evidence of carcinogenicity (5). A TDI of 34 |J,g/kg bw/day was
26 derived. Using standardized conversion assumptions (EPA), this value is equivalent to 0.018
27 ppm (0.12 mg/m3).
28
29 A.2.3. Summary of Inhalation Risk Estimates
30 The following tables summarize the quantitative risk estimates that have been developed
31 by EPA and other agencies. Table A-3 shows the cancer risk values. Table A-4 depicts the risk
32 estimates developed for noncarcinogenic endpoints.
33
34 A.3. QUALITATIVE RISK ASSESSMENTS
3 5 This section contains a brief review of documents that included only a qualitative
36 assessment of tetrachloroethylene toxicity and risk.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-13 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
Table A-3. Estimates of cancer inhalation unit risk using different methods
Dose metric surrogate
(agency)
Dose metabolized (Cal
EPA, 2002)
Dose metabolized (U.S.
EPA, 1998)
Total metabolized dose
(MA DEP, 1998)
Dose metabolized and
administered (NYS DOH,
1997)
Dose metabolized (U.S.
EPA, 1986)
Dose metabolized (U.S.
EPA, 1985
Unit risk value
(Hg/m3)-1
5.9 x 1Q-6
7.1 x ID'7
5.5 x 1Q-5
1 x 1Q-6
2.9 x 10'7to9.5
x 10'7
4.8 x 1Q-7
Studies used
NTP (1986)
NTP (1986)
NCI (1977)
NTP (1986)
NTP (1986)
NCI (1977)
Critical target effect(s)
Liver adenomas and carcinomas in male
mice
Liver adenomas and carcinomas in male
and female mice, mononuclear cell
leukemia in male and female rats
Liver tumors in male and female mice
Liver tumors in male and female mice,
mononuclear cell leukemia in male and
female rats
Liver adenomas in male and female
mice, combined liver adenomas and
carcinomas in male and female mice,
mononuclear cell leukemia in male and
female rats
Liver carcinomas in female mice
4
5
6
7
Table A-4. Inhalation, noncancer endpoints
Agency guideline
Provisional RfC (U.S.
EPA, 1998)
Ambient air criterion
NYS DOH, 1997)
Acute inhalation
minimal risk level
(ATSDR, 1997)
Chronic inhalation
minimal risk level
(ATSDR, 1997)
Tolerable daily intake
(CHWA, 1993)
Standard
0.025 ppm(0. 17 mg/m3)
0.015 ppm(0.1mg/m3)
0.2ppm(1.36mg/m3)
0.04 ppm (0.27 mg/m3)
0.0 1 8 ppm (0.12 mg/m3)
Studies used
Franchini et al.
(1983)
Seeber(1989),
Muttietal. (1992),
Gennari et al.
(1992)
Altmann et al.
(1992)
Ferroni et al.
(1992)
NTP (1986)
Critical target effect(s)
Renal tubule damage in
workers
Motor and cognitive effects,
kidney dysfunction and liver
dysfunction in workers
Neurological effects (VEP,
vigilance, and eye-hand
coordination) in males
Increased reaction times in
female workers
Reduced survival and
hepatotoxic effects in male
mice, lung congestion and
nephrotoxic effects in male
and female mice
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-14 DRAFT—DO NOT CITE OR QUOTE
-------
1 A.3.1. U.S. EPA Qualitative Risk Assessments
2 A.3.1.1. Response to Issues and Data Submission on the Carcinogenicity of
3 Tetrachloroethylene, U.S. EPA, 1991
4 This document discusses issues relating to the classification of tetrachloroethylene as a
5 B2 carcinogen. Lengthy deliberation is given to the specific mechanisms of action that may
6 explain all the tumor endpoints observed after exposure to tetrachloroethylene. In conclusion,
7 EPA stands behind the B2 classification and concludes that sufficient evidence of cancer in
8 animals does exist.
