EPA/635/R-08/011F
February 2012
TOXICOLOGICAL REVIEW
OF
Tetrachloroethylene
(Perchloroethylene)
(CAS No. 127-18-4)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
February 2012
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
11
-------
GUIDE TO READERS OF THIS DOCUMENT
For ease of reading, it is recommended that the Executive Summary, Section 1, and
Section 6 be read prior to Sections 2-5.
Section 1 is the standard introduction to an IRIS Toxicological Review, describing the
purpose of the assessment and the guidelines used in its development.
Section 2 summarizes information about tetrachloroethylene uses, occurrence and
exposure.
Section 3 describes the toxicokinetics and physiologically based pharmacokinetic
(PBPK) modeling of tetrachloroethylene and metabolites.
Section 4 is the hazard characterization of tetrachloroethylene. This section discusses
tetrachloroethylene toxicity on an organ-specific basis. For each of the major organ systems,
human effects are presented first, followed by effects in animals and in in vitro systems. Cancer
and noncancer toxicity and mode of action (MOA) are also included in the discussions. The
order of presentation is as follows: neurotoxicity (refer to Section 4.1); kidney and bladder
toxicity and cancer (refer to Section 4.2); liver toxicity and cancer (refer to Section 4.3);
esophageal cancer (refer to Section 4.4); lung and respiratory cancer (refer to Section 4.5);
immunotoxicity, hematologic toxicity, and cancers of the immune system (refer to Section 4.6);
developmental and reproductive toxicity, and reproductive cancers (refer to Section 4.7);
genotoxicity (refer to Section 4.8); and susceptible populations (refer to Section 4.9). Section
4.10 provides a summary of the hazard identification.
Section 5 is the dose-response assessment of tetrachloroethylene. Section 5.1 and 5.2
describes the dose-response analyses for noncancer effects and Section 5.3 describes the dose-
response analyses for cancer. Appendix D provides details of the cancer dose-response modeling
analyses.
Section 6 is the summary of the major conclusions in the characterization of
tetrachloroethylene hazard and dose response.
Appendix A contains the summary of EPA's response to major external peer review and
public comments.
Appendix B documents the essential design features, exposure assessment approaches,
statistical analyses (including assessment of exposure- or concentration-response), and potential
in
-------
sources of confounding and bias for epidemiologic studies on cancer and tetrachloroethylene.
This analysis supports the discussion of site-specific cancer observations in Sections 4.2-4.7.
Appendix C presents a comparative quantitative analysis of the carcinogenicity of trichloroacetic
acid (including that predicted using PBPK modeling to be produced from tetrachloroethylene)
with the carcinogenicity of tetrachloroethylene.
IV
-------
CONTENTS
TOXICOLOGICAL REVIEW for TETRACHLOROETHYLENE
(PERCHLOROETHYLENE) (CAS No. 127-18-4)
GUIDE TO READERS OF THIS DOCUMENT iii
LIST OF TABLES xiv
LIST OF FIGURES xx
LIST OF ABBREVIATIONS AND ACRONYMS xxiii
FOREWORD xxvi
CONTRIBUTORS AND REVIEWERS xxvii
EXECUTIVE SUMMARY xxxii
1. INTRODUCTION 1-1
2. BACKGROUND 2-1
2.1. USES AND PHYSICAL/CHEMICAL PROPERTIES 2-1
2.2. OCCURRENCE AND EXPOSURE 2-1
2.2.1. Air 2-1
2.2.2. Water 2-4
2.2.3. Food 2-5
2.2.4. Soil 2-6
2.2.5. Breast Milk 2-6
2.2.6. Direct Ingestion 2-7
3. TOXICOKINETICS 3-1
3.1. ABSORPTION 3-1
3.1.1. Inhalation 3-1
3.1.2. Oral 3-2
3.1.3. Dermal 3-2
3.2. DISTRIBUTION AND BODY BURDEN 3-2
3.3. METABOLISM 3-4
3.3.1. Introduction 3-4
3.3.2. Extent of Metabolism 3-6
3.3.3. Pathways of Metabolism 3-7
3.3.3.1. Cytochrome P450-Dependent Oxidation 3-7
3.3.3.2. Glutathione (GSH) Conjugation Pathway 3-10
3.3.3.3. Relative Roles of the Cytochrome P450 (CYP) and
Glutathione (GSH) Pathways 3-15
3.3.4. Susceptibility 3-16
v
-------
3.3.5. Comparison of Tetrachloroethylene Metabolism with Trichloroethylene
Metabolism 3-17
3.3.5.1. Extent of Metabolism 3-17
3.3.5.2. Cytochrome P450 (CYP)-Mediated Oxidation 3-18
3.3.5.3. Glutathione (GSH) Conjugation Pathway 3-19
3.3.5.4. Summary 3-19
3.4. EXCRETION 3-20
3.5. TOXICOKINETIC MODELING 3-22
3.5.1. Choice of Physiologically Based Pharmacokinetic (PBPK) Model for
Use in Dose-Response Modeling 3-23
3.5.1.1. Limitations of Previously Developed Physiologically Based
Pharmacokinetic (PBPK) Models 3-23
3.5.1.2. The Chiu and Ginsberg (2011) Model 3-30
3.5.2. Age and Gender-Related Differences in Tetrachloroethylene
Pharmacokinetics 3-50
3.5.3. Metabolic Interactions with Other Chemicals 3-52
4. HAZARD IDENTIFICATION 4-1
4.1.NEUROTOXICITY 4-2
4.1.1. Human Studies 4-2
4.1.1.1. Chamber Studies 4-3
4.1.1.2. Chronic Exposure Studies 4-6
4.1.1.3. Summary of Neuropsychological Effects in Low- and
Moderate-Exposure Studies 4-35
4.1.2. Animal Studies 4-41
4.1.2.1. Inhalation Studies 4-42
4.1.2.2. Oral and Intraperitoneal Studies 4-49
4.1.3. Mode of Action (MO A) for Neurotoxic Effects 4-53
4.1.3.1. Visual Function Domain 4-56
4.1.3.2. Cognitive Domain 4-56
4.1.3.3. Motor Activity Domain - Reaction Time 4-57
4.1.4. Summary of Neurotoxic Effects in Humans and Animals 4-57
4.2. KIDNEY AND BLADDER TOXICITY AND CANCER 4-60
4.2.1. Human Studies 4-60
4.2.1.1. Kidney Toxicity in Humans 4-60
4.2.1.2. Kidney Cancer in Humans 4-66
4.2.1.3. Bladder Cancer in Humans 4-83
VI
-------
4.2.2. Animal Studies 4-101
4.2.2.1. Kidney Toxicity in Animals 4-101
4.2.2.2. Kidney Cancer in Animals 4-110
4.2.3. Summary of Kidney Effects in Humans and Animals 4-114
4.2.4. Hypothesized Mode(s) of Action for Kidney Carcinogenicity 4-115
4.2.4.1. Role of Metabolism in Kidney Carcinogenicity 4-116
4.2.4.2. a2u-Globulin Accumulation 4-116
4.2.4.3. Genotoxicity 4-120
4.2.4.4. Peroxisome Proliferation 4-121
4.2.4.5. Cytotoxicity/Sustained Chronic Nephrotoxicity Not
Associated with a2u-Globulin Nephropathy 4-123
4.2.4.6. Summary 4-124
4.3. LIVER TOXICITY AND CANCER 4-125
4.3.1. Human Studies 4-125
4.3.1.1. Liver Damage 4-126
4.3.1.2. Liver Cancer 4-129
4.3.2. Animal Studies 4-141
4.3.2.1. Liver Toxicity 4-141
4.3.2.2. Liver Cancer 4-149
4.3.3. Summary of Liver Effects in Humans and Animals 4-152
4.3.4. Mode of Action for Hemangiosarcomas or Hemangiomas in Mice 4-154
4.3.5. Mode of Action for Murine Hepatocellular Tumors 4-155
4.3.5.1. Contribution of Tetrachloroethylene Metabolites to Mode of
Action and Carcinogenicity 4-156
4.3.5.2. Genotoxicity 4-159
4.3.5.3. Altered DNA Methylation 4-161
4.3.5.4. Cytotoxicity and Secondary Oxidative Stress 4-162
4.3.5.5. Peroxisome Proliferator-Activated Receptor (PPAR)
Activation Mode of Action 4-163
4.3.5.6. Mode of Action Conclusions for Hepatocellular Tumors 4-181
4.4. ESOPHAGEAL CANCER 4-184
4.4.1. Consideration of Exposure-Assessment Methodology 4-184
4.4.2. Summary of Results 4-186
4.5. LUNG AND RESPIRATORY CANCER 4-194
4.5.1. Consideration of Exposure-Assessment Methodology 4-195
4.5.2. Summary of Results 4-196
vn
-------
4.6. IMMUNOTOXICITY, HEMATOLOGIC TOXICITY, AND CANCERS OF
THE IMMUNE SYSTEM 4-206
4.6.1. Human Studies 4-206
4.6.1.1. Noncancer Immune and Hematologic Effects 4-206
4.6.1.2. Cancers of the Immune System, Including Childhood
Leukemia 4-215
4.6.2. Animal Studies 4-251
4.6.2.1. Noncancer Effects 4-251
4.6.2.2. Cancer Effects 4-254
4.6.3. Summary and Conclusions 4-265
4.6.3.1. Immunotoxicity, Hematologic Toxicity, and Cancers of the
Immune System in Humans 4-265
4.6.3.2. Immunological and Hematological Toxicity and Mononuclear
Cell Leukemias in Rodents 4-267
4.7. DEVELOPMENTAL AND REPRODUCTIVE TOXICITY AND
REPRODUCTIVE CANCERS 4-271
4.7.1. Development 4-271
4.7.1.1. Human Developmental Toxicity Data 4-271
4.7.1.2. Animal Developmental Toxicity Studies 4-306
4.7.2. Reproduction 4-311
4.7.2.1. Human Reproduction Data 4-311
4.7.2.2. Animal Reproductive Toxicity Studies 4-331
4.7.2.3. Reproductive Cancers in Humans 4-334
4.7.3. Summary of Human and Animal Developmental/Reproductive Studies .... 4-350
4.7.3.1. Summary of Human Data 4-350
4.7.3.2. Summary of Animal Data 4-354
4.7.4. Mode of Action for Developmental Effects 4-358
4.8. GENOTOXICITY 4-359
4.8.1. Tetrachloroethylene (PCE) 4-360
4.8.1.1. Mammalian Systems (Including Human Studies) 4-361
4.8.1.2. DrosophilaMelanogaster 4-373
4.8.1.3. Bacterial and Fungal Systems 4-374
4.8.1.4. Summary 4-375
4.8.2. Trichloroacetic Acid (TCA) 4-377
4.8.2.1. Mammalian Systems (Including Human Studies) 4-377
4.8.2.2. Bacterial Systems 4-382
Vlll
-------
4.8.2.3. Summary 4-383
4.8.3. Dichloroacetic Acid (DCA) 4-384
4.8.3.1. Mammalian Systems 4-384
4.8.3.2. Bacterial Systems 4-389
4.8.3.3. Summary 4-390
4.8.4. Chloral Hydrate 4-390
4.8.4.1. Mammalian Systems (Including Human Studies) 4-390
4.8.4.2. Bacterial and Fungal Systems 4-399
4.8.4.3. Summary 4-400
4.8.5. Trichloroacetyl Chloride 4-401
4.8.5.1. Bacterial Systems 4-401
4.8.6. Tetrachloroethylene (PCE) Epoxide 4-401
4.8.6.1. Bacterial Systems 4-401
4.8.7. Trichloroethanol (TCOH) 4-402
4.8.7.1. Bacterial Systems 4-402
4.8.8. S-(l,2,2-Trichlorovinyl)-Z-Cysteine (1,2-TCVC), S-Trichlorovinyl
Glutathione(TCVG),JV-Acetyl-5-(l,2,2-Trichlorovinyl)-Z-Cysteine
(NAcTCVC) 4-402
4.8.8.1. Bacterial Systems 4-402
4.8.8.2. Mammalian Systems 4-403
4.8.9. TCVC Sulfoxide 4-403
4.8.10. Synthesis and Overall Summary 4-403
4.9. SUSCEPTIBLE POPULATIONS 4-408
4.9.1. Life-Stages 4-408
4.9.1.1. Early Life-Stages 4-409
4.9.1.2. Later Life-Stages 4-427
4.9.2. Other Susceptibility Factors 4-428
4.9.2.1. Gender 4-428
4.9.2.2. Race/Ethnicity 4-430
4.9.2.3. Genetics 4-431
4.9.2.4. Preexisting Disease 4-431
4.9.2.5. Lifestyle Factors and Nutrition Status 4-432
4.9.2.6. Socioeconomic Status 4-433
4.9.2.7. Multiple Exposures and Cumulative Risks 4-433
4.9.3. Uncertainty of Database and Research Needs for Susceptible
Populations 4-435
IX
-------
4.10. SUMMARY OF HAZARD IDENTIFICATION 4-436
4.10.1. Overview of Noncancer and Cancer Hazard 4-436
4.10.2. Characterization of Noncancer Effects 4-437
4.10.2.1.Neurotoxicity 4-437
4.10.2.2. Kidney Toxicity 4-445
4.10.2.3.LiverToxicity 4-445
4.10.2.4. Immunotoxicity and hematologic toxicity 4-446
4.10.2.5. Reproductive and Developmental Toxicity 4-447
4.10.2.6. Summary of Noncancer Toxicities and Identification of
Studies for Dose-Response Analyses 4-451
4.10.3. Characterization of Cancer Hazard 4-453
4.10.4. Synthesis of Epidemiologic Studies 4-454
4.10.5. Synthesis of Rodent Cancer Bioassay Findings 4-459
4.10.5.1. Carcinogenicity Findings in Rats 4-460
4.10.5.2. Carcinogenicity Findings in Mice 4-464
4.10.5.3. Carcinogenic Mode of Action Hypotheses 4-466
5. DOSE-RESPONSE EVALUATION 5-1
5.1. INHALATION REFERENCE CONCENTRATION (RfC) 5-1
5.1.1. Choice of Principal Studies and Critical Effect 5-1
5.1.1.1. Choice of Critical Effect 5-1
5.1.1.2. Overview of Candidate Principal Studies 5-1
5.1.1.3. Selection of Principal Studies 5-2
5.1.2. Additional Analyses: Feasibility of Dose-Response Modeling 5-16
5.1.3. Reference Concentration (RfC) Derivation, Including Application of
Uncertainty Factors 5-17
5.1.4. Dose-Response Analyses for Comparison of Noncancer Effects Other
Than Critical Effects in Neurotoxicity 5-20
5.1.4.1. Sample Reference Concentrations (RfCs) for Kidney Toxicity 5-21
5.1.4.2. Sample Reference Concentrations (RfCs) for Liver Toxicity 5-22
5.1.4.3. Sample Reference Concentrations (RfCs) for Immunotoxicity
and Hematologic Toxicity 5-23
5.1.4.4. Sample Reference Concentrations (RfCs) for Reproductive
and/or Developmental Toxicity 5-24
5.1.4.5. Summary of Sample Reference Concentrations (RfCs) for
Noncancer Endpoints Other Than the Critical Effect 5-24
5.1.5. Previous Inhalation Assessment 5-27
-------
5.1.6. Uncertainties in Inhalation Reference Concentration 5-27
5.2. ORAL REFERENCE DOSE (RfD) 5-28
5.2.1. Choice of Principal Studies and Critical Effects 5-28
5.2.2. Additional Analyses: Route-to-Route Extrapolation Using PBPK
Modeling 5-28
5.2.3. Reference Dose (RfD) Derivation, Including Application of Uncertainty
Factors 5-29
5.2.4. Dose-response Analyses for Noncancer Effects Other Than Critical
Effect of Neurotoxicity 5-33
5.2.4.1. Sample Reference Doses (RfDs) for Kidney Toxicity 5-34
5.2.4.2. Sample Reference Doses (RfDs) for Liver Toxicity 5-35
5.2.4.3. Sample Reference Doses (RfDs) for Immunotoxicity and
Hematologic Toxicity 5-38
5.2.4.4. Sample Reference Doses (RfDs) for Reproductive and
Developmental Toxicity 5-38
5.2.4.5. Summary of Sample Reference Doses (RfDs) for Noncancer
Endpoints Other Than the Critical Effect 5-38
5.2.5. Previous Oral Assessment 5-39
5.2.6. Uncertainties in Oral Reference Dose 5-39
5.3. CANCER DOSE-RESPONSE ASSESSMENT 5-41
5.3.1. Choice of Study/Data with Rationale and Justification 5-41
5.3.2. Dose-Response Data 5-42
5.3.2.1. Liver Tumors in Mice 5-42
5.3.2.2. Other Tumor Sites in Male Mice 5-46
5.3.2.3. Mononuclear Cell Leukemia in Rats 5-46
5.3.2.4. Other Tumor Sites in Rats 5-49
5.3.3. Dose Adjustments and Extrapolation Methods 5-51
5.3.3.1. Estimation of Dose Metrics for Dose-Response Modeling 5-51
5.3.3.2. Extrapolation Methods 5-57
5.3.4. Cancer Risk Values 5-64
5.3.4.1. Dose-Response Modeling Results 5-65
5.3.4.2. Inhalation Unit Risk 5-87
5.3.4.3. Oral Slope Factor 5-89
5.3.4.4. Uncertainties in Human Population Variability and
Quantitative Adjustment for Sensitive Populations (Age-
Dependent Adjustment Factors) 5-93
XI
-------
5.3.4.5. Concordance of Animal and Human Risk Estimates 5-93
5.3.5. Summary of Uncertainties in Cancer Risk Values 5-95
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND
DOSE RESPONSE 6-1
6.1. HUMAN HAZARD POTENTIAL 6-1
6.1.1. Exposure (refer to Section 2) 6-1
6.1.2. Toxicokinetics and Physiologically Based Pharmacokinetic (PBPK)
Modeling (refer to Section 3) 6-2
6.1.3. Noncancer Toxicity (refer to Section 4.10.1) 6-4
6.1.3.1. Neurological Effects (refer to Section 4.1) 6-5
6.1.3.2. Summary of Other Noncancer Adverse Effects (refer to
Sections 4.2, 4.3, 4.6, and 4.7) 6-9
6.1.4. Carcinogenicity (refer to Section 4.10.2) 6-13
6.1.5. Susceptibility (refer to Section 4.9) 6-14
6.2. DOSE-RESPONSE ASSESSMENT 6-15
6.2.1. Noncancer Effects (refer to Section 5.1) 6-15
6.2.1.1. Selection of Principal Studies and Critical Effect (refer to
Section 5.1.1) 6-15
6.2.1.2. Uncertainties and Application of Uncertainty Factors (UFs)
(refer to Sections 5.1.3, 5.2.3) 6-17
6.2.1.3. Reference Concentration (refer to Section 5.1.3) 6-17
6.2.1.4. Reference Dose (refer to Section 5.2) 6-18
6.2.1.5. Dose-Response Analyses for Noncancer Effects Other Than
Critical Effect of Neurotoxicity (refer to Sections 5.1.4 and
5.2.4) 6-19
6.2.2. Cancer (refer to Section 5.2) 6-20
6.2.2.1. Inhalation Unit Risk 6-21
6.2.2.2. Oral Slope Factor 6-22
6.2.2.3. Uncertainties in Cancer Dose-Response Assessment (refer to
Section 5.3.5) 6-22
7. REFERENCES 7-1
APPENDIX A. EPA RESPONSE TO MAJOR EXTERNAL PEER-REVIEW AND
PUBLIC COMMENTS A-l
APPENDIX B. STUDY DESIGN CHARACTERISTICS OF
TETRACHLOROETHYLENE EXPOSURE AND CANCER
EPIDEMIOLOGICAL STUDIES B-l
xn
-------
APPENDIX C. CONSISTENCY OF TETRACHLOROETHYLENE AND
TRICHLOROACETIC ACID HEPATOCARCINOGENICITY C-l
APPENDIX D. CANCER DOSE-RESPONSE MODELING D-l
Xlll
-------
LIST OF TABLES
Table 2-1. Physical and chemical properties of tetrachloroethylene 2-2
Table 3-1. Log-likelihood and parameters after calibration 3-36
Table 3-2. Predictions for area-under-the-curve of tetrachloroethylene in blood
(mg-hr/L-day per ppm in air or mg-hr/L-day per mg/kg-day oral intake)
using posterior mode parameters 3-39
Table 3-3. Predictions for fraction of tetrachloroethylene oxidized by cytochrome P450
(P450s) (mg/kg-day oxidized per mg/kg-day intake) using posterior mode
parameters 3-41
Table 3-4. Predictions for fraction of tetrachloroethylene conjugated with glutathione
(GSH) (mg/kg-day conjugated per mg/kg-day intake) using posterior mode
parameters 3-43
Table 3-5. Predictions for Trichloroacetic acid (TCA) produced systemically (mg/kg-day
systemic TCA per ppm in air or mg/kg-day systemic TCA per mg/kg-day
oral intake) using posterior mode parameters (continued) 3-46
Table 3-6. Summary evaluation of the reliability of tetrachloroethylene dose metrics 3-47
Table 3-7. Ratio of average daily dose at various life-stages to the average daily dose for
a 25-year-old adult: physiologically based pharmacokinetic (PBPK)
simulations 3-51
Table 4-1. Summary of human neurotoxicity studies of occupational or residential
exposures to dry-cleaning facilities using tetrachloroethylene 4-7
Table 4-2. Summary of effects of chronic tetrachloroethylene exposure in humans
observed in studies of neuropsychological functiona 4-36
Table 4-3. Summary of animal inhalation neurotoxicology studies 4-43
Table 4-4. Summary of oral neurotoxicity animal studies 4-50
Table 4-5. Summary of in vitro ion channel effects of tetrachloroethylene and other
chlorinated solvents 4-55
Table 4-6. Summary of human kidney toxicity marker studies of occupational exposures
to dry-cleaning facilities using tetrachloroethylene 4-62
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney
cancer 4-75
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder
cancer 4-87
Table 4-9. Summary of rodent kidney toxicity studies 4-102
Table 4-10. Kidney tumor incidence in laboratory animals exposed to
tetrachloroethylene 4-104
xiv
-------
Table 4-11. Renal a2u-globulin formation in tetrachloroethylene-exposed rodents 4-118
Table 4-12. Renal peroxisome proliferation in tetrachloroethylene-exposed rodents 4-122
Table 4-13. Summary of studies of human liver toxicity 4-128
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer ..4-133
Table 4-15. Summary of inhalation and oral rodent liver toxicity studies 4-142
Table 4-16. Incidence of hepatic tumors in rodents exposed to tetrachloroethylene 4-151
Table 4-17. Hepatocarcinogenicity of TCA in rodent drinking water studies 4-158
Table 4-18. Hepatocarcinogenicity of DC A in rodent drinking water studies 4-158
Table 4-19. Incidence of mouse liver tumors with drinking water administration of TCA
and DCA, alone and in combination 4-159
Table 4-20. Rodent studies of induction of peroxisome proliferation or its markers by
tetrachloroethylene 4-165
Table 4-21. Potency indicators for mouse hepatocarcinogenicity and in vitro
transactivation of mouse PPARa for four PPARa agonistsa 4-174
Table 4-22. Potency indicators for rat hepatocarcinogenicity and common short-term
markers of PPARa-agonism for four PPARa agonists21 4-175
Table 4-23. Summary of human studies on tetrachloroethylene exposure and esophageal
cancer 4-188
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer... 4-198
Table 4-25. Immune and hematological parameters in studies of dry-cleaning workers or
tetrachloroethylene exposure in children 4-208
Table 4-26. Immune-related conditions in studies of dry cleaning or tetrachloroethylene
exposure in humansa 4-214
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and
hematopoietic cancers, including leukemia 4-218
Table 4-28. Summary of epidemiologic studies on tetrachloroethylene exposure and
childhood hematopoietic cancers, including leukemia 4-233
Table 4-29. Results of epidemiologic studies of potential tetrachloroethylene exposure
and adult lymphopoietic cancer and leukemia, by cancer type and study
design 4-239
Table 4-30. Results of epidemiologic studies of potential tetrachloroethylene exposure
and adult non-Hodgkin lymphoma, by study design 4-241
Table 4-31. Results of epidemiologic studies of potential tetrachloroethylene exposure
and adult Hodgkin lymphoma and multiple myeloma, by study design 4-245
Table 4-32. Results of epidemiologic studies of potential tetrachloroethylene exposure
and adult lymphopoeitic cancers, with data pertaining to exposure-response
gradients, by cancer type 4-248
xv
-------
Table 4-33. Mononuclear cell leukemia incidence in rats exposed to tetrachloroethylene....4-255
Table 4-34. Epidemiology studies on reproduction and development 4-274
Table 4-35. Exposure concentrations (ppm) at which effects occurred in a two-
generation study 4-333
Table 4-36. Summary of human studies on tetrachloroethylene exposure and breast
cancer 4-338
Table 4-37. Summary of human studies on tetrachloroethylene exposure and cervical
cancer 4-344
Table 4-38. Summary of mammalian developmental and reproductive toxicity studies for
tetrachloroethylene 4-355
Table 4-39. Genotoxicity of tetrachloroethylene—mammalian systems (in vitro and in
vivo)a 4-362
Table 4-40. Genotoxicity of tetrachloroethylene—bacterial, yeast, and fungal systemsa 4-365
Table 4-41. Genotoxicity of trichloroacetic acid (TCA)—mammalian systems (in vitro
andinvivo)a 4-378
Table 4-42. Genotoxicity of trichloroacetic acid (TCA)—bacterial systems21 4-3 81
Table 4-43. Genotoxicity of dichloroacetic acid (DCA)—mammalian systems (in vitro
and in vivo)a 4-386
Table 4-44. Genotoxicity of dichloroacetic acid (DCA)—bacterial systemsa 4-388
Table 4-45. Genotoxicity of chloral hydrate—mammalian systems (in vitro)a 4-391
Table 4-46. Genotoxicity of chloral hydrate—mammalian systems (in vivo)a 4-393
Table 4-47. Genotoxicity of chloral hydrate—bacterial, yeast, and fungal systemsa 4-395
Table 4-48. Genotoxicity of additional tetrachloroethylene metabolites—all systems 4-403
Table 4-49. Inhalation studies suitable for dose-response analyses 4-438
Table 4-50. NOAELs and LOAELs in selected studies involving oral exposure to
tetrachloroethylene 4-441
Table 4-51. Tumor incidence in rats exposed to tetrachloroethylene 4-462
Table 4-52. Tumor incidence in mice exposed to tetrachloroethylene 4-463
Table 4-53. Renal a2u-globulin accumulation in tetrachloroethylene-exposed rodents 4-469
Table 4-54. Renal peroxisome proliferation in tetrachloroethylene-exposed rodents 4-469
Table 4-55. Rodent studies of induction of hepatic peroxisome proliferation or its
markers by tetrachloroethylene 4-471
Table 4-56. Summary of hypothesized modes of action for tetrachloroethylene-induced
cancer in rodents 4-473
Table 5-1. Neurotoxicological inhalation studies considered in the development of an
RfC 5-3
xvi
-------
Table 5-2. Summary of rationale for identifying studies on tetrachloroethylene for RfC
development 5-7
Table 5-3. Application of uncertainty factors for the neurological endpoints from
the studies used to derive candidate RfCs [Echeverria et al. (1995) and
Cavallerietal. (1994)] 5-17
Table 5-4. Sample RfCs for kidney effects 5-23
Table 5-5. Sample RfCs for liver effects 5-23
Table 5-6. Sample RfCs for immunological and hematological effects 5-25
Table 5-7. Sample RfCs for reproductive and developmental effects 5-25
Table 5-8. Application of uncertainty factors for neurological endpoints from the studies
used to derive candidate RfDs 5-31
Table 5-9. Sample RfDs for kidney effects 5-36
Table 5-10. Sample RfDs for liver effects 5-36
Table 5-11. Sample RfDs for immunological and hematological effects 5-37
Table 5-12. Sample RfDs for reproductive and developmental effects 5-37
Table 5-13. Tumor incidence in mice exposed to tetrachloroethylene 5-44
Table 5-14. Historical control data of the Japan Bioassay Research Center, Crj/BDFl
mouse, 104 week studies 5-45
Table 5-15. Incidence of mononuclear cell leukemia, kidney tumors, and brain gliomas
in rats exposed to tetrachloroethylene by inhalation 5-47
Table 5-16. Historical control data of the Japan Bioassay Research Center, F344/DuCrj
(Fischer) rat, 104 week studies 5-48
Table 5-17. Summary of PBPK-derived dose metric estimates used for dose-response
analysis of rodent tumor data 5-54
Figure 5-9. Dose-response modeling of male mouse hepatocellular tumors associated
with inhalation exposure to tetrachloroethylene, in terms of liver total
oxidative metabolites; response data from JISA (1993) 5-67
Table 5-18. Human equivalent candidate unit risks, derived using PBPK-derived dose
metrics and multistage model; tumor incidence data from JISA (1993) and
NTP(1986) 5-68
Table 5-19. Dose-response summary and candidate unit risk estimates using continuous
equivalent administered tetrachloroethylene levels as dose metric, from NTP
(1986) and JISA (1993) 5-70
Table 5-20. Range of outputs from fitting different BMDS models using continuous
equivalent administered tetrachloroethylene levels as dose metric, from JISA
(1993)a 5-71
xvn
-------
Table 5-21. Human equivalent candidate oral slope factors, derived using primary dose
metrics and multistage model; tumor incidence data from JISA (1993) and
NTP(1986) 5-91
Table 5-22. Summary of uncertainties in tetrachloroethylene cancer unit risk estimate 5-96
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry
workers, and other cohorts) B-28
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site
studies B-124
Table B-3. Summaries of characteristics of case-control studies: single cancer-site
studies B-142
Table B-4. Summaries of characteristics of geographically based and other studies B-l 87
Table C-l. Incidence of hepatocellular adenomas and carcinomas in male B6C3Fi mice
exposed to tetrachloroethylene in two inhalation bioassays C-2
Table C-2. TCA drinking water studies in male mice: incidence of hepatocellular
adenomas and carcinomas C-3
Table C-3. PBPK model-estimated TCA internal dose measures for tetrachloroethylene
and TCA bioassays used in analysis C-4
Table C-4. Logistic regression model fits—beta coefficients and standard errorsa C-7
Table D-l. Model predictions for hepatocellular tumors in male mice (JISA, 1993)a,
using several dose metrics and multistage cancer model D-l
Table D-2. Model predictions for hepatocellular tumors in female mice (JISA, 1993),
using several dose metricsa and multistage cancer model D-8
Table D-3. Model predictions for hemangiomas or hemangiosarcomas in male mice
(JISA, 1993), using tetrachloroethylene AUC in blood and administered
tetrachloroethylene concentration as dose metricsa and multistage cancer
model D-15
Table D-4. Model predictions for male rat mononuclear cell leukemia (MCL) (JISA,
1993), using tetrachloroethylene AUC in blood and administered
tetrachloroethylene concentration as dose metrics21 and multistage model D-21
Table D-5. Model predictions for female rat MCL (JISA, 1993),a using administered
tetrachloroethylene concentration (ppm)c and multistage model D-26
Table D-7. Model predictions for combined male and female rat MCL (JISA, 1993),a
using administered tetrachloroethylene concentrations as dose metric D-32
Table D-8. Model predictions for male rat tumors (NTP, 1986),a using administered
tetrachloroethylene concentration as dose metric and multistage model D-38
Table D-9. Comparison of model predictions for hepatocellular tumors in mice (JISA,
1993), using administered tetrachloroethylene concentration (ppm) as the
dose metric,a across a range of dichotomous models D-52
xvin
-------
Table D-10. Comparison of model predictions for male mice, hemangiomas, or
hemangiosarcomas (JISA, 1993),a using administered tetrachloroethylene
concentration as dose metric,b across a range of dichotomous models D-53
Table D-l 1. Comparison of model predictions for MCL in male rats (JISA, 1993),a using
administered tetrachloroethylene concentration as dose metric,b across a
range of dichotomous models D-54
xix
-------
LIST OF FIGURES
Figure 3-1. Postulated scheme for the metabolism of tetrachloroethylene by the
cytochrome P450 (P450) oxidative pathway and glutathione ^-transferase
(GST)-mediated glutathione (GSH) conjugation pathway 3-5
Figure 3-2. Comparison of model predictions for blood concentration with experiment 3-27
Figure 3-3. Comparison of model predictions for alveolar concentration of
tetrachloroethylene with experimental data on humans 3-28
Figure 3-4. Previously published estimates for the total amount of tetrachloroethylene
metabolized at 0.001 ppm (1 ppb) continuous inhalation exposure 3-29
Figure 3-5. Overall structure of updated physiologically based pharmacokinetic (PBPK)
model for tetrachloroethylene and metabolites 3-31
Figure 3-6. Comparison of mouse (A-B), rat (C-D), and human (E-F) rates of hepatic
oxidation (A, C, and E) or conjugation (B, D, and F) measured in vitro
(symbols) and predicted by the model (lines) 3-38
Figure 3-7. Physiologically based pharmacokinetic (PBPK) simulations of variations
with age and gender in blood concentrations of tetrachloroethylene and its
main metabolite trichloroacetic acid (TCA) 3-51
Figure 4-1. Visual contrast sensitivity functions for control and exposed participants in a
study of workers in a day-care center located in a building with a dry-
cleaning facility (Schreiber et al., 2002) 4-25
Figure 4-2. Visual contrast sensitivity functions for control and exposed participants in
residential exposure study (Schreiber et al., 2002) 4-30
Figure 4-3. Incidences of hepatocellular adenomas (A) and hepatocellular adenomas and
carcinomas (B) in mice exposed toDEFtP 4-172
Figure 5-1. Exposure-response array for neurotoxicological inhalation studies considered
for RfC development (listed in Table 5-1) 5-6
Figure 5-2. Candidate reference concentration values for inhalation exposure to
tetrachloroethylene 5-19
Figure 5-3. Comparison of candidate RfCs (filled squares) supporting the RfC (vertical
line) and sample RfCs (open squares) for effects other the critical effect
(CNS toxicity) 5-26
Figure 5-4. Candidate reference dose values for exposure to tetrachloroethylene 5-32
Figure 5-5. Comparison of candidate RfDs (filled squares) supporting the RfD (vertical
line) and sample RfDs (open squares) for effects other the critical effect
(CNS toxicity) 5-40
Figure 5-6. Mouse liver tumor responses (hepatocellular adenomas or carcinomas), as
additional risk, for two chronic inhalation bioassays (refer to Table 5-13),
xx
-------
plotted against continuous equivalent concentration (ppm), for male and
female mice 5-46
Figure 5-7. Rat mononuclear cell leukemia responses (minus control) in two chronic
bioassays (refer to Table 5-15), plotted against continuous equivalent
exposure (ppm) for (a) male and (b) female rats 5-49
Figure 5-8. Sequence of steps for extrapolating from tetrachloroethylene bioassays in
animals to human-equivalent exposures expected to be associated with
comparable cancer risk (combined interspecies and route-to-route
extrapolation) 5-52
Figure 5-10. Dose-response modeling of female mouse hepatocellular tumors associated
with inhalation exposure to tetrachloroethylene, in terms of liver total
oxidative metabolites; response data from JISA (1993) 5-67
Figure 5-11. Dose-response modeling of male mouse hemangiomas or
hemangiosarcomas associated with inhalation exposure to
tetrachloroethylene, in terms of tetrachloroethylene AUC in blood; response
data from JISA (1993) 5-73
Figure 5-12. Dose-response modeling of female and male rat MCLs associated with
inhalation exposure to tetrachloroethylene, in terms of tetrachloroethylene
AUC in blood; response data from JISA (1993) 5-76
Figure 5-13. Dose-response modeling of male rat tumors—kidney, brain gliomas,
interstitial cell tumors, MCLs—associated with inhalation exposure to
tetrachloroethylene, in terms of tetrachloroethylene AUC in blood; response
data from NTP (1986) 5-82
Figure 5-14. Comparison of inhalation unit risks for tetrachloroethylene derived from
rodent bioassays using PBPK-based dose metrics and administered
concentration 5-88
Figure 5-15. Comparison of oral slope factors for tetrachloroethylene, derived from
rodent bioassays using PBPK-based dose metrics and route-to-route
extrapolation 5-90
Figure C-l. Logistic regression dose-response fits to tetrachloroethylene data [open
circle: JISA (1993); filled circle: NTP (1986)] C-8
Figure C-2. Logistic regression dose-response fits to TCA C-10
Figure D-l One-degree multistage model fit to hepatocellular tumors in male mice
(JISA, 1993), with BMD and BMDL at 10% extra risk, using total oxidative
metabolism in liver (mg/kg°'75-day) D-2
Figure D-2. One-degree multistage model fit to hepatocellular tumors in male mice
(JISA, 1993), with BMD and BMDL at 10% extra risk, using TCA AUC as
dose metric (mg-hr/L-d) D-4
Figure D-3. One-degree multistage model fit to hepatocellular tumors in male mice
(JISA, 1993), with BMD and BMDL at 10% extra risk, using administered
tetrachloroethylene concentration (ppm) D-6
xxi
-------
Figure D-4. Two-degree multistage model fit to hepatocellular tumors in female mice
(JISA, 1993), with BMD and BMDL at 10% extra risk D-9
Figure D-5. Two-degree multistage model fit to hepatocellular tumors in female mice
(JISA, 1993), with BMD and BMDL at 10% extra risk D-ll
Figure D-6. One-degree multistage model fit to hepatocellular tumors in female mice
(JISA, 1993), with BMD and BMDL at 10% extra risk D-13
Figure D-7. One-degree multistage model fit to hemangioma or hemangiosarcoma
incidence in male mice (JISA, 1993), with BMD and BMDL at 10% extra
risk D-16
Figure D-8. One-degree multistage model fit to hemangioma or hemangiosarcoma
incidence in male mice (JISA, 1993), with BMD and BMDL at 10% extra
risk D-18
Figure D-9. One-stage model fit to MCL incidence in male rats (JISA, 1993), with BMD
and BMDL at 10% extra risk D-22
Figure D-10. One-stage model fit to MCL incidence in male rats (JISA, 1993), with
BMD andBMDL at 10% extra risk D-24
Figure D-ll. One-stage multistage model fit to MCL incidence in female rats (JISA,
1993), with BMD and BMDL at 10% extra risk D-28
Figure D-12. Michaelis-Menten model (dichotomous Hill model with exponent fixed at
1) fit to MCL incidence in female rats (JISA, 1993), with BMD and BMDL
at 10% extra risk D-30
Figure D-13. Michaelis-Menten model (dichotomous-Hill model with exponent fixed at
1) fit to MCL incidence in male and female rats (JISA, 1993), with BMD
and BMDL at 10% extra risk D-34
Figure D-14. Michaelis-Menten model (dichotomous-Hill model with exponent fixed at
1) fit to MCL incidence in male and female rats (JISA, 1993), with BMD
and BMDL at 10% extra risk D-36
Figure D-15. Multistage model fits to tumor incidences at multiple sites in male rats—
kidney tumors, brain gliomas, testicular interstitial cell tumors, and MCL
(NTP, 1986) D-39
Figure D-16. Michaelis-Menten model (dichotomous-Hill model with exponent fixed at
1) fit to MCL incidence in male rats (JISA, 1993), with BMD and BMDL at
10% extra risk; with administered concentration as dose metric (top) or
tetrachloroethylene AUC in blood (bottom) D-55
xxn
-------
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 lower bound benchmark concentration
BMD benchmark dose
BMDL lower bound 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 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
GGT gamma-glutamyltransferase
GSH glutathione
GST glutathione S-transferase
GSTx glutathione ^-transferase isoform, where x denotes different isoforms (such as M,
T, P, S, Z)
HEC 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
IPCS International Programme on Chemical Safety
IRIS Integrated Risk Information System
IUGR intrauterine growth restriction
xxin
-------
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
MCA monochloroacetic acid
MCL-5 microsomal epoxide hydrolase
MCL mononuclear cell leukemia
MOA mode of action
MRL minimal risk level
NAG TV-acetyl-p-D-glucosaminidase
NCI National Cancer Institute
NHL non-Hodgkin 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
NYSDOH New York State Department of Health
NYSOAG 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
SES socio-economic status
SGA small for gestational age
SIR standardized incidence ratio
SMR standardized mortality ratio
SSB single-strand break
TCA trichloroacetic acid
TCE trichloroethylene
xxiv
-------
TCOH trichloroethanol
TCVC S-(l,2,2,-trichlorovinyl)-L-cysteine
TCVCSO S-(l,2,2,-tricMorovinyl)-L-cysteine sulfoxide
TCVG ^-(1,2,2-trichlorovinyl) glutathione
TNAP tissue nonspecific 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 visual evoked potential
Vmax maximum velocity
WHO World Health Organization
XXV
-------
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.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose-
Response, is to present the significant conclusions reached in the derivation of the reference
dose, reference concentration, and cancer assessment, where applicable, and to characterize the
overall confidence in the quantitative and qualitative aspects of hazard and dose response by
addressing the quality of data and related uncertainties. 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,
refer to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (e-mail address).
xxvi
-------
CONTRIBUTORS AND REVIEWERS
CHEMICAL MANAGERS
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Kathryn Z. Guy ton
Karen A. Hogan
CONTRIBUTORS
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
AmbujaBale
Stanley Bar one
David Bussard
Jane Caldwell
Weihsueh A. Chiu
Glinda Cooper
Rebecca Brown Dzubow1
Barbara Glenn
Maureen R. Gwinn
Leonid Kopylev
Susan Makris
Robert McGaughy1
Jean Parker1
Deborah Rice1
Cheryl Siegel Scott
Bob Sonawane
Ravi Subramaniam
Larry Valcovic1
Paul White
formerly with the National Center for Environmental Assessment
xxvn
-------
CONTRIBUTORS AND REVIEWERS (continued)
REVIEWERS
This document has been provided for review to EPA scientists, interagency reviewers
from other federal agencies and White House offices, and the public, and peer reviewed by
independent scientists external to EPA. Appendix A provides a summary and EPA's disposition
of the comments received from the independent external peer reviewers and from the public.
INTERNAL PEER REVIEWERS
Hugh Barton, formerly with 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
Vincent 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
Deirdre 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
Brenda Foos, U.S. Environmental Protection Agency, Office of the Administrator, Office of
Children's Health Protection, Washington, DC
xxvin
-------
CONTRIBUTORS AND REVIEWERS (continued)
Bruce Rodan, U.S. Environmental Protection Agency, formerly with National Center for
Environmental Assessment, Washington, DC
Diana M. Wong, U.S. Environmental Protection Agency, formerly with 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, Lincoln University, PA
George Lucier, formerly with National Institute of Environmental Health Sciences, Research
Triangle Park, NC
Robert Park, National Institute for Occupational Safety and Health, Education and Information
Division, Cincinnati, OH
Val Schaeffer, Occupational Safety and Health Administration, Directorate for Health Standards,
Washington, DC
NEUROTOXICITY EXPERT REVIEW PANEL REVIEWERS: Public Workshop,
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
XXIX
-------
CONTRIBUTORS AND REVIEWERS (continued)
EXTERNAL PEER REVIEWERS
Sam Kacew (Chair)
University of Ottawa, Ontario, Canada
Bruce H. Alexander
University of Minnesota School of Public Health, Minneapolis, MN
Margit L. Bleecker
Center for Occupational and Environmental Neurology, Baltimore, MD
Gary P. Carlson
Purdue University, West Lafayette, IN
Linda D. Cowan
University of Oklahoma Health Sciences Center, Oklahoma City, OK
Mary E. Davis
West Virginia University, Morgantown, WV
H. Christopher Frey
North Carolina State University, Raleigh NC
Joseph R. Landolph
University of Southern California, Los Angeles, CA
M.E. (Bette) Meek
University of Ottawa, Ontario, Canada
David C. McMillan
University of Nebraska Medical Center, Omaha, NE
M. ChristopherNewland
Auburn University, Auburn, AL
Julia Quint
California Department of Public Health (retired), Berkeley, CA
Gary L. Rosner
University of Texas M.D. Anderson Cancer Center, Houston, TX
XXX
-------
CONTRIBUTORS AND REVIEWERS (continued)
Ivan Rusyn
University of North Carolina, Chapel Hill, NC
Rolf Schulte-Hermann
Medical University of Vienna, Austria
Irvin R. Schultz
Battelle Pacific Northwest Division, Sequim, WA
Robert Snyder
Rutgers, the State University of New Jersey, Piscataway, NJ
Roberta F. White
Boston University School of Public Health, Boston, MA
Luoping Zhang
University of California, Berkeley, CA
Yiliang Zhu
University of South Florida, Tampa, FL
ACKNOWLEDGMENTS
The authors would like to acknowledge the contributions of the following individuals:
Terri Konoza of NCEA who managed the document production activities; Ellen Lorang of
NCEA, who provided HERO literature database support; Maureen Johnson of NCEA, who
provided website support; Cristopher Broyles of IntelliTech Systems, Inc. and Heidi Glick of
ECFlex, Inc., who provided editorial support; Lana Wood, Debbie Kleiser, and Crystal Lewis of
ECFlex, Inc. and Stacey Herron of IntelliTech Systems, Inc., who provided word processing
support; and Linda Tackett of IntelliTech Systems, Inc., who provided proofing support.
xxxi
-------
EXECUTIVE SUMMARY
Tetrachloroethylene is a widespread contaminant that is present in ambient air, indoor air,
soil, and groundwater. Once exposed, humans, as well as laboratory animal species, rapidly
absorb tetrachloroethylene, which is then distributed to tissues via systemic circulation,
metabolized, and then excreted primarily in breath as unchanged tetrachloroethylene or CC>2, or
in urine as metabolites. Based on the available human epidemiologic data and experimental and
mechanistic studies, it is concluded that tetrachloroethylene poses a potential human health
hazard for noncancer toxicity to the central nervous system, kidney, liver, immune and
hematologic system, and on development and reproduction. Neurotoxicity is identified as a
sensitive endpoint following either oral or inhalation exposure to tetrachloroethylene.
Neurotoxic effects have been characterized in human controlled exposure, occupational and
residential studies, as well as in experimental animal studies, providing evidence that
tetrachloroethylene exposure results in visual changes, increased reaction time, and decrements
in cognition.
Following EPA (2005a) Guidelines for Carcinogen Risk Assessment, tetrachloroethylene
is "Likely to be Carcinogenic to Humans" by all routes of exposure. This characterization is
based on suggestive evidence of carcinogenicity in epidemiologic studies and conclusive
evidence that the administration of tetrachloroethylene, either by ingestion or by inhalation to
sexually mature rats and mice, increases tumor incidence (JISA, 1993; NTP, 1986; NCI, 1977).
In the rodent bioassays, tetrachloroethylene increased the incidence of liver tumors
(hepatocellular adenomas and carcinomas) in male and female mice and of mononuclear cell
leukemia (MCL) in both sexes of rats. These findings were reproducible in multiple lifetime
bioassays employing different rodent strains and, in the case of mouse liver tumors, by the
inhalation and oral exposure routes. Additional tumor findings in rats included significant
increases in the NTP bioassay of testicular interstitial cell tumors and kidney tumors in males,
and brain gliomas in males and females. In mice, hemangiosarcomas in liver, spleen, fat, and
subcutaneous skin were reported in males in the JISA study. The epidemiologic evidence
provides a pattern associating tetrachloroethylene exposure and several types of cancer,
including bladder cancer, non-Hodgkin lymphoma and multiple myeloma. Associations and
exposure-response relationships were reported by studies using more precise
exposure-assessments for tetrachloroethylene. For other sites, including esophageal, kidney,
lung, cervical and breast cancer, more limited data suggestive of an effect are available.
As tetrachloroethylene toxicity and carcinogenicity are generally associated with
tetrachloroethylene metabolism, susceptibility to tetrachloroethylene health effects may be
xxxn
-------
modulated by factors affecting toxicokinetics, including lifestage, gender, genetic
polymorphisms, race/ethnicity, preexisting health status, lifestyle, and nutrition status. These
and other factors (e.g., socioeconomic status and multiple exposures) may contribute to variation
in response to tetrachloroethylene or its metabolites, once produced. In addition, it is not known
how tetrachloroethylene interacts with known risk factors for human diseases.
Dose-response analyses of the noncancer database focused on the neurotoxicity data set
as a basis for derivation of inhalation and oral reference values via the LOAEL/NOAEL
approach. The two principal studies that were used demonstrated color vision changes (Cavalleri
et al., 1994), and cognitive and reaction time changes (Echeverria et al., 1995). Candidate RfCs
derived from these studies span a range from 0.015 to 0.056 mg/m3. The midpoint of this range
(0.04 mg/m3) was chosen as the RfC for tetrachloroethylene.
The RfD for noncancer effects of 6 x 10~3 mg/kg-day was derived through route-to-route
extrapolation of the above inhalation studies. The RfD is equivalent to a drinking water
concentration of 0.21 mg/L, assuming a body weight of 70 kg and a daily water consumption of
2 L. These noncancer reference values are supported by estimates from multiple human
neurotoxicity studies. Additionally, quantitative dose-response analyses of the findings for other
toxicity endpoints (i.e., kidney, liver, immunologic and hematologic, and reproductive and
developmental toxicity), detailed in Section 5 and summarized in Sections 6.2.5 and 6.2.7, are
considered to be supportive of these values.
For cancer, the majority of the NRC peer review panel recommended that the mouse
hepatocellular tumors be used for cancer risk estimation. Therefore, the inhalation unit risk is
2 x 10"3 per ppm or 3 x 10"7 per jig/m3, based on the male mouse hepatocellular tumor data
from the JISA (1993) bioassay. The oral slope factor, developed by PBPK model-derived
route-to-route extrapolation from the same data, is 2 x 10"3 per mg/kg-day.
xxxin
-------
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of
tetrachloroethylene. IRIS Summaries may include oral reference dose (RfD) and inhalation
reference concentration (RfC) values for chronic and other exposure durations, and a
carcinogenicity assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as 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 effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD but provides a continuous inhalation exposure estimate. The RfC
considers toxic effects for both the respiratory system (portal-of-entry) and for effects peripheral
to the respiratory system (extrarespiratory or systemic effects). Reference values are generally
derived for chronic exposures (up to a lifetime) but may also be derived for acute (<24 hours),
short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of lifetime) exposure
durations, all of which are derived based on an assumption of continuous exposure throughout
the duration specified. Unless specified otherwise, the RfD and RfC are derived for chronic
exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question, and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is a
plausible upper bound on the estimate of risk per ug/m3 air breathed.
Development of these hazard identification and dose-response assessments for
tetrachloroethylene has followed the general guidelines for risk assessment set forth by the
National Research Council (NRC, 1994, 1983). EPA Guidelines and Risk Assessment Forum
technical panel reports that may have been used in the development of this assessment include
the following: Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA,
1986c), Guidelines for Mutagenicity Risk Assessment (U.S. EPA, 1986b), Recommendations for
1-1
-------
and Documentation of Biological Values for Use in Risk Assessment (U.S. EPA, 1988),
Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA, 1991b), Interim Policy for
Particle Size and Limit Concentration Issues in Inhalation Toxicity (Whalan and Redden, 1994),
Methods for Derivation of Inhalation Reference Concentrations and Application of Inhalation
Dosimetry (U.S. EPA, 1994), Use of the Benchmark Dose Approach in Health Risk Assessment
(U.S. EPA, 1995), Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996a),
Guidelines for Neurotoxicity Risk Assessment (U.S. EPA, 1998b), Science Policy Council
Handbook: Risk Characterization (U.S. EPA, 2000b), Benchmark Dose Technical Guidance
Document (U.S. EPA, 2000a), Supplementary Guidance for Conducting Health Risk Assessment
of Chemical Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference
Concentration Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S.
EPA, 2005a), Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
Carcinogens (U.S. EPA, 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA,
2006c), and A Framework for Assessing Health Risks of Environmental Exposures to Children
(U.S. EPA, 2006b).
The literature search strategy employed for tetrachloroethylene was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered in
the development of this document. Primary, peer-reviewed literature identified through August 2011
was included where that literature was determined to be critical to the assessment. The relevant
literature included publications on tetrachloroethylene that were identified through Toxicology
Literature Online (TOXLINE), PubMed, the Toxic Substance Control Act Test Submission Database
(TSCATS), the Registry of Toxic Effects of Chemical Substances (RTECS), the Chemical
Carcinogenesis Research Information System (CCRIS), the Developmental and Reproductive
Toxicology/Environmental Teratology Information Center (DART/ETIC), the Hazardous Substances
Data Bank (HSDB), the Genetic Toxicology Data Bank (GENE-TOX), Chemical abstracts, and
Current Contents. Other peer-reviewed information, including health assessments developed by
other organizations, review articles, and independent analyses of the health effects data were
retrieved and may be included in the assessment where appropriate. It should be noted that
references have been added to the Toxicological Review after the external peer review in response to
peer reviewer's comments and for the sake of completeness. These references have not changed the
overall qualitative and quantitative conclusions.
1-2
-------
2. BACKGROUND
2.1. USES AND PHYSICAL/CHEMICAL PROPERTIES
Tetrachloroethylene is a widely used solvent that is produced commercially for use in dry
cleaning, textile processing, and metal-cleaning operations. It has the following use pattern:
55% as a chemical intermediate, 25% for metal cleaning and vapor degreasing, 15% for dry
cleaning and textile processing, and 5% for other unspecified uses (ATSDR, 1997a). Table 2-1
lists the physical and chemical properties of tetrachloroethylene (ATSDR, 1997a).
2.2. OCCURRENCE AND EXPOSURE
Tetrachloroethylene has been detected in ground water and surface water as well as in air,
soil, food, and breast milk. The primary exposure routes of concern are inhalation of vapor and
ingestion of contaminated water. Although dermal exposure is possible via contaminated tap
water during showering, bathing, or swimming, this is generally not considered a major route of
exposure.
2.2.1. Air
Because of its high volatility, there is considerable potential for release of
tetrachloroethylene into the atmosphere. Once in the air, it is not susceptible to wet deposition
because of its hydrophobicity. The primary method for removal is photooxidation to
trichloroacetyl chloride, trichloroacetic acid (TCA), carbon monoxide, ozone, and phosgene
(Gilbert et al., 1982). However, this reaction is very slow, so tetrachloroethylene is not
implicated in the buildup of any of the reaction products in the troposphere. Though the half-life
of tetrachloroethylene can vary based on season and environmental conditions, it has been
estimated at 96 days under typical conditions (ATSDR, 1997a).
Ambient tetrachloroethylene concentrations vary from source to source and with
proximity to the source. Outdoors, the high volatility of tetrachloroethylene leads to increased
ambient air concentrations near points of use (AT SDR, 1997a: U.S. EPA, 1996b). Specific to
early lifestage exposure scenarios, elevated ambient air concentrations include measurements
taken outside of a day care center adjacent to a dry cleaner (NYSDOH, 2005c) and on a
playground near a factory (Monster and Smolders, 1984). It should be noted that outdoor
concentrations can vary widely within a period of a few hours as a function of wind velocity and
direction, precipitation, humidity, and sunlight. ATSDR (1997a) reported mean
tetrachloroethylene concentrations of 8.8 |ig/m3 in areas close to points of release.
2-1
-------
Table 2-1. Physical and chemical properties of tetrachloroethylene
Property
Molecular formula
Molecular weight
Color
Physical state
Melting point
Boiling point
Density at 20° C
Density at 25 °C
Odor
Odor threshold: water
Odor threshold: air
Solubility: water at 25°C
Solubility: organic solvent(s)
Partition coefficients: Log K0w
Partition coefficients: Log K0c
Vapor pressure at 25°C
Henry's law constant at 25 °C
Autoignition temperature
Flashpoint
Flammability limits
Conversion factors, air
Explosive limits
Information
C2C14
165.83
Colorless
Liquid (at room temperature)
-19°C
121°C
1.6227 g/mL
No data
Ethereal
0.3 ppm
1 ppm
150 mg/L
Miscible with alcohol, ether,
chloroform, benzene, solvent
hexane, and most of the fixed
and volatile oils
3.4
2.2-2.7
18.47mmHg
1.8 x 10"2 atm-nvVmol
No data
None
Nonflammable
1 mg/L= 14 1.4 ppm
1 ppm = 6.78 mg/m3
No data
Reference
HSDB (2001)
Lide (1990)
Sax and Lewis (1987)
Sax and Lewis (1987)
Lide (1990)
Lide (1990)
Lide (1990)
HSDB (2001)
U.S. EPA (1987)
U.S. EPA (1987)
HSDB (2001)
HSDB (2001)
HSDB (2001)
Seip et al. (1986)
Zytner et al. (1989)
HSDB (2001)
Gossett (1987)
HSDB (2001)
HSDB (2001)
HSDB (2001)
Source: ATSDR (1997a).
EPA has carried out modeling to characterize the geographic distribution of
tetrachloroethylene for its National-Scale Air Toxics Assessment database (U.S. EPA, 1996b).
Median census tract-based tetrachloroethylene concentrations across the United States were
estimated at about 0.3 |ig/m3 for urban areas and 0.1 |ig/m3 for rural areas (75% upper
percentiles of 0.4 and 0.2 |ig/m3, respectively). The California Air Resources Board (CARS,
1998) reported a statewide median air concentration of 0.3 |ig/m3 in 2001, which represents the
lowest value in what has been a decreasing trend since 1990. Note that these averages, which are
2-2
-------
based on geographic areas, only characterize the likely exposure of individuals who spend an
equal amount of time in all parts of the defined area, and they may, therefore, significantly
underestimate the exposure of individuals who consistently spend time in subareas that have
higher tetrachloroethylene concentrations.
Near points of use, such as dry cleaners or industrial facilities, indoor exposure to
tetrachloroethylene is more significant than outdoor exposure (U.S. EPA, 200la). Adgate and
colleagues measured tetrachloroethylene in outside and indoor air at school, indoor air at home,
and using personal samplers on children, and demonstrated that tetrachloroethylene levels are
lower in homes with greater ventilation and in homes in non-urban settings (Adgate et al., 2004b:
Adgate et al., 2004a). Mean indoor air concentrations in apartments above dry cleaning shops of
4.9 mg/m3 have been reported [Altmann et al. (1995): also refer to McDermott et al. (2005):
Schreiber et al. (2002): Garetano and Gochfeld (2000): Schreiber et al. (1993)1. Measurements
have also been made in a day care center adjacent to a dry cleaners (NYSDOH, 2005a, b, c) and
in a classroom exposed to tetrachloroethylene from an air "emission from a small chemical
factory" (Monster and Smolders, 1984). Mean concentrations inside dry cleaning facilities were
reported to be 454-1390 mg/m3 in the United States and 164 mg/m3 in Nordic countries during
the 1960s and 1970s. Overall levels declined from 95-210 mg/m3 in the 1980s to 20-70 mg/m3
over the next decades in these countries (Lynge et al., 2011: Gold et al., 2008: Lynge et al.,
2006).
The off-gassing of garments that have recently been dry-cleaned may be of concern
[Tichenor et al. (1990): also refer to Thomas et al. (1991)]. In the home, tetrachloroethylene
vapors may off-gas from the clothes of occupationally exposed individuals, or they may come
directly from the exhaled breath of exposed workers [ATSDR (1997a): also refer to Aggazzotti
et al. (1994a: 1994b)]. Relatively high tetrachloroethylene air concentrations have been
measured in the proximity of freshly dry-cleaned clothing stored in small, close spaces. A
residential closet storing newly dry-cleaned clothing had an air concentration of 2.9 mg/m3 after
1 day, which rapidly declined to 0.5 mg/m3 and persisted for several days (Tichenor et al., 1990).
There is one documented mortality case: a 2-year-old boy was found dead after being put to
sleep in a room with curtains that had been incorrectly dry-cleaned (Gamier et al., 1996).
Dry-cleaned garments transported in an automobile may also lead to unexpectedly high
levels of exposure. Park et al. (1998) used simulated driving cycles to estimate the
concentrations of several contaminants emitted from in-vehicle sources; also refer to Gulyas and
Hemmerling (1990). Using dry-cleaned clothes as a source, tetrachloroethylene levels inside a
stationary vehicle after 30 minutes reached 0.230 mg/m3. Approximating these exposures is not
easy because specific exposure levels would depend on many factors: car velocity, wind speed,
ventilation, and time spent in the automobile. Another study demonstrating exposure in a car
2-3
-------
found that transporting a freshly dry-cleaned down jacket in a car resulted in a cabin air
concentration of 24.8 mg/m3 after 108 minutes (Chien, 1997).
Air exposure may also occur during showering or bathing as dissolved
tetrachloroethylene in the warm tap water is volatilized. Rao and Brown (1993) used an adult
physiologically based pharmacokinetic (PBPK) model combined with a microenvironmental
exposure model to estimate the dose received by inhalation exposure during showering and
bathing as well as by dermal exposure to the water. The tap water concentration of
tetrachloroethylene was 1 mg/L, which is probably a higher concentration than exists in most
water supplies. The authors also demonstrated that a majority of the tetrachloroethylene in the
blood, as a result of their bathing scenario, resulted from inhalation exposure, while about 15%
resulted from dermal absorption.
2.2.2. Water
Because of its relatively low aqueous solubility (refer to Table 2-1), it is not likely that
volatilized tetrachloroethylene will enter surface or rain water. However, it has been detected in
drinking water, ground water, and surface water (U.S. EPA, 200 la: AT SDR, 1997'a:
Environment Canada & Health Canada, 1993; Lagakos etal., 1986). Most of this contamination
is probably due to release in water following industrial use or by public use of consumer
products. Therefore, unless a surface water body is in the vicinity of a highly contaminated site,
surface waters are expected to have a lower concentration of tetrachloroethylene than ground
water.
In areas near sources of contamination, ground water and surface water concentrations
can be considerably higher than average. Because the density of tetrachloroethylene is about
60% higher than that of water, tetrachloroethylene is expected to accumulate near the bottom of a
stagnant receiving water body after a large-volume point discharge. Water samples collected
near the bottom of the St. Clair River near Sarnia, Ontario, downstream from several petroleum-
based production facilities, contained tetrachloroethylene concentrations ranging from 0.002 to
34.6 |ig/L (Environment Canada & Health Canada, 1993). The concentrations in 17 samples of
surface water from the lower Niagara River in New York State in 1981 averaged 0.036 |ig/L
(with a maximum of 0.134 |ig/L) (Environment Canada & Health Canada, 1993).
Exposure models have been developed to predict the fate and transport of organic
compounds such as tetrachloroethylene in environmental media, including air, water, and soil.
The outputs from two similar but independently developed environmental exposure models,
CalTOX and FugSONT, were compared for a scenario designed to reproduce a residential area
near an industrial contamination site (Maddalena et al., 1995), in which 75 moles/day are
released into the air and 0.7 moles/day are released into surface water. Although the soil
2-4
-------
predictions differed, the predictions of tetrachloroethylene in air and ground water were similar,
with the concentration of air predicted by CalTOX approximately 6 |ig/m3 and the surface water
concentration, 82 |ig/L. It should be noted that agreement of the models does not confirm the
validity of either one but lends some support to the usefulness of the results.
The off-gassing of tetrachloroethylene from a drinking water supply can result in
exposure. In 1976, EPA measured tetrachloroethylene levels ranging from 800 to 2,000 |ig/L in
drinking water samples in Massachusetts (Paulu et al., 1999). Similar levels were reported
elsewhere in New England. These concentrations were attributed to the vinyl-lined asbestos-
cement pipes that were used to carry water in this area (Webler and Brown, 1993). Letkiewicz et
al. (1982) estimated that 53% of newborn infants are formula-fed from drinking water sources,
and the other 47% receive all of their fluid from breast milk. Taking into account volatilization
during boiling of water, they indicate that the uptake of tetrachloroethylene in formula-fed
infants on a mg/kg-day basis is 10 times higher than in adults with the same level of drinking
water contamination. In addition, incidental water consumption may occur for children when
swimming or bathing (U.S. EPA. 2008).
Although dermal exposure is possible via contaminated tap water during showering,
bathing, or swimming, this is generally not considered a major route of exposure (Poet et al.,
2002: Nakaietal.. 1999: Stewart and Dodd, 1964). Rao and Brown (1993) demonstrated that
only 15% of the tetrachloroethylene in the blood resulted from dermal exposure as compared to
inhalation of vapors.
2.2.3. Food
Certain foods have been found to be contaminated with tetrachloroethylene [U.S. EPA
(200la): also refer to Daft (1988): Heikes and Hopper (1986): McConnell et al. (1975)1.
Because of the lipophilic nature of tetrachloroethylene, it may bind to lipid molecules in such
foods as margarine, oils, meats, and other fatty foods stored in areas where there is
tetrachloroethylene in the air (U.S. EPA. 200la: Schreiber. 1997). In 1988, elevated
tetrachloroethylene levels were observed in margarine and butter samples obtained from grocery
stores located near dry cleaning facilities [Entz and Diachenko (1988): also refer to Uhler and
Miller (1988)1. Further studies confirmed that close proximity to a dry cleaning facility was
associated with elevated tetrachloroethylene levels in butter samples (Kacew and Lambert,
1997). Nonetheless, food is not considered to be a major exposure pathway. Other sources of
information about tetrachloroethylene in foods include the Food and Drug Administration (2003)
and Fleming-Jones and Smith (2003).
2-5
-------
2.2.4. Soil
Where contamination occurs, tetrachloroethylene can be measured in soil (U.S. EPA,
200 la). This pathway for ingestion of tetrachloroethylene has not been directly examined. A
clear need exists to evaluate this pathway, particularly for children with pica, who can ingest
high quantities of contaminated soil through hand-to-mouth activity, as has been shown for lead
(U.S. EPA. 2008).
2.2.5. Breast Milk
Due to its lipid solubility, tetrachloroethylene can concentrate in human breast milk
(Schreiber et al.. 2002: U.S. EPA. 200la: NYSDOH, 2000: Schreiber, 1997. 1993: Sheldon et
al., 1985: Pellizzari etal., 1982: Bagnell and Ellenberger, 1977), as well as in milk from cows
(Wanner et al., 1982), goats (Hamada and Tanaka, 1995), and rats (Byczkowski et al., 1994:
Byczkowski and Fisher, 1994). Breast milk can contain high concentrations of
tetrachloroethylene and some of its toxic metabolites. Reported levels of tetrachloroethylene in
breast milk have ranged up to 43 jig/L in the general population (U.S. EPA, 200la). In one case
study, the breast milk of a woman was found to contain 10 mg/L of tetrachloroethylene 1 hour
following a visit to her husband at his work in a dry cleaning establishment. This concentration
dropped to 3 mg/L after 24 hours. Her child suffered from obstructive jaundice and
hepatomegaly, but these conditions improved when breastfeeding was discontinued (Bagnell and
Ellenberger, 1977).
Physiologically based pharmacokinetic (PBPK) models have been utilized to estimate
tetrachloroethylene doses from milk to the human infant (Fisher et al., 1997: Byczkowski and
Fisher, 1995: Schreiber, 1993), and rat (Byczkowski et al., 1994). Schreiber (1993) used a
PBPK model to estimate the dose a nursing infant might receive from an exposed mother's
breast milk. Using different exposure scenarios, Schreiber (1993) predicted that human breast
milk concentrations could range from 1.5 mg/L for a typical residential scenario, 16-3,000 mg/L
for a residential scenario near a dry cleaner, and to 857-8,440 mg/L for an occupational scenario.
Assuming that a 7.2 kg infant ingests 700 mL of breast milk per day, Schreiber estimated the
dose to the infant could range from 0.0001 to 0.82 mg/kg/day. Actual indoor air concentrations
(24-hr average), as measured in apartments in New York State, were used to predict potential
levels in breast milk in these modeling scenarios. The apartments included one located above a
dry cleaning facility that used an old dry-to-dry machine (average concentration, 45.8 mg/m3),
three located above facilities that used transfer machines (average concentration, 7.7 mg/m3), and
two located above facilities that used newer dry-to-dry machines (average concentration, 0.25
mg/m3) (Schreiber, 1993). The predicted breast milk concentrations in these scenarios ranged
from 16 to 3,000 |ig/L. Assuming that a 7.2 kg infant ingests 700 mL of breast milk per day,
2-6
-------
Schreiber (1993) determined that the infant dose from milk could range from 0.0015 to 0.3
mg/kg-day.
Using the same exposure conditions as Schreiber (1993), Byczkowski and Fisher (1995)
predicted lower doses to the infant (0.0009-0.202 mg/kg-day). Using milk production and
suckling variables, Fisher et al. (1997) estimated the dose that a human infant might receive after
maternal occupational exposure to be 1.4 mg/day.
Ingestion through breast milk and infant exposures is discussed further in Section 4.9.
However, Schreiber (1997) has suggested that if infants live adjacent to or in close proximity to
dry cleaning facilitates, the dose received through breast milk ingestion will be insignificant
when compared with that from their inhalation exposure.
2.2.6. Direct Ingestion
In rare circumstances, direct ingestion of tetrachloroethylene has been documented. A
6-year-old boy who directly ingested 12-16 g tetrachloroethylene experienced drowsiness,
vertigo, agitation, and hallucinations. He then lost consciousness and went into a coma, and later
recovered (Koppel et al., 1985). Follow-up testing on the boy was not reported, so any potential
long-term effects of the exposure are unknown.
2-7
-------
3. TOXICOKINETICS
3.1. ABSORPTION
Tetrachloroethylene is rapidly absorbed into the bloodstream following oral and
inhalation exposures. It can also be absorbed across the skin following dermal exposure to either
pure or diluted solvent or vapors (Poet et al., 2002; Nakai etal., 1999; Stewart and Dodd, 1964).
3.1.1. Inhalation
The major exposure route for tetrachloroethylene is considered to be inhalation.
Pulmonary uptake of tetrachloroethylene is rapid; however, complete tissue equilibrium occurs
only after several hours. Absorption into the systemic circulation through pulmonary uptake is
proportional to the ventilation rate, the duration of exposure, and the concentration in the
inspired air (Monster et al., 1979; Hake and Stewart, 1977).
Chiu et al. (2007) reported that peak levels of tetrachloroethylene in venous blood and air
occurred near the end of a 6-hour inhalation exposure to 1 ppm and declined thereafter. In the
Monster et al. (1979) study, uptake after 4 hours was 75% of its value at the onset of exposure.
Increased physical activity increases uptake but lowers the alveolar partial pressure, thus
removing more tetrachloroethylene from the alveoli, resulting in a longer time to reach tissue
equilibrium (Pezzagno et al., 1988).
The blood:gas partition coefficient for tetrachloroethylene describes how the chemical
will partition itself between the two phases. Specifically, it is the ratio of concentrations at
steady state; i.e., when all rates are constant after equilibrium has been reached. Reported values
for the coefficient in humans range from around 10-20 [e.g.,(Reitz et al., 1996; Byczkowski and
Fisher. 1994: Gearhart et al.. 1993: Hattis et al.. 1990: Ward etal.. 1988: Droz and Guillemin,
1986)1, meaning that if tetrachloroethylene is in equilibrium, the concentration in blood will be
10-20 times higher than the concentration in the alveoli.
Opdam and Smolders (1986) determined concentrations of tetrachloroethylene in alveolar
air for 1-60-second residence times (the time interval from the beginning of an inhalation to the
end of the next inhalation) for six volunteers exposed to 0.5-9.8 ppm of chemical for
1-60 minutes. These investigators found the concentrations of tetrachloroethylene in alveolar air
to decrease with residence times for breaths during exposure periods but to increase during
postexposure for residence times less than 10 seconds. Alveolar air tetrachloroethylene
concentration correlated with the concentrations in pulmonary artery mixed venous blood.
5-1
-------
Like the studies in humans, inhalation studies in laboratory animals provide clear
evidence that tetrachloroethylene is readily absorbed via the lungs into the systemic circulation
(Dallas et al.. 1994b: Peggetal.. 1979).
3.1.2. Oral
Gastric absorption of tetrachloroethylene occurs at a relatively rapid rate and is
essentially complete. Close to 100% of oral doses are absorbed from the gut, according to
reports of several studies conducted in mice, rats, and dogs (Dallas et al., 1995, 1994b: Frantz
andWatanabe, 1983; Schumann et al., 1980; Peggetal., 1979). Absorption into the systemic
circulation was indicated by blood tetrachloroethylene levels of 21.5 |ig/mL following accidental
ingestion of the chemical by a 6-year-old boy (Koppel etal., 1985).
3.1.3. Dermal
Absorption of tetrachloroethylene by humans following dermal exposure to vapors of the
chemical has been reported to be relatively insignificant (only 1%) when compared with
absorption via inhalation of vapors (Nakai et al., 1999; Riihimaki and Pfaffli, 1978). The
amount of chemical absorbed during the immersion of one thumb in liquid tetrachloroethylene is
equivalent to the uptake during inhalation of 10-15 ppm of the compound for the same time
period (Stewart and Dodd, 1964).
Studies in animals confirm that dermal uptake of tetrachloroethylene following vapor
exposure is minimal when compared with pulmonary uptake (McDougal et al., 1990; Tsuruta,
1989), whereas dermal uptake is greater following direct skin application (Jakobson et al., 1982).
Notably, the conclusions of Bogen et al. (1992), based on the results of their study in hairless
guinea pigs, indicate that dermal absorption of tetrachloroethylene from contaminated water
supplies could be an important route of exposure for humans. These investigators estimated that
a standard 70 kg man with 80% of his body immersed in water would completely absorb the
amount of tetrachloroethylene in 2 L of that water.
3.2. DISTRIBUTION AND BODY BURDEN
Once absorbed, tetrachloroethylene is distributed by first-order diffusion processes to all
tissues in the mammalian body. The highest concentrations of tetrachloroethylene are found in
adipose tissue due to the lipophilicity of the compound (Savolainen et al. 1977; Monster et al.
1979; Dallas et al. 1994). Concentrations of tetrachloroethylene reach higher levels in brain and
liver than in many other tissues (Gamier et al., 1996; Levine et al., 1981; Lukaszewski, 1979).
Absolute tissue concentrations are directly proportional to the body burden or exposure dose.
Due to its lipid solubility, tetrachloroethylene is also concentrated in milk, and it has been
5-2
-------
measured in human breast milk (Schreiber et al.. 2002: NYSDOH, 2000: Schreiber, 1997. 1993).
Higher concentrations occur in milk having higher fat content. For example, the milk:blood
partition coefficient is equal to 12 for rats but 2.8 for humans (Byczkowski and Fisher, 1994).
This noticeable difference reflects the higher fat content of rat milk. Tetrachloroethylene readily
crosses both the blood:brain barrier and the placenta. Partition coefficients for various tissues,
relative to blood or air, have been reported by several investigators (Byczkowski and Fisher,
1994: Dallas et al.. 1994b: Gearhart et al.. 1993: Ward et al.. 1988). Section 3.5 presents
examples of these.
Repeated daily inhalation exposures of human volunteers to tetrachloroethylene indicate
accumulation of the compound in the body, which is thought to be due to its high lipid solubility.
Because of its long residence time in adipose tissue, repeated daily exposure results in an
accumulated concentration; tetrachloroethylene from new exposures adds to the residual
concentration from previous exposures until steady state is reached. Blood levels of
tetrachloroethylene increase over several days with continued daily exposures. Following
cessation of these exposures, it is still present in the blood. Exhalation of the compound
continues over a number of days due to its slow release from the adipose tissue (Skender et al.,
1991: Altmann et al., 1990: Stewart et al., 1977). For a given concentration in blood or air, the
half-time—the time necessary to equilibrate the adipose tissue to 50% of its final
concentration—is about 25 hours (Monster, 1979: Fernandez et al., 1976). Therefore, during a
single 8-hour exposure, adipose tissue does not reach steady-state equilibrium.
Tetrachloroethylene uptake by fatty tissue during the working hours of the week is
countered by the elimination that occurs during nonexposure times of nights and weekends; thus,
for persons exposed to tetrachloroethylene on a 5-day-a-week work schedule, an equilibrium is
eventually established, but it requires a time period of 3-4 weeks of exposure for adipose tissue
to reach plateau concentrations.
Animal studies provide clear evidence that tetrachloroethylene distributes widely to all
tissues of the body, readily crossing the blood:brain barrier and the placenta (Dallas et al., 1994a;
Ghantous et al., 1986; Schumann et al., 1980; Savolainen et al., 1977a). Following exposure of
rats to tetrachloroethylene, the compound has been measured in blood, fat, brain, lungs, liver,
kidneys, heart, and skeletal muscle (Dallas et al., 1994a; Savolainen et al., 1977a). The highest
tissue concentrations were found in adipose tissue (60 or more times blood level) and in the brain
and liver (4 and 5 times that found in the blood, respectively), as can be calculated from the rat
tissue-distribution data of (Dallas et al., 1994a; Savolainen et al., 1977a). Dallas et al. (1994a)
found the concentration of tetrachloroethylene in fat to be 9-18 times the concentrations found in
nonfat tissues. Skeletal muscle contained the lowest concentration. In one human fatality case,
the concentration of tetrachloroethylene in the brain was 120 times higher than concentrations
5-3
-------
measured in the lung. In another case, the concentrations in the liver were 8, 3.4, and 3.5 times
higher, respectively, than concentrations measured in the lung, kidney, and brain (Levine et al.,
1981).
3.3. METABOLISM
This section describes the metabolism of tetrachloroethylene, identifying metabolites
thought to be causally associated with toxic responses as well as those used to evaluate the flux
of parent compound through the known metabolic pathways. Sex- and species-dependent
differences in the metabolism of tetrachloroethylene and potential contributors to interindividual
differences are identified. Refer to Section 4.9 for further discussion of how these factors affect
variability and susceptibility.
3.3.1. Introduction
The metabolism of tetrachloroethylene has been studied mostly in mice, rats, and humans
[for reviews, refer to (Lash and Parker. 2001: Anders etal.. 1988: Dekantetal.. 1987)1.
Tetrachloroethylene is metabolized in laboratory animals and in humans through at least
two distinct pathways: oxidative metabolism via the cytochrome P450 (CYP [also abbreviated as
P450]) mixed-function oxidase system and glutathione (GSH) conjugation followed by
subsequent further biotransformation and processing, either through the cysteine conjugate P-
lyase pathway or by other enzymes including flavin-containing monooxygenase 3 (FMO3) and
CYP3A (Lash and Parker. 2001: Lash etal.. 1998: Volkeletal.. 1998: Birneretal.. 1996:
Anders etal.. 1988: Dekantetal.. 1987: Costa and Ivanetich. 1980: Filser and Bolt 1979: Pegg
et al., 1979: Daniel, 1963). The conjugative pathway is lexicologically significant because it
yields relatively potent toxic metabolites (Lash and Parker, 2001: Werner etal., 1996: Vamvakas
etal.. 1989a. c; Vamvakas et al.. 1989d: Anders etal.. 1988: Vamvakas et al.. 1987: Dekant et
al., 1986a: Dekantetal., 1986c). Figure 3-1 depicts the overall scheme of tetrachloroethylene
metabolism. Known metabolites presented in this figure are identified by an asterisk.
5-4
-------
/ V
Cl 5 Cl
O
XX
Cl 6
O
Cl
CO, CO,
Cl
Cl
Cl
Cl.
Cl
Cl
P450
Cl
J-CI
GST
Cl
Cl
Cl
SG
o
J*
CI3C
Cl
H20
CI3C
4" OH
p-lyase
P450I
[Reactive species]
CI2CH
[Reactive species]
— OH
Figure 3-1. Postulated scheme for the metabolism of tetrachloroethylene by
the cytochrome P450 (P450) oxidative pathway and glutathione ^-transferase
(GST)-mediated glutathione (GSH) conjugation pathway.
PCE and identified (*) urinary metabolites: (1) PCE, (2) PCE-Fe-O intermediate, (3)
trichloroacetyl chloride, (4) trichloroacetic acid, (5) PCE oxide, (6) ethandioyl dichloride,
(7) oxalic acid, (8) S-(l,2,2-trichlorovinyl) glutathione (TCVG), (9) S-(l,2,2-trichlorovinyl)-Z-
cysteine (TCVC), (10) A^-acetyl trichlorovinyl cysteine (NAcTCVC), (11) dichloroacetic acid.
Enzymes: cytochrome P450 (P450), GST, gamma-glutamyltransferase (GOT), dipeptidase (DP),
(3-lyase, flavin mono-oxygenase-3 (FMO3), 7V-acetyl transferase (NAT).
Sources: Adapted from Pegg et al. (1979). Costa and Ivanetich (1980). Dekant et al. (1986a). Lash
and Parker (2001). Yoshioka et al. (2002). Chiu et al. (2007).
5-5
-------
3.3.2. Extent of Metabolism
Studies in both animals and humans indicate that overall metabolism of
tetrachloroethylene is relatively limited—particularly at higher exposures [reviewed in (Lash and
Parker, 2001)], as evidenced by the high percentage of absorbed dose excreted in the breath as the
parent molecule (Chiu et al.. 2007: Volkel etal.. 1998: Buben and O'Flahertv, 1985: Frantz and
Watanabe, 1983: Monster et al.. 1983: Ohtsuki et al.. 1983: Schumann et al.. 1980: Filser and
Bolt 1979: Monster. 1979: Monster et al.. 1979: Peggetal.. 1979: Fernandez et al.. 1976: Essing
etal.. 1973: Ikeda and Ohtsuji, 1972: Stewart et al.. 1970: Boettner and Muranko, 1969: Daniel
1963: Stewart et al., 1961: Yllner, 1961). Because of its high lipid solubility, tetrachloroethylene
can be sequestered in fat and, thus, not all metabolism is evident in short sampling time periods.
The extent of metabolism after inhalation exposure in humans has been estimated by
measuring trichloro-compounds excreted in the urine and exhalation of tetrachloroethylene in
expired air (Monster et al., 1983: Monster and Houtkooper, 1979: Monster et al., 1979:
Fernandez et al., 1976: May, 1976: Essing et al., 1973: Bolanowska and Golacka, 1972: Ikeda et
al.. 1972: Stewart et al.. 1970: Boettner and Muranko, 1969: Stewart et al.. 1961). Several
studies reported only about 1-3% of the estimated amounts inhaled were metabolized to
trichloroacetic acid (TCA) and other chlorinated oxidation products, although additional
tetrachloroethylene—as much as 20% or more of the dose—may be metabolized over a longer
period (Bois et al., 1996: Bogen et al., 1992: Monster et al., 1979). For example, Chiu et al.
(2007) noted that although an average of 0.4% of tetrachloroethylene intake (1 ppm for 6 hours)
was recovered in urine as TCA, total recovery in urine and exhaled air accounted for, on average,
only 82% of intake. This would imply that 18% were metabolized, but Chiu et al. (2007) noted
substantial uncertainty and variability in these calculations and concluded that they were
consistent with previous studies at higher exposures. Interestingly, Chiu et al. (2007) also noted
significant variability among the seven subjects and among the four occasions, contributing to
the uncertainty in measurements. A literature review published by Hattis et al. (1990) reported
estimates of the fraction of tetrachloroethylene metabolized at a low dose of 1 ppm to range from
2-86%. Based on data from Monster et al. (1979), Bois and colleagues (Chiu and Bois, 2006:
Bois et al., 1996) used physiologically based pharmacokinetic (PBPK) modeling to predict that
at exposure levels near current occupational standards, a median of approximately 1.5% of
inhaled tetrachloroethylene would be metabolized, whereas, at ambient air levels (0.001 ppm),
the median estimate would be considerably higher (23-36%).
The extent of metabolism in animals has been estimated by conducting excretion-balance
studies using isotopically labeled tetrachloroethylene. In rodents, 2-88% of the dose was
metabolized, depending on dose level and species: the higher the dose, the smaller the percentage
metabolized. Rats metabolized a lower percentage of a given tetrachloroethylene body burden
-------
than did mice (Frantz and Watanabe, 1983; Schumann et al., 1980; Filser and Bolt, 1979; Pegg et
al., 1979; Daniel, 1963; Yllner, 1961). Urinary excretion data from studies by Filser and Bolt
(1979) and Buben and O'Flaherty (1985) suggest that metabolism of tetrachloroethylene is
greater in mice than in rats.
3.3.3. Pathways of Metabolism
The two known biotransformation pathways for tetrachloroethylene metabolism are
(1) oxidation by CYP enzymes and (2) conjugation with GSH followed by further processing of
the conjugate through various pathway bifurcation branches. As shown in Figure 3-1, the initial
step in the metabolism of tetrachloroethylene is formation of an Fe-O intermediate for the
oxidative pathway or conjugation with GSH for the secondary pathway (Yoshioka et al., 2002;
Lash and Parker. 2001: Dekantetal., 1998: Lash etal., 1998: Dekantetal., 1987: Dekant et al..
1986c: Miller and Guengerich, 1983. 1982: Costa and Ivanetich. 1980). It is possible that other
yet unrecognized pathways for tetrachloroethylene could exist in humans (Bois etal., 1996:
Monster et al.. 1979: Sakamoto. 1976).
3.3.3.1. Cytochrome P450-Dependent Oxidation
3.3.3.1.1. Oxidative metabolites
In vivo, the major excretory metabolite of the oxidative pathway, TCA, is excreted in the
urine of all species tested (Volkel et al., 1998: Birner et al.. 1996: DekantetaL. 1987: Ohtsuki et
al.. 1983: Leibman and Ortiz. 1977. 1970: Daniel. 1963: Yllner. 1961). Oxalic acid has been
reported to be a relatively major urinary metabolite in rats (Pegg et al., 1979: Dmitrieva, 1967).
Oxalic acid might either arise from action of microsomal epoxide hydrase on the epoxide
intermediate or may be a separate product from the initial Fe-O intermediate. The oxalate
metabolite excretory product may also be derived from dichloroacetic acid (DCA) or
monochloroacetic acid (Tonget al., 1998a, b). Pulmonary excretion of carbon dioxide (CO2)
amounting to <10% of the administered dose has been identified in exhaled breath from rodents
exposed to 14C-labeled tetrachloroethylene (Frantz and Watanabe, 1983: Schumann et al., 1980:
Pegg etal., 1979), accounting for less than either exhaled tetrachloroethylene or urinary
metabolites.
Trichloroethanol (TCOH) has been detected in the urine of subjects exposed to
tetrachloroethylene in some studies (Schreiber et al., 2002: Birner et al., 1996: Monster et al.,
1983: Weichardt and Lindner. 1975: Ikeda and Ohtsuii. 1972: Ikedaetal.. 1972: Ogata et al..
1971: Tanaka and Ikeda, 1968: Ogata etal., 1962), but it could not be identified by others (Chiu
et al., 2007: Volkel etal., 1998: Buben and O'Flahertv, 1985: Costa and Ivanetich, 1980:
Monster et al., 1979: Hake and Stewart, 1977: Fernandez et al., 1976: Daniel, 1963: Yllner,
5-7
-------
1961). Most of the studies reporting TCOH have involved occupational or environmental
exposures in which there may be simultaneous exposure to trichloroethylene, for which TCOH is
a major urinary metabolite. In vitro, TCA—and, to a lesser extent—oxalic acid, but not TCOH,
are detected in incubations of tetrachloroethylene in rat microsomal protein (e.g., Yoshioka et al.,
2002). Thus, it appears likely that the reports of TCOH following tetrachloroethylene exposure
may have been artifacts of the analytical methodology used or of simultaneous trichloroethylene
exposure. Because TCOH is clearly not a significant metabolite for tetrachloroethylene, very
little, if any, TCA produced from tetrachloroethylene metabolism is likely to come through
chloral, either directly or indirectly through TCOH (Lash and Parker, 2001).
It was initially proposed that the first step in the oxidation of tetrachloroethylene is
hypothesized to yield 1,1,2,2-tetrachloroethylene oxide, a relatively unstable epoxide (Miller and
Guengerich, 1983, 1982; Costa and Ivanetich, 1980). Although an initial epoxide metabolite has
not been unequivocally demonstrated for tetrachloroethylene, evidence for this epoxide does
exist. The epoxide has been chemically synthesized (Kline etal., 1978; Bonse et al., 1975;
Frankel etal., 1957). The potential fates of tetrachloroethylene epoxide include trichloroacetyl
chloride, oxalate dichloride through tetrachloroethylene glycol, trichloroacetyl aminoethanol,
and, possibly, chloral hydrate (in equilibrium with chloral) (Pegg et al., 1979; Henschler and
Bonse, 1977; Bonse and Henschler, 1976).
However, recent data (Yoshioka et al., 2002) favor the hypothesis that the epoxide is not
an obligatory intermediate to formation of trichloroacetyl chloride. In particular, the pattern of
products of tetrachloroethylene oxide hydrolysis reported by Yoshioka et al. (2002) is dominated
by carbon monoxide (CO) and carbon dioxide (CO2), which differs markedly from the products
of oxidation in vivo or in vitro. Because TCA is believed to be derived primarily from
trichloroacetyl chloride (through hydrolysis or through reaction with amino groups of cellular
proteins), this would favor the hypothesis that the epoxide is a minor product of
tetrachloroethylene oxidation. Instead, the Fe-O intermediate is postulated to collapse via
chlorine migration to yield predominantly trichloroacetyl chloride (Yoshioka et al., 2002).
DCA has been identified as a tetrachloroethylene urinary metabolite (Volkel etal., 1998;
Dekant et al., 1987; Yllner, 1961) and may arise as a product of further metabolism of TCA or as
a result of p-lyase bioactivation of GSH conjugation metabolites. The major organ site of DCA
production is likely to differ for each pathway, with DCA arising from oxidative metabolism
primarily in the liver and from GSH-dependent metabolism products mostly in the kidney.
Dechlorination of TCA to DCA is catalyzed by gut contents (ingested food and bacteria) of the
rat and mouse (Moghaddam et al., 1996): isolated mouse microflora have been shown to convert
TCA to DCA (Moghaddam et al., 1997). However, data indicate that this does not contribute to
DCA detected systemically after trichloroethylene exposure, and a similar conclusion is
-------
reasonable for tetrachloroethylene, given the lower rate of formation of TCA from
tetrachloroethylene as compared to trichloroethylene. In addition, data from trichloroethylene
suggest that for that compound, DCA formation is likely dominated by hydrolysis of
dichloroacetyl chloride—rather than dechlorination of TCA. As compared to tetrachloroethylene
exposure, trichloroethylene exposure leads to higher amounts of TCA, in conjunction with the
lower amounts of DCA, detectable in blood or urine. This is inconsistent with dechlorination of
TCA being the origin of DCA detected in urine after tetrachloroethylene exposure and supports
the hypothesis that DCA is derived predominantly from GSH conjugation of tetrachloroethylene.
3.3.3.1.2. Species differences
Although thought to be qualitatively similar, there are clear differences among species in
the quantitative aspects of tetrachloroethylene metabolism (Lash and Parker, 2001; Volkel et al.,
1998; Schumann et al., 1980; Ikeda and Ohtsuji, 1972). These differences are in the relative
yields and kinetic behavior of metabolites (Volkel et al., 1998; Green etal., 1990; Ohtsuki et al.,
1983). Rodents and humans differ in relative rates of tetrachloroethylene metabolism in key
target organs, in the doses at which saturation of metabolism occurs, and in the half-times in the
body.
The rate of metabolism of tetrachloroethylene is faster in rodents than in humans, and
higher metabolite concentrations in blood are obtained in rodents as compared with humans.
The higher blood levels of metabolites in rodents are particularly noticeable at the higher
tetrachloroethylene exposure levels because saturation is approached at lower exposure levels in
humans than in rodents. The half-life in the body of these metabolites is, however, noticeably
longer for humans than for rodents (144 hours in humans vs. approximately 10 hours or less in
rodents). It is for this reason that examinations of tetrachloroethylene concentration and toxicity
associations must reflect both blood concentration and time-integrated dose metrics such as area-
under-the-curve (AUC).
A study of species differences in tetrachloroethylene metabolism conducted by Dekant
and colleagues is presented in Volkel et al. (1998). These investigators compared both oxidative
and GSH-dependent metabolism in rats and humans exposed for 6 hours to 10, 20, or 40 ppm
tetrachloroethylene by inhalation. Rats were also exposed to 400 ppm concentrations. TCA was
the major urinary excretion product in both species; however, the elimination half-time was more
than four times slower in humans than in rats. Blood plasma concentrations of the metabolite
were higher (three-eightfold, depending on the dose) in rats than in humans exposed to identical
air concentration levels. These observations are in agreement with metabolic rates, in general,
which are higher in mice than in rats; rats, in turn, have higher metabolic rates than do larger
animals, including humans. Dekant and his coworkers also reported urinary excretion of DCA
5-9
-------
by rats—but not humans. They concluded most of the DCA resulted from GSH-dependent
metabolism. DCA, however, is further metabolized by GST enzymes, which, in turn, limit its
detectability in urine.
3.3.3.1.3. Cytochrome P450 (CYP) isoforms and genetic polymorphisms
Oxidative metabolism of tetrachloroethylene, irrespective of the route of administration,
occurs predominantly in the liver but also at other sites. For example, the kidneys exhibit
cytochrome P450 enzyme activities, mostly in the proximal tubules, although total activity is
markedly less than in the liver (Lash and Parker, 2001; Lash et al., 2001). The rat kidney
expresses CYP2E1 (Cummings et al., 1999; Speerschneider and Dekant 1995), although the
human kidney has not been shown to do so (Cummings et al., 2000a: Amet et al., 1997).
However, the human kidney expresses other CYP enzymes. CYP enzymes occurring in other
extrahepatic tissues—brain and lungs, for example—may also contribute to oxidative
metabolism of tetrachloroethylene.
Relatively few studies provide information about which specific CYP isoforms play a
role in tetrachloroethylene oxidative metabolism. CYP2E1 is presumed to have an important
role in tetrachloroethylene metabolism (Lash and Parker, 2001): however, the chemical-specific
data are too sparse to provide strong support for this assumption (Doherty et al., 1996).
CYP2B1/2 may also be important for the metabolism of tetrachloroethylene. CYP3A
isoenzymes may contribute to the generation of reactive sulfoxides from metabolites of the GSH
pathway (refer to text below). Costa and Ivanetich (1980) showed increased hepatic metabolism
following treatment with agents now known to induce these isoenzymes specifically.
Genetic polymorphisms are DNA sequence variations that result in changes in the protein
sequence of an enzyme that can alter the enzyme's ability to catalyze a reaction or alter the
expression of an allele. Polymorphisms are known for most of the CYP enzymes including
CYP2E1 (Huetal.. 1999: McCarver et al.. 1998) and CYP3A4 (Sata et al.. 2000: Westlind et al..
1999).
3.3.3.2. Glutathione (GSH) Conjugation Pathway
The GSH conjugation pathway was recognized much later than was the oxidative
pathway, yet it may be lexicologically influential (Lash and Parker, 2001). Similar to
trichloroethylene, GSH conjugation of tetrachloroethylene is associated with renal toxicity (Lash
and Parker. 2001: Lash et al.. 2000: Dekant etal.. 1989: Anders et al.. 1988).
3.3.3.2.1. Glutathione (GSH) conjugation metabolites
The initial conjugation of tetrachloroethylene with GSH occurs mainly in the liver (Green
etal.. 1990: Vamvakas et al.. 1989c: Dekant etal.. 1987: Vamvakas et al.. 1987). with transport
3-10
-------
of the conjugate and its cysteine counterpart to the kidney target organ for further processing.
This first step also occurs within the kidney (Lash et al., 1998). As shown in Figure 3-1,
tetrachloroethylene is initially conjugated with GSH to form ^-(l^l-trichlorovinyl) glutathione
(TCVG). This reaction, which is catalyzed by the GSH-S-transferase (GSTs) enzymes, a group
of enzyme isoforms, was traditionally considered to be a detoxification reaction, leading to more
water-soluble compounds that are more readily excreted. In many cases, however, as with
certain halogenated alkanes and alkenes such as tetrachloroethylene, GSH conjugation can be
important for bioactivation. TCVG is then processed through the cysteinylglycine conjugate
Ł-(1,2,2 trichlorovinyl)-Z-cysteinylglycine to ^-(l^^-trichloroviny^-Z-cysteine (trichlorovinyl
cysteine, or TCVC) by the enzymatic removal of glutamyl and glycine residues by gamma-
glutamyltransferase (GGT) and various membrane-bound dipeptidases known as
cysteinylglycine dipeptidase [reviewed by (Lash and Parker, 2001; Dekant etal., 1989; Anders et
al., 1988)]. These enzymes reside in tissues other than the kidneys (e.g., the brain), indicating a
potential for toxic reactive metabolite formation in those tissues as well. Conversion of TCVG
to TCVC by these cleavage enzymes leads to a critical bifurcation point of the GSH pathway
because the TCVC may be processed by certain enzymes to yield reactive, toxic chemical
species, although it may be metabolized via a different route to yield an excretory product (Lash
and Parker. 2001).
Importantly, the TCVC metabolite may also act as a substrate for renal p-lyases
[reviewed by (Lash and Parker. 2001: Lash et al.. 2000: Dekant etal.. 1989: Anders etal.. 1988:
Dekant etal., 1988)]. Renal P-lyases are known to cleave TCVC to yield an unstable thiol,
1,2,2-trichlorovinylthiol, that may give rise to cytotoxic and mutagenic reactive chemical species
that can form covalent adducts with cellular nucleophiles, including DNA and proteins (Volkel et
al., 1999: Dekant et al., 1990: Dekant etal., 1986a). In addition, DCA is a downstream urinary
excretion product of p-lyase bioactivation of TCVC and has been detected in urine of rats
exposed to tetrachloroethylene (Volkel etal., 1998). P-lyases are a family of pyridoxal
phosphate-containing enzymes that are located in several tissues besides the kidneys, including
liver and brain, and in intestinal flora, although their substrate specificities may vary. Hepatic P-
lyase is distinct from renal P-lyase and has not been found to have a role in TCVC metabolism.
P-lyase activity is higher in the rat kidney than in the human kidney (Cooper, 1994: Lash et al.,
1990), which is consistent with overall metabolic rates being higher in smaller versus larger
mammalian species.
In addition to activation by P-lyases, TCVC may be metabolized by a flavin-containing
monooxygenase, FMO3, or CYP enzymes to TCVC sulfoxide (TCVCSO), another reactive
metabolite (Ripp et al., 1997). TCVCSO is a more potent nephrotoxicant than TCVC (Elfarra
and Krause, 2007). These TCVC sulfoxide and P-lyase cleavage products rearrange, forming a
3-11
-------
thioketene (Ripp etal., 1997; Dekantet al., 1988), which is a potent acylating agent capable of
binding to cellular macromolecules, including DNA (Pahler et al., 1999a: Pahler et al., 1999b:
Volkel et al., 1999; Birner et al., 1996). Interestingly, the thioketene can degrade to form DC A,
potentially making this metabolite a product of both tetrachloroethylene metabolism pathways
(Volkel et al.. 1998: Dekant etal., 1987).
In addition to p-lyase and FMO3/CYP activation of TCVC, reactive sulfoxides can also
be produced by further CYP3A metabolism of 7V-acetyl--S'-(l,2,2 -trichlorovinyl)-Z-cysteine
[NAcTCVC (Werner et al., 1996)1. This tetrachloroethylene-derived mercapturate metabolite
results from TCVC being acetylated via a reversible reaction (Birner et al., 1996; Bartels, 1994;
Duffel and Jakoby, 1982). 7V-acetyl-
-------
3.3.3.2.2. Species differences in gamma-glutamyltransferase (GGT) and p-lyase
Species-dependent differences in GGT (Hinchman and Ballatori, 1990) also are not
thought to be limiting because renal activity is present at high enough levels even in humans so
that GGT activity is not the rate-limiting step in the metabolism. Species-dependent differences
in this enzyme (described below) would have only a very small quantitative effect on the overall
metabolism of TCVG and other similar GSH conjugates. Species differences in GGT activities,
therefore, would not have a major role in species differences in renal toxicity (Lash and Parker,
2001) in affecting transformation of TCVG to TCVC, and, thus, should not be important to
differences in susceptibility to tetrachloroethylene-induced renal toxicity.
GGT is the only enzyme that can split the gamma-glutamyl bond in the GSH conjugates
to form cysteine conjugates (Lash and Parker, 2001). It is this reaction that creates TCVC, the
substrate for the enzymes that generate the toxic metabolites. Therefore, the distribution of GGT
is important. Renal proximal tubular cells have the highest activities of GGT of all tissues,
although GGT activity also occurs in the liver, and the kidney-to-liver ratio of this enzyme varies
among species. In the rat, the specific activity ratio is 875 (Hinchman and Ballatori, 1990). The
ratio is lower in other species that have been studied. The tissue distribution and relative activity
have not been fully studied in humans, but it is known that GGT activity is considerably higher
in the human liver than in the rodent liver (Lash and Parker, 2001). The kidney-to-liver ratio of
GGT for humans is thought to be closer to those of pigs (2) and Macaques (5) than to those of
rats or mice (423). For this reason, use of a rodent model for the processing of the
tetrachloroethylene GSH conjugate to the corresponding cysteine conjugate would overestimate
the contribution of the kidneys and underestimate the contribution of the liver in cleaving TCVG
to TCVC. Even so, the liver excretes most of the cysteine conjugates such as TCVC into the bile
or plasma, where it is cycled to the kidneys and taken up into renal epithelial cells. So, the
TCVC will still end up in the kidneys.
The P-lyase enzyme is among the most important activator of toxic products in the
conjugation pathway—a fact particularly well documented in the kidney. There are some data,
however, that indicate that renal p-lyase-dependent metabolism is greater in rats than in mice or
in humans and greater in male than in female rats (Volkel etal., 1998; Green etal., 1990; Lash et
al., 1990). This is not entirely in keeping with metabolic rates, in general, which are higher in
mice than in rats, which, in turn, have higher metabolic rates than do larger animals, including
humans. Studies that measured only cytoplasmic P-lyase activity did not consider the
importance of mitochondrial P-lyase activity, which may be key to tetrachloroethylene
metabolite toxicity (Lash etal., 2001).
The higher percentage of mercapturate found in rat versus human urine does not indicate
a higher level of production of toxic products in the rat because excreted mercapturate allows no
3-13
-------
estimate of the amount of TCVC or 7V-acetyl TCVC being processed through alternate routes
(Lash and Parker, 2001). The relatively higher percentage of DC A in the urine may, however,
indicate relatively higher p-lyase enzyme activity and higher thioketene production in rats if the
DCA is indeed largely the product of the GST pathway rather than the oxidative pathway
(Volkel etal., 1998). It is not known whether sex-dependent variation of P-lyase activity exists
in humans as it does in rats (Volkel etal., 1998).
And finally, it is important to note that because the enzymes involved in this activation
pathway are also present in other tissues (Alberati-Giani et al., 1995; Malherbe et al., 1995;
Larsen and Stevens, 1986; Tomisawa et al., 1986; Larsen, 1985; Stevens, 1985; Tomisawa et al.,
1984: Stevens and Jakobv, 1983: Dohn and Anders. 1982: Tateishi et al.. 1978). there exists a
potential for formation of the reactive metabolites at sites other than the kidney, e.g., in the brain.
In one carcinogenicity bioassay of tetrachloroethylene, a biologically significant elevation of
gliomas in the rat brain was reported (NTP, 1986). Whether or not toxic metabolites resulting
from P-lyase activity in the brain play a role in the development of the gliomas in the rat has not
been studied. The possibility that such tetrachloroethylene metabolites could be involved in the
mode of tumorigenic action producing gliomas is not unrealistic.
3.3.3.2.2.1. Glutathione S-transferase (GST) isoenzymes/polymorphisms
GSTs are a family of isoenzymes (Mannervik, 1985) found in cytoplasm. A distinct
microsomal GST isoenzyme also exists in most mammalian tissues (Otieno et al., 1997).
Although GST activity occurs in most cell types, the liver is by far the predominant site of GSH
conjugation. GSTa, designated as GSTA in humans, is the predominate isoenzyme expressed in
normal kidney from rodents and humans (Cummings et al., 2000b: Rodilla et al., 1998: Mitchell
etal., 1997: Overby etal., 1994: Campbell etal., 1991). Available data thus far do not indicate
that variability in activity of this isoenzyme is important to differences in individual
susceptibility to toxicity. GST(^ (GSTZ) catalyzes the oxidative metabolism of DCA to
glyoxylate (Tong et al., 1998a, b; Board et al., 1997): however, the tetrachloroethylene
metabolite DCA has been shown to be a potent, irreversible inhibitor of GSTZ activity (Tzeng et
al.. 2000).
There are five human polymorphic variants of this GSTZ isoenzyme (Tzeng et al., 2000).
These genetic polymorphisms may influence tetrachloroethylene metabolism, although human
data regarding this hypothesis are lacking. There are some species differences in the other
three cytoplasmic GSTs relevant to liver and kidney. GSTP expression is the most variable and
appears to be polymorphic in humans (Rodilla et al., 1998). It has been found in rat liver
(Cummings et al., 1999) but only in biliary ducts in humans (Campbell et al., 1991: Terrier et al.,
1990). GSTTi (GSTP) has been detected within the human kidney in various cell types (Terrier et
3-14
-------
al., 1990) but has not been isolated from rat kidney cells (Cummings et al., 1999), although
GSTP has also been detected in the rabbit kidney (Cummings et al., 1999).
Two homodimeric GST9 (GSTT) isoenzymes have been identified in the human kidney
(Cummings et al., 2000a: Veitchet al., 1997). GSTT has been detected in rat and mouse liver
and in mouse but not rat kidney (Cummings et al., 1999; Quondamatteo et al., 1998). GSTji
(GSTM) has been detected in rat kidney distal tubule cells (Cummings et al., 2000b) and in
mouse and rabbit liver and kidney (Mitchell et al., 1997; Overby et al., 1994)—but it was not
detected in human kidney (Cummings et al., 2000a). It is not clear just how the differences in
these isoenzymes are related to species differences in tetrachloroethylene toxicity because the
isoenzyme specificity and reaction rates have not yet been studied with regard to
tetrachloroethylene (Lash and Parker, 2001).
3.3.3.3. Relative Roles of the Cytochrome P450 (CYP) and Glutathione (GSH) Pathways
Although it is clear that in mice and rats, the oxidative CYP pathway is quantitatively
more important than the GSH conjugation pathway, the relative roles of these pathways in
humans has not been determined directly. Moreover, the interorgan patterns for some of the
intermediate metabolites, as well as the relative toxicity of certain key metabolites generated
from these pathways, influence the relative importance of the two pathways in determining
toxicity. It is still not certain which metabolites, alone or in combination, are explicitly
responsible for specific tetrachloroethylene toxicities, and it is likely that different metabolites
contribute to toxicity at different target sites. In general, CYP metabolism is associated with
tetrachloroethylene-induced liver toxicity, whereas GSH conjugation followed by further
processing by p-lyase and other enzymes is associated with tetrachloroethylene-induced renal
toxicity. There is a possibility that P-lyase products could contribute to toxicity in the brain, for
example, and be a factor in the gliomas observed in rats. The parent compound, itself, is also
likely to be a contributing factor to tetrachloroethylene neurotoxicity, particularly central nervous
system effects.
Data from experiments designed to assess the effects of enzyme modulation suggest
competition between the two pathways (Lash et al., 2001; Lash et al., 1999; Volkel et al., 1998;
Dekantetal., 1987). Other data show relatively low urinary excretion of mercapturates as
compared to CYP-derived products. On the basis of these findings, some researchers have
concluded that there is a lack of toxicological significance for the low-affinity, low-activity GSH
pathway except when the high-affinity CYP pathway approaches saturation (Volkel etal., 1998;
Green et al., 1997; Green et al., 1990). However, this conclusion does not consider the relative
toxicological potency or chemical reactivity of the metabolites from the two pathways or the fact
3-15
-------
that the amount of mercapturate excreted is not a valid quantitative indicator of the extent of
conjugative pathway metabolism (Lash and Parker, 2001).
Specific tetrachloroethylene metabolites are known to be associated with certain
toxicities when they are administered directly. Exactly how these same compounds, as
metabolites of tetrachloroethylene, contribute to the various toxicities associated with exposure
to the parent compound is not yet well understood.
3.3.4. Susceptibility
Differences in enzyme activity may lead to variations among individuals in their
sensitivity to tetrachloroethylene toxicities. A 10-fold difference in CYP enzyme metabolic
capacity among humans is a generally accepted norm. Although individual variations in the
CYP2E1 enzymatic activity as high as 20-50-fold have been reported (Lieber, 1997; Stephens et
al., 1994; Yoo et al., 1988), these in vitro measurements would be taken out of physiological
context if used to estimate in vivo interindividual variations. Measurable and obvious
differences in CYP enzymatic activity are observed among various ethnic groups and age groups
(Raunio et al., 1995; Goldstein et al., 1969). No chemical-specific data regarding the manner in
which CYP enzyme isoforms might affect susceptibility to adverse effects are available for
tetrachl oroethy 1 ene.
Diagnoses of polymorphisms in carcinogen-activating and -inactivating enzymes and
cancer susceptibility have been noted (Raucy, 1995; Stephens etal., 1994; Yoo etal., 1988).
Potential strain-dependent differences among rodents and human genetic polymorphisms in
metabolizing enzymes involved in biotransformation of tetrachloroethylene are now known to
exist. Whether CYP polymorphisms could account for interindividual variation in
tetrachloroethylene metabolism among humans—and, thus, differences in susceptibility to
tetrachloroethylene-induced toxicities—is not known.
The GSTs involved in tetrachloroethylene metabolism are described in Section 3.3.3.2.
A potential exists for interindividual variation to occur in tetrachloroethylene metabolism as a
result of variability in GST enzyme expression. It is important to note that GST polymorphism
has been associated with increased risk of kidney cancer in people exposed to trichloroethy 1 ene
(Moore et al, 2011). There are no direct, chemical-distinctive data with regard to the specific
isoenzyme family responsible for TCVG formation in the metabolism of tetrachloroethylene.
There are species-dependent differences as to which isozymes occur in liver and kidney,
although it is unknown how the various enzymes are related to differences in the metabolism of
tetrachloroethylene. The compound is likely a good substrate for GSTA (Lash and Parker,
2001). GSTT and GSTP occur in human kidney, as does GSTA, the primary isozyme in human
kidney, meaning that there is a potential for differences in the ability to produce TCVG. GSTZ
3-16
-------
transforms the tetrachloroethylene metabolite DCA. DCA has also been shown to have a potent
irreversible inhibitory effect on the GSTZ isoenzyme, which is known to have at least
four polymorphic variations.
Inhibition or induction of the enzymes responsible for tetrachloroethylene metabolism
can, and likely does, alter susceptibility to toxicity (Lash and Parker, 2001; IARC, 1995).
Numerous environmental pollutants and therapeutic agents alike have the potential to induce or
inhibit tetrachloroethylene-metabolizing enzymes. For example, tetrachloroethylene metabolism
is increased by inducers of cytochrome CYP enzymes such as toluene, phenobarbital, and
pregnenolone-16a-carbonitrile, whereas CYP inhibitors such as SKF 525A, metyrapone, and
carbon monoxide decrease tetrachloroethylene metabolism (Costa and Ivanetich, 1980; Moslen
et al., 1977; Ikeda and Imamura, 1973). Chronic exposure to tetrachloroethylene has been shown
to cause self-induction of metabolism (Kaemmerer et al., 1982; Savolainen et al., 1977a: Vainio
et al., 1976). Other factors, such as health status or disease state, activity patterns, or
concomitant exposure to other chemicals, can potentially influence tetrachloroethylene
metabolism and its resulting toxicity. Section 4.9 addresses coexposures and cumulative risk in
greater detail.
3.3.5. Comparison of Tetrachloroethylene Metabolism with Trichloroethylene Metabolism
3.3.5.1. Extent of Metabolism
The available data indicate that, overall, tetrachloroethylene is less extensively
metabolized than is the closely related chemical, trichloroethylene. The difference is due to the
fact that a lower fraction of a tetrachloroethylene dose is metabolized via the major oxidative
CYP pathway when compared with an equivalent dose of the trichloroethylene congener (Lash
and Parker, 2001; Volkel et al., 1998; Ohtsuki etal., 1983). For example, in balance studies of
humans, only about 1-3% of the estimated amounts of tetrachloroethylene inhaled were shown
to be metabolized to TCA and other chlorinated metabolites, although these studies fail to
account for total dose (refer above for further discussion). These amounts can be compared to
the 40-75% of trichloroethylene shown to be metabolized in various human balance studies
similar to the ones conducted for tetrachloroethylene.
Because of its higher lipid solubility, tetrachloroethylene may appear to be less well
metabolized than trichloroethylene, at least to a certain degree, simply because it is more slowly
metabolized due to fat sequestration. However, the animal data from studies of the
two compounds provide results similar to those of the human studies regarding the relative extent
of metabolism. For example, the data from Schumann et al. (1980) and Pegg etal. (1979)
indicate that, in rats exposed to 10 and 600 ppm of tetrachloroethylene for 6 hours, the
3-17
-------
percentage of tetrachloroethylene body burdens excreted as unchanged parent compound are 68
and 99%, respectively. By comparison, rats and mice exposed to equivalent 10 and 600 ppm
trichloroethylene doses (Stott et al., 1982) metabolized a higher percentage of this compound,
with mice metabolizing essentially all of the inhaled dose and rats metabolizing 98 and 79% of
the low and high inhaled doses, respectively.
Saturation of metabolism occurs at a higher dose for trichloroethylene than for
tetrachloroethylene; thus, at certain dose levels, the differences in the amounts of the
two compounds metabolized are relatively greater than at other dose levels. Tetrachloroethylene
appears to be a lower-affinity substrate for CYP enzymes than trichloroethylene (Volkel et al.,
1998; Ohtsuki et al., 1983). In vitro, the Michaelis-Menten constant (Km) value for
tetrachloroethylene is reported to be higher than the Km value for trichloroethylene (Lipscomb et
al.. 1998).
Both tetrachloroethylene and trichloroethylene are liver toxicants and cause liver
hepatocellular carcinomas in mice. The liver toxicity, including carcinogenicity, of these
compounds is thought to be due to metabolites. It is interesting to note that although
trichloroethylene appears to be more extensively metabolized—due to greater CYP metabolism
in the liver—the relative cancer potency for liver tumors is similar for the two compounds.
Comparisons of potencies for kidney cancer are more difficult because there is a lack of studies
on these compounds using comparable species/strains and routes of exposure.
3.3.5.2. Cytochrome P450 (CYP)-Mediated Oxidation
TCA, DC A, chloral, and TCOH are reported biotransformation products of both
tetrachloroethylene and trichloroethylene; however, the relative amounts produced and the
precursor intermediates are different for the two compounds. TCA is the major urinary
metabolite for tetrachloroethylene, and it is also an excretion product of trichloroethylene,
whereas TCOH is the major trichloroethylene urinary excretion product. As discussed
previously in Section 3.3.3.1, the formation of chloral and TCOH in metabolism of
tetrachloroethylene is not likely to be significant. Therefore, very little, if any, TCA produced
from tetrachloroethylene metabolism comes through chloral—either directly or indirectly
through TCOH (Lash and Parker, 2001). The TCA from tetrachloroethylene comes through
trichloroacetyl chloride, possibly via the epoxide, but more likely directly from chlorine
migration of the Fe-O intermediate. On the other hand, the TCA produced from
trichloroethylene metabolism is thought to come through chloral—both directly and through
TCOH enterohepatic circulation (Lash et al., 2000).
DCA is a biotransformation product of both tetrachloroethylene and trichloroethylene,
although it is believed that a greater portion of DCA coming from tetrachloroethylene
3-18
-------
metabolism does not arise from CYP metabolism, but rather results from further processing of
TCVC, whereas the DCA coming from trichloroethylene metabolism results from CYP
oxidation.
Quantitatively, the liver is by far the predominant site of tetrachloroethylene and
trichloroethylene oxidative metabolism; although most other tissues contain the CYPs that could
conceivably metabolize these compounds. CYP2E1 has been shown to be important in rodent
metabolism of trichloroethylene; however, the chemical-specific data are sparse with regard to
its role in tetrachloroethylene metabolism (Doherty et al., 1996). Still, assuming that CYP2E1 is
important to tetrachloroethylene metabolism is not unreasonable. CYPS A isoenzymes and
especially CYP2B1/2 may be important for tetrachloroethylene. Costa and Ivanetich (1980)
showed increased/decreased hepatic metabolism following treatment with agents now known to
selectively induce/inhibit CYPS A and/or CY2B specifically.
3.3.5.3. Glutathione (GSH) Conjugation Pathway
The GSH-dependent pathway for tetrachloroethylene exists in both rodents and humans,
and the pathway is also operative for trichloroethylene in these species (Volkel et al., 1998;
Birner etal., 1996). The flux through this pathway at experimental or occupational exposures is
thought to be quantitatively less than that through the P450 pathway, although direct evidence is
lacking, particularly in humans. Toxic metabolites can arise from several sources in the
pathway; however, for tetrachloroethylene, as well as for trichloroethylene, the GSH pathway is
associated with renal toxicity (Lash and Parker, 2001; Lash et al., 2000; Dekant et al., 1989;
Anders et al., 1988). For both compounds, recovery of urinary mercapturates, the stable
end-products of the GSH pathway, comprises 1% or less of the total dose (Lash and Parker,
2001: Dekant etal.. 1986a), but this does not reflect the total flux through the GSH pathway. In
particular, the TCVC metabolite and the corresponding dichlorovinyl cysteine and their
respective 7V-acetylated forms derived from trichloroethylene might also act as substrates for
renal p-lyases and other enzymes such as FMO3 and CYPS A [(Lash and Parker, 2001; Lash et
al.. 2000: Anders etal.. 1988: Dekant etal.. 1988) reviewed by (Dekant et al.. 1989)1 (refer to
Section 3.3.3). It should be noted that a higher cysteine ^-conjugate-to-mercapturate ratio exists
for tetrachloroethylene when compared to trichloroethylene, which could influence the relative
bioactivation and nephrotoxicity of these two compounds (Lash and Parker, 2001).
3.3.5.4. Summary
Tetrachloroethylene is closely related structurally to trichloroethylene, and the
two chemicals cause similar toxic effects, many of which are attributed to metabolic activation of
the parent compounds. Interestingly, although tetrachloroethylene is not as extensively oxidized
3-19
-------
as trichloroethylene, they have similar potency for liver tumors, with which oxidative
metabolism is associated. TCA, DCA, chloral, and TCOH are reported P450 biotransformation
products of both tetrachloroethylene and trichloroethylene; however, only TCA predominates for
tetrachloroethylene, whereas TCOH predominates for trichloroethylene. In addition, DCA is
likely formed via GSH conjugation for tetrachloroethylene and via oxidation for
trichloroethylene. The fact that the two compounds produce different reactive intermediate P450
metabolites is also important to consider. Excretion of urinary mercapturates suggests that,
relative to P450 oxidation, tetrachloroethylene is more extensively metabolized via GSH
conjugation than is trichloroethylene. However, these urinary excretion products do not reflect
the total flux through the GSH pathway because the glutathione and cysteine conjugates of both
chemicals have been shown to undergo further processing to products that are highly reactive.
Thus, regardless of similarities, both the qualitative and the quantitative differences between
tetrachloroethylene and trichloroethylene in metabolite production could have bearing on toxicity
and tumor induction, and the relative importance of various mechanisms, and different modes of
action contributing to their toxic effects, including tumorigenesis, may vary between the
two parent compounds. Recognizing similarities and differences is important in attempting to
understand how each of these two compounds causes its toxic effects.
3.4. EXCRETION
Tetrachloroethylene is eliminated from the body by pulmonary excretion of the parent
compound and urinary excretion of metabolism products, with a small amount of pulmonary
excretion of metabolism products. Tetrachloroethylene that is not metabolized is exhaled
unchanged, and this process is the primary pathway of tetrachloroethylene excretion in humans
for all routes of administration (Opdam and Smolders, 1986; Koppel et al., 1985; Monster et al.,
1979: Stewart et al.. 1977: Guberan and Fernandez. 1974: Stewart et al.. 1974: Stewart et al..
1970: Stewart and Dodd, 1964: Stewart et al., 1961). Pulmonary excretion of (unchanged)
parent compound is also important in animals (Bogen et al., 1992: Frantz and Watanabe, 1983:
Schumann et al., 1980: Pegg et al., 1979: Yllner, 1961). A very small amount of
tetrachloroethylene has been shown to be excreted through the skin (Bolanowska and Golacka,
1972): however, it represents an insignificant percentage of total tetrachloroethylene disposition.
Pulmonary excretion of unchanged tetrachloroethylene and other volatile compounds is
related to ventilation rate, cardiac output, and the solubility of the compound in blood and tissue.
The lung clearance of tetrachloroethylene in six adults exposed at rest to 72 ppm and 144 ppm of
tetrachloroethylene averaged 6.1 L/minute initially and decreased to 3.8 L/minute after 4 hours
(Monster et al., 1979). Lung clearance represents the volume of air from which all
tetrachloroethylene can be removed per unit time. Normal ventilation rates in adults range from
3-20
-------
5-8 L air/minute at rest. Pulmonary excretion of unchanged tetrachloroethylene at the end of
exposure is a first-order diffusion process across the lungs from blood into alveolar air, and it can
be thought of as the inverted equivalent of its uptake from the lungs. Pulmonary excretion
occurs in three first-order phases of desaturation of blood vessel-rich tissues, muscle tissue, and
adipose tissues (Monster et al., 1979; Guberan and Fernandez, 1974). For humans, the
half-times of elimination from these three tissue groups are 12-16, 30-40, and 55-65 hours,
respectively (Monster et al., 1979).
The long half-time of tetrachloroethylene elimination from adipose tissue, due to the high
adipose tissue:blood partition coefficient and the low rate of blood perfusion of the fat tissue
(EgerEI, 1963), is independent of the body burden of tetrachloroethylene, indicated by parallel
blood and exhaled air concentration decay curves. However, the exhaled air or end alveolar air
concentrations and the blood concentrations after exposure and throughout desaturation are
proportional to the acquired body burden or exposure concentration and duration, and they can
serve as a means of estimating body burdens. The half-life of tetrachloroethylene in the human
body, measured as the inverse of the slope of the log-concentration versus the time curve of the
exhaled chemical, varies from 5-20 minutes for the first phase of elimination up to
approximately 50 hours during its extended phase (Chien, 1997; Monster et al., 1979). The long
half-time of tetrachloroethylene pulmonary excretion indicates that a considerable time is
necessary to completely eliminate the compound. This time is greater than five times the
half-life, or about 2 weeks, for humans. For the rat, the half-time of pulmonary elimination is
about 7 hours.
Urinary and pulmonary clearances of metabolism products of tetrachloroethylene provide
other means of excretion. The mean half-time of urinary excretion for total trichloro-compounds
for 13 subjects exposed to tetrachloroethylene was determined to be 144 hours (Ikeda and
Imamura, 1973). When TCA is administered directly, however, the half-life is not that long.
The longer half-life of TCA from tetrachloroethylene metabolism is likely due to constant
metabolic conversion of the parent compound to TCA as tetrachloroethylene is cycled to the
liver over the period of time it is released from adipose tissue.
The urinary excretion of tetrachloroethylene biotransformation products, primarily TCA,
has been thought to represent only a small percentage of the total absorbed dose of
tetrachloroethylene in humans (Volkel et al., 1998; AT SDR, 1997a). Urinary excretion of TCA
(or total trichloro-compounds) was estimated to be only 1-3% in balance studies conducted in
humans (Chiu et al., 2007; Monster et al., 1983; Monster and Houtkooper, 1979; Monster et al.,
1979: Fernandez et al., 1976: Essingetal., 1973: Ikeda etal., 1972: Stewart et al., 1970: Boettner
and Muranko, 1969: Stewart et al., 1961), with urinary excretion of GSH-derived metabolism
products representing an even smaller fraction (Volkel etal., 1998). However, these studies did
3-21
-------
not follow urinary excretion for more than 3-7 days, and it is possible that a larger percentage of
the tetrachloroethylene dose was eventually excreted in urine. In studies that also measured
pulmonary excretion, the entire dose was not always accounted for in the sum of exhaled
tetrachloroethylene and urinary excretion of TCA (Chiu et al., 2007; Monster et al., 1979). Part
of the dose may be metabolized to biotransformation products that were not measured, including
oxidative products such as carbon monoxide, carbon dioxide, or oxalic acid, and GSH
conjugation products such as sulfoxides and reactive thiols (refer to Section 3.3). In addition, the
lowest exposures in these studies were around 1 ppm in air (Chiu et al., 2007), which is several
orders of magnitude higher than ambient environmental exposures.
In laboratory animals, there is both a species- and dose-dependence to the amount of
pulmonary excretion of unchanged tetrachloroethylene that reflects the degree of metabolism
(Dallas et al.. 1994b: Bogen et al.. 1992: Schumann et al.. 1980: Pegg et al.. 1979). As the body
burden of tetrachloroethylene is increased in the rat or mouse, the percentage excreted as
unchanged parent compound increases. Conversely, as metabolism is the other principal route of
elimination of tetrachloroethylene, when the body burden increases, the percentage of the burden
metabolized decreases, although the absolute amount metabolized increases (Schumann et al.,
1980; Pegget al., 1979). These observations suggest that, in the rodent, metabolism of
tetrachloroethylene and urinary excretion of its metabolites are rate limited and dose dependent,
whereas pulmonary excretion is a first-order process and is dose independent, with half-time and
rate constant being independent of the dose. Data from studies by Filser and Bolt (1979) and
Buben and O'Flaherty (1985) suggest that metabolism of tetrachloroethylene is greater in mice
than in rats, so conversely, the amount of pulmonary excretion is greater in rats than in mice.
3.5. TOXICOKINETIC MODELING
Understanding tetrachloroethylene toxicokinetics is critical to both the qualitative and
quantitative assessment of human health risks from environmental exposures. A number of the
neurotoxic effects of tetrachloroethylene appear well correlated with parent compound
concentrations at the target site (Bushnell et al., 2005), so characterizing tetrachloroethylene
blood or tissue concentrations can aid in performing risk assessment-related extrapolations, such
as between rodents and humans or between exposure routes. In addition, understanding
tetrachloroethylene metabolism is especially important lexicologically because specific
metabolites or metabolic pathways are associated with a number of endpoints of observed
toxicity. A more detailed discussion of the evidence for these associations, the specific
metabolites involved, and identification of the most appropriate dose metric are provided in
Section 5.
3-22
-------
3.5.1. Choice of Physiologically Based Pharmacokinetic (PBPK) Model for Use in
Dose-Response Modeling
3.5.1.1. Limitations of Previously Developed Physiologically Based Pharmacokinetic
(PBPK) Models
A large number of PBPK models have been developed for tetrachloroethylene
toxicokinetics in both rodents and humans for various purposes. PBPK models can provide
estimates of tissue concentration as well as total metabolism of tetrachloroethylene. An
overview of the models in literature is provided below—the aim of which is not to exhaustively
cover all of the models in the literature—but rather to capture the different assumptions made,
the range of data that has been used, and to indicate that these assumptions limit the ability of the
models to predict relevant tetrachloroethylene metabolite levels in humans.
Chen and Blancato (1987) developed a PBPK model for rats, mice, and humans. The
metabolic parameters maximum velocity (Vmax) and Km were derived by fitting the model to the
total amount of metabolized tetrachloroethylene. Experimental data on total metabolite were
available for rodents. However, for humans, it was assumed that the urinary metabolite TCA, as
measured by Monster et al. (1979), accounted for 30% of the total metabolite. This percentage
was chosen because it resulted in a better fit.
Reitz et al. (1996) developed a PBPK model for rats, mice, and humans that describes the
total metabolism of tetrachloroethylene using Michaelis-Menten kinetics. The partition
coefficients for the five tissue compartments were measured independently. For rats and mice,
the metabolic parameters Vmax and Km, as well as the volume and blood flow rates of the fat
compartment, were obtained by simultaneously optimizing the fit to three sets of in vivo data
gathered in 6-hour inhalation radiolabeled tetrachloroethylene exposure studies. These data were
(a) concentration of tetrachloroethylene in exhaled breath, (b) radioactive body burden present in
animals at end of exposure, and (c) total postexposure radioactive metabolites recovered from all
excreta and carcass homogenates. The metabolic parameters for humans were estimated using a
"parallelogram approach" (Reitz et al., 1989). First-order constants for the rate of metabolism
were measured in vitro using isolated liver microsomes of all three species. The ratio of these in
vivo and in vitro metabolic rates was assumed to be nearly constant across species, as was found
to be the case for rats and mice. Using this constant ratio, the human in vivo metabolic rate
constant per gram of liver could be determined from the human in vitro value. Km was assumed
to be invariant across species because it is derived solely from the reaction rate constants for the
enzyme-catalyzed metabolic reactions. Reitz et al. (1996) also used a second method for
estimating Vmax, which was based on extrapolation from in vivo animal studies of other
chemicals metabolized by cytochrome P450 enzymes. Vmax, so estimated, was allometrically
3-23
-------
scaled to humans. The values obtained by Reitz et al. (1996) through both these independent
methods were comparable.
Rao and Brown (1993) developed a human PBPK model for the purpose of investigating
neurotoxicological endpoints. 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 Vmax
(allometrically scaled by body-weight374) and Km (considered invariant across species). Other
parameters 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).
Other authors have developed models for tetrachloroethylene that specifically describe
the kinetics of its major metabolite, TCA. Gearhart et al. (1993) developed a model for
tetrachloroethylene that also included the kinetics of TCA, assuming that TCA comprised 60%
of the total tetrachloroethylene metabolized in the rodent and using similar parameters for TCA
as in a model for trichloroethylene. Tetrachloroethylene metabolism parameters for mice were
estimated by fitting the model to the time course of tetrachloroethylene chamber concentration in
gas uptake studies. The model was independently validated at low oral doses (acute oral gavage
of tetrachloroethylene in corn oil) by comparing the time course of blood concentrations of
tetrachloroethylene and TCA in mice.2 The parameters for describing tetrachloroethylene
metabolism in humans were derived by fitting the model to urinary excretion of TCA in
two subjects in a study by Fernandez et al. (1976), assuming the same ratio of TCA to total
metabolite as in the rodent. This value was set to 0.6 and attributed to Dekant et al. (1986c).
The validity of using this value for humans has not been evaluated. Reitz et al. (1996), in their
radiolabeled tetrachloroethylene studies, determined the fraction of urinary to total metabolites to
range from 0.49-0.59 in rats and from 0.56-0.66 in mice for exposure concentrations that varied
by two orders of magnitude.
Clewell et al. (2005) evaluated and extended the Gearhart et al. (1993) model further,
using tetrachloroethylene blood concentrations and urinary and blood TCA data gathered by
Volkel et al. (1998) on human subjects exposed to tetrachloroethylene concentrations of
10-40 ppm for 6 hours. They included metabolism of tetrachloroethylene in the kidney,
allowing for excretion directly into urine. By assuming metabolism in this organ to be at 10% of
the capacity of the liver, they obtained substantial improvement in the agreement with
experimental data on urinary excretion of TCA. An advantage in using the Volkel et al. (1998)
2Details pertaining to the derivation of parameters for metabolism in humans are not provided in the original paper
but are available in a review by Clewell et al. (2005).
3-24
-------
data is that they pertain to exposure concentrations that are lower than those in other studies
[e.g., 72-144 ppm in Monster (1979)1.
Loizou (2001) used a PBPK model that was structurally similar to that of Gearhart et al.
(1993). The model assumes a 15% stoichiometric yield for the total metabolite produced across
various dose levels (i.e., 15% of the parent compound in the liver is metabolized), but the basis
for these assumptions is not substantiated. The above yield is also assumed to hold for the
production of TCA because it is the major metabolite (personal communication from G. Loizou,
Health and Safety Laboratory, UK, to R. Subramaniam, U.S. EPA). Elimination rates of TCA
through blood and urine were chosen by calibrating the model to fit blood and urinary TCA
kinetics and exhaled tetrachloroethylene TCA concentration levels obtained from Monster et al.
(1979).
In addition, a number of PBPK models were developed only in humans, primarily to
characterize uncertainty and/or human variability. To assess intraindividual variability in uptake
and elimination over multiple exposure levels and scenarios, Chien (1997) collected exhaled
breath measurements on a single individual following four different exposure scenarios in a
controlled environmental facility (twelve, 30 or 90 minute exposures ranging from 0.5-3 ppm in
concentration) and following tetrachloroethylene exposure in 22 dry-cleaning facilities, where
ambient levels of tetrachloroethylene were recorded and exposures were carefully timed.
Bois et al. (1996), which was updated by Chiu and Bois (2006), used a Bayesian analysis
in conjunction with a PBPK model that was structurally similar to that used by Reitz et al. (1996)
and that was only calibrated to the parent compound data (blood and exhaled breath) of the
individuals in Monster et al. (1979). The shape of the prior distribution was observed to have
little impact on final results. Model predictions were compared against alveolar concentrations
of subjects in the Opdam and Smolders (1986) study, and all data points were observed to fall
within the 95th percentile envelope of predictions. The exposure concentrations in this study
were 5-100 times lower than those used in the Monster et al. (1979) study; thus, this comparison
provides further weight to the strength of the model.
Covington et al. (2007) applied the same methodology to the Clewell et al. (2005) human
PBPK model, using additional data on the parent compound tetrachloroethylene and urinary
excretion data of its metabolite TCA (Monster et al., 1979; Fernandez et al., 1976), with a range
of exposure concentrations from 10-150 ppm. However, TCA blood concentrations from
Monster et al. (1979) were dropped from the analysis because the authors, in preliminary
calculations using a one-compartment PBPK model for TCE from Clewell et al. (2000), were
unable to reproduce the urinary excretion of TCA using the blood concentration data on TCA
from the same study. In addition, Covington et al. (2007) used only grouped data from both
these studies because the individual data were not available to them.
3-25
-------
The Covington et al. (2007) analysis was revisited by Qiu et al. (2010), with the
following modifications:
1. A brain compartment was added.
2. Human kinetic data from Chiu et al. (2007) and Chien (1997) were used in addition to the
data used by Covington et al. (2007): namely Fernandez et al. (1976), Monster et al.
(1979), and Volkel et al. (1998). Thus, the human exposures used in the Qiu et al. (2010)
modeling range from 0.5-150 ppm.
3. Individual human data were used. However, blood TCA measurements from Volkel
et al. (1998) were not used, which the authors stated was because blood samples could
not be matched with individuals' urine samples and because there were not enough data
points to inform the time course for blood TCA. In addition, none of the TCA data from
Monster et al. (1979) were used.
4. Correlation between parameters (such as between cardiac output and alveolar ventilation
or between Vmax and Km) was addressed by reparameterization.
5. Adjustment factors were added to maintain mass balance among fractional blood flows
and fractional tissue volumes.
These models provide substantially similar estimates of the tetrachloroethylene
concentration in the tissue. For example, as illustrated in Figures 3-2 and 3-3, estimated venous
blood concentration and alveolar concentration of tetrachloroethylene were in agreement to
within a factor of 2.0 among various models and experiment. However, the same models have
different approaches to estimating the metabolic parameters, thereby differing hugely in their
prediction of the amount metabolized at low dose—as shown in Figure 3-4 [adapted from (Chiu
et al., 2007)]. These differences have major implications for the quantitative risk assessment and
represent the key controversy surrounding the application of PBPK models to
tetrachloroethy 1 ene toxicokinetics.
3-26
-------
10 -,
O)
CD
O
c
O
O
CD
0.01
o
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 3-2. Comparison of model predictions for blood concentration with
experiment.
PCE inhaled concentration was 72 ppm. Predictions are at different ventilation-to-perfusion ratios
and at an alveolar ventilation rate of 7 L/minute (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 hours, respectively). Experimental data adapted from Monster et al.
(1979).
3-27
-------
10 -
E
Q.
&.
O
c
o
O
_
o
(D
1 -
O Bois model
V Rao & Brown model
n Reitz model
4 Stewart Expt
o
V
V
n
o
v o
V
n
n
g
R °
M H O
O
o
50
100
150
200
250
300
350
Post-Exposure Time (hrs)
Figure 3-3. Comparison of model predictions for alveolar concentration of
tetrachloroethylene with experimental data on humans.
Inhaled concentration is 100 ppm, 7 hours/day, for 5 days, and predictions assume alveolar
ventilation rate of 5.02 L/minute and a ventilation-to-perfusion ratio of 1.0. Experimental data
show mean alveolar concentration in subjects in Stewart et al. (1970). Some points early in the
time course were deleted because of difficulty in obtaining numerical values from the author's
plot.
3-28
-------
Chen and Blancato (1987)
Wardetal. (1988)
Boisetal. (1990) [1]
Boisetal. (1990) [2]
Rao and Brown (1993)
Reitzetal. (1996)
Boisetal. (1996)
Loizou(2001)
Clewelletal. (2005) [1]
Clewell et al. (2005) [2]
Chiu and Bois (2006)
Covington et al. (2007)
Qiu et al. (2009)
]2.9%
23%
23%
15%
4.4%
16%
36%
1.9%
1.7%
]2.3%
23%
.02%
JO. 97%
0.0% 20.0% 40.0% 60.0% 80.0%
Percent Metabolized at 1 ppb Inhalation Exposure
Figure 3-4. Previously published estimates for the total amount of
tetrachloroethylene metabolized at 0.001 ppm (1 ppb) continuous inhalation
exposure.
All estimates are point estimates except for Bois et al. (1996) and Chiu and Bois (2006). which are
estimates of combined uncertainty and population variability (95% confidence intervals [CIs]),
and Covington et al. (2007) and Qiu et al. (20 IPX which are estimates of uncertainty in the
population mean (90% CIs).
The various analyses in Figure 3-4 for tetrachloroethylene have a number of key
limitations. First, in no case have all the available data in mice, rats, and humans been
considered together in a single analysis. Thus, the extent to which different results reflect use of
different data sets and model structures is unclear. Moreover, while all the models estimate total
metabolism, those estimates are based on different types of data—in some cases, disappearance
of the parent compound, and in other cases, TCA and, therefore, oxidation—none of which
address GSH conjugation. These limitations and the above-mentioned controversy were also
noted in the National Research Council (NRC) report Review of the Environmental Protection
Agency's Draft IRIS Assessment of Tetrachloroethylene (NRC, 2010). In particular, NRC
concluded that, while a number of PBPK models have been developed for tetrachloroethylene,
they all have some key limitations that reduce the confidence with which they can be used for
3-29
-------
risk assessment. NRC (2010) recommended the development of a "harmonized" PBPK model
that would integrate previous models and data. They pointed to the availability of in vitro and in
vivo data relevant to the GSH conjugation pathway, and they recommended exploring the
possibility of adding the GSH pathway to a harmonized PBPK model. This is important because
tetrachloroethylene causes tubular toxicity in mice and rats in the kidney, and is associated with
small increases in the incidences of kidney tumors reported in multiple strains of
tetrachloroethylene-exposed rats (JISA, 1993; NTP, 1986). These effects are thought to be
associated with the tetrachloroethylene metabolism by GSH conjugation based on the production
of nephrotoxic and genotoxic metabolites in the kidney from this pathway (Lash and Parker,
2001).
3.5.1.2. The Chiu and Ginsberg (2011) Model
In response to this advice from the NRC, another PBPK model was developed by Chiu
and Ginsberg (2011). This model was developed to address many of the limitations of the
existing models for tetrachloroethylene, discussed above. Among the most important
improvements are (1) the utilization of all the available toxicokinetic data for tetrachloroethylene
and its metabolites in mice, rats, and humans; (2) the incorporation of available information on
the internal toxicokinetics of TCA derived from the most current PBPK modeling of
trichloroethylene and TCA; and (3) the separate estimation of oxidative and conjugation
metabolism pathways. Therefore, this assessment utilizes the Chiu and Ginsberg (2011) model
to calculate relevant dose metrics to be used in dose-response modeling. An overview of this
model follows below.
In developing this model, first, a comprehensive literature search was made of relevant
toxicokinetic studies, and the available toxicokinetic data were digitized. These data were
further separated into "calibration" and "validation" data sets utilizing a wider range of data than
any previous analysis alone. Second, a harmonized PBPK model structure was developed that
separately tracked tetrachloroethylene oxidation and GSH conjugation. The Chiu and Ginsberg
(2011) model includes a comprehensive analysis of TCA dosimetry originally developed by the
primary author for TCE, and it includes the urinary excretion kinetics of the metabolites
NAcTCVC and DCA. The Chiu and Ginsberg (2011) model is described by the schematic
below. The reader is referred to Chiu and Ginsberg (2011) for further details of the model
structure.
The model structure and parameters (shown in Figure 3-5) used in the Chiu and Ginsberg
(2011) harmonized model differed from other human models along the following lines:
3-30
-------
Inhaled air
(Clnh)
PCE
I Exhaled air
I (CMixExh)
QM'Clnh ^
Respiratory
Tract Lumen
Inhalation
(AlnhResp)
Dresp*(Cresp-
ClnhResp)
\ >
Respiratory
Tract Tissue
__.*!__
Oxidation
(VMaxClara,
Dresp*(Cresp-
CExhResp)
\ >
SL QM'CMixE
Respiratory
Tract Lumen
Exhalation
(AExhResp)
(Dead space)
QF-ClnhResp^
(QM-QP)*ClnhResp
Gas Exchange
jQP*CArt_tmp/PB
QC'CVenT
Venous
Blood
(ABId)
^*S_
^ QRap'CVRap
^ QSIw*CVSIw
^ QFaTCVFat
Q
J,QGut+QLjv)*CVLit
_ QKid-CVKid
Rapidly
Perfused
(ARap)
Slowly
Perfused
(ASIw)
Fat
(AFat)
Gut
(AGut)
Liver
(ALiv)
Kidney
(AKid)
1
_ QRap'CArt \
QSIw'CArt
QSIw'CArt
QFafCArt
,,"""
^'*= = = = = =
QGul'CArt
_ QLjVCArt
^ QKid*CArt
^
^•CArt_tmp_ S^\p^~~
,0^"" Vffix
^OraTN
X^^kStom),/
t
Stomach
(ASton)
^""J t
. Duodenum
(ADuod)
/^p\/^\
X^fkFW/
VMax, KM
^j" Oxidations
1 Conjuo^ation
VMaxTCVG, KMTCVG
VMaxKid, KM Kid
{ "Oxid'ation~&~1
I_Conjuc[ation J
VMaxKidTCVG, KMKidTCVG
Legend
In put expos ure o r dose (fl ux i n mg/h)
"Dynamic" Compartment solved by DDEs (state variable in mg)
"Static" Compartmentatlocal steady-state
Transformation orExcretion
Arrows are fluxes, labeled by magnitude (mg/hr)
I .
Figure 3-5. Overall structure of updated physiologically based
pharmacokinetic (PBPK) model for tetrachloroethylene and metabolites.
Boxes with underlined labels are additions or modifications of the Chiu et al. (2009) model for
trichloroethylene, and are described in the Toxicological Review of Trichloroethylene (U.S. EPA.
20 lib). "Inputs" are exposures/doses by which tetrachloroethylene enters the body, and lead to a
flux of the chemical into various body compartments. Dynamic compartments are those solved by
ordinary differential equations (ODEs). "Static" compartments are those solved by algebraic
equations, and lead to instantaneous changes in the amount of tetracholoroethylene in the
compartment. Transformation or excretion represents fluxes of compounds leaving a
compartment for another compartment (if any) via metabolism or excretion.
3-31
-------
n
11 Oxidation
I _ m Ljy er
Oxidative Metabolism
I Oxidation
[ _ in Lung _
I Oxidation
\ Jn_Kidney_
•acOther.
Other
(kMePALivTCA)
Conjugative Metabolism
I Liver + 1
1 Kidney '
FracNATU
NAcTCVC
urine delay
(ANTCVC'kNAT)
•acDCAUrn
Other
DCA urine
delay
CA1
I NAcTCVC |
I _ -Urine _ _ j
(ADCA'kDCA)
I DCA Urine '
Figure 3-5. (Continued) Overall structure of updated physiologically based
pharmacokinetic (PBPK) model for tetrachloroethylene and metabolites.
3-32
-------
• All the available data on mice, rats, and humans were considered together in a single
analysis.
• The wash-in-wash-out process in the lung was included.
• Oxidative metabolism in the lung was included.
• The model explicitly addressed GSH conjugation of tetrachloroethylene in the liver and
kidney.
• The urinary data on DCA (Volkel et al., 1998) were included so as to be able to consider
separate p-lyase-dependent and p-lyase-independent pathways for the bioactivation of
TCVC in the GSH conjugation pathway.
• An empirical "delay" parameter (whose value was "fitted") was added for urinary
excretion of DCA and NAcTCVC and represented a "lumped" delay in the time course
due to the processes of formation, urinary excretion, and other clearance pathways.
• For tetrachloroethylene oxidation, metabolic parameters were obtained from four in vitro
studies. These consisted of data from microsomes or cells from the liver and microsomes
from the kidney (Lash et al.. 2007: Reitzetal.. 1996: Costa and Ivanetich. 1984. 1980).
• A fraction of tetrachloroethylene oxidation was assumed to form compounds other than
TCA. A baseline value of 10% was used for the fraction not forming TCA.3
• GSH conjugation metabolic parameters were obtained from four studies that measured
tetrachloroethylene GSH conjugation in vitro (Lash et al., 2007: Dekant et al., 1998: Lash
et al., 1998: Green et al., 1990). These studies were utilized to select a baseline value for
metabolic clearance along this pathway in all species.
• The model incorporated all in vivo data considered in the literature for the PBPK
modeling of tetrachloroethylene and metabolites, dividing these data into two groups, one
for model calibration, and the other for model validation. These data included short and
long dosing periods.
• A full Bayesian uncertainty/variability analysis was not performed. The limited Bayesian
analysis involving flat priors and making inferences only using posterior modes was used
for the estimation of a limited number of metabolism parameters for which there was
significant discrepancy between baseline predictions (using baseline values of these
parameters) and in vivo data related to metabolism [refer to Table A-l of Chiu and
Ginsberg (2011) and associated text for rationale]. The Markov Chain Monte Carlo
(MCMC) approach was used for optimization.4
3In vitro data measuring TCA alone; TCA along with chloral hydrate, TCOH, and DCA; and total water soluble
metabolites are all generally found to be consistent with each other.
4The Markov Chain Monte Carlo (MCMC) method provides an algorithm to sample from a desired probability
distribution—in this case, the likelihood function—the output of which is a sequence of samples—the "Markov
3-33
-------
• The model structure allowed it to be used to calculate internal dose metrics for inhaled
and oral exposure to tetrachloroethylene for mice, rats, and humans. Thus, the analysis
could be used for route-to-route extrapolation or interspecies extrapolation, comparison
of parent and metabolite toxicity based on a common internal dose metric, and
investigation of the shape of the dose-response curve. The following dose metrics could
be determined using this model, and the confidence with which it can make predictions
for internal dose metrics of interest was further evaluated by the authors:
o Daily area-under-the-curve of tetrachloroethylene in blood
o Fraction of tetrachloroethylene intake metabolized by oxidation
o Fraction of tetrachloroethylene intake metabolized by GSH conjugation
o Equivalent daily production of TCA per kg body weight.5
3.5.1.2.1. Estimated human parameter values for oxidation and conjugation in Chiu and
Ginsberg (2011)
The results for all estimated parameters are shown in Table 3-1. The estimated
metabolism parameters for oxidation and conjugation, which are of particular interest, are briefly
addressed here, and the reader is referred to the original paper for further details on these and
other parameter estimations. Figure 3-6 compares the in vivo predictions for hepatic metabolism
with available in vitro data. For oxidation, in mice and rats, the optimized values are about an
order of magnitude higher than baseline values, whereas in humans, the optimized values are
quite similar to baseline values. However, they do not appear unreasonable compared to the
limited data available from other related compounds (TCE and some halomethanes), as shown in
Figure 3-6. For example, the linear rates are lower than those for TCE, which is known to be
more extensively oxidized by P450s than tetrachloroethylene. At higher substrate
concentrations, the predicted rate of oxidation of tetrachloroethylene in mice and humans is
greater than that for TCE, but this is an artifact of the assumption of a linear rate necessitated by
KM being unidentifiable. For GSH conjugation, the range of the in vitro data is quite wide,
especially when also taking into considering data from other compounds (refer to Figure 3-6). In
mice and rats, the in vitro data on tetrachloroethylene GSH conjugation (filled symbols in Figure
3-6) span the range of estimates from optimization to in vivo data. For humans, the in vitro data
chain," or "chain" for short. Each "chain" has a random starting point. In order to capture the potential uncertainty
due to different starting points, 24 chains with different starting points were run for mice and rats, and 48 chains
were run for humans. The posterior mode from each chain was determined—i.e., the "chain-specific posterior
modes." Then, the highest posterior model among the 24 (or 48) chains was determined—i.e., the "overall posterior
model," or simply the "posterior mode." GNU MCSim version 5.0.0 was used for computations. A PDF of the
model code is available as Supplementary Materials (Chiu and Ginsberg. 201IX and simulation files are available in
a WinZIP archive (U.S. EPA. 201 la).
5 TCA produced in the kidney and excreted directly to urine was not included, because it does not reach any target
organ (i.e., the liver) or enter systemic circulation.
3-34
-------
only consist of nondetects from Dekant et al. (1998), which, if assumed to be half the detection
limit, are more consistent with the alternative posterior modes. Overall, however, the ranges of
predicted rates for tetrachloroethylene are consistent with the range inferred from halomethanes,
and the in vivo optimized values do not appear to be substantially outside the bounds suggested
by available in vitro data.
Chiu and Ginsberg (2011) observe that, overall, the fits to the data and validation were
generally within threefold of the observed data, and more consistently so for rats and humans,
given the inter- and intraindividual variability [refer to Supplementary Materials (Chiu and
Ginsberg, 2011) in their paper]. In a few cases, the discrepancies were larger, but did not exceed
an order of magnitude. To a large extent, the discrepancies in model fits reflected variability.
There was, for instance, difficulty in fitting the time course of TCA in mice and the fraction of
retained tetrachloroethylene exhaled.
3.5.1.2.2. Dose metric predictions based on posterior modes
Tables 3-2-3-5 summarize the PBPK model dose metric predictions (listed in the
previous subsection) based on the baseline, overall posterior mode, and chain-specific posterior
mode parameters. The uncertainty due to the distribution of chain-specific posterior modes
contributes to the overall uncertainty in the predicted dose metric. The blood tetrachloroethylene
dose metric has by far the least amount of this "sampling" uncertainty. This appears to be true
across all species, routes of exposure, and exposure levels. The dose metrics with the next lower
amount of sampling uncertainty are tetrachloroethylene oxidation and TCA formation. The
predictions for GSH conjugation are more uncertain. In the rat, the ranges of chain-specific
posterior modes span up to twofold, and in mice, up to 10-fold. However, in humans, the ranges
span about 3,000-fold, discussed above.
3.5.1.2.3. Overall pertinent conclusions on tetrachloroethylene dosimetry
Chiu and Ginsberg (2011) also presented detailed sensitivity analyses that enable
determination of the confidence with which a particular dose metric can be estimated [refer to
Table 9 and Supplementary Materials (Chiu and Ginsberg, 2011) in their paper]. These have to
be analyzed together with the residuals for error in calibration and validation (refer to Table 8 of
their paper) and the ranges in the values of the predicted dose metrics (presented above in Tables
3-2-3-5) to obtain perspective on the overall uncertainty in the PBPK model predictions.
Table 3-6 summarizes the various measures that may contribute to this overall uncertainty.
3-35
-------
Table 3-1. Log-likelihood and parameters after calibration
Parameter
Baseline
Postcalibration
(posterior
mode)
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Mouse
Ln (Likelihood)
QP (L/hr)
Vmax (mg/hr) (saturable
oxidation pathway)
Km (L/hr) (saturable
oxidation pathway)
Vmax2/Km2 (L/hr) (linear
oxidation pathway)
VmaxTCVG/KmTCVG
(L/hr) (linear conjugation
pathway)
kMetTCA (/hi)
kUrnTCA (/hr)
-
2.09
0.23
88.6
-
0.656
1.48
2.93
-1,780
2.89
0.026
0.417
0.0188
6.83E-05
0.638
1.26
-
1.03
1.16
1.28
1.05
3.83
1.05
1.05
-1808-1780
2.86-3.22
0.022-0.0369
0.338-0.892
0.0165-0.0207
3.05e-05-0.00179
0.56-0.695
1.11-1.38
Rat
Ln (Likelihood)
QP (L/hr)
Vmax (mg/hr) (saturable
oxidation pathway)
Km (L/hr) (saturable
oxidation pathway)
VmaxTCVG/KmTCVG
(L/hr) (linear conjugation
pathway)
kDCA (/hr)
FracNATUrn
FracDCAUrn
-
10.2
0.256
69.7
2.22
-
-
-
-1314
6.31
0.87
31.1
0.00204
0.129
0.0143
0.702
-
1.02
1.37
1.39
1.27
1.65
1.29
1.26
-1321—1314
6.28-6.68
0.415-1.93
14.8-71.9
0.00131-0.00355
0.0758-0.451
0.00919-0.0253
0.43-0.98
3-36
-------
Table 3-1. Log-likelihood and parameters after calibration (continued)
Parameter
Baseline
Postcalibration
(posterior
mode)
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Human
Ln (Likelihood)
QP (L/hr)
VMax/KM (L/hr) (linear
oxidation pathway)
VMaxKid/KMKid (L/hr)
(linear oxidation pathway)
VMaxTCVG/KMTCVG
(L/hr) (linear conjugation
pathway)
kNAT (/hi)
FracNATUrn
FracDCAUrn
-
372
0.353
0.00076
0.0196
-
-
-
1,828
476
0.454
0.0947
5.26
0.28
0.000482
0.00022
-
1.1
1.08
1.09
17.1
1.07
15.8
18.5
1,790-1,828
450-640
0.346-0.468
0.0702-0.105
0.00194-5.48
0.228-0.293
0.000472-1.00
1.12e-05-0.442
GSD = geometric standard deviation.
3-37
-------
• 7
Q-
A:Mouse Oxidation
o »„
o,
"5 01
B:Mouse Conjugation
10"' 10 1
CVIiver(mmol/L)
10"
10"' 10"
CV liver (mmol/L)
•as
"5 •
Ł
O,.
O O1
C:Rat Oxidation
o o o o o o
*Ł
f b!
§
S'OTJ
"6 O!
D:Rat Conjugation
10"'' 10" 1
CV liver (mmol/L)
10"' 10" 1
CV liver (mmol/L)
f '
!__,
Ł cr?
g-
EiHuman Oxidation
0 o O O 0 O
10"" 10"' 10" 1
CV liver (mmol/L)
10"" 10";> 10"' 1
CV liver (mmol/L)
Figure 3-6. Comparison of mouse (A-B), rat (C-D), and human (E-F) rates of
hepatic oxidation (A, C, and E) or conjugation (B, D, and F) measured in
vitro (symbols) and predicted by the model (lines).
Data shown consist of measurements of tetrachloroethylene in vitro oxidation and conjugation
[solid circle: Dekant et al. (1998): solid square: Green et al. (1990): solid diamond: Lash et al.
(1998): solid triangle: Lash et al. (2007): solid upside-down triangle: Reitz et al. (1996) reported
fits of in vitro tetrachloroethylene Vmax and Km for oxidation [grey-filled circle: Costa and
Ivanetich (1980): grey-filled square: Costa and Ivanetich (1984): grey-filled diamond: Lipscomb
et al. (1998). TCE; grey-filled triangle: Wheeler et al. (2001) CH2I2; grey-filled upside-down
triangle: Wheeler et al. (2001). CH2C12]; and measurements of TCE in vitro conjugation [open
circle: Lash et al. (1998): open square: Lash et al. (1999): open diamond: Green et al. (1997)1.
Model predictions are using baseline parameters (dotted line), overall posterior mode parameters
(solid thick line), and alternative posterior mode parameters (grey lines).
3-38
-------
Table 3-2. Predictions for area-under-the-curve of tetrachloroethylene in
blood (mg-hr/L-day per ppm in air or mg-hr/L-day per mg/kg-day oral
intake) using posterior mode parameters
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Mouse
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
1.2
1.2
1.26
1.73
2.8
2.98
0.0217
0.0218
0.0221
0.0265
0.168
0.296
2.13
2.13
2.18
2.43
2.64
2.68
0.104
0.104
0.105
0.112
0.152
0.178
1.03
1.03
1.02
1.01
1
1
1.06
1.06
1.06
1.05
1.03
1.03
2.11-2.42
2.12-2.42
2.16-2.44
2.39-2.53
2.64-2.68
2.67-2.72
0.0957-0.126
0.0958-0.126
0.0965-0.127
0.103-0.129
0.138-0.152
0.159-0.18
Rat
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
1.03
1.03
1.04
1.11
2
2.4
0.0737
0.0738
0.0744
0.0816
0.23
0.543
2.25
2.25
2.25
2.25
2.29
2.39
0.852
0.852
0.852
0.854
0.864
0.912
1
1
1
1
1
1.01
1.02
1.02
1.02
1.02
1.02
1.02
2.25-2.27
2.25-2.27
2.25-2.27
2.25-2.27
2.28-2.32
2.36-2.42
0.807-0.86
0.807-0.86
0.807-0.86
0.809-0.861
0.821-0.869
0.869-0.919
3-39
-------
Table 3-2. Predictions for area-under-the-curve of tetrachloroethylene in
blood (mg-hour/L-day per ppm in air or mg-hour/L-day per mg/kg-day oral
intake) using posterior mode parameters (continued)
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Human
0.01 ppm
0.1 ppm
1 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
2.35
2.35
2.35
2.35
2.35
2.37
2.71
2.71
2.71
2.71
2.72
2.03
2.03
2.03
2.03
2.03
2.04
1.74
1.74
1.74
1.74
1.74
1.05
1.05
1.05
1.05
1.05
1.05
1.03
1.03
1.03
1.03
1.03
2.01-2.36
2.01-2.36
2.01-2.36
2.01-2.36
2.01-2.36
2.01-2.36
1.58-1.82
1.58-1.82
1.58-1.82
1.58-1.82
1.58-1.82
GSD = geometric standard deviation.
Dose metrics calculated based on daily average given continuous exposure for 100 weeks. Baseline
parameters are based on in vitro data for metabolic parameters, and literature values for other parameters.
Posterior modes are based on the updated parameters listed in Table 3-1.
3-40
-------
Table 3-3. Predictions for fraction of tetrachloroethylene oxidized by
cytochrome P450 (P450s) (mg/kg-day oxidized per mg/kg-day intake) using
posterior mode parameters
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Mouse
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.00252
0.00254
0.00269
0.0062
0.0141
0.00716
0.00367
0.00368
0.00374
0.00445
0.0253
0.0361
0.188
0.187
0.174
0.118
0.0732
0.0664
0.561
0.561
0.557
0.524
0.35
0.239
1.1
1.09
1.08
1.06
1.04
1.05
1.08
1.08
1.07
1.07
1.04
1.03
0.12-0.192
0.12-0.191
0.115-0.179
0.0934-0.124
0.0632-0.075
0.0574-0.0688
0.395-0.574
0.395-0.574
0.394-0.57
0.38-0.535
0.308-0.367
0.216-0.25
Rat
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.000501
0.000502
0.000514
0.000662
0.0025
0.00153
0.00143
0.00144
0.00145
0.00158
0.00431
0.00686
0.0419
0.0419
0.0418
0.0409
0.0331
0.011
0.106
0.106
0.106
0.105
0.0934
0.0434
1.02
1.02
1.02
1.02
1.07
1.27
1.02
1.02
1.02
1.02
1.04
1.2
0.0387-0.042
0.0387-0.042
0.0386-0.0419
0.0379-0.0409
0.0263-0.0358
0.00587-0.0181
0.0988-0.107
0.0988-0.107
0.0987-0.107
0.0977-0.105
0.0817-0.096
0.026-0.0631
3-41
-------
Table 3-3. Predictions for fraction of tetrachloroethylene in oxidized by
cytochrome P450 (P450s) (mg/kg-day oxidized per mg/kg-day intake) using
posterior mode parameters (continued)
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Human
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.00971
0.00971
0.00969
0.00955
0.00828
0.00355
0.0173
0.0173
0.0173
0.0169
0.0138
0.00492
0.0098
0.0098
0.0098
0.0098
0.0098
0.0098
0.0175
0.0175
0.0175
0.0175
0.0175
0.0175
1.12
1.12
1.12
1.12
1.12
1.12
1.09
1.09
1.09
1.09
1.09
1.09
0.00694-0.0104
0.00694-0.0104
0.00694-0.0104
0.00694-0.0104
0.00694-0.0104
0.00693-0.0104
0.0134-0.0184
0.0134-0.0184
0.0134-0.0184
0.0134-0.0184
0.0134-0.0184
0.0133-0.0184
GSD = geometric standard deviation.
Dose metrics calculated based on daily average given continuous exposure for 100 weeks. Baseline
parameters are based on in vitro data for metabolic parameters, and literature values for other parameters.
Posterior modes are based on the updated parameters listed in Table 3-1.
3-42
-------
Table 3-4. Predictions for fraction of tetrachloroethylene conjugated with
glutathione (GSH) (mg/kg-day conjugated per mg/kg-day intake) using
posterior mode parameters
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Mouse
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.348
0.347
0.337
0.244
0.0299
0.00301
0.929
0.929
0.928
0.914
0.454
0.0485
0.000151
0.000152
0.000159
0.000207
0.000251
0.000258
0.000481
0.000481
0.000485
0.000521
0.000706
0.000821
3.87
3.87
3.86
3.81
3.79
3.79
3.89
3.89
3.89
3.87
3.82
3.81
6.39e-05-0.00415
6.43e-05-0.00417
6.83e-05-0.0043
8.95e-05-0.00523
0.000109-0.00642
0.000111-0.00663
0.000208-0.0134
0.000208-0.0134
0.00021-0.0135
0.000229-0.0141
0.00031-0.0181
0.000362-0.0212
Rat
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.303
0.303
0.301
0.286
0.0939
0.0099
0.874
0.874
0.873
0.861
0.608
0.078
0.00308
0.00308
0.00309
0.00309
0.00316
0.00335
0.00783
0.00783
0.00783
0.00785
0.00795
0.00838
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
0.00195-0.00519
0.00195-0.00519
0.00195-0.0052
0.00196-0.00521
0.002-0.00529
0.00213-0.00559
0.00498-0.0133
0.00498-0.0133
0.00498-0.0133
0.00499-0.0133
0.00506-0.0134
0.00535-0.0141
3-43
-------
Table 3-4. Predictions for fraction of tetrachloroethylene in conjugated
with glutathione (GSH) (mg/kg-day conjugated per mg/kg-day intake) using
posterior mode parameters (continued)
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Human
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.000544
0.000543
0.000543
0.000535
0.000468
0.000207
0.000972
0.000972
0.00097
0.00095
0.000788
0.000289
0.0936
0.0936
0.0936
0.0936
0.0935
0.0926
0.177
0.177
0.177
0.177
0.177
0.175
17.5
17.5
17.5
17.5
17.5
17.4
17.1
17.1
17.1
17.1
17.1
17
3.16e-05-0.1
3.16e-05-0.1
3.16e-05-0.1
3.16e-05-0.1
3.16e-05-0.1
3.16e-05-0.0991
6.47e-05-0.188
6.47e-05-0.188
6.47e-05-0.188
6.47e-05-0.188
6.47e-05-0.187
6.47e-05-0.185
GSD = geometric standard deviation.
Dose metrics calculated based on daily average given continuous exposure for 100 weeks. Baseline
parameters are based on in vitro data for metabolic parameters, and literature values for other parameters.
Posterior modes are based on the updated parameters listed in Table 3-1.
3-44
-------
Table 3-5. Predictions for Trichloroacetic acid (TCA) produced systemically
(mg/kg-day systemic TCA per ppm in air or mg/kg-day systemic TCA per
mg/kg-day oral intake) using posterior mode parameters
Species/
continuous exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Mouse
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.0361
0.0363
0.0384
0.0886
0.202
0.103
0.00325
0.00326
0.00331
0.00394
0.0224
0.032
3.74
3.71
3.45
2.34
1.46
1.32
0.497
0.496
0.493
0.464
0.31
0.212
1.08
1.08
1.07
1.04
1.03
1.04
1.08
1.08
1.07
1.07
1.04
1.03
2.63-3.94
2.62-3.9
2.53-3.59
2.05-2.47
1.36-1.55
1.18-1.43
0.35-0.509
0.35-0.508
0.349-0.505
0.337-0.473
0.273-0.325
0.191-0.222
Rat
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.00352
0.00353
0.00361
0.00465
0.0176
0.0108
0.00127
0.00127
0.00128
0.0014
0.00382
0.00607
0.182
0.182
0.181
0.177
0.144
0.0476
0.0941
0.0941
0.094
0.0929
0.0828
0.0385
1.02
1.02
1.02
1.02
1.07
1.26
1.02
1.02
1.02
1.02
1.04
1.2
0.173-0.189
0.173-0.189
0.173-0.189
0.169-0.183
0.117-0.158
0.0261-0.0798
0.0875-0.0952
0.0875-0.0951
0.0874-0.095
0.0866-0.0934
0.0724-0.0851
0.023-0.0559
3-45
-------
Table 3-5. Predictions for Trichloroacetic acid (TCA) produced systemically
(mg/kg-day systemic TCA per ppm in air or mg/kg-day systemic TCA per
mg/kg-day oral intake) using posterior mode parameters (continued)
Species/continuous
exposure
Baseline
Posterior
mode
GSDof
posterior
modes across
chains
Range of posterior
modes across chains
Human
0.01 ppm
0.1 ppm
1 ppm
10 ppm
100 ppm
1,000 ppm
0.01 mg/kg-day
0.1 mg/kg-day
1 mg/kg-day
10 mg/kg-day
100 mg/kg-day
1,000 mg/kg-day
0.0106
0.0106
0.0106
0.0104
0.00906
0.00388
0.0153
0.0153
0.0153
0.015
0.0123
0.00436
0.0125
0.0125
0.0125
0.0125
0.0125
0.0125
0.0145
0.0145
0.0145
0.0145
0.0145
0.0145
1.02
1.02
1.02
1.02
1.02
1.02
1.09
1.09
1.09
1.09
1.09
1.09
0.0117-0.0128
0.0117-0.0128
0.0117-0.0128
0.0117-0.0128
0.0117-0.0128
0.0117-0.0128
0.0111-0.0152
0.0111-0.0152
0.0111-0.0152
0.0111-0.0152
0.0111-0.0152
0.011-0.0152
GSD = geometric standard deviation.
Dose metrics calculated based on daily average given continuous exposure for 100 weeks. Baseline
parameters are based on in vitro data for metabolic parameters, and literature values for other parameters.
Posterior modes are based on the updated parameters listed in Table 3-1.
3-46
-------
Table 3-6. Summary evaluation of the reliability of tetrachloroethylene dose
metrics
Dose metric
species
Calibration
error or
variability
(GSD)a
Validation
error or
variability
(GSD)a
Optimization
runs range"
Additional potential concerns
from sensitivity analysis
AUCCBld
Mouse
Rat
Human
~2-fold
~2-fold
~2-fold
~2-fold
~2-fold
~2-fold
<10%
<10%
<20%
None
None
None
FracOx
Mouse
Rat
Human
~2-fold
~2-fold
~2-fold
~2-fold
~2-fold
~3-fold
<40%
<20%
<1. 5-fold
Some sensitivity to lung
metabolism
None
Some sensitivity to fraction of
oxidation to TCA
FracGSH
Mouse
Rat
Human
NA
~2-fold
~2-fold
NA
NA
NA
-60-fold
<30%
-3,000-fold
None
None
Calibration data cannot
distinguish between modes
TCASys
Mouse
Rat
Human
~2-fold
~2-fold
~2-fold
~2-fold
~2-fold
~3-fold
<30%
<20%
<40%
Some sensitivity to fraction of
oxidation to TCA
Some sensitivity to fraction of
oxidation to TCA
Some sensitivity to fraction of
oxidation to TCA
"Evaluated in rodents at 10 ppm in air by inhalation and 100 mg/kg-day orally, and in humans at 0.01 ppm
in air by inhalation and 0.01 mg/kg-day orally.
GSD = geometric standard deviation.
3-47
-------
The highest confidence dose metric in the Chiu and Ginsberg (2011) analysis is the AUC
of tetrachloroethylene in blood (refer to Table 3-2). The main source of uncertainty in this case
is the residual difference between the model predictions and the calibration and validation
data—a factor of about twofold for each species. Therefore, this dose metric should be
considered reliable for use in risk assessment with the acknowledgement of a possible twofold
residual error.
The next highest confidence as seen from Table 3-6 is in the estimates of
tetrachloroethylene oxidation and TCA formation. Here, the estimates of tetrachloroethylene
oxidation in mice and rats have similar uncertainty to that for AUC of tetrachloroethylene in
blood—predominantly twofold in the residual difference between model predictions and the
calibration and validation data. The range of estimates of tetrachloroethylene oxidation in
humans is largely dominated by interindividual variability—i.e., the differences in urinary
excretion of TCA across individuals. Thus, the central tendencies for the population are well
estimated—even if particular individuals may vary to a fair degree. Thus, at the population
level, these dose metrics should be considered reliable for use in risk assessment with the
acknowledgement of a residual error of about twofold or less.
In terms of predicted interspecies differences, the PBPK model generally predicts the
greatest oxidative metabolism in mice, followed by rats, and then humans. Humans would be
predicted to receive a smaller internal dose of oxidative metabolites for the same applied dose,
whether scaled by body weight or allometrically by body weight to the 3/4 power.
On the other hand, estimates of GSH conjugation appear more uncertain—especially for
humans. In rats, both the calibration data and the range of different optimization runs suggest
about a twofold uncertainty. In mice, there are no data on this pathway other than as a "mass
balance" from total metabolism (e.g., closed-chamber studies). Nonetheless, the range of
estimates based on the different optimization runs is about 60-fold. It is in the human predictions
that the range of estimates becomes extraordinarily large. In particular, there are evidently
two local maxima, each of which gives similar model fits, but for which model predictions differ
by 3,000-fold. This is a reflection not of the calibration data, which are fit quite well regardless,
but of the results of different optimization runs. Therefore, overall, the predictions for rat GSH
conjugation are considered reliable to about twofold, those for the mouse to about 60-fold, and
those for humans vary by about 3,000-fold. At this point, it is not possible to disentangle the
contributions of uncertainty and variability to the very large range of estimates of
tetrachloroethylene GSH conjugation in humans.
Interestingly, the predictions appear to support the default assumption of equivalent
concentrations in air leading to equivalent internal doses, as the estimates of AUC of
tetrachloroethylene in blood are within twofold of each other across species. In addition, at the
3-48
-------
higher oral doses (e.g., 100 mg/kg-day), reseating the AUC in blood by body weight to the
% power leads to estimates across species within threefold of each other. These can be explained
by the sensitivity analysis, which showed AUC in blood to be most sensitive to cardiac output,
alveolar ventilation, and the partition coefficient, all of which either are similar across species or
scale approximately allometrically by body weight to the % power across species.
The implications of these results are quite substantial—particularly for interspecies
extrapolation between rats and humans. In rats, all the evidence appears to support a low amount
(<1% of dose) of GSH metabolism. At environmental exposures, the overall posterior mode
predicts about 15- to 30-fold more GSH conjugation as a fraction of dose in humans relative to
rats, but the uncertainty range in humans overlaps with the rat estimates, so the data are also
consistent with humans having either equal or greater GSH conjugation.
The analysis in Chiu and Ginsberg (2011) appears to have resolved a conflict between
PBPK model-based analyses that predicted high versus low amounts of tetrachloroethylene
metabolized in humans in two key aspects. This makes it particularly suited for use in this
assessment. First, there is now fairly high confidence in the predictions ofoxidative metabolism
across species. Second, these results make it clear that the previously debated uncertainties in
total metabolism can be essentially attributed to uncertainty in GSH conjugation, which is
substantial. Those analyses that concluded low total tetrachloroethylene metabolism all
restricted the fraction of total (not oxidative) metabolism that was TCA to a fairly significant
percentage—30-100% re.g..(Qiu et al.. 2010: Covington et al.. 2007: Clewell et al.. 2005: Chen
and Blancato, 1989)]. Thus, as was noted by the NRC (2010), total metabolism was essentially
only measuring oxidative metabolism. On the other hand, those analyses that concluded high
total tetrachloroethylene metabolism essentially lumped oxidative and GSH conjugation
metabolism together without restrictions as to the fraction producing TCA and/or made
inferences based on disappearance of the parent compound [e.g., (Chiu and Bois, 2006: Bois et
al.. 1996: Reitzetal.. 1996: Boisetal.. 1990: Wardetal.. 1988)1. Thus, the analysis in Chiu and
Ginsberg (2011) essentially reconciles the disparate conclusions as to human tetrachloroethylene
metabolism from previously published PBPK models. First, the conclusion of "low metabolism"
is certainly true for oxidation. Second, the conclusion of "high metabolism" may be true for
GSH conjugation but is highly uncertain. In essence, both conclusions are consistent with the
data if augmented by some additional qualifications: oxidative metabolism is low in humans,
while GSH conjugation metabolism may be high or low in humans, with high uncertainty and/or
variability (Chiu and Ginsberg, 2011).
Results obtained by applying the Chiu and Ginsberg (2011) model for the dose-response
modeling in this assessment are presented in Section 5.
3-49
-------
3.5.2. Age and Gender-Related Differences in Tetrachloroethylene Pharmacokinetics
Age and gender-specific differences in pharmacokinetics can have a substantial 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 were 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 3-7 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 3-7.
Considerable gender differences in blood concentrations of TCA and tetrachloroethylene
were observed 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.
3-50
-------
Table 3-7. Ratio of average daily dose at various life-stages to the average
daily dose for a 25-year-old adult: physiologically based pharmacokinetic
(PBPK) simulations
Dose metric
PCE 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
Source: Clewell et al. (2004).
1.5E-4T
1.0E-4 --
O
tr:
UJ
CL
O
O 5.0E-5 - -
CO
- - 4.0E-8
--3.0E-8 O
PERC - males
PERC - females
TCA - males
TCA - females
O.OE+0 FI 111111111 i 111111 11 I 111111111111 1111111
5.0E-8
- - 2.0E-8
-- 1.0E-8
jj
o
O
o
CD
O.OE+0
0 10 20 30 40 50 60 70 80
Age (years)
Figure 3-7. Physiologically based pharmacokinetic (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 ug/kg-day.
The same group of authors [i.e., (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 nursing infant was modeled using data
for CYP2E1 (Vieira et al., 1996): metabolism in the fetus was not included due to lack of
information pertaining to the development of this pathway during gestation. The dose metrics
were the fetal and infant blood concentrations of tetrachloroethylene and TCA. Changes in fetal
3-51
-------
blood concentrations were not pronounced because changes in tissue composition occurred in
both the mother and the fetus during pregnancy (Gentry et al., 2003). A decrease of nearly
three orders of magnitude of blood concentrations in the lactating infant when compared with
that of the fetus was calculated. This decrease was attributed to the lower exposure rate during
lactation as compared with placental exposure. Concentrations in the lactating infant were
considerably lower, by more than two orders of magnitude, than in the mother. The largest
variation in blood concentration occurred in the early postnatal period.
As the authors indicated, validation of the results in the Clewell et al. (2004) and Gentry
et al. (2003) work and further refinement of the parameters in the models are necessary. It
would, therefore, be premature to consider the results of such analyses for use in risk assessment.
Further investigation of variability in the parameters used in the Clewell et al. (2004) analysis is
needed before the results from Table 3-3 can be used to weigh upon considerations of a
pharmacokinetic uncertainty factor for age and gender variability. Nonetheless, these models
will enable future studies to focus on the key factors that are likely to influence pharmacokinetic
susceptibility.
3.5.3. Metabolic Interactions with Other Chemicals
Fisher et al. (2004) used PBPK modeling and complementary studies in mice to
investigate the effect of coexposures of orally administered carbon tetrachloride (CT) and
tetrachloroethylene on metabolic interactions between the two chemicals. CT is known to inhibit
its own metabolism (referred to as suicide inhibition). TCA was used as a biomarker to assess
the inhibition of the cytochrome P450 system by CT. Oral bolus intubation in the dose range of
1-100 mg/kg of CT was followed by a dose of 100 mg/kg of tetrachloroethylene an hour later. It
was concluded that dose additivity could not be used to predict interactions between the
compounds in this dose range because the metabolic interactions were found to be highly
nonlinear. The inhibition in metabolic capacity of tetrachloroethylene 2 hours after
administration of CT and 1 hour after single dose administration of tetrachloroethylene was
found to be 5, 52, and 90% at CT doses of 1.5, 10, and 19 mg/kg, respectively.
Dobrev et al. (2002) performed gas uptake studies in F344 rats and developed a mixture
PBPK model for humans to study interaction effects during coexposure to mixtures of TCE,
tetrachloroethylene, and methyl chloroform. Corresponding to a 10% increase in TCE blood
concentration, the production rates of toxic conjugative metabolites exceeded 17%, pointing to a
nonlinear interaction effect due to coexposure to TCE.
3-52
-------
4. HAZARD IDENTIFICATION
This section discusses tetrachloroethylene toxicity on an organ-specific basis. For each
of the major organ systems, human effects are presented first, followed by effects in animals and
in in vitro systems. Cancer and noncancer toxicity and mode of action (MOA) are also included
in the discussions. The order of presentation is as follows: neurotoxicity (refer to Section 4.1);
kidney and bladder toxicity and cancer (refer to Section 4.2); liver toxicity and cancer (refer to
Section 4.3); esophageal cancer (refer to Section 4.4); lung and respiratory cancer (refer to
Section 4.5); immunotoxicity, hematologic toxicity, and cancers of the immune system (refer to
Section 4.6); developmental and reproductive toxicity, and reproductive cancers (refer to Section
4.7); genotoxicity (refer to Section 4.8); and susceptible populations (refer to Section 4.9).
Section 4.10 provides a summary of the hazard identification.
The database of published epidemiologic studies on cancer and tetrachloroethylene
exposure was examined to assess its ability to inform the cancer hazard from tetrachloroethylene
exposure. The analysis of epidemiologic studies on cancer and tetrachloroethylene presented in
Appendix B documents each study's essential design features, exposure assessment approaches,
statistical analyses (including assessment of exposure- or concentration-response), and potential
sources of confounding and bias. This analysis supports the discussion of site-specific cancer
observations in Sections 4.2-4.7. In those sections, study findings for site-specific cancers are
presented with an assessment and discussion of their overall weight of evidence. The key
considerations in the weight of evidence are: study design, exposure assessment methodologies,
exposure- or concentration-response, and the potential for alternative explanations, including
bias and confounding. Greater weight is given to studies that (i) employ a cohort or case-control
design, (ii) use exposure assessment methodologies with a relatively high level of sensitivity and
specificity, and a low likelihood of exposure misclassification (iii) show a exposure- or
concentration-response gradient, and (iv) have less potential for alternative explanations.
Sample size (number of cases in a cohort study; number of exposed in a case-control study) was
also considered, but a larger sample size in itself did not outweigh the considerations based on
type of exposure assessment methodology. Studies that are more limited in one or more of these
characteristics are accorded lesser relative weight, but are not necessarily excluded from the
overall weight of evidence.
4-1
-------
4.1. NEUROTOXICITY
4.1.1. Human Studies
A wide range of effects on neurologic function have been observed for both acute and
chronic-duration exposure to tetrachloroethylene in humans, as summarized below. Most of the
reports evaluating neurological function in humans were inhalation chamber or chronic exposure
studies. Study designs, exposure-assessment methods, and results of individual studies are
presented with a discussion of chamber studies in Section 4.1.1.1 and chronic exposure studies in
Section 4.1.1.2. Within the latter section, the studies are further divided by type of exposure
setting (occupational; residential). In residential settings, exposure is more likely to be
continuous and of lower concentrations compared with the more intermittent, higher
concentration, more variable exposure experienced in work settings. Section 4.1.1.3 presents a
summary of neuropsychological and neurobehavioral effects in low- and moderate-exposure
studies with observations across studies discussed by neurological domain, categorized by visual
function, cognitive function, motor function, and neurological and neurobehavioral disorders.
Acute controlled inhalation exposures of 100 ppm and higher induced symptoms
consistent with depression of the central nervous system (CNS), such as dizziness and
drowsiness. Changes in visual function as measured by electroencephalograms (EEGs) have
also been noted with controlled inhalation exposures at this level (Stewart et al., 1977). Acute
exposure to lower levels of tetrachloroethylene (50 ppm for 4 hours/day for 4 days) induced
alterations in neurobehavioral function, with changes indicative of visual system dysfunction
including delayed neuronal processing time (Altmann et al., 1992; Altmann et al., 1990). A wide
range in susceptibility to neurological effects among the participants in these studies was
observed.
Epidemiologic studies of workers or residents with chronic exposure to
tetrachloroethylene show that the nervous system is a target, with most of these studies reporting
decrements in one or more nervous system domains. The visual and cognitive domains are most
commonly affected (NYSDOH. 2010: McDermott et al.. 2005: NYSDOH, 2005a, b; Sharanjeet-
Kaur et al.. 2004: Schreiber et al.. 2002: Gobbaetal., 1998: Spinatonda et al.. 1997: Altmann et
al., 1995: Echeverria et al., 1995: Cavalleri etal., 1994: Echeverria et al., 1994: Ferroni et al.,
1992: Nakatsuka et al.. 1992: Seeber, 1989: Lauwerys et al.. 1983). Other reports (Till et al..
2005: Laslo-Baker et al.. 2004: Till etal.. 200 la: Till et al., 2001b) suggest a vulnerability of the
fetus to organic solvent exposures, including tetrachloroethylene exposure. Deficits in
neurobehavioral parameters and in visual system functioning in young children of mothers
exposed during pregnancy compared with children of unexposed mothers were observed (Till et
al., 2005: Till et al., 200la: Till et al., 2001b). These reports are not discussed further in this
4-2
-------
section because they do not provide specific data pertaining to tetrachloroethylene exposure.
Few studies are available on neurologic diseases such as Parkinson's disease, amyotrophic lateral
sclerosis, and Alzheimer's disease and organic solvents (TOM, 2002), and none of these reports
uniquely assess tetrachloroethylene. The influence of tetrachloroethylene exposure on risk of
these neurological diseases is not addressed in this Toxicological Review.
4.1.1.1. Chamber Studies
Several controlled experiments were conducted in the 1970s examining neurological
effects from short-term exposures (5-7.5 hours per day for 4 or 5 consecutive days) to
tetrachloroethylene at levels up to 100 ppm. There is no description in the published reports of
the informed consent and other human subjects research ethics procedures undertaken in these
studies, but there is no evidence that the conduct of the research was fundamentally unethical or
significantly deficient relative to the ethical standards prevailing at the time the research was
conducted.
In a study by Stewart et al. (1970), 12 healthy adults were exposed to 100 ppm for
7 hours; neurological symptoms including eye and nose irritation was reported by 60% of the
subjects, a slight frontal headache by 26%, mild lightheadedness by 26%, drowsiness by 40%,
and difficulty speaking by 25%. Of five healthy men exposed to 100 ppm, for 7 hours/day, on 5
consecutive days, one reported a mild frontal headache during each exposure, and two
consistently reported mild eye and throat irritation. Individual responses during exposures to 0
ppm were not assessed. Three tests of equilibrium (a modified Romberg test, where an
individual stands on one foot with eyes closed and arms at side; a heel-to-toe test; and a finger-
to-nose test) were performed every 60 minutes during each day of exposure. After 6 hours,
neurobehavioral tests of motor function (the Crawford manual dexterity and Flanagan
coordination tests), cognitive function (arithmetic test), and motor/cognitive function (inspection
test) were also performed. Three of the subjects exhibited impairments to equilibrium within the
first 3 hours of exposure but were able to perform the test normally when given a second chance.
Stewart et al. (1970) concluded that there were CNS effects in some subjects exposed to 100
ppm and that there exists a large range of individual susceptibility to tetrachloroethylene.
In the 6-week study by Hake and Stewart (1977), four healthy men were exposed
7.5 hours/day to 0 ppm (2 days in Week 1, 1 day in Week 3, and 2 days in Week 6), 21 ppm
(4 consecutive days in Week 3), 100 ppm (5 consecutive days in Week 2), and a time-weighted
average (TWA) of 100 ppm (5 consecutive days in Week 4) when exposure levels were more
than 53, 100, or 155 ppm (5 consecutive days in Week 5). In addition, four healthy women were
exposed to 100 ppm for 7.5 hours/day on 5 consecutive days and to 0 ppm on 2 days. The
subjects were told that they would be exposed to various concentrations of tetrachloroethylene,
4-3
-------
but they were not told their sequence of exposures (a single-blind protocol). Reports of
symptoms (e.g., headache) varied among individuals, but overall, complaints during exposures
were similar to those during control conditions—exposures to 0 ppm tetrachloroethylene. The
evaluation of visual function through EEG recordings made during exposure suggested altered
patterns indicative of cortical depression in three of four men and four of five women exposed to
100 ppm (constant or TWA). In five subjects, altered EEG recordings occurred during hours 4 to
7 of exposure; another subject had altered recordings within 10 minutes of exposure, which
gradually returned to normal during continued exposure, and the seventh subject showed changes
between 30 minutes and 6-7 hours of exposure. Recordings of visual evoked potentials (a
measure of visual function) in response to bright flashes of light (i.e., neurophysiological
measurements of the electrical signals generated by the visual system in response to visual
stimuli) and equilibrium tests (Romberg and heel-to-toe) were normal in men and women. The
performance of men on neurobehavioral tests of cognitive function (arithmetic), motor function
(alertness), motor/cognitive function (inspection), and time estimation was not significantly
affected by any exposure. The performance of men on a second test of motor function (Flanagan
coordination) was significantly decreased (p < 0.05) on 1 of 3 days during each of 2 weeks of
exposure to 100 ppm and on 2 of 3 days during the week of exposure to 155 ppm, but Hake and
Stewart (1977) concluded that only the results at 155 ppm were related to tetrachloroethylene. In
women, alertness (the only neurobehavioral endpoint evaluated) was not affected by exposure to
tetrachloroethylene. Hake and Stewart (1977) concluded that (1) there is considerable
interindividual variation in response to tetrachloroethylene vapors, (2) visual function changes
through EEG analysis indicates preliminary signs of narcosis in most subjects exposed to 100
ppm for 7.5 hours, (3) impairment of coordination may occur in subjects exposed to 155 ppm for
7.5 hours, and (4) the effects are likely due to tetrachloroethylene itself, given its slow
metabolism in humans. They also reported that their data suggested that a threshold limit value
of 100 ppm contains no margin of safety for susceptible subjects—both subjectively and
neurologically—to the vapors of tetrachloroethylene.
Altmann et al. (1992: 1990) examined neurological effects of tetrachloroethylene on
healthy adults exposed to 10 ppm or 50 ppm for 4 hours on 4 consecutive days. Visual acuity of
all subjects was normal or corrected to normal. The study was a single-blind study (subjects
were not told their level of exposure), and subjects were randomly assigned to either group.
Sixteen subjects were exposed to 10 ppm, and 12 subjects were exposed to 50 ppm. However,
neurophysiological measurements were made on only 22 subjects (12 at the low-exposure level
and 10 at the high-exposure level). Three neurophysiological measurements evaluating visual
and auditory function were taken on the day before exposure started and on each of the four
exposure days: (1) visual evoked potentials in response to black-and-white checkerboard
4-4
-------
patterns; (2) a visual contrast sensitivity (VCS) test; and (3) recordings of brainstem auditory-
evoked potentials (neurophysiological measurements of the electrical signals generated by the
hearing system in response to auditory stimuli) to evaluate peripheral hearing loss. All
measurements were started 2 hours after a subject entered the chamber and were completed
within 1 hour. A German version of the Neurobehavioral Evaluation System was used to assess
motor, motor/cognitive, and cognitive function of subjects. The battery included nine tests
(finger tapping, eye-hand coordination, simple reaction time, continuous performance, symbol
digit, visual retention, pattern recognition, digit span, and paired associates). A vocabulary test
and a test of emotional state (moods) were also given. Each subject was assessed with a
complete battery of tests during the preexposure baseline assessment and at the end of the study.
Subsets of the battery covering motor function and mood were given at the beginning and end of
each 4-hour exposure period. Tetrachloroethylene was not detected in blood samples collected
before the start of the first exposure period. The detection limit was less than 0.0005 mg/L.
Mean tetrachloroethylene blood levels increased slightly over the 4-day period. Among subjects
exposed to 10 ppm, mean blood levels were 0.33, 0.36, 0.4, and 0.38 mg/L at the end of days 1,
2, 3, and 4 of exposure, respectively. Among subjects exposed to 50 ppm, mean blood levels
were 1.1, 1.2, 1.4, and 1.5 mg/L at the end of days 1, 2, 3, and 4 of exposure, respectively.
The visual evoked potential latencies of subjects during the 3rd hour of exposure to
50 ppm on Days 1, 2, 3, and 4 of exposure were significantly longer (p < 0.05) compared with
those measured on the control day, and the differences became progressively longer on
successive exposure days. One set of visual evoked potential latencies on the day after the end
of the exposure period remained longer than the control day values (statistical significance not
reported). Visual evoked potential latencies in subjects with exposure to 10 ppm were not
statistically significantly longer than those recorded on the control day. There were significant
differences (p < 0.05) between the visual evoked potential latencies of subjects exposed to
10 ppm and those exposed to 50 ppm. Data on visual contrast sensitivity indicated greater
effects at 50 ppm than at 10 ppm; effects were most pronounced on the last day of exposure.
However, statistical analysis was not reported. There were no indications of peripheral hearing
loss at either exposure level. Neurobehavioral tests results were reported only for those tests
given repeatedly on 4 consecutive days (finger tapping, eye-hand coordination test, simple
reaction time, continuous performance, and moods). There were postexposure performance
deficits (p = 0.05) among subjects exposed to 50 ppm when compared with the group exposed to
10 ppm in tests of motor/cognitive function (continuous performance test for vigilance) and
motor function (eye-hand coordination), and a near-significant difference (p = 0.09) on a test of
motor function (simple reaction time). In all cases, the degree of improvement shown by the
subjects exposed to 50 ppm was less than that shown by the subjects exposed to 10 ppm. There
4-5
-------
were no exposure-related effects on the finger-tapping or moods test. Altmann et al. (1990)
concluded that visual function in healthy, young, adult males is mildly affected by
tetrachloroethylene exposures to 50 ppm maintained for 4 hours on each of 4 days and stated that
the impaired performance on tests of motor/cognitive and motor function suggests that 50 ppm
cannot be considered a NOAEL for neurobehavioral endpoints indicative of CNS depression
(Altmann et al.. 1992).
4.1.1.2. Chronic Exposure Studies
Table 4-1 summarizes details of the chronic-duration tetrachloroethylene exposure
studies evaluating neurological function using tests of specific neurological domains in humans.
Most of these are studies of dry-cleaning and laundry workers, but some studies examined
neurobehavioral or visual system effects among residents living in close proximity to a dry-
cleaning establishment (NYSDOH. 2005a. b; Schreiber et al.. 2002: Altmann et al.. 1995) or in
other workers employed in the same building as a dry-cleaning business (Schreiber et al., 2002).
Exposure levels were approximately an order of magnitude higher in occupational settings
compared with residential exposure. Tetrachloroethylene concentrations reported in the dry-
cleaning and laundry worker studies ranged from an 8-hour TWA mean of 6 ppm for dry-cleaner
and ironing workers in Cavalleri et al. (1994) to an 8-hour TWA of 41 ppm for operators of a
wet-transfer dry-cleaning machine in Echeverria et al. (1995). Mean tetrachloroethylene
concentrations in residences near a dry-cleaning business were 0.4 ppm and 0.7 ppm,
respectively, in studies in New York City (Schreiber et al., 2002) and Germany (Altmann et al.,
1995). Two additional studies examining color vision in solvent-exposed workers (Muttray et
al., 1997) and peripheral neuropathy among patients with solvent-induced encephalopathy
(Albers et al., 1999) were identified but are not presented because they involved solvent
mixtures.
4-6
-------
Table 4-1. Summary of human neurotoxicity studies of occupational
residential exposures to dry-cleaning facilities using tetrachloroethyl
or
lene
Subjects, methods
Exposure levels
Results
Reference(s)
Occupational exposures: dry-cleaning settings
Belgium, 26 dry cleaners, 33
unexposed workers (controls),
B, EA, PA, U; not blinded to
exposure status
Germany, 101 dry cleaners
(both sexes), 84 unexposed
workers (controls). PA, AA;
blinded to exposure status
China, 64 dry cleaners,
120 controls (clerical workers
in factories). PA; not blinded to
exposure status
Italy, 60 dry cleaners, 30
controls (hospital launderers, no
solvent use). B, A; blinded to
exposure level but not status
Italy, 22 dry cleaners and 13
ironers, 35 controls. PA, EA;
blinded to exposure level
Italy, 33 dry cleaners and
ironers, self controls [baseline
measurements in Cavalleri et al.
(1994)1. PA; not clear if
blinded
Michigan, 65 dry cleaners,
pressers, clerks; no unexposed
group, PA; blinded to exposure
level
Mean TWA = 21 ppm, mean
duration = 6.4 yr
Low-exposure group (n = 57):
mean TWA =12 ppm, mean
duration = 11.8 yr; high-
exposure group (n = 44): mean
TWA = 53 ppm, mean duration
= 10.6 yr
Geometric mean TWA =15
ppm (males), 1 1 ppm
(females), duration not
reported
Mean TWA =15 ppm, mean
duration = 10.1 yr
Mean TWA = 6 ppm (7.3 ppm,
dry-cleaning workers; 4.8 ppm,
ironers), mean duration = 8.8
yr
Geometric mean TWA ppm:
Group A Group B
(n = 19) (n = 14)
Baseline 1.67 2.95
Follow-up 4.35 0.66
Chronic exposure score based
on work history: low (n = 24;
2.1 yr), moderate (n = 18; 3.9
yr), high (n = 23; 14.6 yr)
Statistically significant differences
for simple reaction time (before
work) and critical flicker fusion
(before and after work); better
scores in exposed workers.
Decrease in information-processing
speed (perceptual threshold, choice
reaction time), visual scanning
(cancellation d2 test), visuospatial
memory (digit reproduction) in dry
cleaners compared with controls;
no statistically significant
difference between high- and low-
exposure groups. No fine motor
function deficits.
No effect on color vision loss
(using less sensitive Lanthony
test).
Impaired performance on simple
reaction time, vigilance, stress. No
fine motor function deficit. No
effects on digit symbol test. No
dose-response patterns observed.
CCI elevated among all workers
(p = 0.025) and dry cleaners (p =
0.007); statistically significant
exposure (TWA)-response
relationship. No effect observed in
ironers.
Increased CCI in Group A
(p < 0.01); no change in Group B.
CCI correlated with exposure
levels (r = 0.38, p< 0.05).
Statistically significant decrease in
high compared with low exposure
on three tests of visuospatial
memory. No effect on digit span.
Lauwerys et
al. (1983)
Seeber (1989)
Nakatsuka et
al. (1992)
Ferroni et al.
(1992)
Cavalleri et
al. (1994)
Gobba et al.
(1998)
[follow up of
Cavalleri et
al. (1994)]
Echeverria et
al. (1995)
4-7
-------
Table 4-1. Summary of human neurotoxicity studies of occupational
residential exposures to dry-cleaning facilities using tetrachloroethyl
(continued)
or
ene
Subjects, methods"
Washington, 45 dry cleaners
matched to 69 laundry workers,
59 pressers, or counter clerks
from the same shop as the dry
cleaner operator. PA; blinded
to exposure level
Italy, 35 dry cleaners, 39 age-
and education-matched
controls. AA; not blinded to
exposure status
Malaysia, 14 dry cleaners, 29
controls (support staff of
Universiti Kebangsaan
Malaysia, control Group 2); not
blinded to exposure status
Israel, 88,820 births,
1964-1976, identified in
Jerusalem Perinatal Study,
linked to national Psychiatric
Registry for hospitalization
with a schizophrenia-related
diagnosis through 1997
Exposure levels
Chronic exposure score groups
based on detailed work history
and estimated measures: mean
= 0,68, and 1,1 50 with
corresponding 8-h TWAs of
<0.2, 3, and 9 ppm. Mean
duration = 2.6 to 1 1 yr for low-
and high-exposure groups,
respectively
Median = 8 ppm, grab sample.
Mean duration of employment
= 10.6 yr (from Figure 2)
No exposure information
presented in paper other than
PCE was used for dry cleaning
Occupation of mother and
father listed as dry cleaner on
birth certificate
Results
Evidence of associations between
chronic exposure and reduced test
performance on three tests of
visuospatial memory: switching (p
= 0.10), pattern memory (p = 0.03),
and pattern recognition (p = 0.09).
Increase in vocal reaction time to
visual stimuli (reading task); dose-
response relationship.
43 and 93% of dry cleaners
compared to 0 controls had errors
on the color vision D-15 test and
FM 100 Hue test, respectively.
Number of errors on FM 100 Hue
test also increased in dry cleaners
(p<0.05).
Four cases were identified in 144
offspring of dry cleaners. RR of
3.4(95%CI: 1.3-9.2) for
schizophrenia in the offspring of
dry cleaners using proportional
hazard modeling.
Reference(s)
Echeverria et
al. (1994)
Spinatonda et
al. (1997)
Sharanjeet-
Kaur et al.
(2004)
Perrin et al.
(2007)
Occupational exposures: other settings
New York, 9 employees of day-
care center located in a building
with a dry-cleaning business, 9
age- and gender-matched
unexposed controls. PA, EA,
B, U; not blinded to exposure
status
New York, 4-yr follow-up of 13
children who had attended a day
care located in a building with a
dry-cleaning business, 13
children matched to exposed
children on age, gender, and
daycare experience; not blinded
to exposure status
Mean = 0.32 ppm (monitoring
before closure of dry cleaners).
No information on duration of
employment
Exposure had ceased 4 yr
earlier
Decreased color discrimination
among exposed but not statistically
significant.
Lower (worse) scores on tests of
visual contrast sensitivity.
No difference in visual function
(VCS, CCI) or neurobehavioral
function between exposed children
and controls.
Schreiber et
al. (2002)
NYSDOH
(2005b)
4-8
-------
Table 4-1. Summary of human neurotoxicity studies of occupational
residential exposures to dry-cleaning facilities using tetrachloroethyl
(continued)
or
ene
Subjects, methods"
Exposure levels
Results
Reference(s)
Residential exposures
Germany, residents near dry-
cleaning business, 14 exposed
and 23 age- and gender-
matched nonexposed controls.
AA, B; not clear if blinded to
exposure status
New York, 17 exposed
(apartment residents living
above dry-cleaning business)
and 17 age- and gender-
matched controls. AA, PA, EA,
B, U; not blinded to exposure
status
New York, 65 households (67
adults and 68 children) in
residential buildings with
colocated dry cleaners, 61
households (61 adults and 71
children) in residential
buildings without dry cleaners.
AA; not blinded to exposure
status
Mean = 7 d monitoring period,
0.7 ppm, mean duration = 10.6
yr
Mean = 0.4 ppm (monitoring
before closure of dry cleaners).
Mean duration of residence = 6
yr
Geometric mean = 5 ppb
(0.005 ppm). Mean duration of
residence = 10 yr
Statistically significant increase in
simple reaction time and decrease in
continuous performance and
visuospatial function. No fine
motor function deficits.
Decreased color discrimination
among exposed, but not statistically
significant.
Lower (worse) scores on tests of
visual contrast sensitivity.
Association (p < 0.05) between PCE
(indoor air and blood) and
performance on test of visual
contrast 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.
Altmann et
al. (1995)
Schreiber et
al. (2002)
NYSDOH
(2005a);
Storm et al.
(2011)
McDermott
et al. (2005)
A = air sample, not specified area or personal sample, AA = area air samples, B = biological monitoring of
blood, CCI = color confusion index, CI = confidence interval, EA = exhaled air samples, PA = Personal air
samples, PCE = tetrachloroethylene, RR = relative risk, U = biological monitoring of urine for
trichloroacetic acid, VCS = visual contrast sensitivity.
Vision testing in the four studies included tests of acuity, tests of spatial vision based on
contrast sensitivity, and tests of color vision. The visual acuity test measured the ability to
discriminate high-frequency (i.e., small) images at high contrast; e.g., reading successively
smaller black-on-white letters as part of an examination for corrective lenses. This measure
typically is dependent on the optics of the eye (and corrective lenses when needed) and is
insensitive to subclinical deficits in neurologic function. Contrast sensitivity measures the least
amount of luminance difference between dark and light bars needed to detect a given pattern
(e.g., a bar pattern). Impairments in color vision, beginning as blue-yellow confusion errors,
have been reported in populations exposed to organic solvents (Campagna et al., 1996;
Campagna et al.. 1995: Mergler et al.. 1991: Mergler et al.. 1988a: Mergler et al.. 1988b: Mergler
andBlain, 1987: Mergler, 1987). The tetrachloroethylene exposure studies that assessed color
4-9
-------
vision relied on various versions of the Lanthony color vision test. This type of test consists of a
series of small round "caps" that the subject is asked to arrange in order by color. The types of
errors made can distinguish specific types of color vision deficiency; e.g., red-green color
confusion errors (blindness) is a common condition in males, mostly but not entirely of
congenital origin, whereas blue-yellow color confusion errors are very rarely due to congenital
conditions and, therefore, are considered as a hallmark of an acquired condition. Test scores are
based on the subject's ability to arrange a set of 15 caps according to a definite chromatic
sequence, with each mistake increasing the score above a perfect score of 1.00. A formula (the
Color Confusion Index [CCI]) based on Total Color Distance Scores can be used for scoring
(Geller, 2001; Bowman, 1982). The Lanthony D-15 desaturated test is more sensitive to mild
and moderate changes in color vision compared with other versions of the test that use more
contrasting hues (Lanthony, 1978). The vision tests are not recommended for epidemiological
studies of children under 5 years of age.
Other types of neurobehavioral effects were assessed in these studies using standardized
tests of cognitive or motor function, such as the digit symbol, digit span, Benton visual memory,
and simple reaction time tests. The standardized neurobehavioral battery has a high rate of
reliability and has been used to assess normal neurological function (Anger et al., 2000).
As with most conditions, age is an important factor that needs to be considered in
interpreting measures of neurological function. Generally, the comparison group within these
studies was age-matched (individually or frequency-matched) to the exposed subjects. Measures
of cognitive function can also be influenced by education (or more broadly, socioeconomic status
variables), by other intelligence measures, and by alcohol use. Thus, these attributes would also
need to be considered in studies using cognitive tests such as visuospatial memory, vigilance,
and information processing. Alcohol use, smoking, certain medications, chronic neurological
conditions, and solvents other than tetrachloroethylene may affect visual contrast sensitivity and
color vision measures (Paramei et al., 2004; Swinker and Burke, 2002). In contrast, color vision
and spatial vision have not been shown to be related to education or socioeconomic status, so
potential confounding by these factors is unlikely.
4.1.1.2.1. Occupational exposure studies: dry-cleaning settings
Lauwerys et al. (1983) studied 266 workers (24 women and 2 men) occupationally
exposed to tetrachloroethylene in six dry-cleaning shops in Belgium for a mean of 6.4 years
(range: 0.1 to 25 years) and 33 controls (31 women and 2 men) working in a chocolate factory
(n = 20) or an occupational health service (n = 13) without occupational exposure to organic
6 Abstract of paper reports 22 subjects were exposed to tetrachloroethylene, but the full text of the paper includes
26 subjects.
4-10
-------
solvents. No information is provided in the paper on the methods used to identify subjects or
their reasons for participating in the study. The level of education was similar in the exposed and
control groups, but the prevalence of smokers was higher among dry-cleaning workers (50%)
compared with the controls (27%). Neurobehavioral tests of motor function (simple and choice
reaction time), sensory function (critical flicker fusion), and cognitive function (sustained
attention test) were given twice to each worker, once before work and once after work. Both
groups were tested in the middle of the workweek. Individuals also were questioned about
chronic neurological symptoms (e.g., fatigue, depression, sleep disturbances). Blood samples
were collected both before and after work. The mean tetrachloroethylene air concentration
(8-hour TWA) was 21 ppm, and the range of TWA values was 9 to 38 ppm, using results from
active sampling of personal air. The mean tetrachloroethylene blood level (30 minutes after the
end of work) was 1.2 mg/L (range of means from the shops was 0.6 to 2.4 mg/L).
Trichloroacetic acid, a metabolite of tetrachloroethylene, was not detected (level of detection
[LOD] not identified in published paper) in urine specimens from exposed subjects. An
evaluation of the subjects was performed at each worksite, so examiners were not blinded to
exposure status. The score of the critical flicker fusion test (a test of sensory function) was
significantly increased (better performance) in the exposed workers compared with controls
when given both before and after work. Decreased simple reaction time was observed among the
exposed workers in the tests performed before work (mean ± standard deviation [SD]: 0.374 ±
0.120 and 0.448 ±0.155 seconds in exposed and nonexposed workers, respectively) but not in
the tests performed after work (mean ± SD: 0.341 ±0.116 and 0.356 ± 0.128 seconds in exposed
and nonexposed workers, respectively). The dry-cleaning workers did not differ from controls
on the other three neurobehavioral tests. The prevalence of abnormal scores (those beyond the
5th or 95th percentile of the control group) did not vary significantly between the two groups.
Seeber (1989)7 evaluated the neurobehavioral effects of tetrachloroethylene in
101 German dry-cleaning workers (machine operators, ironers, touch-up workers, counter
attendants, and other employees) who were employed in coin-operated or while-you-wait shops,
all affiliated with one organization. The workers were separated into a low-exposure group
(50 women, 7 men) and a high-exposure group (39 women, 5 men) based on both activities and
room air measurements. A third group of 84 sales personnel (64 women, 20 men) from several
department stores and receptionists from large hotels served as unexposed controls. No
information was provided on the methods used to identify subjects or their reasons for
participating in the study. Predominant characteristics of both groups included primarily
7 Dr. Seeber provided additional information on this study in written correspondence to the New York State
Department of Health (NYSDOH) dated January 19 and May 20, 1996. This information appears in
NYSDOH (1997).
4-11
-------
standing work, contact with customers, and moderate physical exercise. The authors reported
that 29 service technicians were excluded from the study because of either discontinuous
exposure conditions with peak concentrations or long periods of no exposure, which focused the
investigation on workers with relatively constant exposure levels. Mean tetrachloroethylene
concentrations (8-hour TWA) for the low- and high-exposure groups were 12 (±8) ppm and 53
(±17) ppm, respectively, using results from active sampling of room air and passive sampling of
personal air. The mean durations of occupational exposure for the low- and high-exposure
groups were 11.8 and 10.6 years, respectively.
Several tests of neuropsychological functioning were administered using a German
standardized neurobehavorial battery (Psychologisch-Neurologischer Fragebogen or PNF),
including overall neurological signs, standardized personality tests (emotional lability), tests of
sensorimotor function (including finger tapping and aiming), tests of attention (digit reproduction
and digit symbol), test of visual scanning (cancellations) and tests of information processing
speed (threshold of perceptual speed and choice reaction time) (Seeber, 1989). Some details of
the testing procedures were not provided, and one of the response variables, "delayed reactions,"
was not defined. The typical dependent variable measured in this task—response reaction
time—apparently was not measured; only the number of correct reactions was reported. In
addition, subtests of the Wechsler Intelligence Test (digit span, digit symbol, and cancellations)
were used, as was recognition of words, faces, and digits. Intelligence was assessed using the
logical thinking subtest of the German Performance Test System. The neurobehavioral tests
were given by two specialized clinic staff members who did not question the subjects regarding
exposure status. Test outcomes were reported as means ± SDs.
The control group was younger than the dry-cleaning workers (mean ages: 38.2, 38.4,
and 31.8 years, respectively, in the low-exposure, high-exposure, and control groups,
respectively), and alcohol consumption also differed by group (mean: 8.2, 10.4, and 12.6 g/day
in the low-exposure, high-exposure, and control groups, respectively) (Seeber, 1989). Higher
scores on the intelligence test were observed among the control group (mean + SD: 21.9 + 5.8)
compared with the dry-cleaning workers (mean + SD: 18.3 + 5.0 and 19.2 + 5.2 in the low- and
high-exposure groups, respectively). Age, gender, and intelligence scores were included in the
regression models analyzing the relation between exposure and neurobehavioral test scores;
additional control for group differences in alcohol consumption did not alter the observed results.
Dose-response relationships for several outcomes reported by Seeber (1989) were
suggested by statistical analysis. Performance of both the low-exposure and high-exposure
groups differed significantly from that of the unexposed control group on the threshold of
perceptual speed test, digit reproduction, digit symbol, and cancellations; all pairwise
comparisons had/?-values < 0.01. "Delayed reactions" differed statistically significantly from
4-12
-------
controls in the high-exposure group (p = 0.03), but not in the low-exposure group (p = 0.08). For
other outcomes reported, personality test (emotional lability) and overall performance
(neurological signs), dose-response relationships were less clear. Performance of the low-
exposure group, but not the high-exposure group, differed statistically significantly from that of
the unexposed control group on the personality test (emotional lability) and for neurological
signs (p < 0.05 andp > 0.10 for low-exposure and high-exposure groups, respectively, on both
tests), although there was no statistically significant difference between the exposed groups for
both tests (p > 0.10). Last, the mean scores on "correct reactions" on a choice reaction time task
were not statistically significantly different from control for either exposure group (p > 0.10 for
both).
Characterization of a dose-response in this study is complicated by nonmonotonic
response patterns in the mean scores for several outcomes. Specifically, mean scores of the low-
exposure group for the outcomes of digit symbol and cancellations, as well as for emotional
lability and for overall neurological signs, were worse compared with control than the high-
exposure group scores. However, the wide confidence intervals around the individual means for
each of these outcomes in the exposed groups indicates no statistical differences in outcomes
between the two exposure groups. Further, there is substantial overlap between the exposure
ranges of the two sets of dry cleaning workers, reflecting assignment to groups based partly on
job activities, which also supports the inference that the treated groups are similar. Thus,
exposure misclassification is a possible contributor to the observed pattern. Without individual
test outcomes and exposure measurements, however, it is not possible to evaluate the dose-
response relationships more thoroughly. Seeber (1989) concluded that there was a significant
difference in outcomes between the control and exposed subjects. For some outcomes, NRC
(2010) characterized the study as having discrepant results based on worse mean test scores (for
neurologic signs, emotional lability, choice reaction time, cancellation d2 and digit symbol) in
the low- compared with high-exposure group.
Nakatsuka et al. (1992) evaluated the effects of tetrachloroethylene exposure on visual
function by examining the color vision of 64 dry-cleaning workers (34 women and 30 men) in
China. Control workers (72 women and 48 men) were recruited from the clerical sections of
dry-cleaning shops and from other factories (paint production plants or plants producing
tetrachloroethylene from trichloroethylene). No information is provided in the paper on the
methods used to identify subjects or their reasons for participating in the study. The mean ages
of the dry-cleaning workers (34.2 years for men, 35.3 years for women) were similar to those of
the male controls (34.0 years) but slightly higher than the female controls (32.6 years). The
Lanthony new color test, used for screening color vision, and the Ishihara's color vision test,
used for confirmation of red-green vision loss, were carried out by ophthalmologists or
4-13
-------
occupational health doctors in charge of the factories under one of two lighting conditions
(natural sunlight or a daylight fluorescent light). (This color vision test is not as sensitive as the
Lanthony D-15 test used in the other studies discussed in this section.) The geometric mean air
concentrations of tetrachloroethylene (averaging time not reported) were 15.3 and 10.7 ppm for
the men and women, respectively, using results from passive sampling of personal air. The
overall geometric mean was 13 ppm. The authors reported no significant difference in the
performance of the dry-cleaning workers (or other solvent-exposed groups included in the study)
and unexposed controls on the Lanthony new color vision test, with 60% of the male dry-
cleaning workers and 63% of male controls classified as "normal" color vision. Corresponding
figures for females were 91 and 74% in the dry-cleaning workers and controls, respectively.
Results for the males were not appreciably different when individuals with red-green vision loss
were excluded.8 Nakatsuka et al. (1992) concluded, overall, that they found no distinct color
vision loss among the dry-cleaning workers.
Ferroni et al. (1992)9 evaluated neurobehavioral effects and prolactin levels among
60 female dry cleaners and 30 unexposed female controls. Prolactin secretion by the pituitary is
controlled by hypothalamic dopamine; dopamine is also important to neurotransmitter systems,
and serum prolactin, as a biochemical signal and marker of nervous system function, is a
proposed alternative for assessment of nervous system toxicity (Manzo et al., 1996). The
workers at every dry-cleaning shop in a small town outside of Parma, Italy, were invited to
participate in the study. There were no refusals. Controls were selected from the workers at a
hospital who cleaned clothes using a water-based process. Their jobs were essentially the same
as those of the dry cleaners, but they were not exposed to any organic solvents. Both groups
filled out a questionnaire on their health status, medication (including oral contraceptives),
lifestyle, and current and past jobs. Both groups met the following criteria: no history of
metabolic disorders, no history of psychiatric disorders, and low level of daily alcohol intake.
The dry cleaners and controls were comparable in age (mean ages: 39.7 and 37.6 years,
respectively), vocabulary level, height, weight, body mass index, smoking habits, and use of
medication. Workplace air samples were randomly collected throughout the workweek during
summer and winter to account for variability related to either the work cycle or seasonal
environmental fluctuations. Blood samples were collected during the workday during summer
A statistical analysis of the dry cleaners data using a Fisher's exact test (for differences in proportions with at least
one sparse cell) indicated that tetrachloroethylene-exposed women were more likely to have normal color vision as
compared with unexposed women (p = 0.0423), but no difference was observed among the males (0.83, based on
Chi-squared test); reported in public comments of the Halogenated Solvents Industry Alliance to EPA (Halogenated
Solvents Industry Alliance. 2004) on the Neurotoxicity of Tetrachloroethylene Discussion Paper (U.S. EPA. 2003).
9 Dr. Mutti provided details on the selection process of exposed and control subjects and also clarified reported
results to Dr. Ken Bodgen, NYSDOH, in written correspondence dated July 29 and September 5, 1995 [refer to
NYSDOH( 1997)1.
4-14
-------
and winter. The median tetrachloroethylene air concentration (4-hour TWA) was 15 ppm (range:
1 to 67 ppm). The subjects' range of tetrachloroethylene blood levels was 0.012-0.864 mg/L
[median = 0.145 mg/L; incorrectly expressed in Ferroni et al. (1992) as 12,864 and 145 mg/L;
NYSDOH (1997)1. The mean duration of occupational exposure was 10 years.
Workers and controls were given five neurobehavioral tests (part of the Swedish
Performance Evaluation System, "adapted" Italian version: finger tapping with both dominant
hand and nondominant hand, simple reaction time, digit symbol test, shape comparison-
vigilance, and shape comparison-response to stress) (Ferroni et al., 1992). All subjects were
examined in the morning before their work shift in the same room by the same examiners, using
a standardized testing protocol (NYSDOH, 1997). Although the examiners were not blind to the
status of the subjects (dry cleaner or control), they were blind to the worker's exposure level
(NYSDOH, 1997). Serum prolactin levels were measured in all subjects using a blood sample
taken at the time of the neurobehavioral testing; analysis was limited to those samples obtained
during the proliferative (follicular) phase of the menstrual cycle (41 dry cleaners and
23 controls). Ferroni et al. (1992) did not describe the protocol for determining menstrual cycle
phase, however. Serum samples from dry cleaners and controls were alternated and analyzed in
the same experimental runs (NYSDOH, 1997).
The dry cleaners showed significantly reduced performance when compared with the
unexposed matched controls in three tests (simple reaction time,/? < 0.0001; vigilance,
p < 0.005; and stress, p< 0.005) (Ferroni et al., 1992). Performance on the finger-tapping test
(both hands) and digit symbol test was not affected (NYSDOH, 1997). Additionally, the mean
serum level of prolactin was significantly higher in the workers than in the matched controls
(mean: 12.1 compared with 7.4 \JigfL, p < 0.001). Among the dry cleaners, none of the three
measures of exposure (duration of exposure and air or blood concentration of
tetrachloroethylene) was significantly associated with decreased test scores or increased serum
prolactin levels. Ferroni et al. (1992) concluded that tetrachloroethylene exposure in dry-
cleaning shops may impair performance.
Cavalleri et al. (1994) evaluated the effects of tetrachloroethylene exposure on the color
vision of dry cleaners and a comparison group of matched controls. The investigators compiled
a list of all the dry-cleaning shops in the municipality of Modena, Italy (110 shops employing
189 workers) and randomly selected 60 dry cleaners from 28 premises for recruitment into the
study (Aggazzotti et al., 1994a). Only full-time workers (n = 52) were asked to participate, and
two declined. All 50 workers provided, via questionnaires, information on work history, health
status, occupational and hobby use of solvents, drinking and smoking habits, and drug use.
Thirty-five of the 50 dry cleaners (33 women, 2 men) met the inclusion criteria; others were
excluded for hypertension, smoking more than 30 cigarettes a day, alcohol consumption
4-15
-------
exceeding 50 g of alcohol a day, oculo-visual pathology, or employed at a dry-cleaning facility
for less than 1 year. Another worker was excluded because a matched control could not be
found. The controls were factory workers who were not occupationally exposed to solvents or
other neurotoxic chemicals; they were selected and recruited into the study using the same
methods that were used for dry cleaners. The controls (n = 35) were from factories in the
Modena area and met the same inclusion criteria as the dry cleaners. They were matched to dry
cleaners by gender, age (±3 years), alcohol consumption (±10 g/day), and cigarette use
(±5 cigarettes a day). The mean age of both groups (35 years) and the percentages of each group
that were smokers (43%) or alcohol drinkers (71%) were comparable. All subjects appeared
healthy and met minimal status of visual acuity. None of the subjects reported hobby exposure
to solvents or other substances toxic to the eye. There were no known systematic differences
between exposed and control groups or between machine operators and ironers. Color vision
was assessed using the Lanthony D-15 desaturated panel test. Exposed and control subjects were
tested in random order (NYSDOH, 1997). All subjects were tested at the same time of day (in
the morning, before work) under the same lighting conditions by the same investigator. With
respect to exposed subjects, the investigator was unaware of both the exposure levels and the job
(operator or ironer) of each dry cleaner.
For all dry cleaners, the mean tetrachloroethylene air concentration (8-hour TWA) was
6 ppm, and the range of TWA values was 0.4-31 ppm, using results from passive sampling of
personal air (Cavalleri etal., 1994). For operators (n = 22), the mean air concentration 8-hour
TWA was 7.3 ppm (range 0.4-31 ppm). For ironers (n = 13), mean air concentration (8-hour
TWA) was 4.8 ppm (range 0.5-11 ppm). The mean duration of occupational exposure was
8.8 years. Tetrachloroethylene concentrations were also measured in alveolar air for a subset of
these dry cleaners, with a high correlation observed between tetrachloroethylene concentration in
alveolar air and 8-hour TWA levels in ambient air [r = 0.8,/> < 0.001; Aggazzotti et al. (1994a)].
Only three dry-cleaning workers, as opposed to 13 controls, scored a perfect test score on
the color vision test (p < 0.01). Mistakes were made mainly in the blue-yellow range. Overall,
the workers showed poorer performance on the test as compared to controls, and they had a
significantly higher error rate (mean CCI score: 1.143 and 1.108 in workers and controls,
respectively,/* = 0.03). The effect was observed in dry cleaners (mean CCI score: 1.192 and
1.089 in dry cleaners and their matched controls, respectively, p = 0.007) but not among the
ironers (mean CCI score: 1.061 and 1.073 in ironers and their matched controls, respectively).
There also was a statistically significant positive correlation (p < 0.01) between TWA air
concentrations and the CCI (r = 0.52), which remained after multivariate analysis considered
previous tetrachloroethylene exposure, duration, age, number of cigarettes a day, and daily intake
of alcohol as covariates. The CCI values were not associated with two other measures of
4-16
-------
tetrachloroethylene exposure (mean duration and an integrated index of exposure, yearly TWA
level). The study authors suggested that this may reflect the difficulty in controlling for the
interactive effects of age and exposure and accurately evaluating exposure. The effect on color
vision may not be rapidly reversible; preliminary data showed that the scores of some workers
did not improve when retested after 4 weeks of vacation (NYSDOH, 1997). Moreover, some of
these workers showed poorer performance on this test in the follow-up study by Gobba et al.
(1998), described below, suggesting color vision impairment is a chronic effect.
Gobba et al. (1998) reexamined color vision after a period of 2 years in 33 of the 35 dry
cleaners and ironers examined by Cavalleri et al. (1994). Two subjects had retired during the
2-year period between examinations. These investigators used the Lanthony D-15 test, the test
used by Cavalleri et al. (1994) to assess color vision, and performance was compared with the
subject's score from the initial survey. Tetrachloroethylene concentration in the occupational
setting was determined in the breathing zone using personal passive samplers. Monitoring was
carried out during the afternoon shift, as Cavalleri et al. (1994) did not show any differences
between morning and afternoon samples. Gobba et al. (1998) found that tetrachloroethylene
concentration had increased during the 2-year period for 19 subjects, identified as Group A
(geometric mean, from 1.67 ppm at the first survey to 4.35 ppm at the second survey), and had
decreased for 14 subjects, identified as Group B (geometric mean, from 2.95 ppm to 0.66 ppm).
The decrease in exposures was due to new equipment or other changes to the working
conditions. As found in the first survey, color vision was impaired primarily in the blue-yellow
range of color, with few subjects presenting red-green errors. Color vision performance for the
entire group was related significantly to age (r = 0.45) and tetrachloroethylene concentration
(r = 0.39; p < 0.05). The mean CCI score for Group A subjects showed a statistically significant
difference between the two surveys (arithmetic mean: 1.16 and 1.26 in the first and second
surveys, respectively,p< 0.01). For Group B subjects, who experienced lower exposure
concentrations by the second survey, the CCI score did not change from that of the initial survey
(arithmetic mean: 1.15 and 1.15 in the first and second surveys, respectively). The findings in
Groups A and B were also supported using analysis of variance methods to examine the relation
between CCI score and exposure level (log TWA), adjusting for age, alcohol consumption, or
cigarette smoking between the subgroups.
Echeverria et al. (1995) assessed the performance of 65 dry-cleaning workers on
neurobehavioral tests. The testing was conducted in 1986. The owners of 125 shops in Detroit,
MI, were contacted, and 23 agreed to allow their workers to participate in the study. Within each
shop, operators were matched on education and age (±5 years) with a lower-exposure subject.
The subjects (35 men and 30 women) were grouped into three categories of chronic
tetrachloroethylene exposure (low, moderate, and high), based on type of shop (wet-transfer or
4-17
-------
dry-to-dry), job title (counter clerk, presser, or operator), and years of employment. All the
operators were placed in the high-exposure category. There was no unexposed control group.
Dry-cleaning workers placed in the chronic exposure categories of low, moderate, and high had
been employed at their main job for 2.1, 3.9, and 14.6 years, respectively. Their mean ages were
40.9, 40.6, and 43 years. The three groups were also characterized by estimates of current
exposure (low, medium, and high), which corresponded to mean tetrachloroethylene air
concentrations (8-hour TWA) of 11, 23, and 41 ppm, respectively, for counter clerks, pressers,
and operators in the more common wet-transfer shops (17 of 23 shops). Estimated air
concentrations for counter clerks, pressers, and operators in the dry-to-dry shops were 0.5, 10,
and 11 ppm, respectively. The estimates were based on a relationship between breath and air
concentrations derived from a larger independent study (Solet et al., 1990). These estimates
were comparable to those found in other surveys of dry-cleaning facilities in the United States.
All subjects were tested in a minivan at the worksite in groups of two, in the afternoon
after work on the first or second day of their workweek (Echeverria et al., 1995). Each subject
provided a breath sample and completed a medical, symptom, work history, and hobby
questionnaire. The subjects were administered six neurobehavioral tests, a test of verbal skills,
and questionnaires on emotional states (moods) and CNS symptoms. The neurobehavioral test
battery consisted of one test of motor/cognitive function (symbol digit) and five tests of cognitive
function (digit span, trailmaking A and B, visual reproduction, pattern memory, and pattern
recognition). Multivariate analysis was used to evaluate the relationship between a chronic index
of lifetime exposure and performance on neurobehavioral tests, accounting for the potential
confounding variables of current exposure, age, education, verbal skill, alcohol consumption,
hours of sleep, fatigue, mood, symptoms, medication, and secondary exposures to
neurotoxicants. After adjustment for factors affecting performance, the scores of the dry-
cleaning workers with high chronic exposure were reduced (compared with the low chronic
exposure group) by 4% for pattern recognition, 7% for pattern memory, and 14% for visual
reproduction (all ^-values <0.01). These impairments of visually mediated function were
consistent with the impairment of visuospatial functions observed in four patients who were
diagnosed with tetrachloroethylene encephalopathy who had been previously studied by
Echeverria et al. (1995). Other effects observed in the patients (mood changes and decreased
cognitive function in nonvisual tests) were not found in the dry-cleaning workers with high
lifetime exposures. Among complaints by the dry-cleaning workers, only the number of
complaints of dizziness from standing up rapidly and "solvent-induced dizziness" over the
previous 3 months was significantly elevated (p < 0.04) in the high-exposure group. Echeverria
et al. (1995) concluded that effects on visuospatial function were consistently found in subjects
employed as operators for an average of 14.6 years and exposed to an estimated
4-18
-------
tetrachloroethylene 8-hour TWA air concentration of 41 ppm, suggesting a vulnerability of
visually mediated functions with tetrachloroethylene exposure. This conclusion was based on
the impaired performance of the high-exposure group when compared with a group of dry-
cleaning workers with low lifetime exposure, including workers who were probably clerks in
wet-transfer shops where the mean current exposure level was 11 ppm. This exposure level is
substantially above background ambient levels, and whether the performance of the low-
exposure group was impaired when compared with that of a group without occupational
exposure (i.e., an unexposed control group) is not known.
Echeverria et al. (1994) builds on the results of Echeverria et al. (1995),10 hypothesizing
degradation in behavior (particularly attention, executive function, visuospatial memory, short-
term memory, and mood) is an early indicator of neurotoxicity, leaving motor, language-based
skills, and long-term memory intact. The study was conducted in the Seattle/Tacoma, WA area
from 1989 through 1993, when the area's dry-cleaning industry was switching from wet-transfer
to dry-to-dry machines. Initially, 320 dry-cleaning shops and laundries were sent introductory
letters requesting permission to allow their employees to participate in the study. Of the 181
owners who responded, 39 agreed to participate. The most common reasons for nonparticipation
were disinterest, time constraints, lack of English proficiency, and concerns about pending
regulatory actions concerning tetrachloroethylene. Recruitment ended when a total of 45
operators were enrolled. Each operator was matched with a less-exposed person from the same
shop. The subjects included laundry workers (n = 69), pressers or counter clerks (n = 59), and
operators or former operators (n = 45). The mean ages of the groups were 42.5, 34.2, and 46.2
years, respectively. Women comprised 63% of the study population (109/173). The subjects,
who were paid volunteers, were eligible if they spoke English, had no history of diabetes or CNS
disorders, and had worked for more than 1 year in the trade. The final sample excluded three
subjects because of limited English and reading skills and six subjects who did not wear their
prescription glasses on the day of testing or who were missing covariate information such as
vocabulary test scores.
An index of chronic exposure and measures of subchronic and acute exposure were
developed for each subject. The chronic exposure index was based on a detailed work history,
including consideration of the type of dry-cleaning machine, job title, percentage of time at each
job title, estimated air levels associated with each job title, and employment duration. The
measures of subchronic and acute current exposure were based on mean 8-hour TWA air
concentrations measured on the day of neurobehavioral testing. Mean chronic indices were zero
for the never-exposed group of laundry workers, 68 for the dry-cleaning workers with low
10 Although published a year after this study (Echeverria et al.. 1994X the study by Echeverria et al. (1995),
discussed previously, was conducted in 1986, 3 years before this study.
4-19
-------
exposure (pressers/clerks), and 1,150 for the dry-cleaning workers with high exposure
(operators). Mean exposures (8-hour TWA, using results from passive sampling of personal air)
for workers placed in these chronic exposure categories were <0.2 ppm (laundry workers), 3 ppm
(pressers/clerks), and 9 ppm (operators). Dry-cleaning workers placed in the chronic exposure
categories of low and high had been employed in their current job for 2.6 and 11 years,
respectively. The subjects also were placed in acute and subchronic exposure categories of
<1 ppm (laundry workers and some dry-cleaning workers, e.g., clerks), low (mainly pressers),
and high (operators), with corresponding current tetrachloroethylene 8-hour mean concentrations
of 0.5, 3, and 20 ppm, respectively. Dry-cleaning workers placed in the acute and subchronic
low exposure and high exposure categories had been employed in their current job for 5 and 9
years, respectively. Because of the changes in dry-cleaning practices over the course of the
study, many subjects in the high chronic-exposure category could be found in the low acute- and
low subchronic-exposure categories because these latter two indices were based on air
concentrations on the day of testing.
The test battery included tests of cognitive function, including visuospatial memory,
motor skills, mood, CNS symptoms, and basic verbal and arithmetic skills. The chronic and
subchronic assessment was based on tests given during the morning of each subject's day off and
on preshift scores. Each subject signed a consent form, provided a breath sample at each test
session, and completed a questionnaire covering transient factors that could affect performance
(e.g., headache). This was followed by questionnaires on medical history, medication, drug and
alcohol use, occupational and nonoccupational exposure to chemicals, symptoms, and mood.
Multivariate analysis was used to evaluate the relationship between exposure indices and
levels and performance on neurobehavioral tests after adjusting for the potential confounders of
age, gender, race, vocabulary level (as a surrogate for education and test-taking), and alcohol
consumption. Indications of associations between increased indices of chronic (lifetime)
exposure and reduced test performance were found in three tests of cognitive function: switching
(p = 0.1), pattern memory (p = 0.03), and pattern recognition (p = 0.09). The magnitude of
change attributable to tetrachloroethylene was a 3% loss in function for the latency of pattern
memory and an 11% loss in function for the correct number in visual reproductions. Subjective
measures of mood and symptoms were not significantly associated with exposure. Dry-cleaning
workers scored lower (but not significantly) on all but one of the remaining tests (the digit span
test). Analysis of the association between test scores and measures of subchronic exposure
(8-hour TWA tetrachloroethylene concentrations on the day of testing) confirmed the findings of
the chronic analysis: reduced scores on tests of switching (p = 0.1) and pattern recognition
(p = 0.04) as exposure increased. In summary, Echeverria et al. (1994) detected deficits in
visuospatial function (reduced performance in tests of pattern memory and pattern recognition)
4-20
-------
in the dry-cleaning workers categorized as having high lifetime chronic exposure and whose
current exposure level was 9 ppm, 8-hour TWA. However, the exposure level of 9 ppm is not
representative of past chronic exposure levels because of changes occurring in the industry in the
study area (i.e., switching from wet-transfer to dry-to-dry machine). The investigators attributed
the reduced performance to exposures 3 to 5 years previously that were about two to four times
higher, and they hypothesized that a few years of reduced exposure may not be long enough to
eliminate the residual effects on visuospatial function caused by the exposures associated with
wet-transfer machines.
Spinatonda et al. (1997) assessed the effect of tetrachloroethylene exposure on cognitive
function by measuring vocal reaction times among 35 dry cleaners and 39 unexposed controls.
Controls were matched to exposed individuals by age (mean age of 35 years for both groups) and
education. The published paper did not identify the population from which exposed subjects and
controls were drawn or the inclusion criteria for exposed subjects and controls. Exposure was
assessed by a "grab sample" collected at the time of the neurological testing and is not a TWA.
Exposure monitoring indicated a median concentration of tetrachloroethylene of 8 ppm (range:
2-136 ppm). An index of cumulative exposure to tetrachloroethylene was also developed for
each exposed subject by multiplying the tetrachloroethylene concentration by the number of
years worked. Latency to and duration of vocal response to the stimulus (reading) were
measured in each subject after the presentation of a sequence of words on a computer screen.
For each condition, subjects were asked to say each word immediately or following delays of 0.1
or 0.5 seconds. The test was performed using a random sequence of concrete or meaningless
disyllabic words. These tests were carried out at the place of employment for dry cleaners and in
a clinical setting for controls, indicating that the investigators were not blinded as to a subject's
exposure status. Compared with the control group, the exposed group had statistically
significant longer mean reaction times and/or vocalization durations under all response
conditions (immediate or delayed response) with either real or meaningless words. Furthermore,
statistically significant positive correlations were observed between cumulative
tetrachloroethylene exposure and immediate reading and delayed reading tasks (r = 0.69 and r =
0.73, respectively). No information on alcohol consumption or other potential differences
between exposed subjects and controls was reported, precluding an analysis of how these factors
may have affected the observed association between tetrachloroethylene and reaction time.
Sharanjeet-Kaur et al. (2004) examined visual function effects by assessing color vision
in 14 workers, ages 24-53 years, in three dry-cleaning facilities using tetrachloroethylene in
Malaysia. This study was part of a larger study assessing color vision in two other
occupationally exposed populations (39 workers in a factory producing polyethylene resin plastic
storage containers and 40 workers manufacturing polystyrene plastic bags). The paper does not
4-21
-------
report how facilities were identified or recruitment methods for study subjects. Furthermore, the
paper does not present any information on tetrachloroethylene concentrations,
tetrachloroethylene biomarkers, or exposure levels in this type of work setting in Malaysia,
making it difficult to judge the degree of exposure. Controls (n = 29)11 were recruited from the
support staff of the Universiti Kebangsaan Malaysia and were age-matched to dry-cleaning
workers (mean age: 33 ± 8.5 years and 33 ± 3.9 years in dry cleaners and controls, respectively).
However, dry-cleaning workers differed from controls on several variables: work duration
(mean: 6.7 and 12.6 years in dry cleaners and controls, respectively), hours worked per day
(mean: 9.8 and 8.3 in dry cleaners and controls, respectively), cigarette smoking (36 and 7% in
dry cleaners and controls, respectively), and race (50 and 90% Malays in dry cleaners and
controls, respectively); no information is presented on possible differences between dry cleaners
and controls in socioeconomic status. Consent was obtained from all study participants. Visual
testing was carried out at the factory or dry cleaner, for exposed subjects, and at the Optometry
Clinic in the Universiti Kebangsaan Malaysia for control subjects. Thus, the investigators were
not blinded to exposure status during the testing procedure. Distance visual acuity was measured
using the Snellen chart, and near visual acuity was measured using a reading chart. Subjects
with poor visual acuity or with systemic, ocular, or neurological diseases were excluded; the
number of excluded subjects is not specified in the paper. Color vision was assessed binocularly
using Ishihara plates, the Lanthony D-15 test, and the Farnsworth Munsell (FM) 100 Hue test
under a light box at an illumination of 1,000 lux. None of the controls or dry cleaners had color
vision errors with the Ishihara plates. In contrast, errors on the Lanthony D-15 test and FM 100
Hue test were reported for 6 dry cleaners (43%) and 13 dry cleaners (93%) compared with no
errors reported among controls, respectively. Statistical testing of these differences was not
presented. Total error scores for the FM 100 Hue test differed between dry cleaners and controls
(p < 0.05). It is difficult to interpret these findings due to the lack of information on potential
tetrachloroethylene exposure other than job title, and differences between dry cleaners and
controls regarding test conditions and smoking history.
Perrin et al. (2007) evaluated the risk of schizophrenia among a cohort of 88,829 subjects
born between 1964-1976 in the Jerusalem Perinatal Project, a population-based cohort. Births in
this cohort are linked to the database of Israel's Psychiatric Registry, with cases identified using a
broad definition of schizophrenia-related disorders as recorded as hospital discharge codes.
Diagnoses for individuals with psychosis were validated, and the date of onset was identified as
the date of first psychiatric admission. Of the 88,829 births, 136 offspring were born to parents
identified as having a job title of dry cleaner on the birth certificate, 120 offspring whose fathers
11 An additional control group, Control Group 1, was included in the paper; this group was age matched to the other
factory workers included in the study.
4-22
-------
but not mothers were dry cleaners, 20 whose mothers but not fathers were dry cleaners, and 4
with both parents as dry cleaners; 4 of the 136 births had a later diagnosis of schizophrenia. The
relative risk (crude) between schizophrenia and parental employment in dry cleaning was 3.9
(95% confidence interval [CI]: 1.3-9.2) using proportional hazard methods. The investigators
noted risk estimates did not greatly change when fitting proportional hazard models that adjusted
for a number of potentially confounding variables; although adjusted relative risk (RR) estimates
are not reported in the paper. Variables considered as possible confounders were parents' age,
father's social class, duration of marriage, rural residence, religion, ethnic origin, parental
immigration status, and offspring's birth order, sex, birth weight, and month of birth. Family
history of mental illness was not included as a covariate; rates of schizophrenia are higher among
relatives of patients than in the general population (Mueser and McGurk, 2004).
4.1.1.2.2. Occupational exposure studies: other settings
Schreiber et al. (2002) reported the findings from investigations using visual tests to
assess neurologic function in two populations: apartment residents12 and day-care employees
who had potential environmental tetrachloroethylene exposure due to close proximity to dry-
cleaning facilities. The study of day-care employees will be discussed in this section because
their exposure would have been of a similar pattern to others in an occupational setting. The
day-care facility, located near Albany, NY, was in a building that also housed a business that
performed dry cleaning. Atmospheric monitoring of the day-care facility before closure of the
dry-cleaning business showed airborne concentrations of tetrachloroethylene ranging from 0.27
to 0.35 ppm, with median and mean concentrations of 0.32 ppm. Samples obtained at the time of
visual testing, 5 weeks after removal of the dry-cleaning machines, approached background
concentrations (range: 0.0012-0.0081 ppm).
Objectives of the investigations were to characterize tetrachloroethylene exposure and to
screen for subclinical neurological effects using a battery of visual function tests (Schreiber et
al., 2002). All participants signed consent forms. The study included all of the current staff
members of the day-care center (n = 9, all adult females). Controls were age- and gender-
matched acquaintances of the exposed participants, local retail shop employees, NYSDOH
employees, or staff from other local day-care centers with no known tetrachloroethylene
exposure. All subjects in the exposed and control groups were Caucasian (telephone
communication from K. Hudnell, EPA, to D. Rice, EPA, February 2003). Mean age was
27.7 years for control participants and 27.2 years for day-care workers; mean duration of
employment at the day-care center was 4 years. Sociodemographic data, lifestyle factors (e.g.,
personal and passive smoking, alcohol consumption, and exercise), medical history, and
12 The results of the residential study are summarized in Residential Exposure Studies, following this section.
4-23
-------
neurotoxicant exposure were obtained by questionnaire. Reported alcohol consumption was
similar (low or moderate) in the exposed and control groups.
Visual function testing consisted of near visual acuity, near visual contrast sensitivity,
and color vision (Schreiber et al., 2002). Examiners were not blinded as to a subject's exposure
status. In the contrast sensitivity test, luminance varied between the bars in sine-wave fashion,
and each test pattern represented one size of bars or spatial frequency. The bar patterns were
presented at five different spatial frequencies, thereby breaking spatial visual function into its
essential components. The least amount of luminance contrast needed to detect each bar size
was measured. A strength of this study is that the test of contrast sensitivity employed a forced-
choice procedure, providing better reliability and consistency than other approaches.
Multivariate analysis of variance was used to analyze the visual contrast sensitivity data. Color
vision was assessed using the Lanthony D-15 test, with calculation of color confusion index
(CCI) based on the accuracy of the chip placement. Group differences in the CCI were assessed
using two-tailed Student's Wests for matched-pair analyses.
The mean measure of visual acuity was 20:22.2 in the exposed day-care workers and
20:26.4 in controls (p = 0.16). There was a statistically significant lower group mean visual
contrast sensitivity score across all spatial frequencies when day-care employees were compared
with the control group (refer to Figure 4-1). The mean CCI scores were 1.22 and 1.18 in the
exposed day-care workers and controls, respectively (p = 0.39).
Although it should be noted that the controls came from a different area (a rural area in
upstate New York) compared to the exposed subjects from New York City, there is little
evidence that degree of urbanity would be related to visual contrast sensitivity. Education has
not been found to be related to performance on the visual contrast sensitivity test (NYSDOH,
2005b: U.S. EPA. 2004: Hudnell et al.. 2001: Frenette et al.. 1991: Mergler et al.. 1991).
Additionally, occupation is highly correlated with socioeconomic status (Deonandan et al., 2000)
and is also not likely to confound the visual contrast sensitivity test.
4-24
-------
_ 160
5 140
64 120
+i 100
>• an
I ffl
B
u 40
•
1
1 20
u
"3
u= 10
1.5 3 6 12 1*
Spatial frequency (eye la s/d eg rge)
Figure 4-1. Visual contrast sensitivity functions for control and exposed
participants in a study of workers in a day-care center located in a building
with a dry-cleaning facility (Schreiber et al., 2002).
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). Blue circles (top line) = controls; red circles (bottom line) =
exposed. For any frequency, a higher contrast sensitivity threshold represents better visual
function. Visual contrast sensitivity was significantly lower across all spatial frequencies in
exposed workers at a day-care center colocated with a dry-cleaning facility compared with their
matched controls. Used with permission of the authors.
The Pumpkin Patch Day Care Center Follow-up Evaluation (NYSDOH. 2010. 2005a. b)
examines the effect of tetrachloroethylene exposure on visual function in former students of the
day-care center collocated in a building with a dry-cleaning facility that was studied by Schreiber
et al. (2002). This study is discussed in this section because the children's exposure would have
been of a similar pattern to others in an occupational setting, although exposure ceased 4 years
prior to this study. Children eligible for testing in the current evaluation were enrolled in the
New York State Volatile Organic Chemical (VOC) Registry and had attended the day-care
center. Of the 115 who met this criterion, 27 children with the highest number of hours spent at
the day-care center were invited to participate; 17 children completed vision testing, and
13 children completed some or all of the neurobehavioral assessment. Referents (controls) were
children who attended other day-care centers and were matched to the exposed children by day-
care experience, age, and gender. No information is provided on methods employed for referent
participation. Overall, 17 Pumpkin Patch Day Care Center and 13 comparison children
(13 matched pairs) completed vision testing, and 13 Pumpkin Patch Day Care Center and
13 comparison children (8 matched pairs) completed neurobehavioral testing, consisting of a
battery of tests that assess general intellectual function, attention/information processing speed,
visuospatial ability, reasoning and logical analysis, memory, motor functions, and sensory-
perceptual functions. A parent or guardian completed the Child Behavioral Checklist and a
4-25
-------
background history questionnaire. Neurobehavioral function of the 13 Pumpkin Patch Day Care
Center children evaluated in this follow-up study did not differ from that of the 13 referent
children, and Pumpkin Patch Day Care Center children performed better than referent children
on several tests. Visual function testing consisted of visual acuity, far visual contrast sensitivity,
and color vision. Visual contrast sensitivity was determined using the Functional Acuity
Contrast Test distance chart placed 10 feet from the participant under light conditions specified
by the manufacturer. Scores for each eye were recorded on a graph showing a normal range
(90% CI) of visual contrast sensitivity at each spatial frequency. Color vision was assessed using
both the Farnsworth D15 and Lanthony D-15 tests. Both color vision and contrast sensitivity
tests were performed monocularly. Examiners were not specifically blinded to exposure status,
but this information could have been revealed by the participant during the examination. Using
the Wilcoxon matched-pairs signed-ranks test, Pumpkin Patch Day Care Center children
performed better on the visual contrast sensitivity test compared to referent children. No
significant difference in the proportions of children with abnormal color vision or with children
making major errors, or with CCI scores were observed between Pumpkin Patch Day Care
Center and referent children. Similar results on the vision tests were observed when excluding
two pairs who were <6 years old.
4.1.1.2.3. Residential exposure studies
This section discusses studies of residential exposure scenarios. Residential exposure to
tetrachloroethylene can result in nearly continuous exposure (NYSDOH, 2005b) and is distinct
from the pattern of occupational tetrachloroethylene exposure.
Altmann et al. (1995) examined neurological effects of long-term exposure to
tetrachloroethylene among residents of Mulheim, Germany, who lived near dry-cleaning shops.
A total of 19 exposed subjects were chosen from a population of 92 individuals living in
neighborhoods close to dry-cleaning facilities. Three criteria were used to select subjects: (1) a
tetrachloroethylene blood level above 0.002 mg/L, (2) a period of living above or next to a dry-
cleaning facility for at least 1 year, and (3) no occupational exposure to organic solvents. The
mean age of the exposed subjects was 39.2 years (range: 27-58 years), and the mean duration of
living near a dry-cleaning facility was 10.6 years (range: 1-30 years). Thirty potential controls
(mean age: 37.2 years, range: 24-63 years) were recruited, mainly from the staff of a public
health office or an institute for environmental hygiene. One or two controls, matched for age (±1
year, but ±3 years in one case and ±6 years in another case) and gender, were chosen for each
exposed subject. Consent was obtained from all subjects prior to the initiation of testing. Five
exposed (26%) and seven control subjects (23%) were excluded for various medical reasons,
including impaired vision, diseases with potential neuropathy, hypertension, and joint
4-26
-------
impairment. All subjects met standards for visual acuity and vibration perception. The final
exposed group included 14 subjects (5 men, 9 women), and the control group included 23
subjects (9 men, 14 women). The two groups did not differ with regard to consumption of
alcoholic beverages, regular medication, smoking, or body mass index. Level of education was
divided into three categories, "low," "medium," or "high" (definitions of these categories were
not provided). The numbers of exposed subjects by education group (low, medium, and high)
were 4, 8, and 2, respectively; the numbers of controls in these respective groups were 1, 12, and
10, indicating a considerable imbalance across these strata. The effect of tetrachloroethylene
exposure on the neurophysiological and neurobehavioral measurements was evaluated using
linear regression, adjusting for age, gender, and the three-level education variable.
Visual evoked potentials (a measure of visual function) in response to black-and-white
checkerboard patterns were recorded for all individuals (Altmann et al., 1995). Vibration
perception using a tuning fork—a crude measure of peripheral neuropathy—was assessed at the
ankle. Five tests included in the Neurobehavioral Evaluation System developed in the United
States and adapted for testing on a German population were used: (1) finger-tapping speed with
the index finger of both the dominant and the nondominant hand; (2) hand-eye coordination
using a joystick to follow a sine wave on a computer screen; (3) a continuous performance test
for assessment of vigilance, which requires a response to a specific stimulus appearing on the
computer screen and failure to respond to other stimuli; (4) simple reaction time, which requires
the fastest possible response to a simple visual stimulus (measured twice); and (5) visual memory
on the Benton visual retention test, which requires a match of a previously displayed stimulus out
of several choices after a short delay interval. All testing was completed in a single 3-hour
session; testing times were selected randomly for both exposed or control subjects.
Blood samples were taken in the exam room immediately before testing (all subjects)
and, if possible, once when the exposed subjects were at home (Altmann et al., 1995). The mean
blood level for exposed subjects at the examination was 0.0178 mg/L (standard deviation:
0.469 mg/L). For seven of the nine exposed subjects, blood concentrations in samples collected
at home were higher than those in samples collected at the examination. None of the blood
concentrations in the control group exceeded the detection limit of 0.0005 mg/L. For the
exposed subjects (data from 13 apartments), indoor air sampling indicated that the mean (7-day
TWA) air concentration was 0.7 ppm (standard deviation: 1 ppm), and the median was 0.2 ppm.
For the control group, the mean and median values were 0.0005 ppm (standard deviation: 0.0005
ppm) and 0.0003 ppm, respectively. There was a good correlation between home indoor air
concentrations and blood levels of tetrachloroethylene in the exposed subjects (r = 0.81). The
correlation was much lower when the examination room blood samples were used (r = 0.24).
4-27
-------
Altmann et al. (1995) observed statistically significant differences between the adjusted
mean scores of exposed and control subjects on neurobehavioral tests of simple reaction time
(p < 0.05 for the first test andp < 0.01 for the second test), continuous performance (p < 0.05),
and visual memory as tested with the Benton visual retention test (p < 0.05). In all cases, the
exposed subjects had slower response times or more errors than did the unexposed controls. The
degree of change from control was approximately 15-20% for these tests. The potential for
residual confounding by education should be considered; however, although education level was
independently associated with these measures and its affect on performance adjusted in the
statistical analysis, the use of three categories for education in the multivariate regression
analyses may not fully account for all effects from this covariate. No statistically significant
differences were observed between the performance of the exposed and control groups on the
finger-tapping or hand-eye coordination tests, which are measures of fine motor function; on
visual evoked potentials, which may be less sensitive than direct measurement of visual function;
or on vibration perception at the ankle using a tuning fork.
Schreiber et al. (2002) examined neurologic function as assessed by visual tests among
apartment residents who had potential environmental tetrachloroethylene exposure due to close
proximity to dry-cleaning facilities.13 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 exposed group consisted of 17 subjects (11 adults between the ages of 20 and 50,
2 adults over the age of 60, and 4 children, ages 6-18) from six families residing for a median of
6 years in two apartment buildings in New York City14 (Schreiber et al., 2002). Preliminary
monitoring of these buildings indicated tetrachloroethylene concentrations were elevated
compared to eight other buildings also monitored by the NYSDOH. Exposed residents were
from an affluent, English-speaking, Caucasian population living near New York City's Central
Park (telephone communication from K. Hudnell, EPA, to D. Rice, EPA, February 2003).
Exposed participants were generally unaware of the tetrachloroethylene exposure, although some
study participants noted tetrachloroethylene-like odors prior to the study. Controls were
recruited from among NYSDOH Albany, NY employees and their families. All controls were
Caucasian, except for one Asian individual, and were age- and sex-matched to exposed
apartment residents. In some cases, more than one control participant was matched to an
13 Another study by Schreiber et al. (2002) of day-care staff from a center collocated with a dry-cleaning facility,
using a similar testing protocol, was described in the Occupational Exposure Studies—Other Settings section.
14 Study subjects were identified through several methods: (1) both families in the first building (Building A) had
been referred to the NYSDOH for information about participating in the study by Consumer Union/Hunter College
researchers, (2) one family in the second building (Building B) had previously contacted NYSDOH about exposure
concerns and desired to participate in a study, and (3) three other families in Building B were recruited by a
participating family (NYS OAG. 2004).
4-28
-------
exposed subject. Mean age was 34.5 years for exposed apartment residents and 33.2 years for
control subjects.
The assessment of tetrachloroethylene exposure of residents consisted of concentrations
in indoor air and personal air samples, exhaled breath, and blood, which were collected at the
time of visual testing. Testing was performed during a period of active dry cleaning for four of
the families and 1 month after closure of the facility for the remaining two families in the
residential study. Adult residents also provided urine samples, which were analyzed for
tetrachloroethylene as well as for three products of its metabolism: TCA, trichloroethanol, and
the urinary acetyl metabolite. Ambient concentrations of tetrachloroethylene from 1 to 3 months
before the date of visual testing, when active dry cleaning was occurring in both apartment
buildings, were available for all subjects. Median concentrations in these samples were
0.21 ppm (mean: 0.36 ppm; range: 0.1-0.9 ppm). Airborne tetrachloroethylene concentrations
had decreased in samples collected at the time of visual testing; median tetrachloroethylene
concentration was 0.09 ppm (mean: 0.18 ppm; range: 0.01-0.78 ppm). Tetrachloroethylene
levels in blood correlated well with levels in room air, personal air, and breath.
All participants, or their guardians in the case of children, signed consent forms prior to
study commencement. Information on sociodemographics; lifestyle factors such as exposure to
direct or passive smoke, alcohol consumption, and exercise; medical history; and neurotoxicant
exposure in addition to the visual tests was obtained by questionnaire from both study
populations and their controls. Exposed participants had no known exposure to other
neurotoxicants, ongoing illness, current use of neuroactive drugs, or a medical history indicative
of neurologic dysfunction. Reported alcohol consumption (low to moderate) was similar in the
adult exposed and control groups, and the Profile of Moods test scores of all residential exposed
subjects were within normal limits. However, two of the four children had medically verified
diagnoses of learning disabilities or developmental delays (NYSDOH, 2004).
As described in the previous discussion of Schreiber et al. (2002) (refer to Section
4.1.1.2.2 Occupational Exposure Studies—Other Settings), visual function testing consisted of
near visual acuity, near visual contrast sensitivity, and color vision, and the investigators were
not blinded as to a subject's status as either exposed or nonexposed. The mean measure of visual
acuity was 20:27.7 in exposed residents and 20:22.8 in controls (p = 0.12). Group mean scores
for visual contrast sensitivity across spatial frequencies were statistically significantly lower in
exposed residents than in controls, indicating poorer visual function in the exposed groups (refer
to Figure 4-2). An exposure-response analysis did not show an association between poorer
performance and increasing tetrachloroethylene concentration. CCI scores (a measure of color
vision) of the exposed group were lower than those of controls, but the difference was not
4-29
-------
statistically significant (mean: 1.33 and 1.20 in exposed and control groups, respectively,
p = 0.26).
_ 160
g 140
v> 120
II 100
•I "
g 20
~a
k= 10
• Control (IT =17|
• Exposed In-=17'
1.5 3 6 12 18
Spatial frequency (eye le s-d QQ reg)
Figure 4-2. Visual contrast sensitivity functions for control and exposed
participants in residential exposure study (Schreiber et al., 2002).
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). Blue circles (top line) = controls; red circles (bottom line) =
exposed. For any frequency, a higher contrast sensitivity threshold represents better visual
function. Visual contrast sensitivity was significantly lower across all spatial frequencies in
exposed residents of apartments in building with dry-cleaning facilities compared with their
matched controls. Used with permission of the authors.
A larger study of the effect of tetrachloroethylene exposure on visual function in
residents living in buildings colocated with a dry-cleaning establishment was also conducted
(Storm etal.. 2011 [previously reported in NYSDOH. 20101: NYSDOH. 2005a. b). This study,
the New York City Perc Project, did not include the subjects in Schreiber et al. (2002) and
employed different methods for testing visual contrast sensitivity and color vision. Study design
and protocols were approved by Institutional Review Boards at the NYSDOH and other
collaborating institutes (Mt. Sinai Medical Center and the Centers for Disease Control and
Prevention). Sixty-five households in 24 residential buildings with dry cleaners using
tetrachloroethylene on-site, and 61 households in 36 buildings without dry cleaners were
recruited. Health outcome and tetrachloroethylene concentrations as measured from indoor air
monitoring and in exposed subject's breath and blood were obtained over the period from
2001-2003. McDermott et al. (2005) presents exposure monitoring findings from the dry-
cleaner households.
Subjects were identified in buildings from eleven contiguous zip code areas surrounding
Central Park, New York City. Household eligibility criteria included the presence of at least one
4-30
-------
adult (20-55 years old) and one child (5-14 years old), so as to assess whether residential
tetrachloroethylene exposure would disproportionately affect children. Initial monitoring
indicated few residences in dry-cleaner buildings, with elevated indoor air concentrations of
tetrachloroethylene above the current NYSDOH residential air guideline of 0.015 ppm
(0.1 mg/m3). The study area was broadened to include buildings that had been the subject of a
resident complaint and to include buildings in additional zip codes, primarily characterized by
lower socioeconomic status or higher percentage of minority residents. Of the 1,261 dry-cleaner
and 1,252 reference households contacted, 132 dry-cleaner households and 175 reference
households included age-eligible adult-child pairs. A total of 65 dry cleaner (67 adults,
68 children) and 61 referent households (61 adults, 71 children) participated in the study. The
socioeconomic status characteristics, residence duration, education level, age, and smoking and
alcohol use were similar in the adult residents of reference buildings and the residents of
buildings with dry cleaners. Differences between child residents in gender or residence duration
are not apparent, but the highest exposure group is about a year younger and has about one less
year of education than children in the other exposure groups. All participants or their guardians
signed voluntary consent forms prior to study commencement.
NYSDOH staff visited participants in their residences to collect 24-hour indoor air
samples and breath samples, and to give adult participants a questionnaire seeking information
on residential, occupational, and medical history for themselves and their children. Indoor air
tetrachloroethylene concentrations had decreased since 1997, the period of the pilot study
(Schreiber et al., 2002), and ranged up to around 0.77 ppm (5 mg/m3) with a geometric mean of
0.005 ppm (0.035 mg/m3) in apartment buildings colocated with a dry cleaner. Monitoring was
carried out using passive monitoring badges. In comparison, tetrachloroethylene concentrations
in buildings without dry cleaners ranged up to 0.014 ppm (0.09 mg/m3) with a geometric mean of
0.0004 ppm (0.003 mg/m3). Both breath and blood tetrachloroethylene levels were significantly
(p < 0.05) correlated with indoor air concentrations for adult and child subjects of dry-cleaning
buildings. LODs were 5 |ig/m3 air and 0.048 mg/mL blood. Air, breath, and blood
tetrachloroethylene concentrations were inversely correlated with income and were higher
among minority compared to nonminority subjects. Participants received financial compensation
after completing the home visit ($50.00) and ophthalmology clinic visit ($50.00).
Ophthalmologic examinations and visual function tests were given to study participants
at the Mt. Sinai Medical School of Medicine Department of Ophthalmology research clinic. The
final report does not describe whether examiners were or were not blinded as to a subject's
exposure status (NYSDOH, 2005a). The examination included determination of past ocular and
medical history; measurement of visual acuity, pupil size, extraocular motility, and intraocular
pressure; and anterior and posterior segment exams. Subjects with abnormalities or taking
4-31
-------
medications that could influence visual contrast sensitivity and/or color vision were excluded
from further testing. Furthermore, visual functional tests for some children were excluded from
the statistical analysis because of their young age or because they were identified by their parents
as learning disabled or having attention deficit hyperactivity disorder. Visual contrast sensitivity
was determined using the Functional Acuity Contrast Test (FACT) distance chart placed 10 feet
from the participant under light conditions of 68-240 cd/m2. These testing conditions differ
from those employed by Schreiber et al. (2002) in their residential study where visual testing was
carried out, assessing near-contrast sensitivity.
Adults and children demonstrated a ceiling effect with visual contrast sensitivity
performance, i.e., a maximum score at 1.5, 3, 6, 12, and 18 cycles per degree (cpd) is achieved
by some study participants. Visual contrast sensitivity scores among adults were not correlated
with any socioeconomic status factor or personal characteristics (smoking, alcohol use, education
level, duration of residence). Among all children, poorer visual contrast sensitivity at 1.5, 3, and
6 cpd was significantly correlated with speaking primarily Spanish at home.
Analyses examining relationships between tetrachloroethylene and visual function were
conducted using three categories of exposure: the referent exposure group (background exposure,
living in a building without a dry cleaner, geometric mean: 2.9 ug/m3 [0.0004 ppm], range:
1.5-4.2 |ig/m3 [0.0002-0.0006 ppm]); <100 |ig/m3 [geometric mean: 11.6 |ig/m3 (0.002 ppm},
range: 4.2-42.0 |ig/m3 (0.0002-0.006 ppm}]; and >100 |ig/m3 [geometric mean: 477.9 |ig/m3
(0.07 ppm}, range: 268.9-735.3 |ig/m3 (0.04-0.11 ppm}].15 A decreasing trend (p < 0.05) was
observed across these three exposure groups and the proportion of adults achieving the
maximum contrast sensitivity score at 6 cpd (28.3, 14.3, and 8.3% in the referent, <100, and
>100 |ig/m3 groups, respectively). This pattern was also observed in analyses stratified by race
or ethnicity, or by income, although the smaller sample sizes resulted in larger p-values (from
0.09 to 0.30) for each of the individual strata. In children, decreasing scores were observed at 6
cpd (43.4, 33.3, and 18.2% in the referent, <100, and >100 |ig/m3 groups, respectively, trend: p =
0.05) and 12 cpd (37.7, 33.3, and 0.0% in the referent, <100, and >100 |ig/m3 groups,
respectively, trend: p = 0.02). These effects were limited to minority and low income children in
the ethnicity and income-stratified analyses.
Results from logistic regression analyses further support susceptibility of children—but
not adults—to an adverse effect of tetrachloroethylene exposure on visual contrast sensitivity.
Whereas adult visual contrast sensitivity in the worse eye at 6 or 12 cpd was not significantly
influenced by any measure of tetrachloroethylene exposure, visual contrast sensitivity
performance at 12 cpd among children was significantly influenced (p < 0.05) by
100 ug/m3 =0.015 ppm.
4-32
-------
tetrachloroethylene concentrations in either indoor air or in blood; i.e., a lower percentage of
children achieved a maximum visual contrast sensitivity score with higher tetrachloroethylene
exposure. Odds ratio estimates were 2.64 (95% CI: 1.41, 5.52), 3.37 (95% CI: 1.44, 9.29), and
3.54 (95% CI: 0.94, 17.79) for the association between visual contrast sensitivity performance in
the worse eye at 12 cpd and indoor tetrachloroethylene, exhaled breath tetrachloroethylene at
home, and blood tetrachloroethylene, respectively. The logistic regression models examining
visual contrast sensitivity findings were adjusted for ethnicity or race and age, and, in adults,
smoking and alcohol use.
Color vision was assessed biocularly using both the Farnsworth D-15 test (differentiates
between strong/moderate and mild/normal CCI) and Lanthony D-15 test (differentiates between
normal and mild CCI). Both tests were administered under light conditions specified by the
manufacturer. Analyses were carried out using the proportion of subjects with no errors,
comparing quantitative differences in CCI, and logistic regression modeling to assess
associations between tetrachloroethylene exposure measures and occurrence of any major errors.
A high proportion of adult and child participants scored perfectly on both the Farnsworth and
Lanthony color vision tests. Lower annual household income, being a member of a minority
group, speaking primarily Spanish at home, and fewer years of education were all significantly
associated with increased CCI on both color vision tests. Tetrachloroethylene measures of
exposure were unrelated to color vision performance among adults; however, similar to visual
contrast sensitivity performance, children appear to be a more susceptible population. There
were no differences between exposure groups among adults or children in the percentage of
subjects with major errors on both color vision tests. A comparison of mean CCI between
exposure groups showed that children in the high-exposure category performed worse (mean:
CCI of 1.3, range: 1.0-1.9) compared with children in the low-exposure category (mean: CCI of
1.1, range: 1.0-1.7) and compared with referent children (mean: CCI of 1.2, range: 1.0-2.0) on
the Lanthony test; the test for trend for the three exposure groups was statistically significant
(p < 0.05). Performance (mean CCI) on the less sensitive Farnsworth test was not associated
with tetrachloroethylene exposure in either adults or children. Moreover, for children,
tetrachloroethylene in breath was significantly associated (p < 0.05) with making one or more
major errors on the Lanthony color vision test in logistic regression analyses that adjusted for the
effects of age and gender. Logistic regression analyses examining color vision and other
tetrachloroethylene measures such as indoor tetrachloroethylene concentration or breath
concentration were not discussed in NYSDOH (2005a). The higher mean difference in CCI
between children and adults in the highest exposure category (>0.015 ppm or >100 |ig/m3)
compared with referents was statistically significant. NYSDOH (2010) believed the decreased
mean child-adult difference in CCI was likely influenced by the low adult CCIs in the highest
4-33
-------
exposure group. Children in the high-exposure group, furthermore, were a year younger than in
other exposure groups; age was correlated with CCI and with tetrachloroethylene exposure in
this study. The highly correlated variables and the few numbers of children in the high exposure
group limit analysis of age effects on the association between breath tetrachloroethylene
concentration and CCI.
In summary, this study adopts a different approach than Schreiber et al. (2002) to assess
vision, using far vision methods as opposed to the near vision methods of Schreiber et al. (2002).
For contrast vision, a number of analyses (Storm et al., 2011 [previously reported in NYSDOH,
2010]: NYSDOH, 2005a) are suggestive of vulnerability among children. Exposure to >0.015
ppm (>100 |ig/m3) tetrachloroethylene was highly correlated with race and children's age, and
the sample sizes in the highest exposure group, especially in higher income, nonminority groups,
make it difficult to fully examine possible effects of income, race, and age on vision. However,
association of tetrachloroethylene exposure >0.015 ppm (>100 |ig/m3) with visual deficits
suggests a susceptibility of the children studied.
4.1.1.2.4. Oral exposure studies
Risk of learning and behavioral disorders was evaluated in relation to prenatal and
postnatal exposure to tetrachloroethylene in Cape Cod towns with a contaminated water
distribution system during 1969-1983 (Janulewicz et al., 2008). Mothers reported
developmental and educational histories and learning and behavioral disorders in self-
administered questionnaires returned during 2002-2003. Developmental risks were evaluated in
relation to the amount of tetrachloroethylene delivered to each subject's residence during the
prenatal period (from the month and year of the last menstrual period through the month and year
of the birth) and during the early postnatal period (from the month and year of the birth through
the month and year of the 5th birthday). Prenatal and postnatal exposures were evaluated
separately in generalized estimating equation regression models. After excluding 404 subjects
because they had an attribute with a known association with the outcomes under study, there
were 2,086 children in the final data set. Of these, 842 and 1,244 children had no and any
prenatal exposure, respectively, and 760 and 1,326 children had no and any postnatal exposure,
respectively. Exposed and unexposed children were similar with respect to demographic
characteristics and behaviors. Low- and high-exposure categories were developed for the
9-month prenatal period and 5-year postnatal period using the number of grams of
tetrachloroethylene that corresponded to an average drinking water concentration of 40 ug/L, the
action level used in 1980, as a cutpoint. The authors reported that no meaningful associations
were observed between prenatal exposure and receiving tutoring for reading or math, being
placed on an Individualized Education Plan, or repeating a school grade. Increased odds ratios
4-34
-------
were noted among subjects with low exposure compared to no exposure for receiving a diagnosis
of attention deficit disorder or hyperactivity disorder, special class placement for academic or
behavioral problems, or lower educational attainment (high school graduate or less). However,
odds ratios were not markedly increased for subjects with high exposure (<1.1). For example, in
generalized estimating equations models adjusted for maternal age, race, and education, child's
sex, and prematurity and/or low birth weight, the odds ratio for attention deficit disorder was 1.4
(95% CI: 0.9-2.0) among subjects with low prenatal exposure and 1.0 (95% CI: 0.7-1.6) among
subjects with high prenatal exposure. For postnatal exposure, no associations were observed for
receiving tutoring for reading or math, special class placement for academic or behavioral
problems, repeating a grade in school, or lower educational attainment. The same pattern of risk
with exposure level also was observed for low and high postnatal exposure compared to no
exposure. For example, the adjusted odds ratio for attention deficit disorder was 1.3 (95%
CI: 0.9-1.9) among subjects with low postnatal exposure and 1.0 (95% CI: 0.6-1.7) among
subjects with high postnatal exposure.
4.1.1.3. Summary of Neuropsychological Effects in Low- and Moderate-Exposure Studies
A summary of neuropsychological effects observed in chronic occupational or residential
exposure studies of tetrachloroethylene is shown in Table 4-2 and discussed by domain below.
Several studies (Storm et al.. 2011 [previously reported in NYSDOH, 20101: NYSDOH, 2005a,
b; Schreiber et al., 2002; Altmann et al., 1995; Echeverria et al., 1995) employed multiple
measures of exposure (indoor air monitoring, personal monitoring, and in some cases, biological
monitoring). Although some variation is expected and was observed in individual studies
[Altmann et al. (1995) for example], the correlation between tetrachloroethylene concentration as
assessed from indoor air monitoring or personal monitoring and biological metrics such as blood
tetrachloroethylene concentration was quite strong, suggesting indoor air concentration as a
reasonable exposure metric. Many studies did not include exposure monitoring of individual
subjects, and the statistical analyses compare groups using Mests or %2 tests (Spinatonda et al.,
1997; Ferroni et al., 1992). Dose response and multiple logistic regression analyses are
statistically more powerful, and five studies observed correlations or associations between
various tetrachloroethylene exposure measures and specific neurobehavioral tests (Storm et al.,
2011 [previously reported in NYSDOH, 20101: Altmann et al., 1995: Echeverria et al., 1995:
Cavalleri etal., 1994: Seeber, 1989).
4-35
-------
Table 4-2. Summary of effects of chronic tetrachloroethylene exposure in humans observed in studies of
neuropsychological function3
(Reference), n exposed,
mean or median
exposure(s)
Visual domain"
Spatial
vision
(VCS)
Color
vision1"
VEP
Cognitive domain (executive function, attention)3
Visuo-
spatial
memory0
Vigilance
Trail-
making
Digit
span,
symbol
Cancellation
Information
processing"1
Motor"
Simple
reaction
time
Fine
motor
function
Occupational exposures — dry-cleaning settings
Lauwerys et al. (1983),
n = 26, 21 ppm
Seeber (1989). n = 101,
12 and 53 ppm
Nakatsuka et al. (1992).
n = 64, 13 ppm
Ferroni et al. (1992).
n = 60, 15 ppm
Cavalleri et al. (1994).
n = 35, 6 ppm
Gobba et al. (19981:
Cavalleri et al. (1994)
follow-up, n = 33, 4 ppm
Echeverria et al. (1995).
« = 65, 11, 23, 41 ppm
Echeverria et al. (1994).
n = 173, <0.2, 3, 9 ppm
Spinatonda et al. (1997).
n = 35, 8 ppm
Sharanjeet-Kaur et al.
(2004). n = 14, not
reported
—
+
+
+
+
+
+
+
—
—
+
—
—
—
+
+
—
+
*
+
—
—
-------
Table 4-2. Summary of effects of chronic tetrachloroethylene exposure in humans observed in studies of
neuropsychological function (continued)
(Reference), n exposed,
mean or median
exposure(s)
Visual domain"
Spatial
vision
(VCS)
Color
vision1"
VEP
Cognitive domain (executive function, attention)3
Visuo-
spatial
memory0
Vigilance
Trail-
making
Digit
span,
symbol
Cancellation
Information
processing11
Motor"
Simple
reaction
time
Fine
motor
function
Occupational exposures — other settings
Schreiber et al. (2002).
Day -care workers n = 9,
0.32 ppm
+
Residential exposures
Altmann et al. (1995).
n = 19, 0.7 ppm
Schreiber et al. (2002).
n = 17 (13 adults and 4
children), 0.4 ppm
McDermott et al. (2005):
NYSDOH (Storm etal.
201 1:NYSDOH. 2010).
n = 68 children (C),
n = 67 adults (A), 0.005
ppm
+
+ (C),
— (A)
+
(trend)
-(Q,
— (A)
—
+
+
+
—
a + denotes effects observed (i.e., worse performance) in exposed group; — denotes no effect or better performance in exposed group; —* denotes better
performance in the exposed group (before shift measure); blank cell denotes test not performed.
b Based on Lanthony D-15 test, except for Nakatsuka et al. (1992). who used a less sensitive version of this test.
0 Tests include digit reproduction (Seeber. 1989); switching, pattern memory, and pattern recognition (Echeverriaetal.. 1995; Echeverriaetal.. 1994); and
Benton test (Altmann etal.. 1995).
d Tests include choice reaction time (Seeber. 1989). perceptual threshold (Seeber. 1989). finger tapping (Ferroni et al.. 1992). and vocal reproduction to reading
stimuli (Spinatonda et al.. 1997).
-------
4.1.1.3.1. Visual function domain
Color vision and visual contrast sensitivity are the visual domains that have been
observed to be affected by chronic exposure to tetrachloroethylene (refer to Table 4-2). Only
Schreiber et al. (2002) and NYSDOH (2005a) assessed spatial vision (VCS, visual contrast
sensitivity), an effect reported for exposure to other solvents (Schreiber et al., 2002; Hudnell et
al.. 1996a: Hudnell et al.. 1996b: Broadwell et al.. 1995: Campagna et al.. 1995: Donoghue et al..
1995: Bowler etal.. 1991: Frenette et al.. 1991: Mergler et al.. 1991). In Schreiber et al. (2002).
visual contrast sensitivity deficits in subjects (mostly adults) with normal visual acuity were
observed at low-exposure concentrations in residential populations, and in NYSDOH (2005a):
evidence of these effects were observed in children but not in adults. Exposure levels were
lower in the latter study [mean: 0.4 ppm and geometric mean: 0.005 ppm in Schreiber et al.
(2002) and NYSDOH (2005a), respectively]. Potential bias and confounding could have been
introduced, however, from a lack of blinding of testers and, in the latter study, the inability to
control for socioeconomic and other factors that were highly correlated with higher
tetrachloroethylene exposures.
Deficits in blue-yellow color vision, a well established effect of solvents, were observed
in dry-cleaning workers in Italy in Cavalleri et al. (1994) and in a follow-up study (Gobba et al.,
1998) of this population. Cavalleri et al. (1994) specifically noted that the color vision testing
was conducted by examiners who were blinded to the exposure level of individual study
participants, and the study participants were well-matched in terms of age, smoking, and alcohol
use. Mean TWA exposure levels were approximately 6 ppm among all workers in Cavalleri
et al. (1994). There also was a statistically significant positive correlation (p < 0.01) between
TWA air concentrations and the CCI (r = 0.52), which remained after multivariate analysis
considered previous tetrachloroethylene exposure, duration, age, number of cigarettes a day, and
daily intake of alcohol as covariates. This type of color vision deficit was not observed in the
dry cleaners study by Nakatsuka et al. (1992), but the form of the color vision test used in the
latter study, the Lanthony 15, is less sensitive to mild and moderate changes in color vision
compared with the desaturated version of the test (Lanthony D-15) used in the other studies
(Lanthony, 1978). Effects on color vision were also observed among 14 dry cleaners in the small
study in Malaysia by Sharanjeet-Kaur et al. (2004), but the lack of exposure information (other
than job title), and differences between dry cleaners and controls regarding test conditions and
smoking habits indicate that this study should provide little weight in the overall conclusions
regarding color vision. Two other small studies also reported lower scores on the Lanthony D-15
color vision test in exposed groups compared with controls, but the differences were not
statistically significant: in a study of residents living above dry cleaners (mean
4-38
-------
tetrachloroethylene exposure during active dry cleaning = 0.4 ppm), the mean CCI scores were
1.33 and 1.20 in 17 exposed and 17 control groups, respectively (p = 0.26); in a study of workers
in a day-care center located in a building with a dry-cleaning business (mean tetrachloroethylene
exposure: 0.32 ppm), the mean CCI scores were 1.22 and 1.18 in the exposed day-care workers
and controls, respectively (p = 0.39) (Schreiber et al.. 2002). The follow-up study of NYSDOH
(2005a) further suggests tetrachloroethylene effects on color vision, particularly in children,
although NYSDOH (2010) believed the decreased mean child-adult difference in CCI was likely
influenced by the low adult CCIs in the highest exposure group.
Peer-consultation comments on EPA's earlier draft Neurotoxicity of Tetrachloroethylene
(Perchloroethylene) Discussion Paper (U.S. EPA, 2003) noted that the deficit in contrast
sensitivity could reflect a sensitivity of the visual system to tetrachloroethylene, or it may be that
this test was relatively more sensitive than other vision tests or tests used for other domains (U.S.
EPA, 2004). Furthermore, the peer consultants also suggested that contrast sensitivity loss may
reflect impaired function throughout the brain, because contrast sensitivity is affected by retinal,
optic nerve, or central brain dysfunction (U.S. EPA, 2004). Nonetheless, drawing strong
conclusions from these studies is difficult, particularly in light of the paucity of data on this test
in occupational populations with higher exposure concentrations and in animal studies.
Although Altmann et al. (1992; 1990) reported alterations in visual evoked potentials
(p < 0.05) with 4-hour acute exposure at 10 ppm, they were not altered in residents exposed
chronically to a median of around 1 ppm tetrachloroethylene (Altmann et al., 1995). Acute and
chronic exposures are of different patterns—short-term peak exposure versus longer-duration
exposure—and, therefore, may result in a different pattern of response.
4.1.1.3.2. Cognitive domain
Cognitive domains affected by tetrachloroethylene include visuospatial memory,
attention, vigilance (continuous performance), and speed of information processing (refer to
Table 4-2). Effects on visuospatial memory are of particular interest, given similar results in
studies that examined this type of effect in occupational (Echeverria et al., 1995; Echeverria et
al., 1994; Seeber, 1989) or residential (Altmann et al., 1995) settings, and given similar reports
for other solvents (Daniell et al., 1999; Morrow et al., 1990). Echeverria et al. (1995) found
effects among 23 dry cleaners classified as having a high chronic exposure (based on type of
shop, job title, and years of employment) on tests of pattern memory, visual reproduction, and
pattern recognition in the absence of effects on attention (digit symbol and digit span) or
executive function (trailmaking A and B). Further, Echeverria and colleagues (1994) confirmed
these findings in an independent sample of dry cleaners categorized as having high lifetime
chronic exposure and whose current exposure level was 9 ppm, 8-hour TWA; the exposure level
4-39
-------
of 9 ppm is not representative of past chronic exposure levels because of changes occurring in
the industry (i.e., switching from wet-transfer to dry-to-dry machine). Seeber (1989) also
reported impaired visuospatial recognition in a low exposure (mean: TWA 12 ppm) and a high
exposure group (mean: TWA 53 ppm), and Altmann et al. (1995) observed deficits on a test of
visuospatial function in residents with much lower exposure concentrations (mean: 0.7 ppm)
than those of the occupational studies. All of these studies except Altmann et al. (1995) reported
that investigators were blinded to knowledge of the exposure level of the subject. These studies
provide strong weight, given the numbers of subjects and their use of appropriate statistical
methods, including adjustment for potentially confounding factors that may be relevant for
measures of the cognitive domain. For example, Seeber (1989) adjusted for age, gender, and a
measure of intelligence (alcohol was examined but not shown by these investigators as
confounding the association between tetrachloroethylene and cognitive performance), and a
variety of potential confounders were evaluated by Echeverria et al. (1995; 1994). It should be
noted, however, that residual confounding from education level differences between exposed and
referent subjects may still be present in Altmann et al. (1995).
The results pertaining to cognitive measures other than visuospatial memory are
somewhat mixed. Altmann et al. (1995) and Ferroni et al. (1992) assessed vigilance using a
continuous performance procedure in which the subject faces a screen that presents one of
several different stimuli at random intervals. The subject must make a response to a specified
stimulus and not to the others. This test measures sustained attention and is correlated with
performance on tests of executive function. Both studies found deficits as a result of
tetrachloroethylene exposure on this task. Seeber (1989) found effects on two tests of attention
(digit reproduction and digit symbol) that are subsets of the Wechsler IQ tests and were designed
to be sensitive to performance within the normal range. These investigators also found positive
effects on a visual scanning test (cancellation d2) that is usually used to assess laterality of brain
damage but has also proved sensitive to toxicant (lead) exposure (Bellinger et al., 1994). In
contrast, Echeverria et al. (1995) and Ferroni et al. (19921 as described in NYSDOH (1997) did
not find effects on digit span, which is given as a test of attention and memory, or digit symbol,
despite higher levels of exposure than in Seeber (1989). Speed of information processing was
assessed in two studies: Seeber (1989) and Spinatonda et al. (1997). Seeber used two tasks:
recognition and choice reaction time. Effects were observed in both groups on a task requiring
recognition of briefly presented stimuli (perceptual speed). In the choice reaction time task
(correct reactions), effects were borderline in the lower-exposure group and negative in the
higher-exposure group, with no exposure-response relationship. Spinatonda et al. (1997)
observed longer mean reaction times and/or vocalization durations to vocal and visual stimuli.
4-40
-------
Two studies—an occupational study with relatively higher exposure (Ferroni et al., 1992)
and the Altmann et al. (1995) residential study—also assessed simple reaction time, a task that
uses a motor response and demands a relatively modest amount of attention. In both studies,
lower performance [ranging from an increase in reaction time from 24 (11%, 102 mg/m3)
(Ferroni et al.. 1992)1 to 50 ms [20%, 4.99 mg/m3 (Altmann et al.. 1995)1 was observed among
the exposed workers compared with referents. A third study, Lauwerys et al. (1983), reported
better performance on simple reaction time in exposed workers compared with referents when
measured before a work shift but not when measured after work.
4.1.1.3.3. Motor function domain
Tetrachloroethylene exposure has not been reported to affect fine motor tests. Seeber
(1989), Ferroni et al. (1992), and Altmann et al. (1995) each assessed fine motor control using
various instruments, and all three found no significant decrements in fine motor performance.
4.1.1.3.4. Other clinical tests and conditions
A clinical neurological examination that includes the Romberg test, neuroradiological
examination, neurophysiological tests such as EEGs, and nerve conduction tests or other tests for
peripheral neuropathy have observed limited use for assessing neurotoxicologic effects in
tetrachloroethylene-exposed populations. Mental disease and behavioral disorders of neurologic
origin have not been well studied with respect to environmental factors. Perrin et al. (2007), who
reports an association between schizophrenia and parental exposure in dry cleaning, is the only
such study. A fourfold increased risk of schizophrenia was observed among offspring.
However, in a small study, Janulewicz et al. (2007) did not observe an association between
prenatal or early postnatal drinking water exposure to tetrachloroethylene and disorders of
learning, attention, and behavior. Therefore, other studies are needed to understand the role of
parental tetrachloroethylene exposure in the development of mental disease and behavioral
disorders in children.
4.1.2. Animal Studies
Tetrachloroethylene exposure in experimental studies in animals results in general
CNS-depressant activity (decreased activity, anxiolytic behavior, lethargy), impairment in
balance and motor coordination, cognitive defects, sleep cycle changes, and changes in visual
function and nerve conduction velocity. These changes have been observed following either an
inhalation or oral/intraperitoneal (i.p.) exposure. In addition to these effects, several effects on
brain pathology including DNA- and RNA-level changes, changes in neurotransmitter levels
such as acetylcholine and glutamate, and changes in brain fatty acid composition, have been
4-41
-------
observed. Some studies also document potential developmental neurotoxicity consequences
following exposure to tetrachloroethylene during the gestation period.
4.1.2.1. Inhalation Studies
The animal inhalation neurotoxicity studies are summarized in Table 4-3 and described in
more detail below. Neurobehavioral, neurophysiological, and developmental neurotoxicity
effects have been reported following tetrachloroethylene exposure. Two neurobehavioral studies
observed that there was an increase in motor activity following a 1-hour exposure in NMRI mice
at 90 ppm and higher (Kj ell strand et al., 1985), and there was a decrease in immobility in Swiss
OF1 mice at 649 ppm and higher during 4 hours of exposure (De Ceaurriz et al., 1983). A more
recent neurobehavioral study examined effects of Long-Evans rats in a signal detection test and
reported decreased sustained attention as a measurement of decreased trial completions and
increased reaction time during an hour exposure to 500 ppm or higher (Oshiro et al., 2008). In
F344 rats, significant changes in FEP latency and amplitude following a 12-week repeated
exposure to 800 ppm or higher were reported by Mattsson et al. (1998), and in Long-Evans rats,
changes in visual evoked potential amplitudes during an acute (60-120 minutes) exposure to
250 ppm or higher were reported by Boyes et al. (2009). Developmental neurotoxic effects were
noted in three studies (Szakmary et al., 1997; Tinston, 1994; Nelson et al., 1979) where changes
such as decreases in muscular strength and exploratory behavior as well as other behavioral
habits were significantly different from nonexposed litters. Finally, there were many changes in
brain pathology as noted by decreased brain weight, brain DNA levels, and changes in
neurotransmitter levels (Wang et al., 1993; Kyrklund and Haglid, 1991; Kyrklund et al., 1990,
1988: Karlsson et al., 1987: Kyrklund et al., 1987: Briving et al., 1986: Rosengren et al., 1986:
Kj ell strand etal., 1984: Kyrklund et al., 1984: Honmaetal., 1980a: Honmaetal., 1980b:
Savolainen et al., 1977a: Savolainen et al., 1977b).
4.1.2.1.1. Neurobehavior
De Ceaurriz et al. (1983) exposed male Swiss OF1 mice (n = 10 per exposure group) to
596, 649, 684, or 820 ppm tetrachloroethylene for 4 hours. Immediately following exposure, the
mice were immersed in a cylinder filled with water, and the duration of immobility was observed
for 3 minutes. The term "behavioral despair" has been coined for this initial immobility, and the
length of immobility is shortened by antidepressant administration. Tetrachloroethylene
exposure also shortened the period of immobility, with a no-observed-effect level (NOEL) of
596 ppm.
4-42
-------
Table 4-3. Summary of animal inhalation neurotoxicology studies
Subjects
Effect
NOAEL/LOAEL" (iinni)
Reference
Neurobehavioral studies
Swiss OF1 mice, males
10/dose
NMRI mice, males (n = 27
for 90, 320, 400, 600; n = 14
for 800, 1,200,1,800,3,600
ppm)
Long-Evans rats, males
(n = 12 total; animals served
as own controls)
Decreased duration of immobility
Increased motor activity
Increased number of false alarms,
increased reaction time, and decreased
trial completions in a signal detection
task measuring sustained attention
596. 649. 684, 820; 4 h
2Q, 3,600; 1 h
0, 500. 1,000, 1,500; 60 min
De Ceaurriz
et al. (19831
Kjellstrand et
al. (1985)
Oshiro et al.
(2008)
Neurophysiological studies
F344 rats
Pilot study: male
10/dose
Follow-up study: males and
females
12/sex dose
Long-Evans rats, males
(n = 9-10/exposure)
Changes in FEP, SEP, EEC
Increased amplitude and latency in
late component of FEP
Decreased frequency doubled
amplitude in the steady state VEP
0, SQfl; 4 d, 6 h/d
50, 200. 800;
13 wk, 6 h/d, 5 d/wk
0,25fl, 500, 1,000 for 1.5 h
Mattsson et
al. (1998)
Boyes et al.
(2009)
Developmental neurotoxicity studies
S-D rats
pregnant females
13-21 litters/dose;
males and female
offspring assessed
CFY rats
pregnant females
15 litters/dose;
male and female offspring
assessed
S-D rats, multigeneration
study
28 litters/dose
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 d postnatally
CNS depression in first 2 wk of Fl
and F2 generations, which ceased 2 h
after daily exposures
0, 100. 900 on
CDs 7-13 or on CDs 14-20, 7
h/d
0, 1,500 or 4^500 ms/m3
CDs 1-20 for 8 h/d
0, 100, 300, 1,000;
6 h/d, 5 d/wk, except during
mating, 6 h/d-7 d/wk
Nelson et al.
(1979)
Szakmary et
al. (1997)
Tinston
(1994)
4-43
-------
Table 4-3. Summary of animal inhalation neurotoxicology studies
(continued)
Subjects
Effect
NOAEL/LOAEL" (ppm^
Reference
Brain pathology
S-D rats, males
8/dose
S-D rats, males
10/dose
S-D rats, males
5-6/dose
S-D rats, males
5-6/dose
Mongolian gerbils
males and females
6/sex/dose
Mongolian gerbils
males and females
4/sex/dose
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
Decreased brain weight, DNA, protein
Decreased brain RNA, increased brain
cholinesterase and increased motor
activity
Change in fatty acid composition of
cerebral cortex
Neurotransmitter changes, brain
regions
Decrease in DNA, frontal cortex
Decrease in brain weight
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
300. 600:
4 or 12 wk continuous (24 h/d)
200: 4 d
320; 12 wk
continuous (24 h/d), 30-d
washout period;
320; 4 wk continuous (24 h/d)
200, 400. 800; 4 wk continuous
(24 h/d)
60. 300; 12 wk, continuous
(24 h/d); 16-wk washout period
fiQ;12 wk, continuous (24 h/d)
120:12 mo continuous (24 h/d)
320: 12 wk continuous (24 h/d)
120;52 wk continuous (24 h/d)
Maximum exposure 160; GDs
33 to 65 continuous (24 h/d)
9C, 21, 75, 150; 4 wk
continuous
(24 h/d)
150; 4 wk intermittent-
(1, 2, 4, 8, or 16 h/d)
Wang et al.
(1993)
Savolainen et
al. (1977a:
1977b)
Kyrklund et
al. (1990.
1988)
Honma et al.
(1980a:
1980b)
Rosengren et
al. (1986)
Karlsson et
al. (1987)
Briving et al.
(1986)
Kyrklund et
al. (1987)
Kyrklund et
al. (1984)
Kyrklund and
Hagid (1991)
Kjellstrand et
al. (1984)
Kjellstrand et
al. (1984)
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.
EEG = electroencephalogram; FEP = Flash-evoked potential; GD = Gestational day; S-D = Sprague-Dawley; SEP =
Somatosensory-evoked potential; VEP = Visual Evoked Potential
4-44
-------
The effects of exposure to 90-3,600 ppm tetrachloroethylene for 1 hour on motor activity
were examined in male NMRI mice (n = 14-27 per exposure group ) (Kj ell strand et al., 1985).
A strong odor (cologne) was used as the control condition. Total activity was monitored during
the dark period during exposure and for several hours thereafter. All doses produced increased
activity during exposure; activity decreased over several hours after cessation of exposure.
Although apparently no statistical analyses were performed, it is clear from the figures that the
lowest dose produced an average performance that was well outside the boundary of the 95% CIs
of the cologne-treated controls, and performance was dose dependent.
Male Long-Evans rats (n = 12) previously trained to perform a visual signal detection
task were exposed to 0, 500, 1,000, and 1,500 ppm tetrachloroethylene for 60 minutes (Boyes et
al., 2009). In this learned task, rats are trained to respond to a light stimulus by pressing the
stimulus lever and to press the blank lever when there is no stimulus. Food pellets are provided
to the rat for each correct lever response. The parameters evaluated included measures of (1)
correct responses (pressing stimulus lever with stimulus), (2) correct rejections (pressing blank
lever when stimulus is not presented), (3) false alarms (pressing stimulus lever without stimulus),
and (4) misses (pressing the blank lever when the stimulus is presented). Other endpoints
measured included reaction times from presentation of stimulus to pressing of the lever and if the
rat completed the signal detection task within the allotted period of time (2 minutes).
Tetrachloroethylene (500-1,500 ppm) exposure significantly increased the number of false
alarms, indicative of a decrement in sustained attention. Additionally, the authors reported that
there was a dose-dependent increase in reaction time and decreased trial completions. Rats were
also tested with different signal intensities to evaluate if the changes were partially due to visual
deficits. The number of hits did not significantly change with the signal intensity of the stimulus,
which strongly suggests that the observed effects of tetrachloroethylene in this study are due to
cognitive changes rather than visual effects. The study authors reported a LOAEL of 500 ppm
(60 minute exposure) for effects related to decrements in sustained attention.
4.1.2.1.2. Neurophysiology
Mattsson et al. (1998) studied the effects of acute exposure to tetrachloroethylene for
13 weeks on visual function (flash-evoked potentials [FEPs], EEGs, sensory function
(somatosensory-evoked potentials [SEPs]), and rectal temperature in F344 rats. During the acute
(pilot) study, male rats were exposed to 0 or 800 ppm tetrachloroethylene, for 6 hours/day, for 4
days, and tested before and after exposure on the 4th day. Changes in FEP, SEP, and EEG
components were observed after acute exposure. In the subchronic study, the above evoked
potentials and caudal nerve conduction velocity were determined in male and female rats
4-45
-------
exposed to 0, 50, 200, or 800 ppm, for 6 hours/day, for 13 weeks. Testing was performed during
the week following cessation of exposure. A significant increase in the amplitude and in latency
(-3.0 ms) for the mid-component peak of the FEP was observed at the highest dose (800 ppm).
Several measures of the evoked potential were affected at 50 ppm but not at higher doses. Other
measures were not affected, and no dose response was observed.
Male Long-Evans rats (n = 9-10/group) were exposed to concentrations of
tetrachloroethylene ranging from 0-4,000 ppm in two separate experiments evaluating visual
function by measuring pattern-elicited steady state visual evoked potentials (Boyes et al., 2009).
In the first experiment, rats were exposed to (mean ± SEM in parentheses) 0, 1,000 (1,006 ± 7.4),
2000 (1993 ± 8.3), 3,000 (3,018 ± 6.9), or 4,000 (4,016 ± 19) ppm for 2 hours (0, 1,000, 2,000
ppm), 1.3 hours (3,000 ppm), or 1 hour (4,000 ppm). In the second experiment, rats were
exposed to 0, 250 (249 ±1.1), 500 (488 ± 2.9), or 1,000 (1,053 ± 9.6) ppm for 1.5 hours. In both
experiments, the visual evoked potentials were measured while the animal was exposed to
tetrachloroethylene. The steady state visual evoked potential responses measured from the
animals are sinusoidal in nature, and the potentials were transformed so that amplitudes were
tabulated at the frequency of pattern presentation (Fl) and at double the frequency of pattern
presentation (F2). At all test conditions, tetrachloroethylene significantly decreased the F2
amplitude of the steady state visual evoked potential. The LOAEL for steady state visual evoked
potentials for this study is 250 ppm tetrachloroethylene for 1.5 hours.
4.1.2.1.3. Developmental neurotoxicity
Developmental neurotoxicity is also discussed in Section 4.7.1.2. Nelson et al. (1979)
investigated developmental neurotoxicity in Sprague-Dawley (S-D) rats by exposing pregnant
dams to tetrachloroethylene at concentrations of 100 or 900 ppm during both early pregnancy
(gestation days [GDs] 7 to 13) or late pregnancy (GDs 14 to 20). The investigators made
morphological examinations of the fetuses and performed behavioral testing and neurochemical
analysis of the offspring. There were no alterations in any of the measured parameters in the 100
ppm groups. At 900 ppm, there were no skeletal abnormalities, but the weight gain of the
offspring as compared with controls was depressed about 20% at Weeks 3-5. Developmental
delay was observed in both the early and late pregnancy groups. Offspring of the early
pregnancy-exposed group performed poorly on an ascent test and on a rotarod test (evaluation of
neuromuscular function), whereas those in the late pregnancy group underperformed on the
ascent test only at postnatal day (PND) 14. However, later in development (PNDs 21 and 25),
their performance was higher than that of the controls on the rotarod test. These pups were
markedly more active in the open field test at PNDs 31 and 32.
4-46
-------
There were no effects on running in an activity wheel on PNDs 32 or 33 or avoidance
conditioning on PND 34 and operant conditioning on PNDs 40 to 46. Neurochemical analyses
of whole brain (minus cerebellum) tissue in 21-day-old offspring revealed significant reductions
in acetylcholine levels at both exposure periods, whereas dopamine levels were reduced among
those exposed on GDs 7-13. However, none of the statistics for the 100 ppm treatments were
presented. The authors observed that more behavioral changes occurred in offspring exposed
during late pregnancy than in those exposed during early pregnancy.
Szakmary et al. (1997) exposed CFY rats to tetrachloroethylene via inhalation throughout
gestation (i.e., GDs 1-20) for 8 hours/day at concentrations of 0, 1,500, or 4,500 mg/m3
tetrachloroethylene. The primary focus of the study was prenatal developmental evaluations
(refer to Section 4.7.2). However, a cohort of rats (15 litters/group) was allowed to deliver, and
the offspring (standardized to 8 pups/litter) were maintained on study until PND 100 and
evaluated for growth, development, and neurotoxic effects. The report did not specify whether
the animals were exposed to tetrachloroethylene after birth. Preweaning observations included
weekly body weights, developmental landmarks (pinna detachment, incisor eruption, and eye
opening), and functional assessments (forward movement, surface righting reflex, grasping
ability, swimming ontogeny, rotating activity, auditory startle reflex, and examination of
stereoscopic vision). After weaning, exploratory activity in an open field, motor activity in an
activity wheel, and development of muscle strength were assessed. The study authors reported
that adverse findings included a decreased survival index (details were not provided), a minimal
decrease of exploratory activity and muscular strength in treated offspring (presumably at both
exposure levels) that normalized by PND 51, and significantly increased motor activity on PND
100 of females exposed to 4,500 mg/m3. Litter was evaluated as the statistical unit of measure
for all outcomes. There is no clear indication of group means for postnatal measures reported.
The lack of experimental detail in the postnatal evaluation part of this study reduces the overall
confidence in the findings. There was no evaluation of postnatal histopathology of the nervous
system reported or cognitive testing during the postweaning period or during adulthood.
Tinston (1994) performed a multigeneration study of the effects on rats exposed to
airborne concentrations of tetrachloroethylene. The details of the study are discussed in
Section 4.7.2. The investigators observed several developmental effects. Of interest here were
the signs of CNS depression (decreased activity and reduced response to sound) observed for the
first 2 weeks in both adult generations and when the exposure was resumed on Day 6 postpartum
in the Fl generation (adults and pups). These effects disappeared about 2 hours after cessation
of the daily exposure. Other overt signs of tetrachloroethylene poisoning among the adults
included irregular breathing and piloerection at both 300 and 1,000 ppm. These changes stopped
concurrently with cessation of exposure or shortly thereafter.
4-47
-------
4.1.2.1.4. Brain pathology changes
Wang et al. (1993) exposed male S-D rats to 300 ppm tetrachloroethylene continuously
for 4 weeks or 600 ppm for 4 or 12 weeks. Exposure to 600 ppm at either duration resulted in
reduced brain-weight gain, decreased regional brain weight, and decreased DNA in the frontal
cortex and the brain stem but not the hippocampus. Four specific proteins [S-100 (an astroglial
protein), glial fibrallary acidic protein, neurone specific enolase, and neurofilament (68-kD
polypeptide)] were decreased at 4 and/or 12 weeks exposure to 600 ppm; 300 ppm had no effect
on any endpoint.
The effects of exposure to 200 ppm tetrachloroethylene, for 6 hours/day, for 4 days, in
male S-D rats were examined for a number of endpoints (Savolainen et al., 1977a: Savolainen et
al., 1977b). Rats were euthanized on the 5th day following a further 0-6 hours of exposure.
Tetrachloroethylene levels were highest in fat, followed by liver, cerebrum, cerebellum, lung,
and blood. Tissue levels increased in all tissues over the 6 hours of exposure. Brain RNA
content decreased, and brain nonspecific cholinesterase was increased on the 5th day, although no
statistical comparisons were performed. Locomotion in an open field was increased immediately
following the end of exposure on the 4th day, with no difference 17 hours after exposure,
although no statistical comparisons were made. Brain protein, GSH, and acid proteinase were
unaffected.
A series of experiments were performed on the effects of tetrachloroethylene on brain
lipid patterns. Exposure to 320 ppm for 90 days (Kyrklund et al., 1990) or 30 days (Kyrklund et
al., 1988) in male S-D rats resulted in changes in the fatty acid composition of the cerebral
cortex, which persisted after a 30 day recovery period (Kyrklund et al., 1990). Similar results
were observed in the cerebral cortex and the hippocampus of Mongolian gerbils (sex
unspecified) as well as reduced brain weight after exposure to 320 ppm (Kyrklund et al., 1987).
Exposure of male Mongolian gerbils to 120 ppm for 12 months also resulted in decreases in
long-chain, linolenic acid-derived fatty acids in the cerebral cortex and the hippocampus
(Kyrklund et al.. 1984).
The effect of tetrachloroethylene on neurotransmitter levels in the brain was explored in
male S-D rats exposed continuously to 200, 400, or 800 ppm tetrachloroethylene for a month
(Honmaetal., 1980a: Honmaetal., 1980b). The 800 ppm dose produced a decrease in ACh in
striatum, and there was a dose-related increase in a peak containing glutamine, threonine, and
serine in whole brain preparations. y-Aminobutyric acid (GABA), NE, 5-HT, and other amino
acids were not affected.
In a study from the same laboratory, Mongolian gerbils of both sexes were exposed to 60
or 300 ppm tetrachloroethylene for 3 months, followed by a 4-month solvent-free period.
Changes in both S-100 and DNA concentrations in various brain regions were observed at the
4-48
-------
higher concentration, and decreased DNA in the frontal cortex was observed after exposure to 60
ppm. The higher concentration also produced decreased brain—but not body—weight. The
results at 60 ppm were replicated in a follow-up study (Karlsson et al., 1987).
In a related study (Briving et al., 1986), Mongolian gerbils were exposed to
tetrachloroethylene at 120 ppm for 12 months. At the end of exposure, out of a total of 8 amino
acids assayed, taurine was significantly decreased in the two brain regions assessed
(hippocampus and cerebellum), and glutamine was elevated in the hippocampus. GABA levels
were unaffected, as was uptake of GABA and glutamate.
Kyrklund and Haglid (1991) exposed pregnant guinea pigs to airborne
tetrachloroethylene continuously from GD 33 through GD 65. The exposure was continuous at
160 ppm except for 4 days at the beginning and end of the exposure period, when it was reduced
to 80 ppm. In the control group, there were three dams with litter sizes of four, three, and two
pups, and in the exposed group, there were three dams with litter sizes of two each. The pup
body weights differed between litters. According to the study authors' analysis, the offspring
had a slightly altered brain fatty acid composition, with a statistically significant reduced stearic
acid content in the tetrachloroethylene treatment group, which is consistent with the study
authors' earlier findings in rats. The statistical analysis, however, relied on pups as the
experimental unit rather than the litters, so the ^-values were likely underestimated. The results
also suggested that tetrachloroethylene reduced the litter size, but a much larger study would be
necessary to establish reduced litter size because the effect of tetrachloroethylene in this study
was relatively small, and the reduction was not statistically significant.
Caucasian male and female NMRI mice were exposed to 9, 37, 75, or 150 ppm
tetrachloroethylene continuously for 30 days, to 150 ppm tetrachloroethylene for one of several
exposure periods ranging from 5-30 days, or to 150 ppm tetrachloroethylene for 30 days with
various recovery periods (Kj ell strand et al., 1984). Other groups were exposed intermittently on
several dosing and exposure regimens, which resulted in a TWA of 150 ppm for 30 days. Motor
activity was assessed following exposure. All concentrations of intermittent exposure increased
motor activity. Results of motor activity following continuous exposure were not reported.
4.1.2.2. Oral and Intraperitoneal Studies
Table 4-4 presents a summary of the oral neurotoxicity animal studies, which are
described in greater detail in the sections that follow. For the six oral neurotoxicity studies in
rodents reviewed here, only one (Fredriksson et al., 1993) describes effects lasting more than 1
week. In that study, the effect (increased motor activity) was the same at 5 and 320 mg/kg. The
lowest LOAEL occurring in the four remaining studies is 100 mg/kg for delayed onset of
circadian activity in rats (Motohashi et al., 1993). This LOAEL is based on an i.p.-administered
4-49
-------
dose describing transient neurological effects and is not comparable to inhalation or ingestion
LOAELs without pharmacokinetic modeling of an appropriate dose metric. No information is
available for irreversible neurological effects via the oral route because no studies have evaluated
the potential for neurotoxicity following chronic oral exposure.
Table 4-4. Summary of oral neurotoxicity animal studies
Subjects
Effect
NOAEL/LOAEL" (mg/kg)
Reference
Neurobehavioral studies
S-D rats, male
9/dose
S-D rats, male,
8-10/dose
ICR mice, male
8-10/dose
F344 rats,
female
w/dose
Wistar rats, male
w/dose
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 min
Brain PCE concentrations were the same
at both doses
NOAEL/LOAEL:
| righting reflex, 2,000/4,000
Impaired balance, 1,000/2,000
Operant responses, 1,000/2,000
I punishment responses, 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
Transient delay in circadian activity,
dose-related
Daily dose for 8 wk: 5, 50 mg/kg
Gavage single dose: 0, 160^4Jfl
mg/kg
Single i.p. doses: 0, 500, 1.000. 2.000.
4,000 mg/kg
Single doses: 150 mg/kg is LOAEL
Repeated dosing for 14 d: 1.500
mg/kg is NOAEL
i.p. doses: 0, 100, 500, 1,000 mg/kg-
day for 3 d
Chen et al.
(2002b)
Warren et al.
(1996)
Umezu et al.
(1997)
Moser et al.
(1995)
Motohashi et
al. (1993)
Developmental neurotoxicity study
NMRImale
mice,
postnatal
exposure
12 pups/dose
(derived from 3
litters)
Increased locomotion and decreased
rearing at Day 60 in both dose groups
No effect immediately after treatment
Gavage treatment: 5, 320 mg/kg daily
for PNDs 10-16
Fredriksson et
al. (1993)
a Experimental/observational NOAEL is underlined; LOAEL is double-underlined.
n/dose = Number of animals per dose not clearly defined
4-50
-------
4.1.2.2.1. Neurobehavior
A study in male S-D rats assessed the acute or short-term effects of tetrachloroethylene
by gavage on several screening tests (Chen et al., 2002a). A single dose of 500 mg/kg in adult
rats produced changes on three different tests of pain threshold, locomotor activity, and seizure
susceptibility threshold following pentylenetetrazol infusion, whereas 50 mg/kg resulted in
statistically significant effects only on seizure threshold. In the short-term study, young, 45-50 g
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 required scoring by an observer. The study by Chen et al.
(2002a) does not state whether the observer(s) was blind to the treatment group of the animals, a
condition that is essential for such tests to be valid. Differences in body weight between control
and treated rats add potential bias. Further, the paper does not state whether all animals were
tested by the same person for each task or, if not, whether there was any indication of
interobserver correlation. The potential effect of the difference in weight between the control
and the treated groups on these measures is also unknown. Given that the difference between the
control and the treated groups in response latency to painful stimuli is tenths or hundredths of a
second with no dose response, these issues are of serious concern.
Various behavioral endpoints were assessed in 8-week-old ICR male mice at the
beginning of an experiment by Umezu et al. (1997). Righting reflex was affected after single-
dose i.p. administration of tetrachloroethylene at 4,000 but not at 2,000 mg/kg or less, and ability
to balance on a wooden rod was decreased at 2,000 but not at 1,000 mg/kg or less. Response rate
on a fixed-ratio 20 (FR20) schedule, which requires 20 responses for each reinforcement, was
affected at 2,000 but not at 1,000 mg/kg or less, 30 minutes after administration. In a procedure
in which a thirsty mouse was shocked every 20th lick of a water spout, mice dosed with
500 mg/kg—but not with higher or lower doses—received an increased number of shocks. In an
FR20-FR20 punishment schedule, response in the punishment condition was increased at
1,000 but not at 500 mg/kg or less. A puzzling aspect of the study is the mention in the methods
section of "breeding animals," with no further explanation. If the investigators bred their own
mice, there is no indication of how pups were assigned to treatment groups.
Moser et al. (1995) examined the effects of a number of potentially neurotoxic agents,
including tetrachloroethylene, on a neurotoxicity screening battery in adult female F344 rats
following either a single gavage dose (acute exposure) or repeated gavage doses over 14 days
(subacute exposure). For the acute study, subjects were tested 4 and 24 hours following
exposure. After acute exposure, a LOAEL of 150 mg/kg was identified for increased reactivity
4-51
-------
to being handled 4 hours after dosing, with increased lacrimation, decreased motor activity,
abnormal gait, decreased response to an auditory stimulus, decreased righting ability, and
increased landing foot splay at higher doses at 4 and/or 24 hours postdosing. A NOAEL was not
identified. In the subacute study, no endpoint was significantly different from those of controls
at doses of 50-1,500 mg/kg. This presumably represents behavioral acclimation following
repeated exposure to tetrachloroethylene.
Locomotor activity was monitored in NMRI mice gavaged with 5 or 320 mg/kg
tetrachloroethylene for 7 days beginning at 10 days of age (Fredriksson et al., 1993). Twelve
male pups from three or four litters were assigned to each treatment group. Locomotion, rearing,
and total activity (vibration of the cage) were measured for 60 minutes at 17 and 60 days of age.
A statistically significant increase in locomotor activity and total activity of treated mice in both
dose groups was observed, and rearing behavior decreased as compared with controls for all
three evaluations at 60 days of age, but not at 17 days of age when testing followed shortly after
the last dose. Litter mates were used as independent observations in the statistical analysis,
which tends to underestimates-values and thereby overstate statistical significance [i.e., Holson
(1992): Buelke-Sam (1985)1. However, the magnitude of the effects seen, more than a twofold
increase in locomotion and total activity by the end of the Day 60 evaluation period, and the
persistent effects of subacute developmental exposures in this study raise concern. Locomotor
activity was assessed in 6-week-old male Wister rats following i.p. doses of 100, 500, or 1,000
mg/kg tetrachloroethylene for 3 consecutive days, with activity being monitored for at least 1
week following cessation of administration (Motohashi etal., 1993). Animals were monitored
24 hours/day, and locomotor activity (measured as change in electrical capacitance of a circuit
beneath the floor of the cage) was analyzed by time-series analysis and spectral analysis. All
doses of tetrachloroethylene changed circadian rhythm in a dose-dependent manner, with the
increased activity at the start of the dark period delayed by tetrachloroethylene exposure.
Recovery took 3-5 days after cessation of exposure.
Operant performance on a fixed-ratio 40 schedule of reinforcement was assessed in adult
male S-D rats gavaged with 160 or 480 mg/kg tetrachloroethylene immediately before testing
(Warren et al., 1996). The lower dose produced no effect on response rate over the 90 minute
session, whereas the higher dose produced a transient rate decrease in three of six animals (with
recovery after 20 to 40 minutes) and induced a complete cessation of response in two of the six
animals. Tetrachloroethylene concentrations increased rapidly after administration in blood,
brain, fat, liver, and muscle. For the duration of the 90 minute period of testing, blood
tetrachloroethylene levels were approximately linearly related to the administered dose, but brain
tetrachloroethylene levels were similar for both dose groups. This study did not evaluate the
persistent effects of exposure to tetrachloroethylene on cognitive performance.
4-52
-------
4.1.2.2.2. Developmental neurotoxicity
Evidence of potential developmental neurotoxicity was reported by Fredriksson et al.
(1993). In this study (refer to Section 4.1.2.2), tetrachloroethylene was administered to male
NMRI mice by gavage at dose levels of 0, 5, or 320 mg/kg-day on PNDs 10-16. At PNDs 17
and 60, spontaneous activity (locomotion, rearing, and total activity) was measured over three,
20 minute periods. No treatment-related alterations in activity were observed at 17 days of age;
however, at 60 days of age, all three measures of spontaneous activity were altered.
4.1.3. Mode of Action (MOA) for Neurotoxic Effects
The MOA for the neurotoxic effects of tetrachloroethylene is unknown; however, at
present, the best surrogate for the dose metric for neurotoxicity is blood tetrachloroethylene. The
primary neurobehavioral changes that are observed following tetrachloroethylene exposure are
visual changes, cognitive deficits, and increased reaction time. It is not clear if there are multiple
mechanisms resulting in these outcomes. Additionally, there may be multiple mechanisms or
MO As, which may differ for adult and developmental exposure. The acute effects of
tetrachloroethylene appear to share much in common with those of other chlorinated solvents
such as trichloroethylene and dichloromethane as well as toluene, volatile anesthetics, and
alcohols. It is unknown how these different neurological effects are induced, but there are data
available to help elucidate what areas in the brain and specific molecular targets may be involved
in the resulting neurotoxicological outcome.
Neuropathology and mechanistic studies have been conducted in animal models (rats,
mice, gerbils) to determine how tetrachloroethylene may be producing the observed neurological
effects. Changes in fatty acid composition of the brain following a 30 or 90 day exposure have
been reported, and these changes persist for up to 30 days after the cessation of exposure
(Kyrklund et al.. 1990. 1988. 1987: Kyrklund et al.. 1984). Studies that examined the entire
brains of animals reported decreases in astroglial proteins (GFAP and S-100), decreased brain
RNA content, and decreased levels of glutamine, threonine, and serine (Wang et al., 1993;
Kyrklund et al.. 1990. 1988. 1987: Rosengren et al.. 1986: Kyrklund et al.. 1984: Honma et al..
1980a; Honma et al., 1980b; Savolainen et al., 1977a). Brain regions examined following
tetrachloroethylene exposure included the frontal cortex, the hippocampus, the striatum, and the
cerebellum (Wangetal.. 1993: Karlsson et al.. 1987: Briving et al.. 1986: Kyrklund et al.. 1984:
Honma etal., 1980a; Honma et al., 1980b). Notable changes include decreased DNA content in
the frontal cortex following continuous exposure of 600 ppm for 4 weeks in rats (Wang et al.,
1993) or a 60 ppm exposure for 3 months in Mongolian gerbils (Karlsson et al., 1987).
Decreased taurine levels were noted in both the cerebellum and hippocampus following a
12-month exposure to 120 ppm tetrachloroethylene in Mongolian gerbils, but there were no
4-53
-------
changes in GABA levels or uptake (Briving et al., 1986). Decreased acetylcholine levels in the
striatum were noted in male rats exposed to 800 ppm for 1 month (Honma et al., 1980a: Honma
etal.. 1980b).
Voltage and ligand-gated ion channels have been implicated in many neurological
functions and have been studied as potential neurological targets for tetrachloroethylene and
other structurally related chlorinated solvents (e.g., trichloroethylene, 1,1,1-trichloroethane,
dichloromethane). Table 4-5 summarizes the available in vitro mechanistic studies with
chlorinated solvents. Tetrachloroethylene has been demonstrated to inhibit calcium channel
function (Shafer et al., 2005) and the neuronal nicotinic acetylcholine receptor (Bale et al.,
2005). Based on the structural similarity of tetrachloroethylene to other chlorinated solvents as
well as the similar neurobehavioral and mechanistic findings, it is likely that tetrachloroethylene
also interacts with the other listed targets in Table 4-5. This solvent class has also been shown to
interact with ion channels such as the GABAA and glycine receptors (Lopreato et al., 2003;
Beckstead et al., 2000; Krasowski and Harrison, 2000). Overall, these solvents appear to
potentiate the function of inhibitory receptors and inhibit the function of excitatory receptors
[refer to Bushnell et al. (2007) and Bowen et al. (2006) for a review]. Additionally, this class of
solvents blocks sodium channel (Haydon and Urban, 1983; Shrivastav et al., 1976) and voltage
sensitive calcium channel function (Shafer et al., 2005) when the membrane is held at or near the
resting membrane potential.
Based on these findings as well as other mechanistic studies conducted with
tetrachloroethylene, some neurotransmitter systems may be more favorably involved in
neurotoxicological outcomes than others. Also, based on the number of reported molecular
targets, it is more likely that there are several plausible mechanisms responsible for the resultant
neurotoxicological outcome, and those potential mechanisms (as well as a discussion of
plausibility) are summarized below by the major observed outcomes (visual changes, cognitive
deficits, increased reaction time).
4-54
-------
Table 4-5. Summary of in vitro ion channel effects of tetrachloroethylene
and other chlorinated solvents
Reference
Cellular system
Ion channel/receptor
Concentration
Effects
Tetrachloroethylene
Shafer et al.
(2005)
Bale et al.
(2005)
PC12 cells, primary
cortical neurons
Xenopus oocytes
Voltage Sensitive Calcium
Channels (VSCCs)
Human and rat a4p2, o3p2,
and a? nicotinic acetylcholine
receptors
0-325 uM
0-65 uM
Shift of VSCC activation to a
more hyperpolarizing potential.
Inhibition of VSCCs at a
holding potential of -70 mV
Inhibition of nicotinic
acetylcholine receptor function
Dichloromethane
Hay don and
Urban (1983)
Squid giant axon
Sodium channels
0, 15, 25 mM
Inhibition of inward sodium
channel currents
Trichloroethylene
Shafer et al.
(2005)
Beckstead et
al. (2000)
Lopreato et
al. (2003)
Krasowski
and Harrison
(2000)
Shrivastav et
al. (1976)
PC12 cells, primary
cortical neurons
Xenopus oocytes
Xenopus oocytes
Human embryonic
kidney, 293 cells
Squid giant axon
VSCCs
Human recombinant
glycine receptor al, GABAA
receptors, alpl, alp2y2L
Human recombinant serotonin
3 A receptor
Human recombinant glycine
receptor al,
GAB AA receptors a2pl
Sodium channels
0-2,100 uM
0, 390 uM
0, 390 uM
Not provided
5-80%
saturation
Shift of VSCC activation to a
more hyperpolarizing potential.
Inhibition of VSCCs at a
holding potential of -70 mV
50% potentiation of the
GABAA receptors; 100%
potentiation of the glycine
receptor
Potentiation of serotonin
receptor function
Potentiation of glycine receptor
function with an EC50 of 0.65 ±
0.05 mM.
Potentiation of GAB AA
receptor function with an EC50
of 0.85 ± 0.2 mM
Shift of sodium channel
activation to a more
hyperpolarizing potential.
Inhibition of inward sodium
channel current at -70 mV
1,1,1 -Trichloroethane
Cruz et al.
(2000)
Beckstead et
al. (2000)
Beckstead et
al. (2000)
Xenopus oocytes
Xenopus oocytes
Rat hippocampal
slices
NMDA-glutamate receptor
NR1/2A, NR1/2B
Human recombinant
glycine receptor al, GABAA
receptors, alpl, alp2y2L
GABAA receptor
0.1-10 mM
0.39 mM
0.28 mM
Inhibition of NMDA-glutamate
receptor function
Potentiation of GAB AA and
glycine receptor function
Reversible increase in GABAA-
mediated inhibitory
postsynaptic currents (IPSCs)
4-55
-------
4.1.3.1. Visual Function Domain
Although tetrachloroethylene produces changes in visual evoked potentials, there are no
associated mechanistic studies to indicate what receptor systems may be involved. However,
there is a characterization study evaluating the contribution of specific ligand-gated ion channels
(GABAA, NMDA-glutamate, nicotinic acetylcholine receptors) to the generation of the steady
state visual evoked potential (Bale et al., 2005). The findings suggest that ion channels are
involved in visual function and, specifically, the measured evoked potentials. The only
administered drugs resulting in an effect similar to tetrachloroethylene were NMDA
(NMDA-glutamate receptor agonist) and mecamylamine (nAChR antagonist). Therefore, the
NMDA-glutamate and the nicotinic acetylcholine receptor systems may be more closely
involved in the visual evoked potential changes resulting from solvent exposure.
With respect to the impact on color vision and visual contrast sensitivity following
tetrachloroethylene exposure, the mechanisms behind these effects are unknown. These visual
changes occur at exposures that are lower than the visual evoked potential changes. Cones at the
level of the retina process color vision, and there may be a change in the function and/or
signaling of the retina to the visual center in the CNS. In visual contrast sensitivity, retinal
ganglion cells have been implicated as a sensitive target in processing changes in contrast
(Beaudoin et al., 2008). The available literature suggests that NMDA-glutamate receptors
(Manookin et al., 2010) and calcium channels (Hu et al., 2009) may be involved in visual
contrast sensitivity changes. It is known that tetrachloroethylene exposure affects calcium
channel function in vitro (Shafer et al., 2005), and a related chlorinated solvent,
1,1,1-trichloroethane, has been demonstrated to modulate NMDA-glutamate receptor function
(Cruz et al.. 2000).
4.1.3.2. Cognitive Domain
The hippocampus is involved in cognitive functions, but only changes in taurine levels
were observed in this brain region following tetrachloroethylene exposure in gerbils (Briving et
al., 1986). It was demonstrated that tetrachloroethylene inhibits both human and rat recombinant
nicotinic acetylcholine receptors in vitro (Bale et al., 2005), and perhaps this finding may help
explain why cognitive changes are observed with tetrachloroethylene exposure. However, more
studies need to be conducted with tetrachloroethylene exposure and perhaps incorporating a
challenge with nicotinic agonists and antagonists to determine the involvement of nicotinic
acetylcholine receptors in cognitive function.
4-56
-------
4.1.3.3. Motor Activity Domain - Reaction Time
Reaction time is a general measure of CNS function. With increased reaction time, it can
be surmised that there is a general CNS decrease in movement. Currently, there are no available
mechanistic studies with tetrachloroethylene that have evaluated neurological systems mediating
reaction time activity. There is one study that has reported that decreased CNS function
(anxiolytic profile) observed with tetrachloroethylene may be due to site-specific action on the
GABAA receptors. Chen et al. (2002a) pretreated rats with tetrachloroethylene (50 or 500
mg/kg, oral gavage), and this pretreatment, following both an acute and a subchronic (5 or
50 mg/kg-day, 5 days/week, 8 weeks) schedule, significantly increased the seizure threshold
when challenged with pentylenetetrazole (PTZ), a convulsant that blocks GABAA receptors.
This study suggests that the GABAergic system may be involved in the anxiolytic and general
CNS depressive behavior that is observed following tetrachloroethylene exposure and could be
potentially related to observed increased reaction times in the various tasks.
4.1.4. Summary of Neurotoxic Effects in Humans and Animals
Human and animal studies provide complementary evidence regarding the association of
neurobehavioral deficits and tetrachloroethylene exposure. Tetrachloroethylene exposure in
humans has primarily been shown to affect visual function (including color vision) and
visuospatial memory and other aspects of cognition. Brain-weight changes have been measured
in animal studies. A more in-depth discussion of the human neurotoxicological studies can be
found in Section 4.1.1.3, and the animal inhalation and oral or i.p. exposure studies are discussed
in Sections 4.1.2.1 and 4.1.2.2, respectively.
Visual contrast sensitivity deficits as well as color discrimination deficits are commonly
present prior to detectable pathology in the retina or optic nerve head, making them one of the
earliest signs of disease and potentially more sensitive measures than evoked potentials from
visual stimuli (Regan, 1989). Several independent lines of evidence can be found in the
occupational and residential exposure studies to support an inference of visual deficits following
chronic tetrachloroethylene exposure. The studies that observed effects on color vision using the
Lanthony D-15 color vision test include cross-sectional and longitudinal designs in dry-cleaning
settings (Gobba et al., 1998; Cavalleri et al., 1994) and residential studies (Schreiber et al.,
2002). Decrements in color confusion were reported among all workers exposed to a mean
TWA of 6 ppm for an average of 8.8 years (Cavalleri et al., 1994). A significant dose-response
relationship between CCI value and tetrachloroethylene concentration (r = 0.52,p < 0.01) was
also observed in Cavalleri et al. (1994). As noted previously, the color vision testing in this
study was blinded to exposure level of the study participants, and the study participants were
well matched in terms of age, smoking, and alcohol use. A follow-up of these workers 2 years
4-57
-------
later (Gobbaet al., 1998) showed greater loss in color discrimination in those who were
subsequently exposed to a higher concentration (increase in geometric mean from 1.7 to
4.3 ppm), with no change in those exposed to lower concentrations (decrease in geometric mean
from 2.9 to 0.7 ppm). Although Gobba et al. (1998) demonstrates persistent color confusion
effects in this follow-up evaluation, the study exposures are not clearly characterized over the
course of the 2-year duration. Nakatsuka et al. (1992) did not observe an association with color
vision among dry cleaners in China (n = 64, geometric mean TWA: 11 and 15 ppm in females
and males, respectively), but the relative insensitivity of the specific type of color vision test
used in this study (Lanthony, 1978) is a likely explanation for these results. Effects on color
vision were also observed among 14 dry cleaners in the small study in Malaysia by
Sharanjeet-Kaur et al. (2004), but this study provides little weight to the strength of the evidence
because of the lack of exposure information (other than job title), and differences between dry
cleaners and controls regarding test conditions and smoking habits. Two other small studies also
reported lower scores on the Lanthony D-15 color vision test in much lower exposure settings,
but the differences were not statistically significant: in a study of residents living above dry
cleaners (mean tetrachloroethylene exposure during active dry cleaning = 0.4 ppm), the mean
CCI scores were 1.33 and 1.20 in 17 exposed and 17 controls, respectively (p = 0.26); in a study
of workers in a day-care center located in a building with a dry-cleaning business (mean
tetrachloroethylene exposure 0.32 ppm), the mean CCI scores were 1.22 and 1.18 in the exposed
day-care workers and controls, respectively (p = 0.39) (Schreiber et al., 2002). Overall, the
evidence reveals a high degree of consistency in this aspect of visually mediated function.
Visual contrast sensitivity changes were reported in two NYSDOH residential studies. In
a small pilot study (4 children and 13 adults), mean scores for visual contrast sensitivity (using a
near vision visual contrast sensitivity test) across spatial frequencies were statistically
significantly lower in exposed residents than in controls, indicating poorer visual function in the
exposed groups (Schreiber et al., 2002). Controls were age- and sex-matched to the exposed
group, and both groups were English speaking and predominately Caucasian ethnicity; however,
they were drawn from different geographic areas. In addition, two of the four exposed children
had diagnoses of learning disabilities or developmental delays, which could affect performance
on this type of test. In the larger study (NYSDOH. 2010. 2005a, b), the test (Functional Acuity
Contrast Test, FACT) assessed far vision visual contrast sensitivity, and the test had a low rate of
detecting visual contrast changes. For contrast vision, a number of analyses in NYSDOH (Storm
etal.. 2011 [previously reported in NYSDOH. 20101: NYSDOH. 2005a) suggest a vulnerability
among children. However, exposure to >0.015 ppm (>100 |ig/m3) tetrachloroethylene was
highly correlated with race and children's age, and the sample sizes in the highest exposure
group, especially in higher income, nonminority groups, make it difficult to fully examine
4-58
-------
possible effects of income, race, and age on vision. Therefore, while both studies report visual
contrast sensitivity changes with exposed children being more sensitive, there are concerns with
the methodological and analytic approaches in these studies.
Acute human exposure studies reported increased latencies of up to 3.0 ms in visual
evoked potentials (Altmann et al., 1990) and changes in EEGs (magnitude of effect was not
specified) (Hake and Stewart, 1977; Stewart et al., 1970) at higher exposures ranging from 340
to 680 mg/m3.
In rats, acute inhalation exposure to tetrachloroethylene results in significant changes to
the flash-evoked potential at 800 ppm (Mattsson et al., 1998), and a decrease in F2 amplitudes of
the steady state visual evoked potential at 250 ppm (Boyes et al., 2009). In a subchronic
exposure study (13 weeks, up to 800 ppm tetrachloroethylene), changes in flash-evoked potential
responses were not observed at tetrachloroethylene exposures up to 200 ppm. In the 800 ppm
group, there was a significant increase in the amplitude and a significant increase in latency
(-3.0 ms) of the mid-flash-evoked potential waveform (N3), but histopathological lesions were
not observed in the examination of the visual system brain structures [e.g., visual cortex; optic
nerve; Mattsson et al. (1998)1
Effects on visuospatial memory in humans were also reported in each of the studies that
examined this measure (Altmann et al., 1995; Echeverria et al., 1995; Echeverria et al., 1994;
Seeber, 1989). These effects (increased response times or cognition errors) were observed in
occupational and residential studies, and the occupational studies were quite large, involving
101, 65, and 173 dry-cleaning workers in Seeber (1989), Echeverria et al. (1995), and Echeverria
et al. (1994), respectively. Several different types of tests were used including digit reproduction
(Seeber, 1989), switching, pattern memory, and pattern recognition (Echeverria et al., 1995;
Echeverria et al., 1994), and the Benton test (Altmann et al., 1995). Exposures for the increased
reaction time observations (LOAELs) ranged from 4.99 to 102 mg/m3 (Altmann et al., 1995;
Echeverria et al., 1995; Ferroni etal., 1992). The changes in the cognitive tasks were observed
at exposures (LOAELs) ranging from 53.9 to 364.22 mg/m3 (Spinatonda et al., 1997; Echeverria
etal., 1995; Seeber, 1989). All of these studies except Altmann et al. (1995) indicate that the
neurobehavioral assessment was blinded to knowledge of the exposure level of the subject, and
all of the studies adjusted for potentially confounding factors. It should be noted, however, that
residual confounding from education level differences between exposed and referent subjects
may still be present in Altmann et al. (1995).
Changes in the motor activity domain as measured by increased reaction time, increased
number of false alarms, and decreased trial completions in a signal detection task (measures of
decreased attention) were reported in an acute (60 minutes) exposure (6,782 mg/m3 or higher)
study in rats (Oshiro et al., 2008). Additionally, operant tasks that test cognitive performance
4-59
-------
have demonstrated deficits in rats and mice following acute tetrachloroethylene oral (Warren et
al., 1996) and i.p. (Umezu et al., 1997) exposures. These findings are consistent with observed
effects on cognition and memory in humans. However, no studies, to date, have evaluated the
persistent effects of tetrachloroethylene exposure on cognitive performance deficits in animal
models.
An occupational exposure study (n = 60) (Ferroni et al., 1992) and a residential exposure
study (n = 14) (Altmann et al., 1995), with mean exposure levels of 15 ppm and 0.7 ppm,
respectively, reported significant increases in simple reaction time of 24 ms (11%) (Ferroni et al.,
1992) and 40 and 51.1 ms (15 and 20% increases, respectively, for two separate measurements)
(Altmann et al., 1995) for the exposed subjects. A third study, Lauwerys et al. (1983), reported
better performance on simple reaction time in 21 exposed workers (mean TWA: 21 ppm)
compared with controls measured before a work shift but not after.
The changes in brain weight, DNA/RNA, and neurotransmitter levels that were observed
in the animal studies are highly supportive of the neurobehavioral changes observed with
tetrachloroethylene exposure. Changes in brain DNA, RNA, or protein levels and lipid
composition were altered following inhalation, with changes observed in the cerebellum, the
hippocampus, and the frontal cortex (Wang etal., 1993; Rosengren et al., 1986; Savolainen et al.,
1977a: Savolainen et al., 1977b). The replication of these changes in biochemical parameters
and effects in brain weight in both rats and gerbils is pathognomonic. Changes in
neurotransmitters systems (Briving et al., 1986; Honmaetal., 1980a: Honmaetal., 1980b) and
circadian rhythm (Motohashi etal., 1993) in animal studies are consistent with neuroendocrine
alterations observed in humans (Ferroni et al., 1992).
In conclusion, the weight of evidence across the available studies of humans and animals
exposed to tetrachloroethylene indicates that chronic exposure to tetrachloroethylene can result
in decrements in color vision, visuospatial memory, and possibly other aspects of cognition and
neuropsychological function, including reaction time.
4.2. KIDNEY AND BLADDER TOXICITY AND CANCER
4.2.1. Human Studies
4.2.1.1. Kidney Toxicity in Humans
High concentrations of inhaled tetrachloroethylene given acutely as an anesthetic are
associated with symptoms of renal dysfunction, including proteinuria and hematuria (ATSDR,
1997a: Hake and Stewart, 1977). Controlled inhalation exposure to tetrachloroethylene at levels
of 0, 20, 100, or 150 ppm for up to 11 weeks did not affect a number of urine parameters or
blood urea nitrogen (BUN) (a measure of kidney function) in 12 healthy individuals [Stewart et
4-60
-------
al. (1977), as reported in ATSDR (1997b)]. Whether renal effects would occur from these acute
exposure levels in a larger, more diverse population than the one studied by Stewart et al. (1977)
is not known.
The evidence for kidney effects from chronic inhalation of tetrachloroethylene is limited
to studies of urinary renal proteins as indicator of kidney function. One study has become
available on end stage renal disease (ESRD) incidence in a cohort of dry cleaners (Calvert et al.,
2011). The ATSDR (Lybarger et al., 1999: ATSDR, 1998a) recommends a standard battery of
kidney function tests including serum creatinine, urinalysis with microscopic examination of
urine sediment, albumin, retinol binding protein (RBP), 7V-acetyl-p-D-glucosaminidase (NAG),
alanine aminopeptidase (AAP), osmolality, and urine creatinine (Lybarger et al., 1999). These
indicators evaluate a range of toxicity, from effects on general kidney function to effects on
specific segments of the nephron. For example, the overall integrity of the nephron can be
evaluated from the urinalysis, and albumin is an indicator of the integrity of the glomerulus;
three indicators—RBP, NAG, and AAP—assess damage to the proximal tubules, although it
should be noted that NAG is not a sensitive and specific marker of tubular dysfunction (Lybarger
et al., 1999). The proximal tubules house p-lyase enzymes and are hypothesized to be a target of
tetrachloroethylene toxicity due to the bioactivation of reactive metabolites produced from the
further metabolism of TCVC (refer to Section 3). For this reason, altered urinary indicators of
proximal tubule function are consistent with knowledge of metabolic processing.
The epidemiologic studies are suggestive of subtle damage to the renal tubules.
Table 4-6 summarizes the human kidney function studies. Five studies (Trevisan et al., 2000;
Verplanke et al., 1999: Mutti et al., 1992: Sol et and Robins, 1991: Lauwerys et al., 1983) have
examined the three core indicators of tubule function—RBP, NAG, or AAP—in urine of dry
cleaners. Three studies measured RBP, with two of the studies reporting a statistically
significant elevated prevalence of abnormal values among study participants (Mutti et al., 1992)
or a statistically significant elevated geometric mean concentration of RBP (Verplanke et al.,
1999) for tetrachloroethylene-exposed workers as compared with controls. The mean
concentration of RBP for exposed subjects (75.4 ug/g creatinine) in the Verplanke et al. (1999)
study is within a normal range.16
16 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, respectively, in 118 subjects.
4-61
-------
Table 4-6. Summary of human kidney toxicity marker studies of
occupational exposures to dry-cleaning facilities using tetrachloroethylene
Subjects, methods
Exposure levels
Results
Reference(s)
Occupational exposures: dry-cleaning settings
Belgium, 26 dry cleaners,
33 unexposed workers
(controls), B, EA, PA, U
[before and after shift]
Italy, 57 dry cleaners
(mostly females) (Group
1), 188 painters (mostly
males) (Group 2), 5 1
glass-fiber reinforced boat
workers (Group 3), 212
workers exposed to C5-C7
alkanes (Group 4), 30
unexposed workers
(mostly females) (Control
Group 1) and 81
unexposed workers
(mostly males) (Control
Group 2). U [before and
after shift]
Czech Republic, 22 female
dry cleaners, 15 female
controls (clerical workers).
PA, U [end of shift]
United State, 192 dry
cleaners (mostly females),
no controls. PA, U
[collection time varied by
subject]
Italy, 50 dry cleaners and
ironers (mostly females),
50 controls (blood donors).
B, PA, U [before shift]
Mean TWA = 21
ppm, U-TCA = ND,
mean duration =
6.4 yr
Dry cleaners
(Group 1): mean
TWA = 10 ppm
(extrapolated from
postshift U-TCA
according to Ikeda et
al. (1972). mean
duration = 13.9yr
3 shops with mean
TWA<12ppm, 2
shops with mean
TWA 42 ppm and
47 ppm, mean
duration = 1 1 yr
Mean TWA = 7 ppm,
mean duration =
11.6 yr
Mean TWA = 8.8
ppm, mean duration =
10 yr
No differences in creatinine-adjusted urinary
P2u-globulin, retinol-binding protein and
albumin.
50% increase in creatinine-adjusted geometric
mean concentration of urinary p2-glucuronidase
and 100% increase in geometric mean urinary
lysozyme in dry cleaners compared to either
control group. No difference in total protein or
albumin.
Fourfold elevation in geometric mean creatinine-
adjusted urinary concentration of lysozyme. No
difference in albumin, P2u-globulin and total
protein, or prevalence of subjects whose urinary
proteins above 95th percentile.
No correlation of exposure and creatinine-
adjusted urinary protein, albumin, or jV-acetyl-
p-glucuronidase.
1.5- to 4-fold increase in creatinine-adjusted
mean concentration of 8 urinary renal proteins
(albumin, transferrin, 3 brush border antigens,
tissue nonspecific alkaline phosphatase,
p < 0.05; glycosaminoglycans, Tamm-Horsfall
glycoprotein, p = 0.06) and 2 serum proteins
(anti-glomerular basement membrane, laminin
fragments, p < 0.05) in dry cleaners;
discriminated between dry cleaners and matched
controls (p < 0.05). No difference in 12 other
urinary renal proteins (includes total protein and
7V-acety 1-p-glucuronidase) .
Lauwerys et
al. (1983)
Franchini et
al. (1983)
Vyskocil et
al. (1990)
Solet and
Robins
(1991)
Mutti et al.
(1992)
4-62
-------
Table 4-6. Summary of human kidney toxicity marker studies of
occupational exposures to dry-cleaning facilities using tetrachloroethylene
(continued)
Subjects, methods
Italy, 40 female dry
cleaners, 45 female
controls (ironers). PA, B,
U [before and after shift]
The Netherlands, 101 dry
cleaners (mostly males),
19 controls (seamstresses,
sorters or folders in dry-
cleaning shops or laundry
workers) (mostly females).
PA, U [before shift]
Exposure levels
Mean TWA =14.8
ppm, mean duration =
15 yr
Mean TWA = 8 ppm
(dry cleaners), <2.2
ppm (controls), mean
duration = 3.9 yr
Results
Positive correlation between preshift urinary
PCE and total solutes and total proteins (p <
0.01) and postshift urinary PCE and glutamine
synthetase (p< 0.001). No difference in
creatinine-adjusted mean urinary concentration
of total solutes, total protein angiotensin
converting enzyme, /V-acetyl-p-glucuronidase, or
glutamine synthetase.
Retinol binding protein (creatinine-adjusted
mean concentration) elevated twofold among
dry cleaners (p = 0.01). No difference in
creatinine-adjusted mean of p-galactosidase, N-
acetyl-p-glucuronidase, or alanine
aminopeptidase. No difference in geometric
mean albumin or total protein.
Reference(s)
Trevisan et al.
(2000)
Verplanke et
al. (1999)
A = air sample, not specified area or personal sample; AA = area air samples, B = biological monitoring of
blood, BTX = benzene, toluene, xylene, EA = exhaled air samples, ND = not detectable, PA = personal air
samples, PCE = tetrachloroethylene, U = biological monitoring of urine for trichloroacetic acid.
As a comparison, Nomiyama et al. (1992) suggest a critical level of RBG of 200 ug/g
creatinine as indicative of cadmium-induced kidney toxicity. Exposure levels were to a median
of 15 ppm (range: limit of detection to 85 ppm) in Mutti et al. (1992) and 1.2 ppm (range:
0.3-6.5 ppm) in Verplanke et al. (1999). Lauwerys et al. (1983), the only other study to assess
RBP, did not observe any differences in the geometric mean concentration of RBP between dry
cleaners with a mean tetrachloroethylene exposure of 21 ppm and their controls; however, this
study contained fewer exposed subjects with a shorter duration of exposure than did that of Mutti
et al. (1992).
The four studies that measured urinary excretion of NAG (Trevisan et al., 2000;
Verplanke et al.. 1999: Mutti et al.. 1992: Sol et and Robins. 1991) and the one study that
measured AAP (Verplanke et al., 1999) did not observe any differences between exposed
subjects and controls. These findings are not surprising, given the limitations in terms of
sensitivity and specificity of NAG as a marker of tubular dysfunction (Lybarger et al., 1999).
Mean exposures were 14 ppm in Solet and Robins (1991) and 9 ppm in Trevisan et al. (2000):
both studies assessed exposure from personal monitoring of exhaled breath.
4-63
-------
The above findings are further supported by the observation of elevated urinary excretion
of other proteins that are also indicators of damage to the proximal tubules: P2u-globulin,
intestinal alkaline phosphatase (IAP), tissue nonspecific alkaline phosphatase (TNAP),
lysozyme, p2-glucuronidase, and glutamine synthetase. Both IAP and TNAP are indicators of
proximal tubule brush border integrity (Price et al., 1996), whereas lysozyme and P2u-globulin
indicate a failure of the tubule to reabsorb protein (Lybarger et al., 1999; Kok et al., 1998;
Bernard and Lauwerys, 1995). Glutamine synthetase is a mitochondrial enzyme located in the
proximal tubules and has been recently suggested as a marker of tubular damage in rats exposed
to 1,3-hexachlorobutadiene (Trevisan et al., 1999).
Mutti et al. (1992) observed an elevated prevalence of abnormal values for P2u-globulin
and brush border antigens, a higher geometric mean concentration of brush border antigens in
urine, and a higher concentration of TNAP in urine among 50 exposed dry cleaners as compared
with 50 blood donors matched by sex and age with the exposed subjects. Furthermore, markers
of renal damage were highly predictive of exposure status in discriminant analysis.
P2u-Globulin, however, was not elevated among exposed subjects as compared with controls in
the other two studies that examined this protein ("Vyskocil etal., 1990; Lauwerys et al., 1983),
although the mean concentration of P2u-globulin appeared higher in subjects studied by
Vyskocil et al. (1990) than the mean concentration in controls. Both these studies contained
fewer numbers of exposed subjects than did the study by Mutti et al. (1992), and reduced power
as a consequence of fewer subjects may be a reason for the null observations. Further,
tetrachloroethylene exposure appears to affect reabsorption in the renal tubules. Two studies that
assessed lysozyme or p2-glucuronidase observed a statistically significant elevated mean
concentration of these proteins among dry cleaners as compared with controls (Vyskocil et al.,
1990: Franchini et al., 1983).
It is not clear whether tetrachloroethylene exposure affects other parts of the kidney. The
study by Mutti et al. (1992) is suggestive of damage to the glomerulus; however, the lack of an
elevated excretion of albumin, an indicator of glomerular function (Lvbarger et al., 1999), in the
study by Verplanke et al. (1999) argues for further assessment of possible glomerular effects.
Because some albumin is normally filtered, small increases in the amount of albumin in the urine
may result from tubular damage due to failure to reabsorb the small amount filtered (NRC,
2010).
Calvert et al. (2011) examined the incidence of end stage renal disease (ESRD) in a
cohort of 1,704 dry cleaners assembled by the National Institute of Occupational Safety and
Health (NIOSH), 618 who had worked only in a shop where tetrachloroethylene was the primary
cleaning solvent (tetrachloroethylene-only subcohort) and 1,086 who worked in a shop that used
tetrachloroethylene but who also had a history of employment in shops where the primary
4-64
-------
solvent could not be identified (tetrachloroethylene-plus subcohort) (Ruder etal., 2001, 1994).
All subjects alive as of 1977 were linked to the Renal Management Information System (RMIS),
a database of individuals receiving Medicare benefits for ESRD, and followed to 2004. Thirty
cases of ESRD were identified over the 27-year period (standardized incidence ratio [SIR]: 1.34,
95% CI: 0.90, 1.91), with 12 ESRD cases in the tetrachloroethylene-only subcohort (SIR: 1.30,
95% CI: 0.67, 2.26). Of these cases, eight were due to hypertensive ESRD (SIR: 2.66,
95% CI: 1.15, 5.23), of whom six cases were female subjects (SIR: 2.86, 95% CI: 1.05, 6.23).
The observed risk estimate for hypertensive ESRD among tetrachloroethylene-only subjects
appears larger than that for the tetrachloroethylene-plus subcohort (SIR: 1.53, 95% CI: 0.62,
3.16). An exposure-response pattern was further suggested because hypertensive ESRD risk was
highest among those in the tetrachloroethylene-only subcohort employed for >5 years (SIR: 3.39,
95% CI: 1.10, 7.92). These findings support an association between tetrachloroethylene
exposure and ESRD, particularly hypertensive ESRD. ESRD-observed risk is likely
underestimated using RMIS records. An examination of cause of death among cohort subjects
who had died by 2004 found five additional workers with chronic renal failure listed as an
underlying cause of death. Medical records for three of these five deaths indicated two subjects
with ESRD. Calvert et al. (2011), moreover, found substantial underreporting of chronic renal
disease on death certificates, suggesting incidence as superior to mortality for assessing
tetrachloroethylene exposure and kidney disease. Of the 23 deaths among the 30 ESRD subjects,
cause of death on death certificates for 11 of these subjects was due to chronic renal disease and
3 due to "renal disease not otherwise specified."
Taken together, the epidemiologic studies support an association between
tetrachloroethylene and chronic kidney disease, as measured by urinary excretion of renal
proteins and ESRD incidence. The elevated urinary RBP levels observed in two studies
(Verplanke et al., 1999; Mutti etal., 1992) and lysozyme or p2-glucuronidase in Franchini et al.
(1983) provide some evidence for effects to the proximal tubules from tetrachloroethylene
exposure. Effects are observed in populations of both males and females, and potential
differences in susceptibility due to sex-related differences in rates of metabolism (refer to
Section 3) cannot be determined from the available evidence. Median exposure levels in the
studies that observed alterations in renal enzymes were 9 ppm (Trevisan et al., 2000), 10 ppm
(Franchini et al., 1983), and 15 ppm (Mutti etal., 1992), representing LOAELs for these studies.
Only the study by Trevisan et al. (2000) observed an exposure-response relationship, a
correlation between urinary tetrachloroethylene and the concentration of glutamine synthetase (p
< 0.001). None of the other studies reported exposure-response relationships, which is a
limitation on the inference of an association between tetrachloroethylene and renal damage.
However, as pointed out by Mutti et al. (1992), this is a common finding among solvent-exposed
4-65
-------
populations, and inadequate definition of the dose metric most likely contributes to the null
finding. Table 4-6 summarizes the human kidney toxicity studies. Calvert et al. (2011) supports
an association between inhalation tetrachloroethylene exposure and ESRD, particularly
hypertensive ESRD. They observed a twofold elevated incidence (SIR: 2.66, 95% CI: 1.15,
5.23) among subjects who worked only in a shop where tetrachloroethylene was the primary
cleaning solvent compared to that expected based on U.S. population rates. An exposure-
response pattern was further suggested because hypertensive ESRD risk was highest among
those employed for >5 years (SIR: 3.39, 95% CI: 1.10, 7.92).
4.2.1.2. Kidney Cancer in Humans
Twenty-seven epidemiologic studies reporting data on kidney cancer and
tetrachloroethylene exposure were identified. This set of studies includes 13 cohort or nested
case-control studies within a cohort (Selden and Ahlborg, 2011; Calvert et al., 2011; Pukkala et
al.. 2009: Wilson et al.. 2008: Sung et al.. 2007: Ji et al.. 2005b: Blair etal.. 2003: Chang et al..
2003: Travier et al.. 2002: Anderson et al.. 1999: Boiceetal.. 1999: Anttila et al.. 1995: Lynge
and Thygesen, 1990): 11 case-control studies of occupational exposures ( Lynge et al., 2006:
Parent et al.. 2000: Pesch et al.. 2000a: Dosemeci et al.. 1999: Delahunt et al.. 1995: Mandel et
al.. 1995: Schlehofer et al.. 1995: Auperin et al.. 1994: Mellemgaard et al.. 1994: McCredie and
Stewart, 1993: Asal et al., 1988), and 3 studies of residential exposure through contaminated
drinking water (Ma et al., 2009: Vieira et al., 2005: Aschengrau et al., 1993). Some sets of these
studies represent overlapping study populations. For example, three papers examined cancer risk
among occupational groups defined by census data in Sweden (Wilson et al., 2008: Ji and
Hemminki, 2005a: Travier et al., 2002), one paper used a similar design in Denmark (Lynge and
Thygesen, 1990), two papers were based on census data from Sweden, Denmark, Finland, and
Norway (Lynge et al., 2006: Andersen et al., 1999), and a third paper added data from Iceland
(Pukkala et al., 2009). Cases and controls in another four studies (Dosemeci et al., 1999:
Schlehofer et al., 1995: Mellemgaard et al., 1994: McCredie and Stewart, 1993) were included in
the National Cancer Institute's (NCI's) multicenter international renal cell study (Mandel et al.,
1995).
Generally, cohort studies presented risk estimates for "kidney and other and unspecified
urinary organs," and case-control studies presented risk estimates for renal cell carcinoma, a
histological type included in the broader kidney and other and unspecified urinary organs
category. The exceptions were two studies that presented risk estimates for cancer of the renal
pelvis (Wilson et al., 2008: McCredie and Stewart, 1993) and two studies of the same cohort that
presented risk estimates for kidney and urinary (bladder) organs (Sung et al., 2007: Chang et al.,
2003). These 27 studies represent the core studies evaluated by EPA, as described in more detail
4-66
-------
below. One other cohort study included information on tetrachloroethylene but did not report
risk estimates for kidney cancer (Radican et al., 2008), and one case-control study identified only
one exposed case (as a dry-cleaning operator) and did not provide an estimate of the association
(Partanen et al., 1991), and so were not evaluated further. Appendix B reviews the design,
exposure-assessment approach, and statistical methodology for each study. Most studies were of
the inhalation route of exposure, of occupational exposure, and unable to quantify
tetrachloroethylene exposure.
4.2.1.2.1. Consideration of exposure-assessment methodology
Many studies examine occupational title as dry cleaner, launderer, and presser as
surrogate for tetrachloroethylene, given its widespread use from 1960 onward in the United
States and Europe (Calvert etal.. 2011: Selden and Ahlborg, 2011: Pukkala et al., 2009: Wilson
et al., 2008: Lynge et al., 2006: Ji et al., 2005b: Blair etal., 2003: Travier et al., 2002: Parent et
al., 2000: Andersen et al., 1999: Dosemeci et al., 1999: Delahunt et al., 1995: Mandel et al.,
1995: Auperin et al., 1994: Mellemgaard et al., 1994: McCredie and Stewart, 1993: Lynge and
Thygesen, 1990: Asal et al., 1988). Seven studies conducted in Nordic countries are based on
either the entire Swedish population or combined populations of several Nordic countries;
strengths of these studies are their use of job title as recorded in census databases and
ascertainment of cancer incidence using national cancer registries (Selden and Ahlborg, 2011:
Pukkala et al., 2009: Wilson et al., 2008: Lynge et al., 2006: Ji et al., 2005b: Travier et al., 2002:
Andersen et al., 1999). Some variation can be expected within an occupational group among
countries; however, as Lynge et al. (2006) reported, average tetrachloroethylene usage in
1960-1970 in Sweden was higher than in Finland or Norway. Studies examining mortality
among U.S. dry-cleaner and laundry workers (Calvert et al., 2011; Blair et al., 2003) are of
smaller cohorts than the Nordic studies, with fewer observed kidney cancer events.
The exposure surrogate in studies of dry-cleaners and laundry workers is a broad
category containing jobs of differing potential for tetrachloroethylene exposure. Thus, these
studies have a greater potential for exposure misclassification bias compared to studies with
exposure potential to tetrachloroethylene assigned by exposure matrix approaches applied to
individual subjects. Three studies used additional information pertaining to work environment to
refine the exposure classification (Calvert et al., 2011; Selden and Ahlborg, 2011; Lvnge et al.,
2006). Selden and Ahlborg (2011) obtained information about the dry-cleaning establishment
(e.g., washing techniques, chemicals used, number of employees, and work history of individual
employees) in a questionnaire sent to businesses in Sweden in the 1980s. Lynge et al. (2006),
using job titles reported in the 1970 Census, identified subjects based on the occupational code of
"Laundry and dry-cleaning worker" or industry code of "Laundry and dry cleaning." Additional
4-67
-------
information used to refine this classification was sought for incident kidney cancer cases (and
cases of cancer of the esophagus, gastric cardia, liver, pancreas, cervix, bladder, and
non-Hodgkin lymphoma) within this defined cohort. Five controls, matched to the cases by
country-, sex- age-, and calendar period, were also included in this study. The additional
information sought by Lynge et al. (2006) included handwritten task information from the census
form from Denmark and Norway, pension databases in Denmark and Finland, and next-of-kin
interviews in Norway and Sweden. Exposure classification categories were dry cleaner (defined
as dry cleaners and supporting staff if employed at a business with <10 workers), other job titles
in dry cleaning (launderers and pressers), unexposed (job title reported on 1970 Census was
other than in dry cleaning), or unclassifiable (information was lacking to identify job title of
subject). The unclassifiable category represented 43 of 210 identified kidney cancer cases (20%)
and 241 of the 1,060 controls (22%). Another dry-cleaning study of unionized dry cleaners in
the United States included an analysis of subjects who worked for 1 or more years before 1960 in
a shop known to use tetrachloroethylene as the primary solvent (Calvert et al., 2011; Ruder et al.,
2001, 1994). The cohort was stratified into two groups based on the level of certainty that the
worker was employed only in facilities using tetrachloroethylene as the primary solvent
exposure; tetrachloroethylene-only and tetrachloroethylene-plus. Two of the five observed
kidney cancer deaths were among the tetrachloroethylene-only subset (n = 618) of study
subjects.
Only Blair et al. (2003; 1993) used an exposure metric for semiquantitative cumulative
exposure within a dry-cleaning setting. Four other studies presented risk estimates by
employment duration (Lynge et al., 2006; Ji et al., 2005b: Travier et al., 2002; Mandel et al.,
1995). Because employment duration does not account for variation in exposure levels, it is a
weaker exposure measure (i.e., more subject to misclassification) compared with one defined as
a semiquantitative measure.
One case-control study used a job exposure matrix (JEM) or one with information on
specific tasks, a job-task exposure matrix (ITEM), with semiquantitative exposure assessment
across a variety of jobs (Pesch et al., 2000a), and two study centers (Dosemeci et al., 1999;
Schlehofer et al., 1995) of the large NCI international renal cell carcinoma study used a JEM and
occupations to assign overall tetrachloroethylene exposure potential to individual subjects. In
Pesch et al. (2000a), the use of the German JEM identified approximately three times as many
cases with any potential tetrachloroethylene exposure (38%) compared to the JTEM (12%), and,
in both approaches, few cases were identified with substantial exposure (6% by JEM and 2% by
JTEM). Pesch et al. (2000a) noted, "exposure indices derived from an expert rating of job tasks
can have a higher agent-specificity than indices derived from job titles." For this reason, the
JTEM approach, with consideration of job tasks, is considered a more robust exposure metric for
4-68
-------
examining tetrachloroethylene exposure and kidney carcinoma due to likely reduced potential for
exposure misclassification compared to exposure assignment using only job history and title.
Four other cohorts with potential tetrachloroethylene exposure in manufacturing settings
have been examined. These studies include aerospace workers in the United States (Boice et al.,
1999), workers primarily in the metal industry, workers in Finland (Anttila et al., 1995), and
electronic factory workers in Taiwan (Sung et al., 2007; Chang et al., 2005). Boice et al. (1999)
used an exposure assessment based on a job-exposure matrix, and Anttila et al. (1995) used
biological monitoring of tetrachloroethylene in blood to assign potential tetrachloroethylene
exposure to individual subjects. In contrast, the exposures in the Taiwan studies included
multiple solvents, and tetrachloroethylene exposure was not linked to individual workers. These
cohorts also included white-collar workers, who had an expected lower potential for exposure
(Sung et al., 2007: Chang et al., 2003).
Three geographic studies focused on residential proximity to drinking water sources
contaminated with tetrachloroethylene and other solvents (Ma et al., 2009; Vieira et al., 2005;
Aschengrau et al., 1993). Two other studies in Cape Cod, MA, used either an exposure model
incorporating leaching and characteristics of the community water distribution system to assign a
household-relative dose of tetrachloroethylene (Aschengrau et al., 1993) or residential proximity
to Superfund sites without identifying specific exposures and a generalized additive model that
incorporates smoothing approaches and adjusts for covariates (Vieira et al., 2005). Ma et al.
(2009) is an ecological-designed study examining the rate of hospital discharges with a diagnosis
of kidney cancer and the average number of dry cleaners per square kilometer within New York
City zip codes as an exposure surrogate.
In summary, with respect to exposure-assessment methodologies, nine studies with
kidney cancer data assigned tetrachloroethylene exposure to individuals within the study using a
job exposure matrix (Pesch et al., 2000a: Boice et al., 1999; Dosemeci et al., 1999; Schlehofer et
al., 1995), or semi quantitative metric (Blair et al., 2003), biological samples (Anttila et al.,
1995), an exposure model (Aschengrau et al., 1993), information about working conditions
obtained through a questionnaire (Selden and Ahlborg, 2011), or classifying the cohort by
certainty of tetrachloroethylene exposure (Calvert et al., 2011). One other study based on
occupational census data sought additional information for use in refining potential exposure
within dry-cleaning settings (Lynge et al., 2006). The relative specificity of these exposure-
assessment approaches strengthens their ability to identify cancer hazards compared to studies
with broader and less sensitive exposure-assessment approaches. The least sensitive exposure
assessments are those using very broad definitions such as working in a plant or factory (Sung et
al., 2007; Chang et al., 2003) or density of dry-cleaning establishments by zip code (Ma et al.,
2009).
4-69
-------
4.2.1.2.2. Summary of results
Seven of the 27 studies evaluated by EPA reported estimated relative risks based on a
large number of observed events: 50 or more deaths/incident cases in cohort studies (Pukkala et
al., 2009; Ji and Hemminki, 2005a: Travier et al., 2002; Andersen et al., 1999), or 50 or more
exposed cases in case-control studies (Pesch et al., 2000a: Dosemeci et al., 1999; Mandel et al.,
1995). Two of these studies adopted a relatively high quality exposure-assessment approach to
assign tetrachloroethylene exposure potential to individual subjects (Pesch et al., 2000a:
Dosemeci etal., 1999). Pukkala et al. (2009) updates the analysis of Andersen et al. (1999),
adding data from a 5th country, Iceland, and extending follow-up to 2005, and is preferred over
Andersen et al. (1999) for these reasons.
The three17 cohort studies with findings based on 50 or more events observed
standardized incidence ratios or odds ratio estimates of 1.15 (95% CI: 0.98, 1.35), 0.94 (95% CI:
0.83, 1.07), and 1.11 (95% CI: 0.93, 1.33) in Ji et al. (2005b), Pukkala et al. (2009) and Travier
et al. (2002), respectively, for the association between kidney cancer risk and ever having a job
title of dry-cleaner or laundry worker (refer to Table 4-7). The largest case-control study
(n = 245 cases from Australia, Denmark, Germany, Sweden, and the United States) reported an
odds ratio for the association between renal cell carcinoma and ever exposed to dry-cleaning
solvents of 1.4 (95% CI: 1.1, 1.7) (Mandel et al., 1995). Dosemeci et al. (1999), whose subjects
were included in the larger study of Mandel et al. (1995), reported an odds ratio estimate of 1.07
(95% CI: 0.7, 1.6) for the association between overall tetrachloroethylene exposure and renal cell
carcinoma, based on 50 cases exposed to tetrachlorethylene. The other large case-control study
by Pesch et al. (2000a) also included a high-quality exposure-assessment approach (ITEM) for
tetrachloroethylene. This study observed odds ratio estimates of 1.2 (95% CI: 0.9, 1.7), 1.1 (95%
CI: 0.7, 1.5), and 1.3 (95% CI: 0.7, 2.3) and, 2.2 (95% CI: 0.9, 5.2), 1.5 (95% CI: 0.6, 3.8), and
2.0 (0.5, 7.8) for medium, high, and substantial exposure in males and females, respectively.
This study observed lower odds ratio estimates for the association between kidney cancer and
tetrachloroethylene exposure assigned using a job-exposure-matrix, a less robust exposure-
assessment approach compared to a ITEM.
Differences in risk estimates between males and females were reported in three studies;
two studies observed higher point estimates in females (Ji et al., 2005b: Pesch et al., 2000a), with
a higher risk estimate for males observed in Dosemeci et al. (1999). Pukkala et al. (2009), in
contrast, did not observe differences in kidney cancer risk estimates between male and female
subjects. It is unclear why apparent differences in sex-specific results were observed in some
17 Andersen et al. (1999) is not included in this summary of the data from the individual studies because it was
updated and expanded in the analysis by Pukkala et al. (2009).
4-70
-------
studies, although different exposure potentials, different exposure intensities, chance, or residual
confounding are possible alternative explanations (NRC, 2010; Dosemeci et al., 1999).
In addition to the large cohort and case-control studies, some evidence is found in studies
whose effect estimates are based on fewer observed events and that carry less weight in the
analysis. As expected, the magnitude of the point estimate of the association reported in these
studies is more variable than in the larger studies. Because of the relatively small number of
observed exposed cases in these cohort studies or exposed cases in case-control studies, ranging
from 2 in Antilla et al. (1995) and Boice et al. (1999) to 29 in Selden and Ahlborg (2011). the
statistical power of these lesser-weighted studies is limited. The variation in the association
observed in these studies is consistent with that from studies discussed above that carry greater
weight in the analysis. For the association between kidney cancer and dry cleaning, six studies
reported risk estimates from 0.69 to 0.94 [Andersen et al. (1999): Asal et al. (1988), males;
Pukkala et al. (2009): Lynge et al. (2006): Boice et al. (1999): Lynge and Thygesen (1990)1
three studies reported risk estimates from 1.0 to 1.08 (Selden and Ahlborg, 2011: Blair et al.,
2003: Aschengrau et al., 1993), four studies reported risk estimates from 1.35 to 1.92 (Parent et
al., 2000: Anttila et al., 1995: Delahunt et al., 1995: Calvert, 1976), and four studies reported risk
estimates from 2.3 to 2.8 [Asal et al. (1988), females; Schlehofer et al. (1995): Mellemgaard et
al. (1994): McCredie and Stewart (1993)1.
Several studies had been previously identified based on the relative strengths of their
exposure-assessment methodology. The results from these studies are mixed. Some of these
studies reported no evidence of an increased risk, with relative risks of 0.67 [(Lynge et al., 2006);
dry cleaners], 0.69 [(Boice et al., 1999); routine exposure], 1.04 (Selden and Ahlborg, 2011), and
1.07 [(Dosemeci etal., 1999); tetrachloroethylene exposure]. No cases were observed in the
group above the 90th percentile of exposure based on modeling of residential exposure in
Aschengrau et al. (1993), and the overall relative risk for any tetrachloroethylene exposure was
1.08. In contrast, data from other studies with relatively strong exposure-assessment methods
provide more evidence of an effect, with relative risks of 1.35 [(Calvert, 1976);
tetrachloroethylene-only], 1.5 [(Blair et al., 2003); medium-high exposure], 1.82 (Anttila et al.,
1995); biological samples], and 2.52 [(Schlehofer et al., 1995); tetrachloroethylene or
tetrachlorocarbonate exposure]. The data from Pesch et al. (2000a), as described earlier, do not
indicate a pattern of increasing risk with increasing exposure among males (odds ratio [OR]: 1.2,
1.1, and 1.3 for medium, high, and substantial exposure, respectively), or among females,
although the overall risk pattern is stronger among women (OR: 2.2, 1.5, and 2.0 for medium,
high, and substantial exposure, respectively).
Two studies of the same population, an electronics factory in Taiwan, which did not use
an exposure-assessment approach that allowed individual-level classification of exposure,
4-71
-------
observed standardized mortality ratios (SMRs) for kidney and urinary organ cancer of 1.18
(95% CI: 0.24, 3.44) (Chang et al.. 2003) and 1.10 (95% CI: 0.62, 1.82) (Sung et al.. 20081
respectively. A geographic-based study reported a relatively constant prevalence rate ratio for
the association between hospital discharge diagnoses for kidney cancer and density of dry
cleaners by zip code of residence (Ma et al., 2009).
The two studies reporting findings for cancer of the renal pelvis and dry cleaner and
laundry jobs were each based on 10 or fewer observations, with the standardized incidence ratio
or odds ratio estimates in these studies of 1.23 (95% CI: 0.39, 2.86) and 6.09 (95% CI: 1.95, 8.9)
in Wilson et al. (2008) and McCredie and Stewart (1993), respectively.
Establishment of an exposure or concentration-response relationship can add to the
weight of evidence for identifying a cancer hazard, but only limited data pertaining to exposure-
response relationships for kidney cancer and tetrachloroethylene exposure are available. Seven
studies presented risk estimates for increasing exposure categories. Four studies used exposure
duration as a proxy (Lynge et al.. 2006: Ji et al.. 2005b: Boice et al.. 1999: Mandel etal.. 1995):
one of these included only five cases in three exposure categories (Boice etal., 1999), which
limits the potential of this study to assess trends. Three studies used a semi quantitative exposure
surrogate (Ma et al., 2009: Blair etal., 2003: Pesch et al., 2000a), but one of these was a
relatively nonspecific and nonsensitive measure based on zip code area-based density of dry
cleaners (Ma et al., 2009). A monotonic increasing trend in relative risk with increasing
exposure surrogate was not observed in any of the larger occupational exposure studies with
three or more exposure categories (Lynge et al., 2006: Pesch et al., 2000a: Mandel et al., 1995).
In a smaller study, Blair et al. (2003) reported a higher risk in the higher of two exposure
categories (SMR: 0.3 for little-to-no exposure and 1.5 for medium-to-high exposure). One other
study provided data pertaining to the effect of duration of work. Ji et al. (2005b) reported a
higher, but more imprecise, SIR for females employed as laundry workers and dry cleaners in the
1960 and 1970 Swedish Censuses (SIR: 1.67, 95% CI: 1.07, 2.37) compared to those who were
classified in this type of work only in 1960 (SIR: 1.41, 95% CI: 1.13, 1.71). Neither of the two
studies of renal pelvis cancer reported odds ratio estimates by exposure gradients.
Statistical analyses in all case-control studies except McCredie and Stewart (1993) and
Lynge et al. (2006) controlled for cigarette smoking, a known risk factor for kidney cancer
(Parent et al., 2000: Pesch et al., 2000a: Dosemeci et al., 1999: Delahunt et al., 1995: Mandel et
al., 1995: Schlehofer et al., 1995: Auperin et al., 1994: Mellemgaard et al., 1994: Aschengrau et
al., 1993: Asal etal., 1988). Fewer studies also controlled for body mass index, another risk
factor for kidney cancer (Parent et al., 2000: Dosemeci et al., 1999: Mandel etal., 1995:
Mellemgaard et al., 1994). Direct examination of possible confounders is less common in cohort
studies relying on company-supplied or census work history data compared to case-control
4-72
-------
studies where information is obtained from study subjects or their proxies. In cohort studies,
however, use of internal controls rather than an external referent group (e.g., national mortality
rates) can minimize effects of potential confounding due to smoking or socioeconomic status,
because exposed and referent subjects are drawn from the same target population. However,
only one of the available cohort studies included an analysis using internal controls, and that
study is limited by the observation of only two kidney cancer cases with routine
tetrachloroethylene exposure in the cohort (Boice etal., 1999). Effect of smoking as a possible
confounder may be assessed indirectly through examination of risk ratios for other smoking-
related sites such as lung cancer. Several studies observed roughly a 30% increase in lung
cancer risk among dry cleaners (Calvert et al., 2011; Selden and Ahlborg, 2011; Pukkala et al.,
2009; Ji et al., 2005b; Blair et al., 2003). Any expected contribution of smoking to kidney cancer
risk will be smaller than that for lung cancer.
In conclusion, the epidemiologic data provide limited evidence pertaining to
tetrachloroethylene exposure and kidney cancer risk. The studies that support this finding
include the largest international case-control study (245 exposed cases from Australia, Denmark,
Germany, Sweden, and the United States), which reported a relative risk of 1.4
(95% CI: 1.1, 1.7) for any exposure to dry-cleaning solvents (Mandel etal., 1995). This study
was able to adjust for smoking history, BMI, and other risk factors for kidney cancer. The large
cohort studies, using a more general exposure classification based on national census occupation
data, present more variable results, with relative risks of 0.94, 1.11, and 1.15 in Pukkula et al.
(2009), Travier et al. (2002), and Ji et al. (2005b), respectively. One difference among these
cohort studies is that Travier et al. (2002) and Ji et al. (2005b) were based on data from Sweden,
while Pukkula et al. (2009) used data from Sweden, Denmark, Finland, Norway, and Iceland.
Differences between these countries in tetrachloroethylene usage, as was noted by Lynge et al.
(2006), may have introduced an additional source of exposure misclassification in this
multicountry analysis. In addition to the large cohort studies, evidence also comes from cohort
and case-control studies, whose effect estimates are based on fewer observed events. Smaller
studies that do not also have a more sensitive or specific exposure metric carry lesser weight in
the analysis. Eight studies were identified that used a relatively specific exposure-assessment
approach to refine classification of potential tetrachloroethylene exposure in dry-cleaning
settings (Calvert et al., 2011; Lvnge et al., 2006; Blair et al., 2003), the aerospace industry (Boice
et al., 1999), or within a variety of workplaces (Pesch et al., 2000a; Dosemeci et al., 1999;
Anttila et al., 1995; Schlehofer et al., 1995) or a residential area setting (Aschengrau et al.,
1993). The results from these studies are mixed, with some studies reporting little or no
evidence of an association (Lynge et al., 2006; Pesch et al., 2000a; Boice et al., 1999; Dosemeci
etal., 1999; Aschengrau et al., 1993), and other studies reporting elevated risks (Calvert et al..
4-73
-------
2011: Blair etal.. 2003: Anttilaetal.. 1995: Schlehofer et al.. 1995). An increasing trend in
relative risk with increasing exposure surrogate was not observed in any of the larger
occupational exposure studies with three or more exposure categories (Lynge et al., 2006:
Mandel etal., 1995), but some indication of higher risk with higher exposure (or duration) was
observed in other studies (Blair et al., 2003). As expected, the results from 16 other studies
using a relatively nonspecific exposure measure (broad occupational title of launderers and dry
cleaners, all workers at factory, density of dry-cleaning establishments by zip code) are more
variable and less precise, reflecting a greater potential for misclassification bias.
4-74
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort Studies
Biologically monitored workers
All subjects
1.82 (0.22, 6.56)
2
Aerospace workers (Lockheed)
Routine exposure to PCE
0.69 (0.08, 2.47)
2
Routine-intermittent exposure duration to PCE
0
5yr
1.0a
0.49 (0.07, 3.68)
0.56(0.13,2.41)
0.46(0.10,2.08)
22
1
2
2
Electronic factory workers (Taiwan)
All Subjects
Males
Females
Females
1.18(0.24, 3.44)b
1.10(0.62, 1.82)c
0
l.Slexp
3
10
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
Not reported
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females and
0.7 umol/L in males (median)], follow-up 1974-1992, external
referents (SIR)
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after 1960,
worked at least 1 yr, follow-up 1960-1996, job exposure matrix
without quantitative estimate of PCE intensity, 1987-1988 8-h TWA
PCE concentration (atmospheric monitoring) 3 ppm (mean) and 9.5
ppm (median), external reference for routine exposure (SMR) and
internal references (workers with no chemical exposures) for routine-
intermittent PCE exposure (RR)
Chang et al. (2003): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997, multiple solvents
exposure, does not identify PCE exposure to individual subjects, cancer
mortality, external referents (SMR) (Chang etal. 2003):
63,982 females, follow-up 1979-2001, factory employment proxy for
exposure, multiple solvents exposures and PCE not identified to
individual subjects, cancer incidence, external referents, analyses
lagged 5 yr (SIR) (Sung et al.. 2007)
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever
exposed to mixed solvents, 851 ever-exposed to PCE), employed at
least 1 yrfrom 1952-1956, follow-up 1973-2000, job exposure matrix
(intensity), internal referent (workers with no chemical exposures) (RR)
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Dry -cleaner and laundry workers
All laundry worker and dry cleaners
Males
Females
0.92(0.73, 1.15)
1.03 (0.66, 1.53)
0.88(0.67, 1.15)
81
24
57
All subjects
1.0 (0.4, 2.0)
8
Semiquantitative exposure score
Little to no exposure
Medium to high exposure
0.3 (0.1, 1.6)
1.5(0.6,3.1)
1
7
Laundry workers and dry cleaners in 1960
Census
Males
Females
1.15(0.98, 1.35)
0.90(0.69, 1.14)
1.41(1.13, 1.71)
153
61
92
Laundry workers and dry cleaners in both 1960 and 1970 Censuses
Males
Females
Not reported
1.67 (1.07, 2.37)
26
Laundry workers and dry cleaners in 1960, 1970, and 1980 Censuses
Males
Females
Not reported
1.00(0.90, 1.10)
3
Reference
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or 1970
Census (Denmark, Finland, Norway), follow-up 1971-1987 or 1991,
PCE not identified to individual subjects, external referents (SIR)
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union members
(1945-1978), follow-up 1979-1993, semiquantitative cumulative
exposure surrogate to dry clean solvents, cancer mortality, external
referents (SMR)
Ji et al. (2005b)
9,255 Swedish men and 14,974 Swedish women employed in 1960
(men) or 1970 (women) as laundry workers or dry cleaners, follow-up
1961/1970-2000, PCE not identified to individual subjects, external
referent (SIR) and adjusted for age, period, and socioeconomic status
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Dry -cleaner and laundry workers (continued)
All laundry worker and dry cleaners
Males
Females
0.88 (0.44, 1.58)
1.50 (0.55, 3.27)
0.58(0.19, 1.36)
11
6
5
Launderer and dry cleaner
Male
Female
0.94 (0.83, 1.07)
0.89(0.68, 1.14)
0.96(0.84, 1.10)
263
62
201
All subjects
1.14(0.37,2.67)
5
Exposure duration/time since 1st employment
<5 yr/<20 yr
<5 yr/>20 yr
>5 yr/<20 yr
>5 yr/>20 yr
PCE-only subjects
Not reported
Not reported
Not reported
Not reported
1.35(0.16,4.89)
2
Dry-cleaners and laundry workers
PCE
Laundry
1.04 (0.69, 1.49)
Not reported
Not reported
100
Reference
Lynge and Thygsen (1990)
10,600 Danish men and women, 20-64 yr old, employed in 1970 as
laundry workers, dry cleaners, and textile dye workers, follow-up
1970-1980, external referents (SIR)
Pukkala et al. (2009)
Men and women participating in national census on or before 1990, 5
Nordic countries (Denmark, Finland, Iceland, Norway, Sweden),
30-64 yr, follow-up 2005, occupational title of launderer and dry
cleaner in any census, external referents (SIR)
Calvert et al. (2011)
1,704 U.S. men and women dry-cleaning union members in CA, IL,
MI, NY follow-up 1940-2004 (618 subjects worked for one or more
yr prior to 1960 only at shops where PCE was the primary cleaning
solvent, identified as PCE-only exposure), cancer mortality (SMR)
Selden and Ahlborg (2011)
9,440 Swedish men (n = 2,810) and women (n = 9,440) in 461
washing and dry-cleaning establishments, identified by employer in
mid-1980s, employed 1973-1983, follow-up 1985-2000, exposure
assigned using company serf-reported information on PCE usage —
PCE (dry cleaners and laundries with a proportion of PCE dry
cleaning), laundry (no PCE use), and other (mixed exposures to PCE,
CFCs, TCE, etc.), external referents (SIR)
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
All subjects, 1960 or 1970 Census in laundry and
dry cleaner or related occupation and industry
All subjects in 1960 and 1970 in laundry and dry
cleaner occupation and industry
1.11(0.93, 1.33)
1.20 (0.71, 2.02)
121
14
All subjects, laundry and dry cleaning occupation
Males
Females
Not reported
1.23 (0.39, 2.86)
<2 obs.
5
Reference
Travier et al. (2002)
Swedish men and women identified as laundry worker, dry cleaner, or
presser (occupational title), in the laundry, ironing, or dyeing industry
or related industry in 1960 or 1970 (543,036 person-years); or, as
laundry worker, dry cleaner, or presser (occupational and job title)
(46,933 person-years) in both censuses, follow-up 1971-1989,
external referents (SIR)
Wilson et al. (2008)
16,512 Swedish men (n = 3,375) and women (n = 13,137) identified in
1960 or 1970 as laundry worker or dry cleaner (occupation) or in
laundry, ironing and dyeing industry, follow-up 1971-1989, external
referents (SIR), cancer of the renal pelvis
Case-Control Studies
Dry-cleaning industry
Males
Females
0.7 (0.2, 2.3)
2.8 (0.8, 9.8)
3
8
Upper Cape Cod, MA (United States)
Any PCE
RDD>90thpercentile
1.08 (0.42, 2.79)
6
0
Asal et al. (1988)
315 histologically or radiologically confirmed renal cell carcinoma
cases identified from 29 Oklahoma hospitals, 1981-1984, 336
population controls frequency matched on age and sex and 313
hospital controls matched by age, sex, race, hospital and time of
interview to cases, in-person interview using questionnaire, longest
job held was exposure surrogate, OR adjusted for age, smoking weight
Aschengrau et al. (1993). Vieira et al. (2005)
35 kidney cancer cases, 1983-1986, Massachusetts Cancer Registry,
777 population controls, residential history, ordinal estimate of PCE-
contaminated water (ROD) from exposure model (Aschengrau et al..
1993) or geographical information system and proximity to
groundwater plume (Vieira et al.. 2005). OR adjusted for sex, age at
diagnosis, vital status at interview, education, cigarette smoking, and
urinary tract infection or stone (both studies)
oo
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
10 hospitals (France)
Dry cleaning occupation
Not reported
Population of New Zealand
Launderer and dry cleaner occupation
1.92(0.37, 13.89)
Not
reported
International Renal Cell Cancer Study (Australia, Denmark, Germany, Sweden, United
States)
All Centers (Mandel et al., 1995)
Ever exposed to dry-cleaning solvents
Duration of exposure to dry-cleaning solvents
(yr)
1-7
8-25
26-60
1.4(1.1, 1.7)
0.2 (0.9, 1.8)
1.7(1.2,2.4)
1.2 (0.9, 1.8)
245
70
78
75
Denmark (Mellemgaard et al., 1994)
>1 yr exposure duration in dry-cleaning industry,
10 yr before interview
Males
2.3 (0.2, 27)
2
Reference
Auperin et al. (1994)
151 histologically confirmed renal cell carcinoma hospital cases,
1987-1991, 161 hospital cancer controls and 186 hospital controls
with nonmalignant disease matched on age, sex, and interviewed to
cases, in-person interview, lifetime occupational title as exposure
surrogate, OR adjusted for age, smoking, weight
Delahunt et al. (1995)
710 male histologically confirmed renal cell carcinoma cases, >20 yr
of age, 1978-1986, 12,758 male controls randomly selected from
same cancer registry as cases but with tumor outside urinary tract,
interview method not reported, occupational title (ever employed or
usual job title not reported) as exposure surrogate, Mantel -Haenszel
OR stratified by smoking history and 10 yr age group
Mandel et al. (1995): Dosemeci et al. (1999): McCredie and Stewart
(1993): Mellemgaard et al. (1994): Schlehofer et al. (1995)
1,732 histologically or cytologically confirmed renal cell carcinoma
cases from 6 study centers (Mandel et al.. 1995) [438 renal cell
carcinoma cases from one United States center [Minnesota Cancer
Surveillance System, a SEER reporting site] (Dosemeci etal. 1999).
368 cases from Denmark (Mellemgaard etal.. 1994). 277 renal cell
carcinoma cases from 10 local urology departments near Heidelberg,
Germany (Schlehofer etal.. 1995)1. 20-79 yr (20-75 yr, Heidelberg),
1989-1991, identified from hospital surveillance (Germany) national
cancer registries (all other countries), same birth country and cancer
registry (except Australia and the United States), 2,309 population
controls (all countries, with controls >65 yr in the United States
identified from HCFA roles) (Mandel et al.. 1995) [687 population
controls (Dosemeci etal.. 1999): 396 population referents
(Mellemsaard et al., 1994), 286 population controls (Schlehofer et al..
1995)1. matched on sex, and age, in-person interview with
-^
VO
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
cont
Females
Relative risk
(95% CI)
2.9(0.3,33)
No. obs.
events
2
New South Wales, Australia (McCredie and Stewart, 1993)d
Dry-cleaning industry occupation or job
2.70 (1.08, 6.72)
16
Germany (Heidelberg) (Schlehofer et al., 1995)
PCE and tetrachlorocarbonate
2.52(1.23,5.16)
27
United States (Minnesota) (Dosemeci et al., 1999)
PCE
Male
Female
1.07 (0.7, 1.6)
1.12(0.7, 1.7)
0.82(0.3,2.1)
50
42
8
Nordic Countries (Denmark, Finland, Norway, Sweden)
Unexposed
Dry cleaner
Other in dry-cleaning
Unclassifiable
1.00
0.67 (0.43, 1.05)
1.15(0.52,2.53)
0.76(0.50, 1.16)
129
29
9
3
Dry cleaner, employment duration, 1964-1979
10yr
Unknown
0.24 (0.03, 2.04)
0.86 (0.28, 2.67)
0.70 (0.32, 1.55)
0.75 (0.39, 1.42)
0.70(0.15,3.36)
1
4
8
14
2
Reference
questionnaire inquiry on specific occupations (4 centers)
or Ml occupational history (2 centers); occupation and chemical
grouping as exposure surrogate, OR stratified by sex and adjusted for
age, smoking, BMI, education, and study center, OR reported for
males only (Mandel et al.. 1995). In Mellemgaard et al. (1994). OR
for occupational title/exposure >1 year duration and 10 years before
interview and adjusted for age, BMI and smoking. In Dosemeci et al.
(1999). OR reported for both sexes together and separately and
adjusted for age, smoking hypertension, and/or diuretic use, and/or
anti-hypertension drug use, and BMI. In Schlehofer et al. (1995). OR
for exposure duration >5 years and adjusted for age and smoking
Lynge et al. (2006)
Case-control study among 46,768 Danish, Finnish, Norwegian, and
Swedish men and women employed in 1960 as laundry worker or dry
cleaner, follow-up 1970-1971 to 1997-2001, 210 renal cell
carcinoma cases, 3 controls per case randomly selected from cohort
matched on country, sex, age, calendar period at diagnosis time,
occupational task at 1970 Census proxy for exposure, kidney cancer
incidence, RR adjusted for country, sex, age, calendar period at time
of diagnosis
oo
o
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
New South Wales, Australia
Dry-cleaning industry occupation or job
6.09(1.95, 18.9)e
8
Finland
Dry-cleaning operator
Not reported
1
Germany, 5 regions
PCE, JEM
Medium exposure
High exposure
Substantial exposure
1.1(0.9, 1.4) M
1.2(0.8, 1.8) F
1.1(0.9, 1.4) M
1.3(0.8, 2.0) F
1.3(0.9, 1.8) M
0.8(0.3, 1.9) F
135
28
138
29
55
6
PCE, ITEM
Medium exposure
High exposure
Substantial exposure
1.2(0.9, 1.7) M
2.2 (0.9, 5.2) F
1.1(0.7, 1.5) M
1.5(0.6, 3.8) F
1.3(0.7, 2.3) M
2.0 (0.5, 7.8) F
44
8
39
6
15
3
Reference
McCredie and Stewart (1993)
147 renal pelvic cancer cases, 20-79 yr, 1989-1990, identified from
hospitals and physicians, 523 population controls, in-person or
telephone interview, job title or industry as exposure surrogate, OR
adjusted for age, sex, and method of interview (for renal cell
carcinomas) and age, sex, interview methods and education (for renal
pelvic cancers)
Partanen et al. (1991)
338 renal cell carcinoma cases, 20-95 yr, 1977-1987, identified from
Finnish Cancer Registry, 484 population controls matched on birth
year, sex, and survival status at time of interview, mailed interview,
job title or industry for all jobs held 1926-1968, OR adjusted for
smoking, coffee consumption and obesity
Pesch et al. (2000a)
935 histologically confirmed renal cell carcinomal cancer in men and
women, hospital record study, 1991-1995, 4,298 age-sex-matched
population controls, in-person interview, JEM and ITEM for PCE, OR
adjusted for age, study center, smoking
oo
-------
Table 4-7. Summary of human studies on tetrachloroethylene exposure and kidney cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Montreal, Canada
Launderers and dry cleaners
Any exposure
Substantial exposure
1.7 (0.6, 4.7)
4
0
Reference
Parent, 2000
142 histologically confirmed renal cell carcinoma cancer, 1979-1985,
35-70 yr, 533 population control group and 1,900 cancer control
group, in-person interviews, occupational title, OR adjusted age,
smoking, and BMI
Geographic Studies
New York City, NY (United States)
Zip codes with number of dry cleaners/km2
0-0.47
0.47-0.90
0.90-1.50
1.50-2.70
2.70-16.43
l.Oa
1.14(1.03, 1.27)f
1.09 (0.97, 1.21)f
1.17(1.05, 1.32)f
1.15(1.01, 1.30)f
1,458
2,289
1,838
2,766
2,565
Ma, 2010
10,916 cases with hospital discharge diagnosis of renal or renal pelvis
cancer, 1993-2004, zip code of residential address and dry-cleaner
business number/zip code area as exposure surrogate, crude
prevalence rate ratio (prevalence RR)
oo
to
aReferent.
bFor Chang et al. (2003). SMR for kidney and urinary organs.
Tor Sung et al. (2007). SIR for kidney and urinary organs, 10 yr lag period.
dln McCredie and Stewart (1993). renal cell carcinoma cases from hospitals and physicians in New South Wales, Australia. Of the 489 renal cell carcinoma
cases, 256 were from the Sydney Metropolitan area and were included in the National Cancer Institute's international study (Mandeletal.. 1995).
eln McCredie and Stewart (1993). OR for renal pelvic cancer.
fln Ma et al. (2009). rate ratio from negative binomial regression model with main effect for zip code (crude rate ratio). Rate ratios from models with adjustment
forage, race, sex, population density and median household and effect modifiers that vary by exposure category are 1.0 (referent), 1.15 (95% CI: 1.04, 1.27)
[no effect modification], 1.10 (1.00, 1.24) [effect modification by population density], 1.27 (95% CI: 1.13, 1.42) [effect modification by race], and 1.16 (05%
CI: 1.02, 1.33) [effect modification by mean household income and age], for numbers of dry cleaners of 0 -0.47, 0.47-0.90, 0.90-1.50, 1.50-2.70, and
2.70-16.43/km2, respectively.
JEM = job-exposure matrix, HCFA = Health Care Financing Administration, ITEM = job-task-expo sure-matrix, PCE = tetrachloroethylene, ROD = relative
delivered dose, TWA = time-weighted-average.
-------
4.2.1.3. Bladder Cancer in Humans
Thirty-two epidemiologic studies reporting data on bladder cancer and
tetrachloroethylene exposure were identified. This set of studies includes 13 cohort or nested
case-control studies within a cohort (Calvert et al., 2011; Selden and Ahlborg, 2011; Pukkala et
al.. 2009; Wilson et al.. 2008; Sung et al.. 2007; Lynge et al.. 2006; Chang et al.. 2005; Ji and
Hemminki, 2005a; Blair et al., 2003; Travier et al., 2002; Andersen et al., 1999; Boice et al.,
1999; Lynge and Thygesen, 1990), 16 case-control studies of occupational exposures (Colt et al.,
2011; Drvson et al., 2008; Reulen et al., 2007; Gaertner et al., 2004; Kogevinas et al.. 2003;
Zheng et al., 2002; Pesch et al., 2000b; Teschke et al., 1997; Swanson and Burns, 1995; Burns
and Swanson, 1991; Siemiatycki, 1991; Steineck et al., 1990; Silverman et al., 1989a; Silverman
et al., 1989b; Smith etal., 1985; Schoenberg et al., 1984), and 3 studies of residential exposure
through contaminated drinking water (Vieira et al., 2005; Aschengrau et al., 1993; Mallin, 1990).
These 32 studies represent the core studies evaluated by EPA, as described in more detail below.
Two other cohort studies and one case-control study included information on tetrachloroethylene
but did not report risk estimates for bladder cancer (Radican et al., 2008; Colt et al., 2004;
Anttila et al., 1995), and so were not evaluated further. The peer-reviewed literature also
contains a meta-analysis that examined dry cleaning and bladder cancer (Reulen et al., 2007).
There is some overlap in the study populations among these studies: Travier et al. (2002)
used occupational data from the Swedish national census, and Lynge and Thygsen (1990) used a
similar design in Denmark; Andersen et al. (1999) and Lynge et al. (2006) expanded these
studies to include Denmark, Finland, and Norway in addition to Sweden, and Pukkala et al.
(2009) added Iceland to this set. Pesch et al. (2000b) is a large case-control study examining
urothelial cancers, a grouping of bladder, ureter, and renal pelvis neoplasms, with exposure
information on tetrachloroethylene. Kogevinas et al. (2003), a pooled analysis of 11 studies
conducted in European countries between 1976 and 1996, includes the dry cleaning but not the
tetrachloroethylene exposure observations in males in Pesch et al. (2000b). Kogevinas does not
provide information on women; 't Mannetje et al. (1999) pooled observations in women in these
11 studies but did not report findings on dry-cleaner and laundry workers.
Appendix B reviews the design, exposure-assessment approach, and statistical
methodology for each study. Most studies were of the inhalation route, of occupational
exposure, and unable to quantify tetrachloroethylene exposure.
4.2.1.3.1. Consideration of exposure-assessment methodology
Many studies examine occupational titles such as dry cleaner, launderer, and presser as
surrogate for tetrachloroethylene, given its widespread use from 1960 onward in the United
4-83
-------
States and Europe (Calvert et al.. 2011: Colt etal.. 2011: Pukkala et al.. 2009: Dryson et al.,
2008: Reulen et al.. 2008: Wilson et al.. 2008: Reulen et al.. 2007: Lynge et al.. 2006: Ji and
Hemminki, 2005a: Gaertner et al.. 2004: Blair et al.. 2003: Kogevinas etal.. 2003: Travier et al..
2002: Zheng et al.. 2002: Andersen et al.. 1999: Teschke et al.. 1997: Swanson and Burns. 1995:
Burns and Swanson, 1991: Lynge and Thygesen, 1990: Silverman et al.. 1990: Steineck et al.,
1990: Silverman et al.. 1989a: Silverman et al.. 1989b: Smith etal.. 1985: Schoenberg et al..
1984). Six studies conducted in Nordic countries are either based on the entire Swedish
population or combined populations of several Nordic countries; strengths of these studies are
their use of job titles as recorded in census databases and ascertainment of cancer incidence
using national cancer registries (Pukkala et al., 2009: Wilson et al.. 2008: Lynge et al.. 2006: Ji et
al.. 2005a: Travier et al.. 2002: Andersen et al., 1999). Studies examining mortality among U.S.
dry-cleaner and laundry workers (Calvert et al., 2011: Blair et al.. 2003) are of smaller cohorts
than the Nordic studies, with fewer observed bladder cancer events.
The exposure surrogate in studies of dry-cleaners and laundry workers is a broad
category containing jobs of differing potential for tetrachloroethylene exposure. Thus, these
studies have a greater potential for exposure misclassification bias compared to studies with
exposure potential to tetrachloroethylene assigned by exposure matrix approaches. Two studies
used additional information pertaining to work environment to refine the exposure (Calvert et al.,
2011: Lvnge et al., 2006). Lynge et al. (2006), using job titles reported in the 1970 Census,
identified subjects based on an occupational code of "laundry and dry-cleaning worker" or an
industry code of "laundry and dry cleaning." Additional information to refine this occupational
classification was sought for incident cancer cases, including bladder cancer, within this defined
cohort. Five controls, matched to the cases by country, sex, age, and calendar period, were also
included in the study. The additional information included handwritten task information from
the census forms from Denmark and Norway, pension databases in Denmark and Finland, and
next-of-kin interviews in Norway and Sweden. Exposure classification categories were dry
cleaner (defined as dry cleaners and supporting staff if employed in a business of <10 workers),
other job titles in dry cleaning (launderers and pressers), unexposed (job title reported on 1970
census was other than dry cleaning), or unclassifiable (information was lacking to identify job
title of subject). The unclassifiable category represented 57 of 351 bladder cancer cases (16%)
and 234 out of 1,482 controls (16%). The study by Calvert et al. (2011) included an analysis of
subjects who worked for one or more years before 1960 in one or more shops known to use
tetrachloroethylene as the primary solvent (Calvert et al., 2011). The cohort was stratified into
two groups based on the level of certainty that the worker was employed only in facilities using
tetrachloroethylene as the primary solvent exposure; tetrachloroethylene-only and
tetrachloroethylene plus. However, there were no bladder cancer deaths among this subset
4-84
-------
(n = 618) of tetrachloroethylene-only subjects. Three additional studies used a semi quantitative
or quantitative exposure metric. Blair et al. (2003) used an exposure metric for semi quantitative
cumulative exposure between dry-cleaning and laundry workers. The case-control study by
Siemiatycki (1991) used a JEM based on occupational titles for tetrachloroethylene, and another
case-control study used a JEM and one JEM with information on specific tasks, a JTEM, with
semi quantitative exposure assessment across a variety of jobs (Pesch et al., 2000b).
Two other cohorts with potential tetrachloroethylene exposure in manufacturing settings
have been examined. These studies include aerospace workers in the United States (Boice et al.,
1999) and electronic factory workers in Taiwan (Sung et al., 2007; Chang et al., 2005). Boice et
al. (1999) used an exposure assessment based on a JEM to classify exposures. In contrast, the
exposures in the Taiwan studies included multiple solvents, and tetrachloroethylene exposure
was not linked to individual workers (Sung et al., 2007; Chang et al., 2005).
Three geographic studies focused on residential proximity to drinking water sources
contaminated with tetrachloroethylene and other solvents. Mallin (1990) examined incidence
and mortality by county in Illinois, with the exposure surrogate assigned uniformly to all
subjects. Two other studies in Cape Cod, MA, used either an exposure model incorporating
tetrachloroethylene leaching and characteristics of the community water distribution system
(Aschengrau et al., 1993) or residential proximity to Superfund sites and a generalized additive
model that incorporates smoothing approaches and adjusts for covariates (Vieira et al., 2005).
In summary, four studies with bladder cancer data assigned tetrachloroethylene exposure
to individuals within the study using a job exposure matrix (Blair et al., 2003; Pesch et al.,
2000b: Boice et al., 1999) or an exposure model (Aschengrau et al., 1993). One other study
sought additional data using a questionnaire for use in refining potential exposure within dry-
cleaning settings (Lynge et al., 2006). The relative specificity of these exposure-assessment
approaches strengthens their ability to identify cancer hazards compared to studies with broader
and less sensitive exposure-assessment approaches.
4.2.1.3.2. Summary of results
Seven studies evaluated by EPA reported estimated relative risks based on a large
number of observed events; 50 or more deaths/incident cases in cohort studies (Pukkala et al.,
2009: Wilson et al., 2008: 2005a: Travier et al., 2002: Andersen et al., 1999), or 50 or more
exposed cases in case-control studies (Lynge et al., 2006: Pesch et al., 2000b), with sufficient
power to detect a twofold elevation in estimated risk. Pukkala et al. (2009) updates the analysis
of Andersen et al. (1999) adding data from a 5th country, Iceland, and extending follow-up to
4-85
-------
2005, and is preferred over Andersen et al. (1999) for these reasons. The five18 large cohort
studies observed a standardized incidence ratio or odds ratio estimate of 1.01 (95% CI: 0.86,
1.19), 1.08 (95% CI: 0.98, 1.23), 1.14 (95% CI: 0.89, 1.45), 1.27 (95% CI: 1.08, 1.48), and 1.44
(95% CI: 1.07, 1.93) in Travier et al. (2002). Pukkala et al. (2009). Wilson et al. (2008). Ji et al.
(2005a) and Lynge et al. (2006), respectively, for the association between bladder cancer risk
and ever having a job title of dry cleaner or laundry worker (refer to Table 4-8). The Lynge et al.
(2006) results were slightly higher among the subgroup from Denmark and Norway, in which the
number of unclassifiable data was negligible (relative risk 1.69, 95% CI: 1.18, 2.43). The large
case-control study by Pesch et al. (2000b) reported an odds ratio of 0.8 (95% CI: 0.6, 1.2), 1.3
(95% CI: 0.9, 1.7), and 1.8 (95% CI: 1.2, 2.7) for medium, high, and substantial exposure,
respectively, compared to low exposure, based on the ITEM approach.
Additional evidence is found in studies whose effect estimates are based on fewer
observed events and that carry lesser weight in the analysis. As expected, the magnitude of the
point estimate of the association reported in these studies is more variable than in the larger
studies: 4 studies report relative risks between 0.7 and 0.91 [Colt et al. (2011), males; Dryson et
al. (2008): Boice et al. (1999): Lynge and Thygesen (1990)1, 10 studies report relative risks
between 1.2 and 1.9 [Colt et al. (2011), females; Gaertner et al. (2004): Blair et al. (2003):
Kogevinas et al. (2003); Aschengrau et al. (1993); Burns and Swanson (1991); Siemiatycki
(1991): Steineck et al. (1990): Smith et al. (1985): Schoenberg et al. (1984)1, and 3 studies report
relative risk estimates >2.0 (Reulen et al., 2007; Zheng et al., 2002; Teschke et al., 1997).
Except for the estimate from Reulen et al. (2007) (RR: 2.7, 95% CI: 1.1, 6.6), all of the 95% CIs
of these estimates overlap 1.0. Because of the relatively small number of observed cases in these
cohort studies or exposed cases in case-control studies, ranging from 2 in Boice et al. (1999) to
19 in the pooled study of Kogevinas et al. (2003); the statistical power of these lesser-weighted
studies is limited.
18 Andersen et al. (1999) is not included in this summary of the data from the individual studies because it was
updated and expanded in the analysis by Pukkala et al. (2009).
4-86
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort Studies
Biologically monitored workers
All subjects
Not reported3
Aerospace workers (Lockheed)
Routine exposure to PCE
Routine-Intermittent exposure to PCE
0.70 (0.09, 2.53)
Not reported13
2
Electronic factory workers (Taiwan)
All Subjects
Males
Females
Females
1.06(0.45, 2.08)c
1.09(0.56,1.91)°
0.34 (0.07, 1.00)
8
12
12
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
Not reported
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females
and 0.7 umol/L in males (median)], follow-up 1974-1992,
external referents (SIR)
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after
1960, worked at least 1 yr, follow-up 1960-1996, job exposure
matrix without quantitative estimate of PCE intensity,
1987-1988 8-h TWA PCE concentration (atmospheric
monitoring) 3 ppm [mean] and 9.5 ppm [median], external
reference for routine exposure (SMR) and internal references
(workers with no chemical exposures) for routine-intermittent
PCE exposure (RR)
Chang et al. (2005): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997, multiple
solvents exposure, does not identify PCE exposure to individual
subiects, cancer incidence, external referents (SIR) (Chans et al..
2005);
,,.,., ., /-v^-./-v s\f\f\i r-
63,982 females, follow-up 1979-2001, factory employment
proxy for exposure, multiple solvents exposures and PCE not
identified to individual subjects, cancer incidence, external
referents, analyses lagged 5 yr (SIR) (Sung et al.. 2007)
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever
exposed to mixed solvents, 851 ever-exposed to PCE), employed
at least 1 yrfrom 1952 to 1956, follow-up 1973-2000, job
exposure matrix (intensity), internal referent (workers with no
chemical exposures) (RR)
-^
oo
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Dry -cleaner and laundry workers
All laundry worker and dry cleaners
Males
Females
1.00(0.83, 1.21)
1.14(0.87, 1.46)
0.89(0.68, 1.16)
119
62
57
All subjects
1.3 (0.7, 2.4)
12
Semiquantitative exposure score
Little to no exposure
Medium to high exposure
1.4 (0.4, 3.2)
1.5(0.6,3.1)
5
7
Male laundry workers and dry cleaners in 1960
Census
Male laundry workers and dry cleaners in 1960
Census
Male laundry workers and dry cleaners in both 1960
and 1970 Censuses
Male laundry workers and dry cleaners in 1960, 1970
and 1980 Censuses
Female laundry workers and dry cleaners
1.27(1.08,1.48)
1.13(0.96, 1.31)d
1.03 (0.80, 1.29)d
0.86(0.51, 1.28)d
Not reported
157
157
67
19
All laundry worker and dry cleaners
Males
Females
0.74 (0.41, 1.25)
0.62 (0.23, 1.35)
0.88 (0.38, 1.73)
14
6
8
Reference
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or
1970 Census (Denmark, Finland, Norway), follow-up
1971-1987 or 1991, PCE not identified to individual subjects,
external referents (SIR)
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union
members (1945-1978), follow-up 1979-1993, semiquantitative
cumulative exposure surrogate to dry clean solvents, cancer
mortality, external referents (SMR)
Ji et al. (2005a)
9,255 Swedish men employed in 1960 as laundry worker or dry
cleaner, follow-up 1961-2000, PCE not identified to individual
subjects, external referent (SIR) and adjusted for age, period and
socioeconomic status
Lynge and Thygsen (1990)
10,600 Danish men and women, 20-64 yr old, employed in 1970
as laundry worker, dry cleaners and textile dye workers, follow-
up 1970-1980, external referents (SIR)
oo
oo
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Launderer and dry cleaner
Male
Female
All subjects
Exposure duration/time since 1st employment
<5 yr/<20 yr
<5 yr/>20 yr
>5 yr/<20 yr
>5 yr/>20 yr
PCE-only subjects
Dry-cleaners and laundry workers (females)
All subjects, 1960 or 1970 Census in laundry and dry
cleaner occupation and industry
All subjects in 1960 and 1970 in laundry and dry
cleaner occupation and industry
Relative risk
(95% CI)
1.08 (0.98, 1.23)
1.10(0.95, 1.27)
1.07 (0.95, 1.22)
1.81(0.87,3.33)
0.53(0.03,2.52)
4.08(2.13,7.12)
0.92 (0.65, 1.26)
1.01(0.86, 1.19)
1.00(0.61, 1.63)
No. obs.
events
434
186
248
10
0
1
0
9
0
38
145
16
Reference
Pukkala et al. (2009)
Men and women participating in national census on or before
1990, 5 Nordic countries (Denmark, Finland, Iceland, Norway,
Sweden), 30-64 yr, follow-up 2005, occupational title of
launderer and dry cleaner in any census, external referents (SIR)
Calvert et al. (2011)
1,704 U.S. men and women dry-cleaning union member in CA,
IL, MI, NY follow-up 1940-2004 (618 subjects worked for one
or more years prior to 1960 only at shops where PCE was the
primary cleaning solvent, identified as PCE-only exposure),
cancer mortality (SMR)
Selden and Ahlborg (2011)
9,440 Swedish men (n = 2,810) and women (n = 9,440) in 461
washing and dry-cleaning establishments, identified by employer
in mid-1980s, employed 1973-1983, follow-up 1985-2000
Travier et al. (2002)
Swedish men and women identified in 1960, 1970, or both
Censuses as laundry worker, dry cleaner, or presser
(occupational title) or in the laundry, ironing, or dyeing industry,
follow-up 1971-1989, separates launders and dry cleaners form
pressers, external referents (SIR)
oo
VO
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
All subjects, laundry and dry cleaning occupation
Males
Females
1.14(0.89, 1.45)
1.23 (0.83, 1.74)
1.07 (0.75, 1.47)
68
31
37
Reference
Wilson et al. (2008)
Swedish men and women identified in 1960 or 1970 as laundry
worker or dry cleaner (occupation) or in laundry, ironing and
dyeing industry, follow-up 1971-1989, external referents (SIR),
transitional cell carcinoma
Case-Control Studies
Upper Cape Cod, MA (United States)
Any PCE
ROD >90th percentile
"Hotspof'SWofMMR
1.39(0.67,2.91)
4.03(0.65,25.10)
~2.5(CInot
reported)
13
4
Metropolitan Detroit, MI (United States)
Usual occupation as dry-cleaning workers
Males
Females
Usual industry in dry cleaner and laundry
1.9 (0.7, 4.9)
Not reported
2.0 (0.7, 6.2)
1.2 (0.6, 2.4)
8
2
6
15
New Hampshire (United States)
Launderers and dry cleaners
Males
Females
Not reported
5
0
Aschengrau et al. (1993). Vieira (2005)
63 bladder cancer cases, 1968-1980, Massachusetts Cancer
Registry, 852 population controls, residential history, ordinal
estimate of PCE-contaminated water (ROD) from exposure
model (Aschengrau et al.. 1993) or geographical information
system and proximity to groundwater plume (Vieira etal. 2005).
OR adjusted for sex, age at diagnosis, vital status at interview,
education, cigarette smoking, and urinary tract infection (both
studies), and, past occupational exposure (Aschengrau et al..
1993)
Burns and Swanson (1991): Swanson and Burns (1995)
2,160 histologically confirmed bladder cancer cases in men and
women, 40-84 yr old, Metropolitan Detroit Cancer Surveillance
System, 3.979 rectal or colon cancer controls, telephone
interview, longest period (usual) employed in occupation or
industry, OR adjusted for cigarette smoking, race, sex, and age at
diagnosis
Colt et al. (2004)
459 bladder cancer cases, 1994-1998, New Hampshire State
Cancer Registry, 25-74 yr, 665 populations controls,
1993-1997, occupation as exposure surrogate, OR adjusted for
5-yr age group and smoking
VO
o
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Maine, Vermont, and New Hampshire (United States)
Occupation: Laundering and dry-cleaning machine
operators and tenders
Males
Females
Not reported
0.45 (0.03, 7.46)
5
1
Industry: Laundry, cleaning and garment services
Males
Females
0.91(0.41,2.03)
1.50(0.50,4.50)
14
10
New Zealand
Textile bleaching, dyeing and cleaning machine
operators
0.81(0.19,3.54)
3
Canada, 7 Provinces
Drycleaner
1.24(0.23,6.64)
5
European Pooled Study (Denmark, France, Germany, Greece, Italy, Spain)
Launderers, dry cleaners and pressers
1.24(0.67,2.31)
19
Nordic Countries (Denmark, Finland, Norway, Sweden)
Unexposed
Dry cleaner
1.00
1.44(1.07, 1.93)e
188
93
Reference
Colt et al. (2011)
1,158 patients, aged 30-79, newly diagnosed with histologically
confirmed bladder cancer, 2001-2004, ascertained from hospital
pathology departments, hospital cancer registries and state cancer
registries, 1,402 population controls frequency matched by age
(within 5 yr), state and gender, occupational histories through
interview coded by occupation (SOC 7658) and industry (SIC
721), OR for occupation or industry category compared to other
never employed in that category, adjusted for age, race, Hispanic
ethnicity, state, smoking status, and employment in a high risk
occupation
Dryson et al. (2008)
213 bladder cancer cases, 25-70 yr, 2003-2004, New Zealand
Cancer Registry, 471 population controls, occupational title, OR
adjusted for sex, smoking, SES
Gaertner et al. (2004)
887 histologically confirmed bladder cancer, 20-74 yr, 2,847
population controls, Province Cancer Registry, mailed
questionnaire, occupational title as exposure surrogate, OR
adjusted for age, province, race, smoking status, consumption of
fruit, fried food, and coffee, and past occupational exposure.
Kogevinas et al. (2003)'
Pooled study of 3,346 male bladder cancer cases, 30-79 yr,
study-specific groups of 6,840 controls, occupational title, OR
adjusted for age, smoking, and study center
Lynge et al. (2006)
Case-control study among 46,768 Danish, Finnish, Norwegian,
and Swedish men and women employed in 1960 as laundry
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Other in dry-cleaning
Unclassifiable
Dry cleaner
Other in dry-cleaning
Unclassifiable
Dry cleaner, smoking adjusted
Relative risk
(95% CI)
1.08(0.55, 2.11)e
1.24(0.83, 1.83)e
1.69(1.18, 2.43)e'f
1.13 (0.51, 2.50)^
Not reported6
1.25 (0.79, 1.98)g
No. obs.
events
12
57
15
6
1
Dry cleaner, employment duration, 1964-1979
10yr
Unknown
1.50 (0.57, 3.96)e
2.39(1.09, 5.22)e
0.92 (0.52, 1.59)e
1.57 (1.07 2.29)e
1.97 (0.64, 6.05)e
6
10
17
53
6
Reference
worker or dry cleaner, follow-up 1970-1971 to 1997-2001, 351
bladder cancer cases, 3 controls per case randomly selected from
cohort matched on country, sex, age, calendar period at diagnosis
time, occupational task at 1970 Census proxy for exposure,
bladder cancer incidence (excluding in-situ), RR adjusted for
matching criteria
VO
to
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Germany, 5 regions
PCE, JEM
Medium exposure
High exposure
Substantial exposure
1.1(0.9, 1.3) M
1.8(1.0, 3.0) F
1.2(1.0, 1.5) M
1.0(0.6, 1.9) F
1.4(1.0, 1.9) M
0.7 (0.2, 2.5)
162
21
172
16
71
3
PCE, JTEM
Medium exposure
High exposure
Substantial exposure
0.8 (0.6, 1.2)
1.3 (0.9, 1.7)
1.8(1.2,2.7)
47
74
36
Belgium, Limburg Region
Domestic helpers, cleaners, and launderers
2.7(1.1,6.6)
14
New Jersey (United States)
Dry-cleaning workers
1.33(0.50,3.58)
7
Reference
Pesch et al. (2000b)
1,035 histologically confirmed urothelial cancer in men and
women, hospital record study, 1991-1995, 4,298 population
controls, in-person interview, JEM and JTEM for PCE, OR
adjusted for age, study center, smoking
Reulen et al. (2007)
202 histologically confirmed transitional cell carcinoma cases,
40-96 yr, Limburg Cancer Registry, 390 population controls, in-
person interview, occupational title, OR adjusted for age, sex,
smoking status, number cigarettes, years smoked, education
Schoenberg et al. (1984)
Histologically confirmed bladder cancer cases (658 Caucasian
men), 1978-1979, 21-84 yr, age-stratified population controls
(1,258 Caucasian men) identified through ROD or HCFA
register, in-person interview with questionnaire, industry and job
title surrogate exposure metric, OR adjusted for age and cigarette
smoking
VO
oo
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Montreal, Canada
Launderers and dry cleaners
Any exposure
Substantial exposure
1.6(0.9,3.1)
1.9 (0.9, 4.2)
10
7
National Bladder Cancer study
Laundry and dry cleaners, males and females
Nonsmoker
Former smoker
Current smoker
Laundry and dry cleaners, non-Caucasian males
<5 yr employment duration
>5 yr employment duration
/>-value for linear trend
Laundry and dry cleaners, females
1.31(0.85,2.03)
2.99(1.80,4.97)
3.94(2.39,6.51)
2.8(1.1,7.4)
5.3 (CI not
reported)
1.8(CInot
reported)
;? = 0.016
1.4 (0.8, 2.6)
Not reported
Not reported
Not reported
11
7
4
23
Stockholm, Sweden
Dry cleaner
1.2 (0.2, 9.2)
2
British Columbia, Canada
Laundry and dry -cleaner workers
Exposure surrogate lagged 20 yr
2.3 (0.4, 13.9)
1.8(0.3, 11.3)
5
4
Reference
Siemiatycki (1991)
Histologically confirmed bladder cancer, 1979-1985, 35-70 yr,
population control group and cancer control group, in-person
interviews, occupational title and JEM for PCE, OR adjusted
age, family income, and cigarette index, 90% CI
Silverman et al. (1990: 1989a: 1989b): Smith et al. (1985)
Histologically confirmed bladder cancer cases (2,226 men, 733
women), 1977-1978, 21-64 yr, 5,757 population controls, in-
person interview, occupational title as exposure surrogate, OR
adjusted for smoking (Silverman et al.. 1990) and employment in
other high-risk occupation (Silverman et al.. 1989b) and age, sex,
and smoking (<20/d, >20 to <40/d, >40/d (Smith et al., 1985)
Steineck et al. (1990)
Bladder cancer cases in males, birth years, 191 1-1945 and living
in County of Stockholm 1985-1987, population controls, mailed
questionnaire, occupational title as surrogate, OR adjusted for
birth year and smoking
Teschke et al. (1997)
Histologically confirmed bladder cancer cases (excluding in situ)
from British Columbia Cancer Agency in men and women,
-^
OD
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
Exposure group
Dry cleaners
Relative risk
(95% CI)
Not reported
No. obs.
events
3
Iowa, United States
Laundering and dry cleaning occupation
Males
Females
Duration of employment
<10yr
>10yr
Not reported
9.3 (0.9, 94.8)
2.1(0.1,36.9)
3/1
2/0
1/1
Reference
1990-1991, >19 yr, population controls, in-person or telephone
interviews, occupation and industry as surrogates, OR adjusted
for sex, age, cigarette smoking
Zheng et al. (2002)
Histologically confirmed in situ and invasive bladder cancer
from Iowa state health registry records in men and women,
1986-1989, 40-85 yr, population controls, in-person interview,
occupation and industry as surrogate, OR adjusted for age,
lifetime pack-years of cigarette smoking, and first-degree relative
with bladder cancer
Geographic Studies
Illinois, 8 NW counties
Winnebago County
Males
Females
0.96(0.8, l.l)h
1.39(1.1, 1.7)a
1.03 (0.8, 1.3)h
1.40(1.0, 1.9)a
250
76
96
35
Mallin (1990)
712 bladder cancer cases in Caucasian men and women,
1978-1985, residence as exposure surrogate, solvent-
contaminated municipal drinking water wells in Winnebego
County [multiple solvents including PCE, < 1-5.1 ppb],
incidence and mortality rates of U.S. population as referent (SIR,
SMR)
Meta-analysis
Laundry and dry-cleaning workers
Cohort studies
Case-control studies
1.27 (0.95, 1.71)J
0.82 (0.54, 1.25)
1.66(1.23,2.24)
Reulen et al. (2008)
-^
OD
-------
Table 4-8. Summary of human studies on tetrachloroethylene exposure and bladder cancer (continued)
""Incidence.
bFor Boice et al. (1999), Relative risks for employment duration from Poisson regression with internal referents of factory workers not exposed to any solvent
and with adjustment for date of birth, date first employed, date of finishing employment, race, and sex.
Tor Chang et al. (2005). SIR for urinary organ neoplasms given bladder cancer SIR is not identified separately.
dSmoking-corrected SIR obtained by dividing SIR by 35% of the excess of lung cancer risk (assumed proportion of risk between lung and bladder cancer
associated with smoking 20 cigarettes/d).
eln Lynge et al. (2006), odds ratio from logistic regression adjusted for country, sex, age, and calendar period at time of diagnosis.
f In Lynge et al. (2006J, odds ratio—Norway and Denmark, countries with better exposure information.
8In Lynge et al. (2006). smoking adjusted odds ratio for subjects from Norway and Sweden.
hMortality.
'In Kogevinas et al. (2003) includes the following case-control studies—Claude et al. (1988). Cordier et al. (1993). Gonzalez et al. (1989). Hours et al. (1994).
Jensen et al. (1987). Pesch et al. (2000a). Pohlabeln et al. (1999). Porru et al. (1996). Rebelakos et al. (1985). Serra et al. (2000). and Vineis et al. (1985).
Includes Andersen et al. (1999). Burns et al. (1991). Bouchardy et al. (2002). Colt et al. (2004). Gaertner et al. (2004). Schoenberg et al. (1984). Siemiatycki
(1991). Silverman et al. (1989a). Silverman et al. (1990). Steineck et al. (1990). Swanson et al. (1995). Teschke et al. (1997) Travier et al. (2002). and Zheng et
al. (2002).
HCFA = Health Care Financing Administration, JEM = job-exposure matrix, MMR = Massachusetts Military Reservation, NCI = National Cancer Institute,
PCE = tetrachloroethylene, ROD = random digit dialing, SES = socioeconomic statuts.
-------
Five studies had been previously identified based on the relative strengths of their
exposure-assessment methodology. The results from four of these studies provide additional
evidence of an association, with relative risks of 1.44 (Lynge et al., 2006), 1.5 (Blair et al., 2003)
(medium-high exposure), 4.03 (Aschengrau et al., 1993) (>90th percentile exposure), and the
exposure-response gradient observed in Pesch et al. (2000b). Although a SMR of 2.59 (95%
CI: 1.24, 4.76) was reported among workers with exposure to tetrachloroethylene and possibly
other dry-cleaning solvents (10 exposed cases), no bladder cancer deaths were observed among a
subgroup with a higher certainty of exposure only to tetrachloroethylene (Calvert et al., 2011).
Statistical analyses in all case-control studies controlled for cigarette smoking, a known
risk factor for bladder cancer. The potential effect modification by smoking history is also an
important issue but has been examined in only one study (Smith et al., 1985). In the analysis
stratified by smoking status, adjusted ORs for the association between laundry or dry-cleaning
work (based on occupational title from interview data) and bladder cancer incidence of 1.31
(95% CI: 0.85, 2.03) among nonsmokers, 2.99 (95% CI: 1.80, 4.97) among former smokers, and
3.94 (95% CI: 2.39, 6.51) among current smokers were seen.
Three studies of weaker exposure-assessment approaches observed odds ratio or
standardized incidence ratio estimates of 0.34 (95% CI: 0.07, 1.00), 1.39 (95% CI: 1.1, 1.7;
males) and 1.40 (95% CI: 1.0, 1.9; females), and 2.5 (CI not reported) for the association
between bladder cancer and employment in a manufacturing plant (Sung et al., 2007) or
residential proximity to groundwater contamination (Aschengrau et al., 1993; Mallin, 1990).
These studies carry lower weight in the analyses because of their low level of detail on
tetrachloroethylene exposure.
The Reulen et al. (2008) meta-analysis of occupational titles and bladder cancer included
14 studies reporting relative risk estimates for dry-cleaners and laundry workers. The pooled
relative risk estimate for employment in these industries was 1.27 (95% CI: 0.95, 1.71). While
Reulen et al. (2008) included many of the studies identified above, they do not include the
cohorts of Calvert et al. (2011). Blair et al. (2003). Ji et al. (2005b). and Pukkala et al. (2009). or
the case-control studies of Kogevinas et al. (2003) and Lynge et al. (2006). Other differences
between Reulen et al. (2008) and this analysis are the inclusion of Bouchardy et al. (2002), who
reported a odds ratio estimate for the association between bladder cancer and cleaning, and
personal services—a broad category that included dry cleaners, laundry workers, chimney
sweeps, hairdressers, and other cleaning occupations not included in the EPA analysis due to the
lack of data specific for dry-cleaners and laundry workers. Despite the differences in the specific
studies included in this analysis, the results are similar to that of the EPA's evaluation, indicating
a small (10-40%) increased risk.
4-97
-------
Establishment of an exposure or concentration-response relationship can add to the
weight of evidence for identifying a cancer hazard, but only limited data pertaining to exposure-
response relationships for bladder cancer and tetrachloroethylene exposure are available. As
described previously, effect estimates of 0.8 (95% CI: 0.6, 1.2), 1.3 (95% CI: 0.9, 1.7), and 1.8
(95% CI: 1.2, 2.7) for medium, high, and substantial exposure, respectively, based on ITEM
exposure data were reported in the large case-control study by Pesch et al. (2000b). Some
additional information on exposure-response relationships comes from lesser-weighted studies.
Two of the smaller studies with semi quantitative exposure surrogates observed larger effect
measures for the highest exposure category than for overall exposure. In Aschengrau et al.
(1993), the adjusted OR was 4.03 (95% CI: 0.65, 25.10) for the >90th percentile of the relative
delivered dose, compared with 1.39 (95% CI: 0.67, 2.91) for any tetrachloroethylene exposure.
Siemiatycki (1991) reported an adjusted OR of 1.9 (95% CI: 0.9, 4.2) for substantial exposure
and 1.6 (95% CI: 0.9, 3.1) for any exposure. In the third study with semiquantitative exposure
measurement, the SMR in Blair et al. (2003) was 1.5 (95% CI: 0.6, 3.1) for the medium-to-high
cumulative exposure, 1.4 (95% CI: 0.4, 3.2) for the little-to-no exposure category, and 1.3
(95% CI: 0.7, 2.4) among all cohort members (laundry and dry-cleaning union members). Other
studies examined duration of laundry or dry-cleaning work. Two studies did not observe
increasing patterns of risk with increasing employment durations as measured by census
occupation codes from two or more periods (Ji et al., 2005a: Travier et al., 2002), and one study
observed a lower risk with higher duration of laundry and dry-cleaning work based on
employment duration data collected in interviews with cases and controls [trends-value = 0.016
for the adjusted OR estimate of 5.3 for <5 years and 1.8 for >5 years duration in laundry and
drying cleaning work, respectively (Silverman et al., 1989b)1. Another study using 1960 and
1970 Census data from Nordic countries reported a nonmonotonic pattern of increasing risk, with
adjusted relative risks of 1.50, 2.39, 0.92, and 1.57 for duration of dry-cleaning work from
1964-1979 of <1, 2-4, 4-9, and >10 years, respectively, compared to subjects never employed
as a dry cleaner or in a shop with <10 employees19 (Lynge et al., 2006). For the job held in
1970, Lynge et al. (2006) relied upon a biography of dry-cleaning shop owners, the yellow pages
of local telephone books for self-employed persons, and national pension system records to
assess length of employment for Danish subjects; national pension records for Finnish subjects,20
and self-reported information using questionnaires for subjects from Norway or Sweden.
Several potential sources of exposure misclassification for these data should be noted, however,
such as would be introduced by changing employers, starting dry-cleaning work at a later time
19 Lynge et al. (2006), an analysis based only on the employment periods from 1965 through 1978, gave the
following RRs: 0-1 year = 1.43 (95% CI, 0.52-3.97); 2-4 years = 2.38 (95% CI, 1.08-5.24); 5-9 years =1.21
(95% CI, 0.58-2.50); >10 years = 2.84 (95% CI, 0.97-8.35); unknown = 2.12 (95% CI, 0.65-6.85).
20 Finnish pension records started in 1962 for dry cleaning employees and in 1970 for self-employed persons.
4-98
-------
period, employment during a time period outside the examined range or before recordkeeping
began, or imperfect recall by proxy respondents on questionnaires. Moreover, exposure duration
examined in all of these studies is a poorer surrogate than a semi quantitative or quantitative
exposure metric because it does not account for potential temporal decreases in
tetrachloroethylene intensity resulting from improved tetrachloroethylene recovery and
technological changes (Gold et al., 2008) or for variation in tetrachloroethylene concentration
across shops (Lynge et al., 2006). A fourth study that examined exposure duration and, also,
time since first employment observed statistically significant associations with both increasing
time since first employment and with increasing duration of exposure (Calvert et al., 2011).
Known risk factors for bladder cancer include smoking, aromatic amine dyes, chronic
inflammation, infection with the parasite Schistosoma heamatobium, and pelvic irradiation
(Kaufman et al., 2009). Of these identified risk factors, potential confounding related to smoking
is most important to consider in the evaluation of bladder cancer and tetrachloroethylene in
studies of occupational and residential exposures, as exposure to other known risk factors is
much less common. Statistical control for smoking effects was used in all case-control studies,
including those informing the hazard identification analysis and those contributing lesser weight
(Colt etal.. 2011: Dryson et al.. 2008: Reulen et al.. 2007: Vieira et al.. 2005: Gaertner et al..
2004: Kogevinas etal.. 2003: Zheng et al.. 2002: 2000a: Teschke et al.. 1997: Aschengrau et al..
1993: Burns and Swanson, 1991: Siemiatvcki, 1991: Silverman et al.. 1990: Steineck et al..
1990: Silverman et al.. 1989a: Silverman et al.. 1989b: Smith etal.. 1985: Schoenberg et al..
1984). Lynge et al. (2006), a case-control study with subjects from four Nordic countries,
presented smoking-adjusted and unadjusted effect measures for subjects from two countries for
which smoking histories were obtained through interviews. Adjustment made little difference
(<10%) in the magnitude of the effect measure, indicating that smoking history is not a strong
confounder of the observed risk estimates [smoking unadjusted, 1.34, 95% CI: 0.86, 2.08;
smoking adjusted, 1.25, 95% CI: 0.79, 1.98 (Lynge et al.. 2006)1.
Direct examination of possible confounders is less common in cohort studies relying on
company-supplied or census work history data compared to case-control studies where
information is obtained from study subjects or their proxies. In cohort studies, however, use of
internal controls rather than an external referent group (e.g., national mortality rates) can
minimize effects of potential confounding due to smoking or socioeconomic status, because
exposed and referent subjects are drawn from the same target population. However, only one of
the available cohort studies included an analysis using internal controls, and that study is limited
by the observation of only two bladder cancer cases in the cohort (Boiceetal., 1999). Effect of
smoking as a possible confounder may be assessed indirectly through examination of risk ratios
for other smoking-related sites such as lung cancer. Several studies observed roughly a 30%
4-99
-------
increase in lung cancer risk among dry cleaners (Calvert et al., 2011; Pukkala et al., 2009; Ji and
Hemminki, 2006a: Ji et al., 2005a: Blair et al., 2003) employed a method that assumed smoking
accounted for 35% of their lung cancer observations and adjusted the bladder cancer
standardized incidence ratio by this proportion. This method reduced slightly the effect measure
for dry-cleaner and laundry workers (smoking unadjusted, 1.27, 95% CI: 1.08, 1.48; smoking
adjusted, 1.13, 95% CI: 0.96, 1.31) (Ji et al.. 2005a). Blair et al. (2003) addressed potential
confounding by smoking and noted that if the magnitude of the difference in smoking for dry
cleaners compared with the general population is in the range of 10% of less, confounding from
smoking in their study of dry-cleaners and laundry workers was unlikely to result in increased
excess of over >20%. In the case of bladder cancer in this study, smoking may explain the
excess risk reported for overall exposure (SMR: 1.3). In contrast, the meta-analysis of Reulen et
al. (2008) examined studies that did or did not adjust for smoking and found a stronger effect
estimate with the smoking adjustment: the bladder cancer metarelative risk estimates for
launderers and bladder cancer were 1.72 (95% CI: 1.25, 2.37) in studies that adjusted for
smoking and 0.86 (95% CI: 0.59, 1.26) in studies that did not adjust for smoking. In conclusion,
while smoking may potentially confound, to a small degree, observations in some cohort studies
controlling for its effect in statistical analyses (Reulen et al., 2007; Lynge et al., 2006; Gaertner
et al.. 2004: Kogevinas et al., 2003: Zheng et al.. 2002: Pesch et al.. 2000b: Teschke et al.. 1997:
Aschengrau et al., 1993: Siemiatycki, 1991: Silverman et al., 1989a: Silverman et al., 1989b:
Smith et al., 1985), these studies do provide evidence of an association with tetrachloroethylene
or with holding a job as a dry cleaner or a laundry worker, a surrogate for tetrachloroethylene
exposure potential.
In conclusion, the pattern of results from this collection of studies is consistent with an
elevated risk for tetrachloroethylene of a relatively modest magnitude. The effect estimates from
four of the five studies with the relatively high quality exposure-assessment methodologies
provide evidence of an association, with relative risks of 1.44 to 4.03 (Calvert et al., 2011: Lvnge
et al.. 2006: Blair et al.. 2003: Pesch et al.. 2000b: Aschengrau et al.. 1993). The Lynge et al.
(2006) results were slightly higher among the subgroup from Denmark and Norway, in which the
number of unclassifiable data was negligible (relative risk 1.69, 95% CI: 1.18, 2.43). An
exposure-response gradient was observed in a large case-control study by Pesch et al. (2000b)
using a semi quantitative cumulative exposure assessment. A similar exposure-response pattern
was not observed in the study by Lynge et al. (1995). This study examined exposure duration,
however, rather than a measure that incorporated information on exposure concentration. In
addition, relative risk estimates between bladder cancer risk and ever having a job title of dry-
cleaner or laundry worker in four large cohort studies ranged from 1.01 to 1.44 (Pukkala et al.,
2009: Wilson et al., 2008: Ji et al., 2005a: Travier et al., 2002). Confounding by smoking is an
4-100
-------
unlikely explanation for the findings, given the adjustment for smoking by Pesch et al. (2000a)
and other case-control studies.
4.2.2. Animal Studies
Kidney toxicity and cancer have been observed in laboratory animals exposed to
tetrachloroethylene in multiple studies. The sections below describe studies of kidney toxicity
(refer to Section 4.2.2.1) and cancer (refer to Section 4.2.2.2). These studies are summarized in
Tables 4-9 and 4-10, respectively.
4.2.2.1. Kidney Toxicity in Animals
Tetrachloroethylene causes renal toxicity across multiple species, including several
strains of rats and mice [for reviews, refer to Cal/EPA (200IX ATSDR Q997a), NYSDOH
(1997)]. Adverse effects on the kidney have been observed in studies of animals exposed to high
concentrations of tetrachloroethylene by inhalation, oral intake, and i.p. injection. These effects
increased kidney-to-body-weight ratios, hyaline droplet formation, cast formation, glomerular
"nephrosis," karyomegaly (enlarged nuclei), and other lesions or indicators of renal toxicity.
These nephrotoxic effects mainly occurred following relatively high subchronic (400-800 ppm)
or chronic tetrachloroethylene exposures (100-200 ppm).
4.2.2.1.1. Inhalation
A long-term inhalation study examined the effects of tetrachloroethylene exposure in
male and female rats by observation throughout the lifetime of the animals (0, 300, 600 ppm,
6 hours/day, 5 days/week, for 12 months) (Rampy etal., 1978). No increase in tumors compared
to controls was observed in any animals in this study; however, an increase in mortality related
to renal failure was observed in male rats starting at 5 months exposure in the high-dose group.
No effects were observed in hematologic parameters measured (hemoglobin concentration, WBC
counts) or various urinalysis endpoints (specific gravity, pH, presence of ketones, bilirubin, or
blood, or sugar and albumin concentrations). The authors state that clinical chemistry
measurements are not useful because most animals were deceased or moribund at the end of the
study, and the study details show only measurements in a limited number of animals (1 male per
group, 5 females per group). Although the authors conclude limited tetrachloroethylene toxicity,
due to the large amount of morbidity in this study, it is difficult to make any conclusions as to the
toxicity and/or carcinogen! city of tetrachloroethylene from this study.
4-101
-------
Table 4-9. Summary of rodent kidney toxicity studies
Species/strain/
sex/number
Mouse, B6C3FJ (both
sexes, 49 or 50 of each
sex per dose group, total
of -300 mice)
Rat, F344 (both sexes,
50 of each sex per dose
group, total of -300
mice)
Mouse, Crj/BDFl (both
sexes, 50 of each sex per
dose group, total of
400 mice)
Rat, F344/DuCrj (both
sexes, 50 of each sex per
dose group, total of
400 rats)
Rat, Osborne-Mendel
(both sexes, 50 of each
sex per dose group);
Mouse, B6C3FJ (both
sexes, 50 of each sex per
dose group)
Rat, Sprague-Dawley
(both sexes, 96 per sex
per exposure group;
192 per sex per control
group)
Rat, F344; and mouse,
B6C3FJ (both sexes,
5 of each sex per group)
Mouse, Swiss-Webster
(male, 4/group)
Exposure level/duration
0, 100, 200 ppm for
104 wk, inhalation
0,200, 400 ppm for
104 wk, inhalation
0, 10, 50, 250 ppm for
110 wk, inhalation
0, 50, 200, 600 ppm for
110 wk, inhalation
0, 475, 950 mg/kg-day
(rats); 0, 536,
1,072 mg/kg-day (male
mice); 0, 386,
772 mg/kg-day (female
mice) by oral gavage in
corn oil for 78 wk,
observed for 32 wk (rats) or
12 wk (mice) following
exposure
0, 300, 600 ppm for 6 h/d,
5 d/wk for 12 mo; observed
for the lifetime of the rat
(up to 3 1 mo total)
0, 200 (28 d only), and
400 ppm (14, 21,28 d) for
6 h/d, inhalation
0, 150, 500, and
1,000 mg/kg-day, aqueous
gavage for 30 d
Effects
Karyomegaly and cytomegaly of the
proximal tubules in all exposed mice;
nephrosis was observed in exposed females,
casts increased in all exposed males and in
high-dose females
Karyomegaly and cytomegaly of the
proximal tubules in all exposed rats
Increased relative kidney weights and
karyomegaly in the proximal tubules in 250
ppm exposed male and female mice; atypical
tubular dilation in 250 ppm male and female
mice but was not statistically significant
Increased relative kidney weights and
karyomegaly in the proximal tubules in 200
and 600 ppm exposed male and female rats;
atypical tubular dilation in 600 ppm male
and female rats; exacerbation of chronic
renal disease in male rats only at 600 ppm
Toxic nephropathy observed in all exposed
animal groups, with an increased incidence
in rats as compared to mice
Increased mortality related to renal failure in
male rats exposed to 600 ppm starting at
5 mo of exposure
Analysis in mice was limited to pooled
tissue but showed slight increases in
(3-oxidation in mouse kidney; modest
increases in PCO observed in male rat
kidneys at 200 ppm for 28 d only, but
elevated in female rat kidneys at all doses
and times
No kidney injury or dysfunction was
observed in this study
Reference
NTP (1986)
NTP (1986)
JISA (1993)
JISA (1993)
NCI (1977)
Rampy et al.
(1978)
Odum et al.
(1988)
Philip et al.
(2007)
4-102
-------
Table 4-9. Summary of rodent kidney toxicity studies (continued)
Species/strain/
sex/number
Rat, Wistar (female
only, 10 rats in each
control group; 5 rats in
each treatment group)
Rat, F344 (male only,
5/group) and Mouse,
B6C3FJ (male only,
5/group)
Rat, F344 (both sexes)
Rat, F344 (both sexes,
12 per group)
Mouse, Swiss (both
sexes, 6 groups of 6 each
(1996): male only; 8
groups of 6 each (2001))
Rat, F344 (both sexes)
and Mouse, B6C3FJ
(both sexes) (10 per
group for oral studies, 5
per group for inhalation
studies)
Rat, Sprague-Dawley
(both sexes, 20 per
group)
Rat, Sprague-Dawley
(male only, 4 per group)
Exposure level/duration
0, 600, and 2,400
mg/kg-day for 32 d, corn
oil gavage; alone or in
combination with other
compounds
(trichloroethylene,
hexachloro- 1 ,2-butadiene,
1,1,2-trichloro-
3,3,3 -trifluoropropene)
0 or 1,000 mg/kg-day for
10 d, corn oil gavage
0 or 1,000 mg/kg-day for
10 d, corn oil gavage
0, 500 mg/kg-day daily for
4 wk, corn oil gavage
0 or 3,000 mg/kg-day for
15 d, sesame oil gavage
0, 1,000, or 1,500
mg/kg-day daily by corn oil
gavage for 42 d; 0 or
1,000 ppm for 10 d
0, 14, 400, or 1,400
mg/kg-day for 90 d
0, 115, 230umol/kgof
TCVC or TCVCS in saline
by one i.p. injection,
sacrificed 24 h
postexposure
Effects
Relative kidney weight was increased on
exposure to PCE alone and in combination
with other nephrotoxicants; nephrotoxic
effects noted at high dose (urea, total protein,
albumin, NAG); karyomegaly was also
observed in high dose animals
Increased kidney weight in exposed rats;
increased PCO activity in all exposed mice
Increases in o2u-hyarine droplets in exposed
male, but not female, rats, correlated with
increased cell proliferation and protein
droplet nephropathy
Increases in o2u-hy aline accumulation in
proximal tubule cells
Significant increase in kidney weight;
decreased blood glucose (glucose effects
mitigated by coexposures to 2-deoxy-
D-glucose and vitamin E [1996])
Decreased membrane-bound
Na+K+-ATPases and Mg2+-ATPases activity
but increased Ca-ATPase activity; mitigated
by coexposure to 2-deoxy-D-glucose and
vitamin E, and taurine; hypercellular
glomeruli in PCE-exposed only
Accumulation of o2u-globulin in proximal
tubules of male rats; nephrotoxicity also
observed in male rats (formation of granular
tubular casts and evidence of tubular cell
regeneration)
Inhalation exposure demonstrated formation
of hyaline droplets in kidneys of male rats
Increased kidney weight observed in
exposed animals; nephrotoxicity observed at
400 mg/kg-day
High-dose exposed animals showed visible
kidney necrosis; all other rats showed
histological markers for mild acute tubular
necrosis (TCVC) or severe acute tubular
necrosis (TCVCS); prior exposure to AOAA
increased toxicity
Reference
Jonker et al.
(1996)
Goldsworthy
and Popp
(1987)
Goldsworthy
et al. (1988)
Bergamaschi
et al. (1992)
Ebrahim et al.
(2001: 1996)
Green et al.
(1990)
Hayes et al.
(1986)
Elfarra et al.
(2007)
4-103
-------
Table 4-10. Kidney tumor incidence in laboratory animals exposed to
tetrachloroethylene
Bioassay
NCI (1977)"
B6C3FJ mice
Gavage:
5d/wk,
78 wk
NCI (1977)"
Osborn-Mendel rats
Gavage:
5d/wk,
78 wk
NTP(1986)
B6C3FJ mice
Inhalation:
6h/d,
5d/wk,
104 wk
NTP(1986)
F344/N rats
Inhalation:
6h/d,
5d/wk,
104 wk
JISA (1993)
Crj:BDFl mice
Inhalation:
6h/d,
5d/wk,
104 wk
JISA (1993)
F344/DuCrj rats
Inhalation:
6h/d,
5d/wk,
104 wk
Doses/exposures
Administered
Vehicle control
450 mg/kg-day
900 mg/kg-day
Vehicle control
300 mg/kg-day
600 mg/kg-day
Vehicle control
500 mg/kg-day
1,000 mg/kg-day
Vehicle control
500 mg/kg-day
1,000 mg/kg-day
Oppm
100 ppm
200 ppm
Oppm
100 ppm
200 ppm
Oppm
200 ppm
400 ppm
Oppm
200 ppm
400 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oppm
50 ppm
200 ppm
600 ppm
Oppm
50 ppm
200 ppm
600 ppm
Continuous equivalent
0
332 mg/kg-day
663 mg/kg-day
0
239 mg/kg-day
478 mg/kg-day
0
471 mg/kg-day
941 mg/kg-day
0
474 mg/kg-day
974 mg/kg-day
0
18 ppm
36 ppm
0
18 ppm
36 ppm
0
36 ppm
72 ppm
0
36 ppm
72 ppm
0
1.8 ppm
9.0 ppm
45 ppm
0
1.8 ppm
9.0 ppm
45 ppm
0
9 ppm
36 ppm
108 ppm
0
9 ppm
36 ppm
108 ppm
Sex
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Tumor incidence (%)
Kidney adenomas and carcinomas
0/20 (0)
1/49 (2)
0/48 (0)
0/20 (0)
0/48 (0)
0/45(0)
3/20 (5)
1/49 (2)
0/50 (0)
0/20 (0)
0/50 (0)
1/50 (2)
0/49 (0)
1/49 (2)
0/50 (0)
0/48 (0)
0/50 (0)
0/48 (0)
1/49 (2)
3/47 (6)
4/50 (8)
0/50 (0)
0/50 (0)
0/50 (0)
0/50 (0)
1/50 (2)
1/50 (2)
0/50 (0)
0/50 (0)
0/47 (0)
0/49 (0)
0/50 (0)
1/50 (2)
2/50 (4)
1/50 (2)
2/50 (4)
1/50 (2)
0/50 (0)
0/50 (0)
1/50 (2)
4-104
-------
Acute, subchronic, and chronic exposures to tetrachloroethylene were examined in male
and female F344 rats and B6C3Fi mice (NTP, 1986). Single exposure studies and 14-day
studies were performed, but no kidney effects were observed, with the first kidney effects
observed in the subchronic (13 week) study. Groups of 10 rats and mice of each sex were
exposed to air containing tetrachloroethylene for 6 hours/day, 5 days/week, for 13 weeks (0, 100,
200, 400, 800, or 1,600 ppm). Some rats in the high-dose group died before the end of the
studies (4/10 male, 7/10 female), but no kidney effects were observed. In mice, 2/10 males and
4/10 females in the high-dose group died before the end of the studies, and karyomegaly (nuclear
enlargement) of the renal tubule epithelial cells was observed in all but the lowest dose group.
Toxicity was observed in a 2-year cancer bioassay performed on groups of 50 F344 rats
of each sex (0, 200, or 400 ppm tetrachloroethylene), or groups of 49 or 50 mice (0, 100, or 200
ppm tetrachloroethylene) exposed for 6 hours/day, 5 days/week, for 103 weeks (NTP, 1986).
Karyomegaly and cytomegaly changes were observed in both sexes of rats at all doses but not in
unexposed controls. These lesions were present primarily in the proximal convoluted tubules of
the inner half of the cortex but not limited to this area. In mice, nephrosis (generally defined as
noninflammatory degenerative disease of the kidney) was observed in exposed females, casts
(cylindrical structures formed from cells and protein released from the kidney) were increased in
exposed male and high-dose females, and karyomegaly of the tubular cells was observed in all
dosed mice, with severity of lesions being dose related. Therefore, the LOAEL for renal toxicity
reported in both mice and rats in this study is 100 ppm (678 mg/m3) for inhalation exposure in
mice and 200 ppm (1,356 mg/m3) in rats (NTP. 1986).
Nephrotoxicity was observed in a second, 2-year inhalation cancer bioassay also
performed in 50 male and female Fischer rats (0, 50, 200, or 600 ppm) and Crj:BDFl mice (0,
10, 50, or 250 ppm) in each treatment group (6 hours/day, 5 days/week, for 104 weeks) (JISA,
1993). Survival compared to controls was decreased in all high-dose exposure groups, which
was believed to be treatment related. Relative kidney weight was increased in male and female
rats exposed to tetrachloroethylene (200 or 600 ppm) and in male and female mice (250 ppm).
Karyomegaly in the proximal tubules of the kidneys was observed among males and females
(200 and 600 ppm in male rats [23/50 and 48/50]; 600 ppm in female rats [18/50]; 50 and 250
ppm in male mice [6/50 and 38/50]; 250 ppm in female mice [49/50]), and an increase in atypical
tubular dilation of the proximal tubules [male and female rats, 600 ppm (24/50 males, 6/50
females) and exacerbation of chronic renal disease in male rats only (600 ppm) was observed
with tetrachloroethylene exposure (JISA, 1993). Atypical tubular dilation was also observed in
mice but was not statistically significant (250 ppm in male mice [1/50] and female mice [6/50]).
The role of peroxisome proliferation in tetrachloroethylene-induced kidney toxicity and
cancer was examined in male and female F344 rats and B6C3Fi mice exposed to
4-105
-------
tetrachloroethylene by inhalation (400 ppm, 6 hours/day, for 14, 21, or 28 days, or 200 ppm,
6 hours/day, for 28 days) in a study by Odum et al. (1988). Five animals per group were
exposed. Insufficient mouse kidney tissue limited the analysis to pooled samples. Slight
increases were observed in p-oxidation in mouse kidney (maximum 1.6-fold increase at 21 days,
400 ppm exposure). Modest palmitoyl-CoA oxidation (PCO) increases were observed in the
kidney of male rats at 200 ppm at 28 days (1.3-fold) but not 400 ppm at 14, 21, or 28 days. In
female rat kidney, PCO was elevated (approximately 1.6-fold) at all doses and times. However,
peroxisome proliferation was not observed in rat or mouse kidney upon microscopy, suggesting
that this does not play a role in kidney carcinogenesis. Short-term inhalation exposure to 1,000
ppm tetrachloroethylene for 10 days resulted in the formation of hyaline droplets in the kidneys
of male rats. Although granular casts and tubule cell regeneration were not observed, the time
period may have been too short to allow progression to this stage (Green etal., 1990).
4.2.2.1.2. Oral
Hayes et al. (1986) reported renal effects in rats exposed to 400 mg/kg-day
tetrachloroethylene in drinking water for 90 days. Tetrachloroethylene was administered in the
drinking water at 14, 400, and 1,400 mg/kg per day for 90 days, with no deaths reported before
the end of the study. Increased kidney weight was observed.
A lifetime animal carcinogenicity study in which tetrachloroethylene was administered to
50 of each sex of Osborne-Mendel rats and B6C3Fi mice by oral gavage in corn oil for 78 weeks
resulted in clear evidence of kidney toxicity in both species (NCI, 1977). The TWA doses
(mg/kg-day) used in the bioassay were 471 and 941 for male rats, 474 and 949 for female rats,
536 and 1,072 for male mice, and 386 and 772 for female mice. Animals were observed for 32
weeks (rats) or 12 weeks (mice) following the last dose. Toxic nephropathy was observed in
almost all test animals, with a high incidence observed in treated rats, including those that died
early in the study (as early as Week 20 in male rats, Week 28 in female rats). Similar results
were observed in exposed mice, with no nephropathy observed in control mice. Therefore, the
LOAEL for renal toxicity following oral exposure is 471 mg/kg-day in male rats and
474 mg/kg-day in female rats based on toxic nephropathy. The LOAEL for mice is
536 mg/kg-day for males and 386 mg/kg-day in females based on toxic nephropathy.
In a study by Jonker et al. (1996), tetrachloroethylene nephrotoxicity was observed in
female Wistar rats administered tetrachloroethylene (600 or 2,400 mg/kg-day) in corn oil by
daily oral gavage for 32 days. Relative kidney weight was increased upon exposure to
tetrachloroethylene alone and in combination with other nephrotoxicants (trichloroethylene
[TCE], hexachloro-l,2-butadiene, and l,l,2-trichloro-3,3,3-trifluoropropene [TCTFP]). One
high-dose animal died as a result of tetrachloroethylene treatment, and one animal exposed to the
4-106
-------
high-dose combination of TCE, tetrachloroethylene, and TCTFP also died as a result of
treatment. Nephrotoxic effects were noted at 2,400 mg/kg. Significant changes were observed
following exposure to tetrachloroethylene at 2,400 mg/kg-day in all clinical chemistry markers
related to kidney function (urea, total protein, albumin, NAG) as measured in the urine at the end
of Week 1 or Week 4 except for urinary density, glucose, and creatinine. Karyomegaly was also
observed at the high dose (2,400 mg/kg-day) in four of five animals exposed (p < 0.01) (Jonker
etal.. 1996).
Philip et al. (2007) exposed male 6-7-week-old Swiss Webster mice via aqueous gavage
to three dose levels (150, 500, and 1,000 mg/kg-day) for 30 days. At the highest exposure,
mortality was 10% due to apparent CNS toxicity (tremors and ataxia). Neither kidney injury nor
dysfunction was observed following tetrachloroethylene exposure during the course of this study.
Goldsworthy and Popp (1987) administered tetrachloroethylene (1,000 mg/kg-day) by
corn oil gavage to 5 male F344 rats and 5 male B6C3Fi mice for 10 days. In
tetrachloroethylene-exposed rats, PCO was modestly although not significantly elevated in the
liver (1.4-fold increase) and kidney (1.7-fold increase). In mice, tetrachloroethylene exposure
increased PCO activity 4.3-fold in liver and by 2.3-fold in kidney. Relative liver weight was
increased in rats and mice with tetrachloroethylene exposure, but relative kidney weight was
unaffected. A comparison of corn oil with methyl cellulose revealed no effect of the gavage
vehicle on tetrachloroethylene-induced PCO. A less-than-additive effect of trichloroethylene
(1,000 mg/kg) administered together with tetrachloroethylene on PCO induction was seen.
Oral administration of tetrachloroethylene in sesame oil (3,000 mg/kg-day for 15 days) to
male and female albino Swiss mice caused a significant increase in kidney weight (p < 0.001)
and a decrease in blood glucose levels (p < 0.01) as compared to control animals exposed to
sesame oil alone as well as increases in glomerular nephrosis (Ebrahim et al., 1996). This study
was designed to give support to the beneficial effect of 2-deoxy-D-glucose (2DG) and vitamin E
on tetrachloroethylene-induced kidney damage. Based on previous experimental mouse tumor
studies, administration of 2DG or vitamin E is hypothesized to have a beneficial effect on
tetrachloroethylene-induced kidney damage, either by inhibition of tumor growth (2DG) or the
auto-catalytic process of lipid peroxidation (vitamin E). In this study, concurrent administration
of 2DG (500 mg/kg-day i.p.) or vitamin E (400 mg/kg-day oral gavage) prevented
tetrachloroethylene-induced biochemical and pathological alterations. Tetrachloroethylene
exposure alone led to a decrease in blood glucose levels, which was returned to near normal with
concomitant exposure to 2DG and vitamin E. Elevated levels of glycolytic and gluconeogenic
enzymes following exposure to tetrachloroethylene were also observed to return to near normal
with exposure to 2DG and vitamin E. Histopathology of the kidney showed hypercellular
glomeruli following exposure to tetrachloroethylene, but this was not observed in animals treated
4-107
-------
with tetrachloroethylene and 2DG, or tetrachloroethylene and vitamin E (Ebrahim et al., 1996).
A follow-up study by this group further examined the potential protective properties of 2DG and
vitamin E as well as taurine against tetrachloroethylene-induced membrane damage (Ebrahim et
al., 2001). This study exposed male albino Swiss mice to the same doses used in the previous
study with the addition of a taurine-exposed group (tetrachloroethylene in sesame oil
3,000 mg/kg-day for 15 days orally by intubation; tetrachloroethylene plus 2DG 500 mg/kg-day
by i.p. injection once a day for 15 days; tetrachloroethylene plus vitamin E 400 mg/kg-day by
oral intubation once a day for 15 days; and tetrachloroethylene plus taurine 100 mg/kg-day by
oral intubation once a day for 15 days). As compared to control cells in the kidney, membrane-
bound Na+K+-ATPases and Mg2+-ATPases activity was significantly decreased (p < 0.001),
while Ca-ATPases activity was increased (p < 0.001), following exposure to tetrachloroethylene
alone. These levels remained near normal in the animals exposed to tetrachloroethylene along
with 2DG, vitamin E, or taurine. This return to normal levels following exposure to vitamin E
and taurine may be due to their antioxidant abilities, and reduced oxidative stress in exposed
cells.
Goldsworthy et al. (1988) observed increases in a2u-hyaline droplets in exposed male but
not female F344 rats following 10 days of gavage with 1,000 mg/kg tetrachloroethylene. This
finding was correlated with both protein droplet nephropathy (crystalloid accumulation) and
increases in cellular proliferation. Cell replication was enhanced in the male rats specifically in
damaged P2 segments, suggesting a link between the a2u-globulin accumulation and kidney
tumors. These investigators reported similar findings for pentachloroethane in the same study,
but at a dose of 150 mg/kg for 10 days. Trichloroethylene has a similar structure but did not
cause any a2u-accumulation or increase in protein droplets, nor did it stimulate cellular
proliferation in either male or female rats in this study when a dose of 1,000 mg/kg was
administered for 10 days. Bergamaschi et al. (1992) also demonstrated a2u-accumulation in P2
segments of rat proximal tubule cells resulting from a daily exposure of rats to 500 mg/kg
tetrachloroethylene in corn oil for 4 weeks.
In short-term, high-dose studies, Green et al. (1990) found that the oral administration of
1,000 to 1,500 mg/kg of tetrachloroethylene daily for up to 42 days caused an accumulation of
a2u-globulin in the proximal tubules of male rats. The animals were sacrificed within 24 hours
of the last dose of tetrachloroethylene. The effect was accompanied by evidence of
nephrotoxicity, with the formation of granular tubular casts and evidence of tubular cell
regeneration. These effects were not observed in female rats or in mice.
4-108
-------
4.2.2.1.3. Intraperitoneal injection
The role of the glutathione metabolites, particularly TCVC and TCVCS, in kidney
toxicity was examined by Elfarra et al. (2007) in vivo. This study exposed two groups of four
male Sprague-Dawley rats to a single i.p. injection of TCVC or TCVCS (115 or 230 umol/kg in
saline). Animals were sacrificed 24 hours following exposure. Serum was analyzed for BUN,
and urine samples were analyzed for GGTP activity as markers of nephrotoxicity. Rats exposed
to the high-dose of TCVCS showed visible signs of kidney necrosis, while all other exposed
groups did not. Histologically, kidneys from rats exposed to low-dose TCVC or TCVCS showed
slight-to-mild acute tubular necrosis. Analysis of kidneys at 24 hours postexposure showed
mild-to-moderate acute tubular necrosis in animals exposed to high-dose (230 umol/kg) TCVC,
and severe tubular necrosis in animals exposed to high-dose (230 umol/kg) TCVCS. Similar to
the pattern of toxicity described above, significant increases in BUN (fourfold) were observed in
rats exposed to 230 umol/kg TCVCS as compared to control, but no significant increases were
observed following exposure to TCVC. Variable increases were observed following exposure to
TCVC or TCVCS in urine glucose levels and GGTP activity. A second part of this experiment
involved a preexposure to a p-lyase inhibitor (AOAA) (500 umol/kg bw) by i.p. injection
30 minutes prior to administration of 230 umol/kg TCVC. Exposure to AOAA prior to exposure
to TCVCS resulted in increased toxicity. In a third study, three groups of four rats were exposed
to saline, TCVC, or TCVCS (230 umol/kg) and sacrificed 2 hours after administration. The
kidneys were removed at sacrifice and examined for NPT and NPT disulfide concentrations as a
measure of thiol status in the kidney. Although no changes were observed in NPT status,
histological examination of these kidneys showed scattered foci of mild acute tubular necrosis
(TCVC) or widespread acute tubular necrosis, intratubular casts, and interstitial congestion and
hemorrhage (TCVCS). These results suggest that while both TCVC and TCVCS are
nephrotoxicants, TCVCS is more potent than TCVC.
In summary, exposure to tetrachloroethylene from all routes studied (oral, inhalation, i.p.)
led to nephrotoxicity in multiple strains of rats and mice. These studies demonstrate
karyomegaly, increased kidney weights, and atypical tubular dilation following subchronic high-
dose exposures or lower dose chronic exposures. Limited studies have also examined the
potential role for peroxisome proliferation or a2u-globulin in nephrotoxicity. Exposure to
tetrachloroethylene glutathione conjugation metabolites led to similar effects in rats (mice not
tested). Further, studies examining the impact of concomitant antioxidant exposures with
tetrachloroethylene in mice suggest a role for oxidative stress in tetrachloroethylene-induced
nephrotoxicity.
4-109
-------
4.2.2.2. Kidney Cancer in Animals
4.2.2.2.1. Inhalation
In the studies conducted by NTP [(1986), described above], groups of 50 male and
50 female F344/N rats were exposed for 6 hours/day, 5 days/week, for 103 weeks by inhalation
to atmospheres containing 0, 200, or 400 ppm tetrachloroethylene. Tubule cell hyperplasia was
observed in male rats (control, 0/49; low dose, 3/49; high dose, 5/50) and in one high-dose
female rat. Renal tubule adenomas and adenocarcinomas were observed in male rats (control,
1/49; low dose, 3/49; high dose, 4/50). In the same study (doses described above), one renal
tubule adenocarcinoma was observed in a low-dose male mouse, but no other neoplastic lesions
were observed.
The spontaneous incidence rate for renal tubule tumors in F344/N rats, the strain used in
the NTP bioassay, as well as for other rat strains reported by NTP, was less than 1%. Thus, the
appearance of tubule neoplasms in 8% of the treated animals in the NTP study (low-dose and
high-dose groups combined) provided convincing evidence of a treatment-related effect
Solleveld etal., 1984; Goodman et al., 1979). Also notable is the fact that no malignant renal
tubule neoplasms had ever been observed in any control rats examined by NTP—including
chamber controls from the performing laboratory and the untreated controls and vehicle controls
from gavage studies—whereas two of the tumors observed in high-dose animals in the NTP
study were carcinomas. The probability of two rare carcinomas appearing by chance in a group
of 50 animals has been calculated to be less than 0.001.21 In addition, when compared with
historical control incidences of renal tubule tumors at the NTP, a statistically significant dose-
related positive trend exists, and tumor incidences in both low-dose and high-dose groups are
significantly elevated. Standard statistical analyses of tumor incidence data did not reveal a
significant increase in kidney tumors, and the tumor incidence is not statistically significant
when compared with concurrent controls; however, when the incidences of tubule cell
hyperplasia and neoplasms and tumor severity are all considered, a dose-response relationship is
apparent.
No increase in renal cell cancers was observed in a second 2-year inhalation cancer
bioassay that was also performed in 50 male and female Fischer rats (0, 50, 200, or 600 ppm)
and Crj:BDFl mice (0, 10, 50, or 250 ppm) in each treatment group (6 hours/day, 5 day/week,
for 104 weeks) (JISA, 1993). Survival compared to controls was decreased in all high-dose
exposure groups, which is believed to be treatment related. Renal cell adenoma was observed in
male rats (1/50, control; 2/50, 50 ppm; 1/50, 200 ppm; 2/50, 600 ppm) and male mice (1/50, 50
ppm) but only in control female rats (1/50, control) and not in exposed female mice. Renal cell
21 Assuming a binomial probability distribution, a background rate of 0.2%, and a sample size of 50 animals.
4-110
-------
carcinoma was not observed in male rats or female mice, but was observed in the high-dose
female rats (1/50, 600 ppm) and male mice (1/50, 50 ppm). As described above for the NTP
study (1986), these tumors are rare in Fischer rats, but the reported results are similar to those
historical control rates for this study group (JISA, 1993).
The study authors reported a slight increase in renal tumors with tetrachloroethylene
exposure in a study reporting increased mortality related to renal failure in male rats starting at
5 months exposure in the high-dose group (Rampy et al., 1978). This lifetime observation study
exposed male and female rats to 0, 300, 600 ppm, 6 hours/day, 5 days/week, for 12 months
(Rampy etal., 1978). The authors stated that most animals were deceased or moribund at the
end of study, rendering difficult clear conclusions regarding the renal carcinogenicity of
tetrachl oroethy 1 ene.
4.2.2.2.2. Oral
No significant increased incidence of neoplastic lesions was observed in treated rats
following oral exposure to tetrachloroethylene in a lifetime carcinogenicity bioassay (NCI, 1977;
doses described above). However, a high rate of death occurred in the high-dose groups of both
sexes, so the authors of the study determined carcinogenicity could not be evaluated. Only one
kidney tumor was observed in mice in this study (high dose; doses described above), but this was
a tumor that had metastasized from the liver.
In summary, an increase in rare kidney tumors was reported in one inhalation cancer
bioassay of tetrachloroethylene (0, 200, or 400 ppm) in F344/N rats (NTP. 1986). The JISA
(1993) rat inhalation bioassay of tetrachloroethylene (50, 200, and 600 ppm) reported no
treatment-related increase in the incidence of kidney tubular cell adenoma or carcinoma in
excess of that in the concurrent or historical control animals at administered concentrations.
Another inhalation study, the interpretation of which is limited by high morbidity and mortality,
reported a slight increase in renal tumors in male S-D rats (Rampy et al., 1978). Although the
renal tumors were not significantly increased compared with controls, morbidity related to renal
failure was increased in male rats beginning at 5 months of exposure. The NCI (1977) oral
gavage bioassay of tetrachloroethylene (0, 475, 950 mg/kg-day) reported a high rate of death in
the high-dose groups of both sexes, and, thus, carcinogenicity could not be evaluated in this
study.
Other evidence supporting the conclusion of renal carcinogenicity of tetrachloroethylene
includes low incidences of tubule neoplasms in male rats in NTP bioassay s of other chlorinated
ethanes and ethylenes (NTP. 1990a, 1989. 1988. 1987). In particular, the closely related
compound trichloroethy 1 ene also induces low increases in the incidence of rare renal tumors in
rats and in humans (U.S. EPA. 201 Ib).
4-111
-------
4.2.2.2.3. In vitro
Lash et al. (1998) examined the role of glutathione conjugation of tetrachloroethylene in
rats and mice in isolated renal cortical cells and hepatocytes from male and female F344 rats.
All cells were exposed to tetrachloroethylene (0.5, 1, or 2 mM) and assayed for TCVG formation
at 0, 15, 30, and 60 minutes. This study demonstrated that GSH metabolites from
tetrachloroethylene are formed in kidney cells as well as hepatocytes in both species; however,
the amount of TCVG produced varied depending on sex, species, and tissue assayed. TCVG
formation was higher in male rats and mice as compared to their female counterparts and was
also higher in hepatocytes as compared to kidney cells. Although rats are more susceptible to
nephrocarcinogenicity as compared to mice (refer to Section 4.5.2.2), isolated mouse kidney and
liver cells had a greater amount of TCVG formation (7- to 10-fold and 2- to 5-fold, respectively)
as compared to rat cells (Lash et al., 1998). To further examine the species- and sex-dependent-
differences in tetrachloroethylene cytotoxicity, Lash et al. (2002) measured acute cytotoxicity
following exposure to tetrachloroethylene or TCVG (0.1 to 10 mM) in isolated rat kidney cells
and renal mitochondria from rats and mice. Exposure to tetrachloroethylene or TCVG led to a
marked increase in LDH release in isolated kidney cells from male but not female rats, but no
significant effects were observed in rat hepatocytes from either gender (Lashetal., 1998).
Isolated mitochondria from rats and mice showed a pattern of sensitivity similar to the kidney
cell effects, with increased inhibition of respiration in isolated mitochondria from male rats as
compared to their female counterparts. Inhibition of respiration was observed equally in male
and female mice exposed to tetrachloroethylene or TCVG. The results of this in vitro study
support those of the in vivo studies, which demonstrate increased nephrotoxicity in male rats
following exposure to tetrachloroethylene or TCVG.
Lash et al. (2007) examined the effect of modulation of renal metabolism on toxicity of
tetrachloroethylene in isolated cells and microsomes from male F344 rat kidney and liver.
Oxidative-dependent metabolism of tetrachloroethylene was more than 30-fold increased in liver
microsomes than in kidney. Pretreatment of rats with a P450-inhibitor had little to no effect on
the tetrachloroethylene metabolism in either kidney or liver. Pretreatment of rats with a P450
inducer increased tetrachloroethylene metabolism by over twofold in the kidney microsomes,
with no effect observed in liver. Following exposure to modulating chemicals, lactate
dehydrogenase (LDH) was measured as a marker of cytotoxicity, and the presence of specific
metabolites was documented (TCVG, TCOH, and CH). Tetrachloroethylene metabolism in
kidney cells was slightly (but significantly) increased by the nonspecific inhibitors of P450s but
not affected by the pretreatment with the CYP2E1-specific inhibitor. Increased cytotoxicity in
kidney cells was observed following exposure to tetrachloroethylene (2 or 10 mM, 3 hours), and
this was not affected by pretreatment with CYP inhibitors or inducers. However, increases in
4-112
-------
GSH concentrations in the kidney cells led to increased cytotoxicity following exposure to
tetrachloroethylene, but no effect was observed following pretreatment with GSH inhibitors.
The results of this study highlight the role of different bioactivation pathways needed in both the
kidney and the liver, with the kidney effects being more affected by the GSH conjugation
pathways metabolic products.
Tetrachloroethylene effects in kidney cells have also been demonstrated in a variety of
genotoxicity assays. Exposing kidney cells and/or microsomal fractions from kidneys to
tetrachloroethylene or its some of its metabolites led to low levels of DNA binding (Mazzullo et
al., 1987), micronuclei induction (Wangetal., 2001), single-stranded DNA breaks (Walles,
1986), unscheduled DNA synthesis (Vamvakas et al., 1989c), and gene mutations (Vamvakas et
al., 1989d: Vamvakas et al., 1987; Dekantetal., 1986a). Negative studies were observed in
kidney cells from exposed animals for DNA damage (Cederberg et al., 2010a: Potter et al.,
1996), and DNA adduct formation (Toraason et al., 1999).
Limited DNA binding to calf thymus DNA was observed in the presence of microsomal
fractions from mice and rats (Mazzullo et al., 1987). Binding to DNA in the in vitro study
increased in the presence of microsomal fractions from both mouse and rat liver, but not kidney,
lung, or stomach. Cytosolic fractions from rat and mouse liver, kidney, lung, and stomach, all
induced binding of tetrachloroethylene to calf thymus DNA, with enzymes from both mouse and
rat livers and mouse lung being the most efficient.
Wang et al. (2001) examined micronuclei induction following exposure to
tetrachloroethylene (-63 ppm in culture medium at peak) in vitro in a closed system. Chinese
hamster ovary (CHO-K1) cells were plated in a petri dish surrounding a glass dish of
tetrachloroethylene and incubated for 24 hours. Tetrachloroethylene exposure led to a dose-
dependent significant increase in micronuclei induction (p < 0.001) (Wang et al., 2001).
Vamvakas et al. (1989a) reported concentration-related increases in unscheduled DNA
synthesis (UDS) in LLC-PK1 (a porcine kidney cell line) exposed to TCVC, with the effect
abolished by a p-lyase inhibitor. This effect was observed at exposure to 5 x 10~6-10~5 M
TCVC for 24 hours.
TCVG produced from tetrachloroethylene in isolated perfused rat liver and excreted into
bile, in the presence of a rat kidney fraction, was mutagenic in Salmonella, as was purified
TCVG (Vamvakas et al., 1989d). This study performed the Ames assay in Salmonella
typhimurium TA100, TA98, and TA2638 with tetrachloroethylene, TCVG, and bile from liver
perfusate following tetrachloroethylene exposure in rats and demonstrated that the
GST-metabolites or tetrachloroethylene in the presence of bile containing GST led to gene
mutations in S. typhimurium TA100. Dreessen (2003) also demonstrated for TCVG an
unequivocal dose-dependent mutagenic response in the TA100 strain in the presence of the rat
4-113
-------
kidney S9-protein fraction; TCVC was mutagenic without metabolic activation in this strain. In
a separate study, the tetrachloroethylene metabolite TCVC (1-10 nmol/plate) was also positive
in Salmonella strains TA98 and TA100 but not strain TA2638, and inhibition of p-lyase activity
was blocked by addition of amionoxyacetic acid (AOAA) (Dekant et al., 1986a). A subsequent
study from this same group indicated that Salmonella also was capable of deacetylating the
urinary metabolite NAcTCVC (50-100 nmol/plate) when TA100 showed a clear positive
response in the Ames assay without exogenous activation (Vamvakas et al., 1987).