9
10 A.3.2. Qualitative Risk Assessments Conducted by Non-U.S. EPA Agencies
11 A.3.2.1. Organization for Economic Cooperation and Development (OECD), Screening
12 Information Data Set (SIDS) Initial Assessment Report: Comprehensive Risk Assessment
13 Report for Tetrachloroethylene, 1996
14 The current European Union classification for tetrachloroethylene is Carcinogen
15 Category 3. Category 3 indicates "a substance which causes concern for man owing to possible
16 carcinogenic effect but in respect of which the available information is not adequate for making
17 a satisfactory assessment. There is some evidence from appropriate animal studies, but this is
18 insufficient to place the substance in category 2. "
19 In the summary of carcinogenicity, the report (OECD, 1996) concludes that the liver
20 tumors found in mice and the kidney tumors found in rats following repeated inhalation exposure
21 are almost undoubtedly not of significance in relation to human health. This is based on believed
22 differences in metabolic pathways and mechanisms of action. OECD does not believe that
23 peroxisome proliferation in mice is relevant to human cancer. Similarly, it believes that the
24 human renal beta lyase activity in humans is negligible compared to that in rats. In evaluating
25 human carcinogenicity, OECD determined that the epidemiological studies do not show evidence
26 supporting an increased risk of carcinogenicity in humans.
27
28 A.3.2.2. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 63,
29 1995
30 IARC determined that tetrachloroethylene is probably carcinogenic to humans and has
31 classified tetrachloroethylene as a Group 2A carcinogen. This judgment is based on limited
32 evidence in humans and sufficient evidence of carcinogenicity in experimental animals. In
33 evaluating tetrachloroethylene, the following evidence was considered: although
34 tetrachloroethylene is known to induce peroxisome proliferation in mouse liver, poor quantitative
35 correlation was seen between peroxisome proliferation and tumor formation in the liver after
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-l 5 DRAFT—DO NOT CITE OR QUOTE
-------
1 administration of tetrachloroethylene by inhalation; the spectrum of mutations in proto-
2 oncogenes in liver tumors from mice treated with tetrachloroethylene is different from that in
3 liver tumors from mice treated with trichloroethylene; tetrachloroethylene induced leukemia in
4 rats; and several epidemiological studies showed elevated risks for esophageal cancer, non-
5 Hodgkin's lymphoma and cervical cancer.
6 Evidence of cancer in animal studies is supported by studies that included both oral and
7 inhalation exposures. Cancer endpoints included increases in hepatocellular carcinoma in male
8 and female rats after oral administration of tetrachloroethylene (NCI, 1977), increases in
9 hepatocellular adenoma and carcinoma in male and female mice and increases in mononuclear-
10 cell leukemia in male and female rats after inhalation exposure (NTP, 1986). IARC does not
11 point to a single epidemiological study as being critical, but rather summarizes many studies that
12 support a relationship between cancer and tetrachloroethylene exposures. IARC relies on the
13 consistent positive associations between human exposures to tetrachloroethylene and the risks
14 for esophageal cancer, non-Hodgkin's lymphoma, and cervical cancer.
15
16 A.3.2.3. Report on Carcinogens, Eleventh Edition, NTP, 2005
17 The NTP lists carcinogenic substances in one of two categories: (1) known to be a human
18 carcinogen and (2) reasonably anticipated to be a human carcinogen. They present a brief two-
19 page summary of the evidence for their classification. They classified tetrachloroethylene in
20 Category 2, "reasonably anticipated to be a human carcinogen." It was first listed in the 5th
21 Annual Report on Carcinogens (1989). They based their classification on sufficient evidence of
22 carcinogenicity in experimental animals and limited evidence in humans.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-16 DRAFT—DO NOT CITE OR QUOTE
-------
1 REFERENCES FOR APPENDIX A
2
3 Altmann, L; Weigand, H; Bottger, A; et al. (1992) Neurobehavioral and neurophysiological outcomes of acute
4 repeated perchloroethylene exposure. Applied Psychology: An International Review 41:269-279.
5 Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
6 low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
7 ATSDR (Agency for Toxic Substance and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
8 (update). Prepared for Sciences, International under subcontract to Research Triangle Institute, ATSDR, Atlanta,
9 GA.
10 Bogen, KT; Hall, LC; McKone, TE; et al. (1987) Health risk assessment of tetrachloroethylene (PCE) in California
11 drinking water. DE87-013493. Available from National Technical Information Services, Springfield, VA.
12 Bois, FY; Zeise, L; Tozer, TN. (1990) Precision and sensitivity of pharmacokinetic models for cancer risk
13 assessment: tetrachloroethylene in mice, rats, and humans. Toxicol Appl Pharmacol 102:300-315.
14 Bolanowska, W; Golacka, J. (1972) Absorption and elimination of tetrachloroethylene in humans under
15 experimental conditions (English translation). Medycyna Pracy 23:109-119.
16 Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
17 and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
18 Cal EPA (California Environmental Protection Agency). (2001) Public health goal for tetrachloroethylene in
19 drinking water. Office of Environmental Health Hazard Assessment. Available from online at
20 http://oehha.ca.gov/water/shg/8310 IPHG.htm.