In summary, the limited in vitro studies performed in kidney cells exposed to
tetrachloroethylene or its GSH conjugation metabolites demonstrate an increase in cytotoxicity.
This cytotoxic effect was sex- and species-dependent, with increases observed in male rats and
mice compared to their female counterparts, with rats showing the most cytotoxicity. Limited
genotoxicity studies demonstrated the potential for tetrachloroethylene mutagenicity in
Salmonella strains in the presence of the kidney S9 fraction, or in Salmonella exposed to
GSH-conjugation metabolites (TCVC, TCVG, or NacTCVC) without activation.
4.2.3. Summary of Kidney Effects in Humans and Animals
Taken together, the epidemiologic studies support an association between inhalation
tetrachloroethylene exposure and chronic kidney disease, as measured by urinary excretion of
renal proteins and ESRD. The elevated urinary RBP levels observed in two studies (Verplanke
et al., 1999; Mutti etal., 1992) and lysozyme or p-glucuronidase in Franchini et al. (1983)
provide some evidence for effects to the proximal tubules from tetrachloroethylene exposure.
Exposures in the studies that observed renal toxicity were 1.2 ppm, 10 ppm, and 15 ppm
(means), representing an observational LOAEL for human kidney effects. An exposure-response
relationship was reported in one study (Trevisan et al., 2000) but not in the other human studies
that examined renal function, an important limitation of the available data. However, as pointed
out by Mutti et al. (1992), this is a common finding among solvent-exposed populations, and
inadequate definition of the dose metric most likely contributes to the absence of exposure-
response relationships. Calvert et al. (2011) supports an association between inhalation
tetrachloroethylene exposure and ESRD, particularly hypertensive ESRD. They observed a
twofold elevated incidence (SIR: 2.66, 95% CI: 1.15, 5.23) among subjects who worked only in
a shop where tetrachloroethylene was the primary cleaning solvent compared to that expected
based on U.S. population rates. An exposure-response pattern was further suggested because
hypertensive ESRD risk was highest among those employed for >5 years (SIR: 3.39,
95% CI: 1.10, 7.92). No human studies investigating drinking water or other oral exposures on
kidney toxicity have been published.
4-114
-------
Positive associations between kidney cancer (renal cell carcinoma) and exposure to dry-
cleaning and laundry workers or to tetrachloroethylene specifically were observed in several
well-conducted studies (Mandel et al., 1995). The results from the other studies using a relatively
specific exposure-assessment approach to refine classification of potential tetrachloroethylene
exposure in dry-cleaning settings are mixed, with some studies reporting little or no evidence of
an association (Lynge et al., 2006; Pesch et al., 2000a: Boice et al., 1999; Dosemeci etal., 1999;
Aschengrau et al., 1993), and other studies reporting elevated risks (Calvert et al., 2011; Blair et
al., 2003; Anttila et al., 1995; Schlehofer et al., 1995). An increasing trend in relative risk with
increasing exposure surrogate was not observed in any of the larger occupational exposure
studies with three or more exposure categories (Lynge et al., 2006; Mandel et al., 1995), but
some indication of higher risk with higher exposure (or duration) was observed in other studies
(Blair et al., 2003). As expected, the results from studies using a relatively nonspecific exposure
measure (broad occupational title of launderers and dry cleaners, all workers at a factory, density
of dry-cleaning establishments by zip code) are more variable and less precise, reflecting a
greater potential for misclassification bias.
Adverse effects on the kidney have been observed in studies of animals exposed to high
concentrations of tetrachloroethylene by inhalation (JISA, 1993; NTP, 1986), and oral gavage
(Ebrahim etal., 2001: Ebrahim et al., 1996: Jonkeretal., 1996: Green etal., 1990: Goldsworthv
et al., 1988: NCI, 1977), as well as i.p. injection of tetrachloroethylene metabolites (Elfarra and
Krause, 2007). The nephrotoxic effects include increased kidney-to-body weight ratios, hyaline
droplet formation, glomerular "nephrosis," karyomegaly (enlarged nuclei), cast formation, and
other lesions or indicators of renal toxicity. Increased incidences of relatively rare renal tumors
have been observed in one bioassay of male rats exposed to tetrachloroethylene by inhalation
(NTP, 1986). The renal effects occurred following very high (or chronic, relatively high) doses
of tetrachloroethylene exposures. Overall, multiple lines of evidence support the conclusion that
tetrachloroethylene causes nephrotoxicity in the form of tubular toxicity, mediated potentially
through the tetrachloroethylene GSH conjugation products: TCVC and TCVCS.
4.2.4. Hypothesized Mode(s) of Action for Kidney Carcinogenicity
There are multiple hypothesized MO As for kidney carcinogenicity induced with
tetrachloroethylene exposure, including a2u-globulin accumulation, peroxisome proliferation,
genotoxicity, and cytotoxicity unrelated to a2u-globulin. These MO As are addressed in the
sections that follow.
4-115
-------
4.2.4.1. Role of Metabolism in Kidney Carcinogenicity
Except for a2u-globulin accumulation, which is more likely due to tetrachloroethylene
itself (Lash and Parker, 2001), other mechanisms hypothesized to contributed to
tetrachloroethylene-induced renal carcinogenicity are thought to be mediated by
tetrachloroethylene metabolites rather than by the parent compound. Metabolites from the GSH
conjugation pathway are posited to induce renal tumorigenicity, as opposed to (or to a greater
extent than) the metabolites resulting from oxidative CYP processing. The glutathione
conjugation of tetrachloroethylene in the kidney, discussed in Section 3, leads sequentially to
TCVG and TCVC. TCVC can be further processed by p-lyase to yield an unstable thiol,
1,2,2-trichlorovinylthiol, which may give rise to a highly reactive thioketene, a chemical species
that can form covalent adducts with cellular nucleophiles including DNA. TCVC can also
undergo FMO3- or P450-oxidation to reactive intermediates; additionally, sulfoxidation of both
TCVC and its 7V-acetylated product occurs, resulting in reactive metabolites (Ripp et al., 1999;
1997: Werner etal.. 1996).
4.2.4.2. a2u-Globulin Accumulation
Generally, kidney tumors observed in cancer bioassays are assumed to be relevant for
assessment of human carcinogenic potential. However, male rat-specific kidney tumors that are
caused by the accumulation of a2u-globulin are not generally considered relevant to humans.
Accumulation of a2u-globulin in hyaline droplets initiates a sequence of events that leads to
renal nephropathy and, eventually, renal tubular tumor formation. The phenomenon is unique to
the male rat because female rats and other laboratory mammals administered the same chemicals
do not accumulate a2u-globulin in the kidney and do not subsequently develop renal tubule
tumors (Doi et al.. 2007: Swenberg and Lehman-McKeeman, 1999: U.S. EPA. 199la).
4.2.4.2.1. Identification of key events
The histopathological sequence of events in mature male rats is hypothesized to consist
of the following:
• Excessive accumulation of hyaline droplets containing a2u-globulin in renal proximal
tubules
• Subsequent cytotoxicity and single-cell necrosis of the tubule epithelium
• Sustained regenerative tubule cell proliferation
• Development of intralumenal granular casts from sloughed cellular debris associated with
tubule dilatation and papillary mineralization
4-116
-------
• Foci of tubule hyperplasia in the convoluted proximal tubules
• Renal tubule tumors
4.2.4.2.2. Data requirements for establishing the MOA
The EPA (1991a) Risk Assessment Forum Technical Panel report provides specific
guidance for evaluating chemical exposure-related male rat renal tubule tumors for the purpose
of risk assessment, based on an examination of the potential involvement of a2u-globulin
accumulation. In particular, the following information from adequately conducted studies of
male rats is used for demonstrating that the a2u-globulin process may be a factor in any observed
renal effects. An affirmative response in each of the three categories is required. If data are
lacking for any of the criteria in any one category, the available renal toxicity data should be
analyzed in accordance with standard risk assessment principles. The three categories of
information and criteria are as follows:
• Increased number and size of hyaline droplets in the renal proximal tubule cells of
treated male rats. The abnormal accumulation of hyaline droplets in the P2 segment
helps differentiate a2u-globulin inducers from chemicals that produce renal tubule
tumors by other modes of action.
• Accumulating protein in the hyaline droplets is a2u-globulin. Hyaline droplet
accumulation is a nonspecific response to protein overload, and, thus, it is necessary to
demonstrate that the protein in the droplet is, in fact, a2u-globulin.
• Additional aspects of the pathological sequence of lesions associated with a2u-globulin
nephropathy are present. Typical lesions include single-cell necrosis, exfoliation of
epithelial cells into the proximal tubular lumen, formation of granular casts, linear
mineralization of papillary tubules, and tubule hyperplasia. If the response is mild, not
all of these lesions may be observed. However, some elements consistent with the
pathological sequence must be demonstrated to be present.
4.2.4.2.3. Induction of hypothesized key events by tetrachloroethylene
Three studies show that doses of tetrachloroethylene in excess of those observed to
induce tumorigenesis are capable of precipitating hyaline droplet nephropathy in male rats
(Bergamaschi et al.. 1992: Green etal.. 1990: Goldsworthy et al.. 1988): refer to Table 4-11.
Goldsworthy et al. (1988) observed increases in a2u-hyaline droplets in exposed male—but not
female—F344 rats following 10 days of gavage with 1,000 mg/kg tetrachloroethylene. This
finding was correlated with both protein droplet nephropathy (crystalloid accumulation) and
increases in cellular proliferation. The cell replication was enhanced in the male rats specifically
in damaged P2 segments, suggesting a link between the a2u-globulin accumulation and kidney
tumors. Bergamaschi et al. (1992) also demonstrated a2u-accumulation in P2 segments of rat
proximal tubule cells resulting from a daily exposure of rats to 500 mg/kg tetrachloroethylene in
4-117
-------
corn oil for 4 weeks. In short-term, high-dose studies, Green et al. (1990) found that the oral
administration of 1,000 to 1,500 mg/kg of tetrachloroethylene daily for up to 42 days caused an
accumulation of a2u-globulin in the proximal tubules of male rats. These effects were not
observed in female rats or in mice.
Table 4-11. Renal a2u-globulin formation in tetrachloroethylene-exposed
rodents
Species/strain/
sex/number
Mouse, B6C3Fl5 both
sexes (groups of 49 or
50 mice of each sex per
dose group, total of
-300 mice)
Rat, F344, both sexes
(groups of 50 mice of
each sex per dose group,
total of -300 mice)
Rat, F344 (both sexes,
5 per group)
Rat, F344 (both sexes,
12 per group)
Rat, F344 (both sexes)
and mouse, B6C3FJ (both
sexes) (10 per group for
oral studies, 5 per group
for inhalation studies)
Exposure
level/duration
0, 100, 200 ppm for
104 wk, inhalation
0, 200, 400 ppm for
104 wk, inhalation
0 or 1,000 mg/kg-day for
10 d, corn oil gavage
0, 500 mg/kg-day daily
for 4 wk, corn oil gavage
0, 1,000 or 1,500
mg/kg-day daily by corn
oil gavage for 42 d; 0 or
1,000 ppm for 10 d
Effects
Karyomegaly and cytomegaly of the
proximal tubules in all exposed mice;
nephrosis was observed in exposed
females, casts increased in all exposed
males and in high-dose females
Karyomegaly and cytomegaly of the
proximal tubules in all exposed rats
Increases in o2u-hyaline droplets in
exposed male but not female rats.
Correlated to increased cell proliferation
and protein droplet nephropathy
Increases in o2u-hyaline accumulation in
proximal tubule cells
Accumulation of o2u-globulin in
proximal tubules of male rats;
nephrotoxicity also observed in male rats
(formation of granular tubular casts and
evidence of tubular cell regeneration)
Inhalation exposure demonstrated
formation of hyaline droplets in kidneys
of male rats
Reference
NTP (1986)
NTP (1986)
Goldsworthy
etal. (1988)
Bergamaschi
etal. (1992)
Green et al.
(1990)
Green et al. (1990) tested lower inhaled tetrachloroethylene doses in rats—up to 400 ppm
for 6 hours/day, for 28 days, with the animals being sacrificed within 18 hours of termination of
the final exposure—but found no evidence of hyaline droplet formation; however, there may
have been time for recovery prior to sacrifice. Green et al. (1990) proposed the possibility that
longer-term exposure to the 400 ppm concentration of tetrachloroethylene is required for the
hyaline droplet accumulation in the kidney of rats. a2u-Globulin accumulation can be
demonstrated, however, after only short-term exposures (even a single administration) to several
agents, such as d-limonene, decalin, unleaded gasoline, and trimethylpentane (NTP, 1990b:
Charbonneau et al., 1987).
4-118
-------
Lack of hyaline droplet formation, increase in a2u-globulin, or signs of the characteristic
renal nephropathy at the high dose level of the NTP inhalation study (NTP, 1986) may, thus,
diminish the likelihood that the renal tumors associated with exposure to tetrachloroethylene are
induced through this mechanism (Green et al., 1990). NTP did not report the presence of hyaline
droplets in rats that had been exposed to either 200 or 400 ppm tetrachloroethylene for up to
2 years. These doses were associated with the production of renal tubule neoplasms in male rats.
However, the fact that NTP did not report the presence of hyaline droplets in the 14-day, 90-day,
or 2-year studies is not definitive, because the NTP protocol at that time was not designed
specifically to detect hyaline droplets or a2u-globulin accumulation in the kidney (NTP, 1990b).
Thus, the procedures followed at the time of the study were not necessarily conducive to
detecting hyaline droplets. For example, in the chronic study of tetrachloroethylene, at least 1
week elapsed between the final tetrachloroethylene exposure and the scheduled sacrifice of the
surviving animals. It is possible that had hyaline droplets been present, they could have
regressed. Also, the nephropathy observed at the end of a 2-year bioassay could be difficult to
distinguish from the old-age nephropathy that occurs in these rats.
In contrast, the renal pathology reported in the NTP bioassay is not entirely consistent
with the results generally found for chemicals where there is a2u-globulin accumulation (NTP,
1986) (letter from Scot Eustis, National Toxicology Program, to William Farland, Director,
Office of Health and Environmental Assessment, U.S. EPA, 1988). For example, there was no
mineralization in the inner medulla and papilla of the kidney, a frequent finding in bioassay s of
chemicals that induce a2u-globulin accumulation (e.g., for pentachloroethane, the incidence of
renal papillar mineralization was 8% in controls, 59% in the low-dose group, and 58% in the
high-dose group). In addition, it is important to note that some aspects of toxic tubular
nephropathy were also observed in female rats and male mice exposed to tetrachloroethylene,
clearly contrary to sex and species specificity.
In the NCI gavage study of tetrachloroethylene (NCI, 1977), toxic nephropathy, which
was not detected in the control animals, occurred in both male and female Osborne-Mendel rats
administered tetrachloroethylene. Tetrachloroethylene also clearly caused nephropathy in both
sexes of mice in the study. Unfortunately, animal survival in the rat study was not adequate to
support any conclusions about tetrachloroethylene carcinogenicity.
In summary, although a few studies show an increase in hyaline droplets in the proximal
tubule cells of treated male rats, other studies demonstrate nephrotoxicity in both male and
female rats and mice without hyaline droplet formation. Further, the studies that demonstrate
hyaline droplet formation do not also have additional aspects of nephrotoxicity associated with
a2u-globulin formation. The a2u-globulin response reported following exposure to
tetrachloroethylene is relatively modest, and the fact that renal tumors have been observed at
4-119
-------
doses lower than those shown to cause the a2u-globulin response is inconsistent with this
phenomenon being responsible for tumorigenesis. Chronically induced tetrachloroethylene
nonneoplastic kidney lesions exhibit neither species nor sex specificity. Unlike with other
chemicals that induce a2u-globulin accumulation and have been tested by NTP in chronic
carcinogenicity bioassays, renal lesions occurring in animals exposed to tetrachloroethylene were
not limited to the male rat. Although the female rat did not develop any renal tubule tumors, the
incidence of karyomegaly was significantly elevated in the female rat as well as in the male rat;
1 of 50 female rats exposed at the high dose developed tubule cell hyperplasia. Therefore, based
on the criteria described above, there are insufficient data to demonstrate renal toxicity or
cancers are caused by a2u-globulin formation.
4.2.4.3. Genotoxicity
A hypothesized mutagenic MOA entails the following key events leading to
tetrachloroethylene-induced kidney tumor formation: following metabolism of
tetrachloroethylene to one or more mutagenic intermediates, the genetic material is altered in a
manner that permits changes to be transmitted during cell division through one or more
mechanisms (gene mutations, deletions, translocations, or amplification); the resulting mutations
advance acquisition of the multiple critical traits contributing to carcinogenesis. This MOA may
apply to multiple cancer types.
The genotoxic potential of tetrachloroethylene is addressed in Section 4.8. To
summarize, the results of a large number of in vitro genotoxicity tests in which
tetrachloroethylene was the test agent support the conclusion that tetrachloroethylene does not
exhibit direct mutagenic activity in the absence or presence of the standard S9 fraction
(Watanabe et al.. 1998: DeMarini et al.. 1994: Roldan-Aijonaetal.. 1991: Milman et al.. 1988:
Warner et al.. 1988: NTP. 1986: Connor et al.. 1985: Shimada et al.. 1985: Haworth et al.. 1983:
Hardin et al.. 1981: Kringstad et al.. 1981: Bartsch et al.. 1979: Greim et al.. 1975). However,
the few in vitro mutagenicity studies of tetrachloroethylene under conditions that would generate
the GSH conjugate were positive (Vamvakas et al., 1989c; Vamvakas et al., 1989d). While most
of these intermediates have not been characterized for mutagenic potential, TCVG (Dreessen et
al.. 2003: Vamvakas et al.. 1989d) and JV-acetyl-5-(l,2,2-trichlorovinyl)-Z-cysteine (NAcTCVC)
(Vamvakas et al., 1987) are mutagenic in the presence of activation while TCVC was mutagenic
even in the absence of activation (Dreessen et al., 2003; Dekantetal., 1986a). The metabolite
DCA is the most potent mutagen of the P450-derived metabolites, exhibiting mutagenic activity
in a number of assays. A putative P450-derived metabolite, 1,1,2,2-tetrachloroethylene oxide, is
also mutagenic; the mutagenicity of this epoxide would be predicted from structure-activity
relationships. Studies of chromosomal aberrations following exposure to tetrachloroethylene are
4-120
-------
mostly negative, but positive results have been reported from in vitro studies with enhanced
metabolic activation (Doherty et al., 1996).
The limited in vivo studies of tetrachloroethylene are inconsistent, with only negative
(NTP, 1986: Bronzetti et al.. 1983) or equivocal (Cederberget al.. 2010a: Bellies et al.. 1980)
genotoxicity assay results demonstrated following inhalation or oral exposure. These include the
finding that tetrachloroethylene at higher concentrations induces at most modest increases in
DNA damage in liver tissue (Cederberg et al., 2010a). Following in vivo exposures,
tetrachloroethylene induces SSB and DNA binding in kidney (Potter et al., 1996; Mazzullo et al.,
1987; Walles, 1986). Intraperitoneal injection assays have demonstrated both negative (NTP,
1986) as well as positive results for different genotoxicity endpoints in other tissues (Murakami
and Horikawa, 1995). Assays of clastogenic effects following inhalation exposure in humans
have shown inconsistent results and are suggested to be related to coexposures (Seiji etal., 1990;
Ikedaetal., 1980).
Thus, although tetrachloroethylene has largely yielded negative results in standard
genotoxicity assays, uncertainties remain with respect to the possibility that genotoxicity
contributes to renal carcinogenesis. Not all metabolites have been identified or characterized,
but several known metabolites including those derived from P450 as well as GSH pathways are
mutagenic in the standard battery of tests. Tetrachloroethylene is mutagenic in bacterial assays
in the presence of GST and GSH, whereas the standard S9 fraction has typically yielded negative
results. Tetrachloroethylene at higher concentrations also induces modest increases in DNA
damage and DNA binding in liver tissue (Cederberg et al., 2010a: Murakami and Horikawa,
1995). Given the demonstrated mutagenicity of several tetrachloroethylene metabolites, the
hypothesis that mutagenicity contributes to the MOA for tetrachloroethylene carcinogenesis
cannot be ruled out, although the specific metabolic species or mechanistic effects are not
known.
4.2.4.4. Peroxisome Proliferation
The PPARa-agonism MOA is also hypothesized to induce rat kidney tumorigenesis.
According to this hypothesis, the key events leading to tetrachloroethylene-induced kidney tumor
formation constitute the following, after activation of tetrachloroethylene to one or more reactive
metabolites: the PPARa receptor is activated, which then causes alterations in cell proliferation
and apoptosis, followed by clonal expansion of initiated cells.
Limited data exist to support increased peroxisome proliferation in rodent kidney
following exposure to tetrachloroethylene and are summarized in Table 4-12 (Odum etal., 1988:
Goldsworthy andPopp, 1987). The role of peroxisome proliferation in tetrachloroethylene-
induced kidney toxicity and cancer was examined in male and female F344 rats and B6C3Fi
4-121
-------
mice exposed to tetrachloroethylene by inhalation (400 ppm, 6 hours/day, 14, 21, or 28 days or
200 ppm, 6 hours/day, 28 days) in Odum et al. (1988). Five animals per group were exposed.
Insufficient mouse kidney tissue limited the analysis to pooled samples. Slight increases were
observed in p-oxidation in mouse kidney (maximum 1.6-fold increase at 21 days, 400 ppm
exposure). Modest palmitoyl-CoA oxidation (PCO) increases were observed in the kidney of
male rats at 200 ppm at 28 days (1.3-fold) but not 400 ppm at 14, 21, or 28 days. In female rat
kidney, PCO was elevated (approximately 1.6-fold) at all doses and times. However,
peroxisome proliferation was not observed in rat or mouse kidney upon microscopy, suggesting
that this does not play a role in kidney carcinogenesis.
Table 4-12. Renal peroxisome proliferation in tetrachloroethylene-exposed
rodents
Species/strain/sex/number
Rat, F344; and mouse,
B6C3FJ (both sexes,
5/group)
Odum et al. (1988)
F344 rats (male only,
5/group) and B6C3FJ mice
(male only, 5/group)
Goldsworthy and Popp
(1987)
Effect
Mice of both sexes: Analysis in mice was limited
to pooled tissue, but showed slight increases in
(3-oxidation in mouse kidney
Rats of both sexes: Modest increases in PCO
observed in male rat kidneys at 200 ppm for 28 d
only, but elevated in female rat kidneys at all
doses and times
Mice: Increased PCO activity in all exposed mice
Rats: Increased kidney weight in exposed rats
Dose
200, and
400 ppm,
inhalation
200, and
400 ppm,
inhalation
1,000 mg/kg-day
for 10 d, corn oil
gavage
1,000 mg/kg-day
for 10 d, corn oil
gavage
Time
14, 21, 28 d
14, 21, 28 d
10 d
10 d
Goldsworthy and Popp (1987) administered tetrachloroethylene (1,000 mg/kg-day) by
corn oil gavage to 5 male F344 rats and 5 male B6C3Fi mice for 10 days. In
tetrachloroethylene-exposed rats, PCO was modestly—although not significantly—elevated in
the liver (1.4-fold increase) and kidney (1.7-fold increase). In mice, tetrachloroethylene
exposure increased PCO activity 4.3-fold in liver and by 2.3-fold in kidney. Relative liver
weight was increased in rats and mice with tetrachloroethylene exposure, but relative kidney
weight was unaffected. A comparison of corn oil with methyl cellulose revealed no effect of the
gavage vehicle on tetrachloroethylene-induced PCO. A less-than-additive effect of
trichloroethylene (1,000 mg/kg) administered together with tetrachloroethylene on PCO
induction was seen.
4-122
-------
4.2.4.5. Cytotoxicity/Sustained Chronic Nephrotoxicity Not Associated with a2u-Globulin
Nephropathy
The hypothesis is that renal neoplasms induced by tetrachloroethylene arise secondary to
renal cytotoxicity and subsequent cellular proliferation without regard to a2u-accumulation.
This MOA entails the following key events leading to tetrachloroethylene-induced kidney tumor
formation: following metabolism of tetrachloroethylene to one or more reactive intermediates,
toxicity to the kidney ensues and is sustained; via a variety of potential mechanisms (damage to
and alteration of macromolecules, cell signaling alterations, etc.), the acquisition of the multiple
critical traits contributing to carcinogenesis is advanced.
The kidney is a major target organ for tetrachloroethylene-induced toxicity through the
reactive metabolites produced subsequent to GSH conjugation. Renal tubule neoplasia is
observed to occur only in male rats. This species- and sex-specific response would not be
expected based on the hypothesized MOA because tetrachloroethylene has been reported to
produce nephrotoxicity across species, and in both sexes. Signs of tetrachloroethylene-induced
kidney damage appeared in both rats and mice during the early phases of the NTP inhalation
study, for example, indicating that animals of both species surviving to the scheduled termination
of the study had long-standing nephrotoxicity. Although the female rats did not develop any
renal tubule tumors, the incidence of karyomegaly was significantly elevated in females as well
as in males, and 1/50 female rats exposed at the high dose developed tubule cell hyperplasia
(NTP. 1986).
In the NTP study of the mouse, "nephrosis" was observed at increased incidences in
dosed females, casts were observed at increased incidences in dosed males and high-dose
females, and karyomegaly of the tubule cells was observed at increased incidences in both sexes
of treated mice (NTP, 1986). The severity of the renal lesions was dose related, and one low-
dose male had a renal tubule cell adenocarcinoma. In the NCI gavage study of B6C3Fi mice and
Osborne-Mendel rats exposed to tetrachloroethylene, toxic nephropathy was not detected in
control animals but did occur in both male and female rats as well as in mice (NCI, 1977).
Mechanistic studies of tetrachloroethylene nephrotoxicity are relatively sparse. Most
studies performed to elucidate information related to understanding tetrachloroethylene renal
toxicity have concentrated on the GSH pathway metabolites rather than on the parent chemical;
this is because much of the available data for both tetrachloroethylene and trichloroethylene
suggest that it is flux through this pathway that generates reactive chemical species responsible
for nephrotoxicity. Vamvakas et al. (1989a; 1989b) have shown the tetrachloroethylene
conjugate metabolites TCVG and TCVC to cause dose-related cytotoxicity in renal cell
preparations and prevention of this toxicity by p-lyase enzyme inhibitor. Renal p-lyases are
known to cleave TCVC to yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to
4-123
-------
a highly reactive thioketene, a chemical species that can form covalent adducts with cellular
nucleophiles. Additionally, sulfoxidation of both TCVC and its 7V-acetylated product occurs,
resulting in toxic metabolites (Ripp et al., 1999; 1997; Werner et al., 1996). Findings using in
vitro models studied by Lash et al. (2002) suggest a marked sex difference between male and
female rats in the severity of acute renal toxicity caused by both tetrachloroethylene and its
TCVG metabolite. Tetrachloroethylene and TCVG also produced signs of toxicity in
mitochondria (i.e., mitochondrial dysfunction, such as inhibition of State 3 respiration by specific
inhibition of several sulfhydryl-containing enzymes in both sexes of mice; (Lash et al., 2002;
Lash and Parker. 2001: 2000).
4.2.4.6. Summary
The kidney is a target organ in mammalian species for tetrachloroethylene and other
related chlorinated ethanes and ethylenes, and tetrachloroethylene causes kidney cancer in male
rats. It is likely that several mechanisms contribute to tetrachloroethylene-induced kidney
cancer. Mutagenicity, peroxisome proliferation, a2u-globulin nephropathy, and cytotoxicity not
associated with a2u-globulin accumulation are MO As that have been investigated. Except for
a2u-globulin accumulation, which is more likely due to tetrachloroethylene itself (Lash and
Parker, 2001), other mechanisms hypothesized to contributed to tetrachloroethylene-induced
renal carcinogenicity are thought to be mediated by tetrachloroethylene metabolites rather than
with the parent compound.
Metabolites from the GSH conjugation pathway are posited to induce renal
tumorigenicity, as opposed to or to a greater extent than the metabolites resulting from oxidative
CYP processing. The glutathione conjugation of tetrachloroethylene in the kidney, discussed in
Section 3, leads sequentially to TCVG and TCVC. TCVC can be further processed by p-lyase to
yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a highly reactive
thioketene, a chemical species that can form covalent adducts with cellular nucleophiles
including DNA. TCVC can also undergo FMO3 or P450 oxidation to reactive intermediates;
additionally, sulfoxidation of both TCVC and its 7V-acetylated product occurs, resulting in
reactive metabolites (Ripp et al., 1999; Ripp et al., 1997; Werner etal., 1996). While most of
these intermediates have not been characterized for mutagenic potential, TCVG, TCVC, and
NAcTCVC are clearly mutagenic in Salmonella tests. In addition, tetrachloroethylene exhibited
mutagenicity in Salmonella in the few studies of conditions that could generate GSH-derived
metabolites. Tetrachloroethylene, following in vivo exposures, also binds to kidney DNA and
induces SSB in kidney. Given the known mutagenicity of the GSH-derived tetrachloroethylene
metabolites that are formed in the kidney, and the observed in vitro mutagenicity of
4-124
-------
tetrachloroethylene under conditions that would generate these metabolites, a mutagenic MOA
contributing to the development of the kidney tumors cannot be ruled out.
Due to tetrachloroethylene's nephrotoxic effects, it has been suggested that the low-level
renal tumor production observed in exposed rats is secondary to sustained cytotoxicity and
necrosis leading to activation of repair processes and cellular regeneration. However,
"nephrotoxicity" occurs in both sexes of rats and mice, whereas cell replication and
tumorigenesis occurs in male—but not in female—rats. In addition, tetrachloroethylene induces
kidney tumors at lower doses than those required to cause a2u-globulin accumulation, raising
serious doubt that a2u-globulin plays a key role—especially any major role—in the rat kidney
tumor formation.
Because tetrachloroethylene has been shown to induce peroxisome proliferation, an
indicator of PPAR-activation, the possibility exists that certain responses resulting from
activation of PPAR receptors might be involved in cancer-causing activity leading to
tetrachloroethylene-induced renal tumors. However, the chemical-specific data are limited and
show only modest effects at exposures exceeding those required for renal carcinogenesis. There
is no evidence causally linking PPARa-activation to kidney tumorigenesis for
tetrachloroethylene or other compounds.
In summary, the complete mechanisms of tetrachloroethylene-induced renal
carcinogenesis are not yet understood. Given the known mutagenicity of the GSH-derived
tetrachloroethylene metabolites that are formed in the kidney, and the observed in vitro
mutagenicity of tetrachloroethylene under conditions that would generate these metabolites, a
mutagenic MOA contributing to the development of the kidney tumors cannot be ruled out.
4.3. LIVER TOXICITY AND CANCER
4.3.1. Human Studies
A number of hepatotoxic effects, including hepatomegaly, hepatocellular damage, and
elevations of several hepatic enzymes and bilirubin degradation byproducts, have been observed
after acute high-level exposure to tetrachloroethylene [levels not identified; Meckler and Phelps
(1966): Coler and Rossmiller (1953): Hake and Stewart (1977): Saland (1967): Stewart et al.
(1961), as reported in ATSDR (1997b)1. One case report noted obstructive jaundice and
hepatomegaly in an infant exposed orally to tetrachloroethylene [1 mg/dL; Bagnell and
Ellenberger (1977). as reported in ATSDR (1997b)1.
4-125
-------
4.3.1.1. Liver Damage
Four cross-sectional studies were available that evaluated the prevalence of liver damage
among dry-cleaner populations (Brodkin et al., 1995; Gennari et al., 1992; Cai et al., 1991;
Lauwerys et al., 1983). These studies assessed serum concentration of a number of hepatic
enzymes in dry-cleaner and control populations. Additionally, sonographic changes to hepatic
parenchymal tissue were examined in one study (Brodkin et al., 1995). An elevated
concentration of the serum enzyme GOT and mild hepatic changes were notable observations in
two studies (Brodkin et al.. 1995: Gennari et al.. 1992).
Gennari et al. (1992) measured the electrophoretic fractionation patterns of serum GGT
isozymes among 141 tetrachloroethylene-exposed dry cleaners and 130 nonexposed controls
selected from staff and students from the academic institution of the principal investigators.
Both the exposed subjects and the controls had similar lifestyles (smoking, alcohol consumption)
and clinical medical histories. The TWA tetrachloroethylene concentration in the dry-cleaning
facilities was 11.3 ppm. Total GGT was higher in exposed workers (exposed: mean of 12.4
international units per liter [U/L; standard deviation, 6.9 U/L]; controls: 8.8 U/L [4.9 U/L],
p < 0.01). The GGT-2 isoenzyme component was higher in exposed workers (6.8 U/L [5.7 U/L]
in exposed vs. 3.5 U/L [3.3 U/L] in controls,/? < 0.01), and the GGT-4 component was
detectable in exposed workers but not measurable in controls. The authors regarded a
GGT-2/GGT-3 ratio of greater than 1 as a sensitive index of the reciprocal behavior of the two
isoenzymes. GGT-2 is generally associated with activation of liver microsomal enzymes.
GGT-4 is common in liver diseases and indicates hepato-biliary impairment.
This study excluded individuals who presented values for GGT, or other liver enzymes
above a normal range, and individuals who had past or current liver disease. None of the
workers showed any clinical symptoms of liver disease, and their enzymatic profiles, including
GGT, aspartase amino transaminase (AST), alanine amino transaminase, 5'-nucleotidase, and
alkaline phosphatase, were within the clinically normal reference limits. Given the study's
exclusion criteria, it is not surprising that liver enzyme concentrations were within a normal
range. The authors stated that more research is required to develop this GGT fractionation assay
into a clinically useful method of measuring liver function. Nevertheless, the study showed that
these dry cleaners had markers of tetrachloroethylene oxidative metabolism (GGT-2) and liver
impairment (GGT-4).
The study by Brodkin et al. (1995) examined liver function and carried out sonography
measurements in a population of 27 dry cleaners and 26 nonexposed laundry workers. Dry
cleaners were older and had a longer duration of employment than did laundry workers. The
mean TWA exposure (8 hours) among all dry cleaners was 15.8 ppm (range: 0.4-83 ppm). The
investigators found a higher prevalence of abnormal hepatic sonograms among the dry cleaners
4-126
-------
(67%) than among laundry workers (38%;p < 0.05), the control group. The noninvasive imaged
penetration of ultrasound into liver tissue can reveal the presence of fat accumulation and fibrous
structures. Hepatic parenchymal changes were graded as mild, moderate, or severe. The
prevalence of hepatic parenchymal changes increased both with increasing current concentration
and with cumulative exposure (p < 0.05). Subjects with serological evidence of active hepatitis
infection were excluded from these analyses.
Brodkin et al. (1995) fit logistic regression models to examine possible associations
between mild or greater parenchymal changes and tetrachloroethylene exposure. These analyses
included adjustment for the effects of ethanol consumption within the past 6 months, sex, body
mass index, age, and serological evidence of active and past hepatitis infection. Subjects with
serological evidence of active hepatitis infection were included in the logistic regression analysis
due to the ability of the statistical method to account for the effects associated with this factor.
These analyses showed subjects exposed during older wet or dry-to-dry transfer processes
(average concentration: 19.8 ppm; range: 1.8-83 ppm) was strongly—but imprecisely—
associated with mild or greater sonographic changes (OR: 4.2, 95% CI: 0.9-20.4) as compared
with controls. No association was shown with subacute exposure in new dry-to-dry operations
(OR: 0.7, 95% CI: 0.1-5.9). An inverse dose-response association was found with cumulative
exposure after adjustment for age due to a strong but imprecise association between
tetrachloroethylene exposure and hepatic sonographic changes in younger workers (workers less
than 35 years of age, OR: 15; 95% CI: 1.33-170).
Only 21% of the exposed study subjects who had changes graded as mild or greater had
increases in any hepatic enzyme (Brodkin et al., 1995). Mean concentrations of GGT, AST, and
alanine transferase (ALT) tended to be higher among the dry cleaners as compared with laundry
workers; but, the differences were not statistically significant, and all mean values were within
the normal range of reference values. However, all of the subjects who had elevated ALT
concentrations had moderate or severe sonographic changes. Hence, sonographic imaging of the
liver appears to be a more sensitive indicator of toxicity than measurement of serum hepatic
enzymes.
Lauwerys et al. (1983) performed behavioral, renal, hepatic, and pulmonary tests on 22
subjects exposed to tetrachloroethylene in six dry-cleaning shops and compared the results with
those obtained for 33 subjects nonoccupationally exposed to organic solvents. The mean TWA
concentration was 21 ppm. The investigators found no statistically significant differences in
mean serum hepatic enzyme concentration between exposed subjects and controls, but this study
is poorly reported, and the authors did not describe the statistical methods used to test for
differences between the exposed and control groups.
4-127
-------
Cai et al. (1991) investigated subjective symptoms, hematology, serum biochemistry, and
other clinical signs in 56 dry cleaners exposed to tetrachloroethylene at 20 ppm (as a geometric
mean of 8-hour TWA) and compared the results with findings for 69 nonexposed controls from
the same factories. Exposure-related increases were observed in the prevalence of subjective
symptoms during the workday as well as in the past 3-month period, whereas no significant
changes in hematology were seen. There was no effect on liver and kidney function, as
measured by enzyme activities, blood urea nitrogen (BUN), and creatinine in the serum.
Table 4-13 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 4-13. Summary of studies of human liver toxicity
Subjects
27 PCE-exposed dry cleaners
26 nonexposed laundry
workers
141 PCE-exposed dry cleaners
130 controls
24 PCE-exposed dry cleaners
33 controls nonoccupationally
exposed to organic solvents
56 PCE-exposed dry cleaners
69 nonexposed factory controls
Effects
Sonographic scattering of fat in
liver
Severity greater with higher
cumulative exposure
No liver toxicity
Elevation of total GOT 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 yr
Mean TWA = 11. 3 ppm
Mean duration of exposure = 20 yr
Mean TWA = 21 ppm
Mean duration of exposure = 6 yr
Geometric mean TWA = 20 ppm
Mean duration of exposure = 3 yr
Author
Brodkin et
al. (1995)
Gennari et
al. (19921
Lauwerys et
al. (1983)
Cai et al.
(1991)
4-128
-------
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
effects have occurred (NRC, 1995). The observations of Brodkin et al. (1995) and Gennari et al.
(1992) support the indication that tetrachloroethylene exposure affects liver function; hence, the
lowest-observed-adverse-effect level (LOAEL) for liver effects in humans can be established as
a range from 12 to 16 ppm (TWA).
4.3.1.2. Liver Cancer
Eighteen epidemiologic studies reporting data on liver cancer and tetrachloroethylene
exposure were identified. This set of studies includes 14 cohort or nested case-control studies on
liver cancer (Calvert et al.. 2011: Selden and Ahlborg, 2011: Lindbohm et al.. 2009: Pukkala et
al.. 2009: Sung et al.. 2007: Lynge et al.. 2006: Ji and Hemminki, 2005c: Blair etal.. 2003:
Travier et al.. 2002: Andersen et al.. 1999: Boiceetal.. 1999: Lynge etal.. 1995: Bond et al..
1990: Lynge and Thygesen, 1990), two liver cancer case-control studies of occupational
exposures ( Suarez et al., 1989: Stemhagen et al., 1983), and two liver cancer case-control
studies of residential exposure (Lee etal., 2003: Vartiainen et al., 1993). Two other cohort
studies included information on tetrachloroethylene but did not report risk estimates for liver
cancer (Radican et al., 2008: Anttila et al., 1995), as well as an earlier report of mortality by
Chang et al. (2003) for subjects in Sung et al. (2007), did not provide an estimate of the
association for liver cancer. Additionally, three liver cancer case-control studies that examined
occupational exposure did not report an odds ratio for holding an occupation or for work in a dry
cleaner and laundry (Ferrand et al., 2008: Austin et al., 1987: Houten and Sonnesso, 1980) and
so were not evaluated further. The 18 studies represent the core studies evaluated by EPA, as
described in more detail below. Appendix B reviews the design, exposure-assessment approach,
and statistical methodology for each study. Most studies were of the inhalation route of
exposure, of occupational exposure, and lacked quantitative exposure information.
Thirteen studies reporting risk estimates for liver cancer examine occupational title as dry
cleaner, launderer, and presser as surrogate for tetrachloroethylene, given its widespread use
from 1960 onward in the United States and Europe (Calvert et al., 2011: Selden and Ahlborg,
2011: Lindbohm et al., 2009: Pukkala et al., 2009: Lynge et al., 2006: Ji and Hemminki, 2005c:
Blair etal.,2003: Travier et al., 2002: Andersen et al., 1999: Lyngeetal., 1995: Lynge and
Thygesen, 1990: Suarez etal., 1989: Stemhagen et al., 1983). Seven studies conducted in Nordic
countries are either based on the entire Swedish population or on combined populations of
several Nordic countries; strengths of these studies are their use of job title as recorded in census
databases and ascertainment of cancer incidence using national cancer registries (Selden and
4-129
-------
Ahlborg. 2011: Lindbohm et al.. 2009: Pukkala et al.. 2009: Lynge et al.. 2006: Andersen et al.,
1999: Lynge et al., 1995: Lynge and Thygesen, 1990). Lynge et al. (1995) is a nested case-
control study of subjects in Lynge and Thygsen (1990). Subjects in the multi-Nordic country of
Pukkala et al. (2009) overlapped those of Lynge and Thygesen (1990), Lynge et al. (1995),
Andersen et al. (1999), Lynge et al. (2006), and Selden and Ahlborg (2011). Studies examining
mortality among U.S. dry-cleaner and laundry workers (Blair etal., 2003: Ruder et al., 2001) are
of smaller cohorts than the Nordic studies, with fewer observed liver cancer events.
The exposure surrogate in studies of dry-cleaners and laundry workers is a broad
category containing jobs of differing potential for tetrachloroethylene exposure. Thus, these
studies have a greater potential for exposure misclassification bias compared to studies with
exposure potential to tetrachloroethylene assigned by exposure matrix approaches applied to
individual subjects. One dry-cleaning study included an analysis of subjects who worked for one
or more years before 1960 in one or more shops known to use tetrachloroethylene as the primary
solvent (Calvert et al., 2011). The cohort was stratified into two groups based on the level of
certainty that the worker was employed only in facilities using tetrachloroethylene as the primary
solvent exposure; tetrachloroethylene-only and tetrachloroethylene-plus. Lynge et al. (1995)
separately classifies subjects in Lynge and Thygsen (1990) as either dry cleaners or laundry
workers using occupation and workplace description from 1970 Census records. Lynge et al.
(2006), using job title reported in the 1970 Census, identified subjects as dry cleaner (defined as
dry cleaners and supporting staff if employed in business of <10 workers), other job titles in dry
cleaning (launderers and pressers), unexposed (job title reported on 1970 Census was other than
in dry cleaning), or unclassifiable (information was lacking to identify job title of subject).
Selden and Alhborg (2011) identified subjects as either dry cleaners, assigned with potential for
tetrachloroethylene exposure, or laundry workers, assigned as unexposed, and presented risk
estimates separately by job title. Lindbohm et al. (2009), using a JEM approach based on job
title and exposures, assigned a cumulative exposure index to chlorinated hydrocarbons to
individual subjects. Tetrachloroethylene is one of several chlorinated solvents included in the
broad category, but Lindbohm et al. (2009) do not present risk estimates for tetrachloroethylene-
only subjects.
Three other cohorts with potential tetrachloroethylene exposure in industrial settings have
been examined. These studies include aerospace or aircraft maintenance workers in the United
States (Boice et al., 1999), workers, electronic factory workers in Taiwan (Sung et al., 2007), and
workers at a Dow plant in Michigan (Bond etal., 1990). Boice et al. (1999) used an exposure
assessment based on a JEM, and Bond et al. (1990), a nested case-control study, used company
work history records to assign potential tetrachloroethylene exposure to individual subjects. In
contrast and less sensitive, the exposures in the Taiwan studies included multiple solvents and
4-130
-------
tetrachlorethylene exposure was not linked to individual workers, and cohorts included white-
collar workers, who had an expected lower potential for exposure (Sung et al., 2007).
Two geographical studies focused on residential proximity to drinking water sources
contaminated with tetrachloroethylene and other solvents. Vartiainen et al. (1993) examines
liver cancer incidence in two southern Finnish municipalities, with the exposure surrogate
assigned uniformly to all subjects. Lee et al. (2003), using a mortality odds ratio approach,
examined residence in two communities surrounding the factory whose workers were studied by
Chang et al. (2005; 2003) and Sung et al. (2007). One village upstream from the factory was
considered as unexposed, and another village downstream from the factory was identified as
exposed based on groundwater monitoring of drinking water wells during the period 1999-2000.
In summary, with respect to exposure-assessment methodologies, five studies with liver
cancer data assigned tetrachloroethylene exposure to individuals within the study using a JEM
(Boice et al., 1999; Bond et al., 1990), identified a subcohort whose subjects were employed only
in facilities using tetrachloroethylene as the primary solvent exposure (Calvert et al., 2011), or
restricting analyses to subjects identified as dry cleaners (Selden and Ahlborg, 2011; Lynge et
al., 1995). One other study sought additional data using a questionnaire for use in refining
potential exposure within dry-cleaning settings (Lynge et al., 2006). The relative specificity of
these exposure-assessment approaches strengthens their ability to identify cancer hazards
compared to studies with broader and less sensitive exposure-assessment approaches. The least
sensitive exposure assessments are those using very broad definitions such as working in a plant
or factory (Sung et al., 2007: Chang et al., 2003).
Four22 of the 16 liver cancer studies evaluated by EPA with exposure assessment to
tetrachloroethylene or employment as dry-cleaner or laundry worker reported estimated relative
risks based on 50 or more observed events (Pukkala et al., 2009; Lynge et al., 2006; Ji and
Hemminki, 2005c: Travier et al., 2002). The observed number of liver cancer cases in these
studies ranged from 58 (Lynge et al., 2006) to 113 (Pukkala et al., 2009). The four large studies
observed a standardized incidence ratio of 0.76 (95% CI: 0.38, 1.52), 1.02 (95% CI: 0.84, 1.24),
1.22 (95% CI: 1.03, 1.45), and 1.23 (95% CI: 1.02, 1.49) in Lynge et al. (2006), Travier et al.
(2002), Ji and Hemminki (2005c), and Pukkala et al. (2009), respectively, for the association
between liver cancer risk and ever having a job title of dry-cleaner or laundry worker (refer to
Table 4-14).
In addition to the evidence from the large cohort and case-control studies, 11 other
studies reported effect estimates for liver cancer based on fewer observed events and carry lesser
22 Lynge and Thygsen (1990) and Andersen et al. (1999) are not included in this summary of the data from the
individual studies because they were updated and expanded in the analysis by Lynge et al. (1995) and Pukkala et al.
(2009). respectively.
4-131
-------
weight in the analysis. As expected, the magnitude of the point estimate of the association23
reported in these studies is more variable than in the larger studies: 0.13 to 0.98 (Calvert et al.,
2011: Sung et al.. 2007: Blair et al.. 2001: Vartiainen et al.. 1993: Suarezetal.. 1989), 1.2 to 1.8
(Selden and Ahlborg, 2011: Lindbohm et al.. 2009: Bondetal.. 1990) and 2.05 to 2.57 (Lee et
al.. 2003: Boice et al.. 1999: Stemhagen et al.. 1983). Only the 95% CIs of the risk estimate of
Lee et al. (2003) excluded 1.0.
Establishment of an exposure or concentration-response relationship can add to the
weight of evidence for identifying a cancer hazard, but only limited data pertaining to exposure-
response relationships for liver cancer and tetrachloroethylene exposure are available. Four
studies of liver cancer presented risk estimates for increasing exposure categories using exposure
duration, a proxy inferior to cumulative exposure due to inability to account for temporal
changes in exposure intensity (Selden and Ahlborg, 2011: Lynge et al., 2006: Travier et al.,
2002: Boice et al., 1999). Boice et al. (1999) presents a statistical test for linear trend for
subjects with intermittent-routine tetrachloroethylene exposure, a broader category than that used
to examine overall tetrachloroethylene exposure (comprised of routine-exposed subjects only),
and reported a/>-value of >0.20. In Travier et al. (2002), the standardized incidence ratio
estimate is 1.20 (95% CI: 073, 2.18) for dry-cleaners and laundry workers in both 1960 and 1970
Censuses, compared to 1.02 (95% CI: 0.84, 1.24) for only subjects in one of these census.
Standardized incidence ratio estimates for both males and females with tetrachloroethylene
exposure in Selden and Ahlborg (2011) appeared to decrease monotonically with increasing
employment duration.
23In Lynge et al. (1995), all 17 primary liver cancer deaths occurred among laundry workers, and a risk estimate and
associated 95% CIs were not presented for dry cleaners.
4-132
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort Studies
Biologically monitored workers
All subjects
Not reported
Aerospace workers (Lockheed)
Routine exposure to PCE
Routine-Intermittent exposure duration to PCE
0
5yr
p-value for trend
2.05 (0.83, 4.23)
Not reported
1.0a
1.38 (0.40, 4.69)
1.17(0.39,3.47)
1.29 (0.46, 3.65)
p > 0.20
7
22
3
4
5
Chemical workers
PCE
1.8(0.8,4.3)
6
Electronic factory workers (Taiwan)
All Subjects
Males
Females
Females
Not reported
Not reported
0.79(0.55, 1.10)
0
0.69 exp
0
0.57 exp
36
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females and 0.7
umol/L in males (median)], follow-up 1974-1992, external referents (SIR)
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after 1960,
worked at least 1 yr, follow-up 1960-1996, job exposure matrix without
quantitative estimate of PCE intensity, 1987-1988 8-h TWA PCE
concentration (atmospheric monitoring) 3 ppm [mean] and 9.5 ppm
[median], external reference for routine exposure (SMR) and internal
references (workers with no chemical exposures) for routine-intermittent
PCE exposure (RR), liver and biliary tract (ICD-9, 155, 156)
Bond et al. (1990)
Nested case-control study with cohort (n = 21,437 males), follow-up
1940-1982, 44 liver and biliary tract deaths, unmatched controls randomly
selected from cohort, PCE and 10 other potential exposures assigned to
individual subjects based on company records, Mantel-Haenxzel %2 (OR)
Chang et al. (2003): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997, multiple solvents
exposure, does not identify PCE exposure to individual subjects, cancer
mortality, external referents (SMR) (Chang etal. 2003). primary liver
cancer (A095)
63,982 females, follow-up 1979-2001, factory employment proxy for
exposure, multiple solvents exposures and PCE not identified to individual
subjects, cancer incidence, external referents, analyses lagged 10 yr (SIR),
liver and interhepatic bile ducts (Sung etal.. 2007)
-^
oo
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
Not reported
Dry -cleaner and laundry workers
All laundry worker and dry cleaners
Males
Females
1.30(0.93, 1.78)
1.26 (0.69, 2.21)
1.32(0.88, 1.91)
39
11
28
All subjects
Semiquantitative exposure score
0.8 (0.4, 1.5)
Not reported
10
Laundry workers and dry cleaners in 1960
Census
Males
Females
1.22(1.03, 1.45)b
1.30 (0.97, 1.67)b
1.09 (0.70, 1.56)c
1.52 (0.83, 2.43)d
1.61 (0.88, 2.57)e
1.18(0.94, 1.44)b
1.26 (0.82, 1.81)°
1.05 (0.75, 1.40)d
1.39 (0.87, 2.04)e
138
52
25
14
14
86
25
39
22
Reference
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever exposed to
mixed solvents, 85 1 ever-exposed to PCE), employed at least 1 yr from
1952 to 1956, follow-up 1973-2000, job exposure matrix (intensity),
internal referent (workers with no chemical exposures (RR)
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or 1970
Census (Denmark, Finland, Norway), follow-up 1971-1987 or 1991, PCE
not identified to individual subjects, external referents (SIR), Primary liver
cancer (ICD-7, 155.0)
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union members
(1945-1978), follow-up 1979-1993, semiquantitative cumulative exposure
surrogate to dry clean solvents, cancer mortality, external referents (SMR),
liver and gallbladder (ICDA-8, 155)
Ji and Hemminki. (2005c)
9,255 Swedish men and 14,974 Swedish women employed in 1960 (men) or
1970 (women) as laundry worker or dry cleaner, follow-up
1961/1970-2000, PCE not identified to individual subjects, external
referent (SIR) and adjusted for age, period and socioeconomic status
-^
oo
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer (continued)
Exposure group
Launderers and dry cleaners
Males
Females
Cumulative exposure chlorinated HCs
None
<5 ppm-yr
5-49 ppm-yr
>50 ppm-yr
All laundry worker and dry cleaners
Males
Females
Dry cleaner
Laundry worker
Launderer and dry cleaner
Male
Female
Relative risk
(95% CI)
1.22 (0.56, 2.33)
2.91 (0.35, 4.26)
1.05(0.42,2.16)
1.0a
1.25 (0.80, 1.95)b
1.23 (0.68, 2.24)c
1.13 (0.84, 1.53)b
1.22 (0.83, 1.80)c
2.65(1.38, 5. ll)b
3.59 (1.71, 7.57)c
2.19(0.88,4.51)
3.33(1.34,6.87)
Not reported
Not reported
1.23 (1.02, 1.49)
1.13 (0.76, 1.63)
1.27(1.01, 1.57)
No. obs.
events
9
2
7
1,618
20
11
44
27
9
7
7
0
1.1 exp
7
0 cases
17 cases
113
29
84
Reference
Lindbohm et al. (2009)
Finnish population born 1906-1945 and participated in 1970 Census,
follow-up 1971-1995, Finnish cancer registry, 1,691 males and 783 female
primary liver cancers, longest held occupation reported on 1970 Census,
laundry and dry -cleaner exposure surrogate, external referent for analyses
examining job title (SIR) and all-other job titles for analyses for chlorinated
hydrocarbon (RR) adjusted for age, period, social class, smoking and
alcohol consumption
Lynge and Thygsen (1990): Lynge et al. (1995)
10,600 Danish men and women, 20-64 yr old, employed in 1970 as laundry
worker, dry cleaners and textile dye workers, follow-up 1970-1980,
external referents CSIR). Primary liver cancer CICD-7. 155) CLynse and
Thvsesen, 1990)
Nested case-control study within Lynge and Thygsen (1990). 17 primary
liver cancer cases in men and women, follow-up 1970-1987, 85 controls
randomly selected from within cohort, matched on sex, age, and occupation,
dry cleaner assigned using occupation and workplace on 1970 Census form,
logistic regression (OR) (Lvnse et al., 1995)
Pukkala et al. (2009)
Men and women participating in national census on or before 1990, 5
Nordic countries (Denmark, Finland, Iceland, Norway, Sweden), 30-64 yr,
follow-up 2005, occupational title of launderer and dry cleaner in any
census, external referents (SIR), Primary liver cancer (ICD-7, 155)
-^
oo
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
All subjects
Exposure duration/time since 1st employment
PCE-only subjects
0.13(0.00,0.73)
0.20 (0,00, 1.01)
Not reported
1
1
0
Dry-cleaners and laundry workers
Males
Females
PCE
Males
Duration of employment
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
All subjects, 1960 or 1970 Census in laundry and
dry cleaner or related occupation and industry
All subjects in 1960 and 1970 in laundry and dry
cleaner occupation and industry
1.02 (0.84, 1.24)
1.26(0.73,2.18)
105
13
Reference
Travier et al. (2002)
Swedish men and women identified as laundry worker, dry cleaner, or
presser (occupational title), in the laundry, ironing, or dyeing industry or
related industry in 1960 or 1970 (543,036 person-years); or, as laundry
worker, dry cleaner, or presser (occupational and job title) (46,933 person-
years) in both censuses, follow-up 1971-1989, external referents (SIR),
liver and biliary passages
Case-Control Studies
5 University Hospitals, United States (AL, FL, MA, NC, PA)
Laundry and dry cleaning occupation
Not reported
0
France
Laundry and dry cleaning occupation
Not reported
Laundry and dry-cleaning operatives
Not reported
2
Austin et al. (1987)
80 histologically confirmed hepatocellular carcinoma cases, 18-84 yr, years
not identified, 161 hospital controls matched on sex, age, race, and study
center, unknown interview methods, exposure surrogate jobs held >6 mo,
OR from logistic regression
Ferrand et al. (2008)
125 hepatocellular carcinoma in men, lacking HBV and HCV infection,
identified from four hospitals, <75 yr, 2000-2003, 142 hospital controls in
other departments, face-to-face interview, job title >6 mo as exposure
surrogate, OR from logistic regression model and adjust for hospital, age,
and alcohol consumption
Houten and Sonnesso (1980)
102 primary liver cancer cases in men and women, identified from hospital
records, 1956-1965, controls were all other hospitalized cancer cases, serf-
reported occupation at time of hospitalization, %2 comparing distribution of
job titles
-^
oo
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Nordic Countries (Denmark, Finland, Norway, Sweden)
Unexposed
Dry cleaner
Other in dry cleaning
Unclassifiable
Duration of employment in dry cleaning
10yr
Unknown
l.Oa
0.76 (0.38, 1.52)
0.42 (0.09, 1.89)
1.11(0.59,2.09)
Not reported
Not reported
1.21 (0.43, 3.44)
0.70 (0.26, 1.92)
2.88(0.21,38.81)
58
95
22
121
5
5
1
New Jersey (United States)
Laundering, cleaning, and other garment services
Laundering, cleaning, and other garment services
2.50(1.02, 6.14)f
2.29(0.85, 6.13)g
10
8
Dry-cleaning services
Dry-cleaning operators
0.98 (0.44, 2.20)
0.55(0.17, 1.75)
11
4
Reference
Lynge et al. (2006)
Case-control study among 46,768 Danish, Finnish, Norwegian, and
Swedish men and women employed in 1960 as laundry worker or dry
cleaner, follow-up 1970-1971 to 1997-2001, 72 incident esophageal cancer
cases, 6 controls per case randomly selected from cohort matched on
country, sex, age, calendar period at diagnosis time, occupational task at
1970 Census proxy for exposure, RR adjusted for matching criteria
Stemhagen et al. (1983)
265 histologically confirmed primary liver cancer cases and deaths,
1975-1980, New Jersey State Cancer Registry, 530 hospital controls
matched on age, race, sex, county of residence, vital status, in-person
interview, job title and industry coded to SIC/SOC, OR estimating using
Mantel-Haenszel with matched case-control set and not adjusted for
personal or lifestyle factors
Suarez et al. (1989)
1,742 primary liver cancer deaths, 1969-1980, 1,742 dead controls,
frequency matched on age, sex, race, and year death, Texas vital records,
job tile on death certificate, OR from Mantel-Haenszel analyses for race and
sex separately and adjusted for age
oo
oo
-------
Table 4-14. Summary of human studies on tetrachloroethylene exposure and liver cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events Reference
Geographic-Based Studies
Taoyuan, Taiwan
Residence in upstream village
Residence in downstream village
1.0a
2.57(1.21,5.46)
30
Hausjarvi and Hattula, Finland
Hausjarvi
Hattula
0.7 (0.3, 1.4)
0.6 (0.2, 1.3)
7
6
Lee et al. (2003)
Population of two villages surrounding electronic factory
(Suns et al., 2007; Chans et al., 2005; Chans et al., 2003), 50
liver cancer deaths, primary, underlying, or underlying
condition as cause of death, 1966-1997, residence as
recorded on death certificate, MOR from logistic regression
adjusted for age and period
Vartiainen et al. (1993)
Lymphopoietic cancers, liver cancer and all cancers among
residents with PCE and other solvents in drinking water,
1953-1991, no subject-level exposure information, cancer
rates of Finnish population referent (SIR)
oo
VO
a Referent.
b SIR or RR for liver, biliary tract, and gallbladder cancers.
0 SIR or RR for hepatocellular carcinoma.
d SIR for gallbladder cancer.
e SIR for all other liver cancers.
f In Stemhagen et al. (1983). odds ratio for primary liver cancer and work in laundering, cleaning, and other garment services industry.
g In Stemhagen et al. (1983). odds ratio for hepatocellular carcinoma and work in laundering, cleaning, and other garment services industry.
HBV = hepatitis B virus, HCV = hepatitis C virus, ICD = International Classification of Disease, ICDA = International Classification of Disease, Amended,
ISCO = International Standard Classification of Occupation, ISIC = International Standard Industry Classification, JEM = job-exposure-matrix, MOR = mortality
odds ratio, PCE = tetrachloroethylene, TWA = time-weighted-average.
-------
Risk factors for liver cancer include alcohol and hepatitis B and C viruses, with diabetes
mellitus suggested based on recent epidemiologic studies (El-Serag, 2007). None of the cohort
or case-control studies on liver cancer and tetrachloroethylene controlled for these potential risk
factors.
In conclusion, studies carrying greater weight in the analysis based on a large number of
observed events or exposed cases or a strong exposure-assessment approach, show a mixed
pattern of results. The one case-control study with a large number of exposed liver cancer cases
and a relatively high quality exposure-assessment methodology reported an odds ratio estimate
of 0.76 (95% CI: 0.38, 1.72) for liver cancer and dry cleaning (Lynge et al.. 2006). A recent
multiple Nordic country cohort study and two cohort studies of Swedish subjects with broad
exposure-assessment approaches, and whose subjects overlapped with Lynge et al. (2006),
reported SIRs of 1.02 (95% CI: 0.84, 1.24), 1.22 (95% CI: 1.03, 1.45), and 1.23 (95% CI: 1.02,
1.49) for liver and biliary tract cancer and work as a dry-cleaner or laundry worker (Pukkala et
al.. 2009: Ji and Hemminki, 2005c: Travier et al.. 2002). The study of Lindbohm et al. (2009) of
Finnish dry-cleaner and laundry workers whose subjects overlap the larger multiple-country
study of Pukkala et al. (2009) and that carries less weight in the analysis due to fewer observed
liver and biliary cancer cases supports observations in Swedish or the five Nordic country dry-
cleaner and laundry worker studies (Pukkala et al., 2009; Ji and Hemminki, 2005c). Three other
studies with strong exposure-assessment approaches specific to tetrachloroethylene but whose
risk estimates are based on fewer observed liver cancer cases or deaths provide support for an
association between liver cancer and tetrachloroethylene, risk estimates were 1.21 to 2.05
(Selden and Ahlborg, 2011: Boiceetal., 1999: Bond etal., 1990). However, dry cleaning
workers did not have a higher liver cancer risk estimate than laundry workers or other categories
of dry cleaning workers (Selden and Ahlborg, 2011; Lynge et al., 2006). . An exposure-
response relationship was not observed, and the SIR for tetrachloroethylene-exposed subjects
with the longest employment duration in Selden and Ahlborg (2011) was lower than that for
shorter employment duration. Potential confounding may be an alternative explanation as no
study adjusted for known and suspected risk factors for liver cancer (Selden and Ahlborg, 2011;
Pukkala et al.. 2009: Lynge et al.. 2006: Ji and Hemminki. 2005c: Travier et al.. 2002: Boice et
al., 1999: Bond et al., 1990). Nine other cohort and case-control studies with fewer observed
events and/or a broad exposure-assessment methodology carried less weight in the analysis;
these studies also reported a mixed pattern of results (Calvert et al., 2011: Lindbohm et al., 2009:
Sung et al., 2007: Blair etal., 2003: Lee etal., 2003: Lyngeetal., 1995: Vartiainen et al., 1993:
Suarez etal., 1989: Stemhagen et al., 1983). Lee et al. (2003) reported a risk estimate of 2.57
(95% CI: 1.21, 5.46) for the association between liver cancer and residence in a village with
groundwater contamination, was in region with a high prevalence of HCV and did not control for
4-140
-------
HCV status in the statistical analysis; potential confounding from HCV may be an alternative
explanation for the observed association.
4.3.2. Animal Studies
Liver toxicity and cancer have been observed in laboratory animal studies following
exposure to tetrachloroethylene through multiple routes of exposure. The sections below
describe studies of liver toxicity (refer to Section 4.3.2.1) and cancer (refer to Section 4.3.2.2).
These studies are summarized in Tables 4-15 and 4-16, respectively.
4.3.2.1. Liver Toxicity
Tetrachloroethylene causes hepatic toxicity in multiple species, including several strains
of rats and mice. Adverse effects on the liver have been observed in studies of animals exposed
to tetrachloroethylene by multiple routes of exposure, including inhalation and oral gavage.
Hepatic effects observed after subchronic or chronic inhalation exposure to tetrachloroethylene
include increased liver weight (Kyrklund et al., 1990; Kj ell strand et al., 1984): hypertrophy
(Odum et al., 1988): fatty degeneration (Odum et al., 1988: Kylin et al., 1963): peroxisome
proliferation (Odumetal., 1988): other histological changes (Odumetal., 1988: NTP, 1986:
Kj ell strand etal.. 1984): and degeneration and necrosis (JISA, 1993: NTP, 1986). When
administered by oral gavage, tetrachloroethylene also causes hepatic toxicity, including increased
liver enzymes, increased liver weights, histological changes, degeneration and necrosis,
regenerative repair, and polyploidy (Philip et al., 2007: Ebrahim et al., 1996: Jonker et al., 1996:
Berman et al., 1995: Goldsworthv and Popp, 1987: Buben and O'Flahertv, 1985: NCL 1977).
Table 4-15 presents a summary of inhalation and oral rodent liver toxicity studies, which are
briefly described below. This review focuses on studies that identify critical effects commonly
observed in tetrachloroethylene toxicity studies and could, accordingly, support oral and
inhalation reference values. The database of liver toxicity studies is more extensively reviewed
in prior assessments by ATSDR (1997a), NYSDOH (1997). and CalEPA (2001).
4-141
-------
Table 4-15. Summary of inhalation and oral rodent liver toxicity studies
Species/strain/sex/number
Mouse, B6C3FJ (both
sexes, 49 or 50 per sex per
dose group, total of -300
mice)
Mouse, Crj/BDFl (both
sexes, 50 per sex per dose
group, total of 400 mice)
Rat, F344/DuCrj (both
sexes, 50 per sex per dose
group, total of 400 rats)
Mouse, NMRI (both sexes,
10 per dose group)
Mouse, B6C3Fl5 and rat,
Sprague-Dawley (males
only, 3 per dose group)
Rat, F344; and mouse,
B6C3F! (both sexes, 5 per
dose group)
Rat, Sprague-Dawley
(males only, 8 per group)
Mice, albino (strain not
specified) (females only, 20
per dose group, 240 total)
Mouse, Swiss-Cox, male
(males, 4-6 per 1,500 and
2,000 mg/kg-day doses;
other doses, 12-15
mice/group)
Exposure level/duration
0, 100, 200 ppm for
104 wk, inhalation
0, 10, 50, 250 ppm for
110 wk, inhalation
0, 50, 200, 600 ppm for
110 wk, inhalation
0, 9, 37, 75, 150 ppm, 30
d, inhalation, continuous
(24h);and225(16h/d),
450 (8 h/d), 900 (4 h/d),
1,800 (2 h/d), or 3, 600
(1 h/d), inhalation
Radiolabeled PCE by
inhalation (10 or 600 ppm
for 6 h), or as a single oral
gavage dose (500 mg/kg)
0, 200 ppm (28 d only)
and 400 ppm (14, 21,28 d)
for 6 h/d, inhalation
0 or 320 ppm continuous
for 90 d; 0 or 320 ppm
continuous for 90 d
followed by a 30 d
recovery period, inhalation
0 or 200 ppm
4 h/d, 6 d/w for 1,2, 4 or 8
wk, inhalation
0, 20, 100, 200, 500,
1,000, 1,500, 2,000 mg/kg-
day for 6 wk, gavage
Effects
Liver degeneration and necrosis at >100 ppm
in males and at 200 ppm in females
Focal necrosis in males at >50 ppm;
liver degeneration in males and females at
250 ppm
Spongiosis hepatitis in males at 200 ppm and
higher; hyperplasia in males at 600 ppm
Increase in liver weight (>9 ppm);
morphological changes (>9 ppm); increased
plasma butylcholinesterase (>37 ppm)
Irreversible binding to hepatic
macromolecules at all exposures in male
mice and rats
Increased palmitoyl Co A in mice (3.7-fold)
and rats (1.3 -fold); increased peroxisome
proliferation in mouse liver in all sex, dose
and time groups; mitochondrial proliferation
in male mice at 400 ppm for 28 d;
increased relative liver weight, centrilobular
lipid accumulation in exposed mice of both
sexes
Significantly increased relative liver weight
after exposure; this was decreased following
recovery; decreased cholesterol following the
recovery period
Fatty degeneration after 1 wk; incidence
severity increased with longer exposure
Increased liver/body weight ratio at
100 mg/kg-day; increased triglycerides at
100 mg/kg-day; no change at 20 mg/kg-day
Reference
NTP
(1986)
JISA
(1993)
JISA
(1993)
Kjellstrand
etal.
(1984)
Schumann
etal.
(1980)
Odum et
al. (1988)
Kyrklund
etal.
(1990)
Kylin et al.
(1965)
Buben and
O 'Flaherty
(1985)
4-142
-------
Table 4-15. Summary of inhalation and oral rodent liver toxicity studies
(continued)
Species/strain/sex/number
Mouse, Swiss-Webster,
male (4 per dose group)
Rat, Wistar, female only
(10 rats per each control
group; 5 rats per dose
group)
Rat, F344 (males only, 5
per dose group) and mouse,
B6C3F! (males only, 5 per
dose group)
Mouse, Swiss, both sexes;
6 groups of 6 mice each
(EbrahimetaL 1996);
male only; 8 groups of 6
mice each (EbrahimetaL
2001)
Rat, F344 (females only, 8
per dose group)
Exposure level/duration
0, 150, 500, and
1,000 mg/kg-day, aqueous
gavage for 30 d
0, 600, and 2,400
mg/kg-day for 32 d, corn
oil gavage; alone or in
combination with other
compounds
(trichloroethylene,
hexachloro-l,2-butadiene,
1,1,2-trichloro-
3 ,3 ,3 -trifluoropropene)
0 or 1,000 mg/kg-day for
10 d, corn oil gavage
0 or 3, 000 mg/kg-day for
15 d, sesame oil gavage
0, 50, 150, 500, or 1,500
mg/kg-day, gavage, either
once or for 14 consecutive
days
Effects
Increased plasma ALT 24 hours to 14 d
after initial exposure; mild to moderate fatty
degeneration and necrosis, with focal
inflammatory cell infiltration; increased
mitotic figures and DNA synthesis (peaked
on 7 d, sustained at 14-30 d at all doses);
inhibition of PCE metabolism and TCA
production; no change in CYP2E1; CYP4A
increased at 7 but not 14 d, only at
1,000 mg/kg-day
Relative liver weight increases in animals
exposed to PCE alone or in combination;
hepatotoxicity at 600 mg/kg
Increased relative liver weight in rats and
mice; 4. 3 -fold PCO increase in mice;
modest but not significant (1.4-fold) PCO
increase in rats
Significant increase in liver weight;
degeneration and necrosis of hepatocytes;
decreased blood glucose (glucose effects
mitigated by coexposures to 2-deoxy-
D glucose and vitamin E) (EbrahimetaL
1996);
Decreased membrane-bound
Na+K+-ATPases and Mg2+-ATPases activity
but increased Ca-ATPase activity; mitigated
by coexposure to 2-deoxy-D-glucose and
vitamin E, and taurine
Increased relative liver weight, elevated
ALT and hepatocellular hypertrophy at
1,500 mg/kg-day
Reference
Philip et al.
(2007)
Jonker et al.
(1996)
Goldsworthy
and Popp
(1987)
Ebrahim et
al. (2001:
1996)
Herman et
al. (1995)
4-143
-------
4.3.2.1.1. Inhalation
Hepatic toxicity was observed in chronic lifetime inhalation bioassays of
tetrachloroethylene in mice conducted by the National Toxicology Program (NTP, 1986), and the
Japan Industrial Safety Association (JISA, 1993). The NTP study administered
tetrachloroethylene to groups of 50 F344 rats of each sex (0, 200, or 400 ppm), or groups of 49
or 50 B6C3Fi mice (0, 100, or 200 ppm), for 6 hours/day, 5 days/week, for 103 weeks (NTP.
1986). In addition to liver tumors in mice of both sexes, liver degeneration was reported in 2/49,
8/49, and 14/50 of males and in 1/49, 2/50, and 13/50 of females. Degeneration was
characterized by a variety of histological features, including cytoplasmic vacuolation,
hepatocellular necrosis, inflammatory cell infiltrates, pigment in cells, oval cell hyperplasia, and
regenerative foci. Liver necrosis was observed at increased incidence in dosed males (1/49,
6/49, and 15/50) and in females at 400 ppm (3/48, 5/50, and 9/50). Nuclear inclusions increased
in male mice (2/49, 5/49, and 9/50). No dose-related liver effects were reported in the rats.
In the 13-week NTP study, groups of 10 rats and mice of each sex were exposed to air
containing tetrachloroethylene for 6 hours/day, 5 days/week, for 13 weeks (0, 100, 200, 400,
800, or 1,600 ppm). Some rats in the high-dose group died before the end of the studies (4/10
male, 7/10 female). In mice, 2/10 males and 4/10 females in the high-dose group died before the
end of studies. Tetrachloroethylene (200 ppm and above) increased the incidence of hepatic
congestion in male and female rats. In mice of both sexes, liver lesions (leukocytic infiltration,
centrilobular necrosis, and bile stasis) were observed at 400, 800, or 1,600 ppm. Mitotic
alterations were increased at 200 ppm in male mice. No hepatic effects were reported in the
single exposure or 14-day studies.
In the Japan Industrial Safety Association (1993) study [some results reported in Nagano
et al. (1998)], male and female Crj/BDFl mice were exposed to 0, 10, 50, and 250 ppm
tetrachloroethylene for 104 weeks and sacrificed at 110 weeks. In addition to hepatocellular
carcinomas and adenomas in the mice, telangiectasis (vascular lesions formed by dilation of a
group of small blood vessels) and focal necrosis occurred in males at 50 ppm and above. Liver
degeneration was observed at 250 ppm in both sexes. Hemangiomas or hemangiosarcomas,
occurring primarily in the liver or spleen, were also reported in the male mice. This study also
examined effects in F344/DuCrj rats exposed to 0, 50, 200, and 600 ppm for 104 weeks and
sacrificed at 110 weeks. Male—but not female—rats had excess incidence of spongiosis
hepatitis at 200 ppm and 600 ppm.
The lowest reported level for liver effects by inhalation in laboratory animals is in female
NMRI mice exposed for 30 days at 9 ppm [61 mg/m3; Kjellstrand et al. (1984)1. Significant
increases in liver weight as well as changes in liver morphology were observed in male and
4-144
-------
female mice exposed continuously to 9 ppm and higher concentrations of tetrachloroethylene for
30 days. Livers were enlarged, and vacuolization was evident. Reversible increases in levels of
the blood plasma enzyme butyrylcholinesterase were reported at all tetrachloroethylene
concentration levels at or above 37 ppm. The toxicological significance of the increased serum
cholinesterase is uncertain, and this effect of tetrachloroethylene has not been reported by other
investigators. After a recovery period, liver weight was still slightly elevated at 120 days after
cessation of tetrachloroethylene exposure for 30 days at 150 ppm. Total dose administered in the
continuous exposure experiment is not directly comparable to exposures during intermittent and
pulsed exposure experiments, which also found increased liver weight and increased serum
cholinesterase.
Schumann et al. (1980) administered radiolabeled tetrachloroethylene to male B6C3Fi
mice or Sprague-Dawley rats via inhalation (10 or 600 ppm for 6 hours). In mice, the percentage
metabolized based on recovery of the radiolabeled material was determined to be 88% for a
6-hour inhalation exposure of 10 ppm (as compared to only 17% for a single oral gavage dose of
500 mg/kg). At all dose levels in both rats and mice, irreversible binding of radioactivity to
hepatic macromolecules was observed. DNA binding was not seen. In mice, binding peaked at
the termination of the 6-hour inhalation exposure and 6 hours after the single oral dose. In
contrast, binding in the rat peaked 24 hours after either oral or inhalation exposure.
Odum et al. (1988) exposed groups of male and female F344 rats and B6C3Fi mice by
inhalation for 6 hours/day to 200 ppm (28 days only) or 400 ppm (for 14, 21, or 28 days)
tetrachloroethylene. Five animals per group were exposed. In both sexes, hepatic palmitoyl
coenzyme A (PCO) activity was increased in mice (up to 3.6-fold) and, to a lesser extent, in rats
(up to 1.3-fold). Modest PCO increases were also observed in the kidney of male rats at 200
ppm at 28 days (1.3-fold) but not 400 ppm at 14, 21, or 28 days. In female rat kidney, PCO was
elevated (approximately 1.6-fold) at all doses and times. However, peroxisome proliferation was
not observed in rat kidney upon microscopy. In contrast, hepatic peroxisome proliferation was
noted in mouse liver for all sexes, times, and dose groups on electron microscopy, and the
percentage of cytoplasm occupied by peroxisomes also increased. Catalase, another peroxisomal
enzyme, was unaffected by tetrachloroethylene; male mice exposed at 400 ppm showed the only
moderate (1.4-fold) increase. Mitochondrial proliferation was observed at 28 days in 400 ppm
male mice. In addition, a time-dependent proliferation of smooth endoplasmic reticulum in the
liver of both sexes correlated well with centrilobular hypertrophy. Tetrachloroethylene caused
centrilobular lipid accumulation in male and female mice. Relative liver weight was increased in
mice of both sexes.
Kyrklund et al. (1990) exposed male Sprague-Dawley rats to 320 ppm
tetrachloroethylene continuously for 90 days, followed by a 30 day recovery period. Relative
4-145
-------
liver weight was significantly increased in rats examined at the end of the exposure period. A
slight increase in relative liver weight was also observed in the recovered, solvent-treated group.
Cholesterol was also decreased, but this effect was only statistically significant in the
tetrachloroethylene-exposed group that also included a recovery period.
Kylin et al. (1965) exposed female albino mice (strain not specified) to 200 ppm
tetrachloroethylene for 4 hours daily, 6 days a week, for 1, 2, 4, or 8 weeks. Hepatic effects were
evaluated by histological examination and determination of extractable liver fat. The incidence
and severity of fatty degeneration increased with longer exposure. Neither liver cell necrosis nor
cirrhosis was observed.
4.3.2.1.2. Oral
In addition to studying the effects of inhalation and a single oral gavage dose (500
mg/kg), as described above, Schumann et al. (1980) also administered 100, 250, 500, or
1,000 mg/kg to male B6C3Fi mice or Sprague-Dawley rats as a daily oral dose for 11 days. At
all doses in mice, histopathological evidence of hepatocellular swelling in the centrilobular
region, a decrease in liver DNA content, and an increase in DNA synthesis was observed. At
>250 mg/kg, tetrachloroethylene increased the absolute or relative liver weights in mice. In rats,
no statistically significant treatment-related effects were observed at 100, 250, or 500 mg/kg;
however, increased liver DNA synthesis was observed in one rat in the 250 mg/kg-dose group,
resulting in a large variation in liver DNA synthesis at that exposure level.
Buben and O'Flaherty (1985) exposed male Swiss-Cox mice to tetrachloroethylene doses
of 0, 20, 100, 200, 500, 1,000, 1,500, or 2,000 mg/kg-day, 5 days/week, for 6 weeks.
Liver/body-weight ratios and liver triglycerides were significantly increased at 100 mg/kg-day or
more. Enlarged hepatocytes, karyorrhexis (disintegration of the nucleus), necrosis, polyploidy in
the centrilobular region, and lipid accumulation were evident upon histopathological
examination of mice exposed to 200 or 1,000 mg/kg. Other indices of tetrachloroethylene
hepatotoxicity (decreased glucose-6-phosphatase activity, and increased serum glutamic pyruvic
transaminase activity) were significantly increased at 500 or more mg/kg-day. The liver
response (percentage increase in either liver weight/body-weight ratios or G6P inhibition) was
highly correlated with the amount of tetrachloroethylene metabolized, and a plot of these
measures against total urinary metabolites was linear (r2 = 0.97 and 0.98 for increases in
liver/body weight ratios and G6P inhibition, respectively). The LOAEL was 100 mg/kg-day.
Philip et al. (2007) exposed male 6-7-week old Swiss-Webster mice via aqueous gavage
to three dose levels (150, 500, and 1,000 mg/kg-day) for 30 days. At the highest exposure,
mortality was 10% due to apparent CNS toxicity (tremors and ataxia). Significant liver injury
(as assessed by increased plasma ALT) was evident 24 hours after the first, single exposure at all
4-146
-------
doses. ALT levels decreased transiently to control levels by 30 days thereafter. Histopathology
was consistent with mild-to-moderate fatty degeneration and necrosis. Necrotic hepatocytes had
either pyknotic, karyorrhectic, or karyolitic nuclei. Infiltration of neutrophils and macrophages
was present near necrotic foci. Regenerative repair was evident in the two higher dose groups by
30 days of exposure, with observed increases in mitotic figures, tritiated thymidine incorporation
with pulse-labeling, and PCNA immunostaining. At the two higher dose groups, a robust
increase DNA synthesis peaked on 7 days, was sustained at 14 days, and had returned to control
levels at 30 days of exposure. The amount of blood and liver TCA declined, while
tetrachloroethylene levels increased, from 1 to 30 days. This is consistent with an inhibition of
tetrachloroethylene metabolism. Because CYP2E1 levels and activity were unchanged, a
different CYP isoform is suggested to be critical for tetrachloroethylene metabolism. The study
found a transient increase in hepatic CYP4A expression, a marker of PPARa induction, which
was evident at 7—but not 14—days at the highest dose. This finding suggests that peroxisome
proliferation is not a sustained response in spite of continued tetrachloroethylene exposure.
In a study by Jonker et al. (1996), hepatotoxicity was observed in female Wistar rats
administered tetrachloroethylene (600 or 2,400 mg/kg-day) daily via corn oil oral gavage for
32 days. Relative liver weight was increased on exposure to tetrachloroethylene alone and in
combination with other hepatotoxicants (trichloroethylene, hexachloro-l,2-butadiene, and
l,l,2-trichloro-3,3,3-trifluoropropene). One high-dose animal died as a result of
tetrachloroethylene treatment, and one animal exposed to the high-dose combination also died as
a result of treatment. Hepatotoxic effects were noted at 600 mg/kg.
Goldsworthy and Popp (1987) administered tetrachloroethylene (1,000 mg/kg-day) by
corn oil gavage to 5 male F344 rats and 5 male B6C3Fi mice for 10 days. In
tetrachloroethylene-exposed rats, cyanide-insensitive palmitoyl CoA oxidation (PCO) was
modestly—although not significantly—elevated in the liver (1.4-fold increase) and kidney
(1.7-fold increase). In mice, tetrachloroethylene exposure increased PCO activity by 4.3-fold in
liver and by 2.3-fold in kidney. Relative liver weight was increased in rats and mice with
tetrachloroethylene exposure, but relative kidney weight was unaffected. A comparison of corn
oil with methyl cellulose revealed no effect of the gavage vehicle on tetrachloroethylene-induced
PCO. A less-than-additive effect of trichloroethylene (1,000 mg/kg) administered together with
tetrachloroethylene on PCO induction was seen.
Ebrahim et al. (1996) orally administered 3,000 mg/kg-day tetrachloroethylene in sesame
oil to male and female Swiss mice for 15 days and observed a significant increase in liver weight
and degeneration and necrosis of hepatocytes. These changes occurred simultaneously with a
decrease in blood glucose; elevated activities of enzymes hexokinase, aldolase, and
phosphoglucoisomerase; and decreased activities of gluconeogenic enzymes. Blood glucose
4-147
-------
levels were significantly decreased, and this effect was mitigated by concomitant exposure to
2-deoxy-D-glucose (2DG) and vitamin E. A follow-up study by this group further examined the
potential protective properties of 2DG and vitamin E as well as taurine against
tetrachloroethylene-induced membrane damage (Ebrahim et al., 2001). This study exposed male
albino Swiss mice to the same doses used in the previous study with the addition of a taurine
exposed group (tetrachloroethylene in sesame oil 3,000 mg/kg-day for 15 days orally by
intubation; tetrachloroethylene plus 2DG 500 mg/kg-day by i.p. injection once a day for 15 days;
tetrachloroethylene plus vitamin E 400 mg/kg-day by oral intubation once a day for 15 days; and
tetrachloroethylene plus taurine 100 mg/kg-day by oral intubation once a day for 15 days).
Compared to control cells in the liver, membrane-bound Na+K+-ATPases and Mg2+-ATPases
activity was significantly decreased (p < 0.001), while Ca-ATPases activity was increased
(p < 0.001), following exposure to tetrachloroethylene alone. These levels remained near normal
in the animals exposed to tetrachloroethylene along with 2DG, vitamin E, or taurine. This return
to normal levels following exposure to vitamin E and taurine may be due to their antioxidant
abilities, and reduced oxidative stress in exposed cells.
Berman et al. (1995) reported liver and kidney toxicity in a study of female F344 rats
exposed for 14 days by oral gavage to 0, 50, 150, 500, or 1,500 mg/kg-day tetrachloroethylene.
The reported LOAEL was 1,500 mg/kg-day. Hepatic effects included increased relative liver
weight, elevated ALT, and hepatocellular hypertrophy.
4.3.2.1.3. Intraperitoneal injection
Binding of radiolabelled tetrachloroethylene to hepatic DNA was observed in mice
following i.p. injection (Mazzullo et al., 1987) but not inhalation and oral exposure [Schumann et
al. (1980), described above]. Using a reportedly more sensitive assay, low levels of DNA
binding were observed in vivo in BALB/C mouse liver 22 hours after i.p. injection (1.4 mg/kg),
with 10 fold lower levels observed in Wistar rat liver than mouse liver (Mazzullo et al., 1987).
Still lower levels of DNA binding were observed in the kidney and stomach of mice and rats in
this study. Binding to RNA and protein was always higher than binding to DNA in both mice
and rats. Binding to calf thymus DNA in an in vitro study increased in the presence of
microsomal fractions from both mouse and rat liver, but not kidney, lung or stomach. Cytosolic
fractions from rat and mouse liver, kidney, lung, and stomach all induced binding of
tetrachloroethylene to calf thymus DNA, with enzymes from both mouse and rat livers and
mouse lung being the most efficient. DNA binding in the presence of both cytosolic and
microsomal fractions was similar to cytosolic fraction alone. Phenobarbital pretreatment of
animals increased cytosol-mediated binding, but had only a slight effect on microsomal-mediated
binding. Binding in the presence of rat liver microsomal fraction was also increased (17-fold)
4-148
-------
with addition of GSH but decreased in the presence of superoxide dismutase or mannitol
(Mazzullo et al.. 1987).
4.3.2.2. Liver Cancer
In carcinogenicity bioassays, tetrachloroethylene caused a statistically significant
increase in the incidence of hepatocellular carcinomas in both sexes of B6C3Fi mice following
either oral gavage administration or inhalation exposure (NTP, 1986; NCI, 1977). Both sexes of
Crj:BDFl mice have also been shown to develop an increased incidence of hepatocellular
carcinomas when exposed to tetrachloroethylene by inhalation (Nagano et al., 1998; JISA, 1993).
Additionally, in male Crj:BDFl mice, hemangiosarcomas (reported as malignant
hemangioendotheliomas) in the liver and both hemangiosarcomas and combined
hemangiosarcomas and hemangiomas (reported as benign hemangioendotheliomas) of the spleen
were increased. The studies are presented in Table 4-16 and are briefly summarized here.
4.3.2.2.1. Inhalation
The NTP (1986) inhalation bioassay exposed groups of 50 B6C3Fi mice of each sex to
(epichlorohydrin free) tetrachloroethylene concentrations of 0, 100, or 200 ppm, 6 hours/day,
5 days/week, for 103 weeks. Tetrachloroethylene caused statistically significant dose-related
increases in the incidences of hepatocellular carcinoma and in combined hepatocellular adenoma
and carcinoma in both sexes. Hepatocellular neoplasms (adenomas and carcinomas combined)
were reported in 17/49, 31/49, and 41/50 males, and 4/45, 17/42, and 38/48 females. In male
mice, hepatocellular carcinomas metastasized to the lungs in 2/49, 7/49, and 1/50 animals.
Metastatic hepatocellular carcinomas were found in the lungs of 0/48, 2/50, and 7/50 female
mice.
A Japan bioassay exposed groups of 50 Crj:BDFl mice of each sex to 0, 10, 50, and 250
ppm tetrachloroethylene, 6 hours/day, 5 days/week, for 104 weeks, and the terminal sacrifice
was performed at 110 weeks. Both males and females showed dose-related increased incidences
of liver carcinomas and combined liver adenomas and carcinomas. The incidences of
hepatocellular adenomas were 7/50, 13/50, 8/50, and 26/50 in males and 3/50, 3/47, 7/49, and
26/49 in females in control, 10, 50, and 250 ppm dose groups, respectively. Male hepatocellular
carcinomas also increased, with reported incidences of 7/50, 8/50, 12/50, and 25/50 in males and
0/50, 0/47, 0/49, and 14/49 in females in control, 10, 50, and 250 ppm dose groups, respectively.
Liver hemangiosarcomas (reported as malignant hemangioendotheliomas) were also increased in
males. In the spleen, both hemangiosarcomas and combined hemangiosarcomas and
hemangiomas (reported as benign hemangioendotheliomas) were increased in males.
4-149
-------
4.3.2.2.2. Oral
In the NCI (1977) tetrachloroethylene mouse gavage study, 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). The test sample was estimated to contain
epichlorohydrin concentrations of less than 500 ppm. 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 compared with untreated
controls or vehicle controls. The time to tumor was significantly decreased in treated mice.
4-150
-------
Table 4-16. Incidence of hepatic tumors in rodents exposed to
tetrachloroethylene
Bioassay
NCI (1977)
B6C3FJ miceb
Gavage:
5d/wk,
78 wk
NCI (1977)d
Osborne-Mendel rats
Gavage:
5d/wk,
78 wk
NTP(1986)
B6C3FJ mice
Inhalation:
6h/d,
5d/wk,
104 wk
NTP(1986)
F344/N rats
Inhalation:
6h/d,
5d/wk,
104 wk
Administered
dose/exposure
Vehicle
450 mg/kg-day
900 mg/kg-day
Vehicle
300 mg/kg-daya
600 mg/kg-day
Vehicle
500 mg/kg-day
1,000 mg/kg-day
Vehicle
500 mg/kg-day
1,000 mg/kg-day
Oppm
100 ppm
200 ppm
Oppm
100 ppm
200 ppm
Oppm
200 ppm
400 ppm
Oppm
200 ppm
400 ppm
Continuous
equivalent
exposure
0
332 mg/kg-day
663 mg/kg-day
0
239 mg/kg-day
478 mg/kg-day
0
471 mg/kg-day
941 mg/kg-day
0
474 mg/kg-day
974 mg/kg-day
0
18 ppm
36 ppm
0
18 ppm
36 ppm
0
36 ppm
72 ppm
0
36 ppm
72 ppm
Sex
Male
Female
Male
Female
Male
Female
Male
Female
Hepatocellular
adenomas and
carcinomas
2/20 (10)
32/48 (67)
27/45 (60)
0/20 (0)
19/48 (40)
19/48 (40)
1/20 (0)
1/49 (0)
0/50 (0)
None reported
17/49 (35)
31/49(70)
41/50 (82)
4/45 (9)
17/42(40)
38/48 (79)
0/50 (0)
1/50 (2)
1/50 (2)
0/50
0/50
0/50
Hemangiomas or
hemangiosarcomas3
None reported0
None reported
None reported
None reported
1/49 (2)
0/49 (0)
0/50 (0)
0/48 (0)
3/50 (6)
0/50 (0)
0/50
0/50
0/50
0/50
0/50
0/50
4-151
-------
Table 4-16. Incidence of hepatic tumors in rodents exposed to
tetrachloroethylene (continued)
Bioassay
JISA (1993)
Crj:BDFl mice
Inhalation:
6h/d,
5d/wk,
104 wk
JISA (1993)
F344/DuCrj rats
Inhalation:
6h/d,
5d/wk,
104 wk
Administered
dose/exposure
Oppm
lOppm
50ppm
250 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oppm
50 ppm
200 ppm
600 ppm
Oppm
50 ppm
200 ppm
600 ppm
Continuous
equivalent
exposure
0
1.8 ppm
9.0 ppm
45 ppm
0
1.8 ppm
9.0 ppm
45 ppm
0
9 ppm
36 ppm
108 ppm
0
9 ppm
36 ppm
108 ppm
Sex
Male
Female
Male
Female
Hepatocellular
adenomas and
carcinomas
13/50 (28)
21/50 (43)
19/50 (40)
40/50 (82)
3/50 (6)
3/47 (6)
7/49 (15)
33/49 (67)
4/50
0/50
1/50
2/50
1/50 (2)
0/50 (0)
1/50 (2)
0/50 (0)
Hemangiomas or
hemangiosarcomasa
4/50 (4)
2/50 (2)
7/50 (13)
11/50(18)
1/50
0/47
2/49
3/49
0/50
0/50
0/50
0/50
1/50
0/50
0/50
0/50
a These tumors were reported as hemangioendotheliomas in the JISA (1993) report. The term has been updated to
hemangioma (benign) or hemangiosarcoma (malignant). Note that these incidences do not match those tabulated
in Table 12 of the JISA report summary. The incidences reported here represent a tabulation of
hemangioendotheliomas from the individual animal data provided in the JISA report.
b Administered gavage doses listed were increased after 11 wk by 100 mg/kg-day in each low-dose group or by
200 mg/kg-day in each high-dose group. Mice received the listed TWA daily doses through Week 78, and
surviving mice were observed up to study termination in Week 90.
0 None reported: Individual animal data were not available, and summary data did not include a line item for this
tumor type.
d Gavage doses listed were adjusted several times during the course of the study. Male rats received the listed TWA
daily doses through Week 78, and surviving animals were observed up to study termination in Week 110.
4.3.3. Summary of Liver Effects in Humans and Animals
Two of four studies of occupationally exposed dry cleaners showed indications of liver
toxicity, namely sonographic changes of the liver and altered serum concentrations of liver
enzymes indicative of liver injury. Frank liver disease was not observed among these workers
for a number of possible reasons: individuals with frank liver disease may not have been
included in cross-sectional studies because they had left the workforce due to their conditions,
the healthy worker effect, and other selection biases. LOAELs in these human studies were
between 12 and 16 ppm (TWA).
4-152
-------
Liver toxicity has been reported in multiple animal species by inhalation and oral
exposures to tetrachloroethylene. The effects are characterized by increased liver weight, fatty
changes, necrosis, inflammatory cell infiltration, triglyceride increases, and proliferation. The
mouse has been shown to be more sensitive to hepatic toxicity than the rat in multiple subchronic
and chronic studies [e.g., JISA (1993): NTP (1986): Schumann et al. (1980): NCI (1977)1. After
subchronic or chronic inhalation exposures in mice, liver toxicity is manifested by increased liver
weight (Ki ell strand etal.. 1984). liver enlargement (Odumetal., 1988: Ki ell strand etal.. 1984).
cytoplasmic vacuolation (fatty changes) (Odumetal.. 1988: NTP. 1986: Ki ell strand et al.. 1984).
centrolobular hepatocellular necrosis (JISA, 1993: NTP, 1986), and inflammatory cell infiltrates,
pigment in cells, oval cell hyperplasia, and regenerative foci (NTP, 1986). The LOAEL for the
inhalation studies—9 ppm—is from a 30 day-exposure mouse study reporting increased liver
weight and morphological changes and is supported by a finding of irreversible macromolecular
binding in mouse liver following a single, 6-hour exposure at 10 ppm. The JISA (1993) chronic
mouse inhalation bioassay reported liver necrotic foci at 50 ppm and higher. In two lifetime
inhalation cancer bioassays, increases in liver cancer occurred at 100 ppm and above, and there
was a significant dose-response trend in both studies.
With administration by oral gavage in mice, liver toxicity (increased liver weight,
hepatocellular swelling, necrosis, lipid accumulation, and increased DNA synthesis) has been
observed at 100 mg/kg-day (Buben and O'Flaherty, 1985: Schumann et al., 1980) and above
(Ebrahim et al.. 1996: Jonkeretal.. 1996: Berman et al.. 1995: Goldsworthv andPopp. 1987).
At 150 mg/kg-day administered for 30 days (Philip et al., 2007), tetrachloroethylene increased
ALT levels transiently and stimulated fatty degeneration and necrosis, with ensuing regenerative
repair. These findings support a LOAEL of 100 mg/kg-day and a NOAEL of 20 mg/kg-day.
For liver cancer, epidemiologic studies carrying greater weight in the analysis, based on a
large number of observed events or exposed cases, or a strong exposure-assessment approach,
show a mixed pattern of results. The one case-control study with a large number of exposed
liver cancer cases and a relatively high quality exposure-assessment methodology reported an
odds ratio estimate of 0.76 (95% CI: 0.38, 1.72) for liver cancer and dry cleaning (Lynge et al.,
2006). A recent multiple Nordic country cohort study and two cohort study of Swedish subjects
with broad exposure-assessment approaches and whose subjects overlapped with Lynge et al.
(2006) reported SIRs of 1.02, 1.22, and 1.23 for liver and biliary tract cancer and work as a dry
cleaner or laundry worker in Travier et al. (2002), Ji and Hemminki (2005c), and Pukkala et al.
(2009), respectively. Three other studies with strong exposure-assessment approaches specific to
tetrachloroethylene but whose risk estimates are based on fewer observed liver cancer cases or
deaths reported risk estimates of 1.21 to 2.05 for the association between liver cancer and
tetrachloroethylene (Selden and Ahlborg, 2011: Boice et al., 1999: Bond etal., 1990). However,
4-153
-------
dry cleaning workers did not have a higher liver cancer risk estimate than laundry workers or
other categories of dry cleaning workers (Selden and Ahlborg, 2011; Lynge et al., 2006).
Exposure response was not observed, and the SIR for tetrachloroethylene-exposed subject with
longest employment duration in Selden and Ahlborg (2011) was lower than that for subjects with
shorter employment duration. Potential confounding may be an alternative explanation as no
study adjusted for known and suspected risk factors for liver cancer (Selden and Ahlborg, 2011;
Pukkala et al., 2009; Lvnge et al., 2006; Ji and Hemminki, 2005c: Travier et al., 2002; Boice et
al., 1999; Bond et al., 1990). Nine other cohort and case-control studies with fewer observed
events and broad exposure-assessment methodology carried less weight in the analysis and
reported a pattern of mixed results (Calvert et al., 2011; Lindbohm et al., 2009; Sung et al., 2007;
Blair etal.. 2003: Lee etal.. 2003: Lvnge etal.. 1995: Vartiainen et al.. 1993: Suarezetal.. 1989:
Stemhaeen et al.. 1983). Lee et al. (2003) reported a risk estimate of 2.57 (95% CI: 1.21, 5.46)
for the association between liver cancer and residence in a village with groundwater
contamination, but subjects were from a region with a high prevalence of HCV infection, and
HCV status may confound the observed association.
Tetrachloroethylene caused a statistically significant increase in the incidence of liver
tumors in both sexes of mice in multiple carcinogenicity bioassays. A statistically significant
increase in the incidence of hepatocellular carcinomas in both sexes of B6C3Fi mice was
observed following either oral gavage administration or inhalation exposure (NTP, 1986: NCI,
1977). Both sexes of Crj:BDFl mice also showed an increased incidence of hepatocellular
carcinomas and adenomas when exposed to tetrachloroethylene by inhalation (Nagano et al.,
1998: JISA, 1993). Liver hemangiosarcomas were also increased in males. In the spleen, both
hemangiosarcomas and combined hemangiosarcomas and hemangiomas were increased in
males.
4.3.4. Mode of Action for Hemangiosarcomas or Hemangiomas in Mice
The incidence of hemangiomas or hemangiosarcomas occurring in the liver or spleen
(and to a lesser extent in fat, subcutaneous skin, and the heart) was significantly increased in
male Crj :BDF1 mice exposed to tetrachloroethylene by inhalation (JISA, 1993). This tumor type
is distinct from the hepatocellular adenomas and carcinomas induced by tetrachloroethylene in
male and female Crj:BDFl mice by inhalation exposure (JISA, 1993), and in male and female
B6C3Fi mice by inhalation (NTP, 1986) or oral (NCI, 1977) exposure. No data are available
concerning either the metabolites or the mechanisms that may contribute to the induction of
hemangiosarcomas or hemangiomas occurring in the liver or spleen in male mice. It is
concluded that the mechanisms or modes of action by which tetrachloroethylene induces this
type of tumor are not known.
4-154
-------
4.3.5. Mode of Action for Murine Hepatocellular Tumors
Multiple metabolites formed from tetrachloroethylene are toxic and carcinogenic in the
liver. In particular, it is likely that TCA and DCA contribute to tetrachloroethylene-induced liver
tumors in mice. However, the mode of action through which these (and potentially other)
metabolites elicit the benign and malignant hepatocellular tumors induced with oral or inhalation
exposure to tetrachloroethylene in multiple strains and both sexes of mice remains to be fully
elucidated. As noted by NRC (2010), it is likely that key events from several pathways,
comprising several simultaneous mechanisms, operate in tetrachloroethylene-induced liver
cancer.
The discussion of mechanistic effects addresses the following topics: (1) contribution of
tetrachloroethylene metabolism to hepatocarcinogenicity (refer to Section 4.3.5.1); (2)
genotoxicity (refer to Section 4.3.5.2); (3) epigenetic effects, focusing on DNA hypomethylation
(refer to Section 4.3.5.3); (4) oxidative stress (refer to Section 4.3.5.4); and (5) receptor
activation, focusing on a hypothesized PPARa-activation mode of action (refer to Section
4.3.5.5). Because it has been suggested that hepatocarcinogenesis caused through a PPARa-
activation MOA is not relevant to humans [e.g., Klaunig et al. (2003)], and such a conclusion
would have significant implications for hazard conclusions and dose-response analyses, this
hypothesized MOA is discussed in relatively more detail than other topics. In the NRC review
of EPA's 2008 external review draft of tetrachloroethylene, a dissenting opinion put forth by one
member was that "the weight of evidence strongly favors a key role of PPARa activation in
tetrachloroethylene-induced hepatocarcinogenesis in mice; furthermore, this MOA lacks
relevance for human hepatocarcinogenesis" [refer to Appendix B; NRC (2010)1. However, in
their rebuttal [also presented in Appendix B; NRC (2010)], the committee as a whole did not
support these conclusions. Overall, the committee judged that many gaps in knowledge remain
with regard to the MOA of tetrachloroethylene. They stated that the relevance of the peroxisome
proliferator MOA to tetrachloroethylene-induced mouse hepatic cancer and to
tetrachloroethylene-induced human hepatic cancer remains hypothetical and requires further
rigorous testing. Hence, they concluded that it is premature to draw conclusions on the relevance
of the PPARa MOA to tetrachloroethylene-induced human hepatic carcinogenesis (NRC, 2010).
They encouraged an in-depth presentation of the relevant issues and data, particularly with
respect to tetrachloroethylene studies. The discussion below, especially that in Section 4.3.5.4,
follows these recommendations.
4-155
-------
4.3.5.1. Contribution of Tetrachloroethylene Metabolites to Mode of Action and
Carcinogenicity
Several metabolites of tetrachloroethylene are carcinogenic in mice, and it is thought that
the hepatocarcinogenicity of the parent compound is mediated through the action of one or more
of its metabolites. Oxidative metabolism is thought to predominate in the liver, and TCA is the
major resultant urinary excretion product. As discussed in Section 3, TCA appears to be formed
from spontaneous decomposition of trichloroacetyl chloride, which is known to bind to
macromolecules. DCA may be formed from dechlorination of TCA, but DCA produced from
this pathway is likely to be rapidly metabolized in the liver and not detected in blood or urine.
DCA that has been detected in urine is thought to be the result of kidney-specific p-lyase
metabolism of the results of GSH conjugation of tetrachloroethylene, and DCA produced from
this pathway is presumed to not play a role in liver toxicity or cancer. The potential role of GST
conjugates of tetrachloroethylene in liver carcinogenicity, although unknown, is presumed to be
less important that the role of oxidative metabolites.
The focus of most hypotheses with respect to contributors to tetrachloroethylene
hepatocarcinogenicity has been on TCA and, to a lesser extent, DCA. Data on the
hepatocarcinogenicity of TCA and DCA in rodents, alone and in combination, are summarized in
Tables 4-17, 4-18, and 4-19. In mice, TCA significantly increased the incidence of liver tumors
in male and female B6C3Fi mice exposed via drinking water for 52-104 weeks (DeAngelo et
al.. 2008: Bull et al.. 2002: Pereira. 1996: Bulletal.. 1990: Herren-Freund et al.. 1987).
Incidence of tumors increased with increasing TCA concentrations (DeAngelo et al., 2008: Bull
et al., 2002: Bull et al., 1990). These results were obtained under conditions where the
background incidence of tumors in control animals was generally low. The development of
tumors in animals exposed to TCA progressed rapidly, as evidenced by significant numbers of
tumors in less-than-lifetime studies of 82 weeks or less. Positive evidence for tumor promotion
by TCA (following exposure to known tumor initiators) has been reported for liver tumors in
B6C3Fi mice (Pereira et al., 2001: Pereira et al., 1997) and for GGT-positive foci in livers of
partially hepatectomized Sprague-Dawley rats (Parnell etal., 1988). DCA also causes liver
cancer in mice (DeAngelo et al., 1999: Daniel et al., 1992: Bull et al., 1990: Herren-Freund et al.,
1987). DCA and TCA are also hepatocarcinogenic in mice when coadministered in the drinking
water for 52 weeks (Bull et al., 2002). Treatment-related liver tumors were observed in male
F344/N rats exposed via drinking water to DCA (DeAngelo et al., 1996) but not TCA (DeAngelo
et al., 1997) for 60 or 104 weeks. The carcinogenicity of TCA and DCA has not been evaluated
in female rats or in other species of experimental animals.
Data on tumor phenotype support the view that TCA may not be the sole tumorigenic
metabolite of tetrachloroethylene but also do not provide definitive evidence testing any
4-156
-------
particular hypothesis. For instance, liver tumor genotypes (e.g., with regard to H-ras codon 61
mutation) and phenotypes (e.g., with regard to c-Jun staining) appear to differ among tumors
induced by TCA, DC A, the combination of TCA and DC A, and the structurally related
compound trichloroethylene (Bull et al.. 2002). Bull et al. (2002) suggest that for
trichloroethylene, the data are not consistent with the hypothesis that TCA is the sole active
moiety, but a similar experiment has not been conducted for tetrachloroethylene. However, by
analogy, it is possible that TCA and DCA, in combination with each other (and with other
reactive intermediates produced during the oxidative metabolism of tetrachloroethylene) may
contribute to the production of liver tumors. This appears to be the case for noncancer effects, as
the spectrum of endpoints caused by tetrachloroethylene includes effects broader than that
produced by TCA, and including fatty degeneration, focal necrosis and regenerative repair, some
of which may play a role in liver carcinogenesis (discussed below).
The hepatocarcinogenic potencies of TCA and tetrachloroethylene have not been directly
compared in a single rodent bioassay. Appendix D presents a comparative quantitative analysis
of the carcinogenicity of TCA (including that predicted using PBPK modeling to be produced
from tetrachloroethylene) with the carcinogenicity of tetrachloroethylene. This analysis suggests
that TCA might explain the incidence of carcinomas observed in the available
tetrachloroethylene bioassays, but that a wide range of possible contributions cannot be ruled out
by the available data. Specifically, a contribution of TCA from as little as 12% up to 100%
cannot be ruled out, under the assumptions that the tetrachloroethylene NTP and JISA bioassay
data can be combined, and using the Chiu and Ginsberg (2011) PBPK model for
tetrachloroethylene and the Chiu (2011) PBPK model for TCA and TCA bioavailability. If
either of these assumptions is relaxed—i.e., given that residual uncertainties of about twofold
exist in the PBPK model predictions for TCA internal dose and that there may be some
underlying differences between the NTP and JISA bioassays—then the CIs will be greater.
Furthermore, the high control tumor incidence reported in the TCA bioassay of DeAngelo et al.
(2008) raises questions as to the representativeness of that bioassay for comparison with
tetrachloroethylene bioassays. Overall, as discussed in Chiu (2011) with regards to the
contribution of TCA to TCE-induced hepatomegaly, factors such as study-to-study experimental
variability in kinetics (e.g., metabolism, bioavailability) or in dynamics (e.g., background tumor
rates), different analytical methods used to quantify TCA in blood and tissues, and uncertainty in
TCA dosing patterns in drinking water studies further limit the ability to discern the quantitative
contribution of TCA. A more precise quantitative measure of the relative contribution of TCA to
tetrachloroethylene-induced liver tumors requires an appropriately designed experiment to better
control for these factors.
4-157
-------
Table 4-17. Hepatocarcinogenicity of TCA in rodent drinking water studies
Species (sex)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (F)
F344/N rats (M)
Exposure
0 and 5 g/L in drinking water for
61 wk
0, 1, and 2 g/L for 52 wk
0, 0.05, 0.5, or 5 g/L TCA for 60
wk
0, 0.5, and 2 g/L for 52 wk
0,0.35, 1.2, 3. 5 g/L for 51 wk
0,0.35, 1.2, 3. 5 g/L for 82 wk
0, 0.05, 0.5, 5 g/L for 104 wk
Results
Carcinomas: 0/22, 7/22
Carcinomas: 0/35, 2/11, 4/24
Carcinomas: 7, 4, 21, 38%
Carcinomas: 1/20, 11/20, 9/20
Carcinomas (52 wk): 0/40, 0/40,
0/19, 5/20
Carcinomas (81 wk): 2/90, 0/53,
5/27, 5/18
Carcinomas: 0, 0, 0, 0%
Authors
Herren-Freund et al.
(1987)
Bull et al. (19901
DeAngelo et al. (2008)
Bull et al. (20021
Pereira (1996)
DeAngelo et al. (1997)
Adapted from NRC (2006).
Table 4-18. Hepatocarcinogenicity of DCA in rodent drinking water studies
Species (sex)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (M)
B6C3FJ mice (F)
F344 rats (M)
Exposure
0 and 5 g/L for 61 wk
0 and 2 g/L for 52 wk
0, 0.05, 0.5, 4.5, and 5 g/L for
60-95 wk
0, 0.05 g/L for 60 wk
0,0.5, 1,2, 3.5 g/L for 100 wk
0,0.05 for 60 wk
0,0.28, 0.93, and 2.8 g/L for
52 wk
0,0.28, 0.93, and 2.8 g/L for
81 wk
0,0.05, 0.5, 2.4 g/L for 60 wk
0, 0.05, 0.5 g/L for 104 wk
Results
Carcinomas: 0/22, 21/26
Carcinomas: 0/35, 5/24
Carcinomas: 6.7-10, 22, 38, 98,
55%
Carcinomas (60 wk): 8/12, 25/30
Carcinomas (100 wk): 5/50, 5/24,
16/32, 6/14, 4/8
Carcinomas: 2/20, 15/24
Carcinomas (52 wk): 0/40, 0/40,
0/20, 1/20
Carcinomas (81 wk): 2/90, 0/50,
1/28, 5/19
Carcinomas (60 wk): 0/7, 0/7, 0/7,
1/27
Carcinomas (104 wk): 0/23, 0/26,
2/29
Authors
Herren-Freund et al.
(1987)
Bull et al. (19901
DeAngelo et al. (1991)
DeAngelo et al. (1999)
Daniel et al. (1992)
Pereira (1996)
DeAngelo et al. (1996)
Adapted from NRC (2006).
4-158
-------
Table 4-19. Incidence of mouse liver tumors with drinking water
administration of TCA and DCA, alone and in combination
Species (sex)
B6C3FJ mice (M)
Exposure (52 wk)
0 (drinking water vehicle)
0.5 g/L TCA
2 g/L TCA
0.1 g/L DCA
0.5 g/L DCA
2 g/L DCA
0.1 g/L DCA + 0.5 g/L TCA
0.5 g/L DCA + 0.5 g/L TCA
0.1 g/L DCA + 2 g/L TCA
0.5 g/L DCA + 2 g/L TCA
Liver tumor incidence
1/20
11/20
9/20
2/20
5/20
12/19
9/20
13/19
15/20
13/20
Authors
Bull et al. (20021
4.3.5.2. Genotoxicity
A hypothesized mutagenic MOA entails the following key events leading to
tetrachloroethylene-induced liver tumor formation: following metabolism of tetrachloroethylene
to one or more mutagenic intermediates, the genetic material is altered in a manner that permits
changes to be transmitted during cell division through one or more mechanisms (gene mutations,
deletions, translocations, or amplification); the resulting mutations advance acquisition of the
multiple critical traits contributing to carcinogenesis. This MOA may apply to multiple cancer
types.
The genotoxic potential of tetrachloroethylene is addressed in Section 4.8. To
summarize, the results of a large number of in vitro genotoxicity tests in which
tetrachloroethylene was the test agent support the conclusion that tetrachloroethylene does not
exhibit direct mutagenic activity in the absence or presence of the standard S9 fraction
(Watanabe et al.. 1998: DeMarini et al.. 1994: Roldan-Ariona et aL 1991: Milman et al.. 1988:
Warner etal.. 1988: NTP. 1986: Connor et al.. 1985: Shimada et al.. 1985: Haworth et al.. 1983:
Hardin et al.. 1981: Kringstad et al.. 1981: Bartsch et al.. 1979: Greim et al.. 1975). However,
the few in vitro mutagenicity studies of tetrachloroethylene under conditions that would generate
the GSH conjugate were positive (Vamvakas et al., 1989c: Vamvakas et al., 1989d). Several
other known (DCA) and putative (tetrachloroethylene oxide) P450 metabolites also exhibit in
vitro mutagenicity. Studies of chromosomal aberrations following exposure to
tetrachloroethylene are mostly negative, but positive results have been reported from in vitro
studies with enhanced metabolic activation (Doherty et al., 1996).
4-159
-------
TCA, the primary oxidative metabolite of tetrachloroethylene, exhibits little, if any,
genotoxic activity in vitro. TCA did not induce mutations in S. typhimurium strains in the
absence of metabolic activation or in an alternative protocol using a closed system (Kargalioglu
et al.. 2002: Nelson etal.. 2001b: Gilleretal.. 1997: DeMarini et al.. 1994: Rapsonetal.. 1980:
Waskell, 1978), but a mutagenic response was induced in TA100 in the Ames fluctuation test
(Giller et al., 1997). However, in vitro experiments with TCA should be interpreted with caution
if steps have not been taken to neutralize pH changes caused by the compound (Mackay et al.,
1995). Measures of DNA-repair responses in bacterial systems have shown induction of DNA
repair reported in S. typhimurium but not in E. coli. Mutagenicity in mouse lymphoma cells was
only induced at cytotoxic concentrations (Harrington-Brock et al., 1998). TCA was positive in
some genotoxicity studies in vivo, in mouse, newt, and chick test systems (Giller et al., 1997:
Bhunya and Jena, 1996: Birner et al., 1994: Bhunya and Behera, 1987). DNA unwinding assays
have either shown TCA to be much less potent than DCA (Nelson and Bull, 1988) or negative
(Styles etal., 1991: Nelson etal., 1989). Due to limitations in the genotoxicity database, the
possible contribution of TCA to tetrachloroethylene genotoxicity is unclear.
The limited in vivo studies of tetrachloroethylene are inconsistent, with only negative
(NTP. 1986: Bronzetti et al., 1983) or equivocal (Cederberget al.. 2010a: Bellies et al.. 1980)
genotoxicity assay results demonstrated following inhalation or oral exposure. These include
findings that tetrachloroethylene at higher concentrations induces, at most, modest increases in
DNA damage and DNA binding in liver tissue (Cederberg et al., 2010a: Murakami and
Horikawa, 1995). Intraperitoneal injection assays have demonstrated both negative (NTP, 1986)
as well as positive results for different genotoxicity endpoints (Walles, 1986). Assays of
clastogenic effects following inhalation exposure in humans have shown inconsistent results and
are suggested to be related to coexposures (Seiji et al., 1990: Ikeda et al., 1980).
Thus, although tetrachloroethylene has largely yielded negative results in standard
genotoxicity assays, uncertainties remain with respect to the possibility that genotoxicity
contributes to hepatocarcinogenesis. Not all metabolites have been identified or characterized,
but several known metabolites including those derived from P450 as well as GSH pathways are
clearly mutagenic in the standard battery of tests. Tetrachloroethylene is mutagenic in bacterial
assays in the presence of GST and GSH, whereas the standard S9 fraction has typically yielded
negative results. Tetrachloroethylene at higher concentrations also induces modest increases in
DNA damage and DNA binding in liver tissue (Cederberg et al., 2010a: Murakami and
Horikawa, 1995). The metabolite DCA is the most potent mutagen of the P450-derived
metabolites, exhibiting mutagenic activity in a number of assays. A putative P450 derived
metabolite, 1,1,2,2-tetrachloroethylene oxide, is also mutagenic; the mutagenicity of this epoxide
would be predicted from structure-activity relationships. Given the demonstrated mutagenicity
4-160
-------
of several tetrachloroethylene metabolites, the hypothesis that mutagenicity contributes to the
MOA for tetrachloroethylene carcinogenesis cannot be ruled out, although the specific metabolic
species or mechanistic effects are not known.
4.3.5.3. Altered DNA Methylation
Another hypothesis is that tetrachloroethylene induces hepatocarcinogenesis via the
induction of epigenetic changes, particularly DNA methylation. This MOA entails the following
key events leading to tetrachloroethylene-induced liver tumor formation: following metabolism
of tetrachloroethylene to one or more reactive intermediates, particularly TCA, DC A, and other
reactive species, epigenetic changes ensue; the resulting alterations advance acquisition of the
multiple critical traits contributing to carcinogenesis. This MOA may apply to multiple cancer
types.
No tetrachloroethylene-specific data are available regarding a role of alteration in DNA
methylation in tumorigenesis. However, experimental evidence supports the hypothesis that
hypomethylation of DNA may be related to the carcinogenicity of TCA and DCA in mice. In
female B6C3Fi mice that received an i.p. injection of 7V-methyl-7V-nitrosourea (MNU) and were
then administered TCA or DCA in drinking water, DNA methylation in the resulting
hepatocellular adenomas and carcinomas was about half that observed in noninvolved tissue
from the same animal or from animals given only MNU (Tao et al., 1998). Drinking water
exposure of female B6C3Fi mice to TCA or DCA for 11 days also decreased total liver DNA
methylation by 60% (Tao et al., 1998). The same investigators (Tao et al., 2004) also
demonstrated hypomethylation of a region of the IGF-II gene in liver and tumors from mice
initiated with MNU and subsequently exposed to TCA or DCA. An association between
hypomethylation and cell proliferation in liver of TCA- or DCA-exposed mice was demonstrated
by Ge et al. (2001). An increase in DNA replication (evidenced by increased proliferating cell
nuclear antigen labeling index and mitotic labeling index) was observed 72 hours and 96 hours
after the first daily gavage dose of either TCA or DCA. Hypomethylation of the internal
cytosine of CCGG sites in the promoter region of the c-myc gene began between 48 and 72 hours
from the initiation of TCA or DCA exposure and continued to 96 hours. These observed effects
of TCA and DCA, together with the fact that methylation changes represent common early
molecular events in most tumors (Bavlin etal., 1998; Zingg and Jones, 1997), support the
plausibility of a hypothesis that dysregulation of gene methylation plays a role in
tetrachloroethylene-induced tumorigenesis. However, no data are available specifically testing
this hypothesis for tetrachloroethylene.
4-161
-------
4.3.5.4. Cytotoxicity and Secondary Oxidative Stress
Another hypothesis is that oxidative stress produced secondary to tetrachloroethylene-
induced cytotoxicity plays a critical role in hepatocarcinogenesis. This MOA entails the
following key events leading to tetrachloroethylene-induced liver tumor formation: following
metabolism of tetrachloroethylene to one or more reactive intermediates, toxicity to the liver
ensues; oxidative stress is produced during hepatocyte injury, from infiltrating inflammatory
cells, and/or as part of the intracellular/extracellular repair processes; the resultant oxidative
stress, via a variety of potential mechanisms (damage to and alteration of macromolecules, cell
signaling alterations, etc.), advances acquisition of the multiple critical traits contributing to
carcinogenesis. This MOA may apply to multiple cancer types.
Numerous studies, including chronic bioassays, have demonstrated that
tetrachloroethylene is hepatotoxic. Reported characteristics of the hepatic injury induced by
tetrachloroethylene and the ensuing tissue repair include increased liver weight, fatty changes,
necrosis, inflammatory cell infiltration, triglyceride increases, and proliferation. The NTP
chronic bioassay reported a variety of histological changes, including cytoplasmic vacuolation,
hepatocellular necrosis, inflammatory cell infiltrates, pigment in cells, oval cell hyperplasia, and
regenerative foci. Liver tissue repair is a complex process involving cell division, angiogenesis,
ductulogenesis, cell mobility, and extracellular matrix repair, all in a coordinated manner
(Mehendale, 2005). Reactive oxygen species can play a role in mediating many of these
processes and are produced during hepatocyte injury, from infiltrating inflammatory cells, and/or
as part of the intracellular/extracellular repair processes.
A limited database of studies is available on tetrachloroethylene-induced hepatic
oxidative stress. Two studies by Ebrahim et al. (2001; 1996) have examined the ability of
2-deoxy-glucose (2DG), vitamin E, or taurine to modulate hepatic effects following short-term
exposure. Ebrahim (1996) orally administered 3,000 mg/kg-day tetrachloroethylene in sesame
oil to male and female Swiss mice for 15 days and observed a significant increase in liver weight
and degeneration and necrosis of hepatocytes. These changes occurred simultaneously with a
decrease in blood glucose; elevated activities of enzymes hexokinase, aldolase, and
phosphoglucoisomerase; and decreased activities of gluconeogenic enzymes. Blood glucose
levels were significantly decreased, and this effect was mitigated by concomitant exposure to
2-deoxy-D-glucose and vitamin E.
In a follow-up study, Ebrahim et al. (2001) further examined the potential protective
properties of 2DG and vitamin E as well as taurine against membrane damage induced with a
similar exposure paradigm. This study exposed male albino Swiss mice to the same doses used
in the previous study with the addition of a taurine-exposed group (tetrachloroethylene in sesame
oil 3,000 mg/kg-day for 15 days by oral gavage; tetrachloroethylene plus 2DG 500 mg/kg-day by
4-162
-------
i.p. injection once a day for 15 days; tetrachloroethylene plus vitamin E 400 mg/kg-day by oral
gavage once a day for 15 days; and tetrachloroethylene plus taurine 100 mg/kg-day by oral
gavage once a day for 15 days). Compared to control cells in the liver, membrane-bound
Na+K+-ATPases and Mg2+-ATPases activity was significantly decreased (p < 0.001), while
Ca-ATPases activity was increased (p < 0.001), following exposure to tetrachloroethylene alone.
These levels remained near normal in the animals exposed to tetrachloroethylene along with
2DG, vitamin E, or taurine. This return to normal levels following exposure to vitamin E and
taurine may be due to their antioxidant abilities, and reduced oxidative stress in exposed cells.
A recent in vitro investigation examined tetrachloroethylene-induced gene expression
changes in the HepG2 cultured human hepatoma cell line using an Affymetrix platform (Kawata
et al., 2009). HepG2 cells retain Phase 1 and Phase 2 metabolic enzymes. Tetrachloroethylene
(2 mM) altered the expression of 445 genes, of which, 367 were annotated in Gene Ontology
terms to represent 261 biologic processes. The major processes included cell death, regulation of
metabolic processes, phosphorylation, lipid biosynthesis, steroid metabolism, intracellular
transport, DNA repair, and regulation of cell cycle. Based on KEGG pathway mapping, "cell
cycle" and "MAPK signaling" pathways were prominent; a similar finding was reported for
other chemicals (dimethyl nitrosamine and the phorbol ester 12-O-tetradecanoylphorbol-
13-acetate) and metals (nickel, cadmium, and arsenic). The authors noted that this pathway has
been shown to be activated by reactive oxygen species and metals in earlier studies (Guyton et
al., 1996; Liu et al., 1996) and demonstrated that metal-induced gene changes associated with
this pathway could be inhibited by vitamin C. Upregulation of the oncogene PTT1G was noted
in all exposures. This hypothesis-generating in vitro experiment may aid in elucidating
molecular pathway-based biomarkers of tetrachloroethylene.
4.3.5.5. Peroxisome Proliferator-Activated Receptor (PPAR) Activation Mode of Action
4.3.5.5.1. Description of hypothesized MOA
Another hypothesis is that tetrachloroethylene acts by a PPARa-agonism MOA in
inducing mouse hepatocarcinogenesis. According to this hypothesis, the key events leading to
tetrachloroethylene-induced liver tumor formation constitute the following: tetrachloroethylene
metabolites (primarily the oxidative metabolite, TCA), after being produced in the liver, activate
the PPARa receptor, which then causes alterations in cell proliferation and apoptosis, followed
by clonal expansion of initiated cells. This MOA is assumed to apply only to the liver. This
corresponds to the widely cited version of the hypothesized MOA for hepatocarcinogenesis
induced by PPARa agonists posited by Klaunig et al. (2003), in which three key causal events
were proposed: (1) activation of the receptor, (2) perturbation of hepatocellular apoptosis and
proliferation, and (3) selective clonal expansion. A number of intermediary events were
4-163
-------
considered associative (e.g., expression of peroxisomal and nonperoxisomal genes, peroxisome
proliferation, inhibition of gap junction intracellular communication, hepatocyte oxidative stress
and Kupffer cell-mediated events). The data requirements suggested by Klaunig et al. (2003) for
demonstrating that the PPARa-activation MOA is operative did not comprise all purportedly
causal events; instead, these requirements included PPARa-agonism combined with microscopic
evidence for peroxisome proliferation (or, in lieu of evidence of peroxisome proliferation,
increased liver weight together with in vivo markers such as increases in peroxisomal
p-oxidation, CYP4A, or acyl CoA oxidase). Alterations in proliferation and apoptosis were
considered corroborative evidence.
4.3.5.5.2. Induction of hypothesized key events by tetrachloroethylene and metabolites
4.3.5.5.2.1. Activation of PPARa and associated markers
As summarized in Table 4-20, several in vivo studies have examined the effect of
tetrachloroethylene on peroxisome proliferation or its markers (Philip et al., 2007; Odum et al.,
1988; Goldsworthy and Popp, 1987). Odum et al. (1988) exposed groups of male and female
F344 rats and B6C3Fi mice by inhalation for 6 hours/day to 200 ppm (28 days only) or 400 ppm
(for 14, 21, or 28 days) tetrachloroethylene. Five animals per group were exposed. In both
sexes, hepatic PCO activity was increased in mice (up to 3.6-fold) and, to a lesser extent, in rats
(up to 1.3-fold). Modest PCO increases were also observed in the kidney of male rats at 200
ppm at 28 days (1.3-fold) but not 400 ppm at 14, 21, or 28 days. In female rat kidney, PCO was
elevated (approximately 1.6-fold) at all doses and times. However, peroxisome proliferation was
not observed in rat kidney upon microscopy. In contrast, hepatic peroxisome proliferation was
noted in all exposed mice on electron microscopy, and the percentage of cytoplasm occupied by
peroxisomes also increased in mice. In rats, variable increases in peroxisome volume were noted
at 200 ppm, but results lacked statistical significance. Catalase, another peroxisomal enzyme,
was unaffected by tetrachloroethylene; male mice exposed at 400 ppm showed the only moderate
(1.4-fold) increase. Mitochondrial proliferation was observed at 28 days in 400 ppm male mice.
In addition, a time-dependent proliferation of smooth endoplasmic reticulum in the liver of both
sexes correlated well with centrilobular hypertrophy. Tetrachloroethylene caused centrilobular
lipid accumulation in male and female mice. Relative liver weight was increased in mice of both
sexes.
Goldsworthy and Popp (1987) administered tetrachloroethylene (1,000 mg/kg-day) by
corn oil gavage to 5 male F344 rats and 5 male B6C3Fi mice for 10 days. In
tetrachloroethylene-exposed rats, PCO was modestly—although not significantly—elevated in
the liver (1.4-fold increase) and kidney (1.7-fold increase). In mice, tetrachloroethylene
exposure increased PCO activity 4.3-fold in liver and by 2.3-fold in kidney. Relative liver
4-164
-------
weight was increased in rats and mice with tetrachloroethylene exposure, but relative kidney
weight was unaffected. A comparison of corn oil with methyl cellulose revealed no effect of the
gavage vehicle on tetrachloroethylene-induced PCO. Administration of trichloroethylene (1,000
mg/kg) together with tetrachloroethylene had a less-than-additive effect on PCO induction.
Table 4-20. Rodent studies of induction of peroxisome proliferation or its
markers by tetrachloroethylene
Species/strain/sex/number
Rat, F344; and mouse,
B6C3FJ; both sexes
(5/group)
Odum et al. (1988)
Rat, F344 (male only,
5/group) and mouse,
B6C3FJ (male only,
5/group)
Goldsworthy and Popp
(1987)
Mouse, Swiss-Webster,
male (4/group)
Philip et al. (2007)
Effect
Mice of both sexes: increased relative liver
weight, centrilobular lipid accumulation and
peroxisome proliferation; increased PCO (up
to 3.7-fold)
Male mice: mitochondria! proliferation
Rats of both sexes: increased PCO (up to
1.3 -fold)
Mice: Increased relative liver weight; 4. 3 -fold
PCO increase
Rats: Increased relative liver weight; modest
but not significant (1.4-fold) PCO increase
Increased plasma ALT
Mild-to-moderate fatty degeneration and
necrosis, with focal inflammatory cell
infiltration
Increased mitotic figures and DNA synthesis
CYP4A increased at 7 but not 14 d, only at
1,000 mg/kg-day
Dose
200 and 400 ppm,
inhalation
400 ppm, inhalation
200 and 400 ppm,
inhalation
1,000 mg/kg-day for
10 d, corn oil gavage
1,000 mg/kg-day for
10 d, corn oil gavage
150, 500, and
1,000 mg/kg-day,
aqueous gavage
150, 500, and
1,000 mg/kg-day,
aqueous gavage
150, 500, and
1,000 mg/kg-day,
aqueous gavage
1,000 mg/kg-day,
aqueous gavage
Time
14, 21, 28 d
28 d
14, 21, 28 d
10 d
10 d
24 hours to 14 d
after initial
exposure
24 hours to 30 d
after initial
exposure
Peaked on 7 d,
sustained at
14-30 d
7 but not 14 d
The peroxisome-related effects of tetrachloroethylene are most likely mediated primarily
through TCA based on tetrachloroethylene metabolism producing more TCA than DC A, and the
lower doses of TCA required to elicit a response relative to DCA. Bull (2004) and Bull et al.
(2004) have recently suggested that peroxisome proliferation occurs at higher exposure levels
than those that induce liver tumors for TCA and DCA. They report that a direct comparison of
the no-effect level or low-effect level for induction of liver tumors in the mouse and several other
endpoints shows that, for TCA, liver tumors occur at lower concentrations than peroxisome
4-165
-------
proliferation in vivo but that PPARa-activation occurs at a lower dose than either tumor
formation or peroxisome proliferation. A similar comparison for DCA shows that liver tumor
formation occurs at a much lower exposure level than peroxisome proliferation or PPARa-
activation. In vitro transactivation studies have shown that human and murine versions of
PPARa are activated by TCA and DCA, while tetrachloroethylene itself is relatively inactive in
the in vitro system, at least with mouse PPARa (Maloney and Waxman, 1999; Zhou and
Waxman, 1998). In addition, Laughter et al. (2004) reported that the responses of AGO, PCO,
and CYP4A induction by TCA and DCA were substantially diminished in PPARa null mice.
Therefore, evidence suggests that tetrachloroethylene activates PPARa in vivo, and that the role
of TCA in activating PPARa is likely to predominate at doses relevant to tetrachloroethylene-
induced hepatocarcinogenesis.
4.3.5.5.2.2. Alterations of cell proliferation and apoptosis and clonal expansion of initiated
cells
As discussed above, increased cell proliferation in mice has been reported following
exposure to tetrachloroethylene. However, few data are available to inform the hypothesis that
activation of PPARa after tetrachloroethylene exposure causes alterations in cell proliferation
and apoptosis, followed by clonal expansion of initiated cells. Moreover, available data suggest
that PPARa-activation may not be the predominant cause of the observed cell proliferative
response. For example, transient increases in DNA synthesis and PCNA staining in the liver
were reported by Philip et al. (2007), similar to that observed with other PPARa agonists (with
the exception of WY-14,643, which induces sustained proliferation) (refer to Section
4.3.5.2.4.2). However, Philip et al. (2007) suggest that PPARa-activation is not required for the
observed cell proliferative response, and rather that this is a regenerative response following
cytotoxicity. This is based on evidence of significantly increased CYP4A expression at only the
highest dose (1,000 mg/kg-day) and at the earliest time point (7 days), in contrast to the robust
dose-dependent proliferative response of a more prolonged nature (lasting for 14-30 days post
exposure) observed at the same and lower (150, 500 and 1,000 mg/kg-day) levels of
tetrachloroethylene. The authors concluded that their findings suggest peroxisome proliferation
is not a sustained response in spite of continued tetrachloroethylene exposure and, therefore, are
not supportive of a close mechanistic relationship of carcinogenicity and PPARa induction for
tetrachloroethylene-derived TCA. This interpretation is limited by the possible lack of
sensitivity of CYP4A protein expression as a marker of peroxisome proliferation, and the lack of
other supporting data for the observed absence of sustained peroxisome proliferation in the
context of a robust regenerative proliferative response. Additionally, the sensitivity of the SW
mouse to tetrachloroethylene hepatocarcinogenicity is unknown, somewhat limiting the
significance of these findings for the interpretation of hepatocellular tumor findings in other
4-166
-------
mouse strains. However, other studies of the toxicity of tetrachloroethylene in the B6C3Fi strain
discussed above [e.g., Schumann et al. (1980)] have reported liver toxicity and repair at
100 mg/kg-day, whereas Odum et al. (1988) reported only modest increases in peroxisomal
markers in B6C3Fi mice with repeated exposures to 1,000 mg/kg-day. Another noteworthy
finding in Odum et al. (1988) was the modest increase in peroxisome proliferation observed in
rats.
Data on TCA are also informative of the extent to which tetrachloroethylene alters cell
proliferation and apoptosis through PPARa-activation, as it was concluded above that the
PPARa-agonism following tetrachloroethylene is mostly likely caused by its metabolism to
TCA. Data that inform the hypothesis that activation of PPARa after TCA exposure causes
alterations in cell proliferation and apoptosis, followed by clonal expansion of initiated cells, are
discussed in the EPA Toxicological Review of Trichloroacetic Acid (20lie). To summarize,
several studies have observed hepatocyte proliferation in response to TCA in mice (DeAngelo et
al.. 2008: Stauber and Bull 1997: Pereira. 1996: Dees and Travis. 1994: Sanchez and Bull
1990). For instance, Dees and Travis (1994) observed relatively small (two- to threefold)—but
statistically significant—increases in [3H]thymidine incorporation in hepatic DNA in mice
exposed for 11 days at TCA doses (100-1,000 mg/kg) that increased relative liver weight.
Increased hepatic DNA labeling was observed at doses lower than those associated with
evidence of necrosis, suggesting that TCA-induced cell proliferation is not due to regenerative
hyperplasia. PPARa-null mice exposed to 2 g/L TCA in drinking water for 7 days do not show
the characteristic responses of AGO, PCO, and CYP4A induction associated with PPARa-
activation and peroxisome proliferation in wild-type mice (Laughter et al., 2004). In addition,
the livers from wild-type—but not PPARa-null—mice exposed to TCA developed centrilobular
hepatocyte hypertrophy, although no significant increase in relative liver weight was observed.
Therefore, while there are data associating TCA exposure, PPARa-activation, and cell
proliferation, it is not clear the extent to which PPARa-activation is the cause of the observed
cell proliferation.
Data informing the hypothesis that PPARa-activation following tetrachloroethylene
exposure causes clonal expansion of initiated cells, are limited to studies of its metabolite TCA.
Mechanistic studies reveal that the mode of action for TCA hepatocarcinogenesis is complex and
that TCA may induce tumors by multiple modes of action that may not be mutually exclusive
(U.S. EPA, 201 Ic). In particular, tumor induction by TCA appears to involve perturbation of
cell growth, reduced intercellular communication (Benane etal., 1996), release of cytokines and
oxidants by activated Kupffer cells, and hypomethylation of DNA.
4-167
-------
4.3.5.5.2.3. Conclusions regarding induction of hypothesized key events by
tetrachloroethylene and metabolites
The available evidence from tetrachloroethylene and its metabolites supports the
conclusion that tetrachloroethylene exposure leads to PPARa-activation predominantly through
its metabolite TCA. There is more limited evidence supporting the hypothesis that PPARa-
activation is the cause of the cell proliferative responses observed, and some evidence suggesting
that PPARa-activation is not the cause of these responses. Data informing the hypothesis that
PPARa-activation following tetrachloroethylene exposure causes clonal expansion of initiated
cells are even more limited.
4.3.5.5.3. Are activation of PPAR a and its sequelae key events in tetrachloroethylene-
induced hepatocarcinogenesis?
No tetrachloroethylene-specific data have directly tested the hypothesis that
tetrachloroethylene-induced PPARa-activation, along with its sequelae, are key or causative
events in tetrachloroethylene-induced hepatocarcinogenesis (e.g., bioassays with knockout mice
or involving the blocking of hypothesized key events). With respect to more associative data,
Philip et al. (2007) found increases in CYP4A, a marker for PPARa-activation, to be transient
(only increased at 7 days) rather than sustained, and only occurring at the highest dose (1,000
mg/kg-day). These data are not supportive of PPARa-activation as a key event in
tetrachloroethylene-induced hepatocarcinogenesis for two reasons: (1) chronic activation would
be needed to sustain changes in cell proliferation, apoptosis, and clonal expansion, and
(2) statistically significant increases in liver tumors have been reported at doses around 500
mg/kg-day (NCI, 1977), at which no increased CYP4A activity was reported. However, the SW
strain of mouse used by Philip et al. (2007) may differ in tumor responsiveness from those used
in the cancer bioassays discussed above.
Support for this MOA is based primarily on the hypothesis that TCA induces tumors
through PPARa-activation, and the fact that TCA is formed after in vivo exposure to
tetrachloroethylene. The experimental evidence related to the hypothesis that TCA induces
tumors through PPARa-activation is discussed extensively in the EPA ToxicologicalReview of
TCA (U.S. EPA, 201 Ic). TCA activates PPARa, and induces peroxisome proliferation and
hepatocyte proliferation. However, a number of inconsistencies and data gaps reduce the
confidence in the conclusion that TCA induces hepatocarcinogenesis solely through a PPARa-
activation MOA. First, while TCA induces peroxisome proliferation (a marker for PPARa-
agonism) in both rats and mice, to date, TCA has been shown to be tumorigenic in B6C3Fi mice
but not F344 rats (DeAngelo et al., 1997) (the only strains tested for carcinogenicity). In
addition, the tumor phenotype of TCA-induced mouse liver tumors has been reported to have a
different pattern of H-ras mutation frequency from DCA and other peroxisome proliferators
4-168
-------
(Bull et al.. 2002: Stanley et al.. 1994: HegietaL 1993: Foxetal.. 1990). Other effects of TCA,
including increased c-myc expression and hypomethylation of DNA, are not specific to the
PPARa-activation MO A, and other data (discussed below in Section 4.3.4.2.4) also contribute
uncertainty as to whether PPARa independent mechanisms may be involved in TCA-induced
tumors in mice.
To summarize, based on data from tetrachloroethylene and its metabolites alone, there is
only limited evidence that activation of PPARa and its sequelae are key events in
tetrachloroethylene-induced hepatocarcinogenesis. In all, the modest peroxisome proliferation
observed in response to tetrachloroethylene may lack specificity and consistency with respect to
tissue, species, and dose, and studies of the temporal sequence of events are limited. Given the
limitations in the database of tetrachloroethylene-specific studies, it can be concluded that the
few studies demonstrating activation of PPARa and related markers by tetrachloroethylene are
insufficient to demonstrate a causative role of this effect in the induction of other key events
posited for the PPARa mode-of-action hypothesis, and for hepatocarcinogenesis by
tetrachl oroethy 1 ene.
4.3.5.5.4. Other experimental evidence for the hypothesized MOA
4.3.5.5.4.1. Evidence from PPARa-null mouse bioassays
An apparent reduction was observed in tumor response to an 11-month exposure to the
prototypical agonist 4-chloro-6-(2,3-xylidino)-2-pyrimidyl-thio]acetic acid (Wy-14,643) in
PPARa-null mice in comparison to wild-type mice (Peters et al., 1997). Peters et al. reported the
absence of tumors in nine PPARa-null mice exposed to Wy-14,643 at 11 months, whereas each
of the six similarly exposed wild-type mice had multiple hepatocellular neoplasms.
As has also has been shown for Wy-14,643, the monoester metabolite
(mono-2-ethylhexylphthalate, MEHP) of DEHP activates PPARa in vitro (Maloney and
Waxman, 1999: Issemann and Green, 1990). Other evidence for DEHP includes induction of
peroxisome proliferation (or an increase in peroxisomal enzyme activity), an associative event in
the MOA, by tumorigenic doses of DEHP in the liver of mice and rats and of MEHP in rat
hepatocytes (David et al.. 1999: Hasmall et al.. 1999: Reddvetal.. 1986: Mitchell et al.. 1985:
Mitchell et al.. 1984: Gravetal.. 1983: Gravetal.. 1982). Additionally, an absence of
peroxisomal enzyme induction and peroxisome proliferation in PPARa-null mice exposed to
DEHP for 24 weeks was demonstrated (Wardetal.. 1998).
However, as reviewed recently by Guyton et al. (2009), a 2-year bioassay found that
DEHP (100 or 500 ppm) induces liver tumors in PPARa-null mice (Ito et al., 2007a). Ito et al.
(2007a) reported a significant trend for the observed increase in total liver tumors with DEHP in
PPARa-null male mice with Sv/129 genetic background generated as described in Lee et al.
4-169
-------
(1995). Guyton et al. (2009) performed additional statistical analyses to compare the Ito et al.
(2007a) results with those of a prior DEHP bioassay in B6C3Fi wild-type mice (David et al.,
1999). A pair-wise analysis found that DEHP (500 ppm) significantly increased adenomas in
PPARa-null—but not in companion wild-type—mice compared to their respective controls
(refer to Figure 4-3, single asterisks). In the David et al. study of B6C3Fi mice, DEHP (500
ppm) also significantly increased adenomas and adenomas plus carcinomas (refer to Figure
4-3B, single asterisks). Moreover, a significant dose-response trend for adenomas and for
adenomas plus carcinomas was observed in both the Ito et al. (2007a) PPARa-null mice and the
David et al. B6C3Fi mice after exposure to DEHP (refer to Figure 4-3B, double asterisks).
Additionally, Guyton et al. (2009) found no statistically significant differences between groups at
the same dose, including controls, consistent with mouse strain and PPARa genotype having no
influence on carcinogenicity under the study conditions.
The observed lack of difference in reported control incidences across groups lends
support to the approach of basing comparative analyses on concurrent controls. Historical data
on spontaneous liver tumor incidences in PPARa-null mice are limited; Ito et al. (2007a) is the
largest published 2-year bioassay in PPARa-null mice, reporting findings for 24/25 surviving
unexposed animals at 23 months of age. A different laboratory that had established a distinct
breeding colony reported mouse liver tumor incidences in 12 PPARa-null Svl29/C57BL/6 mice
~2 years of age (Howroyd et al., 2004). Adenomas and carcinomas were reported in 6/12 and
2/12 PPARa-null mice, respectively, compared with adenomas in 5/22 wild-type animals. As
Howroyd et al. note, "The relatively small number of animals available made it difficult to draw
robust conclusions concerning enhancement of spontaneous findings in PPARa-null mice." In
addition, cross-laboratory differences [particularly the low survival of PPARa-null mice in the
Howroyd et al. relative to the Ito et al. (2007a) study] limit statistical comparisons based on this
data set.
In summary, the Ito et al. (2007a) study indicates that DEHP carcinogenesis can occur
independently of PPARa-activation. As noted in a recent National Research Council report on
risk assessment (NRC, 2008), this finding "calls into question" the 2000 IARC conclusions
regarding the carcinogenic risks of DEHP (IARC, 2000). The 2011 IARC Working Group
evaluating DEHP also concluded that "the human relevance of the molecular events leading to
DEHP-induced cancer in several target tissues (e.g., liver and testis) in rats or mice could not be
ruled out, resulting in the evaluation of DEHP as a Group 2B agent, rather than Group 3"
(Grosse et al., 2011). Although new hypotheses are being generated based on more detailed
comparisons between wild-type and PPARa-null mice (Eveillard et al., 2009; Takashima et al.,
2008; Ito et al., 2007a), the available data indicate that the mechanisms of cancer induction by
DEHP are complex.
4-170
-------
4.3.5.5.4.2. Quantitative analyses of hypothesized key events and carcinogenic potency
If potency for PPARa-activation or its attendant sequelae is quantitatively associated with
carcinogenic activity or potency, then it might be possible to predict differences in sensitivity for
carcinogenesis (such as may occur across species) for environmental contaminants that activate
PPARa (e.g., certain phthalates and chloroacetic acids) using quantitative information about the
key events alone. It is, thus, of interest to assess whether potency for inducing these events is
quantitatively related to hepatocarcinogenic potential by these and other compounds that also
activate PPARa. However, there are limitations in the dose-response data available for such
analyses, specifically for precursor events in the proposed PPARa-activation MOA as well as for
liver tumor induction. Most tumor data, including for the best characterized PPARa agonists, are
for exposure concentrations inducing well above 50% tumor incidence, with less-than-lifetime
administration. Precursor events have typically been studied at a single dose, often eliciting a
near maximal response, thus, precluding benchmark-based comparisons across studies. This is
especially true for Wy-14,643, which has been administered most often at only one exposure
concentration (1,000 ppm) that elicits a 100% tumor incidence after 1 year or less (Peters et al.,
1997) and that also appears to be necrogenic (Woods et al., 2007). On the other hand,
hypothesized precursor events such as hepatomegaly, peroxisome proliferation, and increased
DNA synthesis appear to have reached their maximal responses at 50 ppm Wy-14,643, with
some statistically significant responses as low as 5 ppm (Marsman et al., 1992; Wada et al.,
1992). Potencies across compounds have rarely been compared in a single study using the same
experimental paradigm. These deficits in the database notwithstanding, provided below is an
assessment of the quantitative predictive power of the potency for four proposed data elements
for establishing the hypothesized MOA for hepatocarcinogenesis: (1) PPARa-activation in mice;
and (2) hepatomegaly, (3) DNA synthesis, and (4) increased peroxisome proliferation in rats.
4-171
-------
-/-(Itoetal.)
+/+(ltoetal.)
Adenomas
+/+ (David
o d
O
p
ci
0) ^
O d
Is
o
A
i
Adenomas+Carcinomas
o _
q
ci
B
0 100 500 0 100 500 0
ppm DEHP
100 500
Figure 4-3. Incidences of hepatocellular adenomas (A) and hepatocellular
adenomas and carcinomas (B) in mice exposed to DEHP.
Ito et al. (2007a) exposed PPARa null [-/-] and wild-type [+/+] Sv/129 mice for 22 months; David
et al. (1999) exposed B6C3F1 wild-type [+/+] mice for up to 104 weeks. Data are presented as
incidence +/- SD assuming a binomial distribution for each group. Single asterisks (*) indicate a
significant difference from controls of the same genotype in the same study (Fisher's exact test,
p < 0.05). Double asterisks (**) indicate a significant trend with dose in the study (Cochran
Armitage test, p < 0.05). All pair-wise cross-study comparisons between like dose groups (e.g.,
Ito et al. [-/-] 500 ppm compared with David et al. [+/+] 500 ppm) were not significant (Fisher
exact test;? > 0.05). Because David et al. (1999) reported only adenomas and carcinomas, the
cholangiocellular carcinoma reported by Ito et al. (2007a) in DEHP-exposed PPARa null mice
was excluded from analyses. Adapted from Guyton et al. (2009).
4-172
-------
4.3.5.5.4.2.1 PPARa-activation in mice
Table 4-21 presents data for four peroxisome proliferators in order of decreasing potency
for inducing mouse liver tumors. These compounds were selected because of their importance to
environmental human health risk assessments and because data to derive receptor activation
potency indicators were available from a single study (Maloney and Waxman, 1999). The
transactivation potencies of MEHP, Wy-14,643, dichloroacetic acid (DCA), and TCA for the
mouse PPARa were monitored using a luciferase reporter gene containing multiple PPAR
response elements derived from the rat hydratase/dehyrogenase promoter in transiently
transfected COS-1 monkey kidney cells. The derived potency indicators were compared to the
TD50 (i.e., the daily dose inducing tumors in half of the mice that would otherwise have remained
tumor-free) from the Carcinogenic Potency Database (CPDB) of Gold et al. (2005). Note that
for Wy-14,643, the dose listed yielded a maximal response and, thus, represents an upper limit to
the TDso (indicated by "<"). Two estimates of PPARa transactivation potency are given, the first
based on 50% of the maximal response (i.e., ECso) and the second based on the effective
concentration required for a twofold increase in activity (i.e., EC2-f0id) (Maloney and Waxman,
1999). Because unmetabolized DEHP does not exhibit PPARa activity, the transactivation
activity of its metabolite MEHP is given but compared to the hepatocarcinogenic potency
indicator for DEHP. In addition, unmetabolized tetrachloroethylene does not exhibit PPARa
activity, so is not included in the table. No data on the potency for transactivation of rat PPARa
by chemicals in the CPDB were located to enable a similar comparison in rats.
These data clearly show a lack of correlation between the potencies for in vitro PPARa
transactivation and in vivo tumorigenesis across different PPARa agonists. Especially notable is
that MEHP exhibited orders of magnitude more potency for transactivating mouse PPARa than
DCA, but DEHP was sixfold less potent as a mouse hepatocarcinogen. TCA was more similar in
potency to DCA for both outcomes, i.e., was also dramatically less active at transactivating
PPARa than DEHP despite exhibiting comparable hepatocarcinogenic potency. Wy-14,643 and
MEHP activate PPARa at comparable concentrations when directly compared in the
transactivation assay, but the carcinogenic potency of Wy-14,643 was estimated to be at least
70-fold higher than DEHP. This difference cannot be explained by pharmacokinetics (Kessler et
al., 2004; Pollack et al., 1985). Possible explanations for these results include one or more of the
following: (1) the transactivation assay is not an accurate quantitative indicator of in vivo
receptor activation, (2) the rate and nature of effects downstream of PPARa-activation depends
on the ligand or, (3) there are rate-limiting events independent of PPARa-agonism that
contribute to mouse hepatocarcinogenesis by the agonists examined.
4-173
-------
Table 4-21. Potency indicators for mouse hepatocarcinogenicity and in vitro
transactivation of mouse PPARa for four PPARa agonists3
Chemical
Carcinogenic potency
indicators (mg/kg-day)
TD50
Transactivation potency
indicators (uM)
ECso
ECi-foid
Hepatocarcinogens
Wy-14,643
DCA
TCA
DEHP/MEHP
<10.8
119
584
700
0.63
-300
-300
-0.7
-0.4
-300
-300
-0.7
a TD50, the daily dose inducing tumors in half of the mice that would otherwise have remained
tumor-free, estimated from the Carcinogenic Potency Database (Gold et al.. 2005). EC50, the
effective concentration yielding 50% of the maximal response; EC2.f0id, the effective
concentration required for a twofold increase in activity. Transactivation potencies were
estimated from Maloney and Waxman (1999). The "<" symbol denotes an upper limit due to
maximal response. A "~" symbol indicates that the transactivation potency was approximated
from figures in Maloney and Waxman (1999).
Adapted from Guy ton et al. (2009).
4.3.5.5.4.2.1 Hepatomegaly, DNA synthesis, andperoxisomeproliferation in rats
Table 4-22 compares potency indicators for various precursor effects at the TD50 for four
PPARa agonists and rat hepatocarcinogens. The analysis of whether there are consistent levels
of in vivo precursor effect induction across peroxisome proliferators at the TD50 does not include
all of the data from a similar, prior analysis by Ashby et al. (1994) for several reasons. First,
unlike the CPDB, Ashby et al. did not adjust carcinogenicity data for less-than-lifetime dosing,
which is relevant for most compounds. Second, for those mouse carcinogens reported in the
CPDB, only acute data are available regarding DNA synthesis effects from Ashby et al.
Therefore, this analysis was restricted to rat precursor and potency data for the four compounds
Wy-14,643, nafenopin, clofibrate, and DEHP and included both 1-week and 13-week data to
separately address transient and sustained changes in DNA synthesis. Even for this small set of
compounds, several limitations in the rat database were apparent. Because no single study
provided comparative data for the precursor endpoints of interest, four separate reports were
used. In the Wada et al. (1992) and Tanaka et al. (1992) studies of Wy-14,643 and clofibrate,
respectively, administered doses were within 10% of the TD50. However, nafenopin data were
only available at a single dose of 500 ppm (Lake et al., 1993), which was linearly interpolated to
the TD50. The highest administered dose of DEHP was 12,500 ppm (David et al., 1999), a dose
4-174
-------
notably below the TD50, and, thus, a lower limit based on the assumption of monotonicity with
dose is shown. A further data limitation is that in the CPDB, only the TDso for one of the four
compounds—DEHP—incorporates data from studies administering more than one dose for 2
years.
The results shown in Table 4-22 indicate that potency for the occurrence of short-term in
vivo markers of PPARa-agonism varies widely in magnitude and lacks any apparent correlation
with carcinogenic potency. Such differences have been noted previously. Similar to the results
presented in Table 4-22, Marsman et al. (1988) noted that although DEHP (12,000 ppm) and
Wy-14,643 (1,000 ppm) induced a similar extent of hepatomegaly and peroxisome proliferation
(measured either morphologically or biochemically) after 1 year, the frequency of hepatocellular
lesions was over 100-fold higher in Wy-14,643 relative to DEHP-exposed rats. In addition, a
higher labeling index was reported for 12,500 ppm DEHP than the maximal level attained after
50 to 1,000 ppm Wy-14,643 (David et al.. 1999: Tanakaetal.. 1992: Wadaetal.. 1992). Such
differences in response with dose and time observed among PPARa agonists are prominent
enough to prevent displaying dose-response data on a common scale. For instance, labeling
differences in maximal responses were not examined in this analysis.
Table 4-22. Potency indicators for rat hepatocarcinogenicity and common
short-term markers of PPARa-agonism for four PPARa agonists3
Chemical
Wy-14,643
Nafenopin
Clofibrate
DEHP
Tumor TD50
(ppm in diet)
109
275
4,225
17,900
Fold-increase over control at tumor TD50
1 wk
RLW
1.8
1.4
1.4
>1.4
LI
12
3.6
4.4
>19
PCO
13
7.6
4.2
>3.6
13 wk
RLW
2.6
1.5
1.4
>1.9
LI
6.8
1.12
0.95
>1.25
PCO
39
6.7
3.7
>4.9
Tor ease of comparison with precursor effect studies, administered doses for the tumor TD50s in the
Carcinogenic Potency Database were back-converted to equivalent ppm in diet using the formula of
Gold et al. (2005). i.e., TD50 (mg/kg-day) = TD50 (ppm in diet) * 0.04 (for male rats). Administered
doses for precursor data on Wy-14,643 (Wadaetal.. 1992) and clofibrate (Tanakaetal.. 1992) were
within 10% of the TD50. Because nafenopin precursor data were only available at 0 and 500 ppm (Lake
etal.. 1993). these doses were linearly interpolated to the TD50. Because the highest administered dose
of DEHP in precursor effect studies was 12,500 ppm (David etal.. 1999). a lower limit is shown, based
on the assumption of monotonicity with dose. RLW = relative liver weight, LI = labeling index, PCO =
cyanide insensitive palmitoyl CoA oxidation.
Adapted from Guyton et al. (2009).
4-175
-------
Together, these findings underscore the significant chemical-specific quantitative differences in
these markers that limit their utility for predicting carcinogenic dose-response relationships.
4.3.5.5.4.3. Evidence from transgenic animals
Data from transgenic animals suggest the key events in the hypothesized MO A—PPARa-
activation, hepatocellular proliferation, and clonal expansion—are not sufficient to cause tumors.
This suggests that other events not mediated by PPARa-activation, either independently or in
combination with PPARa-activation, are necessary to induce tumors. The discussion below is
based on the review by Guyton et al. (2009).
Yang et al. (2007) raises questions regarding whether PPARa-activation in hepatocytes is
causally linked to hepatocarcinogenesis as a sole operant MOA. The experimental approach
entailed fusing the mouse PPARa to the potent viral transcriptional activator VP16 under control
of the liver enriched activator protein (LAP) promoter, resulting in targeted constitutive
expression of activated PPARa in hepatocytes. In LAP-VP16PPARa transgenic mice, ligand-
independent hepatocyte PPARa-activation evoked many of the same hepatic responses (in type
and magnitude) as observed with PPARa ligand treatment of companion wild-type 129/Sv mice.
For instance, DNA synthesis was increased in LAP-VP16PPARa transgenic mice; the effect was
persistent and still evident at 11 months of age. In addition, increases were reported in markers
of peroxisome proliferation (including increases in expression of peroxisomal membrane protein
70, acyl CoA oxidase and CYP4A family genes, and enhanced cyanide insensitive palmitoyl
CoA oxidation). Other effects included an increase in cell-cycle genes (cyclin Dl and cyclin-
dependent kinases 1 and 4) and a decrease in serum triglycerides and free fatty acids. Together,
these results are consistent with the view that PPARa-activation and its sequelae are alone
sufficient to induce increased hepatocyte DNA synthesis and peroxisome proliferation.
However, constitutive PPARa-activation in hepatocytes in the LAP-VP16PPARa
transgenic mouse model was not sufficient to induce several important hepatic responses
stimulated by PPARa ligand treatment of wild-type mice. Notably, no preneoplastic hepatic
lesions or hepatocellular neoplasia were found in ">20 LAP-VP16PPARa mice at the age of
over 1 year" (Yang et al., 2007). In sharp contrast, wild-type mice exposed to the PPARa
agonist Wy-14,643 for 11 months developed grossly visible lesions consistent with previous
reports of its hepatocarcinogenicity [e.g., Peters et al. (1997)1. Interestingly, nonparenchymal
cell proliferation was observed with Wy-14,643 exposure of wild-type mice but was absent in
the LAP-VP16PPARa transgenic mice. In addition, although liver weight was increased in
LAP-VP16PPARa transgenic mice, the extent of hepatomegaly was reduced in comparison to
Wy-14,643-exposed wild-type mice, and hepatocellular hypertrophy was absent.
4-176
-------
Thus, the Yang et al. (2007) study provides evidence that, by itself, PPARa-activation
(and its sequelae) is not sufficient to induce hepatocarcinogenesis. These data are, therefore,
inconsistent with the hypothesis that effects mediated through PPARa-activation constitute a
complete MOA for carcinogenesis. Notably, key events in the proposed MOA such as the robust
and sustained elevation in hepatocyte proliferation (evidenced by enhanced DNA synthesis),
accompanied by enzyme changes commonly associated with peroxisome proliferation, did not
evoke hepatocarcinogenesis. In fact, a comparable extent of sustained increases in hepatocyte
DNA synthesis was observed with constitutive PPARa-activation in the LAP-VP16PPARa
transgenic mouse model and Wy-14,643 exposure in wild-type mice, but only the latter
developed liver tumors under comparable experimental paradigms.
4.3.5.5.5. Rationale for species differences
Toxicodynamic differences across species, including in the absolute or allometrically
scaled amount or activity of the receptor, may contribute to differences in sensitivity of response
to PPARa agonists. Absolute levels of PPARa are generally thought to be lower in human
compared with rodent liver. However, PPARa amount varies by an order of magnitude among
individuals (Palmer et al., 1998; Tugwood et al., 1996), e.g., 1 of the 6 human samples examined
expressed levels comparable to the mouse in one study (Walgren et al., 2000). The pattern of
PPARa expression across tissues also differs across species (Melnick, 2001; Tugwood et al.,
1996), e.g., human levels are higher in kidney and skeletal muscle than in liver, while the highest
rodent levels are in liver and kidney. In addition, considerable interindividual variation in
PPARa structure and function among humans has been reported (Tugwood et al., 1996), and
polymorphisms have been shown to increase or decrease receptor levels and to modulate
baseline lipid and apolipoprotein levels, atherosclerotic progression, and the presence of diabetes
mellitus and insulin resistance (Tanaka et al., 2007; Tai et al., 2006; Flavell et al., 2005; Foucher
et al., 2004; Flavell et al., 2002; Jamshidi et al., 2002). An impact of PPARa polymorphisms on
preexisting disease status and response to PPARa agonists is also suggested from bezafibrate
[2-(4-(2-[(4-chlorophenyl)formamido]ethyl)phenoxy)-2-methylpropanoic acid] and gemfibrozil
[5-(2,5-dimethylphenoxy)-2,2-dimethyl-pentanoic acid] trials (Tai et al., 2006; Jamshidi et al.,
2002).
The human PPARa is functional in in vitro transactivation assays and is responsive to a
number of PPARa agonists (e.g., nafenopin, clofibrate, and WY-14,643) (Maloney and
Waxman, 1999; Mukherjee et al., 1994; Sheretal., 1993). Compared with the mouse PPARa,
human PPARa is suggested to be 10- to 20-fold less responsive to Wy-14,643 (Maloney and
Waxman, 1999; Palmer et al., 1998; Mukherjee et al., 1994). However, this magnitude of
interspecies difference has not been demonstrated for other compounds. Hurst and Waxman
4-177
-------
(2003) reported a fivefold lower sensitivity to the DEHP metabolite MEHP of human—
compared with mouse—PPARa (ECso = 3.2 uM vs. 0.6 uM) in transfected COS-1 monkey
kidney cells, but acknowledged that they could not quantify the relative amount of each receptor.
Using a similar experimental paradigm, Wolf et al. (2008) found an approximately twofold lower
slope of the dose-response curve for activation of human—compared with mouse—PPARa for
perfluorooctanoic acid and other perfluoroalkyl acids. For other PPARa agonists, including
TCA and DC A, little (
-------
nonparenchymal cells (e.g., Kupffer cells), may limit the in vitro hepatocyte proliferative
response, as observed for other species [e.g., Parzefall et al. (2001)]. The extent of peroxisome
proliferation in human liver following exposure to fibrate drugs (e.g., with clofibrate,
gemfibrozil, or fenofibrate) or dialysis treatment (possibly due to DEHP exposure) is reported to
be generally less than the rodent response (Ganning et al., 1987; Gariot et al., 1987; Panning et
al.. 1984: Bliimcke et al., 1983: Hanefeld et al.. 1983: DeLalglesia et al., 1982: Hanefeld et al..
1980). However, the ability to quantitatively characterize human sensitivity to this effect is
limited (e.g., by the small number of subjects studied).
In sum, despite notable qualitative similarities, quantitative differences in receptor
activation and the subsequent events in the hypothesized MOA are evident across species. The
magnitude of these differences has been best characterized for Wy-14,643, to which rodents
appear to have 10-fold or more greater sensitivity for response (Morimura et al., 2006: Cheung et
al.. 2004: Yu et al.. 2001: Malonev and Waxman, 1999: Palmer etal.. 1998: Mukherjee et al..
1994). Although more limited, studies of other agonists suggest a smaller magnitude of
difference in sensitivity for response across species than is observed for Wy-14,643 (Hurst and
Waxman, 2003: Yu et al., 2001: Malonev and Waxman, 1999). Considerable interindividual
variation in PPARa amount, structure, and function has been reported among humans (Tugwood
et al., 1996), and some studies have suggested variability in human response to PPARa agonists
(Tai et al., 2006: Jamshidi et al., 2002). However, few studies have examined directly how these
factors may affect sensitivity—as well as the potential for heterogeneity of response—to
hepatocarcinogenesis induced by PPARa agonists in humans.
Another consideration is whether human epidemiologic data on fibrates offer an indirect
test of the PPARa-activation MOA hypothesis. Human exposures to exogenous and endogenous
PPARa agonists encompass a broad group of chemicals, including environmental contaminants
known to activate the receptor, as well as a number of therapeutic agents whose molecular target
is one or more receptors in the PPAR family. Indeed, fibrate drugs were developed using rodent
models to treat hyperlipidemia in humans before the receptor was identified. These agents have
varying degrees of affinity for PPARa (Shearer and Hoekstra, 2003), and some have multiple
mechanisms of action. Drugs that have PPARa agonist activity include fibrates or fibric acid
derivatives (which are primarily PPARa agonists), bezafibrate (which also shows PPARy
activity), dual PPARa/y agonists currently under development, the glitazones, and nonsteroid
anti-inflammatory drugs (e.g., ibuprofen) (Sertznig et al., 2007).
Some human data on PPARa agonist effects are available from fibrate clinical trials and
population case-control studies of site-specific cancer (Freeman et al., 2006: Tenkanen et al.,
2006: Keech et al., 2005: Diabetes Atherosclerosis Intervention Study Investigators, 2001:
Meade and clinics, 2001: BIP Study Group, 2000: Rubins etal., 1999: Fricketal., 1997: deFaire
4-179
-------
etal.. 1995: Huttunen et al.. 1994: Rubins etal.. 1993: Fricketal.. 1987: Canneretal.. 1986:
WHO. 1984. 1980. 1978: Coronary Drug Project Research Group. 1977. 1975). These studies
examined a range of human responses to PPARa agonists, which included atherosclerosis,
cardiovascular disease, serum biomarkers of fatty acid metabolism, acute toxicity, and, more
limitedly, organ-specific chronic toxicity, including cancer. However, examination of
hepatotoxicity in the fibrate clinical trials has been limited to alterations in hepatic metabolic
pathways and changes in liver enzymes as assessments of drug tolerance, because the primary
focus of these trials was cardiovascular events.
Reviews of the PPARa-activation MOA hypothesis have generally focused on liver
cancer response in two fibrate clinical trials, the Helsinki Heart Study (Tenkanen et al., 2006:
Huttunen et al., 1994: Frick et al., 1987) and the World Health Organization's Cooperative Trial
on Primary Prevention of Ischemic Heart Disease (WHO, 1984, 1980, 1978), and have
concluded that, while limited, those data did not provide evidence of an increased liver cancer
risk from fibrate exposure (Klaunig etal., 2003: Ashby etal., 1994). However, the available
studies have low power to detect statistical differences in the risk of liver cancer; an estimated
five or fewer liver cancer deaths would have been expected in these studies using data from the
National Cancer Institute's Surveillance, Epidemiology, and End Results database (Ries et al.,
2008). This low statistical power, in addition to the studies' exclusion or removal of subjects
showing signs of liver (or other) toxicity from treatment, precludes a strong conclusion about the
presence or lack of liver cancer risk. These studies and the other fibrate trials did not examine
site-specific causes of mortality or morbidity and did not follow subjects for a sufficient period
to adequately consider cancer latency; in addition, placebo subjects were offered fibrate therapy
at the end of the clinical trials, making analyses after further follow-up difficult to interpret. For
example, the three trials that did assess mortality after a follow-up period longer than 10 years
included liver cancers in a larger category of contiguous sites or in the category of all cancers,
introducing disease misclassification and a downward bias for any site-specific treatment-related
cancers (Tenkanen et al., 2006: Huttunen et al., 1994: Canneretal., 1986: WHO, 1984, 1980,
1978). In voluntary postmarketing safety reports to the U.S. Food and Drug Administration
(FDA), rates of liver adverse event reports for gemfibrozil and fenofibrate (2.6 and 6.9 per
1,000,000 prescriptions, respectively) were similar to that of statins (Holoshitz et al., 2008).
However, an examination of liver cancer is precluded by the general under-reporting of chronic
toxicities to FDA, and the lack of specific FDA reporting requirements for cancer, even
premarketing. Because of these inadequacies, the available epidemiologic data for fibrate drugs
cannot inform conclusions about the relevance of PPARa-activation to human cancer.
4-180
-------
4.3.5.6. Mode of Action Conclusions for Hepatocellular Tumors
There is only limited experimental support for the position that tetrachloroethylene-
induced hepatocarcinogenesis is mediated solely by the hypothesized PPARa-activation MOA.
Chemical-specific data for PPARa-activation support the view that this is not the primary MOA
for hepatocarcinogenesis. Philip et al. (2007) suggest that PPARa-activation is not required for
the observed cell proliferative response. This is based on evidence of significantly increased
CYP4A expression at only the highest dose (1,000 mg/kg-day) and at the earliest time point
(7-days), in contrast to the robust dose-dependent proliferative response of a more prolonged
nature (lasting for 14-30 days post exposure) observed at the same and lower (150, 500, and
1,000 mg/kg-day) levels of tetrachloroethylene. The authors concluded that their findings
suggest peroxisome proliferation is not a sustained response in spite of continued
tetrachloroethylene exposure and, therefore, are not supportive of a close mechanistic
relationship of carcinogen!city and PPARa induction for tetrachloroethylene-derived TCA.
Limitations of this interpretation include the possible lack of sensitivity of CYP4A protein
expression as a marker of peroxisome proliferation, and the unknown sensitivity of the SW
mouse to tetrachloroethylene hepatocarcinogenicity. However, other investigators [e.g.,
Schumann et al. (1980)] have reported liver toxicity and repair at 100 mg/kg-day in the B6C3Fi
strain, whereas repeated exposures to 1,000 mg/kg-day were reported by Philip et al. (2007) and
Odum et al. (1988) to only modestly increase peroxisomal markers in SW and B6C3Fi mice,
respectively. Odum et al. (1988) also observed moderate increases in peroxisome proliferation
in rats, a species insensitive to tetrachloroethylene hepatocarcinogenicity. In all, these findings
indicate that the modest peroxisome proliferation observed in response to tetrachloroethylene
may lack specificity with respect to species, tissue, and dose. Studies of the temporal sequence
of events are limited. Given the limitations in the database of tetrachloroethylene-specific
studies, it can be concluded that the few studies demonstrating peroxisome proliferation by
tetrachloroethylene are insufficient to demonstrate a causative role of this effect in the induction
of other key events posited for the PPARa mode of action hypothesis, and for
hepatocarcinogenesis by tetrachloroethylene.
Other data and analyses more generally support the view that the hypothesized PPARa-
activation MOA is not a sole causative factor in rodent hepatocarcinogenesis. PPARa-agonism
may play a significant role in mouse liver tumor induction by some compounds, such as
Wy-14,643. However, recent studies suggest that DEHP can induce tumors in a PPARa
independent manner without any loss of potency (Ito et al., 2007a), and that PPARa-agonism in
hepatocytes is itself insufficient to cause tumorigenesis (Yang et al., 2007). Additional analyses
presented above demonstrate that peroxisome proliferation and associated markers are poor
quantitative predictors of hepatocarcinogenesis in rats or mice. These data and analyses raise
4-181
-------
serious concerns about basing human health risk assessment conclusions exclusively on evidence
of key events in the hypothesized PPARa-activation MO A, given that other modes, mechanisms,
toxicity pathways, and molecular targets may contribute to or be required for the observed
adverse effects. Indeed, for most PPARa agonists, chemical-specific data to define the range of
effects that may contribute to human carcinogenesis are insufficient. Similarly, the
epidemiologic data are inadequate to inform conclusions of human relevance.
A recent review (Rusyn et al., 2006) addressed other mechanistic effects of the PPARa
agonist DEHP and proposed that tumors arise from a combination of molecular signals and
pathways, rather than from a single event such as PPARa-activation. Indeed, the PPARa
agonists are pleiotropic and have been reported to exhibit a diversity of responses in addition to
the hallmark effect of peroxisome proliferation, including genotoxicity [reviewed by Melnick
(2001)1, epigenetic alterations (e.g., hypomethylation) (Pogribny et al., 2007), oxidative stress
[reviewed in O'Brien et al. (2005)] and effects on other receptors [e.g., Guo et al. (2007)] and
other organelles (e.g., mitochondria) within parenchymal cells (Scatena et al., 2003; Zhou and
Wallace, 1999; Youssef andBadr, 1998; Lundgren et al., 1987). As reviewed above, the
metabolites of tetrachloroethylene have been shown to induce a number of effects that may
contribute to carcinogenicity, including mutagenicity, alterations in DNA methylation, and
oxidative stress. Given the demonstrated mutagenicity of several tetrachloroethylene
metabolites, the hypothesis that mutagenicity contributes to the MOA for tetrachloroethylene
carcinogenesis cannot be ruled out, although the specific metabolic species or mechanistic
effects are not known. Epigenetic effects and oxidative stress, including that produced
secondary to cytotoxicity, may also contribute. Currently, the available database of
tetrachloroethylene-specific studies addressing these mechanisms is very limited and merits
further exploration.
Cancer is a complex, multicausal process that is characterized by the acquisition and/or
activation of multiple critical traits. As described by Hanahan and Weinberg (2000), these traits
or hallmarks comprise six essential features: (1) self-sufficiency in growth signals, (2)
insensitivity to growth-inhibitory (antigrowth) signals, (3) evasion of programmed cell death
(apoptosis), (4) limitless replicative potential, (5) sustained angiogenesis, and (6) tissue invasion
and metastasis. Epigenetic changes (e.g., in the expression of microRNAs that negatively
regulate gene expression by targeting mRNA for translational repression or cleavage) appear to
contribute to many of the observed phenotypic alterations. The acquisition of these six
capabilities can also be facilitated by genomic instability, another feature of the cancer
phenotype. A number of factors, such as inflammation (Grivennikov et al., 2010), diet, and
physiological factors [e.g., obesity (Park et al., 2010)1, can affect the tumor microenvironment in
ways that advance these features of tumor development. Studies of human hepatocarcinogenesis
4-182
-------
reveal significant heterogeneity, with evidence of aberrant signaling in multiple, overlapping
pathways involved in cellular proliferation (e.g., EGF, HGF, RAS/mitogen-activated protein
kinase), survival, differentiation (e.g., Wnt, Hedgehog), and angiogenesis (e.g., VEGF, PDFG,
FGF) [refer to recent review by Hoshida et al. (2010)1. Other studies have provided support for a
hypothesized role of stem cells in hepatocarcinogenesis (Marquardt and Thorgeirsson, 2010). In
contrast to the stochastic cancer model, the cancer stem cell hypothesis posits a hierarchical
model in which a minor cell population possessing sternness undergoes epigenetic changes to
generate heterogeneous tumors [refer to review by Reya et al. (2001). The potential cell types of
origin of liver cancer stem cells include mature hepatocytes possessing stem-like characteristics,
as well as circulating cells (Kim et al., 2009) including bone-marrow derived stem cells
(Marquardt and Thorgeirsson, 2010). Such stem cells have been posited to play a role in liver
development and regeneration in addition to carcinogenesis [refer to review by Kung et al.
(2010)]. Thus, although significant knowledge gaps remain, particularly with respect to the
particular pathways and processes necessary and sufficient for the disease to originate and
develop, the etiology of hepatocarcinogenesis appears complex.
Given the multiple metabolites and mechanisms that may contribute, and the known
complexity and heterogeneity in liver cancer development, in general, it is unlikely that a single
causative metabolite, mechanism, pathway, or mode of action will be identified for
tetrachloroethylene-induced hepatocarcinogenesis. A single, linear sequence of key events does
not seem likely to explain the observed hepatocarcinogenicity, given the multiple cell types and
processes involved. Instead, a plausible hypothesis may be posited of multiple, contributing
mechanistic effects that may, in turn, be affected by multiple modifying factors. Accordingly,
the mechanisms described in this review are not intended to be interpreted as being mutually
exclusive. Altogether, the described mechanistic effects may aid in identifying sources of human
vulnerability, as well as informing the likelihood of other outcomes influenced by the same
mechanisms, pathways, and biological processes. They may be informative of future analysis
integrating data on human "upstream" biomarkers of hepatocarcinogenesis with chemically
induced perturbations. In this manner, the mechanistic data may be informative for addressing
the issues of cumulative assessment across exposures as well as overall population risk.
In summary, as noted by NRC (2010), there are significant gaps in the scientific
knowledge of mechanisms contributing to tetrachloroethylene-induced mouse liver cancer.
Multiple metabolites formed from tetrachloroethylene are toxic and carcinogenic in the liver.
Given this knowledge, and the known complexity and heterogeneity in liver cancer development,
in general, the available evidence supports a hypothesis of multiple, contributing mechanistic
effects that may, in turn, be affected by multiple modifying factors.
4-183
-------
4.4. ESOPHAGEAL CANCER
Thirteen epidemiologic studies reporting data on esophageal cancer and
tetrachloroethylene exposure were identified. This set of publications includes 11 cohort or
nested case-control studies (Calvert et al., 2011; Selden and Ahlborg, 2011; Pukkala et al., 2009;
Sung et al.. 2007: Lynge et al.. 2006: Blair et al.. 2003: Chang et al.. 2003: Travier et al.. 2002:
Andersen et al., 1999: Boice et al., 1999: Lynge and Thygesen, 1990) and two case-control
studies of occupational exposures (Vaughan et al., 1997: Siemiatycki, 1991). No studies of
residential exposure through contaminated drinking water were identified in the literature review.
These 13 studies represent the core studies evaluated by EPA, as described in more detail below.
Two other cohort studies included information on tetrachloroethylene but did not report risk
estimates for esophageal cancer (Radican et al., 2008: Anttila et al., 1995), and one case-control
study did not observe any cases exposed as a dry cleaner (Siemiatycki, 1991), and so were not
evaluated further. There is some overlap in the study populations among these studies: Travier et
al. (2002) used occupational data from the Swedish national census, and Lynge and Thygsen
(1990) used a similar design in Denmark; Andersen et al. (1999) and Lynge et al. (2006)
expanded these studies to include Denmark, Finland, and Norway in addition to Sweden, and
Pukkala et al. (2009) added Iceland to this set. Appendix B reviews the design, exposure-
assessment approach, and statistical methodology for each study. All studies were of the
inhalation route, of occupational exposure, and, except for the case-control study of Vaughan et
al. (1997), unable to quantify tetrachloroethylene exposure.
4.4.1. Consideration of Exposure-Assessment Methodology
Many studies examine occupational title as dry cleaner, launderer, and presser as
surrogate for tetrachloroethylene, given its widespread use from 1960 onward in the United
States and Europe (Calvert et al., 2011: Pukkala et al., 2009: Lynge et al., 2006: Blair et al.,
2003: Travier et al., 2002: Andersen et al., 1999: Lynge and Thygesen, 1990). Six studies
conducted in Nordic countries are based on either the entire Swedish population or on combined
populations of several Nordic countries; strengths of these studies are their use of job title as
recorded in census databases and ascertainment of cancer incidence using national cancer
registries (Selden and Ahlborg, 2011: Pukkala et al., 2009: Lynge et al., 2006: Travier et al.,
2002; Andersen et al., 1999; Lynge and Thygesen, 1990). Studies examining mortality among
U.S. dry-cleaner and laundry workers (Calvert et al., 2011; Blair et al., 2003) are of smaller
cohorts than most Nordic studies, with fewer observed esophageal cancer events.
The exposure surrogate in studies of dry-cleaners and laundry workers is a broad
category containing jobs of differing potential for tetrachloroethylene exposure. Thus, these
studies have a greater potential for exposure misclassification bias compared to studies with
4-184
-------
exposure potential to tetrachloroethylene assigned by exposure matrix approaches. Three studies
used additional information pertaining to work environment to refine the exposure classification
(Calvert etal.. 2011: Selden and Ahlborg, 2011: Lynge et al.. 2006). Selden and Ahlborg (2011)
obtained information about the dry-cleaning establishment (e.g., washing techniques, chemicals
used, number of employees, and work history of individual employees) in a questionnaire sent to
businesses in Sweden in the 1980s. Lynge et al. (2006), using job title reported in the 1970
Census, identified subjects based on occupational code of "laundry and dry-cleaning worker" or
industry code of "laundry and dry cleaning." Additional information to refine this occupational
classification was sought for incident cancer cases, including esophageal cancer, within this
defined cohort. Five controls, matched to the cases by country, sex, age, and calendar period,
were also included in the study. The additional information included handwritten task
information from the census forms from Denmark and Norway, pension databases in Denmark
and Finland, and next-of-kin interviews in Norway and Sweden. Exposure classification
categories were dry cleaner (defined as dry cleaners and supporting staff if employed in business
of <10 workers), other job titles in dry cleaning (launderers and pressers), unexposed (job title
reported on 1970 Census was other than in dry cleaning), or unclassifiable (information was
lacking to identify job title of subject). The unclassifiable category represented 18 out of
72 esophageal cancer cases (25%) and 108 out of 567 controls (19%). The study by Calvert et
al. (2011) of unionized dry cleaners in the United States included an analysis of subjects who
worked for one or more years before 1960 in a shop known to use tetrachloroethylene as the
primary solvent (Calvert et al., 2011: Ruder et al., 2001, 1994). The cohort was stratified into
two groups based on the level of certainty that the worker was employed only in facilities using
tetrachloroethylene as the primary solvent; tetrachloroethylene-only and tetrachloroethylene
plus. There were 6 esophageal cancer deaths among this subset (n = 618) of the study subjects.
Calvert et al. (2011) also presented risk estimates by exposure duration and by latent periods for
the full set of study subjects. Two additional studies used an exposure metric for
semi quantitative or quantitative exposure within a dry-cleaning setting. Blair et al. (2003) used
an exposure metric for semi quantitative cumulative exposure, and the case-control study of
Vaughan et al. (1997) used a JEM with quantitative exposure assessment for dry-cleaning and
laundry jobs.
Two other cohorts with potential tetrachloroethylene exposure in manufacturing settings
have been examined. These studies include aerospace workers in the United States (Boice et al.,
1999) and electronic factory workers in Taiwan (Sung et al., 2007: Chang et al., 2003). Boice et
al. (1999) used an exposure assessment based on a job-exposure matrix to classify exposures. In
contrast, the exposures in the Taiwan studies included multiple solvents, tetrachloroethylene
4-185
-------
exposure was not linked to individual workers, and cohorts included both white- and blue-collar
workers (Sung et al.. 2007: Chang et al.. 2003).
In summary, with respect to exposure-assessment methodologies, five studies with
esophageal cancer data assigned tetrachloroethylene exposure to individuals using a
semi quantitative surrogate or a job exposure matrix (Blair et al., 2003; Boice et al., 1999;
Vaughan et al., 1997), information about working conditions obtained through a questionnaire
(Selden and Ahlborg, 2011), or a classification of the cohort by certainty of tetrachloroethylene
exposure (Calvert et al., 2011). One other study based on occupational census data sought
additional data for use in refining potential exposure within dry-cleaning settings (Lynge et al.,
2006). The relative specificity of these exposure-assessment approaches strengthens their ability
to identify cancer hazards compared to studies with broader and less sensitive exposure-
assessment approaches.
4.4.2. Summary of Results
All studies evaluated by EPA reported estimated relative risks based on a small number
of observed events; 35 or fewer deaths/incident cases in cohort studies (Calvert et al., 2011; Sung
et al., 2007: Lynge et al., 2006: Blair et al., 2003: Chang etal.. 2003: Travier et al., 2002:
Andersen et al., 1999: Boice et al., 1999: Lynge and Thygesen, 1990), except Pukkala et al.
(2009), whose esophageal cancer findings are based on 95 exposed subjects. The few
esophageal cancers in cohort studies and exposed cases in case-control studies contribute to
reduced statistical power and limited ability to inform an evaluation of tetrachloroethylene
exposure, particularly for esophageal cancer, whose estimated incidence is lower than for other
cancer sites discussed in Section 4 (Edwards et al., 2010).
The largest cohort study observed an SIR estimate of 1.18 (95% CI: 0.96, 1.46) (Pukkala
et al., 2009). Some evidence for an association between esophageal cancer risk and ever having
a job title of dry cleaner or laundry worker or routine exposure to tetrachloroethylene is also
found in cohort studies24 whose effect estimates are based on fewer observed events and that
carry lesser weight in the analysis. As expected, the magnitude of the point estimate of the
association reported in these studies is more variable than in the larger study. The smaller cohort
studies reported risks of 0.74 (95% CI: 0.41, 1.25), 1.16 (95% CI: 0.14, 4.20), 1.32 (95% CI:
0.94, 1.85), 1.47 (95% CI: 0.54, 3.21), 2.2 (95% CI: 1.15, 3.3), and 2.44 (95% CI: 1.4, 3.97) in
Lynge and Thygsen (1990), Sung et al. (2007), Travier et al. (2002), Boice et al. (1999), Blair et
al. (2003), and Calvert et al. (2011), respectively (refer to Table 4-23). The 10-year follow-up
period in Lynge and Thygsen (1990) may represent an insufficient latent period with respect to
24 Andersen et al. (1999) is not included in this summary of the data from the individual studies because it was
updated and expanded in the analysis by Pukkala et al. (2009).
4-186
-------
the development of cancer, reducing the study's sensitivity compared to Pukkala et al. (2009),
whose follow-up was >15 years.
The nested case-control study of Lynge et al. (2006) reported an odds ratio of 0.76
(95% CI: 0.34, 1.69) for dry cleaners, with 8 exposed cases, compared to no exposure. In this
study, job title could not be classified for 25% of the cases and 19% of the controls. The odds
ratio for risk cancer in this "unclassifiable" group was 2.04 (95% CI: 0.91, 4.62). Lynge et al.
(2006) carried out sensitivity analyses using different assumptions regarding the true
classification for these subjects. In these analyses, the odds ratio for the association between dry
cleaner and esophageal cancer was 0.66 (95% CI: 0.30, 1.45) assuming all unclassified subjects
were unexposed and 1.19 (95% CI: 0.67, 2.21) assuming all unclassified subjects were dry
cleaners. One other case-control study that adopted a JEM approach to assign exposure reported
odds ratios of 6.5 (95% CI: 0.6, 68.9) and 0.9 (0.1, 10.0) for overall exposure to
tetrachloroethylene, based on two and one exposed case, respectively, for squamous cell
carcinoma and adenocarcinoma of the esophagus (Vaughan et al., 1997).
Several studies had been previously identified based on the relative strengths of their
exposure-assessment methodology. The results from these studies are mixed. Lynge et al.
(2006) reported no evidence of an increased risk among individuals classified as dry cleaners,
with relative risks of 0.76, but a higher risk was observed in the "unclassifiable" group (RR:
2.04). Selden and Ahlborg (2011) reported similar but slightly higher relative risks for laundry
workers (SIR: 1.56) compared with dry cleaners (SIR: 1.25). In contrast, data from other studies
with relatively strong exposure-assessment methods provide more evidence of an effect, with
relative risks of 1.47 [Boice et al. (1999): routine exposure], 2.2 (Blair et al.. 2003). and 2.68
[Calvert et al. (2011): tetrachloroethylene-only workers], and 6.4 (Vaughan et al., 1997).
4-187
-------
Table 4-23. Summary of human studies on tetrachloroethylene exposure and esophageal cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort Studies
Biologically monitored workers
All subjects
Not reported
Aerospace workers (Lockheed)
Routine exposure to PCE
1.47(0.54,3.21)
6
Routine-Intermittent exposure to PCEa
Duration of exposure
Never exposed
5yr
p for trend
1.0b
1.0 (0.30, 3.34)
0.79 (0.27, 2.50)
0.91 (0.13, 1.60)
p = 0.07
28
3
4
3
Electronic factory workers (Taiwan)
All Subjects
Males
Females
Females
1.16(0.14,4.20)
0
0
2
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
Not reported
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females and 0.7
umol/L in males (median)], follow-up 1974-1992, external referents (SIR)
Boice et al. (1999)
77,965 (« = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after 1960,
worked at least 1 yr, follow-up 1960-1996, job exposure matrix without
quantitative estimate of PCE intensity, 1987-1988 8-h TWA PCE
concentration (atmospheric monitoring) 3 ppm [mean] and 9.5 ppm
[median], external reference for routine exposure (SMR) and internal
references (workers with no chemical exposures) for routine -intermittent
PCE exposure (RR)
Chang et al. (2003): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1985-1997, multiple solvents
exposure, does not identify PCE exposure to individual subjects, cancer
mortality, external referents (SMR) (Chang et al.. 2003):
63,982 females, follow-up 1979-2001, factory employment proxy for
exposure, multiple solvents exposures and PCE not identified to individual
subjects, cancer incidence, external referents, analyses lagged 10 yr (SIR)
(Sung etal. 2007)
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever exposed
to mixed solvents, 85 1 ever-exposed to PCE), employed at least 1 yr from
1952 to 1956, follow-up 1973-2000, job exposure matrix (intensity),
internal referent (workers with no chemical exposures) (RR)
oo
oo
-------
Table 4-23. Summary of human studies on tetrachloroethylene exposure and esophageal cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Dry cleaner and laundry workers
All laundry worker and dry cleaners
Males
Females
0.91 (0.57, 1.40)
0.82 (0.33, 1.70)
0.97 (0.53, 1.62)
21
7
14
All subjects
2.2(1.5,3.3)
26
Semiquantitative exposure score
Little to no exposure
Medium to high exposure
2.1(0.9,4.4)
2.2(1.2,3.5)
7
16
All laundry worker and dry cleaners
Males
Females
0.74 (0.41, 1.25)
0.62 (0.23, 1.35)
0.88 (0.38, 1.73)
14
6
8
Launderer and dry cleaner
Male
Female
1.18(0.96, 1.46)
0.99 (0.66, 1.44)
1.29 (1.00, 1.64)
95
28
67
Reference
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or 1970
Census (Denmark, Finland, Norway), follow-up 1971-1987 or 1991,
PCE not identified to individual subjects, external referents (SIR)
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union members
(1945-1978), follow-up 1979-1993, semiquantitative cumulative
exposure surrogate to dry clean solvents, cancer mortality, external
referents (SMR)
Lynge and Thygsen (1990)
10,600 Danish men and women, 20-64 yr old, employed in 1970 as
laundry worker, dry cleaners and textile dye workers, follow-up
1970-1980, external referents (SIR)
Pukkala et al. (2009)
Men and women participating in national census on or before 1990, 5
Nordic countries (Denmark, Finland, Iceland, Norway, Sweden), 30 -64
yr, follow-up 2005, occupational title of launderer and dry cleaner in any
census, external referents (SIR)
oo
VO
-------
Table 4-23. Summary of human studies on tetrachloroethylene exposure and esophageal cancer (continued)
Exposure group
All subjects
Exposure duration/time since 1st employment
<5 yr/<20 yr
<5 yr/>20 yr
>5 yr/<20 yr
>5 yr/>20 yr
PCE-only subjects
Dry-cleaners and laundry workers (females)
PCE (females)
Laundry (females)
All subjects, 1960 or 1970 Census in laundry
and dry cleaner occupation and industry
All subjects in 1960 and 1970 in laundry and dry
cleaner occupation and industry
Relative risk
(95% CI)
2.44 (1.4, 3.97)
2.16(0.85,4.54)
4.78(2.68,7.91)
2.68 (0.98, 5.83)
1.33(0.43,3.10)
1.25 (0.26, 3.25)
1.56(0.19,5.65)
1.32 (0.94, 1.85)
0.34(0.05,2.39)
No. obs.
events
16
0
5
0
11
6
5
3
2
34
1
Reference
Calvert et al. (2011)
1,704 U.S. men and women dry-cleaning union member in CA, IL, MI,
NY follow-up 1940-2004 (618 subjects worked for one or more years
prior to 1960 only at shops where PCE was the primary cleaning solvent,
identified as PCE-only exposure), cancer mortality (SMR)
Selden and Ahlborg (2011)
9,440 Swedish men (n = 2,810) and women (n = 9,440) in 461 washing
and dry-cleaning establishments, identified by employer in mid-1980s,
employed 1973-1983, follow-up 1985-2000, exposure assigned using
company serf-reported information on PCE usage — PCE (dry cleaners
and laundries with a proportion of PCE dry cleaning), laundry (no PCE
use), and other (mixed exposures to PCE, CFCs, TCE, etc.), external
referents (SIR). No observed cases in males
Travier et al. (2002)
Swedish men and women identified in 1960, 1970, or both Censuses as
laundry worker, dry cleaner, or presser (occupational title) or in the
laundry, ironing, or dyeing industry, follow-up 1971-1989, separates
launders and dry cleaners form pressers, external referents (SIR)
VO
o
-------
Table 4-23. Summary of human studies on tetrachloroethylene exposure and esophageal cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Case-Control Studies
Nordic Countries (Denmark, Finland, Norway, Sweden)
Unexposed
Dry cleaner
Assume unclassifiable exposed as dry cleaner
Assume unclassifiable unexposed
Other in dry-cleaning
Unclassifiable
1.00
0.76 (0.34, 1.69)b
1.19 (0.67, 2.21)b
0.66 (0.30, 1.45)b
1.22 (0.41, 3.63)b
2.04 (0.91, 4.62)b
41
8
26
59
5
18
Dry cleaner, employment duration, 1964-1979
Unexposed
10yr
Unknown
1.0
1.20(0.14, 10.41)b
0.66(0.19, 2.29)b
0.70 (0.20, 2.49)b
1.65(0.18, 14.98)b
41
0
1
3
3
1
Montreal, Canada
Launderers and dry cleaners
Any exposure
Substantial exposure
(0.0, 2.4)
(0.0, 4.3)
0
0
Lynge et al. (2006)
Case-control study among 46,768 Danish, Finnish, Norwegian, and
Swedish men and women employed in 1960 as laundry worker or dry
cleaner, follow-up 1970-1971 to 1997-2001, 72 incident esophageal
cancer cases, 6 controls per case randomly selected from cohort matched
on country, sex, age, calendar period at diagnosis time, occupational task
at 1970 Census proxy for exposure, RR adjusted for matching criteria
Siemiatycki (1991)
Histologically confirmed esophageal cancers (n = 99), 1979-1985,
35-70 yr, population control group and cancer control group, in-person
interviews, occupational title, OR adjusted age, family income, and
cigarette index, 90% CI
VO
-------
Table 4-23. Summary of human studies on tetrachloroethylene exposure and esophageal cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Washington State (United States)
Squamous cell carcinoma
Ever exposed to PCE (probable exposure)
6.4 (0.6, 68.9)
2
Cumulative PCE exposure (possible exposure)
1-29 ppm-yr
30+ ppm-yr
11.9(1.1, 124.0)
2
0
Adenocarcinoma
Ever exposed to PCE (Probable exposure)
0.9(0.1, 10.0)
1
Reference
Vaughan et al. (1997)
Esophageal cancer cases (404 cases), 1983 -1987, 20-74 yr, 724
population controls, in-person interview, occupational title and JEM for
PCE, blinded exposure assessment, OR adjusted for age, sex, education,
study period, alcohol consumption and cigarette smoking
VO
to
Tor Boice et al. (1999). relative risks for employment duration from Poisson regression with internal referents of factory workers not exposed to any solvent and
with adjustment for date of birth, date first employed, date of finishing employment, race and sex.
bln Lynge et al. (2006). odds ratio from logistic regression adjusted for country, sex, age, calendar period at time of diagnosis.
JEM = job-exposure matrix, PCE = tetrachloroethylene.
-------
Establishment of an exposure or concentration-response relationship can add to the
weight of evidence for identifying cancer hazard, but only limited data pertaining to exposure-
response relationships for esophageal cancer and tetrachloroethylene exposure are available.
Five studies reported risk by exposure categories using exposure duration (Calvert et al., 2011;
Lynge et al., 2006; Boice et al., 1999) or a semiquantitative or quantitative surrogate (Blair et al.,
2003: Vaughan et al.. 1997). However, Boice et al. (1999) and Vaughan et al. (1997) were based
on relatively few observed cases, with <5 cases in individual exposure categories, greatly
limiting the usefulness of these exposure-response examinations. Boice et al. (1999) presented a
formal statistical test of linear trend (p = 0.07) for exposure duration and esophageal cancer
deaths among workers with routine or intermittent exposure; three of the 10 esophageal cancer
deaths in this group had exposure durations 5 years or longer (RR: 0.91, 95% CI: 0.13, 1.60, with
an internal comparison group of factory workers not exposed to any solvents as the referent).
This analysis included subjects whose exposure was infrequent and likely of lesser certainty than
subjects identified as having routine exposure. The overall SMR for any tetrachloroethylene
exposure in this study was 1.47 (95% CL0.54, 3.21). Both exposed cases in Vaughan et al.
(1997) were identified with lower cumulative exposure, 1-29 ppm-years (OR: 11.9, 95% 1.1,
124.0) compared to no cases with 30+ ppm-years. Effect estimates in one of the two larger
studies that examined exposure duration was not suggestive of a trend (Lynge et al., 2006) (refer
to Table 4-23). However, all 16 exposed esophageal deaths in Calvert et al. (2011) had >20
years since first employment, with effect estimates of 2.16 (95% CI: 0.85, 4.54) and 4.78 (95%
CI: 2.68, 7.91) for <5 years and >5 years exposure duration, respectively. Sixteen of the 26
esophageal cancer deaths in Blair et al. (2003) had medium-to-high cumulative exposure to dry-
cleaning solvents with an effect estimate of 2.2 (95% CI: 1.2, 3.5).
Only Vaughan et al. (1997) directly evaluated possible effects due to smoking or alcohol,
which are risk factors for the squamous cell histologic type of esophageal cancer; all other
studies lacked control for these potential confounders. Both Calvert et al. (2011) and Blair et al.
(2003) noted that the magnitude of the risks for esophageal cancer was greater than could be
explained by smoking alone; any smoking effect was estimated to contribute to no more than a
20% increase in risk. This suggests a further contribution from another risk factor, such as
occupational exposure. The incidence of esophageal cancer is generally higher for non-
Caucasian males than for Caucasian males (Brown etal., 2001; Blot and McLaughlin, 1999). In
contrast, Calvert et al. (2011) observed similar SMRs for esophageal cancer across all race-sex
groupings (supplementary table at http://www.cdc.gov/niosh/dc-mort.html), suggesting the
contribution of another factor such as occupational exposure. However, the inability to adjust for
potential effects of alcohol use in cohort studies is an uncertainty.
4-193
-------
In conclusion, the SIR in the only large cohort study (n = 95 cases), a study using broad
exposure categories, was 1.18 (95% CI: 0.96, 1.46) (Pukkala et al., 2009). The point estimates of
the association in seven of eight smaller studies, four studies with specific exposure assessments
and four other studies with less precise assessments, were between 1.16 and 2.44 (Calvert et al.,
2011: Selden and Ahlborg, 2011: Pukkala et al.. 2009: Sung et al.. 2007: Blair etal.. 2003:
Travier et al., 2002: Boiceetal., 1999: Lynge and Thygesen, 1990). Two small case-control
studies with relatively high quality exposure-assessment approaches, Lynge et al. (2006) and
Vaughan et al. (1997) reported an odds ratio of 0.76 (95% CI: 0.34, 1.69) and of 6.4 (95% CI:
0.6, 68.9), respectively. Some uncertainties in these estimates arise from the lack of job title
information for 25% of the cases and 19% of the controls, and the variability in the results from
the sensitivity analysis using different assumptions regarding the correct classification of
individuals in this group in Lynge et al. (2006) and the small numbers of exposed cases in
Vaughan et al. (1997). One of the two larger studies examining exposure-response suggested a
positive relationship, with SMRs of 2.16 (95% CI: 0.85, 4.54) and 4.78 (95% CI: 2.68, 7.91) for
durations of <5 years and >5 years, respectively (Calvert et al., 2011). None of the cohort
studies can exclude possible confounding from alcohol and smoking—risk factors for squamous
cell carcinoma of the esophagus. Based on smoking rates in blue-collar workers, the twofold
risk estimate reported in Calvert et al. (2011) and Blair et al. (2003) was higher than that
attributable to smoking.
4.5. LUNG AND RESPIRATORY CANCER
Nineteen epidemiologic studies reporting data on lung cancer and tetrachloroethylene
exposure were identified. This set of studies includes 12 cohort or nested case-control studies
within a cohort (Calvert et al.. 2011: Selden and Ahlborg, 2011: Pukkala et al.. 2009: Sung et al..
2007: Ji et al.. 2005b: Blair etal.. 2003: Chang et al.. 2003: Travier et al.. 2002: Andersen et al..
1999: Boice et al., 1999: Anttila et al., 1995: Lynge and Thygesen, 1990), 6 case-control studies
of occupational exposures (Consonni et al., 2010: MacArthur et al., 2009: Richiardi et al., 2004:
Pohlabeln et al., 2000: Brownson et al., 1993: Siemiatycki, 1991), and one case-control study of
residential exposure through contaminated drinking water (Pauluetal., 1999). Some of these
studies represent overlapping populations. For example, Travier et al. (2002) and Lynge and
Thygsen (1990) used occupational data from Sweden and Denmark, respectively; Andersen et al.
(1999) included Denmark, Finland, and Norway in addition to Sweden, and Pukkala et al. (2009)
added Iceland to the study population. Additionally, nonsmoking cases in Richiardi et al. (2004),
whose lung cancer cases included both smokers and nonsmokers, were included in the
International Agency for Research on Cancer (IARC) multicenter study of lung cancer among
nonsmokers (Pohlabeln et al., 2000). These studies represent the core studies evaluated by EPA,
4-194
-------
as described in more detail below. One other cohort study included information on
tetrachloroethylene but did not report risk estimates for lung cancer (Radican et al., 2008). Also,
one other lung cancer case-control study did not identify any cases as a dry cleaner or launderer
(Zeka et al., 2006) and was not evaluated further. Appendix B reviews the design, exposure-
assessment approach, and statistical methodology for each study. Most studies were of the
inhalation route, of occupational exposure, and unable to quantify tetrachloroethylene exposure.
4.5.1. Consideration of Exposure-Assessment Methodology
Most of these studies examine occupational titles such as dry cleaner, launderer, and
presser as surrogates for tetrachloroethylene, given its widespread use from 1960 onward in the
United States and Europe (Calvert etal., 2011: Selden and Ahlborg, 2011: Consonni et al.. 2010:
MacArthur et al.. 2009: Pukkala et al.. 2009: Zeka et al.. 2006: Ji et al.. 2005a: Ji and Hemminki,
2005a: Ji et al.. 2005b: Ji and Hemminkl 2005b, c; Richiardi et al.. 2004: Blair et al.. 2003:
Travier et al.. 2002: Pohlabeln et al.. 2000: Andersen et al.. 1999: Brownson et al.. 1993:
Siemiatycki, 1991: Lynge and Thygesen, 1990). Seven studies conducted in Nordic countries
are based on either the entire Swedish population or on combined populations of several Nordic
countries; the strengths of these studies are their use of job titles as recorded in census databases
and ascertainment of cancer incidence using national cancer registries (Selden and Ahlborg,
2011: Pukkala et al.. 2009: Lynge et al.. 2006: Ji et al.. 2005a: Ji and Hemminkl 2005a: Ji et al..
2005b: Ji and Hemminki, 2005b, c; Travier et al., 2002: Andersen et al., 1999: Lynge and
Thygesen, 1990). Studies examining mortality among U.S. dry-cleaner and laundry workers
(Calvert et al., 2011: Blair et al., 2003) are of smaller cohorts than the Nordic studies, with fewer
observed lung cancer events.
The exposure surrogate in studies of dry-cleaners and laundry workers is a broad
category containing jobs of differing potential for tetrachloroethylene exposure. Thus, these
studies have a greater potential for exposure misclassification bias compared to studies with
exposure potential to tetrachloroethylene assigned by exposure matrix approaches applied to
individual subjects. Three studies used additional information pertaining to work environment to
refine the exposure classification. Selden and Ahlborg (2011) obtained information about the
dry-cleaning establishment (e.g., washing techniques, chemicals used, number of employees, and
work history of individual employees) in a questionnaire sent to businesses in Sweden in the
1980s. Blair et al. (2003) used an exposure metric for semi quantitative cumulative exposure
within the dry-cleaning setting. The study by Calvert et al. (2011) of unionized dry cleaners in
the United States included an analysis of subjects who worked for one or more years before 1960
in a shop known to use tetrachloroethylene as the primary solvent (Calvert et al., 2011: Ruder et
al., 2001, 1994). The cohort was stratified into two groups based on the level of certainty that
4-195
-------
the worker was employed only in facilities using tetrachloroethylene as the primary solvent;
tetrachloroethylene-only and tetrachloroethylene plus. Twenty-six of the 77 observed lung
cancer deaths were among this subset (n = 618) of the study subjects.
Four other cohorts with potential tetrachloroethylene exposure in manufacturing settings
have been examined. These studies include aerospace workers in the United States (Boice et al.,
1999), workers, primarily in the metal industry, in Finland (Anttila et al., 1995) and electronic
factory workers in Taiwan (Sung et al., 2007; Chang et al., 2005). Boice et al. (1999) used an
exposure assessment based on a job-exposure matrix, and Anttila et al. (1995) used biological
monitoring of tetrachloroethylene in blood to assign potential tetrachloroethylene exposure to
individual subjects. In contrast, the exposures in the Taiwan studies included multiple solvents,
and tetrachloroethylene exposure was not linked to individual workers. These cohorts also
included white-collar workers, who had an expected lower potential for exposure (Sung et al.,
2007: Chang et al., 2003).
Paulu et al. (1999) is a case-control study that examined residential proximity to drinking
water sources contaminated with tetrachloroethylene in Cape Cod, MA. This study used an
exposure model incorporating leaching and characteristics of the community water distribution
system to assign a household relative dose of tetrachloroethylene.
In summary, with respect to exposure-assessment methodologies, six studies with lung
cancer data assigned tetrachloroethylene exposure to individuals within the study using
biological monitoring data (Anttila et al., 1995), a job exposure matrix (Boice et al., 1999), a
semiquantitative metric (Blair et al., 2003), an exposure model (Paulu et al., 1999), additional
details pertaining to work environment (Selden and Ahlborg, 2011), or a classification of the
cohort by certainty of tetrachloroethylene exposure (Calvert et al., 2011). The relative specificity
of these exposure-assessment approaches strengthens their ability to identify cancer hazards
compared to studies with broader and less sensitive exposure-assessment approaches. The least
sensitive exposure assessments are those using very broad definitions such as working in a plant
or factory (Sung et al., 2007: Chang et al., 2003).
4.5.2. Summary of Results
Lung cancer is a relatively common cancer, and six of the cohort studies of dry-cleaners
and laundry workers evaluated by EPA reported estimated relative risks based on 100 or more
deaths/incident cases (Selden and Ahlborg, 2011: Pukkala et al., 2009: Ji et al., 2005b: Blair et
al., 2003: Travier et al., 2002: Andersen et al., 1999): Pukkala et al. (2009) was the largest study,
with 965 incident lung cancers. Two other cohort studies, Lynge and Thygsen (1990) and
Calvert et al. (2011), observed 60 and 77 lung cancers, respectively. In contrast, the number of
exposed cases in the case-control studies ranged from 3 cases each of small cell and
4-196
-------
adenocarcinoma histological subtypes in MacArthur et al. (2009) to 30 (all histological types) in
Brownson et al. (1993). The three cohort studies with exposure assessment specific to
tetrachloroethylene observed 5 incident cancer cases, 46 lung cancer deaths, and 125 lung cancer
deaths in Anttila et al. (1995). Boice et al. (1999). and Blair et al. (2003). respectively. The
geographic-based case-control study of Paulu et al. (1999) observed 33 of the 326 lung cancer
cases living in a residence receiving tetrachloroethylene contaminated water, and only 5 of these
cases were identified as highly exposed.
The seven25 cohort studies with findings based on 50 or more events observed a
standardized incidence ratio estimate between 1.15 and 1.4 for the association between lung
cancer risk and ever having a job title of dry-cleaner or laundry worker, each with relatively tight
95% CIs (refer to Table 4-24). These estimates by study were 1.15 (95% CI: 1.02, 1.31) in
Travier et al. (2002). 1.2 (0.9, 1.5) in Lynge and Thygsen (1990). 1.26 (95% CI: 1.18, 1.34) in
Pukkala et al. (2009). 1.32 (1.07, 1.60) in Ji et al. (2005b). 1.32 (95% CI: 1.20, 1.45) in Selden
and Ahlborg (20_U), 1.31(1.04, 1.64) in Calvert et al. (2011). and 1.4(1.1. 1.6) in Blair et al.
(2003), respectively. Selden and Alhborg (2011) examined separately subjects working in a dry
cleaner using tetrachloroethylene (potential tetrachloroethylene exposure) and laundry workers,
subjects without potential tetrachloroethylene exposure. The standardized incidence ratios were
1.16 (95% CI: 0.89, 1.51) and 1.62 (95% CI: 1.15, 2.19) for dry cleaners and for laundry
workers, respectively.
In addition to the large cohort studies, evidence also comes from cohort and case-control
studies whose effect estimates are based on fewer observed events. Smaller studies that do not
also have a more sensitive or specific exposure metric carry lesser weight in the analysis. As
expected, the magnitude of the point estimate of the association reported in these studies is more
variable than in the larger studies: one study reported an odds ratio estimate below 1.0
(Siemiatycki, 1991), four studies reported a relative risk estimate between 1.0 and 1.3 (Consonni
etal.. 2010: MacArthur et al.. 2009: Boice etal.. 1999: Paulu etal.. 1999). three studies reported
relative risks between 1.8 and 2.0 (Pohlabeln et al., 2000; Anttila et al., 1995; Brownson et al.,
1993), and two studies reported odds ratios estimates over 2.0 (MacArthur et al., 2009; Richiardi
et al., 2004). Except for the estimate from Brownson et al. (1993) (OR: 1.8, 95% CI: 1.1, 3.0)
and McArthur et al. (2009) (small cell carcinoma, OR: 3.55, 95% CI: 1.13, 11.17), all of the 95%
CIs of these estimates overlap 1.0.
25 Andersen et al. (1999) is not included in this summary of the data from the individual studies because it was
updated and expanded in the analysis by Pukkala et al. (2009).
4-197
-------
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort studies
Biologically monitored workers
All subjects
1.92 (0.62, 4.48)
5
Aerospace workers (Lockheed)
Routine exposure to PCE
1.08 (0.79, 1.44)
46
Routine-Intermittent exposure duration to PCE
0
5yr
/>-value for linear trend
1.0a
1.15(0.80, 1.66)
1.09 (0.80, 1.48)
0.71 (0.49, 1.02)
p = 0.02
288
33
51
36
Electronic factory workers (Taiwan)
All Subjects
Males
Females
Females
0.97 (0.69, 1.33)
0.90 (0.48, 1.53)
1.01 (0.65, 1.49)
0.92 (0.67, 1.23)
38
13
25
46
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
Not reported
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females and 0.7
umol/L in males (median)], follow-up 1974-1992, external referents (SIR)
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after 1960,
worked at least 1 yr, follow-up 1960-1996, job exposure matrix without
quantitative estimate of PCE intensity, 1987-1988 8-h TWA PCE
concentration (atmospheric monitoring) 3 ppm [mean] and 9.5 ppm
[median], external reference for routine exposure (SMR) and internal
references (workers with no chemical exposures) for routine-intermittent
PCE exposure (RR)
Chang et al. (2003): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997, multiple solvents
exposure, does not identify PCE exposure to individual subjects, cancer
mortality, external referents (SMR) (Chang et al.. 2003):
63,982 females, follow-up 1979-2001, factory employment proxy for
exposure, multiple solvents exposures and PCE not identified to individual
subjects, cancer incidence, external referents, analyses lagged 10 yr (SIR)
(Sung et al.. 2007)
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever exposed
to mixed solvents, 85 1 ever-exposed to PCE), employed at least 1 yr from
1952 to 1956, follow-up 1973-2000, job exposure matrix (intensity),
internal referent (workers with no chemical exposures) (RR)
VO
oo
-------
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Dry -cleaner and laundry workers
All laundry worker and dry cleaners
Males
Females
1.19(1.07, 1.34)
1.24 (1.05, 1.46)
1.16(1.00, 1.35)
313
141
172
All subjects
1.4(1.1,1.6)
125
Semiquantitative exposure score
Little to no exposure
Medium to high exposure
1.0 (0.7, 1.4)
1.5(1.2,1.9)
34
78
Laundry workers and dry cleaners in 1960
Census
Males
Females
1.32 (1.20, 1.46)
1.36 (1.20, 1.54)
1.26 (1.07, 1.47)
403
247
156
Laundry workers and dry cleaners in both 1960 and 1970 Censuses
Males
Females
Not reported
Not reported
Laundry workers and dry cleaners in 1960, 1970, and 1980 Censuses
Males
Females
Not reported
Not reported
Reference
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or 1970
Census (Denmark, Finland, Norway), follow-up 1971-1987 or 1991,
PCE not identified to individual subjects, external referents (SIR)
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union members
(1945-1978), follow-up 1979-1993, semiquantitative cumulative
exposure surrogate to dry clean solvents, cancer mortality, external
referents (SMR)
Ji et al. (2005b);
9,255 Swedish men and 14,974 Swedish women employed in 1960
(men) or 1970 (women) as laundry worker or dry cleaner, follow-up
1961/1970-2000, PCE not identified to individual subjects, external
referent (SIR) and adjusted for age, period and socioeconomic status
VO
VO
-------
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer (continued)
Exposure group
All laundry worker and dry cleaners
Males
Females
Launderer and dry cleaner
Male
Female
All subjects
Exposure duration/time since 1st employment
<5 yr/<20 yr
<5 yr/>20 yr
>5 yr/<20 yr
>5 yr/>20 yr
PCE-only subjects
Dry-cleaners and laundry workers
PCE
Males
Females
Laundry
Males
Females
Relative risk
(95% CI)
1.2 (0.9, 1.5)
1.1 (0.8, 1.7)
0.3 (0.9, 1.8)
1.26(1.18, 1.34)
1.28(1.15, 1.42)
1.25(1.15, 1.35)
1.31 (1.04, 1.64)
0.63 (0.21, 1.44)
1.75(1.33,2.26)
1.27 (0.55, 2.50)
1.08(0.75, 1.51)
1.25 (0.82, 1.83)
1.32 (1.07, 1.60)
1.16(0.89 1.51)
1.30 (0.82, 1.94)
1.09(0.76, 1.51)
1.62(1.15,2.21)
1.60 (0.85, 2.74)
1.63(1.06,2.39)
No. obs.
events
60
28
32
965
353
612
77
4
32
6
26
26
100
58
23
35
13
26
Reference
Lynge and Thygsen (1990)
10,600 Danish men and women, 20-64 yr old, employed in 1970 as
laundry worker, dry cleaners and textile dye workers, follow-up
1070—1080 external referents (STR^
Pukkala et al. (2009)
Men and women participating in national census on or before 1990, 5
Nordic countries (Denmark, Finland, Iceland, Norway, Sweden), 30-64
census, external referents (SIR)
Calvert et al. (2011)
1,704 U.S. men and women dry-cleaning union member in CA, IL, MI,
NY follow-up 1940-2004 (618 subjects worked for one or more years
identified as PCE-only exposure), cancer mortality (SMR)
Selden and Ahlborg (2011)
9,440 Swedish men (n = 2,810) and women (n = 9,440) in 461 washing
and dry-cleaning establishments, identified by employer in mid-1980s,
company serf-reported information on PCE usage — PCE (dry cleaners
and laundries with a proportion of PCE dry cleaning), laundry (no PCE
use), and other (mixed exposures to PCE, CFCs, TCE, etc.), external
referents (SIR)
to
o
o
-------
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
All subjects, 1960 or 1970 Census in laundry
and dry cleaner or related occupation and
industry
All subjects in 1960 and 1970 in laundry and dry
cleaner occupation and industry
1.15(1.02, 1.31)
1.20 (0.84, 1.72)
248
30
Reference
Travier et al. (2002)
Swedish men and women identified as laundry worker, dry cleaner, or
presser (occupational title), in the laundry, ironing, or dyeing industry or
related industry in 1960 or 1970 (543,036 person-years); or, as laundry
worker, dry cleaner, or presser (occupational and job title) (46,933
person-years) in both censuses, follow-up 1971-1989, external referents
(SIR)
Case-Control Studies
Missouri, United States
Dry-cleaning industry
All subjects
Lifetime nonsmokers
Former smokers
1.8(1.1,3.0)
2.1(1.2,3.7)
1.1 (not reported)
30
23
7
Exposure duration
<1.125yr
>1.125yr
0.8 0.2, 1.7)
2.9(1.5,5.4)
Not
reported
Not
reported
Lombardy, Italy (EAGLE study)
Dry-cleaning industry
Males
Females
Not reported
1.26 (0.46, 3.41)
o
J
12
Brownson et al. (1993)
429 female primary lung cancer cases, 30-84 yr, 1986-1991, never
smokers or ex-smokers (>15 yr prior to diagnosis), identified from
Missouri Cancer Registry, 1,021 female population controls matched on
age, identified from state driver's licenses (<65 yr) or HFCA roles
(65-84 yr), telephone and in-person interview using questionnaire, dry
cleaner occupation or job title exposure surrogate, OR adjusted for age,
smoking, and history of previous lung disease
Consonni et al. (2010)
1,943 histologically or cytologically confirmed hospital lung cancer
cases in men and women, 35-79 yr, 2002-2005, and 2,1 16 population
controls matched on residence, sex, and age, in-person and self-
administered questionnaire, job title and industry coded to ISCO and
ISIC surrogate for exposure, dry-cleaning industry identified a priori
suspected lung hazard, OR adjusted for residential area, age, smoking
and number of jobs held
to
o
-------
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
British Columbia, Canada
Dry cleaner and launderer occupation
Squamous cell carcinoma
Adenocarcinoma
Small cell
Large cell
1.25 (0.47, 3.35)
1.28 (0.44, 3.70)
3.55(1.13, 11.17)
4
3
3
0
International Lung Cancer Study (IARC Study) (France, Germany, Italy, Portugal,
Spain, Sweden, United Kingdom)
All Centers
Dry-cleaning industry
Males
Females
Not reported
1.83 (0.98, 3.40)
1
19
Turin and Veneto Regions, Italy
Dry Cleaners and Launderers
Males
Females
1.6 (0.2, 12)
2.1(0.8,5.6)
3
9
Reference
MacArthur et al. (2009)
2,998 male histologically confirmed lung cancer cases, >20 yr,
1983-1990, 10,233 all-other sites-cancer controls matched on age and
diagnosis year, identified from British Columbia Cancer Registry, self-
administered questionnaire, job title and industry coded to Canadian
SOC and Canadian SIC as exposure surrogate, OR adjusted for smoking
duration, respondent status, and education
Pohlabeln et al. (2000)
660 nonsmoking lung cancer cases, <75 yr, 1988-1994, 1,542
nonsmoking controls, 12 study centers in 7 countries, various sources of
nonsmoking controls (community based in 6 centers, hospital-based in 1
center, both sources in 5 centers), hospital controls with diseases not
related to smoking, in-person interview, job title and industry coded to
ISCO and ISIC exposure surrogate, dry-cleaning industry identified a
priori suspected lung hazard, OR adjusted for age and center
Richiardi et al. (2004)
1,132 histologically or cytologically confirmed lung cancer cases, <75
yr, 1990-1991 or 1991-1992, population controls identified from
population registries and matched on sex and age, in-person interview,
job title and industry >6 mo duration coded to ISCO and ISIC exposure
surrogate, dry-cleaning industry identified a priori suspected lung
hazard, OR adjusted for age, study area, cigarette smoking, other tobacco
product use, and number of jobs. Cases and controls included in
international multicenter study of Pohlaban et al. (2000)
to
o
to
-------
Table 4-24. Summary of human studies on tetrachloroethylene exposure and lung cancer (continued)
Relative risk No. obs.
Exposure group (95% CI) events
Montreal, Canada
Launderers and dry cleaners
Any exposure
Substantial exposure
0.8 (0.5, 1.5)b
0.6 (0.2, 1.4)b
12
5
International Lung Cancer Study (IARC Study) (Czech Republic, Hungary, Poland,
Romania, Russia, Slovakia, United Kingdom)
Launderers and dry cleaners
Males
Females
Not reported
Not reported
0
0
Reference
Siemiatycki (1991)
857 histologically confirmed lung cell carcinoma cancer, 1979-1985,
35-70 yr, 533 population control group and 1,900 cancer control group,
in-person interviews, occupational title, OR adjusted age, family income,
ethnic origin, respondent status, cigarette smoking, and alcohol
consumption, 90% CI
Zeka et al. (2006)
223 hospital lung cancer cases, 20-74 yr, 1998-2002, lifetime
nonsmokers, identified from 16 hospitals or clinics in 7 countries,
hospital (14 centers) or population controls (2 centers) frequency-
matched on sex and age, in-person interview, industry and job title
exposure surrogate, dry-cleaning industry identified a priori suspected
lung hazard, OR adjusted for age, sex, and study center, with ETS
exposure included as additional covariate in some analyses
Geographic-Based Studies
Cape Cod, MA
Overall PCE exposure
PCE ROD >90th percentile
1.1 (0.7, 1.7)
2.7(1.0, 11.7)
33
5
Paulu et al. (1999)
326 histologically confirmed lung cancer cases in males and females,
1983-1986, MA Cancer Registry, 2,236 population controls identified
by random digit dialing, vital records for deceased controls, and HCFA
records if >65 yr, telephone interview, algorithm of Webler and Brown
(1993) to estimate mass of PCE in drinking water entering residence was
surrogate exposure metric, OR adjusted for age of diagnosis or index
year, vital status at interview, sex, occupation exposure to PCE, other
solvents, and exposures associated with lung cancer, usual number of
cigarettes smoked, history of cigar/pipe use, living with a smoker
to
o
aReferent.
bln Siemiatycki (1991). 90% CI.
CFC = chloroflourocarbon, HCFA = Health Care Financing Administration, ISCO = International Standard Classification of Occupation, ISIC = International
Standard Industry Classification, JEM = job-exposure-matrix, PCE = tetrachloroethylene, RDD = relative delivered dose, TCE = trichloroethylene, TWA = time-
weighted-average.
-------
Five occupational studies were identified as having a relatively strong exposure-
assessment methodology. The results from four of these studies provide support for an increased
risk in the dry-cleaning cohorts with a relative risk of 1.4 (95% CI: 1.1, 1.6) in Blair et al. (2003),
1.31 (95% CI: 1.04, 1.64) in Calvert et al. (201IX and in other settings, a relative risk of 1.08
(95% CI: 0.79, 1.44) in Boice et al. (1999) and 1.92 (95% CI: 0.62, 4.48) in Anttila et al. (1994).
In contrast, Selden and Ahlborg (2011) reported similar, but slightly higher, relative risks for
laundry workers compared with dry-cleaning workers in their study. Two studies of an
electronics factory using relatively weak exposure-assessment approaches (i.e., no classification
of individuals within the study) observed relative risks or SMRs of 0.97 (95% CI: 0.69, 1.33)
(Chang et al.. 2003) and 0.92 (95% CI: 0.67, 1.23) (Sung et al.. 2008).
Establishment of an exposure or concentration-response relationship can add to the
weight of evidence for identifying a cancer hazard, but only limited data pertaining to exposure-
response relationships for lung cancer and tetrachloroethylene exposure are available. Seven
studies presented risk estimates for increasing exposure categories: three studies using exposure
duration as a proxy (Calvert et al., 2011; Travier et al., 2002; Boice etal., 1999) and four studies
with a semiquantitative exposure surrogate (Blair et al., 2003; Paulu et al., 1999; Brownson et
al., 1993; Siemiatycki, 1991). Boice et al. (1999) was the only study to present a formal
statistical test for trend and reported a statistically significant decreasing trend between lung
cancer risk estimates and duration among subjects with routine and intermittent
tetrachloroethylene exposure (p = 0.02). A monotonic increasing trend in risk estimates and
exposure surrogate was apparent in four studies (Blair et al., 2003; Travier et al., 2002; Paulu et
al.. 1999: Brownson et al.. 1993).
A known risk factor for lung cancer is cigarette smoking (NTP, 2005). Subjects in both
Brownson et al. (1993) and Pohlabeln et al. (2000) were either lifetime nonsmokers or ex-
smokers who had terminated smoking 15 years before cancer diagnosis, reducing any potential
role of confounding from smoking. Furthermore, in the case of Pohlabeln et al. (2000), the
inclusion of occasional smoking (ever smoked occasionally but fewer than 400 cigarettes total)
and exposure to tobacco smoking as possible confounders did not significantly affect the odds
ratio estimate and were not included in the final model. Statistical analyses in all other case-
control studies controlled for cigarette smoking (Consonni et al., 2010; MacArthur et al., 2009;
Richiardi et al., 2004; Paulu etal., 1999; Siemiatycki, 1991). However, both Brownson et al.
(1993) and MacArthur et al. (2009) had a high percentage of surrogate or proxy respondents, 58
and 27%, respectively. Proxy respondents may have motivations to report or not report specific
exposures leading to differential information bias that could result in the relative risk estimate
towards or away from the null depending on whether controls were more or less likely to recall
or report such exposure than cases (Pearce et al., 2007).
4-204
-------
Direct examination of possible confounders is less common or feasible in cohort studies
relying on company-supplied or census work history data compared to case-control studies
where information is obtained from study subjects or their proxies. In cohort studies, however,
use of internal controls rather than an external referent group (e.g., national mortality rates) can
minimize effects of potential confounding due to smoking or socioeconomic status, because
exposed and referent subjects are drawn from the same target population. Only one of the
available cohort studies included an analysis using internal controls and reported a decreasing
trend between lung cancer and tetrachloroethylene exposure duration,/* = 0.02 (Boice et al.,
1999). Blair et al. (2003) considered the potential effect of differences in the prevalence of
smoking in their study of laundry and dry-cleaning workers. Surveys from 1970 to 1990
indicated that smoking rates among dry cleaners were 5-10% higher than the general population.
With this level of difference, confounding from smoking is unlikely to result in a relative risk
greater than 1.2 but may explain most of the observed 40% excess in lung cancer. The
magnitude of relative risk estimates in cohort studies of dry-cleaners and laundry workers
(Calvert et aL 2011: Selden and Ahlborg. 2011: Pukkala et al.. 2009: Ji et al.. 2005b: Travier et
al., 2002: Lynge and Thygesen, 1990) is similar to or less than that of Blair et al. (2003) and
suggests smoking may contribute to the observed association.
In conclusion, the results from seven large cohort studies of dry cleaners are consistent
with an elevated lung cancer risk of 10-40%. Similar results were observed in four of the five
occupational studies that were identified as having a relatively strong exposure-assessment
methodology (Calvert et aL 2011: Blair et al.. 2003: Boice etal.. 1999: Anttila et al.. 1994).
However, Selden and Ahlborg (2011) observed similar, but slightly higher, relative risks for
laundry workers compared with dry-cleaning workers in their study. These studies were unable
to control for potential confounding from cigarette smoking; however, and the magnitude of the
association in these studies is consistent with that expected, assuming the prevalence of smoking
among dry-cleaners and laundry workers was slightly higher (e.g., 10% higher) than among the
general population. Features of the selection of study participants and study analysis in the
available case-control studies reduce the potential for confounding by smoking, however. Two
case-control studies were limited to either nonsmokers or ex-smokers who had ceased smoking
15 years before diagnosis (Pohlabeln et al., 2000: Brownson et al., 1993). Both of these studies
indicate an approximate twofold increased risk with a history of work in the dry-cleaning
industry [OR: 1.8, 95% CI: 1.1, 3.0, in Brownson et al. (1993): and OR: 1.83, 95% CI: 0.98,
3.40, among women in Pohlabeln et al. (2000)]. The other case-control studies adjusted for
smoking history, and the results for these (somewhat smaller studies) are similar to the
previously cited estimates. Among the studies that evaluated exposure-response gradients, the
4-205
-------
evidence for a trend in risk estimates was mixed (Calvert et al., 2011; Blair etal., 2003; Travier
et al.. 2002: Boiceetal.. 1999: Pauluetal.. 1999: Brownson et al.. 1993).
4.6. IMMUNOTOXICITY, HEMATOLOGIC TOXICITY, AND CANCERS OF THE
IMMUNE SYSTEM
Chemical exposures may result in a variety of adverse immune-related effects, including
immunosuppression (decreased host resistance), autoimmunity, and allergy-hypersensitivity, and
may result in specific diseases such as infections, systemic or organ-specific autoimmune
diseases, or asthma. Measures of immune function (e.g., T-cell counts, immunoglobulin [Ig] E
levels, specific autoantibodies, cytokine levels) may provide evidence of an altered immune
response that precedes the development of clinically expressed diseases. This section discusses
effects relating to immunotoxicity and hematotoxicity. It also discusses evidence pertaining to
tetrachloroethylene in relation to lymphoid tissue cancers, including childhood leukemia.
4.6.1. Human Studies
4.6.1.1. Noncancer Immune and Hematologic Effects
Adverse effects on the immune system resultingincluded the following:
immunosuppression (host resistance), immunostimulation, autoimmunity, and allergy-
hypersensitivity. Various immunologic measurements (e.g., T-cell counts, immunoglobulin [Ig]
E levels, specific autoantibodies) may provide evidence of an altered immune response that may
subsequently be related to risk of clinically expressed diseases such as infections, asthma, or
systemic lupus erythematosus. Tetrachloroethylene exposure via air or water may result in
immune-mediated organ-specific or systemic effects, as described in a case report of
hypersensitivity pneumonitis in a 42-year-old female dry-cleaner worker (Tanios et al., 2004).
Another case report described severe fatigue, weight loss, myalgia, arthralgia, cardiac
arrhythmia, decreased T-cell count, high-titer (1:160) antinuclear antibodies, and neurological
symptoms that were linked to chemical sensitivity to tetrachloroethylene in a municipal water
supply (ReaetaL 1991).
4.6.1.1.1. Immunologic and hematologic parameters
Byers et al. (1988) provide data pertaining to immune function from 23 family members
of leukemia patients in Woburn, Massachusetts. In 1979, testing of the wells in this town
revealed that the water in two of the wells was contaminated with a number of solvents,
including tetrachloroethylene (21 ppb) and trichloroethylene (267 ppb) [as cited in Lagakos et al.
(1986)]. These wells had been in operation from 1964 to 1979. Byers et al. collected serum
samples in May and June of 1984 and in November of 1985. They determined the total
4-206
-------
lymphocyte counts and lymphocyte subpopulations (CDS, CD4, CDS), and the CD4:CD8 ratio in
these samples, and in samples from a combined control group of 30 laboratory workers and
40 residents of Boston selected through a randomized probability area sampling process. The
study authors also assessed the presence of autoantibodies (antismooth muscle, antiovarian,
antinuclear, antithyroglobulin, and antimicrosomal antibodies) in the family member samples
and compared the results with laboratory reference values. The lymphocyte subpopulations were
higher, and the CD4:CD8 ratio was lower in the Woburn family members compared to the
controls in both of the samples taken in 1984. In the 1985 samples, however, the subpopulation
levels had decreased and the CD4:CD8 ratio had increased; the values were no longer
statistically different from the controls. None of the family member serum samples had
antithyroglobulin or antimicrosomal antibodies, but 10 family member serum samples (43%) had
antinuclear antibodies (compared to <5% expected based on the reference value). Because the
initial blood sample was taken in 1984, and because of the considerable mixture of exposures
that occurred in this setting, it is not possible to determine the patterns at a time nearer to the
time of the exposure, or to infer the exact role of tetrachloroethylene in alterations of the
immunologic parameters.
Other studies have examined immunological parameters in dry-cleaning workers in the
Czech Republic (Andrys et al.. 1997) and in Egypt (Emara et al.. 2010) (refer to Table 4-25).
Andrys et al. (1997) included 21 dry-cleaning workers (20 women) and 16 office workers in the
dry-cleaning plant (14 women) and compared them to reference values based on samples from
blood donors and "healthy persons in the same region" (n = 14-311, depending on the test). The
mean ages of the exposed workers and office controls were 45.7 years and 31.9 years,
respectively; no information was provided on the age or sex distribution of the reference
controls. The tests included measures of immunoglobulin (Ig) A, IgG, IgM, and IgE levels,
complement (C3 and C4) levels, phagocyte activity, C-reactive protein, a-macroglobulin,
T-lymphocytes, and a blast transformation test. Several differences were observed between the
exposed workers and the office workers (e.g., higher levels of serum complement C3 and C4,
and of salivary IgA in the exposed), and between the exposed workers and the reference controls
(reduced T-lymphocytes, higher phagocytic activity, higher C3 levels in exposed). However,
there were also many differences noted between the office workers and the reference group
(including reduced T-lymphocytes in office workers). The lack of information about the
reference group adds to the difficulty in interpreting these results.
4-207
-------
Table 4-25. Immune and hematological parameters in studies of dry-
cleaning workers or tetrachloroethylene exposure in children
Study details
Adults
Czech Republic, period not
reported. 21 dry-cleaning workers
(20 women; mean age 45.7 yr); 16
office workers in the dry-cleaning
plant (14 women; mean age
3 1.9 yr); reference values based on
samples from blood donors and
"healthy persons in the same
region" (n = 14-3 11, depending on
the test)
Egypt, period not reported. 40 adult
men (ages 20-38 yr), dry-cleaning
workers; 40 healthy male controls
(matched by age and smoking
history): n - 20 in 4 groups
(controls, never smoked; controls,
smoked; PCE-exposed, never
smoked; PCE exposed, smoked).
Amount and duration of smoking
similar among exposed and
nonexposed. Mean years of PCE
exposure 7 yr. Blood PCE levels in
exposed: 1,685 ug/L
Germany, 1995-1996. 121
children (ages 36 mo), selected
based on high risk profile for
allergic diseases, blood sample and
indoor air sampling (child's
bedroom) of 26 volatile organic
chemicals (4 wk around age 36 mo)
Germany 1997-1999. 85
newborns, cord blood and indoor air
sampling (child's bedroom) of 28
volatile organic chemicals (4 wk
immediately after birth)
Measure(s)
Ig (IgA, IgG, IgM) levels,
complement (C3 and C4)
levels, phagocyte activity,
C-reactive protein,
a-macroglobulin,
T-lymphocytes
RBC counts
WBC counts
lymphocyte subpopulations
(CD3+, CD4+, CD8+,
CD3+CD16CD56+, CD19+
cells)
Ig levels (IgA, IgE, IgG,
IgM)
serum and lymphocytic
interferon-y and
interleukin-4
IgE levels
CD3 T-cell subpopulations
from cord blood
Results
Higher levels serum complement
C3 and C4, salivary IgA in
exposed workers compared with
office workers. Reduced
T-lymphocytes, higher phagocytic
activity, higher C3 levels in
exposed workers compared with
reference controls. Reduced
T-lymphocytes in office workers
compared with reference controls
RBC counts and hemoglobin levels
decreased with exposure. No
difference inMCV, MCH, or
MCHC
Total white cell and lymphocyte
counts increased with exposure.
No difference in eosinophils,
monocytes, or platelets
CD4+ and CD8+ T-lymphocytes
and CD3+CD16CD56+ NK cells
increased with exposure
IgE increased with exposure. No
difference in IgA, IgG, or IgM
levels across groups
Interleukin-4 levels increased with
exposure. No differences with
interferon-y
no association between PCE
measures and total IgE or IgE-
specific allergen antibodies
Decreased interferon-y cells
No association with interleukin-4,
interleukin-2, or tumor necrosis
factor-a
Authors
Andrys et
al. (1997)
Emara et al.
(2010)
Lehmann et
al. (2001)
Lehmann et
al. (2002)
Ig = immunoglobulin; MCV = mean corpuscular volume; MCH = mean corpuscular hemoglobin; MCHC = mean
corpuscular hemoglobin concentration; RBC = red blood cells; WBC = white blood cells.
4-208
-------
Emara et al. (2010) examined immunological and hematologic parameters in 80 men,
ages 20-38 years, in Tanta City, Egypt. Forty men were dry-cleaning workers, with a mean
duration of work of 7 years. They were matched by age and smoking history to 40 healthy
controls from the same area. The study, thus, included four groups, each with 20 men: controls
who had never smoked; controls who were smokers; tetrachloroethylene-exposed workers who
had never smoked; and tetrachloroethylene-exposed workers who were smokers. The amount
smoked and duration of smoking were similar in the exposed and nonexposed groups (mean:
17.9 and 17.5 cigarettes per day, respectively; mean: 4.5 and 5.0 years smoking, respectively).
Tetrachloroethylene levels were measured at five sites within each worksite, and blood levels of
tetrachloroethylene were also measured in all study participants. The mean air level was <140
ppm tetrachloroethylene, and mean blood levels were 1,681 and 1,696 |ig/L among nonsmoking
and smoking workers (compared with 0.11 |ig/L in each of the control groups), respectively.
Blood samples were obtained from each study participant for measurement of a differential
blood cell count, serum Ig levels (IgA, IgE, IgG, and IgM), and interferon-y and interleukin-4
levels in serum and lymphocytes.
Red blood cell counts and hemoglobin levels were decreased with exposure (p < 0.05 for
smoking-stratified comparisons), but there was no difference in mean corpuscular volume, mean
corpuscular hemoglobin, or mean corpuscular hemoglobin concentration across groups (Emara et
al., 2010). In contrast, total white cell counts and total lymphocytes increased significantly with
exposure (p < 0.05 for smoking-stratified comparisons). There was no difference in eosinophils,
monocytes, or platelets counts across groups. Neutrophil counts were increased in smokers
compared with nonsmokers but did not differ by tetrachloroethylene-exposure group. CD4+ and
CD8+ T-lymphocytes and natural killer (CD3+CD16CD56+) cells were increased in smoking and
nonsmoking exposed workers (p < 0.05), but CD3+ T-lymphocytes were only increased in the
exposed smoking group. This study demonstrated statistically significant effects of
tetrachloroethylene exposure on hematological parameters including decreased red blood cell
counts, increased white blood cells counts, total lymphocytes, and specific T- and NK cell
subpopulations.
Th2 cytokines (e.g., interleukin-4) stimulate production of IgE, and Thl cytokines (e.g.,
interferon-y) act to inhibit IgE production. The results from Emara et al. (2010) indicate that
tetrachloroethylene exposure results in an increase in serum and lymphocytic interleukin-4
levels, as well as increased IgE levels (p < 0.05 for smoking-stratified comparisons). As
determined from Figure 5 of Emara et al. (2010), the mean levels were approximately 90, 160,
170, and 195 ITJ/mL in nonexposed nonsmokers, nonexposed smokers, exposed nonsmokers, and
4-209
-------
exposed smokers, respectively (p < 0.05 for smoking-stratified comparisons). No difference was
observed in IgA, IgG, or IgM levels across groups.
Two studies examined variation in cytokines and in IgE levels in children (Lehmann et
al., 2002; Lehmann et al., 2001) (refer to Table 4-25). Lehmann et al. (2001) examined IgE
levels and cytokine-producing cells (interferon-y, tumor necrosis factor-a, and interleukin-4) in
relation to indoor levels of volatile organic compounds among children (age 36 months) selected
from a birth cohort study in Leipzip, Germany. The hypothesis underlying this work is that a
shift in Thl to Th2 cytokine profile is a risk factor for IgE-mediated allergic disease in children
(Tang et al., 1994; Warner et al., 1994). Enrollment into the birth cohort occurred between 1995
and 1996. The children in this allergy study represent a higher-risk group for development of
allergic disease, with eligibility criteria that were based on low birth weight (between 1,500 and
2,500 g) or cord blood IgE greater than 0.9 kU/L with a double positive family history of atopy.
These eligibility criteria were met by 429 children; 200 of these children participated in the
allergy study described below, but complete data (IgE and volatile organic compound
measurements) were available for only 121 of the study participants.
Lehmann et al. (2001) measured 26 volatile organic compounds via passive indoor
sampling in the child's bedroom for a period of 4 weeks around the age of 36 months. The
highest exposures were observed for limonene (median: 19.1 ug/m3), a-pinene (median: 16.3
ug/m3), and toluene (median: 13.3 ug/m3). The median exposure of tetrachloroethylene was 2.5
ug/m3 (0.87 ug/m3 and 5.1 ug/m3 for the 25th and 75th percentiles, respectively). The only strong
correlation (r > 0.3) between tetrachloroethylene and the other volatile organic compounds
measured was a correlation of 0.72 with trichloroethylene. Blood samples were taken at the
36-month-study examination and were used to measure the total IgE and specific IgE antibodies
directed to egg white, milk, indoor allergens (house dust mites, cat, molds), and outdoor
allergens (timothy-grass, birch tree). There was no association between tetrachloroethylene
exposure and any of the allergens tested in this study, although some of the other volatile organic
compounds (e.g., toluene, 4-ethyltoluene) were associated with elevated total IgE levels and with
sensitization to milk or eggs.
Another study by Lehmann et al. (2002) examined the relationship between indoor
exposures to volatile organic compounds and T-cell subpopulations measured in cord blood of
newborns (refer to Table 4-25). The study authors randomly selected 85 newborns (43 boys and
42 girls) from a larger cohort study of 997 healthy, full-term babies, recruited between 1997 and
1999 in Germany. Exclusion criteria included a history in the mother of an autoimmune disease
or infectious disease during the pregnancy. Twenty-eight volatile organic compounds were
measured via passive indoor sampling in the child's bedroom for a period of 4 weeks after birth
(a period that is likely to reflect the exposures during the prenatal period close to the time of
4-210
-------
delivery). The levels were generally similar or slightly higher than the levels observed in the
previous study using samples from the bedrooms of the 36-month-old children. The highest
levels of exposure were observed for limonene (median 24.3 ug/m3), a-pinene (median 19.3
ug/m3), and toluene (median 18.3 ug/m3), and the median exposure of tetrachloroethylene was
3.4 ug/m3 (1.8 ug/m3 and 7.3 ug/m3 for the 25th and 75th percentiles, respectively). Flow
cytometry was used to measure the presence of CD3 T-cells obtained from the cord blood
labeled with antibodies against interferon-y, tumor necrosis factor-a, interleukin-2, and
interleukin-4. Tetrachloroethylene was the only one of the measured volatile organic compounds
that was associated with a reduced level of interferon-y. In the univariate analysis, the median
percentages of interferon-y cells were 3.6 and 2.6% in the groups that were below the 75th
percentile and above the 75th percentile of tetrachloroethylene exposure, respectively. The odds
ratio between high (above the 75th percentile) tetrachloroethylene exposure and reduced (less
than the 25th percentile) levels of interferon-y cells was 2.9 (95% CI: 1.0-8.6), adjusting for
family history of atopy, gender, and smoking history of the mother during pregnancy. There was
no association between tetrachloroethylene exposure and interleukin-4 cells, but naphthalene and
methylcyclopentane were associated with elevated levels of interleukin-4 cells.
4.6.1.1.2. Immune-related conditions and diseases
Immunosuppression. In 1982, Lagakos et al. (1986) conducted a telephone survey of
residents of Woburn, Massachusetts, collecting information on residential history and history of
14 types of medically diagnosed conditions. The survey included 4,978 children born since 1960
who lived in Woburn before age 19. Completed surveys were obtained from approximately 57%
of the town residences with listed phone numbers. Lagakos et al. used information from a study
by the Massachusetts Department of Environmental Quality and Engineering to estimate the
contribution of water from the two contaminated wells to the residence of each participant, based
on zones within the town receiving different mixtures of water from various wells, for the period
in which the contaminated wells were operating. This exposure information was used to
estimate a cumulative exposure based on each child's length of residence in Woburn. A higher
cumulative exposure measure was associated with history of kidney and urinary tract disorders
(primarily kidney or urinary tract infections) and with lung and respiratory disorders (asthma,
chronic bronchitis, or pneumonia). There are no other human data that characterize the effects of
tetrachloroethylene-only exposure on immunosuppression, as measured by increased
susceptibility to infections.
Allergy and hyper sensitivity. Allergy and hypersensitivity, as assessed with measures of
immune system parameters or immune function tests (e.g., asthma, atopy) in humans, have not
been extensively studied with respect to the effects of tetrachloroethylene. Delfino et al. (2003a:
4-211
-------
2003b) examined the exacerbation of asthmatic symptoms following exposure to volatile organic
compounds that occurred due to variation in air quality over a 3-month period in 1999-2000 in
Los Angeles. This study included daily repeated exposures to ambient air pollutants and peak
expiratory flow rates over a 3-month period in 21 children (17 males and 4 females) of Hispanic
origin, ages 10-16 years; an additional child participated in the ambient air but not in the exhaled
air portion of the study. Daily diaries were used to record severity of symptoms and asthmatic
episodes. Exposure metrics included exhaled breath measures and ambient levels of eight
volatile organic compounds (benzene, methylene chloride, styrene, toluene, m,p-xy\ene,
o-xylene,/>-dichlorobenzene, and tetrachloroethylene) and eight criteria pollutant gases. An
association between criteria air pollutants and subsequent symptoms of asthma in children in the
Los Angeles area suggests an increased risk of adverse health outcomes with exposure to SC>2
and NC>2 (Delfino et al., 2003a). Although ambient levels of tetrachloroethylene were associated
with bothersome asthma symptoms (OR: 1.37, [95% CI: 1.09, 1.71]) per an interquartile range
change), this association was reduced with the adjustment for 862 or NC>2 (Delfino et al., 2003a).
In the 21 children who participated in the peak expiratory flow measurements, the mean breath
level of tetrachloroethylene was 4.40 ng/L (SD: 10.77 ng/L), the mean ambient level was 3.52
(SD: 2.17) ng/L, and the correlation between the same-day measures was 0.31 [p < 0.01 (Delfino
et al., 2003b)]. There was little relation between asthma symptoms and exhaled breath levels of
tetrachloroethylene. The mean exhalation levels of tetrachloroethylene were 2.50 and 2.69 ng/L,
respectively, in the two groups of asthma symptoms (none or not bothersome; bothersome and
more severe). Stronger associations were reported between asthma symptoms and some of the
other volatile organic chemicals, specifically for benzene, toluene, m,p-xy\ene.
Autoimmune disease. In the 1970s, recognition of a scleroderma-like disease
characterized by skin thickening, Raynaud's phenomenon, and acroosteolysis, and pulmonary
involvement in workers exposed to vinyl chloride (Gama and Meira, 1978) prompted research
pertaining to the role of organic solvents in autoimmune diseases. Exposure to the broad
categories of solvents, organic solvents, or chlorinated solvents has been associated with a two-
to threefold increased risk of systemic sclerosis (scleroderma) in epidemiologic studies
summarized in a recent meta-analysis (Aryal et al., 2001) and in subsequent studies (Maitre et
al., 2004; Garabrant et al., 2003). Similar results were observed in studies of other systemic
autoimmune diseases including undifferentiated connective tissue disease (Lacey et al., 1999),
rheumatoid arthritis (Sverdrup et al., 2005; Lundberg et al., 1994), and antineutrophil-
cytoplasmic antibody (ANCA)-related vasculitis (Beaudreuil et al., 2005; Lane et al., 2003). In
contrast, there was little evidence of an association between solvent exposure and systemic lupus
erythematosus in two recent case-control studies (Finckh et al., 2006; Cooper et al., 2004).
4-212
-------
As described in the preceding paragraph, the epidemiologic data in relation to the role of
solvents, as a broad category, in systemic autoimmune diseases, vary among these conditions.
Much more limited data are available pertaining to specific solvents, including
tetrachloroethylene, and risk of autoimmune diseases. One case report describes a condition
similar to vinyl-chloride induced scleroderma in a man who worked as a presser in a dry-
cleaning plant, and who also helped clean the tetrachloroethylene-containing drums on a weekly
basis (Sparrow, 1977). Another case report describes a localized scleroderma in a man who had
worked with tetrachloroethylene as a metal degreaser, with workplace exposures reported to be
between 10-25 ppm [Hinnen (1995): in German]. Among 279 cases with connective tissue
disease, Goldman (1996) observed a higher frequency of individuals who reported employment
as a dry cleaner among systemic sclerosis patients (4 of 33) compared with patients with other
connective tissue diseases (1 of 246; p < 0.01). Similar patterns were observed with self-
reported history of tetrachloroethylene exposure (3 of 33 systemic sclerosis patients compared
with 2 of 246 other patients,/? < 0.01), but the author noted the difficulty in obtaining this type of
information.
Two registry-linkage studies from Sweden of rheumatoid arthritis (Li et al., 2008;
Lundberg et al., 1994) and three case-control studies of undifferentiated connective tissue disease
(Lacey et al., 1999), scleroderma (Garabrant et al., 2003), and antineutrophil-cytoplasmic
antibody (ANCA) related diseases (Beaudreuil et al., 2005) provide data concerning dry-cleaning
work or tetrachloroethylene exposure (refer to Table 4-26). As expected in population-based
studies, the exposure prevalence is low, with approximately 4% of controls reporting work in dry
cleaning and 1% reporting exposure to tetrachloroethylene. The observed associations are
generally weak for the broad classification of laundry and dry-cleaning work, with odds ratios
for dry cleaning of 1.0 in the largest study of rheumatoid arthritis (Li et al., 2008) and 1.4 in two
studies of scleroderma (Garabrant et al., 2003) and undifferentiated connective tissue disease
(Lacey etal., 1999). None of the individual studies are statistically significant. The studies from
Sweden linking occupational census data to risk of rheumatoid arthritis (Li et al., 2008;
Lundberg et al., 1994) are also limited by the difficulty in defining time of diagnosis for this
disease based on hospitalization data. The results observed for the exposure to
tetrachloroethylene in the three studies that attempted this kind of assessment were more varied
(Beaudreuil et al., 2005: Garabrant et al., 2003: Lacev et al., 1999). Only the study of ANCA-
related diseases resulted in an elevated odds ratio, but again, this estimate was somewhat
imprecise [OR: 2.0, 95% CI: 0.6, 6.9; Beaudreuil et al. (2005)1. These studies are clearly limited
by the low prevalence of and difficulty in accurately characterizing occupational exposure to
tetrachloroethylene in population-based or clinical settings.
4-213
-------
Table 4-26. Immune-related conditions in studies of dry cleaning or
tetrachloroethylene exposure in humans3
Condition and study details
Results
Authors
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
Sweden, hospitalized 1964-2004 (men) or 1970
to 2004 (women). 13,280 male cases and 14,509
female cases; population comparison (full
population), ages >30 yr, Registry linkage to
1960 or 1970 Census occupation data for men
and women, respectively
launderers and dry cleaning
men: 1 exposed cases;
OR: 0.8 (95% CI: 0.1-5.0)
women: 7 exposed cases;
OR: 1.5 (95% CI: 0.7-3.2)
launderers and dry cleaning
men: 57 exposed cases;
OR: 0.8 (95% CI: 0.6-1.0)
women: 204 exposed cases;
OR: 1.0 (95% CI: 0.8-1.1)
Lundberg et
al. (1994)
Li et al. (2008)
Other autoimmune diseases
Undifferentiated connective tissue disease,
Michigan and Ohio, diagnosed 1980-1991
(Michigan) 1980-1992 (Ohio). 205 cases, 2,095
population controls. Women, ages 18 and older.
Structured interview (specific jobs and materials;
jobs held 3 or more mo)
Scleroderma, Michigan and Ohio. Diagnosed
1980-1991 (Michigan), 1980-1992 (Ohio). 660
cases, 2,227 population controls. Women, ages
18 and older. Structured interview (specific jobs
and materials; jobs held 3 or more mo)
ANCA-related diseases,13 France. Diagnosed
1999-2000. 60 patients, 120 hospital controls.
men and women (50% each), mean age 61 yr
dry cleaning
cases: 4.3%, controls 3.8%
OR: 1.4 (95% CI: 0.68, 2.8)
PCE
cases: 0%, controls 1% OR: 0.00
dry cleaning
cases: 4.7%, controls 3.7%
OR: 1.4 (95% CI: 0.9, 2.2)
PCE
serf report cases: 1.1%, controls 1.0%
OR: 1. 4 (95% CI: 0.6, 3.4)
expert review cases: 0.8%, controls 0.8%
OR: 1.1(95%CI:0.4, 2.9)
PCE
cases: 8.3%, controls 4.1%
OR: 2.0 (0.6-6.9)
Lacey et al.
(1999)
Garabrant et
al. (2003)
Beaudreuil et
al. (2005)
Allergy and hypersensitivity
Exacerbation of asthma symptoms, Los Angeles,
1999-2000. 21 children (ages 10-16 yr), 3 mo
diaries, ambient levels and exhaled breath
measures of 8 volatile organic compounds and 8
criteria pollutants
Little evidence of an association between
ambient PCE exposure or exhaled PCE
measures and asthma symptoms
Delfino et al.
(2003a;
2003b)
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).
4-214
-------
4.6.1.1.3. Summary of human noncancer immune and hematologic effects
The strongest study examining immunologic and hematologic effects of
tetrachloroethylene exposure in terms of sample size and use of an appropriately matched control
group is of 40 male dry-cleaning workers (mean exposure levels <140 ppm; mean duration:
7 years) by Emara et al. (2010). Statistically significant decreases in red blood cell count and
hemoglobin levels and increases in total white cell counts and lymphocyte counts were observed
in the exposed workers compared to age- and smoking-matched controls. In addition, increases
in several other immunological parameters, including T-lymphocyte and natural killer cell
subpopulations, IgE, and interleukin-4 levels were observed. These immunologic effects suggest
an augmentation of Th2 responsiveness. However, the limited available data from studies in
children (Delfmo et al.. 2003a: Delfmo et al.. 2003b: Lehmann et al.. 2002: Lehmann et al..
2001) do not provide substantial evidence of an effect of tetrachloroethylene exposure during
childhood on allergic sensitization or exacerbation of asthma symptomology. The observation of
the association between increased tetrachloroethylene exposure and reduced interferon-y in cord
blood samples may reflect a sensitive period of development and points to the current lack of
understanding of the potential immunotoxic effects of prenatal exposures. The available data
pertaining to risk of autoimmune disease in relation to tetrachloroethylene exposure are limited
by issues regarding ascertainment of disease incidence and exposure-assessment difficulties in
population-based studies.
4.6.1.2. Cancers of the Immune System, Including Childhood Leukemia
Forty-one epidemiologic studies report on adult lymphopoietic cancer and
tetrachloroethylene exposure. These publications include numerous cohort or nested case-
control studies (Calvert etal.. 2011: Selden and Ahlborg. 2011: Pukkala et al.. 2009: Radican et
al.. 2008: Sung et al.. 2007: Ji and Hemminki. 2006b: Lynge et al.. 2006: Chang et al.. 2005: Ji
and Hemminki, 2005b: Blair etal., 2003: Travier et al., 2002: Cano and Pollan, 2001: Andersen
etal.. 1999: Boiceetal.. 1999: Blair etal.. 1998: Anttila et al.. 1995: Spirtas et al.. 1991: Lynge
and Thygesen, 1990), case-control studies (Gold et al., 201 Ob: McLean et al., 2009: Schenk et
al.. 2009: 'tMannetie etal.. 2008: Costantini et al.. 2008: Seidler et al.. 2007: Mester et al.. 2006:
Miligi et al.. 2006: Kato et al.. 2005: Costantini etal.. 2001: Fabbro-Perav et al.. 2001: Miligi et
al.. 1999: Clavel et al.. 1998: Aschengrau et al., 1993: Blair etal.. 1993: Scherretal.. 1992:
Siemiatycki, 1991: Mai one et al., 1989: Hardell et al., 1981), and three geographical-based
studies (Cohn et al., 1994: Vartiainen et al., 1993: Morton and Marjanovic, 1984). Some of these
papers represent studies of related populations. For example, three papers examined cancer
incidence or mortality in a cohort of aircraft maintenance workers at an air force base in the
United States, with follow-up through 1982 (Spirtas et al.. 1991). 1990 (Blair etal.. 1998). and
4-215
-------
2000 (Radican et al., 2008). Six papers examined cancer risk among occupational groups
defined by census or employer-provided data in Sweden (Selden and Ahlborg, 2011; Ji and
Hemminki, 2006b, 2005b: Travier et al.. 2002: Cano and Pollan, 2001: Lynge and Thygesen,
1990), two papers were based on census data from Sweden, Denmark, Finland, and Norway
(Lynge et al., 2006: Andersen et al., 1999), and a third paper added data from Iceland (Pukkala et
al., 2009). Four papers examined different subsets of lymphopoietic cancers from a large
population-based case-control study in Italy (Costantini et al., 2008: Miligi et al., 2006:
Costantini et al., 2001: Miligi et al., 1999). Additionally, five epidemiologic studies—one cohort
and four case-control—report on childhood lymphopoietic cancer and tetrachloroethylene
exposure (Sung et al., 2008: Infante-Rivard et al., 2005: Costas et al., 2002: Shuetal., 1999:
Lowengart et al., 1987). Appendix B reviews the design, exposure-assessment approach, and
statistical methodology for each study; the adult lymphopoietic cancer studies are also
summarized in Table 4-27, and the childhood lymphopoietic cancer studies are summarized in
Table 4-28. Most studies were primarily of the inhalation route, of occupational exposure, and,
generally, unable to quantify tetrachloroethylene exposure. Two studies of contaminated
drinking water containing multiple solvents including tetrachloroethylene were available (Cohn
et al., 1994: Vartiainen et al., 1993). Collectively, these studies have varying sensitivities for
identifying cancer hazards.
4.6.1.2.1. Adult lymphopoietic cancer: consideration of exposure assessment
Since the 1960s in Western Europe and the United States, the dry-cleaning industry has
accounted for about 90% of tetrachloroethylene consumption (Gold et al., 2008: Johansen et al.,
2005: IARC, 1995), with more infrequent and lower volume use of trichloroethylene and
CFC-113 for specialized cleaning (IARC, 1995). As described previously, eight publications
used occupational data derived from national census data or by the employer for one or more
northern European countries, focusing on dry cleaners and other laundry workers (Selden and
Ahlborg, 2011: Pukkala et al., 2009: Ji and Hemminki, 2006b: Lynge et al., 2006: Ji and
Hemminki, 2005b: Travier et al., 2002: Cano and Pollan, 2001: Andersen et al., 1999: Lynge and
Thygesen, 1990). Lynge et al. (2006) used national databases and pension schemes to identify
subjects as dry cleaners versus other job titles held in 1970; however, these databases were not
available for subjects from two of the four countries (i.e., Norway and Finland), nor was
information on a subject's workplace and length of employment available for Swedish subjects.
In the absence of national databases, Lynge et al. (2006) collected this information through
interviews, many with a subject's next of kin. A higher likelihood for recall bias is possible with
next of kin or proxy information, particularly for knowledge of solvent exposures as shown by
Boyle et al. (1992). Additionally, workers who may have switched to jobs as dry cleaners after
4-216
-------
1970 would be misclassified using a classification system based on job held in 1970. Two
smaller cohort studies examining mortality using cause of death data from death certificates were
conducted among laundry and dry-cleaning union members in the United States (Calvert et al.,
2011: Blair et aL 2003: Ruder et al.. 2001. 1994).
4-217
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Reference(s) and study description
Cohort Studies
Biologically monitored Finnish workers
All subjects
Lymphopoietic
Non-Hodgkin
Multiple myeloma
Leukemia
1.38(0.28,4.02)
3.76(0.77,11.0)
Not reported
3
3
0
0.38 exp
Aerospace workers (Lockheed)
Routine exposure to PCE
Lymphopoietic
Non-Hodgkin
Hodgkin
Multiple myeloma
Leukemia
1.13(0.62, 1.89)a
1.70 (0.73, 3.34)
0.40(0.01,2.25)
0.55(0.18, 1.29)
14
8
0
0.98 exp
1
5
Routine-intermittent PCE-exposure duration
Oyr
5yr
Test for trend
Non-Hodgkin
1.0b
1.25 (0.43, 3.57)
1.11(0.46,2.70)
1.41 (0.67, 3.00)
p > 0.20
32
4
6
10
Anttila et al. (1995)
849 men and women, blood PCE [0.4 umol/L in
females and 0.7 umol/L in males (median)], follow-up
1974-1992, cancer incidence, external referents (SIR).
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and
n = 3,199 with intermittent-routine PCE exposure),
began work during or after 1960, worked at least 1 yr,
follow-up 1960-1996, JEM without quantitative
estimate of PCE intensity, 1987-1988 8-h TWA PCE
concentration (atmospheric monitoring) 3 ppm [mean]
and 9.5 ppm [median], mortality, external referents for
routine exposure (SMR) and internal referents (workers
with no chemical exposures) for routine-intermittent
PCE exposure (RR).
to
oo
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Oyr
5yr
Test for Trend
Cancer site
Multiple myeloma
Relative risk
(95% CI)
1.0b
0.46 (0.06, 3.48)
1.13(0.38,3.35)
0.24 (0.03, 1.84)
^<0.01
No. obs.
24
1
4
1
Electronic factory workers (Taiwan)
All subjects
Males
Females
Females
Lympho- and hemato-
poietic
Leukemia
0.67 (0.42, 1.01)
0.73 (0.27, 1.60)
0.65 (0.37, 1.05)
0.78(0.49,1.17)
22
6
16
5
Aircraft maintenance workers from Hill Air Force Base
Ever-exposed to PCE
Males
Females
Males
Females
Non-Hodgkin
Multiple myeloma
2.32(0.75, 7. 15)b
2.35 (0.52, 10.71)b
1.71 (0.42, 6.91)b
7.84 (1.43, 43.06)b
5
2
3
2
Reference(s) and study description
Boice et al. (1999) (continued)
Chang et al. (2005): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997,
multiple solvents exposure, does not identify PCE
exposure to individual subjects, lympho- and
hematopoietic cancer incidence, external referents
(SIR) (Chang et al.. 2005): 63,982 females, follow-up
1979-2001, factory employment proxy for exposure,
PCE not identified to individual subjects, leukemia
cancer incidence, external referents, analyses lagged 5
yr (SIR) (Sung et al.. 2007).
Spirtas et al. (1991): Blair et al. (1998): Radican et al.
(2008)
14,066 (10,461 men and 3,605 women) (n = 10,256
ever exposed to mixed solvents, 851 ever-exposed to
PCE), employed at least 1 yr from 1952 to 1956,
follow-up to 2000, PCE used for parachute cleaning,
JEM without quantitative estimate of PCE intensity,
mortality, internal referent (workers with no chemical
exposures) (RR).
to
VO
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Reference(s) and study description
Dry -cleaner and laundry worker
All subjects
Males
Females
All subjects
Males
Females
All subjects
Males
Females
All subjects
Males
Females
All subjects
Males
Females
Lymphopoietic
Non-Hodgkin
Hodgkin
Multiple myeloma
Leukemia
1.0(0.87, 1.15)c
1.05 (0.79, 1.38)c
0.98(0.84, 1.16)c
1.07 (0.86, 1.34)
1.46(0.96,2.13)
0.95 (0.71, 1.23)
1.34(0.81,2.10)
1.88(1.13,2.93)
1.0 (0.73, 1.34)
1.38(0.75,2.31)
0.89 (0.60, 1.26)
0.85(0.65, 1.10)
0.67(0.35, 1.17)
0.90 (0.66, 1.21)
204
53
151
82
27
55
19
0
0.4 exp
19
45
14
31
58
12
46
Andersen et al. (1999)
29,333 men and women identified in 1960 Census
(Sweden) or 1970 Census (Denmark, Finland,
Norway) with occupation as launderers or dry
cleaners, follow-up 1971-1987 or 1991, PCE not
identified to individual subjects, incidence, country -
specific cancer rates referent (SIR).
to
to
o
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
All subjects
Semiquantitative exposure score
Any exposure
Little to no exposure
Medium to high exposure
Males
Females
Males
Females
Males
Females
Males
Females
Cancer site
Non-Hodgkin
Hodgkin lymphoma
Multiple myeloma
Leukemia
Lympho- and hemato-
poietic
Non-Hodgkin
Non-Hodgkinf>g
Multiple myelomaf
Leukemia8
Relative risk
(95% CI)
0.9 (0.5, 1.6)
2.0 (0.6, 4.6)
0.8 (0.3, 1.6)
0.8 (0.4, 1.4)
1.0(0.7, 1.3)
1.0 (0.6, 1.5)
0.9 (0.5, 1.4)
1.76(0.97, 3. 17)d
1.85(0.83,4.12)e
Not reported
0.99 (0.75, 1.26)
1.05 (0.82, 1.32)
0.99 (0.66, 1.38)
1.07 (0.75, 1.45)
0.84 (0.62, 1.90)
1.30(1.03, 1.60)
No. obs.
12
5
7
12
39
18
17
11
6
59
67
52
36
47
80
Reference(s) and study description
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning
union members (1945-1978), follow-up 1979-1993,
PCE exposure potential higher for subcohort entering
union after 1960, semiquantitative cumulative
exposure surrogate to dry clean solvents, cancer
mortality, external referents (SMR).
Cano and Pollan (2001)
Swedish men and women aged 25-64 yr reporting
occupation as "launderers and dry cleaners" in 1970
Census, employed and counted in 1960 Census,
follow-up 1971-1989, NHL incidence from Swedish
Cancer Registry, PCE not identified to individual
subjects, all other occupations referent (RR).
Ji and Hemminki (2QQ6b, 2005b)
9,255 men and 14,974 women reporting laundry and
dry-cleaning work 1970 Swedish Census, follow-up
1960-2002, cases identified from Swedish Cancer
Registry, PCE not assigned to individual subjects,
cancer incidence from Swedish Cancer Registry,
Swedish cancer rates referent (SIR).
to
to
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
All subjects
Launderer and dry cleaner
Male
Female
Launderer and dry cleaner
Male
Female
Launderer and dry cleaner
Male
Female
Launderer and dry cleaner
Male
Female
Launderer and dry cleaner
Male
Female
Cancer site
Non-Hodgkin
Multiple myeloma
Leukemia
Lymphopoietic
Non-Hodgkin
Hodgkin
Multiple myeloma
Leukemia11
Relative risk
(95% CI)
1.03(0.44,2.02)
1.75 (0.70, 3.61)
0.74 (0.30, 1.52)
0.98(0.83, 1.11)
0.94 (0.79, 1.08)
0.99 (0.83, 1.06)
0.98(0.86, 1.10)
0.96 (0.72, 1.25)
0.98(0.86,1.13)
0.97 (0.67, 1.36)
0.77(0.31,1.58)
1.04 (0.68, 1.53)
1.02 (0.86, 1.20)
1.31(0.95,1.78)
0.94 (0.78, 1.33)
0.95 (0.83, 1.09)
0.71(0.50,0.99)
1.03(0.88, 1.19)
No. obs.
8
7
7
653
140
513
264
54
210
33
7
26
152
42
110
204
37
167
Reference(s) and study description
Lynge and Thygesen (1990)
10,600 men and women reporting work in dry cleaner
and laundries in Swedish 1970 Census, follow-up
1970-1980, job title surrogate for exposure, cancer
incidence from Swedish Cancer Registry, Swedish
cancer rates referents (RR).
Pukkala et al. (2009)
15 million men and women participating in national
census on or before 1990, 5 Nordic countries
(Denmark, Finland, Iceland, Norway, Sweden), 30-64
yr, follow-up to 2005, occupational title of launderer
and dry cleaner in any census [n - 8,744 men,
n = 34,752 women], PCE not identified to individual
subjects, cancer incidence from national cancer
registries, national population cancer incidence rates
referent (SIR).
to
to
to
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
All subjects
Exposure duration/time since 1st
employment
PCE-only subjects
All subjects
Exposure duration/time since 1st
employment
PCE-only subjects
Dry-cleaners and laundry workers
PCE
Males
Duration of exposure
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Laundry
Males
Females
Non-Hodgkin
2.33(1.01,4.59)
0.99 (0.43, 1.95)
8
8
All subjects
Males
Females
All subjects
Males
Females
All subjects
Males
Females
Non-Hodgkin1
Hodgkin1
Leukemia
0.86 (0.43, 1.72)
1.32(0.75,2.32)
0.52(0.17,1.61)
2.69(1.01,7.19)
1.58(0.22,11.26)
3.57(1.15,11.13)
1.84(1.11,3.06)
0.93 (0.30, 2.88)
2.53(1.44,4.46)
8
5
3
4
1
3
15
3
12
Reference(s) and study description
Selden and Ahlborg (2011) (continued)
Travier et al. (2002)
Men and women with occupation as dry cleaners,
launderers, and pressers in Swedish 1960 or 1970
Census and employed in laundry, ironing, or dyeing
industries, followed 1971-1989, cancer incidence
from Swedish Cancer Registry, PCE not identified to
individual subjects, all other occupations/industries
referent (RR).
Case-Control Studies
Upper Cape Cod, MA (United States)
Any PCE, no lag
ROD >90th percentile, no lag
Any PCE, >5 yr lag
ROD >90th percentile, >5 yr lag
Leukemia
Leukemia
Leukemia
Leukemia
2.13(0.88,5.19)
8.33(1.53,25.29)
1.96(0.71,5.37)
5.84(1.37,24.91)
7
2
Not
reported
Not
reported
Aschengrau et al. (1993)
34 men and women incident leukemia cases, 737
population controls, stratified by age, vital status, year
of death, sex, telephone or in-person interviews, water
distribution model of Webler and Brown (1993).
adjusted for sex, age, vital status, education, job
exposures (OR).
to
to
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Iowa and Minnesota (United States)
Dry-cleaning industry
Non-Hodgkin
2.0 (0.97, 4.3)
16
Solvents other than benzene
Any exposure
Low intensity
High intensity
Non-Hodgkin
1.1 (0.9, 1.4)
1.1 (0.8, 1.4)
1.4 (0.8, 2.5)
359
334
25
France, 18 provinces
Launderer and dry cleaner
Solvents, more confident exposure
assessment
Hairy cell leukemia
(a type of NHL)
Hairy cell leukemia
(a type of NHL)
3.0 (0.2, 49.2)
0.7 (0.4, 1.2)
1
32
Italy, 12 regions
PCE
Very low/low intensity
Medium/high intensity
Very low/low intensity
Medium/high intensity
Non-Hodgkin +CLL
Leukemia
Leukemia
Hodgkin
0.6 (0.3, 1.2)
1.2 (0.6, 2.5)
0.6 (0.2, 1.6)
1.0 (0.4, 2.7)
Not reported
18
14
6
7
Reference(s) and study description
Blair et al. (1993)
622 histologically confirmed incident NHL cases in
men, 1,245 population controls matched on state, age,
and year deaths [for dead cases], in-person interview,
JEM for solvent group but not PCE individually;
adjusted for age, state, smoking, family history
lymphopoietic disease, agricultural pesticide use, hair
dye use, and proxy respondent (OR).
Clavel et al. (1998)
226 males histologically confirmed hospital HCL
cases, 1980-1990,425 hospital controls from
orthopedic and rheumatological departments matched
on sex, birth date, admission date, residence, self-
administered questionnaire, JEM for solvent
exposures, adjusted for smoking and farming (OR).
Costantini et al. (2001): Miligi et al. (2006): Costantini
et al. (2008)
2,737 incident lymphomas in men and women (1,450
NHL, 365 HD, 652 leukemia, 270 multiple myeloma)
zu /4 yr, iyyi iyyj, i,//y population controls
stratified by sex and age, in-person interview,
exposure proxy of job title and JEM for PCE, adjusted
for sex, age, education, and area (OR).
to
to
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Launderer, dry cleaner, presser
Males
Females
Males
Females
Males
Females
Males
Females
Non-Hodgkin + CLL
Hodgkin
Multiple myeloma
Leukemia
1.6(0.3,9.1)
0.7(0.3, 1.5)
2.5 (0.3, 24.6)
3.5(1.5,8.2)
Not reported
1.0(0.3,3.8)
3.3(0.1,32.4)
1.1(0.4,3.2)
o
J
10
1
7
3
2
5
Languedoc-Roussillon region (France)
Dry-cleaning solvents
Non-Hodgkin
1.0 (0.6, 1.6)
35
Reference(s) and study description
Costantini et al. (2001): Miligi et al. (2006): Costantini
et al. (2008) (continued)
Fabbro-Peray et al. (2001)
445 histologically confirmed Hodgkin and NHL
hospital cases in men and women recruited,
1992-1996, 1,025 population controls stratified on
municipalities size and population distribution, in-
person or telephone interview, serf-reported exposure,
exposed defined as duration >1 yr, 5 yr prior to
diagnosis, information, adjusted for age, sex, urban
setting, education level (OR).
to
to
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Puget Sound-Seattle (Washington State), Detroit (Michigan) (United States)
Ever exposed to PCE
Multiple myeloma
1.5 (0.8, 2.9)1
16
Cumulative exposure (ppm-wk)
Referent
1-353
354-1,430
1,431-4,875
4,876-13,500
/>-value for trend
Textile, apparel, furnishing machine
operators and tenders (includes dry-
cleaning machine operators)
Multiple myeloma
Multiple myeloma
1.0a
0.3 (0.04, 3.0)1
0.5(0.1, 4.4?
1.5 (0.4, 5.4)1
3.3(1.2,9.5)'
p = 0.02
6.0(1.7,21)
164
1
1
4
10
9
Exposure duration
1-5 yr
>5yr
Trend test
Dry-cleaning machine operators
Multiple myeloma
Multiple myeloma
3.6 (0.7, 1.7)
12(1.3,110)
;? = 0.001
Not reported
4
5
5 cases, 3
controls
Umea (Sweden)
Any styrene, TCE, PCE, benzene
exposure
Non-Hodgkin
4.6(1.9, 11.4)
10
Reference(s) and study description
Gold et al. QOlOb)
180 histologically confirmed multiple myeloma cases
in men and women reported to cancer registries,
2000-2002, 481 population controls, ROD or
Medicare/Medicaid services files, in-person interview,
serf-reported or proxy -assisted reply to all jobs held
>12 mo since 1945, adjusted for age, gender, race,
education, study site (OR).
Hardell et al. (1981)
169 men histologically confirmed incident NHL and
Hodgkin cases, 1974-1978, population controls,
25-85 yr, matched for sex, age, and residence, and
death [for dead cases], self-administered questionnaire,
OR from univariate %2 test.
to
to
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
New York (United States)
Dry-cleaning fluids
Non-Hodgkin
1.59(0.49,5.13)
7
Population of Denmark, Finland, Norway, Sweden
Dry cleaner
Other job in DC
Unclassifiable
Dry cleaner, employment duration,
1964-1979
10yr
Unknown
Non-Hodgkin
Non-Hodgkin
Non-Hodgkin
Non-Hodgkin
1.0 (0.7, 1.4)k
0.7 (0.3, 1.6)k
0.9 (0.6, 1.4)k
1.0 (referent)
1.35(0.44,4.14)
0.61(0.17,2.21)
0.92 (0.49, 1.72)
0.66 (0.36, 1.22)
1.47(0.49,4.47)
42
8J
52J
145
5
o
J
14
15
5
Reference(s) and study description
Kato et al. (2005)
376 cases histologically confirmed NHL in women,
20-79 yr, 1995-1998, NY State Cancer registry, 463
population controls stratified on age, telephone
interview, occupation exposure to solvents, dry-
cleaning fluids, adjusted for age, family history
hematologic cancer, education, interview year, proxy
respondent, BMI, prescription/over-counter drugs,
pesticide exposures (OR).
Lynge et al. (2006)
46,768 subjects with occupation "laundry and dry-
cleaning worker" or industry "laundry and dry
cleaning" in 1970 Censuses in Denmark, Finland,
Norway, Sweden followed 1970-1971 through
1997-2001; 247 incident cases NHL, controls
randomly selected from cohort, matched on country,
sex, age, and calendar period at time of diagnosis. Dry
cleaner assigned by job title or employed in shop <10
employees using pension data in Denmark and Finland
or by questionnaire for subjects from Sweden and
Norway; mean PCE during study period, 24 ppb (165
mg/m3), nested case-control study (OR [RR]).
to
to
oo
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
United States (SEER)
Dry cleaner occupation
Chronic lymphocytic
leukemia (a type of
NHL)
1.1(0.6,2.0)
(all respondents)
0.9 (0.4, 1.8)
(self-respondents,
noNOK
information)
14
New Zealand
Textile bleaching, dyeing and
cleaning machine operators
Non-Hodgkin
Leukemia
0.75 (0.24, 2.32)
2.07 (0.70, 6.09)
5
6
Germany, 6 regions
Launderer, dry cleaner, presser
Any exposure
1-10 yr duration
>10 yr duration
Non-Hodgkin and
Hodgkin
1.3 (0.5, 3.2)
0.8 (0.3, 2.5)
3.4 (0.6, 18.5)
11
6
5
PCE
0 ppm-yr
>0-<9.1 ppm-yr
>9.1-<78.8 ppm-yr
>78.8 ppm-yr
Test for trend
Non-Hodgkin and
Hodgkin
1.0 (reference)
1.1(0.5,2.3)
1.0 (0.5, 2.2)
3.4 (0.7, 17.3)
p = 0.12
667
16
14
6
Reference(s) and study description
Malone et al. (1989)
427 men and women incident CLL cases and 1,683
population controls, <80 yr of age, SEER sites,
matched on sex, race, age, education, study site,
questionnaire, chlorinated HC surrogate exposure
metric, adjusted for race, age, education, sex, study
site (OR).
't Mannetje et al. (2008): McLean et al. (2009)
291 NHL cases Ct Mannetie etal. 2008) and 225
leukemia cases (McLean et al.. 2009). in men and
women, 20 or 25-75 yr, 2003-2004, New Zealand
Cancer Registry, 471 population controls frequency
matched on age, in-person interview, occupational title
as surrogate exposure metric, adjusted for age, sex,
and smoking (OR).
Mester et al. (2006): Seidler et al. (2007)
710 histologically confirmed Hodgkin and NHL in
men and women, 18-80 yr, 1998-2003, 710
population controls matched on sex, region, and age,
in-person interviews, exposure assessed by job title
and JEM for semiquantitative intensity metric,
adjusted for smoking and alcohol consumption (OR).
to
to
VO
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
0 ppm-yr
>0-<9.1 ppm-yr
>9.1-<78. 8 ppm-yr
>78.8 ppm-yr
Test for trend
Cancer site
Multiple myeloma
Relative risk
(95% CI)
1.0 (reference)
1.8 (0.5, 6.7)
p = 0.34 (negative)
No. obs.
33
o
J
0
0
4-SEER reporting sites (CA, IO, MI, WA, United States)
Launderers and ironers
Non-Hodgkin
3.89 (1.06, 14.20)
12
Montreal, Canada
Launderer and dry cleaner
Any exposure
Substantial exposure
Non-Hodgkin
0.9 (0.3, 2.4)
(0.00, 1.7)
3
0
Reference(s) and study description
Mester et al. (2006): Seidler et al. (2007) (continued)
Schenk et al. (2009)
2,046 histologically confirmed NHL in men and
women, 20-74 yr, 1998-2000, 1,057 population
controls frequency matched on age, sex, race and
study center, mailed questionnaire, occupational title
exposure surrogate, adjusted for age, group, sex,
ethnicity, and study center (OR).
Siemiatycki (1991)
215 men and women histologically confirmed incident
NHL cases, 1979-1985, 35-70 yr, 533 population
control group and cancer control group, in-person
interviews, occupational title and JEM for PCE,
adjusted age, family income, and cigarette index, 90%
CI (OR).
Geographic-based and Other Studies
Northern New Jersey, 75 Municipalities (United States)
PCE in town water >5 ppb
Males
Females
Males
Females
Non-Hodgkin1
Leukemia
1.20 (0.94, 1.52)
1.38(1.08, 1.70)
0.84 (0.66, 1.06)
1.20 (0.94, 1.52)
78
87
63
56
Cohn et al. (1994)
1,190 leukemia cases identified from NJ State Cancer
Registry, 1979-1987, residence in 1 of 17 NJ
municipalities, PCE and other chlorinated solvents in
municipal water supplies, log-linear regression
adjusted for age, stratified by sex (RR).
to
oo
o
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No. obs.
Portland- Vancouver Metropolitan Area, Oregon (United States)
Dry cleaners and launderers
Males
Females
Males
Females
Males
Females
All leukemia
Lymphatic leukemia
Nonlymphatic leukemia
55.7 per 100,000m
23.7 per 100,000m
27.8 per 100,000m
20.9 per 100,000m
27.8 per 100,000m
9.0 per 100,000m
2
5
1
4
1
2
Hausjarvi and Hattula, Finland
Hausjarvi
Hattula
Hausjarvi
Hattula
Hausjarvi
Hattula
Hausjarvi
Hattula
Non-Hodgkin
Hodgkin
Multiple myeloma
Leukemia
0.6(0.3, 1.1)
1.4(1.0,2.0)
0.8 (0.3, 1.7)
1.4 (0.7, 2.5)
0.7(0.3, 1.3)
0.7 (0.2, 1.3)
1.2 (0.8, 1.7)
0.7(0.4, 1.1)
14
31
6
11
7
6
33
19
Reference(s) and study description
Morton and Marjanovic (1984)
1,622 leukemia cases identified from 24 hospitals and
death certificates, 1963-1977, 16-74 yr, occupational
title as exposure surrogate, 1,611 dry cleaners and
launderers in 1970 population census, age-
standardized rates using 1970 population.
Vartiainen et al. (1993)
Lymphopoeitic cancers, liver cancer and all cancers
among residents with PCE and other solvents in
drinking water, 1953-1991, no subject-level exposure
information, cancer rates of Finnish population
referent (SIR).
to
oo
a For Boice et al. (1999). all lymphopoetic cancers is the sum of ICD 9th Edition, 200-208.
b Internal referent population as comparison.
c For Andersen et al. (1999). all lymphopoeitic cancer is the sum of ICD 7th Edition, 200-204.
d For Cano and Pollan (2001). relative risk for male dry cleaner and launderers in 1970 Census.
e For Cano and Pollan (2001). relative risk for male dry cleaner and launderers in 1960 and 1970 Censuses.
-------
Table 4-27. Summary of epidemiologic studies on tetrachloroethylene exposure and hematopoietic cancers,
including leukemia (continued)
f For Ji and Hemminki (2006b). female subjects reporting occupation as launderers and dry cleaner in two consecutive censuses, 1960 -1970, SIRs for NHL were
0.76 (95% CI: 0.39, 1.25) [n = 12] and 0.87 (95% CI: 0.76, 1.10) [n = 64], respectively, and, for multiple myeloma, 1.01 (0.46, 1.78) [n = 9] and 0.88 (0.60,
1.21) [n = 31], respectively.
8 For Ji and Hemminki (2006b, 2005b), SIR for launderers and dry cleaners in 1960 Census. For lymphopoietic subtypes in launderers and dry cleaners in 1960
Census, for males, SIR: 0.85 (0.51, 1.28) [n = 19] for chronic lymphocytic leukemia, a form of NHL; 0.63 (0.25, 1.18) [n = 7] for acute myelogenous
leukemia; 0.91 (0.29, 1.87) [n = 5] for chronic myelogenous leukemia; and, 1.04 (0.41, 1.96) [n = 7] for polycythemia vera; and, for females, SIR: 1.54 (1.05,
2.12) [n = 32] for chronic lymphocytic leukemia; 0.1.36 (0.83, 2.02) [n = 20] for acute myelogenous leukemia; 0.33 (0.03, 0.94) [n = 2] for chronic
myelogenous leukemia; and, 1.71 (0.93, 2.73) [n = 14] for polycythemia vera.
hFor Pukkala et al. (2009). SIR for chronic lymphatic leukemia, a form of NHL, were 0.90 (95% CI: 0.50-1.49) [males, n = 15 cases] and 1.02 (95% CI: 0.74,
1.36) [females, n = 46 cases].
:For Travier et al. (2002). RRs for subjects reporting occupation as dry cleaners, launderers, or pressers and employed in dry-cleaning industry in 1960 and 1970
Censuses (Group 2). RRs for these subjects for chronic lymphocytic leukemia, a form of NHL, were 0.67 (0.09, 4.76) [males, n = 1] and 2.89 (1.20, 6.96)
[females, n = 5].
1 For Gold et al. (2010b). odds ratio for PCE exposure with jobs assessed with low confidence considered unexposed.
kLynge et al. (2006) is a nested case-control study. RR adjusted for matching criteria (country, sex, 5-yr age group and 5-yr calendar period at the time of
diagnosis of the case).
_^ 'For Cohn et al. (1994). RRs for chronic lymphocytic leukemia, a form of NHL, were 0.98 (0.65, 1.47) [males, n = 28] and 0.93 (0.56, 1.52) [females, n = 19].
^ m For Morton and Marjanovic (1984). age-standardized incidence rate is statistically significantly different from rate for all men or all women.
oo
to
CLL = chronic lymphocytic leukemia; Exp = expected number of cancers; JEM = job-exposure matrix; NOK = next of kin; ROD = relative delivered dose.
-------
Table 4-28. Summary of epidemiologic studies on tetrachloroethylene exposure and childhood hematopoietic
cancers, including leukemia
Exposure group
Cancer site
Relative risk
(95% CI)
No.
obs.
Reference(s) and study description
Cohort Studies
Offspring of Electronic factory workers (Taiwan)
Nonexposed
Exposed pregnancy to organic solvents
All leukemia (ICD 9, 204-208)
1.0
3.83(1.17, 12.55)
9
6
Sung et al. (2009)
40,647 first singleton births among 47,356 women
employed at factory, 1978-2001, 8,506 births
among women employed 3 mo prepregnancy and 3
mo postconception, incident childhood cancers
from national cancer registry, 1979-2001, does not
identify PCE exposure to individual mothers,
Poisson regression adjusted for maternal age,
maternal education, sex and birth year, internal
referents [offspring of subjects not employed
during period] (RR).
Case-Control Studies
Residents of ages <19 in Woburn, MA (United States)
Maternal exposure 2 yr before conception to diagnosis
Never
Least
Most
(p for linear trend)
Acute lymphocytic leukemia
1.00
5.00 (0.75, 33.5)
3.56(0.51,24.8)
>0.05
3
9
7
Maternal exposure 2 yr before conception
Never
Least
Most
(p for linear trend)
Acute lymphocytic leukemia
1.00
2.48 (0.42, 15.2)
2.82 (0.30, 26.4)
>0.05
11
4
5
Costas et al. (2002)
19 leukemia, 1969-1989, identified through
physician or hospital records pre-1982 and MA
Cancer Register 1982 onward, 37 local public
school controls matched on race, sex, birth date,
residential status, in-person interview,
questionnaire to parents included information on
use of public drinking water in the home, hydraulic
mixing model used to estimate fraction of month
that TCE, PCE and other solvents in drinking
water were delivered to residence 1964-1979
(Murphy. 1990). logistic regression with
composite covariate for socioeconomic status,
maternal smoking during pregnancy, maternal age
at birth of child, and breastfeeding (OR).
-------
Table 4-28. Summary of epidemiologic studies on tetrachloroethylene exposure and childhood hematopoietic
cancers, including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No.
obs.
Birth to pregnancy
Never
Least
Most
(p for linear trend)
Acute lymphocytic leukemia
1.00
1.82(0.31, 10.8)
0.90(0.18,4.56)
>0.05
7
7
5
Maternal exposure during pregnancy
Never
Least
Most
(p for linear trend)
Acute lymphocytic leukemia
1.00
3.53(0.22,58.1)
14.3 (0.92, 224)
0.05
9
3
7
Residents of ages <14 yr Quebec (Canada)
Probable/definite exposure to PCE
Maternal exposure 2 yr before
conception to birth
During pregnancy
Acute lymphocytic leukemia
ICD 9 204.0
0.87(0.35-2.18)
0.96 (0.41-2.25)
0.84 (0.30-2.34)
18
11
7
Cumulative exposure score
<4
>4
Acute lymphocytic leukemia
ICD 9 204.0
0.95 (0.35-2.55)
0.55 (0.05-6.34)
Reference(s) and study description
Costas et al. (2002) (continued)
Infante-Rivard et al. (2005)
790 acute lymphoblastic leukemia, 1980-2000,
790 population controls from family stipend
records, 1980-1993, or health insurance records,
1994-2000, matched on sex and age, telephone
interview with questions on maternal occupation,
blinded JEM for PCE, logistic regression stratified
by time period and adjusted for maternal age and
education (OR).
-------
Table 4-28. Summary of epidemiologic studies on tetrachloroethylene exposure and childhood hematopoietic
cancers, including leukemia (continued)
Exposure group
Cancer site
Relative risk
(95% CI)
No.
obs.
Residents of ages <10 yr Los Angeles (CA) Cancer Surveillance Program
Maternal occupational exposure to PCE
Acute lymphatic and
nonlymphatic leukemia
Not reported
Paternal occupational exposure to PCE
1 yr before pregnancy
During pregnancy
After pregnancy
Acute lymphatic and
nonlymphatic leukemia
00(^ = 0.39)
00(^ = 0.39)
oo(0.19-«)
l:0a
l:0a
2:0a
Children's Cancer Group Study (children <15 yrof age) (Australia, Canada, United States)
Maternal occupational exposure to PCE
Anytime
Preconception
During pregnancy
Postnatal
Acute lymphocytic leukemia
0.4(0.1-1.4)
1.4 (0.2-8.6)
1.3 (0.2-8.4)
0.4(0.1-1.5)
4
3
3
4
Paternal occupational exposure to PCE
Anytime
Preconception
During pregnancy
Postnatal
Acute lymphocytic leukemia
0.9 (0.5-1.6)
0.8(0.5-1.5)
0.5(0.2-1.1)
0.5 (0.2-1.2)
25
21
8
10
Reference(s) and study description
Lowengart et al. (1987)
123 case-control pairs — acute lymphocytic and
nonlymphocytic leukemia cases, 1980-1984, and
maternal friend controls or population controls
matched on age, sex, race, nonblinded telephone
interview, serf-reported occupational exposure,
logistic regression (OR).
Shu et al. (1999)
1,842 acute lymphocytic leukemia cases identified
in 37 participating institutions, 1989-1933, 1,986
population controls, ROD, matched on age, race
and telephone area code/exchange, telephone
interview with structured questionnaire to assess
parental exposure to PCE using job-industry title
and self-reported exposure history, logistic
regression adjusted for maternal education, race
and family income (maternal exposures) or
paternal education, race, family income, age and
sex of case (OR).
a For Lowengart et al. (1987). the number of case:control pairs.
Exp = expected number of cancers; JEM = job-exposure-matrix; ROD = relative delivered dose; Obs = observed number of cancers.
-------
The exposure surrogate in studies of dry cleaners and launderers is a broad category and
will have some associated measurement error as this broad category does not account for
individual characteristics that modify one's exposure potential. For example, some variation can
be expected within an occupational group between countries, as Lynge et al. (2006) reported,
average tetrachloroethylene usage in 1960-1970 in Sweden was higher than in Finland or
Norway. The more general the exposure surrogate, such as job title, the greater the likelihood
for misclassification errors, as differences in tasks and exposure conditions within a job title may
be considerable. For some occupations, these differences may be gender related, making it
difficult to interpret differences in relative risk that may be observed between men and women
within a specific occupational group (Messing et al., 1994). Blair et al. (2003) recruited
members of a laundry and dry-cleaning workers union and attempted to increase the specificity
of the classification of tetrachloroethylene exposure by examining a subgroup who entered the
cohort after 1960, a time of widespread tetrachloroethylene use in dry cleaning. However, this
restriction resulted in a considerable decrease in the number of observed cases of lymphopoietic
cancers, from 39 in the full cohort to 2 in the group that joined after 1960. Blair et al. (2003)
also developed a semi quantitative exposure intensity score using published monitoring data. The
available data indicated a high degree of consistency in exposure levels to tetrachloroethylene
between establishments and provided information that could be used to categorize differences in
potential exposures based on types of jobs. Exposure was characterized with respect to distance
from the washers (cleaners were assigned a high-exposure score, pressers, sewers, and counter
clerks were assigned a medium-exposure score, and those who worked at locations that did not
include washing facilities were assigned a no-exposure score) (Blair et al., 2003; Blair et al.,
1990). Another study by Calvert et al. (2011) of unionized dry cleaners in the United States
included an analysis of subjects who worked for one or more years before 1960 in a shop known
to use tetrachloroethylene as the primary solvent. The cohort was stratified into two groups
based on the level of certainty that the worker was employed only in facilities using
tetrachloroethylene as the primary solvent: tetrachloroethylene-only and tetrachloroethylene
plus. Another approach to improving the exposure measure was used by Lynge et al. (2006). In
this study, effect measures were presented for dry cleaners separately from other laundry
workers. Selden and Ahlborg (2011) obtained information about the dry-cleaning establishment
(e.g., washing techniques, chemicals used, number of employees, and work history of individual
employees) in a questionnaire sent to businesses in Sweden in the 1980s to identify subjects as
either dry cleaners or laundry workers. Travier et al. (2002) presented estimates for launderers,
dry cleaners, and pressers, using job classifications based on the 1960 or 1970 Census data, and
for subjects holding a dry-cleaning job in both census years.
4-236
-------
A variety of exposure-assessment approaches have been used in studies in other work
settings and in population-based case-control studies. One occupational study assessed
tetrachloroethylene potential for individual subjects using biological monitoring data (Anttila et
al., 1994). The cohort studies of aerospace workers (Boice etal., 1999) and aircraft maintenance
workers (Radican et al., 2008; Blair et al., 1998; Spirtas et al., 1991) developed a job exposure
matrix referencing historical industrial monitoring data. In case-control studies, attributes that
strengthen the quality of the exposure assessment include ascertainment of a complete job
history (i.e., all jobs held for >6 or 12 months rather than limiting to most recent job or longest-
held job), inclusion of information on job tasks or duties as well as job title, inclusion of
additional modules for specific jobs that collect more detailed information pertaining to exposure
conditions, and blinded exposure assessment and development of job-exposure matrices focusing
on tetrachloroethylene based on this complete set of information. These attributes were used in
the case-control studies in Italy (Costantini et al., 2008; Miligi et al., 2006) and a case-control
study of multiple myeloma in Washington (Gold etal., 2010a). One case-control study of
potential residential tetrachloroethylene exposure used a statistical model of water distribution
system to estimate delivered dose to a subject's home (Aschengrau et al., 1993). Because a
nondifferential misclassification of exposure most often leads to an attenuation of the observed
effect estimates (Dosemeci et al., 1990), the relative specificity of these exposure-assessment
approaches, particularly those that allow assignment of values to individuals within the study,
strengthens their ability to identify cancer hazards compared to studies with broader exposure-
assessment approaches.
4.6.1.2.2. Adult lymphopoietic cancer: consideration of disease subtypes
The broad category of lymphopoietic cancers can be divided into specific types of
cancers, including non-Hodgkin lymphoma, Hodgkin lymphoma, multiple myeloma, and various
types of leukemia (e.g., acute and chronic forms of lymphoblastic and myeloid leukemia). The
classification criteria for these cancers have changed over the past 30 years, reflecting improved
understanding of the underlying stem cell origins of these specific subtypes. For example, hairy
cell leukemia, chronic lymphocytic leukemia, non-Hodgkin lymphoma, and multiple myeloma
may arise from mature B cells. This understanding may help elucidate common etiologic
pathways and exposures. The studies of tetrachloroethylene exposure examine various
outcomes, including the broad category of lymphopoietic cancers, as well as non-Hodgkin
lymphoma, Hodgkin lymphoma, non-Hodgkin lymphoma plus chronic lymphocytic leukemia,
hairy cell leukemia, multiple myeloma, and leukemia.
All of the studies of dry cleaning and other occupations from the Nordic countries
ascertained cancer incidence using national cancer registries. Four other cohort studies from the
United States (Calvert etal., 2011: Radican et al.. 2008: Blair etal.. 2003: Boice etal.. 1999)
4-237
-------
relied on cause-of-death data from death certificates or the National Death Index. For diseases
with a relatively high survival rate such as non-Hodgkin lymphoma (5-year survival: 67.4%
based on 1999-2006 data), use of cause-of-death data may underestimate cancer risk. Most of
the case-control studies relied on histologically confirmed cases of incident cancers in a defined
geographic area, as ascertained from cancer registries.
4.6.1.2.3. Adult lymphopoietic cancer: consideration of potential confounding and other
factors
Common behaviors, such as smoking and use of alcohol, have not been strongly
associated with non-Hodgkin lymphoma and multiple myeloma, so there is little reason to be
concerned about potential confounding of observed results pertaining to specific jobs or
tetrachloroethylene measures by these factors. Smoking is a risk factor for some kinds of
leukemia, however, and so its role as a potential confounder for this outcome should be
considered. Tetrachloroethylene was the primary, or in Nordic countries, the exclusive solvent
used in dry cleaning (Lynge et al., 2006; Johansen et al., 2005). In studies of some types of
occupations, participants may also have been exposed to other solvents.
4.6.1.2.4. Adult lymphopoietic cancer: summary of results
All of the studies examining the broad category of lymphopoietic cancers were cohort
studies, with the number of exposed cases ranging from 3, in a study of biologically monitored
workers in Finland (Anttila et al., 1995), to 653, in a study using occupational census codes in
five Nordic countries (Pukkala et al., 2009) (refer to Table 4-29). The relative risk estimates
among these seven studies ranged from 0.67 (95% CI: 0.42, 1.01) to 1.51 (95% CI: 0.75, 2.70),
with values from the largest studies around 1.0 (Pukkala et al., 2009; Andersen et al., 1999). The
three studies with relative risk estimates greater than 1.0 were studies that used a relatively high
quality exposure-assessment methodology: an standardized incidence ratio (SIR) of 1.39 (95%
CI: 0.28, 4.02) in a small study in Finland examining risk among workers who had been
monitored using blood tetrachloroethylene levels (Anttila et al., 1995), an SMR of 1.51 (95%
CI: 0.75, 2.70) among laundry and dry-cleaning union workers employed prior to 1960 only in
facilities using tetrachloroethylene as the primary solvent (tetrachloroethylene-only) (Calvertet
al., 2011), and an SMR of 1.13 (95% CI: 0.62, 1.89) for routine exposure to tetrachloroethylene,
based on a job exposure matrix, in a cohort study of workers in the aerospace industry (Boice et
al., 1999). In the other study with a relatively detailed exposure-assessment methodology (a
semi quantitative exposure score based on job titles and proximity to washers), no increased risk
was observed (SMR: 0.9, 95% CI: 0.5, 1.4, for the medium/high intensity score group) (Blair et
al., 2003).
4-238
-------
Table 4-29. Results of epidemiologic studies of potential tetrachloroethylene
exposure and adult lymphopoietic cancer and leukemia, by cancer type and
study design
Cancer type,
n exposed
cases
Relative risk
(95% CI)
Lymphopoietic (all)
3
11
14
22
39
204
653
1.38 (0.28,
4.02)
1.51 (0.75,
2.70)
1.13 (0.62,
1.89)
0.67 (0.42,
1.01)
1.0 (0.7, 1.3)
0.9 (0.5, 1.4)
1.0(0.87, 1.15)
0.98 (0.30,
1.52)
Leukemia (all)
5
5
7
12
3
12
15
58
47
0.55(0.18,
1.29)
0.78 (0.49,
1.17)
0.74 (0.30,
1.52)
0.8 (0.4, 1.4)
0.93 (0.30,
2.88)
2.53 (1.44,
4.46)
1.84(1.11,
2.88)
0.85 (0.65, 1.0)
0.84 (0.62,
1.90)
Design, location, exposure assessment"
Cohort
biological monitored workers (SIR), Finland, blood PCEa
laundry and dry-cleaning workers (SMR), United States,
union employment records (PCE-only exposure based on
history of solvent use by shops)
aerospace workers (SMR), United States, job exposure matrix
(PCE routine exposure)3
electronic factory workers (SIR), Taiwan
laundry and dry-cleaning workers (SMR), United States,
union records (all workers)
(medium/high intensity score)3
laundry and dry-cleaning workers (SIR), Sweden, Denmark,
Finland, Norway, census occupation codes
laundry and dry-cleaning workers (SIR), Sweden, Denmark,
Finland, Norway, Iceland, census occupation codes
Cohort
aerospace workers (SMR), United States, job exposure matrix
(PCE routine exposure)3
electronic factory workers (SIR), Taiwan (females)
laundry and dry-cleaning workers (SIR), Sweden, census
occupation codes
laundry and R workers (SMR), United States, union records
(all workers)
laundry and dry-cleaning workers and pressers, Sweden,
census occupation codes, 1960 and 1970 (males)
laundry and dry-cleaning workers and pressers, Sweden,
census occupation codes, 1960 and 1970 (females)
laundry and dry-cleaning workers and pressers, Sweden,
census occupation codes, 1960 and 1970 (males and females)
laundry and dry-cleaning workers (SIR), Sweden, Denmark,
Finland, Norway, census occupation codes
laundry and dry-cleaning workers (SIR), Sweden (males)
Reference
Antilla et al.
(1995)
Calvert et al.
(2011)
Boice et al. (1999)
Chang et al. (2005)
Blair et al. (2003)
Blair et al. (2003)
Andersen et al.
(1999)
Pukkala et al.
(2009)
Boice et al. (1999)
Sung et al. (2007)
Lynge and
Thygesen (1990)
Blair et al. (2003)
Travier et al.
(2002)
Travier et al.
(2002)
Travier et al.
(2002)
Andersen et al.
(1999)
Ji and Hemminki
(2005b)
4-239
-------
Table 4-29. Results of epidemiologic studies of potential
tetrachloroethylene exposure and adult lymphopoietic cancer and leukemia,
by cancer type and study design (continued)
Cancer type,
n exposed
cases
80
204
Relative risk
(95% CI)
1.30(1.03,1.60)
0.95 (0.83, 1.09)
Leukemia (all)
2
5
6
7
3.3 (0.3, 32.4)
1.1(0.4,3.2)
2.07 (0.70, 6.09)
1.0 (0.4, 2.7)
Leukemia (all)
7
19
33
56
64
2.13(0.88,5.19)
0.7(0.4, 1.1)
1.2 (0.8, 1.7)
1.20 (0.94, 1.52)
0.84 (0.66, 1.06)
Design, location, exposure assessment"
laundry and dry-cleaning workers (SIR), Sweden
(females)
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, Iceland, census occupation
codes
Case-control
Italy, job titles (launderer, dry cleaner, presser) (males)
Italy, job titles (launderer, dry cleaner, presser)
(females)
New Zealand, occupational title (textile bleaching,
dyeing and cleaning machine operators)
Italy, job exposure matrix (PCE, medium/high
intensity)3
Geographic based
Massachusetts, water distribution model (any PCE)a
Finland (Hattula), PCE in drinking water
Finland (Hausjarvi), PCE in drinking water
New Jersey, PCE in town water >5 ppb (females)
New Jersey, PCE in town water >5 ppb (males)
Reference
Ji and Hemminki
(2005b)
Pukkala et al. (2009)
Costantini et al.
(2001)
Miligi et al. (1999)
McLean et al. (2009)
Costantini et al.
(2008)
Aschengrau et al.
(1993)
Vartiainen
etal. (1993)
Vartiainen et al.
(1993)
Cohn et al. (1994)
Cohn et al. (1994)
a Studies with relatively high quality exposure assessment methodologies, based on biological monitoring data,
cohort studies with job exposure matrix based on historical industrial monitoring data, or case-control studies with
job exposure matrix focusing on PCE based on information on job title and tasks or duties, and additional modules
for specific jobs, or studies of residential PCE exposure using a statistical model of water distribution system to
estimate delivered dose to a subject's home.
Studies of leukemia risk include occupational cohorts and case-control studies and
geographic-based studies of residential exposure (refer to Table 4-30). The cohort studies range
from 5 to 204 cases. Two studies using Swedish census data on occupation reported elevated
relative risks among women, but not men, who reported jobs as launderers or dry cleaners.
4-240
-------
Table 4-30. Results of epidemiologic studies of potential tetrachloroethylene
exposure and adult non-Hodgkin lymphoma, by study design
Cancer type,
n exposed
cases
Relative Risk
(95% CI)
Adult non-Hodgkin lymphoma
3
2
5
6
8
8
8
11
12
15
18
27
55
59
67
264
3.76(0.77, 11.0)
2.35 (0.52, 10.7)
2.32(0.75,7.15)
2.46 (0.90, 5.36)
1.70 (0.73, 3.34)
1.03 (0.44, 2.02)
0.86 (0.43, 1.72)
1.76(0.97,3.17)
0.9 (0.5, 1.6)
2.02(1.13,3.34)
1.14(0.68, 1.81)
1.46(0.96,2.13)
0.95 (0.71, 1.23)
0.99 (0.75, 1.26)
1.05 (0.82, 1.32)
0.98(0.86, 1.10)
Design, location, exposure assessment"
Cohort
biological monitored workers (SIR), Finland, blood PCEa
aircraft maintenance workers (RR-internal referent),
United States, job exposure matrix (PCE) (females)3
aircraft maintenance workers (RR-internal referent),
United States, job exposure matrix (PCE) (males)3
laundry and dry-cleaning workers (SMR), United States,
union employment records (PCE-only exposure based on
history of solvent use by shops)
aerospace workers (SMR), United States, job exposure
matrix (routine exposure to PCE)3
laundry and dry-cleaning workers (SIR), Sweden, census
occupation codes
laundry and dry-cleaning workers and pressers, Sweden,
census occupation codes
laundry and dry-cleaning workers (SIR), Sweden, census
occupation codes
laundry and dry-cleaning workers (SMR), United States,
union records (all workers)
dry-cleaning workers (SIR), Sweden, census occupation
codes and questionnaire (dry cleaner) (males)3
dry-cleaning workers (SIR), Sweden, census occupation
codes and questionnaire (dry cleaner) (females)3
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, census occupation codes
(males)
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, census occupation codes
(females)
laundry and dry-cleaning workers (SIR), Sweden (males)
laundry and dry-cleaning workers (SIR), Sweden (females)
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, Iceland, census occupation
codes
Reference
Antilla et al.
(1995)
Radican et al.
(2008)
Radican et al.
(2008)
Calvert et al.
(2011)
Boice et al. (1999)
Lynge and
Thygesen (1990)
Travier et al.
(2002)
Cano and Pollan
(2001)
Blair et al. (2003)
Selden and
Ahlborg (2011)
Selden and
Ahlborg (2011)
Andersen et al.
(1999)
Andersen et al.
(1999)
Ji and Hemminki
(2006b)
Ji and Hemminki
(2006b)
Pukkala et al.
(2009)
4-241
-------
Table 4-30. Results of epidemiologic studies of potential
tetrachloroethylene exposure and adult non-Hodgkin lymphoma, by study
design (continued)
Cancer type,
n exposed
cases
42
Relative Risk
(95% CI)
1.0 (0.7, 1.4)
Adult non-Hodgkin lymphoma
1
3
3
5
7
9
10
10
12
14
14
16
35
3.0 (0.2, 49.2)
0.9 (0.3, 2.4)
1.6(0.3,9.1)
0.75 (0.24, 2.32)
1.59(0.49,5.13)
1.6 (0.6, 4.0)
0.7(0.3, 1.5)
4.6(1.9, 11.4)
3.89(1.06, 14.2)
1.2 (0.6, 2.5)
1.1(0.6,2.0)
2.0 (0.97, 4.3)
1.0 (0.6, 1.6)
Adult non-Hodgkin lymphoma
14
31
78
87
0.6(0.3, 1.1)
1.4(1.0,2.0)
1.20 (0.94, 1.52)
1.38(1.08, 1.70)
Design, location, exposure assessment"
nested case-control, Sweden, Denmark, Finland, Norway,
census occupation codes and pension data/questionnaires
(dry cleaners)
Case-control
France, jobs held 6 or more mo, launderer and dry cleanerb
Canada, job exposure matrix forPCE (any exposure)
Italy, job titles (launderer, dry cleaner, presser) (males)
New Zealand, occupational title (textile bleaching, dyeing
and cleaning machine operators)
United States, self-reported exposure to dry-cleaning fluids
United States, laundering, dry cleaning, leather products
fabrication0
Italy, job titles (launderer, dry cleaner, presser) (females)
Sweden, JEM using serf-reported information (any styrene,
TCE, PCE, or benzene exposure)
United States, occupation title (launders and ironers)
Italy, job exposure matrix (PCE, medium/high intensity)3' d
United States, ever employed in dry-cleaning industry6
United States, all jobs held >1 yr (dry-cleaning industry)
France, serf-reported exposure to dry-cleaning solvents
Geographic -based (residential exposure)
Finland (Hausjarvi), PCE and other solvents in drinking
water
Finland (Hattula), PCE and other solvents in drinking
water
New Jersey, PCE in town water >5 ppb (males)
New Jersey, PCE in town water >5 ppb (females)
Reference
Lynge et al. (2006)
Clavel et al. (1998)
Siemiatycki (1991)
Costantini et al.
(2001)
't Mannetje et al.
(2008)
Kato et al. (2005)
Scherr et al. (1992)
Miligi et al. (1999)
Hardell et al.
(1981)
Schenk et al.
(2009)
Miligi et al. (2006)
Malone et al.
(1989)
Blair et al. (1993)
Fabbro-Peray et al.
(2001)
Vartiainen et al.
(1993)
Vartiainen et al.
(1993)
Vartiainen et al.
(1993)
Cohn et al. (1994)
Cohn et al. (1994)
4-242
-------
Table 4-30. Results of epidemiologic studies of potential tetrachloroethylene
exposure and adult non-Hodgkin lymphoma, by study design (continued)
a Studies with relatively high quality exposure-assessment methodologies, based on biological monitoring
data, cohort studies with job exposure matrix based on historical industrial monitoring data, or case-control
studies with job exposure matrix focusing on PCE based on information on job title and tasks or duties, and
additional modules for specific jobs, or studies of residential PCE exposure using a statistical model of
water distribution system to estimate delivered dose to a subject's home.
b Includes patients with hairy cell leukemia.
0 Number of exposed cases estimated based on report of a prevalence of 3% in the population (n cases =
303); job history limited to most recent job, job held 15 yr ago, major occupation, and second most major
occupation.
d Includes patients with non-Hodgkin lymphoma and chronic lymphocytic leukemia.
e Includes patients with chronic lymphocytic leukemia.
Travier et al. (2002) examined cancer incidence from 1971 through 1989. The relative
risk among women who reported work as a launderer, dry cleaner, or presser in the laundry,
ironing, or dyeing industry in 1960 and 1970 was 2.53 (95% CI: 1.44, 4.46), and among men, the
relative risk was 0.93 (95% CI: 0.30, 2.28). Ji and Hemminki (2005b) used a similar approach,
with cancer incidence ascertained through 2002. The start of follow-up began at the time of the
relevant census data (i.e., 1961 for analyses based on jobs held in 1960). The SIR among women
who worked as a launderer or dry cleaner in 1970 was 1.30 (95% CI: 1.03, 1.60), and the SIR
among men who worked as a launderer or dry cleaner in 1960 was 0.84 (95% CI: 0.62, 1.09).
The latter time period was used for women because of the increase of women in the workforce
during the 1960s. A limitation of these studies is the lack of detailed information pertaining to
job tasks for individuals, information that could be particularly useful with respect to the
interpretation of the observed gender-related differences. No increased risk was observed in the
cohort study of aerospace workers using a job exposure matrix to estimate tetrachloroethylene
exposure [SMR: 0.55, 95% CI: 0.18, 1.29 in Boice et al. (1999)1. The number of exposed cases
in the case-control studies range from 2 to 7 leukemia cases. The odds ratio in the study with a
relatively strong exposure-assessment methodology was 1.0 (95% CI: 0.4, 2.7) (Costantini et al.,
2008). The three geographic-based studies of residential exposure involved 7 to 64 exposed
cases. The case-control study in Cape Cod, MA, that estimated exposure using a statistical
model of the water distribution reported an adjusted odds ratio of 2.13 (95% CI: 0.88, 5.19) for
any tetrachloroethylene exposure and 8.33 (95% CI: 1.53, 25.29) for exposures above the 90th
percentile (Aschengrau et al., 1993). Relative risk estimates were lower, ranging from 0.7 to 1.2,
in two other residential studies with poorer quality exposure-assessment approaches (Cohn et al.,
1994: Vartiainen et al., 1993).
The data pertaining to non-Hodgkin lymphoma are more extensive, with 14 cohort
studies ranging in size from 3 (Anttila et al., 1995) to 264 (Pukkala et al., 2009) cases,
13 publications based on case-control studies from six countries ranging in size from 3
4-243
-------
(Siemiatycki, 1991) to 35 exposed cases (Fabbro-Peray et al., 2001), and two geographic-based
studies of residential exposures through drinking water (Cohn et al., 1994; Vartiainen et al.,
1993) (refer to Table 4-30). Six of the relative risk estimates from the cohort studies, including
the four with the largest number of non-Hodgkin lymphoma cases, were between 0.95 and 1.05
(Pukkala et al., 2009: Ji and Hemminki, 2006b, 2005b: Andersen et al., 1999). Among the nine
smaller cohorts (n cases <30) (Calvert et al., 2011: Radican et al., 2008: Blair et al., 2003:
Travier et al., 2002: Cano and Pollan, 2001: Andersen et al., 1999: Boiceetal., 1999: Anttila et
al., 1995: Lynge and Thygesen, 1990), three effect estimates were between 0.86 and 1.03, and
six ranged from 1.46 to 3.76. Five cohort studies using relatively high quality exposure-
assessment methods reported the highest relative risks, but these studies were also based on only
2 to 18 exposed cases, so the estimates are imprecise: RR: 2.35 (95% CI: 0.52, 10.7) for females
and 2.32 (95% CI 0.75, 7.15) for males in Radican et al. (2008): RR: 3.76 (95% CI: 0.77, 11.0)
in Antilla et al. (1995): RR: 1.70 (95% CI: 0.73, 3.34) in Boice et al. (1999): SIR: 2.02 (95% CI:
1.13, 3.34) for males and 1.14 (95% CI: 0.68, 1.68) for females in Selden and Ahlborg (2011):
and SMR: 2.46 (95% CI: 0.90, 5.36) in Calvert et al. (Calvert et al., 2011). Results from the
case-control studies are also quite variable, with ORs ranging from 0.7 to 4.6 (Schenk et al.,
2009: Lynge et al., 2006: Miligi et al., 2006: Fabbro-Perav et al., 2001: Miligi et al., 1999: Blair
etal., 1993: Siemiatvcki, 1991: Malone et al., 1989: Hardell et al., 1981). The studies with the
higher quality exposure estimate reported ORs of 1.2 (95% CI: 0.6, 2.5) and 1.0 (95% CI: 0.7,
1.4) (Lynge et al., 2006: Miligi et al., 2006). Both of the geographic studies provide some
evidence of an association between residential exposures via drinking water. Cohn et al. (1994)
reported RR: 1.38 (95% CI: 1.08, 1.70) in females and RR: 1.20 (95% CI: 0.94. 1.52) for
residence in a town with municipal water supplies containing >5 ppb tetrachloroethylene. In the
second, a study of two towns with tetrachloroethylene and other solvents in the drinking water in
Finland, an association was observed in one town (SIR: 1.4, 95% CI: 1.0, 2.0) but not in the
other (SIR: 0.6, 95% CI: 0.3, 1.1) (Vartiainen et al., 1993). The ability of these studies to
provide clear and specific evidence pertaining to cancer hazard and tetrachloroethylene is limited
by their ecological designs and examination of several solvents in addition to
tetrachl oroethy 1 ene.
Six studies provide data pertaining to tetrachloroethylene and Hodgkin lymphoma (refer
to Table 4-31). Four cohort studies (Pukkala et al., 2009: Blair etal., 2003: Travier et al., 2002:
Andersen et al., 1999) and one case-control study reported in two published papers (Costantini et
al., 2001: Miligi etal., 1999) examine risk among laundry and dry-cleaning workers, and one is a
geographic-based study of drinking water exposure in two towns in Finland (Vartiainen et al.,
1993). No association is observed in the largest cohort, with 33 cases in the cohort of laundry
and dry-cleaning workers from 5 Nordic countries (SIR: 0.97, 95% CI: 0.67, 1.36) (Pukkala et
4-244
-------
al., 2009). A two- to threefold increased risk is observed in each of the smaller occupational
studies, with number of cases ranging from 4 to 19 (Blair et al., 2003; Travier et al., 2002;
Andersen et al., 1999). The exposure-assessment methodology in these studies is relatively
limited, and none were considered to be of high quality.
The studies of multiple myeloma are summarized in Table 4-31. As was observed in the
compilation of studies of other types of lymphopoietic cancers, the larger cohort studies that use
a relatively nonspecific exposure measure (broad occupational title of launderers and dry
cleaners, based on census data) do not report an increased risk, with effect estimates ranging
from 0.99 to 1.07 (Pukkala et al.. 2009: Ji and Hemminkl 2006b: Andersen et al.. 1999). Results
from the cohort and case-control studies with a higher quality exposure-assessment
methodology, with an exposure measure developed specifically for tetrachloroethylene, do
provide evidence of an association, however, with relative risks of 7.84 (95% CI: 1.43, 43.1) in
women and 1.71 (95% CI: 0.42, 6.91) in men in the cohort of aircraft maintenance workers
(Radican et al.. 2008) and 1.5 (95% CI: 0.8, 2.9) in the case-control study (Gold et al.. 2010b).
Boice et al. (1999) also used a relatively high quality exposure measure, but because the results
Table 4-31. Results of epidemiologic studies of potential tetrachloroethylene
exposure and adult Hodgkin lymphoma and multiple myeloma, by study
design
Cancer type,
n exposed
cases
Relative Risk
(95% CI)
Hodgkin
4
5
19
33
2.69(1.01,7.19)
2.0 (0.6, 4.6)
1.88(1.13,2.93)
0.97 (0.67, 1.36)
Hodgkin
1
7
2.5 (0.3, 24.6)
3.5(1.5,8.2)
Hodgkin
6
0.8 (0.3, 1.7)
Design, location, exposure assessment"
Cohort
laundry and dry-cleaning workers and pressers, Sweden,
census occupation codes
laundry and dry-cleaning workers (SMR), United States,
union employment records
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, census occupation codes
(females)
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, Iceland, census occupation
codes
Case-control
Italy, job titles (launderer, dry cleaner, presser) (males)
Italy, job titles (launderer, dry cleaner, presser) (females)
Geographic -based
Finland (Hausjarvi), PCE in drinking water
Reference
Travier et al. (2002)
Blair et al. (2003)
Andersen et al.
(1999)
Pukkala et al. (2009)
Costantini et al.
(2001)
Miligi et al. (1999)
Vartiainen (1993)
4-245
-------
11
1.4 (0.7, 2.5)
Finland (Hattula), PCE in drinking water
Vartiainen (1993)
Table 4-31. Results of epidemiologic studies of potential
tetrachloroethylene exposure and adult Hodgkin lymphoma and multiple
myeloma, by study design (continued)
Cancer type,
n exposed
cases
Relative Risk
(95% CI)
Multiple myeloma
1
2
3
7
7
36
45
52
152
0.40 (0.01, 2.25)
7.84(1.43,43.1)
1.71(0.42,6.91)
0.8 (0.3, 1.6)
1.75 (0.70, 3.61)
1.07 (0.75, 1.45)
1.0(0.73, 1.34)
0.99 (0.66, 1.38)
1.02 (0.86, 1.20)
Multiple myeloma
3
9
16
1.0(0.3,3.8)
6.0(1.7,21)
1.5 (0.8, 2.9)
Multiple myeloma
6
7
0.7 (0.2, 1.3)
0.7(0.3, 1.3)
Design, location, exposure assessment"
Cohort
aerospace workers (SMR), United States, job exposure
matrix (PCE routine exposure)3
aircraft maintenance workers (PJl-internal referent),
United States job exposure matrix (females)3
aircraft maintenance workers (PJl-internal referent),
United States, job exposure matrix (males)3
laundry and dry-cleaning workers (SMR), United States,
union records (all workers)
laundry and dry-cleaning workers (SIR), Sweden, census
occupation codes
laundry and dry-cleaning workers (SIR), Sweden
(females)
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, census occupation codes
laundry and dry-cleaning workers (SIR), Sweden (males)
laundry and dry-cleaning workers (SIR), Sweden,
Denmark, Finland, Norway, Iceland, census occupation
codes
Case-control
Italy, job titles (launderer, dry cleaner, presser) (females)
United States, all jobs held >12 mo (textile, apparel,
furnishing machine operators and tenders)
United States, all jobs held >12 mo, job exposure matrix
(PCE)3'b
Geographic -based
Finland (Hattula), PCE in drinking water
Finland (Hausjarvi), PCE in drinking water
Reference
Boice et al. (1999)
Radican et al. (2008)
Radican et al. (2008)
Blair et al. (2003)
Lynge and Thygesen
(1990)
Ji and Hemminki
(2006b)
Andersen et al.
(1999)
Ji and Hemminki
(2006b)
Pukkala et al. (2009)
Miligi et al. (1999)
Gold et al. (2010a)
Gold et al. (2010b)
Vartiainen (1993)
Vartiainen (1993)
3 Studies with relatively high quality exposure-assessment methodologies, based on biological monitoring
data, cohort studies with job exposure matrix based on historical industrial monitoring data, or case-control
studies with job exposure matrix focusing on PCE based on information on job title and tasks or duties, and
additional modules for specific jobs, or studies of residential PCE exposure using a statistical model of
water distribution system to estimate delivered dose to a subject's home.
4-246
-------
b Results for analysis in which low confidence jobs were considered unexposed. Similar results observed
in the primary analysis in which low confidence jobs were included in the exposure group.
are based on only one observed case, the imprecision of the estimate (RR: 0.40, 95% CI: 0.01,
2.25) limits this study for insights on multiple myeloma and tetrachloroethylene.
Variation in risk in relation to variation in exposure levels is examined in one study of
lymphopoietic cancer (Blair et al., 2003), five studies of non-Hodgkin lymphoma (Lynge et al.,
2006: Miligi et al.. 2006: Boice etal., 1999: Blair etal.. 1993) or of non-Hodgkin combined with
Hodgkin lymphoma (Seidler et al., 2007), four studies of multiple myeloma (Gold etal., 2010a:
Gold etal., 2010b: Seidler et al., 2007: BoiceetaL, 1999) and two studies of leukemia (Miligi et
al., 2006: Aschengrau and Seage, 2003) (refer to Table 4-32). Gold et al. (201 Ob) and Seidler et
al. (2007) examined exposure gradients using a cumulative tetrachloroethylene measure. The
aerospace worker cohort study by Boice et al. (1999), the dry cleaners cohort study by Blair et al.
(2003), and the Italian case-control studies (Costantini et al., 2008: Miligi et al., 2006) used a
semi quantitative measure of exposure intensity or frequency, and two studies used a less-specific
measure of job duration (Gold etal., 2010a: Lynge et al., 2006). Inability to account for
temporal changes in exposure intensity makes duration an inferior exposure surrogate compared
to semi quantitative or quantitative measures. The tetrachloroethylene-based measures in the
non-Hodgkin lymphoma studies (Seidler et al., 2007: Miligi et al., 2006: Boice et al., 1999)
provide evidence of a higher risk at the higher exposure levels, particularly in the highest
category of cumulative exposure (>78.8 ppm-years) in the case-control study by Seidler et al.
(2007). Similar results are observed in one of the multiple myeloma studies (Gold etal., 2010b),
but the smaller study by Seidler et al. (2007) observed no cases among the highest exposure
groups (refer to Table 4-32).
There is considerable variation in the databases (e.g., number of studies, study design,
and quality of the exposure assessment) for the different types of lymphopoietic cancers. In
general, studies with relatively strong exposure assessments are based on a small number of
observed deaths or incident cases, with a relatively low statistical power resulting from few
observed events, or, for population case-control studies, low exposure prevalence. For
non-Hodgkin lymphoma and multiple myeloma, the presence of higher relative risk estimates in
studies with better exposure-assessment methodologies and evidence of an exposure-response
trend in one or more studies provide the basis for considering the collection of studies as
supportive of a role of tetrachloroethylene as a likely carcinogen. The collection of studies for
leukemia, non-Hodgkin lymphoma, Hodgkin lymphoma, and multiple myeloma is summarized
below.
There is little evidence for an association with leukemia. The two studies with a
relatively high quality exposure-assessment methodology had few exposed cases (<7) and did
4-247
-------
not provide evidence of an association (RRs of 0.55 and 1.0 in Boice et al. (1999) and Costantini
et al. (2008), respectively), although a case-control study reported a twofold increased risk of
leukemia with the highest exposure level of tetrachloroethylene-contaminated drinking water
(Aschengrau et al., 1993). The results from studies using more general (i.e., nonspecific)
exposure methods (e.g., occupational codes for laundry or dry-cleaning workers) generally
showed no association with leukemia (i.e., relative risk estimates <1.0 in 6 of the 9 cohorts)
(Pukkala et al.. 2009: Sung et al.. 2007: Blair et al.. 2003: Andersen et al.. 1999: Boice et al..
1999: Lynge and Thygesen, 1990). Two of the increased leukemia relative risks (RR of 2.53 and
1.30) were observed in studies limited to female workers, which may represent a more
homogenous group in terms of potential exposures (Ji and Hemminki, 2005b: Travier et al.,
2002).
Table 4-32. Results of epidemiologic studies of potential tetrachloroethylene
exposure and adult lymphopoeitic cancers, with data pertaining to exposure-
response gradients, by cancer type
Cancer type
Lymphopoeitic
Non-Hodgkin
Exposure measure
Exposure score
Little to no
Medium to high
Job duration (yr)
0
>0-<1
2-4
5-9
>10
PCE (duration, yr)
0
<1
1-4
>5
PCE (intensity)
Very low/low
Medium/high
PCE (duration, yr)
<15
>15
Results
n
18
17
145
5
3
14
15
32
4
6
10
18
14
10
o
6
RR (95% CI)
1.0 (0.6, 1.5)
0.9 (0.5, 1.4)
1.0 (referent)
1.35(0.44,4.14)
0.61(0.17,2.21)
0.92 (0.49, 1.72)
0.66 (0.36, 1.22)
1.0 (referent)
1.25 (0.43, 3.57)
1.11(0.46,2.70)
1.41 (0.67, 3.00)
(trend p> 0.20)
0.6 (0.3, 1.2)
1.2 (0.6, 2.5)
(trend p = 0.12)
1.3(0.5,3.3)
not reported3
Design, location, exposure
assessment
Cohort, laundry and dry-
cleaning workers, union
records (exposure score
based on proximity to
washers)
Nested case-control within
cohort of laundry and dry-
cleaning workers, Sweden,
Denmark, Finland, Norway,
census occupation codes3
Cohort, aerospace workers,
job exposure matrix (routine
or intermittent exposure to
PCE)
Case-control, Italy, job
exposure matrix
Case-control, Italy, job
exposure matrix
Reference
Blair et al.
(2003)
Lynge et al.
(2006)
Boice et al.
(1999)
Miligi et al.
(2006)
Miligi et al.
(2006)
4-248
-------
PCE (cumulative, ppm-yr)
0
>0-<9.1
>9.1-<78.8
>78.8
67
16
14
6
1.0 (referent)
1.1 (0.5,2.3)
1.0 (0.5,2.2)
3.4 (0.7, 17.3)
(trend;? = 0.12)
Case-control, Germany
(PCE, job exposure matrix)b
(Includes non-Hodgkin and
Hodgkin lymphoma; similar
results observed with
B -non-Hodgkin)
Seidler et al.
(2007)
Table 4-32. Results of epidemiologic studies of potential
tetrachloroethylene exposure and adult lymphopoeitic cancers, with data
pertaining to exposure-response gradients, by cancer type (continued)
Cancer type
Multiple
myeloma
Leukemia
Exposure measure
Job duration (yr)
1-5
>5
PCE duration (yr)
0
<1
1-4
>5
PCE duration (yr)
1-4
5-11
12-29
3-51
PCE (cumulative)
1-318
319-2,218
2,219-7,713
7,794-57,000
PCE (cumulative)
0
>0-<9.1 ppm-yr
>9.1-<78. 8 ppm-yr
>78.8 ppm-yr
PCE intensity
Very low/low
Medium/high
Any PCE
ROD >90th percentile
Results
n
4
5
24
1
4
1
3
3
4
6
1
1
4
10
5
6
0
0
6
7
7
2
RR (95% CI)
3.6 (0.7, 1.7)
12 (1.3, 110)
(trend p< 0.01)
1.0 (referent)
0.46 (0.06, 3.48)
1.13(0.38,3.35)
0.24 (0.03, 1.84)
(trend;? < 0.01)
0.9 (0.2,3.5)
2.0 (0.4,9.2)
1.3 (0.3,4.6)
2.1 (0.7,6.8)
(trend;? = 0.18)
0.3 (0.04,3.0)
0.3 (0.1,4.4)
1.5 (0.4,5.4)
3.3 (1.2,9.5)
(trend;? = 0.02)
1.0 (referent)
1.8 (0.5,6.7)
(inverse trend
/> = 0.34)
0.6 (0.2, 1.6)
1.0 (0.4,2.7)
2.13 (0.88,5.19)
8.33 (1.53,25.3)
Design, location, exposure
assessment
Case-control, United States,
all jobs held >12 mo (textile,
apparel, furnishing machine
operators and tenders)
Cohort, aerospace workers,
job exposure matrix (routine
or intermittent exposure to
PCE)
Case-control, United States,
all jobs held >12 mo (PCE,
job exposure matrix)0
Case-control, United States,
all jobs held >12 mo (PCE,
job exposure matrix)0
Case-control, Germany
(PCE, job exposure matrix)b
Case-control, Italy, job
exposure matrix (PCE)
Geographic based, United
States, water distribution
model (any PCE)
Reference
Gold et al.
(2010a)
Boice et al.
(1999)
Gold et al.
(2010b)
Gold et al.
(2010b)
Seidler et al.
(2007)
Costantini et
al. (2008)
Aschengrau
et al. (1993)
4-249
-------
a Relative risk estimates only reported for strata with at least five exposed cases.
b Cumulative score based on summation of the product of intensity score (low, 5 ppm; medium, 50 ppm; high,
200 ppm), frequency score (low, 3%; medium, 7.5%; high, 65%) of workweek, and duration for each job.
0 Results for analysis in which low confidence jobs were considered unexposed. Similar results observed in the
primary analysis in which low confidence jobs were included in the exposure group. Cumulative measure based on
summation of the product of intensity (ppm), frequency (h/wk), and duration (yr) for each job.
The results from the collection of studies pertaining to non-Hodgkin lymphoma indicate
an elevated risk associated with tetrachloroethylene exposure. The results from five cohort
studies that used a relatively high quality exposure-assessment methodology generally reported
relative risks between 1.7 and 3.8 (Calvert et al., 2011; Selden and Ahlborg, 2011; Radican et al.,
2008; Boice etal., 1999; Anttila et al., 1995) and support an association with tetrachloroethylene.
The studies with tetrachloroethylene-specific exposure measures and exposure-response analysis
(based on intensity, duration, or cumulative exposure) (Seidler et al., 2007; Miligi et al., 2006;
Boice etal., 1999) provide further support for an association, reporting higher non-Hodgkin
lymphoma risks in the highest exposure category, with the strongest evidence from the large
case-control study in Germany, in which a relative risk of 3.4 (95% CI: 0.7, 17.3) was observed
in the highest cumulative exposure category (trends-value = 0.12) (Seidler et al., 2007). Lynge
et al. (2006) distinguished dry cleaners from other workers but used an approach with greater
potential for misclassification because exposure was assigned only for jobs held in 1970. This
study did not report an association between dry cleaners and non-Hodgkin lymphoma, nor did
risk estimates increase with exposure duration. Effect estimates in studies with broader exposure
assessments showed a more variable pattern (Selden and Ahlborg, 2011; Pukkala et al., 2009; Ji
and Hemminki, 2006b; Blair et al.. 2003; Travier et al.. 2002; Cano and Pollan, 2001; Lynge and
Thygesen, 1990). Confounding by lifestyle factors are unlikely explanations for the observed
non-Hodgkin lymphoma results because common behaviors, such as smoking and alcohol use,
are not strong risk factors for non-Hodgkin lymphoma (Besson et al., 2006; Morton et al., 2005).
With respect to Hodgkin lymphoma, the data are more limited, with only four cohort
studies (Pukkala et al.. 2009; Blair etal.. 2003; Travier et al.. 2002; Andersen et al.. 1999). one
case-control study from Italy reported in two publications (Costantini et al., 2001; Miligi et al.,
1999), and one geographic-based study from Finland (Vartiainen et al., 1993). None of the
exposure-assessment methods used in these studies were considered to be relatively high quality.
A two- to threefold increased risk is observed in all of the occupational studies except Pukkala et
al. (2009) [SIR: 0.97 (95% CI: 0.67, 1.36)].
The larger cohort studies that use a relatively nonspecific exposure measure (broad
occupational title of launderers and dry cleaners, based on census data) do not report an
increased risk of multiple myeloma, with effect estimates ranging from 0.99 to 1.07 (Pukkala et
al., 2009; Ji and Hemminki, 2006b; Andersen et al., 1999). Some uncertainty in these estimates
4-250
-------
arises from these studies' broader exposure-assessment methodology. Results from the cohort
and case-control studies with a higher quality exposure-assessment methodology, with an
exposure measure developed specifically for tetrachloroethylene, do provide evidence of an
association, however, with relative risks of 7.84 (95% CI: 1.43, 43.1) in women and 1.71 (95%
CI: 0.42, 6.91) in men in the cohort of aircraft maintenance workers (Radican et al., 2008) and
1.5 (95% CI: 0.8, 2.9) in the case-control study in Washington [Gold et al. (201 Ob):
tetrachloroethylene exposure]. Gold et al. (2010a: 201 Ob) also reported increasing risks with
increasing exposure duration [based on job titles (Gold et al., 2010a) and based on a cumulative
tetrachloroethylene exposure metric (Gold et al., 201 Ob)]. A smaller case-control study (n = 76
cases) with tetrachloroethylene-specific exposure measures based on intensity, duration, or
cumulative exposure, Seidler et al. (2007), observed no cases among the highest exposure
groups. A small study by Boice et al. (1999) of aerospace workers observed one death among
routinely exposed subjects and six deaths among subjects with a broader definition of routine or
intermittent exposure.
4.6.1.2.5. Childhood leukemia
One cohort and four case-control studies are available on childhood leukemia (acute
lymphocytic leukemia, ALL) and parental occupational exposure to tetrachloroethylene or to
drinking water contaminated with trichloroethylene, tetrachloroethylene, and other chlorinated
solvents (Sung et al.. 2008: Infante-Rivard et al.. 2005: Costas et al.. 2002: Shuetal., 1999:
Lowengart et al., 1987) (Table 4-28; Appendix B). Some studies suggest a vulnerability for ALL
with maternal exposure either preconception or during pregnancy (Sung et al., 2009: Costas et
al., 2002: ShuetaL, 1999: Lowengart et al., 1987). These studies, however, are insensitive for
assessing association, or lack thereof, between ALL and tetrachloroethylene exposure because
observations are based on a few exposed cases (all studies) or a weak exposure assessment (Sung
et al., 2008). Only Lowengart et al. (1987) and Shu et al. (1999) examined paternal exposure and
tetrachloroethylene exposure with inconsistent observations. Other studies are needed to clarify
the role of tetrachloroethylene in ALL.
4.6.2. Animal Studies
4.6.2.1. Noncancer Effects
4.6.2.1.1. Immunotoxicity
The animal evidence for immunotoxicity following exposure to tetrachloroethylene is
very limited. These studies consist of mixed solvent exposures and some inhalation and oral
studies in which experimental animals were dosed with tetrachloroethylene alone.
4-251
-------
Immune system parameters were altered in a mouse study (female B6C3Fi) administered
tetrachloroethylene by inhalation (maximum concentration: 6.8 ppm) along with a mixture of
24 contaminants frequently found in ground water near Superfund sites. Exposure lasted 14 or
90 days, and mice were sacrificed to assess immune system parameters. Evidence of
immunosuppression was observed, with a dose-related decrease in antibody response to sheep
red blood cells and decreased host resistance following challenge to Plasmodiumyoelli. There
were no changes in lymphocyte number, T-cell subpopulations, NK cell activity, or in response
to challenge to Listeria monocytgens or PYB6 tumor cells. While these findings may be
attributed to B-cell/humoral immunity, these effects cannot be attributed to tetrachloroethylene
alone (Germolec et al., 1989).
Aranyi et al. (1986) studied the effects of acute inhalation exposures to 25 or 50 ppm
tetrachloroethylene on two measures of immune response (susceptibility to respiratory infection
and mortality due to Streptoccocus zooepidemicus exposure and ability of pulmonary
macrophages to clear infection with Klebsiellapneumoniae). Female CD1 mice that were
5-7 weeks of age at the start of the exposure portion of the experiment were used for both
assays. Up to five replicate groups of about 30 mice were challenged with viable
S. zooepidemicus during simultaneous exposure to tetrachloroethylene or to filtered air. Deaths
were recorded over a 14-day observation period. Clearance of 35S-labeled K. pneumoniae by
pulmonary macrophages was determined by measuring the ratio of the viable bacterial counts to
the radioactive counts in each animal's lungs 3 hours after infection; 18 animals were used per
dose group. A single 3-hour exposure to 50 ppm tetrachloroethylene significantly increased the
susceptibility to respiratory infection and greater mortality following exposure to
S. zooepidemicus (p < 0.01). Forty-four deaths occurred in 140 (31.4%) mice challenged during
a 3-hour exposure to 50 ppm tetrachloroethylene; in contrast, 21 deaths occurred in 140 mice
(15.0%) exposed to filtered air. The 3-hour exposure to 50 ppm tetrachloroethylene was
associated with a statistically significant (p < 0.05) 6.6% decrease in pulmonary bactericidal
activity (80.5 and 73.9% of bacteria killed in controls and 50 ppm group, respectively). No
difference was observed in either mortality rate or bactericidal activity in experiments using a
single 3-hour exposure to 25 ppm, or 3-hour exposures to 25 ppm tetrachloroethylene repeated
daily for 5 days compared with control animals exposed to filtered air.
In a study by Hanioka et al. (1995a), atrophy of the spleen and thymus was observed in
rats receiving 2,000 mg/kg-day tetrachloroethylene via corn oil gavage for 5 days. No effect was
observed in the 1,000 mg/kg-day group. In a 14-day corn oil gavage (1,000 mg/kg-day) study of
tetrachloroethylene, no effects were observed on thymus and spleen weights of adult rats at a
dose that produced liver toxicity (Berman et al., 1995). Another study employed 3 daily i.p.
4-252
-------
doses of tetrachloroethylene to mice (Schlichting et al., 1992). No effects were observed on ex
vivo natural killer cell activity or humoral responses of T-cells to exogenous mitogens.
Additional data from inhalation, oral, and dermal exposures of different durations are
needed to assess the potential immunotoxicity of tetrachloroethylene along multiple dimensions,
including immunosuppression, autoimmunity, and allergic sensitization. The data from Aranyi
et al. (1986) suggest that short-term exposures may result in decreased immunological
competence (immunosuppression) in CD-I mice. The relative lack of data, taken together with
the concern that other structurally related solvents (Cooper et al., 2009) have been associated
with immunotoxicity, contributes to uncertainty in the database for tetrachloroethylene.
4.6.2.1.2. Hematologic toxicity
Several studies by Marth (1987) or Marth et al. (1989: 1985a: 1985b) and a study by
Seidel et al. (1992) have demonstrated hematopoietic toxicity of tetrachloroethylene in female
mice. In the Marth (1987) and Marth et al. (1989: 1985a: 1985b) studies, 135 female NMRI
mice were exposed in drinking water to tetrachloroethylene at 0.05 or 0.1 mg/kg per day
beginning at 2 weeks of age for 7 weeks and examined 8 or 16 weeks after exposure cessation.
The mice exhibited a reversible hemolytic anemia and had microscopic evidence of splenic
involvement (Marth et al., 1985a: Marth et al., 1985b). Tetrachloroethylene was found to
accumulate in the spleen to a significantly greater extent than in the liver, brain, or kidney; levels
of tetrachloroethylene were 20-fold higher in spleen than in liver at the end of the exposure
period (Marth, 1987). Tetrachloroethylene was found in the spleen and fatty tissue of test
animals up to 2 months (56 days) after initial exposure (Marth et al., 1989). Reversible body-
weight decreases and increases in the relative weight of the spleen compared with the kidneys
were reported. Serum triglycerides increased, and cholesterol levels decreased. These effects
persisted as long as 16 weeks after cessation of exposure. Liver function (as assessed by serum
protein levels) and hepatic protein synthesis were within normal limits, and there was no
evidence of hepatic fatty accumulation or necrosis. Compared with brain, kidney, or liver, the
erythropoietic system was found to be most susceptible to tetrachloroethylene in these studies.
Seidel et al. (1992) exposed female hybrid mice (C57/BL/6 x DBA/2) to
tetrachloroethylene at 270 ppm (11.5 weeks) and 135 ppm (7.5 weeks), 6 hours/day,
5 days/week. Reductions in the numbers of lymphocytes/monocytes and neutrophils were
observed, with a return to control values over the next 3 weeks. There were no effects on spleen
colony-forming units (CFU-Ss), but evidence of a reduction in red cells was supported by
decreases in erythroid colony-forming units and erythroid burst-forming units and evidence of
reticulocytosis. A partial regeneration was observed in the exposure-free follow-up period of
3 weeks. It was noted that the slight CFU-C depression, which persisted in the exposure-free
4-253
-------
period, could indicate the beginning of a disturbance at all progenitor cell levels. These data
suggest a reversible bone marrow depression.
Hematological parameters were examined following oral administration of
tetrachloroethylene in sesame oil (3,000 mg/kg-day for 15 days) to male albino Swiss mice with
and without concurrent administration of 2-deoxy-D-glucose (2DG; 500 mg/kg-day i.p.), vitamin
E (400 mg/kg-day oral gavage) or taurine (100 mg/kg-day by oral intubation) (Ebrahim et al.,
2001). This study was designed to examine the potential protective properties of 2DG and
vitamin E as well as taurine against tetrachloroethylene-induced cytotoxicity in various organ
systems. Animals exposed to tetrachloroethylene alone demonstrated significantly decreased
hemoglobin and RBC counts (p < 0.01), and significantly decreased HCT (packed cell volume)
and platelet counts (p < 0.001). The WBC count was found to be significantly increased
(p < 0.001). These changes were reverted back to near normal in the animals coexposed to 2DG,
vitamin E, or taurine.
In summary, the limited laboratory animal studies of hematological toxicity demonstrated
an effect of tetrachloroethylene exposure on RBC [decreased RBC (Ebrahim et al., 2001), or
decreased erythrocyte colony-forming units (Seidel et al., 1992)1, with reversible hemolytic
anemia observed in female mice exposed to low drinking water levels (0.05 mg/kg-day) of
tetrachloroethylene beginning at 2 weeks of age in one series of studies (Marth et al., 1989;
Marth, 1987: Marth et al., 1985a: Marth et al., 1985b). Ebrahim et al. (2001) also observed
decreased hemoglobin, platelet counts and packed cell volume, and increased WBC counts.
Although limited studies are available in the peer-reviewed published literature, the results of
these studies support the results observed in the study of dry-cleaning workers by Emara et al.
(2010) described in Section 4.6.1.1.1.
4.6.2.2. Cancer Effects
4.6.2.2.1. Mononuclear cell leukemia in rats
The incidence of mononuclear cell leukemia in rats chronically exposed to
tetrachloroethylene is summarized in Table 4-33. The NCI oral gavage study in
Osborne-Mendel rats was considered to be inconclusive because of the high incidence of
respiratory disease, and high mortality with tetrachloroethylene exposure. Lesions indicative of
pneumonia were observed in almost all rats at necropsy. A high incidence of toxic nephropathy
was evident in tetrachloroethylene-exposed male and female rats. Early mortality was also
observed in tetrachloroethylene-exposed animals; 50% of the high dose males and females had
died by Weeks 44 and 66, respectively. Therefore, this bioassay is not considered further in the
below evaluation of the mononuclear cell leukemia induction by tetrachloroethylene in rats.
4-254
-------
NTP (1986) reported that the chronic inhalation administration of tetrachloroethylene at
concentration levels of 0, 200, and 400 ppm caused statistically significant positive trends in the
incidence of MCL in male (p = 0.004) and female (p = 0.018) F344/N rats. The incidences of
MCL in male and female rats exposed to tetrachloroethylene at 0, 200, and 400 ppm
(6 hours/day, 5 days/week, for 104 weeks) were 56, 77, and 74% and 36, 60, and 58%,
respectively. Interpretation of these data is somewhat complicated by the fact that overall
incidences of MCL in the concurrent chamber control groups were high relative to historical
chamber control groups at the performing laboratory (males: 28/50 [56%] vs. 117/250 [47%];
females: 18/50 [36%] vs. 73/249 [29%]). The concurrent control group rates were also higher
than the NTP program historical rate for untreated control groups (males: 583/1,977 [29%];
females: 375/2,021 [18%]).
Table 4-33. Mononuclear cell leukemia incidence in rats exposed to
tetrachloroethylene
Bioassay
NCI (1977V3
Osborne-Mendel rats
Gavage:
5d/wk,
78 wk
NTP (1986)
F344/N rats
Inhalation:
6h/d,
5d/wk,
104 wk
JISA (1993)
F344/DuCrj rats
Inhalation:
6h/d,
5d/wk,
104 wk
Exposure
Vehicle control
500 mg/kg-day
1,000 mg/kg-day
Vehicle control
500 mg/kg-day
1,000 mg/kg-day
Oppm
200 ppm
400 ppm
Oppm
200 ppm
400 ppm
Oppm
50 ppm
200 ppm
600 ppm
Oppm
50 ppm
200 ppm
600 ppm
Sex
Male
Female
Male
Female
Male
Female
Mononuclear cell leukemia
incidence (%)a
None reported
None reported
28/50 (56)
37/50 (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)
a Reflects the number of animals with MCL reported under "multiple organs," spleen, or liver.
b Gavage doses listed were adjusted several times during the course of the study. Male rats
received the listed TWA daily doses through Week 78, and surviving animals were observed up
to study termination in Week 110.
4-255
-------
To evaluate whether the increased MCL incidence contributed to the increase in early
deaths observed with increasing tetrachloroethylene exposure, NTP (1986) conducted
supplemental analyses according to their standard methods of data evaluation. These analyses
considered the progression of the disease, the effect of tetrachloroethylene on the time of onset of
advanced MCL, and the contribution of MCL to early deaths in control and dosed animals. The
results of these supplemental analyses showed the following:
• In both males and females, tetrachloroethylene produced a dose-related increase in the
severity of MCL.
• Tetrachloroethylene exposure significantly shortened the time to onset of MCL in female
rats.
• Although there was no notable effect of tetrachloroethylene exposure on survival of
female rats, there was an increased incidence of advanced MCL in female rats that died
before the scheduled termination of the study. Thus, statistical analyses of only the
incidences of advanced MCL in rats were considered. Significantly positive trends and
significant increases in the incidences of advanced MCL were observed in both male and
female rats in the high-dose groups.
Thomas et al. (2007) reanalyzed the NTP (1986) dose-response data comparing results
with four statistical methods. In their analysis of MCL incidence in rats exposed to 500
chemicals, tetrachloroethylene was one of five chemicals shown by the authors to produce
"definitive" leukemia effects in both sexes of rats. MCL effects were more often than not
confined to one sex, while tetrachloroethylene induced statistically significant increases in both
sexes of the F344 rat. The findings in Thomas et al. (2007) described in more detail later, are
addressed in the context of other considerations in Section 4.6.2.2.2.
In the JISA (1993) study, F344/DuCrj rats were exposed via inhalation for 104 weeks to
tetrachloroethylene at concentrations of 0, 50, 200, and 600 ppm. As in the NTP (1986) study,
there was a higher control incidence of MCL (22% in males and 20% in females) than the
reported historical rate of MCL for the Japanese laboratory of 147/1,149 [13%] in males and
147/1,048 [14.0%] in females (refer to Table 5-16, Section 5). The incidence of MCL in male
and female rats exposed to tetrachloroethylene at 50, 200, and 600 ppm was 28, 44, and 54% and
34, 32, and 38%, respectively. Both male and female rats displayed a significant dose-dependent
increase in MCL, atp < 0.01 andp = 0.046 (poly-3 test, conducted for this assessment),
respectively. There was decreased latency in MCLs in female rats of the JISA (1993) study, with
first appearance in Week 100 in controls and Weeks 66-70 in treated rats.
4-256
-------
4.6.2.2.2. Additional considerations regarding rodent leukemia findings
Under the conditions of the NTP (1986) and JISA (1993) bioassays, a carcinogenic effect
of tetrachloroethylene in male and female rats was evidenced by significant increases of MCL in
both sexes. The pathology of rat MCL is well characterized and has been well described
(Thomas et al., 2007; Ward et al., 1990; Stromberg, 1985). MCL is among the most common
causes of death in the aging F344 rat and is readily and unequivocally diagnosed by standard
histopathological techniques. However, the utility of observed increases in MCL in the
chemically exposed rat for human carcinogenic risk assessment has been questioned for several
reasons. In particular, the spontaneous background incidence is both high and variable, and,
thus, can obscure chemical-induced increases. As noted in reviews by Caldwell (1999) and
Ishmael and Dugard (2006), the high background rate of MCL in control (untreated) rats can
limit the ability to separate the background response from possible chemically induced
responses, particularly when the chemically induced response above background is low.
Additionally, because high-incidence MCL occurs only in the F344 rat strain and not in mice,
Caldwell (1999) has stated that marginal increases in incidences are of questionable biological
significance. Supplemental analyses, such as have been conducted by NTP for
tetrachloroethylene and summarized in the preceding section, have been endorsed as a means to
aid in data interpretation for these commonly occurring tumors. In the paragraphs that follow,
issues pertinent to the interpretation of evidence that tetrachloroethylene induces MCL in male
and female rats for the purposes of human health risk assessment are addressed. The discussion
summarizes the findings of a recent analysis by Thomas et al. (2007) and considers the available
evidence for tetrachloroethylene in the context of the approach put forth by those authors. Other
considerations identified by NRC (2010) are also addressed, particularly with respect to
uncertainties surrounding the causes of F344 rat MCL, the biology of the disease, including the
cell type of origin, as well as the mechanisms by which tetrachloroethylene may advance
development of this rodent leukemia.
The significance of MCL findings in multiple NTP bioassays that used the F344 rat was
the subject of a recent reanalysis by Thomas et al. (2007). They examined the incidence of
leukemia in 2-yr bioassays that included untreated male and female F344 rats from 1971 to 1998.
They found that background tumor incidence increased substantially, from 7.9 to 52.5% in males
and from 2.1 to 24.2% in females, over that period. The reanalysis also found that MCL
responses are highly variable and subject to substantial modulation by dietary as well as other, as
yet unidentified, factors.
Their review of the disease pathobiology described MCL as a large granular lymphocytic
(LGL) leukemia that is a rapidly progressing and fatal neoplasm, with death typically occurring
within 2 weeks of onset (Thomas et al., 2007). The disease is characterized by splenomegaly
4-257
-------
upon gross pathological examination. Leukemic cell infiltration of the splenic red pulp with
variable lymphoid cell depletion is consistently seen. The tumor is transplantable; its etiological
factor is unknown. The cell of origin appears to reside in and/or require the splenic
microenvironment, and splenectomy dramatically reduces spontaneous MCL incidence
(Molonev and King. 1973).
Thomas et al. (2007) concluded that the exact cell of origin of F344 rat MCL is unknown.
The pathological characteristics of rat MCL are similar in some respects to one of the human
T-cell leukemias (Caldwell, 1999), and some investigators have proposed that MCL can serve as
an experimental model for human T-cell leukemia (Stromberg, 1985). However, MCLs have
been shown to be heterogeneous with respect to cell phenotype and function (e.g., surface
antigen expression, esterase activity, and cytotoxic activity). For example, a study of 10 primary
and 10 transplanted MCLs of aging rats found that natural killer (NK) cell activity was variable
and lacked correlation with surface antigens, with poorly differentiated MCL cells exhibiting less
cytotoxic (i.e., NK-cell) activity (Ward and Reynolds, 1983). These and other investigations
[e.g., Stromberg et al. (1983)1 have provided evidence that MCLs represent a heterogeneous
group of leukemias. Thomas et al. note that the use of specific monoclonal anti-rat NK-cell
antibodies and other rat leukocyte specific markers would aid in establishing the cell type of
origin. The lack of assessment of the rodent tumors according to current classification criteria
[e.g., as specified by Swerdlow et al. (2008)1 hinders ability to identify cell lineage. In
particular, the lack of immunophenotyping data for MCL occurring spontaneously or as the
result of chemical exposure, and the observed heterogeneity in cell phenotype and function of the
spontaneously occurring tumors studied thus far, greatly limit classification of MCL. Based on
the reported heterogeneity in cell phenotype and function, Thomas et al. (2007) stated that MCL
may arise from either mature LGLs or from a variety of individual LGL subpopulations;
alternatively, a pluripotent LGL precursor may be the cell type of origin.
Acknowledging the limitations that arise from the lack of knowledge about the cell type
of origin for MCL, and the observed heterogeneity in phenotype and function among MCL,
Thomas et al. (2007) characterize MCL as having an NK-cell phenotype based on functional
NK-cell activity in most (but not all) MCL cells. They note that human NK-LGL and F344 rat
MCL have "some characteristics in common" and conclude that F344 rat MCL "is comparable to
the aggressive human NK-LGL leukemia on a morphological, functional, and clinical basis."
However, current criteria to identify cell phenotype (e.g., by use of specific monoclonal
antibodies and genomic analysis) were not adopted in this study, and many of the comparison
criteria identify by Thomas et al. (2007) are nonspecific and common to other human leukemia
or lymphoma phenotypes. Although contrary to prior reports that the F344 MCL does not have a
human counterpart [e.g., Caldwell (1999)], a comparable conclusion regarding similarity of F344
4-258
-------
rat MCL to human NK-LGL was reached by Stromberg (1985) and Ishmael and Dugard (2006).
Human NK-LGL is a rare form of LGL. NK-LGL usually occurs in younger patients (median
age: 39), has an aggressive clinical course, and is usually fatal within months of diagnosis
despite multiagent therapy. Epstein Barr virus has been implicated in many of the reported
NK-LGL cases, although the mechanism is unknown. In contrast, the majority of other human
LGLs (i.e., T-cell LGL leukemias) follow a chronic indolent course. Due to the paucity of
available data, mechanisms or modes of action contributing to the MCLs arising in untreated or
chemically exposed F344 rats have not been identified.
Thomas et al. (2007) also evaluated MCL incidence in male and female rats exposed to
500 chemicals. On the basis of 34 NTP studies that yielded evidence of a chemically related
increase in the incidence of leukemia, which included the NTP (1986) study of
tetrachloroethylene, the authors conducted a reanalysis of dose-response data by comparing
results with four statistical methods: Fisher's exact test for pair-wise comparison of leukemia
incidence between a dose group and a control group, the Cochran-Armitage test for incidence
trend, logistic regression for incidence, and life tables for survival-adjusted incidence.
Tetrachloroethylene was one of five chemicals shown by the authors to produce "definitive"
leukemia effects in both sexes of rats. MCL effects were more often than not confined to one
sex, while tetrachloroethylene induced statistically significant increases in both sexes of the F344
rat.
In their analysis, Thomas et al. (2007) employed the rigid statistical criteria suggested in
Food and Drug Administration (FDA) guidance for testing dose-related cancer incidences of
common tumors (p < 0.01 for pairwise comparison;/? < 0.005 for trend test). They noted that
leukemia is generally considered a fatal neoplasm, thus supporting the life table test as more
likely reflecting the true statistical significance of the carcinogenic effect. Life-table analysis
(log-rank test) accounts for time-to-event information, is capable of testing nonlinear dose-
response relationships of arbitrary shapes, and is, therefore, more flexible than the
Cochran-Armitage trend test. The NTP (1986) results in male rats exposed to
tetrachloroethylene revealed a significant dose-response trend when analyzed with a life table
analysis (p = 0.004) assuming that MCL is lethal (a nonsignificant trend with logistic regression
(p = 0.097) resulted if MCL was assumed nonlethal). Pairwise comparisons revealed dose-
related incidences (p = 0.046; Fisher exact test) for both dose groups, and the Cochran-Armitage
trend test yielded ap-va\ue of 0.034; neither met the FDA criteria for statistical significance.
The borderline significance of the trend test and nonsignificance of logistic regression for the
latter two comparisons could be explained, in part, by the fact that the incidences did not follow
an incrementally increasing relationship with dose. In female rats in the NTP (1986) study, use
of a life table (p = 0.053), logistic regression (p = 0.012), a trend test (p = 0.018), and Fisher
4-259
-------
exact test (p = 0.014 and 0.022, respectively, for two doses) all revealed dose-related increases in
incidence that were of borderline significance according to the suggested FDA criteria.
Thomas et al. (2007) note that NTP does not use a rigid statistical rule in interpreting
experimental results, instead relying on consideration of other factors in a weight-of-evidence
approach. These factors include historical control incidences, and whether chemically induced
tumors were sex-specific, dose-responsive, of shorter latency, or of more advanced stage. While
encouraging stringent statistical analysis to reduce false positives, Thomas et al. (2007)
characterized the NTP weight-of-evidence approach as "appropriate" and "rigorous." They
proposed a similar evaluation of the pertinent data, to also include consideration of such factors
as reproducibility of effect across bioassays, and other information to inform biological
plausibility (i.e., evidence of toxic or carcinogenic effects on LGLs or their precursors). An
assessment of the considerations identified by Thomas et al. (2007) and NRC (2010) for
tetrachloroethylene is provided below:
Nature of the dose-response curve in terms of incidence and severity. The NTP (1986)
study found that tetrachloroethylene increased the incidence and severity of MCL in male and
female rats. The JISA (1993) study reported an increasing trend incidence of MCL in both male
and female rats, and overall the number of early deaths attributed to MCL increased with
increasing exposure.
Appropriate historical control data. Historical control data are available from the
laboratory that performed the NTP (1986) study, the NTP program, and from the Japanese
laboratory. A comparison with historical data revealed a higher MCL rate in concurrent controls
in the NTP and Japanese tetrachloroethylene bioassays. Concurrent controls in the NTP studies
were higher than historical chamber control groups at the performing laboratory (males: 28/50
[56%] vs. 117/250 [47%]; females: 18/50 [36%] vs. 73/249 [29%]). The concurrent control
group rates were also higher than the NTP program historical rate for untreated control groups
(males: 583/1,977 [29%]; females: 375/2,021 [18%]). As in the NTP (1986) study, there was a
higher control incidence of MCL (22% in males and 20% in females) than the reported historical
rate of MCL for the Japanese laboratory of 147/1,149 [13%] in males and 147/1,048 [14.0%] in
females (refer to Table 5-16, Section 5).
Reduction in latency time. The NTP (1986) study found that tetrachloroethylene reduced
tumor latency in female rats. In the JISA (1993) study, there was also decreased latency in
MCLs in female rats, with the first appearance in Week 100 in controls and Weeks 66-70 in
treated rats.
Reproducibility in another species and routes of exposure. Tetrachloroethylene has
reproducibly been found to be carcinogenic in rats and mice. Tetrachloroethylene was
carcinogenic when tested in mice in an oral gavage study (NCI, 1977) and in two inhalation
4-260
-------
studies [NTP (1986) and JISA (1993)], inducing hepatic neoplasms. Tetrachloroethylene also
caused other types of tumors in the F344 rat. However, tetrachloroethylene has only been found
to be leukemogenic in F344 rat studies. In the JISA (1993) study, deaths in female mice due to
malignant lymphomas/total dead (or moribund) mice were 6/18, 4/20, 13/27, and 10/33 in the 0,
10, 50, and 250 ppm groups, respectively. Tetrachloroethylene exposure did not affect the
incidence at study termination of malignant lymphomas in the lymph nodes or spleen. The NTP
(1986) study also did not find an effect of tetrachloroethylene on malignant lymphoma incidence
in female mice.
A similar lack of site concordance across rodent bioassays was also observed among
many of the NTP chemicals causing MCL in F344 rats reviewed by Thomas et al. (2007).
Tetrachloroethylene was among six chemicals (the others were allyl isovalerate, bisphenol A,
pyridine, 2,4,6-trichlorophenol, and the benzene metabolite hydroquinone) for which leukemia
was the only neoplastic change for either male or female rats, but for which other sex-species
groups showed evidence of carcinogenicity (Thomas et al., 2007). [Note that, as discussed in
Section 4.10, elevated incidences of other tumors—specifically, brain gliomas and kidney tubule
adenomas and adenocarcinomas—were observed in male F344/N rats in the tetrachloroethylene
NTP (1986) study but were not included in the Thomas et al. (2007) analysis.] For eight other
chemicals evaluated by Thomas et al. (2007), F344 rat MCL was the only carcinogenic effect in
rats or mice. For twenty chemicals, MCL was one of multiple neoplastic changes in F344 rats of
one or both sexes.
Involvement of both sexes. Tetrachloroethylene induced MCL in both sexes of F344 rats
in the NTP (1986) and JISA (1993) inhalation bioassays. In fact, tetrachloroethylene was one of
only 5 chemicals identified in a review of 500 chemicals by Thomas et al. (2007) that were
shown to produce "definitive" leukemia effects in both sexes of rats. Tetrachloroethylene was
also hepatocarcinogenic in both sexes of mice in the available oral (NCI, 1977) and inhalation
bioassays [NTP (1986) and JISA (1993)1. Hence, the carcinogenic effects of tetrachloroethylene
are evident in both male and female rodents across multiple data sets and with tumor sites.
Comparative species metabolism. Species differences in metabolism of
tetrachloroethylene have been noted, as reviewed in Section 3. Although thought to be
qualitatively similar, there are clear differences among species in the quantitative aspects of
tetrachloroethylene metabolism (Lash and Parker, 2001; Volkel et al., 1998; Schumann et al.,
1980; Ikeda and Ohtsuji, 1972). These differences are in the relative yields and kinetic behavior
of metabolites (Volkel et al., 1998: Green et al., 1990: Ohtsuki et al., 1983). Because
metabolites are thought to contribute to the carcinogenicity of tetrachloroethylene, these
differences in metabolism are likely to contribute to species differences in carcinogenic response,
including the types of tumors observed across rodent bioassays.
4-261
-------
The metabolite(s) contributing to the development of MCL from tetrachloroethylene have
not been defined. A role for GSH-derived metabolites was posited based on early reports of fatal
hemorrhagic disease in cattle fed trichloroethylene-extracted soybean oil meal, and the
subsequent finding that the trichloroethylene metabolite ^-(l^-dichloroviny^-Z-cysteine
(generated through the GSH pathway) induces renal toxicity, aplastic anemia, and marked DNA
alteration in bone marrow, lymph nodes, and thymus in calves (Bhattacharya and Schultze, 1972,
1971). However, similar effects were not found in a study that administered TCVC, a
GSH-derived metabolite of tetrachloroethylene, to two calves as a single dose (Lock et al.,
1996). The first calf received 10 mg/kg i.v. (40 umol/kg) and was observed for 25 days and then
given a second dose of 8 mg/kg (36 umol/kg) and observed for a further week. A second calf
was given 18 mg/kg (72 umol/kg) and observed for 20 days. An initial neutropenia was
observed in the first calf during the first few days after dosing. However, no decline in platelet
or neutrophil count, nor elevation in blood urea nitrogen, was observed. Based on clinical and
histopathological evaluation, TCVC was concluded to lack bone marrow or kidney toxicity. The
authors characterized the lack of toxicity in the kidney as "puzzling" given their prior work
demonstrating the nephrotoxicity of comparable TCVC exposures in the rat (Ishmael and Lock,
1986), and their concurrent in vitro studies showing that TCVC, like DCVC, was toxic to renal
transport mechanisms in cortical slices (Lock et al., 1996). Toxicokinetic differences among
species were postulated as an explanation for the observed species differences in TCVC
sensitivity, and the unique sensitivity of the calf to DCVC compared with TCVC and other
haloalkene conjugates. Aside from the Lock et al. (1996) evaluation of bone marrow toxicity of
TCVC in the juvenile cow, a species of unknown sensitivity to tetrachloroethylene-induced
leukemia, other studies aimed at elucidating the active metabolites contributing to leukemic
effects have not been reported. In particular, no such studies are available in the F344 rat, the
species and strain in which leukemic effects have been consistently observed in both sexes.
Analyses of how differences in metabolism may lead to differences in the
leukemogenicity of tetrachloroethylene across species are limited by this lack of knowledge
regarding the putative leukemogenic metabolites. As reviewed in Section 3, tetrachloroethylene
is metabolized by two main pathways, oxidation and GSH conjugation. Species differences in
the extent of metabolism, and in the profile of resultant metabolites, have been observed in both
pathways. Metabolism is higher in mice than in rats, predominantly owing to more extensive
metabolism via the oxidative pathway thought to contribute to hepatic toxicity and
carcinogenicity. Rats, in turn, have higher metabolic rates than do larger animals, including
humans. The half-life of tetrachloroethylene is much longer in humans (>100 hours) than in
rodents (<10 hours). Interindividual differences in metabolism, for instance arising from
variability in activity of GSTs and other metabolic enzymes, may also contribute to interspecies
4-262
-------
differences in metabolism. Overall, the database is insufficient to characterize how these
metabolic differences may impact species sensitivity to the leukemogenic activity of
tetrachl oroethy 1 ene.
Genotoxicity, cytotoxicity, and any other relevant information. Thomas et al. (2007) note
"little evidence to support a mode of action" for F344 rat MCL induced either spontaneously or
by the 34 leukemogens they reviewed, including tetrachl oroethy 1 ene. However, they propose a
review of evidence that may aid in assessing the biological plausibility for tumor induction. The
genotoxicity of tetrachloroethylene is reviewed in Section 4.8. None of the reviewed studies
have specifically investigated the genotoxicity of tetrachl oroethyl ene in the potential target tissue
(bone marrow or spleen) of the F344 rat of either sex. A study in Sprague-Dawley rats found
only marginal effects on chromosomal aberrations and aneuploidy with tetrachl oroethyl ene
exposure by inhalation (100 and 500 ppm) (Beliles etal., 1980). However, the overall
conclusion for tetrachl oroethyl ene genotoxicity supports the view that the contribution of
mutagenicity to one or more carcinogenic outcomes cannot be ruled out.
No studies are available that evaluate the toxicity of tetrachl oroethyl ene in the putative
target tissues (bone marrow and/or spleen) or target cells of MCL in the F344 rat. However, as
reviewed in Section 4.6.2.1.2, several studies by Marth (1987) or Marth et al. (1989; 1985a:
1985b), Seidel et al. (1992), and Ebrahim (2001) have demonstrated hematopoietic toxicity of
tetrachl oroethyl ene in mice. Ebrahim et al. (2001) found that tetrachl oroethyl ene in sesame oil
(3,000 mg/kg-day for 15 days) significantly decreased hemoglobin, RBC counts, decreased HCT
(packed cell volume) and platelet counts, and significantly increased WBC count. These
findings are similar to those observed in studies of tetrachloroethylene-exposed humans (Emara
etal.. 2010). In the Marth (1987) and Marth et al. (1989: 1987: 1985a: 1985b) studies, female
NMRI mice exhibited a reversible hemolytic anemia and had microscopic evidence of splenic
involvement following exposure to low drinking water levels (0.05 mg/kg-day) of
tetrachl oroethyl ene beginning at 2 weeks of age. Seidel et al. (1992) also found evidence of a
reduction in red cells, supported by decreases in erythroid colony-forming units and erythroid
burst-forming units and evidence of reticulocytosis in female hybrid mice (C57/BL/6 x DBA/2)
to tetrachl oroethyl ene at 270 ppm (11.5 weeks) and 135 ppm (7.5 weeks), 6 hours/day, 5
days/week. Reversible reductions in the numbers of lymphocytes/monocytes and neutrophils
were also observed. The slight CFU-C depression, which persisted in the exposure-free period,
could indicate the beginning of a disturbance at all progenitor cell levels. These data suggest a
reversible bone marrow depression.
A number of leukemogens (e.g., benzene) have been reported to inhibit production of
both red cells and various forms of white cells. A decrease in CFU-Ss, an effect not observed
with tetrachloroethylene exposure (Seidel etal., 1992), has commonly been reported.
4-263
-------
Leukemogens also cause a decrease in bone marrow myeloid progenitors CFU-GEMM,
CFU-GM, and CFU-E/BFU-E, the latter of which was also decreased by tetrachloroethylene
(Seidel et al., 1992). Thus, Seidel et al. (1992) provides indirect evidence that
tetrachloroethylene induces effects associated with leukemogens (NRC, 2010).
Other studies that may be relevant to leukemia induction in the F344 rat include those of
the immunotoxicity of tetrachloroethylene. However, the available database of such studies, as
summarized in Section 4.6.2.1.1, is limited for establishing whether tetrachloroethylene affects
immune parameters in a manner indicative of potential for inducing leukemia development.
Immunosuppression was observed in female B6C3Fi mice administered tetrachloroethylene
(maximum concentration: 6.8 ppm) with a mixture of 24 frequent contaminants of ground water
near Superfund sites (Germolec et al., 1989). No changes were evident in lymphocyte number,
T-cell subpopulations, NK cell activity, or with challenge by Listeria monocytgens or PYB6
tumor cells. In a separate inhalation study in mice, exposure to 170 mg/m3 (50 ppm)
tetrachloroethylene for 3 hours increased susceptibility to respiratory streptococcus infection and
significantly decreased pulmonary bactericidal activity (Aranyi etal., 1986).
As reviewed by Thomas et al. (2007), corn oil gavage has been shown to significantly
(p < 0.001) decrease the incidence of MCL in F344 rats, particularly males, by an unknown
mechanism. This complicates interpretation of the few short-term studies in rats administering
tetrachloroethylene in corn oil gavage. These include a finding of atrophy of the spleen and
thymus in rats receiving 2,000 (but not 1,000) mg/kg-day tetrachloroethylene via corn oil gavage
for 5 days (Hanioka et al., 1995a). In a separate 14-day corn oil gavage study,
tetrachloroethylene did not affect thymus and spleen weights of adult rats at a hepatotoxic dose
(1,000 mg/kg-day) (Berman et al.. 1995).
Summary. This assessment of considerations proposed in Thomas et al. (2007) and by
NRC (2010) highlights several findings that add support to the conclusion that
tetrachloroethylene is a leukemogen in the F344 rat. Particularly pertinent are findings of the
evaluation by NTP (1986) of the inhalation bioassay of tetrachloroethylene, demonstrating dose-
related increases in the incidence of MCL in both sexes and in the severity of MCL in both
sexes, as well as a shortened time to onset of MCL in female rats, and an increased incidence of
advanced MCL in female rats that died before the scheduled termination of the study. These
factors are considered the most important in evaluating the significance of the MCL findings for
tetrachl oroethy 1 ene.
Additional factors supporting the carcinogen!city of tetrachloroethylene include the
observation that tetrachloroethylene has also been found to induce other rare tumors besides
MCL in the F344 rat, as well as tumors at other sites in both sexes of the mouse, in both
inhalation and oral gavage bioassays. As noted by Thomas et al. (2007), chemically induced
4-264
-------
MCL has typically been found in only one sex of the F344 rat, and tetrachloroethylene was one
of only 5 chemicals identified in their review of 500 chemicals in the NTP database to
definitively cause the tumor in both males and females. These findings add support to the
conclusion that tetrachloroethylene is a rodent carcinogen. Although limited, studies
demonstrating hemolysis and bone marrow toxicity in mice add some support to the biologic
plausibility of the observed leukemic effects (NRC, 2010). The pharmacokinetics (metabolites)
and pharmacodynamics (biological mechanisms) that contribute to the development of MCL in
the F344 rat, both spontaneously and with chemical exposure, have not been elucidated.
Uncertainties remain regarding the causes of F344 rat MCL, the biology of the disease
including the cell type of origin, as well as the mechanisms by which tetrachloroethylene may
advance development of this rodent leukemia. Further research to clarify the factors that affect
inherent and chemically induced susceptibility to F344 rat MCL is warranted. As proposed by
Stromberg (1985), the F344 rat MCL could serve as a rodent model for human T-cell leukemias,
in which research could be conducted to identify causative factors and disease mechanisms, and
to test and develop novel chemotherapies. Thomas et al. (2007) similarly endorsed additional
research and analyses of F344 leukemogens, such as tetrachloroethylene, to advance
understanding of the mechanisms contributing to the rodent—and by inference, the related
human—diseases.
In summary, although uncertainties remain regarding the pathobiology of MCL and the
mechanisms by which tetrachloroethylene may contribute to disease development and/or
progression, this assessment of additional factors bolsters the support for the finding of
tetrachloroethylene-induced MCL in the F344 rat.
4.6.3. Summary and Conclusions
4.6.3.1. Immunotoxicity, Hematologic Toxicity, and Cancers of the Immune System in
Humans
The strongest epidemiological study examining immunologic and hematopoietic effects
of tetrachloroethylene exposure in terms of sample size and use of an appropriately matched
control group is of 40 male dry-cleaning workers (mean exposure levels <140 ppm; mean
duration: 7 years; mean blood tetrachloroethylene levels: 1,685 |ig/L) by Emara et al. (2010).
Statistically significant decreases in red blood cell count and hemoglobin levels and increases in
total white cell counts and lymphocyte counts were observed in the exposed workers compared
to age- and smoking-matched controls. Similar effects were observed in mice (Ebrahim et al.,
2001). In addition, increases in several other immunological parameters, including
T-lymphocyte and natural killer cell subpopulations, IgE, and interleukin-4 levels were observed
4-265
-------
in tetrachloroethylene-exposed dry-cleaning workers (Emara et al., 2010). These immunologic
effects suggest an augmentation of Th2 responsiveness. However, the limited available data
from studies in children (Delfmo et al., 2003a: Delfmo et al., 2003b: Lehmann et al., 2002;
Lehmann et al., 2001) do not provide substantial evidence of an effect of tetrachloroethylene
exposure during childhood on allergic sensitization or exacerbation of asthma symptomology.
The observation of the association between increased tetrachloroethylene exposure and reduced
interferon-y in cord blood samples may reflect a sensitive period of development, and points to
the current lack of understanding of the potential immunotoxic effects of prenatal exposures.
The available data pertaining to risk of autoimmune disease in relation to tetrachloroethylene
exposure are limited by issues regarding ascertainment of disease incidence and exposure-
assessment difficulties in population-based studies. In summary, there is considerable variation
in the extent and quality of the epidemiologic literature (e.g., number of studies, study design,
and quality of the exposure assessment) for lymphopoeitic cancers. In general, studies with
relatively strong exposure assessments are based on a small number of observed deaths or
incident cases, with a relatively low statistical power. For non-Hodgkin lymphoma and multiple
myeloma, the available studies are considered supportive of a role of tetrachloroethylene as a
likely carcinogen. This is based on the presence of higher effect estimates in studies with better
exposure-assessment methodologies and evidence of an exposure-response trend in one or more
studies.
Among the specific types of lymphopoeitic cancers, there is considerable variation in the
extent and quality of the epidemiologic literature (e.g., number of studies, study design, and
quality of the exposure assessment). In general, studies with relatively strong exposure
assessments are based on a small number of observed deaths or incident cases, with a relatively
low statistical power. For non-Hodgkin lymphoma and multiple myeloma, the presence of
higher relative risk estimates in studies with better exposure-assessment methodologies and
evidence of an exposure-response trend in one or more studies provide the basis for considering
the collection of studies as supportive of a role of tetrachloroethylene as a likely carcinogen.
For non-Hodgkin lymphoma, there is little evidence of an association in the large cohort
studies examining risk in relation to the broad occupational category of work in laundry or dry
cleaning [i.e., relative risk estimates ranging from 0.95 to 1.05 in females in Andersen et al.
(1999), females and males in Ji and Hemminki (2006b), and Pukkala et al. (2009)1. The results
from the four cohort studies that used a relatively higher quality exposure-assessment
methodology, however, reported relative risks between 1.7 and 3.8 (Radican et al., 2008; Boice
et al., 1999; Anttila et al., 1995). There is also some evidence of exposure-response gradients in
studies with tetrachloroethylene-specific exposure measures based on intensity, duration, or
cumulative exposure (Seidler et al., 2007; Miligi et al., 2006; Boice et al., 1999). Higher
4-266
-------
non-Hodgkin lymphoma risks were observed in these studies in the highest exposure categories,
with the strongest evidence from the large case-control study in Germany in which a relative risk
of 3.4 (95% CI: 0.7, 17.3) was observed in the highest cumulative exposure category (trend
/7-value = 0.12) (Seidler et al., 2007). Effect estimates in studies with broader exposure
assessments showed a more variable pattern (Selden and Ahlborg, 2011; Pukkala et al., 2009; Ji
and Hemminki, 2006b: Blair etal., 2003; Travier et al., 2002; Cano and Pollan, 2001; Lynge and
Thygesen, 1990). Confounding by lifestyle factors are unlikely explanations for the observed
results because common behaviors, such as smoking and alcohol use, are not strong risk factors
for non-Hodgkin lymphoma (Besson et al., 2006; Morton and Marjanovic, 1984).
Results from the multiple myeloma studies are based on a smaller set of studies than
those of non-Hodgkin lymphoma, but results are similar. The larger cohort studies that use a
relatively nonspecific exposure measure (broad occupational title of launderers and dry cleaners,
based on census data) do not report an increased risk of multiple myeloma, with effect estimates
ranging from 0.99 to 1.07 (Pukkala et al.. 2009: Ji and Hemminki. 2006b: Andersen et al.. 1999).
Some uncertainty in these estimates arises from these studies' broader exposure-assessment
methodology. Results from the cohort and case-control studies with a higher quality exposure-
assessment methodology, with an exposure measure developed specifically for
tetrachloroethylene, do provide evidence of an association, however, with relative risks of 7.84
(95% CI: 1.43, 43.1) in women and 1.71 (95% CI: 0.42, 6.91) in men in the cohort of aircraft
maintenance workers (Radican et al., 2008) and 1.5 (95% CI: 0.8, 2.9) in a large case-control
study in Washington [Gold et al. (201 Ob): tetrachloroethylene exposure]. Gold et al. (201 Ob)
also reported increasing risks with increasing exposure duration (based on job titles) (Gold et al.,
201 Ob) and based on a cumulative tetrachloroethylene exposure metric (Gold etal., 2010b). A
smaller case-control study (n = 76 cases) with tetrachloroethylene-specific exposure measures
based on intensity, duration, or cumulative exposure, Seidler et al. (2007), observed no cases
among the highest exposure groups. A small cohort study by Boice et al. (1999) of aerospace
workers observed one death among routinely exposed subjects and six deaths among subjects
with a broader definition of routine or intermittent exposure.
4.6.3.2. Immunological and Hematological Toxicity and Mononuclear Cell Leukemias in
Rodents
Additional data from inhalation, oral, and dermal exposures of different durations are
needed to assess the potential immunotoxicity of tetrachloroethylene along multiple dimensions,
including immunosuppression, autoimmunity, and allergic sensitization. The data from Aranyi
et al. (1986) suggest that short-term exposures may result in decreased immunological
competence (immunosuppression) in CD-I mice. The relative lack of data taken together with
4-267
-------
the concern that other structurally related solvents (Cooper et al., 2009) have been associated
with immunotoxicity contributes to uncertainty in the database for tetrachloroethylene.
The limited laboratory animal studies of hematological toxicity demonstrated an effect of
tetrachloroethylene exposure on RBC [decreased RBC (Ebrahim et al., 2001), or decreased
erythrocyte colony forming units (Seidel et al., 1992)] with reversible hemolytic anemia
observed in female mice exposed to low drinking water levels (0.05 mg/kg-day) of
tetrachloroethylene beginning at 2 weeks of age in one series of studies (Marth et al., 1989;
Marth, 1987: Marth et al.. 1985a: Marth et al.. 1985b). Ebrahim et al. (2001) also observed
decreased hemoglobin, platelet counts and packed cell volume, and increased WBC counts.
Cancer findings of primary concern are the statistically significant increases in MCL in
both sexes in the NTP (1986) and JISA (1993) inhalation bioassays. Section 4.6.2.2.2 addresses
issues pertinent to the interpretation of evidence that tetrachloroethylene induces MCL in male
and female rats for the purposes of human health risk assessment. That discussion summarizes
the findings of a recent analysis by Thomas et al. (2007) and considers the available evidence for
tetrachloroethylene in the context of the approach put forth by those authors and by NRC (2010).
This included a summary of the available noncancer studies that may inform the biologic
plausibility of the leukemia findings. In the paragraphs that follow, the findings in and statistical
analyses of the rodent bioassays are presented, and the other factors and data considered in the
analysis presented in Section 4.6.2.2.2 are then summarized. Together, these analyses informed
the conclusions provided concerning the application of the F344 rat leukemia data to human
health risk assessment.
Statistical analysis of the NTP bioassay revealed a statistically significant trend for males
(p = 0.004), and a marginally significant trend for females (p = 0.053). Life table analysis
disclosed statistically significant increases in both the low- and high-dose groups in males. A
significant increase in the low-dose group (p = 0.023) and a marginally significant increase in the
high-dose group (p = 0.053) was observed in females. Additional statistical analyses reported by
Thomas et al. (2007) of the female rat data from the NTP (1986) study found the results
significant by logistic regression (p = 0.012), the Cochran-Armitage trend test (p = 0.018), and
Fisher exact test (p = 0.014 and 0.022, respectively, for the lower and higher doses). Similarly,
additional analyses reported by Thomas et al. (2007) supported the statistical significance of the
male rat NTP data [logistic regression (p = 0.097), the Cochran-Armitage trend test (p = 0.034),
and Fisher exact test (p = 0.046 for the lower and higher doses)]. Notably, these statistical
analyses supported the authors' classification of tetrachloroethylene as one of only five
chemicals of the 500 examined to produce "definitive" leukemia effects in both sexes of rats.
While MCL effects were more often than not confined to one sex, tetrachloroethylene induced
statistically significant increases in both sexes of the F344 rat.
4-268
-------
In the JISA (1993) bioassay, MCL showed a statistically significant increasing trends
with dose in both males (p = 0.002) and females (p = 0.049) by poly-3 test. Because MCL is a
rapidly progressing and fatal neoplasm, Thomas et al. (2007) and NRC (2010) supported the life
table test as more accurately reflecting the statistical significance of the carcinogenic effect.
However, the poly-3 test for trend also addresses the time and cause of death and is comparable
to the life-table test.
Other factors besides statistical analyses can inform interpretation of bioassay data and
the observed effects of chemical exposures. According to NTP practices, as reviewed in Thomas
et al. (2007), bioassay evaluation includes consideration of factors such as historical control
tumor incidences, and whether chemically induced tumors were sex-specific, dose-responsive, of
shorter latency, or of more advanced stage. NTP analyses of the tetrachloroethylene bioassay
results revealed a dose-related increase in the incidence of MCL in both sexes, in the severity of
MCL in both sexes, a shortened time to onset of MCL in female rats, and an increased incidence
of advanced MCL in female rats that died before the scheduled termination of the study. All of
these findings elevate concern that the MCL findings are related to chemical exposure, and
among factors considered, add significant support to the conclusion that tetrachloroethylene is a
leukemogen in F344 rats. An additional consideration in evaluation of the NTP (1986) and JISA
(1993) studies is that a higher MCL incidence was observed in concurrent controls compared
with historical controls. The reason for the reportedly higher MCL incidence in concurrent
controls in these bioassays is not known. However, the finding of a chemically induced effect in
a bioassay with a high background rate, which is more likely to obscure chemically induced
findings, supports the conclusion that the observed tumors are due to tetrachloroethylene
exposure. The independent findings of MCL induction in two bioassays conducted by separate
laboratories also strengthen the conclusions.
Available pharmacokinetic data are insufficient to identify the active metabolite(s) of
tetrachloroethylene that contribute(s) to MCL development. Such data are also insufficient to
inform analyses of how interspecies differences in metabolism may affect leukemic outcomes in
other species. In addition, available mechanistic data are insufficient to characterize the
mechanisms or modes of action contributing to either spontaneously occurring or chemically
induced MCL in the F344 rat (Thomas et al., 2007), including such tumors induced in
tetrachloroethylene-exposed animals. However, the albeit limited studies demonstrating that
tetrachloroethylene induces hemolysis and affects bone marrow function in mice provide indirect
evidence that tetrachloroethylene induces effects associated with MCL and with known
leukemogens (NRC, 2010). These studies support the biological plausibility of
tetrachloroethylene as a leukemogen in rodent species, in general, and provide a basis for
generating hypotheses on how these tumors may be induced. Nonetheless, the paucity of data on
4-269
-------
contributing metabolites and mechanisms, and the lack of similar findings in other species,
contribute to uncertainty in interpreting the MCL data in the F344 rat (NRC, 2010).
Knowledge gaps persist regarding the causes of F344 rat MCL, the biology of the disease
including the cell type of origin, as well as the mechanisms by which tetrachloroethylene may
advance development of this rodent leukemia. Large granular lymphocyte (LGL) cells exist in
humans that are morphologically, biochemically, and functionally similar to the cells involved in
MCL in the F344 rat (Stromberg, 1985). In humans, clonal disorders of LGLs represent a
biologically heterogeneous spectrum of lymphoid malignancies thought as originating either
from mature T-cell or natural killer (NK) cells (Sokol and Loughran, 2006). LGL disorders can
clinically present as indolent (chronic) or aggressive diseases (Sokol and Loughran, 2006). The
indolent form of LGL leukemia is a disease of older adults, with a median age at diagnosis of 60
years. A number of clinical conditions have been observed in patients with LGL leukemia.
These include the following: red cell aplasia and aplastic anemia; other lymphoproliferative
disorders such as NHL, Hodgkin lymphoma, multiple myeloma, hairy cell leukemia, and B-cell
lymphoproliferative disorders; and autoimmune diseases such as rheumatoid arthritis and
systemic lupus erythematosus (Rose and Berliner, 2004). The etiology of LGL disorders is not
known (Sokol and Loughran, 2006; Rose and Berliner, 2004). Several possible etiologies have
been proposed including chronic activation of T-cell by a viral antigen or autoantigen in which
case LGL leukemia could be considered as an autoimmune disorder (Sokol and Loughran, 2006).
Lymphoid tumor pathobiology in rats and humans, its historical and current
classification, and epidemiology, including observations in tetrachloroethylene-exposed
populations, have bearing on examination of the human relevance of rat mononuclear cell
leukemia. Important to any examination are the changes in diagnostic and classification criteria
of human lymphoid tumors and lack of data on molecular markers in the tetrachloroethylene
epidemiologic studies, as discussed above. Diagnostic and classification criteria may not be
uniform across studies and hinder comparison of consistency within epidemiologic studies of
lymphoid cancers and tetrachloroethylene exposure and, also, between human and rat lymphoid
tumor observations. Furthermore, adoption of consensus nomenclatures of human lymphoid
tumors, i.e., the WHO scheme, for rats will facilitate cross-species comparisons, as was recently
conducted by the hematopathology subcommittee of the Mouse Models for Human Cancers
Consortium (Morse et al.. 2002).
Further research to clarify the factors that affect inherent and chemically induced
susceptibility to F344 rat MCL is warranted, particularly given the morphological, functional,
and clinical similarities of this rodent leukemia to human T-cell leukemias. As proposed by
Stromberg (1985), the F344 rat MCL could serve as a rodent model for the human disease, in
which research could be conducted to identify causative factors and disease mechanisms, and to
4-270
-------
test and develop novel chemotherapies. Thomas et al. (2007) similarly endorsed additional
research and analyses of F344 leukemogens, such as tetrachloroethylene, to advance
understanding of the mechanisms contributing to the rodent—and by inference, the related
human—diseases.
In summary, the available bioassay evidence and statistical analyses, together with a
limited database of studies that characterize the biologic plausibility of tetrachloroethylene as a
leukemogen, provide sufficient support of the conclusion that tetrachloroethylene causes MCL in
the F344 rat. No mechanistic or other data are available that would rule out the relevance of the
F344 MCL for assessing potential carcinogenic hazard to humans. The NRC (2010) peer review
panel agreed that there was little information on the mode of action of tetrachloroethylene-
induced rat MCL incidence. The panel, however, had differing opinions about the human
relevance of rat MCL. Some of the reviewers judged that more research was needed to establish
the relevance of the rat MCL to assessing human cancer hazard or risk. Some reviewers believed
that available data were adequate to establish the human relevance of the rat MCL. In the
context of quantitative assessment, a majority of the NRC (2010) panel judged that uncertainties
associated with MCL were too great to support their selection over other tumor types.
4.7. DEVELOPMENTAL AND REPRODUCTIVE TOXICITY AND REPRODUCTIVE
CANCERS
4.7.1. Development
4.7.1.1. Human Developmental Toxicity Data
Epidemiology studies of tetrachloroethylene exposure and effects on reproduction and
development include occupational studies of employment at dry-cleaning establishments in the
Netherlands, Scandinavia, Italy, Canada, and the United States (California) and population-based
studies of exposure through drinking water in the United States (North Carolina, Massachusetts,
and New Jersey). Tetrachloroethylene has been the predominant solvent used in the dry-cleaning
industry in the United States and Europe since the 1970s (Gold et al., 2008; Raisanen et al.,
2001). Other chemical exposures in dry-cleaning establishments are not widespread; individuals
engaged in spot cleaning may use small amounts of trichloroethylene, acetic acid, ketone, and
acetone solvents, petroleum naphthas, or hydrogen fluoride and hydrofluoric acid (Ruder et al.,
2001). Short-term exposure to tetrachloroethylene is highest for dry-cleaning machine operators,
particularly for machines requiring manual transfer of solvent-saturated clothing from a washing
machine to a drying machine. The industry in the United States has gradually switched to dry-
to-dry machines, associated with lower emissions, and in 1993, EPA ruled that all new
4-271
-------
establishments must use these machines. However, existing facilities were required to switch to
dry-to-dry machines only if the older machines became inoperable. Other workplace
characteristics influence exposure levels including adequacy of exhaust systems, level of
equipment maintenance, occurrence of tetrachloroethylene spills, and presence of open
containers (Gold et al.. 2008).
Studies of occupational exposure primarily evaluated employees in dry-cleaning
establishments, but a few studied reproductive and developmental outcomes by occupational
groupings more broadly (Lindbohm et al., 1991; Windham et al., 1991; Taskinen et al., 1989).
Although some studies identified exposed workers based on the industry they worked in, several
developed more precise classifications for tetrachloroethylene exposure levels based on detailed
information on reported job titles, tasks, and work histories obtained through interviews or
questionnaires. Exposure classification using more detailed information is expected to reduce
error in the assessment of exposure and increase confidence in the reported associations with
health outcomes.
Epidemiology studies also have evaluated reproductive and developmental health effects
stemming from incidents of tetrachloroethylene contamination of drinking water in the United
States (Aschengrau et al., 2009a: Aschengrau et al., 2009b: Aschengrau et al., 2008; Janulewicz
et al.. 2008: Sonnenfeld et al.. 2001: ATSDR. 1998b: Bove et al.. 1995: Lagakos et al.. 1986). In
general, drinking water exposures were to multiple pollutants, and most studies were not able to
determine the relative contribution to adverse health effects made by individual substances. In
one incident in Massachusetts, however, investigators were able to evaluate a "natural
experiment" that resulted from scattered water pipe replacements to the water distribution system
in communities and tetrachloroethylene-contaminated water delivered to specific groups of
households (Aschengrau et al., 2009b: Aschengrau et al., 2008; Janulewicz et al., 2008). The
studies of exposure through drinking water are complicated by the occurrence of other water
pollutants, but this literature can provide information about the consistency of health outcomes
reported with those found in the occupational studies.
Studies of developmental effects evaluated low birth weight (Olsen et al., 1990; Bosco et
al., 1987; McDonald et al., 1987), intrauterine growth restriction (IUGR; also known as small for
gestation age [SGA]) (Sonnenfeld etal.. 2001: Boveetal.. 1995). birth defects (Ahlborg. 1990a:
Olsen etal.. 1990: Bosco etal.. 1987: McDonald et al.. 1987). and stillbirth (Olsen etal.. 1990:
McDonald et al., 1987). A brief summary of each study follows, grouped by health outcome,
population (occupational, population-based), and exposure route (inhalation, drinking water).
Table 4-34 summarizes these studies. Two studies evaluated effects on postnatal development
including learning and behavior, and schizophrenia (Janulewicz et al., 2008: Perrin et al., 2007).
4-272
-------
These studies are described in the section on neurotoxicological effects (refer to Section 4.1).
Studies of effects on immunological development and childhood cancer are found in Section 4.6.
Overall, no associations were noted in several studies that assessed maternal or paternal
occupational exposure to tetrachloroethylene and increased incidence of stillbirths, congenital
anomalies, or decreased birth weight (Lindbohm, 1995; Windham et al., 1991; Olsen et al., 1990;
Kyyronen et al., 1989; Taskinen et al., 1989; Bosco et al., 1987). However, the number of
exposed cases for specific types of anomalies was not sufficient to evaluate risk with statistical
precision. When data for adverse birth outcomes identified in Sweden, Norway, and Denmark
were analyzed in relation to low or high tetrachloroethylene exposure among dry cleaners during
their pregnancies, odds ratios for congenital malformation, still birth, and low birthweight
(defined as <1,500 g) were 1.72 (95% CI: 0.40-7.12, 9 cases) for low exposure and 0.87 (95%
CI: 0.20-3.69, 3 cases) for high exposure (Olsenetal., 1990). Kyyronen et al. (1989) reported
an odds ratio for all congenital malformations of 0.8 (95% CI: 0.2-3.5) among 24 cases and 93
controls. The sample size was not large enough to evaluate specific anomalies or conduct
multivariate analyses. A case-control study by Windham et al. (1991) identified one case of
IUGR with prenatal exposure to both tetrachloroethylene and trichloroethylene among their
sample of women with live births. The studies of occupational exposure also evaluated
associations with spontaneous abortion. More detailed descriptions of these studies and analyses
of spontaneous abortions are provided in Section 4.7.2. A study of parental occupational
exposure has also examined schizophrenia in offspring (Perrin et al., 2007) and observed an
increased incidence in offspring of parents who worked in dry-cleaning establishments (RR: 3.4,
95% CI: 1.3-9.2), as discussed in Section 4.1.
4-273
-------
Table 4-34. Epidemiology studies on reproduction and development
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Zielhuis et al. (1989) (letter to editor)
The Netherlands
Cross sectional study of menstrual
disorders among dry-cleaners and
laundry workers (471 of 592, 80%
response). Sampling frame was not
described. After excluding 72 because
of current pregnancy, lactation,
chronic illness, or gynecological
surgery, and 125 exposed and 199
unexposed because they used oral
contraceptives, final data set included
68 exposed and 76 unexposed
Questionnaire responses
Prevalence in referent group
Employment in dry cleaning
compared to employment in
laundries
Amenorrhea
Oligiomenorrhea
Polymenorrhea
Irregular cycle
Unusual cycle length
Intermenstrual blood loss
Menorrhagia
Dysmenorrhea
Premenstrual syndrome
0
10
17
38
30
17
22
29
10
Linear logistic regression
Dry cleaning vs. laundry
OR (95% CI)
Oligiomenorrhea 2.1 (0.9-5.3)
Polymenorrhea 0.8 (0.4-1.7)
Irregular cycle 1.2 (0.7-2.2)
Unusual cycle length
2.3 (1.2-4.4)
Intermenstrual blood loss
1.3 (0.6-2.7)
Menorrhagia 3.0 (1.6-5.6)
Dysmenorrhea 1.9 (1.1-3.5)
Premenstrual syndrome
3.6(1.5-8.6)
Details concerning
study design and
analysis were not
provided.
-^
to
Eskenazi et al. (1991a)
United States
Men in the dry-cleaning industry
compared to men working in laundries
recruited from membership lists of
two union locals in San Francisco Bay
area and Greater Los Angeles.
Included all dry cleaners (n = 85) and
all laundry workers 20-50 yr in Local
3 (n = 119) and random selection of
Local 52 (n = 206). Laundry workers
were frequency matched by age to dry
cleaners from same union local.
Eligible were 20-50 yr of age, current
workers, spoke English or Spanish, no
vasectomy and located by telephone or
mail. Participation: 20 exposed (38%
of 53 eligible) and 56 unexposed (34%
of 166 eligible
Semen quality
Semen samples obtained from
34 exposed and 48 unexposed
Brief physical exam by
physician blind to occupational
status to identify any medical
conditions that might affect
semen quality
Direct (expired air levels) and
indirect (index) measure of PCE
exposure
Exhaled air collected 16-19 h
after the end of a workweek
(except 11, which were corrected
to 16 hours using an elimination
model)
LOD: 2.67 ug/m3, assuming 4 L
breath sample
Exposed: Workers at dry cleaners
or laundries where dry cleaning
was conducted on premises.
Unexposed: Workers at laundries
with no dry cleaning
Confirmed by industrial
hygienists
Analyzed associations with 17
measures of semen quality
Difference in means and
number with abnormal sperm
(<2Q million sperm, >40%
abnormal forms, and < 60%
motile sperm)
Oligospermia (<20
million/mL) approx 25% in
both groups
Average percentage motile
sperm "barely fell within
normal limits" in both groups
Less than 60% motile
Exposed: 44%
Unexposed: 31%,^ = 0.23
Breath samples
reflect exposure in
the last week
Laundry workers
averaged less years
education and had
higher proportion
Hispanic (90 vs.
41%). Smoking and
alcohol use were
comparable.
Laundry workers
reported a higher #
days >80°F
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
-^
to
Eskenazi et al. (1991a) (continued)
In person interviews Work history,
including job tasks and exposures
in preceding week and past 3 mo
Exposure score (0-11): estimate of
exposure during 3 mo period of
spermatogenesis
Exhaled PCE (mean, ug/m3)
Exposed (n = 34)
7,892.9 (1.5-54,949.3)
Unexposed (n = 48)
76.9 (0.6-1,562.4)
Multiple linear regression
(13 sperm measures) within
34 exposed, and all 82 men,
adjusted for several potential
confounders
No association within all 82
men for the 3 exposure
measures and clinical quality
measures: sperm
concentration, total count,
percentage motility, or
percentage abnormal forms
Associations, adjusted for
confounding (p < 0.05) for
ALH, sperm linearity,
percentage round sperm; and
# narrow sperm and at least
one measure of exposure.
ALH and linearity
measure pattern of
sperm motion.
Authors stated
clinical
interpretation is not
yet "fully
established"
Result do not
represent experience
of nonunion workers
(>85% of dry-
cleaning industry)
Eskenazi et al. (1991a; 1991b)
United States
Wives of dry-cleaners and laundry
workers [extension of Eskenazi et al.
(1991a)1
17 of 20 dry cleaners with wives and
32 of 36 laundry workers with wives
participated
# with index pregnancies or trying to
conceive:
14 dry cleaners, 26 laundry workers
Reproductive outcomes:
* Rate of miscarriage: # of
miscarriages during
husband's employment in
industry/total # of
pregnancies during same
period
* Standardized fertility ratio
(SFR): ratio of O/E based on
U.S. national birth
probabilities for race, birth
cohort, parity, and age of
wives for each person-year
Dates of employment in the
industry and exposure to PCE from
interviews (index pregnancies
ended on average 2 yr before
interviews)
Exposure estimates:
* Expired PCE for husband
* Index of exposure
* Occupation: dry-cleaner vs.
laundry worker
SFR: Comparable between
dry-cleaners and laundry
workers
Risk ratio: 1.01, 95% CI:
0.71-2.01
Time to conception (Cox
Proportional Hazard adjusted
for ethnicity and smoking):
Dry cleaners vs. Laundry:
Rate ratio = 0.54 (95% CI:
0.23-1.27)
# pregnancies and
live births similar
between dry-
cleaners and laundry
workers
Power to detect
doubling of SA rate
from 12 to 24% was
0.28
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Eskenazi et al. (1991b) (continued)
Calculated SFR for periods
when the men were
employed and not
employed in the industry
* Time to conception—self
report from wife—number
of months to become
pregnant with index
pregnancy
PCE in expired air was higher
among dry cleaners whose wives
were interviewed (10,245.6 vs.
7,892 ug/m3)
-^
to
Rachootin and Olsen (1983)
Denmark
Case-control study of couples
examined or treated for infertility at
Odense University Hospital, Denmark,
1977-1980. Controls selected from
couples with healthy child conceived
within 1 yr born at same hospital,
1977-1979. Eligible couples,
residents of the island of Funen,
Denmark, identified through hospital
inpatient register (1,069 infertile,
4,305 fertile). Response 87% for both
cases (n = 927) and controls
(n = 3,728)
Infertility
Data on reproductive
history, SES and behaviors
from questionnaire,
medical records of infertile
couples reviewed by
collaborating physician
blind to questionnaire
responses
Serf-report by women through
mailed questionnaire sent Nov
1980-May 1981. Occupation held
in year prior to hospital admission
and longest held job. Classified
based on job title, type of
workplace and description of
duties. Coded using a 5-digit
Danish Occupational Code and a 5-
digit industry code
Exposure defined as contact with
one of 15 specific chemical or
physical agents (included dry-
cleaning chemicals) or performance
of one of 3 work processes a
minimum of one time per week for
at least 1 yr
1. Cases infertile for at least
1 yr compared to controls, all
residing within catchment
area
Dry-cleaning chemicals
OR (95% CI)
* Sperm abnormalities:
1.0 (0.5-2.0)
* Women with hormonal
disturbances
1.3 (0.5-3.3)
* Women with idiopathic
infertility
3.0 (1.2-7.4)
2.7 (1.0-7.1) adjusted for
women's age, education,
residence and parity.
*Men with idiopathic
infertility
0.2 (0.0-1.4)
A higher percentage
of case couples lived
outside the
hospital's catchment
area
Analyzed
associations with 15
chemical or physical
agents, 3 work
processes, noise and
heat
Number of controls
aged >20 yr: <20
Numbers of exposed
cases and controls in
dry cleaning was not
reported
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Rachootin and Olsen (1983)
(continued)
-^
to
2. Within control group
comparison; couples who
gave birth after 1 yr
compared to other controls
Delayed conception
Dry-cleaning chemicals
OR (95% CI)
Men 1.2 (0.7-1.9)
Women 1.6 (0.9-2.9)
Adjusted for women's age,
women's education,
residence, parity, women's
smoking and drinking, and
past use of oral
contraceptives
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
-^
to
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Sallmen et al. (1995) (continued)
-^
to
Exposed: Measurements made
when holding same job and work
tasks implied solvent exposure or
solvent exposure was reported.
High: Handled solvents daily, or
1-4 d/wk and measurements
indicate clear exposure (n = 46)
Low: Handled solvents 1-4 d/wk,
no measurements or low levels, or
handled solvents
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
o
Sallmen et al. (1998), extension of
Taskinen et al. (1989)
Finland
Retrospective time-to-pregnancy study
of paternal exposure to organic
solvents. Wives of workers ever
monitored for organic solvents by
Finnish Institute of Occupational
Health, 1965-1983. Linked ids to
identify wives (n = 1,667) through
Finnish Population Register Centre
and pregnancies (n = 2,687) through
national database of medically
diagnosed pregnancies, treated in
hospital, 1973-1983. Included men in
their first marriage during 1985 with
wives aged 18-40 yr at the end of the
1st trimester of pregnancy.
Serf reported by mothers:
Time-to-pregnancy (TTP)
Included pregnancies begun
during the marriage or up to 9
mo before
Only included pregnancies
identified in register and
reported by participants
Serf-reported paternal exposure to
solvents at time attempt at
pregnancy began
Paternal exposure via mailed
questionnaires (January 1986) to
both spouses re: occupational
exposure related to study
pregnancy—employment,
occupation including work tasks,
and workplace during year of
conception
Use and frequency of any of the
monitored solvents and any other
materials
Biological measurements available
for 60% of men (during TTP
n = 33, same job but not during
TTPw= 161)
141/282 (50%) of men were
highly or frequently exposed
to organic solvents during
TTP, 24.4% (n = 80) were
low or intermediate exposed
Discrete proportional hazards
regression
Paternal exposure to organic
solvents; adj FDR OR (95%
CI)
Low/intermediate (n = 80)
0.74 (0.51-1.06)
High/frequent (n = 141)
0.80(0.57-1.11)
Evaluated several
potential
confounders:
maternal age,
maternal and
paternal alcohol,
maternal and
paternal smoking,
maternal coffee,
recent contraceptive
use, irregular
menstruation,
duration of
menstrual cycle, age
at menarche,
previous induced
abortion or
extrauterine
pregnancy, previous
SA, parity, year of
pregnancy, SA case,
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
Sallmen et al. (1998), extension of
Taskinen et al. (1989) (continued)
Restricted to cases (n = 110) and
controls (« = 332) who participated in
study on pregnancy outcome.
Excluded 1 case and 3 controls.
316 (72%) of wives participated. After
exclusions (n = 21) and inability to
give TTP (n = 13), final population
was 282 couples
Exposure assessment for 80
calendar days preceding study
pregnancy (spermatogenesis) blind
to outcome status. Based on
occupation, job description,
reported solvent or other chemical
use, and biological monitoring data.
New assessment for TTP needed
for 9 men whose job tasks had
changed since last study
Not exposed: Work tasks did not
include handling solvents, worker
did not report exposure and no
biological measurement
Potentially exposed: Work tasks
might have involved solvent use,
but not reported by worker, no
biological measurements
Exposed: Biological measurement
taken while at same job, or tasks
implied solvent exposure, or
solvent exposure reported
Level of Exposure
High: handled solvents daily or
level of biological measurements
above reference value for general
population
Intermediate: Solvent use 1-4 d/wk
and biological measurements
indicate intermediate or low
exposure
Low: Handled solvents <1 d/wk
None: all other
Adjusted for paternal and
maternal smoking, maternal
age, age at menarche >15,
duration of menstrual cycle,
frequency of intercourse,
maternal exposure to organic
solvents, year of pregnancy
and variable for missing
information
Paternal exposure to PCE;
adj FDR; OR (95% CI)
Low (n = 9)
0.86 (0.40-1.84)
Intermediate/High (n = 8)
0.68(0.30-1.53)
Adjusted for short menstrual
cycle, long or irregular
menstrual cycle, older age at
menarche, frequency of
intercourse, maternal age,
maternal exposure to organic
solvents, and variable for
missing information
unplanned
pregnancy,
frequency of
intercourse,
maternal exposure to
organic solvents
Recall: Data
collection on TTP
occurred 8-1 Syr
after pregnancy
Participation: Lower
among women with
>2 previous births
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
McDonald et al.
Canada
Hospital-based survey of maternity
departments, 1982-1984. 56,012
women interviewed in 11 obstetrical
units in Montreal (90% of births);
51,885 with term delivery (90%
interviewed) and 4,127 SA (75% of
those admitted)
to
oo
to
Treatment in hospital of SA
(4,127 women) plus serf report
of previous SA (before Week
28 of pregnancy) (10,910
pregnancies)
Stillbirth without defect:
fetal deaths after the 27th
wk of gestation
Congenital defects:
Information extracted from
medical records at time of
discharge, previous births
obtained from mothers
report at interview and later
review of physician or
hospital records
LEW <2,500 g
Serf-reported occupation at time of
conception for current and previous
pregnancies
2n analysis defined employment
for >30 h/wk at beginning of
pregnancy
Expected numbers calculated
for each occupational
category from effect of
individual factors on
probability of SA using
logistic regression: maternal
age, parity, history of previous
abortion, smoking habit, and
education
Laundry and dry cleaning:
# current pregnancies: 100
#SA:8
O/E: 1.18
# previous pregnancies: 123
#SA:31
O/E: 1.02
2nd analysis combined
current and previous
pregnancies:
# pregnancies: 202
# SA: 36
O/E: 1.05
2nd analysis used maternal
age, gravidity, previous
spontaneous abortion,
smoking, alcohol, education,
and ethnicity
Stillbirth (n = 3) O/E: 1.86
Congenital defects (n = 9)
O/E:
1.41
LEW (n = 15) O/E: 1.17
/>-value >0.05
Potential bias:
* interviewers were
informed of
outcome status
* recall time to first
wk of pregnancy
different for women
with SA vs. term
birth
Dry-cleaning and
laundry workers
likely included
many not exposed to
PCE
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
Taskinen et al. (1989)
Finland
Case-referent study
Workers ever monitored for organic
solvents by Finnish Institute of
Occupational Health, 1965-1983.
Linked IDs to identify wives through
Finnish Population Register Centre
and pregnancy outcomes through
national registers. Included men in
their first marriage during 1985 with
wives aged 18-40 yr at the end of the
1st trimester of pregnancy. Included
pregnancies begun during the marriage
or up to 9 mo before
Cases defined as wives with SA (if
multiple, one randomly selected) or
congenitally malformed child.
Referents selected from wives with
healthy birth 1973-1983 (1:3 for SA,
1:5 malformations), age matched
within 30 mo
Only included pregnancies identified
in register and reported by participants
Response rate of SA: cases 136 of
172, 79.1%; referents 370 of 505,
73.3%
Final data set including eligible
pregnancies for SA case-referent sets:
120 cases and 251 referents
Medically diagnosed
pregnancies from Hospital
Discharge Register
(National Board of Health)
or data on SA treated in
hospital polyclinics,
1973-1983
Congenital malformations
recorded in Finnish
Register of Congenital
Malformations
SA rate among all
recognized pregnancies in
the cohort (including
induced abortions) 8.8%
Paternal exposure via mailed
questionnaires (January 1986) to
both spouses re: occupational
exposure related to study
pregnancy—employment,
occupation including work tasks,
and workplace during year of
conception
Use and frequency of any of the
monitored solvents and any other
materials
Exposure assessment for 80
calendar days preceding study
pregnancy (spermatogenesis) blind
to outcome status. Based on
occupation, job description,
reported solvent or other chemical
use, and biological monitoring data
Not exposed: Work tasks did not
included handling solvents, worker
did not report exposure and no
biological measurement
Potentially exposed: Work tasks
might have involved solvent use,
but not reported by worker, no
biological measurements
Exposed: Biological measurement
taken while at same job, or tasks
implied solvent exposure, or
solvent exposure reported
Categorized into none, low, or high
Conditional logistic
regression
OR for likely exposure to
PCE only presented for
unadjusted model
(controlling for potential
exposure to PCE)
OR (95% CI)
Likely exposed: 4 cases, 17
referents
0.5 (0.2-1.5)
Trichloroethylene
Likely exposed 17 cases, 35
referents
1.0 (0.6-2.0)
Potential
misclassification of
exposure but
nondifferential:
Among men with no
monitoring data,
21.5% of cases and
24.2% of referents
reported exposure to
solvents and were
categorized as
exposure likely
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Taskinen et al. (1989) (continued)
High: handled solvents daily or
level of biological measurements
above reference value for general
population
Intermediate: Solvent use 1-4 d/wk
and biological measurements
indicate intermediate or low
exposure
Low: Handled solvents <1 d/wk
None: all other
to
oo
Lindbohm et al. (1991)
Finland
All pregnancies and outcomes
recorded in nationwide Hospital
Discharge Register and data requested
from outpatient hospital clinics,
1973-1982. Pregnancies 1973-1978
linked to 1975 Census and 1979-1982
to the 1980 Census. Central Statistical
Office of Finland (1975 and 1980)
Census data used for occupation and
industry, SES
For exposure to any
mutagenic agents,
evaluated pregnancies
terminated in 1976 for
exposure in 1975 (to
approximate 80 d prior to
conception) and May 1,
1980-April 20, 1981 for
1980 Census
99,186 pregnancies among
women, 12-49 yr old, with
information on occupation,
industry and woman's SES.
For exposure to specific
agents, included a 2-yr
period close to the census
(Janl, 1976-Dec31, 1977
and May 1, 1980-April 30,
1982)
Paternal exposure classified using
job-exposure matrix developed in
cooperation with 2 industrial
hygienists. Based on occupation
and industry. Assign prevalence of
chemical exposure to job groups
based on monitoring data from
Institute of Occupational Health
Classified into 3 levels for
exposure to mutagens:
Moderate/high: 139
Potential/low: 820
None: 7,772
Prevalence of SA: 8.8%
(Similar to national rate in
Finland: 8.'
Focus of exposure
assessment was on
"mutagens"
PCE: 3 SA and 45
pregnancies defined as
moderate/high exposure
Linear logistic regression
controlling for age only
OR (95% CI)
0.7 (0.2-2.4)
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
Lindbohm et al. (1990)
Finland
Case-control study of women recruited
from Institute of Occupational Health
database of women biologically
monitored for one or more of 6
solvents linked to national registry of
medically recognized pregnancies;
80 cases (78.4% of 102 respondents)
and 286 controls (99.3% of 288) (age
matched 1:3) confirmed pregnancy of
interest
73 cases and 167 controls with
complete information for both cases
and controls
Cases were women with a
spontaneous abortion recorded
in the national register of
pregnancies in Finland and the
Finnish Register of Congenital
Malformations that was
confirmed by the women
Serf-report of employment,
occupation, workplace and
exposure to solvents during first
trimester by mailed questionnaire
Exposure assigned by 2
investigators blind to outcome
status using responses and
biological measurements when
available
Not exposed: Work tasks did not
included handling solvents, worker
did not report exposure and no
biological measurement
Potentially exposed: Work tasks
might have involved solvent use,
but not reported by worker, no
biological measurements
Exposed: Biological measurement
taken while at same job, or tasks
implied solvent exposure, or
solvent exposure reported
Level of Exposure:
High: handled solvents daily or 1-4
d/wk and level of biological or
available industrial hygiene
measurements were high
Low: Handled solvents 1-4 d/wk,
and level of exposure low, or
handled solvents <1 d/wk
None: all other
Conditional logistic
regression controlling for
previous SA, parity,
smoking, use of alcohol, and
exposure to other solvents
OR (95% CI)
All solvents
2.2(1.2-4.1)
PCE (8/15 exposed
cases/controls)
Overall 1.4 (0.5-4.2)
Low 0.5 (0.1-2.9)
High 2.5 (0.6-10.5)
Use of PCE in dry cleaning
(4 cases/5controls)
2.7(0.7-11.2)
Other dry-cleaning work
(1/6)
0.6(0.1-5.5)
Biological
measurements were
available for only
5% of sample
Blood PCE (mean)
at time nearest
pregnancy
Dry cleaners (n = 6)
2.11 iimol/liter
Other workers
(» = 7)
0.43 iimol/liter
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
Windham et al. (1991)
United States
Hospital-based case-control study
697 women ± 18 yr, June 1986-Feb
1987 (81.8% of 852)
1,359 controls (2 per case) randomly
selected from among residents of
Santa Clara County, California with a
live birth, frequency matched by last
menstrual period (± 1 wk) and hospital
(84% of 1,485)
Analysis limited to 1,361 women who
were employed during pregnancy
(70%)
Medically diagnosed SA
defined as 20 wk gestation
with pathology specimen
submitted to one of 11
hospital laboratories in
Santa Clara County,
California; verified by
review of medical charts
Computer-assisted telephone
interview—exposure during
pregnancy (cases) or first 20 wk
(controls)
Asked whether they used or
worked around any of 10 solvents
(including PCE) once per week or
more, plus asked to name any other
solvents or degreasers. For each
product, number hours per week,
weeks of exposure, skin contact,
smelled odors, or experience
symptoms
Unexposed referent did not use any
of the named solvents (n = 847)
Exposure metric: average hours
used/week of pregnancy
249 of 1,361 working women were
exposed to solvents
5 PCE exposed cases, 2
exposed controls
9 TCE exposed cases, 15
exposed controls
Crude OR (95% CI)
PCE
4.7(1.1-21.1)
TCE 3.1(0.55-2.9)
Paint Thinners
2.3 (1.0-5.1)
Paint Strippers
2.1(0.64-6.9)
PCE and/or TCE
3.4(1.0-12.0)
Adjusted OR
PCE adj for hours worked
4.2 (0.86-20.2)
PCE adj for age
6.0 (1.4-25.8)
Intensity: respondents
reported skin contact, odor,
or symptoms (headaches,
dizziness, or forgetfulness)
Yes: ORc: 6.3, ^ = 0.04
None: ORc: 2.l,p = Q.54
Adjustment for
confounders:
Mantel-Haenszel
stratification of
dichotomized
covariates one at a
time: maternal age,
race, education,
prior fetal loss,
smoking, and hours
worked
Cases and controls
worked similar
hours and schedules
4 of 7 women
reporting use of
PCE also used TCE
Adjustment did not
alter OR for other
solvents (TCE,
thinners and
strippers)
No consistent trend
by # hours used per
week
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
Bosco et al. (1987)
Italy
67 women working in 53 of 66 dry-
cleaning shops in 2 neighborhoods in
Rome, Italy (40 dry cleaners and
ironing, 13 ironing service only)
Average age 43 yr employed on
average 20 yr
Serf report by standardized
Interview
SA not defined
Serf-report by standardized
interview
LEW <2,5QO g/live birth
Birth defects/live births
Still births/live births
Serf report by standardized
Interview—work activity prior to
and during pregnancy (dry
cleaning, housewife, other)
Presence of trichloroacetic acid in
24-h urine (53 of 67) Mean (iig/L)
Dry Cleaners 5.01* (n = 40)
Ironers only 1.35 (n = 13)
Controls 1.56(w = 5)
*p = 0.06 compared dry cleaners
with ironers and controls combined
5 SA of 56 pregnancies
reported while employed as
dry cleaner (8.9%)
1 S A of 46 pregnancies
reported while house-wife
Fourfold greater risk,
standardized for age, not
statistically significant
Dry cleaners 51 live births
Housewives 44 live births
n (%)
LEW Dry Cl Housewives
2 (3.9) 9 (6.8)
Birth Defects/LB
2 (3.9) 1 (2.3)
Still births/LB
0 (0) 1 (2.3)
Ascertainment of
exposure and
outcome was not
independent
Asked about
pregnancies
occurring 1>20 yr
previous
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
oo
oo
Olsen et al. (1990)
Scandinavia (Sweden, Finland, and
Denmark)
Nested case-control studies combining
country-specific odds ratios,
1973-1983
Sweden and Denmark: All women
selected from company records of all
active dry-cleaning plants and
laundries (dry cleaners only in
Denmark) working for >1 mo during
1973-1983.
Finland: Dry-cleaning and laundry
workers identified from union
registers and payroll data requested
from all facilities in country,
1973-1983 (% response not
provided). Asked for names of
women employed for at least 3 mo
during 1973-1983
Sweden: 169 women with a registered
pregnancy who worked at laundry or
dry cleaner during all or part of the
year before delivery or 6 mo before
SA, 2 matched controls per case; 84%
respondents of 201 contacted. 61.7%
of identified plants participated
Finland: 720 pregnancies (1 randomly
selected per woman) reported in
hospital discharge register and
reported by woman, 3 age-matched
controls per case; 77.2% respondents
of 932 contacted
Medically recognized SA
recorded in centralized birth
registries and linked to
participants
Sweden: Central medical birth
register (n = 31)
Finland: Nationwide hospital
discharge registry and
polyclinic data on SA
Denmark: Central birth registei
and standard hospital register
(n = 10)
Low birth wt: <1,500 g
(Sweden (n = 5), Norway
(n = 7) and Denmark
(« = 1)
Congenital malformations
(excluding certain minor
malformations)
Sweden: n = 6
Norway: n = 1
Denmark: n = 1
Finland: n = 24
Perinatal death (Sweden and
Norway)
Sweden: n = 5
Exposure during 1st trimester; Serf
report from questionnaires or
interview
Sweden and Denmark:
classification by industrial
hygienist blind to pregnancy
outcome
Finland: classification by study
investigators based on work history
and exposure frequency
Classification:
Unexposed—No exposure to PCE
as defined
Low—worked in dry-cleaning
facility but not high exposure.
High—Conducting dry cleaning or
spot removal >1 h/d
Spontaneous abortion
OR (95% CI)
Combined (weighting by
inverse variance of OR):
Low 1.17 (0.74-1.85)
High 2.88 (0.98-8.44)
Sweden:
Low 1.15 (0.43-3.09)
High 0.82 (0.07-9.86)
Denmark:
Low 0.00
High 2.52 (0.26-24.1)
Finland:
Low 1.18 (0.71-1.97)
High4.53 (1.11-18.5)
Combined outcomes (LEW,
malformations and perinatal
death), All countries
combined, all trimesters
(combined variance
calculated using inverse
variance of the OR) OR
(95% CI)
Low 1.72 (0.4-7.12)
High0.87 (0.2-3.69)
In Sweden and
Denmark only 1
exposed case in high
exposure group, in
Finland 6 exposed
cases in high
category
Models adjusted for
parity, smoking and
drinking habits
(Danish model only
for parity and
smoking)
Analyses using
exposure
information from
employers (55% of
sample) stated to
have similar results
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Olsen et al. (1990) (continued)
Denmark: 143 registered pregnancies
of all women employed at least 1 mo
at listed registered dry cleaners,
1979-1984, 77.3% respondents of 185
in cohort. 74.3% of identified plants
participated
to
oo
VO
Kyyronen et al. (1989)
[Also reported in Olsen et al. (1990)1
Finland
679 women confirmed the pregnancy
contained in the register and provided
exposure information for the 1st
trimester; 25.9% of SA cases did not
report the pregnancy in the register
and were not included along with
matched controls
130 SA reported
289 controls (women with
healthy pregnancy and no
SA during study period),
matched by age within
±2yr
24 cases of congenital
malformation
93 controls
Unexposed—No exposure to PCE
as defined
Low—work tasks included
pressing at a dry cleaners' or spot
removing, or reported handling
PCE less than once per week
High—work tasks included dry
cleaning for at least 1 h daily on
average, or reported handling PCE
at least once per week
Spontaneous abortion:
Multivariate logistic
regression model:
High—3.4 (1-11.2)/?< 0.05
Low exposure was not
included in multivariate
model: unadjusted OR: 0.7
(95% CI not reported)
Model adjusted for frequent
use of solvents other then
PCE, frequent heavy lifting
at work, frequent use of
alcohol
Congenital malformation:
Univariate, matched logistic
regression
PCE (any level) 1st trimester
OR (95% CI)
0.8 (0.2-3.5) 2 exposed cases
6 cases and 6
controls reported
exposure to other
solvents: petroleum
benzene, toluene,
acetone, thinner, and
spot remover
mixtures
Other covariates
(including smoking,
temperature, parity,
febrile disease) were
not associated in
univariate models so
not included
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
VO
o
Ahlborg Q990a) [Complementary
study to Olsen et al. (1990)1
Sweden
Case-referent study: Two cohorts of
women working for >1 mo during
1973-1983 in dry-cleaning or laundry
work
Primary: 2,181 eligible women
selected from company records of 475
active dry-cleaning plants and
laundries, 263 used PCE and had
women as employees
Linked to Medical Birth Registry and
Inpatient Registry for Somatic Care;
identified 2,438 births and 143 SA
955 pregnancies including 66 cases of
SA, perinatal death, congenital
malformation, or low birth weight
involved employment (at least one
week) during year before delivery or 6
mo before SA. Referents matched to
cases (1:2) by mother's age (± 2 yr),
year of pregnancy, and parity (for
deliveries only)
Responses for 158 pregnancies (48
cases (75%, 110 referents (88%)
Complementary: 5,176 female laundry
and dry-cleaning workers registered as
washers/cleaners in the national
census of 1975 and 1980; linked with
medical registers for 2-yr period
following each census—1,136
pregnancies identified
Pregnancies and
hospitalized SA identified
through national registries
occurring 1974-1983
Identified 2,438 births
Cases defined as
spontaneous abortion,
perinatal death, congenital
malformation, or low birth
weight
Exposure during 1st trimester; Serf
reports through mailed
questionnaires; questions re: type
of production (laundry only,
laundry and dry cleaning, or dry
cleaning only), use of specific
agents in dry-cleaning process
(including PCE)
Information obtained from
employers on type of production,
amount of dry cleaning, and use of
specific cleaning agents during
1973-1983, and dates use of PCE
started and ended
Use of PCE:
22 of 48 cases said "don't know,"
19 categorized as exposed by
employer
41 of 110 referents said "don't
know," 30 categorized as exposed
by employer
Exposure classified by 2
investigators blind to case/referent
status
High: Operating dry-cleaning
machines or spot removing with
PCE ± 2 h/wk, or ironing/pressing
dry cleaned cloth >20 h/wk, or
cleaning and filling the machines
>3 times
Low: Other work at workplaces
where dry cleaning with PCE was
performed
Multivariate conditional
logistic regression model
Primary study:
Dry cleaning (Y/N)
Referents did not work in dry
cleaning or were not working
during 1st trimester
All outcomes combined:
OR(95%CI): 1.1 (0.6-2.0)
Serf-report
1.02 (0.47-2.2) (Ahlborg.
1990a)
Employer
1.27 (0.60-2.71)
Use of PCE (Y/N)
OR (95% CI):
Serf-report
0.92 (0.36-2.33)
Employer
0.82 (0.32-2.07)
Adding response from
employer to data self-
reported as "don't know":
1.24 (0.59-2.61)
Highly exposed pregnancies
Primary study: 10 of 55
cases, 27 of 106 referents
Complementary: 9 of 67
cases, 17 of 126 referents
For SA only:
Low 1.0 (0.4-2.2)
High 0.9 (0.4-2.1)
Few highly exposed
pregnancies, few
cases
Validity of self-
reports:
Questionnaire data
compared to
employers
response:.
Dry cleaning Y/N:
sensitivity among
cases 0.97 and
controls 0.96
Specificity among
cases 0.75 and
controls 0.69
PCE Use Y/N:
Sensitivity among
cases: 1.0 and
controls: 0.93;
Specificity among
cases: 1.0 and
controls: 0.94
Large plants
participated in the
primary study (dry
cleaning accounted
for <10% of total
production)—air
concentrations likely
to be lower than for
smaller plants
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Ahlborg Q990a) [Complementary
study to Olsen et al. (1990)1
(continued)
755 pregnancies not found in primary
study, including 55 SA and 28 other
adverse outcomes, response for 68 of
77 cases (88%) and 131 of 150
referents (87%)
Unexposed: Dry cleaning with PCE
was not performed at workplace
Models adjusted for
smoking, alcohol
consumption,
medical
complications, and
history of adverse
pregnancy outcome
-^
to
Doyle et al. (1997)
United Kingdom
7,305 women, 16-45 yr, currently or
previously employed in dry cleaning
or laundry units managed by 4
companies in the UK, 1980-1995
54.5% of 5,712 questionnaires
successfully delivered were returned
completed
Response rate for current dry-cleaning
and laundry workers: 78 and 65%
Previous workers 46 and 40%
Serf report via mailed
questionnaire, self reports
verified with general
practitioner (all women
(114) reporting SA who
worked during pregnancy
and random sample of 58
who reported not working,
comparison for 59).
Distribution of reported
exposures during
pregnancy was similar for
validated vs. not validated;
SA defined as any fetal loss
before 28 wk gestation in a
confirmed pregnancy
Serf report via mailed
questionnaire;
For each pregnancy: Work in dry
cleaning or laundry during
pregnancy or 3 mo prior to
conception
Exposure defined as machine
operator during pregnancy or 3 mo
prior to conception, unexposed as
nonoperator
Unit of analysis: pregnancy
SA rate: # reported SA/#
liveborn, SAs, and stillbirths
408 pregnancies among
operators
#SA:
Operator: 65
Nonoperator: 29
Laundry: 18
Dry cleaning vs. laundry
Pregnancy completed
1980-1995:
Adjusted OR (95% CI):
0.97 (0.55-1.69)
Operator vs. Nonoperator:
1.63 (1.01-2.66)
Compared to unexposed
pregnancies before 1st
exposed pregnancy:
Laundry: 1.49 (0.87-2.58)
Nonoperators: 1.02
(0.65-1.6)
Operators: 1.67(1.17-2.36)
Models adjusted for
maternal age,
pregnancy order,
and year of event
Separate analyses
also restricted to 1st
and last pregnancies
Were dry cleaners
more likely to report
fetal death or
ectopic pregnancy?
No. Current
workers: dry
cleaners vs. laundry
11 vs. 12.9%;
Previous workers:
13.9 vs. 14%
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
VO
to
Perrin et al. (2007)
Israel
Jerusalem Perinatal Study, a
longitudinal study
Examined risk for schizophrenia in a
prospective population-based cohort
of 88,829 offspring born in Jerusalem,
1964-1976, followed from birth to age
21-33 yr (January 1, 1998). Included
all births to mothers in a defined
geographic area and linked to Israel's
national Psychiatric Registry
88,060 with complete information
The Psychiatric Registry
contains diagnoses from
multiple sources, including
inpatient wards in
psychiatric and general
hospitals and psychiatric
day-care facilities.
Definition of
schizophrenia-related
discharge diagnostic codes
F20-F29. Date of onset-
first psychiatric admission
4 offspring of parent dry
cleaners with schizophrenia
(2 male, 2 female); 3 cases
had exposed fathers
Occupation and demographic
information from birth certificate
Dry cleaning = 1 if mother or father
occupation listed on birth
certificate, otherwise 0
144 offspring with one or both
parents a dry cleaner (63 female, 81
male)
Time to schizophrenia using
proportional hazards
methods
Evaluated potential
confounders: parents' age,
father's social class, duration
of marriage, rural residence,
religion, ethnic origin,
parental immigration status,
offspring's birth order sex,
birth weight and month of
birth. Variables included if
changed risk estimate by
>10%. Results presented as
crude because confounding
was minimal
637 diagnosed with
schizophrenia-related
diagnosis; cumulative
incidence = 1%
RR:3.4(95%CI: 1.3-9.2)
Models did not
adjust for family
history of mental
illness
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Drinking Water
to
VO
Lagakos et al. (1986)
United States
Retrospective population-based study
of adverse pregnancy outcomes and
childhood disorders in Woburn,
Massachusetts in relation to drinking
water from two municipal wells
contaminated with chlorinated
organics, 1960-1982. 7,134 of 8,109
possible interviews were completed
(80%). 6,219 distinct residences were
reached and 5,010 interviews were
completed (57% of the towns'
residences with listed telephone
numbers)
235 volunteer interviewers
(approx half were Woburn
residents) conducted a
telephone sample survey of
current and former family
members living in Woburn
household between
1960-1982 using telephone
numbers from the 1982
directory. Interviews were
anonymous and residence
address was not identifiable
For any residents prior to
1979, self-reports on all
pregnancies ending
between 1960 and 1982 for
women born since 1920
SA: loss in the first 6 mo of
pregnancy
Perinatal death: Stillbirth or
livebirth surviving fewer
than 7 d
Low birth weight (LEW): 6
Ibs (2,722 g)
Exposure estimates for water
from Wells G and H using
information on space—time
distribution. Residence history
obtained from 1982 telephone
directory and serf-reported
residence history.
2 of 8 municipal wells (Wells G
and H in eastern Woburn) were
tested in May 1979 and found to
contain volatile organics and the
wells were shut down.
TCE 267 ppb
PCE21ppb
Chloroform 12 ppb
Trichlorotrifluoroethane 23 ppb
Dichloroethylene 28 ppb
Groundwater sampling in 1979,
61 test wells identified
48 EP^4 priority pollutants and 22
metals
MA Dept Environmental Quality
and Engineering estimated
regional temporal distribution of
water from Wells G and H during
October 1964-May 1979 using a
model of the Woburn water
distribution system creating 5
zones of graduated exposure
before and after 1970.
Maximum likelihood logistic
regression model adjusting for
maternal age, smoking status
during pregnancy, year of
pregnancy, SES, sex, and
mother's pregnancy history
4,396 pregnancies, 1960-1982
16% were exposed during year
the pregnancy ended
SA: 12% (n = 520)
Perinatal death: 1.5% (46
stillbirths and 21 deaths before
7d)
LEW among live births >7 d:
6.4% (220/3,462)
Congenital anomalies: 4.6% (n
= 177)
Adjusted OR not presented
SA(p = 0.66)
LEW (p = 0.77)
Perinatal deaths before 1970 (p
= 0.55)
After 1970: OR (p-value)
10 (0.003) (Based on 3 deaths
out of 88 births in highest
exposure quartile, 1970-1982
Rates of adverse
health effects in East
and West Woburn
among unexposed
(during years when
Well G and H were
not operating) were
not statistically
significantly
different
Authors explored
differences between
East and West
Woburn for possible
selection bias, and
completed calls and
refusals. Checked
accuracy of
interviewers
(recontacting) and
respondents
(verified with
medical records)
Did not ask about
perception of
exposure to Wells G
and H in survey
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
VO
Lagakos et al. (1986) (continued)
Medically diagnosed
congenital anomalies
grouped by involved organ
orsystem(ICD):
musculoskeletal,
cardiovascular, and eye/ear
defects. Grouped other
organs/systems with few
cases into a group with
potential environmental
links (CNS, chromosomal,
and oral cleft anomalies)
and "other." Grouped prior
to exposure evaluation
Childhood disorders
grouped into 9 categories
with >20 cases
These data used to estimate the
percentage of annual water
supply from Wells G and H at
each household
Calculated an annual exposure
score corresponding to the
mother's residence in the year the
pregnancy ended
For each child: sum of annual
exposure scores during residence
history in Woburn
Anomalies:
Musculoskeletal (p = 0.78)
Cardiovascular (p = 0.91)
Eye/Ear OR (p-value)
14.9 (O.OOOl)
CNS/chromosomal/oral cleft
OR Rvalue) 4.5 (0.01)
Other (p = 0.62)
Childhood disorders:
Observed vs. expected
cumulative Wells G and H
exposure by disorder
Kidney/urinary tract (p = 0.02)
Lung/respiratory disorders
(p = 0.05)
Study could not
associate effects
with specific
contaminants
Bove et al. (1995)
United States
Cross-sectional study of birth
outcomes and fetal deaths in relation
to total trihalomethanes (TTHM) and
chlorinated organics in public water
supplies in a 4-county area in northern
New Jersey, 1985-1988. 80,938
singleton live births and 594 singleton
fetal deaths (after excluding plural
births, therapeutic abortions and
chromosomal anomalies) from 75 out
of 146 towns primarily served by
public water systems
Live births and fetal deaths
(plus birth weights and
gestational age) identified
through birth or death
certificates occurring
during 1/1/85-12/31/88
LEW <2,500 g among term
births (>37 wk)
SGA: live births below
race-, sex-, and gestational
week-specific S^percentile
weight using NJ data for
1985-1988
Estimated monthly levels of
individual contaminants in each
of 75 towns using tap water
sample data collected by the New
Jersey Dept. of Environmental
Protection and Energy and the
water companies. At least 2
samples per year. Monthly
estimates were assigned to each
gestational month for each live
birth and fetal death. Estimated
independently of birth outcome
data
Linear regression for birth
weight, Logistic regression for
categorical outcomes
Adjusted for maternal age,
maternal race, maternal
education, primipara, previous
stillbirth or miscarriage, sex of
the birth, adequacy of prenatal
care. PCE model also adjusted
for TTHM
Results reported with nested CI
(50, 90, and QQO'
During study period,
birth and death
certificates did not
record maternal
occupation,
smoking, and
alcohol consumption
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Bove et al. (1995) (continued)
-^
to
Preterm birth (<37 wk)
Very low birth wt <1,500 g
Birth weight among "term
births" (>37 wk and <42
wk)
Birth defects ascertained
using NJ Birth Defects
Registry—a population-
based, passive system—
plus fetal death certificates
(>20 wk)
Comparison group
(n = 52,334): all live births
from study population that
were not low birth weight,
SGA, or preterm, and with
no birth defects
Birth defects and fetal deaths in
relation to average exposure
during 1st trimester
PCE
Average 1st trimester: 26 ppb
Average entire pregnancy: 14 ppb
55.6% of study population with
surface water as source of
drinking water, 11.6% had a
mixture of surface and ground
water
82% of comparison group had
PCE concentration in public
water supply <1 ppb, 11.5%
>l-5 ppb, 5.1% >5-10 ppb and
1.4%>10ppb
Adjusted mean decrease in
birth weight among term
births: 27.2 g (50% CI: -13.4-
-41.0) for PCE > 10 ppb
No association with fetal
deaths, LEW, SGA, or preterm
birth
Very LEW: OR, 50% CI: 1.49,
1.13-1.97
All surveillance birth defects:
OR(50%CI): 1.14, >10 ppb
CNS defects: no association
Neural tube defects: PCE >5
ppb: 1.16(0.69-1.83),
association disappeared when
TTHM included in model
Oral cleft defects: PCE
# OR 50% CI
<1 67 ref
>l-5 11 1.170.89-1.53
>5-10 1 0.240.05-0.63
>10 4 3.54 2.12-5.57
No monotonic trend
Major cardiac defects: PCE
>5ppm:OR: 1.13
Information on these
risk factors was
obtained for a small
number of mothers
by phone interview.
For these women,
adjustment for these
risk factors did not
change the
contaminant specific
ORsby>15%
Authors noted that
nondifferential
misclassification
could result in
underestimate or
overestimates of the
true effect for
middle exposure
categories
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
VO
Sonnenfeld et al. (2001)
United States
Retrospective study of birth outcome
among singleton liveborn and stillborn
infants of >20 wk gestation, and
exposure to volatile organic
compounds in drinking water at the
U.S. Marine Corps Base at Camp
Lejeune, North Carolina, 1968-1985.
Included births to mothers living in
base family housing at delivery and
for at least 1 wk prior. Excluded 2
groups of residents exposed to TCE
through a different water system and
residents in trailer parks because
housing records were incomplete
Outcome data obtained
from birth and fetal death
certificates:
* Mean birth weight
* Small for gestational age:
gestational age calculated
from last menstrual period.
Weight less than the 10th
percentile based on sex-
specific growth curves
* Preterm birth: live births
<37-wk gestation (12
weighing >3,600 g were
recorded as full term)
Birth certificate data were
matched to Camp Lejeune
housing records to confirm
address and that pregnancy
occurred during occupancy
Well, dug in 1958, supplying
residents at Tarawa Terrace
Housing Areas I and II was
contaminated with PCE and other
volatile organic compounds from
a dry-cleaning business that
opened in 1954. Business
practices did not change between
1960 and 1985, when 3
contaminated wells were
disconnected from the TT water
distribution system (February 8).
Data on concentrations available
for 1982 and later. One well
(TT26) of 6 had detectable
contamination and proportion of
water from TT26 varied daily.
Water from all wells was mixed
prior to distribution
Concentration (ppb) in finished
water samples, 1982-85
May-June 1982
PCE 76-1,580
TCE ND-57
Exposed: TT residents
Unexposed: Remaining base
family housing units (minus
exclusions)—based on water
samples from supply wells and
finished water in 1984 and 1985
Potential confounders: infant's
sex and year of birth, mother's
race, age, educational level,
parity, adequacy of prenatal
care, marital status, and history
of fetal death, father's age,
educational level, and military
pay grade. Variable selection
by backward elimination
Exposed vs. unexposed
Difference in mean birth wt:
-26 g (90% CI: -43, -9)
SGA
OR(90%CI): 1.2(1.0, 1.3)
Preterm birth
1.0(0.9, 1.1)
No discernable pattern with
duration of exposure estimated
by length of residence at TT
prior to giving birth
Adjustment for
confounders did not
alter risk estimates
for exposure
Did not control for
maternal smoking,
alcohol and height
No data on
concentration at tap
in individual homes,
water consumption
or showering
Exposure
misclassification:
Unexposed group
was exposed to PCE
prior to 1972
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
VO
Aschengrau et al. (2009a; 2009b:
2008)
United States
Population-based retrospective cohort
study of exposure to PCE in drinking
water after installation of water
distribution pipes lined with PCE-
impregnated vinyl liners (VL),
selected all births (index birth),
1969-1983, from birth certificates
with addresses in one of 8 Cape Cod
towns with some VL/asbestos cement
(AC) water distribution pipes at the
birth. Selected 1,492 with addresses
with exposure to VL/AC pipe and
1,704 frequency matched to "exposed"
by month and year of birth. 959
(64.3% of selected, 70.5% of located)
of exposed and 1,087 of referents
(63.8% of selected, 69.3% of located)
were enrolled
Included only pregnancies with
completely geocoded residential
histories (94.2% of reported
pregnancies)
Clinically recognized
pregnancy outcomes:
* miscarriages, stillbirths
up to Dec 1990 by self
report, self-administered
questionnaire
Final analysis included
5,567 pregnancies from
1,891 women
prevalence of loss among
eligible pregnancies: 11.8%
* Birthweight and
gestational age among
single healthy infants from
birth certificates
* Low birth weight
(<2,500 g)
* Premature birth (gestation
<31 wk)
* Intrauterine growth
retardation (IGR) (Birth
weight
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
Aschengrau et al. (2009a; 2009b:
2008) (continued)
During the LMP year: 213 losses
and 1,743 live births with some
exposure; 446 losses, 3,165 live
births with no exposure
Increased odds ratios for any
exposure and neural tube
defects (3.5, 95% CI:
0.8-14.0), oral clefts (3.2, 95%
CI: 0.7-15.0), gastrointestinal
(1.8, 95% CI: 0.7-4.4), and
genitourinary malformations
(1.6, 95% CI: 0.6-3.8)
No increased odds ratios for
cardiac and musculoskeletal
malformations
-^
to
-------
Table 4-34. Epidemiology studies on reproduction and development (continued)
Reference, population, study design
Outcomes
Exposure assessment
Key results
Notes
to
VO
VO
Janulewicz et al. (2008) (continued)
Included only pregnancies with
completely geocoded residential
histories (94.2% of reported
pregnancies)
2,125 subjects in final data set
Exposure:
• Cumulative prenatal: from
month and year of last
menstrual period to the month
and year of birth.
• Cumulative postnatal: from
month and year of birth
through month and year of the
child'55th birthday
Final data set using refined
exposure assessment
Exposed: 1,349
Nonexposed: 737
Exposure variables divided into
quartiles
Final model for gestational
age: maternal education, race,
prior preterm delivery,
obstetric complications in the
current pregnancy,
occupational exposure to
solvents, use of serf-service dry
cleaning, and proximity of any
residences to dry-cleaning
establishments
No associations withprenatal or
postnatal exposure and
outcomes; some increased OR
in low exposure groups. For
example, ADD (OR [ 95%
CI]):
Low: 1.4 [0.9-2.0]
High: 1.0 [0.7-1.6]
Did not use
information on
water consumption
and bathing habits
by residence—
estimates are not a
direct measure of
PCE intake by
individuals
-------
Several studies in the United States of tetrachloroethylene in drinking water have
evaluated developmental risks (Aschengrau et al., 2009b: Aschengrau et al., 2008; Janulewicz et
al.. 2008: Sonnenfeld etal.. 2001: Boveetal.. 1995: Lagakos et al.. 1986). Lagakos et al. (1986)
reported the results of a population-based study in Woburn, Massachusetts, among residents
whose drinking water source was two wells contaminated with chlorinated organic substances
from 1960 to 1982 (refer to previous study description in discussion of spontaneous abortion).
Of the 3,809 infants that survived more than 7 days, 220 had low birth weights defined as 6
pounds (not the typical definition of 2,500 g). The 177 medically diagnosed congenital
anomalies (4.6%) were grouped by the involved organ or system using ICD codes. Sufficient
cases existed for musculoskeletal (n = 55), cardiovascular (n = 43), and eye/ear defects (n= 18)
for separate analyses. CNS, chromosomal, and oral cleft anomalies were grouped together
because they contained few cases. The authors felt there was evidence from previous studies to
suggest that these anomalies may be associated with exposure to environmental contaminants.
The rest of the anomalies were grouped into a category called "other." Childhood disorders were
compiled into nine categories. Incidence of childhood leukemia in relation to exposure also was
assessed and is described in the Section 4.6.1.2.5.
Logistic regression analyses, controlling for other risk factors, found no statistically
significant associations between the annual exposure score for the year a pregnancy ended and
musculoskeletal, cardiovascular, or "other" birth anomalies. However, an association was
observed for eye/ear anomalies (OR: 14.9, p < 0.0001) and CNS/chromosomal/oral cleft
anomalies (OR: 4.5, p = 0.01). In an effort to evaluate potential recall bias, the authors checked
66 of 96 disorders (perinatal death post-1970, eye/ear, or CNS/chromosomal/oral cleft anomaly,
other childhood disorders) that had been confirmed in a second interview with medical records.
Of the 66 events, the authors were able to verify 62 using medical records. No relation of
reporting accuracy with exposure was found, thus, there was no evidence of recall bias, although
the authors did not attempt to check birth records among controls.
A prevalence study in four counties in New Jersey evaluated organic contaminants
monitored in the public water supply in relation to birth outcomes (Bove etal., 1995). All live
births and fetal deaths reported on birth or death certificates between January 1, 1985, and
December 31, 1988, among residents of 75 out of 146 towns were ascertained. The final data set
included 80,938 singleton live births and 594 fetal deaths that were not therapeutic abortions or
chromosomal anomalies. Birth weights and gestational age were obtained from birth or death
certificates. Birth defects for live births were obtained from the New Jersey Birth Defects
Registry, a population-based, passive reporting system. Additional birth defects were
ascertained from fetal death certificates (>20 weeks gestation). Categorical outcomes were
4-300
-------
compared to all live full-term births in the study population that were normal weight and had no
birth defects (n = 52,334).
Monthly levels of the contaminants of interest in each town were estimated from
sampling data (at least one sample per 6-month period) obtained from the New Jersey
Department of Environmental Protection and Energy and the 49 water companies that served the
communities. The monthly estimates were assigned to each gestational month for each live birth
and fetal death. Fetal death and birth defects were evaluated in relation to levels averaged over
the first trimester. Other birth outcomes were analyzed in relation to levels averaged over the
entire pregnancy. Average tetrachloroethylene concentrations during the first trimester for all
live births and fetal deaths were 26 ppb.
Tetrachloroethylene concentrations during the first trimester were <1 ppb among 82% of
the comparison group. Concentrations were >l-5 ppb, >5-10 ppb, and >10 ppb for 11.5, 5.1,
and 1.4% of the comparison group, respectively. Infants in the >10 ppb group were 27.2 g
lighter (50% CI: -13.4-41.0). The regression models were adjusted for maternal age, race and
education, primipara, previous stillbirth or miscarriage, sex of the birth, and adequacy of prenatal
care, plus total trihalomethane levels. The odds ratio for very low birth weight was 1.49
(50% CI: 1.13-1.97) among term births in the >10 ppb group. An odds ratio of 1.16
(50% CI: 0.69-1.83) was observed for neural tube defects among singleton live births and fetal
deaths in the >5 ppb group. The odds ratio for oral clefts in the >10 ppb group was 3.54
(50% CI: 2.12-5.57). There were 67, 11, 1, and 4 oral cleft cases in the <1 ppb (referent),
>l-5 ppb (OR: 1.17, 50% CI: 0.89-1.53), >5-10 ppb (OR: 0.24, 50% CI: 0.05-0.63), and >10
ppb tetrachloroethylene exposure groups, respectively. The authors also reported 90 and 99%
CIs for odds ratios over 1.5. For oral clefts, the 90 and 99% CIs for the odds ratio in the >10 ppb
group were 1.28-8.78 and 0.82-12.15, respectively. When multipollutant models including all
contaminants with associations were evaluated, the authors stated that tetrachloroethylene was no
longer associated with neural tube defects, and the odds ratio for oral cleft defects was reduced to
2.0 (CIs were not presented). In the multipollutant model, the odds ratios for trichloroethylene
and total trihalomethanes increased to 3.5. Therefore, while tetrachloroethylene appeared to
increase risk for very low birth weight, neural tube defects, and oral clefts, other monitored
drinking water contaminants also were associated with increased risk, and the contribution of
individual substances cannot be determined.
A study of birth outcomes among singleton liveborn and stillborn infants, >20 weeks,
was conducted at the U.S. Marine Corps Base at Camp Lejeune in North Carolina for the period
1968-1985 (Sonnenfeldetal.. 2001: ATSDR. 1998b). Tetrachloroethylene and other volatile
organic compounds used by a nearby dry-cleaning business contaminated drinking water
supplied to two housing areas on the base (Tarawa Terrace I and II) until the contaminated wells
4-301
-------
were disconnected in 1985. Water concentrations measured in samples taken between 1982 and
1985 ranged from 76 to 1,580 ppb for tetrachloroethylene and from not detected (<10 ppb) to
57 ppb for trichloroethylene. The study population included births to mothers living in base
family housing at delivery and for at least 1 week prior. Residents of Tarawa Terrace I and II
were defined as exposed (n = 6,117 births). On the basis of water samples collected from wells
and finished water during 1984 and 1985, residents of the remaining base family housing units
were defined as unexposed (n = 5,681 births). Information on birth weight, gestational age, and
preterm birth (live births less than 37 weeks gestation) was obtained from North Carolina birth
records. To define small for gestational age, a gestational age specific birth weight distribution
for a Caucasian population in California (Williams et al., 1982) was found to best describe the
distribution of live births among the nonexposed group. Because standard birth weight
distributions for military populations were not available, the California reference was used to
identify a weight that classified 10% of births as small for gestational age in the nonexposed
group. In models including a term for gestational age, mean birth weight among exposed infants
was 26 g lower than the nonexposed infants (95% CI: -43, -9). The odds ratios for small for
gestational age and preterm birth were 1.2 (95% CI: 1.0-1.3) and 1.0 (95% CI: 0.9-1.1),
respectively. Regression models included several covariates to evaluate confounding, which
were retained after backward elimination; however, some known factors associated with birth
were not evaluated (maternal smoking, alcohol consumption, or height). Because exposure
status was associated with mother and father's education, father's military pay grade, and
mother's age, the unexamined risk factors also may have been associated with exposure and may
have acted as confounders. Final models for mean birth weight included mother's age, history of
one previous fetal loss, history of two or more fetal losses, gestational age, mother's race, living
in an officer's or warrant officer's household, year of birth, and sex of the infant. Final models
for small for gestational age included mother's age, mother's history of one previous fetal loss,
history of two or more previous fetal losses, primiparity, living in an officer's or warrant
officer's household, year of birth, and mother's education. The authors also reported the results
of regression models containing cross-product terms for exposure and maternal age (<35 years,
>35 years) or number of previous fetal losses (none, 1, >2). Among mothers 35 years of age or
older, infants of exposed mothers weighed 104 g less than infants of unexposed mothers (90%
CI: -236, -23). Birth weights of infants born to women less than 35 years of age were not
different between exposure groups. In addition, among women with >2 previous fetal losses,
exposed infants were 104 g lighter than unexposed infants (90% CI: -174, -34). Mother's age
and history of previous fetal loss also appeared to modify the tetrachloroethylene risk for small
for gestational age. The odds ratios for small for gestational age were 1.1 (90% CI: 0.9-1.2) and
2.1 (90% CI: 0.9-4.9) among women <35 and >35 years of age, respectively. There were only
4-302
-------
11 exposed and 8 unexposed small for gestational age infants among mothers older than
35 resulting in effect estimates with lower precision. Odds ratios were 1.1 (90% CLO.9-1.2), 1.5
(90% CI: 1.1-2.0), and 2.5 (90% CI: 1.5-4.3) among women with none, 1, and >2 previous fetal
losses, respectively. There were 43 exposed and 14 unexposed small for gestational age infants
among mothers with >2 previous fetal losses. The authors did not present tests for interaction.
The study found small differences in birth weight and a small increased risk of small for
gestational age among live births to mothers living in two housing areas at the military base with
exposure to tetrachloroethylene and other volatile organic compounds in their drinking water.
Although the impact of residual confounding by unmeasured covariates is not known, a possibly
larger problem may be exposure misclassification. Samples were collected over the last 3 years
of the 17-year study period, although the dry-cleaning business operated during the entire period,
and no operational changes occurred. Water pumped from the contaminated well was mixed
with water from five other wells, but the proportion of water provided from the individual wells
varied from day to day. Variation in concentrations delivered to the tap, as well as individual
consumption and exposure through bathing, could not be evaluated in this study. Further, any
movement on the base prior to delivery was not accounted for. During the course of an exposure
reconstruction study, ATSDR learned that some of the cohort initially considered to be
unexposed were in fact supplied with contaminated water from the Hadnot distribution system
between 1968 and 1972 and for a 2-week period in the winter of 1985 IYNRC. 2009):
www.atsdr.cdc.gov/HS/lejeune/erratum.htmll. Exposed pregnancies during 1968-1972 were
erroneously classified as unexposed. This calls into question the findings in Sonnenfeld et al.
(2001): however, it is likely that as a result of the misclassification, any associations with birth
outcome, if they exist, would have been biased toward the null. Aschengrau et al. (2008) did not
observe an association of tetrachloroethylene in drinking water with either birth weight or
gestational duration. This study, described previously in the discussion of spontaneous abortion,
evaluated effects on pregnancy and development from tetrachloroethylene in drinking water
delivered to homes in the Cape Cod region in Massachusetts between 1968 and 1980. A group
of 1,910 children (1,862 singleton, 24 sets of twins) were born between 1969 and 1983 to
mothers living in one of several Cape Cod towns where tetrachloroethylene leached into drinking
water from vinyl-lined pipes in the water distribution system. Children initially designated as
unexposed (1,853 singleton, 37 sets of twins) were randomly chosen from the remaining resident
births and were frequency matched to the exposed group by month and year of birth. Response
among mothers who were successfully located was comparable between the exposed and
unexposed groups (70%); in the end, 56.4% of selected births designated as exposed were
included, and 54.4% of selected births designated as unexposed were included. After exposure
modeling, 1,353 exposed and 772 unexposed healthy, singleton births were identified.
4-303
-------
The prevalence of prior low birth weight infants in the cohort was low: 5% (n = 68)
among the exposed and 3.4% (n = 26) among the unexposed group. No differences in mean
birth weight or odds ratios for low birth weight (<2,500 g) or intrauterine growth retardation
(<10th percentile based on U.S. age-, sex-, and race-specific cut-offs, 1970-1976) were observed
by exposure status. Generalized estimating equation regression models for birth weight
differences adjusted for gestational age, maternal race, educational level, history of a low-birth-
weight child, occupational exposure to solvents, use of self-service dry cleaning, and proximity
of any residences to dry-cleaning establishments. Mean birth weights were slightly greater
among exposed infants in almost all quartiles for all of the three exposure measures, but the
estimates were statistically imprecise, and no pattern by exposure amount was observed.
Average monthly maternal exposure during the year of the last menstrual period in quartiles was
associated with increases in birth weight of 20.9, 6.2, 30.1, and 15.2 g compared to no exposure.
Models of gestational age were adjusted for maternal race, educational level, prior preterm
delivery, obstetric complications in the current pregnancy, occupational exposure to solvents, use
of self-service dry cleaning, and the proximity of any residences to dry-cleaning establishments.
Estimates of the difference in duration of gestation with increasing quartiles of exposure during
the year of the last menstrual period were -0.2, 0.1, -0.1, and -0.2 weeks. CIs were wide,
included the null, and did not indicate a pattern by exposure amount.
The study of exposure from leaching tetrachloroethylene in water distribution pipes
installed between 1968 and 1980 in the Cape Cod region in Massachusetts also assessed the risk
of congenital anomalies reported by participants (Aschengrau et al., 2009b). Congenital
anomalies were coded by two study investigators, blind to exposure status, in consultation with a
pediatrician using guidelines from the Metropolitan Atlanta Congenital Defects Program. Of the
total of 4,657 children reported by the mothers, 643 were excluded because they were born after
1990, were missing prenatal information, were from multiple pregnancies, were exposed to
known teratogens, mothers smoked marijuana daily or weekly, or drank 7 or more alcoholic
drinks during pregnancy. There were 61 children with congenital anomalies among the 1,658
with prenatal exposure, and 95 children with congenital anomalies among the 2,999 with no
prenatal exposure. The unadjusted odds ratio (generalized estimating equation regression) for all
congenital anomalies was 1.1 (95% CI: 0.8-1.6) for any prenatal exposure to
tetrachloroethylene. Simultaneous control for maternal and paternal age did not change the odds
ratio. This also was true when other potential confounders were included one at a time (calendar
year of birth, mother's educational level, cigarette smoking, alcoholic beverage consumption,
prior pregnancy loss, and child's gender). Among children with an average monthly prenatal
exposure greater than or equal to the 75th percentile (2.3 g), the odds ratio was 1.5 (95%
CI: 0.9-2.5). Although case numbers were low, increased odds ratios were observed for several
4-304
-------
organ systems, diagnostic groups, and any prenatal exposure compared to none. These included
neural tube defects (3.5, 95% CI: 0.8-14.0, n = 6 exposed cases), oral clefts (3.2, 95%
CI: 0.7-15.0, n = 5 exposed cases), gastrointestinal malformations (1.8, 95% CI: 0.7-4.4, n=\\
exposed cases), and genitourinary malformations (1.6, 95% CI: 0.6-3.8, n = 11 exposed cases).
Odds ratios for cardiac (0.9, 95% CI: 0.4-2.0, n = 9 exposed cases) and musculoskeletal
malformations (0.9, 95% CI: 0.5-1.6, n = 19 exposed cases) were not increased, and risk was not
estimated for eye, ear, respiratory, and other malformations because the number of cases was too
low.
As discussed previously, nondifferential exposure misclassification was likely given the
lack of individual level exposure information, which may have resulted in lower observed risk
estimates. In addition, the authors stated that the prevalence of anomalies, particularly minor
ones, may have been underreported by the mothers because it was lower in the study population
than reported by other monitoring programs. This would affect the statistical power of the study.
The authors did not believe that recall was differential with respect to exposure status because
most of the respondents did not know whether or not they were exposed.
Risk of learning and behavioral disorders was evaluated in relation to prenatal and
postnatal exposure to tetrachloroethylene in the Cape Cod towns with a contaminated water
distribution system (Janulewicz et al., 2008). The authors did not observe an association with
increasing amount of exposure among children born between 1969-1983 whose mothers lived in
one of the towns with vinyl-lined asbestos-cement pipes at the time of birth. The study is
discussed in detail in Section 4.1.
In summary, some studies of tetrachloroethylene in drinking water suggest that exposure
during pregnancy is associated with low birth weight (Bove et al., 1995; Lagakos etal., 1986),
eye/ear anomalies (Lagakos et al., 1986), and oral clefts (Aschengrau et al., 2009b: Bove et al.,
1995; Lagakos etal., 1986). No associations with tetrachloroethylene exposure were reported
for small for gestational age (Bove etal., 1995) or other classifications of congenital anomalies
(e.g., musculoskeletal, cardiovascular) (Aschengrau et al., 2009b: Lagakos et al., 1986).
Although a small increase in risk of small for gestational age was reported for infants exposed
prenatally to tetrachloroethylene at the Camp Lejeune military base, the finding remains
inconclusive until ATSDR completes its reanalysis (Sonnenfeld et al., 2001). Aschengrau et al.
(2008) did not observe associations with birth weight or gestational age in a Cape Cod
population exposed to a wide range of tetrachloroethylene concentrations in drinking water.
Occupational studies of dry-cleaning and laundry workers in Scandinavia could not evaluate
specific congenital anomalies because few cases were identified (Lindbohm, 1995; Ahlborg,
1990a: Olsen et al., 1990: Kyyronen et al., 1989: Taskinen et al., 1989). The number of cases
with birth anomalies in specific diagnostic groups was very small in all of the studies, and CIs
4-305
-------
often included one. In addition, imprecise exposure estimates likely resulted in nondifferential
misclassification and a bias of risk estimates toward the null. Participants in the studies were
exposed to multiple contaminants, and it was not possible to analyze substance-specific risks.
Finally, a more than threefold risk of schizophrenia was associated with dry cleaning as a
surrogate for prenatal tetrachloroethylene exposure [Perrin et al. (2007), discussed in
Section 4.1]. The longitudinal design and use of a national registry to identify psychiatric
diagnoses were strengths of the study, but tetrachloroethylene exposure was not directly
analyzed.
4.7.1.2. Animal Developmental Toxicity Studies
Evaluation of the developmental effects of tetrachloroethylene exposure in mammalian
animal models is based on several studies of in utero exposures to maternal animals during
specific periods of pregnancy. Additionally, evaluations of the developmental neurotoxic
potential of tetrachloroethylene have been conducted in rats. These studies are described below.
4.7.1.2.1. In vitro developmental toxicity assay
Saillenfait et al. (1995), using a rat whole embryo (Day 10) culture system, found
tetrachloroethylene-induced embryo toxicity, including mortality, malformations, and delayed
growth and differentiation. No adverse effect was produced at the 2.5 mM concentration, but
concentration-related trends of increasing toxicity occurred from 3.5 through 15 mM. Statistical
tests for a concentration-related trend were not reported. The investigators found that
trichloroethylene produced similar effects, with potency somewhat less than that of
tetrachloroethylene. They also found that TCA and DCA caused a variety of abnormalities in
this culture system.
4.7.1.2.2. Nonmammalian developmental toxicity assay
Spencer et al. (2002) evaluated the effects of tetrachloroethylene on the embryonic
development of Japanese medaka. In this study, 1-day-old in ovo embryos were exposed to
concentrations of 0, 20, 40, 60, or 80 mg/L for 96 hours or to concentrations of 0, 1.5, 3, 6, 12, or
25 mg/L for 10 days. Viability, hatchability, and morphological/developmental abnormalities
were evaluated. A 96-hour LCso of 27.0 mg/L was identified for egg viability. Following
10 days of exposure, hatchability and larval survival were significantly decreased, and
developmental abnormalities were significantly increased in a concentration-dependent manner.
At the lowest concentration tested (1.5 mg/L), developmental findings included abnormalities of
the circulatory system, yolk-sac edema, pericardial edema, scoliosis, hemorrhaging, blood
pooling, and cardiac morphological defects. The study authors concluded that
tetrachloroethylene is teratogenic to the Japanese medaka.
4-306
-------
4.7.1.2.3. In vivo mammalian screening study
In a developmental toxicity screening study, timed-pregnant F344 rats were treated by
gavage with tetrachloroethylene at doses of 900 or 1,200 mg/kg-day in corn oil vehicle on GDs
6-19 (Narotsky and Kavlock, 1995). There were 17 dams in each of the tetrachloroethylene-
treated groups and 21 in the control groups. The dams were allowed to deliver, and their litters
were examined on PNDs 1, 3, and 6. At 1,200 mg/kg, no live pups were delivered on GD 22. At
900 mg/kg-day, there was maternal ataxia, and weight gain was markedly less than in the
controls. The number of pups per litter was reduced (p < 0.01) as compared with the controls at
GD 22. On PND 6, the number of pups per litter was reduced (p < 0.001) as compared with the
controls. The investigators noted that full-litter resorptions were not observed with other
chemicals they tested in the presence of maternal toxicity. An increase in micro/anophthalmia
was found in the offspring. There was no evaluation for skeletal changes, and not all available
pups were examined for soft tissue changes.
4.7.1.2.4. In vivo prenatal developmental toxicity studies
Schwetz et al. (1975) conducted an inhalation developmental toxicity study, in which
25-30 Sprague-Dawley rats and 30-40 Swiss-Webster mice were exposed to airborne
tetrachloroethylene at 300 ppm, 7 hours/day, on GDs 6-15. Following laparohysterectomy on
GDs 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 authors concluded that there was no significant maternal, fetal, or embryo
toxicity for any of the solvents tested, the maternal and fetal data demonstrated a number of
statistically significant differences from control values following gestational exposures to
tetrachloroethylene in rats and mice. In the rats, exposures to tetrachloroethylene produced
slight, but statistically significant, maternal toxicity (4-5% reductions in mean maternal body-
weight gains) and embryotoxicity (increased resorptions; 9% in treated vs. 4% in controls). In
the mice, maternal toxicity consisted of a significant 21% increase in mean relative liver weight
as compared with controls. The mean fetal weight in mice was significantly (9%) less than in the
concurrent control, and the percentage of litters with delayed ossification of the skull bones,
delayed ossification of the sternebra, and subcutaneous edema was significantly increased. Due
to the single exposure level used in this study, a dose response could not be determined.