21 Cal EPA (California Environmental Protection Agency). (2002) Technical support document for describing
22 available cancer potency factors. Air Toxics Hot Spots Program Risk Assessment Guidelines Part II. Available
23 online at http://www.oehha.ca.gov/air/hot_spots/pdf/TSDNov2002.pdf.
24 Chen, C; Blancato, J. (1987) Role of pharmacokinetic modeling in risk assessment: perchloroethylene as an
25 example. In: Safe Drinking Water Committee SoPNRC (ed) Pharmacokinetics in risk assessment. Washington, DC:
26 National Academy Press; pp. 367-388.
27 CHWA (Canada Health and Welfare Agency). (1993) Tetrachloroethylene: priority substances list report. Cat. No.
28 En 40-215/28E, Health-Related Sections: Canada Communication Group: Ottawa.
29 Ferroni, C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
30 perchloroethylene. Neurotoxicology 13:243-247.
31 Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
32 solvents. Int Arch Occup Environ Health 52:1-9.
33 Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
34 tetrachloroethylene as neonates. Toxicol Lett 66:13-19.
3 5 Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
36 tetrachloroethylene. Am JIndMed 21:661-671.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-17 DRAFT—DO NOT CITE OR QUOTE
-------
1 IARC (International Agency for Research on Cancer). (1985) Tetrachloroethylene. In: IARC monographs on the
2 evaluation of carcinogenic risks to humans; vol. 63: dry cleaning, some chlorinated solvents and others. Lyon,
3 France.
4 MA DEP (Massachusetts Department of Environmental Protection). (1998) Inhalation unit risk for
5 perchloroethylene. Memorandum from Marion Harnios, ORS, to Carol West, December 1988.
6 Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-
7 cleaners. Lancet 340:189-193.
8 NCI (National Cancer Institute). (1977) Bioassay of tetrachloroethylene for possible carcinogenicity. DHEW Pub.
9 (NIH) 77-813. U.S. Department of Health, Education and Welfare, Public Health Service.
10 NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
11 (perchloroethylene) (CAS No. 127-18-4) inF344/N rats andB6C3Fl mice. 311,1-190. U.S. Department of Health
12 and Human Services. Technical Report Series.
13 NTP (National Toxicology Program). (2005) Report on carcinogens, eleventh edition. U.S. Department of Health
14 and Human Services, Public Health Service, Research Triangle Park, NC. Available online at http://ntp-
15 server.niehs.nih.gov.
16 NYS DOH (New York State Department of Health). (1997) Tetrachloroethene ambient air criteria document. Final
17 Report. Albany, NY.
18 OECD (Organization for Economic Co-operation and Development). (1996) Comprehensive risk assessment report
19 on tetrachloroethylene, SIDS initial assessment report. OECD Secreteriat. Paris.
20 Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
21 11:579-583.
22 Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
23 of subjects chronically exposed to organic solvents (II)]. G Ital Med Lav Ergon 19:85-88.
24 U.S. EPA (Environmental Protection Agency). (1985) Health assessment document for tetrachloroethylene
25 (perchloroethylene). National Center for Environmental Assessment, Washington, DC; EPA/600/8-82/005F.
26 Available from: National Technical Information Service, Springfield, VA; PB-85-249696/AS.
27 U.S. EPA (Environmental Protection Agency). (1986) Addendum to the health assessment document for
28 tetrachloroethylene (perchloroethylene) [review draft]. National Center for Environmental Assessment, Washington,
29 DC; EPA/600/8-82/05FA.
30 U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
31 carcinogenicity of tetrachloroethylene (perchloroethylene). Office of health and Environmental Assessment,
32 Washington, DC; EPA/600/6-91/002F. Pages 1-73.
33 U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
34 and application of inhalation dosimetry. Environmental Criteria and Assessment Office, Office of Health and
3 5 Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F.
36 U.S. EPA (Environmental Protection Agency). (1998) Cleaner technologies substitutes assessment: professional
3 7 fabricare processes. Office of Pollution Prevention and Toxics, Washington, DC; EPA 744-B-OO1.
38 U.S. EPA (Environmental Protection Agency). (2005) Integrated Risk Information System (tetrachloroethylene file).
39 National Center for Environmental Assessment, Washington, DC. Available online at http://www.epa.gov/iris.
This document is a draft for review purposes only and does not constitute Agency policy
06/06/08 A-l 8 DRAFT—DO NOT CITE OR QUOTE
------- |