Szakmary et al. (1997) exposed CFY rats to tetrachloroethylene via inhalation throughout
gestation (i.e., GDs 1-20) for 8 hours/day at concentrations of 1,500, 4,500, or 8,500 mg/m3. In
the same study, the study authors exposed C57B1 mice via inhalation on GDs 7-15 (i.e., during
4-307
-------
the period of organogenesis) to a concentration of 1,500 mg/m3 and New Zealand white rabbits
during organogenesis (GDs 7-20) to a concentration of 4,500 mg/m3. Maternal animals were
killed approximately 1 day prior to expected delivery; a gross necropsy was conducted, organ
weights were recorded, blood was taken by aorta puncture for hematology and clinical chemistry
evaluations, ovarian corpora lutea were counted, and uterine contents were examined (number
and position of living, dead, or resorbed fetuses; and fetal and placental observations and
weights). The numbers of litters available for evaluation were as follows: 20 control and 21 or
22 per treated group in the rat, 77 control and 10 treated in the mice, and 10 control and
16 treated in the rabbit. One-half of the fetuses from each litter were evaluated for visceral
abnormalities, and the other half were evaluated for skeletal development. The study authors
reported that the organs of five dams and five embryos from each group were also evaluated by
routine histological methods. To evaluate the concentration of tetrachloroethylene in maternal
and fetal blood and in amniotic fluid, another subset of rats (number not specified) was studied.
(For the 1,500 and 8,500 mg/m3 exposure levels, maternal blood concentrations of
tetrachloroethylene were 17.8 + 8.9 and 86.2 + 13.0 uL/mL, respectively. Concentrations in the
fetal blood were 66 and 30% of maternal blood concentrations, and amniotic fluid concentrations
were 33 and 20% of maternal blood concentrations.) In the rat, at 4,500 and 8,500 mg/m3,
maternal body-weight gain during gestation was significantly decreased (37 and 40%,
respectively), relative maternal liver mass was significantly increased (10 and 6%, respectively),
and serum aspartate amino transferase activity was increased (data not provided) as compared to
controls. Percentage preimplantation loss was significantly increased from controls by 133 and
117% at these exposure levels, while percentage postimplantation loss was increased
nonsignificantly from controls by 80% in each group. Also, at 4,500 and 8,500 mg/m3, fetal
weight was significantly decreased in 98.5 and 100% of all fetuses, the number of fetuses with
skeletal retardation was significantly increased in 98.5 and 100% of fetuses, and the percentage
of fetuses with malformations was both significantly increased to 6.4 and 15.7% as compared to
the control incidence of 2.0%. Although the study authors judged the 1,500 mg/m3 exposure
level to be the NOAEL for the rat study, it is noted that there were concentration-dependent
nonsignificant decreases in maternal body-weight gain (13% lower than control), and increases
in pre- and postimplantation loss (49 and 38% greater than control, respectively). The
percentage of weight-retarded fetuses increased to 3.4 times the control incidence, and the
incidences of fetuses with skeletal retardation (48% increased) or total malformations increased
by 2.3 times the control incidence observed at the low-exposure level of 1,500 mg/m3.
Therefore, these findings are judged to be adverse consequences of treatment. The attribution of
these findings to treatment, and the designation of 1,500 mg/m3 as the study LOAEL is
consistent with the adverse developmental findings of Schwetz et al. (1975). In mice
4-308
-------
(1,500 mg/m3) and rabbits (4,500 mg/m3), relative liver mass was significantly increased;
decreased maternal body-weight gain was also observed in the rabbits. In the mice, a
significantly increased number of fetuses with visceral malformations (details not specified) was
observed, while in the rabbits, 2/16 does aborted, total resorption of four litters was reported, and
the percentage of postimplantation loss was significantly increased. The percentage of rabbit
fetuses with malformations (details not provided in the report) was also increased, although not
significantly.
Hardin et al. (1981) [also refer to Beliles et al. (1980)1 exposed Sprague-Dawley rats
(30/group) and New Zealand white rabbits (20/group) via inhalation to 500 ppm of
tetrachloroethylene for 7 hours/day, 5 days/week. Tetrachloroethylene was administered with
and without 3-week pregestation exposures and with both full-term and terminal two-thirds-term
exposure. No maternal or developmental toxicity was identified.
In a developmental toxicity study, Carney et al. (2006) investigated the effects of whole-
body inhalation exposures to pregnant Sprague-Dawley rats at nominal concentrations of 0, 75,
250, or 600 ppm (actual chamber concentrations of 0, 65, 249, or 600 ppm) tetrachloroethylene
for 6 hours/day, 7 days/week on GDs 6-19. This study was conducted under Good Laboratory
Practice (GLP) regulations according to current EPA and OECD regulatory testing guidelines.
Maternal toxicity consisted of slight, but statistically significant, decreases in body-weight gain
during the first 3 days of exposure to 600 ppm, establishing a no-adverse-effect concentration of
249 ppm for dams. A slight, statistically significant decrease in gravid uterine weight at 600
ppm correlated with significant reductions in mean fetal body weight (9.4%) and placental
weight (15.8%) at GD 20 cesarean section. At >249 ppm, mean fetal and placental weights were
significantly decreased by 4.3 and 12.3% from control, respectively. A significant increase in
the incidence of incomplete ossification of the thoracic vertebral centra at this exposure level was
consistent with fetal growth retardation. No treatment-related alterations in fetal growth or
development were noted at 65 ppm. Therefore, the LOAEL for this study is 249 ppm.
4.7.1.2.5. Developmental neurotoxicity
Developmental neurotoxicity data are also discussed in Section 4.1.2.
A cohort of rats from the Szakmary et al. (1997) study (15 litters/group at exposure levels
of 1,500 or 4,500 mg/m3 tetrachloroethylene) was allowed to deliver, and the offspring
(standardized to 8 pups/litter) were maintained on study to PND 100. It was not clearly specified
in the report whether the daily inhalation exposures continued throughout the postnatal period.
Preweaning observations included weekly body weights, developmental landmarks (pinna
detachment, incisor eruption, and eye opening), and functional assessments (forward movement,
surface righting reflex, grasping ability, swimming ontogeny, rotating activity, auditory startle
4-309
-------
reflex, and examination of stereoscopic vision). After weaning, exploratory activity in an open
field, motor activity in an activity wheel, and development of muscle strength were assessed.
The study authors reported that adverse findings included a decreased survival index (details not
provided), minimally decreased exploratory activity and muscular strength in treated offspring
(presumably at both exposure levels) that normalized by PND 51, and significantly increased
motor activity on PND 100 of females exposed to 4,500 mg/m3 of tetrachloroethylene.
Nelson et al. (1979) investigated developmental neurotoxicity in Sprague-Dawley rats by
exposing pregnant dams (13-21/group) to tetrachloroethylene at concentrations of 100 ppm or
900 ppm during either early pregnancy (GDs 7 to 13) or late pregnancy (GDs 14 to 20).
Morphological examination of the fetuses (gross, visceral, and skeletal) was performed, and
behavioral testing and neurochemical analyses of the offspring were conducted. There were no
alterations in any of the measured parameters in the 100 ppm groups. At 900 ppm, there were no
skeletal abnormalities, but the weight gain of the offspring as compared with controls was
depressed approximately 20% at postnatal Weeks 3-5. Developmental delays were observed in
both the groups exposed during early and late pregnancy. Offspring of the early pregnancy-
exposed group performed poorly on an ascent test and on a rotorod test, whereas those in the late
pregnancy group underperformed on the ascent test at only PND 14. However, later in
development (Days 21 and 25), their performance was higher than that of the controls on the
rotorod test. These pups were markedly more active in the open field test at Days 31 and 32.
Activity wheel testing on Days 32 and 33 did not reveal statistically significant changes.
Avoidance conditioning on Day 34 and operant conditioning on Days 40-46 did not identify
treatment-related effects. Neurochemical analyses of whole brain (minus cerebellum) tissue in
21-day-old offspring revealed significant reductions in acetylcholine levels at both exposure
periods, whereas dopamine levels were reduced among those exposed on GDs 7-13. All of the
described effects in the 900 ppm group were statistically significant as compared with controls.
Unfortunately, none of the statistics for the 100 ppm treatments were presented. The authors
observed that more behavioral changes occurred in offspring exposed during late pregnancy than
in those exposed during early pregnancy.
Additional evidence of potential developmental neurotoxicity was reported by
Fredriksson et al. (1993). In this study (refer to Section 4.1.2.2), tetrachloroethylene was
administered to male NMRI mice by gavage at dose levels of 0, 5, or 320 mg/kg-day on PNDs
10-16. At PND 17 and 60, spontaneous activity (locomotion, rearing, and total activity) was
measured over three, 20 minute periods. No treatment-related alterations in activity were
observed at 17 days of age; however, at 60 days of age, all three measures of spontaneous
activity were altered.
4-310
-------
4.7.2. Reproduction
4.7.2.1. Human Reproduction Data
Studies of tetrachloroethylene exposure have evaluated several outcomes including
effects on menstrual disorders (Zielhuis etal., 1989), semen quality (Eskenazi etal., 1991a),
fertility (Eskenazi et al., 1991b: Rachootin and Olsen, 1983), time to pregnancy (Sallmen et al.,
1998; Sallmen et al., 1995), and spontaneous abortion (Doyle et al., 1997; Lindbohm et al., 1991;
Windham et al., 1991: Ahlborg, 1990a: Lindbohm et al., 1990: Olsen etal., 1990: Kyyronen et
al., 1989: Taskinen et al., 1989: Boscoetal., 1987: McDonald et al., 1987: McDonald et al.,
1986). Many of the studies evaluated exposure during a specific critical window for
development, usually the first trimester.
In a letter to the editor, Zielhuis et al. (1989) described the results of a cross sectional
study of menstrual disorders among dry-cleaners and laundry workers in the Netherlands. A
total of 471 of 592 women returned a mailed questionnaire (80%). The sampling frame for
recruitment was not described. After excluding 72 respondents because the woman was
currently pregnant or lactating at the time of administering the questionnaire or reported a
chronic illness or gynecological surgery, and excluding another 324 respondents because the
woman reported use of oral contraceptives, the final data set included 68 exposed and 76
unexposed women. Exposure was defined on the basis of occupation (dry cleaners versus
laundry workers). The authors reported that the exposed and unexposed groups were similar
with respect to age, lifestyle, work conditions, and personal characteristics (body mass index,
number of children, and use of contraceptives). Risk of specific menstrual characteristics by
occupation was evaluated using linear logistic regression adjusting for age, body mass index,
substantive weight changes, number of children, history of diseases, sporting activities, life
events, smoking, alcohol consumption, medical drugs, and work conditions other than exposure
to tetrachloroethylene. Prevalence of menstrual conditions in the population varied between
10% (oligomenorrhea, premenstrual syndrome) to 30% (unusual cycle length) and occurred with
greater frequency among dry cleaners compared to laundry workers for all symptoms except for
one (polymenorrhea). There were no reports of amenorrhea. Elevated odds ratios were observed
for several of the symptoms including oligomenorrhea (2.1, 90% CI: 0.9-5.3), unusual cycle
length (2.3, 90% CI: 1.2-4.4), menorrhagia (3.0, 90% CI: 1.6-5.6), dysmenorrhea (1.9,
90% CI: 1.1-3.5), and premenstrual syndrome (3.6, 90% CI: 1.5-8.6). This study indicates that
working in dry cleaning may adversely affect menstruation, but the lack of detail in reporting
precludes a thorough assessment of selection bias or confounding. In addition, the assignment of
exposure status by industry also precludes a definitive conclusion regarding a potential
association with tetrachloroethylene.
4-311
-------
Semen quality was evaluated among men who worked in the dry-cleaning industry
compared to men working in laundries in California (Eskenazi et al., 199la). The population,
recruited from membership lists of the Laundry and Dry Cleaners Union Locals 3 (San Francisco
Bay area) and 52 (Greater Los Angeles), included all dry cleaners (n = 85) and all laundry
workers, 20-50 years of age, in Local 3 (n = 119) and a randomly selected sample of Local 52
members (n = 206). Laundry workers were frequency matched by age to dry cleaners from the
same Local. Dry cleaners also were recruited from nonunion shops in the San Francisco area.
Eligible individuals were 20-50 years of age, current workers in the industry, spoke English or
Spanish, had not had a vasectomy, and were located by telephone or mail. Respondents included
20 union drycleaners (38% of 53 eligible) and 56 union laundry workers (34% of 166 eligible),
plus 13 nonunion dry cleaners. Men were considered exposed if they worked in the dry-cleaning
industry or a laundry where dry cleaning was performed. The unexposed group included laundry
workers at businesses where dry cleaning was not conducted. After exposure was assessed, the
final data set included 34 exposed workers and 48 unexposed workers with adequate semen
samples and confirmed type of establishment. Information on sociodemographic characteristics,
reproductive and medical history, and personal habits was collected by interview. In addition, a
detailed work history including job tasks and exposures during the previous week and the past
3 months was obtained. A physical exam was conducted by a study physician blind to exposure
status, and participants returned a semen sample collected after at least 2 days of abstinence.
The semen was analyzed for sperm concentration, morphology, and motility. Each sperm
measure was evaluated in relation to three measures of exposure: dry cleaning versus laundry,
tetrachloroethylene in exhaled breath (limit of quantitation: 2.67 |ig/m3), and an exposure index
encompassing the entire period of spermatogenesis (approximately 3 months). Exhaled air was
measured 16-19 hours after the end of the workweek or was corrected to 16 hours using an
elimination model (11 samples). An industrial hygienist assigned an exposure score using
responses to the questionnaire concerning job task (e.g., machine operator, presser, etc.), the type
of dry-cleaning machine used (e.g., wet to dry transfer, dry-to-dry) and other tasks and attributes
known to influence the level of exposure to tetrachloroethylene. The exposure score ranged
from 0 among unexposed men to 11 among the exposed group. The association of semen
parameters with tetrachloroethylene exposure was analyzed using multiple linear regression with
adjustment for potential confounding variables that were associated with both the semen
parameter and any of the exposure measures. Models of three clinically relevant measures of
semen quality, oligospermia (<20 million/mL), >40% abnormal forms, and <60% motile sperm,
were not associated with any exposure measure among the entire cohort. Of four measures of
sperm motility, Ln median amplitude of lateral head displacement was associated with Ln
tetrachloroethylene in exhaled air among all 82 participants (t = 2.0, p = 0.05), adjusting for
4-312
-------
ethnicity, education, religion, and physical abnormalities found on exam. Exposure scores and
industry group were not statistically significant predictors of this semen parameter. However, Ln
tetrachloroethylene levels (t = 2.14,p = 0.04) and exposure score (t = 3.07,/> = 0.005) were
predictors of amplitude of lateral head displacement among the 34 participants in the exposed
group. Sperm linearity was inversely associated with exposure score in both analytic groups
(t = -2.57, p = 0.02). Percentage of round sperm was statistically significantly associated with
all three exposure measures, controlling for history of STD and working in temperatures over
100°F among all participants but not in the dry-cleaning group alone. Percentage of narrow
sperm was inversely associated with all three exposure measures controlling for ethnicity,
number of days working in temperatures greater than 80°F, and use of marijuana among all
participants. Among the dry cleaners, Ln percentage narrow sperm was inversely related to Ln
tetrachloroethylene levels (t = -2.29,p = 0.03) but not by exposure score (t = 0.92., p = 0.36).
Tetrachloroethylene exposure appeared to alter sperm quality in this population of
unionized dry cleaners. However, the effects were subtle, and the clinical significance of the
semen parameters associated with tetrachloroethylene exposure is not clear. The low response
rate in the primarily unionized cohort limits generalizations to the industry as a whole.
Reproductive outcomes also were evaluated among the wives of the men who participated in the
study of semen quality (Eskenazi etal., 199la). Telephone interviews were conducted with 17
wives of the 20 married dry cleaners (85%) and 32 wives of the 36 married laundry workers
(89%) in the original cohort. Pregnancies and miscarriages during the years of their husbands'
employment in the industry were identified among 14 wives of dry cleaners and 26 wives of
laundry workers. Standardized fertility ratios were calculated using the U.S. national birth rates
during periods of employment in the industries and periods when the men were not employed in
the industries as a comparison. Investigators also analyzed the number of months to conception
for the last pregnancy during the period of employment in the industries. The wives of laundry
workers were more likely to be Hispanic, Catholic, to have smoked during the year of the index
pregnancy, and to have a history of reproductive disease or surgery. They had fewer years of
education, and a greater proportion weighed more. The wives also were more likely to work in
dry cleaning and laundries, confounding the source of exposure.
Fertility rates among the wives of dry-cleaners and laundry workers were higher than the
national average for women of the same race, parity, birth cohort, and age. The standardized
fertility ratios were comparable in both industry groups. However, it took longer for the wives of
dry cleaners to achieve the index pregnancy compared to the wives of laundry workers
(8.2 ± 10.2 months versus 4.1 ± 5.8 months, respectively,/? = 0.08). In Cox Proportional
Hazards Models with adjustments for ethnicity (Hispanic vs. non-Hispanic) and smoking, the
per-cycle pregnancy rate of wives of dry cleaners was approximately one-half that of the wives
4-313
-------
of laundry workers (rate ratio = 0.54, 95% CI: 0.23-1.27). A rate ratio of less than 1 also was
indicated in models using husbands' exhaled tetrachloroethylene (rate ratio = 0.94, 95% CI:
0.85-1.04) and husbands' exposure index (rate ratio = 0.90, 95% CI: 0.78-1.03). The latter two
exposure indices may not have estimated exposure during the sensitive window for the index
pregnancy, however. The small sample size resulted in CIs that included the null hypothesis.
The authors noted that to detect a halving of risk for pregnancy with 80% power (a = 0.05), over
50 women per group would have been required.
A Danish case-control study of couples examined or treated for infertility during
1977-1980 reported evidence of idiopathic infertility among women reporting exposure to dry-
cleaning chemicals (Rachootin and Olsen, 1983). Controls were couples with a healthy child
born at the same hospital during 1977-1979. Information about occupational and reproductive
history was obtained from 87% of both cases and controls who returned a mailed, self-
administered questionnaire during November 1980 to May 1981. Participants were defined as
exposed if they reported contact with any of 15 types of chemicals and physical agents
(including dry cleaning) and three specific work processes a minimum of once per week for at
least 1 year in the period prior to hospital admission. The medical records of infertile couples
were reviewed by a collaborating physician who had no knowledge of exposures. Three analytic
approaches were used to evaluate subgroups of couples with a medical history anticipated to be
related to occupational exposures. Reported exposure to dry-cleaning chemicals was associated
with idiopathic infertility among women compared to fertile couples with a healthy child
conceived within 1 year (OR: 2.7, 95% CI: 1.0-7.1). The statistical method was not described,
but the authors stated that the odds ratio was adjusted for the women's age, education, residence,
and parity. Cases and controls lived within the catchment area of the hospital. Exposure to dry-
cleaning chemicals was not associated with sperm abnormalities or idiopathic infertility among
male partners or with hormonal disturbances among women. The odds ratio for idiopathic
infertility among women with exposure to dry-cleaning chemicals also was increased when
couples who had been infertile for at least 1 year were compared to other infertile couples with
conditions believed to be unrelated to occupational exposures (crude odds ratio [ORc] = 1.8,
95% CI: 0.5-5.8). A third analysis involved comparison within the control group; couples who
experienced a delay in conception of more than 1 year but who gave birth to a healthy child were
compared to couples who conceived a healthy child in less than 1 year. Again, women reporting
exposure to dry-cleaning chemicals had an increased odds ratio for delayed conception
(ORc: 1.6, 95% CI: 0.9-2.9). Although two of the risk estimates did not reach statistical
significance, all three were greater than 1.5. The consistent increased odds ratios observed using
three different comparison groups suggest an effect of exposure to dry-cleaning chemicals on
conception among women. The study evaluated a large number of chemicals and physical
4-314
-------
exposures. The authors did not present the number of cases by subgroup, or the number of
controls who reported exposure to dry-cleaning chemicals, so it is difficult to assess the impact
of sample size on the precision of the effect estimates. Other chemical exposures, as well as
noise, also were associated with idiopathic infertility among the women. In addition, the
statistical analyses for dry-cleaning chemicals did not control for exposure to other chemical or
physical agents.
Sallmen et al. (1995) conducted a retrospective time-to-pregnancy study among Finnish
women biologically monitored at the Institute of Occupational Health in 1965-1983 for one or
more of six solvents (styrene, toluene, xylene, tetrachloroethylene, trichloroethylene, and
1,1,1-trichloroethane). This study was an extension of an investigation of the risk of
spontaneous abortion in the same study population. That study is described later in this section
(Lindbohm et al., 1990). Pregnancies and their outcomes (live birth, spontaneous abortion, or
fetal loss) between 1973 and 1983 among the women had been identified using a national
register of pregnancies in Finland and the Finnish Register of Congenital Malformations. Time-
to-pregnancy information was obtained through questionnaires mailed to 355 women who were
the cases and controls in the previous study. Information about exposure during the preceding
12 months before each woman's pregnancy began was collected. The response rate was 66%,
and the final data set contained 197 women who had been attempting to become pregnant, had
no other risk factors for infertility, for whom complete information was available on exposure
and time-to-pregnancy. Time-to-pregnancy was defined as the number of menstrual cycles
required to become pregnant and is a measure of fertility, the per cycle probability of conceiving
a clinically detectable pregnancy. Increased time-to-pregnancy can indicate a loss during
pregnancy during any stage from gametogenesis to fertilization to the clinical stage of
pregnancy, including early stage spontaneous abortions.
The same exposure-assessment procedure as was used in the previous study was adopted
for this study, and if the subject reported working in the same job, their previous exposure
classification was used. Self-reported work tasks during the 12 months prior to conception were
assigned to an exposure classification by likelihood and level of exposure for 84 women whose
jobs or exposures were different than reported previously for the first trimester. Classifications
were made without knowledge of reproductive history and were checked by an independent,
experienced industrial hygienist. The three categories for likelihood of exposure were not
exposed, potentially exposed, and exposed. Subjects were grouped according to high (n = 46),
low (n = 59), and none (n = 92) for level of exposure [refer to description of Lindbohm et al.
(1990)].
Exposure to organic solvents during their time-to-pregnancy was reported by more than
one-half of the women (105 out of 197). Incidence density ratios, indicating the likelihood that
4-315
-------
exposed women will achieve a clinical pregnancy during the fertile period in each menstrual
cycle class (e.g., 1st menstrual cycle, 2nd, 3rd and 4th, 5th, 6th, etc.) compared to an unexposed
woman, were estimated using discrete proportional hazards regression. Incidence density ratios
(IDRs) were reported for women exposed to tetrachloroethylene (n = 20) or working in dry
cleaning (n = 17). Compared to women with no exposure, the IDRs for low and high exposure
were 0.63 (95% CI: 0.34-1.17) and 0.69 (95% CI: 0.31-1.52), respectively. The statistical
models controlled for exposure to other solvents, recent contraceptive use, and age at menarche.
For workers in dry cleaning, the IDR for 11 women with low or high exposure combined was
0.44 (95% CI: 0.22-0.86) and for 6 women with high exposure was 0.57 (95% CI: 0.24-1.34).
These models controlled for low and high exposure to solvents in other industries, recent use of
IUD/spermicides, and age at menarche. The model for high exposure also adjusted for low
exposure to organic solvents. The authors noted that only 1 of the 11 women who worked in dry
cleaning reported exposure to other solvents in addition to tetrachloroethylene. These results
suggest that exposure to tetrachloroethylene may affect fecundability, however, because the
focus was on a broad range of solvent exposures and industries, the sample size for assessing
tetrachloroethylene was small, and statistical precision was low. However, the study had several
strengths, including collection of detailed work histories. Exposure classifications were based on
the frequency of solvent use, not just reported use ever or job title. In addition, several potential
confounders were assessed, and statistical models controlled for exposure to other solvents. It
was not clear if the models for individual solvents were assessed for confounding by case status
(i.e., pregnancy ended in a spontaneous abortion). However, reduced fecundability was
associated with exposure to organic solvents combined in separate analyses of cases and
controls. The low response rate overall, and evidence that response was higher among cases and
exposed controls, particularly those with lower parity, raises the possibility of selection bias.
Time-to-pregnancy also was evaluated among the wives of men exposed to organic
solvents and monitored by the Finnish Institute of Occupational Health during 1965-1983
(Sallmen et al., 1998). This was an extension of an earlier case-referent study of risk of
spontaneous abortion [refer to description later in this section of Taskinen et al. (1989)]. The
investigators used a similar approach as that used in Sallmen et al. (1995), described above.
Cases (n = 110) and referents (n = 332) that participated in Taskinen et al. (1989) were recruited.
Time-to-pregnancy information was obtained through questionnaires mailed to 355 women who
were the cases and controls in the previous study. A detailed history of occupation and work
tasks during the year the pregnancy started had been obtained from the husbands in the previous
study. A similar history was now requested of the wives, focusing on the preceding 12 months
before the pregnancy. The response rate was 72%, and the final data set contained 282 women
who had been attempting to become pregnant, had no other risk factors for infertility, for whom
4-316
-------
complete information was available on exposure and time-to-pregnancy. The same exposure-
assessment procedure as was used in the previous study was adopted for this study, and if the
subject reported working in the same job at the beginning of the pregnancy, their previous
exposure classification was used. A new exposure classification was required only for nine men
whose jobs or exposures were different than reported previously for the first 3 months before the
pregnancy began. Classifications were made without knowledge of reproductive history and
were checked by an independent, experienced industrial hygienist (Taskinen et al., 1989). The
three categories for likelihood of exposure were, not exposed, potentially exposed, and exposed.
Subjects were grouped according to high/frequent (n= 141), intermediate/low (n = 80), and
unexposed (n = 61) for level of exposure to organic solvents during the time-to-pregnancy
period.
Incidence density ratios (IDRs) were reported for exposure to all organic solvents
combined and for specific solvents. The IDRs for low (n = 9) and combined intermediate/high
(n = 8) exposure to tetrachloroethylene were 0.86 (95% CI: 0.4-1.84) and 0.68 (95%
CI: 0.30-1.53). The discrete proportional hazards regression models were adjusted for short
menstrual cycle, long or irregular menstrual cycle, older age at menarche, frequency of
intercourse, maternal age, maternal exposure to organic solvents, and a variable for missing
information. Fecundity appeared most reduced among the wives whose husbands had a high
level and/or frequency of tetrachloroethylene exposure compared to low or no exposure.
However, the study was limited by low statistical precision because of small sample size. Time-
to-pregnancy information and exposures were collected 8 to 18 years after the pregnancy of
interest, which likely resulted in some misclassification. It is less likely that recall bias affected
the risk estimates because the exposures were assigned based on information collected for the
earlier study of spontaneous abortion.
Among studies evaluating effects of tetrachloroethylene on reproduction and
development, the majority of studies assessed effects on risk of spontaneous abortion. These
studies defined spontaneous abortion as a fetal loss prior to 20-28 weeks gestation, although one
study included all fetal loss during the first 6 months of pregnancy (Lagakos et al., 1986).
Several studies included only clinically recognized spontaneous abortions reported in birth
registers (Lindbohm et al.. 1991: Windham et al.. 1991: Ahlborg, 1990a: Lindbohm et al.. 1990:
Olsenetal.. 1990: Kyyronen et al.. 1989: Taskinen et al.. 1989: McDonald et al.. 1987:
McDonald et al., 1986), while some included spontaneous abortions reported by participants
(Aschengrau et al.. 2009a: Aschengrau et al.. 2008: Doyle etal.. 1997: Eskenazi et al.. 199la:
Bosco etal., 1987: Lagakos etal., 1986). It should be noted that it is not possible to identify all
spontaneous abortions that occur in populations because a woman may not recognize very early
events and/or may not seek treatment.
4-317
-------
McDonald et al. (1987; 1986) conducted a large survey of occupation and reproductive
outcomes among 56,012 women in 11 large obstetrical units in Montreal, Canada, over a 2-year
period from May 11, 1982 to May 10, 1984. Interviews were conducted with 51,885 women
with a term delivery and 4,127 women treated in the hospital for a spontaneous abortion, defined
in this study as a fetal loss <28 weeks of gestation. The 11 hospitals included in the survey
treated approximately 90% of all births in Montreal. As part of the interview, women were
asked to describe all previous pregnancies that ended in a spontaneous abortion, and 10,910 were
identified. Interviews were completed for 90% of the women with term births, and 75% of
women admitted for a spontaneous abortion. Information also was collected about occupation at
the time of conception for the current and any previous pregnancies. Nine occupational groups
in the Canadian Classification and Dictionary of Occupations were reduced to six major
groupings and included 42 categories that the investigators concluded were homogenous.
Logistic regression was used to evaluate risk of spontaneous abortion for five nonoccupational
factors: maternal age, parity, history of a previous abortion, smoking habit, and highest
educational level reached, and the expected number of spontaneous abortions for each
occupational group was calculated. The ratio of observed to expected numbers was evaluated for
each occupational group. Among women in the laundry and dry-cleaning occupational grouping,
there were 8 spontaneous abortions out of 100 recent pregnancies (O/E: 1.18;/> > 0.1 [CI not
reported]) and 31 out of 123 previous pregnancies (O/E: 1.02). Subsequent analysis of the data
included women who worked at their jobs for at least 30 hours weekly at the beginning of
pregnancy (McDonald et al., 1987). In this analysis, 36 combined current and previous
spontaneous abortions were observed out of 202 pregnancies. An O/E ratio of 1.05 (p > 0.1; CI
not reported) was reported. The expected number was determined from a logistic regression
model of spontaneous abortion risk including maternal age, parity, history of a previous abortion,
smoking habit, and alcohol consumption. This study is not very informative regarding
tetrachloroethylene risk because the group of dry-cleaners and laundry workers likely included
individuals with no exposure to the solvent.
A case-referent study of adverse pregnancy outcome was conducted among the wives of
male workers who had been monitored for organic solvents by the Finnish Institute of
Occupational Health between 1965 and 1983 (Taskinen et al., 1989). The cohort included men
in their first marriage during 1985 with wives who were 18-40 years old at the end of the 1st
trimester of pregnancy. Pregnancies and outcomes were identified through national registers.
Eligible pregnancies began during the marriage or up to 9 months before. Cases were defined as
wives with a spontaneous abortion (if multiple, one randomly selected) or a congenitally
malformed child. Referents were selected from wives with a healthy birth between 1973 and
1983 (1:3 for spontaneous abortions, 1:5 for malformations), age matched within 30 months. A
4-318
-------
total of 136 of 172 selected cases (79.1%) and 370 of 505 selected referents (73.3%) responded
to a questionnaire mailed in January 1986. Only pregnancies that were identified in the register
and reported by participants were included. Because of this, and because a matched response
was required, the final data set included 120 cases and 251 referents. Information on occupation
and exposure to solvents during the year of conception was requested of the men. Information
on occupational and other exposures during the first trimester of pregnancy was solicited from
the wives. Exposure classifications were made blind to pregnancy outcome. Solvent exposure
for the men was assessed for an 80-day period that preceded the pregnancy, the relevant period
of spermatogenesis, using information on occupation, job description, reported solvent or other
chemical use, and biological monitoring data. Workers were classified as not exposed if work
tasks did not include handling solvents and no exposure was reported, and no biological
measurement for a particular solvent was made. Workers were classified as potentially exposed
if work tasks might have involved solvent use, but use was not reported by the worker, and no
biological measurements for a particular solvent were made. Workers were classified as exposed
if biological measurements for a solvent were taken while at the same job for the reported
pregnancy, reported tasks implied solvent exposure, or solvent exposure was reported. Exposure
was categorized into none, low, intermediate, or high. Workers with high exposure handled
solvents daily, or their biological measurements were above the reference value for the general
population. Workers with intermediate exposure used solvents 1-4 days per week, and
biological measurements indicated intermediate or low exposure. Workers with low exposure
handled solvents <1 day per week. All other scenarios were classified as no exposure.
A spontaneous abortion rate of 8.8% was observed among all recognized pregnancies, a
rate within the range reported for Finland between 1973 and 1983 (Lindbohm and Hemminki,
1988). The unadjusted odds ratio for risk of spontaneous abortion in relation to likely paternal
exposure to tetrachloroethylene was 0.5 (95% CI: 0.2-1.5) using conditional logistic regression.
Likely exposure was assigned to 4 cases and 17 referents. Adjusted odds ratios controlling for
potential paternal exposure to the solvent, likely paternal exposure to other organic solvents and
dusts, maternal exposure to solvents, maternal heavy lifting, and history of previous spontaneous
abortion were presented only for likely exposure to all halogenated hydrocarbons. In addition to
exposure to tetrachloroethylene, this group included exposure to trichloroethylene and
1,1,1-trichloroethane. The adjusted odds ratios for low/rare, intermediate, and high/frequent
exposure were 1.1 (95% CI: 0.5-2.6), 1.3 (95% CI: 0.5-3.1), and 0.8 (95% CI: 0.3-2.2),
respectively. The exposure assessment encompassed a broad range of solvents, and only a small
number reported exposure to tetrachloroethylene. In addition, exposure to multiple chemicals
was possible for much of the cohort, and this was not controlled for in the chemical-specific
models.
4-319
-------
A subsequent study of paternal occupational exposure and spontaneous abortions
attempted to identify all medically recognized pregnancies (spontaneous abortion, induced
abortion, and healthy births) between 1973 and 1982 through the Finnish nationwide Hospital
Discharge Register and from outpatient hospital clinics (Lindbohm et al., 1991). Information on
occupation was obtained from 1975 and 1980 national census records. Pregnancies during 1973
to 1978 were linked to the 1975 Census, and pregnancies during 1979 to 1982 were linked to the
1980 Census. A job-exposure classification, developed in cooperation with two industrial
hygienists, assigned chemical exposures commonly used by job groups within industries.
Exposures were assigned to job groupings using a list of 78 exposures, including specific
substances, mixtures, and nonspecific exposures, plus industrial hygiene measurements made by
the Institute of Occupational Health and the Finnish register of employees occupationally
exposed to carcinogens. Exposure assessment focused on mutagens, and three levels were
defined: moderate/high, potential/low, and none.
The susceptible exposure period of interest was an 80-day period prior to conception
corresponding to spermatogenesis. Because the investigators did not have temporally resolved
exposure information, pregnancies that were terminated during a 2-year period close to the
census were selected (January 1, 1976-December 31, 1977 for the 1975 Census, and May 1,
1980-April 30, 1982 for the 1980 Census). A total of 99,186 pregnancies to women aged
12-50 years with complete information about occupation, industry, and socioeconomic status
occurred during these time periods. There were three spontaneous abortions among the wives of
men with moderate or high exposure to tetrachloroethylene (out of 45 pregnancies). The odds
ratio was 0.7 (95% CI: 0.2-2.4) in a linear logistic regression model adjusting only for age. This
large occupational survey was meant to evaluate reproductive risks associated with paternal
exposures to a wide array of substances and mixtures believed to be mutagens. While the focus
was on exposure to mutagens as a whole, specific exposures also were analyzed, and a broad
2-year time period was used to identify pregnancies related to occupation listed in the 1975 or
1980 censuses. The nonspecific exposure window and use of a crude exposure assignment
method based on occupational title in a census, along with the small number of cases, limit the
ability to draw conclusions concerning paternal tetrachloroethylene exposure and risk of
spontaneous abortion.
A case-control study in Finland evaluated the association of medically diagnosed
spontaneous abortions and maternal occupational exposure to specific solvents (Lindbohm et al.,
1990). The sampling frame was a database of women biologically monitored at the Institute of
Occupational Health in 1965-1983 for one or more of six solvents (styrene, toluene, xylene,
tetrachloroethylene, trichloroethylene, and 1,1,1-trichloroethane). Pregnancies and their
outcomes between 1973 and 1983 among the women were identified using a national register of
4-320
-------
pregnancies in Finland and the Finnish Register of Congenital Malformations. Cases were
women with a spontaneous abortion recorded in the database. One to three controls per case
were selected from among women with a live birth (congenital malformations were not included)
matched for age (± 2.5 years). Among the 456 women, overall response to a mailed
questionnaire was 85% for both cases and controls. A lower proportion of cases (78%) than
controls (99%) confirmed the pregnancy selected from the register. The final data set contained
73 cases and 167 controls with complete information about their occupational history and solvent
exposures during the first trimester of pregnancy.
Likelihood and level of exposure to specific solvents was determined by two
investigators blind to pregnancy outcome using responses to the questionnaires and biological
measurements when available. Women were defined as not exposed if work tasks did not
include handling solvents, the worker did not report exposure, and no biological measurements
were available. Women were defined as potentially exposed if work tasks might have involved
solvent use, but exposures were not reported by the worker, and no biological measurements
were available. Women were defined as exposed if biological measurements were taken while at
the same job, reported tasks implied solvent exposure, or solvent exposure was reported. The
level of exposure was categorized into none, low, or high. High exposure involved handling
solvents daily or 1-4 days per week and high-recorded concentrations for biological or available
industrial hygiene measurements. Low exposure involved handling solvents 1-4 days per week
with low biological concentrations, or solvents were handled <1 day per week. All other
exposure scenarios were defined as none. Biological measurements during the first trimester
were available for only 5% of the population, and, therefore, exposure assignments were based
primarily on reports of work tasks and reported solvent use. Exposure classifications were
checked by an experienced industrial hygienist.
Among the exposed women, there were 8 cases and 15 controls with exposure to
tetrachloroethylene. An odds ratio of 1.4 (95% CI: 0.5-4.2) was observed using conditional
logistic regression with adjustment for previous spontaneous abortions, parity, smoking, use of
alcohol, and exposure to other solvents. The adjusted odds ratios for low and high exposure
were 0.5 (95% CI: 0.1-2.9) and 2.5 (95% CI: 0.6-10.5), respectively. Among four cases and
five controls who reported tetrachloroethylene exposure and whose work tasks involved dry
cleaning, the odds ratio for spontaneous abortion, controlling for exposure to other solvents, was
2.7 (95% CI: 0.7-11.2). The odds ratio for women who reported tetrachloroethylene exposure
but who conducted other work in dry cleaners (1 case and 6 controls) was 0.6 (95% CI: 0.1-5.5).
Blood tetrachloroethylene measurements taken closest to the pregnancy were available for six
women who worked in dry cleaning and seven women in other occupations. The mean
concentration was higher among dry cleaners (2.11 umol/L versus 0.43 umol/L). The authors
4-321
-------
reported that the proportion of study subjects who did not report exposure to a specific solvent in
contrast to a biological measurement that indicated that they were exposed was 18% among
cases and 20% among controls, suggesting that recall was not different by exposure. The study,
which is limited by small sample size and a low prevalence of exposure to tetrachloroethylene,
suggests that exposure during the first trimester may increase risk of spontaneous abortion.
Moreover, odds ratios increased in size when the analysis was restricted to more homogenous
exposure groups representing high exposures.
A case-control study in Santa Clara County, California, also focused on occupational
exposure to solvents, including tetrachloroethylene (Windham et al., 1991). Selection of cases
was hospital based; spontaneous abortions, defined in this study as <20 weeks gestation, among
women 18 years of age or older that occurred between June 1986 to February 1987 were
identified through records of pathology specimens submitted to the 11 hospital laboratories
located in the county. Investigators reviewed medical charts to differentiate spontaneous
abortions from induced abortions. Controls, two per case, were randomly selected from women
with live births, frequency matched by last menstrual period and hospital. A total of 697 of
772 eligible cases (90.3%) and 1,359 of 1,485 controls (91.5%) participated. The analysis was
limited to 1,361 women who were employed during their pregnancy. A higher proportion of
cases was over 35 years of age, reported a prior fetal loss, and consumed more alcohol per week.
Information on exposure during the first 20 weeks of pregnancy or for the duration of the
pregnancy for cases was obtained through a computer-assisted telephone interview. The women
provided detailed information about industry and occupation, job tasks and use of 10 solvents,
plus reported exposure to any other solvents or degreasers. Among the women who reported that
they used tetrachloroethylene during the first weeks of pregnancy, 5 were cases, and 2 were
controls (ORc: 4.7, 95% CI: 1.1-21.1, calculated using Haldane's method for small samples).
Unexposed participants reported no use of any named solvents and did not work in the
microelectronics industry (n = 847). Four of the women exposed to tetrachloroethylene also
reported use of trichloroethylene. The unadjusted odds ratio for use of tetrachloroethylene
and/or trichloroethylene was 3.4 (95% CI: 1.0-12.0). Odds ratios also were calculated in
stratified analyses using Mantel-Haenszel estimation for each of six dichotomous variables
individually (age, race, education, prior fetal loss, smoking, and hours worked). This limited
evaluation of potential confounding does not appear to have resulted in a large decrease of the
summary odds ratios compared to the crude odds ratio, although the adjusted odds ratios were
presented only as a range (e.g., 4.2 [95% CI: 0.86-20.2] controlling for hours worked to 6.0
[95% CI: 1.4-25.8] controlling for age). Estimated risk increased with a higher level or intensity
of exposure when the analyses were stratified by whether exposed participants reported
symptoms, skin contact, or odor versus none (6.3,/>-value for Fisher exact test (1-tail) = 0.04
4-322
-------
compared to 2.1, p-value = 0.54). Despite the small numbers with tetrachloroethylene exposure,
the results suggest an elevated risk for spontaneous abortion. However, several of the exposed
women also were exposed to other solvents, including trichloroethylene, and a detailed
evaluation of potential confounding was precluded by small numbers.
One of the first studies to evaluate adverse reproductive outcomes, including spontaneous
abortions, stillbirths, birth defects, and low birth weight, among female dry cleaners evaluated 53
of 66 small establishments (40 dry cleaning and ironing and 13 ironing only) in two
neighborhoods in Rome, Italy (Bosco et al., 1987). The study population included all of the 67
women who worked in the participating shops. The women averaged 43 years of age and had
been employed an average of 20 years. Information on the work setting and operations and
reproductive histories were collected through a standardized interview. Participants reported if
they had worked in dry cleaning, as a housewife, or other job prior to and during their
pregnancies. In addition, a 24-hour urine sample was collected on a Friday at the end of the
workweek from 53 of the women. Trichloroacetic acid concentrations were higher among
40 dry cleaners (5.01 ug/L) compared to 13 ironers (1.35 ug/L) and 5 controls (1.56 ug/L). Of
56 pregnancies reported during employment as a dry cleaner, 5 ended in a spontaneous abortion
(8.9%). One spontaneous abortion was reported among the 46 pregnancies that occurred while
the women were working at home. The fourfold higher incidence of spontaneous abortion
suggests a tetrachloroethylene-related risk among the dry cleaners. However, individual
characteristics and behaviors that may pose a risk of spontaneous abortion were not presented by
exposure status during pregnancy, and potential confounding was not assessed in this very small
study.
A study that used a common protocol to evaluate reproductive outcomes among dry
cleaners in Denmark, Finland, Norway, and Sweden employed a more precise definition of
tetrachloroethylene exposure (Olsen et al., 1990). All women who had worked at identified
laundries and dry-cleaning plants for at least 1 month during 1973-1983 were included, and a
nested case-referent study was conducted in each country. Identification numbers were linked to
national birth registers and hospital discharge registers to obtain information on births and
outcomes, including spontaneous abortions, in the cohort. In Denmark, all women in the cohort
and every pregnancy that occurred during the study period were included. In Sweden and
Finland, two and three controls per case, matched on maternal age (±2 years), year of pregnancy,
and parity (for Denmark and Sweden), were selected from women with a healthy newborn. In
Norway, information on spontaneous abortions was not available. Women were identified
through company records of active dry-cleaning plants (Sweden and Denmark) and laundries
(Sweden). Approximately 62 and 74% of dry-cleaning plants in Sweden and Norway
participated, respectively. The final study sample consisted of 31 spontaneous abortions and 53
4-323
-------
referents in Sweden (84% response) and 10 spontaneous abortions and 119 referents in Denmark
(77.3% response). In Finland, laundry and dry-cleaning workers on the rolls of the Union of
Chemical Workers and the Municipal Workers of Finland and or included in payroll data from
employers for 1973-1983 were identified and linked with the nationwide hospital discharge
register and polyclinic data for information on pregnancies. One pregnancy for each woman was
randomly selected for analysis. The final data set included 118 spontaneous abortions and
264 referents (77.2% response). Information on exposure to tetrachloroethylene was obtained
from the interviews and questionnaires and was classified by an industrial hygienist blinded to
pregnancy status (Sweden and Denmark). The Finnish investigators had more detailed
information and used reported work history and exposure frequency to classify exposure status.
Exposure was categorized into three groups: unexposed (no dry cleaning), low (worked in dry
cleaning but not high exposures), and high (workers who conducted dry cleaning or spot removal
for at least 1 hour per day). Risk of spontaneous abortion in relation to exposure during the first
trimester was analyzed using conditional logistic regression for matched Swedish and Finnish
data, and unconditional logistic regression for the Danish data set. Models were adjusted for
parity, smoking, and alcohol consumption (Sweden and Finland only).
Odds ratios greater than 1 were observed for the high exposure group in Denmark (OR:
2.52, 95% CI: 0.26-24.1) and Finland (OR: 4.53, 95% CI: 1.11-18.5). The high exposure group
contained small numbers of cases and controls with one case each in Sweden and Denmark, and
six cases in Finland. The odds ratios were combined using the inverse variance of the odds ratio.
The odds ratios for low and high exposure (95% CI) were 1.17 (0.74-1.85) and 2.88
(0.98-8.44), respectively. The authors stated that similar results were obtained when exposure
information provided by the employers (55% of sample) was used instead of responses from the
participants.
A separate report of the Finnish study population was published, evaluating 130 cases of
spontaneous abortions and 289 controls matched for maternal age (Kvyronen et al., 1989).
Slightly different categorizations were used to define exposure. High exposure included women
whose tasks included dry cleaning at least 1 hour daily, and who handled tetrachloroethylene at
least once a week (n = 15). Low exposure included women whose work tasks involved pressing
at a dry cleaners or spot removing, or who handled tetrachloroethylene less than once a week
(n = 31). Blood tetrachloroethylene measurements, taken within 10 months of the first trimester
of pregnancy, were available for seven of the participants (except for one more distant
measurement). These data corresponded well to their reported exposure. Exposure to other
solvents, including petroleum, benzene, toluene, acetone, thinner, and spot remover mixtures,
was reported by six cases (5.9% of women who worked during their pregnancy) and six controls
(2.9%). The odds ratio for high exposure to tetrachloroethylene was 3.4 (95% CI: 1.0-11.2,
4-324
-------
p < 0.05) in a multivariate model adjusted for frequent use of solvents other than
tetrachloroethylene (OR: 1.5, 95% CI: 0.4-5.4), frequent heavy lifting at work (OR: 1.9;
95% CI: 1.0-2.8), and frequent use of alcohol (OR: 2.0, 95% CI: 1.0-4.0). Selection bias did
not appear to be a major factor; when exposure information obtained from employers was used
to classify eight cases and six controls instead of self-reports, the proportion returning the
questionnaire was similar (0.25 and 0.17, respectively).
Ahlborg et al. (1990a) published the Swedish results separately along with a
complementary study designed to be more representative of the entire dry-cleaning and laundry
sector. Laundry and dry-cleaning establishments, identified from the Swedish Post Address
Register in 1984, were mailed a questionnaire to obtain names and contact information for all
women employees who had worked for at least 1 month during 1974 and 1983. Cases of
spontaneous abortion (defined in this study as fetal death at <28 weeks gestation), perinatal
death, congenital malformation, or low birth weight (<1,500 g) were identified among deliveries
during 1974-1983 recorded in the Medical Birth Registry, the Swedish Registry of Congenital
Malformations, and the Inpatient Registry for Somatic Care (spontaneous abortion treated in a
hospital). Dates of delivery or spontaneous abortion were used to identify women who had been
working while they were pregnant (at least 1 week of the year before delivery or 6 months before
a spontaneous abortion). A total of 67 cases were identified among 955 pregnancies, and two
referents per case were selected, matched on mother's age, year of pregnancy, and parity (only
for deliveries). Responses were received from 48 cases (75%) and 110 referents (88%).
Recruitment for the complementary study involved the identification of women registered as
washers/cleaners via an occupational code in the 1975 and 1980 Censuses. A total of
755 additional pregnancies were identified via linkage with the medical registers for the 2-year
period after each census. Responses to the mailed questionnaire were received from 68 cases
(88%) and 131 referents (87%). Exposure to tetrachloroethylene during the first trimester was
classified independently by two investigators who were unaware of the worker's case/control
status. High exposure included operating a dry-cleaning machine or conducting spot removing
using tetrachloroethylene at least 2 hours per week, or ironing/pressing dry-cleaned cloth for
over 20 hours per week, or cleaning and filling the machines at least three times. Low exposure
included other work in dry-cleaning businesses where tetrachloroethylene was used. Unexposed
workers were employed in companies that did not dry clean using tetrachloroethylene. In the
combined data set, 31 and 19 cases (all outcomes) were classified as having low and high
exposure, respectively. The numbers of spontaneous abortions by exposure category were not
reported. Odds ratios for spontaneous abortion among workers with low and high exposure
using conditional logistic regression were 1.0 (95% CI: 0.4-2.2) and 0.9 (95% CI: 0.4-2.1),
respectively. The models adjusted for smoking, alcohol consumption, medical complications,
4-325
-------
and history of adverse pregnancy outcome. This study did not find an increased risk of
spontaneous abortion among workers reporting tetrachloroethylene exposure during the first
trimester.
A relatively large study in the United Kingdom evaluated the risk of spontaneous
abortions among current and former employees of dry-cleaning and laundry establishments
managed by four companies between 1980 and 1995 (Doyle etal., 1997). Information about
workplace exposure and reproductive history were obtained in 1995-1996 via mailed
questionnaires sent to 7,301 women, aged 16-45 years, who were identified by the employers.
Of the 5,712 questionnaires successfully delivered, 54.5% were completed (n = 3,110). The
responses by current dry-cleaners and laundry workers were 78 and 65%, respectively, but were
lower among former workers (46.1 and 39.7%, respectively). The authors reported that the age
distribution of responders was comparable to that of nonresponders. The final data set included
3,092 respondents with complete information about 3,517 total pregnancies. Pregnancies were
included in the analysis if the women reported that it had been confirmed by a doctor, hospital
treatment was required, or it ended in a live birth. The rate of spontaneous abortions was
evaluated in relation to the woman's employment during her pregnancy or the 3 months prior.
Work at a dry cleaner and as a dry-cleaning machine operator was used as an exposure surrogate
for tetrachloroethylene. This was compared to work at a laundry or no employment at a dry
cleaners or laundry during the pregnancy or 3 months prior.
Spontaneous abortions were compared to total pregnancies (spontaneous abortions,
stillbirths, and live births) excluding ectopic and molar pregnancies and induced abortions. For
the 325 reported spontaneous abortions between 1980-1995, the odds ratio for dry cleaning
compared to laundry work was 0.97 (0.55-1.69). However, among 93 spontaneous abortions to
women employed in dry cleaning, machine operators had a 63% higher risk of spontaneous
abortion compared to nonoperators (OR: 1.63, 95% CI: 1.01-2.66). The unconditional logistic
regression models controlled for maternal age, pregnancy order, and year of birth. A similar
pattern of risk was observed when the analyses were restricted to the women's first or last
pregnancies. These latter analyses were meant as a check to address the lack of independence of
multiple pregnancies reported by the same woman. For example, among dry-cleaning machine
operators, when the last exposed pregnancy was compared to pregnancies that occurred later
during periods with no exposure to tetrachloroethylene, risk of spontaneous abortion was 82%
higher (OR: 1.82, 95% CI: 1.09-3.05). An elevated risk also was observed when pregnancies
during work as a dry-cleaning machine operator were compared to unexposed pregnancies before
the first exposed pregnancy. Laundry workers also experienced more spontaneous abortions
when employed in laundries compared to periods when they had other employment or were not
employed; however, the CIs included one. The investigators were not able to compare risks
4-326
-------
between dry cleaning generally and laundry work because the number of spontaneous abortions
reported for pregnancies while working in a laundry was low (n = 19). Doyle et al. (1997) found
an elevated risk of spontaneous abortion for work as a dry-cleaning machine operator during or 3
months before a pregnancy compared to work in other dry-cleaning jobs or work in other
industries or in the home during this sensitive period.
The rate of self-reported spontaneous abortions was comparable among the wives of dry
cleaners (n = 14) and laundry workers (n = 26) in a cohort of primarily unionized men in
northern and southern California who participated in a study of semen quality (Eskenazi et al.,
199la). Rates of spontaneous abortion during the time periods when their husbands worked in
the industry were 11.1 and 15.2% among the wives of dry-cleaners and laundry workers,
respectively (X2 = 0.32,/? = 0.57). Although the authors presented the rates as spontaneous
abortion rates, it does not appear that the fetal deaths reported were limited to <28 weeks of
gestation. The rate was calculated as the total number of miscarriages during the husband's
employment in the industry divided by the total number of pregnancies during the same time
period, multiplied by 100. It was not stated how many years the women, whose average age was
mid-thirties, had to recall previous miscarriages.
A population-based study in Woburn, Massachusetts, evaluated outcomes during
pregnancy and effects in children among residents whose drinking water source was two wells
contaminated with chlorinated organic substances from 1960 to 1982 (Lagakos et al., 1986). The
two wells were operated as a single water source. The contamination of the two wells, located in
eastern Woburn, was discovered in May, 1979. Levels of trichloroethylene (267 ppb),
tetrachloroethylene (21 ppb), and chloroform (12 ppb) were detected, and the wells were shut
down. The other six wells that supplied Woburn were located in the southwest part of town, and
testing did not find levels above state and federal standards. Information was collected through a
telephone survey of former and current family members residing in Woburn from 1960-1982
and listed in the 1982 town directory. The survey was conducted by 235 volunteers trained in
interview techniques who successfully contacted 6,219 residences. In the end, 5,010 completed
interviews were obtained, approximately 57% of the town's residences with listed telephone
numbers. All pregnancies ending between 1960 and 1982 to women born since 1920 were
ascertained, and information was collected on pregnancy outcomes and the health of offspring,
maternal characteristics, and residence history. Regional and temporal distribution of the water
from the two contaminated wells was determined by the Massachusetts Department of
Environmental Quality and Engineering during October 1964 to May 1979. The town was
partitioned into five zones of graduated exposure to water from the wells. The study
investigators estimated the proportion of each household's annual water supply that came from
the two wells. Each pregnancy was assigned an annual exposure score using the mother's
4-327
-------
residence during the year the pregnancy ended. An exposure history was constructed for each
child consisting of the sum of annual scores accumulated during their residence in Woburn.
Of the 4,396 pregnancies that occurred during 1960 to 1982, 16% were exposed during
the year the pregnancy ended. There were 520 spontaneous abortions (12%), defined in this
study as a fetal loss in the first 6 months, and 67 perinatal deaths (1.5%), defined as a stillbirth or
a live birth that survived fewer than 7 days. Logistic regression analyses, controlling for other
risk factors, found no statistically significant associations between the annual exposure score for
the year a pregnancy ended and spontaneous abortion, or perinatal deaths before 1970. An odds
ratio of 10 (p = 0.003) was observed for perinatal deaths after 1970, when changes in industrial
water demand occurred, and a different set of five zones representing exposure to water from the
contaminated wells was constructed. This was due to 3 perinatal deaths that occurred in
households with the highest exposure score category of 0.51-1.0.
A population-based retrospective study of tetrachloroethylene in drinking water evaluated
effects on pregnancy and development from exposure resulting from leaching of
tetrachloroethylene from vinyl linings in water distribution pipes installed between 1968 and
1980 in the Cape Cod region in Massachusetts (Aschengrau et al., 2009a: Aschengrau et al.,
2009b: Aschengrau et al., 2008). Because the pipes were used to replace existing pipes or to
extend the distribution system to serve a growing population, population exposure was
irregularly distributed, and a wide range of tetrachloroethylene concentrations were detected in
samples collected in 1980. In addition, only one town used a chlorinated surface water supply,
resulting in a low probability that drinking water was contaminated with chlorinated byproducts.
Water concentrations ranged from 1.5 to 80 ug/L along main streets, and from 1,600 to 7,750
ug/L along dead end streets where water flow was low. All births between 1969 and 1983 were
identified from birth certificates, and women residing in one of eight Cape Cod towns with vinyl-
lined water distribution pipes at the time of the index birth were eligible for the study. A total of
1,492 women with addresses along streets where the pipes had been installed or with connections
to such pipes were initially defined as exposed. A comparison group of 1,704 births, frequency
matched to the exposed group by month and year of birth, was selected. Follow-up of the
selected individuals occurred during 2002-2003. The final data set contained 959 women with
potential exposure and 1,087 potentially unexposed women who returned a self-administered
questionnaire, comprising 64% of the selected sample and 69% of those who were located.
Response did not vary by potential exposure status. The study population was primarily
Caucasian, with an average age of 27 years, and most had adequate prenatal care (72-73%). The
annual mass of tetrachloroethylene delivered to each address before and during pregnancy was
estimated using self-reported residential histories mapped using GIS (94% of reported
pregnancies), a leaching and transport model developed for the study, and EPA's EPANET
4-328
-------
modeling software estimating water flow and direction. Estimated water concentrations of
tetrachloroethylene ranged between 1 and 5,197 ug/L.
Self-reported clinically recognized pregnancy loss (659 spontaneous abortions and
stillbirths) and 4,908 live births up to December 1990 were eligible for analysis. Pregnancy
outcomes were analyzed in relation to three measures of exposure: cumulative exposure up to the
month and year of the last menstrual period (prepregnancy window), peak exposure up to the last
menstrual period year of the pregnancy (prepregnancy window), and average monthly exposure
during the year containing the last menstrual period (time of conception). Risk of pregnancy loss
associated with exposure measures, divided into quartiles, was evaluated using generalized
estimating equations to account for lack of independence of multiple pregnancies by the same
woman. Risk estimates for pregnancy loss by increasing quartiles of exposure were similar
across the three exposure measures. For example, the multivariate GEE odds ratios for average
monthly exposure in increasing quartiles during the year of the last menstrual period were 1.1
(95% CI: 0.8-1.6), 0.7 (95% CI: 0.5-1.1), 0.8 (95% CI: 0.6-1.2), and 0.7 (95% CI: 0.5-1.0),
respectively. Several covariates were evaluated for potential confounding, including risk factors
for pregnancy loss, those associated with tetrachloroethylene exposure, and nondrinking water
sources of solvent exposure. Maternal age, year of pregnancy, paternal age, maternal history of
gynecologic infections, and the number of prior live births were included in the final models.
The authors checked the validity of self-reported birth outcomes by comparing the reproductive
histories reported by the women for all of the index pregnancies with information from birth
certificates. Further, information from medical records about pregnancies reported by 60 women
also was compared to self-reported histories. The authors reported good-to-excellent agreement
including for gestational duration and birth weight, prenatal cigarette smoking, number of prior
live births, and spontaneous and induced abortions. The study evaluated a large number of
pregnancy losses using a detailed exposure model and carefully assessed potential confounding.
It is important to note, however, that exposure estimates were not based on household
measurements, and individual consumption was not known. Therefore, exposure
misclassification may not have allowed detection of a small increase in risk. Finally, use of
exposure prior to the last menstrual period or during that year may not have had the required
precision to identify a risk associated with a particular susceptible window for pregnancy loss
(e.g., the first trimester).
In summary, the literature contains few studies of effects on spermatogenesis or
menstruation among subjects with exposure to tetrachloroethylene. One study of primarily
unionized workers in the dry-cleaning and laundry industries in California observed subtle
deficits in sperm quality in relation to tetrachloroethylene in exhaled breath, an exposure index,
and occupational group (dry-cleaning or laundry worker) (Eskenazi et al., 199la). However,
4-329
-------
three clinically recognized measures of sperm quality were not associated with exposure in the
study population. The results of Eskenazi et al. (1991a) are compelling, but more studies are
needed to understand the spectrum of effects on sperm and their impact on fecundity. Two other
studies that evaluated effects on sperm, hormonal disturbances, or menstruation among men and
women with occupational exposure were not adequate to draw conclusions concerning the
association (Zielhuis et al., 1989; Rachootin and Olsen, 1983). Some studies that relied on
detailed work histories and monitoring data to classify exposure suggested that maternal or
paternal exposure to tetrachloroethylene or work in dry cleaning reduces fertility or delays
conception (Sallmen et al., 1998; Sallmen et al., 1995; Eskenazi etal., 1991b). However, the risk
estimates were imprecise because the number of participants reporting exposure to
tetrachloroethylene was small. As a consequence, the existing literature is inconclusive
concerning effects of tetrachloroethylene on reproduction and fertility.
A number of studies have evaluated the risk of spontaneous abortions in relation to
maternal and paternal occupational exposure to tetrachloroethylene. Results of several studies of
maternal occupational exposure to tetrachloroethylene suggest an increased risk of spontaneous
abortion, particularly at higher levels (Doyle et al., 1997; Windham et al., 1991; Lindbohm et al.,
1990; Olsen etal., 1990; Kyyronen et al., 1989). Most of the studies evaluated exposure during
the first trimester of pregnancy. Some of the studies observed an increased odds ratio ranging
between 1.4 to 4.7, but had low statistical power because the cohort contained small numbers of
exposed cases and controls, and were limited in their ability to evaluate potential confounding
(Windham et al.. 1991: Lindbohm et al.. 1990: Olsen et al.. 1990: Bosco et al.. 1987). In general,
the studies that used a more precise definition of exposure, or categorized exposure into levels of
increasing dose or intensity, observed higher risk estimates (Doyle etal., 1997; Windham et al.,
1991; Lindbohm et al., 1990; Olsen et al., 1990; Kyyronen et al., 1989). Increased risks were not
found among dry cleaners in Sweden (Ahlborg, 1990a: Olsen etal., 1990). Three studies of
paternal occupational exposure prior to the beginning of the pregnancy did not observe an
association (Eskenazi et al., 1991a: Lindbohm et al., 1991; Taskinen et al., 1989). Two of these
surveyed occupational exposure to a broad array of substances and, consequently, had low
statistical power for chemical-specific analyses (Lindbohm et al., 1991; Taskinen et al., 1989).
Although there is no evidence of an increased risk associated with paternal exposure, the studies
were not of sufficient size or detail in exposure estimates to draw conclusions. No associations
with incidence of spontaneous abortion were observed among two populations exposed to
tetrachlorethylene in drinking water (Aschengrau et al., 2009a: Aschengrau et al., 2008; Lagakos
etal., 1986). The populations were likely exposed to lower levels compared to the occupational
populations. In addition, the window of exposure used to assess risk in both studies may not
have had been precise enough to detect a small elevation in risk for spontaneous abortion.
4-330
-------
4.7.2.2. Animal Reproductive Toxicity Studies
Evaluation of the reproductive effects of tetrachloroethylene exposure in mammalian
animal models is based on a two-generation reproduction studies in rats, an in vivo sperm assay,
and an in vitro oocyte fertilization assay following in vivo exposure of adult female rats. These
studies are described below.
4.7.2.2.1. In vitro fertilization assay
In a study designed to examine the fertilizability of rat oocytes, female rats were exposed
to inhaled tetrachloroethylene at 12,000 mg/m3 (2 hours/day, 5 days/week) for 2 weeks (Berger
and Horner, 2003). The percentage of extracted oocytes that were fertilized in vitro was reduced
for tetrachloroethylene-treated females as compared with controls.
4.7.2.2.2. In vivo reproductive toxicity studies
Beliles et al. (1980) described an experiment in which male rats and mice (12/group)
were exposed via inhalation to tetrachloroethylene concentrations of 100 and 500 ppm, for
7 hours/day, for 5 days. Sperm head abnormalities and abnormal sperm were evaluated at 1, 4,
and 10 weeks after the last dose. Rats were unaffected. In mice, at 4 weeks, but not at 1 or 10
weeks after exposure, there was a significant increase (p < 0.05) in the percentage of males with
abnormal sperm heads (19.7%) in the 500 ppm exposure group. For the 100 ppm and control
groups, the percentages were 10.3 and 6% (not statistically significant at thep < 0.05 level),
respectively. A positive control group administered triethylene melanime was adversely affected
(11.1%). The authors suggested that the temporal appearance of the abnormal sperm heads
indicated that the spermatocyte and/or spermatogonia were the stages most sensitive to the
effects of inhaled tetrachloroethylene. In this study, the NOAEL was 100 ppm, and the LOAEL
was 500 ppm.
A multigeneration study of the effects on rats of exposure to airborne concentrations of
tetrachloroethylene was performed by Tinston (1994). It was conducted under GLP standards
and received frequent quality assurance audits. In this study, weanling male and female
(Alpk: APfSD) rats (FO) (24/sex/group) were exposed to airborne tetrachloroethylene
concentrations of 0, 100, 300, or 1,000 ppm, 6 hours/day, 5 days/week, for 11 weeks prior to
mating and then for 6 hours/day during mating and through GD 20. There were no exposures
from GD 21 through Day 5 postpartum. One litter was produced in the first generation (Fl A).
The first-generation dams and their litters were exposed to tetrachloroethylene from PND 6
through 29, at which time, parental animals for the second generation were selected. The
second-generation parents (Fl) were then exposed 5 days/week during the 11-week premating
period. In the second generation, three litters were produced: F2A, F2B, and F2C. The F2A
4-331
-------
dams and litters were exposed from Days 6 to 29 (control and 100 ppm) or Days 7 to 29
(300 ppm). The 1,000 ppm exposure for the Fl dams stopped after the F2A littering.
F2B litters were generated by mating the Fl parental males and females in the control,
300, and 1,000 ppm groups; the dams and F2B litters were not exposed to tetrachloroethylene
during lactation. An F2C litter was produced by mating Fl males exposed to 1,000 ppm with
unexposed females. These females and the F2C litters were killed on PND 5 and discarded
without further examination. Overall, the FO males were exposed for 19 weeks, and the Fl
males were exposed up to 35 weeks. Postmortem evaluation in adults and selected weanlings
included organ weight and histopathology examination of liver, kidney, and reproductive organs;
sperm measures were not assessed.
Table 4-35 summarizes the results of the Tinston study. Signs of CNS depression
(decreased activity and reduced response to sound) were observed at 1,000 ppm for the first 2
weeks in both adult generations and again when the exposure was resumed on Day 6 postpartum
in the Fl generation (adults and pups). Other signs of overt tetrachloroethylene toxicity in the
adults included irregular breathing and piloerection at both 1,000 and 300 ppm and salivation
and tip-toe gait (in one Fl female) at 1,000 ppm. These changes stopped with the cessation of
exposure or within approximately 30 minutes thereafter.
There were a number of changes relative to controls that were of minor biological
significance. One change, transient statistically significant reductions of mean body weights
(originating from treated males and nontreated females), suggests the absence of male-mediated
effects on reproductive outcome. Nevertheless, the alterations in testes weight cannot be
discounted as a possible effect of treatment.
In females, dystocia was noted in one FO dam at 100 ppm, two Fl dams at 300 ppm, and
a total of four dams (two each FO and Fl) at 1,000 ppm; these dams were terminated without
completion of delivery. From the data for surviving dams and litters, it can be assumed that the
difficulties in parturition were not associated with or attributable to alterations in mean gestation
length or increased mean pup or litter weights. In fact, mean pup body weights showed a
statistically significant decrease throughout the lactation period at 300 and 1,000 ppm for Fl A
litters and in early lactation for F2A and F2B litters. Additionally, mean Fl A male pup body
weight was significantly decreased (5% less than controls; p < 0.05) at 100 ppm on PND 29.
These PND 29 mean body-weight deficits in all treated groups were observed in the animals
selected as parents of the second generation, but by the second week of the Fl premating period,
mean body weights were similar to those of controls for both 100 and 300 ppm-animals.
4-332
-------
Table 4-35. 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
(percentage)
Decreased mean
percentage pup survival
Days 1-5
Decreased mean
percentage 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
a Not exposed after delivery.
V<0.05.
°/7<0.01.
d trend;? < 0.05.
NA = Not applicable (pups terminated on Day 5 postnatal).
Source: Adapted from Tinston (1994).
4-333
-------
Mean litter size was decreased at 1,000 ppm for F2A and F2B litters. Statistically
significant decreases in the number of live pups on PND 1 (25 and 37% lower than controls for
F2A and F2B, respectively) are suggestive of either an adverse effect on fertilization or on in
utero survival. Early postnatal survival (i.e., on PND 1 and between PNDs 1 and 5) was also
compromised in F2A and F2B pups at 1,000 ppm, with mean litter sizes decreasing to 48% and
53% of those of controls, respectively. The number of dead pups and litters with dead pups was
also increased, although not significantly, at 300 ppm for F2A litters. Clinical observations data
for 1,000 ppm litters reported an increased incidence of F2A and F2B pups that were found dead,
were killed in extremis, or were missing and presumed dead. The apparent increase in adverse
survival findings at 300 and 1,000 ppm in the second generation as compared with the first
generation could not be definitively attributed to any particular aspect of study design or conduct
(e.g., differences in the duration of treatment), although it is noted that, unlike the second
generation (Fl) parental animals, the first generation (FO) rats were not exposed to
tetrachloroethylene during preconception and in utero development.
A deficiency of the Tinston study is that the pregnant rats were not exposed from
gestation Day 21 through lactation Day 6 or 7, and the exposure at the 1,000 ppm treatment level
stopped for the Fl dams at the littering of the F2B pups. The F2B pups were not exposed
postnatally. It is additionally noted that this study was conducted according to the pre-1998 EPA
harmonized two-generation reproduction study guideline and, thus, did not assess a number of
sensitive endpoints such as estrous cyclicity, sperm measures, age of sexual maturation, and
enhanced reproductive organ pathology.
A summary of the doses at which treatment-related effects were observed in the Tinston
(1994) study is presented in Table 4-35. Overall, the parental systemic toxicity was observed at
300 and 1,000 ppm, with a NOAEL of 100 ppm. For offspring, the LOAEL of 100 ppm was
based upon decreased body weight in Fl A pups at PND 21; no NOAEL was established. There
was no evidence of treatment-related effects on reproductive function at any exposure level
tested.
4.7.2.3. Reproductive Cancers in Humans
Thirteen epidemiologic studies reporting data on breast cancer and tetrachloroethylene
exposure and 12 epidemiologic studies reporting data on cervical cancer and tetrachloroethylene
exposure were identified. This set of studies includes 10 cohort studies on breast and cervical
cancers (Calvert etal.. 2011: Selden and Ahlborg, 2011: Pukkala et al.. 2009: Sung et al.. 2007:
Chang et al.. 2005: Blair et al.. 2003: Ruder etal.. 2001: Andersen et al.. 1999: Boice et al..
1999: Lynge and Thygesen, 1990), one study reporting on breast cancer but not cervical cancer
(Radican et al., 2008), two studies reporting on cervical cancer but not breast cancer (Travier et
4-334
-------
al., 2002; Anttila et al., 1995), two breast cancer case-control studies of occupational exposures
(Peplonska et al., 2007; Band et al., 2000), one cervical cancer nested case-control study of
occupational exposure (Lynge et al., 2006), and one breast cancer case-control study of
residential exposure through contaminated drinking water (Aschengrau et al., 2003).
Aschengrau et al. (2003) extended Aschengrau et al. (1998), adding additional breast cancer
cases from 1987-1993, and presenting odds ratios for the combined 10-year study period,
1983-1993. Most breast cancer studies examined females (Radican et al., 2008; Peplonska et
al., 2007: Sung et al., 2007: Aschengrau etal., 2003: Blair et al., 2003: Band et al., 2000) or
males and females combined (Calvert et al., 2011: Boice etal., 1999). Five studies, mostly of
Nordic subjects, presented risk estimates for male subjects separately (Selden and Ahlborg,
2011: Pukkala et al., 2009: Chang et al., 2005: Andersen et al., 1999: Lynge and Thygesen,
1990). These studies represent the core studies evaluated by EPA, as described in more detail
below. Appendix B reviews the design, exposure-assessment approach, and statistical
methodology for each study. Most studies were of the inhalation route of exposure, of
occupational exposure, and lacked quantitative exposure information. Nine studies reporting risk
estimates for breast or cervical cancer examine occupational titles such as dry cleaner, launderer,
and presser as surrogates for tetrachloroethylene, given its widespread use from 1960 onward in
the United States and Europe (Calvert et al., 2011: Selden and Ahlborg, 2011: Pukkala et al.,
2009: Peplonska et al., 2007: Lynge et al., 2006: Blair et al., 2003: Ruder etal., 2001: Band et
al., 2000: Andersen et al., 1999: Lynge and Thygesen, 1990). Five studies conducted in Nordic
countries are either based on either the entire Swedish population or on combined populations of
several Nordic countries; strengths of these studies are their use of job title as recorded in census
databases and ascertainment of cancer incidence using national cancer registries (Selden and
Ahlborg, 2011: Pukkala et al., 2009: Lynge et al., 2006: Andersen et al., 1999: Lynge and
Thygesen, 1990). Subjects in the multi -Nordic country study of Pukkala et al. (2009) overlapped
those of Lynge and Thygesen (1990), Andersen et al. (1999), Lynge et al. (2006), and Selden and
Ahlborg (2011). Studies examining mortality among U.S. dry-cleaner and laundry workers
(Blair etal., 2003: Ruder etal., 2001) are of smaller cohorts than the Nordic studies, with fewer
observed lung cancer events.
The exposure surrogate in studies of dry-cleaners and laundry workers is a broad
category containing jobs of differing potential for tetrachloroethylene exposure. Thus, these
studies have a greater potential for exposure misclassification bias compared to studies with
exposure potential to tetrachloroethylene assigned by exposure matrix approaches applied to
individual subjects. Calvert et al. (2011) studied unionized dry cleaners in the United States in
California, Illinois, Michigan, and New York who worked for one or more years before 1960 in
one or more shops known to use tetrachloroethylene as the primary solvent (Calvert et al., 2011:
4-335
-------
Ruder etal., 2001, 1994). The cohort was stratified into two groups based on the level of
certainty that the worker was employed only in facilities using tetrachloroethylene as the primary
solvent; tetrachloroethylene-only and tetrachloroethylene plus. Lynge et al. (2006), using job
titles reported in the 1970 Census, identified subjects as dry cleaners (defined as dry cleaners and
supporting staff if employed in a business of <10 workers), other job titles in dry cleaning
(launderers and pressers), unexposed (job title reported on 1970 Census was other than in dry
cleaning), or unclassifiable (information was lacking to identify job title of subject). Selden and
Ahlborg (2011) identified subjects as either dry cleaners or laundry workers and presented risk
estimates separately by job title.
Four other cohorts with potential tetrachloroethylene exposure in industrial settings have
been examined. These studies include aerospace or aircraft maintenance workers in the United
States (Radican et al., 2008; Boice etal., 1999), workers, in Finland, primarily in the metal
industry (Anttila et al., 1995) and electronic factory workers in Taiwan (Sung et al., 2007; Chang
et al., 2005). Boice et al. (1999) and Radican et al. (2008) used an exposure assessment based on
a job-exposure matrix, and Anttila et al. (1995) used biological monitoring in blood to assign
potential tetrachloroethylene exposure to individual subjects. In contrast and less sensitive, the
exposures in the Taiwan studies included multiple solvents and tetrachloroethylene exposure was
not linked to individual workers. Additionally, cohorts included white-collar workers, who had
an expected lower potential for exposure (Sung et al., 2007; Chang et al., 2005).
Aschengrau et al. (2003) is a case-control study that examined residential proximity to
drinking water sources contaminated with tetrachloroethylene in Cape Cod, MA, and used an
exposure model incorporating leaching and characteristics of the community water distribution
system to assign quantitative estimates of a household relative dose of tetrachloroethylene.
In summary, with respect to exposure-assessment methodologies, four studies with breast
or cervical cancer data assigned tetrachloroethylene exposure to individuals within the study
using a job exposure matrix (Boice etal., 1999; Anttila et al., 1995), an exposure model
(Aschengrau et al., 2003), a classification of the cohort by certainty of tetrachloroethylene
exposure (Calvert et al., 2011), or restricting analyses to subjects identified as dry cleaners
(Selden and Ahlborg, 2011; Lynge et al., 2006). The relative specificity of these exposure-
assessment approaches strengthens their ability to identify cancer hazards compared to studies
with broader and less sensitive exposure-assessment approaches. The least sensitive exposure
assessments are those using very broad definitions such as working in a plant or a factory (Sung
et al., 2007: Chang et al., 2003).
4-336
-------
Five26 of the nine breast cancer studies evaluated by EPA with exposure assessment to
tetrachloroethylene or employment as dry-cleaner or laundry worker reported estimated relative
risks based on 50 or more observed events (Selden and Ahlborg, 2011; Pukkala et al., 2009;
Aschengrau et al., 2003; Blair etal., 2003; Lynge and Thygesen, 1990): the observed number of
breast cancer cases or deaths ranged from 56 (Blair et al., 2003) to 1,757 (Pukkala et al., 2009).
The largest cohort of breast cancer cases in female dry-cleaners and laundry workers (n = 1,757)
observed a standardized incidence ratio of 0.89 (95% CI: 0.85, 0.94) (Pukkala et al.. 2010).
Three other studies of dry-cleaners and laundry workers with findings based on between 68 and
219 cases or deaths observed a standardized incidence ratio or SMR estimate of 0.88 (95% CI:
0.77, 1.01) (Selden and Ahlborg. 2011). 1.0 (95% CI: 0.8, 1.3) (Blair et al.. 2003). and 1.11 (95%
CI: 0.90, 1.34) (Lynge and Thygesen, 1990) for the association between breast cancer risk and
ever having a job title of dry-cleaner or laundry worker (refer to Table 4-36). A case-control
study with findings based on 50 or more exposed cases observed an odds ratio of 1.2 (95% CI:
0.9, 1.7) for living in a residence receiving contaminated water with a relative delivered dose of
tetrachloroethylene above the median value (median: 2.1, range: 0.001-243.8) compared to
controls (Aschengrau et al., 2003). SMRs or standardized incidence ratios for breast cancer were
similar for subjects identified as dry cleaners compared to laundry workers or for the subcohort
of females whose starting date of employment was after 1960 compared to the larger cohort
(Selden and Ahlborg. 2011).
In addition to the evidence from the large cohort and case-control studies, evidence is
found in five other studies whose effect estimates for breast cancer are based on fewer observed
events and that carry lesser weight in the analysis. As expected, the magnitude of the point
estimate of the association reported in these studies is more variable than in the larger studies:
0.48 (Radican et al.. 2008). 1.1 to 1.5 (Calvert etal., 2011: Peplonska et al.. 2007: Boice et al..
1999). and >2.0 (Band et al.. 2000). Of these five studies, only risk estimates of Band et al.
(2000) excluded 1.0. Chang et al. (2005) and Sung et al. (2008). a follow-up study of the same
population, reported standardized incidence ratios of 1.19 (95% CI: 1.03, 1.36) and 1.09 (95%
CI: 0.96, 1.22). Both studies observed over 200 breast cancer incident cases; however, these
studies carry lesser weight in the analysis, given their low level of detail of the exposure
assessment.
26 Andersen et al. (1999) is not included in this summary of the data from the individual studies because it was
updated and expanded in the analysis by Pukkala et al. (2009).
4-337
-------
Table 4-36. Summary of human studies on tetrachloroethylene exposure and breast cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort Studies
Biologically monitored workers
All subjects
Not reported
Aerospace workers (Lockheed)
Routine exposure to PCE
Routine-Intermittent exposure duration to PCE
1.16(0.32,2.97)
Not reported
4
Electronic factory workers (Taiwan)
All Subjects
Males
Females
Females
0.90 (0.48, 1.53)
1. 19(1.03, 1.36)
1.09 (0.96, 1.22)
0
0.11
exp
215
286
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females and 0.7
umol/L in males (median)], follow-up 1974-1992, external referents (SIR)
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after 1960.
worked at least 1 yr, follow-up 1960-1996, job exposure matrix without
quantitative estimate of PCE intensity, 1987-1988 8-h TWA PCE
concentration (atmospheric monitoring) 3 ppm [mean] and 9.5 ppm
[median], external reference for routine exposure (SMR) and internal
references (workers with no chemical exposures) for routine -intermittent
PCE exposure (RR), male (ICD-9, 175) and female breast cancer (ICD-9,
174)
Chang et al. (2005): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997, multiple solvents
exposure, does not identify PCE exposure to individual subjects, cancer
mortality, external referents (SIR) (Chang et al.. 2005):
63,982 females, follow-up 1979-2001, factory employment proxy for
exposure, multiple solvents exposures and PCE not identified to individual
subjects, cancer incidence, external referents, analyses lagged 15 yr (SIR)
(Sung etal. 2007)
oo
oo
-------
Table 4-36. Summary of human studies on tetrachloroethylene exposure and breast cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
0.48 (0.07, 3.50)
1
Dry -cleaner and laundry workers
All laundry worker and dry cleaners
Males
Females
(0, 3.41)
0.89 (0.83, 0.97)
0
634
All subjects
1.0(0.8, 1.3)
68
Semiquantitative exposure score
Little to no exposure
Medium to high exposure
0.8 (0.6, 1.2)
1.2 (0.8, 1.7)
30
29
Laundry workers and dry cleaners in 1960 Census
Males
Females
Not reported
Not reported
All laundry worker and dry cleaners
Males
Females
1.11(0.90, 1.34)
0
0.2 exp
94
Reference
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever exposed
to mixed solvents, 85 1 ever-exposed to PCE), employed at least 1 yr from
1952 to 1956, follow-up 1973-2000, job exposure matrix (intensity),
internal referent (workers with no chemical exposures [RR]), female breast
cancer (ICD-A8, -9, 174; ICD-10, C50)
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or 1970
Census (Denmark, Finland, Norway), follow-up 1971-1987 or 1991, PCE
not identified to individual subjects, external referents (SIR), ICD-7, 170
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union members
(1945-1978), follow-up 1979-1993, semiquantitative cumulative exposure
surrogate to dry clean solvents, cancer mortality, external referents (SMR),
female breast (ICDA-8, 174).
Ji et al. (2005b)
9,255 Swedish men and 14,974 Swedish women employed in 1960 (men)
or 1970 (women) as laundry worker or dry cleaner, follow-up
1961/1970-2000, PCE not identified to individual subjects, external
referent (SIR) and adjusted for age, period and socioeconomic status.
Lynge and Thygsen (1990)
10,600 Danish men and women, 20-64 yr old, employed in 1970 as
laundry worker, dry cleaners and textile dye workers, follow-up
1970-1980, external referents (SIR), ICD-7, 170.
oo
VO
-------
Table 4-36. Summary of human studies on tetrachloroethylene exposure and breast cancer (continued)
Exposure group
Launderer and dry cleaner
Male
Female
All subjects
Exposure duration/time since 1st employment
PCE-only subjects
Dry-cleaners and laundry workers
Males
Females
PCE
Males
Females
Laundry
Males
Females
Relative risk
(95% CI)
0.86(0.18,2.50)
0.89 (0.85, 0.94)
1.05 (0.70, 1.52)
Not reported
1.06(0.51,1.94)
(0.00, 7.68)
0.88 (0.77, 1.01)
0.85 (0.72, 1.00)
0.96 (0.76, 1.21)
No. obs.
events
o
3
1,757
28
10
0
219
0
140
0
76
Reference
Pukkala et al. (2009)
Men and women participating in national census on or before 1990, 5
Nordic countries (Denmark, Finland, Iceland, Norway, Sweden), 30 -64 yr,
follow-up 2005, occupational title of launderer and dry cleaner in any
census, external referents (SIR), ICD-7, 170.
Calvert et al. (2011)
1,704 U.S. men and women dry-cleaning union member in CA, IL, MI,
NY follow-up 1940-2004 (618 subjects worked for one or more years
prior to 1960 only at shops where PCE was the primary cleaning solvent,
identified as PCE-only exposure), cancer mortality (SMR), female and
male breast cancer (ICD-9, 174, 175)
Selden and Ahlborg (2011)
9,440 Swedish men (n = 2,810) and women (n = 9,440) in 461 washing
and dry-cleaning establishments, identified by employer in mid-1980s,
employed 1973-1983, follow-up 1985-2000, exposure assigned using
company serf-reported information on PCE usage — PCE (dry cleaners and
laundries with a proportion of PCE dry cleaning), laundry (no PCE use),
and other (mixed exposures to PCE, CFCs, TCE, etc.), external referents
(SIR), ICD-7, 170
-^
o
-------
Table 4-36. Summary of human studies on tetrachloroethylene exposure and breast cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
All subjects, 1960 or 1970 Census in laundry and
dry cleaner or related occupation and industry
All subjects in 1960 and 1970 in laundry and dry
cleaner occupation and industry
Not reported
Not reported
Reference
Travier et al. (2002)
Swedish men and women identified as laundry worker, dry cleaner, or
presser (occupational title), in the laundry, ironing, or dyeing industry or
related industry in 1960 or 1970 (543,036 person-years); or, as laundry
worker, dry cleaner, or presser (occupational and job title) (46,933 person-
years) in both censuses, follow-up 1971-1989, external referents (SIR)
Case-control studies
British Columbia, Canada
Laundry and dry cleaning occupation
Pre- and postmenopausal
Usual occupation
5.24(1.41,19.5)
9
Postmenopausal
Usual occupation
4.85 (1.26, 18.7)
8
Power laundries and dry cleaners industry
Pre- and postmenopausal
Usual occupation
2.00(0.78,5.13)
9
Postmenopausal
Usual occupation
1.57(0.68,3.61)
10
Band et al. (2000)
995 breast cancer cases, females,75 yr, 1988-1989, identified fromBritish
Columbia Cancer Registry, Canadian citizens and British Columbia
residents, English speaking, 1,020 population controls matched on age and
sex, self -administered questionnaire, job title and industry coded to
Canadian SOC and Canadian SIC as exposure surrogate, OR for
postmenopausal subjects, adjusted for body weight in 1986, family history
of breast cancer, history of benign breast disease, cumulative alcohol
score. OR for pre- and postmenopausal subjects also adjusted for smoking
pack-years
-------
Table 4-36. Summary of human studies on tetrachloroethylene exposure and breast cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Poland, 2 regions (Warsaw and Lodz)
Laundry, cleaning and garment services industry
1.2 (0.7, 1.9)
28
Exposure duration
<10yr
>10yr
1.5 (0.8, 2.8)
0.5 (0.2, 16)
23
5
Reference
Peplonska et al. (2007)
2,275 histologically confirmed in situ or invasive breast cancers in female
residents of Warsaw and Lodz, 20-74 yr, 2000-2003, population controls,
identified from the Polish Electronic System of Population Evidence and
matched to cases by city of residence and age within 5-yr age groups, in-
person interview, structured questionnaire, lifetime occupational history,
employed >6 mo in relevant industry exposure surrogate, OR adjusted for
age, education, age of menarche, menopausal status, age at menopause,
number of full-time births, MBI, family breast cancer history, and previous
screening mammography
Geographic-based studies
Cape Cod, MA
PCC ROD < median
PCE ROD > median
PCE ROD >90th percentile
1.0 (0.7, 1.3)a
0.9 (0.6, 1.3)b
1.2 (0.9, 1.7)a
1.3 (0.9, 1.9)b
1.3 (0.7, 2.6)a
1.7 (0.8, 4.4)b
91
59
100
69
4
16
Aschengrau et al. (2003: 1998)
334 histologically confirmed breast cancer cases in males and females,
1983-1986, 2,236 population controls identified by random digit dialing,
vital records for deceased controls, and HCFA records if >65 yr
( Aschengrau etal. 1998): 672 histologically confirmed primary or
recurrent breast cancer cases infemales, 1987-1993, 616 population
controls identified by random digit dialing, vital records for deceased
controls, and HCFA records if >65 yr (Aschengrau et al.. 2003): MA
Cancer Registry, telephone interview, algorithm of (1993) to estimate mass
of PCE in drinking water entering residence was surrogate exposure metric
[90th percentile, 53.4], OR adjusted for age of diagnosis or index year, vital
status at interview, family history of breast cancer, age at first live birth,
personal history of prior breast cancer and benign breast disease, and
occupational exposure to solvents (PCE, benzene, other solvents),
statistically analyses also explored effect of different latent periods
(0,5,7,9, 11, 13, and 15 yr)
to
a In Aschengrau et al. (2003). odds ratios for breast cancer are presented for combined data from Aschengrau et al.
bOdds ratios considering a 7-yr latent period.
HCFA = Health Care Financing Administration, ISCO = International Standard Classification of Occupation, ISIC = International Standard Industry
Classification, JEM = job-exposure-matrix, RDD = relative delivered dose, TWA = time-weighted-average.
-------
No male breast cancer cases were observed in four of the five studies reporting risk
estimates for males separately from that of females (Selden and Ahlborg, 2011; Chang et al.,
2005; Anderson et al., 1990; Lynge and Thygesen, 1990). Not surprising given the low
background rate of male breast cancer, less than one case was expected in each study. Pukkala et
al. (2010) reported three observed cases among a cohort of 8,744 male dry-cleaners and laundry
workers.
Two27 of the eight cervical cancer studies evaluated by EPA with exposure assessment to
tetrachloroethylene or employment as dry-cleaner or laundry worker reported estimated relative
risks based on 50 or more observed events. Estimates of the standardized incidence ratio or
SMR in these studies were 1.34 (95% CI: 1.12, 1.60) and 1.20 (95% CI: 1.08, 1.34) in Travier et
al. (2002) and Pukkala et al. (2009), respectively. In addition to the evidence from the two large
cohort studies, additional evidence is found in six other studies whose effect estimates are based
on fewer observed events and that carry lesser weight in the analysis. As expected, the
magnitude of the point estimate of the association reported in these studies is more variable than
in the larger studies: 0.40 to 0.98 (Lynge et al.. 2006: Lynge and Thygesen, 1990). 1.1 to 1.5
(Selden and Ahlborg. 2011). 1.6 to 2.0 (Calvert et al.. 2011: Blair etal.. 2003: Ruder et al..
2001). and >3.0 (Anttila et al.. 1995). Chang et al. (2005) and Sung et al. (2008). a follow-up
study of the same population, observed over 200 cervical cancer incident cases and reported
standardized incidence ratios of 1.06 (95% CI: 0.95, 1.18) and 0.69 (95% CI: 0.87, 1.06).
Although based on a large number of observed events, these studies carry lesser weight in the
analysis given their lower level exposure-assessment approach. SMRs or standardized incidence
ratios for cervical cancer were lower for subjects identified as dry cleaners compared to laundry
workers or for the subcohort of females whose starting date of employment was after 1960
compared to the larger cohort (Selden and Ahlborg, 2011: Lynge et al., 2006) (refer to Table
4-37).
27 In addition to Andersen et al. (1999), Boice et al. (1999) is not counted because no cervical deaths are observed
among tetrachloroethylene-exposed female subjects.
4-343
-------
Table 4-37. Summary of human studies on tetrachloroethylene exposure and cervical cancer
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Cohort studies
Biologically monitored workers
All subjects
3.20(0.39, 11.6)
2
Aerospace workers (Lockheed)
Routine exposure to PCE
Routine-Intermittent exposure duration to PCE
(0.00, 7.77)
Not reported
0
0.47 exp
Electronic factory workers (Taiwan)
All Subjects
Females
Females
1.06(0.95, 1.18)
0.96 (0.87, 1.06)
337
337
Aircraft maintenance workers from Hill Air Force Base
Any PCE exposure
Not reported
Anttila et al. (1995)
849 Finnish men and women, blood PCE [0.4 umol/L in females
and 0.7 umol/L in males (median)], follow-up 1974-1992,
external referents (SIR)
Boice et al. (1999)
77,965 (n = 2,63 1 with routine PCE exposure and n = 3,199 with
intermittent-routine PCE exposure), began work during or after
1960, worked at least 1 yr, follow-up 1960-1996, job exposure
matrix without quantitative estimate of PCE intensity, 1987-1988
8-h TWA PCE concentration (atmospheric monitoring) 3 ppm
[mean] and 9.5 ppm [median], external reference for routine
exposure (SMR) and internal references (workers with no
chemical exposures) for routine-intermittent PCE exposure (RR)
Chang et al. (2005): Sung et al. (2007)
86,868 (n = 70,735 female), follow-up 1979-1997, multiple
solvents exposure, does not identify PCE exposure to individual
subjects, cancer mortality, external referents (SIR); female genital
organs (Chang et al.. 2005):
63,982 females, follow-up 1979-2001, factory employment proxy
for exposure, multiple solvents exposures and PCE not identified
to individual subjects, cancer incidence, external referents,
analyses lagged 15 yr (SIR) (Sung et al.. 2007)
Radican et al. (2008)
10,461 men and 3,605 women (total n = 14,066, n = 10,256 ever
exposed to mixed solvents, 85 1 ever-exposed to PCE), employed
at least 1 yr from 1952 to 1956, follow-up 1973-2000, job
exposure matrix (intensity), internal referent (workers with no
chemical exposures) (RR)
-------
Table 4-37. Summary of human studies on tetrachloroethylene exposure and cervical cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Dry -cleaner and laundry workers
All laundry worker and dry cleaners
1.18(1.01, 1.38)
155
All subjects
1.6(1.0,2.3)
27
Semiquantitative exposure score
Little to no exposure
Medium to high exposure
1.5 (0.8, 2.7)
1.4 (0.7, 1.7)
12
11
Laundry workers and dry cleaners in 1960
Census
Not reported
Laundry worker and dry cleaners
0.40 (0.28, 0.52)
34
Launderer and dry cleaner
1.20(1.08, 1.34)
332
Reference
Andersen et al. (1999)
29,333 men and women identified in 1960 Census (Sweden) or
1970 Census (Denmark, Finland, Norway), follow-up 1971-1987
or 1991, PCE not identified to individual subjects, external
referents (SIR)
Blair et al. (2003)
5,369 U.S. men and women laundry and dry-cleaning union
members (1945-1978), follow-up 1979-1993, semiquantitative
cumulative exposure surrogate to dry clean solvents, cancer
mortality, external referents (SMR)
Ji et al. (2005a, b); Ji and Hemminki (2005a, b, c)
9,255 Swedish men and 14,974 Swedish women employed in 1960
(men) or 1970 (women) as laundry worker or dry cleaner, follow-
up 1961/1970-2000, PCE not identified to individual subjects,
external referent (SIR) and adjusted for age, period and
socioeconomic status
Lynge and Thygesen (1990)
10,600 Danish men and women, 20-64 yr old, employed in 1970
as laundry worker, dry cleaners and textile dye workers, follow-up
1970-1980, external referents (SIR)
Pukkala et al. (2009)
Men and women participating in national census on or before
1990, 5 Nordic countries (Denmark, Finland, Iceland, Norway,
Sweden), 30-64 yr, follow-up 2005, occupational title of launderer
and dry cleaner in any census, external referents (SIR)
-------
Table 4-37. Summary of human studies on tetrachloroethylene exposure and cervical cancer (continued)
Exposure group
All subjects
Exposure duration/time since 1st employment
<5 yr/<20 yr
>5 yr/<20 yr
<5 yr/>20 yr
>5 yr/>20 yr
PCE subcohort
Dry-cleaners and laundry workers
PCE
Duration of employment
-------
Table 4-37. Summary of human studies on tetrachloroethylene exposure and cervical cancer (continued)
Exposure group
Relative risk
(95% CI)
No. obs.
events
Reference
Case-control studies
Nordic Countries (Denmark, Finland, Norway, Sweden)
Unexposed
Dry cleaner
Other in dry-cleaning
Unclassifiable
1.00
0.98 (0.65, 1.47)
1.72 (1.00, 2.97)
1.11(0.72, 1.71)
105
36
22
44
Dry cleaner, employment duration, 1964-1979
10yr
Unknown
2.68(0.89,8.11)
0.78(0.31, 1.94)
0.47(0.20, 1.13)
1.18(0.64,2.15)
1.14(0.12, 11.00)
7
6
6
16
1
Lynge et al. (2006)
Case-control study among 46,768 Danish, Finnish, Norwegian,
and Swedish men and women employed in 1960 as laundry
worker or dry cleaner, follow-up 1970-1971 to 1997-2001, 102
cervical cancer cases, 3 controls per case randomly selected from
cohort matched on country, sex, age, calendar period at diagnosis
time, occupational task at 1970 Census proxy for exposure,
cervical cancer incidence, RR adjusted for matching criteria
HCFA = Health Care Financing Administration, ISCO = International Standard Classification of Occupation, ISIC = International Standard Industry
Classification, JEM = job-exposure-matrix, TWA = time-weighted-average.
-------
Establishment of an exposure or concentration-response relationship can add to the
weight of evidence for identifying a cancer hazard, but only limited data pertaining to exposure-
response relationships for lung cancer and tetrachloroethylene exposure are available. Three
studies of breast cancer presented risk estimates for increasing exposure categories; one study
using exposure duration as a proxy (Peplonska et al., 2007) and two studies with a
semiquantitative or quantitative exposure surrogate (Aschengrau et al., 2003; Blair et al., 2003).
Risk estimates are larger for highest exposure groups compared to overall exposure or to a no or
low exposed group in one cohort study that use a semiquantitative or quantitative exposure-
assessment approach (Blair et al., 2003), and in one study when latent periods are considered
(Aschengrau et al., 2003). One other study with an exposure assessment based on exposure
duration reported a lower risk estimate with >10 years longer exposure duration than the risk
estimate for <10 years (Peplonska et al., 2007).
With respect to cervical cancer, five studies presented risk estimates for increasing
exposure categories using exposure duration (Calvert et al., 2011; Selden and Ahlborg, 2011;
Lvnge et al.. 2006: Blair etal.. 2003: Travier et al.. 2002). Calvert et al. (2011) was the only
study to report a higher risk estimate for cervical cancer for the group with longest exposure
duration (<5 years versus 5+ years).
All three case-control studies of breast cancer controlled for associated risk factors
(Peplonska et al., 2007: Aschengrau et al., 2003: Band et al., 2000). Direct examination of
possible confounders is less common in cohort studies examining breast cancer compared to
case-control studies where information is obtained from study subjects or their proxies. None of
the cohort studies of cervical cancer considered socioeconomic or lifestyle factors such as
smoking or exposure to the human papilloma virus (HPV), a known risk factor for cervical
cancer and correlated with socioeconomic status, particularly with the squamous cell subtype
(NCI, 2010: Pukkala et al.. 2010). The case-control study of Lynge et al. (2006) included
controls similar in socioeconomic status as cases, and the odds ratio estimate in this study for dry
cleaners did not support an association with tetrachloroethylene.
In conclusion, most studies examined breast cancer in females (Radican et al., 2008:
Peplonska et al.. 2007: Sung et al.. 2007: Aschengrau et al.. 2003: Blair etal.. 2003: Band et al..
2000): or males and females combined (Calvert et al., 2011: Ruder et al., 2001: Boice et al.,
1999). Five studies, mostly of Nordic subjects, presented risk estimates for male subjects
separately (Selden and Ahlborg, 2011: Pukkala et al., 2009: Chang et al., 2005: Anderson et al.,
1990: Lynge and Thygesen, 1990). The results from the large studies of breast cancer risk in
women in relation to tetrachloroethylene exposure are mixed. The largest, based on 1,757 breast
cancer cases in female dry-cleaners and laundry workers, reported a statistically significant
deficit in the risk of breast cancer incidence compared to the populations of Nordic countries
4-348
-------
(Pukkala et al., 2009). Findings in the other six studies were based on fewer events or exposed
cases; two of four studies with nonspecific exposure-assessment methodology provided evidence
for association between breast cancer in females and tetrachloroethylene exposure (Sung et al.,
2007; Chang et al., 2005; Aschengrau et al., 2003; Anderson et al., 1990; Lynge and Thygesen,
1990) but effects were not observed in two other large cohort studies with a relatively high
quality exposure-assessment methodology to tetrachloroethylene (Selden and Ahlborg, 2011;
Blair et al., 2003). Small studies observed mixed findings (Calvert et al., 2011; Radican et al.,
2008: Peplonska et al., 2007: Sung et al., 2007: Chang et al., 2005: Aschengrau etal., 2003:
Band et al., 2000: Boiceetal., 1999). Band et al. (2000), but not other less-weighted studies,
excluded chance as an alternative explanation. Although cohort studies were unable to control
for potential confounding from reproductive history or menopausal status, observations in case-
control studies controlled for these potential confounders in statistical analyses and provided
support of an association between female breast cancer and tetrachloroethylene compared to
controls (Peplonska et al., 2007: Aschengrau et al., 2003: Band et al., 2000). Three studies
examined exposure response, with risk estimates in females monotonically increased in higher
exposure groups in two studies with semiquantitative or quantitative exposure-assessment
approaches (Aschengrau et al., 2003: Blair et al., 2003). A third study examining exposure
duration observed an inverse relation (Peplonska et al., 2007). Exposure duration is more
uncertain than use of a semiquantitative surrogate given increased potential for bias associated
with exposure misclassification. Because of the limitation in statistical power, none of the five
studies reporting on male breast cancer is adequate to examine tetrachloroethylene exposure. All
studies of male breast cancer are sufficiently underpowered; no male breast cancer cases were
observed in four of the five studies (Selden and Ahlborg, 2011; Pukkala et al., 2010; Chang et
al., 2005: Anderson et al., 1990: Lynge and Thygesen, 1990).
For cervical cancer, the results from the two large cohort studies of dry cleaners are
consistent with an elevated cervical cancer risk of 20-30% (Pukkala et al., 2009; Travier et al.,
2002). Results from four smaller cohort and case-control studies with a relatively high quality
exposure-assessment methodology presented a pattern of more variable results, with relative
risks of 0.98 (95% CI: 0.65, 1.47), 1.19 (95% CI: 0.64, 1.93), 2.10 (95% CI: 0.68, 4.90), and 3.20
(95% CI: 0.39, 11.6) in Lynge et al. (2006), Selden and Ahlborg (2011), Calvert et al. (2011),
and Anttila et al. (1995), respectively. A fourth study with higher quality exposure-assessment
specific to tetrachloroethylene did not observe any cervical cancer deaths among women, but less
than one death was expected (Boiceetal., 1999). Calvert et al. (2011) was the only study to
report an exposure response gradient with employment duration. Dry cleaning workers did not
have higher cervical cancer risks compared with laundry workers or other categories of dry
cleaning workers (Selden and Ahlborg, 2011; Lynge et al., 2006). Lack of data on
4-349
-------
socioeconomic status—a proxy for exposure to the human papilloma virus, a known risk factor
for cervical cancer—indicates great uncertainty for asserting this association with
tetrachloroethylene exposure. Potential confounding by socioeconomic status is an alternative
explanation, with some support provided by Lynge et al. (2006), a case-control study with
controls of similar socioeconomic status as cases, and who did not observe an association
between cervical cancer and dry cleaning.
4.7.3. Summary of Human and Animal Developmental/Reproductive Studies
4.7.3.1. Summary of Human Data
Studies of tetrachloroethylene exposure have evaluated several reproductive outcomes
including effects on menstrual disorders (Zielhuis et al., 1989), semen quality (Eskenazi et al.,
1991a: Eskenazi et al., 1991b), fertility (Eskenazi et al., 199la: Rachootin and Olsen, 1983), time
to pregnancy (Sallmen et al., 1998; Sallmen et al., 1995), and risk of adverse pregnancy
outcomes including spontaneous abortion (Aschengrau et al., 2009a: Doyle et al., 1997;
Lindbohm et al., 1991: Windham et al., 1991: Ahlborg, 1990b: Lindbohm et al., 1990: Olsen et
al., 1990: Kyyronen et al., 1989: Taskinen et al., 1989: Boscoetal., 1987: McDonald et al.,
1987: McDonald et al., 1986), low birth weight or gestational age (Aschengrau et al., 2008:
Olsen etal., 1990: Boscoetal., 1987: McDonald et al., 1987), birth anomalies (Aschengrau et
al., 2009b: Ahlborg, 1990b: Olsen etal., 1990: Boscoetal., 1987: McDonald et al., 1987), and
stillbirth (Olsen etal., 1990: McDonald et al., 1987). A few studies evaluated effects of prenatal
exposure to tetrachloroethylene on postnatal development including learning and behavior, and
schizophrenia (Janulewicz et al., 2008: Perrin et al., 2007). Many of the studies evaluated
exposure during a specific critical window relevant to the health endpoint under study, for
example, the period before conception or during the first trimester.
Some studies that relied on detailed work histories and monitoring data to classify
exposure were suggestive that maternal or paternal exposure to tetrachloroethylene or work in
dry cleaning reduces fertility or delays conception (Sallmen et al., 1998: Sallmen et al., 1995:
Eskenazi et al., 1988). However, the risk estimates were imprecise because the number of
participants reporting exposure to tetrachloroethylene was small. One small study of primarily
unionized workers in the dry-cleaning and laundry industries in California observed subtle
deficits in sperm quality in relation to tetrachloroethylene exposure (Eskenazi etal., 1988).
However, three clinically recognized measures of sperm quality were not associated with
exposure in the study population. A study of occupational exposures among a group of infertile
couples who sought treatment found no association between either a diagnosis of sperm
4-350
-------
abnormalities among male partners, or a diagnosis of hormonal disturbances among female
partners with self-reported exposure to dry-cleaning chemicals (Rachootin and Olsen, 1983).
The results of Eskenazi et al. (1988) are compelling, but more studies are needed to
conclude if exposure to tetrachloroethylene is associated with adverse effects on male and female
reproduction.
Results of several studies of maternal occupational exposure to tetrachloroethylene
suggest an increased risk of spontaneous abortion, particularly at higher levels (Doyle et al.,
1997: Windham et al.. 1991: Lindbohm et al.. 1990: Olsen etal.. 1990: Kwronen et al.. 1989).
Most of the studies evaluated exposure during the first trimester of pregnancy. Some of the
studies observed an increased odds ratio ranging between 1.4 to 4.7, but had low statistical power
because the cohort contained small numbers of exposed cases and controls, and were limited in
their ability to evaluate potential confounding (Windham et al., 1991: Lindbohm et al., 1990:
Olsen et al., 1990: Bosco et al., 1987). In general, the studies that used a more precise definition
of exposure, or categorized exposure into levels of increasing dose or intensity, observed higher
risk estimates. For example, two reports of occupational exposure in the dry-cleaning and
laundry industries in Finland observed a dose-related increase in risk among employees
classified into risk levels based on whether or not their work tasks involved dry cleaning (Olsen
et al., 1990: Kyyronen et al., 1989). Odds ratios for low and high exposure compared to no
exposure were 1.18 (95% CI: 0.71-1.97) and 4.53 (95% CI: 1.11-18.5), respectively. The
Finnish studies controlled for reported exposure to other substances in the workplace as well as
for several potential confounders. They also found agreement between self-reported exposures
and biological measurements taken close to the time of pregnancy for a small subset of the
cohorts. A relatively large study of workers in the United Kingdom classified exposure among
current and former employees at dry-cleaning and laundry establishments by job tasks (machine
operator versus other tasks) and analyzed risk of spontaneous abortions among all pregnancies
reported between 1980 and 1995 (Doyle et al., 1997). Machine operators had a 63% higher risk
of spontaneous abortion compared to nonoperators adjusting for several potential confounders
(OR: 1.63, 95% CI: 1.09-3.05). These findings are consistent with breathing zone
measurements of tetrachloroethylene in dry-cleaning establishments, indicating that machine
operators have the highest exposures (Gold et al., 2008).
Increased risks were not found among dry cleaners in Sweden using a comparable study
design (Ahlborg, 1990a: Olsen et al., 1990). Further, three studies of paternal occupational
exposure prior to the beginning of the pregnancy did not observe an association (Eskenazi et al.,
1991a: Lindbohm et al., 1991: Taskinen et al., 1989). Two of these surveyed occupational
exposure to a broad array of substances and, consequently, had low statistical power for
chemical-specific analyses (Lindbohm et al., 1991: Taskinen et al., 1989). Although there is no
4-351
-------
evidence of an increased risk associated with paternal exposure, the studies were not of sufficient
size, nor did they provide adequate detail regarding exposure estimates to allow definitive
conclusions. Finally, no associations with incidence of spontaneous abortion were observed
among two populations exposed to tetrachlorethylene in drinking water (Aschengrau et al.,
2009a: Aschengrau et al., 2008; Lagakos et al., 1986). The studies of drinking water
contamination evaluated populations with much lower exposures compared to the occupational
cohorts.
Studies of tetrachloroethylene in drinking water have reported that exposure during
pregnancy is associated with low birth weight (Sonnenfeld et al., 2001; Bove etal., 1995;
Lagakos et al., 1986), eye/ear anomalies (Lagakos et al., 1986), and oral clefts (Aschengrau et
al., 2009b: Bove et al., 1995; Lagakos et al., 1986). However, the number of cases with birth
anomalies in specific diagnostic groups was very small, and CIs often included one. In addition,
imprecise exposure estimates likely resulted in nondifferential misclassification, biasing risk
estimates toward the null. Participants in the studies were exposed to multiple contaminants, and
it was not possible to disentangle substance-specific risks.
Aschengrau et al. (2008) evaluated a unique exposure event in a population in eight Cape
Cod towns exposed to a wide range of tetrachloroethylene concentrations in an irregular pattern
throughout the region (1.5-7,750 ug/L). It is less likely that the population was exposed to
sizable concentrations of other halogenated substances. A detailed exposure model was used to
estimate the distribution of contaminated water to the homes of residents. Birth weight and
gestational age were not associated with exposure to tetrachlorethylene. Effect estimates for
some congenital anomalies were increased, although the number of infants with anomalies was
very small, and statistical power was low. The small increased risk is consistent with the other
studies of drinking water exposure to mixtures of halogenated pollutants. Diagnoses of attention
deficit disorder, hyperactive disorder or educational histories reported by the mothers about their
children were not increased in relation to the amount of tetrachloroethylene delivered to the
homes during pregnancy or childhood (Janulewicz et al., 2008). On the other hand, a more than
threefold risk of schizophrenia was associated with dry cleaning as a surrogate for prenatal
tetrachloroethylene exposure (Perrin et al., 2007). The longitudinal design and use of a national
registry to identify psychiatric diagnoses were strengths of the study, but tetrachloroethylene
exposure was not directly analyzed. In conclusion, the literature is insufficient to draw
conclusions regarding effects of tetrachloroethylene exposure on development in infants and
children.
Most epidemiologic studies examined breast cancer in females (Radican et al., 2008;
Peplonska et al., 2007: Sung et al., 2007: Aschengrau et al., 2003: Blair etal., 2003: Band et al.,
2000) or males and females combined (Calvert et al., 2011: Ruder etal., 2001: Boice et al..
4-352
-------
1999): five studies presented risk estimates for male subjects separately (Selden and Ahlborg,
2011: Pukkala et al.. 2009: Chang et al.. 2005: Anderson et al.. 1990: Lynge and Thygesen,
1990). The largest study, based on 1,757 breast cancer cases in female dry-cleaners and laundry
workers, reported a statistically significant deficit in the risk of breast cancer incidence compared
to the populations of Nordic countries. Findings in the other four large studies were based on
fewer events or exposed cases with mixed findings (Selden and Ahlborg, 2011: Aschengrau et
al., 2003: Blair etal., 2003: Lynge and Thygesen, 1990). Additional studies carrying less weight
also observed mixed findings (Radican et al., 2008: Peplonska et al., 2007: Sung et al., 2007:
Chang et al.. 2005: Ruder et al.. 2001: Band et al.. 2000: Boice et al.. 1999). Three studies
examined exposure-response, with risk estimates in females monotonically increased in higher
exposure groups in two studies with semiquantitative or quantitative exposure-assessment
approaches (Aschengrau et al., 2003: Blair et al., 2003). A third study examining exposure
duration observed an inverse direction (Peplonska et al., 2007). Exposure duration is more
uncertain than use of a semiquantitative approach because there is increased potential for bias
associated with exposure misclassification. None of the five studies reporting on male breast
cancer is adequate to examine tetrachloroethylene exposure. All studies of male breast cancer
are statistically underpowered; no male breast cancer cases were observed in four of the five
studies (Selden and Ahlborg, 2011: Chang et al., 2005: Anderson et al., 1990: Lynge and
Thygesen, 1990), less than one case was expected in each study, and Pukkala et al. (2009)
observed three cases among a cohort of 8,744 male dry-cleaners and laundry workers.
For cervical cancer, the results from the two large cohort studies with broad exposure
assessment is consistent with an elevated cervical cancer risk of 20-30% (Pukkala et al., 2009:
Travier et al., 2002). Results from four smaller cohort and case-control studies with a higher
quality exposure-assessment methodology presented a pattern of more variable results, with
relative risks of 0.98 (95% CI: 0.65, 1.47), 1.19 (95% CI: 0.64, 1.93), 2.10 (95% CI: 0.68, 4.90),
and 3.20 (95% CI: 0.39, 11.6) in Lynge et al. (2006). Selden and Ahlborg (2011). Calvert et al.
(2011), and Anttila et al. (1995), respectively. A fourth study with high quality exposure
assessment specific to tetrachloroethylene did not observe any cervical cancer deaths among
women and was insensitive, as less than one death was expected (Boice et al., 1999). Calvert et
al. (2011) was the only study to report an exposure response gradient. Dry cleaning workers did
not have higher cervical cancer risks compared with laundry workers or other categories of dry
cleaning workers (Selden and Ahlborg, 2011: Lynge et al., 2006). Lack of data on
socioeconomic status—a proxy for exposure to the human papilloma virus, a known risk factor
for cervical cancer—indicates great uncertainty for asserting this association with
tetrachloroethylene exposure. Potential confounding by socioeconomic status is an alternative
explanation with some support provided by Lynge et al. (2006), a case-control study with
4-353
-------
controls of similar socioeconomic status as cases, and who did not observe an association
between cervical cancer and dry cleaning.
4.7.3.2. Summary of Animal Data
Table 4-38 summarizes the findings of the animal developmental and reproductive
toxicity studies described in Sections 4.7.2.1 to 4.7.2.3. The inhalation study database includes
assessments of developmental toxicity in rats, mice, and rabbits following exposures during
gestation, assessments of developmental neurotoxicity in rats following pre- and/or postnatal
exposures of the offspring, and evaluation of reproductive and fertility outcomes in rats and
mice. Additional supportive studies include in vitro assays of embryo development and oocyte
fertilizability, a developmental assay in Japanese medaka, and two oral gavage studies that
assessed developmental toxicity in rats and developmental neurotoxicity in mice.
Limitations of the inhalation developmental and reproductive toxicity studies are
described in the individual study summaries above. These limitations include the lack of dose-
response information due to the use of a single treatment level in the prenatal developmental
toxicity assessment by Schwetz et al. (1975): the lack of either maternal or developmental
toxicity in Hardin et al. (1981): absence of methodological details in study reporting (Szakmary
et al., 1997): and a concern about a short peri-parturition exposure gap in Tinston (1994).
Additionally, the studies were conducted in accordance with standard EPA and OECD
toxicological study guidelines in place at the time but did not assess endpoints that are included
in the guidelines that were revised and harmonized in 1998 [e.g., refer to Tinston (1994)].
Maternal toxicity, when observed, did not compromise the evaluation or interpretation of
treatment-related findings in the offspring.
4-354
-------
Table 4-38. Summary of mammalian developmental and reproductive
toxicity studies for tetrachloroethylene
Subjects
Effects
Concentration
Authors
Developmental toxicity studies
Rat (whole
embryo culture)
Mortality, malformations, delayed
growth and differentiation
No effect at 2.5 mM, effects at 3.5
mM and higher
Saillenfait et al.
(1995)
Japanese
medaka
Decreased egg viability at 96-h
(LC50 = 27 mg/L); at 10 d: decreased
hatchability and larval survival,
increased developmental abnormalities
10d:0, 1.5,3,6, 12, 25 mg/L
LOAEL= 1.5 mg/L
Spencer et al.
(2002)
SW Mice
Maternal toxicity (statistically
significant, 21% increase in mean
relative liver weight), decreased fetal
weight, delayed ossification, 9%
decrease in birth weight
Inhalation: 0, 300 ppm on CDs 6-15
Schwetz et al.
(1975)
S-D Rats
Maternal toxicity (slight, but
statistically significant, 4-5%
reductions in maternal body weight
gains), increased resorptions (fetal
death)
Inhalation: 0, 300 ppm on Days
6-15
Schwetz et al.
(1975)
S-D Rats, NZW
Rabbits
No developmental toxicity
Inhalation: Exposures throughout
gestation
NOAEL = 500 ppm
Hardin et al.
(1981)
F344 Rats
100% mortality at 1,200 mg/kg-day,
increased mortality and micro-
/anophthalmia at 900 mg/kg-day; soft
tissues not examined
Gavage, 0, 900, 1,200 mg/kg-day on
CDs 6-19
Narotsky and
Kavlock (1995)
CFY Rats
Maternal toxicity (statistically
significant, 37-40%; decreased body
weight gain; slight, but statistically
significant, 6-10% increased liver
weight; and increased serum enzymes);
increased pre- and postimplantation
loss, skeletal retardation, and total
malformations; decreased fetal weight
Inhalation: 0, 1,500, 4,500, 8,500
mg/m3onGDsl-20
LOAEL = 1,500 mg/m3
Szakmary et al.
(1997)"
C57B1 Mice
Maternal toxicity (statistically
significant increased liver weight);
visceral malformations
Inhalation: 0, 1,500 mg/m3 on CDs
7-15
LOAEL = 1,500 mg/m3
Szakmary et al.
(1997)
4-355
-------
Table 4-38. Summary of mammalian developmental and reproductive
toxicity studies for tetrachloroethylene (continued)
Subjects
NZW Rabbits
S-D Rats
Effects
Maternal toxicity (decreased body
weight gain, increased liver weight);
abortions, total litter resorptions,
increased postimplantation loss,
malformations
Maternal toxicity (slight, but
statistically significant, decreased body
weight gain; decreased gravid uterine
weight); fetal body weight and
placental weight decrements, increased
delays in thoracic vertebral ossification
Concentration
Inhalation: 0, 4,500 mg/m3 on CDs
7-20
LOAEL = 4,500 mg/m3
Inhalation: 0, 75, 250, or 600 ppm
(actual concentrations: 0, 66, 249,
600 ppm), 6 h/d, 7 d/wk, on CDs
0-19
Maternal LOAEL = 600 ppm
Fetal LOAEL = 250 ppm
Authors
Szakmary et al.
(1997)
Carney et al.
(2006)
Developmental neurotoxicity assessments
CFY Rats
S-D Rats
S-D Rats, two-
generation
study
NMRI Mice
Decreased postnatal survival, minimal
transient decreases in exploratory
activity and muscular strength, and
increased motor activity in females on
PND 100
Decreased weight gain, behavioral
changes (more extensive for late
pregnancy exposure), decreased brain
acetylcholine
Behavioral effects (decreased activity;
reduced response to sound) in Fl pups
Alterations in spontaneous motor
activity (locomotion, rearing, and total
activity) at PND 60
Inhalation: 0, 1,500, 4,500 mg/m3 on
GDs 1-20 (and perhaps postnatally
to PND 100)
LOAEL = 1,500 mg/m3
Inhalation: 0, 100, 900 ppm on Days
7-13 or on Days 14-20
NOAEL = 100 ppm
LOAEL = 900 ppm
Inhalation: 0, 100, 300, 1,000 ppm
NOAEL = 300 ppm
LOAEL = 1,000 ppm
Gavage: 0, 5, 320 mg/kg-day on
PNDs 10-16
LOAEL = 5 mg/kg-day
Szakmary et al.
(1997)"
Nelson et al.
(1979)
Tinston (1994)b
Fredriksson et al.
(1993)
Reproductive toxicity studies
Rat (in vitro)
CD-I Mice
Reduced fertilizability of extracted
oocytes
Abnormal sperm heads at 500 ppm but
not at 100 ppm, spermatogonia or
spermatocyte stage affected
12,000 mg/m3, 2 hours/d, 5 d/wk for
2wk
Inhalation: 0, 100, 500 ppm for 5 d
LOAEL = 500 ppm
Berger and Horner
(2003)
Beliles et al.
(1980)
4-356
-------
Table 4-38. Summary of mammalian developmental and reproductive
toxicity studies for tetrachloroethylene (continued)
Subjects
S-D Rats, two-
generation
study
Effects
Increased death of Fl A and F2A and
F2B pups, decreased body weight
Concentration
Inhalation: 0, 100, 300, 1,000 ppm
NOAEL = 100 ppm for
body weight reduction
Authors
Tinston (19941b
a The Szakmary et al. (1997) study in CFY rats assessed both developmental toxicity and developmental
neurotoxicity outcomes.
b The Tinston (1994) study in S-D rats demonstrated both developmental neurotoxicity and reproductive
toxicity outcomes.
The tetrachloroethylene database included assessments of the various potential
manifestations of developmental toxicity, i.e., alterations in survival, growth, morphology, and
functional development. Indications of effects on prenatal survival following in utero exposure
included increased pre- and/or postimplantation loss in rats, mice, and rabbits (Szakmary et al.,
1997; Schwetz et al., 1975). These findings were supported by evidence of embryo mortality in
a rat whole embryo culture (WEC) assay (Saillenfait et al., 1995) and decreased viability in a
Japanese medaka assay (Spencer et al., 2002). Decreased prenatal growth was observed in mice
(Schwetz et al., 1975) and rats (Szakmary et al., 1997). Morphological alterations associated
with prenatal exposures to tetrachloroethylene included delays in skeletal ossification in mice
(Schwetz et al., 1975) and rats (Carney et al., 2006; Szakmary et al., 1997), which were often
associated with fetal weight decrements, and increased incidences of malformations in mice, rats,
and rabbits (Szakmary et al., 1997). Evidence of tetrachloroethylene exposure-related
malformations was also observed in the rat WEC and Japanese medaka assays (Spencer et al.,
2002; Saillenfait et al., 1995) and in a gavage prenatal developmental toxicity screening study in
rats (Narotsky and Kavlock, 1995). Alterations in neurological function following pre- and/or
postnatal inhalation exposures to tetrachloroethylene were observed in rats by Szakmary et al.
(1997), Nelson et al. (1979), and Tinston (1994). These findings were supported by a study that
found altered spontaneous motor activity in young adult rats that had been treated orally with
tetrachloroethylene postnatally during a critical period of nervous system development
(Fredriksson et al., 1993). Additionally, reductions in brain acetylcholine and dopamine were
observed in rat offspring following gestational tetrachloroethylene exposures (Nelson et al.,
1979).
An assessment of fertility and reproductive function in rats exposed to
tetrachloroethylene via inhalation over the course of two generations was conducted by Tinston
(1994). Effects on offspring included decreased pup weights and postnatal survival in both
4-357
-------
generations, as well as behavioral alterations in the Fl pups. Decreased mean testes weight was
observed in Fla males; however, no effects on male or female fertility or other evidence of
alterations in reproductive function were observed. For males, this finding is supported by the
results of a study by Beliles et al. (1980), who found no sperm abnormalities in rats following up
to 10 weeks of tetrachloroethylene inhalation exposures. While the Beliles et al. (1980) study
identified an increase in abnormal sperm heads in mice after 4 weeks of exposure, no other
reproductive toxicity data in mice were available to aid in the interpretation of this finding.
In conclusion, based upon a consideration of the entire available database of animal
developmental and reproductive toxicity studies for tetrachloroethylene, the overall inhalation
NOAEL is 100 ppm, based on Tinston (1994). The overall inhalation 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.
Overall, the developmental and reproductive toxicity database for tetrachloroethylene
was judged to include a range of data from appropriate well-conducted studies in several
laboratory animal species plus limited human data and was considered sufficient for hazard
characterization and dose-response assessment, based upon EPA risk assessment guidelines
(U.S. EPA, 1996a, 1991b).
4.7.4. Mode of Action for Developmental Effects
Because of its lipid solubility, tetrachloroethylene can cross both the blood:brain barrier
and the placental barrier and, therefore, it can be present in all tissues, including the brain, during
development.
Peroxidation of the lipids of the cell membranes (Cojocel etal., 1989), alteration of
regulation of fatty acid composition of the membrane (Kyrklund and Haglid, 1991), disturbances
in the properties of the nerve membrane (Juntunen, 1986), and progressively increased activity in
one or more of the phosphoinositide-linked neurotransmitters (Subramoniam et al., 1989) have
all been suggested as MO As for neurotoxic effects. These mechanisms could be involved during
development phases, as well as in adults.
The metabolite TCA may be a causative agent or contribute to developmental toxicity
expressed as morphological changes, lethality, or growth reductions. Evidence in support of this
speculative position is presented in the following discussion. TCA is a weak organic acid, as are
many developmental toxicants, such as ethylhexanoic acid and valproic acid. These materials
accumulate to a greater extent in the embryo/fetal compartment than in the mother, based on the
pKa of the acid and the pH gradient between the maternal plasma and the embryo compartments
(O'Flaherty et al., 1992). TCA could induce developmental toxicity by changing the intracellular
4-358
-------
pH or through peroxisome proliferation. Ghantous et al. (1986) detected TCA in the amniotic
fluid of pregnant mice exposed to tetrachloroethylene via inhalation.
Smith et al. (1989) found that oral gavage doses of TCA (330, 800, 1,200, and
1,800 mg/kg-day) delivered on GDs 6-15 to pregnant Long-Evans rats produced soft tissue
malformations, principally in the cardiovascular system. Johnson et al. (1998) found cardiac
defects in rat fetuses whose mothers received 2,730 ppm TCA in drinking water during the
period of cardiac development. Saillenfait et al. (1995), using the rat whole embryo (Day 10)
culture system, found that both tetrachloroethylene and TCA induced embryo toxicity, including
mortality, malformations, and delayed growth and differentiation. TCA produced a reduction in
the first branchial arch as well as other morphological changes at a lower concentration (2.5 mM)
than that at which tetrachloroethylene induced no adverse effect (3.5 mM). TCA also induced a
reduction of the yolk sac diameter at 1 mM.
Arguments counter to the involvement of TCA in the MOA for tetrachloroethylene
developmental toxicity include that the types of malformations associated with TCA [i.e., cardiac
malformations reported by Smith et al. (1989) and Johnson et al. (1998)1 or other weak acid
exposures [e.g., valproic acid and ethylhexanoic acid (Scott et al., 1994)1 are not consistent with
those observed in tetrachloroethylene studies. Additionally, relatively high concentrations of
TCA are required to cause developmental toxicity compared with the concentration expected to
result from metabolism of tetrachloroethylene in vivo, which may account for the differences in
the type of developmental effects resulting from tetrachloroethylene exposure. There is also a
lack of information on the availability of metabolized TCA to the developing fetus and the
potential differences related to oral-versus-inhalation exposure in tetrachloroethylene studies.
4.8. GENOTOXICITY
Tetrachloroethylene and its metabolites have been extensively studied for genotoxic
activity in a variety of in vitro assay systems such as bacteria, yeast, and mammalian cells [Refer
to reviews by IARC (1995). WHO (2006). and ATSDR (1997a)1. This section discusses the
genotoxic potential of tetrachloroethylene and its known or postulated metabolites (TCA, DC A,
CH, TCVC, TCVG, NAcTCVC, tetrachloroethylene epoxide), with a summary provided at the
end of each section for tetrachloroethylene or its metabolite for their mutagenic potential, in
addition to an overall synthesis summary at the end of this section. TCVC sulfoxide does not
appear to have been investigated for genotoxicity.
The application of genotoxicity data to predict potential carcinogenicity is based on the
principle that genetic alterations are found in all cancers. Genotoxicity is the ability of chemicals
to alter genetic material in a manner that permits changes to be transmitted during cell division.
Although most tests for mutagenicity detect changes in DNA or chromosomes, some specific
4-359
-------
modifications of the epigenome, which includes proteins associated with DNA or RNA, can also
cause transmissible changes. Genetic alterations can occur through a variety of mechanisms
including gene mutations, deletions, translocations, or amplifications; evidence of mutagenesis
provides mechanistic support for the inference of potential for carcinogenicity in humans.
Evaluation of genotoxicity data entails a weight-of-evidence approach that includes
consideration of the various types of genetic damage that can occur. In acknowledging that
genotoxicity tests are, by design, complementary evaluations of different mechanisms of
genotoxicity, a recent IPCS publication (Eastmond et al., 2009) notes that "multiple negative
results may not be sufficient to remove concern for mutagenicity raised by a clear positive result
in a single mutagenicity assay." These considerations inform the present approach. In addition,
consistent with EPA's Guidelines on Carcinogenic Risk Assessment and Supplemental Guidance
for Assessing Susceptibility from Early-Life Exposure to Carcinogens (U.S. EPA, 2005a), the
approach does not address relative potency (e.g., among tetrachloroethylene metabolites, or of
such metabolites with other known genotoxic carcinogens) per se, nor does it consider
quantitative issues related to the probable production of these metabolites in vivo. Instead, the
analysis of genetic toxicity data presented here focuses on the identification of a genotoxic
hazard of these metabolites; a quantitative analysis of tetrachloroethylene metabolism to reactive
intermediates, via PBPK modeling, is presented in Section 3.
Below, the genotoxicity data for tetrachloroethylene and its metabolites are briefly
reviewed, with detailed study information in the corresponding tables. The contributions of
these data are twofold. First, to the extent that these metabolites may be formed in the in vitro
and in vivo test systems for tetrachloroethylene, these data provide insight into what agent or
agents may contribute to the limited activity observed with tetrachloroethylene in these
genotoxicity assays. Second, because the in vitro systems do not necessarily fully recapitulate in
vivo metabolism, the demonstration of in vitro genotoxicity by the known in vivo metabolites
themselves provides information regarding the expected genotoxicity of tetrachloroethylene
following in vivo exposure.
4.8.1. Tetrachloroethylene (PCE)
Limited studies have been performed examining tetrachloroethylene genotoxicity in vivo.
These and in vitro genotoxicity studies of tetrachloroethylene are described below and listed in
Tables 4-39 and 4-40.
4-360
-------
4.8.1.1. Mammalian Systems (Including Human Studies)
4.8.1.1.1. Gene mutation
Tetrachloroethylene was negative for increased frequency of mutations of thymidine
kinase locus in L5178Y/TK +/- mouse lymphoma cells both with and without S9 activation
(F344 rat liver) (NTP, 1986). Experiments were performed twice, with replicates of all doses.
L5178Y/TK +/- mouse lymphoma cells were exposed to tetrachloroethylene in 1%
dimethylsufoxide for 4 hours at 37°C in medium; cells were then washed and resuspended in
fresh medium for 48 hours at 37°C. TK mutation frequency was determined by plating cells in
medium supplemented with trifluorothymidine. Overall cell viability was determined by plating
cells in nonselective medium. Mutation frequency was not above background for any dose
tested (6.25, 12.50, 25, 50, 100 nL/mL in the presence of S9; 12.5, 25, 50, 75, and 150 nL/mL in
the absence of S9). Positive controls in both the presence and absence of S9 activation
[3-methylcholanthrene (2.5 ug/mL) and ethyl methanesulfonate (250 ug/mL), respectively]
showed significant increases in mutation frequencies (p < 0.001, Mest) (NTP, 1986).
Gene mutations were induced in a host-mediated assay, using S. typhimurium strain
TA98 implanted into the peritoneal cavity of male and female CD-I mice that were previously
exposed to tetrachloroethylene by inhalation (100 or 500 ppm, 7 hours/day, for 5 days) (Beliles
et al., 1980). Positive results were observed in male mice at 100 (but not 500) ppm, and in female
mice at 500 (but not 100) ppm. Although no explanation was given for the variability in the dose
response, the authors conclude that tetrachloroethylene is an active frameshift mutagen using in
vivo activation.
In summary, the in vitro thymidine kinase gene mutation assay in mammalian cells was
negative for gene mutations in the presence and absence of S9 (F344 rat liver) metabolic
activation (NTP, 1986). Positive results for frameshift mutagenicity were observed in a host-
mediated assay by implanting S. typhimurium into mice exposed to tetrachloroethylene, but
without a clear dose-response effect (Beliles et al., 1980).
4-361
-------
Table 4-39. Genotoxicity of tetrachloroethylen
vitro and in vivo)a
mammalian systems (in
Test system/endpoint
Unscheduled DNA synthesis, rat primary
hepatocytes in vitro
Unscheduled DNA synthesis, Osborne Mendel
rat primary hepatocytes in vitro
Unscheduled DNA synthesis, B6C3F! mouse
primary hepatocytes in vitro
Gene mutation, mouse lymphoma L5178Y cells,
tk locus
Sister chromatid exchange, Chinese hamster
ovary (CHO) cells in vitro
Chromosomal aberrations, Chinese hamster lung
(CHL) cells in vitro
Chromosomal aberrations, Chinese hamster
ovary (CHO) cells in vitro
Cell transformation, RLV/Fischer rat embryo
F1706 cells in vitro
BALB/C-3T3 mouse cells, cell transformation in
vitro
Rat and mouse hepatocyte, DNA damage
(unscheduled DNA synthesis)
Human fibroblast cells, DNA damage
(unscheduled DNA synthesis)
Host mediated assay — S. typhimurium implanted
in CD-I mice
Chinese hamster ovary cells, sister chromatid
exchange
Chinese hamster ovary (CHO-K1) cells,
increased frequency of micronuclei
Cy tochalasin B -blocked micronucleus AHH- 1
assay using human lymphoblastoid
cell lines with enhanced metabolic TT~F ,
activity, increased frequency of
micronuclei
MCL-5
Human white blood cells, length of DNA
migration
Human lymphocytes, sister chromatid exchange
Doses
(LED or HID)b
166 (vapor)
NA
NA
245
164
500
136
16
250
2.5mM
O.lnL/mL
100 ppm (male
mice;500 ppm
(female mice)
164 ug/mL
~63 ppm
5mM
ImM
ImM
5 x 10~3 M
Results0
With
activation
NT
NT
NT
-
-
-
-
NT
NT
NT
(+/-)
+
-
NT
NT
NT
NT
Without
activation
_d
-
-
-
-
-
-
+
-
—
(+/-)
NT
-
+
+
+
+
Reference
Shimada et al.
(1985)
Milman et al.
(1988)
Milman et al.
(1988)
NTP (1986)
Galloway et al.
(1987)
Sofuni et al.
(1985)
Galloway et al.
(1987)
Price et al. (1978)
Tu et al. (1985)
Costa and
Ivanetich (1984)
Beliles et al.
(1980)
Beliles et al.
(1980)
NTP (1986)
Wang et al.
(2001)
Doherty et al.
(1996)
Doherty et al.
(1996)
Doherty et al.
(1996)
Hartmann and
Speit (1995)
4-362
-------
Table 4-39. Genotoxicity of tetrachloroethylene—mammalian systems (in
vitro and in vivo)a (continued)
Test system/endpoint
Gene conversion and reverse mutation in
S. cerevisiae D7 recovered from liver, lungs, and
kidneys of CD-I mice
Gene conversion and reverse mutation in
S. cerevisiae D7 recovered from liver, lungs, and
kidneys of CD-I mice
DNA single-strand breaks (alkaline unwinding)
in liver and kidney of male NMRI mice in vivo
Sister chromatid exchange, human lymphocytes
in vivo
Chromosomal aberrations, human lymphocytes
in vivo
Binding (covalent) to calf thymus DNA in vitro
Binding (covalent) to DNA in male B6C3Fi
mouse liver in vivo
Binding (covalent) to DNA in male B6C3Fi
mouse liver in vivo
Binding (covalent) to DNA in male BALB/c
mouse and Wistar rat liver, kidney, lung, and
stomach in vivo
Binding (covalent) to RNA and protein in male
BALB/c mouse and Wistar rat liver, kidney,
lung, and stomach in vivo
Human lymphocytes, sister chromatid exchange
Mouse, reticulocytes, micronucleus
Mouse, hepatocytes, micronucleus
Before partial hepatectomy
After partial hepatectomy
Mouse, induction of DNA damage in hepatocytes
(alkaline Comet assay)
Mouse, induction of DNA damage in kidney
(alkaline Comet assay)
Rat bone marrow cells, chromosomal aberrations
Doses
(LED or HID)b
11,000 p.o. x 1
2,000 p.o. x 12
660 i.p. x 1
1,500 mg/m3
inhaled
92 ppm inhaled
2.5 uCi 14C-PCE
1,400 inhaled 6 h
600 ppm
500 p.o. x 1
1.4 i.p. x l
22 h
1.4 i.p. x l
22 h
10 ppm (geometric
mean)
2,000 mg/kg
1,000 mg/kg
1,000 mg/kg-day
2,000 mg/kg-day
1,000 mg/kg-day
2,000 mg/kg-day
100 and 500 ppm
Results0
With
activation
NT
NT
NT
NT
+
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
Without
activation
NT
+e
—
—
Data not
shown
-
-
+
+
-
+
+/-
+/-
_
-
Reference
Bronzetti et al.
(1983)
Bronzetti et al.
(1983)
Walles (1986)
Ikeda et al.
(1980)
Ikeda et al.
(1980)
Mazzullo et al.
(1987)
Schumann et
al. (1980)
Schumann et
al. (1980)
Mazzullo et al.
(1987)
Mazzullo et al.
(1987)
Seiji et al.
(1990)
Murakami and
Horikawa
(1995)
Murakami and
Horikawa
(1995)
Cederberg et al.
(2010a)
Cederberg et al.
(2010a)
Beliles et al.
(1980)
4-363
-------
Table 4-39. Genotoxicity of tetrachloroethylene—mammalian systems (in
vitro and in vivo)a (continued)
Test system/endpoint
Enzyme-altered foci in male Osborne Mendel rat
liver in vivo, promotion protocol, with or without
N-nitrosodiethylamine as an initiator
Enzyme-altered foci in male Osborne Mendel rat
liver in vivo, initiation protocol, phenobarbital as
a promoter
Micronucleus induction (Chinese hamster lung
cell line)
Gap Junction Intercellular Communication (rat
liver cells)
DNA damage (8-OHdG) in urine and leukocytes
of dry cleaners (female only)
DNA damage (8-OHdG) in Fischer rats
measured in urine, lymphocytes, and liver
Human lymphocytes in vitro
(unscheduled DNA synthesis)
Human lymphocytes in vivo
(Chromosomal aberrations)
DNA single-strand breaks
Doses
(LED or HID)b
1,000, 5d/wkfor7
wk
1,000
250 ug/mL
0.1 mM
3.8±5.3ppm
(TWA)
100-1,000 mg/kg
ImM
144 mg/m3 (but
contaminated with
trichloroethylene)
1,000 mg/kg p.o.
Results0
With
activation
NT
NT
-
NT
NT
NT
-
NT
NT
With
activation
+
-
+
-
(Substantial
morbidity at
all doses
limits
interpretation.)
-
+
-
Reference
Milman et al.
(1988)
Milman et al.
(1988)
Matsushima et
al. (1999)
Benane et al.
(1996)
Toraason et al.
(2003)
Toraason et al.
(1999)
Perocco et al.
(1983)
Fender (1993)
Potter et al.
(1996)
aTable adapted from ATSDR (1997a) and I ARC monograph (1995) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests; mg/kg for in vivo
tests unless otherwise specified; i.p. = intraperitoneal; p.o. = oral; NA = not available.
'Results: + = positive; (+) = weakly positive; (+/-) = mixed results; - = negative; NT = not tested.
dPCE with stabilizers was positive with and without metabolic activation.
eNegative in lung.
4-364
-------
Table 4-40. Genotoxicity of tetrachloroethylene—bacterial, yeast, and fungal
systems3
Test system/endpoint
SOS chromotest, E. coli PQ37
SOS chromotest, E. coli PQ37
A Prophage induction, E. coli WP2
S. typhimurium BAL13, forward mutation
(am test)
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1537, reverse mutation
S. typhimurium TA1537, reverse mutation
S. typhimurium, gene mutation TA100,
TA1535, TA1537, TA98
S. typhimurium TA98, reverse mutation
S. typhimurium TA98, reverse mutation
S. typhimurium TA98, reverse mutation
S. typhimurium UTH8413, reverse
mutation
S. typhimurium UTH8414, reverse
mutation
S. typhimurium TA102, TA2638
E. coli WP2/pKM101, WP2
wvrA/pKMlOl, gene mutation
Doses
(LEDorHID)b
8,150
NA
10,000
76
660
167
1,000
166 (vapor)
NA
332
1.3 (vapor)
50
167
66 (vapor)
NA
167
NA
333 ug/plate
167
1,000
NA
1,000
1,000
1,250 ug/plate
Results0
With
activation
-
—
-
—
-
-
-
-
-
+e
-
NT
-
(+)
-
-
-
—
-
-
-
-
-
Without
activation
-
—
-
—
-
-
-
_d
-
-
-
-
-
_d
-
-
-
—
-
-
-
-
-
NT
Reference
Mersch-Sundermann et
al. (1994)
von der Hude et al.
(1988)
DeMarini et al. (1994)
Roldan-Arjona et al.
(1991)
Bartsch et al. (1979)
Haworth et al. (1983)
Connor et al. (1985)
Shimada et al. (1985)
Milman et al. (1988)
Vamvakas et al.
(1989d)
DeMarini et al. (1994)
Kringstad et al. (1981)
Haworth et al. (1983)
Shimada et al. (1985)
Milman et al. (1988)
Haworth et al. (1983)
Milman et al. (1988)
NTP (1986)
Haworth et al. (1983)
Connor et al. (1985)
Milman et al. (1988)
Connor et al. (1985)
Connor et al. (1985)
Watanabe et al. (1998)
4-365
-------
Table 4-40. Genotoxicity of tetrachloroethylene—bacterial, yeast, and
fungal systems3 (continued)
Test system/endpoint
S. typhimurium, YGVlOSpinSERbs, gene
mutation (strain is methyltransferase
deficient and stably expresses complete
electron transport chain including P450
reductase, cytochrome b5 and CYP2E1)
E. coli K12, forward mutation
E. coli K12, reverse mutation (arg*)
E. coli K12, reverse mutation (gal*)
E. coli K12, reverse mutation (nad*)
S. cerevisiae D7, log-phase cultures, gene
conversion
S. cerevisiae D7, gene conversion
S. cerevisiae D7, log-phase and stationary
cultures, gene conversion
S. cerevisiae D7, log-phase cultures,
mitotic recombination or other genetic
alterations (ade2)
S. cerevisiae D7, mitotic recombination
S. cerevisiae D7, log-phase cultures,
reverse mutation
S. cerevisiae D7, reverse mutation
S. cerevisiae D7, log-phase and stationary
cultures, reverse mutation
S. cerevisiae D61.M, growing cells,
aneuploidy
D. melanogaster , sex-linked recessive
lethal mutation
D. melanogaster, sex-linked recessive
lethal mutation
Doses
(LED or HID)b
200 ug/plate
150
150
150
150
1,100
9,960
2,440
1,100
9,960
810
9,960
2,440
810
4,000 ppm p.o.
1,000 ppm injection
3,400 mg/m3, 7 h
Results0
With
activation
NT
-
-
-
-
NT
-
—
NT
—
NT
—
—
(+)
NT
NT
Without
activation
-
-
-
-
+
-
—
+
—
(+)
—
—
(+)
—
—
Reference
Emmert et al. (2006)
Greim et al. (1975)
Greim et al. (1975)
Greim et al. (1975)
Greim et al. (1975)
Callenetal. (1980)
Bronzetti et al.
(1983)
Koch etal. (1988)
Callenetal. (1980)
Bronzetti et al.
(1983)
Callenetal. (1980)
Bronzetti et al.
(1983)
Koch et al. (1988)
Koch et al. (1988)
NTP (1986)
Beliles et al. (1980)
aTable adapted from ATSDR (1997a) and IARC monograph (1995) and modified/updated for newer
references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests unless
otherwise specified; NA = not available.
'Results: + = positive; (+) = weakly positive; - = negative; NT = not tested.
dPCE with stabilizers was positive with and without metabolic activation.
eWeak increase in activity with rat liver S9, rat kidney microsomes and glutathione (GSH): fourfold
increase with rat kidney microsomes, GSH and GSH S-transferase.
4-366
-------
4.8.1.1.2. DNA binding
Schumann et al. (1980) assessed hepatic macromolecular binding in both rats and mice
exposed to radiolabeled tetrachloroethylene by inhalation (10 or 600 ppm, 6 hours; binding
measured at 6, 24, 48, and 72 hours postexposure) or a single oral gavage (500 mg/kg in corn oil;
binding measured at 1, 6, 12, 24, 48, and 72 hours). In mice, tetrachloroethylene binding to
macromolecules in liver peaked at the termination of the inhalation exposure or 6 hours postoral
exposure. In rats, hepatic macromolecular binding peaked 24 hours after either oral or inhalation
exposure. At these peak times, no DNA binding was observed in the mouse (rat data not
reported). Using a more sensitive assay, Mazzullo et al. (1987) reported low levels of DNA
binding (2.9 pmol/mg) in mouse liver 22 hours after i.p. injection (1.4 mg/kg bw). Levels of
DNA binding were 6- to 10-fold lower in rat liver and in the kidney, lung, and stomach of mice
and rats. Binding to RNA or protein was considerably higher than binding to DNA in both mice
and rats. This raises the concern that possible contamination with RNA or protein might have
contributed to the DNA results. Protein binding levels were highest in mouse liver and rat
kidney. In a companion in vitro study, binding to calf thymus DNA was increased by
microsomal fractions from rat or mouse liver, but not kidney, lung, or stomach. Cytosolic
fractions from rat or mouse liver, kidney, lung, or stomach also enhanced DNA binding in vitro,
with mouse and rat liver and mouse lung fractions being the most efficient. Cytosolic and
microsomal fractions, when combined, enhanced DNA binding to a comparable extent as
cytosolic fractions alone. Phenobarbital pretreatment of animals increased cytosol-mediated
binding but minimally affected microsomal-mediated binding. DNA binding by rat liver
microsomal fraction was enhanced 17-fold by GSH but decreased by superoxide dismutase or
mannitol (Mazzullo et al., 1987).
In summary, DNA binding was not observed in one assay in mice exposed to
tetrachloroethylene by inhalation and oral routes, while protein and RNA binding was observed
(Schumann et al., 1980). Low levels of DNA binding in mouse liver, and yet lower levels in
mouse kidney or rat and mouse stomach, were observed after i.p. injection using a more sensitive
assay (Mazzullo et al., 1987). In vitro binding to calf thymus DNA was enhanced by
microsomal and cytosolic fractions from various mouse and rat tissues. These results suggest a
role for metabolic activation of the parent compound in DNA binding in vitro.
4.8.1.1.3. Chromosomal aberrations
Beliles et al. (1980) assessed bone marrow chromosomal aberrations and aneuploidy in
male and female Sprague-Dawley rats after acute (sacrificed 6, 24, or 48 hours after dosing) and
subchronic (7 hours a day, for 5 days; sacrificed 6 hours after last exposure) exposures to
tetrachloroethylene by inhalation (100 and 500 ppm). The only effect reported with acute
4-367
-------
exposure was a slight increase in the percentage of cells with aberrations and aneuploidy (peak
of 3.3% compared to 0.7% in controls with 500 ppm tetrachloroethylene) in male, but not
female, rats. No significant effects were observed in any subchronically exposed groups, but
female rats showed a nonsignificant increase in cells with aberrations (Beliles et al., 1980). NTP
(1986) did not observe chromosomal aberrations in Chinese hamster ovary cells exposed to
tetrachloroethylene (17, 34.1, 68.1, and 136.3 ug/mL without activation or 17, 34.1, and
68.1 ug/mL with activation by Sprague-Dawley rat liver S9).
4.8.1.1.3.1. Micronucleus induction
Tetrachloroethylene exposure increased the frequency of micronuclei in hepatocytes, but
not peripheral blood reticulocytes, of ddY mice given single i.p. injections of 1,000 or 2,000
mg/kg tetrachloroethylene after, but not prior to, partial hepatectomy (Murakami and Horikawa,
1995). This twofold increase in micronuclei in hepatocytes after partial hepatectomy was
statistically significant but was not evident at the lower dose of 500 mg/kg. Conflicting results
of other studies of tetrachloroethylene micronuclei induction have also been reported in cultured
Chinese hamster cells (Wang et al., 2001; Matsushima et al., 1999) and in human cells (White et
al., 2001; Doherty et al., 1996). Micronucleus induction was not observed in a Chinese hamster
lung cell line (CHL/IU) following exposure to high doses of tetrachloroethylene
(125-250 ug/mL) as part of a test validation assay, but some induction (not statistically
significant) was observed at the lower dose (75 ug/mL) in the presence of S9 fraction
(Matsushima et al., 1999). Details from this study are limited. Wang et al. (2001) examined
micronuclei induction following in vitro exposure to tetrachloroethylene (-63 ppm in culture
medium at peak) in a closed system. Chinese hamster ovary (CHO-K1) cells were plated in a
petri dish surrounding a glass dish of tetrachloroethylene and incubated for 24 hours.
Tetrachloroethylene exposure led to a dose-dependent significant increase in micronuclei
induction (p < 0.001) (Wang et al., 2001). Similar results were also observed in human cell lines
in other studies.
Micronucleus induction was enhanced by tetrachloroethylene exposure in AHH-1
parental human lymphoblastoid cells, and in two daughter cell lines (h2El and MCL-5) stably
expressing human metabolic enzymes lines (Doherty et al., 1996). Parental AHH-1 cells possess
native, albeit low, CYP1A1 activity but considerable glutathione-S-transferase activity; h2El
cells stably express human CYP2E1; and MCL-5 cells stably express human CYP1A2, 2A6,
3A4, 2E1, and microsomal epoxide hydrolase. Tetrachloroethylene (5 mM) induced a threefold
increase in micronuclei in AHH-1 cells and ninefold increases in h2El and MCL-5 cells,
respectively (Doherty et al., 1996). White et al. (2001) similarly observed dose-dependent
4-368
-------
increases in micronuclei induction after 24 hours incubation (p < 0.05) with tetrachloroethylene
(0, 0.01, 0.05, 0.1, 0.25, 0.5, 1.0, 2.0 mM) in theMCL-5 cell line.
4.8.1.1.3.2. Sister chromatid exchanges (SCEs)
Limited studies of sister chromatid exchanges demonstrate conflicting results. No
differences were observed in the frequency of chromosomal aberrations and SCE between
unexposed workers, workers exposed to moderate levels of tetrachloroethylene (70-280mg/m3),
and those exposed to high doses (200-1,500 mg/m3) (Ikeda et al., 1980). Although an exposure
assessment was performed in this study, the results are limited by the small number of subjects
(total n= 19). Another study from this group had similar limitations (total n = 10) and also
found no sister chromatid exchanges in lymphocytes in workers occupationally exposed to either
high-dose tetrachloroethylene (92 ppm, geometric mean) or low-dose tetrachloroethylene (10-40
ppm range) (Ikeda et al., 1980). Similarly, no differences were observed between exposed and
controls in a larger Japanese study, which examined SCE in 27 occupationally exposed workers
(Seiji etal., 1990), or a German study on dry-cleaning workers (Bottger and Elstermeier, 1989).
Increased chromosomal aberrations were observed in another occupational study following
exposure to tetrachloroethylene (144-348 mg/m3); however, exposure also included a small
amount of trichloroethylene (0.11-0.43% by wt), so interpretation of the results relative to
tetrachloroethylene alone may be limited (Fender, 1993)
Tetrachloroethylene-induced damage was also not observed in the sister chromatid
exchange (SCE) assay or in the single-cell gel test (i.e., the Comet assay) in cultured human
blood exposed to up to 5 mM (-830 mg/L) tetrachloroethylene, a dose that reduced viability by
40% due to cytotoxicity (Hartmann and Speit 1995). Neither chromosome aberrations nor SCE
were induced in Chinese hamster ovary cells following in vitro exposure to tetrachloroethylene
(Galloway et al., 1987; Sofuni et al., 1985) as summarized in NRC (2010). Chinese hamster
ovary cells exposed to tetrachloroethylene (16.4, 54.5, or 164 ug/mL) in the presence and
absence of S9 activation (Sprague-Dawley rat livers) showed no increase in frequency of sister
chromatid exchanges following exposure to tetrachloroethylene (NTP, 1986).
In summary, the majority of studies of chromosomal aberrations, micronuclei induction,
and sister chromatid exchange following exposure to tetrachloroethylene are negative. Positive
micronuclei induction was observed following partial hepatectomy at high doses
(2,000 mg/kg-day i.p.) in ddY mice (Murakami and Horikawa, 1995). Increased micronuclei
induction was observed in CHO cells in vitro when exposed to tetrachloroethylene in a closed
system (Wang et al., 2001) but not in CHL cells when exposed in an open system (Matsushima
et al., 1999), suggesting the need to control for loss of tetrachloroethylene via vaporization in in
vitro assays. Dose-dependent increases in micronuclei were observed in human lymphoblastoid
4-369
-------
cell lines, an effect enhanced by stable expression of CYP450 enzymes (White etal., 2001;
Doherty et al., 1996): however, these cell lines are not generally considered part of the standard
genotoxicity testing battery. No in vitro studies of tetrachloroethylene (Hartmann and Speit
1995; Galloway et al., 1987: NTP, 1986) and only one occupational exposure study of exposures
to tetrachloroethylene and trichloroethylene (Fender, 1993) reported sister chromatid exchanges.
4.8.1.1.4. Unscheduled DNA synthesis
Human fibroblasts (WI-38 cells) were assayed for unscheduled DNA synthesis following
exposure to tetrachloroethylene (0.1 to 5.0 uL/mL), but the results were equivocal, with results at
low doses similar to the positive controls and negative results at high doses, but it is noted that
the high doses yielded considerable cytotoxicity (Beliles et al., 1980). The positive controls
were only weakly positive, as described based on the laboratory criteria (criteria details not
given). No evidence of unscheduled DNA synthesis was observed in human lymphocytes,
human fibroblasts, or rat and mouse hepatocytes (Milman etal., 1988: Shimada et al., 1985:
Costa and Ivanetich, 1984: Perocco et al., 1983). In summary, UDS was not statistically
significantly increased in any published studies, although some increases were observed in one
study (Beliles et al.. 1980).
4.8.1.1.5. DNA strand breaks
An increased level of DNA single-strand breaks (SSB), as assessed by a DNA unwinding
technique, was observed in liver and kidney tissues but not in the lung tissue of male NMRI mice
1 hour after single i.p. injections in Tween 80 of 4-8 mmol/kg (663-1,326 mg/kg) of
tetrachloroethylene (Walles, 1986). This effect was reversible as early as 24 hours postexposure,
presumably by DNA repair. Limitations of i.p. injection include the potential inflammatory
effect at the site of injection, which could, in turn, lead to production of reactive oxygen species
and other inflammatory mediators. These could lead to an increase in DNA damage unrelated to
the specific exposure. Potter et al. (1996) found no increases in DNA strand breaks, when
assessed by an alkaline unwinding procedure, in kidneys of male F344 rats assessed after daily
oral gavage treatment with 1,000 mg/kg tetrachloroethylene for 7 days. A more recent study
(Cederberg et al., 2010a) found oral gavage exposure to tetrachloroethylene (1,000 or
2,000 mg/kg-day given as two administrations, 24 hours apart, in corn oil) led to slight increases
(1.3- and 1.4-fold as compared to control) in DNA damage in liver (but not kidney) of CD1 mice
as measured by the alkaline Comet assay when tissues were sampled 3 hours after the last
administration. Others have interpreted these data to demonstrate a lack of DNA damage in the
liver and kidney of CD1 mice after oral tetrachloroethylene exposure [presented in Dreessen
(2003) and in Lillford (2010)]. Cederberg et al. (2010a) reported a statistically significant dose-
related increase in tail intensity (p = 0.041; one-sided Jonckheere-Terpstra test using exact
4-370
-------
permutation) in the liver following exposure to PCE. The authors note that 8 of
12 tetrachloroethylene-exposed animals had higher tail intensity values than the highest value in
the controls, a finding significant by the Fisher exact probability test (p = 0.013). No statistically
significant effects were observed for tail moment in the liver, or for either tail intensity or
moment in the kidney. The alternative interpretation is that the variability between mice in the
treatment groups and the low magnitude of the response in the tetrachloroethylene-dosed animals
does not support the conclusion that tetrachloroethylene induced DNA damage in this study.
This interpretation is supported by the lack of statistical significance when the results are
analyzed by Dunnett's test for pairwise comparisons. Cederberg et al. (2010a) argue that the
interindividual animal variability is not exceptionally large, and that the Dunnett's pairwise test
has less power than the trend test of Jonckheere-Terpstra (Cederberg et al., 2010a). Further
discussion of this publication in the literature is ongoing (Lillford et al., 2010; Lovell 2010).
Lillford et al. (2010) give additional details on the alternative interpretation described in the
original paper, stating also that the limited biological significance of these slight increases in tail
intensity needs to be taken into account. This paper states that the results described in the
original study are within the range of historical controls in the study laboratory. Lillford et al.
(2010) endorse the use of the parametric test for statistical analysis (Dunnett's), which showed
no statistical significance for the results reported in Cederberg. The third publication discusses
the use of various statistical analyses used in the two interpretations (Lovell, 2010). Overall,
Lovell (2010) states that it is not a question of one statistical analysis being right and the other
wrong; it is more a question of using the best statistical analysis for the hypothesis being tested.
The different approaches show a contrast between a powerful trend test and a more conservative
pairwise comparison. Lovell (2010) also commented on the magnitude of the response as it
relates to biological relevance. Further studies, as suggested by Cederberg et al. (201 Ob), may or
may not address this issue if carried out the same way as the original study. Finally, both
Lillford et al. (2010) and Lovell (2010) agree that the statistical analysis utilized should not be
used as the sole determinant of how the results of this, or any study, are interpreted.
In summary, the results of the limited DNA strand break assays following exposure to
tetrachloroethylene are equivocal. Walles (1986) demonstrated DNA single-strand breaks in the
liver and kidney of male mice exposed by i.p. injection, but this was reversible within 24 hours.
A second study examined DNA strand breaks after 1 week oral exposure to tetrachloroethylene
and demonstrated no DNA damage (Potter et al., 1996). A recently published report on DNA
strand breaks showed a marginal increase in only one parameter from the Comet assay (tail
length) following oral exposure to tetrachloroethylene in mice (Cederberg et al., 2010a), but the
statistical and biological significance of this result has been disputed (Cederberg et al., 2010b:
Lillford et al.. 2010: Lovell 2010).
4-371
-------
4.8.1.1.6. DNA damage related to oxidative stress
Toraason et al. (2003) reported no increase in leukocyte 8-OHdG in 18 dry-cleaner
workers compared with 20 launderers, and reported no increase in urinary 8-OHdG among the
dry-cleaner workers sampled pre- and postshift work (time-weighted average [TWA]
concentration of tetrachloroethylene was 3.8 ± 5.3 ppm). Under the conditions of this study, no
evidence of oxidative DNA damage was found. Toraason et al. (1999) measured 8-OHdG and a
"free radical-catalyzed isomer of arachidonic acid and marker of oxidative damage to cell
membranes, 8-Epi-prostaglandin F2a (8-epiPGF)," excretion in the urine, and TEARS (as an
assessment of malondialdehyde and marker of lipid peroxidation) in the liver and kidney of male
Fischer rats exposed to single i.p. injections of tetrachloroethylene in Alkamuls vehicle. Male
Fischer rats sacrificed 24 hours after a single i.p. injection of tetrachloroethylene (0, 100, 500, or
1,000 mg/kg) showed no significant increases in 8-OHdG in liver, lymphocytes, or urine
(Toraason et al., 1999). Lipid peroxidation of the liver (as measured by TEARS) was also not
observed following a single exposure to tetrachloroethylene. However, the authors reported
morbidity and mortality with a single 500 mg/kg tetrachloroethylene exposure inducing Stage II
anesthesia (loss of righting reflex but maintained reflex response) and a single 1,000 mg/kg
tetrachloroethylene exposure inducing Level III or IV (absence of reflex response) anesthesia
and burgundy-colored urine during the first 12 hours of collection. Although none of the rats
exposed to 1,000 mg/kg tetrachloroethylene died from treatment, the authors state that some in
this high-dose group would not have survived another 24 hours. Thus, using this paradigm, there
was significant toxicity and additional issues related to route of exposure. Urine volume
declined significantly during the first 12 hours of treatment, and while water consumption was
not measured, it was suggested by the authors to be decreased due to the moribundity of the rats.
Although the authors suggest that evidence of oxidative damage was equivocal, the effects on
urine volume and water consumption, as well as the profound toxicity induced by this exposure
paradigm, limit interpretation of these data. In summary, the limited studies examining DNA
adduct formation related to oxidative stress are inconclusive, with no results in the urine or
leukocytes of occupationally exposed individuals and limited utility of the animal study due to
significant toxicity in the exposed animals.
4.8.1.1.7. Cell transformation
Tetrachloroethylene exposure did not lead to cell transformation in BALB/C-3T3 cells
after 3-day exposure (0, 1, 10, 100, and 250 ug/mL) followed by a 30-day incubation period (Tu
et al., 1985). Exposure to tetrachloroethylene (study details not given) was also negative for cell
transformation in BALB/C-3T3 cells (Milman et al., 1988). However, Fischer rat embryo cells
were transformed in the absence of metabolic activation (Price et al., 1978).
4-372
-------
4.8.1.1.8. Gap junction intercellular communication
One assay examined gap junction intercellular communication following exposure to
tetrachloroethylene in rat liver cells (0, 0.01, 0.1, and 1 mM at 0, 1, 4, 6, 24, 48, and 168 hours)
(Benane et al., 1996). Communication was inhibited following exposure to 0.1 -mM
tetrachloroethylene at 48 hours and continued at the final time point tested (168 hours). This
study also examined tetrachloroethylene metabolites, including DCA, TCA, CH, and
trichloroethanol. These metabolites also led to decreases in intercellular communication, but to
varying levels.
4.8.1.1.9. Tumor initiation
Milman et al. (1988) reported a statistically significant increase (p < 0.01) in
y-glutamyltranspeptidase-positive liver foci in a promotion, but not in an initiation, test protocol
in male Osborne-Mendel rats. Initiation capacity was tested by exposing 10 rats to 1,000 mg/kg
tetrachloroethylene after partial hepatectomy, followed by phenobarbital promotion for 7 weeks.
In the promotion test, rats were initiated with DEN after partial hepatectomy, followed by
promotion with tetrachloroethylene for 7 weeks. In a separate initiation study of neonatal female
Wistar rats exposed to 2,000 ppm, 8 hours/day, 5 days/week, for 10 weeks [described in Bolt et
al. (1982), as reported in NRC (2010)1, preneoplastic liver foci were reportedly not observed.
4.8.1.2. Drosophila Melanogaster
Limited tetrachloroethylene genotoxicity studies have been performed in Drosophila
melanogaster. One study was negative for both the induction of sex-linked recessive lethal
mutations and chromosomal aberrations following inhalation exposure to tetrachloroethylene in
D. melanogaster (up to 3,400 mg/m3 for 7 hours) (Beliles et al., 1980). The frequencies of the
sex-linked recessive lethal mutations were 0 and 0.10% for the low- and high-dose exposures,
respectively, which was not significantly different from the negative control (0.11%). This study
also showed no chromosomal aberrations, as there were no significant loses of the long arm of
the Y chromosome for either the low (0.11%) or high (0.02%) doses as compare to the negative
control (0.02%). A second study, also negative for sex-linked recessive lethal mutations,
exposed male Drosophila by feeding tetrachloroethylene (4,000 ppm) or by injection (1,000
ppm) before successive mating with untreated females for 3 days (NTP, 1986; Valencia et al.,
1985). Fl heterozygous daughters were mated to their siblings. Analysis of the data after 17
days demonstrated no significant increase in sex-linked recessive lethal mutations following
exposure to tetrachloroethylene.
4-373
-------
4.8.1.3. Bacterial and Fungal Systems
Cells of Saccharomyces cerevisiae contain cytochrome P450 monooxygenase system and
are capable of metabolizing promutagens to genetically active products. Tetrachloroethylene
alone was positive for mitotic recombination in yeast following 1 hour exposure to 6.6 mM
tetrachloroethylene (Callen et al., 1980) but negative in yeast exposed in suspension with
metabolic activation or in the intrasinguineous hose-mediated assay (Koch et al., 1988; Bronzetti
et al., 1983). Results were negative in the same assay for tetrachloroethylene, but the high level
of cytotoxicity in this assay at the dose used (9.8 mM) limits the interpretation of these results
(Koch etal., 1988). Bronzetti et al. (1983) also demonstrated negative results both in vitro (0, 5,
10, 20, 60, and 85 mM) with and without S9 activation. There also appeared to be high
cytotoxicity in yeast cells exposed to high dose tetrachloroethylene based on decreasing
percentage survival in this study, which may also limit the interpretation of these data.
A number of in vitro genotoxicity assays have been performed using prokaryotic cells.
Studies of mutagenicity on Escherichia coli have been negative (Greim et al., 1975; also reported
in Henschler, 1977). Most Ames tests using S. typhimurium have indicated that
tetrachloroethylene in the absence of metabolic activation or in the presence of the standard S9
fraction is not a mutagen (Watanabe et al., 1998; DeMarini et al., 1994; Roldan-Arjona et al.,
1991: Milmanetal., 1988: Warner etal.. 1988: NTP, 1986: Connor etal.. 1985: Shimada et al..
1985: Haworth et al.. 1983: Hardinetal., 1981: Kringstad et al., 1981: Bartsch et al.. 1979:
Greim etal., 1975). However when incubated with rat liver GST, GSH, and a rat kidney
fraction, tetrachloroethylene exhibited a clear dose response (Vamvakas et al., 1989d).
Specifically, this study demonstrated the mutagenicity in S. typhimurium (primarily strain
TA100) of tetrachloroethylene that had been preincubated with rat liver GST, GSH, and rat
kidney microsomes, and of TCVG that had been preincubated with rat kidney microsomes.
Additionally, the bacterial mutagenicity of bile from liver perfusate following
tetrachloroethylene exposure in rats was demonstrated (Vamvakas et al., 1989d). These results
support a role for GSH conjugation in the genotoxicity of tetrachloroethylene.
A more recent study examined genotoxicity of tetrachloroethylene in an S. typhimurium
strain (YG7108pin3ERb5) with enhanced metabolic activity (transformed with CYP2E1,
cytochrome P450 reductase, and cytochrome b5),which led to microcolony formation believed to
be from toxicity of tetrachloroethylene metabolites formed at 200 and 1,000 ug doses (but not at
the higher doses of 2,000 or 3,000 ug) (Emmert et al., 2006). Tetrachloroethylene was negative
in the parent strain (YG7108) at all doses in the presence of S9. These results support a role for
CYP2E1-derived metabolites in the toxicity of tetrachloroethylene, but not the mutagenicity of
tetrachl oroethy 1 ene.
4-374
-------
In summary, gene mutations were not observed following exposure to tetrachloroethylene
in E. coli or S. typhimurium cells in the absence of metabolic activation. Addition of standard S9
fraction also did not lead to mutagenicity, but exposure to bacterial cells with enhanced
metabolic activity (CYP2E1 GSH) led to positive Ames test results. These support a role of
metabolic activation of tetrachloroethylene in its genotoxicity. Results in yeast cells are
conflicting, with one positive study (Callen et al., 1980) and two negative studies (Koch et al.,
1988; Bronzetti et al., 1983). However, tetrachloroethylene led to cytotoxicity of S. cerevisiae at
the doses tested, making interpretation of these results difficult. These results, although limited,
suggest tetrachloroethylene exposure can lead to genotoxicity in the presence of appropriate
metabolic activation.
4.8.1.4. Summary
The in vitro thymidine kinase gene mutation assay in mammalian cells was negative in
the presence and absence of S9 (F344 rat liver) metabolic activation (NTP, 1986). Positive
results for frameshift mutation were observed in a host-mediated assay by implanting
S. typhimurium into mice exposed to tetrachloroethylene, but without a clear dose-response
effect (Bellies etal., 1980). Studies of mutagenicity on E. coli have been negative (Greim et al.,
1975) and also reported in Henschler (1977). A number of mutagenicity studies in S.
typhimurium indicate that, in the absence of metabolic activation or in the presence of the
standard S9 fraction, tetrachloroethylene is not a mutagen (Emmert et al., 2006; Watanabe et al.,
1998: DeMarini et al., 1994: Rol dan-Arjona etal., 1991: Milman et al., 1988: Warner et al.,
1988: NTP, 1986: Connor etal., 1985: Shimada et al., 1985: Haworth et al., 1983: Hardin et al.,
1981: Kringstad et al., 1981: Bartsch et al., 1979: Greim et al., 1975). However, when
tetrachloroethylene was activated with rat liver GST, GSH, and a rat kidney fraction,
tetrachloroethylene exhibited a clear dose-response (Vamvakas et al., 1989d). These findings
support a role of metabolic activation of tetrachloroethylene in its in vitro genotoxicity. Results
in yeast cells are conflicting, with one positive study (Callen et al., 1980) and two negative
studies (Koch etal., 1988: Bronzetti et al., 1983). However, tetrachloroethylene led to
cytotoxicity of S. cerevisiae at the doses tested, making interpretation of these results difficult.
These results, although limited, suggest tetrachloroethylene exposure can lead to genotoxicity in
the presence of appropriate metabolic activation.
DNA binding was not observed in one assay in mice exposed to tetrachloroethylene by
inhalation and oral routes, while protein and RNA binding was observed (Schumann et al.,
1980). With a more sensitive assay, low levels of DNA binding were observed in mouse liver,
and even lower levels in mouse kidney and rat and mouse stomach after i.p. injection exposure
(Mazzullo et al., 1987). In vitro binding to calf thymus DNA occurred in the presence of various
4-375
-------
microsomal fractions, as well as in the presence of cytosolic fractions from mice and rats. These
results suggest a role for metabolic activation of the parent compound in DNA binding.
The majority of studies of chromosomal aberrations, micronuclei induction, and sister
chromatid exchange following exposure to tetrachloroethylene are negative. Positive
micronuclei induction was observed following partial hepatectomy at high doses
(2,000 mg/kg-day) in mice (Murakami and Horikawa, 1995). Increased micronuclei induction
was observed in CHO cells in vitro when exposed to tetrachloroethylene in a closed system
(Wang et al., 2001) but not in CHL cells when exposed in an open system (Matsushima et al.,
1999). Dose-dependent increases were observed in human lymphoblastoid cell lines that were
enhanced by stable expression of CYP450 enzymes (White et al., 2001; Doherty et al., 1996).
Sister chromatid exchanges were not observed in any in vitro studies (Hartmann and Speit 1995;
Galloway et al., 1987; NTP, 1986) and were observed in only one occupational exposure study,
but exposures were contaminated with trichloroethylene, so the interpretation of these results is
limited (Fender. 1993).
Although some increases were observed in UDS following exposure, these were not
statistically significant (NTP, 1986). The results of DNA strand break assays following exposure
to tetrachloroethylene are equivocal. Walles (1986) demonstrated DNA single-strand breaks in
the liver and kidney of male mice exposed by i.p. injection, but this was reversible within 24
hours. A second study examined DNA strand breaks after 1 week oral exposure to
tetrachloroethylene, and demonstrated no DNA damage (Potter et al., 1996). A study of DNA
strand breaks showed a marginal increase in only one parameter from the alkaline Comet assay
(tail intensity) in the liver but not the kidney following oral exposure to tetrachloroethylene in
mice (Cederberg et al., 2010a), but the statistical and biological significance of this result has
been disputed (Cederberg et al.. 2010b: Lillfordetal.. 2010: Lovell. 2010).
Studies examining DNA adduct formation related to oxidative stress are inconclusive,
with no results in the urine or leukocytes of occupationally exposed individuals (Toraason et al.,
2003) and limited utility of the animal study due to significant toxicity in the exposed animals
(Toraason et al., 1999). Tumor initiation was not observed in Milman et al. (1988) or Bolt et al.
(1982), but the former study reported significant increases in liver foci in a tumor promotion
study. A study examining inhibition of gap junction intercellular communication was positive
(Benane et al., 1996). Negative results were found for a limited number of other genotoxicity
endpoints including cell transformations (Tu et al., 1985) and sex-linked recessive lethal
mutation assay in Drosophila [(NTP, 1986; Beliles et al., 1980): also reported in Valencia et al.
(1985)].
Overall, evidence from a number of different analyses with various genetic endpoints
indicates that tetrachloroethylene has the potential to induce damage to the structure of the
4-376
-------
chromosome in a number of targets but has little-to-no ability to induce mutation in bacterial
systems in the absence of metabolic activation or with the standard S9 fraction. However,
metabolic activation via GSH conjugation or cytochrome P450s yields positive results in
bacterial mutagenicity assays.
4.8.2. Trichloroacetic Acid (TCA)
The tetrachloroethylene metabolite TCA has been studied using a variety of genotoxicity
assays for its genotoxic potential [refer to International Agency for Research on Cancer (2004)
and U.S. EPA (20 lie) for additional information]. Evaluation of in vitro studies of TCA must
consider toxicity and acidification of medium resulting in precipitation of proteins, as TCA is
commonly used as a reagent to precipitate proteins. These studies are summarized in Tables
4-41 and 4-42.
4.8.2.1. Mammalian Systems (Including Human Studies)
4.8.2.1.1. Gene mutations
The mutagenicity of TCA has also been tested in cultured mammalian cells (refer to
Table 4-41). Harrington-Brock et al. (1998) examined the potential of TCA to induce mutations
in L5178Y/TK+/- -3.7.2C mouse lymphoma cells. In this study, mouse lymphoma cells were
incubated in a culture medium treated with TCA concentrations up to 2,150 |ig/mL in the
presence of S9 metabolic activation and up to 3,400 |ig/mL in the absence of S9 mixture. In the
presence of S9, a doubling of mutant frequency was observed at concentrations of 2,250 |ig/mL
and higher, including several concentrations with survival >10%. In the absence of S9, TCA
increased the mutant frequency by twofold or greater only at concentrations of 2,000 |ig/mL or
higher. These results were obtained at <11% survival rates. The authors noted that the mutants
included both large-colony and small-colony mutants. The small-colony mutants are indicative
of chromosomal damage. It should be noted that no rigorous statistical evaluation was
conducted on these data.
4-377
-------
Table 4-41. Genotoxicity of trichloroacetic acid (TCA)—mammalian systems
(in vitro and in vivo)a
Test system/endpoint
Gene mutation, mouse lymphoma L5178Y/TK+/-
cells, in vitro
DNA strand breaks, B6C3F! mouse and Fischer 344
rat hepatocytes, in vitro
DNA strand breaks, human CCRF-CEM
lymphoblastic cells, in vitro
DNA damage, Chinese hamster ovary cells, in vitro,
comet assay
DNA strand breaks, B6C3F! mouse liver, in vivo
DNA strand breaks, B6C3F! mouse liver, in vivo
DNA strand breaks, B6C3FJ mouse liver, in vivo
DNA strand breaks, B6C3FJ mouse liver and
epithelial cells from stomach and duodenum, in vivo
DNA strand breaks, male B6C3FJ mice, in vivo
DNA strand breaks, male B6C3F! mouse liver, in
vivo
Micronucleus formation, Swiss mice, in vivo
Micronucleus formation, female
C57BL/6JfBL10/Alpk mouse bone-marrow
erythrocytes, in vivo
Micronucleus formation, male C57BL/6JfBL10/Alpk
mouse bone-marrow erythrocytes, in vivo
Micronucleus formation, Pleumdeles waltl larvae
peripheral erythrocytes, in vivo
Chromosomal aberrations, Swiss mouse bone-marrow
cells in vivo
Chromosomal aberrations, Swiss mouse bone-marrow
cells in vivo
Chromosomal aberrations, Swiss mouse bone-marrow
cells in vivo
Chromosomal aberrations, chicken Callus domesticus
bone marrow, in vivo
Doses
(LED or
HID)b
3,000
1,630
1,630
3mM
1.0, P.O., xl
500, p.o., xl
500, p.o., 10
repeats
1,630, p.o., xl
500
(neutralized)
300, p.o.
125, i.p., x2
1,300, i.p., x2
1,080, i.p., x2
80
125, i.p., xl
100, i.p., x5
500, p.o., xl
200, i.p., xl
Results0
With
activation
(+)
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
Without
activation
7
-
-
-
+
+
-
-
—
+
+
—
+
+
+
+
+
Reference
Harrington-Brock
et al. (19981
Chang et al. (1992)
Chang et al. (1992)
Plewa et al. (2002)
Nelson and Bull
(1988)
Nelson et al.
(1989)
Nelson et al.
(1989)
Chang et al. (1992)
Styles et al. (1991)
Hassoun and Dey
(2008)
Bhunya and
Behera (1987)
Mackay et al.
(1995)
Mackay et al.
(1995)
Ciller et al. (1997)
Bhunya and
Behera (1987)
Bhunya and
Behera (1987)
Bhunya and
Behera (1987)
Bhunya and Jena
(1996)
4-378
-------
Table 4-41. Genotoxicity of trichloroacetic acid (TCA)—mammalian
systems (in vitro and in vivo)a (continued)
Test system/endpoint
Chromosomal aberrations, human lymphocytes, in
vitro
Sperm morphology, Swiss mouse, in vivo
Increased detection of M\G and 8-OHdG adducts,
B6C3Fi neonatal mouse liver DNA, in vivo
Doses
(LED or
HID)b
5,000
(neutralized)
125,i.p.,x5
2,000 nmol
Results0
With
activation
NT
NT
NT
Without
activation
—
+
+
Reference
Mackay et al.
(1995)
Bhunya and
Behera (1987)
Von Tungeln et al.
(2002)
aTable adapted from ATSDR (1997a) and IARC monograph (1995) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests; mg/kg for in vivo
tests unless specified.
'Results: + = positive; (+) = weakly positive; - = negative; NT = not tested; ? = inconclusive.
4.8.2.1.2. Chromosomal aberrations
Mackay et al. (1995) investigated the ability of TCA to induce chromosomal damage in an in
vitro chromosomal aberration assay using cultured human cells. The authors treated the cells
with TCA as free acid, both in the presence and absence of metabolic activation. TCA induced
chromosomal damage in cultured human peripheral lymphocytes at concentrations (2,000 and
3,500 |ig/mL) that significantly reduced the pH of the medium. However, exposure of cells to
neutralized TCA did not have any effect, even at a cytotoxic concentration of 5,000 |ig/mL. It is
possible that the reduced pH was responsible for the TCA-induced clastogenicity in this study.
To further evaluate the role of pH changes in the induction of chromosome damage, the authors
isolated liver-cell nuclei from B6C3Fi mice and suspended the isolates in a buffer at various pH
levels. The cells were stained with chromatin-reactive (fluorescein isothiocyanate) and
DNA-reactive (propidium iodide) fluorescent dyes. A decrease in chromatin staining intensity
was observed with the decrease in pH, suggesting that pH changes, independent of TCA
exposure, can alter chromatin conformation. It was concluded by the authors that TCA-induced
pH changes are likely to be responsible for the chromosomal damage induced by unneutralized
TCA. In another in vitro study, Plewa et al. (2002) evaluated the induction of DNA strand
breaks by TCA (1-25 mM) in CHO cells and did not observe any genotoxicity.
4.8.2.1.2.1. Micronucleus induction
Genotoxicity of TCA was tested in a mouse in vivo system using three different
cytogenetic assays (bone marrow chromosomal aberrations, micronucleus and sperm-head
4-379
-------
abnormalities) (Bhunya and Behera, 1987) and for chromosomal aberrations in chicken (Bhunya
and Jena, 1996). TC A induced a variety of anomalies including micronucleus in the bone
marrow of mice and chicken. A small increase in the frequency of micrenucleated erythrocytes
at 80 |ig/mL in a newt (Pleurodeles waltl larvae) micronucleus test was observed in response to
TCA exposure (Giller et al.. 1997). Mackay et al. (1995) investigated the ability of TCA to
induce chromosomal DNA damage in the in vivo bone-marrow micronucleus assay in mice.
C57BL mice were given TCA i.p. at doses of 0, 337, 675, or 1,080 mg/kg-day for males and 0,
405, 810, or 1,300 mg/kg-day for females for 2 consecutive days, and bone-marrow samples
were collected 6 and 24 hours after the last dose. The administered doses represented 25, 50, and
80% of the median lethal dose, respectively. No treatment-related increase in micronucleated
polychromatic erythrocytes was observed.
4.8.2.1.2.2. DNA damage studies
DNA unwinding assays have been used as indicators of single-strand breaks. Studies
were conducted on the ability of TCA to induce DNA single-strand breaks [(Chang et al., 1992;
Styles etal.. 1991: Nelson etal.. 1989: Nelson and Bull 1988): Table 4-12]. Nelson and Bull
(1988) evaluated the ability of TCA and other compounds to induce DNA single-strand breaks in
vivo in Sprague-Dawley rats and B6C3Fi mice. Single oral doses were administered to three
groups of three animals, with an additional group as a vehicle control. Animals were sacrificed
after 4 hours, and 10% liver suspensions were analyzed for DNA single-strand breaks by the
alkaline unwinding assay. Dose-dependent increases in DNA single-strand breaks were induced
in both rats and mice, with mice being more susceptible than rats. The lowest dose of TCA that
produced significant SSBs was 0.6 mmol/kg (98 mg/kg) in rats but 0.006 mmol/kg (0.98 mg/kg)
in mice.
4-380
-------
Table 4-42. Genotoxicity of trichloroacetic acid (TCA)—bacterial systems3
Test system/endpoint
A, Prophage induction, E. coli WP2s
SOS chromotest, E. coli PQ37
S. typhimurium TA1535, 1536, 1537, 1538,
reverse mutation
S. typhimurium TA100, 98, reverse mutation
S. typhimurium TA100, 1535, reverse
mutation
S. typhimurium TA1537, 1538, 98, reverse
mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, 98, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation,
liquid medium
S. typhimurium TA104, reverse mutation,
microsuspension
S. typhimurium TA100, RSJ100, reverse
mutation
S. typhimurium TA98, reverse mutation
S. typhimurium TA1535, SOS DNA repair
Doses
(LED or HID)b
10,000
10,000
20 ug/plate
450 ug/plate
4,000 ug/plate
2,000 ug/plate
520 ug/plate
5,000 ug/plate
600 ppm
1,750
250 ug/plate
16,300
13,100
NA
Results0
With
activation
-
-
NT
-
—
—
NT
-
-
+
-
-
-
+
Without
activation
-
-
—
-
—
—
-
-
-
+
-
-
-
-
Reference
DeMarini et al.
(1994)
Ciller etal. (1997)
Shirasu et al. (1976)
Waskell (1978)
Nestmann et al.
(1980)
Nestmann et al.
(1980)
Rapson et al. (1980)
Moriya et al. (1983)
DeMarini et al.
(1994)
Ciller et al. (1997)
Nelson et al. (200 Ib)
Kargalioglu et al.
(2002)
Kargalioglu et al.
(2002)
Ono et al. (1991)
"Table adapted from IARC monograph (2004) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests, unless otherwise
specified.
'Results: + = positive; - = negative; NT = not tested.
However, in a follow-up study (Nelson et al., 1989), no significant differences from
controls in DNA single-strand breaks in whole liver homogenates were observed in male
B6C3Fi mice exposed to 500 mg/kg TCA. Moreover, DCA increased single-strand breaks but
with no dose response between 10 and 500 mg/kg, raising concerns about the reliability of the
DNA unwinding assay used in these studies. In an additional follow-up experiment with a
similar experimental paradigm, Styles et al. (1991) tested TCA for its ability to induce strand
4-381
-------
breaks in male B6C3Fi mice in the presence and absence of liver growth induction. The test
animals were given 1, 2, or 3 daily doses of neutralized TCA (500 mg/kg) by gavage and killed 1
hour after the final dose. Additional mice were given a single 500 mg/kg gavage dose and
sacrificed 24 hours after treatment. Liver nuclei DNA were isolated, and the induction of single-
strand breaks was evaluated using the alkaline unwinding assay. Exposure to TCA did not
induce strand breaks under the conditions tested in this assay. In a study by Chang et al. (1992),
administration of single oral doses of TCA (1 to 10 mmol/kg) to B6C3Fi mice did not induce
DNA strand breaks in a dose-related manner as determined by the alkaline unwinding assay. No
genotoxic activity (evidence for strand breakage) was detected in F344 rats administered by
gavage up to 5 mmol/kg (817 mg/kg).
In summary, Nelson and Bull (1988) reported that DC A and TCA enhance DNA
unwinding in mice, with DCA having the highest activity and TCA the lowest. However, Nelson
et al. (1989) reported no effect for TCA and a lack of dose response for the effect of DCA (with
10 and 500 mg/kg DCA inducing the same magnitude of effect). Moreover, Styles et al. (1991)
did not report a positive result for TCA using the same paradigm as Nelson and Bull (1988) and
Nelson et al. (1989). Furthermore, Chang et al. (1992) also did not find increased DNA single-
strand breaks for TCA exposure in rats.
4.8.2.2. Bacterial Systems
4.8.2.2.1. Gene mutations
TCA has been evaluated in a number of in vitro test systems including the bacterial
assays (Ames) using different S. typhimurium strains such as TA98, TA100, TA104, TA1535,
and RSJ100 (refer to Table 4-42). The majority of these studies did not report positive findings
for genotoxicity (Kargalioglu et al., 2002; Nelson et al., 2001b: DeMarini etal., 1994; Moriya et
al.. 1983: Nestmann et al.. 1980: Rapson et al.. 1980: Waskell. 1978: Shirasu et al.. 1976).
Waskell (1978) studied the effect of TCA (0.45 mg/plate) on bacterial strains TA98 and TA100
both in the presence and absence of S9. The author did not find any revertants at the maximum
nontoxic dose tested. Following exposure to TCA, Rapson et al. (1980) reported no change in
mutagenic activity in strain TA100 in the absence of S9. DeMarini et al. (1994) performed
different studies to evaluate the genotoxicity of TCA, including the Microscreen prophage-
induction assay (TCA concentrations: 0 to 10 mg/mL) and use of the S. typhimurium TA100
strain using bag vaporization technique (TCA concentrations: 0-100 ppm), neither of which
yielded positive results. Nelson et al. (200Ib) reported no positive findings with TCA using a S.
typhimurium microsuspension bioassay (S. typhimurium strain TA104) following incubation of
TCA for various lengths of time, with or without rat cecal microbiota. Similarly, no activity was
observed in a study conducted by Kargalioglu et al. (2002) where S. typhimurium strains TA98,
4-382
-------
TA100, and RSJ100 were exposed to TCA (0.1-100 mM) either in the presence or absence of S9
(Kargalioglu et al.. 2002).
TCA was also negative in other bacterial systems. The SOS chromotest (which measures
DNA damage and induction of the SOS repair system) inE. coll PQ37, with and without S9
(Giller et al., 1997), evaluated the genotoxic activity of TCA ranging from 10 to 10,000 |ig/mL,
and no response was reported. Similarly, TCA was not genotoxic in the Microscreen prophage-
induction assay in E. coll with TCA concentrations ranging from 0 to 10,000 |ig/mL, with and
without S9 activation (DeMarini et al., 1994).
However, TCA induced a small increase in SOS DNA repair (an inducible error-prone
repair system) in S. typhimurium strain TA1535 in the presence of S9 (Ono et al., 1991).
Furthermore, Giller et al. (1997) reported that TCA demonstrated genotoxic activity in an Ames
fluctuation test in S. typhimurium TA100 in the absence of S9 at noncytotoxic concentrations
ranging from 1,750 to 2,250 |ig/mL. The addition of S9 decreased the genotoxic response, with
effects observed at 3,000-7,500 |ig/mL. Cytotoxic concentrations in the Ames fluctuation assay
were 2,500 and 10,000 |ig/mL, without and with microsomal activation, respectively.
4.8.2.3. Summary
TCA, an oxidative metabolite of tetrachloroethylene, exhibits little, if any genotoxic
activity in vitro. TCA did not induce mutations in S. typhimurium strains in the absence of
metabolic activation or in an alternative protocol using a closed system (Kargalioglu et al., 2002;
Nelson etal.. 2001b: Giller etal.. 1997: DeMarini et al.. 1994: Rapsonetal.. 1980: Waskell.
1978), but a mutagenic response was induced in TA100 in the Ames fluctuation test (Giller et al.,
1997). However, in vitro experiments with TCA should be interpreted with caution if steps have
not been taken to neutralize pH changes caused by the compound (Mackay et al., 1995).
Measures of DNA-repair responses in bacterial systems have shown induction of DNA repair
reported in S. typhimurium but not in E. coll. Mutagenicity in mouse lymphoma cells was only
induced at cytotoxic concentrations (Harrington-Brock et al., 1998). TCA was positive in some
genotoxicity studies in vivo mouse, newt, and chick test systems (Giller et al., 1997: Bhunya and
Jena, 1996: Birner et al., 1994: Bhunya and Behera, 1987). DNA unwinding assays have either
shown TCA to be much less potent than DCA (Nelson and Bull, 1988) or negative (Styles et al.,
1991: Nelson etal., 1989). Due to limitations in the genotoxicity database, the possible
contribution of TCA to tetrachloroethylene genotoxicity is unclear.
4-383
-------
4.8.3. Dichloroacetic Acid (DCA)
DCA is another metabolite of tetrachloroethylene that has been studied using a variety of
genotoxicity assays for its genotoxic potential [refer to Tables 4-43 and 4-44; refer to IARC
(2004) for additional information].
4.8.3.1. Mammalian Systems
4.8.3.1.1. Gene mutations
The mutagenicity of DCA has been tested in mammalian systems, particularly, mouse
lymphoma cell lines in vitro (Harrington-Brock et al., 1998; Fox et al., 1996) and lacl transgenic
mice in vivo (Leavitt et al., 1997). Harrington-Brock et al. (1998) evaluated DCA for mutagenic
activity in L5178Y/TK +/- (-) 3.7.2C mouse lymphoma cells. A dose-related increase in
mutation (and cytotoxic) frequency was observed at concentrations between 100 and 800 |ig/mL.
Most mutagenic activity of DCA at the Tk locus was due to the production of small-colony Tk
mutants (indicating chromosomal mutations). Different pH levels were tested in induction of
mutant frequencies, and it was determined that the mutagenic effect observed was due to the
chemical and not pH effects.
Mutation frequencies were studied in male transgenic B6C3Fi mice harboring the
bacterial lacl gene administered DCA at either 1.0 or 3.5 g/L in drinking water (Leavitt et al.,
1997). No significant difference in mutant frequency was observed after 4 or 10 weeks of
treatment in both the doses tested as compared to control. However, at 60 weeks, mice treated
with 1.0 g/L DCA showed a slight increase (1.3-fold) in the mutant frequency over the control,
but mice treated with 3.5 g/L DCA had a 2.3-fold increase in the mutant frequency. Mutational
spectra analysis revealed that -33% had G:C-A:T transitions and 21% had G:C-T:A
transversions, and this mutation spectra was different than that was observed in the untreated
animals, indicating that the mutations were likely induced by the DCA treatment. The authors
conclude that these results are consistent with the previous observation that the proportion of
mutations at T:A sites in Codon 61 of the H-ras gene was increased in DCA-induced liver tumors
in B6C3Fi mice (Leavitt et al.. 1997).
4.8.3.1.2. Chromosomal aberrations and micronucleus induction
Harrington-Brock et al. (1998) evaluated DCA for its potential to induce chromosomal
aberrations in DCA-treated (0, 600, and 800 |ig/mL) mouse lymphoma cells. A clearly positive
induction of aberrations was observed at both concentrations tested. No significant increase in
micronucleus was observed in DCA-treated (0, 600, and 800 |ig/mL) mouse lymphoma cells
(Harrington-Brock et al., 1998). However, no chromosomal aberrations were found in Chinese
hamster ovary cells exposed to DCA (Fox et al., 1996).
4-384
-------
Fuscoe et al. (1996) investigated in vivo genotoxic potential of DC A in bone marrow and
blood leukocytes using the peripheral-blood-erythrocyte micronucleus assay (to detect
chromosome breakage and/or malsegregation) and the alkaline single cell gel electrophoresis
(comet) assay, respectively. Mice were exposed to DCA in drinking water, available ad libitum,
for up to 31 weeks. A statistically significant dose-related increase in the frequency of
micrenucleated PCEs was observed following subchronic exposure to DCA for 9 days.
Similarly, a significant increase was also observed when mice were exposed for >10 weeks,
particularly at the highest dose of DCA tested (3.5 g/L). DNA cross-linking was observed in
blood leukocytes in mice exposed to 3.5 g/L DCA for 28 days. These data provide evidence that
DCA may have some potential to induce chromosome damage when animals are exposed to
concentrations similar to those used in the rodent bioassay.
4.8.3.1.3. DNA damage studies
Nelson and Bull (1988) and Nelson et al. (1989) have been described above in
Sections 4.2.2.4 and 4.2.3. Nelson and Bull (1988) reported positive results for DNA unwinding
for DCA, although Nelson et al. (1989) reported the same response at 10 and 500 mg/kg in mice,
raising concerns about the reliability of the assay in these studies. Chang et al. (1992) conducted
both in vitro and in vivo studies to determine the ability of DCA to cause DNA damage. Primary
rat (Fischer 344) hepatocytes and primary mouse hepatocytes treated with DCA for 4 hours did
not induce DNA single-strand breaks as detected by the alkaline DNA unwinding assay. No
DNA single-strand breaks were observed in human CCRF-CEM lymphoblastoid cells in vitro
exposed to DCA. Similarly, analysis of the DNA single-strand breaks in mice killed 1 hour after
a single dose of 1, 5, or 10 mM/kg DCA did not cause DNA damage. None of the Fischer 344
rats killed 4 hours after a single gavage treatment (1-10 mM/kg) produced any detectable DNA
damage.
4-385
-------
Table 4-43. Genotoxicity of dichloroacetic acid (DCA)—mammalian systems
(in vitro and in vivo)a
Test system/endpoint
Gene mutation, mouse lymphoma cell line
L5178Y/TK+/- in vitro
Gene mutation, mouse lymphoma cell line
L5178Y/TK+/- -3.7.2C in vitro
DNA strand breaks and alkali-labile damage, Chinese
hamster ovary cells in vitro (single-cell gel
electrophoresis assay)
DNA strand breaks, B6C3F! mouse hepatocytes in
vitro
DNA strand breaks, Fischer 344 rat hepatocytes in
vitro
Micronucleus formation, mouse lymphoma
L5178Y/TK+/- -3.7.2C cell line in vitro
Micronucleus induction, peripheral blood erythrocytes,
Tg. AC hemizygous mouse, dermal application in vivo
Micronucleus induction, peripheral blood erythrocytes,
Tg. AC hemizygous mouse, drinking water, in vivo
Micronucleus induction, peripheral blood erythrocytes,
p53 haploinsufficient mouse, drinking water, in vivo
Micronucleus induction, peripheral blood erythrocytes,
B6C3FJ mouse, drinking water, in vivo
Chromosomal aberrations, Chinese hamster ovary in
vitro
Chromosomal aberrations, mouse lymphoma
L5178Y/Tk+/- -3.7.2C cell line in vitro
Aneuploidy, mouse lymphoma L5 178Y/Tk+/- -3.7.2C
cell line in vitro
DNA strand breaks, human CCRF-CEM
lymphoblastoid cells in vitro
DNA strand breaks, male B6C3F! mouse liver in vivo
DNA strand breaks, male B6C3F! mouse liver in vivo
DNA strand breaks, male B6C3F! mouse liver in vivo
Doses
(LED or
HID)b
5,000
400
3,225 ug/mL
2,580
1,290
800
500 mg/kg
2,000 mg/L
2,000 mg/L
67 mg/L
5,000
600
800
1,290
13, p.o., xl
10, p.o., xl
1,290, p.o.,
xl
Results0
With
activation
-
NT
NT
NT
NT
NT
NT
NT
NT
NT
—
NT
NT
NT
NT
NT
NT
Without
activation
-
+
-
-
-
-
-
-
- (male)
equivocal
(female)
—
+
—
—
+
+
—
Reference
Fox et al. (1996)
Harrington-Brock et
al. (1998)
Plewa et al. (2002)
Chang et al. (1992)
Chang et al. (1992)
Harrington-Brock et
al. (1998)
NTP (2007)
NTP (2007)
NTP (2007)
NTP (2007)
Fox et al. (1996)
Harrington-Brock et
al. (1998)
Harrington-Brock et
al. (1998)
Chang et al. (1992)
Nelson and Bull
Q988)
Nelson et al. (1989)
Chang et al. (1992)
4-386
-------
Table 4-43. Genotoxicity of dichloroacetic acid (DCA)—mammalian systems
(in vitro and in vivo)a (continued)
Test system/endpoint
DNA strand breaks, male B6C3FJ mouse splenocytes in
vivo
DNA strand breaks, male B6C3FJ mouse epithelial cells
from stomach and duodenum in vivo
DNA strand breaks, male B6C3FJ mouse liver in vivo
DNA strand breaks, male B6C3FJ mouse liver, in vivo
DNA strand breaks, alkali-labile sites, cross linking,
male B6C3FJ mouse blood leukocytes in vivo (single-
cell gel electrophoresis assay)
DNA strand breaks, male Sprague-Dawley rat liver in
vivo
DNA strand breaks, male Fischer 344 rat liver in vivo
DNA strand breaks, male Fischer 344 rat liver in vivo
Gene mutation, lacl transgenic male B6C3F! mouse
liver assay in vivo
Altered gene expression, male B6C3F! mouse liver
assay in vivo
Micronucleus formation, male B6C3Fi mouse
peripheral erythrocytes in vivo
Micronucleus formation, male B6C3Fi mouse
peripheral erythrocytes in vivo
Micronucleus formation, male B6C3Fi mouse
peripheral erythrocytes in vivo
Micronucleus formation, male and female Crl:CD (S-D)
BR rat bone-marrow erythrocytes in vivo
Micronucleus formation, Pleurodeles waltl larvae
peripheral erythrocytes in vivo
Doses
(LED or
HID)b
1,290, p.o.,
xl
1,290, p.o.,
xl
5,000, dw,
x?-14d
300, p.o.
3,500, dw,
x28d
30, p.o., xl
645, p.o., xl
2,000, dw,
x30wk
1,000, dw,
x60 wk
2,000, dw,
x4 wk
3,500, dw, x9
d
3,500, dw,
x28d
3,500, dw,
xlOwk
1,100, i.v.,
x3
80 d
Results0
With
activation
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
Without
activation
—
—
—
+
+
+
-
-
+
+
+
-
+
-
—
Reference
Chang et al. (1992)
Chang et al. (1992)
Chang et al. (1992)
Hassoun and Dey
(2008)
Fuscoe et al. (1996)
Nelson and Bull
(1988)
Chang et al. (1992)
Chang et al. (1992)
Leavitt et al. (1997)
Thai et al. (2003)
Fuscoe et al. (1996)
Fuscoe et al. (1996)
Fuscoe et al. (1996)
Fox et al. (1996)
Ciller et al. (1997)
"Table adapted from IARC monograph (2004) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests; mg/kg for in vivo
tests unless specified; dw = drinking-water (in mg/L); i.v. = intravenous.
'Results: + = positive; - = negative; NT = not tested.
4-387
-------
Table 4-44. Genotoxicity of dichloroacetic acid (DCA)—bacterial systems3
Test system/endpoint
A, Prophage induction, E. coli WP2s
SOS chromotest, E. coli PQ37
S. typhimurium, DNA repair-deficient strains
TS24, TA2322, TA1950
S. typhimurium TA100, TA1535, TA1537,
TA1538, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, TA1535, TA1537,
TA98, reverse mutation
S. typhimurium TA100, reverse mutation, liquid
medium
S. typhimurium RSJ100, reverse mutation
S. typhimurium TA104, reverse mutation,
microsuspension
S. typhimurium TA98, reverse mutation
S. typhimurium TA98, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA98, gene mutation
S. typhimurium TA100, gene mutation
S. typhimurium TA1535, gene mutation
E. coli WP2uvrA, reverse mutation
Doses
(LED or HID)b
2,500
500
31,000
NA
50
5,000
100
1,935
150 ug/plate
10 ug/plate
5,160
1,935
3 ug/plate
333 ug/plate
333 ug/plate
5,000
Results0
With
activation
+
-
-
-
+
-
+
-
-
(+)
—
+
-
-
-
-
Without
activation
-
(+)
-
-
+
-
+
+
-
-
+
+
-
+
+
-
Reference
DeMarini et al.
(1994)
Ciller et al. (1997)
Waskell (1978)
Herbert et al. (1980)
DeMarini et al.
(1994)
Fox et al. (1996)
Ciller et al. (1997)
Kargalioglu et al.
(2002)
Nelson et al. (200 Ib)
Herbert et al. (1980)
Kargalioglu et al.
(2002)
Kargalioglu et al.
(2002)
NTPQ007)
NTP (2007)
NTPQ007)
Fox et al. (1996)
aTable adapted from IARC monograph (2004) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests, unless otherwise
specified; NA = not available.
'Results: + = positive; (+) = weakly positive; - = negative.
4-388
-------
4.8.3.2. Bacterial Systems
4.8.3.2.1. Gene mutations
Studies were conducted to evaluate mutagenicity of DC A in different S. typhimurium and
E. coli strains lYKargalioglu et al.. 2002: Nelson etal.. 2001b: Gilleretal.. 1997: Fox et al..
1996: DeMarini et al.. 1994: Herbert et al.. 1980: Waskell. 1978): summarized in Table 4-44].
DCA was mutagenic in three strains of S. typhimurium: strain TA100 in three of five studies,
strain RSJ100 in a single study, and strain TA98 in two of three studies. DCA failed to induce
point mutations in other strains of S. typhimurium (TA104, TA1535, TA1537, and TA1538) or in
E. coli strain WP2uvrA. In one study, DCA caused a weak induction of SOS repair in E. coli
strain PQ37 (Gilleretal.. 1997).
DeMarini et al. (1994), in the same study as described in the TCA section (refer to
Section 4.8.2), also studied DCA as one of their compounds for analysis. In the prophage-
induction assay using E. coli, DCA, in the presence of S9, was genotoxic, producing 6.6-7.2
plaque-forming units (PFU)/mM and slightly less than threefold increase in PFU/plate in the
absence of S9. In the second set of studies, which involved the evaluation of DCA at
concentrations of 0-600 ppm for mutagenicity in S. typhimurium TA100 strain, DCA was
mutagenic both in the presence and absence of S9, producing three- to fivefold increases in the
revertants/plate compared to the background. The lowest effective concentration for DCA
without S9 was 100 ppm and 50 ppm in the presence of S9. In the third and most important
study, mutation spectra of DCA were determined at the base-substitution allele hisG46 of S.
typhimurium TA100. DCA-induced revertants were chosen for further molecular analysis at
concentrations that produced mutant yields that were two- to fivefold greater than the
background. The mutation spectra of DCA were significantly different from the background
mutation spectrum. Thus, despite the modest increase in the mutant yields (3-5 times) produced
by DCA, the mutation spectra confirm that DCA is mutagenic. DCA primarily induced GC-AT
transitions.
Kargalioglu et al. (2002) analyzed the cytotoxicity and mutagenicity of the drinking
water disinfection by-products including DCA in S. typhimurium strains TA98, TA100, and
RSJ100 +/- S9. DCA was mutagenic in this test, although the response was low when compared
to other disinfection by-products tested in strain TA100. This study was also summarized in a
review by Plewa et al. (2002). Nelson et al. (200 Ib) investigated the mutagenicity of DCA using
a S. typhimurium microsuspension bioassay following incubation of DCA for various lengths of
time, with or without rat cecal microbiota. No mutagenic activity was detected for DCA with
S. typhimurium strain TA104. Although the data are limited, it appears that DCA has mutagenic
activity in the S. typhimurium strains, particularly TA100.
4-389
-------
4.8.3.3. Summary
DCA, a chloroacid metabolite of tetrachloroethylene, has also been studied using
different types of genotoxicity assays. Although studies are limited for different genetic
endpoints, DCA has demonstrated mutagenicity at high doses in some strains in S. typhimurium
assays (Kargalioglu et al., 2002; Plewa et al., 2002; DeMarini et al., 1994), a mouse lymphoma
assay (Harrington-Brock et al., 1998), in vivo cytogenetic tests (Leavitt et al., 1997; Fuscoe et
al., 1996), the micronucleus induction test, the Big Blue mouse system, and other tests
(Harrington-Brock et al., 1998: Leavitt et al., 1997: Fuscoe etal., 1996: DeMarini et al., 1994:
Chang etal., 1989: Nelson etal., 1989: Nelson and Bull, 1988). DCA can cause DNA strand
breaks in mouse and rat liver cells following in vivo exposures (Fuscoe et al., 1996). Because of
uncertainties as to the extent of DCA formed from tetrachloroethylene exposure, inferences as to
the possible contribution from DCA genotoxicity to tetrachloroethylene toxicity are difficult to
make.
4.8.4. Chloral Hydrate
Although chloral hydrate is postulated as a metabolite of tetrachloroethylene, this is not
widely accepted. However, to be inclusive of all known genotoxicity information, chloral
hydrate genotoxicity studies have been reviewed in the following section. Chloral hydrate has
been evaluated for its genotoxic potential using a variety of genotoxicity assays (refer to Tables
4.45, 4.46, and 4-47).
4.8.4.1. Mammalian Systems (Including Human Studies)
4.8.4.1.1. Gene mutations
Harrington-Brock (1998) noted that chloral hydrate-induced concentration related
cytotoxicity in TK+/- mouse lymphoma cell lines without S9 activation. A nonstatistical
increase in mutant frequency was observed in cells treated with chloral hydrate. The mutants
were primarily small colony TK mutants, indicating that most chloral hydrate-induced mutants
resulted from chromosomal mutations rather than point mutations. It should be noted that in
most concentrations tested (350-1,600 ug/mL), cytotoxicity was observed. Percentage cell
survival ranged from 96 to 4%.
4-390
-------
Table 4-45. Genotoxicity of chloral hydrate—mammalian systems (in vitro)a
Test system/endpoint
DNA -protein cross-links, rat nuclei in vitro
DNA single-strand breaks, rat primary hepatocytes in
vitro
Gene mutation, mouse lymphoma L5178Y/TK+/-, in
vitro
Sister chromatid exchange, CHO cells, in vitro
Micronucleus formation (kinetochore-positive),
Chinese hamster Cl cells, in vitro
Micronucleus formation (kinetochore-negative),
Chinese hamster Cl cells, in vitro
Micronucleus formation (kinetochore-positive),
Chinese hamster LUC2 cells, in vitro
Micronucleus formation (kinetochore-positive),
Chinese hamster LUC2 cells, in vitro
Micronucleus formation, Chinese hamster V79 cells,
in vitro
Micronucleus formation, mouse lymphoma
L5178Y/TK+/-, in vitro
Micronucleus formation, mouse lymphoma
L5178Y/TK+/-, in vitro
Chromosomal aberrations, Chinese Hamster cells, in
vitro
Chromosomal aberrations, Chinese Hamster ovary
cells, in vitro
Chromosomal aberrations, mouse lymphoma
L5178Y/TK +/- cells line, in vitro
Aneuploidy, Chinese hamster CHED cells, in vitro
Aneuploidy, primary Chinese hamster embryonic
cells, in vitro
Aneuploidy, Chinese hamster LUC2p4 cells, in vitro
Aneuploidy, mouse lymphoma L5178Y/TK+/- , in
vitro
Tetraploidy and endoredupliation, Chinese hamster
LUC2p4cells, in vitro
Cell transformation, Syrian hamster embryo cells
(24-h treatment)
Doses
(LED or
HID)b
41,250
1,650
1,000
100
165
250
400
400
316
1,300
500
20
1,000
1,250
10
250
250
1,300
500
350
Results0
With
activation
NT
NT
NT
+
NT
NT
NT
NT
NT
NT
NT
NT
+
NT
NT
NT
NT
NT
NT
NT
Without
activation
-
—
(+)
+
+
—
+
+
+
—
+
+
+
(+)
+
+
+
-
+
+
Reference
Keller and Heck
(1988)
Chang et al. (1992)
Harrington-Brock et
al. (1998)
Beland (1999)
Degrassi and
Tanzarella (1988)
Degrassi and
Tanzarella (1988)
Parry et al. (1990)
Lynch and Parry
Q993)
Seelbach et al. (1993)
Harrington-Brock et
al. (1998)
Nesslany and Marzin
(1999)
Furnus et al. (1990)
Beland (1999)
Harrington-Brock et
al. (1998)
Furnus et al. (1990)
Natarajan et al.
(1993)
Warr et al. (1993)
Harrington-Brock et
al. (1998)
Warr et al. (1993)
Gibson et al. (1995)
4-391
-------
Table 4-45. Genotoxicity of chloral hydrate—mammalian systems (in
vitro)" (continued)
Test system/endpoint
Cell transformation, Syrian hamster dermal cell line
(24-h treatment)
DNA single-strand breaks, human lymphoblastoid
cells, in vitro
Gene mutation, tk and hprt locus, human
lymphoblastoid
Sister chromatid exchanges, human lymphocytes, in
vitro
Micronucleus formation, human lymphocytes, in vitro
Micronucleus formation, human lymphoblastoid
AHH-1 cell line, in vitro
Micronucleus formation, human lymphoblastoid
MCL-5 cell line, in vitro
Micronucleus formation (kinetochore-positive),
human diploid LEO fibroblasts, in vitro
Aneuploidy (double Y induction), human
lymphocytes, in vitro
Aneuploidy (hyperdiploidy and hypodiploidy), human
lymphocytes in vitro
Polyploidy, human lymphocytes, in vitro
C-Mitosis, human lymphocytes, in vitro
Doses
(LED or
HID)b
50
1,650
1,000
54
100
100
500
120
250
50
137
75
Results0
With
activation
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
Without
activation
+
-
+
(+)
+
+
-
+
+
+
+
+
Reference
Parry et al. (1996)
Chang et al. (1992)
Beland (1999)
Gu et al. (1981)
Van Hummelen and
Kirsch-Volders
(1992)
Parry et al. (1996)
Parry et al. (1996)
Bonatti et al. (1992)
Vagnarelli et al.
(1990)
Sbrana et al. (1993)
Sbrana et al. (1993)
Sbrana et al. (1993)
aTable adapted from IARC monograph (2004) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests unless otherwise
specified.
'Results: + = positive; (+) = weakly positive; - = negative; NT = not tested.
4-392
-------
Table 4-46. Genotoxicity of chloral hydrate—mammalian systems (in vivo)a
Test system/endpoint
DNA single-strand breaks, male Sprague-Dawley rat liver
DNA single-strand breaks, male Fischer 344 rat liver
DNA single-strand breaks, male B6C3Fi mouse liver
DNA single-strand breaks, male B6C3Fi mouse liver
Increased detection of MjG and 8-OHdG adducts, B6C3FJ
neonatal mouse liver DNA, in vivo, i.p. injection
Micronucleus formation, male and female NMRI mice,
bone-marrow erythrocytes
Micronucleus formation, BALB/c mouse spermatids
Micronucleus formation, male BALB/c mouse bone-marrow
erythrocytes and early spermatids
Micronucleus formation, male BALB/c mouse bone-marrow
erythrocytes
Micronucleus formation, male Fl mouse bone-marrow
erythrocytes
Micronucleus formation, C57B 1 mouse spermatids
Micronucleus formation, male Swiss CD-I mouse bone-
marrow erythrocytes
Micronucleus formation, B6C3F! mouse spermatids after
spermatogonial stem-cell treatment
Micronucleus formation, B6C3F! mouse spermatids after
meiotic cell treatment
Micronucleus formation, male Fl, BALB/c mouse
peripheral-blood erythrocytes
Micronucleus formation, male B6C3F! mouse bone-marrow
erythrocytes
Micronucleus formation, infants, peripheral lymphocytes
Chromosomal aberrations, male and female Fl mouse bone
marrow cells
Chromosomal aberrations, male and female Sprague-
Dawley rat bone-marrow cells
Chromosomal aberrations, BALB/c mouse spermatogonia
treated
Chromosomal aberrations, Fl mouse secondary
spermatocytes
Chromosomal aberrations, male Swiss CD-I mouse bone-
marrow erythrocytes
Doses
(LED or HID)b
300, p.o.
1,650, p.o.
100, p.o.
825, p.o.
2,000 nmol
500, i.p.
83, i.p.
83, i.p.
200, i.p.
400, i.p.
41, i.p.
200, i.p.
165, i.p.
413, i.p.
200, i.p.
500, i.p., x3
50, p.o.
600, i.p.
1,000, p.o.
83, i.p.
82.7, i.p.
400, i.p.
Results0
+
-
+
-
+
-
-
+
+
-
+
+
+
-
-
+
+
-
-
-
+
Reference
Nelson and Bull (1988)
Chang et al. (1992)
Nelson and Bull (1988)
Chang et al. (1992)
Von Tungeln et al. (2002)
Leuschner and Leuschner
(1991)
Russo and Levis (1992a)
Russo and Levis (1992b)
Russo et al. (1992)
Leopardi et al. (1993)
Allen et al. (1994)
Marrazzini et al. (1994)
Nutley et al. (1996)
Nutley et al. (1996)
Grawe et al. (1997)
Beland (1999)
Ikbal et al. (2004)
Xu and Alder (1990)
Leuschner and Leuschner
(1991)
Russo and Levis (1992a)
Russo et al. (1984)
Marrazzini et al. (1994)
4-393
-------
Table 4-46. Genotoxicity of chloral hydrate—mammalian systems (in
vivo)a (continued)
Test system/endpoint
Chromosomal aberrations, ICR mouse oocytes
Micronucleus formation, infants, peripheral lymphocytes
Polyploidy, male and female Fl, mouse bone-marrow cells
Aneuploidy Fl mouse secondary spermatocytes
Aneuploidy, male Fl mouse secondary spermatocytes
Hyperploidy, male Swiss CD-I mouse bone-marrow
erythrocytes
Doses
(LED or HID)b
600, i.p.
50, p.o.
600, i.p.
200, i.p.
400, i.p.
200, i.p.
Results0
+
+
+
Reference
Mailhes et al. (1993)
Ikbal et al. (20041
Xu and Adler (1990)
Miller and Adler
(1992)
Leopardi et al. (1993)
Marrazzini et al.
(1994)
aTable adapted from IARC monograph (2004) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in mg/kg for in vivo tests unless otherwise
specified; i.p. = intraperitoneally, p.o. = orally.
'Results: + = positive; - = negative
4-394
-------
Table 4-47. Genotoxicity of chloral hydrate—bacterial, yeast, and fungal
systems3
Test system/endpoint
SOS chromotest, E. coli PQ37
S. typhimurium TA100, TA1535, TA98, reverse
mutation
S. typhimurium TA100, TA1537, TA1538, TA98,
reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation
S. typhimurium TA100, reverse mutation, liquid
medium
S. typhimurium TA100, TA104, reverse mutation
S. typhimurium TA104, reverse mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA1535, TA1537 reverse
mutation
S. typhimurium TA1535, reverse mutation
S. typhimurium TA98, reverse mutation
S. typhimurium TA98, reverse mutation
A. nidulans, diploid strain 35X17, mitotic
crossovers
A. nidulans, diploid strain 30, mitotic crossovers
A. nidulans, diploid strain NH, mitotic crossovers
A. nidulans, diploid strain PI, mitotic crossovers
A. nidulans, diploid strain 35X17,
nondisjunctions
A. nidulans, diploid strain 30, aneuploidy
A. nidulans, haploid conidia, aneuploidy,
polyploidy
A. nidulans, diploid strain NH, nondisjunctions
A. nidulans, diploid strain PI, nondisjunctions
A. nidulans, haploid strain 35, hyperploidy
Doses
(LED or HID)b
10,000
10,000
1,000
5,000 ug/plate
2,000 ug/plate
300
1,000 ug/plate
1,000 ug/plate
1,850
6,667
10,000
7,500
10,000 ug/plate
1,650
6,600
1,000
990
825
825
1,650
450
660
2,640
Results0
With
activation
-
—
+
—
+
+
+
+
-
-
-
-
-
NT
NT
NT
NT
NT
NT
NT
NT
NT
NT
Without
activatio
n
-
—
+
—
+
-
+
+
-
-
-
-
+
—
-
-
-
+
+
+
+
+
+
Reference
Ciller et al. (1995)
Waskell (1978)
Haworth et al. (1983)
Leuschner and
Leuschner (1991)
Ni et al. (1994)
Ciller et al. (1995)
Beland (1999)
Ni et al. (1994)
Leuschner and
Leuschner (1991)
Haworth et al. (1983)
Beland (1999)
Haworth et al. (1983)
Beland (1999)
Crebelli et al. (1985)
Kafer (1986)
Kappas (1989)
Crebelli et al. (1991)
Crebelli et al. (1985)
Kafer (1986)
Kafer (1986)
Kappas (1989)
Crebelli et al. (1991)
Crebelli et al. (1991)
4-395
-------
Table 4-47. Genotoxicity of chloral hydrate—bacterial, yeast, and fungal
systems" (continued)
Test system/endpoint
S. cerevisiae, meiotic recombination
S. cerevisiae, disomy in meiosis
S. cerevisiae, disomy in meiosis
S. cerevisiae, D61.M, mitotic chr. malsegregation
D. melanogaster, somatic mutation wing spot test
D. melanogaster, induction of sex-linked lethal
mutation
D. melanogaster, induction of sex-linked lethal
mutation
Doses
(LED or HID)b
3,300
2,500
3,300
1,000
825
37.2 feed
67.5 injection
Results0
With
activation
NT
NT
NT
NT
NT
NT
NT
Without
activation
7
+
+
+
+
7
—
Reference
Sora and Agostini
Carbone (1987)
Sora and Agostini
Carbone (1987)
Sora and Agostini
Carbone (1987)
Albertini (1990)
Zordan et al.
(1994)
Beland (1999)
Beland (1999)
"Table adapted from IARC monograph (2004) and modified/updated for newer references.
bLED, lowest effective dose; HID, highest ineffective dose; doses are in ug/mL for in vitro tests; inj =
injection.
'Results: + = positive; - = negative; NT = not tested; ? = inconclusive.
4.8.4.1.2. DNA binding studies
Limited analysis has been performed examining the DNA binding potential of chloral
hydrate (Von Tungeln et al.. 2002: Ni et al.. 1995: Keller and Heck. 1988). Keller and Heck
(1988) conducted both in vitro and in vivo experiments using the B6C3F1 mouse strain. The
mice were pretreated with 1,500 mg/kg TCE for 10 days and then given 800 mg/kg [14C]
chloral. No detectable covalent binding of 14C to DNA in the liver was observed. Another
study with in vivo exposures to nonradioactive chloral hydrate at a concentration of 1,000 and
2,000 nmol in B6C3F1 mice demonstrated an increase in malondialdehyde-derived and
8-oxo-2'-deoxyguanosine adducts in liver DNA (Von Tungeln et al., 2002). Ni et al. (1995)
observed malondialdehyde adducts in calf thymus DNA when exposed to chloral hydrate and
microsomes from male B6C3Fi mouse liver.
Keller and Heck (1988) investigated the potential of chloral to form DNA-protein cross-
links in rat liver nuclei using concentrations of 25, 100, or 250 mM. No statistically significant
increase in DNA-protein cross-links was observed. DNA and RNA isolated from the [14C]
4-396
-------
chloral-treated nuclei did not have any detectable 14C bound. However, the proteins from
choral-treated nuclei did have a concentration-related binding of 14C.
4.8.4.1.3. Chromosomal aberrations
Chloral hydrate induced aneuploidy in vitro in multiple Chinese hamster cell lines
(Natarajan et al., 1993; Warr etal., 1993; Furnus etal., 1990) and human lymphocytes (Sbrana et
al., 1993; Vagnarelli etal., 1990) but not mouse lymphoma cells (Harrington-Brock et al., 1998).
In vivo studies performed in various mouse strains led to increased aneuploidy in spermatocytes
(Miller and Adler, 1992; Liang and Pacchierotti, 1988; Russo et al., 1984) but not oocytes
(Mailhes etal., 1988) or bone marrow cells (Leopardi et al., 1993; Xu and Adler, 1990).
The potential of chloral hydrate to induce aneuploidy in mammalian germ cells has been
of particular interest since Russo et al. (1984) first demonstrated that chloral hydrate treatment of
male mice results in a significant increase in frequencies of hyperploidy in metaphase II cells.
This hyperploidy was thought to have arisen from chromosomal nondisjunction in
premeiotic/meiotic cell division and may be a consequence of chloral hydrate interfering with
spindle formation [reviewed by Russo et al. (1984) and Liang and Brinkley (1985)1. Chloral
hydrate also causes meiotic delay, which may be associated with aneuploidy (Miller and Adler,
1992). Chloral hydrate has been shown to induce micronuclei but not structural chromosomal
aberrations in mouse bone-marrow cells. Micronuclei induced by nonclastogenic agents are
generally believed to represent intact chromosomes that failed to segregate into either daughter-
cell nucleus at cell division (Russo and Levis, 1992b: Xu and Adler, 1990). Furthermore, chloral
hydrate-induced micronuclei in mouse bone-marrow cells (Russo et al., 1992) and in cultured
mammalian cells (Bonatti etal., 1992; Degrassi and Tanzarella, 1988) have shown to be
predominantly kinetochore-positive in composition upon analysis with immunofluorescent
methods. The presence of a kinetochore in a micronucleus is considered evidence that the
micronucleus contains a whole chromosome lost at cell division (Eastmond and Tucker, 1989;
Degrassi and Tanzarella, 1988; Hennig et al., 1988). Therefore, both TCE and chloral hydrate
appear to increase the frequency of micronuclei.
Allen et al. (1994) exposed male C57B1/6J mice to a single i.p. injection of 0, 41, 83, or
165 mg/kg chloral hydrate. Spermatids were harvested at 22 hours and 11, 13.5, and 49 days
following exposure (Allen et al., 1994). Harvested spermatids were processed to identify both
kinetochore-positive micronucleus (aneugenicity) and kinetochore-negative micronucleus
(clastogenicity). All chloral hydrate doses administered 49 days prior to cell harvest were
associated with significantly increased frequencies of kinetochore-negative micronuclei in
spermatids; however, dose dependence was not observed. This study is in contrast with other
4-397
-------
studies (Bonatti et al., 1992; Degrassi and Tanzarella, 1988) that demonstrated predominantly
kinetochore-positive micronucleus.
The ability of chloral hydrate to induce aneuploidy and polyploidy was tested in human
lymphocyte cultures established from blood samples obtained from two healthy nonsmoking
donors (Sbrana et al., 1993). Cells were exposed for 72 and 96 hours at doses between 50 and
250 |ig/mL. No increase in percentage hyperdiploid, tetraploid, or endoreduplicated cells was
observed when cells were exposed for 72 hours at any doses tested. However, at 96 hours of
exposure, a significant increase in hyperdiploid was observed at one dose (150 ug/mL) and was
not dose dependent. Tetraploidy was significantly increased at 137 mg/mL, again without dose-
dependency.
Ikbal et al. (2004) assessed genotoxicity (i.e., induction of micronuclei) in cultured
peripheral blood lymphocytes of 18 infants (age range of 31-55 days) before and after
administration of a single dose of chloral hydrate (50 mg/kg of body weight) for sedation before
a hearing test. A significant increase in micronuclei frequency was observed after administration
of chloral hydrate.
Analysis of chloral hydrate treated mouse lymphoma cell lines for chromosomal
aberrations resulted in a nonsignificant increase in chromosomal aberrations (Harrington-Brock
et al., 1998). However, it should be noted that the concentrations tested (1,250 and
1,300 |ig/mL) were cytotoxic (with a cell survival of 11 and 7%, respectively). Chinese hamster
embryo cells were also exposed to 0.001, 0.002, and 0.003% chloral hydrate for 1.5 hours
(Furnus et al., 1990). A nonstatistically significant increase in frequency of chromosomal
aberrations was observed only at 0.002 and 0.003% concentrations, with the increase not being
dose dependent. In this study, it should be noted that the cells were only exposed for 1.5 hours to
chloral hydrate and cells were allowed to grow for 48 hours (two cell cycles) to obtain similar
mitotic indices before analyzing for chromosomal aberrations. No information on cytotoxicity
was provided except that higher doses decreased the frequency of mitotic cells at the time of
fixation.
In vivo chromosome aberration studies have mostly reported negative or null results
(Mailhes et al., 1993; Russo and Levis, 1992a: Leuschner and Leuschner, 1991; Xu and Adler,
1990; Liang and Pacchierotti, 1988) with the exception of one study (Russo et al., 1984) in an Fl
cross of mouse strain between C57Bl/Cne x C3H/Cne.
4.8.4.1.3.1. Micronucleus induction
Micronuclei induction following exposure to chloral hydrate is positive in most test
systems in both in vitro and in vivo assays, although some negative tests do also exist (Ikbal et
al., 2004; Beland, 1999: Nesslany and Marzin, 1999: Harrington-Brock et al., 1998: Grawe et al..
4-398
-------
1997: Nutley etal.. 1996: Parry etal.. 1996: Gilleretal.. 1995: Allen etal.. 1994: Marrazzini et
al.. 1994: Leopard! et al.. 1993: Lynch and Parry. 1993: Seelbach et al.. 1993: Bonatti et al..
1992: Russo and Levis, 1992a, b; Van Hummelen and Kirsch-Volders, 1992: Leuschner and
Leuschner, 1991: Degrassi and Tanzarella, 1988). Some studies have attempted to make
inferences regarding aneuploidy induction or clastogenicity as an effect of chloral hydrate.
Aneuploidy results from defects in chromosome segregation during mitosis and is a common
cytogenetic feature of cancer cells. Ciller et al. (1995) studied chloral hydrate genotoxicity in
three short-term tests. Chloral hydrate caused a significant increase in the frequency of
micrenucleated erythrocytes following in vivo exposure of the amphibian Pleurodeles waltl
larvae.
4.8.4.1.3.2. Sister chromatid exchanges (SCEs)
SCEs were assessed by Ikbal et al. (2004) in cultured peripheral blood lymphocytes of
18 infants (age range of 31-55 days) before and after administration of a single dose of chloral
hydrate (50 mg/kg of body weight) for sedation before a hearing test. The authors report a
significant increase in the mean number of SCEs, from before administration (7.03 ±0.18
SCEs/cell) and after administration (7.90 ± 0.19 SCEs/cell), with each of the 18 individuals
showing an increase with treatment. Micronuclei were also significantly increased. SCEs were
also assessed by Gu et al. (1981) in human lymphocytes exposed in vitro with inconclusive
results, although positive results were observed by Beland (1999) in Chinese hamster ovary cells
exposed in vitro with and without an exogenous metabolic system.
4.8.4.1.4. Cell transformation
Chloral hydrate was positive in the two studies designed to measure cellular
transformation (Parry etal., 1996: Gibson etal., 1995). Both studies exposed Syrian hamster
cells (embryo and dermal) to chloral hydrate, which induced cellular transformation.
4.8.4.2. Bacterial and Fungal Systems
4.8.4.2.1. Gene mutations
Chloral hydrate induced gene mutations in S. typhimurium TA100 and TA104 strains but
not in most other strains assayed. Four of six studies of chloral hydrate exposure in
S. typhimurium TA100 and two of two studies in S. typhimurium TA104 were positive for
revertants (Beland. 1999: Gilleretal.. 1995: Metal.. 1994: Haworth et al.. 1983). Waskell
(1978) studied the effect of chloral hydrate along with TCE and its other metabolites. Chloral
hydrate was tested at different doses (1.0-13 mg/plate) in different S. typhimurium strains
(TA98, TA100, TA1535) for gene mutations using the Ames assay. No revertant colonies were
observed in strains TA98 or TA1535 both in the presence and absence of S9 mix. Similar results
4-399
-------
were obtained by Leuschner and Leuschner (1991). However, in TA100, a dose-dependent
statistically significant increase in revertant colonies was obtained both in the presence and
absence of S9. It should be noted that chloral hydrate that was purchased from Sigma was
recrystallized from one to six times from chloroform, and the authors describe this as crude
chloral hydrate. However, this positive result is consistent with other studies in this strain as
noted above. Furthermore, Giller et al. (1995) studied chloral hydrate genotoxicity in three
short-term tests. Chloral-induced mutations in strain TA100 of S. typhimurium (fluctuation test).
Similar results were obtained by Haworth et al. (1983). These are consistent with several studies
of TCE, in which low, but positive, responses were observed in the TA100 strain in the presence
of S9 metabolic activation, even when genotoxic stabilizers were not present.
A significant increase in mitotic segregation was observed mAspergillus nidulans when
exposed to 5 and 10 mM chloral hydrate (Crebelli et al., 1985). Studies of mitotic crossing-over
in A. nidulans have been negative, while these same studies were positive for aneuploidy
(Crebelli et al.. 1991: Kappas. 1989: Kafer, 1986: Crebelli et al.. 1985).
Two studies were conducted in S. cerevisiae to understand the chromosomal
malsegregation as a result of exposure to chloral hydrate (Albertini, 1990: Sora and Agostini
Carbone, 1987). Chloral hydrate (1-25 mM) was dissolved in sporulation medium, and the
frequencies of various meiotic events such as recombination and disomy were analyzed. Chloral
hydrate inhibited sporulation as a function of dose and increased diploid and disomic clones.
Chloral hydrate was also tested for mitotic chromosome malsegregation using S. cerevisiae
D61.M (Albertini, 1990). The tester strain was exposed to a dose range of 1-8 mg/mL. An
increase in the frequency of chromosomal malsegregation was observed as a result of exposure
to chloral hydrate.
Limited analysis of chloral hydrate mutagenicity has been performed mDrosophila
(Beland, 1999: Zordan etal., 1994). Of these two studies, chloral hydrate was positive in the
somatic mutation wing spot test (Zordan et al., 1994), equivocal in the induction of sex-linked
lethal mutation when administered in feed, but negative when exposed via injection (Beland,
1999).
4.8.4.3. Summary
Chloral hydrate has been reported to induce micronuclei formation, aneuploidy, and
mutations in multiple in vitro systems and in vivo. In vivo studies are limited to increased
micronuclei formation mainly in mouse spermatocytes. CH is positive in some in vitro
genotoxicity assays that detect point mutations, micronuclei induction, chromosomal aberrations,
and/or aneuploidy. The in vivo data exhibit mixed results (Leuschner and Beuscher, 1998:
Nutlev et al., 1996: Allen et al., 1994: Mailhes et al., 1993: Xu and Adler, 1990). Most of the
4-400
-------
positive studies show that chloral hydrate induces aneuploidy. Based on the existing array of
data, CH has the potential to be genotoxic, particularly when aneuploidy is considered in the
weight of evidence for genotoxic potential. Some have suggested that chloral hydrate may act
through a mechanism of spindle poisoning, resulting in numerical changes in the chromosomes,
but some data also suggest induction of chromosomal aberrations. These results are consistent
with tetrachloroethylene, albeit there are more limited data on tetrachloroethylene for these
genotoxic endpoints.
4.8.5. Trichloroacetyl Chloride
Trichloroacetyl chloride results from oxidative metabolism of tetrachloroethylene. The
limited genotoxicity studies of this metabolite are described below and listed in Table 4-48.
4.8.5.1. Bacterial Systems
4.8.5.1.1. Gene mutation
The genotoxicity of trichloroacetyl chloride has been studied in S. typhimurium with
inconsistent results. Reichert et al. (1983) found no mutagenicity of trichloroacetyl chloride
exposed in a liquid suspension to S. typhimurium TA98 and TA100 strains with and without S9
activation. A second study (DeMarini et al., 1994) evaluated genotoxicity in S. typhimurium
TA100 in the vapor state and found trichloroacetyl chloride to be positive in the presence and
absence of S9 activation, but inducing predominantly GC-to-TA transversions (the predominant
background mutation). Trichloroacetyl chloride was negative for prophage induction in E. coli
in the same study (DeMarini et al., 1994).
4.8.6. Tetrachloroethylene (PCE) Epoxide
Tetrachloroethylene epoxide, a hypothesized intermediate in tetrachloroethylene P450
oxidative metabolism (Henschler and Bonse, 1977; Henschler, 1977), has been investigated in
only one published study. This study is described below and listed in Table 4-48.
4.8.6.1. Bacterial Systems
4.8.6.1.1. Gene mutation
In a study examining the genotoxicity of multiple chloroepoxides, tetrachloroethylene
epoxide (0, 0.5, 1.3, 2.5, 5.0, 25.0 mM, closed system) was mutagenic in S. typhimurium
TA1535 but not in E. coli WP2 uvrA (Kline etal., 1982). Mutagenicity was observed at the
lower doses in S. typhimurium, but not at higher doses, most likely due to cytotoxicity at the high
doses.
4-401
-------
4.8.7. Trichloroethanol (TCOH)
4.8.7.1. Bacterial Systems
4.8.7.1.1. Gene mutation
Limited studies are available on the effect of TCOH on genotoxicity (refer to Table
4-47). TCOH is negative in the S. typhimurium assay using the TA100 strain (DeMarini et al.,
1994: Bignami et al.. 1980: Waskell. 1978). A study by Beland (1999) using S. typhimurium
strain TA104 did not induce reverse mutations without exogenous metabolic activation,
however, did increase mutant frequency in the presence of exogenous metabolic activation at a
dose above 2,500 jig/plate. TCOH has not been evaluated in other recommended screening
assays. Therefore, the database is limited for the determination of TCOH genotoxicity
(summarized in Table 4-48).
4.8.8. Ł-(l,2,2-Trichlorovinyl)-L-Cysteine (1,2-TCVC), Ł-Trichlorovinyl Glutathione
(TCVG), 7V-Acetyl-S-(l,2,2-Trichlorovinyl)-L-Cysteine (NAcTCVC)
Limited studies have been performed examining the genotoxicity of three metabolites
from the GSH-conjugation metabolic pathway of tetrachloroethylene. The results for all three
are described below and summarized in Table 4-48.
4.8.8.1. Bacterial Systems
4.8.8.1.1. Gene mutation
TCVG produced from tetrachloroethylene in isolated perfused rat liver and excreted into
bile, in the presence of a rat kidney fraction, was mutagenic in Salmonella, as was purified
TCVG (Vamvakas et al., 1989d). This study performed the Ames assay in S. typhimurium
TA100, TA98, and TA2638 with tetrachloroethylene, TCVG, and bile from liver perfusate
following tetrachloroethylene exposure in rats, demonstrated that the GST-metabolites or
tetrachloroethylene in the presence of bile containing GST led to gene mutations in
S. typhimurium TA100. Dreessen et al. (2003) also demonstrated for TCVG an unequivocal
dose-dependent mutagenic response in the TA100 strain in the presence of the rat kidney
S9-protein fraction; TCVC was mutagenic without metabolic activation in this strain. In a
separate study, the tetrachloroethylene metabolite TCVC (1-10 nmol/plate) was also positive in
Salmonella strains TA98 and TA100 but not strain TA2638, and inhibition of p-lyase activity
was blocked by the addition of aminooxyacetic acid (AOAA) (Dekant et al., 1986a). A
subsequent study from this same group indicated that Salmonella also were capable of
deacetylating the urinary metabolite NAcTCVC (50-100 nmol/plate) when TA100 showed a
clear positive response in the Ames assay without exogenous activation (Vamvakas et al., 1987).
4-402
-------
Addition of cytosolic protein increased this mutagenicity, while addition of a p-lyase inhibitor
(AOAA) decreased it.
4.8.8.2. Mammalian Systems
4.8.8.2.1. Unscheduled DNA synthesis
Vamvakas et al. (1989a) reported concentration-related increases in unscheduled DNA
synthesis (UDS) in LLC-PK1 (a porcine kidney cell line) exposed to TCVC, with the effect
abolished by a P-lyase inhibitor. This effect was observed at exposure to 5 x 1CT6-1CT5 M
TCVC for 24 hours. This study also measured LDH release to determine cytotoxicity at the
same doses, and no increases in LDH were observed at these doses.
4.8.9. TCVC Sulfoxide
TCVC sulfoxide does not appear to have been investigated for genotoxicity.
4.8.10. Synthesis and Overall Summary
Tetrachloroethylene and its metabolites (TCA, DCA, CH, TCVC, TCVG, and
NAcTCVC) have been evaluated to varying degrees for their genotoxic activity in several of in
vitro systems such as bacteria, yeast, and mammalian cells and, also, in in vivo systems.
Genotoxicity studies of other metabolites (e.g., TCVC sulfoxide, tetrachloroethylene epoxide,
trichloroacetyl chloride, trichloroethanol) are limited or nonexistent but are discussed where
available.
The results of a large number of in vitro genotoxicity tests in which tetrachloroethylene
was the test agent do not clearly support the conclusion that tetrachloroethylene exhibits direct
mutagenic activity in the absence or presence of the standard S9 fraction [(Watanabe et al., 1998;
DeMarini et al.. 1994: Roldan-Ariona et aL 1991: Milman et al.. 1988: Warner etal.. 1988:
NTP. 1986: Connor etal.. 1985: Shimada et al.. 1985: Haworth et al.. 1983: Hardinetal.. 1981:
Kringstad et al.. 1981: Bartsch et al.. 1979: Greimetal.. 1975): summarized in Table 4-40]. A
more recent study demonstrated cytotoxicity but not genotoxicity of tetrachloroethylene in an
S. typhimurium strain (YG7108pin3ERb5) with enhanced metabolic activity (transformed with
CYP2E1, cytochrome P450 reductase, and cytochrome b5) (Emmert et al., 2006). PCE was
negative in the parent strain (YG7108) at all doses in the presence of S9. However, when
tetrachloroethylene was activated with rat liver GST, GSH, and a rat kidney fraction,
tetrachloroethylene exhibited a clear dose response (Vamvakas et al., 1989d). These findings
support a role of metabolic activation of tetrachloroethylene in its in vitro genotoxicity.
4-403
-------
Table 4-48. Genotoxicity of additional tetrachloroethylene metabolites—all
systems
Metabolite
Chloral
Oxalic acid
Trichloroethanol
(TCOH)
Trichloroacetyl
chloride
Trichlorovinyl-
glutathione
(TCVG)
Test system/endpoint
S. typhimurium TA100, increased
mutation frequency
Sclerotinia sclerotiorum, DNA
fragmentation
Madin-Darby cultured canine
kidney cells, renal prothrombin
fragment- 1 mRNA expression
Crepis capillaris, chromosomal
aberrations
S. typhimurium TA100, 98,
reverse mutation
S. typhimurium TA100, reverse
mutation
S. typhimurium TA104, reverse
mutation
S. typhimurium TA100, 1535
reverse mutation
Sister chromatid exchanges
PRB, A, Prophage induction, E.
coli WP2
SAO, S. typhimurium TA100,
reverse mutation
S. typhimurium TA100, increased
mutation frequency
S. typhimurium TA100, reverse
mutation
S. typhimurium TA100, increased
mutation frequency
Cultured porcine LLC-PK1
(kidney) cells, unscheduled DNA
synthesis, in vitro
Doses
(LED or
HID)a
NA
10 mM
0.09 mM
1.0 mM
7,500 ug/plate
0.5 ug/cm3
vapor
2,500 ug/plate
NA
NA
10,000
2.6
5 ug/mL
100 nmol/plate
25 nmol/plate
(with)
250-500
nmol/plate
(without)
7.5 x 10~6M
Results'1
With
activation
+
NT
NT
NT
-
-
+
-
NA
-
+
-
+
+
NT
Without
activation
possible
+
+
(+)
-
-
-
-
+
-
+
-
-
(+)
+
Reference
Sato et al.
(1985)
Kim et al.
(2008)
Moryama et al.
(2005)
Shevchenko et
al. (1985)
Waskell (1978)
DeMarini et al.
(1994)
Beland (1999)
Bignami et al.
(1980)
Gu et al. (1981)
DeMarini et al.
(1994)
DeMarini et al.
(1994)
Reichert et al.
(1983)
Dreessen et al.
(2003)
Vamvakas et al.
(1989c)
Vamvakas et al.
(1989d)
4-404
-------
Table 4-48. Genotoxicity of additional tetrachloroethylene metabolites—all
systems (continued)
Metabolite
Trichlorovinyl-
cysteine (TCVC)
NAcTCVC
PCE oxide
Test system/endpoint
S. typhimurium TA100, reverse
mutation
Cultured porcine LLC-PK1
(kidney) cells, unscheduled DNA
synthesis, in vitro
S. typhimurium TA100, increased
mutation frequency
S. typhimurium TA1535, reverse
mutation
E. coli WP2 uvrA, reverse
mutation
Doses
(LED or
HID)a
50 nmol/plate
5 x 1(T6 M
<50 nrnol0
2.5 mM
25 mM
Results'1
With
activation
NT
NT
+
NT
NT
Without
activation
+
+
+
+
—
Reference
Dreessen et al.
(2003)
Vamvakas et al.
Q989a)
Vamvakas et al.
(1987)
Kline et al.
(1982)
Kline et al.
(1982)
aLED, lowest effective dose; HID, highest ineffective dose; NA = not available.
bResults: + = positive; (+) = weakly positive; - = negative; NT = not tested.
'Lower-level concentrations that indicate mutagenicity are not specified in Vamvakas et al. (1987).
Limited in vivo studies of tetrachloroethylene are inconsistent, with only negative (NTP,
1986: Bronzetti et al.. 1983) or equivocal (Cederberg etal.. 2010a: Bellies et al.. 1980)
genotoxicity assay results demonstrated following inhalation or oral exposure to
tetrachloroethylene in animals (refer to Table 4-39). Intraperitoneal injection assays have
demonstrated both negative (NTP, 1986) as well as positive results for different genotoxicity
endpoints (Walles, 1986). Assays of clastogenic effects following inhalation exposure in
humans have shown inconsistent results and are suggested to be related to coexposures (Seiji et
al., 1990; Ikeda et al., 1980). Studies of chromosomal aberrations following exposure to
tetrachloroethylene are mostly negative (Galloway et al., 1987: NTP, 1986: Sofuni et al., 1985),
but positive results have been observed in vivo (Murakami andHorikawa, 1995) and in vitro
studies with enhanced metabolic activation (Doherty et al., 1996).
TCA, an oxidative metabolite of tetrachloroethylene, exhibits little, if any, genotoxic
activity in vitro (refer to Tables 4-41 and 4-42). TCA did not induce mutations in S.
typhimurium strains in the absence of metabolic activation or in an alternative protocol using a
closed system (Kargalioglu et al., 2002: Nelson etal., 200 la: Giller etal., 1997: DeMarini et al.,
1994: Rapson et al., 1980: Waskell, 1978), but a mutagenic response was induced in TA100 in
the Ames fluctuation test (Giller et al., 1997). However, in vitro experiments with TCA should
be interpreted with caution if steps have not been taken to neutralize pH changes caused by the
4-405
-------
compound (Mackay et al., 1995). Measures of DNA-repair responses in bacterial systems have
shown induction of DNA repair reported in S. typhimurium but not in E. coli. Mutagenicity in
mouse lymphoma cells was only induced at cytotoxic concentrations (Harrington-Brock et al.,
1998). TCA was positive in some genotoxicity studies in in vivo mouse, newt, and chick test
systems (Giller et al., 1997; Bhunya and Jena, 1996; Birner et al., 1994; Bhunya and Behera,
1987). DNA unwinding assays have either shown TCA to be much less potent than DCA
(Nelson and Bull 1988) or negative (Styles et al.. 1991: Nelson etal.. 1989). Due to limitations
in the genotoxicity database, the possible contribution of TCA to tetrachloroethylene
genotoxicity is unclear.
DCA, a chloroacid metabolite of tetrachloroethylene, has also been studied using
different types of genotoxicity assays (refer to Tables 4-43 and 4-44). Although limited studies
are conducted for different genetic endpoints, DCA has been demonstrated to be mutagenic in
the S. typhimurium assays, in vitro (Kargalioglu et al., 2002; Plewa et al., 2002; DeMarini et al.,
1994) in some strains, in a mouse lymphoma assay (Harrington-Brock et al., 1998), in vivo
cytogenetic tests (Leavitt et al., 1997; Fuscoe etal., 1996), in the micronucleus induction test,
using the Big Blue mouse system, and in other tests (Harrington-Brock et al., 1998; Leavitt et al.,
1997: Fuscoe etal., 1996: DeMarini et al., 1994: Chang etal., 1989: Nelson etal., 1989: Nelson
and Bull, 1988: Gu et al., 1981). DCA can cause DNA strand breaks in mouse and rat liver cells
following in vivo exposure in mice and rats (Fuscoe et al., 1996). Because of uncertainties as to
the extent of DCA formed from tetrachloroethylene exposure, inferences as to the possible
contribution from DCA genotoxicity to tetrachloroethylene toxicity are difficult to make.
Chloral hydrate is mutagenic in the standard battery of screening assays (refer to Tables
4-45, 4-46, and 4-47). Effects include positive results in bacterial mutation tests for point
mutations and in the mouse lymphoma assay for mutagenicity at the Tk locus (Haworth et al.,
1983). In vitro tests showed that CH also induced micronuclei and aneuploidy in human
peripheral blood lymphocytes and Chinese hamster pulmonary cell lines. Micronuclei were also
induced in Chinese hamster embryonic fibroblasts. Several studies demonstrate that chloral
hydrate induces aneuploidy (loss or gain of whole chromosomes) in both mitotic and meiotic
cells, including yeast (Gualandi, 1987: Sora and Agostini Carbone, 1987: Kafer, 1986: Singh and
Sinha, 1979, 1976), cultured mammalian somatic cells (Degrassi and Tanzarella, 1988), and
spermatocytes of mice (Liang and Pacchierotti, 1988: Russo etal., 1984). Chloral hydrate was
negative for sex-linked recessive lethal mutations in Drosophila (Yoon et al., 1985). It induces
SSB in hepatic DNA of mice and rats (Nelson and Bull, 1988) and mitotic gene conversion in
yeast (Bronzetti etal., 1984). Schatten and Chakrabarti (1998) showed that chloral hydrate
affects centrosome structure, which results in the inability to reform normal microtubule
formations and causes abnormal fertilization and mitosis of sea urchin embryos. Based on the
4-406
-------
existing array of data, CH has the potential to be genotoxic, particularly when aneuploidy is
considered in the weight of evidence for genotoxic potential. Chloral hydrate appears to act
through a mechanism of spindle poisoning, resulting in numerical changes in the chromosomes.
These results are consistent with tetrachloroethylene, albeit there are limited data on
tetrachloroethylene for these genotoxic endpoints.
The genotoxicity analysis of other metabolites (e.g., trichloroacetyl chloride,
tetrachloroethylene epoxide, trichloroethanol) is limited (refer to Table 4-48). Trichloroacetyl
chloride was found to be mutagenic in S. typhimurium when exposed in vapor phase (DeMarini
etal., 1994) but not in liquid phase (Reichert et al., 1983): tetrachloroethylene epoxide was
found to be mutagenic in S. typhimurium but not E. coli (Kline et al., 1982): and trichloroethanol
was found to be negative in three (DeMarini etal., 1994: Bignami etal., 1980: Waskell, 1978) of
four mutagenicity studies (Beland, 1999). These results are limited, and further studies are
needed to make any conclusions on the genotoxicity of these metabolites.
Although also limited, genotoxicity tests for the GSH conjugation metabolites are
positive (refer to Table 4-48). These include 1,2-TCVC, TCVG, and NAcTCVC. In the one
mammalian study, unscheduled DNA synthesis in porcine kidney cells was observed to increase
in a dose-dependent manner following exposure to TCVC (Vamvakas et al., 1989c).
Mutagenicity assays found TCVG (Dreessen et al., 2003: Vamvakas et al., 1989d) and
NAcTCVC (Vamvakas et al., 1987) to be mutagenic in the presence of activation, while TCVC
was mutagenic even in the absence of activation (Dreessen et al., 2003: Dekant etal., 1986a).
In summary, tetrachloroethylene has been shown to induce some genotoxic effects
(micronuclei induction following in vitro exposure, DNA binding, and SSBs in tumor tissue), but
these result are inconsistent. A number of in vitro mutagenicity (Ames) tests of
tetrachloroethylene have largely been negative in the absence or presence of the standard S9
fractions. Positive results have been observed in tests of conditions where metabolites of the
GSH pathway are generated. These support a role of metabolic activation of tetrachloroethylene
in its genotoxicity. Consistent with this view, positive results have been reported when the GSH
metabolites were used as the test agent, and certain of the oxidative metabolites are also
mutagenic. TCVC is the most potent bacterial mutagen of the tetrachloroethylene metabolites
and induces UDS in a porcine kidney cell line; TCVG and NAcTCVC are also mutagenic in
bacteria.
There are several challenges in interpreting the genotoxicity results obtained from
tetrachloroethylene exposure. Because of the volatile nature of tetrachloroethylene, there could
be false negative results if proper precautions are not taken to limit evaporation, such as the use
of a closed sealed system. The adequacy of the enzyme-mediated activation of
tetrachloroethylene in vitro tests is another consideration. For example, it is not clear if standard
4-407
-------
S9 fractions can adequately recapitulate the complex in vivo metabolism of tetrachloroethylene
to reactive intermediates, which, in some cases, entails multiple sequential steps involving
multiple enzyme systems (e.g., CYP, GST, etc.). In addition, the relative potency of the
metabolites in vitro may not necessarily inform their relative contribution to the overall
mechanistic effects of the parent compound, tetrachloroethylene. Furthermore, although
different assays provided data relevant to different types of genotoxic endpoints, not all effects
that are relevant for carcinogenesis are encompassed. The standard battery of prokaryotic as
well as mammalian genotoxicity test protocols typically specify the inclusion of significantly
cytotoxic concentrations of the test compound.
In conclusion, uncertainties with regard to the characterization of tetrachloroethylene
genotoxicity remain. This is primarily because in vivo tests of tetrachloroethylene have been
equivocal, with at most, modest evidence of genotoxic effects in rodent tumor tissues examined
(including mouse liver and rat kidney) following exposure at tumorigenic doses. However, no
evidence is available regarding the potential contribution of tetrachloroethylene genotoxicity to
other rodent tumor types (particularly, MCL, testes, and brain). Ames assays of
tetrachloroethylene have yielded largely negative results. The tetrachloroethylene metabolites
TCVG, TCVC, NAcTCVC, tetrachloroethylene oxide, and DCA are genotoxic, but not all such
metabolites have been sufficiently tested in the standard screening battery to support clear
conclusions about their genotoxic potential. However, the predominance of positive data for
these metabolites supports their potential genotoxicity following in situ production and/or
bioactivation. This, in turn, supports the view that contribution of genotoxicity to
tetrachloroethylene carcinogenesis cannot be ruled out for one or more target organs. Additional
testing of the genotoxicity of tetrachloroethylene and its metabolites (particularly those from the
GSH conjugation pathway) using state-of-the-art methods and in a more comprehensive panel
of tumor tissues is warranted.
4.9. SUSCEPTIBLE POPULATIONS
Variation in response to tetrachloroethylene may be due to age, gender, genetics, and
race/ethnicity, as well as differences in lifestyle factors, nutrition, preexisting disease status,
socioeconomic status, and multiple exposures. These could be potential modifying risk factors
that play an important role in determining an individual's susceptibility to chemical exposures
and are discussed below.
4.9.1. Life-Stages
Individuals in one life-stage are physiologically, anatomically, and biochemically unique
from individuals in another life-stage. Early and later life-stages differ greatly from mid-life-
4-408
-------
stages in body composition, organ function, and many other physiological parameters that can
influence the toxicokinetics of parent chemicals and their metabolites from the body (Guzelian et
al., 1992b). This section presents and evaluates the pertinent published literature available to
assess how individuals of early life-stages (refer to Section 4.9.1.1) and later life-stages (refer to
Section 4.9.1.2) may respond differently to tetrachloroethylene than adults. The limited data on
tetrachloroethylene exposure suggest that these populations—particularly individuals in early
life-stages—may have greater susceptibility than does the general population.
4.9.1.1. Early Life-Stages
4.9.1.1.1. Early life-stage-specific exposures
Section 2.2 describes the various exposure routes of concern for tetrachloroethylene. For
all postnatal life-stages, the primary exposure routes of concern include inhalation (refer to
Section 2.2.1) and contaminated water (refer to Section 2.2.2). Ingestion of contaminated food
or soil is also a possible exposure route (refer to Section 2.2.3), as is direct ingestion (refer to
Section 2.2.6). In addition, certain exposure pathways to tetrachloroethylene are unique to early
life-stages, such as through placental transfer or via breast milk ingestion (refer to Section 2.2.4),
or may be increased during early or later life-stages. Other reviews of the reproductive and
developmental effects of tetrachloroethylene exist (Brown Dzubow et al., 2010; Beliles, 2002;
Bove et al.. 2002: Danielsson, 1990: van der Gulden and Zielhuis, 1989: Tabacova, 1986).
Prenatal. In utero, lipophilic substances are known to cross the placental barrier (Herrera
et al., 2006). There is biological plausibility of transfer of tetrachloroethylene across the human
placental barrier as tetrachloroethylene has been measured in fetal blood and amniotic fluid in
rodents (Szakmary et al., 1997: Ghantous et al., 1986). Fetal blood concentrations have been
modeled for human exposure (Gentry et al., 2003).
Inhalation. Inhalation exposures may be altered for early life-stages compared to adults,
because children have increased ventilation rates (both intake and exhalation) per kg body
weight compared to adults (U.S. EPA, 2008: NRC, 1993). These populations spend the majority
of their time indoors (Bateson and Schwartz. 2008: U.S. EPA. 2008: NRC, 1993). where
increased concentrations of tetrachloroethylene have been found compared to those measured
outdoors (U.S. EPA, 200la). Increased indoor air concentrations have been measured in places
where children may spend time: inside apartments containing dry-cleaned clothing (Thomas et
al., 1991: Tichenor et al., 1990), in the homes of dry-cleaning employees (ATSDR, 1997a:
Aggazzotti et al., 1994a: Aggazzotti etal., 1994b), in apartments above or adjacent to dry
cleaners (Storm etal.. 2011 [previously reported in NYSDOH. 20101: McDermott et al.. 2005:
Schreiber et al.. 2002: Garetano and Gochfeld. 2000: Chi en. 1997: Altmann et al.. 1995:
Schreiber, 1993: Verberk and Scheffers, 1980), in day care centers adjacent to dry cleaners
4-409
-------
(NYSDOH, 2005b), in a classroom exposed to tetrachloroethylene from an air "emission from a
small chemical factory" (Monster and Smolders, 1984), and in automobiles containing dry-
cleaned clothing (Park et al., 1998; Gulyas and Hemmerling, 1990). Similarly, increased
ambient air concentrations have been measured in places where children may spend time: outside
of a day care center adjacent to a dry cleaner (NYSDOH, 2005c), and on a playground near a
factory (Monster and Smolders, 1984). Adgate and colleagues (Adgate et al., 2004b: Adgate et
al., 2004a) measured tetrachloroethylene in outside and indoor air at school, indoor air at home,
and using personal samplers on children, and demonstrated that tetrachloroethylene levels are
lower in homes with greater ventilation and in homes in nonurban settings (Adgate et al., 2004b:
Adgate et al., 2004a). In addition, inhalation may also occur indoors during showering or
bathing as dissolved tetrachloroethylene in the warm tap water is volatilized, although dermal
exposure is also relevant during these scenarios (Rao and Brown, 1993).
Ingestion. Due to its lipophilicity, tetrachloroethylene has been found in human breast
milk samples (Schreiber et al.. 2002: U.S. EPA. 200la: Schreiber, 1997. 1993: Sheldon et al..
1985: Pellizzari et al., 1982: Bagnell and Ellenberger, 1977), as well as in milk from cows
(Wanner et al., 1982), goats (Hamada and Tanaka, 1995), and rats (Byczkowski et al., 1994:
Byczkowski and Fisher, 1994). The breast milk of one woman was found to contain 10 mg/L
tetrachloroethylene 1 hour following a visit to her spouse working at a dry-cleaning
establishment, dropping to 3 mg/L after 24 hours (Bagnell and Ellenberger, 1977).
Tetrachloroethylene has also been measured in the breast milk of two women living in
apartments colocated with a dry-cleaning facility (NYSDOH, 2005c: Schreiber et al., 2002).
PBPK models have been used to estimate the dose a nursing infant might receive from an
exposed mother's breast milk (Gentry et al., 2003: Fisher etal., 1997: Byczkowski and Fisher,
1995: Bvczkowski et al., 1994: Schreiber, 1993). A PBPK model was also developed and
validated for breast milk ingestion in nursing rats after maternal inhalation exposure (Fisher,
1994). Using different exposure scenarios, Schreiber (1993) predicted that breast milk
concentrations could range from 1.5 ug/L for a typical residential scenario, 16-3,000 ug/L for a
residential scenario near a dry cleaner, to 857-8,440 ug/L for an occupational scenario.
Assuming that a 7.2-kg infant ingests 700 mL of breast milk per day, Schreiber estimated dose to
the infant could range from 0.0001 to 0.82 mg/kg-day (Schreiber et al., 1993). Byczkowski and
Fisher (1995) refined the approach used by Schreiber (1993) and found that with the same
residential exposure conditions, the results predicted lower doses to the infant (0.0009-0.202
mg/kg-day). Using milk production and suckling variables, Fisher et al. (1997) estimated the
dose that a human infant might receive after maternal occupational exposure to be 25 ppm/day.
Gentry et al. (2003) modeled a rapid decline in concentration of tetrachloroethylene and TCA
during lactation in humans. Although ingestion of tetrachloroethylene through breast milk may
4-410
-------
be a significant pathway of exposure for some infants, it has been suggested that if these infants
live adjacent to or in close proximity of dry-cleaning facilities, the dose received through
ingestion of breast milk will become less important when compared with the dose resulting from
inhalation exposure (Schreiber, 1997; McKone and Daniels, 1991).
Children ingest higher amounts of water per body weight than adults (U.S. EPA, 2008;
NRC, 1993). For infants on formula, ingestion of tetrachloroethylene-contaminated water may
be of concern. Taking into account tetrachloroethylene volatilization in boiling water,
Letkiewicz et al. (1982) estimated that 22% of formula-fed infants received fluids contaminated
with tetrachloroethylene levels found in the water supply. Data showed that about 11%
(0.5 x 22%) of formula-fed infants could receive an increased exposure as compared with adults
on a mg/kg-basis through drinking contaminated water. In addition, incidental water
consumption may occur for children when swimming or bathing (U.S. EPA, 2008).
Children consume a higher quantity of food per body weight compared to adults,
specifically dairy and other foods with high fat content (U.S. EPA, 2008) that have been found to
have elevated concentrations of tetrachloroethylene (refer to Section 2.2.3). Assuming 100
mg/kg represents the average tetrachloroethylene concentration in fatty foods such as butter, and
using daily total fat intake rates by age (U.S. EPA, 2008), the daily dose would be 0.46
mg/kg-day for a 10 kg 1-year-old compared to the daily dose of 0.12 mg/kg-day for a 70 kg
adult. Therefore, there may be concern for ingestion of contaminated dairy products in early
life-stages, although this exposure route for tetrachloroethylene has not been well characterized
for any life-stage.
Where contamination occurs, tetrachloroethylene can be measured in soil (U.S. EPA,
200 la). This pathway for ingestion of tetrachloroethylene has not been directly examined. A
clear need exists to evaluate this pathway because children, particularly those with pica, can
ingest high quantities of contaminated soil through hand-to-mouth activity, as has been shown
for lead (U.S. EPA, 2008).
Rare instances of direct ingestion of tetrachloroethylene have been documented,
including a 6-year-old boy who directly ingested 12-16 g of tetrachloroethylene (Koppel et al.,
1985).
Dermal. Dermal exposures may be increased for both early life-stages, because infants
have increased surface area-per-body weight-ratio than adults (U.S. EPA, 2008; NRC, 1993).
Although an infant's skin has similar permeability to adults, a premature infant may have
increased permeability (Guzelian et al., 1992a). Dermal exposure for children may occur in a
residential setting from showering, bathing, or swimming in contaminated water, although
inhalation exposure is also relevant during these scenarios (U.S. EPA, 200la: Rao and Brown,
4-411
-------
1993). While dermal exposure is generally not considered a major route of exposure, this route
of exposure is not well characterized for early life-stages (prenatal or postnatal).
4.9.1.1.2. Early life-stage-specific toxicokinetics
Section 3 describes the toxicokinetics of tetrachloroethylene. However, children may
have differential exposure to tetrachloroethylene compared to adults due to age-related
physiological differences. These include body composition, organ function, and many other
physiological parameters that can influence the toxicokinetics of chemicals and their metabolites
from the body (Renwick, 1998; Guzelian et al., 1992b). Early life-stage-specific information
regarding toxicokinetics needs to be considered for a child-specific and chemical-specific PBPK
model. To adequately address the risk to infants and children, age-specific parameters for these
values should be used in PBPK models that can approximate the internal dose an infant or child
receives based on a specific exposure level [(Clewell et al., 2004; Gentry et al., 2003;
Byczkowski and Fisher, 1994; Rao and Brown, 1993): refer to Section 3.5].
Absorption. As discussed in Section 3.1, exposure may occur via inhalation, ingestion,
and skin absorption. In addition, prenatal exposure may result in absorption via the
transplacental route. For lipophilic compounds such as tetrachloroethylene, percentage adipose
tissue, which varies with age (NRC, 1993), will affect absorption and retention of the absorbed
dose. Absorption into the lungs via inhalation is related to the ventilation rate per body weight,
which is higher in children than in adults (U.S. EPA. 2008: WHO. 2006: NRC, 1993). with an
increased alveolar surface area per kg body weight for the first 2 years (NRC, 1993). Absorption
into the gut from oral ingestion may be altered by gastric pH levels, which are higher in infants
than in adults (WHO, 2006). Absorption during dermal exposure may be affected by the ratio of
surface area, which is higher in infants than in adults (U.S. EPA. 2008: WHO. 2006).
Distribution. The distribution of tetrachloroethylene to specific organs will depend on
organ blood flow and the lipid and water content of the organ, which may vary between life-
stages (WHO, 2006: NRC, 1993). Due to its high lipophilicity, tetrachloroethylene has been
found to distribute widely to all tissues in the body as observed in early lifestages of humans
(Gamier et al., 1996: Gaillard et al., 1995: Koppel et al., 1985) and early lifestages of animals
(Szakmary et al., 1997: Dallas etal.. 1994a: Ghantous et al.. 1986: Schumann et al.. 1980:
Savolainen et al., 1977b): however, this is true for adults as well, and it is not clear whether
distribution may vary differentially with life-stage. It should be noted that the total body burden
of tetrachloroethylene increases with age (Clewell et al., 2004), as would be expected, given that
adult body weight is generally positively correlated with age.
Rodent studies demonstrate that tetrachloroethylene crosses the placental barrier when
pregnant dams are exposed (Szakmary et al., 1997: Ghantous et al., 1986), and in humans, it has
4-412
-------
been shown that during lactation, tetrachloroethylene distributes to breast milk (NYSDOH,
2005c: Schreiber et al., 1993; Sheldon et al., 1985). However, a noticeable difference exists
between the milk:blood partition coefficients for rats (12) and for humans (2.8) (Byczkowski and
Fisher, 1994), reflecting the higher fat content of rat milk.
Tetrachloroethylene or its metabolites have been measured in blood of children (Storm et
al.. 2011 [previously reported in NYSDOH. 20101: NYSDOH. 2005a: Poppet al.. 1992). A
longitudinal study of blood concentrations of 11 volatile organic chemicals (VOCs) measured in
more than 150 poor, minority children in Minneapolis, MN, found the mean blood
tetrachloroethylene levels to be 0.06 ng/mL (Sexton et al., 2005). When compared to adult data
from NHANES III, the blood level in children was lower (Sexton et al., 2005). However, these
results do not necessarily represent TK differences between lifestages because the study did not
control for exposure differences between these two cohorts. Lower estimated blood
concentrations of tetrachloroethylene in children compared to adults have also been described in
Clewell et al. (2004), although the variability of the parameters used as well as the results have
not been validated.
Tetrachloroethylene can also cross the blood:brain barrier during both prenatal and
postnatal development; this may occur, to a greater extent, in younger children. Based on the
modeled dose of tetrachloroethylene to the brain after a showering/bathing scenario, a study by
Rao and Brown (1993) showed that for a given set of exposures, the younger a person is, the
greater the estimated concentration of tetrachloroethylene in the brain. Modeling showed that
after a 30-minute bathing scenario, a 3-year-old child accumulated higher brain tissue
concentrations of tetrachloroethylene as compared with a 10-year-old and an adult. An autopsy
conducted on the previously mentioned 2-year-old boy found dead after exposure to dry-cleaned
curtains revealed the highest levels of tetrachloroethylene in the brain, 77 mg/kg. Levels in his
blood, heart, and lungs were 66 mg/L, 31 mg/kg, and 46 mg/kg, respectively (Gamier et al.,
1996:Gaillardetal., 1995).
Metabolism. Section 3.3.3 describes the enzymes involved in the metabolism of
tetrachloroethylene. In general, expression of CYP enzymes changes during various stages of
fetal development (Hakkola et al., 1998b: Hakkola et al., 1996a: Hakkola et al., 1996b) and
during postnatal development (Shao et al., 2007: Clewell et al., 2004: Hakkola et al., 1998a:
Hakkola et al., 1998b: Tateishi et al., 1997: Hakkola et al., 1996a: Hakkola et al., 1996b: George
et al., 1995). In addition, production of GST enzymes varies significantly during early postnatal
lifestages (Shao et al., 2007: McCarver and Hines, 2002: Dome etal., 2001: Raijmakers et al.,
2001: Nakasaetal., 1997: Meraetal., 1994).
After maternal oral exposure to tetrachloroethylene it was observed that fetus and infant
blood levels were higher for TCA than for tetrachloroethylene (Gentry et al., 2003),
4-413
-------
demonstrating that metabolism of tetrachloroethylene does occur during these lifestages. In
addition, there is in vitro evidence of an age-related increase in metabolism of
tetrachloroethylene as estimated in the blood (Clewell et al., 2004; Sarangapani et al., 2003),
associated with age-related activation of oxidative metabolism pathways, suggesting a decreased
ability to metabolize tetrachloroethylene during early lifestages compared to during adulthood.
One study modeled the role of the age-dependent development of CYP2E1 in oxidative
metabolism (TCA) in the mother and lactating infant (Vieira et al., 1996). A number of other
human studies suggest that CYP2B6 may also play a role in the metabolism of
tetrachloroethylene (White et al., 2001), although this enzyme was not detected in placental or
fetal liver samples (Hakkola et al., 1996a: Hakkola et al., 1996b), and differences between a
group of 10 prenatal and infant patients showed significantly lower CYP2B6 protein expression
in placental hepatic microsomes as compared with an adult group (Tateishi et al., 1997). These
findings need to be validated in studies of target tissues in addition to blood to better evaluate
any role of variation and heterogeneity.
Excretion. The major processes of excretion of tetrachloroethylene and its metabolites
are discussed in Sections 3.3 and 3.4, respectively. Excretion profile differences in exhaled
breath and urinary excretion are likely between children and adults. This is due to differences in
ventilation rate, activity level, and the solubility of the compound in blood and tissue, as well as
differences in amounts of water ingested per body weight (U.S. EPA, 2008; NRC, 1993).
Tetrachloroethylene or its metabolites have been measured in exhaled breath (Storm et
al., 2011 [previously reported in NYSDOH, 20101: NYSDOH, 2005a: Delfmo et al., 2003b:
Schreiber et al., 2002: Monster and Smolders, 1984), and urine (NYSDOH, 2005c: Schreiber et
al., 2002: Popp etal., 1992) of children. However, these studies do not provide clear information
whether excretion levels in children differ from those of adults for a similar exposure
concentration.
PBPKModels. A number of PBPK models present toxicokinetic variation between early
lifestages and adulthood for tetrachloroethylene and its metabolites for both humans and animals.
Early lifestage-specific exposure scenarios considered in these models include fetal exposure
(Gentry et al., 2003) and breast milk exposure (Gentry et al., 2003: Fisher et al., 1997:
Bvczkowski and Fisher, 1995: Byczkowski et al., 1994: Schreiber et al., 1993). Other PBPK
models have addressed comparisons of early lifestage toxicokinetics with those in adulthood for
inhalation (Mahle et al., 2007: Rodriguez et al., 2007: Sarangapani et al., 2003: Pelekis et al.,
2001), drinking water (Clewell et al., 2004), and bathing and showering (Rao and Brown, 1993).
When considering inhalation exposure, Mahle et al. (2007) found no difference in the blood:air
partition coefficient for tetrachloroethylene for children aged 3-10 years compared to adults
(420 years old). This same study reported that rats at PND 10 and at 2 months (adult) have an
4-414
-------
age-dependent difference in fatair, muscle:air, and brain:air partition coefficients, but not for
blood:air, liver:air, or kidney :air (Mahle et al., 2007). Another study of rats found higher peak
concentrations of tetrachloroethylene in the blood at PND 10 compared to 2 months (adult) after
inhalation exposure, likely due to the lower metabolic capacity of the young rats as observed in
the liver (Rodriguez et al., 2007). Pelekis et al. (2001) found little difference in the suggested
intraspecies uncertainty factor when including lifestage-specific pharmacokinetics. Sarangapani
et al. (2003) also found no age-related difference in tetrachloroethylene blood concentration;
however, this study found that metabolite concentrations were lowest in infancy and increased
with age. For drinking water exposure, Clewell et al. (2004) found an age-related trend in the
average daily dose and cumulative lifetime dose of tetrachloroethylene and its metabolites, with
lower levels of metabolites observed in children compared to higher levels of metabolites
observed in adulthood. In a showering/bathing scenario, Rao and Brown (1993) found that
tetrachloroethylene accumulates in the brain at higher levels in younger versus older children.
Validation and further refinement of the parameters in these PBPK models are necessary, in
particular, modeling of fetal and breast milk exposure, and child-adult differences in partition
coefficients after inhalation, drinking water, and bathing scenarios.
4.9.1.1.3. Early life-stage-specific effects
Although limited data exist on tetrachloroethylene toxicity as it relates to early life-
stages, there is enough information to discuss the qualitative differences. In addition to the
evidence described below, Section 4.7 contains information on both human and animal evidence
for reproductive and developmental outcomes such as spontaneous abortion/fetal loss, low birth
weight, IUGR, SGA, congenital abnormalities, sperm quality, developmental delays, and
behavioral changes. Together, Section 4.4 on liver toxicity, Section 4.5 on kidney toxicity,
Section 4.6 on neurotoxicity, and Section 4.8 on toxic effects in other organ systems characterize
a wide array of postnatal developmental effects.
4.9.1.1.3.1. Preconception
Exposures occurring prior to conception may result in adverse reproductive outcomes.
For tetrachloroethylene exposure, adverse outcomes assessed prior to conception include reduced
fertility, altered sperm, and altered reproductive hormones.
Fertility. In humans, limited evidence exists on impacts to fertility. A study of couples
seeking treatment for infertility found that employment in dry cleaning was significantly
associated with infertility among women but not among men, although exposure to
tetrachloroethylene was inferred but not documented (Rachootin and Olsen, 1983). Another
study observed no impacts on the number of pregnancies or fertility ratio among wives of men
employed as dry cleaners compared to wives employed as laundry workers, although wives of
4-415
-------
dry cleaners took longer to become pregnant compared to wives of laundry workers (Eskenazi et
al., 1991b). Other epidemiological studies have not shown any association between reduced
fertility and working in dry cleaning or exposed to tetrachloroethylene, although these results
were imprecise because the prevalence of exposure was low (Sallmen et al., 1998; Sallmen et al.,
1995). A review of the data by the National Research Council regarding exposures to
tetrachloroethylene, trichloroethylene, or solvent mixtures in drinking water at Camp Lejeune,
NC, found limited/suggestive evidence of an association for female infertility with concurrent
exposure to solvent mixtures, but inadequate/insufficient evidence to determine whether an
association exists for female infertility after exposure cessation, and inadequate/insufficient
evidence to determine whether an association exists for male infertility (NRC, 2009).
In experimental animals, a study found that the percentage of fertilized oocytes in vitro
was reduced in tetrachloroethylene-treated female rats as compared with controls, although this
study found no effect from exposure in drinking water (Berger and Horner, 2003). Other studies
in rats also found no change in fertility (Carney et al., 2006; Tinston, 1994), and one earlier study
reported an increase in fertility of female rats exposed to tetrachloroethylene (Carpenter, 1937).
Sperm. Few studies in either humans or animals have examined altered sperm quality,
generally with no observed adverse or consistent effects. Eskenazi and colleagues found that
tetrachloroethylene can have subtle effects on sperm quality (Eskenazi etal., 199la): however,
they also reported that altered sperm parameters did not appear to affect reproduction because
wives did not have fewer pregnancies as compared with a national standard (Eskenazi et al.,
1991b). A study of couples treated for infertility also examined sperm abnormalities among dry
cleaners but did not observe an elevated prevalence of sperm alterations, suggesting that the
observed reduced fertility rate among these couples was related to other reasons (Rachootin and
Olsen, 1983). One rodent study demonstrated inconsistent effects (abnormal sperm at 4 weeks
but not 1 or 10 weeks after exposure) in mice, but no adverse effect was observed in rats (Beliles,
2002). Additionally, reduced testes weight was observed in the offspring of rats after inhalation
exposure, although these were not significant after adjusting for body weight (Tinston, 1994).
Reproductive Hormones. Few studies in either humans or animals have examined altered
hormones related to reproduction, generally with no observed adverse or consistent effects. The
study discussed above of couples seeking treatment for infertility examined employment in dry
cleaning and found inconsistent results for "a female diagnosis indicating hormonal
disturbances" among three analyses (Rachootin and Olsen, 1983). An exploratory study of
menstrual disorders among dry-cleaning workers found associations with unusual cycle length,
menorrhagia, dysmenorrhea, and premenstrual syndrome, but not with oligomenorrhea,
polymenorrhea, irregular cycle, and intermenstrual blood loss (Zielhuis et al., 1989).
4-416
-------
A study of rats exposed to 1,700 ppm tetrachloroethylene did not affect progesterone
levels (Berger and Horner, 2003). The few studies on altered reproductive hormones suggest this
as an area for further research, both in females and males.
4.9.1.1.3.2. Prenatal and birth outcomes
Prenatal and birth outcomes resulting from exposure occurring prior to conception or
during fetal development include fetal death (i.e., spontaneous abortion, perinatal death), birth
defects, and decreased birth weight. It is important to note that maternal toxicity (e.g., reduced
maternal body-weight gain) may influence adverse outcomes in the offspring and was assessed
in a number of experimental animal studies of tetrachloroethylene exposure (Szakmary et al.,
1997: Narotsky and Kavlock, 1995: Tinston, 1994: Hardin et al.. 1981: Schwetz et al.. 1975).
Pregnancy Loss. Human and animal studies examining pregnancy loss are discussed in
detail in Section 4.7. For humans, both occupational and drinking water studies have examined
fetal loss, an outcome for which there is good retrospective recall, and any bias would result in
an underestimation of the true risk (Wilcox and Horney, 1984). However, the available studies
may be limited by selection bias and small sample sizes.
A number of occupational studies have shown spontaneous abortion or perinatal loss
among women employed as dry cleaners (Doyle et al., 1997: Olsen et al., 1990: Kyyronen et al.,
1989: Bosco et al., 1987), or otherwise exposed occupationally (Lindbohm et al., 1991:
Windham et al., 1991). An increased risk of spontaneous abortion was not observed in other
studies of women who were dry cleaners or wives of dry cleaners (Eskenazi etal., 1991b:
Lindbohm et al., 1991: Ahlborg, 1990a: Taskinen et al., 1989: McDonald et al., 1987: McDonald
etal., 1986).
A few residential studies have examined spontaneous abortion or perinatal loss among
women drinking contaminated water (Aschengrau et al., 2009a: ATSDR, 1998b: Bove, 1996:
Boveetal., 1995: Lagakos et al., 1986) or inhaling VOCs (ATSDR, 2008), with no conclusive
results. Lagakos et al. (1986) found no association with drinking contaminated water and risk of
spontaneous abortion and no association for risk of perinatal death prior to 1970; however, a
positive association was observed for perinatal death since 1970. No association was observed
in Aschengrau et al. (2009a), but the authors note that the differences between occupational and
residential studies may be due to the exposure levels. The National Research Council
determined that there is limited/suggestive evidence of an association for miscarriage with
tetrachloroethylene-contaminated drinking water exposure at Camp Lejeune during pregnancy
(NRC, 2009). No increased risk was observed among women living in a community concerned
about vapor intrusion from VOCs including tetrachloroethylene (ATSDR, 2008).
4-417
-------
Fetal loss in experimental animals correlates with the observation of spontaneous
abortions in humans, with varying tendencies for fetal loss depending on species (rodents have a
very low propensity to abort, while rabbits and primates have higher rates). There is evidence of
increased preimplantation loss in rats (Szakmary et al., 1997), increased resorption of pups after
maternal inhalation in rats and rabbits (Szakmary et al., 1997; Schwetz et al., 1975), reduction in
litter size and pup survival in rats and guinea pigs (Szakmary et al., 1997; Narotsky and Kavlock,
1995; Tinston, 1994; Kvrklund and Haglid, 1991), spontaneous abortion in rabbits (Szakmary et
al., 1997), and litters with dead pups (Tinston, 1994). However, fetal loss was not observed in
other in vivo studies (Carney et al., 2006; Hardin et al., 1981). In vitro studies of exposure to
tetrachloroethylene show decreased fertilized oocytes (Berger and Horner, 2003), and increased
mortality, malformations, and delayed growth and differentiation of embryos (Saillenfait et al.,
1995).
Birth Defects. After residential exposure to contaminated drinking water, birth defects
related to in utero exposure in humans include eye/ear anomalies and CNS/chromosomal/oral
cleft anomalies (Lagakos et al., 1986). A study of residents living in a community with vapor
intrusion including tetrachloroethylene examined birth outcomes and observed a significantly
higher prevalence of total and major cardiac defects (ATSDR, 2006): a follow-up study of this
cohort noted that conotruncal heart malformations were particularly elevated (ATSDR, 2008). A
recent study in Massachusetts of maternal exposure to drinking water contaminated with
tetrachloroethylene reported a 20% increased risk (95% CI: 0.8-1.7) between any maternal
exposure at the time of conception and congenital anomalies (oral cleft anomalies, neural tube
defects, and gastrointestinal and genitourinary malformations) in the offspring after adjustment
for maternal and paternal ages (Aschengrau et al., 2009b): however, this study is inconclusive
due to limited adjustment for potential confounding factors and low statistical power. A
hypothesis-generating ecological study found a 3.5-fold increased risk of oral cleft defects in
New Jersey towns with 410 ppb tetrachloroethylene in drinking water (Bove, 1996; Bove et al.,
1995), although a case-control study of oral cleft defects from a larger area in New Jersey
designed to test this hypothesis did not confirm the earlier observation (Bove, 1996). Three
overlapping studies similarly did not observe any association with birth defects among women
who were dry cleaners or laundry workers, although the number of exposed cases was very small
(Olsenetal., 1990: Kyyronen et al., 1989: Taskinen et al., 1989). While the NAS has
determined that there is inadequate/insufficient evidence to determine whether an association
exists between drinking water at Camp Lejeune, NC, and congenital malformations (NRC,
2009), a follow-up study is currently underway to examine the incidences of neural tube defects
and oral cleft anomalies (NRC, 2009: ATSDR, 2003).
4-418
-------
In experimental animals, an increase in microphthalmia or anophthalmia in rat offspring
was observed after maternal gavage exposure, but no other evaluation of birth defects was
undertaken in this study (Narotsky and Kavlock, 1995). Delayed ossification was observed in
mice but not in rats exposed prenatally (Schwetz et al., 1975): for skeletal retardation, no
significant differences were observed for exposed mice in another study (Szakmary et al., 1997).
Skeletal malformations were increased in mice pups after maternal inhalation exposure, but no
additional details were given regarding type of malformation (Szakmary et al., 1997), and no
significant differences were observed in other studies (Carney et al., 2006; Schwetz et al., 1975).
Internal organ malformations were significantly increased in mice exposed in utero (Szakmary et
al., 1997; Schwetz et al., 1975), and an in vitro study of rat embryos exposed to
tetrachloroethylene showed increased malformations (Saillenfait et al., 1995). No birth defects
were observed in other studies of rats (Hardin etal., 1981; Nelson et al., 1979; Schwetz et al.,
1975) or rabbits (Hardin et al., 1981).
Conclusions about the association of birth defects with exposure to tetrachloroethylene
cannot be drawn from the available epidemiological studies, which contain a number of
deficiencies and uncertainties that may introduce a positive or negative bias on observations. A
clear need exists for better studies of tetrachloroethylene exposure and birth defects. In
particular, given the evidence for heart defects reported in animal studies with exposure to TCE
and its metabolites, TCA (Johnson et al., 1998: Smith etal., 1989) and DCA [(Epstein et al.,
1992): refer to Sections 4.6.2, 4.7.2, and 4.8.2], there is a need for additional studies of heart
defects after exposure to tetrachloroethylene.
Birth Weight. The epidemiological studies reported equivocal findings on birth weight.
At the military base of Camp Lejeune, NC, babies born to women living in housing that received
drinking water containing VOCs including tetrachloroethylene had a slight decrease in mean
birth weight (-26 g, 90% CI: -43, -9) and an increase in small for gestational age (SGA, 22
weeks gestation) (OR: 1.2, 90% CI: 1.0-1.3), most notably among women who had two or more
prior fetal losses (OR: 2.5, 90% CI: 1.5-4.3), compared to unexposed women; no increase in
preterm births was observed (OR: 1.0, 90% CI: 0.9-1.1) (Sonnenfeld etal., 2001: AT SDR,
1998b). The NAS determined that there is inadequate/insufficient evidence to determine
whether an association exists between contaminated drinking water and decreased birth weight at
Camp Lejeune, NC (NRC, 2009).
Risk of intrauterine growth restriction (IUGR) was observed in an occupational study
(OR: 12.5, no CI given) based on one case exposed to tetra- and trichloroethylene (Windham et
al., 1991). A second residential study of a community with VOC exposure from vapor intrusion
reported that low birth weight was slightly but statistically elevated (OR: 1.26, 95%
CI: 1.00-1.59), as was SGA (OR: 1.22, 95% CI: 1.02-1.45) and full-term low birth weight
4-419
-------
(OR: 1.41, 95% CI: 1.01-1.95) (ATSDR, 2006). However, the analysis did not adjust for
smoking and sociodemographic factors, which are known to also cause birth weight reductions.
Other residential drinking water (Aschengrau et al., 2008; Lagakos et al., 1986) and occupational
(Olsenet al., 1990) studies showed no association between exposure to tetrachloroethylene and
low birth weight.
In experimental animals, exposure to tetrachloroethylene caused decreased birth weight
(Tinston, 1994) and decreased fetal body weight in some studies of rats (Carney et al., 2006;
Szakmary et al., 1997) and mice (Schwetz et al., 1975). However, no effect on birth weight was
found in other studies of mice (Szakmary et al., 1997), rats (Hardin et al., 1981; Schwetz et al.,
1975), and rabbits (Szakmary et al., 1997; Hardin et al., 1981). Experimental animal studies also
observed decreased weight gain after either pre- or postnatal tetrachloroethylene exposure. A
study in rats demonstrated a reduction in overall pup body weight after preconception, prenatal,
and postnatal inhalation exposure (0-1,000 ppm) through 29 days of age (Tinston, 1994).
Another study found that the offspring of rats exposed to tetrachloroethylene (0-900 ppm)
during late pregnancy (GDs 14-20) had reduced weight gain at postnatal Weeks 3-5, but the
same effect was not observed in those exposed earlier in pregnancy (GDs 7-13) (Nelson et al.,
1979).
4.9.1.1.3.3. Developmental neurotoxicity
Neurotoxicological effects have been reported after low exposure levels to
tetrachloroethylene in children (refer to Section 4.6 and Table 4-4) and in animals after prenatal
exposure (refer to Sections 4.6.2 and 4.7.2). Both human and animal evidence supports an
association between neurodevelopmental effects and tetrachloroethylene exposure. While other
neurotoxic effects are observed in adults (refer to Table 4-5), decreased VCS has been the main
observation in children.
Visual deficits. Recent studies have examined the visual system as a target of
tetrachloroethylene toxicity in both children and adults. Subjects were New York City apartment
residents (Storm etal., 2011 [previously reported in NYSDOH, 20101: NYSDOH, 2005a:
Schreiber et al., 2002) and employees and children at a daycare center (NYSDOH, 2005a, b, c;
Schreiber et al., 2002) exposed to tetrachloroethylene by proximity to dry cleaners. Exposure
was measured in indoor air, exhaled air, and blood levels, and the visual system was assessed by
visual contrast sensitivity (VCS) and color confusion index (CCI).
In the day-care studies, visual tests were not conducted on children at the time of
exposure due to their young age (NYSDOH, 2005c: Schreiber et al., 2002), and a follow-up
evaluation 4 to 5 years after the colocated dry cleaner closed showed no residual changes in VCS
or CCI (NYSDOH, 2005a, b). There is a possibility that the results of these test results for
4-420
-------
children could be due to a learning disability or a developmental delay (Storm and Mazor, 2004),
although these data were not available for the control children (Hudnell and Schreiber, 2004).
The residential studies were designed to assess vision in children and adults living in the
same household colocated near dry cleaners (Storm et al., 2011 [previously reported in
NYSDOH, 20101: NYSDOH, 2005a: Schreiber et al.. 2002). Investigators found that children
generally performed better than adults for both VCS and CCI. Children exposed to
tetrachloroethylene performed worse than adults for VCS for the highest category of exposure
compared to both child and adult reference subjects (NYSDOH, 2010, 2005a), indicating there
may be increased susceptibility for children. Poorer CCI scores were associated with levels of
tetrachloroethylene-in exhaled breath in children but not in adults (NYSDOH, 2005a), but a later
study found that CCI was not associated with levels of tetrachloroethylene exposure in either
children or adults (NYSDOH, 2010). The investigators noted that exposure to
tetrachloroethylene was highly correlated with race and income, but small sample sizes made it
difficult to fully examine this correlation (NYSDOH, 2010, 2005a).
Additionally, a case study reported reduced VCS in a 2.5-year-old boy after prenatal
exposure to tetrachloroethylene (Till et al., 2003), as do reports from Till et al. (2005; 200la:
2001b) and Laslo-Baker et al. (2004) showing visual system functioning deficits in young
children of mothers exposed to multiple solvents during pregnancy, although exposure to
tetrachloroethylene was not uniquely identified. An important factor to consider in the testing of
visual function in children is the requirement for sustained attention and cognition (Tschopp et
al., 1998; Scharre et al., 1990). For this reason, visual testing of young children, particularly,
contrast sensitivity in children younger than 6 years of age, is difficult, and responses of young
children are more variable than those of adults (Scharre et al., 1990). A need exists for
developing methods to better evaluate contrast sensitivity effects in the very young-aged child.
Acute Neurotoxicity. Acute neurotoxicity has been observed in children exposed to
tetrachloroethylene. A case study by Koppel et al. (1985) reported that a 6-year-old boy who
directly ingested 12-16 g of tetrachloroethylene suffered from drowsiness, vertigo, agitation, and
hallucinations before lapsing into a coma. One hour after ingestion, his blood
tetrachloroethylene concentration was 21.5 mg/L. He recovered, but because follow-up testing
was not conducted, any potential long-term effects of the exposure are unknown (Koppel et al.,
1985). Gamier et al. (1996) reported mild CNS depression (dizziness and drowsiness were the
most common symptoms, along with nausea, vomiting, headache, tinnitus, unconsciousness)
after exposure to coin-operated dry-cleaned items in 5 cases of children and 24 cases of adults
but did not separate the analysis by age group. Gamier et al. (1996) also described two
additional reports (published in Danish) of unconsciousness in a 9-year-old boy who died after
4-421
-------
using his dry-cleaned sleeping bag (Korn, 1977), and in a 7-year-old girl who was left in a car
with dry-cleaned clothing (Larsen et al., 1977).
Brain neurochemistry. There are no studies in humans measuring brain neurochemistry
after exposure to tetrachloroethylene, in either children or adults. In experimental animals,
altered brain biochemistry (fatty acid composition) was observed in the offspring after
gestational exposure to rats and guinea pigs (Kyrklund and Haglid, 1991; Nelson etal., 1979).
These studies do not necessarily indicate effects on brain neurochemistry after gestational
exposure compared to adult exposure.
Neurobehavior. Two cohorts examined behavior in children after exposure to
tetrachloroethylene, with neither finding any association. In the daycare study described above,
18 children were examined for neurobehavioral deficits using a battery of tests for both
neurological and behavioral function. Tests were conducted approximately 5 weeks after
exposure ceased (at ages 4-5 years old) (NYSDOH, 2005c), and again in 13 children at a follow-
up evaluation 4-5 years later (NYSDOH, 2005a) and reported no functional change at either
examination. A large retrospective cohort study in Cape Cod, MA, examined prenatal and
postnatal exposure to drinking water contaminated by tetrachloroethylene leaching into water
distribution pipes (Janulewicz et al., 2008). Children born in 1969-1983 were included in the
analysis (n = 2,086), and followed during 2002-2003. Data were collected from birth
certificates and self-administered questionnaires including information on medical history for the
mother and child, potential solvent exposure, and water use. Cumulative exposure during the
prenatal period was estimated to be 4 x 10~5 to 1,328 g, and exposure during the postnatal period
was estimated to be 2.9 x io~4 to 3,310 g. No statistically significant association was observed
with attention, learning, or behavioral functions.
Rats exposure to tetrachloroethylene during pregnancy resulted in developmental delay as
measured by the ascent test and rotorod test (Nelson etal., 1979), although another study found
no adverse effects for running wheel activity, avoidance behaviors, or operant conditioning
(Nelson et al., 1979). Other effects observed include altered motor activity (Szakmary et al.,
1997; Tinston, 1994), decreased muscular strength (Szakmary et al., 1997), and short-term
reduced response to sound in pups (Tinston, 1994).
Young animals have also been directly exposed postnatally to tetrachloroethylene. Daily
exposure of rats to 1,000 ppm tetrachloroethylene on PNDs 6-29 resulted in sedation and
hypothermia, but the effect ceased 2 hours or less after exposure ended (Tinston, 1994). One
gavage study on young 45-50 gram rats showed behavioral and locomotor effects (Chen et al.,
2002a). One study of mice showed no neurobehavioral effects immediately after exposure
ceased at PND 17, but the mice exhibited increased locomotion and total activity and decreased
rearing at PND 60 (Fredriksson et al., 1993). Following i.p. dosing, 8-week-old male mice
4-422
-------
showed effects on the righting reflex and balancing (Umezuetal., 1997), and 6-week-old rats
showed effects on locomotor activity (Motohashi et al., 1993).
Autism spectrum disorder: One case-control study examined the relationship between
autism spectrum disorder (ASD) for births in 1994 in the San Francisco Bay Area and estimates
of 19 hazardous air pollutant concentrations for the census tract of the birth residence (Windham
et al., 2006). Risk estimates for the upper 3rd quartile and upper 4th quartile of
tetrachloroethylene exposure were OR: 1.31 (95% CI: 0.93-1.84) and OR: 1.11 (95%
CI: 0.78-1.59), respectively, with no suggestion of a linear concentration-response pattern. The
low level of exposure detail for individual subjects in the study does not provide sufficient
information either for or against an association between tetrachloroethylene and ASD. The
causes of autism are unknown, but environmental factors have been hypothesized (Grandjean
and Landrigan, 2006). Epidemiologic studies of analytical designs and with more sensitive
exposure-assessment approaches are needed to more clearly define any role of
tetrachloroethylene and other air pollutants.
4.9.1.1.3.4. Developmental immunotoxicity
Section 4.8.1.1.1 and Table 4-38 describe studies relating tetrachloroethylene to immune
response in children. The developing immune system is an area of potential susceptibility
(Dietert, 2008), although there are few published studies relating to immune response after
tetrachloroethylene exposure to either children or adults. The childhood studies examined a
relationship with tetrachloroethylene exposure and allergy, asthma, and infection—immunotoxic
outcomes not reported in any of the studies of adults. In addition, family members of children
diagnosed with leukemia from Woburn, MA, exhibited altered lymphocyte (CD3, CD4, CDS)
and CD4/CD8 ratios (Byers et al., 1988), though this was a mixed exposure to other
contaminants in addition to tetrachloroethylene. Other immunological conditions have been
observed in adults, but these are distinct from those observed in children discussed below. This
is an area for future research.
Allergy. Lehmann et al. (2002) examined cord blood samples from healthy, full-term
neonates for T-cell populations and associated them with indoor exposure to VOCs measured 4
weeks after birth (likely to reflect late-prenatal exposures) and observed a significant association
of tetrachloroethylene exposure with a reduction of interferon-g-producing Type 1 T-cells.
However, another study examining indoor exposure to VOCs and allergic sensitization and
cytokine secretion in 3-year-old children at high risk for development of allergic disease (low
birth weight, high cord blood IgE, family history of atopy) found no significant association
between tetrachloroethylene exposure and allergic sensitization to egg white and milk (Lehmann
et al., 2001). No studies of allergy after exposure to tetrachloroethylene were reported in adults.
4-423
-------
However, tetrachloroethylene has been demonstrated to adversely affect IL-4 and TNF-a in
rodent mast cells (Seo et al., 2008a) and passive cutaneous anaphylaxis in rats exposed i.p. (Seo
et al.. 2008a) and in drinking water (Seo et al.. 2008b).
Asthma. In a study of inhalation exposure, Delfmo et al. (2003a: 2003b) measured the
concentration of ambient air pollutants, including tetrachloroethylene, and correlated it with
subsequent symptoms of asthma in children in the Los Angeles, CA area. These results
suggested an increased risk with exposure to tetrachloroethylene (Delfmo et al., 2003a).
However, another analysis of the data examined the amount of tetrachloroethylene and other
volatile organic compounds in exhaled breath of asthmatic children (Delfmo et al., 2003b).
Although there was a significant correlation between ambient and exhaled concentrations, the
investigators did not find any association with exhalation concentrations and asthma symptoms
or ambient air concentrations and asthma symptoms, although the OR for exhaled breath was
larger than for ambient air exposure [OR: 1.94, 95% CI: 0.8-4.7; Delfmo et al. (2003b)1. An
18-year-old without personal or family history of bronchial asthma developed respiratory
symptoms (cough, dyspnea, altered forced expiratory volume) after maintaining dry-cleaning
machines (Boulet 1988).
Susceptibility to Infection. Only one report on tetrachloroethylene exposure and
childhood infection was found in the published literature. Higher prevalences of kidney and
urinary tract disorders (primarily infection) and lung and respiratory disorders (asthma, chronic
bronchitis, or pneumonia) in children were reported by mothers living in a community with a
past history of VOC-contaminated drinking water compared to prevalences reported by mothers
living in uncontaminated areas (Lagakos et al., 1986).
4.9.1.1.3.5. Hepatotoxicity
Bagnell and Ellenberger (1977) reported that a child suffered from obstructive jaundice
and hepatomegaly after consuming tetrachloroethylene-contaminated breast milk (10 mg/L),
with conditions improving when breastfeeding was discontinued.
4.9.1.1.3.6. Fatality
A case report found that vapors off-gassing from dry-cleaned fabrics were implicated in
causing the death of a 2-year-old boy who had slept in a room with multiple curtains that had
been incorrectly dry cleaned (Gaillard et al., 1995) and retained 6 kg of tetrachloroethylene as
estimated by a later experiment repeating the conditions (Gamier et al., 1996). Another case
reported a death in a 17-year-old employed at a plastics manufacturing plant and using
tetrachloroethylene to clean the inside of a metal mold (NIOSH, 1994).
In the one case of a child's direct ingestion of tetrachloroethylene, a 6-year-old boy who
swallowed 12-16 g tetrachloroethylene lost consciousness and lapsed into a coma (Koppel et al..
4-424
-------
1985). This 6-year-old also experienced drowsiness, vertigo, agitation, and hallucinations, but
he later recovered. Follow-up testing on the boy was not reported; therefore, any potential long-
term effects of the exposure are unknown (refer to Section 2.2.5). Due to the rarity of these
cases, there are little data to support any hypothesis regarding increased susceptibility for acute
mortality in childhood compared to adulthood.
4.9.1.1.3.7. Childhood cancer
The epidemiologic and experimental animal evidence is limited regarding susceptibility
to cancer from exposure to tetrachloroethylene during early life-stages. Generally speaking,
there may be developmental susceptibility for early lifestage exposure to chemicals and cancer
(Anderson et al., 2000; Olshan et al., 2000). The human epidemiological evidence is
summarized above for cancer in the liver (refer to Section 4.4.1.2), kidney (refer to Section
4.5.1.2), and other organ systems (refer to Section 4.8.1.2). The experimental animal research is
summarized above for cancer in the liver (refer to Section 4.4.2.2), kidney (refer to Section
4.5.2.2), and other organ systems (refer to Section 4.8.2). Few studies have examined cancer in
children after exposure to tetrachloroethylene; those few have examined total childhood cancer,
leukemia, and brain tumors. A recent review of the data related to exposure to
tetrachloroethylene, trichloroethylene, or solvent mixtures found inadequate/insufficient
evidence to determine whether an association exists for childhood leukemia, neuroblastoma, or
brain cancer (NRC, 2009).
Total Childhood Cancer. One study examined childhood cancers in an area in Endicott,
NY, for which vapor intrusion into homes was of concern. Many VOCs were identified in
samples and included trichloroethylene and tetrachloroethylene (ATSDR, 2006). This study
found fewer than six cases of cancer over a 20-year period, in children up to 19 years of age,
which did not exceed expected cases or types.
Childhood Leukemia. Leukemia has been observed in a few studies after exposure to
tetrachloroethylene in adults and children. However, the studies are limited by small sample
sizes, lack of exposure measurements, exposure to multiple contaminants, and possible
participation bias.
A small case-control study of children residing in Woburn, MA, found a strong but
imprecise association between maternal exposure during pregnancy and drinking water
contaminated with multiple solvents including tetrachloroethylene and childhood leukemia, with
a positive dose-response trend, when compared with exposure prior to pregnancy or postnatal
exposure to the infant via lactation [(Costas et al., 2002; MDPH, 1997); refer to Section
4.9.1.2.4]. However, it is difficult to uniquely identify tetrachloroethylene as the causative agent
given the higher concentrations of trichloroethylene reported. Other population case-control
4-425
-------
studies of childhood leukemia have not shown an increased risk from paternal (Shu etal., 1999;
Lowengart et al., 1987) or maternal (Infante-Rivard et al., 2005; Shu et al., 1999) occupational
exposure to tetrachloroethylene, possibly due to the relatively small sample size. Another study
population is currently being further examined to determine any association between maternal
ingestion of contaminated water and the incidence of childhood cancers (ATSDR, 2003). One in
vitro study of human mononuclear cord blood cells exposed to tetrachloroethylene found that
pathways involved in cancer induction were affected through altered gene expression of
inflammatory responses, tumor and metastatic progression, and the apoptotic process (Diodovich
et al., 2005). In addition, a follow-up study of children from Camp Lejeune, NC, is currently
being conducted to determine any association between maternal ingestion of contaminated water
and the incidence of childhood leukemia and non-Hodgkin lymphoma (NRC, 2009; AT SDR,
2003). No data are available on cancer risk in animals from early lifestage tetrachloroethylene
exposure.
Childhood Brain Cancer. Very few studies of tetrachloroethylene exposure have
reported brain tumors, and these are generally quite limited. One study of parental occupational
exposure to tetrachloroethylene (8 cases, 11 controls) found no risk of neuroblastoma in the
offspring (OR: 0.5, 95% CI: 0.2-1.4) (DeRoos et al., 2001). This study, like those on childhood
leukemia, is quite limited for examining parental exposure to tetrachloroethylene and childhood
cancer.
4.9.1.1.3.8. Age-dependent adjustment factors (ADAFs)
According to EPA's Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens (U.S. EPA, 2005b) there may be increased susceptibility to early-life
exposures for carcinogens with a mutagenic MOA. Although the contribution of genotoxicity to
tetrachloroethylene carcinogenesis cannot be ruled out for one or more target organs,
uncertainties with regard to the characterization of tetrachloroethylene genotoxicity remain. This
is primarily because in vivo tests of tetrachloroethylene have been equivocal, with at most,
modest evidence of genotoxic effects in rodent tumor tissues examined (including mouse liver
and rat kidney) following exposure at tumorigenic doses. Additionally, no evidence is available
regarding the potential contribution of tetrachloroethylene genotoxicity to other rodent tumor
types (particularly, MCL, testes, and brain) or to human cancers. Ames assays of
tetrachloroethylene have yielded largely negative results. Certain tetrachloroethylene
metabolites (TCVG, TCVC, NAcTCVC, tetrachloroethylene oxide, and DCA) exhibit
genotoxicity, the database of available studies is limited, and not all metabolites have been
sufficiently tested to support clear conclusions about their genotoxic potential. Additionally, the
specific active moiety(ies) that contribute to tetrachloroethylene carcinogenesis are not known.
4-426
-------
Thus, because the specific active moiety(ies), mechanisms, or modes of action by which
tetrachloroethylene induces carcinogenesis are not known, early-life susceptibility is not
assumed, and the application of ADAFs is not recommended.
4.9.1.1.3.9. Early lifestage exposure and outcomes in adulthood
Many additional studies have described adverse outcomes in adults only, mainly based on
the assumption of exposure occurring in adulthood; whether or not early lifestage exposures
might have occurred is often not considered. Only one identified study reports an examination of
early lifestage exposure to tetrachloroethylene and latent outcomes in adults. A large prospective
study of the offspring of dry cleaners found a significant increased risk for schizophrenia at
21-33 years of age (Perrin et al., 2007). This is a preliminary report that did not adjust for
family history of mental disease, a risk factor for schizophrenia.
4.9.1.2. Later Life-Stages
Due to changes in physiology, in the elderly, exposure levels may be distinct from those
observed in younger adults. The elderly have increased ventilation rates per kg body weight
compared to adults (U.S. EPA, 2006a) and spend the majority of their time indoors, where
increased concentrations of tetrachloroethylene have been found compared to those measured
outdoors (U.S. EPA, 200la). The elderly also experience changes in skin permeability (U.S.
EPA, 2006a), which may lead to increased exposure while showering, bathing, or swimming in
contaminated water (U.S. EPA, 200la: Rao and Brown, 1993). While dermal exposure is
generally not considered a major route of exposure, this route of exposure is not well
characterized for later life-stages.
Toxicokinetics in later lifestages can be distinct in younger adults (Benedetti et al., 2007;
U.S. EPA, 2006b: Ginsberg et al., 2005), although there is only limited evidence showing a
possible age-related difference in CYP expression (Dome and Renwick, 2005; Parkinson et al.,
2004; George et al., 1995). GST expression has been observed to decrease with age in human
lymphocytes, with the lowest expression in those aged 60-80 years old (van Lieshout and Peters,
1998).
Few studies examined the exposure to tetrachloroethylene in elderly adults (>65 years
old). One study found elevated blood tetrachloroethylene levels (310-1,770 ug/L) and urine
trichloroacetic acid levels (22-1,650 ug/L) in an elderly couple living above a dry-cleaning
facility (Popp et al., 1992).
Similarly, few studies examine the effects of tetrachloroethylene exposure in elderly
adults. Another residential study examined two individuals over the age of 60 years and found
4-427
-------
that the mean scores of VCS were lower than the 12th percentile of all control subjects (Schreiber
et al.. 2002).
One PBPK modeled tetrachloroethylene in adults aged 65, 75, and 85 years old and
predicted lower concentrations in all compartments for older adults compared to younger adults,
and similar predictions for TCA in older and younger adults (Yokley and Evans, 2007). The
authors noted that these results indicate that increased susceptibility is likely among older adults
due to metabolic changes associated with aging. Another model predicted a decrease in alveolar
concentration of tetrachloroethylene in 65-year-olds versus 25-year-olds, which the authors
attribute to age-related decreases in cardiac output and ventilation (Guberan and Fernandez,
1974).
These very limited studies suggest that older adults may experience increased exposure to
tetrachloroethylene and resulting increased VCS deficits compared to younger adults. However,
there is no further evidence of effects for older adults exposed to tetrachloroethylene beyond
these studies.
4.9.2. Other Susceptibility Factors
Aside from age, many other factors may affect susceptibility to tetrachloroethylene
toxicity. A partial list of these factors includes gender, genetic polymorphisms, pre-existing
disease status, nutritional status, diet, and previous or concurrent exposures to other chemicals.
The toxicity that results due to changes in multiple factors may be quite variable, depending on
the exposed population and the type of exposure. Qualitatively, the presence of multiple
susceptibility factors will increase the variability that is observed in a population response to
tetrachloroethylene toxicity.
4.9.2.1. Gender
Individuals of different genders are physiologically, anatomically, and biochemically
different. Males and females can differ greatly in many physiological parameters such as body
composition, organ function, ventilation rate, and metabolic enzyme expression, which can
influence the toxicokinetics of chemicals and their metabolites in the body (Gochfeld, 2007;
Gandhi et al., 2004; Parkinson et al., 2004). In the case of tetrachloroethylene, there is some
indication that tetrachloroethylene metabolism is different between males and females. One
PBPK model found gender-specific differences that were small (although significant) in
tetrachloroethylene blood concentrations but considerable (twofold at age 40) with regard to
TCA blood concentration levels [(Clewell et al., 2004): refer to Section 3.5.2 and Figure 3-7].
Opdam and Smolders (1986) exposed six human subjects to concentrations ranging from 0.5-9
ppm and found alveolar concentrations in male subjects to be only slightly less than those in
4-428
-------
females (refer to Figures 3-6a, b). It is not known whether gender variation of p-lyase activity
(refer to Section 3.3.3.2.3), the most important activator of toxic products in the conjugation
pathway, exists in humans as it does in rats, with metabolism in males being faster than in
females (Volkel etal., 1998), although there seems to be little gender difference in the
concentrations of metabolites in blood, regardless of age (Sarangapani et al., 2003).
In humans, there have been a few studies demonstrating sex-specific effects (refer to
Section 4.7.2.3), but it has not been determined whether there is a gender difference in response
to exposure to tetrachloroethylene. Among former residents of Camp Lejeune, NC, exposed to
contaminated drinking water, there is limited/suggestive evidence of an association between
breast cancer and tetrachloroethylene, and inadequate/insufficient evidence to determine whether
an association exists for cervical, ovarian/uterine, or prostate cancer (NRC, 2009). Male breast
cancer has also been reported by former residents of Camp Lejeune exposed to contaminated
drinking water; however, this association has not been investigated sufficiently to draw any
conclusions (NRC. 2009).
Ferroni et al. (1992) evaluated neurological effects of tetrachloroethylene exposure
among female dry cleaners and concluded that tetrachloroethylene exposure in dry-cleaning
shops may impair neurobehavioral performance and affect pituitary function. The pituitary is
controlled, in part, by hypothalamic dopamine, which is important to neurotransmission. Study
participants were tested during the proliferation phase of menstruation, which may better capture
changes in prolactin secretion but also may potentially confound findings if there are individual
differences in severity of menstruation and in the timing of a test session relative to the day of
menstruation [(U.S. EPA. 2004): refer to Section 4.6.1.2.5].
Some studies have observed an increased risk for NHL, Hodgkin lymphoma, chronic
lymphocytic leukemia or multiple myeloma in females compared to males [(Radican et al., 2008;
Ji and Hemminki, 2006b: Miligi et al.. 2006: Ji and Hemminki, 2005b: Blair etal.. 2003:
Andersen et al.. 1999: Cohnetal., 1994: Spirtas et al.. 1991: Morton and Marianovic, 1984):
refer to Section 4.6.1.2], whereas other studies observed an increase in both males and females
(Travier et al., 2002) or no increase in either males or females (Lynge et al., 2006: Boice et al.,
1999). Other studies did not examine the outcome in both sexes. Some of these studies are
limited by lack of quantitative exposure information, ecological design, or exposure to mixtures,
differences in exposure potential and level of exposure may explain the difference in risk
between women and men. Differences in physiological parameters may also explain the
observed gender difference in risk.
The studies by Pesch et al. (2000a) and Dosemeci et al. (1999) suggest that there may be
gender differences in risk to renal cell carcinoma with occupational exposure to
tetrachloroethylene; in both studies, the risks were higher in males than in females (refer to
4-429
-------
Section 4.5.1.2). In a rat inhalation study, tubule cell hyperplasia was observed in eight males at
various doses but in only one female at the high dose. Also, renal tubule adenomas and
adenocarcinomas were observed only in males; however, chronically induced tetrachloroethylene
neoplastic kidney lesions do not exhibit sex specificity (NTP, 1986). In a rat gavage study, there
was no gender difference for toxic nephropathy (NCI, 1977). A marked gender difference was
observed between male and female rats in the severity of acute renal toxicity, with male rats
being more affected than female rats (Lash et al., 2002), but otherwise, no gender variation was
observed for chronic nephrotoxicity not associated with a2u-globulin nephropathy (refer to
Sections 4.5.2.2 and 4.5.4.3.3).
In the liver, male rats showed an increased incidence of spongiosis hepatitis as compared
with females, but there was no gender difference in hepatocellular adenomas and carcinomas;
however, the spleen showed increased effects in males versus females IYJISA, 1993); refer to
Sections 4.4.2.1 and 4.4.2.2].
4.9.2.2. Race/Ethnicity
Race/ethnicity can often be observed as an important consideration, and may be due to
actual increased exposure or to variation in expression of metabolic enzymes due to genetic
variability (Garte et al., 2001). In particular, ethnic variability in expression has been reported
for CYP (Neafsev et al.. 2009: Dome and Renwick, 2005: Parkinson et al.. 2004: McCarver et
al.. 1998: Shimada et al.. 1994: Stephens et al.. 1994) and GST (Ginsberg et al.. 2009: Nelson et
al.. 1995).
Studies of VCS in residents in apartments colocated with dry cleaners in New York, NY,
found that participants of minority status and low income (<$60,000) were more likely to have
high indoor air levels of tetrachloroethylene (>100 |ig/m3), but analyses of this small sample size
of participants in this exposure category could not definitively separate minority status from
VCS performance (Storm et al.. 2011 [previously reported in NYSDOH, 20101).
Oxidative damage among female dry cleaners appeared to be increased among black
workers compared to female Caucasian workers, although female dry cleaners had decreased
levels of oxidative damage compared to female launderers (Toraason et al., 2003). In a follow-
up study on the mortality of a cohort of dry cleaners, bladder cancer was elevated among
Caucasian men and women, and kidney cancer was elevated among black men and women;
however, these associations were not strongly related to duration or estimated level of exposure
to tetrachloroethylene (Blair et al., 2003). One study found that following tetrachloroethylene
exposure, TCA concentration in the urine of six Asian subjects was no different from the levels
found in six Caucasians; however, this study was confounded by significant differences in
alcohol consumption between the Caucasian and Asian populations (Jang and Droz, 1997).
4-430
-------
4.9.2.3. Genetics
Human variation in response to tetrachloroethylene exposure may be associated with
genetic variation. For example, in a study of six adults, Monster et al. (1979) found that the
mean coefficient of interindividual variation for tetrachloroethylene uptake was 17%. Human
genetic polymorphisms in metabolizing enzymes involved in biotransformation of
tetrachloroethylene are known to exist: Section 3.3.3.1.5 discusses CYP isoforms and genetic
polymorphisms, Section 3.3.3.2.1 covers GST isoenzymes and polymorphisms, and Section 3.3.4
describes differences in enzymatic activity.
Reitz et al. (1996) examined tetrachloroethylene metabolism in seven adult human liver
samples and found a fivefold difference in the rate of tetrachloroethylene metabolism between
the 50th and 99th percentiles. Opdam (1989a) found a twofold spread in tetrachloroethylene
blood concentrations in a study population of nine adult human subjects. In this study, the
amount of fat and the blood concentrations seemed to be positively correlated but could not be
confirmed; the author suggested that if the subjects had a wider range of body fat levels (range in
this study was only 7-22 kg), a larger amount of interindividual variation would be expected.
Computer modeling was used to examine the toxicokinetic variability of
tetrachloroethylene (Chiu and Bois, 2006; Bois et al., 1996). However, whether CYP or GSH
polymorphisms account for interindividual variation in tetrachloroethylene metabolism among
humans, and, thus, differences in susceptibility to tetrachloroethylene-induced toxicities, is not
known.
4.9.2.4. Preexisting Disease
It is known that kidney and liver diseases can affect the clearance of chemicals from the
body, and, therefore, poor health may lead to increased half-lives for tetrachloroethylene and its
metabolites. There are limited data indicating that certain diseases may alter susceptibility to
tetrachloroethylene exposure, mainly through altered metabolism. Presence of cancer likely
alters tetrachloroethylene metabolism, because increased CYP2E1 expression has been observed
in these individuals (Neafsey et al., 2009). Cirrhosis of the liver likely alters tetrachloroethylene
metabolism, because increased CYP2E1 expression has been observed in these individuals
[(Neafsey et al., 2009): also refer to Section 4.9.2.5.1]. Tetrachloroethylene is lipophilic and
stored in adipose tissue (Monster and Houtkooper, 1979): therefore, obese individuals may
experience altered toxicokinetics of tetrachloroethylene compared to nonobese individuals.
Obesity also likely alters tetrachloroethylene metabolism, because increased CYP2E1 expression
has been observed in obese individuals, compared to nonobese individuals (Neafsey et al., 2009:
McCarver et al., 1998). For obese individuals, a model predicted a decrease in alveolar
4-431
-------
concentration of tetrachloroethylene during exposure and a decrease in elimination, compared to
nonobese individuals (Guberan and Fernandez, 1974).
4.9.2.5. Lifestyle Factors and Nutrition Status
4.9.2.5.1. Alcohol intake
Alcohol is generally regarded as a confounder, although the additive or interactive effects
of these exposures along with tetrachloroethylene are not well characterized. Alcohol intake
likely alters tetrachloroethylene metabolism and causes higher toxicity, because increased
CYP2E1 expression has been observed in individuals who consume alcohol, compared to those
who do not (Neafsey et al., 2009; Liangpunsakul et al., 2005; Parkinson et al., 2004; Meskar et
al.. 2001: McCarver et al.. 1998: Lieber, 1997: Perrot et al.. 1989). Those exposed to both
tetrachloroethylene and TCE and consumed alcohol demonstrated an elevated color confusion
index (Valic et al.. 1997).
4.9.2.5.2. Tobacco smoking
Smoking, or the number of factors correlated to smoking (e.g., socioeconomic status,
diet, alcohol consumption), is generally regarded as a confounder in epidemiological studies
(Ruder, 2006), although the additive or interactive effects of these exposures along with
tetrachloroethylene are not well characterized. Immunotoxicity and hematotoxicity were
observed in tetrachloroethylene-exposed dry cleaners, particularly for those who were smokers
(Emara et al., 2010). Sister chromatid exchange in peripheral lymphocytes was observed more
frequently in male smokers exposed to tetrachloroethylene alone or in combination with TCE
(Seiji etal., 1990). No increase in oxidative damage among tetrachloroethylene-exposed dry
cleaners was observed among smokers compared to nonsmokers (Toraason et al., 2003).
Regarding esophageal cancer, occupational observations suggest that the magnitude of the risks
for several smoking-related cancers among dry cleaners was greater than could be explained by
smoking alone, suggesting a further contribution from another risk factor, such as occupational
exposure [(Blair et al., 2003: Ruder et al., 2001): refer to Section 4.8.1.2.2].
4.9.2.5.3. Nutritional status
Vegetable or vitamin intake may decrease susceptibility to tetrachloroethylene because
CYP2E1 inhibition has been observed in individuals who consume various vegetables, herbs,
and teas, and increased expression in those consuming high-fat diets (Neafsey et al., 2009).
Coexposure to a-tocopherol (vitamin E) along with tetrachloroethylene resulted in decreased rat
(Costa et al., 2004) and mouse (Ebrahim et al., 2001: Ebrahim etal., 1996) liver cell toxicity. A
similar protective effect was also observed with coexposure to 2-deoxy-D-glucose in mice
(Ebrahim et al., 2001: Ebrahim et al., 1996) and taurine in mice (Ebrahim et al., 2001). An in
4-432
-------
vitro study of cultured normal human epidermal keratinocytes demonstrated an increase in lipid
peroxidation in a dose-dependent manner after exposure to tetrachloroethylene, which was then
attenuated by exposure to vitamin E (Ding et al., 2006). However, no associations were found
for blood levels of vitamin E and p-carotene in rats [(Toraason et al., 2003): refer to Sections 4.3
and 4.4.4.4.3].
4.9.2.5.4. Physical activity
Studies and models have examined the effect of increased workloads on the
toxicokinetics of inhaled tetrachloroethylene alone (Droz et al., 1989a: Droz et al., 1989b:
Imbriani et al.. 1988: Jakubowski and Wieczorek, 1988: Pezzagno et al., 1988) or with TCE
(Opdam, 1989a, b). These studies are equivocal on whether an increase in pulmonary ventilation
increases the amount of tetrachloroethylene taken up during exposure. A model predicted an
increase in alveolar concentration of tetrachloroethylene after exercise, which the authors
attribute to increased cardiac output and ventilation (Guberan and Fernandez, 1974).
4.9.2.6. Socioeconomic Status
Socioeconomic status (SES) can be an indicator for a number of coexposures, such as
increased tobacco smoking, poor diet, education, income, and health care access, which may play
a role in the results observed in the health effects of tetrachloroethylene exposure.
Children's exposure to tetrachloroethylene was measured in a low SES community, as
characterized by income, educational level, and receipt of free or reduced cost school meals
(Sexton et al., 2005): however, this study did not compare data to a higher SES community, nor
examine health effects. Studies of VCS measured in child and adult residents in apartments
colocated with dry cleaners in New York, NY, found that the study participants more likely to be
exposed to high indoor air levels of tetrachloroethylene (>100 |ig/m3) were of minority status,
low income (<$60,000), or, for adults, had significantly lower level of education (Storm et al.,
2011 [previously reported in NYSDOH, 20101: NYSDOH, 2005a). However, analyses of the
small sample size in this exposure category could not definitively separate race/ethnicity or SES
from VCS performance.
4.9.2.7. Multiple Exposures and Cumulative Risks
When considering health risks, it is important to consider the cumulative impact of
effects that may be due to multiple routes of exposure. EPA published a Framework for
Cumulative Risk Assessment (U.S. EPA, 2003) to address these issues. A human aggregate
exposure model developed by McKone and Daniels (1991) incorporated likely exposures from
air, water, and soil media through inhalation, ingestion, and dermal contact. They asserted that
4-433
-------
the aggregate exposure may be age dependent but did not present any data for persons of
differing life-stages.
The limited data summarized by the ATSDR in its draft interaction profile on
tetrachloroethylene, trichloroethylene, 1,1-dichloroethane, and 1,1,1-trichloroethane suggest that
additive joint action is plausible (ATSDR, 2004). Coexposure to other pollutants, including
trichloroethylene and methylchloroform, which produce some of the same metabolites and
similar health effects as tetrachloroethylene, is likely to occur in occupational settings as well as
in nonoccupational sources such as in ground water contamination (e.g., Bove et al., 2002;
Sonnenfeld et al.. 2001: ATSDR. 1998a: MDPH, 1997: Lagakos et al.. 1986). However, no
evidence from the available studies indicates greater-than-additive effects for liver and kidney
toxicity.
Numerous environmental pollutants and therapeutic agents have the potential to induce or
inhibit tetrachloroethylene-metabolizing enzymes. For example, tetrachloroethylene metabolism
is increased by inducers of CYP enzymes such as toluene, phenobarbital, and pregnenolone-
16-a-carbonitrile, whereas CYP inhibitors such as SKF 525A, metyrapone, and carbon monoxide
decrease tetrachloroethylene metabolism (Costa and Ivanetich, 1980; Moslen et al., 1977; Ikeda
andlmamura, 1973). Likewise, tetrachloroethylene exposure may increase the effects of
exposures to other chemicals or stressors. For instance, adverse effects due to exposure to
chlorinated solvents and alcohol may be increased because tetrachloroethylene may induce
shared metabolic enzymes (refer to Section 3.3.4).
The acute effects of tetrachloroethylene share much in common functionally with those
of other solvents (e.g., toluene, volatile anesthetics, and alcohols) such as changes in reaction
time, nerve conduction velocity, and sensory deficits. There is emerging evidence that such
agents act on the ligand-gated ion channel superfamily in vitro (Shafer et al., 2005), particularly
on the inhibitory amino acids NMD A, nicotinic, and GABA receptors in vivo (Bale et al., 2005).
Other organic solvents induce effects on memory and color vision (Hudnell et al., 1996a:
Hudnell etal., 1996b: Altmann et al., 1995; Mergler et al., 1991). The consistency of these
observations suggests a common MOA of organic solvents to altered vision pattern. Hence, a
concern exists for neurobehavioral effects from interaction or competitive inhibition between
tetrachloroethylene and exposures with similarly hypothesized MO As.
The interaction between tetrachloroethylene, trichloroethylene, and 1,1,1-trichloroethane
(methylchloroform) was modeled in rats (Dobrev et al., 2001) and in computer models for
humans (Dobrev et al., 2002) and was shown to compete for metabolic capacity. The interaction
between tetrachloroethylene and trichloroethylene showed a less-than-additive effect on the liver
and kidney through inhibition of TCA formation (Pohl et al., 2003). Similarly, when exposed to
tetrachloroethylene, rat liver cells had increased toxicity when coexposed to peroxidation drugs
4-434
-------
such as cyclosporine A, valproic acid, and amiodarone (Costa et al., 2004), and w-hexane and
ethylbenzene inhibited the metabolism of tetrachloroethylene in rats (Skowron et al., 2001).
4.9.3. Uncertainty of Database and Research Needs for Susceptible Populations
There is some evidence that certain populations may be more susceptible to exposure to
tetrachloroethylene. The factors examined for tetrachloroethylene include age, gender,
race/ethnicity, genetics, preexisting disease, lifestyle factors, nutritional status, socioeconomic
status, and multiple exposures and cumulative risk. Areas where the database is currently
insufficient for characterizing the impact of tetrachloroethylene on susceptible populations are
identified below, along with research needs.
While there is more information on early life exposure to tetrachloroethylene than on
other potentially susceptible populations, there remain a number of uncertainties regarding
childhood susceptibility. Although inhalation is believed to be of most concern for
tetrachloroethylene, pathways of exposure for children are not well characterized. It is not clear
to what extent tetrachloroethylene may pass through the placenta in humans, as shown in rodent
studies (Szakmary et al., 1997; Ghantous et al., 1986): for some infants, the primary route of
exposure may be through breast milk ingestion (refer to Sections 2.2.4 and 3.2), while for other
infants, the dose received through ingestion of breast milk will become insignificant when
compared with the inhalation exposure (Schreiber, 1997). The amount of tetrachloroethylene
ingested from food is not well described; and it is not known to what extent tetrachloroethylene
is absorbed by a child and to which organs tetrachloroethylene and its metabolites may be
distributed. The neurological effects of tetrachloroethylene constitute the most sensitive
endpoints of concern for noncancer effects, and limited data show that early life-stages may be
more susceptible to visual deficits than are adults (Storm et al., 2011 [previously reported in
NYSDOH, 20101: NYSDOH, 2005a: Schreiber et al., 2002), yet developmental neurotoxic
effects, particularly in the developing fetus, need further evaluation using age-appropriate testing
for assessment. There are a number of adverse health effects observed uniquely in early
lifestages, with no comparable observations in adults to determine relative sensitivity (e.g., birth
outcomes, autism, allergy); conversely, there are some adverse outcomes that have been
observed only in adults.
There is suggestive evidence that there may be greater susceptibility for exposures to the
elderly, but the available data are much more limited with related uncertainties. Improved PBPK
modeling that contains physiologic information for infants and children (including, for example,
the effects of maternal inhalation exposure and the resulting concentration in breast milk) and for
older adults, and validation of these models, will aid in determining differences in life-stage
toxicokinetics of tetrachloroethylene. There may be a true difference in outcome after exposure
4-435
-------
during one life-stage compared to another, a lack of assessment of these outcomes in all life-
stages, or a lack of assessment of effects of exposures in one life-stage and latent outcomes.
More studies specifically designed to evaluate effects in early and later life-stages are needed in
order to more fully characterize potential life-stage-related tetrachloroethylene toxicity.
For other susceptibility factors, the data are more limited and based mainly on
nonchemical specific data that provide information on variation in physiology, exposure, and
toxicokinetics. Until quantitative conclusions can be made for each susceptibility factor, it will
be very hard to consider the impacts of changes in multiple susceptibility factors. In addition,
further evaluation of the effects of aggregate exposure to tetrachloroethylene from multiple
routes and pathways is needed. Similarly, the effects due to coexposures to other compounds
with similar or different MO As need to be evaluated.
4.10. SUMMARY OF HAZARD IDENTIFICATION
4.10.1. Overview of Noncancer and Cancer Hazard
This section summarizes the noncancer and cancer hazard findings for
tetrachloroethylene. This summary is based on the analyses presented in the preceding sections,
which discussed tetrachloroethylene toxicity on an organ-specific basis, in the following order of
presentation: neurotoxicity (refer to Section 4.1); kidney and bladder toxicity and cancer (refer to
Section 4.2); liver toxicity and cancer (refer to Section 4.3); esophageal cancer (refer to Section
4.4); lung and respiratory cancer (refer to Section 4.5); immunotoxicity, hematologic toxicity,
and cancers of the immune system (refer to Section 4.6); and developmental and reproductive
toxicity and reproductive cancers (refer to Section 4.7). Section 4.8 discusses genotoxicity, and
susceptible populations are addressed in Section 4.9.
The noncancer hazard characterization for tetrachloroethylene is presented in
Section 4.10.2. Effects that were noted in humans and in experimental animals (i.e.,
neurotoxicity [refer to Section 4.10.2.1], kidney toxicity [refer to Section 4.10.2.2], liver toxicity
[refer to Section 4.10.2.3], immunotoxicity and hematologic toxicity [refer to Section 4.10.2.4],
and reproductive and developmental toxicity [refer to Section 4.10.2.5]) are first summarized. A
tabular summary of the inhalation (refer to Table 4-49) and oral (refer to Table 4-50) studies that
are suitable for dose-response analysis, considering all studies across toxicologic endpoints, is
then presented in Section 4.10.2.6. Neurotoxicity is identified as a sensitive endpoint following
either oral or inhalation exposure to tetrachloroethylene. Section 5 presents dose-response
analyses of the neurotoxicity data set as a basis for derivation of inhalation and oral reference
values. Quantitative dose-response analyses of the findings for other toxicological effects (i.e.,
kidney, liver, reproductive and developmental toxicity) are also presented in Section 5.
4-436
-------
The cancer hazard characterization for tetrachloroethylene is presented in Section 4.10.3.
Section 4.10.3.1 presents the hazard descriptor, characterizing tetrachloroethylene as "likely to
be carcinogenic to humans." Section 4.10.3.2 synthesizes the epidemiologic data pertaining to
tetrachloroethylene and several cancer types, including non-Hodgkin lymphoma, multiple
myeloma, bladder, esophageal, kidney, lung, cervical, and breast cancer. Section 4.10.3.3
summarizes the results from three chronic bioassays that identified tetrachloroethylene-induced
rodent cancer, including mononuclear cell leukemia, kidney, and brain tumors in rats and liver
tumors in mice. The available mode-of-action information for the carcinogenicity of
tetrachloroethylene is presented in Section 4.10.3.4. Section 5 presents dose-response analyses
of the rodent bioassay data as a basis for derivation of inhalation and oral cancer slope factors.
4.10.2. Characterization of Noncancer Effects
4.10.2.1. Neurotoxicity
Human and animal studies provide complementary evidence regarding the association of
neurobehavioral deficits and tetrachloroethylene exposure. Tetrachloroethylene exposure in
humans has primarily been shown to affect visual function (including color vision) and
visuospatial memory and other aspects of cognition. Brain-weight changes have been measured
in animal studies. A more in-depth discussion of the human neurotoxicological studies can be
found in Section 4.1.1.3. The animal inhalation and oral or i.p. exposure studies are discussed in
Sections 4.1.2.1 and 4.1.2.2, respectively.
4-437
-------
Table 4-49. Inhalation studies suitable for dose-response analyses
Organ/
system
CNS
Study
Schreiber et al. (2002)
Schreiber et al. (2002)
NYSDOH (20101
Cavalleri et al. (1994):
Gobba et al. (19981
Spinatonda et al.
(1997)
Seeber (1989)
Ferroni et al. (1992)
Echeverria et al.
(1995)
Altmann et al. (1990)
Hake and Stewart
(1977)
Kjellstrand et al.
(1985)
Rosengren et al.
(1986)
Mattsson et al. (1998)
Wang et al. (1993)
Oshiro et al. (2008)
Species
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Mouse
Gerbil
Rat
Rat
Rat
Duration/dosing
4 yr mean duration
5. Syr (mean),
continuous
10 yr mean
duration
8.8 yr mean
duration
No information on
duration
>10yrmean
duration
10.6 yr mean
duration
15 (high-exposure
group) yr mean
duration
4-h exposure each
day for 4 d.
7.5 h exposure each
day for 5 d.
Acute (1 h)
Subchronic (12 wk,
with 16-wk follow-
up) continuous
Subchronic (13 wk)
6 h/d, 5 d/wk
Subchronic (12 wk)
continuous
60 min
NOAEL/LOAEL"
(ppm)
0.3 (daycare workers,
mean and median)
0.1 (residents, median
and mean), maybe as
high as 0.4 (mean) and
0.3 (median) b
0.002, 0.05 (children)
0.002, 0.07 (adults)
I^Cavalleri et al.
Ł1994}
& (median)
12,53
15
11° (operators)
10,5Q
20, 100. 150
0, 90, 320, 400, 600,
800, 1,200, 1,800,
3,600
0, 6fl, 300
0, 50, 200. 800
0. 300. 600
500. 1.000. 1,500
500, 1,000, 1,500
500, 1.000. L500
Effect
Visual contrast sensitivity
Visual contrast
sensitivity13
Visual contrast sensitivity
Dyschromatopsia
Reaction time
Visuospatial function,
information processing
speed
Reaction time,
continuous performance
Visuospatial function
Visual Evoked Potentials
EEGs
Increased motor activity
Brain: protein, DNA
concentration
Flash-evoked potential
Reduced brain weight,
DNA, protein
Reaction time
False alarms
Trial completions —
Signal Detection Task)
4-438
-------
Table 4-49. Inhalation studies suitable for dose-response analyses (continued)
Organ/
system
CNS
(continue
d)
Kidney
Liver
Immune
and
hematolo
gic
toxicity
Reproduct
ive and
develop-
mental
toxicity
Study
Boyes et al. (2009)
Mutti et al. (1992)
NTPQ986)
JISA (19931
JISA (1993)
Kjellstrand et al.
(19841
NTP (19861
JISA (1993)
Emara et al. (2010)
Sallmen et al. (1995)
Eskenazi et al.
(1991a)
Species
Rat
Human
Rat
Rat
Mouse
Mouse
Mouse
Mouse
Human
Human
Human
Duration/dosing
90 minutes
120 minutes
10 yr duration
Chronic bioassay
(104 wk)
Chronic
(104 wk)
Chronic
(104 wk)
Subchronic (4 wk)
continuous
Chronic bioassay
(104 wk)
Chronic
(104 wk)
Mean duration 7 yr
Exposure during
year before
initiation of
pregnancy,
occupational,
1973-1983
Wives of exposed
men working as
dry cleaners, 1980s
NOAEL/LOAEL3
(ppm)
250, 500, 1,000
1.000. 2,000, 3,000,
4,000
_LJ (median)
0, 200. 400
0, 50, 200. 600
0, 10, 50, 250
0, Ł, 37, 75, 150
0, 100. 200
0, 10, 5Q, 250
Mean exposure levels
<140 ppm; mean blood
levels 1,685 ug/L
Mean concentration
for dry cleaners in
Nordic countries,
1964-1979 = 24 ppm
[from Lynge et al.
(2006)d1
11 ppm average
concentration,
personal samples
(n = 208), any job title,
all sample durations
[Table II, Gold et al.
(2006)]
Effect
Impairment in steady
state visual evoked
potential
Impairment in steady
state visual evoked
potential
Urine and serum markers
of nephrotoxicity
Increased karyomegaly
(74%), megalonuclear-
cytosis
Increased relative kidney
weight; karyomegaly in
proximal tubules
Increased relative kidney
weight; karyomegaly in
proximal tubules
Increased liver weight
Increased liver
degeneration, necrosis
Increased angiectasis
Reduced RBC count,
reduced hemoglobin,
increased WBC count,
increased lymphocytes,
increased IgE
Time to pregnancy
Time to conception
4-439
-------
Table 4-49. Inhalation studies suitable for dose-response analyses (continued)
Organ/
system
Reproduct
ive and
develop-
mental
toxicity
(continue
d)
Study
Olsen et al. (1990):
Kyyronen et al. (1989)
Nelson et al. (1979)
Bellies et al. (1980)
Tinston (1994)
Carney et al. (2006)
Species
Human
Rat
Mouse
Rat
Rat
Duration/dosing
1st trimester,
occupational,
1973-1983
7 h/d on CDs 7-13
or 14-20
5 d exposure;
1, 4, and 10 wk
follow-up
Developmental —
multigeneration; 6
h/d, 5 d/wk
Developmental — 6
h/d on CDs 6-19
NOAEL/LOAEL3
(ppm)
4=Jppmd
0. 100. 900
0. 100. 500
0, 100. 300. 1,000
0, 65, 250. 600
Effect
Spontaneous abortion
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 vertebral centra
"Experimental/observational NOAEL is underlined; LOAEL is double-underlined.
bSchreiber et al. (2002) found mean PCE 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 mo 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).
°Echeverria et al. (1995)—the lowest exposure group is chosen to represent the LOAEL; (3 coefficient for lifetime or
chronic PCE exposure was positive and statistically significant for pattern memory, pattern recognition, and
pattern reproduction.
Low group (working at dry cleaners but not operator or spot removal >1 h/d); Calculated from mean concentration
for dry cleaners 1964-1979 [24 ppm, Lynge et al. (2006)1 divided by ratio of exposure for operators versus other
work in dry cleaners. Chose a ratio of 5:1 as an intermediate level between 7:1 from Gold et al. (2008) (pg. 816)
that included transfer type machines in the United States and 3.5:1 from Raisanen et al. (2001) which included
only dry to dry, primarily nonvented machines in Finland.
4-440
-------
Table 4-50. NOAELs and LOAELs in selected studies involving oral
exposure to tetrachloroethylene
Organ/
System
CNS
Kidney
Liver, kidney
Liver
Liver
Hematologic
toxicity
Development
Study
Fredriksson et
al. (1993)
NCI (1977)
Jonker et al.
(1996)
Herman et al.
(1995)
Buben and
O'Flaherty
(1985)
Marth et al.
(1989: 1985a;
1985b): Marth
(1987)
Bove et al.
(1995)
Species
Mouse
Mouse, Rat
Rat
Rat
Mouse
(40 g)
Mouse
(2 wk old,
20 g)
Human
Duration/exposure
Route
PND 10-16/oral
gavage
Chronic
(78 wk)/oral gavage
4 wk/oral gavage
14 d/oral gavage
6 wk/oral gavage
7 wk/drinking water
1st trimester,
drinking water,
1985-1988
NOAEL/LOAEL"
(mg/kg-day)
0, 5, 320
0, 536. 1,072 (male mice);
0, 386. 772 (female mice);
0, 475. 950 (male and
female rats)
0, 600, 2,400
0, 50, 150, 500, 1.500.
5,000
0, 20, 1QQ, 200, 500,
1,000, 1,500, 2,000
0,0.05,0.1
<1.3.5. 7.5 and >10 ug/Lb
Effect
Day 60: Increased
locomotion,
decreased rearing
Toxic
nephropathy
Liver weight,
enzyme levels;
kidney weight,
kidney enzyme
levels
Liver weight,
ALT
Liver weight,
triglycerides
Reversible
hemolytic anemia,
increased serum
triglycerides,
decreased
cholesterol
Oral clefts
"NOAELs are underlined once; LOAELs are double-underlined.
bBove et al. (1995) reported risks for categories of drinking water concentration of <1, >l-5, >5-10, and >10 ug/L.
Exposure levels are the midpoints of these exposure categories. Supported by Aschengrau et al. (2009b) who
observed an increased risk of oral clefts associated with any exposure to PCE versus no exposure (1-5, 197 ug/L).
Visual contrast sensitivity deficits as well as color discrimination deficits are commonly
present prior to detectable pathology in the retina or optic nerve head. These deficits are, thus,
among the earliest signs of disease and potentially more sensitive measures than evoked
potentials from visual stimuli (Regan, 1989). Several independent lines of evidence can be
found in the occupational and residential exposure studies to support an inference of visual
deficits following chronic tetrachloroethylene exposure. The studies that observed effects on
color vision using the Lanthony D-15 color vision test include cross-sectional and longitudinal
4-441
-------
designs in dry cleaning (Gobba etal., 1998; Cavalleri etal., 1994) and residential (Schreiber et
al., 2002) settings. Decrements in color confusion were reported among all workers exposed to a
mean TWA of 6 ppm for an average of 8.8 years (Cavalleri et al., 1994). A significant dose-
response relationship between CCI value and tetrachloroethylene concentration (p < 0.01) was
also observed in Cavalleri et al. (1994). As noted previously, the color vision testing in this
study was blinded to exposure level of the study participants, and the study participants were
well matched in terms of age, smoking, and alcohol use. A follow-up of these workers 2 years
later (Gobba et al., 1998) showed greater loss in color discrimination in those who were
subsequently exposed to a higher concentration [increase in geometric mean from 1.7 to 4.3
ppm], with no change in those exposed to lower concentrations [decrease in geometric mean
from 2.9 to 0.7 ppm]). Although Gobba et al. (1998) demonstrated persistent color confusion
effects in this follow up evaluation, the study exposures are not clearly characterized over the
course of the 2-year duration. Nakatsuka et al. (1992) did not observe an association with color
vision among dry cleaners in China (n = 64, geometric mean: TWA 11 and 15 ppm in females
and males, respectively), but the relative insensitivity of the specific type of color vision test
used in this study (Lanthony, 1978) is a likely explanation for these results. Effects on color
vision were also observed among 14 dry cleaners in the small study in Malaysia by
Sharanjeet-Kaur et al. (2004), but this study provides little weight to the strength of the evidence
because of the lack of exposure information (other than job title), and differences between dry
cleaners and controls regarding test conditions and smoking habits. Two other small studies also
reported lower scores on the Lanthony D-15 color vision test in much lower exposure settings,
but the differences were not statistically significant. A study of residents living above dry
cleaners (mean tetrachloroethylene exposure during active dry cleaning = 0.4 ppm), reported
mean CCI scores of 1.33 and 1.20 for 17 exposed and 17 controls, respectively (p = 0.26). A
study of workers in a day care center located in a building with a dry-cleaning business (mean
tetrachloroethylene exposure: 0.32 ppm) reported mean CCI scores of 1.22 and 1.18 in the
exposed daycare workers and controls, respectively (p = 0.39) (Schreiber et al., 2002). Overall,
the evidence reveals a high degree of consistency in this aspect of visually mediated function.
Visual contrast sensitivity changes were reported in two NYSDOH residential studies. In
a small pilot study (4 children and 13 adults), mean scores for visual contrast sensitivity (using a
near vision visual contrast sensitivity test) across spatial frequencies were statistically
significantly lower in exposed residents than in controls, indicating poorer visual function in the
exposed groups (Schreiber et al., 2002). Controls were age- and sex-matched to the exposed
group, and both groups were English speaking and of predominately Caucasian ethnicity;
however, they were drawn from different geographic areas. In addition, two of the four exposed
children had diagnoses of learning disabilities or developmental delays, which could affect
4-442
-------
performance on this type of test. In the larger study (NYSDOH, 2010. 2005a, b), the test
(Functional Acuity Contrast Test, FACT) assessed far vision visual contrast sensitivity, and the
test had a low rate of detecting visual contrast changes. For contrast vision, a number of
analyses in NYSDOH (Storm et al.. 2011 [previously reported in NYSDOH, 20101: NYSDOH,
suggest a vulnerability among children. However, exposure to >0.015 ppm (>100 |ig/m3)
tetrachloroethylene was highly correlated with race and children's age. Additionally, the sample
sizes in the highest exposure group, especially in higher income, nonminority groups, makes it
difficult to fully examine possible effects of income, race, and age on vision. Therefore, while
both studies report visual contrast sensitivity changes, with exposed children being more
sensitive, there are concerns with the methodological and analytic approaches in these studies.
Acute human exposure studies reported increased latencies of up to 3.0 ms in visual
evoked potentials (Altmann et al., 1990) and changes in EEGs (magnitude of effect was not
specified) (Hake and Stewart, 1977; Stewart et al., 1970) at higher exposures ranging from 340
to 680 mg/m3.
In rats, acute inhalation exposure to tetrachloroethylene results in significant changes to
the flash-evoked potential at 800 ppm (Mattsson et al., 1998) and a decrease in F2 amplitudes of
the steady state visual evoked potential at 250 ppm (Boyes et al., 2009). In a subchronic
exposure study (13 weeks, up to 800 ppm tetrachloroethylene), changes in flash-evoked potential
responses were not observed at tetrachloroethylene exposures up to 200 ppm. In the 800 ppm
group, there was a significant increase in the amplitude and a significant increase in latency
(-3.0 ms) of the mid-flash-evoked potential waveform (N3), but histopathological lesions were
not observed in the examination of the visual system brain structures [e.g., visual cortex; optic
nerve; Mattsson et al. (1998)1.
Effects on visuospatial memory in humans were also reported in each of the studies that
examined this measure (Altmann et al., 1995; Echeverria et al., 1995; Echeverria et al., 1994;
Seeber, 1989). These effects (increased response times or cognition errors) were observed in
occupational and residential studies, and the occupational studies were quite large, involving
101, 65, and 173 dry-cleaning workers in Seeber (1989), Echeverria et al. (1995), and Echeverria
et al. (1994), respectively. Several different types of tests were used including digit reproduction
(Seeber, 1989), switching, pattern memory, and pattern recognition (Echeverria et al., 1995;
Echeverria et al., 1994), and the Benton test (Altmann et al., 1995). Exposure ranges for the
increased reaction time observations (LOAELs) ranged from 4.99 to 102 mg/m3 (Altmann et al.,
1995; Echeverria et al., 1995; Ferroni et al., 1992). The changes in the cognitive tasks were
observed at exposures (LOAELs) ranging from 53.9 to 364.22 mg/m3 (Spinatonda et al., 1997;
Echeverria et al., 1995; Seeber, 1989). All of these studies except Altmann et al. (1995) indicate
that the neurobehavioral assessment was blinded to knowledge of the exposure level of the
4-443
-------
subject, and all of the studies adjusted for potentially confounding factors. It should be noted,
however, that residual confounding from education-level differences between exposed and
referent subjects may still be present in Altmann et al. (1995).
Increased reaction time, increased number of false alarms, and decreased trial
completions in a signal detection task (measures of decreased attention) were reported in an
acute (60 minutes) exposure (6,782 mg/m3 or higher) study in rats (Oshiro et al., 2008).
Additionally, operant tasks that test cognitive performance have demonstrated performance
deficits in rats and mice following acute tetrachloroethylene oral (Warren et al., 1996) and i.p.
(Umezu et al., 1997) exposures. These findings are consistent with observed effects on cognition
and memory in humans. However, no studies, to date, have evaluated the persistent effects of
tetrachloroethylene exposure on cognitive performance deficits in animal models.
An occupational exposure study (n=60) (Ferroni et al., 1992) and a residential exposure
study (n = 14) (Altmann et al., 1995), with mean exposure levels of 15 and 0.7 ppm,
respectively, reported significant increases in simple reaction time of 24 ms (11% increase)
(Ferroni etal., 1992) and 40 and 51.1 ms (15 and 20% increases', respectively', for two separate
measurements) (Altmann et al., 1995) for the exposed subjects. A third study, Lauwerys et al.
(1983), reported better performance on simple reaction time in 21 exposed workers (mean TWA:
21 ppm) compared with controls when measured before a work shift but not when measured
after work.
The changes in brain weight, DNA/RNA, and neurotransmitter levels that were observed
in the animal studies are highly supportive of the neurobehavioral changes observed with
tetrachloroethylene exposure. Changes in brain DNA, RNA, or protein levels and lipid
composition were altered following inhalation, with changes observed in cerebellum,
hippocampus, and frontal cortex (Wang et al., 1993; Rosengren et al., 1986; Savolainen et al.,
1977a: Savolainen et al., 1977b). The replication of these changes in biochemical parameters
and effects in brain weight in both rats and gerbils is pathognomonic. Changes in
neurotransmitters systems (Briving et al., 1986; Honmaetal., 1980a: Honmaetal., 1980b) and
circadian rhythm (Motohashi etal., 1993) in animal studies are consistent with neuroendocrine
alterations observed in humans (Ferroni et al., 1992).
In conclusion, the weight of evidence across the available studies of humans and animals
exposed to tetrachloroethylene indicates that chronic exposure to tetrachloroethylene can result
in decrements in color vision, visuospatial memory, and possibly other aspects of cognition and
neuropsychological function, including reaction time.
4-444
-------
4.10.2.2. Kidney Toxicity
The epidemiologic studies support association between inhalation tetrachloroethylene
exposure and chronic kidney disease, as measured by urinary excretion of renal proteins and
ESRD. The elevated urinary RBP levels observed in two studies (Verplanke et al., 1999; Mutti
et al., 1992) and lysozyme or p-glucuronidase in Franchini et al. (1983) provide some evidence
for effects to the proximal tubules from tetrachloroethylene exposure. Exposures in the studies
that observed renal toxicity were time-weighted averages of 8, 10, and 15 ppm. None of the
reviewed studies reported exposure-response relationships, and this is an important limitation of
the available data. Calvert et al. (2011) supports an association between inhalation
tetrachloroethylene exposure and ESRD, particularly hypertensive ESRD, and observed a
twofold elevated incidence among subjects who worked only in a shop where tetrachloroethylene
was the primary cleaning solvent compared to that expected based on U.S. population rates. An
exposure-response pattern was further suggested because hypertensive ESRD risk was highest
among those with longest employment durations. No human studies investigating drinking water
or other oral tetrachloroethylene exposures on kidney toxicity have been published.
Adverse effects on the kidney have been observed in studies of rodents exposed to high
concentrations of tetrachloroethylene by inhalation (JISA, 1993; NTP, 1986), oral gavage
(Ebrahim et al.. 2001: Ebrahim et al.. 1996: Jonkeretal., 1996: Green etal.. 1990: Goldsworthy
et al., 1988: NCI, 1977), and i.p. injection of tetrachloroethylene metabolites (Elfarra and
Krause, 2007). The nephrotoxic effects include increased kidney-to-body weight ratios, hyaline
droplet formation, glomerular "nephrosis," karyomegaly (enlarged nuclei), cast formation, and
other lesions or indicators of renal toxicity. The male rat has been shown to be more sensitive to
nephrotoxicity following exposure to tetrachloroethylene. These findings support a LOAEL of
200 ppm and a NOAEL of 50 ppm. Overall, multiple lines of evidence support the conclusion
that tetrachloroethylene causes nephrotoxicity in the form of tubular toxicity, mediated
potentially through the tetrachloroethylene GSH conjugation products TCVC and TCVCS.
4.10.2.3. Liver Toxicity
Two of four studies of occupationally exposed dry cleaners showed early indications of
liver toxicity, namely sonographic changes of the liver and altered serum concentrations of one
enzyme indicative of liver injury (Brodkin et al., 1995: Gennari et al., 1992). Frank liver disease
was not observed among these workers nor were changes in other biomarkers indicative of liver
toxicity (e.g., serum transaminases), which was not unexpected, given subjects with signs of liver
disease were excluded in both studies. LOAELs in these human studies were between 12 and
16 ppm (TWA).
4-445
-------
Liver toxicity has been reported in multiple animal species by inhalation and oral
exposures to tetrachloroethylene. The effects are characterized by increased liver weight, fatty
changes, necrosis, inflammatory cell infiltration, triglyceride increases, and proliferation. The
mouse has been shown to be more sensitive to hepatic toxicity than the rat in multiple subchronic
and chronic studies [e.g., JISA (1993): NTP (1986): Schumann et al. (1980): NCI (1977)1. After
subchronic or chronic inhalation exposures in mice, liver toxicity is manifested by increased liver
weight (Ki ell strand etal.. 1984). liver enlargement (Odumetal., 1988: Ki ell strand etal.. 1984).
cytoplasmic vacuolation (fatty changes) (Odumetal.. 1988: NTP. 1986: Ki ell strand et al.. 1984).
centrilobular hepatocellular necrosis (JISA, 1993: NTP, 1986), and inflammatory cell infiltrates,
pigment in cells, oval cell hyperplasia, and regenerative foci (NTP, 1986). The LOAEL for the
inhalation studies, 9 ppm, is from a 30-day continuous exposure mouse study reporting increased
liver weight and morphological changes, and is supported by a finding of irreversible
macromolecular binding in mouse liver following a single, 6-hour exposure at 10 ppm. The
JISA (1993) chronic mouse inhalation bioassay reported liver necrotic foci at 50 ppm and higher.
With oral administration in mice, liver toxicity (increased liver weight, hepatocellular
swelling, necrosis, lipid accumulation, and increased DNA synthesis) has been observed at
100 mg/kg-day (Buben and O'Flaherty, 1985: Schumann et al., 1980) and above (Ebrahim et al.,
1996: Jonkeretal.. 1996: Berman et al.. 1995: Goldsworthv andPopp. 1987). At 150 mg/kg-day
administered for 30 days (Philip et al., 2007), tetrachloroethylene increased ALT levels
transiently and stimulated fatty degeneration and necrosis, with ensuing regenerative repair.
These findings support a LOAEL of 100 mg/kg-day and a NOAEL of 20 mg/kg-day.
4.10.2.4. Immunotoxicity and hematologic toxicity
The strongest human study examining immunologic and hematologic effects of
tetrachloroethylene exposure in terms of sample size and use of an appropriately matched control
group is of 40 male dry-cleaning workers (mean exposure levels: <140 ppm; mean duration:
7 years; mean blood tetrachloroethylene levels: 1,685 |ig/L) by Emara et al. (2010). Statistically
significant decreases in red blood cell count and hemoglobin levels and increases in total white
cell counts and lymphocyte counts were observed in the exposed workers compared to age- and
smoking-matched controls. Similar effects were observed in mice (Ebrahim et al., 2001). In
addition, increases in several other immunological parameters, including T-lymphocyte and
natural killer cell subpopulations, IgE, and interleukin-4 levels were observed in
tetrachloroethylene-exposed dry-cleaning workers (Emara et al., 2010). These immunologic
effects suggest an augmentation of Th2 responsiveness. However, the limited available data
from studies in children (Delfmo et al., 2003a; Lehmann et al., 2002; Lehmann et al., 2001) do
not provide substantial evidence of an effect of tetrachloroethylene exposure during childhood on
4-446
-------
allergic sensitization or exacerbation of asthma symptomology. The observation of the
association between increased tetrachloroethylene exposure and reduced interferon-y in cord
blood samples may reflect a sensitive period of development and points to the current lack of
understanding of the potential immunotoxic effects of prenatal exposures. The available data
pertaining to risk of autoimmune disease in relation to tetrachloroethylene exposure are limited
for ascertainment of disease incidence and exposure-assessment difficulties in population-based
studies.
The available data from experimental studies assessing immunotoxic responses in
animals are very limited (Hanioka et al., 1995b: Germolec et al., 1989; Aranyi et al., 1986).
Additional data from inhalation, oral, and dermal exposures of different durations are needed to
assess the potential immunotoxicity of tetrachloroethylene along multiple dimensions, including
immunosuppression, autoimmunity, and allergic sensitization. The data from Aranyi et al.
(1986) suggest that short-term exposures may result in decreased immunological competence
(immunosuppression) in CD-I mice. The relative lack of data, taken together with the concern
that other structurally related solvents have been associated with immunotoxicity, particularly
relating to autoimmune disease (Cooper et al., 2009), contributes to uncertainty in the database
for tetrachloroethylene. The limited laboratory animal studies of hematological toxicity
demonstrated an effect of tetrachloroethylene exposure on red blood cells (decreased RBCs)
(Ebrahim et al., 2001), or decreased erythrocyte colony-forming units (Seidel et al., 1992), with
reversible hemolytic anemia observed in female mice exposed to low drinking water levels
(0.05 mg/kg-day) of tetrachloroethylene beginning at 2 weeks of age in one series of studies
(Marth et al., 1989: Marth, 1987: Marth et al., 1985a: Marth et al., 1985b). Ebrahim et al. (2001)
also observed decreased hemoglobin, platelet counts, and packed cell volume, and increased
WBC counts. Although limited experimental animal studies examining the immunotoxicity and
hematologic toxicity of tetrachloroethylene are available in the peer-reviewed published
literature, the results of these studies support the human epidemiology studies described above.
4.10.2.5. Reproductive and Developmental Toxicity
4.10.2.5.1. Reproductive toxicity
The literature contains few studies of effects on reproduction among subjects with
exposure to tetrachloroethylene. One study of primarily unionized workers in the dry-cleaning
and laundry industries in California observed subtle deficits in sperm quality in relation to
increasing levels of three measures of exposure, including tetrachloroethylene in exhaled breath
(Eskenazi etal., 199la). However, three clinically recognized measures of sperm quality were
not associated with exposure in the study population. The results of Eskenazi et al. (1991a) are
compelling, but more studies are needed to understand the spectrum of effects on sperm and their
4-447
-------
impact on fecundity. Some studies that relied on detailed work histories and monitoring data to
classify exposure suggested that maternal or paternal exposure to tetrachloroethylene or work in
dry cleaning reduces fertility or delays conception (Sallmen et al., 1998; Sallmen et al., 1995;
Eskenazi et al., 1991b). However, the risk estimates were imprecise because the number of
participants reporting exposure to tetrachloroethylene was small. As a consequence, the existing
literature is limited and inconclusive concerning effects of tetrachloroethylene on reproduction
and fertility.
Results of several studies of maternal occupational exposure to tetrachloroethylene
suggest an increased risk of spontaneous abortion, particularly at higher levels of exposure
(Doyle etal.. 1997: Windham et al.. 1991: Lindbohm et al.. 1990: Olsenetal.. 1990: Kyyronen
et al., 1989). Most of the studies evaluated exposure during the first trimester of pregnancy.
Some of the studies observed an increased odds ratio ranging from 1.4 to 4.7, but risk estimates
were statistically imprecise, and some studies were limited in their ability to evaluate potential
confounding (Windham et al., 1991: Lindbohm et al., 1990: Olsen et al., 1990: Bosco et al.,
1987). In general, the studies that used a more precise definition of exposure, or categorized
exposure into levels of increasing dose or intensity, observed higher risk estimates (Doyle et al.,
1997: Lindbohm et al., 1990: Olsenetal., 1990: Kyyronen et al.. 1989). The Finnish studies
controlled for reported exposure to other substances in the workplace as well as for several
potential confounders. Increased risks were not found among dry cleaners in Sweden using a
similar study design (Ahlborg, 1990a: Olsen et al., 1990). Although there is no evidence of an
increased risk associated with paternal exposure, the studies were not of sufficient size or detail
in exposure estimates to draw conclusions (Eskenazi etal., 199 la: Lindbohm et al., 1991:
Taskinen et al., 1989). No associations with incidence of spontaneous abortion were observed
between two populations exposed to tetrachlorethylene in drinking water (Aschengrau et al.,
2008: Lagakos etal., 1986). The populations were likely exposed to lower levels compared to
the occupational populations. In addition, the window of exposure used to assess risk in both
studies may not have been precise enough to detect a small elevation in risk for spontaneous
abortion.
The database of experimental animal studies for tetrachloroethylene includes evaluations
of reproductive and fertility outcomes in rats and mice following inhalation exposures.
Additionally, an in vitro assay of oocyte fertilizability is available. An assessment of fertility
and reproductive function in rats exposed to tetrachloroethylene via inhalation over the course of
two generations was conducted by Tinston (1994). Effects on offspring included decreased pup
weights and postnatal survival in both generations, as well as behavioral alterations in the Fl
pups. Decreased mean testes weight was observed in Fla males; however, no effects on male or
female fertility or other evidence of alterations in reproductive function were observed. For
4-448
-------
males, this finding is supported by the results of a study by Beliles et al. (1980), who found no
sperm abnormalities in rats following up to 10 weeks of tetrachloroethylene inhalation
exposures. While Beliles et al. (1980) identified an increase in abnormal sperm heads in mice
after 4 weeks of exposure, no other reproductive toxicity data in mice were available to aid in the
interpretation of this finding. A limitation of the Tinston (1994) study included a concern about
a short peri-parturition exposure gap. Additionally, the study was conducted in accordance with
standard EPA and OECD toxicological study guidelines in place at the time but did not assess
endpoints that are included in the guidelines that were revised and harmonized in 1998.
4.10.2.5.2. Developmental toxicity
A few epidemiologic studies evaluated developmental toxicity endpoints such as
decreased birth weight, intrauterine growth restriction (IUGR; also known as small for gestation
age [SGA]), and congenital anomalies. Overall, no associations were noted in several studies
that assessed maternal or paternal occupational exposure to tetrachloroethylene and increased
incidence of stillbirths, congenital anomalies, or decreased birth weight (Lindbohm, 1995;
Windham et al.. 1991: Olsenetal., 1990: Kyyronen et al.. 1989: Taskinen et al.. 1989: Boscoet
al., 1987). However, congenital anomalies were analyzed as a combined group, and the number
of exposed cases for specific types of anomalies was not sufficient to evaluate risk with
statistical precision. Some studies of tetrachloroethylene in drinking water reported that
exposure during pregnancy is associated with low birth weight (Bove et al., 1995: Lagakos et al.,
1986), eye/ear anomalies (Lagakos et al., 1986), and oral clefts (Aschengrau et al., 2009b: Bove
et al., 1995: Lagakos et al., 1986). No associations with tetrachloroethylene exposure were
reported for small for gestational age (Bove et al., 1995) or other classifications of congenital
anomalies [e.g., musculoskeletal, cardiovascular (Lagakos et al., 1986)]. Although a small
increase in risk of small for gestational age was reported for infants exposed prenatally to
tetrachloroethylene at the Camp Lejeune military base, the finding remains inconclusive until
ATSDR completes its reanalysis. Aschengrau et al. (2008) did not observe associations with
birth weight or gestational age in a Cape Cod population living in communities receiving
drinking water containing a wide range of tetrachloroethylene concentrations. Participants in
some of the studies of drinking water contamination were exposed to multiple pollutants (Bove
et al., 1995: Lagakos etal., 1986), and it was not possible to disentangle substance-specific risks.
Diagnoses of attention deficit or educational histories reported by the mothers were not increased
in relation to the amount of tetrachloroethylene delivered to the homes during pregnancy or
childhood (Janulewicz et al., 2008). Finally, a more than threefold risk of schizophrenia was
associated with dry cleaning as a surrogate for prenatal tetrachloroethylene exposure (Perrin et
al., 2007). The longitudinal design and use of a national registry to identify psychiatric
4-449
-------
diagnoses were strengths of the study, but tetrachloroethylene exposure was not directly
analyzed.
The animal inhalation study database includes assessments of developmental toxicity in
rats, mice, and rabbits following exposures during gestation and assessments of developmental
neurotoxicity in rats following pre- and/or postnatal exposures of the offspring. Additional
supportive studies include in vitro assays of embryo development and oocyte fertilizability, a
developmental assay in Japanese medaka, and two oral gavage studies that assessed
developmental toxicity in rats and developmental neurotoxicity in mice. The tetrachloroethylene
database included assessments of the various potential manifestations of developmental toxicity,
i.e., alterations in survival, growth, morphology, and functional development. Indications of
effects on prenatal survival following in utero exposure included increased pre- and/or
postimplantation loss in rats, mice, and rabbits (Szakmary et al., 1997; Schwetz etal., 1975).
These findings were supported by evidence of embryo mortality in a rat whole embryo culture
(WEC) assay (Saillenfait et al., 1995) and decreased viability in a Japanese medaka assay
(Spencer et al., 2002). Decreased prenatal growth was observed in mice (Schwetz et al., 1975)
and rats (Szakmary et al., 1997). Morphological alterations associated with prenatal exposures to
tetrachloroethylene included delays in skeletal ossification in mice (Schwetz et al., 1975) and
rats (Carney et al., 2006; Szakmary et al., 1997), which were often associated with fetal weight
decrements, and increased incidences of malformations in mice, rats, and rabbits (Szakmary et
al., 1997). Evidence of tetrachloroethylene exposure-related malformations was also observed in
the rat WEC and medaka assays (Spencer et al., 2002; Saillenfait et al., 1995) and in a gavage
prenatal developmental toxicity screening study in rats (Narotsky and Kavlock, 1995).
Alterations in neurological function following pre- and/or postnatal inhalation exposures to
tetrachloroethylene were observed in rats by Szakmary et al. (1997), Nelson et al. (1979), and
Tinston (1994). These findings were supported by a study that found altered spontaneous motor
activity in young adult rats that had been exposed orally to tetrachloroethylene postnatally during
a critical period of nervous system development (Fredriksson et al., 1993). Additionally,
reductions in brain acetylcholine and dopamine were observed in rat offspring following
gestational tetrachloroethylene exposures (Nelson et al., 1979). Limitations of the inhalation
developmental toxicity studies include the lack of dose-response information due to the use of a
single treatment level in the prenatal developmental toxicity assessment by Schwetz et al. (1975):
the lack of either maternal or developmental toxicity in Hardin et al. (1981): and absence of
methodological details in study reporting (Szakmary et al., 1997).
4-450
-------
4.10.2.5.3. Synthesis of human and animal reproductive and developmental toxicity
The finding of spontaneous abortions in several human studies of dry cleaners is
supported by the occurrence of reduced birth weight and mortality in several animal studies.
Although not a consistent finding in epidemiology studies, the finding of low birth weight in a
study of contaminants in drinking water (Bove et al., 1995) is supported by reduced birth weight
in five animal studies (Carney et al., 2006; Szakmary et al., 1997; Nelson et al., 1979; Schwetz et
al., 1975) and in the Fl generation but not the F2 generation of Tinston (1994). There are no
human observations of behavioral changes to compare with the animal evidence of CNS effects.
The subtle effects on sperm observed in humans (Eskenazi et al., 199la) correspond to one
report of abnormal sperm in mice. Overall, the developmental and reproductive toxicity
database for tetrachloroethylene was judged to include a range of data from appropriate well-
conducted studies in several laboratory animal species plus limited human data and was
considered sufficient for hazard characterization and dose-response assessment, based upon EPA
risk assessment guidelines (U.S. EPA, 2006b, 1991b). Based upon a consideration of the
available database of animal developmental and reproductive toxicity studies for
tetrachloroethylene, the overall inhalation NOAEL is 100 ppm, based on Tinston (1994). The
overall inhalation 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.
4.10.2.6. Summary of Noncancer Toxicities and Identification of Studies for Dose-Response
Analyses
Noncancer effects of tetrachloroethylene identified in exposed humans and animals
include toxicity to the central nervous, renal, hepatic, immune, and hematologic systems, and on
development and reproduction. Neurotoxic effects have been characterized in human
occupational and residential studies, as well as in experimental animal studies, providing
evidence of an association between tetrachloroethylene exposure and neurobehavioral deficits.
Tetrachloroethylene exposure primarily results in visual changes, increased reaction time, and
decrements in cognition in humans; in animal studies, effects on vision, visual-spatial function,
and reaction time, as well as brain-weight changes were also seen. Adverse effects on the kidney
in the form of tubular toxicity, potentially mediated through the tetrachloroethylene GSH
conjugation products TCVC and TCVCS, have been reported in numerous well-conducted
animal studies. Although epidemiological studies have not systematically investigated
nephrotoxicity, an association between inhalation tetrachloroethylene exposure and chronic
kidney disease, as measured by urinary excretion of renal proteins and ESRD, is supported. The
developmental and reproductive toxicity database for tetrachloroethylene includes a range of
data from appropriate well-conducted studies in several laboratory animal species plus limited
4-451
-------
human data. Evidence of liver toxicity is primarily from several well-conducted rodent studies,
including chronic bioassays.
Other toxicity endpoints are less well characterized. The few published reports of
experimental studies examining immune or hematologic system toxicity are consistent with the
limited findings in the human occupational exposure studies. The limited laboratory animal
studies of hematological toxicity demonstrated an effect of tetrachloroethylene exposure on red
blood cells (decreased RBCs (Ebrahim et al., 2001), or decreased erythrocyte colony-forming
units (Seidel et al., 1992)), with reversible hemolytic anemia observed in female mice exposed to
low drinking water levels (0.05 mg/kg-day) of tetrachloroethylene beginning at 2 weeks of age in
one series of studies (Marth et al., 1989: Marth, 1987: Marth et al., 1985a: Marth et al., 1985b).
Ebrahim et al. (2001) also observed decreased hemoglobin, platelet counts, and packed cell
volume, and increased WBC counts. The relative lack of additional data, including confirmatory
reports of immunotoxic or hematologic toxicity with low continuous exposures beginning in
early lifestages, taken together with evidence of immunotoxicity from structurally related
solvents (Cooper et al., 2009), contributes to uncertainty in the database for tetrachloroethylene.
No epidemiological studies identified potential noncancer respiratory toxicities, and no lung
effects in rodents were reported in chronic bioassays (NTP, 1986: NCI, 1977) or other published
reports.
The tables above present the inhalation (refer to Table 4-49) and oral (refer to Table
4-50) findings of tetrachloroethylene toxicity, arranged by organ, which are suitable for dose-
response analyses. The NOAELs and LOAELs from candidate dose-response studies are
identified. In examining the studies judged to be suitable for dose-response analyses, it is
evident that the neurotoxicological findings consistently occur at the lowest exposure levels.
Additionally, the database for neurotoxicity comprises a number of both occupational and
residential human studies as well as animal studies that are suitable for dose-response analyses.
Residential inhalation exposures to tetrachloroethylene resulted in visual contrast sensitivity
changes and cognitive and motor changes at exposures approximately 5- to 10-fold lower than
the lowest sensitive exposure for other toxicological endpoints. Similarly, with oral doses,
developmental neurotoxicity effects were observed at levels at least fivefold lower (Fredriksson
et al., 1993). Therefore, the CNS effects are identified as a sensitive endpoint following either
oral or inhalation exposure to tetrachloroethylene. Section 5 presents dose-response analyses of
the neurotoxicity data set as a basis for derivation of inhalation and oral reference values.
Quantitative dose-response analyses of the findings for other endpoints (i.e., kidney, liver,
reproductive and developmental toxicity) are also presented in Section 5. In addition to
providing information regarding the relative sensitivity of different organs/systems to
4-452
-------
tetrachloroethylene, such quantitative analyses may be useful for cumulative risk assessment in
which multiple chemicals have a common target organ/system other than the CNS.
4.10.3. Characterization of Cancer Hazard
Following EPA (2005a) Guidelines for Carcinogen Risk Assessment, tetrachloroethylene
is "likely to be carcinogenic in humans by all routes of exposure." This characterization is based
on suggestive evidence of carcinogenicity in epidemiologic studies and conclusive evidence that
the administration of tetrachloroethylene, either by ingestion or by inhalation to sexually mature
rats and mice, increases tumor incidence (JISA, 1993; NTP, 1986; NCI, 1977). Several rodent
tumor types were significantly increased with tetrachloroethylene administration in at least two
studies. Mouse liver tumors (hepatocellular adenomas and carcinomas) and rat mononuclear cell
leukemia were reported in both sexes in two lifetime inhalation bioassays employing different
rodent strains, and mouse liver tumors were also reported in both sexes in an oral bioassay (NCI,
1977). Tumors reported in single inhalation bioassays include kidney and testicular interstitial
cell tumors in male F344 rats (NTP, 1986), brain gliomas in male and female F344 rats (NTP,
1986), and hemangiomas or hemangiosarcomas in male Crj :BDF1 mice (JISA, 1993). Several
metabolites of tetrachloroethylene have also been analyzed for rodent carcinogenicity. TCA and
DCA produce liver tumors in mice, and DCA also induces liver tumors in rats. Other
tetrachloroethylene metabolites have not been tested in a rodent bioassay.
The specific active moiety(ies) and mode(s) of action involved in the carcinogenicity of
tetrachloroethylene and its metabolites are not known. For rat kidney tumors, it is generally
believed that metabolites resulting from GSH conjugation of tetrachloroethylene are involved.
The hypothesized modes of action for this endpoint include mutagenicity, peroxisome
proliferation, a2u-globulin nephropathy, and cytotoxicity not associated with a2u-accumulation.
For mouse liver tumors, it is generally believed that metabolites resulting from P450-mediated
oxidation of tetrachloroethylene are involved. The mode of action (MO A) hypotheses for this
endpoint concern mutagenicity, epigenetic effects (especially DNA hypomethylation), oxidative
stress, and receptor activation (focusing on a hypothesized PPARa-activation MO A). However,
the available evidence is insufficient to support the conclusion that either rat kidney or mouse
liver tumors are mediated solely by one of these hypothesized modes of action. In addition, no
data are available concerning the metabolites or the mechanisms that may contribute to the
induction of other rodent tumors (including mononuclear cell leukemia, brain gliomas, or
testicular interstitial cell tumors in exposed rats and hemangiosarcomas in exposed mice).
Furthermore, no mechanistic hypotheses have been advanced for the human cancers suggested to
be increased with tetrachloroethylene exposure in epidemiologic studies, including bladder
cancer, non-Hodgkin lymphoma and multiple myeloma. Although tetrachloroethylene is largely
4-453
-------
negative in genotoxicity assays including in the Ames mutagenicity test, tetrachloroethylene has
been shown to induce modest genotoxic effects (micronuclei induction following in vitro or in
vivo exposure, and DNA binding and single-strand breaks in tumor tissue) and mutagenic effects
under certain metabolic activation conditions. In addition, some tetrachloroethylene metabolites
have been shown to be mutagenic. Thus, the hypothesis that mutagenicity contributes to the
tetrachloroethylene carcinogenesis cannot be ruled out for one or more target organs, although
the specific metabolic species or mechanistic effects are not known.
4.10.4. Synthesis of Epidemiologic Studies
The available epidemiologic studies provide a pattern of evidence associating
tetrachloroethylene exposure and several types of cancer, specifically bladder cancer,
non-Hodgkin lymphoma, and multiple myeloma. Associations and exposure-response
relationships for these cancers were reported in studies using higher quality (more precise)
exposure assessment methodologies for tetrachloroethylene. Confounding by common lifestyle
factors such as smoking are unlikely explanations for the observed results. For other sites,
including esophageal, kidney, lung, liver, cervical, and breast cancer, more limited data
supporting a suggestive effect are available.
With respect to bladder cancer, the pattern of results from this collection of studies is
consistent with an elevated risk for tetrachloroethylene of a relatively modest magnitude (i.e., a
10-40% increased risk). The effect estimates from five of the six studies with relatively high
quality exposure-assessment methodologies ranged from 1.44 to 4.03 [Calvert et al. (2011):
Lynge et al. (2006): Blair et al. (2003): Pesch et al. (2000b): Aschengrau et al. (1993)1. An
exposure-response gradient was observed in a large case-control study by Pesch et al. (2000b),
using a semi quantitative cumulative exposure assessment, with an adjusted odds ratio of 0.8
(95% CI: 0.6, 1.2), 1.3 (95% CI: 0.9, 1.7), and 1.8 (95% CI: 1.2, 2.7) for medium, high, and
substantial exposure, respectively, compared to low exposure. A similar exposure-response
pattern was not observed in the study by Lynge et al. (1995). This study examined exposure
duration, however, rather than a measure that incorporated information on exposure
concentration. In addition, relative risk estimates between bladder cancer risk and ever having a
job title of dry-cleaner or laundry worker in four large cohort studies ranged from 1.01 to 1.44
(Pukkala et al.. 2009: Wilson et al.. 2008: Ji et al.. 2005a: Travier et al.. 2002). As expected, the
results from the smaller studies are more variable and less precise, reflecting their reduced
statistical power. Confounding by smoking is an unlikely explanation for the findings, given the
adjustment for smoking by Pesch et al. (2000b) and in other case-control studies.
The results from the collection of studies pertaining to non-Hodgkin lymphoma also
indicate an elevated risk for tetrachloroethylene. The results from five cohort studies that used a
4-454
-------
relatively high quality exposure-assessment methodology generally reported relative risks
between 1.7 and 3.8 (Calvert et al.. 2011: Selden and Ahlborg. 2011: Radican et al.. 2008: Boice
et al., 1999: Anttila et al., 1995). There is also some evidence of exposure-response gradients in
studies with tetrachloroethylene-specific exposure measures based on intensity, duration, or
cumulative exposure (Seidler et al., 2007: Miligi et al., 2006: Boice et al., 1999). Higher
non-Hodgkin lymphoma risks were observed in these studies in the highest exposure categories,
with the strongest evidence from the large case-control study in Germany in which a relative risk
of 3.4 (95% CI: 0.7, 17.3) was observed in the highest cumulative exposure category (trend
/7-value = 0.12) (Seidler et al., 2007). Effect estimates in studies with broader exposure
assessments showed a more variable pattern (Selden and Ahlborg, 2011: Pukkala et al., 2009: !i
and Hemminki, 2006b: Blair etal.. 2003: Travier et al.. 2002: Cano and Pollan. 2001: Lynge and
Thygesen, 1990). Confounding by life-style factors are unlikely explanations for the observed
results because common behaviors, such as smoking and alcohol use, are not strong risk factors
for non-Hodgkin lymphoma (Besson et al., 2006: Morton et al., 2005).
Results from the multiple myeloma studies are based on a smaller set of studies than
those of non-Hodgkin lymphoma, but results are similar. The larger cohort studies that use a
relatively nonspecific exposure measure (broad occupational title of launderers and dry cleaners,
based on census data) do not report an increased risk of multiple myeloma, with effect estimates
ranging from 0.99 to 1.07 (Pukkala et al.. 2009: Ji and Hemminkl 2006b: Andersen et al.. 1999).
Some uncertainty in these estimates arises from these studies' broader exposure-assessment
methodology. Results from the cohort and case-control studies with a higher quality exposure-
assessment methodology, with an exposure measure developed specifically for
tetrachloroethylene, do provide evidence of an association, however, with relative risks of 7.84
(95% CI: 1.43, 43.1) in women and 1.71 (95% CI: 0.42, 6.91) in men in the cohort of aircraft
maintenance workers (Radican et al., 2008) and 1.5 (95% CI: 0.8, 2.9) in a case-control study in
Washington (Gold etal., 201 Ob): tetrachloroethylene exposure). Gold et al. also reported
increasing risks with increasing exposure duration (based on job titles) (Gold et al., 2010a) and
based on a cumulative tetrachloroethylene exposure metric (Gold etal., 2010b). A smaller case-
control study (n = 76 cases) with tetrachloroethylene-specific exposure measures based on
intensity, duration, or cumulative exposure, Seidler et al. (2007), observed no cases among the
highest exposure groups. A small cohort study by Boice et al. (1999) of aerospace workers
observed one death among routinely exposed subjects and six deaths among subjects with a
broader definition of routine or intermittent exposure.
Suggestive but limited evidence was also observed in the collection of epidemiologic
studies pertaining to tetrachloroethylene exposure and esophageal, kidney, lung, liver, cervical,
and breast cancer. One difference between these sets of data and the data for bladder cancer,
4-455
-------
non-Hodgkin lymphoma, and multiple myeloma is a more mixed pattern of observed risk
estimates and an absence of exposure-response data from the studies using a quantitative
tetrachloroethylene-specific cumulative exposure measure.
For esophageal cancer, the SIR in the only large cohort study (n = 95 cases), a study
using broad exposure categories, was 1.18 (95% CI: 0.96, 1.46) (Pukkala et al., 2009). The point
estimates of the association in seven of eight smaller studies, four studies with specific exposure
assessments, and four other studies with less precise assessments were between 1.16 and 2.44
(Calvert et al., 2011: Selden and Ahlborg, 2011: Pukkala et al.. 2009: Sung et al.. 2007: Blair et
al., 2003: Travier et al., 2002: Boiceetal., 1999: Lynge and Thygesen, 1990). Two small case-
control studies with relatively high quality exposure-assessment approaches, Lynge et al. (2006)
and Vaughan et al. (1997) reported an odds ratio of 0.76 (95% CI: 0.34, 1.69) and of 6.4 (95%
CI: 0.6, 68.9), respectively. Some uncertainties in these estimates arise from the lack of job title
information for 25% of the cases and 19% of the controls, and the variability in the results from
the sensitivity analysis using different assumptions regarding the correct classification of
individuals in this group or the small number of exposed cases. One of the two larger studies
examining exposure-response suggested a positive relationship (Calvert et al., 2011). Based on
smoking rates in blue-collar workers, the twofold risk estimate reported in Calvert et al. (2011)
and Blair et al. (2003) was higher than that which could reasonably be attributable to smoking.
One primary study that supports an association between tetrachloroethylene exposure and
kidney cancer, the largest international case-control study (245 exposed cases from Australia,
Denmark, Germany, Sweden, and the United States), reported a relative risk of 1.4 (95% CI: 1.1,
1.7) for any exposure to dry-cleaning solvents (Mandel etal., 1995). This study was able to
adjust for smoking history, BMI, and other risk factors for kidney cancer. Results from the large
cohort studies, using a more general exposure classification based on national census occupation
data, present more variable results, with relative risks of 0.94, 1.11, and 1.15 in Pukkula et al.
(2009), Travier et al. (2002), and Ji et al. (2005b), respectively. The results from the smaller
studies using a relatively specific exposure-assessment approach to refine classification of
potential tetrachloroethylene exposure in dry-cleaning settings are mixed, with some studies
reporting little or no evidence of an association (Lynge et al., 2006: Pesch et al., 2000b: Boice et
al., 1999: Dosemeci et al., 1999: Aschengrau et al., 1993), and other studies reporting elevated
risks (Calvert et al., 2011: Blair et al., 2003: Anttila et al., 1995: Schlehofer et al., 1995). An
increasing trend in relative risk with increasing exposure surrogate was not observed in any of
the larger occupational exposure studies with three or more exposure categories (Lynge et al.,
2006: Mandel etal., 1995), but some indication of higher risk with higher exposure (or duration)
was observed in other studies (Blair et al., 2003).
4-456
-------
For lung cancer risk, the results from seven large cohort studies of dry cleaners are
consistent with an elevated lung cancer risk of 10-40% (Calvert et al., 2011; S el den and
Ahlborg. 2011: Pukkala et al.. 2009; Ji et al.. 2005b; Blair et al.. 2003; Travier et al.. 2002;
Schlehofer et al., 1995; Lynge and Thygesen, 1990). Similar results were observed in four of the
five occupational studies that were identified as having a relatively strong exposure-assessment
methodology (Calvert et al.. 2011; Blair et al.. 2003; Boiceetal.. 1999; Anttila et al.. 1994).
However, Selden and Ahlborg (2011) observed similar, but slightly higher, relative risks for
laundry workers compared with dry-cleaning workers in their study. These studies were unable
to control for potential confounding from cigarette smoking; however, and the magnitude of the
association in these studies is consistent with that expected assuming the prevalence of smoking
among dry-cleaners and laundry workers was slightly higher (e.g., 10% higher) than among the
general population. Features of the selection of study participants and study analysis in the
available case-control studies reduce the potential for confounding by smoking, however. Two
case-control studies were limited to either nonsmokers or ex-smokers and both of these studies
indicate an approximate twofold increased risk with a history of work in the dry-cleaning
industry [OR: 1.8, 95% CI: 1.1, 3.0 in Brownson et al. (1993). and OR: 1.83, 95% CI: 0.98, 3.40
among women in Pohlabeln et al. (2000)]. The other case-control studies adjusted for smoking
history, and the results for these (somewhat smaller studies) are similar to the previously cited
estimates. Among the studies that evaluated exposure-response gradients, the evidence for a
trend in risk estimates was mixed (Calvert et al., 2011; Blair etal., 2003; Travier et al., 2002;
Boiceetal.. 1999; Pauluetal.. 1999; Brownson et al.. 1993).
For liver cancer, studies carrying greater weight in the analysis based on the large number
of observed events or exposed cases, or based on a strong exposure-assessment approach show a
mixed pattern of results. The one case-control study with a large number of exposed liver cancer
cases and a relatively high quality exposure-assessment methodology reported an odds ratio
estimate of 0.76 (95% CI: 0.38, 1.72) for liver cancer and dry cleaning (Lynge et al.. 2006). A
recent multiple Nordic country cohort study and two cohort studies of Swedish subjects with
broad exposure-assessment approaches, and whose subjects overlapped with Lynge et al. (2006),
reported SIRs of 1.02 (95% CI: 0.84, 1.24), 1.22 (95% CI: 1.03, 1.45), and 1.23 (95% CI: 1.02,
1.49) for liver and biliary tract cancer and work as a dry-cleaner or laundry worker in Travier et
al. (2002), Ji and Hemminki (2005c), and Pukkala et al. (2009), respectively. Three other studies
with strong exposure-assessment approaches specific to tetrachloroethylene, but whose risk
estimates are based on fewer observed liver cancer cases or deaths, reported risk estimates of
1.21 to 2.05 for the association between liver cancer and tetrachloroethylene (Selden and
Ahlborg, 2011; Boiceetal., 1999; Bond etal., 1990; Blair etal., 1979). However, dry cleaning
workers did not have a higher liver cancer risk estimate than laundry workers or other categories
4-457
-------
of dry cleaning workers (Selden and Ahlborg, 2011; Lynge et al., 2006). Exposure response was
not observed, and the SIR for tetrachloroethylene-exposed subjects with the longest employment
duration in Selden and Ahlborg (2011) was lower than that for subjects with shorter employment
duration. Potential confounding may be an alternative explanation as no study adjusted for
known and suspected risk factors for liver cancer (Selden and Ahlborg, 2011; Pukkala et al.,
2009: Lynge et al.. 2006: Ji and Hemminki, 2005c: Travier et al.. 2002: Boiceetal.. 1999: Bond
et al., 1990). Nine other cohort and case-control studies with fewer observed events and/or a
broad exposure-assessment methodology carried less weight in the analysis and reported a mixed
pattern of results (Calvert et al., 2011: Lindbohm et al.. 2009: Sung et al.. 2007: Blair et al..
2003: Lee etal.. 2003: Lynge etal.. 1995: Vartiainen et al.. 1993: Suarezetal.. 1989: Stemhagen
etal.. 1983). Lee et al. (2003) reported a risk estimate of 2.57 (95% CI: 1.21, 5.46) for the
association between liver cancer and residence in a village with ground water contamination, but
subjects were from a region with a high prevalence of HCV infection, and HCV status may
confound the observed association.
For cervical cancer, the results from the two large cohort studies with a broad exposure
assessment are consistent with an elevated cervical cancer risk of 20-30% (Pukkala et al.. 2009:
Travier et al.. 2002). Results from four smaller cohort and case-control studies with a relatively
high quality exposure-assessment methodology presented a pattern of more variable results, with
relative risks of 0.98 (95% CI: 0.65, 1.47), 1.19 (95% CI: 0.64, 1.93), 2.10 (95% CI: 0.68, 4.90),
and 3.20 (95% CI: 0.39, 11.6) in Lynge et al. (2006). Selden and Ahlborg (2011). Calvert et al.
(2011), and Anttila et al. (1995), respectively. A fourth study with higher quality exposure-
assessment specific to tetrachloroethylene did not observe any cervical cancer deaths among
women, but less than one death was expected (Boice etal., 1999). Calvert et al. (2011) was the
only study to report an exposure response gradient with employment duration. Dry cleaning
workers did not have higher cervical cancer risks compared with laundry workers or other
categories of dry cleaning workers (Selden and Ahlborg, 2011: Lynge et al., 2006). Lack of data
on socioeconomic status—a proxy for exposure to the human papilloma virus, a known risk
factor for cervical cancer—indicates great uncertainty for asserting this association with
tetrachloroethylene exposure. Potential confounding by socioeconomic status is an alternative
explanation with some support provided by Lynge et al. (2006), a case-control study with
controls of similar socioeconomic status as cases and that did not observe an association between
cervical cancer and dry cleaning.
The results from the large studies of breast cancer risk in women in relation to
tetrachloroethylene exposure are mixed. The largest, based on 1,757 breast cancer cases in
female dry-cleaners and laundry workers, reported a statistically significant deficit in the risk of
breast cancer incidence compared to the populations of Nordic countries (Pukkala et al., 2009).
4-458
-------
Findings in the other four studies were based on fewer events or exposed cases; two of four
studies with a nonspecific exposure-assessment methodology provided evidence for association
between breast cancer in females and tetrachloroethylene exposure (Sung et al., 2007; Chang,
2005; Aschengrau et al., 2003; Anderson et al., 1999; Lynge and Thygesen, 1990), but no
association was observed in two other large cohort studies with a relatively high quality
exposure-assessment methodology to tetrachloroethylene (Selden and Ahlborg, 2011; Blair et al.,
2003). Small studies also observed mixed findings (Calvert et al., 2011; Radican et al., 2008;
Peplonska et al., 2007: Sung et al., 2007: Aschengrau et al., 2003: Band et al., 2000: Boice et al.,
1999). Although cohort studies were unable to control for potential confounding from
reproductive history or menopausal status, observations in case-control studies controlled for
these potential confounders in statistical analyses and provided support of an association between
female breast cancer and tetrachloroethylene compared to controls (Peplonska et al., 2007:
Aschengrau et al., 2003: Band et al., 2000). Three studies examined exposure response, and two
of these studies with semi quantitative or quantitative exposure-assessment approaches reported
risk estimates in females monotonically increased in higher exposure groups (Aschengrau et al.,
2003: Blair et al., 2003). A third study examining exposure duration observed an inverse relation
(Peplonska et al., 2007). Exposure duration is more uncertain than use of a semiquantitative
surrogate given increased potential for bias associated with exposure misclassification. Because
of the limitation in statistical power, none of the five studies reporting on male breast cancer is
adequate to examine tetrachloroethylene exposure (Selden and Ahlborg, 2011: Pukkala et al.,
2009: Chang et al., 2005: Anderson et al., 1990: Lynge and Thygesen, 1990).
4.10.5. Synthesis of Rodent Cancer Bioassay Findings
One oral gavage (NCI, 1977) and two inhalation (JISA, 1993: NTP, 1986) cancer
bioassays provide evidence of tetrachloroethylene carcinogenicity in rats and mice. In male and
female rats, inhalation exposure to tetrachloroethylene significantly increased the incidence of
mononuclear cell leukemia (MCL) in independent bioassays of the F344/N (JISA, 1993: NTP,
1986) or F344/DuCrj (JISA, 1993) strain. Tetrachloroethylene reduced MCL latency in females
in both studies. In addition, the NTP bioassay reported dose-related increases in the severity of
MCL in males and females. Additional tumor findings in rats included significant increases in
the NTP bioassay of two rare tumor types, kidney tumors in males, and brain gliomas in males
and females. Additionally, the NTP (1986) bioassay reported increases in the rate of testicular
interstitial cell tumors, a tumor type of high incidence in unexposed male F344 rats. Other
evidence, including that brain gliomas occurred earlier with tetrachloroethylene exposure than in
control animals, and that the related compound trichloroethylene is a kidney carcinogen in rats
and humans and a testicular carcinogen in rats, support the significance of these findings. A
4-459
-------
third rat bioassay, of oral gavage exposure in Osborne-Mendel rats, was inconclusive with
respect to carcinogenicity due to a high incidence of respiratory disease in all animals and
shortened survival in tetrachloroethylene-exposed animals (NCI, 1977).
In male and female mice, tetrachloroethylene exposure via inhalation (JISA, 1993; NTP,
1986) or oral gavage (NCI, 1977) significantly increased the incidence of hepatocellular
adenomas and carcinomas. The NCI and NTP studies employed the B6C3Fi strain, while the
JISA study examined the Crj :BDF1 strain. The JISA study reported increases in hemangiomas
or hemangiosarcomas of the liver, spleen, fat, and subcutaneous skin in exposed male Crj:BDFl
mice.
In summary, tetrachloroethylene increased the incidence of liver tumors (hepatocellular
adenomas and carcinomas) in male and female mice and of MCL in both sexes of rats. These
findings were reproducible in multiple lifetime bioassays employing different rodent strains and,
in the case of mouse liver tumors, by inhalation and oral exposure routes. Additional tumor
findings in rats included significant increases in the NTP bioassay of testicular interstitial cell
tumors and kidney tumors in males, and brain gliomas in males and females. In mice,
hemangiosarcomas in liver, spleen, fat, and subcutaneous skin were reported in males in the
JISA study. The rat and mouse findings are summarized in Tables 4-51 and 4-52, respectively,
and in the sections below.
4.10.5.1. Carcinogenicity Findings in Rats
The NCI oral gavage study in Osborne-Mendel rats was considered to be inconclusive
because of the high incidence of respiratory disease, and high mortality with tetrachloroethylene
exposure. Lesions indicative of pneumonia were observed in almost all rats at necropsy. A high
incidence of toxic nephropathy was evident in tetrachloroethylene-exposed male and female rats.
Early mortality was also observed in tetrachloroethylene-exposed animals; 50% of the high dose
males and females had died by Weeks 44 and 66, respectively. Therefore, this bioassay is not
considered further in the below evaluation of the carcinogenicity of tetrachloroethylene in rats.
The NTP (1986) and JISA (1993) inhalation bioassays reported increases in the incidence
of mononuclear cell leukemia (MCL) in male and female F344/N or F344/DuCrj rats.
Supplemental analyses by NTP indicated that tetrachloroethylene produced a dose-related
increase in the severity of MCL in both males and females. Additionally, NTP found that
tetrachloroethylene exposure significantly shortened the time to onset of MCL in females.
Although survival was unaffected, the incidence of advanced MCL increased in female rats that
died before the scheduled study termination. MCL incidences were higher in the concurrent than
in the historical chamber control groups at the performing laboratory (males: 28/50 [56%] vs.
117/250 [47%]; females: 18/50 [36%] vs. 73/249 [29%]). The concurrent control rates were also
4-460
-------
higher than the NTP program historical rate for untreated control groups (males: 583/1,977
[29%]; females: 375/2,021 [18%]).
The Japanese bioassay (JISA, 1993) also reported a significant dose-dependent increase
in MCL in male and female F344/DuCrj rats exposed for 104 weeks to 50, 200, and 600 ppm
tetrachloroethylene. MCL latency was decreased in female rats, with the first appearance in
Week 100 in controls and Weeks 66-70 in exposed rats. As in the NTP study, there was a
higher control incidence of MCL (22% in males and 20% in females) than the reported historical
rate of MCL for the Japanese laboratory of 147/1,149 [13%] in males and 147/1,048 [14.0%] in
females.
Additional tumor findings in rats included a significant increase in the NTP bioassay of
two rare tumor types, kidney tumors in males and brain gliomas in both sexes of exposed F344/N
rats. Kidney tumors rarely occur in unexposed F344/N male rats, with historical incidences
reported to be 0.2% in 1968 controls. The reported incidences with 0, 200, or 400 ppm
tetrachloroethylene exposure were 1/49, 3/47, and 4/50, respectively. Additional support for the
significance of the kidney tumors comes from evidence that the related chemical
trichloroethylene induces this tumor type in humans and in male rats (U.S. EPA, 20 lib). For
brain gliomas, the laboratory and overall program historical control incidences were 2/247
(0.8%) and 4/1971 (0.2%), respectively. Reported incidence with 0, 200, or 400 ppm
tetrachloroethylene exposure was 2/50, 0/48, and 4/50 in males and 1/50, 0/50, and 2/50 in
females, respectively. The significance of the brain tumor findings is supported by the earlier
occurrence with tetrachloroethylene exposure, suggesting an effect on latency. In males,
tetrachloroethylene-induced brain tumors were observed beginning at Week 88 compared with
Week 99 in controls. Female brain tumors were first observed at 75 weeks in
tetrachloroethylene-exposed animals compared with 104 weeks in control group females.
4-461
-------
Table 4-51. Tumor incidence in rats exposed to tetrachloroethylene
Bioassay
NCI (1977)d
Osborne-Mendel
rats
Gavage:
5d/wk,
78 wk
NTP(1986)
F344/N rats
Inhalation:
6h/d,
5d/wk,
104 wk
JISA (1993)
F344/DuCrj rats
Inhalation:
6h/d,
5d/wk,
104 wk
Doses/exposures
Admin.
Vehicle
500 mg/kg-day
1,000 mg/kg-day
Vehicle
500 mg/kg-dayf
1,000 mg/kg-day
0
200 ppm
400 ppm
0
200 ppm
400 ppm
0
50 ppm
200 ppm
600 ppm
0
50 ppm
200 ppm
600 ppm
Continuous
equivalent
Oe
471 mg/kg-day
941 mg/kg-day
Of
474 mg/kg-day
974 mg/kg-day
0
36 ppm
72 ppm
0
36 ppm
72 ppm
0
9 ppm
36 ppm
108 ppm
0
9 ppm
36 ppm
108 ppm
Sex
Male
Female
Male
Female
Male
Female
Reported cumulative tumor incidencea(%)
Hepatocellular
adenomas or
carcinomas
None reported3
None reported
0/50 (0)
1/50 (2)
1/50 (2)
0/50
0/50
0/50
0/50
0/50
0/50
1/50 (2)
0/50 (0)
1/50 (2)
0/50 (0)
Hemangioma
or hemangio-
sarcomasb
1/20
1/49
0/50
None reported
0/50
0/50
0/50
0/50
0/50
0/50
0/50
0/50
0/50
0/50
1/50
0/50
0/50
0/50
Renal
adenomas or
carcinomas
2f/20 (5)
lf/49 (2)
0/50 (0)
0/20 (0)
0/50 (0)
0/50 (2)
1/49 (2)
3/49 (6)
4/50 (8)
0/47
0/44
0/46
1/50 (2)
2/50 (4)
1/50 (2)
2/50 (4)
0/50 (2)
0/50 (0)
0/50 (0)
1/50 (2)
Mononuclear
cell leukemia0
None reported
None reported
28/50 (56)
37/50 (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)
Testicular
interstitial
cell tumors
None
reported
N/A
36/50 (76)
39/49 (80)
41/50 (82)
N/A
47/50 (94)
46/50 (92)
45/50 (90)
48/50 (96)
N/A
Brain
gliomas
None
reported
None
reported
1/50 (2)
0/50 (0)
4/50 (8)
1/50 (2)
0/50 (0)
2/50 (4)
2/50 (4)
0/50 (0)
0/50 (0)
0/50 (0)
0/50
0/50
1/50
0/50
to
aNone reported: Individual animal data were not available, and summary data did not include a line item for this tumor type.
bThese tumors were reported as hemangioendotheliomas in the JISA (1993) report. The term has been updated to hemangioma (benign) or hemangiosarcoma
(malignant). Note that these incidences do not match those tabulated in Table 12 of the JISA report summary. The incidences reported here represent a
tabulation of hemangioendotheliomas from the individual animal data provided in the JISA report.
'Reflects the number of animals with MCL reported under "multiple organs," spleen, or liver.
dThis study was inconclusive with respect to carcinogenicity due to a high incidence of respiratory disease in all animals and shortened survival in PCE-exposed
animals.
eGavage doses listed were adjusted several times during the course of the study. Male rats received the listed TWA daily doses through Week 78, and surviving
animals were observed up to study termination in Week 110.
f "Mixed tumor, malignant" (NCI. 1977).
-------
Table 4-52. Tumor incidence in mice exposed to tetrachloroethylene
Bioassay
NCI (197Z)b
B6C3FJ mice
Gavage:
5d/wk,
78 wk
NTP(1986)
B6C3FJ mice
Inhalation:
6h/d,
5d/wk,
104 wk
JISA (1993)
Crj:BDFl mice
Inhalation:
6h/d,
5d/wk,
104 wk
Doses/exposures
Administered
exposure
Vehicle
450 mg/kg-day
900 mg/kg-day
Vehicle
300 mg/kg-dayd
600 mg/kg-day
Oppm
100 ppm
200 ppm
Oppm
100 ppm
200 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oppm
10 ppm
50 ppm
250 ppm
Continuous
equivalent
exposures
0
536 mg/kg-day
1,072 mg/kg-
day
0
386 mg/kg-day
772 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
Sex
Male
Female
Male
Female
Male
Female
Reported cumulative tumor incidence (%)
Hepatocellular
adenomas or
carcinomas
2/20 (10)
32/48 (67)
27/45 (60)
0/20 (0)
19/48 (40)
19/48(40)
17/49 (35)
31/49(70)
41/50 (82)
4/48 (8)
17/50(38)
38/50 (76)
13/50 (28)
21/50 (43)
19/50 (40)
40/50 (82)
3/50 (6)
3/47 (6)
7/49 (15)
33/49 (67)
Hemangioma
or hemangio-
sarcoma"
None reported0
None reported
1/49 (2)
0/49 (0)
0/50 (0)
0/48 (0)
3/50 (6)
0/50 (0)
4/50 (4)
2/50 (2)
7/50 (13)
9/50 (18)
1/50
0/47
2/49
3/49
Renal
adenomas or
carcinomas
0/20 (0)
1/49 (2)
0/48 (0)
None
reported
0/49 (0)
1/49 (2)
0/50 (0)
None
0/50
1/50
1/50
0/50
0/50
0/47
0/49
0/49
Malignant
lymphoma
None
None
None
None
9/50
7/50
7/50
9/50
14/50
10/47
16/49
10/49
Testicular
interstitial cell
tumors
None reported
N/A
1/49 (2)
0/48 (0)
0/49 (0)
N/A
3750
0/50
0/50
1750
N/A
Brain
gliomas
None
reported
None
reported
None
1/48 (2)
0/49 (0)
0/50 (0)
0/50
0/50
0/50
0/50
0/50
0/47
0/49
0/49
""Administered gavage doses listed were increased after 11 wk by 100 mg/kg-day in each low-dose group or by 200 mg/kg-day in each high-dose group. Animals
received the listed TWA daily doses through Week 78, and surviving animals were observed up to study termination in Week 90.
bThese tumors were reported as hemangioendotheliomas in the JISA (1993) report. The term has been updated to hemangioma (benign) or hemangiosarcoma
(malignant). Note that these incidences do not match those tabulated in Table 12 of the JISA report summary. The incidences reported here represent a
tabulation of hemangioendotheliomas from the individual animal data provided in the JISA report.
°None reported: Individual animal data were not available, and summary data did not include a line item for this tumor type.
dHistiocytic sarcomas, epididymides, or seminal vesicles.
-------
The NTP (1986) study also reported an increase in the rate of testicular interstitial cell
tumors, a tumor type of high incidence in unexposed F344 rats. The reported incidences of
testicular interstitial cell tumors in male rates exposed to 0, 200, or 400 ppm tetrachloroethylene
were 36/50, 39/49, and 41/50, respectively. A higher incidence (47/50, or 92%) was observed in
control rats in the JISA (1993) study than in the NTP (1986) study. In the JISA study, exposure
to 0, 50, 200 or 600 ppm tetrachloroethylene resulted in incidences of 47/50, 46/50, 45/50, and
48/50, respectively. Support for the significance of the testicular interstitial cell tumors comes
from evidence that the related chemical trichloroethylene induces this tumor type in rats.
Trichloroethylene did not induce increases in testicular interstitial cell tumors in the F344 rat in a
bioassay with a reported incidence of 47/48 (98%) in the vehicle control. However, increases
were observed in male Marshall rats, in which the incidences were 16/46, 17/46, 21/33, and
32/39 in the untreated control, vehicle control and 500, or 1,000 mg/kg-day trichloroethylene
exposure groups, respectively.
In conclusion, evidence for the carcinogenicity of tetrachloroethylene in rats was
provided by increases in MCL incidence in both sexes in two inhalation bioassays. Rare kidney
tumors in males and rare brain gliomas in males and females were increased in a single bioassay
(NTP, 1986). Additionally, the NTP (1986) bioassay reported increases in the rate of testicular
interstitial cell tumors, a tumor type of high incidence in unexposed male F344 rats. The
available oral gavage cancer bioassay was inconclusive due to respiratory infection in all groups
and high mortality in tetrachloroethylene-exposed animals.
4.10.5.2. Carcinogenicity Findings in Mice
In both sexes of mice, tetrachloroethylene increased the incidence of liver tumors in
multiple bioassays. In male and female B6C3Fi mice exposed for 2 years by oral gavage,
significant increases were noted in hepatocellular carcinomas and adenomas (NCI, 1977). The
reported incidence with 0, 500, and 1,000 mg/kg-day tetrachloroethylene were 2/20, 32/48, and
27/45 in males and 0/20, 19/48, and 19/49 in females, respectively. Tumor latency was
significantly decreased with tetrachloroethylene exposure. A significant association between
increased mortality and dose of tetrachloroethylene was seen, with liver tumors found in many of
the mice that died early. In lifetime inhalation studies of B6C3Fi (NTP. 1986) and Crj :BDF1
mice, tetrachloroethylene similarly increased liver tumors. Statistically significant, dose-related
increases in the incidence of hepatocellular carcinoma and in combined hepatocellular adenoma
and carcinoma were observed in both sexes. The reported incidence of liver carcinomas and
adenomas with 0, 100, and 200 ppm tetrachloroethylene in the NTP inhalation bioassay were
17/49, 31/49, and 41/50 in males and 4/45, 17/42, and 38/48 in females, respectively. In male
mice, hepatocellular carcinomas metastasized to the lungs in 2/49, 7/49, and 1/50 animals.
4-464
-------
Metastatic hepatocellular carcinomas were found in the lungs of 0/48, 2/50, and 7/50 female
mice. In the JISA study, the reported incidence of liver carcinomas and adenomas with 0, 10, 50,
and 250 ppm tetrachloroethylene were 13/50, 21/50, 19/50, and 40/50 in males and 3/50, 3/47,
7/49, and 33/49 in females, respectively.
Additional evidence of carcinogenicity from the lifetime bioassays in mice included a
significant increase in the incidence of hemangiosarcomas (reported as malignant
hemangioendotheliomas) or hemangiomas (reported as benign hemangioendotheliomas) of the
liver, spleen, fat, and subcutaneous skin in JISA study males. This tumor type was not reported
in the NCI oral gavage bioassay, and no increase was reported in the NTP inhalation bioassay.
Other findings in the JISA study were Harderian gland adenomas and enlargement of the nucleus
in the kidney proximal tubular cells in male mice at the highest exposure.
Other supporting evidence for carcinogenicity is the known hepatocarcinogenicity of
tetrachloroethylene metabolites. The major urinary metabolite of tetrachloroethylene in humans
and rodents, TCA, is hepatocarcinogenic in mice. TCA significantly increased the incidence of
liver tumors in male and female B6C3Fi mice exposed via drinking water for 52-104 weeks
(DeAngelo et al.. 2008: Bull et al.. 2002: Pereira, 1996: Bulletal.. 1990: Herren-Freund et al..
1987). Incidence of tumors increased with increasing TCA concentrations (DeAngelo et al.,
2008: Bull et al.. 2002: Pereira, 1996: Bulletal.. 1990). The development of tumors in animals
exposed to TCA progressed rapidly, as evidenced by significant numbers of tumors in less-than-
lifetime studies of 82 weeks or less. The tetrachloroethylene metabolite DCA also causes liver
cancer in mice (DeAngelo et al., 1999: Daniel et al., 1992: Bull et al., 1990: Herren-Freund et al.,
1987). Additionally, DCA and TCA are hepatocarcinogenic in mice when coadministered in the
drinking water for 52 weeks (Bull et al., 2002). Treatment-related liver tumors were observed in
male F344/N rats exposed via drinking water to DCA (DeAngelo et al., 1996) but not TCA
(DeAngelo et al., 1997) for 60 or 104 weeks. However, the extent to which DCA is available to
the liver following tetrachloroethylene exposure is unclear, because it is thought to be formed in
the kidney following p-lyase processing of TCVC and may be largely excreted in urine without
circulating systemically. The carcinogenicity of TCA and DCA has not been evaluated in female
rats or in other species of experimental animals.
In conclusion, evidence for the carcinogenicity of tetrachloroethylene in mice is provided
by increases in hepatocellular carcinomas and adenomas in both sexes of mice in a gavage
bioassay (B6C3Fi mice) and in two inhalation bioassays (one of the B6C3Fi strain and the other
of the Crj:BDFl strain). In male Crj:BDFl mice, hemangiosarcomas or hemangiomas of the
liver, spleen, fat, and subcutaneous skin were increased (JISA, 1993). Supporting evidence
includes the hepatocarcinogenicity of tetrachloroethylene metabolites TCA and DCA, alone and
in combination.
4-465
-------
4.10.5.3. Carcinogenic Mode of Action Hypotheses
This section summarizes the supporting evidence for the modes of action posited for the
rat and mouse tumors presented in Table 4-51 and 4-52. The discussion focuses on
tetrachloroethylene-specific studies, for which the database is especially limited. Evidence from
studies of metabolites of tetrachloroethylene is also summarized. A tabular summary of the
hypothesized MOA and key events, and the supporting evidence from studies of
tetrachloroethylene and its metabolites, are provided in Table 4-56. Overall, these findings
support the conclusion that the mechanisms by which tetrachloroethylene induces rodent
carcinogenesis are not yet fully characterized, completely tested, or understood.
4.10.5.3.1. Hypothesized modes of action for rat tumors
4.10.5.3.1.1. Testicular interstitial cell tumors
No data are available concerning either the metabolites or the mechanisms that may
contribute to the induction of testicular interstitial cell tumors occurring in exposed rats.
Evidence for the related compound trichloroethylene, while suggestive of a MOA involving
hormonal disruption, is inadequate to specify and test a hypothesized sequence of key events. It
is concluded that the specific active moiety(ies), mechanisms, or modes of action by which
tetrachloroethylene induces this type of tumor is not known.
4.10.5.3.1.2. Brain gliomas
No data are available concerning either the metabolites or the mechanisms that may
contribute to the induction of rare brain gliomas occurring in exposed rats. It is concluded that
the specific active moiety(ies), mechanisms, or modes of action by which tetrachloroethylene
induces this type of tumor are not known.
4.10.5.3.1.3. Mononuclear cell leukemia
Regarding the metabolites that potentially contribute to MCL development, a role for
GSH-derived intermediates was posited based on findings for the related compound
trichloroethylene. However, TCVC, a GSH-derived metabolite of tetrachloroethylene, induced
no kidney or bone marrow effects when administered to two calves as a single dose (Lock et al.,
1996). Aside from this evaluation of bone marrow toxicity of TCVC in the juvenile cow, a
species of unknown sensitivity to tetrachloroethylene-induced leukemia, other studies aimed at
elucidating the active metabolites contributing to leukemic effects have not been reported. In
particular, no such studies are available in the F344 rat, the species and strain in which leukemic
effects have been consistently observed in both sexes. Additionally, no data are available
concerning the contributing mechanisms. It is, thus, concluded that the specific active
4-466
-------
moiety(ies), mechanisms, or modes of action by which tetrachloroethylene induces this type of
tumor are not known.
4.10.5.3.1.4. Renal tumors
It is likely that several mechanisms contribute to tetrachloroethylene-induced kidney
cancer. Mutagenicity, peroxisome proliferation, a2u-globulin nephropathy, and cytotoxicity not
associated with a2u-globulin accumulation are MO As that have been investigated. Except for
a2u-globulin accumulation, which is more likely due to tetrachloroethylene itself (Lash and
Parker, 2001), other mechanisms hypothesized to contribute to tetrachloroethylene-induced renal
carcinogenicity are thought to be mediated by tetrachloroethylene metabolites rather than with
the parent compound. Metabolites from the GSH conjugation pathway are posited to induce
renal tumorigenicity, as opposed to, or to a greater extent, than the metabolites resulting from
oxidative CYP processing. The glutathione conjugation of tetrachloroethylene in the kidney,
discussed in Section 3, leads sequentially to TCVG and TCVC. TCVC can be further processed
by p-lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a highly
reactive thioketene, a chemical species that can form covalent adducts with cellular nucleophiles
including DNA. TCVC can also undergo FMO3 or P450 oxidation to reactive intermediates;
additionally, sulfoxidation of both TCVC and its 7V-acetylated product occurs, resulting in
reactive metabolites (Rippetal., 1999: Rippetal.. 1997: Werner et al.. 1996). TCVG, TCVC,
and NAcTCVC are mutagenic in Salmonella tests, as is tetrachloroethylene in the few studies of
conditions that could generate GSH-derived metabolites (Dreessen et al., 2003: Vamvakas et al.,
1989c: Vamvakas et al., 1989d: Vamvakas et al., 1987: Dekant et al., 1986a). Evidence of in
vivo genotoxicity in the kidney is limited to reports of modest effects following i.p. exposures,
including low level binding to rat kidney DNA (Mazzullo et al., 1987) and DNA single-strand
breaks in mouse kidney (Walles, 1986). Given the known mutagenicity of the GSH-derived
tetrachloroethylene metabolites that are formed in the kidney, and the observed in vitro
mutagenicity of tetrachloroethylene under conditions that would generate these metabolites, a
mutagenic MO A contributing to the development of the kidney tumors cannot be ruled out.
It has been suggested that the low-level renal tumor production observed in exposed rats
is secondary to sustained cytotoxicity and necrosis leading to activation of repair processes and
cellular regeneration. However, nephrotoxicity occurs in both sexes of rats and mice, whereas
cell replication and tumorigenesis occurs only in male rats. In addition, tetrachloroethylene
induces kidney tumors at lower doses than those required to cause a2u-globulin accumulation,
raising serious doubt that a2u-globulin plays a key role—especially any major role—in rat
kidney tumor formation. Rodent studies of tetrachloroethylene addressing renal a2u-globulin
accumulation are summarized in Table 4-53.
4-467
-------
Because tetrachloroethylene has been shown to induce peroxisome proliferation, an
indicator of PPARa-activation, the possibility exists that certain responses resulting from
activation of this receptor might be involved in cancer-causing activity leading to
tetrachloroethylene-induced renal tumors. However, as summarized in Table 4-54, chemical-
specific studies are limited and show only modest effects at exposures exceeding those required
for renal carcinogenesis. There is no evidence causally linking PPARa-activation to kidney
tumorigenesis for tetrachloroethylene or other compounds.
In summary, the complete mechanisms of tetrachloroethylene-induced renal
carcinogenesis are not yet understood. Given the known mutagenicity of the GSH-derived
tetrachloroethylene metabolites that are formed in the kidney, and the observed in vitro
mutagenicity of tetrachloroethylene under conditions that would generate these metabolites, a
mutagenic MO A contributing to the development of the kidney tumors cannot be ruled out.
4.10.5.3.2. Hypothesized modes of action for mouse tumors
4.10.5.3.2.1. Hemangiosarcomas
No data are available concerning either the metabolites or the mechanisms that may
contribute to the induction of hemangiosarcomas or hemangiomas observed in the liver, spleen,
fat, and subcutaneous skin in male mice. It is concluded that the mechanisms or modes of action
by which tetrachloroethylene induces this type of tumor are not known.
4.10.5.3.2.2. Hepatocellular tumors
As noted by NRC (2010), it is likely that key events from several pathways, comprising
several simultaneous mechanisms, operate in tetrachloroethylene-induced liver cancer. MOA
hypotheses for mouse liver tumors concern genotoxicity, epigenetic effects (especially DNA
hypomethylation), oxidative stress, and receptor activation (i.e., a hypothesized PPARa-
activation MOA). Because it has been suggested that hepatocarcinogenesis caused through a
PPARa-activation MOA is not relevant to humans [e.g., Klaunig et al. (2003)], and such a
conclusion would have significant implications for hazard conclusions and dose-response
analyses, this hypothesized MOA is discussed in relatively more detail than other topics.
4-468
-------
Table 4-53. Renal a2u-globulin accumulation in tetrachloroethylene-exposed
rodents
Species/strain/
sex/number
Mouse, B6C3FJ, both
sexes (49 or 50 mice per
sex per dose group)
Rat, F344, both sexes (50
mice per sex per dose
group)
Rat, F344 (both sexes, 5
per group)
Rat, F344 (both sexes, 12
per group)
Rat, F344 (both sexes)
and B6C3F! mice (both
sexes); 10 per group for
oral studies, 5 per group
for inhalation studies
Exposure level/duration
0, 100, 200 ppmfor 104 wk,
inhalation
0, 200, 400 ppm for 104 wk,
inhalation
0 or 1,000 mg/kg-day for
10 d, corn oil gavage
0, 500 mg/kg-day daily for 4
wk, corn oil gavage
0, 1,000 or 1,500 mg/kg-day
daily by corn oil gavage for
42 d; 0 or 1,000 ppm for 10 d
Effects
Karyomegaly and cytomegaly of the
proximal tubules in all exposed mice;
nephrosis in exposed females, casts
increased in all exposed males and in
high-dose females.
Karyomegaly and cytomegaly of the
proximal tubules in all exposed rats.
Increases in o2u-hyaline droplets in
exposed males but not females. Correlated
to increased cell proliferation and protein
droplet nephropathy.
Increases in o2u-globulin accumulation in
proximal tubule cells.
Accumulation of o2u-globulin in proximal
tubules of male rats; nephrotoxicity in
male rats (formation of granular tubular
casts and evidence of tubular cell
regeneration).
Inhalation exposure demonstrated
formation of hyaline droplets in kidneys of
male rats.
Reference
NTP (1986)
NTP (1986)
Goldsworthy
et al. (1988)
Bergamaschi
et al. (1992)
Green et al.
(1990)
Table 4-54. Renal peroxisome proliferation in tetrachloroethylene-exposed
rodents
Species/strain/sex/number
Rat, F344; and mouse,
B6C3Fi; both sexes
(5/group)
Odum et al. (1988)
Rat, F344 (male only,
5/group) and mouse,
B6C3FJ (male only,
5/group)
Goldsworthy and Popp
(1987)
Effect
Mice of both sexes: Analysis in mice was
limited to pooled tissue, but showed slight
increases in (3-oxidation in mouse kidney
Rats: Modest increases in PCO in male rat
kidneys at 200 ppm for 28 d only, but elevated
in female rat kidney at all doses and times.
Mice: Increased PCO activity
Rats: Increased kidney weight
Dose
200, and 400 ppm,
inhalation
200, and 400 ppm,
inhalation
1,000 mg/kg-day
for 10 d, corn oil
gavage
1,000 mg/kg-day
for 10 d, corn oil
gavage
Time
14, 21, 28 d
14, 21, 28 d
10 d
10 d
4-469
-------
The limited tetrachloroethylene-specific data for PPARa-activation support the view that
this is not the primary MOA for hepatocarcinogenesis (refer to Table 4-55). Philip et al. (2007)
reported significantly increased expression of CYP4A, a marker of PPARa-activation, in SW
mice at only the highest dose (1,000 mg/kg-day) and at the earliest time point (7 days), in
contrast to the robust dose-dependent proliferative response of a more prolonged nature (lasting
for 14-30 days post exposure) observed at the same and lower (150, 500, and 1,000 mg/kg-day)
levels of tetrachloroethylene. The authors suggested that these data are not supportive of a close
mechanistic relationship of carcinogenicity and PPARa-activation for tetrachloroethylene-
derived TCA. Limitations of this interpretation include the possible lack of sensitivity of
CYP4A protein expression as a marker of peroxisome proliferation, and the unknown sensitivity
of the SW mouse to tetrachloroethylene hepatocarcinogenicity. Other investigators [e.g.,
Schumann et al. (1980)1 have reported liver toxicity and repair at 100 mg/kg-day in the B6C3Fi
strain, whereas repeated exposures to 1,000 mg/kg-day were reported by Philip et al. (2007) and
Odum et al. (1988) to only modestly increased peroxisomal markers in SW and B6C3Fi mice,
respectively. Odum et al. (1988) also observed moderate increases in peroxisome proliferation
in rats, a species insensitive to tetrachloroethylene hepatocarcinogenicity. In all, these findings
indicate that the modest peroxisome proliferation observed in response to tetrachloroethylene
may lack specificity with respect to species, tissue, and dose. Studies of the temporal sequence
of events are limited. Given the limitations in the database of tetrachloroethylene-specific
studies, it can be concluded that the few studies demonstrating peroxisome proliferation by
tetrachloroethylene are insufficient to demonstrate a causative role of this effect in the induction
of other key events posited for the PPARa-activation MOA hypothesis, and for
hepatocarcinogenesis by tetrachloroethylene.
Studies of other PPARa agonists, and of transgenic models of PPARa-activation, more
generally support the view that the hypothesized PPARa-activation MOA may not be a limiting
factor in rodent hepatocarcinogenesis (refer to Section 4.3.5.5). PPARa-activation may play a
significant role in mouse liver tumor induction by some compounds, such as Wy-14,643.
However, recent studies suggest that DEHP can induce tumors in a PPARa independent manner
without any loss of potency (Ito et al., 2007a), and that PPARa-activation in hepatocytes is itself
insufficient to cause tumorigenesis (Yang et al., 2007). Additional analyses, presented in
Section 4.3.5.3.2, demonstrate that peroxisome proliferation and associated markers are poor
quantitative predictors of hepatocarcinogenesis in rats or mice. These findings raise serious
concerns about human health risk assessment MOA conclusions based exclusively on evidence
of PPARa-agonism and other key events in the hypothesized PPARa-activation MOA, given that
other modes, mechanisms, toxicity pathways, and molecular targets may contribute to or be
required for the observed adverse effects. Indeed, for tetrachloroethylene and most other PPARa
4-470
-------
agonists, chemical-specific data to define the range of effects that may contribute to human
carcinogenesis are insufficient. Similarly, the epidemiologic data are inadequate to inform
conclusions of human relevance (Guvton et al., 2009).
Table 4-55. Rodent studies of induction of hepatic peroxisome proliferation
or its markers by tetrachloroethylene
Species/strain/sex/number
Rat, F344; and mouse, E6C3Ąl;
both sexes (5/group)
Odum et al. (1988)
Rat, F344 (male only, 5/group)
and mouse, B6C3FJ (male only,
5/group)
Goldsworthy and Popp (1987)
Mouse, Swiss-Webster, male (4
mice/group)
Philip et al. (2007)
Effect
Mice of both sexes: increased relative
liver weight, centrilobular lipid
accumulation and peroxisome
proliferation; increased PCO
(up to 3. 7-fold)
Male mice: mitochondria! proliferation
Rats of both sexes: increased PCO (up to
1.3 -fold)
Mice: Increased relative liver weight;
4.3 -fold PCO increase
Rats: Increased relative liver weight;
modest but not significant (1.4-fold) PCO
increase
Increased plasma ALT
Mild to moderate fatty degeneration and
necrosis, with focal inflammatory cell
infiltration
Increased mitotic figures and DNA
synthesis
CYP4A increased at 7 but not 14 d, only
at 1,000 mg/kg-day
Dose
200, and 400 ppm,
inhalation
400 ppm, inhalation
200, and 400 ppm,
inhalation
1,000 mg/kg-day
for 10 d, corn oil
gavage
1,000 mg/kg-day
for 10 d, corn oil
gavage
150, 500, and 1,000
mg/kg-day,
aqueous gavage
150, 500, and 1,000
mg/kg-day,
aqueous gavage
150, 500, and 1,000
mg/kg-day,
aqueous gavage
1,000 mg/kg-day,
aqueous gavage
Time
14, 21,28 d
28 d
14, 21,28 d
10 d
10 d
24 hours to
14 d after
initial
exposure
24 hours to
30 d after
initial
exposure
Peaked on 7
d, sustained
at 14-30 d
7 but not 14 d
A recent review (Rusyn et al., 2006) addressed other mechanistic effects of the PPARa
agonist DEHP and proposed that tumors arise from a combination of molecular signals and
pathways, rather than from a single event such as PPARa-activation. As reviewed in
Section 4.3.5.1, the metabolites of tetrachloroethylene have been shown to induce a number of
effects that may contribute to carcinogenicity, including mutagenicity, alterations in DNA
methylation, and oxidative stress. Given the demonstrated mutagenicity of several
4-471
-------
tetrachloroethylene metabolites, the hypothesis that mutagenicity contributes to the MOA for
tetrachloroethylene hepatocarcinogenesis cannot be ruled out, although the specific metabolic
species or mechanistic effects are not known. Epigenetic effects and oxidative stress, including
those produced secondary to cytotoxicity, may also contribute. Currently, the available database
of tetrachloroethylene-specific studies addressing these mechanisms is very limited.
4.10.5.3.3. Mode-of-action summary
Table 4-56 reviews the hypothesized modes of action for tetrachloroethylene-induced
cancer in rodents, which are not intended to be interpreted as being mutually exclusive. The
evidence summarized in this table supports the view that there are significant gaps in the
scientific knowledge of mechanisms contributing to tetrachloroethylene-induced cancer.
Multiple metabolites formed from tetrachloroethylene are toxic and carcinogenic in rodents.
Given this knowledge, and the known complexity and heterogeneity in cancer development, in
general, the available evidence supports a hypothesis of multiple, contributing mechanistic
effects that may, in turn, be affected by multiple modifying factors.
4-472
-------
Table 4-56. Summary of hypothesized modes of action for tetrachloroethylene-induced cancer in rodents
Tumor type, sex, strain,
species
Testicular interstitial cell
tumors in male F344/N rats
Brain gliomas in male and
female F344/N rats
Mononuclear cell leukemia in
male and female F344/N and
F344/DuCrj rats
Hypothesized MOA and key events
None hypothesized
None hypothesized
None hypothesized
Evidence that PCE or PCE metabolites
induces key events
N/A
N/A
N/A
Necessity of key
events for
carcinogenesis
N/A
N/A
N/A
Sufficiency of
MOA for
carcinogenesis
N/A
N/A
N/A
-------
Table 4-56. Summary of hypothesized modes of action for tetrachloroethylene-induced cancer in rodents
(continued)
Tumor type, sex, strain,
species
Hypothesized MOA and key events
Evidence that PCE or PCE metabolites
induces key events
Necessity of key
events for
carcinogenesis
Sufficiency of
MOA for
carcinogenesis
Kidney adenocarcinoma in
male F344/N rats
Mutagenicity induced by GSH-
derived metabolites advances
acquisition of the multiple critical
traits contributing to carcinogenesis
PCE lacks mutagenicity in Salmonella
(Ames), other genotoxicity tests [Emmert et
al. (2006): Watanabe et al. (1998) (refer to
Table 4-40); DeMarini et al. (1994): Roldan-
Arjona et al. (1991): Milman et al. (1988):
Warner et al. (1988): NTP (1986): Connor et
al. (1985): Shimada et al. (1985): Haworth et
al. (1983): Hardin et al. (1981): Kringstad et
al. (1981): Bartsch et al. (1979): Greim et al.
(1975)1
Limited studies of PCE genotoxicity in
rodent kidney: low levels of DNA binding in
Wistar rats with 1.4 mg/kg i.p. (Mazzullo et
al.. 1987): DNA single-strand breaks in
NMRI mice with 660 mg/kg i.p. (Walles.
1986)
Mutagenicity of TCVG, TCVC, NAcTCVC
(and PCE under conditions that could
generate these GSH-derived metabolites) in
Ames assays (Dreessen et al.. 2003:
Vamvakas et al.. 1989c: Vamvakas et al..
1989d: Vamvakas et al.. 1987: Dekant et al..
1986a)
No PCE-specific
studies3
No PCE-specific
studies;
Mutagenicity is
assumed to cause
cancer, as a
sufficient cause
-------
Table 4-56. Summary of hypothesized modes of action for tetrachloroethylene-induced cancer in rodents
(continued)
Tumor type, sex, strain,
species
Hypothesized MOA and key events
Evidence that PCE or PCE metabolites
induces key events
Necessity of key
events for
carcinogenesis
Sufficiency of
MOA for
carcinogenesis
Kidney adenocarcinoma in
male F344/N rats (continued)
Tubular cell necrosis and
nephrotoxicity followed by
hyperplasia
Nephrotoxicity of PCE reported in multiple
studies in both sexes of rats and mice at
carcinogenic doses [e.g., NTP (1986)1
No PCE-specific
studies3
No PCE-specific
studies
o2u-globulin accumulation:
• Excessive accumulation of hyaline
droplets containing o2u-globulin
in renal proximal tubules
• Subsequent cytotoxicity and
necrosis
• Sustained regenerative tubule cell
proliferation
• Development of intralumenal
granular casts from sloughed
cellular debris associated with
tubule dilatation and papillary
mineralization
• Foci of tubule hyperplasia in the
convoluted proximal tubules
• Renal tubule tumors
In F344 rats, PCE induced hyaline droplets at
500 mg/kg-day for 4 wk (Bergamaschi et al.
1992). or>l,000 mg/kg-day for 10
(Goldsworthy et al.. 1988) or 42 d (Green et
al.. 1990)
No PCE-specific
studies3
No PCE-specific
studies
No evidence of mineralization in PCE
bioassays (JISA. 1993: NTP. 1986) or of
hyaline droplets with <400 ppm for 28 d
(Green etal.. 1990) in F344 rats
PPARa-activation:
• Metabolites (e.g., TCA) activate
PPARa
• Alterations in cell proliferation
and apoptosis
• Clonal expansion of initiated cells
In F344 rat kidney, PCE increased PCO in
males only at 200 ppm for 28 d (PCO
increased in females at 200 and 400 ppm, at
14, 21 and 28 d) (Odumetal.. 1988): in
B6C3F! male mouse kidney, PCE increased
PCO with 1,000 mg/kg-day p.o. for 10 d
(Goldsworthy and Popp. 1987)
No PCE-specific
studies
No data from other
chemicals on PPARa
involvement in
kidney tumors.
No PCE-specific
studies
Hemangiosarcomas in male
Crj:BDF! mice
None hypothesized
N/A
N/A
N/A
-------
Table 4-56. Summary of hypothesized modes of action for tetrachloroethylene-induced cancer in rodents
(continued)
Tumor type, sex, strain,
species
Hypothesized MOA and key events
Evidence that PCE or PCE metabolites
induces key events
Necessity of key
events for
carcinogenesis
Sufficiency of
MOA for
carcinogenesis
Liver hepatocellular carcinoma
in male and female B6C3Fi
and Crj :BDP! mice
Mutagenicity induced by one or more
metabolites advances acquisition of
multiple critical traits contributing to
carcinogenesis
PCE lacks mutagenicity in Salmonella
(Ames), other genotoxicity tests [Emmert et
al. (2006): Watanabe et al. (1998) (refer to
Table 4-40); DeMarini et al. (1994): Roldan-
Arjona et al. (1991): Milman et al. (1988):
Warner et al. (1988): NTP (1986): Connor et
al. (1985): Shimada et al. (1985): Haworth et
al. (1983): Hardin et al. (1981): Kringstad et
al. (1981): Bartsch et al. (1979): Greim et al.
(1975)]
Limited PCE genotoxicity studies in mouse
liver: Positive/equivocal Comet assay in GDI
mice (Cederberg et al.. 2010a). positive
micronucleus assay in ddY mice post (but
not pre) partial hepatectomy (Murakami and
Horikawa. 1995) at 1,000 mg/kg-day; DNA
binding in male Balb/c mice at 1.4 mg/kg i.p.
(Mazzullo et al.. 1987): DNA single-strand
breaks in NMRI mice with 660 mg/kg i.p.
(Walles. 1986)
Certain metabolites of PCE are mutagenic in
vitro and in vivo (refer to Tables 4-41 and
4-42)
No PCE-specific
studies3
No PCE-specific
studies;
Mutagenicity is
assumed to
cause cancer, as
a sufficient
cause
-------
Tumor type, sex, strain,
species
Hypothesized MOA and key events
Evidence that PCE or PCE metabolites
induces key events
Necessity of key
events for
carcinogenesis
Sufficiency of
MOA for
carcinogenesis
Liver hepatocellular carcinoma
in male and female B6C3F!
and Crj :BDP! mice (continued)
Epigenetic changes, particularly
DNA methylation, induced by one or
more metabolites (TCA, DCA, and
other reactive species) advance
acquisition of multiple critical traits
contributing to carcinogenesis
No PCE-specific studies
In mouse liver, TCA and DCA decrease
global DNA methylation and promoter
hypomethylation (e.g., of c-myc) (Ge etal..
2001:Taoetal.. 1998)
No PCE-specific
studies3
Cytotoxicity and secondary oxidative
stress:
• One or more reactive
intermediates induce
hepatotoxicity
• Oxidative stress results (from
hepatocyte injury, from infiltrating
inflammatory cells and/or as part
of the intra- and/or intercellular
repair processes)
• Oxidative stress advances
acquisition of multiple critical
traits contributing to
carcinogenesis
PCE induces hepatotoxicity characterized by
increased liver weight, fatty changes,
necrosis, inflammatory cell infiltration, and
proliferation [e.g., NTP (1986)1
No PCE-specific
studies3
No PCE-specific
studies; dys-
regulation of
methylation
represents a
common early
molecular event
in most tumors
and is
hypothesized to
cause cancer
No PCE-specific
studies
-------
Tumor type, sex, strain,
species
Hypothesized MOA and key events
Evidence that PCE or PCE metabolites
induces key events
Necessity of key
events for
carcinogenesis
Sufficiency of
MOA for
carcinogenesis
Liver hepatocellular carcinoma
in male and female B6C3F!
and Crj :BDP! mice (continued)
PPARa-activation:
• TCA, after being produced in the
liver, activates PPARa
• Alterations in cell proliferation
and apoptosis
• Clonal expansion of initiated cells
oo
In B6C3FJ mouse liver, PCE increased PCO
(three- to fourfold) with 200 and 400 ppm
(OdumetaL 1988) or 1,000 mg/kg-day p.o.
(Goldsworthy and Popp. 1987)
In SW mouse liver, PCE increased CYP4A at
7 but not 14 d, at 1,000 mg/kg-day; increased
mitotic figures and DNA synthesis at 7-30 d
with 150, 500, and 1,000 mg/kg-day (Philip
etal.,2007)
TCA activates PPARa, induces peroxisome
proliferation and hepatocyte proliferation in
mice and rats [e.g., DeAngelo et al. (2008):
Laughter et al. (2004): Stauber and Bull
(1997): Pereira and Phelps (1996): Dees and
Travis (1994): Sanchez and Bull (1990)1
No PCE-specific
studies; liver tumor
response from WY
dramatically
diminished in
PPARa-null mice
(Peters etal. 1997);
liver tumor response
fromDEHP
unchanged in
PPARa-null mice
(Ito et al.. 2007a).
No inference
possible with PCE.
No PCE-specific
studies; PPARa-
activation in a
transgenic
mouse model
caused all the
key events in the
MOA, but not
carcinogenesis,
suggesting that
the MOA is not
sufficient for
carcinogenesis
(Yang etal..
2007).
Consistent with
hypothesis that
PCE liver
carcinogenesis
involves
multiple
mechanisms.
a Associations [e.g., per Hill (1965) considerations] noted for some chemicals between hypothesized sequence of key events and carcinogenesis.
-------
5. DOSE-RESPONSE EVALUATION
5.1. INHALATION REFERENCE CONCENTRATION (RfC)
This section presents quantitative risk estimates for chronic noncancer inhalation
tetrachloroethylene exposure. Although the RfD is commonly presented first in the IRIS
toxicological reviews, the RfC is presented in Section 5.1 and the RfD in Section 5.2 because the
RfD was developed by route-to-route extrapolation of the RfC to the oral route of exposure. The
analysis is based on the noncancer hazard characterization for tetrachloroethylene presented in
Section 4.10.2, which identified neurotoxicity as a sensitive endpoint following either inhalation
or oral exposure to tetrachloroethylene. Neurotoxicity is thus selected as the critical effect for
deriving the noncancer inhalation RfC. All neurotoxicity studies suitable for dose-response
analysis are evaluated in the selection of principal studies.
5.1.1. Choice of Principal Studies and Critical Effect
5.1.1.1. Choice of Critical Effect
The database of human and animal studies on inhalation toxicity of tetrachloroethylene is
adequate to support derivation of inhalation reference values. As summarized in Section 4.10, a
number of targets of toxicity from chronic exposure to tetrachloroethylene have been identified
in published animal and human studies. These targets include the central nervous system (CNS),
kidney, liver, immune and hematologic systems, and development and reproduction. In general,
neurological effects were judged to be associated with lower tetrachloroethylene concentrations
compared with other noncancer endpoints of toxicity.
5.1.1.2. Overview of Candidate Principal Studies
The evidence for neurotoxicity in humans includes controlled experimental chamber
(Altmann et al., 1990; Hake and Stewart, 1977) and epidemiologic (Spinatonda et al., 1997;
Altmann et al.. 1995: Echeverria et al.. 1995: Ferroni etal.. 1992: Seeber. 1989: Hake and
Stewart, 1977) studies that used standardized neurobehavioral batteries or employed assessment
of visual function (Storm etal., 2011 [previously reported in NYSDOH, 20101: Schreiber et al.,
2002: Gobbaetal., 1998: Cavalleri etal., 1994), a neurological outcome known to be sensitive to
volatile organic compounds. Of the 12 candidate studies in humans, seven epidemiological
studies of tetrachloroethylene examined occupational exposure (Schreiber et al., 2002: Gobba et
al., 1998: Spinatonda et al., 1997: Echeverria et al., 1995: Cavalleri et al., 1994: Ferroni et al.,
1992: Seeber, 1989), three epidemiological studies examined residential exposure to
5-1
-------
tetrachloroethylene (Storm et al., 2011 [previously reported in NYSDOH, 2010]: Schreiber et al.,
2002; Altmann et al., 1995), and 2 were acute experimental chamber studies (Altmann et al.,
1990: Hake and Stewart, 1977). Together, the epidemiologic evidence supports an inference of a
broad range of cognitive, motor, behavioral, and visual functional deficits following
tetrachloroethylene exposure (U.S. EPA, 2004).
The research in animal models comprises acute and subchronic studies of the effects of
tetrachloroethylene on functional neurological endpoints (functional observation battery, motor
activity) (Oshiro et al., 2008: Kjellstrand et al., 1985), on sensory system function as assessed by
evoked potential (Boyes et al., 2009: Mattsson et al., 1998) or pathological changes in the brain
(Wang et al., 1993). The studies in animal models support the human studies, with notable
effects on motor activity and motor function following exposure to tetrachloroethylene during
either adulthood or the developmental period. Changes in evoked potentials following acute and
subchronic exposures were also seen. In addition, postmortem effects in animals were observed
with pathological alterations in brain DNA, RNA, or protein levels and brain-weight changes.
The studies considered for derivation of the RfC are summarized in the following
sections and in Table 5-1 and Figure 5-1. Table 5-1 identifies the species, exposure duration,
and ambient (experimental) concentrations. For epidemiologic studies, the reported
concentrations, and the observed effect and its magnitude associated with the NOAEL or the
LOAEL are provided. Additionally, human equivalent concentrations (FtECs) for LOAELs or
NOAELs are presented to better allow examination of effect levels across studies and species.
FtECs are calculated using the RfC methodology for a Category 3 gas, extrathoracic effects, and
adjusted to equivalent continuous exposure (U.S. EPA, 1994).28 The studies in Table 5-1 are
listed in order of increasing HEC and displayed graphically in Figure 5-1.
5.1.1.3. Selection of Principal Studies
The candidate principal studies of CNS effects listed in Table 5-1 were evaluated
according to study characteristics identified in Table 5-2. Human studies were preferred to
animal studies, as were studies of chronic duration. Certain human studies are considered as
more methodologically sound based on study quality attributes identified in Table 5-2 and are
preferred for supporting an RfC. The sections below summarize the evaluation of these studies.
NOAEL* [me] = NOAEL* [ADJ] (pprri) x (Hb/g)A/Hb/g)H: where, NOAEL* [HEC] = the NOAEL or analogous effect
level such as the benchmark concentration (BMC), NOAEL*[ADj] = the NOAEL or analogous effect level adjusted
for duration of experimental regimen; experimental exposure times duration (number of hours exposed/24 hours)
times week (number of days of exposure/7 days), and (Hb/g)A/Hb/g)H = the ratio of the blood/gas (air) partition
coefficient of the chemical for the laboratory animal species to the human value. The value of one is used for the
ratio if (Hb/g)A > Hb/g)H.
5-2
-------
Table 5-1. Neurotoxicological inhalation studies considered in the development of an RfC
Study
(Storm etal.. 2011
[previously reported
inNYSDOH. 2010]
Schreiber et al. (2002)
Schreiber et al. (2002)
Altmann et al. (1995)
Cavalleri et al. (1994):
Gobba et al. (1998)
Spinatonda et al.
(1997)
Species
Human
Human
Human
Human
Human
Human
Duration
10 yr (mean),
continuous
4 yr (mean),
occupational
5.8 yr (mean),
continuous
10.6 yr (median)
continuous
8.8 yr (mean),
occupational
Inhalation (no
duration
information),
occupational
NOAEL/LOAELa
ppm
0.002. 0.05 (children)
0.002. 0.07 (adults)
0.3 (daycare workers,
mean and median)
0.1 (residents, median,
and mean), maybe as
high as 0.4 (mean) and
0.3 (median)
0.7 (mean)
0.2 (median)
6 (Cavalleri et al..
1994)
j| (median)
Effect
(effect magnitude)
at LOAEL
Visual contrast sensitivity (6%
| in children)
Visual contrast sensitivity0
Visual contrast sensitivity0
Cognitive function (14% |),
reaction time (15%-20 |)
visual memory (15% j)
Dyschromatopsia (color
vision) (6% |)d
Reaction time (15% t)
Human equivalent continuous
concentrations'5
NOAEL/
LOAEL
NOAEL
LOAEL
LOAEL
LOAEL
LOAEL
LOAEL
ppm
0.002
0.1
0.4d
0.7
2
3
mg/m3
0.01
0.7
3d
5
15
19
-------
Table 5-1. Neurotoxicological inhalation studies considered in the development of an RfC (continued)
Study
Seeber (1989)
Ferroni et al. (1992)
Echeverria et al.
(1995)
Altmann et al. (1990)
Mattsson et al.
(1998)
Rosengren et al.
(1986)
Kjellstrand et al.
(1985)
Boyes et al. (2009)
Species
Human
Human
Human
Human
Rat
Gerbil
Mouse
Rat
Duration
>10yr (mean),
occupational
10.6 yr (mean),
occupational
15 yr (high-exposure
group; mean),
occupational
4 hr/d for 4 d
Subchronic (13 wk)
6 hr/d, 5 d/wk
Subchronic (12 wk,
with 16-wk follow-
up) continuous
60 min
90min
120 min
NOAEL/LQAEJ/
ppm
12,53
H
11,21,41
10, 5J
0. 50. 200. 800
0, M, 300
0, 90, 320, 400, 600,
800, 1,200, 1,800,
3,600
0. 250. 500, 1,000
0. 1.000. 2,000,
3,000, 4,000
Effect
(effect magnitude)
at LOAEL
Visuospatial function and
information processing speed
(5-30% change depending on
subtest)
Reaction time (10% t),
continuous performance (7-1 1%
4)
Cognitive and visuospatial
measures (4-14% change
depending on subtest)
Visual evoked potentials (2-3
mst)
Flash-evoked potential
(3 ms t)
Brain: protein, DNA
concentration (10-15% change
depending on brain region;
there were both | and J,)
Increased locomotor activity
(20% t)
Impairment in steady state
visual evoked potential (10% j)
Impairment in steady state
visual evoked potential (20% J,)
Human equivalent continuous
concentrations'5
NOAEL/
LOAEL
LOAEL
LOAEL
LOAEL
NOAEL
NOAEL
LOAEL
LOAEL
LOAEL
LOAEL
ppm
4
5
8
4
36
60
90e
250e
l,000e
mg/m3
29
36
56
24
240
410
6,100e
l,700e
6,800e
-------
Table 5-1. Neurotoxicological inhalation studies considered in the development of an RfC (continued)
Study
Wang et al. (1993)
O V /
Oshiro et al. (2008)
Species
Rat
Rat
Duration
Subchronic (12 wk)
continuous
60 min
NOAEL/LQAEL8
ppm
0. 300. 600
0. 500. 1,000, 1,500
0. 500. 1.000. 1,500
Effect
(effect magnitude)
at LOAEL
Reduced brain weight
(j0.10g),DNA
(10.05-0.06 mg), protein
(42.5-3.5 mg)
False alarms (10%|)
Reaction time (200 ms |)
Human equivalent continuous
concentrations'5
NOAEL/
LOAEL
NOAEL
LOAEL
NOAEL
ppm
300e
500e
500e
mg/m3
2,000e
3,400e
3,400e
Note: Studies from which candidate RfCs were derived shaded in blue. 1 ppm = 6.78 mg/m .
"Experimental/observational NOAEL is underlined, LOAEL is double-underlined.
bCalculated using RfC methodology for a Category 3 gas, extrathoracic effects, and adjusted to equivalent continuous exposure. Occupational exposures were
multiplied by 5/7(d) x 10/20 (m3/d, breathing rate), and experimental exposures were multiplied by hours exposed/24 (hr) x 5/7(d).
°Effect magnitude could not be determined from information in published paper.
dAtmospheric 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) for four families living in apartments above active dry cleaning facilities 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) and is used as the LOAEL for this study.
eHECs are the human equivalent concentrations for the same duration as in the experiments, not adjusted to continuous daily exposures.
-------
NYSDOH (2010)-Visual contrast sensitivity; children, residential
Schreiber et al(2002) - Visual contrast sensitivity; adult, occupational
Schreiberet al(2002)-Visual contrast sensitivity; adult, residential
Altmann etal (1995)- Simple reaction tune; adults, residential
CavallerietaU1994); Gobba etal (1998^- Color confusion; adult, occupational
HUMAN CHRONIC STUDIES Spmatonda e
J * Jlq^Jlf
Seeber (1989) — Visuospatialfunction; adult, occupational
POD
NOAEL - • Acute Exposure Studies - *
LOAEL- O
Studies from which candidate RfCs were derived
are underlined and shaded + 2SE
Ferroniet al (1992) — Reaction tune, continuous performance; adult, occupational
^^^^^^^^^^^^^^^^^^^^^^^^ 36 mgfrnr
Echeverria. etal (1995) — Cognithre and visuospatialitieasures; adult, occupational
HUMAN ACUTE STUDIES
Altmann et al (1990) - VEPs; humans, 4 days*
Hake and Stewart (1977) -J.E Gs, humans, 5 days*
242mBm
ANIMAL STUDIES
Mattsson et al(19981-FEP; rat, 12 wte
242 mg/m:
Rosengren et al (1P86) - Brainweight; gerbil, 3 mths
408 Ji^nr
Kjellstrand et al (1985) — Loco motor activity; mice, 60 ntin '
6102lS&im
Boyesetal(2009)-,VEP; rats, 90mm*
169Fmetm
Boyesetal(2009)-VEP; rats, 120 min*
STSUn^nv
Wang etal (1993) -Brain weight; rat, 12wfcs
2034n^nr
Oshiro et al (2008) - False alarms; rats, 60 min*
Oshiro et al (2008) - Reaction time; rats, 60 min*
3390m6firf
OJ001
0.01
0.1
1 10
[PERC] mg m3
100
1000
10000
Figure 5-1. Exposure-response array for neurotoxicological inhalation studies considered for RfC development
(listed in Table 5-1).
PODs (HEC for LOAELs and NOAELs) are displayed and labeled by study, effect, and duration. Studies from which candidate RfCs were
derived are shaded in blue, and the POD ranges (± 2SE) are presented.
-------
Table 5-2. Summary of rationale for identifying studies on
tetrachloroethylene for RfC development
Consideration
Data characteristics
Decision context
Species studied
Animal and human
neurotoxicity studies
Human data are preferred to reduce interspecies extrapolation
uncertainties. Animal data are considered as supporting studies
when adequate human studies are available.
Relevance of
exposure paradigm
scenario
Acute, subchronic,
and chronic exposure
durations
Peak and chronic
exposure intensities
Subchronic or chronic studies, if adequate, are preferred over studies
of acute exposure durations.
Studies of residential exposures, if available and of adequate quality,
are preferred. In residential settings, exposure is more likely to be
continuous, and of lower concentrations compared with the more
intermittent, higher concentration exposure experienced in work
settings. The potential influence of peak or intensity concentrations
is more common with occupational than residential exposures.
Study quality attributes for human toxicity studies
Study populations
Comparability of
referent and exposed
groups
Referent and exposed groups were evaluated and compared. In
addition to age, potential confounders for neurobehavioral measures
including education, lifestyle factors such as alcohol consumption,
and SES are controlled for to limit selection bias and confounding.
Use of a study design (e.g., matching procedures) or analysis
(procedures for statistical adjustment) that adequately addresses the
relevant sources of potential confounding for a given outcome adds
weight to the consideration of the study as principal rather than
supportive.
Measurement of
exposure
Area or individual
measures of exposure
Stronger studies have exposure estimates that are supported by
ambient monitoring and/or biological monitoring. Measurement or
assignment of exposure should not be influenced by knowledge of
results of tests of neurobehavioral function. Higher quality
assessment strategies in occupational studies are based on
assignment of exposure potential to individual subjects considering
individual job titles and tasks with consideration of changes over
time.
Use of higher quality assessment strategies adds weight to the
consideration of the study as principal rather than supportive.
5-7
-------
Table 5-2. Summary of rationale for identifying studies on
tetrachloroethylene for RfC development (continued)
Consideration
Measurement of
effect(s)
Data characteristics
Standardized
neurological tests:
validity and
reliability
Decision context
Neurobehavioral function (reaction time measures, cognitive
function, and motor activity) assessed using a standardized test
battery (e.g., Neurobehavioral Evaluation System) is preferred,
because wide administration to populations in different settings has
resulted in a high degree of validity within the context of potential
population norms. WHO and ATSDR recommend these test
methods to evaluate nervous system deficits in adults and children.
Other standardized methods were used to evaluate color vision and
visual contrast sensitivity.
Administration or interpretation of the test should not be influenced
by knowledge of exposure status. This information adds weight to
the consideration of the study as principal rather than supportive.
Use of standardized neurological tests and sensitive methods to
detect neurological changes adds weight to the consideration of the
study as principal rather than supportive.
Study quality attributes for animal toxicity studies
Study populations
Measurement of
effect(s)
Comparability of
animal models to
effects observed in
humans
Validity and
comparability of
neurological tests
Studies in animal models reporting effects concordant to observed
solvent-associated effects in humans were considered preferable.
Neurological tests and methods that have been validated in animal
models were preferred. Endpoints in animals that were concordant
or comparable with evaluated endpoints in humans were the most
preferred.
5.1.1.3.1. Evaluation of epidemiologic studies of residentially exposed populations
Three epidemiological studies of residential exposures were examined as candidate
principal studies for deriving a RfC (Storm et al., 2011 [previously reported in NYSDOH, 20101:
Schreiber et al., 2002; Altmann et al., 1995). As outlined in Table 5-2, residential exposures
come closest to the chronic, continuous exposures addressed by reference values. The exposed
populations in these studies lived in buildings colocated with dry cleaners. Additional strengths
of all of these studies included high quality exposure assessment, matching of controls by age
and sex, and use of standardized testing. In addition, statistical analyses adjusted for
race/ethnicity, age, and other covariates such as smoking or alcohol use. On the other hand,
there were differences in comparability between referent and exposed groups in each of these
studies for which statistical analyses could not sufficiently adjust, limiting their use as principal
-------
studies. Section 4 describes the studies in detail; study-specific issues relevant to principal study
selection are summarized below.
The NYSDOH pilot study (Schreiber et al., 2002) reported deficits in visual contrast
sensitivity (VCS) in residents exposed to tetrachloroethylene compared to controls. Schreiber
et al. (2002) evaluated 17 exposed subjects, including four children (in New York City) and
17 control subjects (recruited from among NYSDOH employees living in Albany, NY) and
reported reduced group-mean VCS scores in residents compared to unexposed referents at a
human equivalent LOAEL (LOAELnEc) of 3 mg/m3 (arithmetic mean concentration). A key
limitation of this study, in addition to its small sample size and potential for selection bias owing
to health department employees being referents for exposed residents, was that vision testing was
not blinded to exposure classification.
NYSDOH (2010) and Storm et al. (2011) report on a larger study of 104 exposed adult
and children residents of 24 buildings with colocated dry cleaners using tetrachloroethylene and
101 unexposed adults and children in 36 buildings without colocated dry cleaners. High quality
exposure assessment addressed some of the concerns of selection bias in the previous study of
Schreiber et al. (2002): for example, the study employed a larger number of subjects and
referents from the same geographical area. Additionally, exposure and effects were assessed in
family units, allowing comparison of parents and children in the same household. Storm et al.
(2011) identified a human equivalent NOAEL (NOAELHEc) of 0.01 mg/m3 (median
concentration) in children and aNOAELnEc of 0.48 mg/m3 (median concentration) in adults.
However, there are other concerns as to the comparability of referent and exposed subjects.
Those living in households with higher levels of tetrachloroethylene were more likely to be of
minority race and of lower income status compared to referent families. Additionally, exposed
subjects were younger (p < 0.05) and of lower educational attainment (p < 0.05) than those in
referent buildings. Another concern is that, although a standardized visual test (Functional
Acuity Contrast Test [FACT]) was used, it was of far distance VCS only. The test was also less
sensitive than that employed in other studies because the response was scored as either maximum
(perfect) or less than maximum, with no gradations of reduced response. Statistical analyses
appropriately examined the association between these exposure metrics and vision and adjusted
for a number of relevant covariates. However, the small number of nonminority and high
income subjects in the highest tetrachloroethylene exposure group, and the lower mean education
level of the high exposure group, limit conclusions that observed effects were completely
independent of education level, race/ethnicity, or income. This raises concerns about the
comparability between exposed and referent subjects. Consequently, due to the ceiling effect of
the testing method and potential confounding of education level, race/ethnicity, or income,
NYSDOH (2010) and Storm et al. (2011) were not selected as principal studies.
5-9
-------
Altmann et al. (1995) reported visuospatial and cognitive deficits (from two tests of
simple reaction time, continuous performance, and visual memory) among 19 residents
compared to 30 unexposed referents at a LOAELHEc of 5 mg/m3 (arithmetic mean
concentration). Statistical analyses appropriately adjusted for covariates and possible
confounders of age, gender, and education in logistic regression models; however, the paper
lacked reporting of logistic regression coefficients and effect magnitudes, limiting a clear
assessment of the effects observed. Furthermore, the referent group in Altmann et al. (1995) had
a higher educational attainment than tetrachloroethylene-exposed subjects. Altmann et al. (1995)
adjusted for a potential effect of education, a surrogate for socioeconomic status, and on
visuospatial test performance in multiple regression models. However, the National Research
Council (NRC, 2010) noted the potential for residual confounding, as education was examined as
a categorical, not continuous, variable using three groups, which might affect interpretation of
cognitive testing of continuous performance and visual memory. Nonetheless, effects of
tetrachloroethylene exposure were observed on reaction time, an endpoint that is not influenced
by education level. There was potential bias in subject selection: 19 of 95 potentially eligible
subjects participated in the study, and the study did not identify reasons for excluding the
remaining 76 subjects. Altmann et al. (1995) was not selected as a principal study given the
limited reporting, concern about potential selection bias, and concern about residual confounding
for some of the adverse outcomes observed.
In sum, none of these residential studies was selected as a principal study. These studies
nonetheless provide qualitative evidence for hazard identification of neurological deficits in
visual function, reaction time, and cognitive function. The database of residential studies also
adds support for the choice of key endpoints in principal studies and informs uncertainty factor
(UF) selection, as described in Section 5.1.3.
5.1.1.3.2. Evaluation of epidemiologic studies of occupationally exposed populations
Seven occupational studies assessed visual function or other neurobehavioral effects and
were considered as candidate studies for deriving the RfC (Schreiber et al., 2002; Gobba et al.,
1998: Spinatonda et al.. 1997: Echeverria et al.. 1995: Cavalleri et al.. 1994: Ferroni et al.. 1992:
Seeber, 1989). The primary strength of each of these studies is their use of standardized test
methodology to evaluate neurobehavioral or visual function. Additional details regarding the
evaluation of occupational study characteristics that informed selection of candidate studies are
provided below.
Ferroni et al. (1992) was a prevalence study of 60 female dry cleaners or other
dry-cleaning workers and 30 sex-, age-, and vocabulary test score-matched controls from an
industrial cleaning plant that did not use organic solvents. Compared to responses in referents,
5-10
-------
dry cleaners had a 10% increased simple reaction time and decrements in response on
two subtests of the shape comparison test, one of vigilance (7% decrease) and one of stress
(11% decrease) at the LOAEL of 102 mg/m3 [LOAELHEc = 36 mg/m3] (median concentration).
Study details are sparsely reported, and results are not accurately reported in the published paper.
Ferroni et al. (1992) does not clearly identify whether age-matching was for individual subjects,
or for the group's average age. A crude exposure assessment was used based on ambient
monitoring data assigned to the group of dry cleaners, and statistical analyses did not control
adequately for confounding characteristics among participants. As compared to the other
occupational studies, this study had poorer quality in terms of comparability of referent and
exposed groups and measurement of exposure and analysis methods, in part because of poor
reporting of study details and results and, therefore, was not selected as a principal study.
Spinatonda et al. (1997) was a prevalence study of 35 dry cleaners and 39 age- and
education-matched unexposed subjects that reported a 15% increased latency to a vocal response
time at a LOAEL of 54 mg/m3 [LOAELHEc =19 mg/m3] (median concentration). The study
design is sparsely reported, and the paper lacks details of subject selection, including the
population from which controls were drawn, and demographic information for evaluation of
comparability of dry cleaners and controls. Exposure was assessed by a "grab sample" which is
inferior to a time-weighted average estimate. The study developed an index of cumulative
exposure to tetrachloroethylene for each exposed subject by multiplying the tetrachloroethylene
concentration by the number of years worked. Statistical analyses comprised Wests comparing
average latency in dry cleaner and control groups, and regression models fit to responses of
exposed subjects only, a weaker approach than fitting multiple logistic regression models to data
from all subjects. Additionally, the statistical analyses did not control for alcohol consumption,
which is also associated with response time, indicating a greater potential for confounding. As
compared to the other occupational studies, this study had poorer quality in terms of
comparability of referent and exposed groups and measurement of exposure, in part because of
poor reporting of study details and results, as well as less robust statistical analyses controlling
for alcohol consumption. Therefore, Spinatonda et al. (1997) was not selected as a principal
study.
Schreiber et al. (2002) was a small study examining nine adult staff at a day-care facility
colocated in the same building as a dry cleaner, comparing group mean visual contrast values to
age- and sex-matched referents values and identifying a LOAEL of 2 mg/m3
[LOAELHEc = 0.7 mg/m3] (arithmetic mean concentration). Referents in this study were
acquaintances, local retail shop employees, staff of other local day-care centers, or NYSDOH
employees. Exposed and referent subjects were similar on sex and age; however, the paper lacks
any details of whether referents were of similar education or socioeconomic status. Use of
5-11
-------
NYSDOH employees located in Albany, NY, may indicate referents and exposed subjects may
be different on education and other variables. Exposure assignment to subjects was based on
ambient monitoring during time of active dry cleaning; no personal monitoring was conducted.
Schreiber et al. (2002) used a standardized test (FACT) for near vision; however, a shortcoming
is that assessment of vision was 6 weeks after exposure ceased, when measured
tetrachloroethylene concentrations were 100-fold lower than during active dry cleaning. While
Schreiber et al. (2002) adopted a valid and sensitive test to measure vision, it was not selected as
a principal study due to its few subjects, concern that testers were not blinded to exposure
classification, concern about comparability of exposed and referent subjects, and lack of
concurrent exposure and outcome assessment.
Seeber (1989) evaluated the neurobehavioral effects of tetrachloroethylene on
101 dry-cleaning workers (employed in coin-operated or while-you-wait shops) and reported
effects on several measures of cognition at a LOAEL of 83 mg/m3 [LO AELnEc = 29 mg/m3]
(time-weighted average mean concentration), compared to referents from several department
stores and receptionists from large hotels (refer to section 4.1.1.2 for more details). Statistically
significant changes from control of 5-30% were seen at the LOAEL for the measures of
perceptual speed, digit reproduction, cancellation, and digit symbol. The exposure assessment
used estimates of long-term exposure from interview data, active sampling of room air, and
passive sampling of personal air to assign dry cleaners to two exposed groups (mean ± SD: 83 ±
53 and 364 ±114 mg/m3). Strengths of the study included the relatively large sample sizes for
all three groups (57, 44, and 84 subjects in the lowest, highest, and referent groups, respectively),
measurement of effects using recognized methods (standardized tests of symptoms and
personality; tests of sensorimotor function, including finger tapping and aiming; digit
reproduction and digit symbol) and use of examiners blinded to subjects' exposure status.
Stratified regression analysis was used to statistically control for the influence of potentially
confounding factors—gender ratios, age, and scores on the intelligence test—on test scores.
Additional adjustment for group differences in alcohol consumption were considered, and did
not alter the results. A limitation of the study was that no information was provided on the
methods used to identify subjects or their reasons for participating in the study, but this was
offset by the ascertainment of potentially confounding factors and the use of multiple regression
to adjust for these factors. Another limitation was the lack of individual data to clarify dose-
response relationships, as test outcomes and exposures were reported only as summary measures
(mean ± SD) for two substantially overlapping exposure groups. There is some uncertainty in
the quantitative dose-response relationships for this study, given the substantial overlap in
exposure estimates between the low- and high-exposure groups and the wide confidence
intervals around the mean test results. For many test outcomes, one or both exposure groups was
5-12
-------
statistically significantly different from controls, but no significant differences between the low
and high exposure were reported. For some outcomes, NRC (2010) characterized the study as
having discrepant results based on worse mean test scores (for neurologic signs, emotional
lability, choice reaction time, cancellation d2 and digit symbol) in the low- compared with high-
exposure group. This study was not among those recommended by NRC for consideration in
deriving the RfC.
Cavalleri et al. (1994) and Gobba et al. (1998) are two studies of the same exposed
population. Cavalleri et al. (1994) reported poorer performance (6% decrement on average) on a
test of color vision among 35 dry cleaning and laundry workers compared to 35 controls matched
on age, alcohol consumption, and smoking. The LOAEL for all workers in this study was
42 mg/m3 [LOAELHEc =15 mg/m3] (time-weighted average mean concentration). Controls were
not matched on education or intelligence, but these factors have not been shown to be associated
with color vision. Exposure was assessed for individual subjects from personal monitoring over
the full work shift and represented an 8-hour time-weighted average. Standard testing methods,
including an established protocol, were used to detect changes in color vision, which were
assessed by the Lanthony D-15 Hue desaturated panel. The investigators' statistical analyses
included comparison of group mean Color Confusion Indexes (CCIs) by the arithmetic mean of
three exposure groupings: all workers (42 mg/m3), dry cleaners (49 mg/m3), and ironers
(33 mg/m3), and multiple logistic regression analyses which adjusted for effects of age, alcohol
consumption, and smoking.
Gobba et al. (1998) examined color vision in 33 of these 35 dry cleaners and laundry
workers after a 2-year period and reported a further decrement in color vision (9% decrement on
average) among 19 subjects whose geometric mean exposure had increased from 12 mg/m3 to
29 mg/m3 over the 2-year period. No improvement was observed among 14 subjects whose
geometric mean exposure had decreased from 20 mg/m3 to 5 mg/m3. The mean responses of
both subgroups supported a persistence of deficits in visual function and suggested a worsening
of effects when exposure increased for individuals. A strength of Gobba et al. (1998) is subjects
serving as their self-controls, with scores on the test of color vision compared from the initial and
follow-up studies. Given the vision deficits reported by Cavalleri et al. (1994), Gobba et al.
(1998) serves to confirm and extend those findings.
Cavalleri et al. (1994) is preferred to Gobba et al. (1998) for candidate reference value
derivation for several reasons. First, the earlier study more clearly associated a deficit in color
vision with tetrachloroethylene exposure through comparison to a suitable and well
characterized, unexposed reference group. The Gobba et al. (1998) study did not include
unexposed controls and, therefore, cannot distinguish the possible impact of age on the CCI
scores of subjects who were 2 years older at the second evaluation. Second, the Gobba et al.
5-13
-------
(1998) study suggests that the earlier exposure was sufficient to cause the CCI deficit in at least
those subjects (n= 14) whose exposure decreased after the earlier evaluation. While the Gobba
et al. (1998) study also demonstrated further deficits in workers whose exposure increased after
the first study (n = 19), it is unclear how to relate the higher measurement to the incremental
deficit, given the lack of improvement in the subset with decreased exposure and the lack of
information concerning the other confounding variables considered in the first evaluation—
absolute age, smoking, and alcohol status. In any case, a deficit existed in this subset before the
follow-up period, at a lower exposure than that of the second evaluation. Third, the exposures in
Cavalleri et al. (1994) were reported as time-weighted average arithmetic means, which are
expected to represent total risk better than time-weighted average geometric means [as reported
in Gobba et al. (1998)1 when data are grouped (Crump, 1998). The point of departure (POD)
was, therefore, taken from the Cavalleri et al. (1994) study. The exposure level for the full study
sample is used as the LOAEL for several reasons. Although no apparent CCI deficit was
observed in ironers, their reported exposure range (0.52-11.28 ppm, or 3.5-76 mg/m3) was
completely contained within the range of exposures for dry cleaners (0.38-31.19 ppm, or
2.6-210 mg/m3). Yet elevated CCI scores were observed at exposures lower than the mean
exposure of the ironers (4.8 ppm, or 33 mg/m3), indicating that the mean exposure of the ironers
cannot be considered a NOAEL. For these reasons, Cavalleri et al. (1994) is used to derive a
candidate RfC.
Echeverria et al. (1995) examined 65 dry cleaners in Detroit, MI, using a standardized
neurobehavioral battery and found changes in cognitive and visuospatial function. A LOAEL of
156 mg/m3 [LOAELHEc = 56 mg/m3] (time-weighted average mean concentration) was
identified, based on comparison of the two higher exposure categories with an internal referent
group comprising mainly counter clerks, who were matched to exposed dry cleaners on age and
education. Changes of 4-14% from internal referent levels, depending on subtest, were observed
at the LOAEL. The study had a high quality exposure-assessment approach and appropriate
statistical analyses that adjusted for covariates including alcohol. A potential selection bias may
have resulted from the 18% participation rate among dry-cleaning shop owners, if the low
participation could be explained by the health status of employees. The study also lacked an
unexposed referent group; subjects were categorized into three exposure groups. Without an
unexposed control group, the exposure level for the lowest exposure group (i.e., the internal
referent group) cannot be classified as a NOAEL or a LOAEL. This study was of relatively good
quality in terms of the comparability of referent and exposed groups, measurement of effect, and
measurement of exposure and, although there are concerns about the lack of an unexposed
referent group, this study was used to derive a candidate RfC.
5-14
-------
5.1.1.3.3. Evaluation of experimental human exposure studies
The two human controlled exposure studies (Altmann et al., 1990; Hake and Stewart,
1977) were of fewer subjects and shorter exposure durations, and effects were observed at higher
exposure concentrations than chronic studies of residential and occupational exposure. While
subjects in Altmann et al. (1990) could serve as their own controls, there was not an unexposed
group. Therefore, neither study was selected as a principal study given the availability of
suitable human data of chronic duration. These studies do provide qualitative evidence for
hazard identification of neurological deficits in visual function and neurological function and add
support for choice of key endpoints in principal studies.
5.1.1.3.4. Evaluation of animal neurotoxicity studies
The animal neurotoxicity studies mostly consist of acute duration studies (Boyes et al.,
2009; Oshiro et al., 2008; Kj ell strand etal., 1985) and subchronic (repeated dosing) studies,
which generally involve lower exposures than the acute animal studies (Mattsson et al., 1998;
Wang etal., 1993; Rosengren et al., 1986). However, these studies covered shorter exposure
duration periods than the available human studies and require extrapolation of animal
observations to humans. They were not considered principal studies given the availability of
suitable human data from chronic exposures. The findings in the animal studies contribute to the
weight of evidence that tetrachloroethylene exposure results in neurological deficits and is
considered supportive of the human studies in terms of hazard identification.
5.1.1.3.5. Selection of studies
To summarize, three studies (Echeverria et al., 1995; Cavalleri et al., 1994; Seeber, 1989)
had more of the preferred qualities listed in Table 5-2 compared to other epidemiologic studies
of occupational and residential exposure. However, NRC (2010) characterized the Seeber (1989)
study as having discrepant results based on worse mean test scores (for neurologic signs,
emotional lability, choice reaction time, cancellation d2 and digit symbol) in the low- compared
with high-exposure group. Therefore, Seeber (1989) was not among those recommended by
NRC for consideration in deriving the RfC.
NRC (2010) recommended five studies for consideration in deriving the RfC (Altmann et
al., 1990: Boves et al., 2009: Echeverria et al., 1995: Cavalleri et al., 1994: and Gobba et al.,
1998). Two acute studies recommended for consideration by NRC [the human chamber study of
Altmann et al. (1990) and the rodent study of Boyes et al. (2009)] were judged by EPA to be
supportive, but were not considered further for deriving candidate RfCs because of the
preference to use quality studies of chronic, human exposures over studies of acute exposures.
In addition, two of the other studies recommended by NRC (2010), Cavalleri et al. (1994), and
5-15
-------
Gobba et al. (1998), evaluated the same cohort, and the earlier study was preferred by EPA due
to its use of a control group and the clearer identification of a POD (refer to section 5.1.1.3.2).
Thus, two studies—Cavalleri et al. (1994) and Echeverria et al. (1995)—are considered principal
studies by EPA for the derivation of the RfC. Endpoints selected for the candidate RfCs were
reaction time measures (Echeverria et al., 1995), cognitive changes (Echeverria et al., 1995), and
visual function changes (Cavalleri et al., 1994).
5.1.2. Additional Analyses: Feasibility of Dose-Response Modeling
The present analysis defines a POD using the traditional NOAEL/LOAEL approach. The
NOAELs/LOAELs were adjusted to an equivalent continuous exposure (U.S. EPA, 1994), and
described in Section 5.1.1) so that comparisons could be made between studies. Ambient
(inhaled) concentration of tetrachloroethylene was used as the dose metric in deriving the RfC.
Because the application of dose-response modeling offers advantages over the use of
NOAELs/LOAELs, the data sets from the endpoints in the two studies (refer to Table 5-3)
(Echeverria et al., 1995; Cavalleri et al., 1994) were evaluated for feasibility of dose-response
modeling. In both studies, it was determined that PODs could not be derived using dose-
response modeling, for varying reasons as detailed below.
In evaluating the CCIs in Cavalleri et al. (1994), normative data for color confusion
(Lomax et al., 2004; Iregren et al., 2002) were considered. However, the normal ranges are
influenced strongly by age. Although the investigators reported statistical significance of the
cleaning workers' CCIs using analyses which adjusted for age (as well as smoking and alcohol
consumption), no individual ages were provided with the individual CCIs and exposure
measurements, and no individual measurements for the control subjects were provided.
Insufficient information was available to make use of the investigators' model, including
response measures adjusted for age.
Echeverria et al. (1995) identified three exposure groups, but there was no unexposed
group for comparison. Historical control data from the Echeverria group were unavailable,
precluding the derivation of PODs from the logistic regression they reported. In some cases,
particularly with individual exposure and response data available, a control level might be
inferred by extrapolating from the observed data. However, only summary values for exposure
and responses were available. Therefore, projecting a control response and biologically relevant
level of change from that point was judged to be too uncertain.
5-16
-------
Table 5-3. Application of uncertainty factors for the neurological endpoints
from the studies used to derive candidate RfCs [Echeverria et al. (1995) and
Cavalleri et al. (1994)1
Neurological endpoint
Human
equivalent
NOAEL/LOAEL
(mg/m3)
Uncertainty factors (UFs)
Composite
UF
UFA
UFH
UFS
UFD
UFL
Candidate
RfC
(mg/m3)
Reference
Cognitive Domain
Visual reproduction,
pattern memory, pattern
recognition — adult,
occupational
56
(LOAEL)
1,000
1
10
1
10
10
0.056
Echeverria
etal.
(1995)
Reaction Time Domain
Reaction time in pattern
memory — adult,
occupational
56
(LOAEL)
1,000
1
10
1
10
10
0.056
Echeverria
etal.
(1995)
Visual Function Domain
Color confusion —
adults, occupational
15
(LOAEL)
1,000
1
10
1
10
10
0.015
Cavalleri
etal.
(1994)
5.1.3. Reference Concentration (RfC) Derivation, Including Application of Uncertainty
Factors
The RfC is the midpoint of the range of candidate reference values from two principal
studies. Candidate RfCs for tetrachloroethylene were derived by dividing the PODs of 15 mg/m3
(Cavalleri etal.. 1994) and 56 mg/m3 (Echeverria et al.. 1995) by a total UF of 1,000, comprised
of 10 for interindividual variability, 10 for extrapolation from a LOAEL to a NOAEL, and 10 for
database uncertainty. The application of uncertainty factors is based on EPA's A Review of the
Reference Dose and Reference Concentration Processes [(U.S. EPA, 2002): Section 4.4.5],
which address five areas of uncertainty.
• An UF of 10 was applied to account for human variability in the effects that were used
for the derivation of the RfC. The principal studies are based on occupationally exposed
subjects, who are generally healthier than the overall population, and, thus, provide no
data to determine the relative effects of susceptible population including children, elderly,
and/or people with compromised health. Additionally, no information was presented in
the human studies with which to examine variation among subjects. Quantitative
analyses have been carried out by Clewell et al. (2004) and Pelekis et al. (2001)
evaluating pharmacokinetic variation between adults and children for
tetrachloroethylene and its metabolites using physiologically based pharmacokinetic
5-17
-------
(PBPK) models. However, validation of these results for various life-stages and further
refinement of the parameters in the model have not been conducted.
• An UF of 1 was applied to account for interspecies variability in extrapolation from
laboratory animals to humans because the principal studies and critical endpoints were
from human studies.
• An UF of 1 was applied for the use of data from subchronic study to assess potential
effects from chronic exposure because the PODs are based on studies involving chronic
exposure (refer to Section 5.1.3).
• An UF of 10 was applied for the extrapolation from a LOAEL to a NOAEL because the
PODs from the studies were LOAELs.
• An UF of 10 was applied to address the lack of data to adequately characterize the hazard
and dose response in the human population. The following critical data gaps have been
identified: uncertainties associated with database deficiencies on neurological,
developmental, and immunological effects. The two studies (Echeverria et al., 1995;
Cavalleri etal., 1994) used to derive the RfC evaluated neurotoxicity following
occupational exposures with PODs 3- to 100-fold higher than those identified from
residential studies (Storm et al., 2011 [previously reported in NYSDOH, 20101: Schreiber
et al., 2002; Altmann et al., 1995). In comparison to the occupational studies, the
available residential studies were judged to be more limited for developing an RfC, based
on consideration of the study design (population comparability) and/or selection of
neurological methods. However, they provide human evidence of neurotoxicity
following tetrachloroethylene exposure in a residential setting, with reaction time deficits,
visual system dysfunction, and cognitive performance deficits.
In addition, data characterizing dose-response relationships and chronic visuospatial
functional deficits and the cognitive effects of tetrachloroethylene exposure under
controlled laboratory conditions are lacking. Data from acute studies in animals (Oshiro
et al., 2008: Umezu et al., 1997: Warren et al., 1996) suggest that cognitive function is
affected by exposure to tetrachloroethylene. These studies do not address the exposure-
response relationship for subchronic and chronic tetrachloroethylene exposures on
cognitive functional deficits observed in humans (e.g., Altmann et al., 1995: Echeverria
et al., 1995: Seeber, 1989). There is also a lack of cognitive testing following exposures
of longer than acute duration, including during development. Visual system dysfunction
and processing of visuospatial information are sensitive endpoints in human studies. The
exposure-response relationship of these functional deficits could be evaluated more
definitively with studies using homologous methods that examine retinal and visual
function in experimental animals. However, there has been a limited evaluation of
effects of chronic exposure to tetrachloroethylene on visual function in rodents, with the
exception of the evoked potential studies by Mattsson et al. (1998). These types of
studies could help determine whether there are both peripheral and central effects of
tetrachloroethylene exposure on visual perception, and they could be used as an animal
model to better define the exposure-response relationships in humans.
5-18
-------
Finally, additional data are needed to assess the potential hematological and
immunological effects of tetrachloroethylene. In humans, Emara et al., (2010) reported
changes in various standard hematological measures in subjects with mean
tetrachloroethylene blood levels of 1.685 mg/L. The limited laboratory animal studies of
hematological toxicity demonstrated an effect of tetrachloroethylene exposure on red
blood cells (decreased RBCs (Ebrahim et al., 2001), or decreased erythrocyte colony-
forming units (Seidel etal., 1992)), with reversible hemolytic anemia observed in female
mice exposed to low drinking water levels (0.05 mg/kg-day) of tetrachloroethylene
beginning at 2 weeks of age in one series of studies (Marth etal., 1989; Marth, 1987;
Marth et al., 1985a: Marth et al., 1985b). Ebrahim et al. (2001) also observed decreased
hemoglobin, platelet counts, and packed cell volume, and increased WBC counts.
Although additional corroborating studies are lacking, the observation of an effect at a
low exposure level raises additional concern about hematological and immunological
effects. The fact that other solvents [e.g., toluene, and the structurally similar solvent
trichloroethylene (Cooper et al., 2009)] have been associated with immunotoxicity
contributes further concern about this gap in the database for tetrachloroethylene.
These UFs were applied to each of the following endpoints from the selected
neurotoxicological studies of occupational tetrachloroethylene exposure: color vision changes
(Cavalleri etal., 1994): and cognitive and reaction time changes (Echeverria et al., 1995). The
UFs for each study and endpoint are presented in Table 5-3 as well as in Figure 5-2. The,
candidate RfCs from these studies span a range from 0.015 to 0.056 mg/m3. The RfC for
tetrachloroethylene is 0.04 mg/m3, the midpoint of this range rounded to one significant figure.
Echeverria etal (1995)-Cognitive measures; adult, occupational- LOAEL
0.056 mg/m3<^^^^^^^^^^EJ-jyyyyyyyyaggggggggggggg^^^^ULJLJLJLJLJLJLJ 1 > 56 mg/m3
Echeverria et al (1995) - Reaction time measures; adult, occupational - LOAEL
0.056 mg/m^-p-r I I I I I • • tfMWlllllllllllllllllll<<<<<&-'ULf-f-f-f-f-fJ* 56 mg/m3
Cavalleri et al (1994) - Color confusion; adult, occupational - LOAEL
• Point of Departure
Q TJEA - Interspecies; animaltohuman
• TJEs - Subchronic to chronic exposure duration
H UFL - LOAEL to NOAEL
^ TJEn-Intraspecies;humanvariability
BUFD- Database
V Candidate Reference Concentration
[PERC] mg/m3
Figure 5-2. Candidate reference concentration values for inhalation exposure to
tetrachloroethylene.
5-19
-------
5.1.4. Dose-Response Analyses for Comparison of Noncancer Effects Other Than Critical
Effects in Neurotoxicity
This section presents inhalation dose-response analyses for noncancer effects other than
the critical effect of neurotoxicity. The purpose of these analyses is twofold: (1) to provide a
quantitative characterization of the relative sensitivity of different organs/systems to
tetrachloroethylene, and (2), to provide information that may be useful for cumulative risk
assessment in which multiple chemicals have a common target organ/system other than the
central nervous system. Therefore, for each organ/system, "sample reference concentrations"
(sRfCs) are calculated based on the same methodology as is used for the critical effect of
neurotoxicity. These sRfCs are based on an evaluation of studies identified in Section 4.10 as
suitable for dose-response analysis.
The method of analysis is the same as that described above for neurotoxicity, using the
NOAEL/LOAEL approach. Benchmark dose modeling was not performed because these sample
RfCs are meant for comparison purposes only (across organs/tissues or across chemicals). HECs
are derived using either (1) the RfC methodology for a Category 3 gas, extrathoracic effects,
adjusted for equivalent continuous exposure; or (2) the PBPK model with an appropriate dose
metric. For liver effects, the dose metric of liver oxidative metabolism was used, based on the
view that oxidative metabolites are involved in tetrachloroethylene-induced liver effects. For
kidney effects, while it is generally thought that GSH conjugation metabolites are involved, the
large uncertainty in estimates of human GSH conjugation preclude use of that dose metric.
Instead, the AUC of tetrachloroethylene in blood is used as a surrogate. For the other non-cancer
effects, the AUC of tetrachloroethylene in blood was used as the preferred dose metric due to the
lack of data on what the active toxic moeity(ies) may be for those effects. In addition, the PBPK
model is being used to perform route-to-route extrapolation from oral to inhalation exposure, so
both inhalation and oral studies are considered together here. The HEC is then treated as a POD
to which the following uncertainty factors may be applied:
• An UF of 10 was applied for human variation to all PODs. The rationale is the same as
described above for neurotoxicity. Furthermore, there is some indication that human
variability (at least for one endpoint) may be substantially more than that implied by the
default UF. Kidney toxicity is thought to be associated with metabolism of
tetrachloroethylene along the glutathione (GSH) conjugation pathway. As described in
Section 3.5, PBPK model predictions for GSH conjugation span a wide range that may be
due to uncertainty, variability, or both. Glutathione S-transferases (GSTs) are known to
be polymorphic in the human population, with some isoforms exhibiting a substantial
population of null phenotypes.
5-20
-------
• An UF of 3 was applied to the PODs from all rodent studies to account for interspecies
variability in extrapolation from laboratory animals to humans. The PODs from rats and
mice are expressed as HECs calculated using either the Methods for Derivation of
Inhalation Reference Concentrations and Application of Inhalation Dosimetry (hereafter
referred to as the RfC Methodology (U.S. EPA. 1994) or the PBPK model. Therefore,
the UF of 3 was applied for animal-to-human uncertainty to the PODs from these studies
of rats and mice to account for potential pharmacodynamic differences. This factor is not
applied to PODs from human studies.
• An UF of 10 was applied to PODs of studies of subchronic or shorter duration to address
the potential for additional or more severe toxicity from chronic or lifetime exposure.
• An UF of 10 was applied when a LOAEL is used due to a lack of a NOAEL. This factor
was applied to the PODs of studies that identified a LOAEL but not a NOAEL.
• A database UF of 10 was applied to all PODs to address the lack of data to adequately
characterize the hazard and dose response. The rationale is the same as described above
for neurotoxicity.
5.1.4.1. Sample Reference Concentrations (RfCs) for Kidney Toxicity
As discussed in Section 4, numerous studies have reported adverse effects in the kidney
from tetrachloroethylene. Five studies reporting kidney toxicity were identified in Section 4.10
as suitable for dose-response analysis. The only human study was Mutti et al. (1992), which
reported statistically significant increases in retinol binding protein (RBP), p2n-globulin, and
albumin in urine among dry cleaners as compared to matched controls. In addition, for
seven different urinary markers, the prevalence of individuals with abnormal values
(>95th percentile of controls) was four- to fivefold greater in the exposed group. This study was
in humans chronically exposed and was, thus, used to calculate an sRfC. Of the rodent studies
reporting nephrotoxicity, only JISA (1993) identified a chronic NOAEL, with the other
three rodent studies reporting subchronic (Jonker et al., 1996) or chronic LOAELs (NTP, 1986;
NCL 1977).
Therefore, among the rodent studies, only JISA (1993), which reported effects in both
mice and rats, was used in sRfC calculations. A summary of the PODs and UFs applied is in
Table 5-4. The resulting sRfCs range from 0.05-0.2 mg/m3 based on nuclear enlargement
(karyomegaly) in the proximal tubules of chronically exposed mice and rats (JISA, 1993) with a
slightly lower sRfC of 0.03 mg/m3 based on urinary markers of nephrotoxicity in occupationally
exposed humans (Mutti etal., 1992).
5-21
-------
5.1.4.2. Sample Reference Concentrations (RfCs) for Liver Toxicity
As discussed in Section 4, numerous studies have reported adverse effects in the liver
from tetrachloroethylene. Six studies, none in humans, reporting liver toxicity were identified in
Section 4.10 as suitable for dose-response analysis. Only JISA (1993) reported a chronic
NOAEL, and so was carried forward for derivation of an sRfC. However, it is unclear whether
the reported effect of angiectasis, or enlargement of the blood vessels, is related to the other liver
effects of tetrachloroethylene, which generally involve hepatocytes. Therefore, two other studies
were utilized, one of which reported a chronic LOAEL for liver degeneration and necrosis (NTP,
1986), and the other of which reported a NOAEL for liver weight increases after 6-week
exposures (Buben and O'Flahertv, 1985). The remaining studies either only reported a LOAEL
(Jonker et al., 1996; Kjellstrand et al., 1984), or reported a NOAEL for a very short duration [14
days, Berman et al. (1995)], and were, therefore, not considered further.
Therefore, JISA (1993), NTP (1986), and (Buben and O'Flahertv, 1985) were used to
calculate sRfCs. In addition, PBPK modeling was used to calculate the total rate of oxidative
metabolism in the liver as a dose metric for deriving the HECs.29 Table 5-5 summarizes the
PODs and UFs applied. The resulting sRfCs range from 0.09 mg/m3 based on increased
liver/body-weight ratios after 6-week exposures (Buben and O'Flaherty, 1985) to 0.7 mg/m3
based on liver effects after chronic exposures (JISA, 1993; NTP, 1986). It should also be noted
that in the chronic studies, increased liver tumors were observed at the lowest doses tested.
Therefore, under chronic exposure conditions, cancer effects are likely to be more
important than noncancer effects in the liver.
29 The MOA for tetrachloroethylene-induced liver toxicity is not clear. It appears that TCA as the sole contributory
metabolite cannot explain tetrachloroethylene-induced hepatotoxicity (ClewelletaL 2005: Buben and O'Flahertv.
1985). It is not known whether reactive intermediates such as tetrachloroethylene oxide and trichloroacetyl chloride
are involved in induced liver toxicity. In consideration of these uncertainties, it appears more appropriate to use
total rate of oxidative metabolism as the dose metric for tetrachloroethylene-induced liver toxicity. This quantity is
then scaled by body weight to the 374th power so as to enable extrapolation of risk across species.
5-22
-------
Table 5-4. Sample RfCs for kidney effects
Kidney endpoint
(species)
Urinary markers of
nephrotoxicity (human)
Nuclear enlargement in
proximal tubules (rat)
Nuclear enlargement in
proximal tubules
(mouse)
HECin
mg/m3
(LOAEL/
NOAEL)
34 (LOAEL)
61 (NOAEL)
14 (NOAEL)
Uncertainty factors (UFs)
Composite
UF
1,000
300
300
UFA
1
3
3
UFH
10
10
10
UFS
1
1
1
UFD
10
10
10
UFL
10
1
1
Sample
RfC
(mg/m3)
0.03
0.2
0.05
Reference
Mutti et al.
(1992)
JISA0993)
JISA0993)
Table 5-5. Sample RfCs for liver effects
Liver endpoint (species)
Increased angiectasis
(mouse)
Increased liver
degeneration/necrosis
(mouse)
Increased liver/body-weight
ratio (mouse)
HECain
mg/m3
(LOAEL/
NOAEL)
210
(NOAEL)
2,100
(LOAEL)
270b
(NOAEL)
Uncertainty factors (UFs)
Composite
UF
300
3,000
3,000
UFA
3
3
3
UFH
10
10
10
UFS
1
1
10
UFD
10
10
10
UFL
1
10
1
Sample
RfC
(mg/m3)
0.7
0.7
0.09
Reference
JISA
(1993)
NTP
(1986)
Buben &
O'Flaherty
(1985)
""Calculated with PBPK model using the dose metric of liver oxidative metabolism.
bRoute-to-route extrapolation from oral exposure.
5.1.4.3. Sample Reference Concentrations (RfCs) for Immunotoxicity and Hematologic
Toxicity
As discussed in Section 4, a number of studies have reported changes in hematologic or
immunologic parameters with tetrachloroethylene exposure. Two studies reporting hematologic
effects were identified in Section 4.10 as suitable for dose-response analysis. The human study
(Emaraetal., 2010) reported changes in various standard hematological measures in subjects
with mean blood levels of 1.685 mg/L. Application of the PBPK model gives an air
concentration estimate during exposure of 18 ppm corresponding to this blood level, assuming
constant concentration during exposure. Adjustment to equivalent continuous exposure gives an
5-23
-------
HEC of 6.4 ppm, or 43 mg/m3. This can be treated as a chronic LOAEL, given the 7-year mean
exposure duration (>10% of lifespan). The other study (Marth, 1987) reported reversible
hemolytic anemia after drinking water exposure to 2-week-old mice for 7 weeks. Because only a
LOAEL was identified, the exposures were subchronic, and the effect has not been reproduced at
such low exposure in other studies, Marth (1987) was not considered further for sRfC derivation.
However, it should be noted that the LOAEL identified was low—0.05 mg/kg-day—and may be
a cause for additional concern about hematologic effects. Therefore, Emara et al. (2010) was
used to calculate an sRfC. A summary of the POD and UFs applied is in Table 5-6. The result is
an sRfC of 0.04 mg/m3.
5.1.4.4. Sample Reference Concentrations (RfCs) for Reproductive and/or Developmental
Toxicity
As discussed in Section 4, a number of studies have reported reproductive and
developmental effects from tetrachloroethylene exposure. Four studies, none in humans,
reporting reproductive or developmental effects, were identified in Section 4.10 as suitable for
dose-response analysis. All of these studies reported NOAELs. The developmental studies were
all of appropriate duration for detecting those effects. The reproductive study (Beliles et al.,
1980) was short term (5 days exposure), but was the only suitable study for reproductive toxicity,
and assessment was limited to males. Therefore, all four studies were used to calculate sRfCs.
A summary of the PODs and UFs applied is in Table 5-7. For all these endpoints, the UF for
subchronic to chronic extrapolation was not used because the studies sufficiently covered the
developmental window or window of sperm development. The resulting sRfCs range from
0.4-0.7 mg/m3 for different developmental effects (Carney et al., 2006; Tinston, 1994; Nelson et
al., 1979), with an intermediate value of 0.5 mg/m3 for reduced sperm quality (Beliles et al.,
1980).
5.1.4.5. Summary of Sample Reference Concentrations (RfCs) for Noncancer Endpoints
Other Than the Critical Effect
The lowest sRfCs for these noncancer endpoints are similar to the values calculated based
on the critical effect of neurotoxicity (refer to Figure 5-3), therefore supporting the selection of
the critical effect: 0.03 mg/m3 from Mutti et al. (1992) and 0.04 mg/m3 from Emara et al. (2010).
The other sRfCs are less than 20-fold greater than the RfC. This suggests that multiple effects
may begin to occur as exposure rises above those at which tetrachloroethylene begins to induce
neurotoxicity. These results also suggest that it is important to take into account effects from
tetrachloroethylene other than neurotoxicity when assessing the cumulative effects of multiple
exposures.
5-24
-------
Table 5-6. Sample RfCs for immunological and hematological effects
Immunotoxicity/hematotoxicity
endpoint (species)
Reduced RBC, hemoglobin;
increased WBC, lymphocytes,
IgE (human)
HECin
mg/m3
(LOAEL/
NOAEL)
43
(LOAEL)
Uncertainty factors (UFs)
Composite
UF
1,000
UFA
1
UFH
10
UFS
1
UFD
10
UFL
10
Sample
RfC
(mg/m3)
0.04
Reference
Emara et al.
(2010)
RBC = red blood cells. WBC = white blood cells.
Table 5-7. Sample RfCs for reproductive and developmental effects
Reproductive/developmental
endpoint (species)
Decreased weight gain; altered
behavior, brain acetylcholine (rat)
Reduced sperm quality (mouse)
Increased F2A pup deaths by Day
29; CNS depression in Fl and F2
Decreased fetal and placental
weight; skeletal effects (rat)
HECin
mg/m3
(LOAEL/
NOAEL)
200
(NOAEL)
140
(NOAEL)
122
(NOAEL)
110
(NOAEL)
Uncertainty factors (UFs)
Composite
UF
300
300
300
300
UFA
3
3
3
3
UFH
10
10
10
10
UFS
1
1
1
1
UFD
10
10
10
10
UFL
1
1
1
1
Sample
RfC
(mg/m3)
0.7
0.5
0.4
0.4
Reference
Nelson et al.
(1979)
Beliles et al.
(1980)
Tinston et al.
(1994)
Carney et al.
(2006)
to
CNS = central nervous system.
-------
fj\
to
cc
CMS
_ UFcomp=1000
UFcomp=1000 _
Kidney
!— i UFcomp=1000 »
°
1
•— 1 UFcomp=300
BUFcomp=300 ~
Liver
.—, UFcomp=300
al. (1994) [H]
Jtti et al. (1992) [H]
[M]
.— I UFcomp=3000 _ L
QUFcomp=3000
Immunolog
— i UFcomp=1000 _
ical/Hematological
Emaraet al. (2010) [H]
Reproductive/Developmental
Q UFcomp=300 ^
i— i UFcomp=300
.— 1 UFcomp=300
•— | UFcomp=300
10'
D2
10
D1
1 fl-
mg/
m
Figure 5-3. Comparison of candidate RfCs (filled squares) supporting the RfC (vertical line) and sample RfCs (open
squares) for effects other the critical effect (CNS toxicity).
Filled circles = study/endpoint LOAEL in terms of human equivalent concentrations. Open circles = study/endpoint NOAEL in terms of
human equivalent concentrations. Species in each study is shown in brackets after the reference (mouse: M; rat: R; human: H).
-------
5.1.5. Previous Inhalation Assessment
There is no previous IRIS RfC for tetrachloroethylene.
5.1.6. Uncertainties in Inhalation Reference Concentration
As presented above, the uncertainty factor approach was applied to PODs consisting of
LOAELs from two epidemiologic studies of neurological effects. These studies were considered
to be methodologically sound based on study quality attributes, including study population
selection, exposure measurement methods, and endpoint measurement methods. Other strengths
are that they are human studies of chronic duration, obviating the need for extrapolation across
species and exposure duration. However, there are a number of uncertainties deriving the RfC
from POD, as discussed below.
First, there is uncertainty in the POD, which was based on a LOAEL. In particular, a
LOAEL or a NOAEL is a reflection of the particular exposure concentrations or doses at which a
study was conducted. These lack characterization of the dose-response curve and for this reason
is less informative than a POD obtained from dose-response modeling. In addition, the PODs
are all LOAELs because a NOAEL was not identified in the studies and benchmark dose-
response modeling was not feasible. An UF of 10 is applied as an adjustment to the LOAEL, but
the actual extent of adjustment necessary may be larger or smaller.
Second, there is uncertainty related to human variability. Subjects in the principal studies
comprise a population of occupationally exposed adult subjects, and there is uncertainty in
extrapolating doses to a larger, more diverse population. In the absence of tetrachloroethylene-
specific data on human variation, a factor of 10 was used to account for human variation. Actual
human variation in tetrachloroethylene susceptibility may be larger or smaller; however, there
are inadequate tetrachloroethylene-specific data to examine the potential magnitude of any over-
or under-estimation.
Finally, critical data gaps have been identified that also contribute to uncertainty in the
RfC. In particular, there is a need for high quality epidemiologic studies of residential exposures
and chronic-duration animal studies (including in developing animals). A fuller characterization
is also needed of the noncancer effects other than the critical effect of neurotoxicity, particularly
immunological and hematological effects. Given these limitations, a factor of 10 was used to
account for database limitations, but the actual factor necessary may be larger or smaller.
5-27
-------
5.2. ORAL REFERENCE DOSE (RfD)
5.2.1. Choice of Principal Studies and Critical Effects
Generally, the studies of greatest duration of exposure and conducted via the oral route of
exposure have the most confidence for derivation of an RfD.30 However, the application of
pharmacokinetic models for a route-to-route extrapolation of the inhalation studies expands the
database of studies suitable for RfD calculation.
As discussed in Section 5.1.1, based on evidence that neurological effects were
associated with lower tetrachloroethylene concentrations, neurotoxicity is selected as the critical
noncancer health effect of tetrachloroethylene. The nervous system is an expected target
following oral exposure, because tetrachloroethylene and metabolites produced from inhalation
exposures will also reach the target tissue via oral exposure. In addition, other organ systems
such as the liver and kidney are common targets associated with both inhalation and oral routes
of exposure, which supports the use of route extrapolation to compare PODs for oral and
inhalation exposure. In addition, differences in first-pass metabolism between oral and
inhalation exposures can be adequately accounted for by the PBPK model. For these reasons,
the inhalation neurotoxicity studies used in deriving the candidate RfCs are chosen for deriving
candidate RfDs.
5.2.2. Additional Analyses: Route-to-Route Extrapolation Using PBPK Modeling
The present analysis defines a POD using the traditional NOAEL/LOAEL approach. As
discussed in Section 5.1.2, dose-response modeling was not feasible with these studies. This
assessment has attempted to expand the database for derivation of an RfD using relevant
inhalation data and route-to-route extrapolation with the aid of a PBPK model (refer to
Section 3.5). Several factors support the use of route-to-route extrapolation for
tetrachloroethylene. Tetrachloroethylene has been shown to be rapidly and well absorbed by
both the oral and inhalation routes of exposure (ATSDR, 1997b). Additionally, the metabolic
pathways and kinetics of excretion with oral exposure are similar to those of inhalation exposure
(ATSDR, 1997b). Furthermore, the data for oral administration indicate a pattern of effects
similar to that of inhalation exposure. PBPK modeling was also used with suitable studies in
animals in order to extrapolate to human equivalent doses (HEDs). It is not clear if the
noncancer effects observed in humans are the result of tetrachloroethylene itself and/or one or
more metabolites. However, tetrachloroethylene in the blood can safely be presumed to be a step
30 The RfD is expressed in units of milligrams per kilogram of 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.
5-28
-------
in the toxicity pathway. Therefore, area under the curve (AUC) of blood tetrachloroethylene
concentration derived from PBPK modeling is considered the best surrogate for an internal dose.
The use of blood tetrachloroethylene provides some attempt to account for breathing rates and to
adjust for processes related to tetrachloroethylene toxicokinetics, and it is assumed to better
reflect tetrachloroethylene toxicokinetics than use of default methodologies. Moreover, based on
the results of the harmonized PBPK model (Chiu and Ginsberg, 2011), the sensitivity to the
choice of dose metric for route-to-route extrapolation is low, with alternative dose metrics such
as GSH metabolism, oxidative metabolism, or trichloroacetic acid (TCA) in blood giving route-
to-route conversions within 1.4-fold of the conversion based on tetrachloroethylene in blood.
Importantly, the PBPK model accounts for the potential first-pass effect of liver metabolism
from oral exposure, which was found to be minimal.
The harmonized PBPK model of Chiu and Ginsberg (2011) was used to derive the
continuous oral dose (i.e., in mg/kg-day) that would result in the same tetrachloroethylene in
blood AUC as that following a continuous inhalation exposure from the two studies (Echeverria
etal., 1995; Cavalleri etal., 1994. The route-to-route extrapolation starts with the estimation of
the average venous blood tetrachloroethylene AUC resulting from continuous inhalation
exposure at the adjusted LOAELs from the neurological endpoints in the two studies (Echeverria
et al., 1995; Cavalleri etal., 1994:). The venous blood tetrachloroethylene AUC at steady state
resulting from continuous exposure to these tetrachloroethylene concentrations was estimated to
range from 4.5 to 17 mg-hr/L-day, according to the Chiu and Ginsberg (2011) harmonized
model. While the model utilizes data from some healthy adult volunteers, it cannot be
considered to address pharmacokinetic variation in the full human population. The oral exposure
scenario was also modeled as continuous (i.e., a constant oral dose rate in mg/kg-day), because at
these exposure levels, the AUC of tetrachloroethylene in blood is insensitive to the exposure
pattern. The route-to-route extrapolation oral ingestion values at the LOAELs were
2.6 mg/kg-day for Cavalleri et al. (1994) and 9.7 mg/kg-day for Echeverria et al. (1995). The
results are presented in Table 5-8.
5.2.3. Reference Dose (RfD) Derivation, Including Application of Uncertainty Factors
The RfD is the midpoint of the range of candidate reference values from two principal
studies. Candidate RfDs for tetrachloroethylene were derived by dividing the route-to-route
extrapolated PODs of 2.6 mg/kg-day (Cavalleri et al., 1994) and 9.7 mg/kg-day (Echeverria et
al., 1995) by a total UF of 1,000, comprised of 10 for interindividual variability, 10 for
extrapolation from a LOAEL to a NOAEL, and 10 for database uncertainty. The application of
UFs was similar to that for the different endpoints used to derive the candidate RfCs. The
application of uncertainty factors is based on EPA's A Review of the Reference Dose and
5-29
-------
Reference Concentration Processes [(U.S. EPA, 2002): Section 4.4.5], which address five areas
of uncertainty.
• An UF of 10 was applied for human variation in the effects that were chosen for the
derivation of the RfD. As indicated in the RfC discussion (Section 5.1.3), the studies
selected do not include evaluation of potential sensitive populations including children,
elderly, and immune-compromised individuals.
• An UF of 1 was applied to account for interspecies variability in extrapolation from
laboratory animals to humans because the studies and critical endpoints were from human
studies.
• An UF of 1 was applied for the use of data from subchronic study to assess potential
effects from chronic exposure because, as with the RfC derivation described in
Section 5.1.3 for the human studies, the PODs are based on studies involving chronic
exposure.
• An UF of 10 was applied for the extrapolation from a LOAEL to a NOAEL because the
PODs from the studies were LOAELs.
• An UF of 10 was applied to address the lack of data to adequately characterize the hazard
and dose response in the human population as was done for the derivation of the
inhalation RfC. The rationale is the same as described above for the RfC.
UFs for the different endpoints were applied similarly to that for the RfC. The PODs
from each neurological endpoint were derived from a route-to-route extrapolation using a PBPK
model to obtain oral exposure equivalents. A composite UF of 1,000 was applied to the PODs
for the critical endpoints. A summary for each endpoint can be found in Table 5-8 and
Figure 5-4.
In summary, candidate RfDs for tetrachloroethylene were developed through route-to-
route extrapolation from the PODs for the following endpoints from neurotoxicological studies
of occupational tetrachloroethylene exposure: color vision changes (Cavalleri etal., 1994): and
cognitive and reaction time changes (Echeverria et al., 1995). The oral exposure POD equivalent
to the continuous inhalation exposure NOAELs or LOAELs was estimated via PBPK modeling.
The resulting PODs were 2.6 mg/kg-day (Cavalleri et al., 1994) and 9.7 mg/kg-day (Echeverria
et al., 1995). The same composite UF of 1,000 that was used for the RfC derivation was applied
to each of these PODs. The candidate RfDs from these studies span a range from 2.6 x io~3 to
9.7 x 1Q~3 mg/kg-day. The RfD for tetrachloroethylene is 6 x 10~3 mg/kg-day, the midpoint of
this range rounded to one significant figure. This RfD is equivalent to a drinking water
concentration of 0.21 mg/L, assuming a body weight of 70 kg and a daily water consumption of
2L.
5-30
-------
Table 5-8. Application of uncertainty factors for neurological endpoints from the studies used to derive
candidate RfDs
Neurological endpoint
Oral
human equivalent
dose", mg/kg-day
(NOAEL/LOAEL)
Uncertainty factors (UFs)
Composite
UF
UFA
UFH
UFS
UFD
UFL
Candidate
RfD
mg/kg-day
Reference
Cognitive domain
Visual reproduction, pattern memory,
pattern recognition — adult, occupational
9.7
(LOAEL)
1,000
1
10
1
10
10
0.0097
Echeverria
et al. (1995)
Reaction time domain
Reaction time in pattern memory, adult,
occupational
9.7
(LOAEL)
1,000
1
10
1
10
10
0.0097
Echeverria
et al. (1995)
Visual function domain
Color confusion — adults, occupational
2.6
(LOAEL)
1,000
1
10
1
10
10
0.0026
Cavalleri
et al. (1994)
"Equivalent oral exposure from application of the PBPK model on the basis of equivalent AUC of blood tetrachloroethylene for humans.
-------
Echeverriaetal (1995)-Cognitive measures; adult, occupational-LOAEL
Echeverria et al (1995)- Reaction time measures; adult, occupational - LOAEL
Cavalleri et al (1994) - Color confusion; adult, occupational - LOAEL
0.0026mg/kg-day <11.1.1.1.1.1.1.1.1.1.1.1.tMMMMM^^^^lBJJJJJJJJJJJ* l.fi mg/kg-day
1E-03
1E-02
1E-01
IE-GO
1E+01
• Point of Departure
Q UFA - Interspecies; animalto human
• UFS - Subchronic to chronic exposure duration
H UFL - LOAEL to NOAEL
n UFH - Intra species; human variability
HUFD-Database
V Candidate Reference Dose
[PERC] mg/kg-day
Figure 5-4. Candidate reference dose values for exposure to tetrachloroethylene.
to
-------
5.2.4. Dose-response Analyses for Noncancer Effects Other Than Critical Effect of
Neurotoxicity
This section presents oral dose-response analyses for noncancer effects other than the
critical effect of neurotoxicity. The purpose of these analyses is twofold: (1) to provide a
quantitative characterization of the relative sensitivity of different organs/systems to
tetrachloroethylene, and (2), to provide information that may be useful for cumulative risk
assessment in which multiple chemicals have a common target organ/system other than the
central nervous system. Therefore, for each organ/system, "sample reference doses" (sRfDs) are
calculated based on the same methodology as is used for the critical effect of neurotoxicity.
These sRfDs are based on an evaluation of studies identified in Section 4.10 as suitable for dose-
response analysis.
The method of analysis is the same as that described above for neurotoxicity, using the
NOAEL/LOAEL approach. Benchmark dose modeling was not performed because these sample
RfDs are meant for comparison purposes only (across organs/tissues or across chemicals). HEDs
are derived using either (1) mg/kg-day dose adjusted for equivalent continuous exposure; or
(2) the PBPK model with an appropriate dose metric. For liver effects, the dose metric of liver
oxidative metabolism was used, based on the view that oxidative metabolites are involved in
tetrachloroethylene-induced liver effects. For kidney effects, while it is generally thought that
GSH conjugation metabolites are involved, the large uncertainty in estimates of human GSH
conjugation preclude use of that dose metric. Instead, the AUC of tetrachloroethylene in blood is
used as a surrogate. For the other non-cancer effects, the AUC of tetrachloroethylene in blood
was used as the preferred dose metric due to the lack of data on what the active toxic moeity(ies)
may be for those effects. In addition, the PBPK model is being used to perform route-to-route
extrapolation from inhalation to oral exposure, so both inhalation and oral studies are considered
together here. For each endpoint where PBPK modeling is used, the dose metric used to derive
the HED is the same as that used to derive the HEC. The HED is then treated as a POD to which
the following UFs may be applied:
• An UF of 10 was applied for human variation to all PODs. The rationale is the same as
described above for neurotoxicity. Furthermore, there is some indication that human
variability (at least for one endpoint) may be substantially more than that implied by the
default UF. Kidney toxicity is thought to be associated with metabolism of
tetrachloroethylene along the glutathione (GSH) conjugation pathway. As described in
Section 3.5, PBPK model predictions for GSH conjugation span a wide range that may be
due to uncertainty, variability, or both. Glutathione S-transferases (GSTs) are known to
be polymorphic in the human population, with some isoforms exhibiting a substantial
population of null phenotypes.
5-33
-------
• An UF of 3 was applied to the PODs from all rodent studies to account for interspecies
variability in extrapolation from laboratory animals to humans. The PODs from studies
of rats and mice are expressed as HEDs calculated using the PBPK model. Therefore, an
UF of 3 was applied for animal-to-human uncertainty to the PODs from these rodent
studies to account for potential pharmacodynamic differences. This factor was not
applied to PODs from human studies.
• An UF of 10 was applied to PODs of studies of subchronic or shorter duration to address
the potential for additional or more severe toxicity from chronic or lifetime exposure.
• An UF of 10 is applied when a LOAEL is used due to a lack of a NOAEL. This factor
was applied to the PODs of studies that identified a LOAEL but not a NOAEL.
• A database UF of 10 is applied to all PODs to address the lack of data to adequately
characterize the hazard and dose response. The rationale is the same as described above
for neurotoxicity.
5.2.4.1. Sample Reference Doses (RfDs) for Kidney Toxicity
As discussed in Section 4, numerous studies have reported adverse effects in the kidney
from tetrachloroethylene. Five studies reporting kidney toxicity were identified in Section 4.10
as suitable for dose-response analysis. The only human study was Mutti et al. (1992), which
reported statistically significant increases in RBP, P2u-globulin, and albumin in urine among
chronically exposed dry cleaners as compared to matched controls. In addition, for
seven different urinary markers, the prevalence of individuals with abnormal values
(>95th percentile of controls) was four- to fivefold greater in the exposed group. This study was
considered adequate to derive an sRfD. Of the rodent studies reporting nephrotoxicity, only
JISA (1993) identified a chronic NOAEL, with the other three rodent studies reporting
subchronic (Jonker et al.. 1996) or chronic LOAELs (NTP, 1986: NCL 1977).
Therefore, among the rodent studies, only JISA (1993), which reported effects in both
mice and rats, was carried forward to calculate sRfDs. Because all the studies are inhalation
studies, route-to-route extrapolation was performed using the PBPK model with the AUC of
tetrachloroethylene in venous blood dose metric. A summary of the extrapolated PODs and UFs
applied is in Table 5-9. The resulting sRfDs range from 0.007-0.03 mg/kg-day, based on
nuclear enlargement in the proximal tubules of chronically exposed mice and rats (JISA, 1993),
with a slightly lower sRfD of 0.005 mg/kg-day based on urinary markers of nephrotoxicity in
occupationally exposed humans (Mutti et al., 1992).
5-34
-------
5.2.4.2. Sample Reference Doses (RfDs) for Liver Toxicity
As discussed in Section 4, numerous studies have reported adverse effects in the liver
from tetrachloroethylene. Six studies reporting liver toxicity, none in humans, were identified in
Section 4.10 as suitable for dose-response analysis. Only JISA (1993) reported a chronic
NOAEL, and so was carried forward for derivation of an sRfD. However, it is unclear whether
the reported effect of angiectasis, or enlargement of the blood vessels, is related to the other liver
effects of tetrachloroethylene, which generally involve hepatocytes. Therefore, two other studies
were included at this stage, one of which reported a chronic LOAEL for liver degeneration and
necrosis (NTP, 1986), and the other of which reported a NOAEL for liver-weight increases after
6-week exposures (Buben and O'Flaherty, 1985). The remaining studies either only reported a
LOAEL (Jonker et al., 1996: Ki ell strand et al.. 1984). or reported a NOAEL for a very short
duration [14 days, Berman et al. (1995)], and were, therefore, not considered further.
Therefore, JISA (1993), NTP (1986), and (Buben and O'Flahertv, 1985) were used to
calculate sRfDs. In addition, PBPK modeling was applied using the liver oxidative metabolism
dose metric to derive the HEDs. A summary of the PODs and UFs applied is in Table 5-10. The
resulting sRfDs range from 0.01 mg/kg-day based on increased liver/body-weight ratios after 6-
week exposures (Buben and O'Flaherty, 1985) to 0.08 mg/kg-day based on liver effects after
chronic exposures (JISA, 1993; NTP, 1986). It should also be noted that in the chronic studies,
increased liver tumors were observed at the lowest doses tested. Therefore, under chronic
exposure conditions in this organ, liver cancers are likely to be more important than noncancer
effects in the liver.
5-35
-------
Table 5-9. Sample RfDs for kidney effects
Kidney endpoint (species)
Urinary markers of
nephrotoxicity (human)
Nuclear enlargement in
proximal tubules (rat)
Nuclear enlargement in
proximal tubules (mouse)
HEDa in
mg/kg-day
(LOAEL/
NOAEL)
5.4 (LOAEL)
9.5 (NOAEL)
2.2 (NOAEL)
Uncertainty factors (UFs)
Composite
UF
1,000
300
300
UFA
1
O
O
UFH
10
10
10
UFS
1
1
1
UFD
10
10
10
UFL
10
1
1
Sample
RfD
(mg/kg-day)
0.005
0.03
0.007
Reference
Mutti et al.
(1992)
JISA (1993)
JISA (1993)
""Calculated with PBPK model using the dose metric of AUC of tetrachloroethylene in venous blood.
Table 5-10. Sample RfDs for liver effects
Liver endpoint (species)
Increased angiectasis
(mouse)
Increased liver
degeneration/necrosis
(mouse)
Increased liver/body-
weight ratio (mouse)
HEDa in
mg/kg-day
(LOAEL/
NOAEL)
24.5 (NOAEL)
252 (LOAEL)
32 (NOAEL)
Uncertainty factors (UFs)
Composite
UF
300
3,000
3,000
UFA
3
3
3
UFH
10
10
10
UFS
1
1
10
UFD
10
10
10
UFL
1
10
1
Sample
RfD
(mg/kg-day)
0.08
0.08
0.01
Reference
JISA (1993)
NTP (1986)
Buben &
O'Flaherty
(1985)
aCalculated with PBPK model using the dose metric of liver oxidative metabolism.
-------
Table 5-11. Sample RfDs for immunological and hematological effects
Immunotoxicity/
hematotoxicity endpoint
(species)
Reduced RBC, hemoglobin;
increased WBC, lymphocytes,
IgE (human)
HEDa in
mg/kg-day
(LOAEL/
NOAEL)
6.8
(LOAEL)
Uncertainty factors (UFs)
Composite
UF
1,000
UFA
1
UFH
10
UFS
1
UFD
10
UFL
10
Sample
RfD
(mg/kg-day)
0.007
Reference
Emara et al.
(2010)
""Calculated with PBPK model using the dose metric of AUC of tetrachloroethylene in venous blood.
RBC = red blood cells; WBC = white blood cells.
Table 5-12. Sample RfDs for reproductive and developmental effects
Reproductive/developmental
endpoint (species)
Reduced sperm quality
(mouse)
HEDa in
mg/kg-day
(LOAEL/
NOAEL)
22
(NOAEL)
Uncertainty factors (UFs)
Composite
UF
1000
UFA
3
UFH
10
UFS
O
UFD
10
UFL
1
Sample
RfD
(mg/kg-day)
0.02
Reference
Beliles
et al. (1980)
aCalculated with PBPK model using the dose metric of AUC of tetrachloroethylene in venous blood.
-------
5.2.4.3. Sample Reference Doses (RfDs) for Immunotoxicity and Hematologic Toxicity
As discussed in Section 4, a number of studies have reported changes in hematologic or
immunologic parameters with tetrachloroethylene exposure. Two studies reporting hematologic
effects were identified in Section 4.10 as suitable for dose-response analysis. The human study
(Emara etal., 2010) reported changes in various standard hematological measures in subjects
with mean blood levels of 1.685 mg/L. Application of the PBPK model provides an estimated
HED of 6.8 mg/kg-day. This was treated as a chronic LOAEL, given the 7-year mean exposure
duration (>10% of lifespan) and is carried forward to calculate an sRfD. The other study (Marth,
1987) reported reversible hemolytic anemia in mice after 7 weeks drinking water exposure to
2-week-old mice for 7 weeks. Although Marth (1987) was not considered further, as
summarized in Section 5.1.4.3, it should be noted that the LOAEL identified was
low—0.05 mg/kg-day—and may be a cause for additional concern about hematologic effects.
Therefore, Emara et al. (2010) was used to calculate an sRfD. A summary of the POD
and UFs applied is in Table 5-11. The result is an sRfD of 0.007 mg/kg-day.
5.2.4.4. Sample Reference Doses (RfDs) for Reproductive and Developmental Toxicity
As discussed in Section 4, a number of studies have reported reproductive and
developmental effects from tetrachloroethylene exposure. Four studies, none in humans,
reporting reproductive or developmental effects, were identified in Section 4.10 as suitable for
dose-response analysis. All of these studies reported NOAELs. The developmental studies were
all of appropriate duration for detecting those effects. The reproductive study (Beliles et al.,
1980) was short term (5 days exposure) but was the only suitable study for reproductive toxicity.
The PBPK model does not include gestational, fetal, or neonate compartments, so none of the
inhalation studies could be converted to oral equivalents. However, the reproductive study was
performed in mature male mice, for which the PBPK model could be used.
Therefore, only Beliles et al. (1980) was used to calculate an sRfD. A summary of the
POD and UFs applied is in Table 5-12. The UF for subchronic-to-chronic extrapolation was not
used because the study period sufficiently covered the window of sperm production. The
resulting sRfD is 0.07 mg/kg-day for reduced sperm quality (Beliles et al., 1980).
5.2.4.5. Summary of Sample Reference Doses (RfDs) for Noncancer Endpoints Other Than
the Critical Effect
The lowest sRfDs for these noncancer endpoints are similar to the values calculated
based on the critical effect of neurotoxicity (refer to Figure 5-5): 0.005 mg/kg-day from Mutti
et al. (1992). and 0.007 mg/kg-day from both JISA (1993) and Emara et al. (2010). All of the
5-38
-------
other sRfDs are within about 10-fold of the RfD. This suggests that multiple effects may occur
at about the same exposures at which tetrachloroethylene begins to induce neurotoxicity. These
results also suggest that it is important to take into account effects from tetrachloroethylene other
than neurotoxicity when assessing the cumulative effects of multiple exposures.
5.2.5. Previous Oral Assessment
The previous RfD of 1 x 1CT2 mg/kg-day was posted on the IRIS database on March 1,
1988. It was based on a NOAEL of 14 mg/kg-day (Buben and O'Flahertv 1985). and a
composite UF of 1,000 (10 for extrapolation from rats to humans, 10 for human variation, and 10
for extrapolating to chronic exposure conditions).
5.2.6. Uncertainties in Oral Reference Dose
As presented above, the uncertainty factor approach was applied to oral equivalent PODs
extrapolated from inhalation LOAELs from two epidemiologic studies of neurological effects.
These studies were considered to be methodologically sound based on study quality attributes,
including study population selection, exposure measurement methods, and endpoint
measurement methods. Other strengths are that they are human studies of chronic duration,
obviating the need for extrapolation across species and exposure duration. Because of the
adequacy of the PBPK model (Chiu and Ginsberg, 2011) for extrapolating from inhalation to
oral exposures, there is little uncertainty in the use of inhalation studies for deriving the RfD. For
instance, the sensitivity to the choice of dose metric for route-to-route extrapolation is low, with
alternative dose metrics giving route-to-route conversions within 1.4-fold of the conversion
selected based on tetrachloroethylene in blood. Other uncertainties in deriving the RfD are the
same as those for the RfC, discussed above in Section 5.1.6.
5-39
-------
.
*
CMS
UFcomp=1000
I Echeverria et al.
UFcomp=1000
I Cavalleriet al. (1994) [H]
Kidney
UFcomp=1000
UFcomp=300
Muttiet al. (1992) [H]
0 JISA(1993) [R]
UFcomp=300
-0JISA(1993) I
Liver
UFcomp=300
UFcomp=3000
I NTP (1986) [M]
-0 Buben & O'Flaherty (1985) [
Immunological/Hematological
UFcomp=1000
I Emara et al. (2010) [H]
Reproductive/Developmental
UFcomp=300
-0 Bellies et al. (1980) [M]
10
D1
1
10
10
10
mg/kgDd
Figure 5-5. Comparison of candidate RfDs (filled squares) supporting the RfD (vertical line) and sample RfDs (open squares)
for effects other the critical effect (CNS toxicity).
Filled circles = study/endpoint LOAEL in terms of human equivalent dose. Open circles = study/endpoint NOAEL in terms of human equivalent dose.
Species in each study is shown in brackets after the reference (mouse: M; rat: R; human: H).
-------
5.3. CANCER DOSE-RESPONSE ASSESSMENT
The following dose-response assessment was developed following the Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005a). As discussed in Section 4.10.2,
tetrachloroethylene is characterized as "likely to be carcinogenic in humans by all routes of
exposure," based on suggestive epidemiologic evidence and conclusive evidence in mice and
rats. No available human studies of cancer were found to be suitable for dose-response
assessment. Therefore, the following dose-response assessment is based on data from rodent
bioassays. Because the MO As for tetrachloroethylene carcinogenicity are not known, the tumors
reported in rodent bioassays are considered relevant to humans, and a low-dose linear
extrapolation is used to estimate human cancer risk from rodent dose-response data.
5.3.1. Choice of Study/Data with Rationale and Justification
As discussed in Section 4, the several chronic exposure studies in rats and mice include
an oral gavage study in mice and female rats by the National Cancer Institute (NCI, 1977) and
two inhalation studies in mice and rats (JISA, 1993; NTP, 1986). These studies established that
the administration of tetrachloroethylene, either by ingestion or by inhalation to sexually mature
rats and mice, results in increased incidence of tumors. Mouse liver tumors (hepatocellular
adenomas and carcinomas) and rat mononuclear cell leukemia (MCL) were reported in both
sexes in two lifetime inhalation bioassays employing different rodent strains, and mouse liver
tumors were also reported in both sexes in an oral bioassay (NCI, 1977). Tumors reported in a
single inhalation bioassay include kidney and testicular interstitial cell tumors in male F344 rats
(NTP, 1986), brain gliomas in male and female F344 rats (NTP, 1986), and hemangiomas or
hemangiosarcomas in male Crj:BDFl mice (JISA, 1993).
This analysis considers all three bioassays but focuses primarily on the JISA (1993) study
results. The NCI (1977) oral gavage study in Osborne-Mendel rats was considered to be
inconclusive because of the high incidence of respiratory disease, and high mortality with
tetrachloroethylene exposure. Lesions indicative of pneumonia were observed in almost all rats
at necropsy. A high incidence of toxic nephropathy was evident in tetrachloroethylene-exposed
male and female rats. Early mortality was also observed in tetrachloroethylene-exposed animals;
50% of the high dose males and females had died by Weeks 44 and 66, respectively. Regarding
the NCI (1977) gavage study in mice, several issues contribute to judging the results to be less
useful for quantitative risk assessment than the inhalation studies. First, dosing lasted 78 weeks
rather than 104 weeks as in the inhalation studies. Thus, in making direct comparisons, it might
be expected that the observed tumor incidence in the NCI (1977) study would underestimate the
incidence associated with 104 weeks of exposure. Second, the dosing schedule was variable, and
5-41
-------
doses were increased by 100 mg/kg-day in the low-dose group and by 200 mg/kg-day in the
high-dose group after 11 weeks of study. Consequently, while time-weighted averages and
PBPK modeling provide means for estimating the effective level of exposure, the actual
correspondence of exposure with the observed effects is less clear. Further, mortality was
significantly increased in both treated groups over that of controls, suggesting that the maximum
tolerated dose had been exceeded. Therefore, dose-response modeling of the NCI (1977) rat and
mouse bioassay data was not conducted.
The JISA (1993) bioassay was used for dose-response modeling of rodent cancer
endpoints also observed with higher exposures in the earlier NTP (1986) bioassay. The lower
exposure of both mice and rats in the JISA bioassay and the use of three—rather than two—
exposure groups provides a stronger basis for deriving dose-response relationships for risk
assessment purposes, insofar as all other aspects of these studies can be considered comparable.
For mice, the lowest and mid-dose exposure concentrations in the JISA (1993) study were 10-
and twofold lower, respectively, than the lower exposure concentration (100 ppm) in the NTP
(1986) inhalation study. For rats, the low-exposure concentration in the JISA (1993) study was
fourfold lower than in the NTP study (200 ppm). The JISA (1993) bioassay was also used for
dose-response modeling of the increased hemangiomas and hemangiosarcomas primarily in
spleen, liver, skin, and adipose tissue of male mice because it was the only bioassay that reported
this tumor type. Therefore, for most endpoints, including liver tumors, MCL, and
hemangiosarcomas, the JISA (1993) study was used for dose-response modeling. The NTP
(1986) study was utilized for modeling the increased incidence in renal cancers, brain cancers,
and testicular tumors with treatment reported only in this bioassay. The sections below
summarize the rodent tumor findings and additional considerations for data set selection.
5.3.2. Dose-Response Data
5.3.2.1. Liver Tumors in Mice
All three bioassays showed increases in hepatocellular tumors in male and female mice.
Table 5-13 summarizes these incidence patterns. Because hepatic adenomas and carcinomas are
considered part of the same continuum of tumor development, and adenomas may be
differentiated from carcinomas only on the basis of size, this analysis emphasizes the combined
incidence of these two tumor types. Historical data from the Japan Bioassay Research Center
(JBRC), where the JISA (1993) study was conducted, indicate that the liver tumor incidences in
the control group were fairly typical for this laboratory (refer to Table 5-14). Specifically, the
5-42
-------
incidence in controls was 28% for males and 6% for females; the averages for the laboratory
were 23 and 2%, and the upper bounds were 42 and 8%, respectively, for carcinomas.31
The liver tumor results of the two inhalation studies are reasonably concordant for both
male and female mice when adjusted for background tumor incidence (refer to Figure 5-6). The
incidence among male mice in the JISA (1993) study did not follow a clearly monotonic pattern,
with a higher response in the lowest dose group than that in the next higher dose group.
However, when considering the degree of expected variability given the number of animals in
each dose group, this pattern appeared consistent with the overall supralinear dose-response
patterns for the male and female mice in both the NTP (1986) and JISA (1993) studies.
The NCI (1977) study, in addition to the dosing and duration limitations noted above,
only reported hepatocellular carcinomas but not adenomas. This was consistent with other NCI
study reports of that time. Because, as stated above, hepatic adenomas and carcinomas are
considered part of the same continuum of tumor development, the other two bioassays provide a
more complete evaluation of hepatocarcinogenesis associated with tetrachloroethylene exposure.
Combined historical incidence of adenomas or carcinomas was not available. Presumably the incidence of
carcinomas slightly underestimates the overall incidence of adenomas or carcinomas.
5-43
-------
Table 5-13. Tumor incidence in mice exposed to tetrachloroethylene
Bioassay
Doses/exposures
Administered
Continuous
equivalent
Sex
Body
weight"
(kg)
Survival-adjusted tumor
incidence1" (%)
Hepatocellular adenomas or carcinomas
NCI(1977)C
B6C3FJ mice
Gavage:
5d/wk,
78 wk
NTP (1986)
B6C3FJ mice
Inhalation:
6hr/d,
5d/wk,
104 wk
JISA (1993)
Crj:BDFl mice
inhalation:
6 hr/d,
5 d/wk,
104 wk
Vehicle control
450 mg/kg-day
900
Vehicle control
300 mg/kg-dayd
600
Oppm
100
200
Oppm
100
200
Oppm
10
50
250
Oppm
10
50
250
Oe mg/kg-day
332
663
Oe mg/kg-day
239
478
Oppm
18
36
Oppm
18
36
Oppm
1.8
9.0
45
Oppm
1.8
9.0
45
Male
Female
Male
Female
Male
Female
0.030
0.025
0.037
0.032
0.048
0.035
2/20
32/48
27/45
0/20
19/48
19/45
17/49
31/47
41/50
4/45
17/42
38/48
13/46
21/49
19/48
40/49
3/50
3/47
7/48
33/49
(10)
(67)
(60)
(0)
(40)
(42)
(35)
(70)
(82)
(9)
(40)
(79)
(28)
(43)
(40)
(82)
(6)
(6)
(15)
(67)
Hemangiosarcomas", liver or spleen
JISA (1993)
Oppm
10
50
250
Oppm
1.8
9.0
45
Male
0.048
4/46
2/49
7/48
11/49
(4)
(2)
(13)
(18)
Note: Data sets carried through dose-response modeling shown in bold.
aAverage body weight reached during adulthood.
bAnimals 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 not to have adequate time on study to develop tumors.
°No adenomas were reported in this study.
dGavage doses listed were increased after 11 weeks by 100 mg/kg-day in each low-dose group or by 200 mg/kg-day
in each high-dose group. Animals surviving the 78-week exposure period were observed until Week 90 study
termination. Lifetime average daily (administered) doses (LADDs) were calculated as follows:
LADD (mg/kg-day) = Cumulative administered dose (mg/kg)/(total days on study)
= {[(initial dose rate x 11 weeks) + (later dose rate x 67 weeks)]/90 weeks}
x 5/7 (days)
eThese tumors were reported as hemangioendotheliomas in the JISA (1993) report. The term has been updated to
hemangiosarcoma. Note that these incidences do not match those tabulated in Tables 11 and 12 of the JISA report
summary. The incidences reported here represent a tabulation of hemangioendotheliomas in liver or spleen from
the individual animal data provided in the JISA report.
5-44
-------
Table 5-14. 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
hemangiomaa
hemangiosarcomaa
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
hemangiomaa
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
aThe 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.
5-45
-------
OS-
Ul
.2 0.8 -
" 0.7-
C
0
^ 0.6 -
«
•O 0.5-
-2.
a> 0.4-
§03-
Q.
-------
Table 5-15. 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
weight8
(kg)
Su rvival-adj u sted
tumor incidence15
(%)
Mononuclear cell leukemia
NTP (1986)
F344/N rats
inhalation
6 hr/d,
5 d/wk,
104 wk
JISA (1993)
F344/DuCrj rats
inhalation
6 hr/d,
5 d/wk,
104 wk
0
200
400
0
200
400
0
50
200
600
0
50
200
600
0
36
71
0
36
71
0
8.9
36
110
0
8.9
36
110
Male
Female
Male
Female
0.44
0.32
0.45
0.3
28/50
37/48
37/50
18/50
30/50
29/50
11/50
14/50
22/50
27/50
10/50
17/50
16/50
19/50
(56)
(77)
(74)
(36)
(58)
(60)
(22)
(28)
(44)
(54)
(20)
(34)
(32)
(38)
Kidney: tubular cell adenoma or adenocarcinoma
NTP (1986)
0
200
400
0
36
71
Male
0.44
1/49
3/47
4/50
(2)
(6)
(8)
Brain gliomas
NTP (1986)
0
200
400
0
36
71
Male
0.44
1/50
0/48
4/50
(2)
(0)
(8)
Testicular interstitial cell tumors
NTP (1986)
0
200
400
0
36
71
Male
0.44
35/50
39/47
41/50
(70)
(83)
(82)
Note: Data sets carried through dose-response analysis shown in bold.
"Average body weight reached during adulthood.
bAnimals 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 (19861 and JISA (1993).
5-47
-------
Table 5-16. 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
147/1,149(12.8)
2/1,149(0.2)
2/1,149(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
147/1,048 (14.0)
1/1,048(0.1)
0/1,048 (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
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 (g.epa.gov.
The responses in the NTP (1986) study were approximately twofold higher than for the
corresponding groups in the JISA (1993) study in all groups, including controls. Control groups
for both laboratories were consistent with their respective historical controls (refer to Table 5-16
for the JISA historical controls). Like the hepatocellular tumor results in mice (refer to
Section 5.3.2.1), the MCL results from the NTP and JISA studies were plotted in terms of
additional risk versus administered concentration to evaluate relative increases in tumor
incidence (refer to Figure 5-7). The NTP and JISA studies are consistent for male rats at the
administered concentration of 200 ppm (36 ppm continuous equivalent) in terms of the relative
increases in tumors over background incidences. For female rats, the dose-response patterns are
less similar. A higher overall response is observed in the NTP study. However, the JISA female
rats have a steeper increase at the lowest exposure level (50 ppm administered concentration,
9 ppm continuous equivalent) than would be expected based on the NTP study, which did not
include that exposure level. Both studies suggest some degree of saturation of effects in the
range of exposures considered (refer to Figure 5-7).
5-48
-------
2"
V)
^03
^
c
O
5
•o
TJ0.2
S.
o
L
o
00 S
-A- NTP, 1986
-0-JISA, 1993
S"
^^~~° ™
^___~— ~~~~"~^
^^_^-^-^^"^
^^-~~~~~^
^^~-~~~~^
/8^1
// — — —- —
// ~~ — ~A
//
/
gy
/ male rats
/
Z 0.3
ro
c
O
^
^
"O 0.2
S.
Ol
c
a. 0.1
&
n n f
-i- NTP, 1986
^JISA, 1993
A - _
/ - ^
^^"
^
/
/ / female rats
//'
20 40 60 80 100
Continuous equivalent concentration (ppm)
20 40 60 80 100
Continuous equivalent concentration (ppm)
Figure 5-7. Rat mononuclear cell leukemia responses (minus control) in
two chronic bioassays (refer to Table 5-15), plotted against continuous
equivalent exposure (ppm) for (a) male and (b) female rats.
Overall, the NTP and JISA studies show concordant MCL responses for both male and
female F344 rats. F344 rats were used in both studies, so residual differences could be
attributable to the specific lines of animals used at each laboratory and to laboratory-specific
procedures. As discussed in Section 5.3.1, the JISA study rather than the NTP study was
selected for dose-response modeling because it provides data on tumor incidences at lower
exposure, and the use of three exposures provides a stronger basis for dose-response analyses.
5.3.2.4. Other Tumor Sites in Rats
Additional tumor findings in rats included a significant increase in the NTP bioassay of
two rare tumor types, kidney tumors in males, and brain gliomas in both sexes of exposed
F344/N rats. The NTP (1986) bioassay also reported increases in the rate of testicular interstitial
cell tumors, a tumor type of high incidence in unexposed male F344 rats. Table 5-15
summarizes the incidence data for these tumor sites.
The potential significance of the NTP brain tumor findings is supported by their relative
rarity (evidenced by a low historical control incidence) and earlier occurrence with increasing
tetrachloroethylene exposure, indicating an effect of exposure on latency. In males,
tetrachloroethylene-induced brain tumors were observed beginning at Week 88 compared with
Week 99 in controls. Female brain tumors were first observed at 75 weeks in
tetrachloroethylene-exposed animals compared with 104 weeks in control group females.
Additionally, the nervous system is known to be a target of tetrachloroethylene exposure in
humans and animals (refer to Sections 4.1 and 5.1.1). Therefore, although the overall incidences
5-49
-------
are low relative to other tumor sites, and the finding was not replicated in the JISA study, the
rarity of rat brain tumors in control animals and the additional data suggesting biological
plausibility support dose-response modeling of this tumor type.
The evidence for kidney tubule cell adenomas and adenocarcinomas differed slightly
between the two bioassays (refer to Table 5-15). The JISA study showed no apparent trend
among incidences compared with either concurrent or historical controls (refer to Table 5-16).
In contrast, the elevation in exposed male rats in the NTP study, while not statistically significant
when compared with concurrent controls, was significant when compared using a trend test with
the historical control rate for the same facility (p = 0.0002, Cochran-Armitage, two-sided trend
test). The investigators noted the relative rarity of these tumors, with incidences of 1/549 among
historical controls for the study facility, and of about 0.2% in 1968 untreated controls in the NTP
program overall. Further support for the significance of the kidney tumors comes from evidence
that the related chemical trichloroethylene induces this tumor type in humans and in male rats
(U.S. EPA, 201 Ib). Additional biological plausibility for this endpoint includes toxicokinetic
data that nephrotoxic and mutagenic metabolites are formed in the kidney following
tetrachloroethylene exposure. Therefore, although the overall incidences are low relative to
other tumor sites, the rarity of rat kidney tumors in control animals and the additional data
suggesting biological plausibility support dose-response modeling of this tumor type. The NTP
(1986) study was better suited for modeling because it had a stronger trend, and was, therefore,
selected for dose-response modeling.
The NTP (1986) study also reported an increase in the rate of testicular interstitial cell
tumors, a tumor type of high incidence in unexposed F344 rats. The reported incidences of
testicular interstitial cell tumors in male rates exposed to 0, 200, or 400 ppm tetrachloroethylene
were 36/50, 39/49, and 41/50, respectively. A higher incidence (47/50, or 92%) was observed in
control rats in the JISA (1993) study than in the NTP (1986) study. In the JISA study, exposure
to 0, 50, 200, or 600 ppm tetrachloroethylene resulted in incidences of 47/50, 46/50, 45/50, and
48/50, respectively. Support for the significance of the testicular interstitial cell tumors comes
from evidence that the related chemical trichloroethylene induces this tumor type in rats.
Trichloroethylene did not induce increases in testicular interstitial cell tumors in the F344 rat in a
bioassay with a reported incidence of 47/48 (98%) in the vehicle control. However, increases
were observed in male Marshall rats, in which the incidences were 16/46, 17/46, 21/33, and
32/39 in untreated, vehicle control, 500, or 1,000 mg/kg-day trichloroethylene, respectively.
Therefore, although the overall increases in incidence are low relative to other tumor sites, the
additional data suggesting biological plausibility support dose-response modeling of this tumor
type.
5-50
-------
5.3.3. Dose Adjustments and Extrapolation Methods
This section provides details of the dose-response modeling carried out for developing
cancer risk values. The steps include estimation of dose metrics using relevant PBPK modeling
[refer to Section 3; Chiu and Ginsberg (2011)], suitable adjustment to continuous daily exposures
from intermittent bioassay exposures, dose-response modeling in the range of observation,
interspecies extrapolation, extrapolation to low exposures, and route-to-extrapolation. An
overview of these steps is provided in Figure 5-8. The schematic also addresses route-to-route
extrapolation using the Chiu and Ginsberg (2011) PBPK model because after the slope factor is
expressed in terms of risk per unit of internal human dose, the PBPK model can be used to
estimate the risk per unit of oral or inhalation exposure, regardless of the route of administration
in the original study.
5.3.3.1. Estimation of Dose Metrics for Dose-Response Modeling
Several factors inform the criteria for selection of dose metrics in this assessment: the
association of the metric with the toxic moiety relevant to the endpoint under consideration, the
availability of data and models for estimating that metric, and whether the resulting estimate is
sufficiently robust. When PBPK modeling is used, it is generally preferable to use a single
model for estimating all the dose metrics for dose-response modeling.
5.3.3.1.1. Hepatocellular tumors
Several metabolites of tetrachloroethylene are carcinogenic in mice, and it is thought that
the hepatocarcinogenicity of the parent compound is mediated through the action of one or more
of its metabolites. Oxidative metabolism is thought to predominate in the liver, and TCA is the
major resultant urinary excretion product. As discussed in Section 3, TCA appears to be formed
from spontaneous decomposition of trichloroacetyl chloride, which is known to bind to
macromolecules. Dichloroacetic acid (DCA) may be formed from dechlorination of TCA, but
DCA produced from this pathway is likely to be rapidly metabolized in the liver and not detected
in blood or urine. DCA that has been detected in urine is thought to be the result of kidney-
specific p-lyase metabolism of the results of GSH conjugation of tetrachloroethylene, and DCA
produced from this pathway is presumed to not play a role in liver toxicity or cancer. The
potential role of GST conjugates of tetrachloroethylene in liver carcinogenicity, although
unknown, is presumed to be less important than the role of oxidative metabolites.
5-51
-------
Administered dose in
inhalation/oral animal
bioassay
I
Animal PBPK model
Lifetime average daily dose metric
Preferred Dose Metrics Alternative Dose Metrics
Rate of
liver
oxidation
AUCof
tetra-
chloro-
ethylene
in blood
AUCof
TCAin
blood
Rate of
kidney
GSH
conjuga-
tion
I
Fit dose response model
to observed response
POD in units of lifetime
average daily dose
metric
I
BMR-HPOD
Slope Factor in units of
risk/(lifetime average
daily dose metric)
If dose metric is
rate of oxidation or
of conjugation,
apply BW3/4 scaling.
Otherwise assume
equal AUCs.
Slope Factors as
risk/(Human Equivalent
lifetime daily dose metric)
I
Human PBPK model
Slope Factor or Unit Risk as risk/(Human
Equivalent continuous inhalation or oral
environmental exposure level)
Figure 5-8 Sequence of steps for extrapolating from tetrachloroethylene
bioassays in animals to human-equivalent exposures expected to be
associated with comparable cancer risk (combined interspecies and route-to-
route extrapolation).
Refer to Table 5-17 for units.
5-52
-------
The focus of most hypotheses with respect to contributors to tetrachloroethylene
hepatocarcinogenicity has been on TCA and, to a lesser extent, DCA. Data on the
hepatocarcinogenicity of TCA and DCA in rodents, alone and in combination, are summarized in
Tables 4-17, 4-18, and 4-19. TCA statistically significantly increased the incidence of liver
tumors in male and female B6C3Fi mice exposed via drinking water for 52-104 weeks
(DeAngelo et al.. 2008: Bull et al.. 2002: Pereira, 1996: Bulletal.. 1990: Herren-Freund et al..
1987). Incidence of tumors increased with increasing TCA concentrations (DeAngelo et al.,
2008: Bull et al.. 2002: Pereira, 1996: Bulletal.. 1990). The development of tumors in animals
exposed to TCA progressed rapidly, as evidenced by significant numbers of tumors in less-than-
lifetime studies of 82 weeks or less. The tetrachloroethylene metabolite DCA also causes liver
cancer in mice (DeAngelo et al., 1999: Daniel et al., 1992: Bull et al., 1990: Herren-Freund et al.,
1987). Additionally, DCA and TCA are hepatocarcinogenic in mice when coadministered in the
drinking water for 52 weeks (Bull et al., 2002). Treatment-related liver tumors were observed in
male F344/N rats exposed via drinking water to DCA (DeAngelo et al., 1996) but not TCA
(DeAngelo et al., 1997) for 60 or 104 weeks. However, the extent to which DCA is available to
the liver following tetrachloroethylene exposure is unclear, because it is thought to be formed in
the kidney following p-lyase processing of TCVC and may be largely excreted in urine without
circulating systemically. The carcinogenicity of TCA and DCA has not been evaluated in female
rats or in other species of experimental animals.
Data on tumor phenotype support the view that TCA may not be the sole tumorigenic
metabolite of tetrachloroethylene but also do not provide definitive evidence testing any
particular hypothesis. For instance, liver tumor genotypes (e.g., with regard to H-ras codon
61 mutation) and phenotypes (e.g., with regard to c-Jun staining) appear to differ among tumors
induced by TCA, DCA, the combination of TCA and DCA, and the structurally related
compound trichloroethylene (Bull et al., 2002). Bull et al. (2002) suggest that for
trichloroethylene, the data are not consistent with the hypothesis that TCA is the sole active
moiety, but a similar experiment has not been conducted for tetrachloroethylene. However, by
analogy, it is possible that TCA and DCA, in combination with each other (and with other
reactive intermediates produced during the oxidative metabolism of tetrachloroethylene) may
contribute to the production of liver tumors. This appears to be the case for noncancer effects, as
the spectrum of endpoints caused by tetrachloroethylene includes effects broader than that
produced by TCA, and including fatty degeneration, focal necrosis, and regenerative repair,
some of which may play a role in liver carcinogenesis (refer to Section 4.3.5).
The hepatocarcinogenic potencies of TCA and tetrachloroethylene have not been directly
compared in a single rodent bioassay. Appendix C presents a comparative quantitative analysis
of the carcinogenicity of TCA (including that predicted using PBPK modeling to be produced
5-53
-------
from tetrachloroethylene) with the carcinogenicity of tetrachloroethylene. Statistically, this
analysis did not reject the hypothesis of equivalent carcinogenic potencies of TCA and the
internal dose of TCA resulting from tetrachloroethylene exposure. However, the analysis also
concluded that a contribution of TCA of as little as 12% could be not ruled out. In addition,
several factors, including the much higher control incidence of liver tumors and the relatively
high body weights of the animals in the TCA bioassay, limit the direct comparability of the
tetrachloroethylene and TCA bioassay data. Therefore, this analysis is only of limited utility in
elucidating the contribution of TCA to tetrachloroethylene hepatocarcinogenic potency.
In consideration of these uncertainties, total rate of oxidative metabolism in the liver is
the most relevant dose-metric for tetrachloroethylene-induced liver toxicity. AUC for TCA in
the liver is also presented as a plausible alternative dose metric. The PBPK-derived estimates of
liver total oxidative metabolism and TCA AUC corresponding to the JISA bioassay exposures
for male and female mice are provided in Table 5-17.
Table 5-17. Summary of PBPK-derived dose metric estimates used for
dose-response analysis of rodent tumor data
Study
group
Mice,
JISA
(1993)
Rats,
JISA
(1993)
Rats,
NTP
(1986)
Administered
concentration
(ppm)
0
10
50
250
0
50
200
600
0
200
400
Liver total oxidative
metabolism
(mg/kg%-daya)
Males
0
2.25
8.25
33.6
Females
0
2.13
7.75
31.6
Not used
Not used
Tetrachloroethylene
AUC in blood
(mg-hr/L-db)
Males
0
4.11
22.3
116
0
20.0
80.9
247
0
81.0
164
Females
0
4.18
22.6
117
0
20.1
81.4
248
0
81.3
164
TCA AUC in liver
(mg-hr/L-dc)
Males
0
78.5
280
1120
Females
0
77.0
111
1090
Not used
Not used
Total GSH
metabolism
(mg/kg3/4-dayd)
Males
Females
Not used
Not used
0
0.303
0.615
Not used
aPrimary dose metric for mouse hepatocellular tumors.
bPrimary dose metric for mouse hemangiomas or hemangiosarcomas, rat MCLs, rat kidney tumors, rat brain
gliomas, and rat testicular tumors.
Alternative dose metric for mouse hepatocellular tumors.
Alternative dose metric for rat kidney tumors.
5-54
-------
5.3.3.1.2. Mononuclear Cell Leukemia
Mononuclear cell leukemia has been observed in rats following exposure to
tetrachloroethylene. Regarding the metabolites that potentially contribute to MCL development,
a role for GSH-derived intermediates was posited based on findings for the related compound
trichloroethylene in bovine species. However ^-(l^^-trichloroviny^-Z-cysteine (TCVC), a
GSH-derived metabolite of tetrachloroethylene, induced no kidney or bone marrow effects when
administered to two calves as a single dose (Locket al., 1996). Aside from this evaluation of
bone marrow toxicity of TCVC in the juvenile cow, a species of unknown sensitivity to
tetrachloroethylene-induced leukemia, other studies aimed at elucidating the active metabolites
contributing to leukemic effects have not been reported. In particular, no such studies are
available in the F344 rat, the species and strain in which leukemic effects have been consistently
observed in both sexes. It is thus concluded that the specific active moiety(ies) by which
tetrachloroethylene induces this type of tumor are not known.
In summary, because considerable uncertainty surrounds the identification of the
causative chemical species, AUC of the parent compound in the blood is considered a viable
dose metric for MCL and has the advantage of being a more proximal dose than administered
dose. The estimates of tetrachloroethylene AUC in blood corresponding to the JISA bioassay
exposures for male and female rats are provided in Table 5-17.
5.3.3.1.3. Kidney tumors
Tetrachloroethylene causes tubular toxicity in mice and rats and is associated with small
increases in the incidences of kidney tumors reported in multiple strains of rats (JISA, 1993;
NTP, 1986). These effects, including kidney cancer, are thought to be associated with
tetrachloroethylene metabolism by GSH conjugation, based on the production in the kidney of
nephrotoxic and genotoxic metabolites from this pathway (Lash and Parker, 2001). As noted in
Section 3, the PBPK model by Chiu and Ginsberg (2011) allows calculation of this dose metric.
GSH conjugation occurs in the kidney as well as in the liver from where the metabolic products
may be transported to the kidney. Therefore, the most appropriate dose metric for kidney
toxicity would be the total rate of metabolism of tetrachloroethylene via the GSH conjugation
pathway.
However, overall, the estimates of GSH conjugation in Chiu and Ginsberg (2011) were
highly uncertain and/or variable, and to a very different extent across species (also refer to
Section 3). Uncertainty in this estimate was the least, roughly twofold, in rats. In mice, the
range of estimates based on the different optimization runs was about 10-fold. In the human, the
range of predicted estimates spanned several orders of magnitude. In particular, two local
maxima were observed for the posterior modes, each of which the fit to the data was good and
5-55
-------
substantially similar. However, the model predictions corresponding to each estimate differed by
3,000-fold. It was not clear as to whether this 3,000-fold spread represented uncertainty or
variability in the form of a bimodal distribution for human GSH conjugation or both (refer to
Section 3 for a discussion of plausible reasons for a multimodal distribution).
In view of this large uncertainty/variability, and the inability to differentiate uncertainty
from variability, it appears more prudent to use AUC of the parent compound in the blood as a
preferred dose metric for kidney toxicity. This has the advantage of being a more proximal dose
to the kidney than administered dose. Total rate of metabolism of tetrachloroethylene via the
GSH conjugation pathway is also used as an alternative dose metric. PBPK-derived estimates of
tetrachloroethylene AUC in blood and total GSH metabolism corresponding to the JISA (1993)
male rat exposures are provided in Table 5-17.
5.3.3.1.4. Other dose metrics
No data are available concerning the metabolites that may contribute to the induction of
other rodent tumor types, including hemangiosarcomas or hemangiomas in male mice, kidney
tumors and testicular interstitial cell tumors in male rats, or brain gliomas in male and female
rats. It is concluded that the specific active moiety(ies), mechanisms, or modes of action by
which tetrachloroethylene induces these rodent tumor are not known. Accordingly, AUC of
tetrachloroethylene in the blood was used for these tumors because it is more proximal to the
target tissues than administered dose (refer to Table 5-17 for dose estimates used for dose-
response modeling).
In addition, all tumor sites considered for modeling were also modeled using
administered inhalation concentration, for comparison purposes. These concentrations (in ppm)
were adjusted for continuous exposure by averaging the five 6-hour daily exposures over the full
week, by multiplying by 6 hours/24 hours x 5 days/7 days (0.179) to yield equivalent continuous
concentrations. Tables 5-13 and 5-15 provide these adjusted concentrations.
5.3.3.1.5. Uncertainties in PBPK modeling and dose metrics
A detailed discussion of uncertainties in the dosimetric estimates, derived using a PBPK
model that considered all the available tetrachloroethylene PK data in the literature, was
provided in Sections 3.5.1.2.2 and 3.5.1.2.3. A full Bayesian analysis of the
uncertainty/variability was not performed. Nonetheless, the range of posterior modes provided
for the various dose metrics in Section 3.5.1.2.2 gives an estimate of the range of uncertainty
associated with each dose metric, which, in turn, results in a range of human unit risk estimates
associated with each dose metric used for any given end point in Tables 5-18 and 5-20.
In particular, the predictions for GSH conjugation in humans were found to be highly
uncertain. In the rat, the ranges of chain-specific posterior modes for GSH conjugation spanned
5-56
-------
up to twofold, and in mice, up to 10-fold. However, in humans, the ranges spanned several
orders of magnitude, reflecting the two "clusters" of posterior modes with estimates of GSH
conjugation clearance differing by up to 3,000-fold. Tetrachloroethylene AUC was associated
with a twofold pharmacokinetic uncertainty/variability. The range in estimates of
tetrachloroethylene oxidation in humans was found to be largely dominated by a twofold
interindividual variability.
In terms of the selection of dose metric, tetrachloroethylene is metabolized to several
intermediates with carcinogenic potential. Although much data exist for TCA, they are
inadequate to support the conclusion that TCA alone is able to explain the hepatocarcinogenicity
associated with tetrachloroethylene exposure. Whether total oxidative metabolism, total GSH
metabolism, or tetrachloroethylene AUC in blood—either as measures of a precursor or
intermediate or as surrogates directly proportional to the toxic agent(s)—are adequate indicators
of potential risk is unclear. A role for the parent compound has not been ruled out nor is it clear
whether the specific active moiety(ies) are proportional to administered concentration.
5.3.3.2. Extrapolation Methods
5.3.3.2.1. Dose-response models and extrapolation to low doses
As discussed in Section 4.10.3, the available body of MO A information is not sufficient
to derive biologically based quantitative models for low-dose extrapolation. No key events in
the tumor development process for tetrachloroethylene have been identified that would
determine the overall dynamics of such a model nor are there experimental data specific to
tetrachloroethylene describing any of the underlying toxicodynamic processes, such as cell
replication rates.
The multistage model has been used by EPA in the majority of quantitative cancer
assessments, initially because of its parallelism to the multistage carcinogenic process. A benefit
of the multistage model is its flexibility in fitting a broad array of dose-response patterns,
including allowing linearity at low dose. Occasionally, the multistage model does not fit the
available data, in which case, alternate models should be considered. The related multistage-
Weibull model has been the preferred model when individual data are available for time-to-
tumor modeling, which incorporates more of the information about response than does the
simpler dichotomous response model. Use of this decision scheme has contributed to greater
consistency among cancer risk assessments.
5-57
-------
The multistage model is given by
P(d) = q0+(l- q0) x [l-exp(-Z, - 1 .•« q, * d}} (5-1)
where:
d = exposure level (including internal dose metric) and
qt = parameters estimated in fitting the model, q; > 0; n is degree of the model
The multistage model in BMDS [Benchmark Dose Software, version 2.1.1 (U.S. EPA,
2009)] was used initially to fit all data sets. Using the method of maximum likelihood, all
feasible orders of the multistage model up to the number of dose groups (n) less one were
evaluated for fit. Model fits with goodness-of-fit ^-values >0.05 are generally considered
acceptable, with good visual fit and evaluation of standardized residuals for the control group
and points near the benchmark dose (the dose corresponding to a predetermined increase above
control levels, or BMD) also important. Among the model fits satisfying these criteria, the most
parsimonious model fit was generally selected.
Two tumor sites with statistically significantly decreased time to tumor were noted: brain
gliomas in NTP male rats and MCL in the NTP female rats, especially for the most severe stage
of leukemia observed (Stage 3). The multistage-Weibull model, given by the following
equation, was also used to evaluate the importance of decreased time to tumor and intercurrent
mortality in interpreting these responses.
P(d,f) = q0 + (l- <7o) x [l-e*3P(-X/= i :nqt * d) x f] (5-2)
where:
d = exposure level (or dose metric)
t = time to observation of the tumor
qt, z = parameters estimated in fitting the model; q, > 0, z > 1; n is degree of the model
The multistage-Weibull model is the same as the multistage model when z = 0. MSW (U.S.
EPA, 2010) was used for all multistage-Weibull model fits.
Following dose-response modeling in the range of observation, the cancer risk values for
extrapolation to low doses were derived from the lower bound on the concentration (BMCL)
associated with a level of risk from the low end of the observed range, usually 10% extra risk.
Extra risk has been used consistently throughout EPA risk assessments and is given by
5-58
-------
Extra risk = [P(d) - P(0)] / [1 - P(0)] (5-3)
where:
P(d) = estimated response at exposure d and
P(0) = estimated response in the control group
The slope factor (risk per mg/kg-day for oral exposure, risk per dose metric unit for PBPK-
modeled dose metrics) and risk per unit concentration (risk per mg/L for drinking water
exposure, or per |ig/m3 for inhalation exposure) are estimated using linear extrapolation from the
PODs because of the lack of information supporting another extrapolation approach (U.S. EPA,
2005a), by dividing the risk level by its associated BMCL:
Risk/(unit of exposure) = Extra risk/BMCL. (5-4)
5.3.3.2.2. Uncertainties in low-dose extrapolation approach
The MOA is a key consideration in clarifying how risks should be estimated for low-dose
exposure. However, MOA data are lacking or limited for all candidate cancer endpoints for
tetrachloroethylene (i.e., rat MCL, brain, testicular and kidney tumors, mouse hepatocellular
tumors and hemangiosarcomas). When the MOA cannot be established, EPA uses a linear
approach to estimate low-exposure risk as outlined in EPA's Guidelines for Carcinogen Risk
Assessment (U.S. EPA, 2005a). The overall uncertainty in low-dose risk estimation could be
reduced to some degree if the MOA for tetrachloroethylene were known. However, even in such
a case, incorporation of MOA into dose-response modeling might not be straightforward and
might not significantly reduce the uncertainty about low-dose extrapolation. This is because in
addition to the MOA, other factors, such as human response variability, may influence the dose-
response function in humans.
A number of different biological motivations have been put forward to support functional
forms that might be used to estimate risks from low-dose exposure to carcinogens or other toxic
substances. For cancer, the most prominent class of models, including the multistage model used
in this assessment, treats tumorigenesis as a multievent 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.
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 NRC report on
5-59
-------
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 of
dose-response functions can be obtained using distributions of thresholds models: "In fact any
monotonic dose-response model, including the linearized multistage model, can be defined
solely in terms of a tolerance distribution without resorting to mechanistic arguments. These
considerations suggest that one must consider both the role of mode of action and the role of
response variability among humans in determining the likely shape of the dose-response
function" (p. 323).
The discussion from NRC (2006) emphasizes some key points in risk assessment.
Variability in the human population will have an important influence on the shapes of the dose-
response relationships for that population. This is distinct from the amount of variability that
may be observed in inbred animal strains. As noted in the NRC report, "One might expect these
individual tolerances to vary extensively in humans depending on genetics, coincident exposures,
nutritional status, and various other susceptibility factors..." (p. 320). Thus, if a distribution-of-
thresholds approach is considered for a carcinogen risk assessment, application would depend on
the ability of modeling to reflect the degree of variability in response in human populations. By
design, most cancer bioassays are conducted in inbred rodent strains; accordingly, the parameters
provided by curve fits of distribution-of-thresholds models to bioassay data would not be
predicted to reflect the dose-response patterns in diverse human populations. It is important to
note that the NRC text has no recommendation for an approach where a tolerance distribution
model for humans is estimated by a statistical fit to rodent bioassay data.
The question of whether a tolerance distribution model is indeed an appropriate basis for
a risk assessment also warrants consideration. Low-dose linearity can arise in other contexts
distinct from effects of population variability and may be directly appropriate to a MOA. Low-
dose linearity can also arise due to additivity of a chemical's effect on top of background
chemical exposures and biological processes. In the case of chemicals such as
tetrachloroethylene, basic biological data do not exist to support the appropriateness of an
individual threshold model above models having inherent low-dose linearity. However, if
5-60
-------
distribution of thresholds modeling was supported, it would need to be developed based on an
examination of predicted variability within the human population.
Given the current state of scientific knowledge about tetrachloroethylene carcinogenicity,
the straight line-based risk estimates presented above form the preferred recommendation for
estimating a plausible upper-bound estimate of potential human risks from tetrachloroethylene.
This approach is supported by both general scientific considerations, including those supporting
the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), as well as chemical-specific
findings. The former include the scientific principles articulated above—the expectation that a
chemical functions additively to background exposures, diseases, and processes; that variability
within the human population would broaden the dose-response curve and may eliminate
population thresholds if present; and that the approach provides consistency across assessments,
facilitating direct comparison of the derived risk values.
5.3.3.2.3. Extrapolation to human equivalent environmental exposure
For extrapolation of risk to humans, this assessment used two approaches that were
dependent on the relevant dose metric: the EPA RfC methodology (U.S. EPA, 1994), which
applies when chemical-specific pharmacokinetic data are lacking, and EPA's cross-species
scaling methodology (U.S. EPA, 1992). The discussions below include a consideration of
uncertainties inherent in each of these approaches.
5.3.3.2.3.1. Internal dose metrics
Because of the availability of PBPK modeling to estimate a plausible dose metric either
in terms of specific metabolites or metabolic pathways or blood concentration of the parent
compound in both laboratory rodents and humans, extrapolation to human equivalent
environmental exposure entailed the steps as shown in Figure 5-8. First, consistent with the
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), EPA's methodology for cross-
species scaling (U.S. EPA, 1992) was considered when toxicological equivalence for the relevant
tumor sites was addressed, in order to convert the slope factor to units of risk per unit of human
equivalent internal dose metric. Then the slope factor was converted to units of risk per unit of
human equivalent environmental exposure by using the relationship between continuous human
exposure and internal dose metric estimated via the human PBPK model. These last
two considerations are further described below.32
32 Typically, the POD would be expressed in terms of a human equivalent exposure. However, in this case, it is
expressed in terms of the internal dose metric. This is because the relationship between exposure and internal dose
may be nonlinear at the POD, even if the relationship between risk and internal dose is assumed to be linear below
the POD. Therefore, the slope factor is first expressed in terms of internal dose, reflecting the assumption of low-
dose linearity in internal dose. Then, provided the slope factor is applied at exposures well below the POD, where
the relationship between exposure and internal dose is linear, it can be converted to a risk per unit exposure.
5-61
-------
EPA's cross-species scaling methodology, grounded in general principles of allometric
variation of biologic processes, was used for describing toxicological equivalence because of the
extensive empirical evidence supporting it (U.S. EPA, 1992; Crump et al., 1989; Allen et al.,
1988). Briefly, in the absence of adequate information to the contrary, the methodology
determines toxicological equivalence across species through equal average lifetime
concentrations, or AUCs, of the carcinogen. One typical application of this methodology is to
oral exposures in mg/kg-day in the absence of pharmacokinetic or pharmacodynamic
information. However, the same principles apply to the parent compound and metabolites (U.S.
EPA. 1992).
For the orally administered dose, the correspondence of equal AUCs is equivalent to
considering the exposures in terms of mg/kg3/4-day and is achieved by multiplying animal
exposures by (BWanimai/BWhuman)1/4, based on the principle that clearance, on average, scales
allometrically according to BW 1/4 across species (U.S. EPA, 2005a). Note that this equivalence
across species entails the cross-species correspondence of internal doses in terms of AUCs or
mg/kg3/4-day, which is implicit in the frequent default case, i.e., oral carcinogens without
chemical-specific pharmacokinetic data. In other words, each time a carcinogen is scaled from
animals to humans on the basis of mg/kg3/4-day, an implicit assumption is that internal doses are
equipotent in terms of mg/kg3/4-day ("cross-species scaling"), not mg/kg-day ("body-weight
scaling").
Accordingly, when pharmacokinetic data are available that relate administered
concentration to enzymatically derived metabolites of the carcinogen, this methodology is still
applicable; internal doses, as a fraction of administered dose, should still tend to produce
equivalent effects when considered in terms of AUCs (when clearance of a specific metabolite is
specifically modeled) or mg/kg3/4-day (when rate of metabolism is calculated) because
metabolites are also subject to scale-affected clearance processes. The equivalence of
considering equal AUCs of a metabolite to scaling the rate of metabolism by BW3/4 can be easily
understood if one assumes clearance rates^br the metabolite scales allometrically adjusted by the
fraction metabolized. There is a wide body of empirical evidence that metabolic rates associated
with enzymatic processes scale with body weight to the % power (U.S. EPA, 2005a), or BW 1/4;
if one thinks of the scaling as applied to the administered dose (U.S. EPA, 1992). Furthermore,
when this scaling is applied to an internal dose expressed as a rate of production of metabolite(s),
it is applicable regardless of the route of exposure. As an example, in EPA's trichloroethylene
assessment, the human equivalent risk for liver and kidney effects was estimated using BW3/4
scaling of the daily rate of the lexicologically relevant metabolic pathway (U.S. EPA, 201 Ib).
As discussed earlier in this subsection, rates of liver oxidative metabolism and total GSH
metabolism are considered plausible dose metrics for the liver and kidney, respectively. In order
5-62
-------
to estimate equivalent toxic effects in humans using the cross-species scaling methodology,
3/4
tetrachloroethylene metabolized via either of these pathways was scaled using BW so that the
3/4
dose metric was expressed as mg/kg -day. As explained earlier, the AUC of TCA in the liver,
the predominant metabolite along the oxidative pathway, is also presented as a plausible dose
metric for liver cancer. No additional scaling was needed as the average concentration of TCA
so determined was assumed to be equipotent when applied continuously over a lifetime in either
species. Likewise, AUC of tetrachloroethylene, used as the preferred dose metric for MCL and
kidney tumors, was not scaled further to extrapolate to humans.
Note that the involvement of reactive metabolites cleared nonenzymatically through
which all other metabolites may follow has been hypothesized in many cases, and scaling by BW
as opposed to BW3/4 has been proposed to be more appropriate in such cases. However, scaling
by BW was not considered pertinent for tetrachloroethylene because the possible reactive
metabolites cleared nonenzymatically have not been identified and because the majority of the
metabolites formed are thought to be sufficiently stable to be cleared enzymatically.
In the last step of the extrapolation to risk per human equivalent exposure, the slope
factors in terms of internal dose metrics (associated with parent or metabolites) were converted
to slope factors or unit risks in terms of human equivalent environmental inhalation and oral
exposures using pharmacokinetic modeling. Refer to footnote c in Tables 5-18 and 5-21 for the
inhalation and oral conversion factors. For animals, the study-specific body weights were used
(refer to Tables 5-13 and 5-15), and for humans, the default of 70 kg was used.
In summary, an adjustment for cross-species scaling (BW3/4) was applied to address
toxicological equivalence of internal doses between each rodent species and humans for
two dose metrics, total liver oxidative metabolism and total GSH metabolism, consistent with the
Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a). It is assumed that, without data
to the contrary, equal risks result from equivalent constant exposures. While the true
correspondence of equipotent tetrachloroethylene exposures across species is unknown, the use
of BW3/4 scaling is expected neither to over- or underestimate human risk, based on allometry
(U.S. EPA. 1992).
5.3.3.2.3.2. Administered inhaled concentration as dose metric
For those sites for which pharmacokinetic-adjusted doses were not available or not
otherwise relevant, EPA's default RfC methodology was used (U.S. EPA, 1994).
Tetrachloroethylene is considered a Category 3 gas because it is water soluble and perfusion
limited, and it has systemic (extrarespiratory) effects. Because the ratio of blood/air partition
coefficients for the experimental animal species relative to humans is greater than or equal to one
(for F344 rats, 15.1/14.7= 1.03; forB6C3Fi mice, 18.6/14.7= 1.3), a default value of one was
5-63
-------
used for this ratio (U.S. EPA, 1994). Consequently, when administered inhalation
concentrations were used as the dose metric, the concentrations were considered equipotent
across species for extrapolating risk to humans. Therefore, no further extrapolation was
necessary, with the resulting PODs in the units of human equivalent environmental exposure
levels.
In summary, for MCL, hemangiomas or hemangiosarcomas, and brain, testicular, and
kidney tumors, tetrachloroethylene AUC in blood was judged to be more proximal than
administered tetrachloroethylene concentration to the adverse effect, and, therefore, more
relevant for estimating unit risks. Also based on allometry, average daily AUCs are expected to
be equipotent across species without any additional scaling involved. The true correspondence is
unknown, and risk may be higher or lower in humans than in rodents to an unknown degree.
5.3.4. Cancer Risk Values
Human cancer risk was assessed using four different sex-species animal data sets and a
PBPK model for interspecies and route-to-route extrapolation. In all cases, linear extrapolation
from the PODs was carried out because of the lack of information supporting another
extrapolation approach (U.S. EPA, 2005a). For each data set, multistage modeling (refer to
Section 5.3.3.2.1) using preferred and alternative (if available) PBPK model-based dose metrics
was conducted, in addition to multistage modeling using administered concentration.
In addition, the NRC (2010) peer review recommended more extensive quantitative
evaluation of the uncertainty due to different forms of dose-response models. Moreover, NRC
(2010) agreed that for several datasets, the multistage model does not fit the data at lower doses,
noting evidence of supralinearity in the underlying dose-response relationship. NRC (2010) also
noted that in such cases, low-dose linear extrapolation is not conservative, and the external
review draft of the Toxicological Review did not present the full ranges of variation and
uncertainty in relation to model choice. Therefore, for the JISA (1993) datasets, additional
analyses were conducted using administered concentration and the range of dichotomous models
included in BMDS. In addition to the multistage model, these include the gamma, Weibull, log-
logistic, log-probit, dichotomous Hill, probit, and logistic models. For the dichotomous Hill
model, the slope was fixed at 1, making it equivalent to a Michaelis-Menten model. Statistically,
a simpler model (fewer free parameters) was needed so that goodness-of-fit statistics could be
derived given the number of dose groups (three exposed plus one control). Biologically, the
Michaelis-Menten model is a natural choice for saturable biological processes, such as enzyme
kinetics, that are not accounted for in the selected dose metrics. Hereafter, the dichotomous Hill
model with slope fixed at 1 is referred to as the Michaelis-Menten model.
5-64
-------
The results of the suite of models were evaluated for goodness-of-fit. For datasets
exhibiting supralinearity, models that led to both a better fit to the supralinear shape and a stable
BMDL were considered for further application using PBPK model-based dose metrics.
The sections below provide the results of the dose-response modeling using the male and
female mouse and rat data from the JISA (1993) inhalation bioassay and male and female rat
data from the NTP (1986) inhalation bioassay. Route-to-route extrapolation for estimating
human cancer risk via oral exposure to tetrachloroethylene is then presented. Finally,
quantitative and qualitative uncertainties underlying the risk estimation process are discussed.
5.3.4.1. Dose-Response Modeling Results
5.3.4.1.1. Hepatocellular tumors, male mice
In accordance with standard practice in the absence of MO A data supporting a particular
dose-response model form, multistage modeling of the JISA bioassay data was carried out, using
the preferred and alternative dose metrics of total liver oxidative metabolism and TCA AUC in
liver. Modeling for both dose metrics generated fits for one-, two-, and three-stage models
(details in Appendix D). All model fits had adequate goodness-of-fit/^-values (p > 0.05), and
overall adequate fit given the nonmonotonicity in the observed dose-response range (with
standardized residuals within ± 2). There was no statistical improvement (by likelihood ratio
test) in adding higher order terms to the first-order term, and a one-stage model was selected
(refer to Figure 5-9 for the fit using total oxidative metabolism).
Extrapolation to humans using total oxidative metabolism led to a BMDio of 2.9, and its
lower bound benchmark dose (BMDLio) was 1.4-fold lower at 2.1 mg/kg3/4-day liver oxidative
metabolism (refer to Figure 5-9). Linear extrapolation from the POD to low internal dose,
followed by conversion to human exposures, led to a human equivalent unit risk of 1.8 x 10
per ppm.
Extrapolation to humans using TCA AUC in liver led to a human equivalent internal dose
POD (BMCLio) of 69 mg-hr/L-day TCA in blood. The corresponding central tendency estimate
was approximately 1.5-fold higher, at 97 mg-hr/L-day. Linear extrapolation from the POD to
low internal dose, followed by conversion to human exposures, led to a human equivalent unit
risk of 1.5 x 10 per ppm, slightly lower than the estimate using total liver oxidative
metabolism.
Dose-response modeling of the male mouse liver tumor data using administered exposure
fit the data points similarly to when using total oxidative metabolism or TCA AUC in liver
(details in Appendix D). The result was directly interpretable as a human equivalent POD
(BMCLio), at 2.7 ppm tetrachloroethylene in air. The corresponding central tendency estimate
5-65
-------
was nearly twofold higher, at 3.9 ppm. Linear extrapolation from this POD led to a human
equivalent unit risk of 37 x 10 per ppm, more than an order of magnitude higher than using
either PBPK-estimated dose metric.
The NRC (2010) peer review recommended more extensive quantitative evaluation of the
uncertainty due to different forms of dose-response models. The analysis was conducted using
administered concentration and the range of dichotomous models included in BMDS. Among
the models fitted, five models fit worse than the multistage (gamma, Weibull, log-logistic, log-
probit, and Michaelis-Menten), and two models fit better than the multistage (probit and
logistic). However, the multistage model had the lowest residual for the control group,
indicating that the alternative models were no better than the multistage model in addressing the
supralinear shape in this data set. Nonetheless, the estimated BMCLi0s from the better-fitting
models were less than twofold different than that using the multistage model.
Therefore, due to the limited sensitivity to the selection of dose-response models and the
finding that none of the alternative models was clearly superior to the standard multistage model
for addressing this data set's supralinearity at the lower doses, the multistage model results were
carried forward to support cancer risk estimates (refer to Table 5-18). Due to the data supporting
oxidative metabolism as being involved in hepatocellular tumors, the estimates carried forward
were those using total oxidative metabolism as the dose metric (preferred), and those using TCA
AUC in liver as the dose metric (alternative). The remaining analyses (refer to Tables 5-19 and
5-20) of administered concentration using multistage and other dose-response models are
retained only to better characterize the range of results from different dose-response models.
5.3.4.1.2. Hepatocellular tumors, female mice
As was done for the male mouse hepatocellular tumors, in accordance with standard
practice in the absence of MO A data supporting a particular dose-response model form,
multistage modeling of the JISA bioassay data was carried out, using the preferred and
alternative dose metrics of total liver oxidative metabolism and TCA AUC in liver. Modeling
for both dose metrics included one-, two-, and three-stage models. Adequate fits were obtained
with all three models, with adequate goodness-of-fit ^-values (p > 0.05), and overall adequate
visual fit (refer to details in Appendix D). The second-order term led to a statistically significant
improvement in fit, but there was no statistical improvement with the third-order term, as it was
estimated to be zero. Therefore a two-stage model was selected for both dose metrics (refer to
Figure 5-10 for the fit using total oxidative metabolism).
5-66
-------
Mu Itistage Can cer Model with 0.95 Confidence Level
I
Multistage Can cer
Linear extrapolation
40
45
Figure 5-9. Dose-response modeling of male mouse hepatocellular tumors
associated with inhalation exposure to tetrachloroethylene, in terms of liver
total oxidative metabolites; response data from JISA (1993).
Details in Appendix D.
Multistage Can cer Model with 0.95 Confidence Level
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Linear extrapolation
BMDL
10
15
dose
20
25
30
Figure 5-10. Dose-response modeling of female mouse hepatocellular tumors
associated with inhalation exposure to tetrachloroethylene, in terms of liver
total oxidative metabolites; response data from JISA (1993).
Details in Appendix D.
5-67
-------
Table 5-18. Human equivalent candidate unit risks, derived using PBPK-
derived dose metrics and multistage model; tumor incidence data from JISA
(1993) and NTP (1986)
Study Group
Tumor type
(multistage model with
all dose groups unless
otherwise specified)
Human Equivalents
PODa, in internal dose units
SFxlO'3
/internal
dose
unitb
Candidate
IlIRxlO'3
/ppm (PBPK
range)0
Primary dose metrics
Male mice
JISA (1993)
Female mice
JISA (19931
Male rats
JISA (1993)
Female rats
JISA (1993)
Female and male
rats JISA (1993)
Male rats
NTP (1986)
Hepatocellular adenomas
or carcinomas
Hemangiomas,
hemangiosarcomas,
Hepatocellular adenomas
or carcinomas
MCL
MCL (Michaelis-
Menten)
MCL
MCL (control and low
dose groups only)
MCL (Michaelis-
Menten)
Kidney tumors
Brain gliomas
Testicular interstitial cell
tumors
MCL
Total risk for any of
above four tumor types
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
2.9
2.1
63
34
8.4
4.0
46
30
20
5.0
136
61
11
5.2
17
3.0
250
110
410
170
30
14
28
15
13
8.1
Total liver oxidative
metabolism, mg/kg°75-d
PCEAUC in blood,
mg-hr/L-d
Total liver oxidative
metabolism, mg/kg°75-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
49
2.9
25
3.4
20
1.6
19
33
0.90
0.62
7.1
7.2
12
1.8
(1.6-1.8)
5.9
(5.9-6.9)
0.90
(0.84-0.93)
6.8
(6.8-8.0)
40
(40-47)
3.3
(3.3-3.9)
39
(39-45)
68
(67-71)
1.8
(1.8-2.1)
1.2
(1.2-1.4)
14
(14-17)
15
(14-17)
25
(25-29)
Alternate Dose Metrics
Male mice
JISA (1993)
Female mice
JISA (1993)
Male rats
NTP (1986)
Hepatocellular aden-
omas or carcinomas
Hepatocellular aden-
omas or carcinomas
Kidney tumors
BMD10
BMDL10
BMD10
BMDL10
BMD05
BMDL05
97
69
292
141
0.46
0.21
TCA AUC in liver,
mg-hr/L-d
TCA AUC in liver,
mg-hr/L-d
Total GSH metabolism,
mg/kg°75-d
1.5
0.72
243
1.5
(1.4-1.5)
0.72
(0.68-0.74)
100
(0.047-110)
5-68
-------
Table 5-18. Human equivalent candidate unit risks, derived using PBPK-
derived dose metrics and multistage model; tumor incidence data from JISA
(1993) and NTP (1986) (continued)
SF = Slope Factor; IUR = Inhalation Unit Risk; MCL= Mononuclear cell leukemias.
aPODs were estimated at the indicated BMRs in terms of extra risk; i.e., BMDL10 = lower bound for the level of the
internal dose metric associated with 10% extra risk. Dose metric units are in the first column and include cross-
species scaling to a human equivalent internal dose metric. Refer to Appendix D for dose-response modeling
details.
bSlope Factor = BMR/BMDLBMR in units of risk per dose metric unit (as given in the first column).
Inhalation unit risk (IUR) is given by the product of the slope factor in units of risk per dose metric unit and an
inhalation dose metric conversion factor (DMCFppm): IUR = BMR/BMDLBMR x DMCFppm, where the DMCFppm is
derived from the PBPK model. The DMCFppm for each dose metric is shown below:
Dose metric
Total liver oxidative metabolism
Tetrachloroethylene blood AUC
TCA AUC in liver
Total GSH metabolism
DMCFppm
Overall posterior mode
0.0363
2.03
1.02
0.428
Range of posterior modes
0.0339-0.0372
2.01-2.36
0.956-1.04
0.00019-0.44
Values in bold correspond to using the overall posterior mode and are carried forward for consideration in the
derivation of the IUR. The difference between the overall and alternative posterior modes is negligible (relative to
other uncertainties) except for the Total GSH metabolism dose metric.
5-69
-------
Table 5-19. Dose-response summary and candidate unit risk estimates using
continuous equivalent administered tetrachloroethylene levels as dose metric,
from NTP (1986) and JISA (1993)
Study group
Male mice
JISA (19931
Female mice
JISA (1993)
Male rats
JISA (19931
Female rats
JISA (19931
Female and male
rats
JISA (19931
Male rats
NTP (19861
Tumor type
(multistage model and all
dose groups unless
otherwise specified)
Hepatocellular adenomas
or carcinomas
Hemangiomas or
hemangiosarcomas
Overall risk of either
tumor type above0
Hepatocellular adenomas
or carcinomas
MCL
MCL (Michaelis-Menten
model)
MCL
MCL (control and low-
dose groups only)
MCL
MCL (Michaelis-Menten
model)
Kidney tumors
Brain gliomas
Testicular interstitial cell
tumors
Mononuclear cell
leukemia
Overall risk for any of
above four tumor types0
POD (ppm)
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
3.9
2.7
24
13
3.3
2.4
5.0
3.8
21
13
8.6
2.2
60
27
4.9
2.3
32
21
7.7
1.4
110
50
180
73
13
6.1
12
6.5
5.7
3.5
Candidate
Unit risk a'b
x 10~3/ppm
37
7.5
42
27
7.6
45
3.7
43
4.8
71
2.0
1.4
16
15
29
MCL = Mononuclear cell leukemia.
aUsing dose coefficients in terms of administered ppm of tetrachloroethylene adjusted to equivalent continuous
exposure, consistent with RfC methodology (U.S. EPA. 1994). BMCs/BMCLs estimated in terms of extra risk.
'Unit risks, which are approximations for extrapolation to lower doses, should not be used with exposures greater
than the POD from which they were derived without considering the curvature of the dose-response function (refer
to Appendix D for modeling details).
°Overall risk estimated using maximum likelihood method. Refer to Appendix D.3.1 for details.
Data source: Refer to Tables 5-13 and 5-15 and Appendix D.
5-70
-------
Table 5-20. Range of outputs from fitting different BMDS models using
continuous equivalent administered tetrachloroethylene levels as dose metric,
from JISA H993)a
Study group
Male mice
JISA (19931
Female mice
JISA (1993)
Male rats
JISA (19931
Female and male
rats JISA (19931
Tumor type
(all dose groups unless
otherwise specified)
Hepatocellular adenomas
or carcinomas
Hemangiomas or
hemangiosarcomas
Hepatocellular adenomas
or carcinomas
MCL
MCL
Mononuclear cell
leukemia
Range of PODs (ppm)
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
2.5-11
0.4-4.8
16-32
4.1-22
5.0-13
3.8-11
6.9-30
0.062 - 22
4.9 - 88
0-37
4.5-42
0.001-31
Range of
0.1/BMCL10
x 10~3/ppm
21-250
4.5-24
9.4-27
4.5 - 1600
2.7 - +00
3.3-71
MCL = Mononuclear cell leukemia
aUsing dose coefficients in terms of administered ppm of tetrachloroethylene adjusted to equivalent continuous
exposure, consistent with RfC methodology (U.S. EPA. 1994). BMCs/BMCLs estimated in terms of extra risk.
Range from use of different dose-response models (gamma, Weibull, log-logistic, log-probit, Michaelis-Menten,
probit, logistic, and multistage) using all dose groups, only including models with goodness-of-fit^-values > 0.1.
Extrapolation to humans using total liver oxidative metabolism led to a human equivalent
internal dose POD (BMCLio) of 3.9 mg/kg3/4-day liver oxidative metabolism. The corresponding
central tendency estimate was 2.2-fold higher, at 8.4 mg/kg3/4-day. Linear extrapolation from the
POD to low internal dose, followed by conversion to human exposures led to a human equivalent
unit risk of 0.92 x 10 per ppm.
Extrapolation to humans using TCA AUC in liver led to a human equivalent POD
(BMCLio) of 139 mg-hr/L-day TCA in blood. The corresponding central tendency estimate was
approximately 2.1-fold higher, at 292 mg-hr/L-day. Linear extrapolation from the POD to low
internal dose, followed by conversion to human exposures, led to a human equivalent unit risk of
0.73 x 10 per ppm, slightly lower than the estimate using total liver oxidative metabolism.
Dose-response modeling using administered exposure fit the data points similarly to
when using total oxidative metabolism, or TCA AUC, in liver (details in Appendix D). The
5-71
-------
result was directly interpretable as a human equivalent POD (BMCLio), at 3.8 ppm
tetrachloroethylene in air. The corresponding central tendency estimate was approximately
twofold higher, at 5.0 ppm. Linear extrapolation from this POD led to a human equivalent unit
risk of 27 x 10 per ppm, more than an order of magnitude higher than using either PBPK-
estimated dose metric.
The NRC (2010) peer review recommended more extensive quantitative evaluation of the
uncertainty due to different forms of dose-response models. The analysis was conducted using
administered concentration using the range of dichotomous models included in BMDS. All the
models (gamma, Weibull, log-logistic, Michaelis-Menten, log-probit, probit, and logistic) fit
similar to or better than the multistage. The estimated BMCLios from the better-fitting models
were less than threefold different than that using the standard multistage model.
Therefore, due to the limited sensitivity to the selection of dose-response models, the
multistage model results were carried forward to support cancer risk estimates (refer to
Table 5-18). Due to the data supporting oxidative metabolism in hepatocellular tumors, the
estimates carried forward were those using total oxidative metabolism as the dose metric
(preferred), and those using TCA AUC in liver as the dose metric (alternative). The remaining
analyses (refer to Tables 5-19 and 5-20) using administered concentration using multistage and
other dose-response models are retained only to better characterize the range of results from
different dose-response models.
5.3.4.1.3. Hemangiosarcomas, male mice
Hemangiosarcomas were also observed in the JISA male mice: in liver, spleen, fat, and
subcutaneous skin. Because these tumors differ etiologically from the hepatocellular adenomas
and carcinomas, they were modeled separately. In accordance with standard practice in the
absence of MO A data supporting a particular dose-response model form, multistage modeling of
the JISA bioassay data was carried out, using the preferred dose metric of tetrachloroethylene
AUC in blood, including fits for one-, two-, and three-stage models (details in Appendix D). A
one-stage model was found to be sufficient, with an adequate goodness-of-fit/?-value (p = 0.38),
and overall adequate visual fit (refer to Figure 5-11). There was no statistical improvement in
fitting higher order models, as all the higher order parameters were estimated to be zero.
Extrapolation to humans led to an internal dose POD (BMCLio) of 34 mg-hr/L-day
tetrachloroethylene in blood (refer to Table 5-18). The corresponding central tendency estimate
was nearly twofold higher, at 63 mg-hr/L-day. Linear extrapolation from the POD to low
internal dose, followed by conversion to human exposures led to a human equivalent unit risk of
5.9 x 10 per ppm.
5-72
-------
••5
Multistage Cancer Model with 0.95 Confidence Level
0.4
0.35
0.3
0.25
02
0.15
0.1
0.05
0
Multistage Cancer
Linear extrapolation
BMDL
BMD
20
40
60
dose
80
100
120
Figure 5-11. Dose-response modeling of male mouse hemangiomas or
hemangiosarcomas associated with inhalation exposure to
tetrachloroethylene, in terms of tetrachloroethylene AUC in blood; response
data from JISA (1993).
Details in Appendix D.
Dose-response modeling using administered exposure fit the data points similarly to when
using tetrachloroethylene AUC in blood (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 13 ppm tetrachloroethylene in air (refer
to Table 5-19). The corresponding central tendency estimate was approximately twofold higher,
at 24 ppm. Linear extrapolation from this POD led to a human equivalent unit risk of 7.5x10
per ppm, slightly higher than using tetrachloroethylene AUC in blood.
These results raise some concern that total cancer risk based on the male mice data may
be underestimated by considering only one site. Methods for estimating overall risk from sites
with very different dose metrics are not currently available. However, when an analysis using
administered concentration as the dose metric for both sites was carried out, using a method
based on maximum likelihood estimation,33 the overall risk was estimated to be only slightly
-3
An approach suggested in the EPA Guidelines for Carcinogen Risk Assessment (2005a) to characterize total risk
from multiple tumor sites would be to estimate cancer risk from tumor-bearing animals. EPA traditionally used this
approach until Science and Judgment in Risk Assessment (NRC. 1994) made a case that this approach would tend to
underestimate composite risk when tumor types occur in a statistically independent manner— that is, that the
occurrence of a hemangiosarcoma, say, would not be dependent on whether there was a hepatocellular tumor. This
assumption cannot currently be verified and if not correct could lead to an overestimate of risk from combining
5-73
-------
higher than that using hepatocellular tumors alone (refer to Table 5-19). The analysis yielded an
overall risk value of 0.042 per ppm, compared with the unit risk of 0.037 based on hepatocellular
tumors alone. On the other hand, using administered concentration for the hepatocellular tumors
may substantially overestimate human equivalent risk as compared to that estimated by using
total liver metabolism, under the assumption that oxidative metabolism is likely an important
component of this process. Refer to Appendix D. 1.1.3 for a summary of the calculations.
The NRC (2010) peer review recommended more extensive quantitative evaluation of the
uncertainty due to different forms of dose-response models. The analysis was conducted using
administered concentration using the range of dichotomous models included in BMDS. All of
the models had similar or worse fits than the multistage (gamma, Weibull, log-logistic, log-
probit, Michaelis-Menten, probit, and logistic). The estimated BMCLi0s ranged from 3.2-fold
less to 1.7-fold more than that using the multistage model.
Therefore, due to the limited sensitivity to the selection of dose-response models, the
multistage model result was carried forward to support cancer risk estimates (refer to Table 5-
18). Due to the lack of data on the active moiety for this endpoints, the result carried forward
used AUC of tetrachloroethylene in blood as the preferred dose metric. The remaining analyses
(refer to Table 5-19) using administered concentration using multistage and other dose-response
models are retained only to better characterize the range of results from different dose-response
models.
5.3.4.1.4. Mononuclear cell leukemia (MCL), male rat
In accordance with standard practice in the absence of MO A data supporting a particular
dose-response model form, multistage modeling of the JISA bioassay data was carried out
considering fits for one-, two-, and three-stage models (details in Appendix D). Using the
preferred dose metric of tetrachloroethylene AUC in blood, a one-stage model had a goodness-
of-fitp-value = 0.52, generally considered adequate, and the standardized residuals were within
the recommended limit of ±2 units (refer to Figure 5-12a). There was no statistical improvement
in fitting higher order models, as all the higher order parameters were estimated to be zero.
Extrapolation to humans led to an internal dose POD (BMCLio) of 30 mg-hr/L-day
tetrachloroethylene in blood (refer to Table 5-18). The corresponding central tendency estimate
was less than twofold higher, at 46 mg-hr/L-day. Linear extrapolation from the POD to low
across tumor sites. However, NRC (1994) argued that a general assumption of statistical independence of tumor-
type occurrences within animals was not likely to introduce substantial error in assessing carcinogenic potency from
rodent bioassay data.
5-74
-------
exposures, followed by conversion to human exposures, led to a human equivalent unit risk of
6.8 x 10 per ppm.
Dose-response modeling using administered exposure fit the data points similarly to that
using tetrachloroethylene AUC in blood (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 13 ppm tetrachloroethylene in air (refer
to Table 5-19). The corresponding central tendency estimate was approximately twofold higher,
at 21 ppm. Linear extrapolation from this POD led to a human equivalent unit risk of 7.6 x 10
per ppm, very similar to that using tetrachloroethylene AUC in blood.
To address NRC (2010) peer review comments, additional dose-response models were
evaluated for this data set in order to obtain a better model fit, particularly at lower doses where
the data set exhibited some supralinearity (refer to Appendix D, Table D-l 1). The analysis was
conducted using administered concentration using the range of dichotomous models included in
BMDS. Among the models fitted, five models fit better than the multistage (gamma, Weibull,
log-logistic, log-probit, and Michaelis-Menten), with two models leading to worse fits than the
multistage (probit and logistic). Visually, the Michaelis-Menten model better captured the
supralinear dose-response shape of the data. Because of the better dose-response fit, the
Michaelis-Menten model was preferred over the standard multistage model for this data set using
administered concentration. The human equivalent POD (BMCLio) was 2.2 ppm
tetrachloroethylene in air (refer to Table 5-19), with the corresponding central tendency estimate
3.8-fold higher at 8.6 ppm. Linear extrapolation from this POD led to a human equivalent unit
risk of 45 x 10 per ppm, sixfold higher than using the multistage model.
Based on this analysis, the Michaelis-Menten model was also fitted using the preferred
dose metric of tetrachloroethylene AUC in blood (the analysis was not conducted using other
dose-response models because of the near proportionality between this dose metric and
administered tetrachloroethylene). Extrapolation to humans led to an internal dose POD
(BMCLio) of 5 mg-hr/L-day tetrachloroethylene in blood (refer to Table 5-18). The
corresponding central tendency estimate was about 4-fold higher, at 20 mg-hr/L-day. Linear
extrapolation from the POD to low exposures, followed by conversion to human exposures, led
to a human equivalent unit risk of 40 x 10 per ppm.
5-75
-------
Mu Itistage Can cer Model with 0.95 Con fiden ce Level
BMDL | feMD
0 50 100 150 200 250
dose
a. One-degree multistage model fit to male
rat MCL data, all dose groups.
Dichotomous-Hill Model with 0.95 Confidence Level
50 100 150 200 250
b. One-degree multistage model fit to
female rat MCL data, all dose groups.
Multistage Can cer Model with 0.95 Confidence Level
< 0.3
100 150 200 250
c. Michaelis-Menten model fit to male rat
MCL data, all dose groups.
d. Multistage model fit to female rat MCL
data, control and lowest dose group only.
e. Michaelis-Menten model fit to female and
male rat MCL data, all dose groups.
Figure 5-12. Dose-response modeling of female and male rat MCLs
associated with inhalation exposure to tetrachloroethylene, in terms of
tetrachloroethylene AUC in blood; response data from JISA (1993).
Details in Appendix D.
5-76
-------
Therefore, two approaches were carried forward to support cancer risk estimates: the
standard approach using the multistage model and a better-fitting approach using the Michaelis-
Menten model, both on the basis of AUC of tetrachloroethylene in blood. The remaining
analyses using administered concentration using these and (less preferred) alternative approaches
are retained only to better characterize the range of results from different dose-response models.
5.3.4.1.5. Mononuclear cell leukemia (MCL), female rat
In accordance with standard practice in the absence of MO A data supporting a particular
dose-response model form, multistage modeling of the JISA bioassay data was carried out
considering fits for one-, two-, and three-stage models (details in Appendix D). Using the
preferred dose metric of tetrachloroethylene AUC in blood, a one-stage model had a goodness-
of-fitp-value (p = 0.34) generally considered adequate, and the standardized residuals were
within the recommended limit of two units (refer to Figure 5-12b). There was no statistical
improvement in fitting higher order models, as all the higher order parameters were estimated to
be zero. Extrapolation to humans led to an internal dose POD (BMCLio) of 61 mg-hr/L-day
tetrachloroethylene in blood. The corresponding central tendency estimate was about twofold
higher, at 136 mg-hr/L-day. Linear extrapolation from the POD to low exposures, followed by
conversion to human exposures, led to a human equivalent unit risk of 3.4 x 10 per ppm.
Dose-response modeling using administered exposure fit the data points similarly to that
using tetrachloroethylene AUC in blood (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 27 ppm tetrachloroethylene in air. The
corresponding central tendency estimate was approximately twofold higher, at 60 ppm. Linear
extrapolation from this POD led to a human equivalent unit risk of 3.7 x 10 per ppm,
essentially the same as using tetrachloroethylene AUC in blood.
To address NRC (2010) peer review comments, additional options were evaluated for this
data set in order to obtain a better model fit, particularly for lower doses at which the data set
exhibited some supralinearity (refer to Appendix D, Table D-6). These analyses were conducted
using administered concentration, due to its close proportionality with AUC of
tetrachloroethylene in blood. This case was the most extreme among the supralinear datasets,
with the multistage model estimate of the control incidence markedly above the data, and the
estimate of the lowest dose group markedly below the data. Briefly, use of a wider range of
dose-response models [as suggested by NRC (2010)] for the full data set was considered first.
When those attempts proved unsuccessful, incorporation of historical controls and exclusion of
higher exposure groups were also considered. These approaches are described in more detail
below.
5-77
-------
First, the range of dichotomous models included in BMDS was considered. Among the
models fitted, four models fit better than the multistage (gamma, Weibull, log-logistic, and
Michaelis-Menten), two models fit similarly to the multistage (probit and logistic), and one
model fit worse than the multistage (log-probit). However, for the better-fitting models, the
predicted response rate became virtually infinite in slope, approaching zero dose. Thus, no
BMCLio could be estimated (refer to Appendix D), indicating that the statistical uncertainty is
too great to support the BMC estimates. While data are lacking to inform the dose-response
relationship below 50 ppm in female rats, these fits are consistent with the possibility that a
response plateau extends below the lowest observed response. Therefore, none of these options
were successful in both improving upon the multistage model fit and estimating a BMCL.
The next strategy for obtaining an adequate fit to the female rat MCL data involved
focusing model fitting on the low-exposure range. First, the sensitivity of the fit to the use of
historical controls was examined in an attempt to constrain the estimated control response at a
level representative of previously observed values. Thus, the concurrent control was replaced
with the overall historical control incidence for inhalation studies in this laboratory (66/448
among control female rats in inhalation studies; refer to Table 5-16), and all models above were
fitted. None of these fits was both adequate and an improvement on the fits obtained with
concurrent controls (results not shown).
Next, exposure groups were excluded from analysis, starting with the highest exposure
group (600 ppm). All models used above were considered, as was the use of either the
concurrent or historical controls. All model fits were essentially the same as when using the full
data set (refer to Appendix D). Consequently, the next highest exposure group's data (200 ppm)
were also excluded. Only the multistage model was fit to the two remaining data points (control
and 50 ppm) because the other models use more parameters and need more data points. The
BMCLio was 2.3 ppm, and the BMCio was about twofold higher at 4.9 ppm (refer to
Figure 5-12d; details in Appendix D). Linear extrapolation from the POD to low exposures,
followed by conversion to human exposures, led to a human equivalent unit risk of 43 x 10 per
ppm. In sum, dose-response modeling of the full female rat MCL data set was only superior to
the multistage model for models that could not provide a lower bound estimate for a POD. The
only method that both led to a better fit to the control data and provided a lower bound BMC
estimate for a POD was use of just the concurrent control and lowest female rat exposure group.
This analysis is, therefore, consistent with the suggestion by the NRC (2010) that use of the
multistage model for the full datasets is not likely to provide a conservative upper bound
estimate of risk for this data set, and may, therefore, underestimate risk.
Based on this analysis, the multistage model was also fitted to only the concurrent control
and lowest exposure group using the preferred dose metric of tetrachloroethylene AUC in blood
5-78
-------
(the analysis was not conducted using other models because of the near proportionality between
this dose metric and administered tetrachloroethylene). Extrapolation to humans led to an
internal dose POD (BMCLio) of 5.2-mg-hr/L-day tetrachloroethylene in blood. The
corresponding central tendency estimate was about twofold higher, at 11 mg-hr/L-day. Linear
extrapolation from the POD to low exposures, followed by conversion to human exposures led to
a human equivalent unit risk of 39 x 10 per ppm, essentially the same as the result using
administered concentration.
Therefore, two approaches were carried forward to support cancer risk estimates: the
standard approach using the multistage model and the full data set, and the only available better-
fitting approach using the multistage model, and only the control and lowest dose group data,
both on the basis of AUC of tetrachloroethylene in blood. However, neither method fully
captures the potential extent of supralinearity into the region below the lowest dose. The
remaining analyses using administered concentration using these and (less preferred) alternative
approaches are retained only to better characterize the range of results from different dose-
response models.
5.3.4.1.6. Mononuclear cell leukemia (MCL), combined female and male rat
The MCL data for male rats and especially female rats were challenging to fit because of
the apparent supralinearity at lower doses. It was hypothesized that the male and female MCL
responses reflect the same underlying dose response to tetrachloroethylene. The presence of a
supralinear shape to the dose response for both male and female rats, in both the NTP (1986) and
JISA (1993) bioassays (refer to Figure 5-7), and the similar background MCL rates between
sexes in the JISA rats, are consistent with this hypothesis. Combining the datasets would
increase statistical power and, thus, perhaps better stabilize the BMDL estimates while being
able to fit the supralinear shape.
Two analyses were conducted to evaluate the consistency of the two JISA datasets. A
test described by Stiteler et al. (1993) evaluates whether two datasets are consistent with an
underlying dose-response model. In this case, the Michaelis-Menten model was used, given its
relative success at fitting both datasets in the low-dose region. The test involves comparing the
maximum log-likelihoods for the separate and combined datasets. The resulting p-va\ue was
0.54, indicating insufficient reason to conclude that the datasets differ from one underlying
model. The other analysis used a logistic regression to test whether the datasets differed
significantly between males and females. The advantage of this approach is that it does not
require assuming a specific functional form to represent the dose-response relationship. This
5-79
-------
analysis yielded a/?-value of 0.197, indicating no significant relationship of sex in the pattern of
responses. Refer to Appendix D for more details of both analyses.
The analysis began with fitting all dichotomous models to the combined male and female
MCL data on the basis of administered concentration. As compared to the sex-specific analyses,
only the Michaelis-Menten model provided an overall improved fit to all dose groups relative to
the multistage model (refer to Figure 5-12d). The resulting BMCio was 7.7 ppm, and the
BMCLio was about sixfold lower at 1.4 ppm (refer to Table 5-19). Thus, combining the male
and female rat MCL data generated a result with slightly greater statistical uncertainty (shown in
the wider confidence interval) than POD estimates for the sex-specific results. Linear
extrapolation from this POD to low exposures led to a human equivalent unit risk of 71 x 10
per ppm.
Based on this analysis, the Michaelis-Menten model was also fitted using the preferred
dose metric of tetrachloroethylene AUC in blood (the analysis was not conducted using other
models because of the near proportionality between this dose metric and administered
tetrachloroethylene). The result was a human equivalent POD (BMCLio) of 3.0 mg-hr/L-day.
The corresponding central tendency estimate was approximately sixfold higher, at 17 mg-hr/L-
day. Linear extrapolation from this POD led to a human equivalent unit risk of 68 x 10 per
ppm, essentially the same as the estimates using administered tetrachloroethylene.
Therefore, the approach carried forward to support cancer risk estimates was the
Michaelis-Menten model on the basis of AUC of tetrachloroethylene in blood. The remaining
analyses using administered concentration are retained only to better characterize the range of
results from different dose-response models.
5.3.4.1.7. Other tumors in male rats
As discussed in Section 5.3.1, tumors occurred at multiple sites in male rats exposed to
tetrachloroethylene in the NTP (1986) bioassay. While the design of NTP study is less suitable
than the JISA study for developing risk estimates, due to the higher exposures and the fewer dose
groups, dose-response modeling of these data was conducted to address variability in responses
across animal strains and bioassays. Estimates were developed for the risk of each tumor type
individually, as well as for the risk of any combination of tumor types. Because these analyses
are considered less preferred alternatives to those based on the JISA study, additional analyses
with respect to dose-response model selection were not conducted for these data.
5-80
-------
5.3.4.1.7.1. Kidney tumors, male rat
As discussed in Section 5.3.3.3 regarding selection of dose metrics, metabolism of
tetrachloroethylene via the GSH conjugation pathway was calculated as a dose metric relevant
for effects in the kidney. Multistage modeling of the NTP bioassay was carried out in units of
tetrachloroethylene conjugated with GSH per kg body weight to the % power per day
considering fits for one- and two-stage models. A one-stage model was found to be sufficient,
with an adequate goodness-of-fit^-value (p = 0.75) and overall adequate visual fit (refer to
Figure 5-13, details not provided). There was no statistical improvement in fitting higher order
models, as all the higher order parameters were estimated to be zero. Extrapolation to humans
led to an internal POD (BMDLio) of 0.21 mg/kg°75-day in blood (refer to Table 5-18). The
corresponding central tendency estimate was about twofold higher, at 0.46 mg/kg°'75-day. Linear
extrapolation from the POD to low internal dose, followed by conversion to human exposures,
led to a human equivalent unit risk of 100 x 10 per ppm. However, using the range of posterior
modes for the PBPK model predictions led to human equivalent risks of 0.047 x 10 to 110 x
10 , a range of more than 2000-fold. In view of this large range (much larger than the range for
any of the other endpoints), and the inability to discern from the toxicokinetic data whether this
spread represented uncertainty or variability or both [refer to Section 3; Chiu and Ginsberg
(2011)1, AUC of the parent compound in the blood was preferred as the dose metric for kidney
toxicity, while carrying forward the results of using the GSH conjugation dose metric for
comparison.
Thus, multistage modeling of the kidney tumor data was also carried out in units of
tetrachloroethylene AUC in blood, considering fits for one-, two-, and three-stage models. A
one-stage model had an adequate goodness-of-fit/?-value (p = 0.74) and overall adequate visual
fit. There was no statistical improvement in fitting higher order models, as all the higher order
parameters were estimated to be zero. Extrapolation to humans led to an internal POD
(BMDLio) of 110 mg-hr/L-day tetrachloroethylene in blood (refer to Table 5-18). The
corresponding central tendency estimate was about twofold higher, at 250 mg-hr/L-day. Linear
extrapolation from the POD to low internal dose, followed by conversion to human exposures,
led to a human equivalent unit risk of 1.8 x 10 per ppm.
Dose-response modeling using administered exposure fit the data points similarly to when
tetrachloroethylene AUC in blood was used (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 50 ppm tetrachloroethylene in air (refer
to Table 5-19). The corresponding central tendency estimate was approximately twofold higher,
at 110 ppm. Linear extrapolation from this POD led to a human equivalent unit risk of
2.1 x 10 per ppm, essentially the same as the estimate using tetrachloroethylene AUC in blood.
5-81
-------
Multistage Cancer Model with 0.95 Confidence Level
MultistageCancer -
Lin ear extrapolation
BMDL
100 150
dose
200 250
Multistage Cancer Model with 0.95 Confidence Level
MultistageCancer
Lin ear extrapolation
BMpL
0 50 100 150 200 250 300 350 400
dose
a. One-stage model fit to kidney tumors.
b. One-stage model fit to brain gliomas.
Multistage Cancer Model wrth 0.95 Confidence Level
Multistage Cancer Model wfth 0.95 Confidence Level
0 20 40
80 100 120 140 160
dose
c. One-stage model fit to testicular
interstitial cell tumors.
d. One-stage model fit to MCLs.
Figure 5-13. Dose-response modeling of male rat tumors—kidney, brain
gliomas, interstitial cell tumors, MCLs—associated with inhalation exposure
to tetrachloroethylene, in terms of tetrachloroethylene AUC in blood;
response data from NTP (1986).
Details in Appendix D.
5-82
-------
Two multistage model results were carried forward to support cancer risk estimates (refer
to Table 5-18): that using AUC of tetrachloroethylene in blood as the dose metric (preferred),
and those using GSH conjugation metabolism as the dose metric (alternative). For the
alternative dose metric, it is also noted that the range of PBPK model-based estimates is carried
forward to characterize the impact of uncertainty in GSH conjugation metabolism in humans.
5.3.4.1.7.2. Brain tumors, male rat
Multistage modeling of the NTP bioassay data for brain gliomas in male rats was carried
out in units of tetrachloroethylene AUC in blood, considering fits for one- and two-stage models.
A one-stage model was found to be sufficient, with an adequate goodness-of-fit^-value (p =
0.11) and overall adequate visual fit (refer to Figure 5-13). There was no statistical improvement
in fitting higher order models, as all the higher order parameters were estimated to be zero.
Extrapolation to humans led to an internal POD (BMDLio) of 170 mg-hr/L-day
tetrachloroethylene AUC in blood (refer to Table 5-18). The corresponding central tendency
estimate was less than twofold higher, at 410 mg-hr/L-day. Linear extrapolation from the POD
to low internal dose, followed by conversion to human exposures, led to a human equivalent unit
risk of 1.3 x 10~3 per ppm.
Dose-response modeling using administered exposure fit the data points similarly to when
tetrachloroethylene AUC in blood was used (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 73 ppm tetrachloroethylene in air (refer
to Table 5-19). The corresponding central tendency estimate was about twofold higher, at
180 ppm. Linear extrapolation from this POD led to a human equivalent unit risk of
1.4 x io~3 per ppm, essentially the same as the estimate using tetrachloroethylene AUC in blood.
The multistage modeling result using tetrachloroethylene AUC in blood was carried
forward to support cancer risk estimates (refer to Table 5-18).
5.3.4.1.7.3. Testicular tumors, male rat
Multistage modeling of the NTP bioassay data for testicular tumors was carried out in
units of tetrachloroethylene AUC in blood, considering fits for one- and two-stage models. A
one-stage model had an adequate goodness-of-fit^-value (p = 0.40) and overall adequate visual
fit (refer to Figure 13c). There was no statistical improvement in fitting higher order models, as
all the higher order parameters were estimated to be zero.
Extrapolation to humans led to an internal POD (BMDLio) of 14 mg-hr/L-day
tetrachloroethylene in blood (refer to Table 5-18). The corresponding central tendency estimate
was about twofold higher, at 30 mg-hr/L-day. Linear extrapolation from the POD to low internal
5-83
-------
dose, followed by conversion to human exposures, led to a human equivalent unit risk of
14 x 10 per ppm.
Dose-response modeling using administered concentration fit the data points similarly to
when tetrachloroethylene AUC in blood was used (details in Appendix D). The result was
directly interpretable as a human equivalent POD (BMCLio), at 6.1 ppm tetrachloroethylene in
air (refer to Table 5-19). The corresponding central tendency estimate was approximately
twofold higher, at 13 ppm. Linear extrapolation from this POD led to a human equivalent unit
risk of 16 x 10 per ppm, the same as the higher estimate using tetrachloroethylene AUC in
blood.
The multistage modeling result using tetrachloroethylene AUC in blood was carried
forward to support cancer risk estimates (refer to Table 5-18).
5.3.4.1.7.4. Mononuclear cell leukemia, male rat
Multistage modeling of the NTP bioassay data for male rat MCL was carried out in units
of tetrachloroethylene AUC in blood, considering fits for one- and two-stage models. A
one-stage model had an adequate goodness-of-fit^-value (p = 0.18) and overall adequate visual
fit (refer to Figure 5-13d). There was no statistical improvement in fitting higher order models,
as all the higher order parameters were estimated to be zero. Extrapolation to humans led to an
internal POD (BMDLio) of 15 mg-hr/L-day tetrachloroethylene in blood, and a corresponding
central tendency estimate about twofold higher, at 28 mg-hr/L-day. Linear extrapolation from
the POD to low internal dose, followed by conversion to human exposures, led to a human
equivalent unit risk of 15 x 10 per ppm.
Dose-response modeling using administered exposure fit the data points similarly to when
tetrachloroethylene AUC in blood was used (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 6.5 ppm tetrachloroethylene in air (refer
to Table 5-19). The corresponding central tendency estimate was approximately twofold higher,
at 12 ppm. Linear extrapolation from this POD led to a human equivalent unit risk of 15 x 10
per ppm, essentially the same as the estimate using tetrachloroethylene AUC in blood.
The multistage modeling result using tetrachloroethylene AUC in blood was carried
forward to support cancer risk estimates (refer to Table 5-18).
5.3.4.1.7.5. Total risk estimate for NTP (1986) male rats
The increased incidences of kidney, brain, and testicular interstitial cell tumors observed
in the NTP (1986) male rats led to unit risks that ranged from about 1 x 10"3 to 15 x 10"3 per
ppm, all lower than the unit risk based on male rats in the JISA (1993) study using the Michaelis-
5-84
-------
Menten model. In order to compare the results of both studies more equitably, the overall impact
of these multiple tumor types, or the risk of developing any combination of the four tumor types,
was estimated. First, the tumor types were judged likely to occur independently of each other or
not only in the presence of one of the other tumor types. The individual risk estimates developed
above were combined for an overall estimate of risk of any combination of these four tumor
types, using the approach based on maximum likelihood estimation described in Section
5.3.4.1.3.
In terms of tetrachloroethylene AUC, the POD (BMDLio) was 8.1-mg-hr/L-day
tetrachloroethylene in blood (refer to Table 5-18). The corresponding central tendency estimate
was almost twofold higher, at 13 mg-hr/L-day. Linear extrapolation from the POD to low
internal dose, followed by conversion to human exposures, led to a human equivalent unit risk of
25 x 10 per ppm.
Using administered exposure, the estimated overall risk was similar to when
tetrachloroethylene AUC in blood was used (details in Appendix D). The result was directly
interpretable as a human equivalent POD (BMCLio), at 3.5 ppm tetrachloroethylene in air (refer
to Table 5-19). The corresponding central tendency estimate was approximately twofold higher,
at 6.1 ppm. Linear extrapolation from this POD led to a human equivalent unit risk of
29 x 10 per ppm, essentially the same as the higher estimate using tetrachloroethylene AUC in
blood.
The combined overall risk using tetrachloroethylene AUC in blood as the dose metric for
each tumor type was carried forward to support cancer risk estimates (refer to Table 5-18).
Overall, the combined unit risk estimate was less than twofold higher than the highest individual
unit risk. While this bioassay is less ideal for low-dose extrapolation than the JISA bioassay, it is
still notable that the combined risk estimate supports the JISA study results, less than threefold
lower than the highest JISA study estimate of-70 x 10 per ppm.
5.3.4.1.8. Summary and discussion of site-specific dose-response modeling
The standard approach of applying the multistage model to the candidate data sets, using
PBPK model-based dose metrics, yielded results that were considered adequate according to
several criteria, including goodness-of-fit ^-values > 0.05 and standardized residuals within ±2.
However, the NRC (2010) peer review report recommended a more extensive quantitative
evaluation of uncertainty due to different forms of dose-response models. In particular, NRC
(2010) agreed that for several datasets, the multistage model does not fit the data at lower doses,
owing to the supralinear shape in the data. Furthermore, they noted that lack of significance in
goodness-of-fit tests can result from a small number of animals in each dose group, and use of
5-85
-------
such tests to justify a selection of a dose-response model can be misleading. Therefore, for the
datasets from JISA (1993), additional analyses were performed to examine whether alternative
dose-response models better accounted for datasets that exhibited supralinearity, and to more
generally characterize the range that would result from applying different dose-response models.
The discussion here focuses on the JISA (1993) data, because these were selected as the primary
source of dose-response data.
For mouse hepatocellular tumors and hemangiomas and hemangiosarcomas, the
alternative analyses did not lead to better fits and did not suggest a wide range of possible results
from alternative dose-response models. Therefore, for those datasets, the results from the
standard multistage approach were carried forward for consideration (in some cases including an
alternative dose metric in addition to the preferred one).
For male and female rat MCLs, some of the analyses yielded model results that
substantially improved fit to the datasets' supralinearity. For male rat MCLs, the preferred result
carried forward used the Michaelis-Menten model, with the standard multistage approach also
carried forward as an alternative for comparison. However, application of the range of
alternative dose-response models led to a wide (>300-fold) range of BMCL estimates, indicating
that the data have difficulty supporting a robust statistical lower bound on the BMC.
For female rat MCLs, the only approach that was successful in both addressing the
supralinearity and estimating a BMCL was multistage modeling of only the control and low-dose
group. Moreover, for this data set, the standard multistage approach using the entire data set had
the most pronounced inaccuracy with respect to the supralinearity in the data. These two results-
the multistage model using the full data set and using only the control and low-dose group-were
carried forward because they were the best available for this data set. The fit to the full data set
likely substantially overestimates the BMD, due to the markedly high estimate for the control
incidence and the markedly low estimate for the low-dose incidence. However, while the fit to
only the control and low-dose groups leads to a good fit to those data, it cannot quantitatively
address the possibility that the supralinearity extends below the lowest dose group. Finally,
application of the range of alternative dose-response models led to an unbounded range of
BMCL estimates, with some models unable to estimate a statistical lower bound on the BMC.
Because of these difficulties in fitting the individual rat MCL datasets, a subsequent
analysis was performed using the combined male and female datasets. There are no biological
data suggesting that the male and female rats would not reflect the same underlying toxicological
dose response, and statistical tests indicated that these data could be combined. In fitting the
range of available dichotomous models, it was found that the Michaelis-Menten model led to the
best fit and was able to account for the supralinearity in the full data set. Moreover, the range of
alternative dose-response models leads to stable estimates for the lower bound on the BMD.
5-86
-------
Therefore, the analysis using the Michaelis-Menten model on the full, combined male and
female rat MCL data set was carried forward for consideration.
Figure 5-14 shows the relative magnitudes of the unit risks associated with each tumor
site. Also shown are the unit risks estimated using alternate dose metrics, including administered
concentration, the range of estimates based on alternative PBPK model parameters, and the range
of estimates based on the range of dose-response models available in BMDS. Finally, this figure
also includes estimates based on dose-response modeling of the NTP (1986) bioassay. In terms
of preferred dose metrics (refer to Section 5.3.3.2.3.1), the unit risks, rounded to one significant
figure, ranged from 0.9 to 70 x 10 per ppm, about an 80-fold range.
5.3.4.2. Inhalation Unit Risk
Human inhalation cancer risk has been assessed using several different gender-species
animal tumor data sets and a newly developed "harmonized" PBPK model. These results, and
their uncertainties, have been discussed above and are summarized in Figure 5-14.
The majority of the NRC peer review panel recommended that the mouse hepatocellular
tumors be used for cancer risk estimation. Therefore, the inhalation unit risk is 2 x 10~3 per ppm
or 3 x 10~7 per ug/m3 (rounding to one significant digit), based on the male mouse hepatocellular
tumor data from the JISA (1993) bioassay. The inhalation unit risk should not be used with
exposures exceeding 60 ppm, or 400 mg/m3 (the equivalent ambient exposures corresponding to
the POD for male mouse hepatocellular tumors), because above this exposure level, the dose-
response relationship is not linear, and the unit risk would tend to overestimate risk. The slope
of the linear extrapolation from the human equivalent central estimate BMCio is 1.9 x 10~7 per
ug/m3 [0.1/(5.4 x io5 ug/m3)]. Some members of the NRC peer review panel recommended that
the MCL data be used for cancer risk estimation. The inhalation unit risk would be 7 x 10~2 per
ppm, or 1 x 1CT5 per ug/m3 (rounding to one significant digit) if it were based on the male and
female rat MCL data from the JISA (1993) bioassay.
5-87
-------
permg/m3
1CT6 10"5 10^
i i i i
a
A
H
(I
T3 '-
d =5
C •=
SZ O
(— • —
m J=
Ct .E
4
10"3 10"2 10~1
I I I
I----X
L
V **"Jtft
V ^4*i
*
>
.j<-~,
*
Female mice, hepatocellular tumors
Male mice, hepatocellular tumors
Female and male rats. MCL
Male rats, MCL
Male rats (NTP), kidney tumors
Male mice, hemangiomas or hernangiosarcornas
Male rats (NTP): brain, kidney, testes, MCL
I I I I I
10"5 10"* 10~3 10"2 10"1 1
Human equivalent PCE unit risk estimates (per ppm)
Administered PCE,
multistage, all dose groups
Administered PCE,
other analyses
PCE AOC in blood
° TCA AUC in liver
O Total GSH metab.
A Total liver oxidative metab.
(filled) primary dose metric,
multistage, all dose groups
(filled) primary dose metric,
other analyses
Preferred estimate for dataset
Figure 5-14. Comparison of inhalation unit risks for tetrachloroethylene
derived from rodent bioassays using PBPK-based dose metrics and
administered concentration.
Symbols represent results using the posterior mode PBPK model results, with filled symbols
representing the preferred dose metrics (refer to Tables 5-18 and 5-19). Red-filled symbols use
the multistage model with all dose groups; green-filled symbols use a different dose-response
approach in response to NRC (2010) comments. Solid error bars show the range of estimates
using the range of posterior modes for the human PBPK model-based conversion to a human
equivalent unit risk (refer to Table 5-18). Dashed error bars show the range of unit risk estimates
(based on administered concentration) using alternative dose-response models with goodness-of-
fit p-values > 0.10 (refer to Table 5-20). The preferred estimate for each dataset is circled in pink.
-------
5.3.4.3. Oral Slope Factor
The oral slope factor was developed from inhalation data because the only available oral
bioassay had several limitations for extrapolating to lifetime risk in humans (also refer to
Section 5.3.1). First, the oral study (NCI, 1977) was conducted by gavage at relatively high
doses. Human exposures are less likely to occur in boluses, and high doses are associated with
saturable metabolism processes, which may involve a different profile of toxicological processes
than those prevalent at more likely environmental exposure levels. Also, the animals were dosed
for only approximately 75% of the more usual 2-year period, making the oral study less useful
for estimating lifetime risk. Route-to-route extrapolation from the inhalation PODs developed
from the JISA study (refer to Table 5-18) was carried out using the human pharmacokinetic
models described in Section 3.5. The total tumor risks from multiple sites (brain, kidney, testes,
and MCL in rats and hepatocellular tumors and hemangiomas or hemangiosarcomas in mice)
were estimated using the same methods as were used for the inhalation unit risk estimates, with
results of 20 x 10 per mg/kg-day for rats in NTP (1986) and 18^10 per mg/kg-day for mice
in JISA (1993). Table 5-21 and Figure 5-15 summarize all of the resulting candidate oral slope
factors.
The majority of the NRC peer review panel recommended that the mouse hepatocellular
tumors be used for cancer risk estimation. Therefore, the oral slope factor is 2 x 10~3 per mg/kg-
day (rounding to one significant digit) based on the male mouse hepatocellular tumor data from
the JISA (1993) bioassay. The oral slope factor should not be used with exposures exceeding 50
mg/kg-day (the equivalent ambient exposure corresponding to the POD for male mouse
hepatocellular tumors), because above this exposure level, the dose-response relationship is not
linear, and the slope factor would tend to overestimate risk. The slope of the linear extrapolation
from the human equivalent central estimate BMDio is 1.5 x 10"3 per mg/kg-day [0.1/(67 mg/kg-
day)]. Some members of the NRC peer review panel recommended that the MCL data be used
for cancer risk estimation. The oral slope factor would be 6 x 10~2 per mg/kg-day (rounding to
one significant digit) if it were based on the male and female rat MCL data from the JISA (1993)
bioassay.
5-89
-------
he
ha
A
o —
oj ro
CC o
10 10 10 J 10 10 1
Human equivalent PCE slope factor estimates (per mg/kg-d)
Female mice, hepatocellular tumors
Male mice, hepatocellular tumors
Female and male rats. MCL
Male rats, MCL
Male rats (NTP), kidney tumors
Male mice, hemangiomas or hemangiosarcomas
Male rats (NTP): brain, kidney, testes, MCL
„ Administered PCE, _
X multistage, all dose groups D
^ Administered PCE,
other analyses
o PCE AUC in blood
in liver
0 Total GSH rnetab.
A Total liver oxidative metab.
(filled) primary dose metric,
multistage, all dose groups
(filled) primary dose metric,
other analyses
Preferred estimate for dataset
Figure 5-15. Comparison of oral slope factors for tetrachloroethylene,
derived from rodent bioassays using PBPK-based dose metrics and route-to-
route extrapolation.
Symbols represent results using the posterior mode PBPK model results, with filled symbols
representing the preferred dose metrics (refer to Table 5-21). Red-filled symbols use the
multistage model with all dose groups; green-filled symbols use a different dose-response
approach in response to NRC (2010) comments. Solid error bars show the range of estimates
using the range of posterior modes for the human PBPK model-based conversion to a human
equivalent unit risk (refer to Table 5-21). The preferred estimate for each dataset is circled in
pink.
5-90
-------
Table 5-21. Human equivalent candidate oral slope factors, derived using
primary dose metrics and multistage model; tumor incidence data from JISA
(1993) and NTP (1986)
Study Group
Tumor type
(multistage model with
all dose groups unless
otherwise specified)
Human Equivalents
PODa, in internal dose units
SFxlO'3
/internal
dose
unitb
Candidate
OSFxlO'3/
mg/kg-day
(PBPK
range)0
Primary dose metrics
Male mice
JISA (1993)
Female mice
JISA (1993)
Male rats
JISA (1993)
Female rats
JISA (1993)
Female and male
rats JISA (1993)
Male rats
NTP (1986)
Hepatocellular adenomas
or carcinomas
Hemangiomas,
hemangiosarcomas,
Hepatocellular adenomas
or carcinomas
MCL
MCL (Michaelis-
Menten)
MCL
MCL (control and low
dose groups only)
MCL (Michaelis-
Menten)
Kidney tumors
Brain gliomas
Testicular interstitial cell
tumors
MCL
Total risk for any of
above four tumor types
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
BMD10
BMDL10
2.9
2.1
63
34
8.4
4.0
46
30
20
5.0
136
61
11
5.2
17
3.0
246
110
400
192
31
14
28
15
14
8.2
Total liver oxidative
metabolism, mg/kg°75-d
PCEAUC in blood,
mg-hr/L-d
Total liver oxidative
metabolism, mg/kg°75-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
PCEAUC in blood,
mg-hr/L-d
49
2.9
25
3.4
20
1.6
19
33
0.90
0.62
7.1
6.6
12
2.1
(1.6-2.6)
5.1
(4.6-5.3)
1.1
(0.84-1.3)
5.9
(5.3-6.1)
35
(31-36)
2.8
(2.5-2.9)
33
(30-35)
58
(53-61)
1.6
(1.4-1.6)
1.1
(1.0-1.1)
12
(11-13)
12
(10-12)
21
(19-22)
Alternate Dose Metrics
Male mice
JISA (1993)
Female mice
JISA (1993)
Male rats
NTP (1986)
Hepatocellular aden-
omas or carcinomas
Hepatocellular aden-
omas or carcinomas
Kidney tumors
BMD10
BMDL10
BMD10
BMDL10
BMD05
BMDL05
97
69
292
141
0.46
0.21
TCA AUC in liver,
mg-hr/L-d
TCA AUC in liver,
mg-hr/L-d
Total GSH metabolism,
mg/kg0.75-d
1.5
0.72
243
1.7
(1.3-1.8)
0.85
(0.65-0.89)
120
(0.045-140)
5-91
-------
Table 5-21. Human equivalent candidate oral slope factors, derived using
primary dose metrics and multistage model; tumor incidence data from JISA
(1993) and NTP (1986) (continued)
SF = slope factor; OSF = oral slope factor.
aPODs were estimated at the indicated BMRs in terms of extra risk; i.e., BMDL10 is the lower bound for the internal
dose metric on the level associated with 10% extra risk. Dose units are in the first column, which include cross-
species scaling to a human equivalent internal dose metric. Refer to Appendix D for dose-response modeling
details.
Slope Factor = BMR/BMDLBMR in units of risk per dose metric unit.
°The oral slope factor is given by the product of the slope factor in units of risk per dose metric unit and an oral
dose-metric conversion factor (DMCF^^.^y): Oral Slope Factor = BMR/BMDLBMR x DMCFmg/kg.day, where the
DMCFmg/kg.day is derived from the PBPK model. The DMCFmg/lcg.day for each dose metric is a constant factor
shown:
Dose metric
Total liver oxidative metabolism
Tetrachloroethylene blood AUC
TCA AUC in liver
Total GSH metabolism
DMCF^/k,.^
Overall posterior mode
0.0438
1.74
1.18
0.512
Range of posterior modes
0.0334-0.0459
1.58-1.82
0.903-1.24
0.00019-0.543
Values in bold correspond to using the overall posterior mode and are carried forward for consideration in the
derivation of the cancer slope factor. The difference between the overall and alternative posterior modes is
negligible (relative to other uncertainties) except for the Total GSH metabolism dose metric.
5-92
-------
5.3.4.4. Uncertainties in Human Population Variability and Quantitative Adjustment for
Sensitive Populations (Age-Dependent Adjustment Factors)
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. The extent of interindividual variability in tetrachloroethylene metabolism
has not been characterized. As noted above, several enzymes of the oxidative and GSH
metabolism, notably Cytochrome 2E1 (CYP2E1), CYP3A4, GSTZ, GSTA, GSTM, and GSTT,
show genetic polymorphisms with the potential for variation in production of specific
metabolites. Inducers of CYP450 enzymes such as toluene, phenobarbital, and
pregnenolone-16a-carbonitrile have been shown to increase tetrachloroethylene metabolism,
whereas CYP enzyme inhibitors such as SKF 525A, metyrapone, and carbon monoxide have
been shown to decrease tetrachloroethylene metabolism. Additionally, chronic exposure to
tetrachloroethylene has been shown to cause self-induction of metabolism. Human population
variability has also been discussed in Section 3.
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 heterogeneous response that may
function additively to ongoing or background exposures, diseases, and biological processes. As
noted in 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. These considerations strengthen the scientific support
for the choice of a linear nonthreshold extrapolation approach. However, because chemical-
specific life-stage susceptibility data are not available, and the MOA for tetrachloroethylene has
not been established, the application of age-derived adjustment factors for early life exposures,
as discussed in Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
Carcinogens (U.S. EPA, 2005b), is not recommended.
5.3.4.5. Concordance of Animal and Human Risk Estimates
Sufficient human health outcome data with quality exposure characterizations linked to
individual study subjects or epidemiologic studies with characterization of exposure-response
using a quantitative surrogate of tetrachloroethylene exposure are not available to derive cancer
risk values. An analysis of epidemiologic studies provides some limited perspectives on the
human cancer risk values estimated from animal bioassays (van Wijngaarden and Hertz -
5-93
-------
Picciotto, 2004). The analysis assigned an exposure surrogate of average tetrachloroethylene
concentration to all exposed subjects based on information in the published literature
(van Wijngaarden and Hertz-Picciotto, 2004). EPA prefers that the exposure-assessment
approach of epidemiologic studies used for estimating lifetime cancer risk represent not only the
relevant conditions and exposures (e.g., through a job-exposure matrix or exposure model), but
also subject-specific quantitative estimates of exposure. The epidemiologic study (Vaughan et
al., 1997) in the analysis did not meet these criteria; the study did not assign a unique exposure
estimate to individual subjects, nor did it examine exposure response using a quantitative
exposure surrogate. Although not sufficient to serve as a primary basis for dose-response
assessment, this study does provide information without extrapolation from animals to humans.
Van Wijngaarden and Hertz-Picciotto (2004) demonstrated a simple methodology using
epidemiologic data for four chemical exposures including tetrachloroethylene. For
tetrachloroethylene specifically, a linear dose-response model was fit to laryngeal cancer
observations in the upper airway cancer case-control study of Vaughan et al. (1997). Van
Wijngaarden and Hertz-Picciotto (2004) presented both an ED0i and LED0i (effective dose for a
1% additional lifetime risk over background and the lower confidence interval on this dose,
called the TD1 and LCL1 in their paper) for humans exposed for 45 years, 240 days/year, a
standard occupational exposure scenario. The EDoi was 228.40 mg/day, and the LEDoi was
60.16 mg/day. In order to compare these results with those derived from the JISA (1993) study,
we assumed a continuous lifetime exposure (70 years, 365 days/year, and 20-m3/day breathing
rate), resulting in an equivalent EDoi of 4.8 mg/m3 and LEDoi of 1.3 mg/m3. Using the
continuous lifetime equivalent LED0i as the POD and a low-dose linear approach, a unit risk
based upon Vaughan et al. (1997) is 0.01/1.3 x 103 |ig/m3 = 8 x 10~6 per |ig/m3 (0.05 per ppm).
A cancer risk estimate from human data using the EDoi as the POD isO.01/4.8 x 103 |ig/m3 = 2
x 10~6 per |ig/m3 (0.01 per ppm). These estimates overlap with the cancer risk estimates from
combined male and female rat MCL tumors in JISA (1993), and from the combined male rat
brain, kidney, testes, and MCL tumors in NTP (1986).
These estimates are based on extrapolated exposure estimates, assume that laryngeal
cancer is the only carcinogenic hazard in humans, and may be subject to other sources of bias.
Thus, they should only be viewed as order-of-magnitude estimates.
5-94
-------
5.3.5. Summary of Uncertainties in Cancer Risk Values
A number of uncertainties underlie the cancer unit risk for tetrachloroethylene, including
the choice of study, PBPK modeling and dose metrics, cross-species scaling, low-dose
extrapolation, model uncertainty, statistical uncertainty in the POD, the species/gender/tumor
type combination selected, and sensitive subpopulations (refer to Table 5-22). Some suggest
risks could be higher than was estimated (e.g., selection of MCL rather than mouse liver tumors,
sensitive subpopulations), while others would decrease risk estimates (e.g., use of central
tendency instead of lower 95% confidence bound on the POD), or have an impact of an uncertain
direction. Several uncertainties are quantitatively characterized for the significantly increased
rodent tumors. These include the statistical uncertainty in the POD, the range of uncertainty in
PBPK modeling and dose metrics, dose-response model uncertainty, and the
species/gender/tumor type combination selected. The latter three of these could either increase
or decrease risk estimates. Due to limitations in the data, particularly regarding the mode of
action and relative human sensitivity and variability, the quantitative impact of other
uncertainties, which may have equal or greater impact, has not been explored.
5-95
-------
Table 5-22. Summary of uncertainties in tetrachloroethylene cancer unit
risk estimate
Consideration/
approach
(Section)
Bioassay (5.4.1)
PBPK modeling
and dose metrics
(5.4.3.1.5)
Cross-species
scaling
(5.4.3.2.2.3)
Low-dose
extrapolation
procedure
(5.4.3.2.3)
Model
uncertainty
Statistical
uncertainty at
POD (5.4.4.1.7)
Impact on unit
risk
t unit risk up to
twofold if NTP
study used
Alternatives could
t or 4 unit risk by
an unknown extent
Alternatives could
4 or t unit risk
(e.g., 3.5-fold 4
[scaling by BW]
or t twofold
[scaling by
BW2/3])
Departure from
EPA's Guidelines
for Carcinogen
Risk Assessment
OJ.S. EPA. 2005a)
POD paradigm, if
justified, could 4
or 1 unit risk an
unknown extent
Alternatives could
4 or t unit risk
4 unit risk 1.4-fold
ifBMC10used
rather than
BMCL10
Decision
JISA study
Relied on total liver
oxidative metabolism and
tetrachloroethylene AUC,
in addition to
administered
concentration
(default approach) for
total oxidative or GSH
metabolism; direct
animal-to-human
correspondence when the
dose metric was an AUC
Multistage model to
determine POD, linear
low-dose extrapolation
from POD (default
approach)
Multistage model for all
tumor sites except
Michaelis-Menten model
for MCLs from JISA
(1993) male rats, and
male and female rats
combined
BMCL (default approach
for calculating plausible
upper bound)
Justification
JISA study used the lowest experimental
exposures (reduces extrapolation uncertainty)
and used three treated groups.
Experimental evidence supports a role for
metabolism in toxicity, but actual responsible
metabolites are not clearly identified.
There are no data to support alternatives. Use
of BW3/4 for metabolism rates and no scaling
for dose metrics expressed as AUCs are
consistent treatments of the available dose
metrics. While the true human
correspondence is unknown, this overall
approach is expected neither to over- or
underestimate human equivalent risks.
Available MOA data do not inform selection
of dose-response model but do not support
nonlinearity (mutagenicity is plausible
contributor and cannot be ruled out); linear
approach in absence of clear support for an
alternative is generally supported by scientific
deliberations supporting EPA's Guidelines for
Carcinogen Risk Assessment.
No biologically based models available; no a
priori basis for selecting a model other than
multistage. Selected options tended to be
intermediate among the available alternatives.
Refer to Appendix D.
Limited size of bioassay results in sampling
variability; lower bound is 95% confidence
interval on concentration.
5-96
-------
Table 5-22. Summary of uncertainties in tetrachloroethylene cancer unit
risk estimate (continued)
Consideration/
approach
(section)
Species/gender/
tumor type
combination
(5.4.4.2,
Figure 5-14)
Human
population
variability
sensitive
subpopulations
(5.4.4.5)
Impact on unit
risk
Human risk could
4 or t more than
an order of
magnitude,
depending on
selected tumor
type and relative
species sensitivity
Low-dose risk f to
an unknown extent
Decision
Male mouse
hepatocellular tumors.
Considered qualitatively
Justification
Recommended by majority of the NRC (2010)
peer review panel.
No data to support range of human
variability /sensitivity in metabolism or
response, including whether children are more
sensitive. Mutagenic MO A, which cannot be
ruled out, would indicate increased early -life
susceptibility.
5-97
-------
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE
6.1. HUMAN HAZARD POTENTIAL
This section summarizes the human hazard potential for tetrachloroethylene. For
extensive discussions and references, refer to Section 2 for Exposure, Section 3 for
toxicokinetics and physiologically based pharmacokinetic (PBPK) modeling, and
Sections 4.1-4.8 for the epidemiologic and experimental studies of noncancer toxicity and
carcinogenicity. Section 4.9 summarizes information on susceptibility, and Section 4.10
provides a more detailed summary of noncancer toxicity and carcinogenicity.
6.1.1. Exposure (refer to Section 2)
Tetrachloroethylene is a volatile compound with relatively low water solubility. It is
widely used for dry cleaning of fabrics, for metal degreasing, and in manufacturing some
consumer products and other chemicals. Tetrachloroethylene has been detected in drinking,
ground, and surface water as well as in air, soil, food, and breast milk. The primary exposure
routes of concern are vapor inhalation and ingestion of contaminated water. Inhalation exposure
is the predominant route of exposure compared with ingestion, including from breast milk.
The highest environmental releases are to the air. Ambient tetrachloroethylene
concentrations vary from source to source and with proximity to the source. Outdoors, the high
volatility leads to increased ambient air concentrations near points of use (AT SDR, 1997a: U.S.
EPA, 1996b). The U.S. Environmental Protection Agency (EPA) has carried out modeling to
characterize the geographic distribution of tetrachloroethylene for its National-Scale Air Toxics
Assessment database (U.S. EPA, 1996b). Median census tract-based tetrachloroethylene
concentrations across the United States were estimated at about 0.3 |ig/m3 for urban areas and
0.1 |ig/m3 for rural areas (75% upper percentiles of 0.4 and 0.2 |ig/m3, respectively). Air
exposure may also occur from vapor intrusion, or during showering or bathing as dissolved
tetrachloroethylene in the warm tap water is volatilized.
Near points of use, such as dry cleaners or industrial facilities, indoor exposure to
tetrachloroethylene is more significant than outdoor exposure (U.S. EPA, 200la). Adgate et al.
(2004a) measured tetrachloroethylene in outside and indoor air at school, indoor air at home, and
using personal samplers on children, and demonstrated that levels are lower in homes with
greater ventilation (Adgate et al., 2004a) and in homes in nonurban settings (Adgate et al.,
2004b: Adgate et al., 2004a). Mean indoor air concentrations in apartments above dry-cleaning
shops of 4.9 mg/m3 have been reported [Altmann (1995): also refer to Garetano and Gochfeld
6-1
-------
(2000): McDermott (2005): Schreiber (1993): Schreiber (2002)]. Measurements have also been
made in a day care center adjacent to a dry cleaner (NYSDOH, 2005a, b, c), and in a classroom
exposed to tetrachloroethylene from an air "emission from a small chemical factory" (Monster
and Smolders, 1984). Mean concentrations inside dry-cleaning facilities were reported to be
454—1,390 mg/m3 in the United States and 164 mg/m3 in Nordic countries during the 1960s and
1970s. Overall levels declined from 95-210 mg/m3 in the 1980s to 20-70 mg/m3 over the next
decades in these countries (Lvnge et al., 2011; Gold et al., 2008; Lynge et al., 2006).
The off-gassing of garments that have recently been dry-cleaned may be of concern
[Tichenor (1990); also refer to Thomas et al. (1991)]. Relatively high tetrachloroethylene air
concentrations have been measured in closets and automobiles containing freshly dry-cleaned
clothing. Using dry-cleaned clothes as a source, tetrachloroethylene levels inside a stationary
vehicle after 30 minutes reached 0.230 mg/m3 (Park et al., 1998). A residential closet storing
newly dry-cleaned clothing had an air concentration of 2.9 mg/m3 after 1 day, which rapidly
declined to 0.5 mg/m3 and persisted for several days (Tichenor et al., 1990). There is
a documented mortality case: a 2-year-old boy was found dead after being put to sleep in a room
with curtains that had been incorrectly dry-cleaned (Gamier et al., 1996).
Exposure to related compounds—including metabolites and other parent compounds that
produce similar metabolites—can alter or enhance tetrachloroethylene metabolism and toxicity
by generating higher internal metabolite concentrations than would result from
tetrachloroethylene exposure by itself.
6.1.2. Toxicokinetics and Physiologically Based Pharmacokinetic (PBPK) Modeling (refer
to Section 3)
Tetrachloroethylene is a lipophilic compound that readily crosses biological membranes.
Tetrachloroethylene is rapidly absorbed into the bloodstream following oral and inhalation
exposures. It can also be absorbed across the skin following dermal exposure to either pure or
diluted solvent or vapors (Poet et al.. 2002: Nakaietal., 1999: Stewart and Dodd, 1964).
Additionally, tetrachloroethylene can be transferred transplacentally and through breast milk
ingestion. Refer to Section 3.1 for additional discussion of tetrachloroethylene absorption.
Once absorbed, tetrachloroethylene is distributed by first-order diffusion processes.
Animal studies provide clear evidence that tetrachloroethylene distributes widely to all tissues of
the body, readily crossing the blood:brain barrier and the placenta (Dallas et al., 1994b:
Ghantous et al., 1986; Schumann et al., 1980; Savolainen et al., 1977b). The highest tissue
concentrations were found in adipose tissue (60 or more times blood level) and in brain and liver
6-2
-------
(4 and 5 times blood level, respectively). Refer to Section 3.2 for additional discussion of
tetrachl oroethy 1 ene di stributi on.
The metabolism of tetrachl oroethy 1 ene is an important determinant of its toxicity.
Metabolites are generally thought to be responsible for toxicity—especially to the liver and
kidney. Tetrachl oroethy 1 ene is metabolized in laboratory animals and in humans through at least
two distinct pathways: oxidative metabolism via the cytochrome P450 (CYP [also abbreviated as
P450 and CYP 450]) mixed-function oxidase system and glutathione (GSH) conjugation
followed by further biotransformation and processing, either through the cysteine conjugate
P4yase pathway or by other enzymes including FMO3 and CYP3 A (Lash and Parker, 2001;
Lashetal.. 1998: Volkeletal.. 1998: Birneretal.. 1996: Dekantetal.. 1989: Anders et al.. 1988:
Dekantetal.. 1987: Costa and Ivanetich. 1980: Filser and Bolt 1979: Peggetal.. 1979: Daniel
1963). The conjugative pathway is lexicologically significant because it yields relatively potent
toxic metabolites (Lash and Parker, 2001: Werner et al., 1996: Dekant et al., 1989: Vamvakas et
al.. 1989c: Vamvakas et al.. 1989d: Anders et al.. 1988: Vamvakas et al.. 1987: Dekant et al..
1986b: Dekantetal., 1986d). Studies in both animals and humans indicate that overall
metabolism of tetrachl oroethy 1 ene is relatively limited, particularly at higher exposures
[reviewed in Lash and Parker (2001)]. Although thought to be qualitatively similar, there are
clear differences among species in the quantitative aspects of tetrachl oroethy 1 ene metabolism
(Lash and Parker. 2001: Volkeletal.. 1998: Schumann et al.. 1980: Ikeda and Ohtsuii. 1972).
Refer to Section 3.3 for additional discussion of tetrachl oroethy 1 ene metabolism.
Tetrachl oroethy 1 ene is excreted from the body by pulmonary excretion of the parent
compound and urinary excretion of metabolism products, with a small amount of pulmonary
excretion of metabolism products. Tetrachl oroethy 1 ene that is not metabolized is exhaled
unchanged, and this process is the primary pathway of tetrachl oroethy 1 ene excretion in humans
for all routes of administration (Opdam and Smolders, 1986: Koppel et al., 1985: Monster, 1979:
Stewart et al., 1977: Guberan and Fernandez, 1974: Stewart et al., 1974: Stewart and Dodd,
1964). Pulmonary excretion of (unchanged) parent compound is also important in animals
(Bogenetal.. 1992: Frantz and Watanabe. 1983: Schumann et al.. 1980: Peggetal.. 1979:
Yllner, 1961). A small amount of tetrachl oroethy 1 ene has been shown to be excreted through the
skin (Bolanowska and Golacka, 1972): however, it represents an insignificant percent of total
tetrachl oroethy 1 ene disposition. Refer to Section 3.4 for additional discussion of
tetrachl oroethy 1 ene excretion.
As part of this assessment, a PBPK model-based analysis of the toxicokinetics of
tetrachl oroethy 1 ene and its metabolites was developed in mice, rats, and humans [also reported in
Chiu and Ginsberg (2011)1. This model was developed to address many of the limitations of the
existing models for tetrachl oroethy 1 ene. Among the most important improvements are (1) the
6-3
-------
utilization of all the available toxicokinetic data for tetrachloroethylene and its metabolites in
mice, rats, and humans; (2) the incorporation of available information on the internal
toxicokinetics of TCA derived from the most current PBPK modeling of trichloroethylene and
TCA; and (3) the separate estimation of oxidative and conjugation metabolism pathways. This
"harmonized" PBPK model used a limited Bayesian analysis implemented using Markov chain
Monte Carlo approach for parameter calibration. As expected, the major route of elimination of
absorbed tetrachloroethylene is predicted to be exhalation as parent compound, with metabolism
accounting for less than 20% of intake except in the case of mice exposed orally, in which
metabolism is predicted to be slightly over 50% at lower exposures. In all three species, the
concentration in blood, the extent of oxidation, and the amount of TCA production is well
estimated, with residual uncertainties of-twofold. However, the resulting range of estimates for
the amount of GSH conjugation is quite wide in humans (~3,000-fold) and mice (~60-fold).
While even high-end estimates of GSH conjugation in mice are lower than estimates of
oxidation, in humans the estimated rates range from much lower to much higher than rates for
tetrachloroethylene oxidation. It is unclear to what extent this range reflects uncertainty,
variability, or a combination. Importantly, by separating total tetrachloroethylene metabolism
into separate oxidative and conjugative pathway, this analysis reconciles the disparity between
those previously published PBPK models that predicted either low or high metabolism in
humans. In essence, both conclusions are consistent with the data if augmented with some
additional qualifications: in humans, oxidative metabolism is low, while GSH conjugation
metabolism may be high or low, with uncertainty and/or interindividual variability spanning
three orders of magnitude. More direct data on the internal kinetics of tetrachloroethylene and
GSH conjugation, such as trichlorovinyl glutathione or trichlorovinyl cysteine levels in blood
and/or tissues, would be needed to better characterize the uncertainty and variability in GSH
conjugation in humans. Because of the substantial refinements from previous PBPK models, this
assessment utilizes the Chiu and Ginsberg (2011) model to calculate relevant dose-metrics that
were then used in dose-response modeling. Refer to Section 3.5 for additional discussion of and
details about PBPK modeling of tetrachloroethylene and metabolites.
6.1.3. Noncancer Toxicity (refer to Section 4.10.1)
Noncancer effects of tetrachloroethylene identified in exposed humans and animals
include toxicity to the central nervous system, kidney, liver, immune and hematologic system,
and on development and reproduction. Neurotoxic effects have been characterized in human
controlled exposure, occupational and residential studies, as well as in experimental animal
studies, providing evidence of an association between tetrachloroethylene exposure and
neurological deficits. Tetrachloroethylene exposure primarily results in visual changes,
6-4
-------
increased reaction time, and cognitive decrements in humans; in animal studies, effects on
vision, visual-spatial function, and reaction time, as well as brain weight changes were also seen.
Adverse effects on the kidney in the form of tubular toxicity, potentially mediated through
tetrachloroethylene GSH conjugation, have been reported in numerous well-conducted animal
studies. Although human studies have not systematically investigated nephrotoxicity, an
association between tetrachloroethylene exposure via inhalation and chronic kidney disease, as
measured by urinary excretion of renal proteins and end-stage renal disease, is supported. The
developmental and reproductive toxicity database for tetrachloroethylene includes a range of
data from appropriate well-conducted studies in several laboratory animal species plus limited
human data. Evidence of liver toxicity is primarily from several well-conducted rodent studies,
including chronic bioassays.
Other toxicity endpoints are less well characterized. The few published reports of
experimental studies examining immune or hematologic system toxicity are consistent with the
limited findings in the human occupational studies. These include evidence of an effect of
tetrachloroethylene exposure on red blood cells [decreased RBCs (Ebrahim et al., 2001), or
decreased erythrocyte colony-forming units (Seidel et al., 1992)1, with reversible hemolytic
anemia observed in female mice exposed to low drinking water levels (0.05 mg/kg-day) of
tetrachloroethylene beginning at 2 weeks of age in one series of studies (Marth et al., 1989;
Marth, 1987: Marth et al., 1985a: Marth et al., 1985b). Ebrahim et al. (2001) also observed
decreased hemoglobin, platelet counts, and packed cell volume, and increased WBC counts. The
relative lack of additional data, including confirmatory reports of immunotoxic or hematologic
toxicity with low continuous exposures beginning in early lifestages, taken together with
evidence of immunotoxicity from structurally related solvents (Cooper et al., 2009), contributes
to uncertainty in the database for tetrachloroethylene. No human studies identified adverse
effects on the respiratory tract, and no lung toxicities in rodents were reported in chronic
bioassays (NTP, 1986; NCI, 1977) or other published reports.
6.1.3.1. Neurological Effects (refer to Section 4.1)
Human and animal studies provide complementary evidence regarding the association of
neurobehavioral deficits and tetrachloroethylene exposure. Tetrachloroethylene exposure in
humans has primarily been shown to affect visual function (including color vision) and
visuospatial memory and other aspects of cognition. Brain-weight changes have been measured
in animal studies. A more in-depth discussion of the human neurotoxicological studies can be
found in Section 4.1.1.3, and the animal inhalation and oral or i.p. exposure studies are discussed
in Sections 4.1.2.1 and 4.1.2.2, respectively.
6-5
-------
Visual contrast sensitivity deficits as well as color discrimination deficits are commonly
present prior to detectable pathology in the retina or optic nerve head. These deficits are, thus,
among the earliest signs of disease and potentially more sensitive measures than evoked
potentials from visual stimuli (Regan, 1989). Several independent lines of evidence can be
found in the occupational and residential exposure studies to support an inference of visual
deficits following chronic tetrachloroethylene exposure. The studies that observed effects on
color vision using the Lanthony D-15 color vision test include cross-sectional and longitudinal
designs in dry cleaning (Gobba et al., 1998; Cavalleri et al., 1994) and residential (Schreiber et
al., 2002) settings. Decrements in color confusion were reported among all workers exposed to a
mean TWA of 6 ppm for an average of 8.8 years (Cavalleri et al., 1994). A significant dose-
response relationship between CCI value and tetrachloroethylene concentration (p < 0.01) was
also observed in Cavalleri et al. (1994). As noted previously, the color vision testing in this
study was blinded to exposure level of the study participants, and the study participants were
well matched in terms of age, smoking, and alcohol use. A follow-up of these workers 2 years
later (Gobba et al., 1998) showed greater loss in color discrimination in those who were
subsequently exposed to a higher concentration [increase in geometric mean from 1.7 to 4.3
ppm], with no change in those exposed to lower concentrations [decrease in geometric mean
from 2.9 to 0.7 ppm]). Although Gobba et al. (1998) demonstrated persistent color confusion
effects in this follow up evaluation, the study exposures are not clearly characterized over the
course of the 2-year duration. Nakatsuka et al. (1992) did not observe an association with color
vision among dry cleaners in China (n = 64, geometric mean: TWA 11 and 15 ppm in females
and males, respectively), but the relative insensitivity of the specific type of color vision test
used in this study (Lanthony, 1978) is a likely explanation for these results. Effects on color
vision were also observed among 14 dry cleaners in the small study in Malaysia by
Sharanjeet-Kaur et al. (2004), but this study provides little weight to the strength of the evidence
because of the lack of exposure information (other than job title), and differences between dry
cleaners and controls regarding test conditions and smoking habits. Two other small studies also
reported lower scores on the Lanthony D-15 color vision test in much lower exposure settings,
but the differences were not statistically significant. A study of residents living above dry
cleaners (mean tetrachloroethylene exposure during active dry cleaning = 0.4 ppm), reported
mean CCI scores of 1.33 and 1.20 for 17 exposed and 17 controls, respectively (p = 0.26). A
study of workers in a day care center located in a building with a dry-cleaning business (mean
tetrachloroethylene exposure: 0.32 ppm) reported mean CCI scores of 1.22 and 1.18 in the
exposed daycare workers and controls, respectively (p = 0.39) (Schreiber et al., 2002). Overall,
the evidence reveals a high degree of consistency in this aspect of visually mediated function.
6-6
-------
Visual contrast sensitivity changes were reported in two NYSDOH residential studies. In
a small pilot study (4 children and 13 adults), mean scores for visual contrast sensitivity (using a
near vision visual contrast sensitivity test) across spatial frequencies were statistically
significantly lower in exposed residents than in controls, indicating poorer visual function in the
exposed groups (Schreiber et al., 2002). Controls were age- and sex-matched to the exposed
group, and both groups were English speaking and of predominately Caucasian ethnicity;
however, they were drawn from different geographic areas. In addition, two of the four exposed
children had diagnoses of learning disabilities or developmental delays, which could affect
performance on this type of test. In the larger study (NYSDOH. 2010. 2005a, b), the test
(Functional Acuity Contrast Test, FACT) assessed far vision visual contrast sensitivity, and the
test had a low rate of detecting visual contrast changes. For contrast vision, a number of
analyses in NYSDOH (Storm et al.. 2011 [previously reported in NYSDOH. 20101: NYSDOH.
suggest a vulnerability among children. However, exposure to >0.015 ppm (>100 |ig/m3)
tetrachloroethylene was highly correlated with race and children's age. Additionally, the sample
sizes in the highest exposure group, especially in higher income, nonminority groups, makes it
difficult to fully examine possible effects of income, race, and age on vision. Therefore, while
both studies report visual contrast sensitivity changes, with exposed children being more
sensitive, there are concerns with the methodological and analytic approaches in these studies.
Acute human exposure studies reported increased latencies of up to 3.0 ms in visual
evoked potentials (Altmann et al., 1990) and changes in EEGs (magnitude of effect was not
specified) (Hake and Stewart, 1977; Stewart et al., 1970) at higher exposures ranging from 340
to 680 mg/m3.
In rats, acute inhalation exposure to tetrachloroethylene results in significant changes to
the flash-evoked potential at 800 ppm (Mattsson et al., 1998) and a decrease in F2 amplitudes of
the steady state visual evoked potential at 250 ppm (Boyes et al., 2009). In a subchronic
exposure study (13 weeks, up to 800 ppm tetrachloroethylene), changes in flash-evoked potential
responses were not observed at tetrachloroethylene exposures up to 200 ppm. In the 800 ppm
group, there was a significant increase in the amplitude and a significant increase in latency
(-3.0 ms) of the mid-flash-evoked potential waveform (N3), but histopathological lesions were
not observed in the examination of the visual system brain structures [e.g., visual cortex; optic
nerve; Mattsson et al. (1998)1.
Effects on visuospatial memory in humans were also reported in each of the studies that
examined this measure (Altmann et al., 1995; Echeverria et al., 1995; Echeverria et al., 1994;
Seeber, 1989). These effects (increased response times or cognition errors) were observed in
occupational and residential studies, and the occupational studies were quite large, involving
101, 65, and 173 dry-cleaning workers in Seeber (1989), Echeverria et al. (1995), and Echeverria
6-7
-------
et al. (1994), respectively. Several different types of tests were used including digit reproduction
(Seeber, 1989), switching, pattern memory, and pattern recognition (Echeverria et al., 1995;
Echeverria et al., 1994), and the Benton test (Altmann et al., 1995). Exposure ranges for the
increased reaction time observations (LOAELs) ranged from 4.99 to 102 mg/m3 (Altmann et al.,
1995; Echeverria et al., 1995; Ferroni et al., 1992). The changes in the cognitive tasks were
observed at exposures (LOAELs) ranging from 53.9 to 364.22 mg/m3 (Spinatonda et al., 1997;
Echeverria et al., 1995; Seeber, 1989). All of these studies except Altmann et al. (1995) indicate
that the neurobehavioral assessment was blinded to knowledge of the exposure level of the
subject, and all of the studies adjusted for potentially confounding factors. It should be noted,
however, that residual confounding from education-level differences between exposed and
referent subjects may still be present in Altmann et al. (1995).
Increased reaction time, increased number of false alarms, and decreased trial
completions in a signal detection task (measures of decreased attention) were reported in an
acute (60 minutes) exposure (6,782 mg/m3 or higher) study in rats (Oshiro et al., 2008).
Additionally, operant tasks that test cognitive performance have demonstrated performance
deficits in rats and mice following acute tetrachloroethylene oral (Warren et al., 1996) and i.p.
(Umezu et al., 1997) exposures. These findings are consistent with observed effects on cognition
and memory in humans. However, no studies, to date, have evaluated the persistent effects of
tetrachloroethylene exposure on cognitive performance deficits in animal models.
An occupational exposure study (n=60) (Ferroni et al., 1992) and a residential exposure
study (n = 14) (Altmann et al., 1995), with mean exposure levels of 15 and 0.7 ppm,
respectively, reported significant increases in simple reaction time of 24 ms (11% increase)
(Ferroni etal., 1992) and 40 and 51.1 ms (15 and 20% increases", respectively", for two separate
measurements) (Altmann et al., 1995) for the exposed subjects. A third study, Lauwerys et al.
(1983), reported better performance on simple reaction time in 21 exposed workers (mean TWA:
21 ppm) compared with controls when measured before a work shift but not when measured
after work.
The changes in brain weight, DNA/RNA, and neurotransmitter levels that were observed
in the animal studies are highly supportive of the neurobehavioral changes observed with
tetrachloroethylene exposure. Changes in brain DNA, RNA, or protein levels and lipid
composition were altered following inhalation, with changes observed in cerebellum,
hippocampus, and frontal cortex (Wang et al., 1993; Rosengren et al., 1986; Savolainen et al.,
1977a: Savolainen et al., 1977b). The replication of these changes in biochemical parameters
and effects in brain weight in both rats and gerbils is pathognomonic. Changes in
neurotransmitters systems (Briving et al., 1986; Honmaetal., 1980a: Honmaetal., 1980b) and
6-8
-------
circadian rhythm (Motohashi etal., 1993) in animal studies are consistent with neuroendocrine
alterations observed in humans (Ferroni et al., 1992).
In conclusion, the weight of evidence across the available studies of humans and animals
exposed to tetrachloroethylene indicates that chronic exposure to tetrachloroethylene can result
in decrements in color vision, visuospatial memory, and possibly other aspects of cognition and
neuropsychological function, including reaction time.
6.1.3.2. Summary of Other Noncancer Adverse Effects (refer to Sections 4.2, 4.3, 4.6, and
4.7)
In addition to evidence of toxicity to the central nervous system, tetrachloroethylene has
been shown to adversely affect the kidney, liver, immune and hematologic systems, as well as
development and reproduction. The human and animal evidence for these effects is summarized
in the paragraphs below.
6.1.3.2.1. Kidney Toxicity (refer to Section 4.2)
The human evidence for kidney effects is limited. Most available reports do not include
information on the standard battery of tests for kidney function and only one study (Calvert et al.,
2011) reported on end-stage renal disease (ESRD). However, an association between
tetrachloroethylene exposure via inhalation and chronic kidney disease is supported by evidence
of urinary excretion of renal proteins (Verplanke et al., 1999; Mutti et al., 1992) and higher
ESRD, particularly hypertensive ESRD, with higher exposures (Calvert et al., 2011). Mutti et al.
(1992) reported statistically significant increases in retinol binding protein, P2u-globulin, and
albumin in urine among dry cleaners as compared with matched controls. In addition, for
seven different urinary markers, the prevalence of individuals with abnormal values (>95th
percentile of controls) was four- to fivefold greater in the exposed group. Adverse effects on the
kidney have been observed in studies of animals exposed to high concentrations of
tetrachloroethylene by inhalation (JISA, 1993: NTP, 1986). oral gavage (Ebrahim et al., 2001:
Ebrahim et al.. 1996: Jonkeretal.. 1996: Green etal.. 1990: Goldsworthv et al.. 1988: NCI.
1977) and by intraperitoneal injection of tetrachloroethylene metabolites (Elfarra and Krause,
2007). The nephrotoxic effects include increased kidney-to-body weight ratios, hyaline droplet
formation, glomerular "nephrosis," karyomegaly (enlarged nuclei), cast formation, and other
lesions or indicators of renal toxicity. Overall, multiple lines of evidence support the conclusion
that tetrachloroethylene causes nephrotoxicity in the form of tubular toxicity, mediated
potentially through GSH conjugation products. Limitations to the database include the lack of
human studies investigating drinking water or other oral tetrachloroethylene exposures on kidney
toxicity.
6-9
-------
6.1.3.2.2. Liver Toxicity (refer to Section 4.3)
Two of four studies of occupationally exposed dry cleaners showed early indications of
liver toxicity, namely sonographic changes of the liver and altered serum concentrations of
one liver enzyme indicative of liver injury (Brodkin et al., 1995; Gennari et al., 1992). Frank
liver disease was not observed among these workers, nor were changes in other biomarkers
indicative of liver toxicity (e.g., serum transaminases), not unexpected given that subjects with
signs of liver disease were excluded in both studies. Liver toxicity was reported in multiple
animal species exposed to tetrachloroethylene via inhalation and oral routes of exposure. The
effects were characterized by increased liver weight, fatty changes, necrosis, inflammatory cell
infiltration, triglyceride increases and proliferation (Philip et al., 2007; Ebrahim etal., 1996;
Jonkeretal.. 1996: Bermanetal.. 1995: JISA, 1993: Odumetal.. 1988: Goldsworthy and Popp,
1987: NTP, 1986: Buben and O'Flahertv, 1985: Kj ell strand et al.. 1984: Schumann et al.. 1980).
6.1.3.2.3. Iniiminologic and Hematopoietic Toxicity (refer to Section 4.6)
The strongest human study examining immunologic and hematologic effects of
tetrachloroethylene exposure in terms of sample size and use of an appropriately matched control
group is the study of 40 male dry-cleaning workers (mean exposure levels <140 ppm; mean
duration 7 years; mean blood tetrachloroethylene levels 1,685 |ig/L) by Emara et al. (2010).
Statistically significant decreases in red blood cell count and hemoglobin levels and increases in
total white cell counts and lymphocyte counts were observed in the exposed workers compared
to age- and smoking-matched controls. Similar effects were observed in mice (Ebrahim et al.,
2001). In addition, increases in several other immunological parameters, including T
lymphocyte and natural killer cell subpopulations, IgE, and interleukin-4 levels were observed in
tetrachloroethylene-exposed dry-cleaning workers (Emara et al., 2010). These immunologic
effects suggest an augmentation of Th2 responsiveness. The available data from experimental
studies assessing immunotoxic responses in animals are limited (Hanioka et al., 1995b:
Germolec et al., 1989: Aranyi et al., 1986), with one study (Aranyi et al., 1986) suggesting that
short-term exposures may result in decreased immunological competence (immunosuppression)
in CD-I mice. The limited laboratory animal studies of hematological toxicity demonstrated an
effect on red blood cells [decreased RBC (Ebrahim et al., 2001), or decreased erythrocyte colony
forming units (Seidel et al., 1992)], with reversible hemolytic anemia observed in female mice
exposed to low drinking water levels (0.05 mg/kg-day) of tetrachloroethylene beginning at
2 weeks of age in one series of studies (Marth et al., 1989: Marth, 1987: Marth et al., 1985a:
Marth et al., 1985b). Ebrahim et al. (2001) also observed decreased hemoglobin, platelet counts
and packed cell volume, and increased WBC counts. The results of these studies, while limited,
support the human epidemiology studies. Additional data from inhalation, oral, and dermal
6-10
-------
exposures of different durations are needed to assess the potential immunotoxicity of
tetrachloroethylene along multiple dimensions—including immunosuppression, autoimmunity,
and allergic sensitization. The relative lack of additional data, including confirmatory reports of
immunotoxic or hematologic toxicity with low continuous exposures beginning in early
lifestages, taken together with evidence of immunotoxicity from structurally related solvents
(Cooper et al., 2009), contributes to uncertainty in the database for tetrachloroethylene.
6.1.3.2.4. Reproductive Toxicity (refer to Section 4.7)
The epidemiologic database is inconclusive concerning potential effects of
tetrachloroethylene exposure on spermatogenesis, menstruation, fertility or delayed conception
(Sallmen et al.. 1998: Sallmen et al.. 1995: Eskenazi et al.. 199la: Eskenazi et al.. 1991b:
Zielhuis et al., 1989: Rachootin and Olsen, 1983). One study of primarily unionized workers in
the dry-cleaning and laundry industries in California observed subtle deficits in sperm quality in
relation to increasing levels of three measures of exposure, including tetrachloroethylene in
exhaled breath (Eskenazi et al., 199la). This observation is supported by one report of abnormal
sperm in mice (Beliles et al., 1980). Several studies of maternal occupational exposure to
tetrachloroethylene suggest an increased risk of spontaneous abortion, particularly at higher
levels (Doyle etal.. 1997: Windham et al.. 1991: Lindbohm et al.. 1990: Olsen etal.. 1990:
Kvyronen et al., 1989), but other studies did not report an association with maternal (Ahlborg,
1990a: Olsen etal., 1990) or paternal (Eskenazi etal., 199 la: Lindbohm et al., 1991: Taskinen et
al., 1989) exposure. Some studies observed an increased odds ratio ranging from 1.4 to 4.7, but
risk estimates were statistically imprecise and some studies were limited in their ability to
evaluate potential confounding (Windham et al., 1991: Lindbohm et al., 1990: Olsen etal., 1990:
Bosco et al., 1987). In general, the studies that used a more precise definition of exposure, or
categorized exposure into levels of increasing dose or intensity, observed higher risk estimates
(Doyle et al., 1997: Lindbohm et al., 1990: Olsen et al., 1990: Kyyronen et al., 1989). No
associations with incidence of spontaneous abortion were observed among two populations
exposed to tetrachloroethylene in drinking water, although the window of exposure used to
assess risk in both studies may not have had been precise enough to detect a small elevation in
risk (Aschengrau et al., 2009a: Aschengrau et al., 2008: Lagakos et al., 1986). The finding of
spontaneous abortions in several human studies of dry cleaners is supported by the occurrence of
reduced birth weight and mortality in several animal studies (Carney et al., 2006: Szakmary et
al., 1997: Nelson etal., 1979: Schwetz et al., 1975) [and in the Fl generation but not the F2
generation of Tinston (1994)1.
6-11
-------
6.1.3.2.5. Developmental Toxicity (refer to Section 4.7)
Stillbirths, congenital anomalies, or decreased birth weight were not associated with
maternal or paternal occupational exposure to tetrachloroethylene in several epidemiologic
studies (Lindbohm, 1995: Windham et al.. 1991: Olsenetal.. 1990: Kyyronen et al.. 1989:
Taskinen et al., 1989: Bosco et al., 1987). However, the studies analyzed congenital anomalies
in a combined category, and the number of exposed cases for specific types of anomalies was not
sufficient to evaluate risk with statistical precision. Some studies of tetrachloroethylene in
drinking water reported that exposure during pregnancy is associated with low birth weight
(Bove et al., 1995: Lagakos et al., 1986), eye/ear anomalies (Lagakos et al., 1986), and oral clefts
(Aschengrau et al., 2009b: Bove et al., 1995: Lagakos et al., 1986). No associations with
prenatal tetrachloroethylene exposure in drinking water were reported for small for gestational
age (Aschengrau et al., 2008: Bove etal., 1995), other classifications of congenital anomalies
[e.g., musculoskeletal, cardiovascular (Lagakos et al., 1986)], or deficits in attention or
educational performance (Janulewicz et al., 2008). Although a small increase in risk of small for
gestational age was reported for infants exposed prenatally to tetrachloroethylene at the Camp
Lejeune military base (Sonnenfeld et al., 2001), the finding remains inconclusive until ATSDR
completes its reanalysis. Participants in some of the studies of drinking water contamination
were exposed to multiple pollutants (Bove etal., 1995: Lagakos etal., 1986), and it was not
possible to disentangle substance-specific risks. In animals, the developmental toxicity database
provides evidence of decreased prenatal survival, decreased fetal growth, delays in skeletal
ossification, and increased incidences of malformations following in utero exposure in rats, mice,
and/or rabbits (Carney et al., 2006: Szakmary et al., 1997: Narotsky and Kavlock, 1995: Schwetz
etal., 1975). The decreased survival and malformation findings in laboratory mammals were
supported by data from whole embryo culture (Saillenfait et al., 1995) and Japanese medaka
assays (Spencer et al., 2002: Saillenfait et al., 1995). Alterations in neurological function
following pre- and/or postnatal inhalation exposures to tetrachloroethylene were observed in rats
by Szakmary et al. (1997), Nelson et al. (1979), Fredriksson et al. (1993), and Tinston (1994).
These findings were supported by a study that found reductions in brain acetylcholine and
dopamine in rat offspring following gestational tetrachloroethylene exposures (Nelson et al.,
1979). Limitations of the inhalation developmental toxicity studies include the lack of
dose-response information due to the use of a single treatment level in the prenatal
developmental toxicity assessment by Schwetz et al. (1975): the lack of either maternal or
developmental toxicity in Hardin et al. (1981): and absence of methodological details in study
reporting (Szakmary et al., 1997).
6-12
-------
6.1.4. Carcinogenicity (refer to Section 4.10.2)
Following EPA (2005a) Guidelines for Carcinogen Risk Assessment., tetrachloroethylene
is "likely to be carcinogenic in humans by all routes of exposure." This characterization is based
on suggestive evidence of carcinogenicity in epidemiologic studies and conclusive evidence that
the administration of tetrachloroethylene, either by ingestion or by inhalation to sexually mature
rats and mice, increases tumor incidence (7ISA. 1993: NTP. 1986: NCL 1977).
Tetrachloroethylene increased the incidence of liver tumors (hepatocellular adenomas and
carcinomas) in male and female mice and of mononuclear cell leukemia (MCL) in both sexes of
rats. These findings were reproducible in multiple lifetime bioassays employing different rodent
strains and, in the case of mouse liver tumors, by inhalation and oral exposure routes. Additional
tumor findings in rats included significant increases in the NTP bioassay of testicular interstitial
cell tumors and kidney tumors in males, and brain gliomas in males and females. In mice,
hemangiosarcomas in liver, spleen, fat, and subcutaneous skin were reported in males in the
JISA study. The available epidemiologic studies provide a pattern of evidence associating
tetrachloroethylene exposure and several types of cancer, specifically bladder cancer,
non-Hodgkin lymphoma, and multiple myeloma. Associations and exposure-response
relationships for these cancers were reported in studies using higher quality (more precise)
exposure-assessment methodologies for tetrachloroethylene. Confounding by common lifestyle
factors such as smoking are unlikely explanations for the observed results. For other sites,
including esophageal, kidney, lung, liver, cervical, and breast cancer, more limited data are
available.
The specific active moiety(ies) and mode(s) of action involved in the carcinogenicity of
tetrachloroethylene and its metabolites are not fully characterized. For rat kidney tumors, it is
generally believed that metabolites resulting from GSH conjugation of tetrachloroethylene are
involved. The hypothesized modes of action for this endpoint include mutagenicity, peroxisome
proliferation, a2u-globulin nephropathy, and cytotoxicity not associated with a2u-globulin
accumulation. For mouse liver tumors, it is generally believed that metabolites resulting from
P450-mediated oxidation of tetrachloroethylene are involved. The mode of action (MO A)
hypotheses for this endpoint concern mutagenicity, epigenetic effects (especially DNA
hypomethylation), oxidative stress, and receptor activation (focusing on a hypothesized PPARa
activation MO A). However, the available evidence is insufficient to support the conclusion that
either rat kidney or mouse liver tumors are mediated solely by one of these hypothesized modes
of action. In addition, no data are available concerning the metabolites or the mechanisms that
may contribute to the induction of other rodent tumors (including mononuclear cell leukemia,
brain gliomas, or testicular interstitial cell tumors in exposed rats and hemangiosarcomas in
exposed mice). Furthermore, no mechanistic hypotheses have been advanced for the human
6-13
-------
cancers suggested to be increased with tetrachloroethylene exposure in epidemiologic studies,
including bladder cancer, non-Hodgkin lymphoma and multiple myeloma. Although
tetrachloroethylene is largely negative in genotoxicity assays—including in the Ames
mutagenicity test—tetrachloroethylene has been shown to induce modest genotoxic effects (e.g.,
micronuclei induction following in vitro or in vivo exposure, and DNA binding and single strand
breaks in tumor tissue) and mutagenic effects under certain metabolic activation conditions. In
addition, some tetrachloroethylene metabolites have been shown to be mutagenic. Thus, the
hypothesis that mutagenicity contributes to the tetrachloroethylene carcinogenesis cannot be
ruled out for one or more target organs, although the specific metabolic species or mechanistic
effects are not known.
6.1.5. Susceptibility (refer to Section 4.9)
There is some evidence that certain populations might be more susceptible to exposure to
tetrachloroethylene. Attributes that may increase susceptibility to tetrachloroethylene include
age, gender, race/ethnicity, genetics, preexisting disease, lifestyle factors, nutritional status,
socioeconomic status, and multiple exposures and cumulative risk. Although there is more
information on early life exposure to tetrachloroethylene than on other potentially susceptible
populations, there remain a number of uncertainties regarding childhood susceptibility.
Although inhalation of tetrachloroethylene is believed to be of most concern, pathways of
exposure for children are not well characterized. Tetrachloroethylene has been shown to pass
through the placenta in rodent studies (Szakmary et al., 1997; Ghantous et al., 1986), but the
extent to which this occurs in humans is not known. For some infants the primary route of
exposure may be through breast milk ingestion (refer to Sections 2.2.4 and 3.2), while for other
infants the dose received through ingestion of breast milk will become insignificant when
compared with inhalation exposure (Schreiber, 1997). The amount of tetrachloroethylene
ingested from food is not well described; and it is not known to what extent tetrachloroethylene
is absorbed by a child and to which organs tetrachloroethylene and its metabolites may be
distributed. The neurological effects of tetrachloroethylene may constitute the most sensitive
endpoints of concern for noncancer effects, and limited data show that early life-stages may be
more susceptible to visual deficits than are adults (Storm et al., 2011 [previously reported in
NYSDOH, 20101: NYSDOH, 2005a: Schreiber et al., 2002), yet developmental neurotoxic
effects, particularly in the developing fetus, need further evaluation using age-appropriate testing
for assessment. There are a number of adverse health effects observed uniquely in early
lifestages, with no comparable observations in adults to determine relative sensitivity (e.g., birth
outcomes, autism, or allergy); conversely, there are some adverse outcomes that have been
observed only in adults.
6-14
-------
There is suggestive evidence that there may be greater susceptibility among the elderly,
but the available data are much more limited with related uncertainties. Improved PBPK
modeling that contains physiologic parameter information for infants and children (including, for
example, the effects of maternal inhalation exposure and the resulting concentration in breast
milk) and for older adults, and validation of these models, will aid in determining differences in
life stage toxicokinetics of tetrachloroethylene. The differences reported in the literature may
reflect a true difference in susceptibility by life stage, an incomplete assessment of these
outcomes in all life-stages, or latent outcomes associated with earlier exposure. More studies
specifically designed to evaluate effects in early and later life-stages are needed in order to more
fully characterize potential life stage-related tetrachloroethylene toxicity.
For other susceptibility factors, the data are more limited and based mainly on
nonchemical specific data that provides information on variation in physiology, exposure, and
toxicokinetics. Until quantitative conclusions can be made for each susceptibility factor, it will
be difficult to consider the impacts of changes in multiple susceptibility factors. In addition,
further evaluation of the effects of aggregate exposure to tetrachloroethylene from multiple
routes and pathways is needed. Similarly, the effects due to coexposures to other compounds
with similar or different MO As need to be evaluated.
6.2. DOSE-RESPONSE ASSESSMENT
This section summarizes the major conclusions of the dose-response analysis for
tetrachloroethylene noncancer effects and carcinogenicity, with more detailed discussions in
Section 5.
6.2.1. Noncancer Effects (refer to Section 5.1)
The database of human and animal studies on inhalation toxicity of tetrachloroethylene is
adequate to support derivation of inhalation and oral reference values. A number of targets of
toxicity from chronic exposure to tetrachloroethylene have been identified in published animal
and human studies. These targets include the central nervous system, kidney, liver, immune and
hematologic system, and development and reproduction. In general, neurological effects were
judged to be associated with lower tetrachloroethylene exposures.
6.2.1.1. Selection of Principal Studies and Critical Effect (refer to Section 5.1.1)
The evidence for human neurotoxicity includes 12 well-conducted epidemiological
studies of tetrachloroethylene exposure. Of these, seven examined occupational exposure (i.e.,
Schreiber et al.. 2002: Gobbaetal.. 1998: Spinatonda et al.. 1997: Echeverria et al.. 1995:
Cavalleri et al., 1994: Ferroni et al., 1992: Seeber, 1989), several examined residential exposure
6-15
-------
(i.e., Storm etal.. 2011 [previously reported in NYSDOH, 20101: Schreiber et al.. 2002: Altmann
et al., 1995) and two were acute-duration experimental chamber studies (i.e., Altmann et al.,
1990: Hake and Stewart, 1977). The animal database comprises acute-duration and
sub chronic-duration studies of the effects of tetrachloroethylene on functional neurological
endpoints (functional observation battery, motor activity) (i.e., Oshiro et al., 2008: Kj ell strand et
al., 1985), on sensory system function as assessed by evoked potential (i.e., Boyes et al., 2009:
Mattsson et al., 1998), or pathological changes in the brain (i.e., Wang et al., 1993).
Principal study selection from these candidate studies of central nervous system effects
involved evaluation of study characteristics as identified in Table 5-2. To summarize, human
studies are preferred to animal studies, as are studies of chronic duration and in residential
settings, if available and of adequate quality. Study quality attributes evaluated include the
comparability of study populations and the quality of exposure information (in human studies),
and effect measurements. Three human studies—Seeber (1989), Cavalleri et al. (1994), and
Echeverria et al. (1995)—were considered to be more methodologically sound based on study
quality attributes, including study population selection, exposure measurement methods, and
endpoint measurement methods listed in Table 5-2. However, the NRC (2010) characterized the
Seeber (1989) study as having discrepant results based on worse mean test scores (for neurologic
signs, emotional lability, choice reaction time, cancellation d2 and digit symbol) in the low-
compared with high-exposure group. Therefore, Seeber (1989) was not among those
recommended by NRC for consideration in deriving the RfC.
NRC (2010) recommended five studies for consideration in deriving the RfC (Altmann et
al., 1990: Boves et al., 2009: (Echeverria et al., 1995): Cavalleri et al., 1994: and Gobba et al.,
1998). Two acute studies recommended for consideration by NRC [the human chamber study of
Altmann et al. (1990) and the rodent study of Boyes et al. (2009)] were judged by EPA to be
supportive, but were not considered further for deriving candidate RfCs because EPA gives
preference to quality studies of chronic, human exposures over studies of acute exposures. In
addition, two of the other studies recommended by NRC (2010), Cavalleri et al. (1994), and
Gobba et al. (1998), evaluated the same cohort, and the earlier study was preferred by EPA due
to its use of a control group and the clearer identification of a POD (refer to section 5.1.1.3.2).
Thus, two studies—Cavalleri et al. (1994) and Echeverria et al. (1995)—are considered principal
studies by EPA for the RfC. Endpoints selected for the candidate RfCs were reaction time
measures (Echeverria et al., 1995), cognitive changes (Echeverria et al., 1995), and visual
function changes (Cavalleri et al., 1994).
6-16
-------
6.2.1.2. Uncertainties and Application of Uncertainty Factors (UFs) (refer to Sections 5.1.3,
5.2.3)
For the studies from which candidate reference values were derived, it was determined
that PODs could not be derived using dose-response modeling, due to no control group
(Echeverria et al., 1995) or lack of an important covariate (age) (Cavalleri et al., 1994). Each of
the candidate studies provided lowest-observed-adverse-effect levels (LOAELs) that were
selected as PODs. The adjusted LOAELs are as follows: 56 mg/m3 [for either visual
reproduction, pattern memory, and pattern recognition, or reaction time in pattern memory in
Echeverria et al. (1995)1 and 15 mg/m3 [for color confusion in Cavalleri et al. (1994)1. The
application of uncertainty factors is based on EPA's A Review of the Reference Dose and
Reference Concentration Processes [(U.S. EPA, 2002): Section 4.4.5], which address five areas
of uncertainty. No adjustment of the PODs was needed for animal-to-human extrapolation
uncertainty. Additionally, no adjustment was needed for subchronic-to-chronic uncertainty
because these studies involved chronic exposures. An overall uncertainty factor of 1,000 was
applied to each selected POD, comprised of the following uncertainty factors (UFs): An UF of
10 was applied to account for human variability in the effects that were used for the derivation of
the RfC. An UF of 10 was applied for the extrapolation from a LOAEL to a NOAEL because
the PODs from the studies were LOAELs. An UF of 10 was applied to address the lack of data
to adequately characterize the hazard and dose response in the human population. The following
critical data gaps have been identified: uncertainties associated with database deficiencies on
neurological, developmental, and immunological effects.
6.2.1.3. Reference Concentration (refer to Section 5.1.3)
An uncertainty factor of 1,000 was applied to the PODs for the following endpoints from
the two principal neurotoxicological studies: color vision changes (Cavalleri et al., 1994): and
cognitive and reaction time changes (Echeverria et al., 1995). The candidate RfCs derived from
these endpoints span a range from 0.015 to 0.056 mg/m3. The RfC for tetrachloroethylene is
0.04 mg/m3, the midpoint of this range rounded to one significant figure.
A confidence level of high, medium, or low is assigned to the study used to derive the
RfC, the overall database, and the RfC itself, as described in Section 4.3.9.2 of EPA's Methods
for Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry
(U.S. EPA, 1994). The overall confidence in the RfC is medium. Although the confidence in the
evidence of neurotoxicological hazard is high, the estimates from studies for which candidate
RfCs were calculated are of medium confidence. These studies were considered to be
methodologically sound based on study quality attributes, including study population selection,
exposure measurement methods, and endpoint measurement methods. Other strengths are that
6-17
-------
they are human studies of chronic duration, obviating the need for extrapolation across species
and exposure duration. However, high confidence was not attained for the studies for which
candidate RfCs were calculated because they identified a LOAEL rather than a NOAEL, and
dose-response modeling could not be used for POD derivation due to lack of sufficient data [e.g.,
no control group (Echeverria et al., 1995) or lack of an important covariate (age) (Cavalleri et al.,
1994)]. Additionally, the studies for which candidate RfCs were calculated are of occupationally
exposed subjects; no data concerning potential susceptibility or variability among subjects were
available.
Medium confidence in the database is based on a number of limitations of both the
human and animal literature. Regarding neurotoxicity, there is a need for high quality
epidemiologic studies of residential exposures and chronic-duration animal studies (including in
developing animals). A fuller characterization is also needed of the noncancer effects other than
the critical effect of neurotoxicity, particularly immunological and hematological effects.
6.2.1.4. Reference Dose (refer to Section 5.2)
Candidate RfDs for tetrachloroethylene were developed through route-to-route
extrapolation from the inhalation PODs from two neurotoxicological studies of occupational
tetrachloroethylene exposure (i.e., Echeverria et al., 1995; Cavalleri etal., 1994). First, the
harmonized PBPK model of Chiu and Ginsberg (2011) was used to predict the
tetrachloroethylene in blood area under the curve (AUC) at the inhalation PODs from the two
principal studies. Then, using the same PBPK model, the oral equivalent POD for each study
was derived as the oral dose that would result in the same tetrachloroethylene in blood AUC.
Although it is not clear if the noncancer effects observed in humans are the result of
tetrachloroethylene itself and/or one or more metabolites, it is reasonable to assume that the
appearance of tetrachloroethylene in the blood is a step in the toxicity pathway. Moreover, the
sensitivity to the choice of dose metric for route-to-route extrapolation is low, with alternative
dose metrics giving route-to-route conversions within 1.4-fold of the conversion based on
tetrachloroethylene in blood. The resulting PODs were 2.6 mg/kg-day (Cavalleri et al., 1994)
and 9.7 mg/kg-day (Echeverria et al., 1995), respectively, for each of the critical endpoints. The
composite UF of 1,000 that was used for the RfC derivation as described above was applied to
each of these PODs. Candidate RfDs span a range from 2.6 x io~3 to 9.7 x io~3 mg/kg-day. The
RfD for tetrachloroethylene is 6 x 10~3 mg/kg-day, the midpoint of this range rounded to one
significant figure. This RfD is equivalent to a drinking water concentration of 0.21 mg/L,
assuming a body weight of 70 kg and a daily water consumption of 2 L.
A confidence level of high, medium, or low is assigned to the study used to derive the
RfD, the overall database, and the RfD itself, as described in Section 4.3.9.2 of EPA's Methods
6-18
-------
for Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry
(U.S. EPA, 1994). The overall confidence in the RfD is medium. Although the confidence in the
evidence of neurotoxicological hazard is high, the estimates from studies for which candidate
RfDs were calculated are of medium confidence. These studies were considered to be
methodologically sound based on study quality attributes, including study population selection,
exposure measurement methods, and endpoint measurement methods. Other strengths are that
they are human studies of chronic duration, obviating the need for extrapolation across species
and exposure duration. However, high confidence was not attained for the studies for which
candidate RfDs were calculated because they identified a LOAEL rather than a NOAEL, and
dose-response modeling could not be used for POD derivation due to lack of sufficient data [e.g.,
no control group (Echeverria et al., 1995) or lack of an important covariate (age) (Cavalleri et al.,
1994)1. Additionally, the studies for which candidate RfDs were calculated are of occupationally
exposed subjects; no data concerning potential susceptibility or variability among subjects were
available. Because of the adequacy of the PBPK model (Chiu and Ginsberg, 2011) for
extrapolating from inhalation to oral exposures, the use of inhalation studies for deriving the RfD
did not decrease confidence.
Medium confidence in the database is based on a number of limitations of both the
human and animal literature. Regarding neurotoxicity, there is a need for high quality
epidemiologic studies of residential exposures and chronic-duration animal studies (including in
developing animals). A fuller characterization is also needed of the noncancer effects other than
the critical effect of neurotoxicity, particularly immunological and hematological effects.
6.2.1.5. Dose-Response Analyses for Noncancer Effects Other Than Critical Effect of
Neurotoxicity (refer to Sections 5.1.4 and 5.2.4)
Inhalation and oral dose-response analyses for noncancer effects other than the critical
effect of neurotoxicity were also conducted. The purpose of these analyses is twofold: (1) to
provide a quantitative characterization of the relative sensitivity of different organs/systems to
tetrachloroethylene, and (2), to provide information that may be useful for cumulative risk
assessment in which multiple chemicals have a common target organ/system other than the
central nervous system. The method of analysis is analogous to that described above for
neurotoxicity, using the NOAEL/LOAEL approach and the application of uncertainty factors to
studies of kidney, liver, immunologic and hematologic, and reproductive and developmental
toxicity. Specifically, human equivalent concentrations [HECs] and human equivalent doses
[HEDs] are derived using either (1) for inhalation exposure, the RfC methodology for a
category 3 gas with extrarespiratory effects, adjusted for equivalent continuous exposure; (2) for
oral exposure, mg/kg-day dose adjusted for equivalent continuous exposure; or (3) for either
6-19
-------
route of exposure, the PBPK model with an appropriate dose metric. The HECs and HEDs are
then treated as PODs to which uncertainty factors are applied.
The sample values for renal and hematologic toxicity overlap with the range of values
based on the critical effect of neurotoxicity. Specifically, for renal effects, the resulting values
range from 0.03-0.2 mg/m3 for inhalation and 0.005-0.03 mg/kg-day for oral exposure, based
on effects in chronically exposed mice and rats (JISA, 1993) and occupationally exposed humans
(Mutti et al., 1992). For hematologic toxicity, the resulting values were 0.04 mg/m3 for
inhalation and 0.007 mg/kg-day for oral exposure, based on changes in hematological measures
in occupationally exposed humans (Emara et al., 2010). These overlap with the ranges of
0.02-0.06 mg/m3 for inhalation and 0.003-0.01 mg/kg-day for oral exposure based on the
critical effect of neurotoxicity, and thereby provide additional support for the RfC and RfD. The
sample values from the other endpoints are up to 20-fold greater than the RfC, and up to 10-fold
greater than the RfD. This suggests that multiple effects may occur at about the same exposure
levels at which tetrachloroethylene begins to induce neurotoxicity. These results suggest that it
is important to take into account effects from tetrachloroethylene other than neurotoxicity,
particularly when assessing the cumulative effects of multiple exposures.
6.2.2. Cancer (refer to Section 5.2)
As summarized above, following EPA (2005a) Guidelines for Carcinogen Risk
Assessment., tetrachloroethylene is characterized as "Likely to be carcinogenic to humans" by all
routes of exposure based on suggestive epidemiologic evidence of carcinogenicity and
conclusive evidence of carcinogenicity in mice and rats. No available epidemiologic studies of
cancer were found to be suitable for dose-response modeling. Therefore, cancer risk estimation
is based on data from rodent bioassays. Because the mode(s) of action for tetrachloroethylene
carcinogenicity has not been adequately established, the tumors reported in rodent bioassays are
considered relevant to humans and a low-dose linear extrapolation is used to estimate human
cancer risk from rodent dose-response data, in accordance with EPA's Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005a).
Chronic studies in rats and mice include an oral gavage study in mice and rats by NCI
(1977) and two inhalation studies in mice and rats (JISA. 1993: NTP, 1986). The NCI (1977) rat
and mouse oral gavage study had a number of limitations that made it less suitable for
dose-response modeling as compared to the other studies, including significantly higher early
noncancer morbidity and mortality in treated groups, a variable dosing schedule, and the study
duration was substantially less than the other available bioassays. With respect to the other
two bioassays, the JISA (1993) bioassay included lower exposures of both mice and rats than the
NTP (1986) study, and it included three exposure groups as compared to two exposure groups in
6-20
-------
the NTP (1986) study. Therefore, JISA (1993) provides a stronger basis for deriving
dose-response relationships for risk assessment purposes, insofar as all other aspects of these
studies can be considered comparable. Thus, for endpoints which were reported to be
tetrachloroethylene-related in multiple studies—i.e., liver tumors and MCLs—the JISA (1993)
study was used for dose-response modeling. The JISA (1993) bioassay was also used for
dose-response modeling of the increased hemangiomas and hemangiosarcomas in male mice
because it was the only bioassay that reported this tumor type. The NTP (1986) study was
utilized for modeling the increased incidence of renal cancers, brain cancers, and testicular
tumors in male rats, which were reported only in this bioassay. In male mice and male rats,
multiple treatment-related tumors were reported in the same study [(JISA, 1993) and (NTP,
1986), respectively]; thus, dose-response analyses of the combined risk of multiple tumors for
those experiments were also conducted.
The harmonized PBPK model of Chiu and Ginsberg (2011) was used to perform the
interspecies extrapolation from rodents to humans, and for route-to-route extrapolation of the
inhalation bioassay results to oral exposures. The choice of the preferred dose-metric to use for
each endpoint was based on the strength of its association with the toxic moiety relevant to the
endpoint and an evaluation of uncertainties in the calculation of that dose-metric. For cancer,
total rate of oxidative metabolism in the liver was considered the most relevant dose metric for
tetrachloroethylene-induced liver tumors, and AUC of the parent compound in the blood was
considered the preferred dose metric for all other sites, including MCL. Alternative dose-metrics
were also used for the purposes of comparison. These include the AUC of TCA in the liver for
mouse liver tumors and the rate of GSH conjugation for rat kidney tumors.
6.2.2.1. Inhalation Unit Risk
Several animal tumor data sets were analyzed for estimating cancer risk values. The
majority of the NRC peer review panel recommended that the mouse hepatocellular tumors be
used for cancer risk estimation. Therefore, the inhalation unit risk is 2 x 10~3 per ppm or
3 x io~7 per ug/m3, based on the male mouse hepatocellular tumor data from the JISA (1993)
bioassay. The unit risk should not be used with exposures exceeding 60 ppm, or 400 mg/m3 (the
equivalent ambient exposures corresponding to the POD for male mouse hepatocellular tumors),
because above this exposure level, the dose-response relationship is not linear, and the unit risk
would tend to overestimate risk. Some members of the NRC peer review panel recommended
that the MCL data be used for cancer risk estimation. The inhalation unit risk would be 7 x 10~2
per ppm, or 1 x io~5 per |ig/m3 if it were based on the male and female rat MCL data from the
JISA (1993) bioassay.
6-21
-------
6.2.2.2. Oral Slope Factor
The oral slope factor was developed from inhalation data because the only available oral
bioassay had several limitations for extrapolating to lifetime risk in humans (see also
Section 5.3.1). Route-to-route extrapolation from the inhalation PODs developed from the JISA
study (see Table 5-18) was carried out using the harmonized PBPK model (Section 3.5). The
oral slope factor is 2 x 10~3 per mg/kg-day based on the male mouse hepatocellular tumor data
from the JISA (1993) bioassay. The unit risk should not be used with exposures exceeding 50
mg/kg-day (the equivalent ambient exposure corresponding to the POD for male mouse
hepatocellular tumors), because above this exposure level, the dose-response relationship is not
linear, and the slope factor would tend to overestimate risk. The oral slope factor would be
6 x 1CT2 per mg/kg-day if it were based on the male and female rat MCL data from the JISA
(1993) bioassay.
6.2.2.3. Uncertainties in Cancer Dose-Response Assessment (refer to Section 5.3.5)
A number of uncertainties underlie the cancer unit risk for tetrachloroethylene, including
the choice of study, PBPK modeling and dose metrics, cross-species scaling, low-dose
extrapolation, model uncertainty, statistical uncertainty in the POD, the species/gender/tumor
type combination selected, and sensitive subpopulations. Some suggest risks could be higher
than was estimated (e.g., selection of MCL rather than mouse liver tumors, sensitive
subpopulations), while others would decrease risk estimates (e.g., use of central tendency instead
of lower 95% confidence bound on the POD), or have an impact of an uncertain direction.
Several uncertainties are quantitatively characterized for the significantly increased rodent
tumors. These include the statistical uncertainty in the POD, the range of uncertainty in PBPK
modeling and dose metrics, dose-response model uncertainty, and the species/gender/tumor type
combination selected. The latter three of these could either increase or decrease risk estimates.
Due to limitations in the data, particularly regarding the mode of action and relative human
sensitivity and variability, the quantitative impact of other uncertainties, which may have equal
or greater impact, has not been explored.
6-22
-------
7. REFERENCES
'tMannetje, A; Kogevinas, M; Chang-Claude, J: Cordier, S; Gonzalez, CA; Hours, M; Jockel
KH; Bolm-Audorff U; Lynge, E; Porru, S; Donato, F; Ranft, U; Serra, C; Tzonou, A;
Vineis, P; Wahrendorf J: Boffetta, P. (1999). Occupation and bladder cancer in European
women. Cancer Causes Control 10: 209-217.
http://dx.doi.Org/10.1023/A:1008852127139.
't Mannetje, A; Dryson, E; Walls, C: McLean, D; McKenzie, F; Maule, M: Cheng, S:
Cunningham. C: Kromhout H: Boffetta. P: Blair. A: Pearce. N. (2008). High risk
occupations for non-Hodgkin's lymphoma in New Zealand: Case-control study. Occup
Environ Med 65: 354-363. http://dx.doi.org/10.1136/oem.2007.035014.
Adgate, J: Eberly, L; Stroebel C: Pellizzari, E: Sexton, K. (2004a). Personal, indoor, and
outdoor VOC exposures in a probability sample of children. J Expo Anal Environ
Epidemiol 14: S4-S13. http://dx.doi.org/10.1038/si.iea.7500353.
Adgate, JL; Church, TR; Ryan, AD; Ramachandran, G: Fredrickson, AL; Stock, TH; Morandi,
MT; Sexton, K. (2004b). Outdoor, indoor, and personal exposure to VOCs in children.
Environ Health Perspect 112: 1386-1392. http://dx.doi.org/10.1289/ehp.7107.
Aggazzotti, G: Fantuzzi, G: Predieri, G: Righi, E: Moscardelli, S. (1994a). Indoor exposure to
perchloroethylene (PCE) in individuals living with dry-cleaning workers. Sci Total
Environ 156: 133-137. http://dx.doi.org/10.1016/0048-9697(94)90349-2.
Aggazzotti, G: Fantuzzi, G: Righi, E: Predieri, G: Gobba, FM; Paltrinieri, M: Cavalleri, A.
(1994b). Occupational and environmental exposure to perchloroethylene (PCE) in dry
cleaners and their family members. Arch Environ Occup Health 49: 487-493.
Ahlborg, G, Jr. (1990a). Pregnancy outcome among women working in laundries and dry-
cleaning shops using tetrachloroethylene. Am J Ind Med 17: 567-575.
http://dx.doi.org/10.1002/aiim.4700170503.
Ahlborg, G, Jr. (1990b). Validity of exposure data obtained by questionnaire. Two examples
from occupational reproductive studies. Scand J Work Environ Health 16: 284-288.
Alberati-Gianl D: Malherbe. P: Kohler. C: Lang. G: Kiefer. V: Lahm. HW: Cesura. AM. (1995).
Cloning and characterization of a soluble kynurenine aminotransferase from rat brain:
Identity with kidney cysteine conjugate [beta]-lyase. JNeurochem 64: 1448-1455.
http://dx.doi.0rg/10.1046/i.1471-4159.1995.64041448.x.
Albers, JW: Wald, JJ: Werner, RA; Franzblau, A; Berent, S. (1999). Absence of polyneuropathy
among workers previously diagnosed with solvent-induced toxic encephalopathy. J
Occup Environ Med 41: 500-509.
Albertini, S. (1990). Analysis of nine known or suspected spindle poisons for mitotic
chromosome malsegregation using Saccharomyces cerevisiae D61.M. Mutagenesis 5:
453-459.
Allen, BC: Crump, KS: Shipp, AM. (1988). Correlation between carcinogenic potency of
chemicals in animals and humans. Risk Anal 8: 531-544.
http://dx.doi.org/10.1111/i. 1539-6924.1988.tb01193.x.
Allen, JW: Collins, BW: Evansky, PA. (1994). Spermatid micronucleus analyses of
trichloroethylene and chloral hydrate effects in mice. Mutat Res Lett 323: 81-88.
http://dx.doi.org/10.1016/0165-7992(94)90049-3.
7-1
-------
Altmann, L; Bottger, A; Wiegand, H. (1990). Neurophysiological and psychophysical
measurements reveal effects of acute low-level organic solvent exposure in humans. Int
Arch Occup Environ Health 62: 493-499. http://dx.doi.org/10.1007/BF00381179.
Altmann, L; Wiegand, H; Bottger, A; Elstermeier, F; Winneke, G. (1992). Neurobehavioural and
neurophysiological outcomes of acute repeated perchloroethylene exposure. Appl
Psychol 41: 269-279. http://dx.doi.Org/10.llll/i.1464-0597.1992.tb00705.x.
Altmann, L: Neuhann, HF; Kramer, U; Witten, J: Jermann, E. (1995). Neurobehavioral and
neurophysiological outcome of chronic low-level tetrachloroethene exposure measured in
neighborhoods of dry cleaning shops. Environ Res 69: 83-89.
http://dx.doi.org/10.1006/enrs.1995.1028.
Amet Y: Berthou, F: Fournier. G: Dreano, Y: Bardou, L: Cledes, J: Menez, JF. (1997).
Cytochrome P450 4A and 2E1 expression in human kidney microsomes. Biochem
Pharmacol 53: 765-771. http://dx.doi.org/10.1016/50006-2952(96)00821-0.
Anders. MW: Lash. L: Dekant W: Elfarra, AA: Dohn, PR: Reed. DJ. (1988). Biosynthesis and
biotransformation of glutathione S-conjugates to toxic metabolites. CritRev Toxicol 18:
311-341. http://dx.doi.org/10.3109/10408448809037470.
Andersen, A; Barlow, L: Engeland, A; Kjaerheim, K; Lynge, E; Pukkala, E. (1999). Work-
related cancer in the Nordic countries. Scand J Work Environ Health 25: 1-116.
Anderson. BE: Zeiger, E: Shelby. MD: Resnick, MA: Gulati. DK: Ivett JL: Lovedav, KS.
(1990). Chromosome aberration and sister chromatid exchange test results with 42
chemicals. Environ Mol Mutagen 16: 55-137. http://dx.doi.org/10.1002/em.2850160505.
Anderson, LM; Diwan, BA; Fear, NT; Roman, E. (2000). Critical windows of exposure for
children's health: Cancer in human epidemiological studies and neoplasms in
experimental animal models. Environ Health Perspect 108: 573-594.
Anderson, WB; Board, PG: Gargano, B; Anders, MW. (1999). Inactivation of glutathione
transferase zeta by dichloroacetic acid and other fluorine-lacking [alpha]-haloalkanoic
acids. ChemRes Toxicol 12: 1144-1149. http://dx.doi.org/10.1021/tx9900851.
Andrys, C; Hanovcova, I; Chylkova, V: Tejral, J: Eminger, S; Prochazkova, J. (1997).
Immunological monitoring of dry-cleaning shop workers - exposure to
tetrachloroethylene. Cent Eur J Public Health 5: 136-142.
Anger. WK: Liang. YX: Nell V: Kang, SK: Cole. D: Bazylewicz-Walczak, B: Rohlman, PS:
Sizemore, OJ. (2000). Lessons learned - 15 years of the WHO-NCTB: A review
[Review]. Neurotoxicology 21: 837-846.
Anttila, A; Pukkala, E: Sallmen, M; Hernberg, S: Hemminki, K. (1995). Cancer incidence among
Finnish workers exposed to halogenated hydrocarbons. J Occup Environ Med 37: 797-
806.
Anttila, S: Hirvonen, A; Husgafvel-Pursiainen, K: Karjalainen, A; Nurminen, T; Vainio, H.
(1994). Combined effect of CYP1A1 inducibility and GSTM1 polymorphism on
histological type of lung cancer. Carcinogenesis 15: 1133-1135.
http://dx.doi.0rg/10.1093/carcin/15.6.1133.
Aranyi, C: O'Shea, WJ: Graham, JA; Miller, FJ. (1986). The effects of inhalation of organic
chemical air contaminants on murine lung host defenses. Fundam Appl Toxicol 6: 713-
720. http://dx.doi.org/10.1016/0272-0590(86)90184-3.
Aryal BK; Khuder, SA; Schaub, EA. (2001). Meta-analysis of systemic sclerosis and exposure
to solvents. Am J Ind Med 40: 271-274. http://dx.doi.org/10.1002/aiim.1098.
7-2
-------
Asal NR: Gever. JR: Risser. PR: Lee. ET: Kadamani. S: Cherng. N. (1988). Risk factors in
renal cell carcinoma. II. Medical history, occupation, multivariate analysis, and
conclusions. Cancer Detect Prev 13: 263-279.
Aschengrau, A; Ozonoff, D; Paulu, C: Coogan, P; Vezina, R: Heeren, T; Zhang, Y. (1993).
Cancer risk and tetrachloroethylene-contaminated drinking water in Massachusetts. Arch
Environ Health 48: 284-292.
Aschengrau, A; Paulu, C: Ozonoff, D. (1998). Tetrachloroethylene-contaminated drinking water
and the risk of breast cancer. Environ Health Perspect 106: 947-953.
Aschengrau, A; Seage, GR, III. (2003). Essentials of Epidemiology in Public Health. Sudbury,
MA: Jones & Bartlett Publishers.
Aschengrau, A; Rogers, S: Ozonoff, D. (2003). Perchloroethylene-contaminated drinking water
and the risk of breast cancer: Additional results from Cape Cod, Massachusetts, USA.
Environ Health Perspect 111: 167-173.
Aschengrau, A; Weinberg, J: Rogers, S: Gallagher, L; Winter, M; Vieira, V: Webster, T;
Ozonoff D. (2008). Prenatal exposure to tetrachloroethylene-contaminated drinking
water and the risk of adverse birth outcomes. Environ Health Perspect 116: 814-820.
http://dx.doi.org/10.1289/ehp.10414.
Aschengrau, A; Weinberg, JM: Gallagher, LG: Winter, MR; Vieira, VM; Webster, TF: Ozonoff
DM. (2009a). Exposure to tetrachloroethylene-contaminated drinking water and the risk
of pregnancy loss. Water Qual Expo Health 1: 23-34. http://dx.doi.org/10.1007/sl2403-
009-0003-x.
Aschengrau, A; Weinberg, JM: Janulewicz, PA; Gallagher, LG: Winter, MR; Vieira, VM:
Webster, TF: Ozonoff, DM. (2009b). Prenatal exposure to tetrachloroethylene-
contaminated drinking water and the risk of congenital anomalies: A retrospective cohort
study. Environ Health 8: 44. http://dx.doi.org/10.1186/1476-069X-8-44.
Ashbv, J: Brady. A: Elcombe, CR: Elliott BM: Ishmael J: Odum, J: Tugwood, JD: Kettle. S:
Purchase, IF. (1994). Mechanistically-based human hazard assessment of peroxisome
proliferator-inducedhepatocarcinogenesis. HumExp Toxicol 13: SI-SI 17.
http://dx.doi.org/10.1177/096032719401300201.
AT SDR (Agency for Toxic Substances and Disease Registry). (1997a). Toxicol ogical profile for
tetrachloroethylene. Atlanta, GA: U.S. Department of Health and Humans Services.
http://www.atsdr.cdc. gov/toxprofiles/tp.asp?id=265&tid=48.
AT SDR (Agency for Toxic Substances and Disease Registry). (1997b). Toxicol ogical profile for
trichloroethylene (update). (PB98-101165). Atlanta, GA: U.S. Department of Health and
Human Services, Public Health Service, Agency for Toxic Substances and Disease
Registry, http://www.atsdr.cdc.gov/toxprofiles/tp 19-p.pdf
AT SDR (Agency for Toxic Substances and Disease Registry). (1998a). Biomarkers of kidney
function for environmental health field studies. (PB98-142128; ATSDRHS9896).
Atlanta, GA. http://www.ntis.gov/search/product.aspx?ABBR=PB98142128.
ATSDR (Agency for Toxic Substances and Disease Registry). (1998b). Volatile organic
compounds in drinking water and adverse pregnancy outcomes. United States Marine
Corps Base, Camp Lejeune, NC. (PB98-156540; ATSDRHS9897). Atlanta, GA.
http://www.atsdr.cdc.gov/hs/lejeune/.
ATSDR (Agency for Toxic Substances and Disease Registry). (2003). Survey of specific
childhood cancers and birth defects among children whose mothers were pregnant while
7-3
-------
living at U.S. Marine Corps Base Camp Lejeune, North Carolina, 1968-1985. Atlanta,
GA. http://www.atsdr.cdc.gov/sites/lejeune/survey_full.html.
AT SDR (Agency for Toxic Substances and Disease Registry). (2004). Interaction profile: 1,1,1-
trichloroethane, 1,1-dichloroethane, trichloroethylene, and tetrachloroethylene. Atlanta,
GA. http://www.atsdr.cdc.gov/interactionprofiles/ip02.html.
AT SDR (Agency for Toxic Substances and Disease Registry). (2006). Toxicological profile for
1,1,1-trichloroethane. Atlanta, GA: U.S. Department of Health and Human Services.
http://www.atsdr.cdc.gov/ToxProfiles/tp.asp?id=432&tid=76.
AT SDR (Agency for Toxic Substances and Disease Registry). (2008). Toxicological profile for
1,1,2,2 tetrachloroethane. Atlanta, GA: U.S. Department of Health and Humans Services.
http://www.atsdr.cdc.gov/ToxProfiles/tp.asp?id=801&tid=l 56.
Auperin, A; Benhamou, S; Ory-Paoletti, C: Flamant, R. (1994). Occupational risk factors for
renal cell carcinoma: A case-control study. Occup Environ Med 51: 426-428.
http://dx.doi.0rg/10.1136/oem.51.6.426.
Austin. H: Delzell E: Grufferman. S: Levine. R: Morrison. AS: Stolley. PD: Cole. P. (1987).
Case-control study of hepatocellular carcinoma, occupation, and chemical exposures. J
Occup Med 29: 665-669.
Bagnell, PC: Ellenberger, HA. (1977). Obstructive jaundice due to a chlorinated hydrocarbon in
breast milk. Can Med Assoc J 117: 1047-1048.
Bale. AS: Adams. TL: Bushnell. PJ: Shafer. TJ: Boves. WK. (2005). Role of NMD A, nicotinic,
and GABA receptors in the steady-state visual-evoked potential in rats. Pharmacol
Biochem Behav 82: 635-645. http://dx.doi.Org/10.1016/i.pbb.2005.ll.003.
Band. PR: Le. ND: Fang. R: Deschamps. M: Gallagher. RP: Yang. P. (2000). Identification of
occupational cancer risks in British Columbia: A population-based case-control study of
995 incident breast cancer cases by menopausal status, controlling for confounding
factors. J Occup Environ Med 42: 284-310.
Bartels, MJ. (1994). Quantitation of the tetrachloroethylene metabolite N-acetyl-S-
(trichlorovinyl)cysteine in rat urine via negative ion chemical ionization gas
chromatography/tandem mass spectrometry. Biological Mass Spectrometry 23: 689-694.
http://dx.doi.org/10.1002/bms.1200231107.
Bartsch, H; Malaveille, C: Barbin, A; Planche, G. (1979). Mutagenic and alkylating metabolites
of halo-ethylenes, chlorobutadienes and dichlorobutenes produced by rodent or human
liver tissues: Evidence for oxirane formation by P450-linked microsomal mono-
oxygenases. Arch Toxicol 41: 249-277. http://dx.doi.org/10.1007/BF00296896.
Bateson, TF; Schwartz, J. (2008). Children's response to air pollutants. J Toxicol Environ Health
A 71: 238-243. http://dx.doi.org/10.1080/15287390701598234.
Bavlin. SB: Herman. JG: Graff. JR: Vertino. PM: Issa. JP. (1998). Alterations in DNA
methylation: A fundamental aspect of neoplasia. Adv Cancer Res 72: 141-196.
http://dx.doi.org/10.1016/S0065-230X(08)60702-2.
Beaudoin, PL; Manookin, MB; Demb, JB. (2008). Distinct expressions of contrast gain control
in parallel synaptic pathways converging on a retinal ganglion cell. J Physiol 586: 5487-
5502. http://dx.doi.org/10.1113/iphvsiol.2008.156224.
Beaudreuil, S: Lasfargues, G: Laueriere, L; El Ghoul, Z; Fourquet F; Longuet C: Halimi, JM;
Nivet H: Buchler, M. (2005). Occupational exposure in ANCA-positive patients: A case-
control study. Kidney Int 67: 1961-1966. http://dx.doi.Org/10.l 11 l/j.1523-
1755.2005.00295.x.
7-4
-------
Beckstead. MJ: Weiner. JL: Eger. El: Gong. DH: Mihic. SJ. (2000). Glycine and [gamma]-
aminobutyric acid(A) receptor function is enhanced by inhaled drugs of abuse. Mol
Pharmacol57: 1199-1205.
Bel and, F. (1999). NTP technical report on the toxicity and metabolism studies of chloral hydrate
(CAS No. 302-17-0). Administered by gavage to F344/N rats and B6C3F1 mice. U.S.
Dept. of Health and Human Services, National Institutes of Health: Rockville, MD.
Beliles, RP: Brusick, DJ: Mecler, FJ. (1980). Teratogenic-mutagenic risk of workplace
contaminants: trichloroethylene, perchloroethylene, and carbon disulfide. (210-77-0047).
Cincinnati, OH: National Institute for Occupation Safety and Health.
Beliles, RP. (2002). Concordance across species in the reproductive and developmental toxicity
of tetrachloroethylene. Toxicol Ind Health 18: 91-106.
http://dx.doi.org/10.1191/0748233702thl37oa.
Bellinger, D; Hu, H; Titlebaum, L; Needleman, HL. (1994). Attentional correlates of dentin and
bone lead levels in adolescents. Arch Environ Health 49: 98-105.
http://dx.doi.org/10.1080/00039896.1994.9937461.
Benane, S: Blackman, C: House, D. (1996). Effect of perchloroethylene and its metabolites on
intercellular communication in clone 9 rat liver cells. J Toxicol Environ Health 48: 427-
437.
Benedetti, M; Martuccio, G: Fattorini, D: Canapa, A; Barucca, M; Nigro, M; Regoli, F. (2007).
Oxidative and modulatory effects of trace metals on metabolism of poly cyclic aromatic
hydrocarbons in the Antarctic fish Trematomus bernacchii. Aquat Toxicol 85: 167-175.
http://dx.doi.0rg/10.1016/i.aquatox.2007.08.009.
Bergamaschi, E; Mutti, A; Bocchi, MC: Alinovi, R; Olivetti, G: Ghiggeri, GM; Franchini, I.
(1992). Rat model of perchloroethylene-induced renal dysfunctions. Environ Res 59:
427-439. http://dx.doi.org/10.1016/50013-9351(05)80046-5.
Berger, T; Horner, CM. (2003). In vivo exposure of female rats to toxicants may affect oocyte
quality. Reprod Toxicol 17: 273-281. http://dx.doi.org/10.1016/50890-6238(03)00009-1.
Berman, E; Schlicht, M: Moser, VC: MacPhail, RC. (1995). A multidisciplinary approach to
toxicological screening: I. Systemic toxicity. J Toxicol Environ Health 45: 127-143.
http://dx.doi.org/10.1080/15287399509531986.
Bernard, A; Lauwerys, R. (1995). Low-molecular-weight proteins as markers of organ toxicity
with special reference to Clara cell protein. Toxicol Lett 77: 145-151.
http://dx.doi.org/10.1016/0378-4274(95)03284-3.
Besson, H; Brennan, P; Becker, N; Meters, A; De Sanjose, S: Font, R: Maynadie, M: Foretova,
L: Cocco, PL; Staines, A; Vornanen, M: Boffetta, P. (2006). Tobacco smoking, alcohol
drinking and non-Hodgkin's lymphoma: A European multicenter case-control study
(Epilymph). Int J Cancer 119: 901-908. http://dx.doi.org/10.1002/ijc.21913.
Bhattacharya, RK; Schultze, MO. (1971). Properties of DNA isolated from tissues of calves
treated with S-(l,2-dichlorovinyl)-L-cysteine: II. Primer-template activity for bacterial
DNA polymerases. Arch Biochem Biophys 145: 575-582.
http://dx.doi.org/10.1016/S0003-9861(71)80017-6.
Bhattacharya. RK: Schultze. MO. (1972). Properties of DNA treated with S-(l,2-dichlorovinyl)-
L-cysteine and a lyase. Arch Biochem Biophys 153: 105-115.
http://dx.doi.org/10.1016/0003-9861(72)90426-2.
Bhunya, SP: Behera, BC. (1987). Relative genotoxicity of trichloroacetic acid (TCA) as revealed
by different cytogenetic assays: Bone marrow chromosome aberration, micronucleus and
7-5
-------
sperm-head abnormality in the mouse. Mutat Res Genet Toxicol Environ Mutagen 188:
215-221. http://dx.doi.org/10.1016/0165-1218(87)90092-9.
Bhunya, SP: Jena, GB. (1996). The evaluation of clastogenic potential of trichloroacetic acid
(TCA) in chick in vivo test system. Mutat Res Genet Toxicol Environ Mutagen 367: 253-
259. http://dx.doi.org/10.1016/S0165-1218(96)90085-3.
Bignami, M; Conti, G: Conti, L; Crebelli, R: Misuraca, F; Puglia, AM; Randazzo, R;
Sciandrello, G: Carere, A. (1980). Mutagenicity of halogenated aliphatic hydrocarbons in
Salmonella typhimurium, Streptomyces coelicolor and Aspergillus nidulans. Chem Biol
Interact 30: 9-23. http://dx.doi.org/10.1016/0009-2797(80)90110-6.
BIP Study Group (Bezafibrate Infarction Prevention Study Group). (2000). Secondary
prevention by raising HDL cholesterol and reducing triglycerides in patients with
coronary artery disease: The Bezafibrate Infarction Prevention (BIP) study. Circulation
102: 21-27.
Birner, G: Richling, C: Henschler, D: Anders. MW: Dekant W. (1994). Metabolism of
tetrachloroethene in rats: Identification of N[epsilon]-(dichloroacetyl)-L-lysine and
N[epsilon]-(trichloroacetyl)-L-lysine as protein adducts. Chem Res Toxicol 7: 724-732.
http://dx.doi.org/10.1021/tx00042a003.
Birner, G: Rutkowska, A; Dekant, W. (1996). N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and
2,2,2-trichloroethanol: Two novel metabolites of tetrachloroethene in humans after
occupational exposure. Drug Metab Dispos 24: 41-48.
Blair, A; Decoufle, P; Grauman, D. (1979). Causes of death among laundry and dry cleaning
workers. Am J Public Health 69: 508-511.
Blair. A: Stewart. PA: Tolbert PE: Grauman. D: Moran, FX: Vaught J: Ravner, J. (1990).
Cancer and other causes of death among a cohort of dry cleaners. Br J Ind Med 47: 162-
168. http://dx.doi.0rg/10.1136/oem.47.3.162.
Blair, A; Linos, A; Stewart, PA; Burmeister, LF; Gibson, R: Everett G: Schuman, L; Cantor,
KP. (1993). Evaluation of risks for non-Hodgkin's lymphoma by occupation and industry
exposures from a case-control study. Am J Ind Med 23: 301-312.
http://dx.doi.org/10.1002/aiim.4700230207.
Blair, A; Hartge, P: Stewart, PA; McAdams, M; Lubin, J. (1998). Mortality and cancer incidence
of aircraft maintenance workers exposed to trichloroethylene and other organic solvents
and chemicals: Extended follow-up. Occup Environ Med 55: 161-171.
http://dx.doi.0rg/10.1136/oem.55.3.161.
Blair. A: Zheng. T: Linos. A: Stewart. PA: Zhang. YW: Cantor. KP. (2001). Occupation and
leukemia: A population-based case-control study in Iowa and Minnesota. Am J Ind Med
40: 3-14. http://dx.doi.org/10.1002/aiim.1066.
Blair, A; Petralia, SA; Stewart, PA. (2003). Extended mortality follow-up of a cohort of dry
cleaners. AnnEpidemiol 13: 50-56. http://dx.doi.org/10.1016/S1047-2797(02)00250-8.
Blot WJ: McLaughlin, JK. (1999). The changing epidemiology of esophageal cancer. Semin
Oncol 26: 2-8.
Bliimcke, S: Schwartzkopff, W: Lobeck, H: Edmondson, NA: Prentice. DE: Blane, GF. (1983).
Influence of fenofibrate on cellular and subcellular liver structure in hyperlipidemic
patients. Atherosclerosis 46: 105-116. http://dx.doi.org/10.1016/0021-9150(83)90169-7.
Board, PG: Baker, RT; Chelvanayagam, G: Jermiin, LS. (1997). Zeta, a novel class of
glutathione transferases in a range of species from plants to humans. Biochem J 328: 929-
935.
7-6
-------
Boettner, EA; Muranko, HJ. (1969). Animal breath data for estimating the exposure of humans
to chlorinated hydrocarbons. Am Ind Hyg Assoc J 30: 437-442.
Bogen, KT; Colston, BW, Jr: Machicao, LK. (1992). Dermal absorption of dilute aqueous
chloroform, trichloroethylene, and tetrachloroethylene in hairless guinea pigs. Fundam
Appl Toxicol 18: 30-39. http://dx.doi.org/10.1016/0272-0590(92)90192-K.
Boice, JD, Jr; Marano, D; Fryzek, J: Sadler, C; McLaughlin, JK. (1999). Mortality among
aircraft manufacturing workers. Occup Environ Med 56: 581-597.
http://dx.doi.0rg/10.1136/oem.56.9.581.
Bois, FY; Zeise, L; Tozer, TN. (1990). Precision and sensitivity of pharmacokinetic models for
cancer risk assessment: Tetrachloroethylene in mice, rats, and humans. Toxicol Appl
Pharmacol 102: 300-315. http://dx.doi.org/10.1016/0041-008X(90)90029-T.
Bois, FY; Gelman, A; Jiang, J: Maszle, PR: Zeise, L; Alexeef, G. (1996). Population
toxicokinetics of tetrachloroethylene. Arch Toxicol 70: 347-355.
http://dx.doi.org/10.1007/s002040050284.
Bolanowska, W; Golacka, J. (1972). [Inhalation and excretion of tetrachloroethylene in men in
experimental conditions]. Med Pr 23: 109-119.
Bolt HM; Laib, RJ: Filser, JG. (1982). Reactive metabolites and carcinogenicity of halogenated
ethylenes. Biochem Pharmacol 31: 1-4. http://dx.doi.org/10.1016/0006-2952(82)90227-1.
Bonatti, S: Cavalieri, Z; Viaggi, S: Abbondandolo, A. (1992). The analysis of 10 potential
spindle poisons for their ability to induce CREST-positive micronuclei in human diploid
fibroblasts. Mutagenesis 7: 111-114. http://dx.doi.org/10.1093/mutage/7.2. 111.
Bond, G: McLaren, E; Sabel, F; Bodner, K; Lipps, T; Cook, R. (1990). Liver and biliary tract
cancer among chemical workers. Am J Ind Med 18: 19-24.
Bonse, G: Urban, T; Reichert, D: Henschler, D. (1975). Chemical reactivity, metabolic oxirane
formation and biological reactivity of chlorinated ethylenes in the isolated perfused rat
liver preparation. Biochem Pharmacol 24: 1829-1834. http://dx.doi.org/10.1016/0006-
2952(75)90468-2.
Bonse, G: Henschler, D. (1976). Chemical reactivity, biotransformation, and toxicity of
polychlorinated aliphatic compounds. CritRev Toxicol 4: 395-409.
Bosco, MG: Figa-Talamanca, I; Salerno, S. (1987). Health and reproductive status of female
workers in dry cleaning shops. Int Arch Occup Environ Health 59: 295-301.
http://dx.doi.org/10.1007/BF00377741.
Bottger, A; Elstermeier, F. (1989). Belastung der bevolkerung durch organische losungsmittel.
Dessau-RoBlau, Germany: Umweltbundesamt.
Bouchardy, C: Schiiler, G: Minder, C: Hotz, P; Bousquet A; Levi, F: Fisch, T: Torhorst J:
Raymond, L. (2002). Cancer risk by occupation and socioeconomic group among men—A
study by the Association of Swiss Cancer Registries. Scand J Work Environ Health 28: 1-
88.
Boulet LP. (1988). Increases in airway responsiveness following acute exposure to respiratory
irritants: Reactive airway dysfunction syndrome or occupational asthma? Chest 94: 476-
481. http://dx.doi.0rg/10.1378/chest.94.3.476.
Bove, F: Shim, Y; Zeitz, P. (2002). Drinking water contaminants and adverse pregnancy
outcomes: A review. Environ Health Perspect 110: 61-74.
Bove. FJ: Fulcomer. MC: Klotz. JB: Esmart J: Dufficv. EM: Savrin. JE. (1995). Public drinking
water contamination and birth outcomes. Am J Epidemiol 141: 850-862.
7-7
-------
Bove, FJ. (1996). Public drinking water contamination and birthweight, prematurity, fetal deaths,
and birth defects. Toxicol Ind Health 12: 255-266.
Bowen, SE; Batis, JC: Paez-Martinez, N; Cruz, SL. (2006). The last decade of solvent research
in animal models of abuse: Mechanistic and behavioral studies. Neurotoxicol Teratol 28:
636-647. http://dx.doi.0rg/10.1016/i.ntt.2006.09.005.
Bowler, RM; Mergler, D; Huel, G: Harrison, R; Cone, J. (1991). Neuropsychological impairment
among former microelectronics workers. Neurotoxicology 12: 87-103.
Bowman, KJ. (1982). A method for quantitative scoring of the Farnsworth Panel D-15. Acta
Ophthalmol 60: 907-916. http://dx.doi.Org/10.llll/i.1755-3768.1982.tb00621.x.
Boyes, WK: Bercegeay, M; Oshiro, WM; Krantz, OT; Kenyon, EM; Bushnell, PJ: Benignus,
VA. (2009). Acute perchloroethylene exposure alters rat visual-evoked potentials in
relation to brain concentrations. Toxicol Sci 108: 159-172.
http://dx.doi.org/10.1093/toxsci/kfn265.
Boyle, CA; Brann, EA. (1992). Proxy respondents and the validity of occupational and other
exposure data. The Selected Cancers Cooperative Study Group. Am J Epidemiol 136:
712-721.
Briving, C; Jacobson, I; Hamberger, A; Kjellstrand, P; Haglid, KG: Rosengren, LE. (1986).
Chronic effects of perchloroethylene and trichloroethylene on the gerbil brain amino
acids and glutathione. Neurotoxicology 7: 101-108.
Broadwell. DK: Darcev. DJ: Hudnell. HK: Otto. DA: Boves. WK. (1995). Work-site clinical and
neurobehavioral assessment of solvent-exposed microelectronics workers. Am J Ind Med
27: 677-698. http://dx.doi.org/10.1002/aiim.4700270505.
Brodkin, CA: Daniell, W; Checkoway, H; Echeverria, D: Johnson, J: Wang, K; Sohaey, R:
Green, D; Redlich, C: Gretch, D. (1995). Hepatic ultrasonic changes in workers exposed
to perchloroethylene. Occup Environ Med 52: 679-685.
Bronzettl G: Bauer. C: Corsl C: Del Carratore. R: Galli. A: Nierl R: Paolinl M. (1983).
Genetic and biochemical studies on perchloroethylene 'in vitro' and 'in vivo'. Mutat Res
Genet Toxicol Environ Mutagen 116: 323-331. http://dx.doi.org/10.1016/0165-
1218(83)90070-8.
Bronzettl G: Galli. A: Corsl C: Cundarl E: Del Carratore. R: Nieri. R: Paolinl M. (1984).
Genetic and biochemical investigation on chloral hydrate in vitro and in vivo. Mutat Res
141: 19-22. http://dx.doi.org/10.1016/0165-7992(84)90031-9.
Brown, LM; Hoover, R: Silverman, D: Baris, D: Hayes, R: Swanson, GM; Schoenberg, J:
Greenberg, R: Liff J: Schwartz, A; Dosemeci, M: Pottern, L; Fraumeni JF, J, r. (2001).
Excess incidence of squamous cell esophageal cancer among US Black men: Role of
social class and other risk factors. Am J Epidemiol 153: 114-122.
http://dx.doi.0rg/10.1093/aie/153.2.114.
Brown Dzubow, R: Makris, S; Siegel Scott C: Barone, S, Jr. (2010). Early lifestage exposure
and potential developmental susceptibility to tetrachloroethylene. Birth Defects Res B
Dev Reprod Toxicol 89: 50-65. http://dx.doi.org/10.1002/bdrb.20222.
Brownson, RC: Alavanja, MC: Chang, JC. (1993). Occupational risk factors for lung cancer
among nonsmoking women: A case-control study in Missouri (United States). Cancer
Causes Control 4: 449-454.
Buben, JA; O'Flaherty, EJ. (1985). Delineation of the role of metabolism in the hepatotoxicity of
trichloroethylene and perchloroethylene: A dose-effect study. Toxicol Appl Pharmacol
78: 105-122.
-------
Buelke-Sam. J: Kimtnel CA: Adams. J: Nelson. CJ: Vorhees. CV: Wright DC: St Omen V:
Korol, BA; Butcher, RE; Geyer, MA. (1985). Collaborative behavioral teratology study:
Results. Neurobehav Toxicol Teratol 7: 591-624.
Bull, RJ: Sanchez, IM; Nelson, MA; Larson, JL; Lansing, AJ. (1990). Liver tumor induction in
B6C3F1 mice by dichloroacetate and trichloroacetate. Toxicology 63: 341-359.
http://dx.doi.org/10.1016/0300-483X(90)90195-M.
Bull RJ: Orner. GA: Cheng. RS: Stillwell L: Stauber. AJ: Sasser. LB: Lingohr. MK: Thrall BD.
(2002). Contribution of dichloroacetate and trichloroacetate to liver tumor induction in
mice by trichloroethylene. Toxicol Appl Pharmacol 182: 55-65.
http://dx.doi.org/10.1006/taap.2002.9427.
Bull, RJ: Sasser, LB: Lei, XC. (2004). Interactions in the tumor-promoting activity of carbon
tetrachloride, trichloroacetate, and dichloroacetate in the liver of male B6C3F1 mice.
Toxicology 199: 169-183. http://dx.doi.Org/10.1016/j.tox.2004.02.018.
Bull, RJ. (2004). Trichloroethylene and liver tumors in mice. In Proceedings of New Scientific
Research Related to the Health Effects of Trichlorothylene: U.S. Environmental
Protection Agency.
http://oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=432706.
Burns, PB: Swanson, GM. (1991). Risk of urinary bladder cancer among blacks and whites: The
role of cigarette use and occupation. Cancer Causes Control 2: 371-379.
http://dx.doi.org/10.1007/BF00054297.
Bushnell PJ: Shafer. TJ: Bale. AS: Boves. WK: Simmons. JE: Eklund. C: Jackson. TL. (2005).
Developing an exposure-dose-response model for the acute neurotoxicity of organic
solvents: overview and progress on in vitro models and dosimetry. Environ Toxicol
Pharmacol 19: 607-614. http://dx.doi.Org/10.1016/i.etap.2004.12.026.
Bushnell PJ: Oshiro. WM: Samsam. TE: Benignus. VA: Krantz. QT: Kenyon. EM. (2007). A
Dosimetric Analysis of the Acute Behavioral Effects of Inhaled Toluene in Rats. Toxicol
Sci 99: 181-189. http://dx.doi.org/10.1093/toxsci/kfml46.
Bvczkowskl JZ: Kinkead. ER: Leahy. HF: Randall GM: Fisher. JW. (1994). Computer
simulation of the lactational transfer of tetrachloroethylene in rats using a physiologically
based model. Toxicol Appl Pharmacol 125: 228-236.
http://dx.doi.org/10.1006/taap.1994.1068.
Byczkowski, JZ: Fisher, JW. (1994). Lactational transfer of tetrachloroethylene in rats. Risk
Anal 14: 339-349. http://dx.doi.Org/10.llll/i.1539-6924.1994.tb00250.x.
Byczkowski, JZ: Fisher, JW. (1995). A computer program linking physiologically based
pharmacokinetic model with cancer risk assessment for breast-fed infants. Comput
Methods Programs Biomed 46: 155-163. http://dx.doi.org/10.1016/0169-2607(94)01616-
N.
Byers, VS: Levin, AS: Ozonoff DM; Baldwin, RW. (1988). Association between clinical
symptoms and lymphocyte abnormalities in a population with chronic domestic exposure
to industrial solvent-contaminated domestic water supply and a high incidence of
leukaemia. Cancer Immunol Immunother 27: 77-81.
Cai. SX: Huang. MY: Chen. Z: Liu. YT: Jin. C: Watanabe. T: Nakatsuka. H: Seiii. K: Inoue. O:
Ikeda, M. (1991). Subjective symptom increase among dry-cleaning workers exposed to
tetrachloroethylene vapor. Ind Health 29: 111-121.
http://dx.doi.0rg/10.2486/indhealth.29.l 11.
7-9
-------
Cal EPA (California Environmental Protection Agency). (2001). Public health goal for
tetrachloroethylene in drinking water. Sacramento, CA: Office of Environmental Health
Hazard Assessment. http://www.oehha.org/water/phg/pdf/PCEAug2001.pdf.
Caldwell, DJ. (1999). Review of mononuclear cell leukemia in F-344 rat bioassays and its
significance to human cancer risk: A case study using alkyl phthalates. Regul Toxicol
Pharmacol 30: 45-53. http://dx.doi.org/10.1006/rtph.1999.1305.
Callen. DF: Wolf CR: Philpot RM. (1980). Cytochrome P-450 mediated genetic activity and
cytotoxicity of seven halogenated aliphatic hydrocarbons in Saccharomyces cerevisiae.
MutatRes 77: 55-63. http://dx.doi.org/10.1016/0165-1218(80)90120-2.
Calvert GM; Ruder, AM; Petersen, MR. (2011). Mortality and end-stage renal disease incidence
among dry cleaning workers. Occup Environ Med 68: 709-716.
http://dx.doi.org/10.1136/oem.2010.060665.
Calvert, JG. (1976). Hydrocarbon involvement in photochemical smog formation in Los Angeles
atmosphere. Environ Sci Technol 10: 256-262. http://dx.doi.org/10.1021/es60114a003.
Campagna. D: Mergler. D: Huel G: Belanger. S: Truchon. G: Ostiguv. C: Drolet D. (1995).
Visual dysfunction among styrene-exposed workers. Scand J Work Environ Health 21:
382-390.
Campagna, D: Gobba, F; Mergler, D: Moreau, T; Galassi, C: Cavalleri, A; Huel, G. (1996).
Color vision loss among styrene-exposed workers neurotoxicological threshold
assessment. Neurotoxicology 17: 367-374.
Campbell, JA; Corrigall, AV: Guy, A; Kirsch, RE. (1991). Immunohistologic localization of
alpha, mu, and pi class glutathione S-transferases in human tissues. Cancer 67: 1608-
1613. http://dx.doi.org/10.1002/1097-0142(19910315)67:6<1608::AID-
CNCR2820670623>3.0.CO:2-S.
Canner, PL; Berge, KG: Wenger, NK; Stamler, J: Friedman, L; Prineas, RJ: Friedewald, W.
(1986). Fifteen year mortality in Coronary Drug Project patients: long-term benefit with
niacin. J Am Coll Cardiol 8: 1245-1255.
Cano, MI; Pollan, M. (2001). Non-Hodgkin's lymphomas and occupation in Sweden. Int Arch
Occup Environ Health 74: 443-449. http://dx.doi.org/10.1007/s004200100248.
CARS (California Air Resources Board). (1998). 1990-1996 statewide perchloroethylene
summary, ppb [Database]. Sacramento, CA: California Environmental Protection
Agency. Retrieved from http://www.arb.ca.gov/aqd/perc/pcstate.htm
Carney, EW: Thorsrud, BA; Dugard, PH; Zablotny, CL. (2006). Developmental toxicity studies
in Crl:CD (SD) rats following inhalation exposure to trichloroethylene and
perchloroethylene. Birth Defects Res B Dev Reprod Toxicol 77: 405-412.
http://dx.doi.org/10.1002/bdrb.20091.
Carpenter, CP. (1937). The chronic toxicity of tetrachlorethylene. J Ind Hyg Toxicol 19: 323-
336.
Cavalleri. A: Gobba. F: Paltrinieri. M: Fantuzzi. G: Righi. E: Aggazzottl G. (1994).
Perchloroethylene exposure can induce colour vision loss. Neurosci Lett 179: 162-166.
http://dx.doi.org/10.1016/0304-3940(94)90959-8.
Cederberg, H; Henriksson, J: Binderup, ML. (2010a). DNA damage detected by the alkaline
comet assay in the liver of mice after oral administration of tetrachloroethylene.
Mutagenesis25: 133-138. http://dx.doi.org/10.1093/mutage/gep051.
Cederberg. H: Henriksson. J: Binderup. M. -L. (201 Ob). RE: DNA damage detected by the
alkaline comet assay in the liver of mice after oral administration of tetrachloroethylene.
7-10
-------
(Mutagenesis, 25, 133-138, 2010). Mutagenesis 25: 429-430.
http://dx.doi.org/10.1093/tnutage/geq021.
Chang, F. (2005). Strategies for benzodiazepine withdrawal in seniors. CPJ 138: 38-40.
Chang. LW: DeAngelo. AB: Daniel FB. (1989). Evaluation of DNA strand breaks (SB) by
chloroacetic acids (CAs) and chloroacetaldehydes (CADs) in vivo and in vitro. Proc Ann
Mtg Am Assoc Cane Res 30: 146.
Chang, LW: Daniel FB: DeAngelo, AB. (1992). Analysis of DNA strand breaks induced in
rodent liver in vivo, hepatocytes in primary culture, and a human cell line by chlorinated
acetic acids and chlorinated acetaldehydes. Environ Mol Mutagen 20: 277-288.
http://dx.doi.org/10.1002/em.2850200406.
Chang. YM: Tai. CF: Yang. SC: Chen. CJ: Shih. TS: Lin. R: Liou. SH. (2003). A cohort
mortality study of workers exposed to chlorinated organic solvents in Taiwan. Ann
Epidemiol 13: 652-660. http://dx.doi.org/10.1016/S1047-2797(03)00038-3.
Chang. YM: Tai. CF: Yang. SC: Lin. RS: Sung. FC: Shih. TS: Liou. SH. (2005). Cancer
incidence among workers potentially exposed to chlorinated solvents in an electronics
factory. J Occup Health 47: 171-180. http://dx.doi.org/10.1539/ioh.47.171.
Charbonneau, M; Short, BG: Lock, EA; Swenberg, JA. (1987). Mechanism of petroleum induced
sex-specific protein droplet nephropathy and renal cell proliferation in Fischer-344 rats:
Relevance to humans. In Trace substances in environmental health, XXI: Proceedings of
the University of Missouri's 21st Annual Conference on Trace Substances in
Environmental Health. Columbia, MO: University of Missouri.
Chen, C: Blancato, J. (1987). Role of pharmacokinetic modeling in risk assessment:
Perchloroethylene as an example. In Pharmacokinetics in Risk Assessment: Drinking
Water and Health. Washington, DC: National Academy Press.
Chen, CW: Blancato, JN. (1989). Incorporation of biological information in cancer risk
assessment: Example—vinyl chloride. Cell Biol Toxicol 5: 417-444.
Chen, HH; Chan, MH; Fu, SH. (2002a). Behavioural effects of tetrachloroethylene exposure in
rats: Acute and subchronic studies. Toxicology 170: 201-209.
http://dx.doi.org/10.1016/S0300-483X(01)00544-3.
Chen. SJ: Wang. JL: Chen. JH: Huang. RN. (2002b). Possible involvement of glutathione and
p53 in trichloroethylene- and perchloroethylene-induced lipid peroxidation and apoptosis
in human lung cancer cells. Free Radic Biol Med 33: 464-472.
http://dx.doi.org/10.1016/S0891-5849(02)00817-l.
Cheung, C: Akiyama, TE; Ward, JM; Nicol, CJ: Feigenbaum, L; Vinson, C: Gonzalez, FJ.
(2004). Diminished hepatocellular proliferation in mice humanized for the nuclear
receptor peroxisome proliferator-activated receptor alpha. Cancer Res 64: 3849-3854.
http://dx.doi.org/10.1158/0008-5472.CAN-04-0322.
Chien, YC. (1997) The influences of exposure pattern and duration on elimination kinetics and
exposure assessment of tetrachloroethylene in humans [PhD]. (Doctoral Dissertation).
New Brunswick, NJ, Rutgers University.
Chiu, WA; Bois, FY. (2006). Revisiting the population toxicokinetics of tetrachloroethylene.
Arch Toxicol 80: 382-385. http://dx.doi.org/10.1007/s00204-006-0061-9.
Chiu. WA: Micallef. S: Monster. AC: Bois. FY. (2007). Toxicokinetics of inhaled
trichloroethylene and tetrachloroethylene in humans at 1 ppm: Empirical results and
comparisons with previous studies. Toxicol Sci 95: 23-36.
http://dx.doi.org/10.1093/toxsci/kfll29.
7-11
-------
Chiu, WA; Okino, MS: Evans, MV. (2009). Characterizing uncertainty and population
variability in the toxicokinetics of trichloroethylene and metabolites in mice, rats, and
humans using an updated database, physiologically based pharmacokinetic (PBPK)
model, and Bayesian approach. Toxicol Appl Pharmacol 241: 36-60.
http://dx.doi.0rg/10.1016/j.taap.2009.07.032.
Chiu, WA; Ginsberg, GL. (2011). Development and evaluation of a harmonized physiologically
based pharmacokinetic (PBPK) model for perchloroethylene toxicokinetics in mice, rats,
and humans. Toxicol Appl Pharmacol 253: 203-234.
http://dx.doi.0rg/10.1016/i.taap.2011.03.020.
Chiu, WA. (2011). Trichloroacetic acid: Updated estimates of its bioavailability and its
contribution to trichloroethylene-induced mouse hepatomegaly. Toxicology 285: 114-
125. http://dx.doi.0rg/10.1016/i.tox.2011.04.009.
Claude, JC: Frentzel-Beyme, RR: Kunze, E. (1988). Occupation and risk of cancer of the lower
urinary tract among men: A case-control study. Int J Cancer 41: 371-379.
http://dx.doi.org/10.1002/iic.2910410309.
Clavel, J: Mandereau, L; Conso, F; Limasset JC: Pourmir, I; Flandrin, G: Hemon, D. (1998).
Occupational exposure to solvents and hairy cell leukaemia. Occup Environ Med 55: 59-
64. http://dx.doi.0rg/10.1136/oem.55.l.59.
Clewell, HJ: Gentry, PR: Covington, TR: Sarangapani, R; Teeguarden, JG. (2004). Evaluation of
the potential impact of age- and gender-specific pharmacokinetic differences on tissue
dosimetry. Toxicol Sci 79: 381-393. http://dx.doi.org/10.1093/toxsci/kfhl09.
Clewell, HJ: Gentry, PR: Kester, JE; Andersen, ME. (2005). Evaluation of physiologically based
pharmacokinetic models in risk assessment: An example with perchloroethylene. Crit
Rev Toxicol 35: 413-433. http://dx.doi.org/10.1080/10408440590931994.
Clewell HJ. Ill: Gentry. PR: Covington. TR: Gearhart JM. (2000). Development of a
physiologically based pharmacokinetic model of trichloroethylene and its metabolites for
use in risk assessment. Environ Health Perspect 108: 283-305.
http://dx.doi.org/10.1289/ehp.00108s2283.
Cohn, P; Klotz, J: Bove, F; Berkowitz, M; Fagliano, J. (1994). Drinking water contamination and
the incidence of leukemia and non-Hodgkin's lymphoma. Environ Health Perspect 102:
556-561.
Cojocel, C; Beuter, W; Miiller, W; Mayer, D. (1989). Lipid peroxidation: A possible mechanism
of trichloroethylene-induced nephrotoxi city. Toxicology 55: 131-141.
http://dx.doi.org/10.1016/0300-483X(89)90180-7.
Coler, HR; Rossmiller, HR. (1953). Tetrachlorethylene exposure in a small industry. AMA Arch
Ind Hyg Occup Med 8: 227-238.
Colt J: Baris, D: Stewart. P: Schned, A: Heanev, J: Mott L: Silverman, D: Karagas, M. (2004).
Occupation and bladder cancer risk in a population-based case-control study in New
Hampshire. Cancer Causes Control 15: 759-769.
http://dx.doi.Org/10.1023/B:CACO.0000043426.28741.a2.
Colt JS: Karagas, MR; Schwenn, M: Baris, D: Johnson, A; Stewart, P: Verrill, C: Moore, LE;
Lubin, J: Ward, MH; Samanic, C: Rothman, N; Cantor, KP; Beane Freeman, LE; Schned,
A; Cherala, S; Silverman, DT. (2011). Occupation and bladder cancer in a population-
based case-control study in Northern New England. Occup Environ Med 68: 239-249.
http://dx.doi.org/10.1136/oem.2009.052571.
7-12
-------
Connor. TH: Theiss. JC: Hanna. HA: Monteith. DK: Matnev. TS. (1985). Genotoxicity of
organic chemicals frequently found in the air of mobile homes. Toxicol Lett 25: 33-40.
http://dx.doi.org/10.1016/0378-4274(85)90097-9.
Consonni, D; De Matteis, S; Lubin, JH; Wacholder, S; Tucker, M; Pesatori, AC: Caporaso, NE;
Bertazzi, PA; Landi, MT. (2010). Lung cancer and occupation in a population-based
case-control study. Am J Epidemiol 171: 323-333. http://dx.doi.org/10.1093/aje/kwp391.
Cooper, AJ. (1994). Enzymology of cysteine S-conjugate beta-lyases. In MW Anders; W Dekant
(Eds.), Conjugation-dependent carcinogenicity and toxicity of foreign compounds (pp.
71-113). San Diego, CA: Academic Press.
Cooper. GS: Parks. CG: Treadwell. EL: St Clair. EW: Gilkeson. GS: Doolev. MA. (2004).
Occupational risk factors for the development of systemic lupus erythematosus. J
RheumatolSl: 1928-1933.
Cooper, GS: Makris, SL; Nietert, PJ: Jinot J. (2009). Evidence of autoimmune-related effects of
trichloroethylene exposure from studies in mice and humans. Environ Health Perspect
117: 696-702. http://dx.doi.org/10.1289/ehp.11782.
Cordier, S: Clavel J: Limasset JC: Boccon-Gibod, L; Le Moual N; Mandereau, L; Hemon, D.
(1993). Occupational risks of bladder cancer in France: A multicentre case-control study.
Int J Epidemiol 22: 403-411. http://dx.doi.Org/10.1093/iie/22.3.403.
Coronary Drug Project Research Group. (1975). Clofibrate and niacin in coronary heart disease.
JAMA 231: 360-381.
Coronary Drug Project Research Group. (1977). Gallbladder disease as a side effect of drugs
influencing lipid metabolism: Experience in the Coronary Drug Project. N Engl J Med
296: 1185-1190. http://dx.doi.org/10.1056/NEJM197705262962101.
Costa, AK; Ivanetich, KM. (1980). Tetrachloroethylene metabolism by the hepatic microsomal
cytochrome P-450 system. Biochem Pharmacol 29: 2863-2869.
http://dx.doi.org/10.1016/0006-2952(80)90023-4.
Costa, AK; Ivanetich, KM. (1984). Chlorinated ethylenes: their metabolism and effect on DNA
repair in rat hepatocytes. Carcinogenesis 5: 1629-1636.
http://dx.doi.0rg/10.1093/carcin/5.12.1629.
Costa, C; Barbaro, M; Catania, S: Silvari, V: Germano, MP. (2004). Cytotoxicity evaluation after
coexposure to perchloroethylene and selected peroxidant drugs in rat hepatocytes.
Toxicol In Vitro 18: 37-44. http://dx.doi.org/10.1016/S0887-2333(03)00133-4.
Costantini, AS: Miligi, L: Kriebel, D: Ramazzotti, V: Rodella, S: Scarpi, E; Stagnaro, E;
Tumino, R; Fontana, A; Masala, G: Vigano, C: Vindigni, C: Crosignani, P; Benvenuti, A;
Vineis, P. (2001). A multicenter case-control study in Italy on hematolymphopoietic
neoplasms and occupation. Epidemiology 12: 78-87.
Costantini, AS: Benvenuti, A; Vineis, P: Kriebel, D: Tumino, R; Ramazzotti, V: Rodella, S:
Stagnaro, E: Crosignani, P; Amadori, D: Mirabelli, D: Sommani, L: Belletti, I; Troschel,
L: Romeo. L: Micell G: Tozzl G: Mendico. I: Maltoni. S: Miligi. L. (2008). Risk of
leukemia and multiple myeloma associated with exposure to benzene and other organic
solvents: Evidence from the Italian Multicenter Case-control study. Am J Ind Med 51:
803-811. http://dx.doi.org/10.1002/aiim.20592.
Costas, K; Knorr, RS: Condon, SK. (2002). A case-control study of childhood leukemia in
Woburn, Massachusetts: The relationship between leukemia incidence and exposure to
public drinking water. Sci Total Environ 300: 23-35. http://dx.doi.org/10.1016/S0048-
9697(02)00169-9.
7-13
-------
Covington, TR; Gentry, PR: Van Landingham, CB; Andersen, ME; Kester, JE; Clewell, HJ.
(2007). The use of Markov chain Monte Carlo uncertainty analysis to support a public
health goal for perchloroethylene. Regul Toxicol Pharmacol 47: 1-18.
http://dx.doi.0rg/10.1016/i.vrtph.2006.06.008.
Crebelli, R: Conti, G: Conti, L; Carere, A. (1985). Mutagenicity of trichloroethylene,
trichloroethanol and chloral hydrate in Aspergillus nidulans. Mutat Res Genet Toxicol
Environ Mutagen 155: 105-111. http://dx.doi.org/10.1016/0165-1218(85)90126-0.
Crebelli, R: Conti, G: Conti, L; Carere, A. (1991). In vitro studies with nine known or suspected
spindle poisons: results in tests for chromosome malsegregation in Aspergillus nidulans.
Mutagenesis 6: 131-136. http://dx.doi.Org/10.1093/mutage/6.2.131.
Crump, K; Allen, B; Shipp, A. (1989). Choice of dose measure for extrapolating carcinogenic
risk from animals to humans: An empirical investigation of 23 chemicals. Health Phys
57: 387-393.
Crump, KS. (1998). On summarizing group exposures in risk assessment: Is an arithmetic mean
or a geometric mean more appropriate? Risk Anal 18: 293-297.
Cruz, SL; Balster, RL: Woodward, JJ. (2000). Effects of volatile solvents on recombinant N-
methyl-D-aspartate receptors expressed in Xenopus oocytes. Br J Pharmacol 131: 1303-
1308. http://dx.doi.org/10.1038/si.bip.0703666.
Cummings, BS: Zangar, RC: Novak, RF: Lash, LH. (1999). Cellular distribution of cytochromes
P-450 in the rat kidney. Drug Metab Dispos 27: 542-548.
Cummings, BS: Lasker, JM: Lash, LH. (2000a). Expression of glutathione-dependent enzymes
and cytochrome P450s in freshly isolated and primary cultures of proximal tubular cells
from human kidney. J Pharmacol Exp Ther 293: 677-685.
Cummings, BS: Parker, JC: Lash, LH. (2000b). Role of cytochrome P450 and glutathione S-
transferase alpha in the metabolism and cytotoxicity of trichloroethylene in rat kidney.
Biochem Pharmacol 59: 531-543. http://dx.doi.org/10.1016/S0006-2952(99)00374-3.
Daft JL. (1988). Rapid determination of fumigant and industrial chemical residues in food. J
Assoc Offic Anal Chern 71: 748-760.
Dallas. CE: Chen. XM: O'Barr, K: Muralidhara, S: Varkonyi, P: Bruckner. JV. (1994a).
Development of a physiologically based pharmacokinetic model for perchloroethylene
using tissue concentration-time data. Toxicol Appl Pharmacol 128: 50-59.
http://dx.doi.org/10.1006/taap.1994.1179.
Dallas. CE: Chen. XM: Muralidhara, S: Varkonvl P: Tackett RL: Bruckner. JV. (1994b). Use of
tissue disposition data from rats and dogs to determine species differences in input
parameters for a physiological model for perchloroethylene. Environ Res 67: 54-67.
http://dx.doi.org/10.1006/enrs.1994.1064.
Dallas. CE: Chen. XM: Muralidhara, S: Varkonvl P: Tackett RL: Bruckner. JV. (1995).
Physiologically based pharmacokinetic model useful in prediction of the influence of
species, dose, and exposure route on perchloroethylene pharmacokinetics. J Toxicol
Environ Health 44: 301-317. http://dx.doi.org/10.1006/taap.1994.1179.
Daniel FB: DeAngelo, AB: Stober, JA: Olson. GR: Page. NP. (1992). Hepatocarcinogenicity of
chloral hydrate, 2-chloroacetaldehyde, and dichloroacetic acid in the male B6C3F1
mouse. Fundam Appl Toxicol 19: 159-168. http://dx.doi.org/10.1016/0272-
0590(92)90147-A.
7-14
-------
Daniel, JW. (1963). The metabolism of 36Cl-labelled trichloroethylene and tetrachloroethylene
in the rat. Biochem Pharmacol 12: 795-802. http://dx.doi.org/10.1016/0006-
2952(63)90109-6.
Daniell WE: Claypoole. KH: Checkowav. H: Smith-Weller. T: Pager. SR: Townes. BD:
Rosenstock, L. (1999). Neuropsychological function in retired workers with previous
long-term occupational exposure to solvents. Occup Environ Med 56: 93-105.
Danielsson, BR. (1990). Effects on reproduction of tri- and tetrachloroethylene. Part II:
Tetrachloroethylene. Summary and evaluation of effects on reproduction. (KEMI Report
Series No 3/90:33-49). Solna, Sweden: Nordic Council of Ministers.
David, RM; Moore, MR; Cifone, MA; Finney, DC: Guest D. (1999). Chronic peroxisome
proliferation and hepatomegaly associated with the hepatocellular tumorigenesis of di(2-
ethylhexyl)phthalate and the effects of recovery. Toxicol Sci 50: 195-205.
De Ceaurriz. J: Desiles. JP: Bonnet P: Marignac. B: Mullen J: Guenier. JP. (1983).
Concentration-dependent behavioral changes in mice following short-term inhalation
exposure to various industrial solvents. Toxicol Appl Pharmacol 67: 383-389.
http://dx.doi.org/10.1016/0041-008X(83)90322-8.
de Faire, U; Ericsson, CG: Hamsten, A; Nilsson, J. (1995). Design features of a five-year
Bezafibrate Coronary Atherosclerosis Intervention Trial (BECAIT). Drugs under
Research 21: 105-124.
De La Iglesia, FA; Lewis, JE; Buchanan, RA; Marcus, EL; McMahon, G. (1982). Light and
electron microscopy of liver in hyperlipoproteinemic patients under long-term
gemfibrozil treatment. Atherosclerosis 43: 19-37. http://dx.doi.org/10.1016/0021-
9150(82)90096-X.
De Roos. A: Olshan. A: Teschke. K: Poole. C: Savitz. D: Blatt J: Bondv. M: Pollock. B. (2001).
Parental occupational exposures to chemicals and incidence of neuroblastoma in
offspring. Am JEpidemiol 154: 106-114. http://dx.doi.Org/10.1093/aie/154.2.106.
DeAngelo. AB: Daniel FB: Stober. JA: Olson. GR. (1991). The carcinogenicity of
dichloroacetic acid in the male B6C3F1 mouse. Fundam Appl Toxicol 16: 337-347.
http://dx.doi.org/10.1016/0272-0590(91)90118-N.
DeAngelo, AB: Daniel, FB: Most BM: Olson, GR. (1996). The carcinogenicity of dichloroacetic
acid in the male fischer 344 rat. Toxicology 114: 207-221.
http://dx.doi.org/10.1016/S0300-483X(96)03510-X.
DeAngelo, AB: Daniel, FB: Most BM: Olson, GR. (1997). Failure of monochloroacetic acid and
trichloroacetic acid administered in the drinking water to produce liver cancer in male
F344/N rats. J Toxicol Environ Health 52: 425-445.
http://dx.doi.org/10.1080/00984109708984074.
DeAngelo, AB: George, MH: House, DE. (1999). Hepatocarcinogenicity in the male B6C3F1
mouse following a lifetime exposure to dichloroacetic acid in the drinking water: Dose-
response determination and modes of action. J Toxicol Environ Health A 58: 485-507.
DeAngelo. AB: Daniel FB: Wong. DM: George. MH. (2008). The induction of hepatocellular
neoplasia by trichloroacetic acid administered in the drinking water of the male B6C3F1
mouse. J Toxicol Environ Health A 71: 1056-1068.
http://dx.doi.org/10.1080/15287390802111952.
Dees, C: Travis, C. (1994). Trichloroacetate stimulation of liver DNA synthesis in male and
female mice. Toxicol Lett 70: 343-355. http://dx.doi.org/10.1016/0378-4274(94)90129-5.
7-15
-------
Degrassi, F; Tanzarella, C. (1988). Immunofluorescent staining of kinetochores in micronuclei:
A new assay for the detection of aneuploidy. Mutat Res 203: 339-345.
http://dx.doi.org/10.1016/0165-1161(88)90030-1.
Dekant W: Vamvakas. S: Berthold. K: Schmidt S: Wild. D: Henschler. D. (1986a). Bacterial
beta-lyase mediated cleavage and mutagenicity of cysteine conjugates derived from the
nephrocarcinogenic alkenes trichloroethylene, tetrachloroethylene and
hexachlorobutadiene. Chem Biol Interact 60: 31-45. http://dx.doi.org/10.1016/0009-
2797(86)90015-3.
Dekant W: Vamakas. S: Berthold. K: Schmitt S: Henschler. D. (1986b). Cysteine conjugate
beta-lyase mediated cleavage and mutagenicity of the tetrachloroethylene metabolite s-
1,1,2-trichlorovinyl-cysteine. Toxicol Lett 31: 209.
Dekant W: Metzler, M; Henschler, D. (1986c). Identification of S-l,2-dichlorovinyl-N-acetyl-
cysteine as a urinary metabolite of trichloroethylene: A possible explanation for its
nephrocarcinogenicity in male rats. Biochem Pharmacol 35: 2455-2458.
http://dx.doi.org/10.1016/0006-2952(86)90039-0.
Dekant W: Metzler, M; Henschler, D. (1986d). Identification of 8-1,2,2-trichlorovinyl-N-
acetylcysteine as a urinary metabolite of tetrachloroethylene: Bioactivation through
glutathione conjugation as a possible explanation of its nephrocarcinogenicity. J Biochem
Toxicol 1: 57-72. http://dx.doi.org/10.1002/ibt.2570010206.
Dekant W: Martens, G: Vamvakas, S: Metzler, M: Henschler, D. (1987). Bioactivation of
tetrachloroethylene: role of glutathione S-transferase-catalyzed conjugation versus
Cytochrome P-450-dependent phospholipid alkylation. Drug Metab Dispos 15: 702-709.
Dekant W: Berthold, K: Vamvakas. S: Henschler. D: Anders. MW. (1988). Thioacylating
intermediates as metabolites of S-(l,2-dichlorovinyl)-L-cysteine and S-(l,2,2-
trichlorovinyl)-L-cysteine formed by cysteine conjugate beta-lyase. Chem Res Toxicol 1:
175-178. http://dx.doi.org/10.1021/tx00003a008.
Dekant W: Vamvakas, S: Anders, MW. (1989). Bioactivation of nephrotoxic haloalkenes by
glutathione conjugation: formation of toxic and mutagenic intermediates by cysteine
conjugate beta-lyase. Drug Metab Rev 20: 43-83.
http://dx.doi.org/10.3109/03602538908994144.
Dekant W: Vamvakas, S: Anders, MW. (1990). Biosynthesis, bioactivation, and mutagenicity of
S-conjugates. Toxicol Lett 53: 53-58. http://dx.doi.org/10.1016/0378-4274(90)90094-3.
Dekant, W: Birner, G: Werner, M: Parker, J. (1998). Glutathione conjugation of perchloroethene
in subcellular fractions from rodent and human liver and kidney. Chem Biol Interact 116:
31-43. http://dx.doi.org/10.1080/20018091094565.
Delahunt, B; Bethwaite, PB; Nacey, JN. (1995). Occupational risk for renal cell carcinoma. A
case-control study based on the New Zealand Cancer Registry. Br J Urol 75: 578-582.
http://dx.doi.Org/10.llll/i.1464-410X.1995.tb07410.x.
Delfino, RJ: Gone, H; Linn, WS: Pellizzari, ED; Hu, Y. (2003a). Asthma symptoms in Hispanic
children and daily ambient exposures to toxic and criteria air pollutants. Environ Health
Perspect 111: 647-656. http://dx.doi.org/10.1289/ehp.5992.
Delfino, RJ: Gong, H; Linn, WS: Hu, Y; Pellizzari, ED. (2003b). Respiratory symptoms and
peak expiratory flow in children with asthma in relation to volatile organic compounds in
exhaled breath and ambient air. J Expo Anal Environ Epidemiol 13: 348-363.
http://dx.doi.org/10.1038/si.iea.7500287.
7-16
-------
DeMarini, DM; Perry, E; Shelton, ML. (1994). Dichloroacetic acid and related compounds:
Induction of prophage in E. coli and mutagenicity and mutation spectra in Salmonella
TA100. Mutagenesis 9: 429-437. http://dx.doi.Org/10.1093/mutage/9.5.429.
Deonandan, R; Campbell, K; Ostbye, T; Tummon, I; Robertson, J. (2000). A comparison of
methods for measuring socio-economic status by occupation or postal area. Chronic Dis
Can 21: 114-118.
Diabetes Atherosclerosis Intervention Study Investigators. (2001). Effect of fenofibrate on
progression of coronary-artery disease in type 2 diabetes: The Diabetes Atherosclerosis
Intervention Study, a randomised study. Lancet 357: 905-910.
http://dx.doi.org/10.1016/S0140-6736(00)04209-4.
Dietert, RR. (2008). Developmental immunotoxicology (DIT): windows of vulnerability,
immune dysfunction and safety assessment. J Immunotoxicol 5: 401-412.
http://dx.doi.org/10.1080/15476910802483324.
Ding. R: Shen. T: Zhu. QX. (2006). [Effects of TCE and PCE on cultured human keratinocyte
lipid peroxidation and protective effect of vitamin E on it]. Zhonghua Lao Dong Wei
Sheng Zhi Ye Bing Za Zhi 24: 662-665.
Diodovich, C; Ferrario, D; Casati, B; Malerba, I; Marafante, E: Parent-Massin, D; Gribaldo, L.
(2005). Sensitivity of human cord blood cells to tetrachloroethylene: Cellular and
molecular endpoints. Arch Toxicol 79: 508-514. http://dx.doi.org/10.1007/s00204-005-
0662-8.
Dmitri eva, NV. (1967). [On the metabolism of tetrachloroethylene]. Gig Tr Prof Zabol 11: 54-
56.
Dobrev, ID: Andersen, ME; Yang, RS. (2001). Assessing interaction thresholds for
trichloroethylene in combination with tetrachloroethylene and 1,1,1-trichloroethane using
gas uptake studies and PBPK modeling. Arch Toxicol 75: 134-144.
http://dx.doi.org/10.1007/s002040100216.
Dobrev, ID: Andersen, ME; Yang, RS. (2002). In silico toxicology: Simulating interaction
thresholds for human exposure to mixtures of trichloroethylene, tetrachloroethylene, and
1,1,1-trichloroethane. Environ Health Perspect 110: 1031-1039.
Doherty, AT; Ellard, S; Parry, EM; Parry, JM. (1996). An investigation into the activation and
deactivation of chlorinated hydrocarbons to genotoxins in metabolically competent
human cells. Mutagenesis 11: 247-274. http://dx.doi.Org/10.1093/mutage/ll.3.247.
Dohn, PR; Anders, MW. (1982). Assay of cysteine conjugate beta-lyase activity with S-(2-
benzothiazolyl)cysteine as the substrate. Anal Biochem 120: 379-386.
http://dx.doi.org/10.1016/0003-2697(82)90361-X.
Pol A: Hill G: Seelv. J: Hailev. J: Kissling. G: Bucher. J. (2007). alpha 2u-globulin
nephropathy and renal tumors in national toxicology program studies. Toxicol Pathol 35:
533-540. http://dx.doi.org/10.1080/01926230701338941.
Donoghue, AM; Dryson, EW: Wynn-Williams, G. (1995). Contrast sensitivity in organic-
solvent-induced chronic toxic encephalopathy. J Occup Environ Med 37: 1357-1363.
Dome, J: Walton, K; Renwick, A. (2001). Human variability in glucuronidation in relation to
uncertainty factors for risk assessment. Food Chem Toxicol 39: 1153-1173.
Dome, JLC, M; Renwick, AG. (2005). The refinement of uncertainty/safety factors in risk
assessment by the incorporation of data on toxicokinetic variability in humans. Toxicol
Sci 86: 20-26. http://dx.doi.org/10.1093/toxsci/kfil60.
7-17
-------
Dosemeci, M; Wacholder, S: Lubin, JH. (1990). Does nondifferential misclassification of
exposure always bias a true effect toward the null value? Am J Epidemiol 132: 746-748.
Dosemeci, M; Cocco, P; Chow, WH. (1999). Gender differences in risk of renal cell carcinoma
and occupational exposures to chlorinated aliphatic hydrocarbons. Am J Ind Med 36: 54-
59. http://dx.doi.org/10.1002/(SICni097-0274(199907)36:l<54::AID-AJIM8>3.0.CO:2-
0.
Doyle, P: Roman, E; Beral, V: Brookes, M. (1997). Spontaneous abortion in dry cleaning
workers potentially exposed to perchloroethylene. Occup Environ Med 54: 848-853.
http://dx.doi.org/10.1136/oem.54.12.848.
Dreessen, B; Westphal G: Biinger, J: Hallier, E; Miiller, M. (2003). Mutagenicity of the
glutathione and cysteine S-conjugates of the haloalkenes l,l,2-trichloro-3,3,3-trifluoro-l-
propene and trichlorofluoroethene in the Ames test in comparison with the
tetrachloroethene-analogues. Mutat Res Genet Toxicol Environ Mutagen 539: 157-166.
http://dx.doi.org/10.1016/S1383-5718(03)00160-8.
Droz, PO: Guillemin, MP. (1986). Occupational exposure monitoring using breath analysis. J
Occup Med 28: 593-602.
Droz, PO: Wu, MM; Cumberland, WG. (1989a). Variability in biological monitoring of organic
solvent exposure II application of a population physiological model. Occup Environ Med
46: 547-558.
Droz, PO: Wu, MM; Cumberland, WG: Berode, M. (1989b). Variability in biological monitoring
of solvent exposure. I. Development of a population physiological model. Br J Ind Med
46: 447-460.
Dryson, E: 't Mannetje, A; Walls, C: McLean, D; McKenzie, F; Maule, M: Cheng, S:
Cunningham, C; Kromhout H; Boffetta, P: Blair, A; Pearce, N. (2008). Case-control
study of high risk occupations for bladder cancer in New Zealand. Int J Cancer 122:
1340-1346. http://dx.doi.org/10.1002/iic.23194.
Duffel, MW: Jakoby, WB. (1982). Cysteine S-conjugate N-acetyltransferase from rat kidney
microsomes. Mol Pharmacol 21: 444-448.
Eastmond, DA; Tucker, JD. (1989). Kinetochore localization in micronucleated cytokinesis-
blocked Chinese hamster ovary cells: A new and rapid assay for identifying aneuploidy-
inducing agents. Mutat Res 224: 517-525. http://dx.doi.org/10.1016/0165-
1218(89)90079-7.
Eastmond, DA; Hartwig, A; Anderson, D; Anwar, WA; Cimino, MC: Dobrev, I; Douglas, GR:
Nohmi, T: Phillips, DH; Vickers, C. (2009). Mutagenicity testing for chemical risk
assessment: Update of the WHO/IPCS harmonized scheme. Mutagenesis 24: 341-349.
http://dx.doi.org/10.1093/mutage/gep014.
Ebrahim, AS: Babakrishnan, K; Sakthisekaran, D. (1996). Perchloroethylene-induced alterations
in glucose metabolism and their prevention by 2-deoxy-D-glucose and vitamin E in mice.
J Appl Toxicol 16: 339-348. http://dx.doi.org/10.1002/(SICIU099-
1263O99607)16:4<339::AID-JAT352>3.0.CO:2-3.
Ebrahim, AS: Babu, E: Thirunavukkarasu, C: Sakthisekaran, D. (2001). Protective role of
vitamin E, 2-deoxy-D-glucose, and taurine on perchloroethylene induced alterations in
ATPases. Drug Chem Toxicol 24: 429-437. http://dx.doi.org/10.1081/DCT-100106267.
Echeverria, D: Heyer, N: Checkoway, H; Brodkin, CA; Bittner, A, Jr; Toutonghi, G: Ronhovde,
K (1994). A behavioral investigation of occupational exposures to solvents:
Perchloroethylene among dry cleaners, and styrene among reinforced fiberglass
7-18
-------
laminators. (BSRC-100/94/040). Seattle, WA: Battelle Centers for Public Health
Research and Evaluation.
Echeverria, D: White, RF; Sampaio, C. (1995). A behavioral evaluation of PCE exposure in
patients and dry cleaners: A possible relationship between clinical and preclinical effects.
J Occup Environ Med 37: 667-680.
Edwards, BK; Ward, E; Kohler, BA; Eheman, C: Zauber, AG: Anderson, RN: Jemal A;
Schymura, MJ: Lansdorp-Vogelaar, I; Seeff, LC: van Ballegooijen, M; Goede, SL; Ries,
LAG. (2010). Annual report to the nation on the status of cancer, 1975-2006, featuring
colorectal cancer trends and impact of interventions (risk factors, screening, and
treatment) to reduce future rates. Cancer 116: 544-573.
http://dx.doi.org/10.1002/cncr.24760.
Eger El, II. (1963). A mathematical model of uptake and distribution. In EM Papper; RJ Kitz
(Eds.), Uptake and distribution of anesthetic agents (pp. 72-87). New York, NY:
McGraw-Hill Book Company, Inc.
El-Serag, HB. (2007). Epidemiology of hepatocellular carcinoma in USA. Hepatology Research
37: S88-S94. http://dx.doi.Org/10.llll/i.1872-034X.2007.00168.x.
Elfarra, AA: Krause, RJ. (2007). S-(l,2,2-trichlorovinyl)-L-cysteine sulfoxide, a reactive
metabolite of S-(l,2,2-Trichlorovinyl)-L-cysteine formed in rat liver and kidney
microsomes, is a potent nephrotoxicant. J Pharmacol Exp Ther 321: 1095-1101.
http://dx.doi.org/10.1124/ipet.107.120444.
Emara, AM; Abo El-Noor, MM; Hassan, NA; Wagih, AA. (2010). Immunotoxicity and
hematotoxicity induced by tetrachloroethylene in egyptian dry cleaning workers. Inhal
Toxicol22: 117-124. http://dx.doi.org/10.3109/08958370902934894.
Emmert B: Bunger. J: Keuch. K: Mullen M: Emmert S: Hallier. E: Westphal GA. (2006).
Mutagenicity of cytochrome P450 2E1 substrates in the Ames test with the metabolic
competent S. typhimurium strain YG7108pin3ERb5. Toxicology 228: 66-76.
http://dx.doi.0rg/10.1016/i.tox.2006.08.013.
Entz, RC: Diachenko, GW. (1988). Residues of volatile halocarbons in margarines. Food Addit
Contam 5: 267-276.
Environment Canada & Health Canada. (1993). Tetrachloroethylene. Ottawa, Canada: Minister
of Supply and Services Canada, http://www.hc-sc.gc.ca/ewh-
semt/pubs/contaminants/psl 1 -Isp 1/trichloroethylene/index-eng.php.
Epstein. PL: Nolen. GA: Randall JL: Christ SA: Read. EJ: Stober. JA: Smith. MK. (1992).
Cardiopathic effects of dichloroacetate in the fetal Long-Evans rat. Teratology 46: 225-
235. http://dx.doi.org/10.1002/tera.1420460306.
Eskenazi, B: Bracken, MB; Holford, TR; Grady, J. (1988). Exposure to organic solvents and
hypertensive disorders of pregnancy. Am J Ind Med 14: 177-188.
http://dx.doi.org/10.1002/aiim.4700140208.
Eskenazi, B: Wyrobek, AJ: Fenster, L; Katz, DF; Sadler, M: Lee, J: Hudes, M: Rempel DM.
(1991a). A study of the effect of perchloroethylene exposure on semen quality in dry
cleaning workers. Am J Ind Med 20: 575-591.
http://dx.doi.org/10.1002/aiim.4700200502.
Eskenazi. B: Fenster. L: Hudes. M: Wvrobek. AJ: Katz. DF: Gerson. J: Rempel DM. (1991b). A
study of the effect of perchloroethylene exposure on the reproductive outcomes of wives
of dry-cleaning workers. Am J Ind Med 20: 593-600.
http://dx.doi.org/10.1002/aiim.4700200503.
7-19
-------
Essing, HG: Schacke, G: Schaller, KH: Valentin, H. (1973). [Occupational-medical studies on
the dynamics of tetrachloroethylene in the organism]. 24: 242-244.
Eveillard, A; Mselli-Lakhal, L: Mogha, A; Lasserre, F; Polizzi, A; Pascussi, JM: Guillou, H:
Martin. PG: Pineau. T. (2009). Di-(2-ethylhexyl)-phthalate (DEHP) activates the
constitutive androstane receptor (CAR): a novel signalling pathway sensitive to
phthalates. Biochem Pharmacol 77: 1735-1746.
http://dx.doi.0rg/10.1016/i.bcp.2009.02.023.
Fabbro-Peray, P; Daures, JP: Rossi, JF. (2001). Environmental risk factors for non-Hodgkin's
lymphoma: A population-based case-control study in Languedoc-Roussillon, France.
Cancer Causes Control 12: 201-212. http://dx.doi.Org/10.1023/A:1011274922701.
FDA (U.S. Food and Drug Administration). (2003). Food and Drug Administration total diet
study: Summary of residues found ordered by pesticide market baskets 91-3-01-4.
Washington, DC. http://www.ntis.gov/search/product.aspx?ABBR=PB2005108087.
Fender, H. (1993). Chromosomenanalytische undersuchungen bei textilreinigern. In D Arndt; G
Obe (Eds.), Cytogenetische methoden im rahmen des populationsmonitoring (pp. 71-76).
Miinchen, Germany: MMV Verlag.
Fernandez, J: Guberan, E: Caperos, J. (1976). Experimental human exposures to
tetrachloroethylene vapor and elimination in breath after inhalation. Am Ind Hyg Assoc J
36: 143-150.
F errand, JF: Cenee, S; Laurent-Puig, P: Loriot MA; Trinchet JC: Degos, F; Bronovicky, JP:
Pelletier, G: Stacker, I. (2008). Hepatocellular carcinoma and occupation in men: A case-
control study. J Occup Environ Med 50: 212-220.
http://dx.doi.org/10.1097/JOM.Ob013e31815d88e2.
Ferroni, C; Sells, L: Mutti, A; Folli, D: Bergamaschi, E: Franchini, I. (1992). Neurobehavioral
and neuroendocrine effects of occupational exposure to perchloroethylene.
Neurotoxicology 13: 243-247.
Filser, JG: Bolt FfJVI. (1979). Pharmacokinetics of halogenated ethylenes in rats. Arch Toxicol
42: 123-136. http://dx.doi.org/10.1007/BF00316492.
Finckh, A: Cooper. GS: Chibnik, LB: Costenbader, KH: Watts. J: Pankev, H: Fraser. PA:
Karlson, EW. (2006). Occupational silica and solvent exposures and risk of systemic
lupus erythematosus in urban women. Arthritis Rheum 54: 3648-3654.
http://dx.doi.org/10.1002/art.22210.
Fisher, J: Mahle, D: Bankston, L: Greene, R: Gearhart, J. (1997). Lactational transfer of volatile
chemicals in breast milk. Am Ind Hyg Assoc J 58: 425-431.
http://dx.doi.org/10.1080/15428119791012667.
Fisher. J: Lumpkin. M: Bovd. J: Mahle. D: Bruckner. JV: El-Masri. HA. (2004). PBPK modeling
of the metabolic interactions of carbon tetrachloride and tetrachloroethylene in B6C3F1
mice. Environ Toxicol Pharmacol 16: 93-15.
http://dx.doi.0rg/10.1016/j.etap.2003.10.006.
Fisher, JW. (1994). Estimating human health risks for trichloroethylene, a water contaminant,
using physiological models. In RGM Wang (Ed.), Water contamination and health:
Integration of exposure assessment, toxicology, and risk assessment (pp. 393 - 406). New
York: Marcel Dekker.
Flavell, DM: Jamshidi, Y; Hawe, E: Pineda Torra, I; Taskinen, MR: Frick, MH: Nieminen, MS:
Kesaniemi, YA: Pasternack, A: Staels, B: Miller, G: Humphries, SE: Talmud, PJ:
Svvanne, M. (2002). Peroxisome proliferator-activated receptor alpha gene variants
7-20
-------
influence progression of coronary atherosclerosis and risk of coronary artery disease.
Circulation 105: 1440-1445.
Flavell DM; Ireland, H; Stephens, JW: Hawe, E; Acharya, J: Mather, H; Hurel SJ: Humphries,
SE. (2005). Peroxisome proliferator-activated receptor alpha gene variation influences
age of onset and progression of type 2 diabetes. Diabetes 54: 582-586.
Fleming-Jones, ME; Smith, RE. (2003). Volatile organic compounds in foods: A five year study.
JAgricFoodChemSl: 8120-8127. http://dx.doi.org/10.1021/if0303159.
Foucher, C; Rattier, S; Flavell DM; Talmud, PJ: Humphries, SE: Kastelein, JJ: Avyobi, A;
Pimstone, S; Frohlich, J: Ansquer, JC: Steiner, G: investigators, D. (2004). Response to
micronized fenofibrate treatment is associated with the peroxisome-proliferator-activated
receptors alpha G/C intron? polymorphism in subjects with type 2 diabetes.
Pharmacogenetics 14: 823-829.
Fox. AW: Yang. X: Murli. H: Lawlor. TE: Cifone. MA: Reno. FE. (1996). Absence of
mutagenic effects of sodium dichloroacetate. Fundam Appl Toxicol 32: 87-95.
http://dx.doi.org/10.1006/faat.1996.0110.
Fox. TR: Schumann. AM: Watanabe. PG: Yano. BL: Maher. VM: McCormick. JJ. (1990).
Mutational analysis of the H-ras oncogene in spontaneous C57BL/6 x C3H/He mouse
liver tumors and tumors induced with genotoxic and nongenotoxic hepatocarcinogens.
Cancer Res 50: 4014-4019.
Franchini, I; Cavatorta, A; Falzoi, M; Lucertini, S: Mutti, A. (1983). Early indicators of renal
damage in workers exposed to organic solvents. Int Arch Occup Environ Health 52: 1-9.
http://dx.doi.org/10.1007/BF00380601.
Frankel, DM; Johnson, CE; Pitt HM. (1957). Preparation and properties of tetrachloroethylene
oxide. J Org Chem 22: 1119-1120.
Frantz, SW: Watanabe, PG. (1983). Tetrachloroethylene: Balance and tissue distribution in male
Sprague-Dawley rats by drinking-water administration. Toxicol Appl Pharmacol 69: 66-
72. http://dx.doi.org/10.1016/0041-008X(83)90120-5.
Fredriksson, A; Danielsson, BRG: Eriksson, P. (1993). Altered behaviour in adult mice orally
exposed to tri- and tetrachloroethylene as neonates. Toxicol Lett 66: 13-19.
http://dx.doi.org/10.1016/0378-4274(93)90074-8.
Freeman. SR: Drake. AL: Heilig. LF: Graber. M: McNealv. K: Schilling. LM: Dellavalle. RP.
(2006). Statins, fibrates, and melanoma risk: a systematic review and meta-analysis. 98:
1538-1546. http://dx.doi.org/10.1093/inci/dii412.
Frenette, B; Mergler, D; Bowler, R. (1991). Contrast-sensitivity loss in a group of former
microelectronics workers with normal visual acuity. Optometry and Vision Science 68:
556-560.
Frick, M: Elo, O: Haapa, K; Heinonen, O: Heinsalmi, P; Helo, P; Huttunen, J: Kaitaniemi, P;
Koskinen, P; Manninen, V. (1987). Helsinki Heart Study: Primary-prevention trial with
gemfibrozil in middle-aged men with dyslipidemia: Safety of treatment, changes in risk
factors, and incidence of coronary heart disease. NEngl JMed 317: 1237-1245.
http://dx.doi.org/10.1056/NEJM198711123172001.
Frick, MH; Svvanne, M: Nieminen, MS: Kauma, H: Majahalme, S: Virtanen, V: Kesaniemi, YA;
Pasternack, A; Taskinen, MR. (1997). Prevention of the angiographic progression of
coronary and vein-graft atherosclerosis by gemfibrozil after coronary bypass surgery in
men with low levels of HDL cholesterol. Lopid Coronary Angiography Trial (LOCAT)
Study Group. Circulation 96: 2137-2143.
7-21
-------
Furnus, CC: Ulrich, MA; Terreros, MC: Dulout FN. (1990). The induction of aneuploidy in
cultured Chinese hamster cells by propionaldehyde and chloral hydrate. Mutagenesis 5:
323-326.
Fuscoe. JC: Afshari. AJ: George. MH: DeAngelo. AB: Tice. RR: Salman. T: Allen. JW. (1996).
In vivo genotoxicity of dichloroacetic acid: Evaluation with the mouse peripheral blood
micronucleus assay and the single cell gel assay. Environ Mol Mutagen 27: 1-9.
http://dx.doi.org/10.1002/(SICB1098-2280(1996)27:Kl::AID-EMl>3.0.CO:2-L.
Gaertner, RRW: Trpeski, L; Johnson, KC. (2004). A case-control study of occupational risk
factors for bladder cancer in Canada. Cancer Causes Control 15: 1007-1019.
http://dx.doi.org/10.1007/sl0552-004-1448-7.
Gaillard, Y; Billault F; Pepin, G. (1995). Tetrachloroethylene fatality: Case report and simple
gas chromatographic determination in blood and tissues. Forensic Sci Int 76: 161-168.
http://dx.doi.org/10.1016/0379-0738(95)01813-l.
Galloway, SM; Armstrong, MJ: Reuben, C; Colman, S; Brown, B; Cannon, C; Bloom, AD;
Nakamura, F; Ahmed, M; Duk, S; Rimpo, J; Margolin, BH; Resnick, MA; Anderson, B;
Zeiger, E. (1987). Chromosome aberrations and sister chromatid exchanges in Chinese
hamster ovary cells: Evaluations of 108 chemicals. Environ Mol Mutagen 10: 1-175.
http://dx.doi.org/10.1002/em.2850100502.
Gama, C: Meira, JB. (1978). Occupational acro-osteolysis. J Bone Joint Surg Am 60: 86-90.
Gandhi. M: Aweeka, F: Greenblatt RM: Blaschke, TF. (2004). Sex differences in
pharmacokinetics and pharmacodynamics. Annu Rev Pharmacol Toxicol 44: 499-523.
http://dx.doi.org/10.1146/annurev.pharmtox.44.101802.121453.
Panning, A; Brunk, U; Dallner, G. (1984). Phthalate esters and their effect on the liver.
Hepatology4: 541-547.
Ganning, AE; Brunk, U: Edlund, C: Elhammer, A; Dallner, G. (1987). Effects of prolonged
administration of phthalate ester on the liver. Environ Health Perspect 73: 251-258.
Garabrant DH: Lacev JV. J. r: Laing, TJ: Gillespie, BW: Maves, MD: Cooper. BC:
Schottenfeld, D. (2003). Scleroderma and solvent exposure among women. Am J
Epidemiol 157: 493-500. http://dx.doi.org/10.1093/aje/kwf223.
Garetano, G: Gochfeld, M. (2000). Factors influencing tetrachloroethylene concentrations in
residences above dry-cleaning establishments. Arch Environ Health 55: 59-68.
http://dx.doi.org/10.1080/00039890009603387.
Gariot P: Barret E: Drouin, P: Genton, P: Pointel JP: Foliguet B: Kolopp, M: Debry, G.
(1987). Morphometric study of human hepatic cell modifications induced by fenofibrate.
Metabolism 36: 203-210.
Gamier, R; Bedouin, J: Pepin, G: Gaillard, Y. (1996). Coin-operated dry cleaning machines may
be responsible for acute tetrachloroethylene poisoning: Report of 26 cases including one
death. Clin Toxicol 34: 191-197.
Garte, S: Gaspari, L; Alexandrie, AK; Ambrosone, C: Autrup, H; Autrup, JL; Baranova, H;
Bathum, L; Benhamou, S: Boffetta, P: Bouchardy, C: Breskvar, K; Brockmoller, J:
Cascorbi. I: Clapper. ML: Coutelle. C: Daly. A: DeU'Omo. M: Dolzan. V: Dresler. CM:
Fryer, A; Haugen, A; Hein, DW: Hildesheim, A; Hirvonen, A; Hsieh, LL; Ingelman-
Sundberg, M: Kalina, I; Kang, D; Kihara, M: Kiyohara, C: Kremers, P: Lazarus, P: Le
Marchand, L: Lechner, MC: van Lieshout EM; London, S: Manni, JJ: Maugard, CM;
Morita. S: Nazar-Stewart. V: Noda. K: Oda. Y: Parl. FF: Pastorelli. R: Persson. I: Peters.
WH; Rannug, A; Rebbeck, T; Risch, A; Roelandt L: Romkes, M: Ryberg, D: Salagovic,
7-22
-------
J: Schoket B: Seidegard. J: Shields. PG: Sim. E: Sinnet D: Strange. RC: Stacker. I:
Sugimura, H; To-Figueras, J: Vineis, P: Yu, MC: Taioli, E. (2001). Metabolic gene
polymorphism frequencies in control populations. Cancer Epidemiol Biomarkers Prev 10:
1239-1248.
Ge, R: Yang, S: Kramer, PM; Tao, L; Pereira, MA. (2001). The effect of dichloroacetic acid and
trichloroacetic acid on DNA methylation and cell proliferation in B6C3F1 mice. J
BiochemMol Toxicol 15: 100-106. http://dx.doi.Org/10.1002/jbt.5.
Gearhait JM: Mahle. DA: Greene. RJ: Seckel CS: Flemming. CD: Fisher. JW: IIL CH. (1993).
Variability of physiologically based pharmacokinetic (PBPK) model parameters and their
effects on PBPK model predictions in a risk assessment for perchloroethylene (PCE).
Toxicol Lett 68: 131-144. http://dx.doi.org/10.1016/0378-4274(93)90126-1.
Geller, AM. (2001). A table of color distance scores for quantitative scoring of the Lanthony
Desaturate color vision test. Neurotoxicol Teratol 23: 265-267.
http://dx.doi.org/10.1016/S0892-0362(01)00139-8.
Gennari. P: Naldl M: Motta. R: Nuccl MC: Giacominl C: Violante. FS: Raffi. GB. (1992).
gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
tetrachloroethylene. Am J Ind Med 21: 661-671.
http://dx.doi.org/10.1002/aiim.4700210506.
Gentry. PR: Covington. TR: Clewell HJ. III. (2003). Evaluation of the potential impact of
pharmacokinetic differences on tissue dosimetry in offspring during pregnancy and
lactation. Regul Toxicol Pharmacol 38: 1-16. http://dx.doi.org/10.1016/S0273-
2300(03)00047-3.
George. JD: Price. CJ: Navarro. HA: Marr. MC: Myers. CB: Hunter ES Schwetz. BA: Shelby.
MD. (1995). Developmental toxicity of thiophenol (THIO) in rats and rabbits [Abstract].
Toxicol ogist 15: 160.
Germolec, PR: Yang, RSH; Ackermann, MF; Rosenthal, GJ: Boorman, GA; Blair, P: Luster,
MI. (1989). Toxicology studies of a chemical mixture of 25 groundwater contaminants:
II. Immunosuppression in B6C3F1 mice. Fundam Appl Toxicol 13: 377-387.
http://dx.doi.org/10.1016/0272-0590(89)90275-3.
Ghantous, H; Danielsson, BRG: Dencker, L; Gorczak, J: Vesterberg, O. (1986). Trichloroacetic
acid accumulates in murine amniotic fluid after tri- and tetrachloroethylene inhalation.
ActaPharmacol Toxicol 58: 105-114. http://dx.doi.Org/10.l 11 l/i.1600-
0773.1986.tb00078.x.
Gibson, D; Aardema, M; Kerckaert, G: Carr, G: Brauninger, R: LeBoeuf, R. (1995). Detection of
aneuploidy-inducing carcinogens in the Syrian hamster embryo (SHE) cell
transformation assay. MutatRes 343: 7-24. http://dx.doi.org/10.1016/0165-
1218(95)90058-6.
Gilbert. D: Gover. M: Lyman. W: Magil G: Walker. P: Wallace. D: Wechsler. A: Yee. J. (1982).
An exposure and risk assessment for tetrachloroethylene. (EPA-440/4-85-015).
Washington, DC: U.S. Environmental Protection Agency, Office of Water Regulations
and Standards. http://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=2000LLOH.txt.
Giller, S: Le Curieux, F; Gauthier, L: Erb, F; Marzin, D. (1995). Genotoxicity assay of chloral
hydrate and chloropicrine. MutatRes 348: 147-152. http://dx.doi.org/10.1016/0165-
7992(95)90002-0.
Giller, S: Le Curieux, F: Erb, F: Marzin, D. (1997). Comparative genotoxicity of halogenated
acetic acids found in drinking water. Mutagenesis 12: 321-328.
7-23
-------
Ginsberg, G: Hattis, D: Russ, A; Sonawane, B. (2005). Pharmacokinetic and pharmacodynamic
factors that can affect sensitivity to neurotoxic sequelae in elderly individuals. Environ
Health Perspect 113: 1243-1249. http://dx.doi.org/10.1289/ehp.7568.
Ginsberg. G: Smolensk!. S: Hattis. D: Guvton. KZ: Johns. DO: B. S. (2009). Genetic
polymorphism in glutathione transferases (GST): Population distribution of GSTM1, Tl,
and PI conjugating activity. J Toxicol Environ Health B Crit Rev 12: 389-439.
http://dx.doi.org/10.1080/10937400903158375.
Gobba, F; Righi, E; Fantuzzi, G: Predieri, G: Cavazzuti, L; Aggazzotti, G. (1998). Two-year
evolution of perchloroethylene-induced col or-vision loss. Arch Environ Health 53: 196-
198.
Gochfeld, M. (2007). Framework for gender differences in human and animal toxicology.
Environ Res 104: 4-21. http://dx.doi.Org/10.1016/i.envres.2005.12.005.
Gold. LS: Manlev. NB: Slone. TH: Rohrbach. L: Garfmkel. GB. (2005). Supplement to the
Carcinogenic Potency Database (CPDB): results of animal bioassays published in the
general literature through 1997 and by the National Toxicology Program in 1997-1998.
Toxicol Sci 85: 747-808. http://dx.doi.org/10.1093/toxsci/kfil61.
Gold, LS: De Roos, AJ: Waters, M; Stewart, P. (2008). Systematic literature review of uses and
levels of occupational exposure to tetrachloroethylene [Review]. J Occup Environ Hyg 5:
807-839. http://dx.doi.org/10.1080/15459620802510866.
Gold. LS: Milliken. K: Stewart. P: Purdue. M: Severson. R: Seixas. N: Blair. A: Davis. S:
Hartge, P; De Roos, AJ. (2010a). Occupation and multiple myeloma: An occupation and
industry analysis. Am J Ind Med 53: 768-779. http://dx.doi.org/10.1002/aiim.20857.
Gold. LS: Stewart. PA: Milliken. K: Purdue. M: Severson. R: Seixas. N: Blair. A: Hartge. P:
Davis, S: De Roos, AJ. (201 Ob). The relationship between multiple myeloma and
occupational exposure to six chlorinated solvents. Occup Environ Med 68: 391-399.
http://dx.doi.org/10.1136/oem.2009.054809.
Goldman, JA. (1996). Connective tissue disease in people exposed to organic chemical solvents:
Systemic sclerosis (scleroderma) in dry cleaning plant and aircraft industry workers. J
Clin Rheumatol 2: 185-190.
Goldstein, A; Aronow, L: Kalman, SM. (1969). Principles of drug action: The basis of
pharmacology. New York, NY: Harper and Row.
Goldsworthy, TL: Popp, JA. (1987). Chlorinated hydrocarbon-induced peroxisomal enzyme
activity in relation to species and organ carcinogenicity. Toxicol Appl Pharmacol 88:
225-233. http://dx.doi.org/10.1016/0041-008X(87)90008-1.
Goldsworthy. TL: Lyght O: Burnett. VL: Popp. JA. (1988). Potential role of [alpha]-2 [mu]-
globulin, protein droplet accumulation, and cell replication in the renal carcinogenicity of
rats exposed to trichloroethylene, perchloroethylene, and pentachloroethane. Toxicol
Appl Pharmacol 96: 367-379. http://dx.doi.org/10.1016/0041-008X(88)90095-6.
Gonzalez, CA; Lopez-Abente, G: Errezola, M: Escolar, A; Riboli, E; Izarzugaza, I; Nebot M.
(1989). Occupation and bladder cancer in Spain: A multi-centre case-control study. Int J
Epidemiol 18: 569-577. http://dx.doi.Org/10.1093/ije/18.3.569.
Goodman. DG: Ward. JM: Squire. RA: Chu. KC: Linhait MS. (1979). Neoplastic and
nonneoplastic lesions in aging F344 rats. Toxicol Appl Pharmacol 48: 237-248.
http://dx.doi.org/10.1016/0041-008X(79)90029-2.
Gossett JM. (1987). Measurement of Henry's law constants for Cl and C2 chlorinated
hydrocarbons. Environ Sci Technol 21: 202-208. http://dx.doi.org/10.1021/es00156a012.
7-24
-------
Grand] ean, P; Landrigan, PJ. (2006). Developmental neurotoxicity of industrial chemicals.
Lancet 368: 2167-2178. http://dx.doi.org/10.1016/s0140-6736(06)69665-7.
Grawe, J: Niisse, M; Adler, ID. (1997). Quantitative and qualitative studies of micronucleus
induction in mouse erythrocytes using flow cytometry: I: Measurement of micronucleus
induction in peripheral blood polychromatic erythrocytes by chemicals with known and
suspected genotoxicity. Mutagenesis 12: 1-8. http://dx.doi.Org/10.1093/mutage/12.l.l.
Gray, TJ: Beamand, JA: Lake, BG: Foster, JR: Gangolli, SD. (1982). Peroxisome proliferation in
cultured rat hepatocytes produced by clofibrate and phthalate ester metabolites. Toxicol
Lett 10: 273-279. http://dx.doi.org/10.1016/0378-4274(82)90087-X.
Gray, TJ: Lake, BG: Beamand, JA: Foster, JR: Gangolli, SD. (1983). Peroxisomal effects of
phthalate esters in primary cultures of rat hepatocytes. Toxicology 28: 167-179.
http://dx.doi.org/10.1016/0300-483X(83)90115-4.
Green, T; Odum, J: Nash, JA: Foster, JR. (1990). Perchloroethylene-induced rat kidney tumors:
An investigation of the mechanisms involved and their relevance to humans. Toxicol
Appl Pharmacol 103: 77-89. http://dx.doi.org/10.1016/0041-008X(90)90264-U.
Green, T; Dow, J: Ellis, MK: Foster, JR: Odum, J. (1997). The role of glutathione conjugation in
the development of kidney tumours in rats exposed to trichloroethylene. Chem Biol
Interact 105: 99-117.
Greim, H: Bonse, G: Radwan, Z: Reichert, D: Henschler, D. (1975). Mutagenicity in vitro and
potential carcinogenicity of chlorinated ethylenes as a function of metabolic oxirane
formation. Biochem Pharmacol 24: 2013-2017. http://dx.doi.org/10.1016/0006-
2952(75)90396-2.
Grivennikov, SI; Greten, FR: Karin, M. (2010). Immunity, inflammation, and cancer. Cell 140:
883-899. http://dx.doi.0rg/10.1016/i.cell.2010.01.025.
Grosse, Y; Baan, R: Secretan-lauby, B: El Ghissassi, F; Bouvard, V: Benbrahim-Tallaa, L; Guha,
N; Islami, F; Galichet L; Straif K. (2011). Carcinogenicity of chemicals in industrial and
consumer products, food contaminants and flavourings, and water chlorination
byproducts. Lancet Oncol 12: 328-329. http://dx.doi.org/10.1016/S1470-2045a 1V70088-
2.
Gu. ZW: Sele. B: Chmara, D: Jalbeit P: Vincent M: Vincent F: Marka, C: Faure, J. (1981).
[Effects of trichlorethylene and its metabolites on the rate of sister chromatid exchange.
In vivo and in vitro study on the human lymphocytes]. Ann Genet 24: 105-106.
Gualandi, G. (1987). Use of alpha- and beta-tubulin mutants for the study of spontaneous and
induced chromosomal mis-distribution in Aspergillus nidulans. Mutat Res-Fundam Mol
Mech Mutagen 178: 33-41. http://dx.doi.org/10.1016/0027-5107(87)90083-2.
Guberan, E; Fernandez, J. (1974). Control of industrial exposure to tetrachloroethylene by
measuring alveolar concentrations: theoretical approach using a mathematical model.
Occup Environ Med 31: 159-167.
Gulyas, H; Hemmerling, L. (1990). Tetrachloroethene air pollution originating from coin-
operated dry cleaning establishments. Environ Res 53: 90-99.
Guo. D: Sarkar. J: Suino-Powell. K: Xu. Y: Matsumoto. K: Jia. Y: Yu. S: Khare. S: Haldar. K:
Rao. MS: Foreman. JE: Monga. SP: Peters. JM: Xu. HE: Reddv. JK. (2007). Induction of
nuclear translocation of constitutive androstane receptor by peroxisome proliferator-
activated receptor alpha synthetic ligands in mouse liver. J Biol Chem 282: 36766-36776.
http://dx.doi.org/10.1074/ibc.M707183200.
7-25
-------
Guyton, KZ: Liu, Y; Gorospe, M; Xu, 0: Holbrook, NJ. (1996). Activation of mitogen-activated
protein kinase by H2O2. Role in cell survival following oxidant injury. J Biol Chem 271:
4138-4142.
Guvton. KZ: Chiu. WA: Bateson. TF: Jinot J: Scott CS: Brown. RC: Caldwell. JC. (2009). A
reexamination of the PPAR-alpha activation mode of action as a basis for assessing
human cancer risks of environmental contaminants. Environ Health Perspect 117: 1664-
1672. http://dx.doi.org/10.1289/ehp.0900758.
Guzelian. PS: Henry. CJ: Farland. WH: Olin. SS. (1992a). Developmental toxicants risk
assessment. Pediatrics 89: 353.
Guzelian, PS: Henry, CJ: Olin, SS (Eds.). (1992b). Similarities and differences between children
and adults: Implications for risk assessment. New York: ILSI Press.
Hake, CL: Stewart, RD. (1977). Human exposure to tetrachloroethylene: Inhalation and skin
contact. Environ Health Perspect 21: 231-238.
Hakkola, J: Raunio, H: Purkunen, R: Pelkonen, O: Saarikoski, S; Cresteil, T; Pasanen, M.
(1996a). Detection of cytochrome P450 gene expression in human placenta in first
trimester of pregnancy. Biochem Pharmacol 52: 379-383. http://dx.doi.org/10.1016/0006-
2952(96)00216-X.
Hakkola, J: Pasanen, M: Hukkanen, J: Pelkonen, O: Maenpaa, J: Edwards, RJ: Boobis, AR:
Raunio, H. (1996b). Expression of xenobiotic-metabolizing cytochrome P450 forms in
human full-term placenta. Biochem Pharmacol 51: 403-411.
http://dx.doi.org/10.1016/0006-2952(95)02184-l.
Hakkola, J: Tanaka, E: Pelkonen, O. (1998a). Developmental expression of cytochrome P450
enzymes in human liver. Pharmacol Toxicol 82: 209-217.
http://dx.doi.0rg/10.llll/i.1600-0773.1998.tb01427.x.
Hakkola, J: Pelkonen, O: Pasanen, M: Raunio, H. (1998b). Xenobiotic-metabolizing cytochrome
P450 enzymes in the human feto-placental unit: Role in intrauterine toxicity. Crit Rev
Toxicol 28: 35-72. http://dx.doi.org/10.1080/10408449891344173.
Halogenated Solvents Industry Alliance. (2004). Comment entitled "evaluation of EPAs
neurotoxicity of tetrachloroethylene" submitted by Paul H. Dugard, Ph.D, Halogenated
Solvents Industry Alliance, Inc. for EPA docket ORD-2003-0014. Arlington, VA.
http://www.regulations.gov/#!documentDetail:D=EPA-HQ-ORD-2003-0014-0006.
Hamada, T; Tanaka, H. (1995). Transfer of methyl chloroform, trichloroethylene and
tetrachloroethylene to milk, tissues and expired air following intraruminal or oral
administration in lactating goats and milk-fed kids. Environ Pollut 87: 313-318.
http://dx.doi.org/10.1016/0269-7491(94)P4163-I.
Hanahan, D: Weinberg, RA. (2000). The hallmarks of cancer. Cell 100: 57-70.
http://dx.doi.org/10.1016/S0092-8674(00)81683-9.
Hanefeld, M: Kemmer, C: Leonhardt W: Kunze, KD: Jaross, W: Haller. H. (1980). Effects of p-
chlorophenoxyisobutyric acid (CPIB) on the human liver. Atherosclerosis 36: 159-172.
Hanefeld, M: Kemmer, C: Kadner, E. (1983). Relationship between morphological changes and
lipid-lowering action of p-chlorphenoxyisobutyric acid (CPIB) on hepatic mitochondria
and peroxisomes in man. Atherosclerosis 46: 239-246. http://dx.doi.org/10.1016/0021-
9150(83)90115-6.
Hanioka, N; Jinno, H: Toyo'oka, T: Nishimura, T: Ando, M. (1995a). Induction of rat liver drug-
metabolizing enzymes by tetrachloroethylene. Arch Environ Contam Toxicol 28: 273-
280. http://dx.doi.org/10.1007/BF00213102.
7-26
-------
Hanioka, N; Jinno, H; Takahashi, A; Nakano, K; Yoda, R: Nishimura, T; Ando, M. (1995b).
Interaction of tetrachloroethylene with rat hepatic microsomal P450-dependent
monooxygenases. Xenobiotica25: 151-165.
http://dx.doi.org/10.3109/00498259509061841.
Hardell, L; Eriksson, M; Lenner, P; Lundgren, E. (1981). Malignant lymphoma and exposure to
chemicals, especially organic solvents, chlorophenols and phenoxy acids: A case-control
study. Br J Cancer 43: 169-176.
Hardin. BD: Bond. GP: Sikov. MR: Andrew. FD: Bellies. RP: Niemeier. RW. (1981). Testing of
selected workplace chemicals for teratogenic potential. Scand J Work Environ Health 7:
66-75.
Harrington-Brock, K; Doerr, CL; Moore, MM. (1998). Mutagenicity of three disinfection by-
products: Di- and trichloroacetic acid and chloral hydrate in L5178Y(+/-) --3.7.2C mouse
lymphoma cells. Mutat Res Genet Toxicol Environ Mutagen 413: 265-276.
Hartmann, A; Speit G. (1995). Genotoxic effects of chemicals in the single cell gel (SCG) test
with human blood cells in relation to the induction of sister-chromatid exchanges (SCE).
Mutat Res Lett 346: 49-56. http://dx.doi.org/10.1016/0165-7992(95)90068-3.
Hasmall SC: James. NH: Macdonald. N: West D: Chevalier. S: Cosulich. SC: Roberts. RA.
(1999). Suppression of apoptosis and induction of DNA synthesis in vitro by the
phthalate plasticizers monoethylhexylphthalate (MEHP) and diisononylphthalate (DINP):
a comparison of rat and human hepatocytes in vitro. Arch Toxicol 73: 451-456.
http://dx.doi.org/10.1007/s002040050634.
Hassoun, EA; Dey, S. (2008). Dichloroacetate- and trichloroacetate-induced phagocytic
activation and production of oxidative stress in the hepatic tissues of mice after acute
exposure. J Biochem Mol Toxicol 22: 27-34. http://dx.doi.org/10.1002/ibt.20210.
Hattis, D: White, P: Marmorstein, L; Koch, P. (1990). Uncertainties in pharmacokinetic
modeling for perchloroethylene. I. Comparison of model structure, parameters, and
predictions for low-dose metabolism rates for models derived by different authors. Risk
Anal 10: 449-458. http://dx.doi.org/10.1111/i. 1539-6924.199Q.tbOQ528.x.
Haworth, S: Lawlor, T; Mortelmans, K: Speck, W; Zeiger, E. (1983). Salmonella mutagenicity
test results for 250 chemicals. Environ Mutagen 5: 3-142.
http://dx.doi.org/10.1002/em.2860050703.
Haydon, DA; Urban, BW. (1983). The action of alcohols and other non-ionic surface active
substances on the sodium current of the squid giant axon. J Physiol 341: 411-427.
Haves, JR: Condie, LW, Jr; Borzelleca, JF. (1986). The sub chronic toxicity of
tetrachloroethylene (perchloroethylene) administered in the drinking water of rats.
Fundam Appl Toxicol 7: 119-125.
Hegi, ME; Fox, TR: Belinsky, SA; Devereux, TR: Anderson, MW. (1993). Analysis of activated
protooncogenes in B6C3F1 mouse liver tumors induced by ciprofibrate, a potent
peroxisome proliferator. Carcinogenesis 14: 145-149.
http://dx.doi.0rg/10.1093/carcin/14.l.145.
Heikes, PL; Hopper, ML. (1986). Purge and trap method for determination of fumigants in
whole grains, milled grain products, and intermediate grain-based foods. J Assoc Offic
Anal Chem 69: 990-998.
Hennig, UG: Rudd, NL; Hoar, DI. (1988). Kinetochore immunofluorescence in micronuclei: A
rapid method for the in situ detection of aneuploidy and chromosome breakage in human
fibroblasts. Mutat Res 203: 405-414.
7-27
-------
Henschler, D; Bonse, G. (1977). Metabolic activation of chlorinated ethylenes: Dependence of
mutagenic effect on electrophilic reactivity of the metabolically formed epoxides. Arch
Toxicol 39: 7-12. http://dx.doi.org/10.1007/BF00343270.
Henschler, D. (1977). Metabolism and mutagenicity of halogenated olefms - A comparison of
structure and activity. Environ Health Perspect 21: 61-64.
Herbert, V: Gardner, A; Colman, N. (1980). Mutagenicity of dichloroacetate, an ingredient of
some formulations of pangamic acid (trade-named "vitamin B15"). Am J Clin Nutr 33:
1179-1182.
Herren-Freund, SL; Pereira, MA; Khoury, MD; Olson, G. (1987). The carcinogenicity of
trichloroethylene and its metabolites, trichloroacetic acid and dichloroacetic acid, in
mouse liver. Toxicol Appl Pharmacol 90: 183-189. http://dx.doi.org/10.1016/0041-
008X(87)90325-5.
Herrera, E; Amusquivar, E; Lopez-Soldado, I; Ortega, H. (2006). Maternal lipid metabolism and
placental lipid transfer. Horm Res 65: 59-64. http://dx.doi.org/10.1159/000091507.
Hill, AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58:
295-300.
Hinchman, CA; Ballatori, N. (1990). Glutathione-degrading capacities of liver and kidney in
different species. Biochem Pharmacol 40: 1131-1135. http://dx.doi.org/10.1016/0006-
2952(90)90503-0.
Hinnen, U; Schmid-Grendelmeier, P; Muller, E: Eisner, P. (1995). Exposure to solvents in
scleroderma: Disseminated circumscribed scleroderma (morphea) in a painter exposed to
perchloroethylene. Schweizerische Medizinische Wochenschrift 125: 2433-2437.
Hoivik. DJ: Quails. CW. Jr: Mirabile. RC: Cariello. NF: Kimbrough. CL: Colton. HM:
Anderson, SP: Santostefano, MJ: Morgan, RJO: Dahl, RR: Brown, AR; Zhao, Z; Mudd,
PN. Jr: Oliver. WB. Jr: Brown. HR: Miller. RT. (2004). Fibrates induce hepatic
peroxisome and mitochondrial proliferation without overt evidence of cellular
proliferation and oxidative stress in cynomolgus monkeys. Carcinogenesis 25: 1757-
1769. http://dx.doi.org/10.1093/carcin/bghl82.
Holoshitz, N: Alsheikh-Ali, AA; Karas, RH. (2008). Relative safety of gemfibrozil and
fenofibrate in the absence of concomitant cerivastatin use. Am J Cardiol 101: 95-97.
http://dx.doi.0rg/10.1016/j.amjcard.2007.07.057.
Hoi son, RR: Pearce, B. (1992). Principles and pitfalls in the analysis of prenatal treatment effects
in multiparous species. Neurotoxicol Teratol 14: 221-228.
http://dx.doi.org/10.1016/0892-0362(92)90020-B.
Honma, T; Hasegawa, H; Sato, M; Sudo, A. (1980a). Changes of free amino acid content in rat
brain after exposure to trichloroethylene and tetrachloroethylene. Ind Health 18: 1-7.
http://dx.doi.0rg/10.2486/indhealth.18.l.
Honma, T; Sudo, A; Miyagawa, M; Sato, M; Hasegawa, H. (1980b). Effects of exposure to
trichloroethylene and tetrachloroethylene on the contents of acetylcholine, dopamine,
norepinephrine and serotonin in rat brain. Ind Health 18: 171-178.
http://dx.doi.org/10.2486/indhealth.18.171.
Hoshida, Y; Toffanin, S; Lachenmayer, A; Villanueva, A; Minguez, B; Llovet JM. (2010).
Molecular classification and novel targets in hepatocellular carcinoma: recent
advancements. Semin Liver Dis 30: 35-51. http://dx.doi.org/10.1055/s-0030-1247131.
7-28
-------
Hours, M; Dananche, B; Fevotte, J: Bergeret A; Ayzac, L; Cardis, E; Etard, JF: Fallen, C; Roy,
P; Fabry, J. (1994). Bladder cancer and occupational exposures. Scand J Work Environ
Health 20: 322-330.
Houten, L; Sonnesso, G. (1980). Occupational exposure and cancer of the liver. Arch Environ
Health 35: 51-53.
Howroyd, P: Swanson, C: Dunn, C: Cattley, RC: Gorton, JC. (2004). Decreased longevity and
enhancement of age-dependent lesions in mice lacking the nuclear receptor peroxisome
proliferator-activated receptor alpha (PPARalpha). Toxicol Pathol 32: 591-599.
HSDB (Hazardous Substances Data Bank). (2001). Tetrachloroethylene. Bethesda, Maryland:
National Library of Medicine, National Toxicology Program, Hazardous Substances Data
Bank.
Hu, C: Bi, A; Pan, ZH. (2009). Differential expression of three T-type calcium channels in
retinal bipolar cells in rats. Vis Neurosci 26: 177-187.
http://dx.doi.org/10.1017/S0952523809090026.
Hu, Y; Hakkola, J: Oscarson, M; Ingelman-Sundberg, M. (1999). Structural and functional
characterization of the 5'-flanking region of the rat and human cytochrome P450 2E1
genes: Identification of a polymorphic repeat in the human gene. Biochem Biophys Res
Commun 263: 286-293. http://dx.doi.org/10.1006/bbrc.1999.1362.
Hudnell HK: Boves, WK: Otto. DA: House. DE: Creason, JP: Geller. AM: Darcev, DJ:
Broadwell, DK. (1996a). Battery of neurobehavioral tests recommended to ATSDR:
Solvent-induced deficits in microelectronic workers. Toxicol Ind Health 12: 235-243.
Hudnell, HK: Otto, DA; House, DE. (1996b). The influence of vision on computerized
neurobehavioral test scores: A proposal for improving test protocols. Neurotoxicol
Teratol 18: 391-400. http://dx.doi.org/10.1016/0892-0362(96)00040-2.
Hudnell HK: House. D: Schmid, J: Koltai. D: Stopford, W: Wilkins, J: Savitz, DA: Swinker. M:
Music, S. (2001). Human visual function in the North Carolina clinical study on possible
estuary-associated syndrome. J Toxicol Environ Health A 62: 575-594.
http://dx.doi.org/10.1080/15287390151079633.
Hudnell HK: Schreiber, JS. (2004). Residential tetrachloroethylene exposure: Response.
Environ Health Perspect 112: A864-A865.
Hurst, CH; Waxman, DJ. (2003). Activation of PPARalpha and PPARgamma by environmental
phthalate monoesters. Toxicol Sci 74: 297-308. http://dx.doi.org/10.1093/toxsci/kfgl45.
Huttunen, J: Heinonen, O: Manninen, V: Koskinen, P: Hakulinen, T; Teppo, L: Manttari, M:
Frick, M. (1994). The Helsinki Heart Study: An 8.5-year safety and mortality follow-up.
JInternMed235: 31-39. http://dx.doi.Org/10.llll/i.1365-2796.1994.tb01029.x.
IARC (International Agency for Research on Cancer). (1995). Tetrachloroethylene. In Dry
cleaning, some chlorinated solvents and other industrial chemicals (pp. 159-221). Lyon,
France. http://monographs.iarc.fr/ENG/Monographs/vol63/index.php.
IARC (International Agency for Research on Cancer). (2000). Di(2-ethylhexyl) phthalate. In
Some industrial chemicals (pp. 41-148). Lyon, France.
IARC (International Agency for Research on Cancer). (2004). Some drinking-water disinfectants
and contaminants, including arsenic. Lyon, France.
http://monographs.iarc.fr/ENG/Monographs/vol84/index.php.
Ikbal, M: Tastekin, A; Dogan, H; Pirim, I; Ors, R. (2004). The assessment of genotoxic effects in
lymphocyte cultures of infants treated with chloral hydrate. Mutat Res 564: 159-164.
http://dx.doi.0rg/10.1016/i.mrgentox.2004.08.007.
7-29
-------
Ikeda, M; Ohtsuji, H. (1972). A comparative study of the excretion of Fujiwara reaction-positive
substances in urine of humans and rodents given trichloro- or tetrachloro-derivatives of
ethane and ethylene. Occup Environ Med 29: 99-104.
Ikeda, M; Ohtsuji, H; Imamura, T; Komoike, Y. (1972). Urinary excretion of total trichloro-
compounds, trichloroethanol, and trichloroacetic acid as a measure of exposure to
trichloroethylene and tetrachloroethylene. Br J Ind Med 29: 328-333.
http://dx.doi.0rg/10.1136/oem.29.3.328.
Ikeda, M: Imamura, T. (1973). Biological half-life of trichloroethylene and tetrachloroethylene
in human subjects. Int Arch Occup Environ Health 31: 209-224.
http://dx.doi.org/10.1007/BF00539241.
Ikeda, M: Koizumi, A; Watanabe, T: Endo, A; Sato, K. (1980). Cytogenetic and cytokinetic
investigations on lymphocytes from workers occupationally exposed to
tetrachloroethylene. Toxicol Lett 5: 251-256. http://dx.doi.org/10.1016/0378-
4274(80)90068-5.
Imbriani, M: Ghittori, S: Pezzagno, G: Capodaglio, E. (1988). Urinary excretion of
tetrachloroethylene (perchloroethylene) in experimental and occupational exposure. Arch
Environ Occup Health 43: 292-298.
Infante-Rivard, C: Siemiatycki, J: Lakhani, R; Nadon, L. (2005). Maternal exposure to
occupational solvents and childhood leukemia. Environ Health Perspect 113: 787-792.
http://dx.doi.org/10.1289/ehp.7707.
IOM (Institute of Medicine). (2002). Gulf War and health. Vol. 2: Insecticides and solvents.
Washington, DC: National Academies Press.
Iregren, A; Andersson, M: Nylen, P. (2002). Color vision and occupational chemical exposures,
II: Visual functions in non-exposed subjects. Neurotoxicology 23: 735-745.
Ishmael, J: Lock, EA. (1986). Nephrotoxicity of hexachlorobutadiene and its glutathione-derived
conjugates. Toxicol Pathol 14: 258-262.
Ishmael, J: Dugard, PH. (2006). A review of perchloroethylene and rat mononuclear cell
leukemia [Review]. Regul Toxicol Pharmacol 45: 178-184.
http://dx.doi.0rg/10.1016/j.vrtph.2006.02.009.
Issemann, I; Green, S. (1990). Activation of a member of the steroid hormone receptor
superfamily by peroxisome proliferators. Nature 347: 645-650.
Ito, Y: Yamanoshita, O: Asaeda, N; Tagawa, Y; Lee, CH; Aoyama, T: Ichihara, G: Furuhashi, K;
Kamijima, M: Gonzalez, FJ: Nakajima, T. (2007a). Di(2-ethylhexyl)phthalate induces
hepatic tumorigenesis through a peroxisome proliferator-activated receptor alpha-
independent pathway. J Occup Health 49: 172-182.
Ito, Y: Yamanoshita, O: Kurata, Y: Kamijima, M: Aoyama, T: Nakajima, T. (2007b). Induction
of peroxisome proliferator-activated receptor alpha (PPARalpha)-related enzymes by
di(2-ethylhexyl) phthalate (DEHP) treatment in mice and rats, but not marmosets. Arch
Toxicol 81: 219-226. http://dx.doi.org/10.1007/s00204-006-0141-x.
Jakobson, I; Wahlberg, JE; Holmberg, B; Johansson, G. (1982). Uptake via the blood and
elimination of 10 organic solvents following epicutaneous exposure of anesthetized
guinea pigs. Toxicol Appl Pharmacol 63: 181-187. http://dx.doi.org/10.1016/0041-
008X(82)90039-4.
Jakubowski, M: Wieczorek, H. (1988). The effects of physical effort on pulmonary uptake of
selected organic compound vapours. Pol J Occup Med 1: 62-71.
7-30
-------
Jamshidi. Y: Montgomery. HE: Hense. HW: Myerson. SG: Torra. IP: Staels. B: World. MJ:
Doering, A; Erdmann, J: Hengstenberg, C: Humphries, SE; Schunkert, H; Flavell, DM.
(2002). Peroxisome proliferator—activated receptor alpha gene regulates left ventricular
growth in response to exercise and hypertension. Circulation 105: 950-955.
Jang, JY; Droz, PO. (1997). Ethnic differences in biological monitoring of several organic
solvents: II. A simulation study with a physiologically based pharmacokinetic model. Int
Arch Occup Environ Health 70: 41-50. http://dx.doi.org/10.1007/s004200050184.
Janulewicz, PA; White, RF: Winter, M; Weinberg, J: Gallagher, L; Vieira, V: Webster, T;
Aschengrau, A. (2007). Prenatal exposure to tetrachloroethylene(PCE)-contaminated
drinking water and the risk of learning and attention disabilities. Neurotoxicol Teratol 29:
406. http://dx.doi.0rg/10.1016/i.ntt.2007.03.037.
Janulewicz. PA: White. RF: Winter. MR: Weinberg. JM: Gallagher. LE: Vieira. V: Webster. TF:
Aschengrau, A. (2008). Risk of learning and behavioral disorders following prenatal and
early postnatal exposure to tetrachloroethylene (PCE)-contaminated drinking water.
Neurotoxicol Teratol 30: 175-185. http://dx.doi.Org/10.1016/j.ntt.2008.01.007.
Jensen, OM; Wahrendorf, J: Knudsen, JB: Sorensen, BL. (1987). The Copenhagen case-referent
study on bladder cancer: Risks among drivers, painters, and certain other occupations.
Scand J Work Environ Health 13: 129-134.
Ji, J: Granstrom, C: Hemminki, K. (2005a). Occupation and bladder cancer: A cohort study in
Sweden. Br J Cancer 92: 1276-1278. http://dx.doi.org/10.1038/si.bic.6602473.
Ji, J: Hemminki, K. (2005a). Occupation and upper aerodigestive tract cancers: A follow-up
study in Sweden. J Occup Environ Med 47: 785-795.
http://dx.doi.org/10.1097/01.iom.0000165798.28569.b5.
Ji, J: Granstrom, C: Hemminki, K. (2005b). Occupational risk factors for kidney cancer: A
cohort study in Sweden. World Journal of Urology 23: 271-278.
http://dx.doi.org/10.1007/s00345-005-0007-5.
Ji, J: Hemminki, K. (2005b). Occurrences of leukemia subtypes by socioeconomic and
occupational groups in Sweden. J Occup Environ Med 47: 1131-1140.
http://dx.doi.org/10.1097/01.jom.0000174302.63621.e8.
Ji, J: Hemminki, K. (2005c). Variation in the risk for liver and gallbladder cancers in
socioeconomic and occupational groups in Sweden with etiological implications. Int
Arch Occup Environ Health 78: 641-649. http://dx.doi.org/10.1007/s00420-005-0015-l.
Ji, J: Hemminki, K. (2006a). Socioeconomic and occupational risk factors for pancreatic cancer:
A cohort study in Sweden. J Occup Environ Med 48: 283-288.
http://dx.doi.org/10.1097/01.jom.0000199511.14305.dc.
Ji, J: Hemminki, K. (2006b). Socioeconomic/occupational risk factors for lymphoproliferative
diseases in Sweden. Ann Epidemiol 16: 370-376.
http://dx.doi.0rg/10.1016/i.annepidem.2005.09.002.
JISA (Japan Industrial Safety Association). (1993). Carcinogenicity study of tetrachloroethylene
by inhalation in rats and mice. Hadano, Japan.
Johansen, K: Tinnerberg, H: Lynge, E. (2005). Use of history science methods in exposure
assessment for occupational health studies. Occup Environ Med 62: 434-441.
http://dx.doi.org/10.1136/oem.2004.016493.
Johnson, PD; Dawson, BV: Goldberg, SJ. (1998). Cardiac teratogenicity of trichloroethylene
metabolites. J Am Coll Cardiol 32: 540-545. http://dx.doi.org/10.1016/S0735-
1097(98)00232-0.
7-31
-------
Jonker, D: Woutersen, RA: Feron, VJ. (1996). Toxicity of mixtures of nephrotoxicants with
similar or dissimilar mode of action. Food Chem Toxicol 34: 1075-1082.
http://dx.doi.org/10.1016/S0278-6915(97)00077-X.
Juntunen, J. (1986). Occupational toxicology of trichloroethylene with special reference to
neurotoxicity. In PL Chambers; P Gehring; F Sakai (Eds.), New concepts and
developments in toxicology (pp. 189-200). Amsterdam, Netherlands: Elsevier Science
Publishers.
Kacew, S; Lambert, GH (Eds.). (1997). Environmental toxicology and pharmacology of human
development. Washington, DC: Taylor and Francis.
Kaemmerer, K; Fink, J: Keitzmann, M. (1982). Studies on the pharmacodynamics of
perchloroethylene. Prakt Tierarzt 63: 171-182.
Kafer, E. (1986). Tests which distinguish induced crossing-over and aneuploidy from secondary
segregation in Aspergillus treated with chloral hydrate or gamma-rays. Mutat Res 164:
145-166. http://dx.doi.org/10.1016/0165-l 161(86)90006-3.
Kappas, A. (1989). On the mechanisms of induced aneuploidy in Aspergillus nidulans and
validation of tests for genomic mutations. In MA Resnick; BK Vig (Eds.), Mechanisms
of Chromosome Distribution and Aneuploidy (pp. 377-384). New York, NY: Wiley.
Kargalioglu, Y; McMillan, BJ: Minear, RA: Plewa, MJ. (2002). Analysis of the cytotoxicity and
mutagenicity of drinking water disinfection by-products in Salmonella typhimurium.
Teratog Carcinog Mutagen 22: 113-128. http://dx.doi.org/10.1002/tcm.10010.
Karlsson, JE; Rosengren, LE; Kjellstrand, P: Haglid, KG. (1987). Effects of low-dose inhalation
of three chlorinated aliphatic organic solvents on deoxyribonucleic acid in gerbil brain.
Scand J Work Environ Health 13: 453-458.
Kato, I; Koenig, KL; Watanabe-Meserve, H; Baptiste, MS: Lillquist PP: Frizzera, G: Burke, JS:
Moseson, M; Shore, RE. (2005). Personal and occupational exposure to organic solvents
and risk of non-Hodgkin's lymphoma (NHL) in women (United States). Cancer Causes
Control 16: 1215-1224.
Kaufman. PS: Shipley. WU: Feldman. AS. (2009). Bladder cancer. Lancet 374: 239-249.
http://dx.doi.org/10.1016/S0140-6736(09)60491-8.
Kawata, K; Shimazaki, R; Okabe, S. (2009). Comparison of gene expression profiles in HepG2
cells exposed to arsenic, cadmium, nickel, and three model carcinogens for investigating
the mechanisms of metal carcinogenesis. Environ Mol Mutagen 50: 46-59.
http://dx.doi.org/10.1002/em.20438.
Keech. A: Simes. RJ: Barter. P: Best J: Scott R: Taskinen. MR: Forder. P: Filial. A: Davis. T:
Glasziou, P: Drury, P: Kesaniemi, YA; Sullivan, D; Hunt, D: Colman, P: d'Emden, M:
Whiting, M: Ehnholm, C; Laakso, M. (2005). Effects of long-term fenofibrate therapy on
cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study):
Randomised controlled trial. Lancet 366: 1849-1861. http://dx.doi.org/10.1016/S0140-
6736(05)67667-2.
Keller, DA; Heck, HP. (1988). Mechanistic studies on chloral toxicity: Relationship to
trichloroethylene carcinogenesis. Toxicol Lett 42: 183-191.
http://dx.doi.org/10.1016/0378-4274(88)90076-8.
Kessler. W: Numtip. W: Grote. K: Csanadv. GA: Chahoud. I: Filser. JG. (2004). Blood burden
of di(2-ethylhexyl) phthalate and its primary metabolite mono(2-ethylhexyl) phthalate in
pregnant and nonpregnant rats and marmosets. Toxicol Appl Pharmacol 195: 142-153.
http://dx.doi.0rg/10.1016/i.taap.2003.ll.014.
7-32
-------
Kim, KS: Min, JY; Dickman, MB. (2008). Oxalic acid is an elicitor of plant programmed cell
death during Sclerotinia sclerotiorum disease development. Mol Plant Microbe Interact
21: 605-612. http://dx.doi.org/10.1094/MPMI-21-5-0605.
Kim, MY; Oskarsson, T; Acharyya, S; Nguyen, DX; Zhang, XH; Norton, L; Massague, J.
(2009). Tumor self-seeding by circulating cancer cells. Cell 139: 1315-1326.
http://dx.doi.0rg/10.1016/i.cell.2009.ll.025.
Kjellstrand, P; Holmquist B; Kanje, M; Aim, P; Romare, S: Jonsson, I; Mansson, L; Bjerkemo,
M. (1984). Perchloroethylene: Effects on body and organ weights and plasma
butyrylcholinesterase activity in mice. Acta Pharmacol Toxicol 54: 414-424.
http://dx.doi.0rg/10.llll/i.1600-0773.1984.tb01951.x.
Kjellstrand, P: Holmquist B; Jonsson, I; Romare, S: Mansson, L. (1985). Effects of organic
solvents on motor activity in mice. Toxicology 35: 35-46. http://dx.doi.org/10.1016/0300-
483X(85)90130-1.
Klaunig, JE: Babich, MA: Baetcke, KP: Cook. JC: Gorton. JC: David. RM: DeLuca, JG: Lai.
DY: McKee, RH: Peters. JM: Roberts. RA: Fenner-Crisp, PA. (2003). PPARalpha
agonist-induced rodent tumors: Modes of action and human relevance [Review]. Crit Rev
Toxicol 33: 655-780. http://dx.doi.org/10.1080/713608372.
Kline, SA: Solomon, JJ: Van Duuren, BL. (1978). Synthesis and reactions of chloroalkene
epoxides. J Org Chem 43: 3596-3600.
Kline. SA: McCoy. EC: Rosenkranz, HS: Van Duuren. BL. (1982). Mutagenicity of
chloroalkene epoxides in bacterial systems. MutatRes 101: 115-125.
http://dx.doi.org/10.1016/0165-1218(82)90002-7.
Koch, R; Schlegelmilch, R; Wolf HU. (1988). Genetic effects of chlorinated ethylenes in the
yeast Saccharomyces cerevisiae. Mutat Res 206: 209-216.
http://dx.doi.org/10.1016/0165-1218(88)90162-0.
Kogevinas, M: 'tMannetje, A; Cordier, S: Ranft, U; Gonzalez, C; Vineis, P: Chang-Claude, J:
Lynge, E; Wahrendorf J: Tzonou, A; Jockel, K; Serra, C; Porru, S: Hours, M: Greiser, E;
Boffetta, P. (2003). Occupation and bladder cancer among men in Western Europe.
Cancer Causes Control 14: 907-914.
http://dx.doi.Org/10.1023/B:CACO.0000007962.19066.9c.
Kok, RJ: Haas, M: Moolenaar, F; de Zeeuw, D: Meijer, DK. (1998). Drug delivery to the
kidneys and the bladder with the low molecular weight protein lysozyme. Ren Fail 20:
211-217.
Koppel, C: Arndt L Arendt U: Koeppe, P. (1985). Acute tetrachloroethylene poisoning - Blood
elimination kinetics during hyperventilation therapy. Clin Toxicol 23: 103-115.
http://dx.doi.org/10.3109/15563658508990621.
Korn, J. (1977). [How many more? Perchlorate poisoning after dry-cleaning]. Ugeskr Laeger
139: 303-304.
Krasowski, M: Harrison, N. (2000). The actions of ether, alcohol and alkane general anaesthetics
on GABAA and glycine receptors and the effects of TM2 and TM3 mutations. Br J
Pharmacol 129: 731-743. http://dx.doi.org/10.1038/sj.bjp.0703087.
Kringstad, KP: Ljungquist PO: de Sousa, F; Stromberg, LM. (1981). Identification and
mutagenic properties of some chlorinated aliphatic compounds in the spent liquor from
kraft pulp chlorination. Environ Sci Technol 15: 562-566.
http://dx.doi.org/10.1021/es00087a006.
7-33
-------
Kung, JW: Currie, IS: Forbes, SJ: Ross, JA. (2010). Liver development, regeneration, and
carcinogenesis. J Biomed Biotechnol 2010: 984248.
http://dx.doi.org/10.1155/2010/984248.
Kurata. Y: Kidachi. F: Yokovama. M: Toyota. N: Tsuchitani. M: Katoh. M. (1998). Subchronic
toxicity of Di(2-ethylhexyl)phthalate in common marmosets: lack of hepatic peroxisome
proliferation, testicular atrophy, or pancreatic acinar cell hyperplasia. Toxicol Sci 42: 49-
56.
Kylin, B; Reichard, H; Sumegi, I; Yllner, S. (1963). Hepatotoxicity of inhaled trichloroethylene,
tetrachloroethylene and chloroform Single exposure. Basic Clin Pharmacol Toxicol 20:
16-26. http://dx.doi.0rg/10.llll/i.1600-0773.1963.tb01714.x.
Kylin, B; Sumegi, I; Yllner, S. (1965). Hepatotoxicity of inhaled trichloroethylene and
tetrachloroethylene Long-term exposure. Acta Pharmacol Toxicol 22: 379-385.
http://dx.doi.0rg/10.llll/i.1600-0773.1965.tb01833.x.
Kyrklund, T: Ailing. C: Kiellstrand, P: Haglid, KG. (1984). Chronic effects of perchloroethylene
on the composition of lipid and acyl groups in cerebral cortex and hippocampus of the
gerbil. Toxicol Lett 22: 343-349. http://dx.doi.org/10.1016/0378-4274(84)90112-7.
Kyrklund, T; Kiellstrand, P: Haglid, KG. (1987). Lipid composition and fatty acid pattern of the
gerbil brain after exposure to perchloroethylene. Arch Toxicol 60: 397-400.
http://dx.doi.org/10.1007/BF00295762.
Kyrklund, T: Kiellstrand, P: Haglid. KG. (1988). Effects of exposure to Freon 11, 1,1,1-
trichloroethane or perchloroethylene on the lipid and fatty-acid composition of rat
cerebral cortex. Scand J Work Environ Health 14: 91-94.
Kyrklund, T: Kjellstrand, P: Haglid, KG. (1990). Long-term exposure of rats to
perchloroethylene, with and without a post-exposure solvent-free recovery period: Effects
on brain lipids. Toxicol Lett 52: 279-285. http://dx.doi.org/10.1016/0378-4274(90)90037-
M
Kyrklund, T: Haglid, K. (1991). Brain lipid composition in guinea pigs after intrauterine
exposure to perchloroethylene. Pharmacol Toxicol 68: 146-148.
http://dx.doi.0rg/10.llll/i.1600-0773.1991.tb02054.x.
Kyyronen, P: Taskinen, H; Lindbohm, ML; Hemminki, K; Heinonen, OP. (1989). Spontaneous
abortions and congenital malformations among women exposed to tetrachloroethylene in
dry cleaning. J Epidemiol Community Health 43: 346-351.
http://dx.doi.0rg/10.1136/jech.43.4.346.
Lacev, JV. Jr: Garabrant DH: Laing, TJ: Gillespie, BW: Maves, MD: Cooper. BC: Schottenfeld,
L\ (1999). Petroleum distillate solvents as risk factors for undifferentiated connective
tissue disease (UCTD). Am J Epidemiol 149: 761-770.
Lagakos, SW: Wessen, BJ: Zelen, M. (1986). An analysis of contaminated well water and health
effects in Woburn, Massachusetts. J Am Stat Assoc 81: 583-596.
Lake, BG: Evans, JG: Cunninghame, ME; Price, RJ. (1993). Comparison of the hepatic effects of
nafenopin and WY-14,643 on peroxisome proliferation and cell replication in the rat and
Syrian hamster. Environ Health Perspect 101: 241-247.
Lane. SE: Watts. RA: Bentham. G: Innes. NJ: Scott DGI. (2003). Are environmental factors
important in primary systemic vasculitis? A case-control study. Arthritis Rheum 48: 814-
823. http://dx.doi.org/10.1002/art.10830.
Lanthony. P. (1978). The desaturated panel D-15. Doc Ophthalmol 46: 185-189.
http://dx.doi.org/10.1007/BF00174107.
7-34
-------
Lapsley, M; Akers, K; Norden, AG. (1998). Sensitive assays for urinary retinol-binding protein
and beta-2-glycoprotein-l based on commercially available standards. Ann Clin Biochem
35: 115-119.
Larsen, GL. (1985). Distribution of cysteine conjugate beta-lyase in gastrointestinal bacteria and
in the environment. Xenobiotica 15: 199-209.
http://dx.doi.org/10.3109/00498258509045350.
Larsen, GL: Stevens, JL. (1986). Cysteine conjugate beta-lyase in the gastrointestinal bacterium
Eubacterium limosum. Mol Pharmacol 29: 97-103.
Larsen, NL; Nielsen, B; Rayn-Nielsen, A. (1977). [Perchloroethylene intoxication. A hazard in
the use of coin laundries]. Ugeskr Laeger 139: 270-275.
Lash, LH; Nelson, RM; Van Dyke, RA; Anders, MW. (1990). Purification and characterization
of human kidney cytosolic cysteine conjugate beta-lyase activity. Drug Metab Dispos 18:
50-54.
Lash. LH: Qian. W: Putt DA: Desal K: Elfarra. AA: Sicurl AR: Parker. JC. (1998). Glutathione
conjugation of perchloroethylene in rats and mice in vitro: Sex-, species-, and tissue-
dependent differences. Toxicol Appl Pharmacol 150: 49-57.
http://dx.doi.org/10.1006/taap.1998.8402.
Lash. LH: Lipscomb. JC: Putt DA: Parker. JC. (1999). Glutathione conjugation of
trichloroethylene in human liver and kidney: Kinetics and individual variation. Drug
Metab Dispos 27: 351-359.
Lash, LH: Fisher, JW: Lipscomb, JC: Parker, JC. (2000). Metabolism of trichloroethylene.
Environ Health Perspect 108: 177-200.
Lash, LH: Parker, JC. (2001). Hepatic and renal toxicities associated with perchloroethylene.
Pharmacol Rev 53: 177-208.
Lash. LH: Qian. W: Putt DA: Huenl SE: Elfarra. AA: Krause. RJ: Parker. JC. (2001). Renal and
hepatic toxicity of trichloroethylene and its glutathione-derived metabolites in rats and
mice: Sex-, species-, and tissue-dependent differences. J Pharmacol Exp Ther 297: 155-
164.
Lash. LH: Qian. W: Putt DA: Huenl SE: Elfarra. AA: Sicurl AR: Parker. JC. (2002). Renal
toxicity of perchloroethylene and S-(l,2,2-trichlorovinyl)glutathione in rats and mice:
sex- and species-dependent differences. Toxicol Appl Pharmacol 179: 163-171.
http://dx.doi.org/10.1006/taap.2001.9358.
Lash. LH: Putt. DA: Huang. P: Huenl SE: Parker. JC. (2007). Modulation of hepatic and renal
metabolism and toxicity of trichloroethylene and perchloroethylene by alterations in
status of cytochrome P450 and glutathi one. Toxicology 235: 11-26.
http://dx.doi.0rg/10.1016/i.tox.2007.03.001.
Laslo-Baker, D; Barrera, M; Knittel-Keren, D; Kozer, E; Wolpin, J: Khattak, S: Hackman, R;
Rovet J: Koren, G. (2004). Child neurodevelopmental outcome and maternal
occupational exposure to solvents. Arch Pediatr Adolesc Med 158: 956-961.
http://dx.doi.org/10.1001/archpedi.158.10.956.
Laughter. AR: Dunn. CS: Swanson. CL: Howrovd. P: Cattlev. RC: Gorton. JC. (2004). Role of
the peroxisome proliferator-activated receptor alpha (PPARalpha) in responses to
trichloroethylene and metabolites, trichloroacetate and dichloroacetate in mouse liver.
Toxicology 203: 83-98. http://dx.doi.Org/10.1016/i.tox.2004.06.014.
7-35
-------
Lauwerys, R: Herbrand, J: Buchet JP: Bernard, A; Gaussin, J. (1983). Health surveillance of
workers exposed to tetrachloroethylene in dry-cleaning shops. Int Arch Occup Environ
Health 52: 69-77. http://dx.doi.org/10.1007/BF00380609.
Leavitt SA; DeAngelo, AB; George, MH; Ross, JA. (1997). Assessment of the mutagenicity of
dichloroacetic acid in lacl transgenic B6C3F1 mouse liver. Carcinogenesis 18: 2101-
2106. http://dx.doi.org/10.1093/carcin/18.ll.2101.
Lee. LJ. -H: Chung. CW: Ma. YC: Wang. GS: Chen. PC: Hwang. YH: Wang. JD. (2003).
Increased mortality odds ratio of male liver cancer in a community contaminated by
chlorinated hydrocarbons in groundwater. Occup Environ Med 60: 364-369.
http://dx.doi.0rg/10.1136/oem.60.5.364.
Lee, SS: Pineau, T; Drago, J: Lee, EJ: Owens, JW: Kroetz, PL; Fernandez-Salguero, PM;
Westphal H; Gonzalez, FJ. (1995). Targeted disruption of the alpha isoform of the
peroxisome proliferator-activated receptor gene in mice results in abolishment of the
pleiotropic effects of peroxisome proliferators. Mol Cell Biol 15: 3012-3022.
Lehmann, I; Rehwagen, M; Diez, U; Seiffart, A; Rolle-Kampczyk, U; Richter, M; Wetzig, H:
Borte, M; Herbarth, O. (2001). Enhanced in vivo IgE production and T cell polarization
toward the type 2 phenotype in association with indoor exposure to VOC: Results of the
LARS study. Int J Hyg Environ Health 204: 211-221. http://dx.doi.org/10.1078/1438-
4639-00100.
Lehmann, I; Thoelke, A; Rehwagen, M: Rolle-Kampczyk, U: Schlink, U: Schulz, R; Borte, M:
Diez, U: Herbarth, O. (2002). The influence of maternal exposure to volatile organic
compounds on the cytokine secretion profile of neonatal T cells. Environ Toxicol 17:
203-210. http://dx.doi.org/10.1002/tox.10055.
Leibman, KC: Ortiz, E. (1970). Epoxide intermediates in microsomal oxidation of olefms to
glycols. J Pharmacol Exp Ther 173: 242-246.
Leibman, KC: Ortiz, E. (1977). Metabolism of halogenated ethylenes. Environ Health Perspect
21: 91-97.
Leopardi, P; Zijno, A; Bassani, B; Pacchierotti, F. (1993). In vivo studies on chemically induced
aneuploidy in mouse somatic and germinal cells. Mutat Res 287: 119-130.
http://dx.doi.org/10.1016/0027-5107(93)90151-5.
Letkiewicz, F; Johnston, P; Macaluso, C; Elder, R: Yu, W. (1982). Occurrence in
tetrachloroethylene (perchloroethylene) in drinking water, food and air.
(JRB28130385229). McLean, VA: JRB Associates, Inc.
Leuschner, J: Leuschner, F. (1991). Evaluation of the mutagenicity of chloral hydrate in vitro
and in vivo. Arzneimittelforschung 41: 1101-1103.
Leuschner, J: Beuscher, N. (1998). Studies on the mutagenic and carcinogenic potential of
chloral hydrate. Arzneimittelforschung 48: 961-968.
Levine, B; Fierro, MF; Goza, SW: Valentour, JC. (1981). A tetrachloroethylene fatality. J
Forensic Sci 26: 206-209. http://dx.doi.org/10.1520/JFS11350J.
Li, X; Sundquist J: Sundquist K. (2008). Socioeconomic and occupational risk factors for
rheumatoid arthritis: A nationwide study based on hospitalizations in Sweden. J
Rheumatol35: 986-991.
Liang, JC: Brinkley, BR. (1985). Chemical probes and possible targets for the induction of
aneuploidy. Basic Life Sci 36: 491-505.
7-36
-------
Liang, JC: Pacchierotti, F. (1988). Cytogenetic investigation of chemically-induced aneuploidy
in mouse spermatocytes. Mutat Res-Fundam Mol Mech Mutagen 201: 325-335.
http://dx.doi.org/10.1016/0027-5107(88)90021-8.
Liangpunsakul S; Kolwankar, D; Pinto, A; Gorski, JC: Hall SD: Chalasani, N. (2005). Activity
of CYP2E1 and CYP3A enzymes in adults with moderate alcohol consumption: A
comparison with nonalcoholics. Hepatology 41: 1144-1150.
http://dx.doi.org/10.1002/hep.20673.
Lide. D. . R. (Ed.). (1990). CRC Handbook of Chemistry and Physics. Boca Raton, FL: CRC
Press.
Lieber, CS. (1997). Cytochrome P-4502E1: Its physiological and pathological role. Physiol Rev
77: 517-544.
Lillford. L: Beevers. C: Bowen. D: Kirkland. D. (2010). RE: DNA damage detected by the
alkaline comet assay in the liver of mice after oral administration of tetrachloroethylene.
(Mutagenesis, 25, 133-138, 2010). Mutagenesis 25: 427-428.
http://dx.doi.org/10.1093/mutage/geq020.
Lindbohm, ML; Hemminki, K. (1988). Nationwide data base on medically diagnosed
spontaneous abortions in Finland. Int J Epidemiol 17: 568-573.
http://dx.doi.0rg/10.1093/iie/17.3.568.
Lindbohm, ML; Taskinen, H; Sallmen, M; Hemminki, K. (1990). Spontaneous abortions among
women exposed to organic solvents. Am J Ind Med 17: 449-463.
http://dx.doi.org/10.1002/ajim.4700170404.
Lindbohm, ML; Hemminki, K: Bonhomme, MG: Anttila, A; Rantala, K: Heikkila, P; Rosenberg,
MJ. (1991). Effects of paternal occupational exposure on spontaneous abortions. Am J
Public Health 81: 1029-1033.
Lindbohm, ML. (1995). Effects of parental exposure to solvents on pregnancy outcome. J Occup
Environ Med 37: 908-914.
Lindbohm. ML: Sallmen. M: Kyyronen, P: Kauppinen, T: Pukkala, E. (2009). Risk of liver
cancer and exposure to organic solvents and gasoline vapors among Finnish workers. Int
J Cancer 124: 2954-2959. http://dx.doi.org/10.1002/iic.24309.
Lipscomb, JC: Fisher, JW: Confer, PD; Byczkowski, JZ. (1998). In vitro to in vivo extrapolation
for trichloroethylene metabolism in humans. Toxicol Appl Pharmacol 152: 376-387.
http://dx.doi.org/10.1006/taap.1998.8485.
Liu. Y: Guvton, KZ: Gorospe, M: Xu. Q: Lee. JC: Holbrook, NJ. (1996). Differential activation
of ERK, JNK/SAPK and P38/CSBP/RK map kinase family members during the cellular
response to arsenite. Free Radic Biol Med 21: 771-781.
Lock. EA: Sani, Y: Moore. RB: Finkelstein, MB: Anders. MW: Seawright AA. (1996). Bone
marrow and renal injury associated with haloalkene cysteine conjugates in calves. Arch
Toxicol 70: 607-619. http://dx.doi.org/10.1007/s002040050319.
Loizou, GD. (2001). The application of physiologically based pharmacokinetic modelling in the
analysis of occupational exposure to perchloroethylene. Toxicol Lett 124: 59-69.
http://dx.doi.org/10.1016/S0378-4274(00)00283-6.
Lomax, RB: Ridgway, P: Meldrum, M. (2004). Does occupational exposure to organic solvents
affect colour discrimination? Toxicol Rev 23: 91-121.
Lopreato, GF: Phelan, R: Borghese. CM: Beckstead. MJ: Mihic. SJ. (2003). Inhaled drugs of
abuse enhance serotonin-3 receptor function. Drug Alcohol Depend 70: 11-15.
http://dx.doi.org/10.1016/S0376-8716(02)00330-7.
7-37
-------
Lovell, D. (2010). Is tetrachloroethylene genotoxic or not. Mutagenesis 25: 443-446.
http://dx.doi.org/10.1093/mutage/geq036.
Lowengart, RA; Peters, JM; Cicioni, C: Buckley, J: Bernstein, L; Preston-Martin, S: Rappaport,
K (1987). Childhood leukemia and parents' occupational and home exposures. J Natl
Cancer Inst 79: 39-46.
Lukaszewski, T. (1979). Acute tetrachloroethylene fatality. Clin Toxicol 15: 411-415.
http://dx.doi.org/10.3109/15563657908989895.
Lundberg, I; Alfredsson, L; Plato, N; Sverdrup, B; Klareskog, L; Kleinau, S. (1994). Occupation,
occupational exposure to chemicals and rheumatological disease: A register based cohort
study. Scand J Rheumatol 23: 305-310. http://dx.doi.org/10.3109/03009749409099278.
Lundgren, B; Meijer, J: DePierre, JW. (1987). Induction of cytosolic and microsomal epoxide
hydrolases and proliferation of peroxisomes and mitochondria in mouse liver after dietary
exposure to p-chlorophenoxyacetic acid, 2,4-dichlorophenoxyacetic acid and 2,4,5-
trichlorophenoxyacetic acid. Biochem Pharmacol 36: 815-821.
http://dx.doi.org/10.1016/0006-2952(87)90169-9.
Lybarger, JA: Lichtveld, MY; Amler, RW. (1999). Biomedical testing of the kidney for persons
exposed to hazardous substances in the environment. Ren Fail 21: 263-274.
http://dx.doi.org/10.3109/08860229909085088.
Lynch, A; Parry, J. (1993). The cytochalasin-B micronucleus/kinetochore assay in vitro: studies
with 10 suspected aneugens. Mutat Res 287: 71-86. http://dx.doi.org/10.1016/0027-
5107(93)90146-7.
Lynge, E; Thygesen, L. (1990). Primary liver cancer among women in laundry and dry-cleaning
work in Denmark. Scand J Work Environ Health 16: 108-112.
Lynge, E; Carstensen, B: Andersen, O. (1995). Primary liver cancer and renal cell carcinoma in
laundry and dry-cleaning workers in Denmark. Scand J Work Environ Health 21: 293-
295.
Lynge, E: Andersen, A; Rylander, L: Tinnerberg, H; Lindbohm, ML; Pukkala, E: Romundstad,
P; Jensen, P; Clausen, LB: Johansen, K. (2006). Cancer in persons working in dry
cleaning in the Nordic countries. Environ Health Perspect 114: 213-219.
http://dx.doi.org/10.1289/ehp.8425.
Lynge, E. (2008). Comments on toxicological review of tetrachloroethylene CAS no. 127-18-4
EPA/635/R-08/011A external draft review. Available online at (accessed
Lynge, E: Tinnerberg, H: Rylander, L: Romundstad, P: Johansen, K: Lindbohm, ML; Heikkila,
P: Westberg. H: Clausen. LB: Piombino. A: Thorsted. BL. (2011). Exposure to
tetrachloroethylene in dry cleaning shops in the Nordic countries. Ann Occup Hyg 55:
387-396. http://dx.doi.org/10.1093/annhyg/meql01.
Ma, J: Lessner, L: Schreiber, J: Carpenter, DO. (2009). Association between residential
proximity to PERC dry cleaning establishments and kidney cancer in New York City. J
Environ Public Health 2009: 183920. http://dx.doi.org/10.1155/2009/183920.
MacArthur, A; Le, N: Fang, R; Band, P. (2009). Identification of occupational cancer risk in
British Columbia: A population-based case-control study of 2,998 lung cancers by
histopathological subtype. Am J Ind Med 52: 221-232.
http://dx.doi.org/10.1002/ajim.20663.
Mackav. JM: Fox. V: Griffiths. K: Fox. DA: Howard. CA: Coutts. C: Wvatt I: Styles. JA.
(1995). Trichloroacetic acid: Investigation into the mechanism of chromosomal damage
in the in virto human lymphocyte cytogenetic assay and the mouse bone marrow
7-38
-------
micronucleus test. Carcinogenesis 16: 1127-1133.
http://dx.doi.0rg/10.1093/carcin/16.5.1127.
Maddalena, RL; McKone, TE; Lay ton, DW: Hsieh, DP. (1995). Comparison of multi-media
transport and transformation models: regional fugacity model vs. CalTOX. Chemosphere
30: 869-889. http://dx.doi.org/10.1016/0045-6535(94)00447-3.
Mahle. DA: Gearhait JM: Grigsbv. CC: Mattie. PR: Barton. HA: Lipscomb. JC: Cook. RS.
(2007). Age-dependent partition coefficients for a mixture of volatile organic solvents in
Sprague-Dawley rats and humans. J Toxicol Environ Health A 70: 1745-1751.
http://dx.doi.org/10.1080/15287390701458991.
Mailhes, J: Aardema, M; Marchetti, F. (1993). Investigation of aneuploidy induction in mouse
oocytes following exposure to vinblastine-sulfate, pyrimethamine, diethylstilbestrol
diphosphate, or chloral hydrate. Environ Mol Mutagen 22: 107-114.
http://dx.doi.org/10.1002/em.2850220209.
Mailhes, JB: Preston, RJ: Yuan, ZP; Payne, HS. (1988). Analysis of mouse metaphase II oocytes
as an assay for chemically induced aneuploidy. Mutat Res 198: 145-152.
http://dx.doi.org/10.1016/0027-5107(88)90049-8.
Maitre, A; Hours, M; Bonneterre, V: Arnaud, J: Arslan, MT; Carpentier, P; Bergeret A; de
Gaudemaris, R. (2004). Systemic sclerosis and occupational risk factors: role of solvents
and cleaning products. J Rheumatol 31: 2395-2401.
Malherbe, P: Alberati-Giani, D; Kohler, C; Cesura, AM. (1995). Identification of a
mitochondrial form of kynurenine aminotransferase/glutamine transaminase K from rat
brain. FEES Lett 367: 141-144. http://dx.doi.org/10.1016/0014-5793(95)00546-L.
Mallin, K. (1990). Investigation of a bladder cancer cluster in northwestern Illinois. Am J
Epidemiol 132: S96-S106.
Malone. KE: Koepsell. TD: Paling. JR: Weiss. NS: Morris. PD: Taylor. JW: Swanson. GM:
Lyon, JL. (1989). Chronic lymphocytic leukemia in relation to chemical exposures. Am J
Epidemiol 130: 1152-1158.
Maloney, EK; Waxman, DJ. (1999). trans-Activation of PPARalpha and PPARgamma by
structurally diverse environmental chemicals. Toxicol Appl Pharmacol 161: 209-218.
http://dx.doi.org/10.1006/taap.1999.8809.
Mandel JS: McLaughlin, JK; Schlehofer, B; Mellemgaard, A; Helmert, U; Lindblad, P:
McCredie, M: Adami, HO. (1995). International renal-cell cancer study. IV. Occupation.
Int J Cancer 61:601-605. http://dx.doi.org/10.1002/iic.2910610503.
Mannervik, B. (1985). The isoenzymes of glutathione transferase. Adv Enzymol Relat Areas
Mol Biol 57: 357-417.
Manookin. MB: Weick. M: Stafford. BK: Demb. JB. (2010). NMD A receptor contributions to
visual contrast coding. Neuron 67: 280-293.
http://dx.doi.0rg/10.1016/i.neuron.2010.06.020.
Manzo, L; Artigas, F; Martinez, E; Mutti, A; Bergamaschi, E; Nicotera, P: Tonini, M: Candura,
SM; Ray, DE: Costa, LG. (1996). Biochemical markers of neurotoxicity A review of
mechanistic studies and applications [Review]. Hum Exp Toxicol 1]: S20-S35.
Marquardt JU: Thorgeirsson, SS. (2010). Stem cells in hepatocarcinogenesis: Evidence from
genomic data. Semin Liver Dis 30: 26-34. http://dx.doi.org/10.1055/s-0030-1247130.
Marrazzini, A; Betti, C: Bernacchi, F: Barrai, I; Barale, R. (1994). Micronucleus test and
metaphase analyses in mice exposed to known and suspected spindle poisons.
Mutagenesis 9: 505-515. http://dx.doi.Org/10.1093/mutage/9.6.505.
7-39
-------
Marsman, DS: Cattley, RC: Conway, JG: Popp, JA. (1988). Relationship of hepatic peroxisome
proliferation and replicative DNA synthesis to the hepatocarcinogenicity of the
peroxisome proliferators di(2-ethylhexyl)phthalate and [4-chloro-6-(2,3-xylidino)-2-
pyrimidinylthio]acetic acid (Wy-14,643) in rats. Cancer Res 48: 6739-6744.
Marsman, DS: Goldsworthy, TL: Popp, JA. (1992). Contrasting hepatocytic peroxisome
proliferation, lipofuscin accumulation and cell turnover for the hepatocarcinogens Wy-
14,643 and clofibric acid. Carcinogenesis 13: 1011-1017.
http://dx.doi.0rg/10.1093/carcin/13.6.1011.
Marth, E; Stuenzner, D: Binder, H: Moese, JR. (1985a). [Tetrachloroethylene: A study of the
effect of low concentrations of 1,1,2,2-tetrachloroethylene on the organism of the mouse.
II. Examinations of tetrachloroethylene-residues in various organs and establishment of
the examined organs]. Zentralbl Bakteriol Mikrobiol Hyg 181: 541-547.
Marth. E: Stunzner, D: Binder. H: Mose, JR. O985b). [Tetrachloroethylene: effect of low
concentrations of 1,1,2,2-tetrachloroethylene (perchloroethylene) on organisms in the
mouse. I. Laboratory chemical research]. Zentralbl Bakteriol Mikrobiol Hyg 181: 525-
540.
Marth, E. (1987). Metabolic changes following oral exposure to tetrachloroethylene in subtoxic
concentrations. Arch Toxicol 60: 293-299.
Marth, E: Stunzner, D: Kock, M: Mose, JR. (1989). Toxicokinetics of chlorinated hydrocarbons.
J Hyg Epidemiol Microbiol Immunol 33: 514-520.
Matsushima, T: Hayashi, M: Matsuoka, A; Ishidate, M: Miura, KF: Shimizu, H: Suzuki, Y;
Morimoto, K: Ogura, H: Mure. K: Koshi, K: Sofuni, T. (1999). Validation study of the in
vitro micronucleus test in a Chinese hamster lung cell line (CHL/IU). Mutagenesis 14:
569-580. http://dx.doi.0rg/10.1093/mutage/14.6.569.
Mattsson, J: Albee, RR: Yano, BL: Bradley, GJ: PJ, S. (1998). Neurotoxicologic examination of
rats exposed to 1,1,2,2-tetrachloroethylene (perchloroethylene) vapor for 13 weeks.
Neurotoxicol Teratol 20: 83-98.
May, R. (1976) Aufnahme Metabolisierung und Ausscheidung von Tetrachlorathylen und
Beeinflussung durch gleichzeitige Aufnahme von Athanol beim Menschen. (Doctoral
Dissertation). Universitat Wurzburg, Wurzburg, Germany.
Mazzullo, M: Grilli. S: Lattanzi, G: Prodi. G: Turina, MP: Colacci, A. (1987). Evidence of DNA
binding activity of perchloroethylene. Res Comm Chem Pathol Pharmacol 58: 215-235.
McCarver, DG: Byun, R: Hines. RN: Hichme. M: Wegenek. W. (1998). A genetic
polymorphism in the regulatory sequences of human CYP2E1: Association with
increased chlorzoxazone hydroxylation in the presence of obesity and ethanol intake.
Toxicol Appl Pharmacol 152: 276-281. http://dx.doi.org/10.1006/taap.1998.8532.
McCarver, DG: Hines, RN. (2002). The ontogeny of human drug-metabolizing enzymes: Phase
II: Conjugation enzymes and regulatory mechanisms. J Pharmacol Exp Ther 300: 361-
366. http://dx.doi.0rg/10.1124/ipet.300.2.361.
McConnell, G: Ferguson, DM; Pearson, CR. (1975). Chlorinated hydrocarbons and the
environment. Endeavour 34: 13-18.
McCredie, M: Stewart, JH. (1993). Risk factors for kidney cancer in New South Wales. IV.
Occupation. Br J Ind Med 50: 349-354.
McDermott MJ: Mazon KA: Shost SJ: Narang. RS: Aldous. KM: Storm. JE. (2005).
Tetrachloroethylene (PCE, Perc) levels in residential dry cleaner buildings in diverse
communities in New York City. Environ Health Perspect 113: 1336-1343.
7-40
-------
McDonald, AD; Armstrong, B; Cherry, NM: Delorme, C: Diodati-Nolin, A; McDonald, JC:
Robert, D. (1986). Spontaneous abortion and occupation. J Occup Med 28: 1232-1238.
McDonald, AD; McDonald, JC: Armstrong, B; Cherry, N; Delorme, C: Nolin, A; Robert, D.
(1987). Occupation and pregnancy outcome. Br J Ind Med 44: 521-526.
McDougal, JN: Jepson, GW: III, CH; Gargas, ML; Andersen, ME. (1990). Dermal absorption of
organic chemical vapors in rats and humans. Toxicol Sci 14: 299-308.
McKone, TE: Daniels, JI. (1991). Estimating human exposure through multiple pathways from
air, water, and soil. Regul Toxicol Pharmacol 13: 36-61.
McKone, TE: Bogen, KT. (1992). Uncertainties in health-risk assessment: an integrated case
study based on tetrachloroethylene in California groundwater. Regul Toxicol Pharmacol
15: 86-103.
McLean, D: Mannetje, A; Dryson, E; Walls, C: McKenzie, F; Maule, M; Cheng, S;
Cunningham, C: Kromhout H; Boffetta, P; Blair, A; Pearce, N. (2009). Leukaemia and
occupation: A New Zealand Cancer Registry-based case-control Study. Int J Epidemiol
38: 594-606. http://dx.doi.org/10.1093/iie/dvn220.
MDPH (Massachusetts Department of Public Health). (1997). Woburn childhood leukemia
follow-up study: Information booklet. Available online at
http://www.mass.gov/Eeohhs2/docs/dph/environmental/investigations/woburn childhood
_leukemia_follow.pdf (accessed
Meade, TW: clinics, FtBMRCGPRFapv. (2001). Design and intermediate results of the Lower
Extremity Arterial Disease Event Reduction (LEADER)* trial of bezafibrate in men with
lower extremity arterial disease [ISRCTN4119421]. Current Controlled Trials in
Cardiovascular Medicine 2: 195-204. http://dx.doi.org/10.1186/cvm-2-4-195.
Meckler, LC: Phelps, DK. (1966). Liver disease secondary to tetrachloroethylene exposure: A
case report. JAMA 197: 662-663.
Mehendale, HM. (2005). Tissue repair: An important determinant of final outcome of toxicant-
induced injury. Toxicol Pathol 33: 41-51. http://dx.doi.org/10.1080/01926230590881808.
Mellemgaard, A; Engholm, G: McLaughlin, JK; Olsen, JH. (1994). Occupational risk factors for
renal-cell carcinoma in Denmark. Scand J Work Environ Health 20: 160-165.
Melnick, RL. (2001). Is peroxisome proliferation an obligatory precursor step in the
carcinogenicity of di(2-ethylhexyl)phthalate (DEHP)? Environ Health Perspect 109: 437-
442.
Mera, N: Ohmori, S: Itahashi, K: Kiuchi, M: Igarashi, T: Rikihisa, T: Kitada, M. (1994).
Immunochemical evidence for the occurrence of Mu class glutathione S-transferase in
human fetal livers. J Biochem 116: 315-320.
Mergler, D: Blain, L. (1987). Assessing color vision loss among solvent-exposed workers. Am J
Ind Med 12: 195-203. http://dx.doi.org/10.1002/aiim.4700120208.
Mergler, D. (1987). Worker participation in occupational health research: Theory and practice.
Int J Health Serv 17: 151-167.
Mergler, D: Belanger, S: De Grosbois, S: Vachon, N. (1988a). Chromal focus of acquired
chromatic discrimination loss and solvent exposure among printshop workers.
Toxicology 49: 341-348. http://dx.doi.org/10.1016/0300-483X(88)90017-0.
Mergler, D: Blain, L; Lemaire, J: Lalande, F. (1988b). Colour vision impairment and alcohol
consumption. Neurotoxicol Teratol 10: 255-260. http://dx.doi.org/10.1016/0892-
0362(88)90025-6.
7-41
-------
Mergler, D; Huel, G: Bowler, R; Frenette, B; Cone, J. (1991). Visual dysfunction among former
microelectronics assembly workers. Arch Environ Health 46: 326-334.
http://dx.doi.org/10.1080/00039896.1991.9934398.
Mersch-Sundermann, V: Schneider, U; Klopman, G: Rosenkranz, HS. (1994). SOS induction in
Escherichia coli and Salmonella mutagenicity: A comparison using 330 compounds.
Mutagenesis 9: 205-224. http://dx.doi.Org/10.1093/mutage/9.3.205.
Meskar, A; Plee-Gautier, E; Amet Y; Berthou, F; Lucas, D. (2001). Interactions alcool-
xenobiotiques. Role du cytochrome P450 2E1 [Alcohol-xenobiotic interactions. Role of
cytochrome P450 2E1]. Pathol Biol 49: 696-702. http://dx.doi.org/10.1016/S0369-
8114(01)00235-8.
Messing, K; Dumais, L; Courville, J: Seifert, AM; Boucher, M. (1994). Evaluation of exposure
data from men and women with the same job title. J Occup Med 36: 913-917.
Mester, B; Meters, A; Deeg, E; Eisner, G: Becker, N; Seidler, A. (2006). Occupation and
malignant lymphoma: A population based case control study in Germany. Occup Environ
Med 63: 17-26. http://dx.doi.org/10.1136/oem.2005.020453.
Miligi, L: Seniori, CA: Crosignani, P; Fontana, A; Masala, G: Nanni, O; Ramazzotti, V: Rodella,
S; Stagnaro, E: Tumino, R; Vigano, C; Vindigni, C; Vineis, P. (1999). Occupational,
environmental, and life-style factors associated with the risk of hematolymphopoietic
malignancies in women. Am J Ind Med 36: 60-69. http://dx.doi.org/10.1002/(SICI)1097-
0274(199907)36:1<60::AID-AJIM9>3.0.CO:2-Z.
Miligi, L: Costantini, AS: Benvenuti, A; Kriebel, D: Bolejack, V: Tumino, R: Ramazzotti, V:
Rodella, S: Stagnaro, E: Crosignani, P; Amadori, D: Mirabelli, D: Sommani, L: Belletti,
I: Troschel L: Romeo. L: Micell G: Tozzi. GA: Mendico. I: Vineis. P. (2006).
Occupational exposure to solvents and the risk of lymphomas. Epidemiology 17: 552-
561. http://dx.doi.org/10.1097/01.ede.0000231279.30988.4d.
Miller, BM; Adler, ID. (1992). Aneuploidy induction in mouse spermatocytes. Mutagenesis 7:
69-76. http://dx.doi.0rg/10.1093/mutage/7.l.69.
Miller, RE; Guengerich, FP. (1982). Oxidation of trichloroethylene by liver microsomal
cytochrome P-450: Evidence for chlorine migration in a transition state not involving
trichloroethylene oxide. Biochemistry 21: 1090-1097.
http://dx.doi.org/10.1021/bi00534a041.
Miller, RE; Guengerich, FP. (1983). Metabolism of trichloroethylene in isolated hepatocytes,
microsomes, and reconstituted enzyme systems containing cytochrome P-450. Cancer
Res 43:1145-1152.
Milman. HA: Story. PL: Riccio. ES: Sivak. A: Tu. AS: Williams. GM: long. C: Tyson. CA.
(1988). Rat liver foci and in vitro assays to detect initiating and promoting effects of
chlorinated ethanes and ethylenes. Ann N Y Acad Sci 534: 521-530.
Mitchell, AE: Morin, D: Lakritz, J: Jones, AD. (1997). Quantitative profiling of tissue- and
gender-related expression of glutathione S-transferase isoenzymes in the mouse. Biochem
1325:207-216.
Mitchell, AM; Bridges, JW: Elcombe, CR. (1984). Factors influencing peroxisome proliferation
in cultured rat hepatocytes. Arch Toxicol 55: 239-246.
Mitchell AM: Lhuguenot JC: Bridges. JW: Elcombe. CR. (1985). Identification of the
proximate peroxisome proliferator(s) derived from di(2-ethylhexyl) phthalate. Toxicol
Appl Pharmacol 80: 23-32.
7-42
-------
Moghaddam, AP: Abbas, R: Fisher, JW: Stavrou, S; Lipscomb, JC. (1996). Formation of
dichloroacetic acid by rat and mouse gut microflora, an in vitro study. Biochem Biophys
Res Commun 228: 639-645. http://dx.doi.org/10.1006/bbrc.1996.1709.
Moghaddam, AP: Abbas, R: Fisher, JW: Lipscomb, JC. (1997). The role of mouse intestinal
microflora in the metabolism of trichloroethylene, an in vivo study. Hum Exp Toxicol 16:
629-635. http://dx.doi.org/10.1177/096032719701601101.
Moloney, WC: King, VP. (1973). Reduction of leukemia incidence following splenectomy in the
rat. Cancer Res 33: 573-574.
Monster, A; Regouin-Peeters, W; Van Schijndel, A; Van der Tuin, J. (1983). Biological
monitoring of occupational exposure to tetrachloroethene. Scand J Work Environ Health
9:273-281.
Monster, AC. (1979). Difference in uptake, elimination, and metabolism in exposure to
trichloroethylene, 1,1,1-trichloroethane, and tetrachloroethylene. Int Arch Occup Environ
Health 42: 311-317. http://dx.doi.org/10.1007/BF00377785.
Monster, AC: Houtkooper, JM. (1979). Estimation of individual uptake of trichloroethylene,
1,1,1-trichloroethane and tetrachloroethylene from biological parameters. Int Arch Occup
Environ Health 42: 319-323. http://dx.doi.org/10.1007/BF00377786.
Monster, AC: Boersma, G: Steenweg, H. (1979). Kinetics of tetrachloroethylene in volunteers;
influence of exposure concentration and work load. Int Arch Occup Environ Health 42:
303-309.
Monster, AC: Smolders, JF. (1984). Tetrachloroethene in exhaled air of persons living near
pollution sources. Int Arch Occup Environ Health 53: 331-336.
http://dx.doi.org/10.1007/BF00380672.
Moore. LE: Boffetta. P: Karami, S: Brennan, P: Stewart. PS: Hung. R: Zaridze, D: Matveev, V:
Janout V: Kollarova, H; Bencko, V: Navratilova, M; Szeszenia-Dabrowska, N; Mates,
D: Gromiec, J: Holcatova, I: Merino. M: Chanock, S: Chow. WH: Rothman, N. (2010).
Occupational trichloroethylene exposure and renal carcinoma risk: Evidence of genetic
susceptibility by reductive metabolism gene variants. Cancer Research 70: 6527-6536.
Morimura, K: Cheung. C: Ward. JM: Reddv, JK: Gonzalez. FJ. (2006). Differential
susceptibility of mice humanized for peroxisome proliferator-activated receptor alpha to
Wy-14,643-induced liver turnorigenesis. Carcinogenesis 27: 1074-1080.
http://dx.doi.org/10.1093/carcin/bgi329.
Moriva, M: Ohta, T: Watanabe, K: Miyazawa, T: Kato. K: Shirasu, Y. (1983). Further
mutagenicity studies on pesticides in bacterial reversion assay systems. MutatRes 116:
185-216.
Morrow, LA; Ryan, CM; Hodgson, MJ: Robin, N. (1990). Alterations in cognitive and
psychological functioning after organic solvent exposure. J Occup Med 32: 444-450.
Morse. HC. Ill: Anver. MR: Fredrickson, TN: Haines, DC: Harris. AW: Harris. NL: Jaffe. ES:
Kogan, SC: MacLennan, 1C: Pattengale, PK: Ward. JM. (2002). Bethesda proposals for
classification of lymphoid neoplasms in mice. Neoplasia 100: 246-258.
Morton. LM: Hartge, P: Holford, TR: Holly. EA: Chiu, BC: Vineis, P: Stagnaro, E: Willett EV:
Franceschi, S: La Vecchia, C: Hughes, AM; Cozen, W: Davis, S: Severson, RK;
Bernstein, L; Mayne, ST; Dee, FR; Cerhan, JR; Zheng, T. (2005). Cigarette smoking and
risk of non-Hodgkin lymphoma: A pooled analysis from the International Lymphoma
Epidemiology Consortium (InterLymph). Cancer Epidemiol Biomarkers Prev 14: 925-
933. http://dx.doi.org/10.1158/1055-9965.EPI-04-0693.
7-43
-------
Morton, W; Marjanovic, D. (1984). Leukemia incidence by occupation in the Portland-
Vancouver metropolitan area. Am J Ind Med 6: 185-205.
Moryama, MT: Domiki, C: Miyazawa, K; Tanaka, T; Suzuki, K. (2005). Effects of oxalate
exposure on Madin-Darby canine kidney cells in culture: Renal prothrombin fragment-1
mRNA expression. Urol Res 33: 470-475. http://dx.doi.org/10.1007/s00240-005-0510-6.
Moser, VC: Cheek, BM; MacPhail, RC. (1995). A multidisciplinary approach to toxicological
screening: III. Neurobehavioral toxicity. J Toxicol Environ Health A 45: 173-210.
http://dx.doi.org/10.1080/15287399509531988.
Moslen, MT: Reynolds, ES: Szabo, S. (1977). Enhancement of the metabolism and
hepatotoxicity of trichloroethylene and perchloroethylene. Biochem Pharmacol 26: 369-
375. http://dx.doi.org/10.1016/0006-2952(77)90193-9.
Motohashi, Y; Miyazaki, Y; Takano, T. (1993). Assessment of behavioral effects of
tetrachloroethylene using a set of time-series analyses. Neurotoxicol Teratol 15: 3-10.
Mueser. KT: McGurk. SR. (2004). Schizophrenia. Lancet 363: 2063-2072.
http://dx.doi.org/10.1016/80140-6736(04)16458-1.
Mukherjee, R: Jow, L; Noonan, D; McDonnell, DP. (1994). Human and rat peroxisome
proliferator activated receptors (PPARs) demonstrate similar tissue distribution but
different responsiveness to PPAR activators. J Steroid Biochem Mol Biol 51: 157-166.
http://dx.doi.org/10.1016/0960-0760(94)90089-2.
Murakami, K: Horikawa, K. (1995). The induction of micronuclei in mice hepatocytes and
reticulocytes by tetrachloroethylene. Chemosphere 31: 3733-3739.
http://dx.doi.org/10.1016/0045-6535(95)00222-T.
Murphy, JP. (1990). Exposure to Wells G and H in Woburn, Massachusetts. Boston, MA:
Massachusetts Health Research Institute, Inc. and Massachusetts Department of Public
Health.
http://www.mass.gov/Eeohhs2/docs/dph/environmental/investigations/woburn exposure
wells.pdf.
Mutti, A; Alinovi, R; Bergamaschi, E; Biagini, C: Cavazzini, S: Franchini, I; Lauwerys, RR:
Bernard. AM: Roels. H: Gelpi. E: Rosello. J: Ramis. I: Price. RG: Taylor. SA: de Broe.
M: Nuvts. GD: Stolte. H: Pels. LM: Herb oil C. (1992). Nephropathies and exposure to
perchloroethylene in dry-cleaners. Lancet 330: 189-193. http://dx.doi.org/10.1016/0140-
6736(92)90463-0.
Muttray, A; Wolff, U; Jung, D: Konietzko, J. (1997). Blue-yellow deficiency in workers exposed
to low concentrations of organic solvents. Int Arch Occup Environ Health 70: 407-412.
Nagano, K: Nishizawa, T: Yamamoto, S: Matsushima, T. (1998). Inhalation carcinogenesis
studies of six halogenated hydrocarbons in rats and mice. In K Chiyotani; Y Hosoda; Y
Aizawa (Eds.), Advances in Prevention of Occupational Respiratory Diseases (pp. 741-
746). Amsterdam: Elsevier Science.
Nakal JS: Stathopulos. PB: Campbell GL: Chu. I: Li-Muller. A: Aucoin. R. (1999). Penetration
of chloroform, trichloroethylene, and tetrachloroethylene through human skin. J Toxicol
Environ Health A 58: 157-170. http://dx.doi.org/10.1080/009841099157368.
Nakasa. H: Mera. N: Ohmorl S: Itahashl K: Igarashl T: Ishii. I: Kitada. M. (1997). Nucleotide
sequence of pi class glutathione S-transferase in human fetal liver. Res Commun Mol
Pathol Pharmacol 97: 67-78.
Nakatsuka, H: Watanabe, T: Takeuchi, Y: Hisanaga, N: Shibata, E: Suzuki, H: Huang, MY;
Chen, Z; Ou, OS: Ikeda, M. (1992). Absence of blue-yellow color vision loss among
7-44
-------
workers exposed to toluene or tetrachloroethylene, mostly at levels below occupational
exposure limits. Int Arch Occup Environ Health 64: 113-117.
http://dx.doi.org/10.1007/BF00381478.
Narotsky, MG: Kavlock, RJ. (1995). A multidisciplinary approach to toxicological screening: II.
Developmental toxicity. J Toxicol Environ Health 45: 145-171.
http://dx.doi.org/10.1080/15287399509531987.
Natarajan, A; Duivenvoorden, W; Meijers, M; Zwanenburg, T. (1993). Induction of mitotic
aneuploidy using Chinese hamster primary embryonic cells. Test results of 10 chemicals.
Mutat Res 287: 47-56. http://dx.doi.org/10.1016/0027-5107(93)90144-5.
NCI (National Cancer Institute). (1977). Bioassay of tetrachloroethylene for possible
carcinogenicity. (NCI-CGTR-13; DHEW Publication No. (Nffl) 77-813). Bethesda, Md:
National Institutes of Health. http://ntp.niehs.nih.gov/ntp/htdocs/LT rpts/trO 13.pdf
NCI (National Cancer Institute). (2010). What you need to know about cervical cancer. (NIH
Publication No. 08-2407). Washington, D.C.: U.S. Department of Health and Human
Services, http://www.cancer.gov/cancertopics/wyntk/cervix.pdf
Neafsev. P: Ginsberg. G: Hattis. D: Johns. DO: Guvton. KZ: Sonawane. B. (2009). Genetic
polymorphism in CYP2E1: Population distribution of CYP2E1 activity. J Toxicol
Environ Health B Crit Rev 12: 362-388. http://dx.doi.org/10.1080/10937400903158359.
Nelson, A; Mendoza, T; Hoyle, G: Brody, A; Fermin, C: Morris, G. (200la). Enhancement of
fibrogenesis by the p53 tumor suppressor protein in asbestos-exposed rodents. Chest 120:
S33-S34. http://dx.doi.org/10.1378/chest.120.l_suppl.S33.
Nelson. BK: Taylor. BJ: Setzer. JV: Hornung. RW. (1979). Behavioral teratology of
perchloroethylene in rats. J Environ Pathol Toxicol Oncol 3: 233-250.
Nelson. GM: Swank. AE: Brooks. LR: Bailey. KC: George. SE. (200Ib). Metabolism,
microflora effects, and genotoxicity in haloacetic acid-treated cultures of rat cecal
microbiota. Toxicol Sci 60: 232-241. http://dx.doi.Org/10.1093/toxsci/60.2.232.
Nelson. HH: Wiencke. JK: Christianl DC: Cheng. TJ: Zuo. ZF: Schwartz. BS: K. LB-: Spitz.
MR; Wang, M; Xu, X; Kelsey, KT. (1995). Ethnic differences in the prevalence of the
homozygous deleted genotype of glutathione S-transferase theta. Carcinogenesis 16:
1243-1245.
Nelson, MA; Bull, RJ. (1988). Induction of strand breaks in DNA by trichloroethylene and
metabolites in rat and mouse liver in vivo. Toxicol Appl Pharmacol 94: 45-54.
http://dx.doi.org/10.1016/0041-008X(88)90335-3.
Nelson, MA; Lansing, AJ: Sanchez, IM; Bull, RJ: Springer, PL. (1989). Dichloroacetic acid and
trichloroacetic acid-induced DNA strand breaks are independent of peroxisome
proliferation. Toxicology 58: 239-248. http://dx.doi.org/10.1016/0300-483X(89)90139-
X
Nesslany, F; Marzin, D. (1999). A micromethod for the in vitro micronucleus assay. Mutagenesis
14: 403-410. http://dx.doi.Org/10.1093/mutage/14.4.403.
Nestmann, ER; Chu, I; Kowbel, DJ: Matula, TI. (1980). Short-lived mutagen in Salmonella
produced by reaction of trichloroacetic acid and dimethyl sulphoxide. Can J Genet Cytol
22: 35-40. http://dx.doi.org/10.1139/g80-006.
Ni, Y; Wong, T: Kadlubar, F; Fu, P. (1994). Hepatic metabolism of chloral hydrate to free
radical(s) and induction of lipid peroxidation. Biochem Biophys Res Commun 204: 937-
943. http://dx.doi.org/10.1006/bbrc.1994.2550.
7-45
-------
Ni. YC: Kadlubar. FF: Fu. PP. (1995). Formation of malondialdehyde-modified 2'-
deoxyguanosinyl adduct from metabolism of chloral hydrate by mouse liver microsomes.
Biochem Biophys Res Commun 216: 1110-1117.
http://dx.doi.org/10.1006/bbrc.1995.2735.
NIOSH (National Institute for Occupational Safety and Health). (1994). Fatality assessment and
control evaluation (FACE) for Colorado: 17-year-old worker at a plastic products
manufacturing plant died as a result of an overexposure to tetrachloroethylene (also
known as perchloroethylene). (94CO006A). Morgantown, West Virginia.
http://www.cdc.gov/niosh/face/stateface/co/94co006.html.
Nomiyama, K; Liu, SJ: Nomiyama, H. (1992). Critical levels of blood and urinary cadmium,
urinary beta 2-microglobulin and retinol-binding protein for monitoring cadmium health
effects. In GF Nordberg; RFM Herber; L Alessio (Eds.), Cadmium in the human
environment: Toxicity and carcinogenicity (pp. 325-340). Lyon, France: International
Agency for Research on Cancer.
NRC (National Research Council). (1983). Risk assessment in the federal government:
Managing the process. Washington, DC: National Academies Press.
http://www.nap.edu/openbook.php?record_id=366&page=Rl.
NRC (National Research Council). (1993). Pesticides in the diets of infants and children.
Washington, DC: National Academy Press.
http://www.nap.edu/openbook.php?isbn=0309048753.
NRC (National Research Council). (1994). Science and judgment in risk assessment.
Washington, DC: National Academy Press.
http://www.nap.edu/openbook.php?isbn=030904894X.
NRC (National Research Council). (1995). Biologic markers in urinary toxicology. Washington,
DC: The National Academies Press.
http://www.nap. edu/openbook.php?isbn=0309052289.
NRC (National Research Council). (2006). Assessing the human health risks of
trichloroethylene: Key scientific issues. Washington, DC: The National Academies Press.
http://nae.edu/nae/naepcms.nsf/weblinks/MKEZ-6SSHPD7OpenDocument.
NRC (National Research Council). (2008). Phthalates and cumulative risk assessment: The task
ahead. Washington, DC: National Academies Press.
http://www.nap.edu/catalog.php7record id=12528.
NRC (National Research Council). (2009). Contaminated water supplies at Camp Lejeune:
Assessing potential health effects. Washington, DC: National Academies Press.
http://www.nap.edu/catalog.php?record_id=12618.
NRC (National Research Council). (2010). Review of the Environmental Protection Agency's
draft IRIS assessment of tetrachloroethylene. Washington, DC: National Academies
Press.
NTP (National Toxicology Program). (1986). Toxicology and carcinogenesis studies of
tetrachloroethylene (perchloroethylene) (CAS no. 127-18-4) in F344/N rats and B6C3F1
mice (inhalation studies). (NTP TR 311). Research Triangle Park, NC: U.S. Department
of Health and Human Services, National Toxicology Program.
http://ntp.mehs.nih.gov/ntp/htdocs/LT_rpts/tr311 .pdf
NTP (National Toxicology Program). (1987). Toxicology and carcinogenesis studies of dimethyl
methylphosponate (CAS No. 756-79-6) in F344/N rats and B6C3F 1 mice. Research
7-46
-------
Triangle Park, NC: Public Health Service, U.S. Department of Health and Human
Services. http://ntp.niehs.nih.gov/ntp/htdocs/LT rpts7tr323.pdf.
NTP (National Toxicology Program). (1988). Toxicology and carcinogenesis studies of
trichloroethylene (CAS No. 79-01-6) in four strains of rats (ACT, August, Marshall,
Osborne-Mendel) (gavage studies). Research Triangle Park, NC: U.S. Department of
Health and Human Services, Public Health Service, National Institutes of Health.
http://ntp.ni ehs.nih.gov/ntp/htdocs/LT_rpts/tr273.pdf
NTP (National Toxicology Program). (1989). Toxicology and carcinogenesis studies of
hexachloroethane (CAS no. 67-72-1) in F344/N rats (gavage studies). (NTP TR 361).
Research Triangle Park, NC: U.S. Department of Health and Human Services, National
Toxicology Program, http://ntp.mehs.nih.gov/ntp/htdocs/LT_rpts/tr361.pdf
NTP (National Toxicology Program). (1990a). Carcinogenesis studies of trichloroethylene
(without epichlorohydrin) (CAS No. 79-01-6) in F344/N rats and B6C3F1 mice (gavage
studies). Research Triangle Park, NC: U.S. Department of Health and Human Services,
Public Health Service, National Institutes of Health.
http://ntp.niehs.nih.gov/ntp/htdocs/LT rpts/tr243.pdf
NTP (National Toxicology Program). (1990b). Toxicology and carcinogenesis studies of d-
limonene (CAS NO. 5989-27-5) in F344/N rats and B6C3F1 mice (gavage studies).
(PB90231416). Research Triangle Park, NC: U.S. Department of Health and Human
Services, Public Health Service, National Institutes of Health.
http://ntp.mehs. nih.gov/ntp/htdocs/LT_rpts/tr347.pdf
NTP (National Toxicology Program). (2005). Report on carcinogens, eleventh edition. Research
Triangle Park, NC: U.S. Department of Health and Human Services, National Toxicology
Program. http://ntp.niehs.nih.gov/index.cfm?objectid=32BA9724-FlF6-975E-
7FCE50709CB4C932.
NTP (National Toxicology Program). (2007). NTP report on the toxicology studies of
dichloroacetic acid (CAS No. 79-43-6) in genetically modified (FVB Tg.AC
hemizygous) mice (dermal and drinking water studies) and carcinogenicity studies of
dichloroacetic acid in genetically modified [B6.129-Trp53tmlBrd(N5) haploinsufficient]
mice (drinking water studies). (PB2008-109740; NTP GMM 11; NIH PUB 07-4428).
Research Triangle Park.
http://www.ntis. gov/search/product.aspx?ABBR=PB2008109740.
Nutley, E; Tcheong, A; Allen, J; Collins, B; Ma, M; Lowe, X; Bishop, J; Moore, D; Wyrobek, A.
(1996). Micronuclei induced in round spermatids of mice after stem-cell treatment with
chloral hydrate: evaluations with centromeric DNA probes and kinetochore antibodies.
Environ Mol Mutagen 28: 80-89. http://dx.doi.org/10.1002/(SICni098-
2280(1996)28:2<80::AID-EM3>3.0.CO:2-I.
NYS OAG (New York State Office of the Attorney General). (2004). Comment focusing on
"neurotoxicity of tetrachloroethylene (perchloroethylene) discussion paper" submitted by
Judith S. Schreiber, Ph.D., Environmental Protection Bureau for EPA docket ORD-2003-
0014 (March 1, 2004). Albany, NY: State of New York Office of the Attorney General.
http://www.regulations.gov/#!documentDetail:D=EPA-HQ-ORD-2003-0014-0003.
NYSDOH (New York State Department of Health). (1997). Tetrachloroethene ambient air
criteria document. Final report. Albany, NY.
NYSDOH (New York State Department of Health). (2000). Evaluation of residential exposure to
tetrachloroethene using biomarkers of dose and neurological tests. Albany, NY.
7-47
-------
NYSDOH (New York State Department of Health). (2004). Comment focusing on
"neurotoxicity of tetrachloroethylene (perchloroethylene) discussion paper" submitted by
Jan E. Storm and Kimberly A. Mazor, New York State Department of Health for EPA
docket ORD-2003-0014 (March 3, 2004). Troy, NY.
http://www.regulations.gov/#!documentDetail:D=EPA-HQ-ORD-2003-0014-0004.
NYSDOH (New York State Department of Health). (2005a). Improving human risk assessment
for tetrachloroethylene by using biomakers and neurobehavioral testing. (R827446).
Troy, NY: New York State Department of Health, Bureau of Toxic Substance
Assessment.
NYSDOH (New York State Department of Health). (2005b). Pumpkin patch day care center
follow-up evaluation (final report). Troy, NY: Center for Environmental Health, Bureau
of Toxic Substance Assessment.
NYSDOH (New York State Department of Health). (2005c). Pumpkin patch day care center
investigation (final report). Troy, NY: Center for Environmental Health, Bureau of Toxic
Substance Assessment.
NYSDOH (New York State Department of Health). (2010). Tetrachloroethylene (perc) exposure
and visual contrast sensitivity (VCS) test performance in adults and children residing in
buildings with or without a dry cleaner. Troy, NY.
O'Brien, M; Spear, B; Glauert, H. (2005). Role of oxidative stress in peroxisome proliferator-
mediated carcinogenesis. CritRev Toxicol 35: 61-88.
O'Flahertv. EJ: Scott W: Schreiner. C: Bellies. RP. (1992). A physiologically based kinetic
model of rat and mouse gestation: disposition of a weak acid. Toxicol Appl Pharmacol
112: 245-256. http://dx.doi.org/10.1016/0041-008X(92)90194-W.
Odum, J: Green, T; Foster, JR; Hext, PM. (1988). The role of trichloroacetic acid and
peroxisome proliferation in the differences in carcinogenicity of perchloroethylene in the
mouse and rat. Toxicol Appl Pharmacol 92: 103-112. http://dx.doi.org/10.1016/0041-
008X(88)90232-3.
Ogata, M: Sugiyama, K; Kuroda, YI. (1962). [Investigation of a dry-cleaning shop using
tetrachloroethylene, with special reference to Fujiwara's substance in the urine of the
employees]. Okayama Igakkai Zasshi 74: 247-253.
Ogata, M: Takatsuka, Y; Tomokuni, K. (1971). Excretion of organic chlorine compounds in the
urine of persons exposed to vapours of trichloroethylene and tetrachloroethylene. Occup
Environ Med 28: 386-391. http://dx.doi.Org/10.1136/oem.28.4.386.
Ohtsuki, T: Sato, K: Koizumi, A; Kumai, M: Ikeda, M. (1983). Limited capacity of humans to
metabolize tetrachloroethylene. Int Arch Occup Environ Health 51: 381-390.
http://dx.doi.org/10.1007/BF00378352.
Olsen, J: Hemminki, K: Ahlborg, G: Bjerkedal, T: Kyyronen, P; Taskinen, H; Lindbohm, ML;
Heinonen, OP; Brandt L; Kolstad, H; Halvorsen, BA: Egenaes, J. (1990). Low
birthweight, congenital malformations, and spontaneous abortions among dry-cleaning
workers in Scandinavia. Scand J Work Environ Health 16: 163-168.
Olshan. AF: Anderson. L: Roman. E: Fear. N: Wolff M: Whyatt R: Vu. V: Diwan. BA:
Potischman, N. (2000). Workshop to identify critical windows of exposure for children's
health: cancer work group summary. Environ Health Perspect 108: 595-597.
Ono, Y: Somiya, I; Kawamura, M. (1991). Genotoxicity of the by-products produced in
chlorination and ozonation processes. Mutat Res Environ Mutagen Relat Subj 252: 102.
http://dx.doi.org/10.1016/0165-116K9n90301-N.
7-48
-------
Opdam, JJ: Smolders, JF. (1986). Alveolar sampling and fast kinetics of tetrachloroethene in
man. I. Alveolar sampling. Br J Ind Med 43: 814-824.
http://dx.doi.org/10.1136/oem.43.12.814.
Opdam, JJ. (1989a). Intra and interindividual variability in the kinetics of a poorly and highly
metabolising solvent. Br J Ind Med 46: 831-845.
http://dx.doi.org/10.1136/oem.46.12.831.
Opdam, JJ. (1989b). Respiratory input in inhalation experiments. Occup Environ Med 46: 145-
156. http://dx.doi.0rg/10.1136/oem.46.3.145.
Oshiro. WM: Krantz. QT: Bushnell PJ. (2008). Characterization of the effects of inhaled
perchloroethylene on sustained attention in rats performing a visual signal detection task.
Neurotoxicol Teratol 30: 167-174. http://dx.doi.Org/10.1016/j.ntt.2008.01.002.
Otieno, MA; Baggs, RB; Haves, JD: Anders, MW. (1997). Immunolocalization of microsomal
glutathione S-transferase in rat tissues. Drug Metab Dispos 25: 12-20.
Overbv. LH: Gardlik. S: Philpot RM: Serabiit-Singh. CJ. (1994). Unique distribution profiles of
glutathione S-transferases in regions of kidney, ureter, and bladder of rabbit. Lab Invest
70: 468-478.
Pahler, A; Parker, J: Dekant, W. (1999a). Dose-dependent protein adduct formation in kidney,
liver, and blood of rats and in human blood after perchloroethene inhalation. Toxicol Sci
48: 5-13.
Pahler, A; Volkel, W: Dekant W. (1999b). Quantitation of N epsilon-(dichloroacetyl)-L-lysine
in proteins after perchloroethene exposure by gas chromatography-mass spectrometry
using chemical ionization and negative ion detection following immunoaffinity
chromatography. J Chromatogr A 847: 25-34. http://dx.doi.org/10.1016/S0021-
9673(99)00276-9.
Palmer, CN: Hsu, MH; Griffin, KJ: Raucy, JL; Johnson, EF. (1998). Peroxisome proliferator
activated receptor-alpha expression in human liver. Mol Pharmacol 53: 14-22.
Paramei, GV: Meyer-Baron, M; Seeber, A. (2004). Impairments of colour vision induced by
organic solvents: A meta-analysis study. Neurotoxicology 25: 803-816.
http://dx.doi.0rg/10.1016/j.neuro.2004.01.006.
Parent ME; Siemiatycki, J: Fritschi, L. (2000). Workplace exposures and oesophageal cancer.
Occup Environ Med 57: 325-334.
Park. EJ: Lee. JH: Yu, GY: He. G: All SR: Holzer, RG: Osterreicher, CH: Takahashi, H: Karin,
M. (2010). Dietary and genetic obesity promote liver inflammation and tumorigenesis by
enhancing IL-6 and TNF expression. Cell 140: 197-208.
http://dx.doi.0rg/10.1016/i.cell.2009.12.052.
Park. JH: Spengler, JD: Yoon, DW: Dumyahn, T: Lee. K: Ozkavnak, H. (1998). Measurement of
air exchange rate of stationary vehicles and estimation of in-vehicle exposure. J Expo Sci
Environ Epidemiol 8: 65-78.
Parkinson, A; Mudra, PR: Johnson, C: Dwyer, A; Carroll, KM. (2004). The effects of gender,
age, ethnicity, and liver cirrhosis on cytochrome P450 enzyme activity in human liver
microsomes and inducibility in cultured human hepatocytes. Toxicol Appl Pharmacol
199: 193-209. http://dx.doi.Org/10.1016/i.taap.2004.01.010.
Parnell, M: Exon, J: Roller, L. (1988). Assessment of hepatic initiation-promotion properties of
trichloroacetic acid. Arch Environ Contam Toxicol 17: 429-436.
http://dx.doi.org/10.1007/BF01055507.
7-49
-------
Parry, JM: Parry, EM; Warr, T; Lynch, A; James, S. (1990). The detection of aneugens using
yeasts and cultured mammalian cells. In Mutation and the environment, Part B:
Metabolism, testing methods, and chromosomes: Proceedings of the Fifth International
Conference on Environmental Mutagens, held in Cleveland, Ohio, July 10-15, 1989. New
York, NY: Wiley-Liss.
Parry, JM: Parry, EM; Bourner, R; Doherty, A; Ellard, S: O'Donovan, J: Hoebee, B; de
Stoppelaar, JM: Mohn, GR: Onfelt A; Renglin, A; Schultz, N; Soderpalm-Berndes, C:
Jensen, KG: Kirsch-Volders, M; Elhajouji, A; Van Hummelen, P; Degrassi, F; Antoccia,
A; Cimini, D; Izzo, M; Tanzarella, C: Adler, ID; Kliesch, U; Schriever-Schwemmer, G:
Gasser, P; Crebelli, R: Carere, A; Andreoli, C: Benigni, R; Leopardi, P: Marcon, F;
Zinjo, Z; Natarajan, AT; Boei, JJW, A; Kappas, A; Voutsinas, G: Zarani, FE; Patrinelli,
A; Pachierotti, F: Tiveron, C: Hess, P. (1996). The detection and evaluation of aneugenic
chemicals. MutatRes 353: 11-46. http://dx.doi.org/10.1016/0027-5107(95)00242-1.
Partanen, T; Heikkila, P: Hernberg, S: Kauppinen, T; Moneta, G: Ojajarvi, A. (1991). Renal cell
cancer and occupational exposure to chemical agents. Scand J Work Environ Health 17:
231-239.
Parzefall, W; Berger, W; Kainzbauer, E; Teufelhofer, O: Schulte-Hermann, R: Thurman, R.
(2001). Peroxisome proliferators do not increase DNA synthesis in purified rat
hepatocytes. Carcinogenesis 22: 519-523.
Paulu, C: Aschengrau, A; Ozonoff D. (1999). Tetrachloroethylene-contaminated drinking water
in Massachusetts and the risk of colon-rectum, lung, and other cancers. Environ Health
Perspect 107: 265-271.
Pearce, N: Checkoway, H; Kriebel, D. (2007). Bias in occupational epidemiology studies. Occup
Environ Med 64: 562-568. http://dx.doi.org/10.1136/oem.2006.026690.
Pegg. DG: Zempel. JA: Braun. WH: Watanabe. PG. (1979). Disposition of
tetrachloro(14C)ethylene following oral and inhalation exposure in rats. Toxicol Appl
PharmacolSl: 465-474.
Pelekis, M: Gephart, LA; Lerman, SE. (2001). Physiological-model-based derivation of the adult
and child pharmacokinetic intraspecies uncertainty factors for volatile organic
compounds. Regul Toxicol Pharmacol 33: 12-20.
http://dx.doi.org/10.1006/rtph.2000.1436.
Pellizzarl ED: Hartwell. TD: Harris. BS. Ill: Waddell. RD: Whitaker. DA: Erickson. MD.
(1982). Purgeable organic compounds in mother's milk. Bull Environ Contam Toxicol
28: 322-328.
Peplonska, B: Stewart, P: Szeszenia-Dabrowska, N: Rusiecki, J: Garcia-Closas, M: Lissowska, J:
Bardin-Mikolajczak, A; Zatonski, W: Gromiec, J: Brzeznicki, S: Brinton, L; Blair, A.
(2007). Occupation and breast cancer risk in Polish women: A population-based case-
control study. Am J Ind Med 50: 97-111. http://dx.doi.org/10.1002/aiim.20420.
Pereira, MA. (1996). Carcinogenic activity of dichloroacetic acid and trichloroacetic acid in the
liver of female B6C3F1 mice. Fundam Appl Toxicol 31: 192-199.
http://dx.doi.org/10.1006/faat.1996.0091.
Pereira, MA; Phelps, JB. (1996). Promotion by dichloroacetic acid and trichloroacetic acid of N-
methyl-N-nitrosourea-initiated cancer in the liver of female B6C3F1 mice. Cancer Lett
102: 133-141. http://dx.doi.org/10.1016/0304-3835(96)04156-0.
Pereira, MA; Li, K; Kramer, PM. (1997). Promotion by mixtures of dichloroacetic acid and
trichloroacetic acid of N-methyl-N-nitrosourea-initiated cancer in the liver of female
7-50
-------
B6C3F1 mice. Cancer Lett 115: 15-23. http://dx.doi.org/10.1016/50304-3835(97)04699-
5.
Pereira, MA; Kramer, PM; Conran, PB; Tao, L. (2001). Effect of chloroform on dichloroacetic
acid and trichloroacetic acid-induced hypomethylation and expression of the c-myc gene
and on their promotion of liver and kidney tumors in mice. Carcinogenesis 22: 1511-
1519. http://dx.doi.0rg/10.1093/carcin/22.9.1511.
Perocco, P: Bolognesi, S: Alberghini, W. (1983). Toxic activity of seventeen industrial solvents
and halogenated compounds on human lymphocytes cultured in vitro. Toxicol Lett 16:
69-75. http://dx.doi.org/10.1016/0378-4274(83)90012-7.
Perrin, MC: Opler, MG: Harlap, S: Harkavy-Friedman, J: Kleinhaus, K; Nation, D; Fennig, S:
Susser, ES: Malaspina, D. (2007). Tetrachloroethylene exposure and risk of
schizophrenia: offspring of dry cleaners in a population birth cohort, preliminary
findings. Schizophr Res 90: 251-254.
Perrot N; Nalpas, B; Yang, CS: Beaune, PH. (1989). Modulation of cytochrome P450 isozymes
in human liver, by ethanol and drug intake. Eur J Clin Invest 19: 549-555.
http://dx.doi.0rg/10.llll/i.1365-2362.1989.tb00273.x.
Pesch. B: Haerting. J: Ranft U: Klimpel A: Oelschlagel B: Schill W. (2000a). Occupational
risk factors for renal cell carcinoma: Agent-specific results from a case-control study in
Germany. Int J Epidemiol 29: 1014-1024. http://dx.doi.Org/10.1093/ije/29.6.1014.
Pesch. B: Haerting. J: Ranft U: Klimpel A: Oelschlagel B: Schill W. (2000b). Occupational
risk factors for urothelial carcinoma: Agent-specific results from a case-control study in
Germany. Int J Epidemiol 29: 238-247. http://dx.doi.Org/10.1093/iie/29.2.238.
Peters. JM: Cattlev. RC: Gonzalez. FJ. (1997). Role of PPAR alpha in the mechanism of action
of the nongenotoxic carcinogen and peroxisome proliferatorWy-14,643. Carcinogenesis
18: 2029-2033. http://dx.doi.org/10.1093/carcin/18.ll.2029.
Peters, JM: Cheung, C: Gonzalez, FJ. (2005). Peroxisome proliferator-activated receptor-alpha
and liver cancer: where do we stand. J Mol Med 83: 774-785.
http://dx.doi.org/10.1007/s00109-005-0678-9.
Pezzagno, G: Imbriani, M; Ghittori, S: Capodaglio, E. (1988). Urinary concentration,
environmental concentration, and respiratory uptake of some solvents: Effect of the work
load. Am Ind Hyg Assoc J 49: 546-552.
Philip, BK; Mumtaz, MM; Latendresse, JR: Mehendale, HM. (2007). Impact of repeated
exposure on toxicity of perchloroethylene in Swiss Webster mice. Toxicology 232: 1-14.
http://dx.doi.0rg/10.1016/i.tox.2006.12.018.
Plewa, MJ: Kargalioglu, Y; Vankerk, D: Minear, RA: Wagner, ED. (2002). Mammalian cell
cytotoxicity and genotoxicity analysis of drinking water disinfection by-products.
Environ Mol Mutagen 40: 134-142. http://dx.doi.org/10.1002/em.10092.
Poet TS: Weitz. KK: Gies. RA: Edwards. JA: Thrall KD: Corlev. RA: Tanoio. H: Hul X:
Maibach, HI; Wester, RC. (2002). PBPK modeling of the percutaneous absorption of
perchloroethylene from a soil matrix in rats and humans. Toxicol Sci 67: 17-31.
http://dx.doi.0rg/10.1093/toxsci/67.l.17.
Pogribnv. IP: Tryndvak. VP: Woods. CG: Witt SE: Rusyn. I. (2007). Epigenetic effects of the
continuous exposure to peroxisome proliferator WY-14,643 in mouse liver are dependent
upon peroxisome proliferator activated receptor alpha. Mutat Res-Fundam Mol Mech
Mutagen 625: 62-71. http://dx.doi.Org/10.1016/j.mrfmmm.2007.05.004.
7-51
-------
Pohl HR: Ronev. N: Wilbur. S: Hansen. H: De Rosa. CT. (2003). Six interaction profiles for
simple mixtures. Chemosphere 53: 183-197. http://dx.doi.org/10.1016/S0045-
6535(03)00436-3.
Pohlabeln, H; Jockel, KH: Bolm-Audorff, U. (1999). Non-occupational risk factors for cancer of
the lower urinary tract in Germany. Eur J Epidemiol 15: 411-419.
http://dx.doi.Org/10.1023/A:1007595809278.
Pohlabeln, H: Boffetta, P; Ahrens, W; Merletti, F; Agudo, A; Benhamou, E; Benhamou, S:
Bruske-Hohlfeld, I; Ferro, G: Fortes, C; Kreuzer, M; Mendes, A; Nyberg, F; Pershagen,
G: Saracci, R: Schmid, G: Siemiatycki, J: Simonato, L; Whitley, E; Wichmann, HE;
Winck, C: Zambon, P; Jockel, KH. (2000). Occupational risks for lung cancer among
nonsmokers. Epidemiology 11: 532-538.
Pollack. GM: Li. RC: Ermer, JC: Shen, DP. (1985). Effects of route of administration and
repetitive dosing on the disposition kinetics of di(2-ethylhexyl) phthalate and its mono-
de-esterified metabolite in rats. Toxicol Appl Pharmacol 79: 246-256.
http://dx.doi.org/10.1016/0041-008X(85)90346-1.
Popp, W: Muller, G: Baltes-Schmitz, B; Wehner, B; Vahrenholz, C: Schmieding, W:
Benninghoff, M; Norpoth, K. (1992). Concentrations of tetrachloroethene in blood and
trichloroacetic acid in urine in workers and neighbours of dry-cleaning shops. Int Arch
Occup Environ Health 63: 393-395.
Porru, S: Aulenti, V: Donato, F: Boffetta, P: Fazioli, R; Cosciani Cunico, S: Alessio, L. (1996).
Bladder cancer and occupation: A case-control study in northern Italy. Occup Environ
Med53: 6-10.
Potter. CL: Chang. LW: DeAngelo. AB: Daniel FB. (1996). Effects of four trihalomethanes on
DNA strand breaks, renal hyaline droplet formation and serum testosterone in male F-344
rats. Cancer Lett 106: 235-242.
Price, PJ: Hassett CM; Mansfield, JI. (1978). Transforming activities of trichloroethylene and
proposed industrial alternatives. In Vitro 14: 290-293.
Price, RG: Taylor, SA; Chivers, I; Arce-Tomas, M: Crutcher, E: Franchini, I; Slinovi, R:
Cavazzini, S: Bergamaschi, E: Mutti, A; Vettori, MV: Lauwerys, R: Bernard, A;
Kabanda, A; Roels, H: Thielemans, N: Hotz, P: De Broe, ME; Elseviers, MM; Nuyts,
GD: Gelpi, E: Hotter, G: Rosello, J: Ramis, I; Stolte, H: Fels, LM; Eisenberger, U.
(1996). Development and validation of new screening tests for nephrotoxic effects. Hum
Exp Toxicol 15: S10-S19.
Pukkala, E: Martinsen, J: Lynge, E: Gunnarsdottir, H: Sparen, P: Tryggvadottir, L: Weiderpass,
E: Kjaerheim, K. (2009). Occupation and cancer - follow-up of 15 million people in five
Nordic countries. Acta Oncol 48: 646-790.
http://dx.doi.org/10.1080/02841860902913546.
Pukkala, E: Malila, N: Hakama, M. (2010). Socioeconomic differences in incidence of cervical
cancer in Finland by cell type. Acta Oncol 49: 180-184.
http://dx.doi.org/10.3109/02841860903386390.
Qiu, J: Chien, YC: Bruckner, JV: Fisher, JW. (2010). Bayesian analysis of a physiologically
based pharmacokinetic model for perchloroethylene in humans. J Toxicol Environ Health
A 73: 74-91. http://dx.doi.org/10.1080/15287390903249099.
Quondamatteo, F: Schulz, TG: Bunzel, N: Hallier, E: Herken, R. (1998). Immunohistochemical
localization of glutathione S-transferase-Tl in murine kidney, liver, and lung. Histochem
CellBiol 110: 417-423. http://dx.doi.org/10.1007/s004180050302.
7-52
-------
Rachootin, P; Olsen, J. (1983). The risk of infertility and delayed conception associated with
exposures in the Danish workplace. J Occup Med 25: 394-402.
Radican, L; Blair, A; Stewart, P: Wartenberg, D. (2008). Mortality of aircraft maintenance
workers exposed to trichloroethylene and other hydrocarbons and chemicals: Extended
follow-up. J Occup Environ Med 50: 1306-1319.
http://dx.doi.org/10.1097/JOM.Ob013e3181845f7f
Raiimakers. MTM: Steegers. EAP: Peters. WHM. (2001). Glutathione S-transferases and thiol
concentrations in embryonic and early fetal tissues. Hum Reprod 16: 2445-2450.
http://dx.doi.org/10.1093/humrep/16.ll.2445.
Raisanen, J: Niemela, R: Rosenberg, C. (2001). Tetrachloroethylene emissions and exposure in
dry cleaning. J Air Waste Manag Assoc 51: 1671-1675.
Rampy. LW: Quasi JF: Balmer. MF: Leong. BKJ: Gehring. PJ. (1978). Results of a long-term
inhalation toxicity study on rats of a perchloroethylene in rats (tetrachloroethylene)
formulation. Midland, MI: Dow Chemical.
Rao, HV: Brown, DR. (1993). A physiologically based pharmacokinetic assessment of
tetrachloroethylene in groundwater for a bathing and showering determination. Risk Anal
13: 37-49. http://dx.doi.Org/10.llll/i.1539-6924.1993.tb00727.x.
Rapson, WH; Nazar, MA; Butsky, VV. (1980). Mutagenicity produced by aqueous chlorination
of organic compounds. Bull Environ Contam Toxicol 24: 590-596.
http://dx.doi.org/10.1007/BF01608160.
Raucy, JL. (1995). Risk assessment: Toxicity from chemical exposure resulting from enhanced
expression of CYP2E1. Toxicology 105: 217-224. http://dx.doi.org/10.1016/0300-
483X(95)03216-3.
Raunio, H; Husgafvel-Pursiainen, K; Anttila, S: Hietanen, E; Hirvonen, A; Pelkonen, O. (1995).
Diagnosis of polymorphisms in carcinogen-activating and inactivating enzymes and
cancer susceptibility—a review [Review]. Gene 159: 113-121.
Rea, WJ: Ross, GH; Johnson, AR; Smiley, RE; Fenyes, EJ. (1991). Chemical sensitivity in
physicians. Bol Asoc Med P R 83: 383-388.
Rebelakos, A; Trichopoulos, D: Tzonou, A; Zavitsanos, X; Velonakis, E; Trichopoulos, A.
(1985). Tobacco smoking, coffee drinking, and occupation as risk factors for bladder
cancer in Greece. J Natl Cancer Inst 75: 455-461.
Reddv, JK: Reddv, MK: Usman, MI: Lalwani, ND: Rao. MS. (1986). Comparison of hepatic
peroxisome proliferative effect and its implication for hepatocarcinogenicity of phthalate
esters, di(2-ethylhexyl) phthalate, and di(2-ethylhexyl) adipate with a hypolipidemic
drug. Environ Health Perspect 65: 317-327.
Regan, D. (1989). Human brain eletrophysiology. New York: Elsevier Science.
Reichert, D: Neudecker, T; Spengler, U; Henschler, D. (1983). Mutagenicity of
dichloroacetylene and its degradation products trichloroacetyl chloride, trichloroacryloyl
chloride and hexachlorobutadiene. Mutat Res Genet Toxicol Environ Mutagen 117: 21-
29. http://dx.doi.org/10.1016/0165-1218(83)90149-0.
Reitz, RH; Mendrala, AL; Guengerich, FP. (1989). In vitro metabolism of methylene chloride in
human and animal tissues: Use in physiologically based pharmacokinetic models.
Toxicol Appl Pharmacol 97: 230-246. http://dx.doi.org/10.1016/0041-008X(89)90328-1.
Reitz, RH: Gargas, ML; Mendrala, AL: Schumann, AM. (1996). In vivo and in vitro studies of
perchloroethylene metabolism for physiologically based pharmacokinetic modeling in
7-53
-------
rats, mice, and humans. Toxicol Appl Pharmacol 136: 289-306.
http://dx.doi.org/10.1006/taap.1996.0036.
Renwick, AG. (1998). Toxicokinetics in infants and children in relation to the ADI and TDI.
Food Addit Contam 15: 17-35.
Reulen, R; Kellen, E; Buntinx, F; Zeegers, M. (2007). Bladder cancer and occupation: A report
from the Belgian case-control study on bladder cancer risk. Am J Ind Med 50: 449-454.
http://dx.doi.org/10.1002/ajim.20469.
Reulen, RC: Kellen, E; Buntinx, F; Brinkman, M; Zeegers, MP. (2008). A meta-analysis on the
association between bladder cancer and occupation. Scand J Urol Nephrol Suppl64-78.
http://dx.doi.org/10.1080/03008880802325192.
Reya, T; Morrison, SJ: Clarke, MF; Weissman, IL. (2001). Stem cells, cancer, and cancer stem
cells [Review]. Nature 414: 105-111. http://dx.doi.org/10.1038/35102167.
Richiardi, L; Boffetta, P; Simonato, L; Forastiere, F: Zambon, P; Fortes, C; Gaborieau, V:
Merletti, F. (2004). Occupational risk factors for lung cancer in men and women: A
population-based case-control study in Italy. Cancer Causes Control 15: 285-294.
http://dx.doi.Org/10.1023/B:CACO.0000024223.91059.ed.
Ries, LAG: Melbert, D: Krapcho, M: Stinchcomb, DG: Howlader, N; Horner, MJ: Mariotto, A;
Miller, BA; Feuer, EJ: Altekruse, SF; Lewis, PR: Clegg, L: Eisner, MP: Reichman, M:
Edwards. BK. (2008). SEER cancer statistics review, 1975-2005. Bethesda, MD: U.S.
National Institutes of Health, National Cancer Institute.
http://seer. cancer. gov/csr/1975_2005.
Riihimaki, V: Pfaffli, P. (1978). Percutaneous absorption of solvent vapors in man. Scand J
Work Environ Health 4: 73-85.
Ripp, SL; Overby, LH; Philpot RM: Elfarra, AA. (1997). Oxidation of cysteine S-conjugates by
rabbit liver microsomes and cDNA-expressed flavin-containing mono-oxygenases:
Studies with S-(l,2-dichlorovinyl)-L-cysteine, S-(l,2,2-trichlorovinyl)-L-cysteine, S-
allyl-L-cysteine, and S-benzyl-L-cysteine. Mol Pharmacol 51: 507-515.
Ripp, SL; Itagaki, K; Philpot RM: Elfarra, AA. (1999). Methionine S-oxidation in human and
rabbit liver microsomes: evidence for a high-affinity methionine S-oxidase activity that is
distinct from flavin-containing monooxygenase 3. Arch Biochem Biophys 367: 322-332.
http://dx.doi.org/10.1006/abbi.1999.1247.
Rodilla. V: Benzie. AA: Veitch. JM: Murray. GI: Rowe. JD: Hawksworth. GM. (1998).
Glutathione S-transferases in human renal cortex and neoplastic tissue: Enzymatic
activity, isoenzyme profile and immunohistochemical localization. Xenobiotica 28: 443-
456. http://dx.doi.org/10.1080/004982598239371.
Rodriguez, CE; Mahle, DA; Gearhart, JM: Mattie, PR: Lipscomb, JC: Cook, RS: Barton, HA.
(2007). Predicting age-appropriate pharmacokinetics of six volatile organic compounds in
the rat utilizing physiologically based pharmacokinetic modeling. Toxicol Sci 98: 43-56.
http://dx.doi.org/10.1093/toxsci/kfm082.
Roldan-Arjona, T; Garcia-Pedrajas, MD: Luque-Romero, FL; Hera, C: Pueyo, C. (1991). An
association between mutagenicity of the Ara test of Salmonella typhimurium and
carcinogenicity in rodents for 16 halogenated aliphatic hydrocarbons. Mutagenesis 6:
199-205. http://dx.doi.0rg/10.1093/mutage/6.3.199.
Rose, MG: Berliner, N. (2004). T-cell large granular lymphocyte leukemia and related disorders.
Oncologist 9: 247-258.
7-54
-------
Rosengren. LE: Kjellstrand. P: Haglid. KG. (1986). Tetrachloroethylene: Levels of DNA and S-
100 in the gerbil CNS after chronic exposure. Neurobehav Toxicol Teratol 8: 201-206.
Rubins. HB: Robins. SJ: Iwane. MK: Boden. WE: Elam. MB: Fye. CL: Gordon. DJ: Schaefen
EJ: Schectman, G: Wittes, JT. (1993). Rationale and design of the Department of
Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (HIT) for
secondary prevention of coronary artery disease in men with low high-density lipoprotein
cholesterol and desirable low-density lipoprotein cholesterol. Am J Cardiol 71: 45-52.
Rubins, HB: Robins, SJ: Collins, D: Fye, CL: Anderson, JW: Elam, MB; Faas, FH; Linares, E;
Schaefer. EJ: Schectman. G: Wilt TJ: Wittes. J. (1999). Gemfibrozil for the secondary
prevention of coronary heart disease in men with low levels of high-density lipoprotein
cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial
Study Group. N Engl J Med 341: 410-418.
http://dx.doi.org/10.1056/NEJM199908053410604.
Ruder, A. (2006). Potential health effects of occupational chlorinated solvent exposure. Ann N Y
Acad Sci 1076: 207-227. http://dx.doi.org/10.1196/annals.1371.050.
Ruder, AM; Ward, EM; Brown, DP. (1994). Cancer mortality in female and male dry-cleaning
workers. J Occup Med 36: 867-874.
Ruder, AM; Ward, EM; Brown, DP. (2001). Mortality in dry-cleaning workers: An update. Am J
IndMed39: 121-132. http://dx.doi.org/10.1002/1097-0274(200102)39:2<121::AID-
AJIM1000>3.0.CO:2-H.
Russo, A; Pacchierotti, F: Metalli, P. (1984). Nondisjunction induced in mouse spermatogenesis
by chloral hydrate, a metabolite of trichloroethylene. Environ Mutagen 6: 695-703.
http://dx.doi.org/10.1002/em.2860060507.
Russo, A; Levis, AG. (1992a). Detection of aneuploidy in male germ cells of mice by means of a
meiotic micronucleus assay. Mutat Res 281: 187-191. http://dx.doi.org/10.1016/0165-
7992(92)90007-5.
Russo, A; Levis, AG. (1992b). Further evidence for the aneuploidogenic properties of chelating
agents: Induction of micronuclei in mouse male germ cells by EDTA. Environ Mol
Mutagen 19: 125-131.
Russo, A; Stocco, A; Majone, F. (1992). Identification of kinetochore-containing (CREST+)
micronuclei in mouse bone marrow erythrocytes. Mutagenesis 7: 195-197.
http://dx.doi.0rg/10.1093/mutage/7.3.195.
Rusyn, I; Peters, JM; Cunningham, ML. (2006). Modes of action and species-specific effects of
di-(2-ethylhexyl)phthalate in the liver. Crit Rev Toxicol 36: 459-479.
http://dx.doi.org/10.1080/10408440600779065.
Saillenfait AM; Langonne, I: Sabate, JP. (1995). Developmental toxicity of trichloroethylene,
tetrachloroethylene and four of their metabolites in rat whole embryo culture. Arch
Toxicol 70: 71-82. http://dx.doi.org/10.1007/BF02733666.
Sakamoto, N. (1976). [Metabolism of tetrachloroethylene in guinea pigs (author's transl)].
Sangyolgaku 18: 11-16.
Saland, G. (1967). Accidental exposure to perchloroethylene. N Y State J Med 67: 2359-2361.
Sallmen, M; Lindbohm, ML; Kyyronen, P; Nykyri, E; Anttila, A; Taskinen, H; Hemminki, K.
(1995). Reduced fertility among women exposed to organic solvents. Am J Ind Med 27:
699-713. http://dx.doi.org/10.1002/aiim.4700270506.
7-55
-------
Sallmen, M; Lindbohm, ML; Anttila, A; Kyyronen, P; Taskinen, H; Nykyri, E; Hemminki, K.
(1998). Time to pregnancy among the wives of men exposed to organic solvents. Occup
Environ Med 55: 24-30.
Sanchez, IM; Bull, RJ. (1990). Early induction of reparative hyperplasia in the liver of B6C3F1
mice treated with dichloroacetate and trichloroacetate. Toxicology 64: 33-46.
http://dx.doi.org/10.1016/0300-483X(90)90097-Z.
Sarangapanl R: Gentry. PR: Covington. TR: Teeguarden. JG: Clewell HJ. III. (2003).
Evaluation of the potential impact of age- and gender-specific lung morphology and
ventilation rate on the dosimetry of vapors. Inhal Toxicol 15: 987-1016.
Sata. F: Sapone. A: Elizondo. G: Stocker. P: Miller. VP: Zheng. W: Raunio. H: Crespi. CL:
Gonzalez, FJ. (2000). CYP3A4 allelic variants with amino acid substitutions in exons 7
and 12: Evidence for an allelic variant with altered catalytic activity. Clin Pharmacol
Ther 67: 48-56. http://dx.doi.org/10.1067/mcp.2000.104391.
Sato, T; Mukaida, M; Ose, Y; Nagase, H: Ishikawa, T. (1985). Mutagenicity of chlorinated
products from soil humic substances. Sci Total Environ 46: 229-241.
http://dx.doi.org/10.1016/0048-9697(85)90296-7.
Savolainen, H: Pfaffli, P: Tengen, M: Vainio, H. (1977a). Biochemical and behavioural effects
of inhalation exposure to tetrachlorethylene and dichlormethane. J Neuropathol Exp
Neurol 36: 941-949.
Savolainen, H: Pfaffli, P: Tengen, M: Vainio, H. (1977b). Trichloroethylene and 1,1,1-
trichloroethane: Effects on brain and liver after five days intermittent inhalation. Arch
Toxicol 38: 229-237. http://dx.doi.org/10.1007/BF00293657.
Sax, NI: Lewis, RJ, Sr. (1987). Hazardous chemicals desk reference. New York, NY: Van
Nostrand Reinhold.
Sbrana, I; Di Sibio, A; Lomi, A; Scarcelli, V. (1993). C-mitosis and numerical chromosome
aberration analyses in human lymphocytes: 10 known or suspected spindle poisons.
Mutat Res 287: 57-70. http://dx.doi.org/10.1016/0027-5107(93)90145-6.
Scatena, R: Bottom, P: Vincenzoni, F; Messana, I; Martorana, G: Nocca, G: De Sole, P:
Maggiano, N; Castagnola, M: Giardina, B. (2003). Bezafibrate induces a mitochondrial
derangement in human cell lines: a PPAR-independent mechanism for a peroxisome
proliferator. Chem Res Toxicol 16: 1440-1447. http://dx.doi.org/10.1021/tx0341052.
Scharre. JE: Cotter. SA: Block. SS: Kelly. SA. (1990). Normative contrast sensitivity data for
young children. Optometry and Vision Science 67: 826-832.
Schatten, H: Chakrabarti, A. (1998). Centrosome structure and function is altered by chloral
hydrate and diazepam during the first reproductive cell cycles in sea urchin eggs. Eur J
Cell Biol 75: 9-20.
Schenk. M: Purdue. MP: Colt JS: Hartge. P: Blair. A: Stewart. P: Cerhan. JR: De Roos. AJ:
Cozen, W: Severson, RK. (2009). Occupation/industry and risk of non-Hodgkin's
lymphoma in the United States. Occup Environ Med 66: 23-31.
http://dx.doi.org/10.1136/oem.2007.036723.
Scherr, PA; Hutchison, GB; Neiman, RS. (1992). Non-Hodgkin's lymphoma and occupational
exposure. Cancer Res 52: 5503s-5509s.
Schlehofer, B; Heuer, C; Blettner, M: Niehoff D; Wahrendorf J. (1995). Occupation, smoking
and demographic factors, and renal cell carcinoma in Germany. Int J Epidemiol 24: 51-
57. http://dx.doi.0rg/10.1093/iie/24.l.51.
7-56
-------
Schlichting, LM; Wright PF; Stacey, NH. (1992). Effects of tetrachloroethylene on hepatic and
splenic lymphocytotoxic activities in rodents. Toxicol Ind Health 8: 255-266.
Schoenberg. JB: Stemhagen. A: Mogielnicki. AP: Altman. R: Abe. T: Mason. TJ. (1984). Case-
control study of bladder cancer in New Jersey. I. Occupational exposures in white males.
J Natl Cancer Inst 72: 973-981.
Schreiber. JS: House. S: Prohonic. E: Smead. G: Hudson. C: Stvk. M: Lauber. J. (1993). An
investigation of indoor air contamination in residences above dry cleaners. Risk Anal 13:
335-344. http://dx.doi.0rg/10.llll/i.1539-6924.1993.tb01085.x.
Schreiber, JS. (1993). Predicted infant exposure to tetrachloroethene in human breastmilk. Risk
Anal 13: 515-524. http://dx.doi.Org/10.llll/i.1539-6924.1993.tb00010.x.
Schreiber. JS: Hudnell. HK: Geller. AM: House. DE: Aldous. KM: Force. MS: Langguth. K:
Prohonic, EJ: Parker, JC. (2002). Apartment residents' and day care workers' exposures to
tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect
110: 655-664.
Schreiber, JS. (1997). Transport of organic chemicals to breast milk: Tetrachloroethene case
study. In S Kacew; G Lambert (Eds.), Environmental toxicology and pharmacology of
human development (pp. 95-143). Washington, DC: Taylor and Francis.
Schumann, AM; Quasi JF: Watanabe, PG. (1980). The pharmacokinetics and macromolecular
interactions of perchloroethylene in mice and rats as related to oncogenicity. Toxicol
Appl Pharmacol 55: 207-219. http://dx.doi.org/10.1016/0041-008X(80)90082-4.
Schwetz, BA; Leong, BKJ: Gehring, PJ. (1975). The effect of maternally inhaled
trichloroethylene, perchloroethylene, methyl chloroform, and methylene chloride on
embryonal and fetal development in mice and rats. Toxicol Appl Pharmacol 32: 84-96.
http://dx.doi.org/10.1016/0041-008X(75)90197-0.
Scott WJ, Jr; Collins, MD; Nau, H. (1994). Pharmacokinetic determinants of embryotoxicity in
rats associated with organic acids. Environ Health Perspect 102: 97-101.
Seeber, A. (1989). Neurobehavioral toxicity of long-term exposure to tetrachloroethylene.
Neurotoxicol Teratol 11: 579-583. http://dx.doi.org/10.1016/0892-0362(89)90041-X.
Seelbach, A; Fissler, B; Madle, S. (1993). Further evaluation of a modified micronucleus assay
with V79 cells for detection of aneugenic effects. Mutat Res 303: 163-169.
Seidel, HJ: Weber, L; Barthel, E. (1992). Hematological toxicity of tetrachloroethylene in mice.
Arch Toxicol 66: 228-230. http://dx.doi.org/10.1007/BF01974021.
Seidler, A; Mohner, M; Berger, J: Mester, B; Deeg, E: Eisner, G: Meters, A; Becker, N. (2007).
Solvent exposure and malignant lymphoma: A population-based case-control study in
Germany. J Occup Med Toxicol 2: 2. http://dx.doi.org/10.1186/1745-6673-2-2.
Seiji, K: Jin, C: Watanabe, T; Nakatsuka, H; Ikeda, M. (1990). Sister chromatid exchanges in
peripheral lymphocytes of workers exposed to benzene, trichloroethylene, or
tetrachloroethylene, with reference to smoking habits. Int Arch Occup Environ Health 62:
171-176. http://dx.doi.org/10.1007/BF00383594.
Seip, H: Alstad, J: Carlberg, G: Martinsen, K: Skaane, R. (1986). Measurement of mobility of
organic compounds in soils. Sci Total Environ 50: 87-101.
http://dx.doi.org/10.1016/0048-9697(86)90353-0.
Selden, AI; Ahlborg, G. (2011). Cancer morbidity in Swedish dry-cleaners and laundry workers:
Historically prospective cohort study. Int Arch Occup Environ Health 84: 435-443.
http://dx.doi.org/10.1007/s00420-010-0582-7.
7-57
-------
Seo. M: Ikeda. K: Okamura. T: Kida. K: Satoh. M: Inagakl N: Nagai. H: Nagase. H. (2008a).
Enhancing effect of chlorinated organic solvents on histamine release and inflammatory
mediator production. Toxicology 243: 75-83. http://dx.doi.Org/10.1016/j.tox.2007.09.024.
Seo, M: Yamagiwa, T; Kobayashi, R; Ikeda, K: Satoh, M: Inagaki, N: Nagai, H: Nagase, H.
(2008b). A small amount of tetrachloroethylene ingestion from drinking water accelerates
antigen-stimulated allergic responses. Immunobiology 213: 663-669.
http://dx.doi.0rg/10.1016/i.imbio.2008.01.009.
Serra, C: Bonfill, X; Sunyer, J: Urrutia, G: Turuguet D: Bastus, R: Rogue, M: 'tMannetje, A;
Kogevinas. M: Occidental WGotSoBCitCoV. (2000). Bladder cancer in the textile
industry. Scand J Work Environ Health 26: 476-481.
Sertznig, P; Seifert, M: Tilgen, W; Reichrath, J. (2007). Present concepts and future outlook:
Function of peroxisome proliferator-activated receptors (PPARs) for pathogenesis,
progression, and therapy of cancer. J Cell Physiol 212: 1-12.
http://dx.doi.org/10.1002/icp.20998.
Sexton, K: Adgate, JL; Church, TR; Ashley, PL; Needham, LL; Ramachandran, G: Fredrickson,
AL: Ryan, AD. (2005). Children's exposure to volatile organic compounds as determined
by longitudinal measurements in blood. Environ Health Perspect 113: 342-349.
http://dx.doi.org/10.1289/ehp.7412.
Shafer, TJ: Bushnell, PJ: Benignus, VA; Woodward, JJ. (2005). Perturbation of voltage-sensitive
Ca2+ channel function by volatile organic solvents. J Pharmacol Exp Ther 315: 1109-
1118. http://dx.doi.0rg/10.l 124/jpet. 105.090027.
Shao. J: Stapleton. PL: Lin. YS: Gallagher. EP. (2007). Cytochrome p450 and glutathione s-
transferase mRNA expression in human fetal liver hematopoietic stem cells. Drug Metab
Dispos35: 168-175. http://dx.doi.org/10.1124/dmd.106.012757.
Sharanjeet-Kaur; Mursyid, A; Kamaruddin, A; Ariffm, A. (2004). Effect of petroleum
derivatives and solvents on colour perception. Clin Exp Optom 87: 339-343.
http://dx.doi.0rg/10.llll/i.1444-0938.2004.tb05064.x.
Shearer, BG: Hoekstra, WJ. (2003). Recent advances in peroxisome proliferator-activated
receptor science. CurrMed Chem 10: 267-280.
http://dx.doi.org/10.2174/1567204052930970.
Sheldon, LS: Handy, RW: Hartwell, TD; Leininger, C: Zelon, H. (1985). Human exposure
assessment to environmental chemicals: nursing mothers study. Final report. U.S.
Environmental Protection Agency.
Sher, T: Yi, HF; McBride, OW: Gonzalez, FJ. (1993). cDNA cloning, chromosomal mapping,
and functional characterization of the human peroxisome proliferator activated receptor.
Biochemistry 32: 5598-5604. http://dx.doi.org/10.1021/bi00072a015.
Shevchenko, VV: Grinikh, LI; Alekseenok, AY. (1985). Clastogenic effect of oxalic acid and its
specificity. Sov Genet 21: 606-610.
Shimada, T: Swanson, AF; Leber, P; Williams, GM. (1985). Activities of chlorinated ethane and
ethylene compounds in the Salmonella/rat microsome mutagenesis and rat
hepatocyte/DNA repair assays under vapor phase exposure conditions. Cell Biol Toxicol
1: 159-179. http://dx.doi.org/10.1007/BF00120162.
Shimada, T: Yamazaki, H: Mimura, M: Inui, Y; Guengerich, FP. (1994). Interindividual
variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs,
carcinogens and toxic chemicals: Studies with liver microsomes of 30 Japanese and 30
Caucasians. J Pharmacol Exp Ther 270: 414-423.
7-58
-------
Shirasu, Y; Moriya, M; Kato, K; Furuhashi, A; Kada, T. (1976). Mutagenicity screening of
pesticides in the microbial system. Mutat Res 40: 19-30. http://dx.doi.org/10.1016/0165-
1218(76)90018-5.
Shrivastav, BB; Narahashi, T; Kitz, RJ: Roberts, JD. (1976). Mode of action of trichloroethylene
on squid axon membranes. J Pharmacol Exp Ther 199: 179-188.
Shu. XO: Stewart. P: Wen. WQ: Han. D: Potter. JD: Buckley. JD: Heineman. E: Robison. LL.
(1999). Parental occupational exposure to hydrocarbons and risk of acute lymphocytic
leukemia in offspring. Cancer Epidemiol Biomarkers Prev 8: 783-791.
Siemiatycki, J, . (Ed.). (1991). Risk factors for cancer in the workplace. Boca Raton, FL: CRC
Press.
Silverman, DT: Levin, LI; Hoover, RN: Hartge, P. (1989a). Occupational risks of bladder cancer
in the United States: I. White men. J Natl Cancer Inst 81: 1472-1480.
http://dx.doi.org/10.1093/inci/81.19.1472.
Silverman, DT: Levin, LI; Hoover, RN. (1989b). Occupational risks of bladder cancer in the
United States: II. Nonwhite men. JNatl Cancer Inst 81: 1480-1483.
http://dx.doi.org/10.1093/inci/81.19.1480.
Silverman, DT: Levin, LI; Hoover, RN. (1990). Occupational risks of bladder cancer among
white women in the United States. Am J Epidemiol 132: 453-461.
Singh, M: Sinha, U. (1976). Chloral hydrate induced haploidization in Aspergillus nidulans.
Experientia 32: 1144-1145. http://dx.doi.org/10.1007/BF01927592.
Singh, M: Sinha, U. (1979). Mitotic haploidization and growth of Aspergillus nidulans on media
containing chloral hydrate. J Cytol Genet 14: 1-4.
Skender, LJ: Karacic, V: Prpic-Majic, D. (1991). A comparative study of human levels of
trichloroethylene and tetrachloroethylene after occupational exposure. Arch Environ
Health 46: 174-178. http://dx.doi.org/10.1080/15459620802510866.
Skowron, J: Miranowicz-Dzierzawska, K; Zapor, L; Golofit-Szymczak, M: Starek, A. (2001).
Interactions of some organic solvents: Hydrocarbons and chloroalkene. Int J Occup Saf
Ergon 7: 35-47.
Smith, EM; Miller, ER; Woolson, RF; Brown, CK. (1985). Bladder cancer risk among laundry
workers, dry cleaners, and others in chemically-related occupations. J Occup Med 27:
295-297.
Smith, MK; Randall JL; Read, EJ: Stober, JA. (1989). Teratogenic activity of trichloroacetic
acid in the rat. Teratology 40: 445-451. http://dx.doi.org/10.1002/tera.1420400506.
Sofuni, T: Hayashi, M: Matsuoka, A; Sawada, M: Hatanaka, M: M Jr, I. (1985). [Mutagenicity
tests on organic chemical contaminants in city water and related compounds. II.
Chromosome aberration tests in cultured mammalian cells]. Kokuritsu lyakuhin
Shokuhin Eisei Kenkyusho Hokoku 103: 64-75.
Sokol, L: Loughran, TP, Jr. (2006). Large granular lymphocyte leukemia. Oncologist 11: 263-
273. http://dx.doi.org/10.1634/theoncologist.ll-3-263.
Solet D: Robins, TG: Sampaio, C. (1990). Perchloroethylene exposure assessment among dry
cleaning workers. AIHA J 51: 566-574.
Solet D: Robins, TG. (1991). Renal function in dry cleaning workers exposed to
perchloroethylene. Am J Ind Med 20: 601-614.
Solleveld, HA; Haseman, JK; McConnell, EE. (1984). Natural history of body weight gain,
survival, and neoplasia in the F344 rat. J Natl Cancer Inst 72: 929-940.
7-59
-------
Sonnenfeld, N; Hertz-Picciotto, I; Kaye, WE. (2001). Tetrachloroethylene in drinking water and
birth outcomes at the US Marine Corps Base at Camp Lejeune, North Carolina. Am J
Epidemiol 154: 902-908. http://dx.doi.org/10.1093/aie/154.10.902.
Sora, S; Agostini Carbone, ML. (1987). Chloral hydrate, methylmercury hydroxide and ethidium
bromide affect chromosomal segregation during meiosis of Saccharomyces cerevisiae.
MutatRes Lett 190: 13-17. http://dx.doi.org/10.1016/0165-7992(87)90075-3.
Sparrow, GP. (1977). A connective tissue disorder similar to vinyl chloride disease in a patient
exposed to perchlorethylene. Clin Exp Dermatol 2: 17-22.
Speerschneider, P; Dekant, W. (1995). Renal tumorigenicity of 1,1-dichloroethene in mice: the
role of male-specific expression of cytochrome P450 2E1 in the renal bioactivation of
1,1-dichloroethene. Toxicol Appl Pharmacol 130: 48-56.
Spencer, HB; Hussein, WR: Tchounwou, PB. (2002). Effects of tetrachloroethylene on the
viability and development of embryos of the Japanese medaka, Oryzias latipes. Arch
Environ Contam Toxicol 42: 463-469. http://dx.doi.org/10.1007/s00244-001-0050-l.
Spinatonda, G: Colombo, R: Capodaglio, EM; Imbriani, M; Pasetti, C; Minuco, G: Pinelli, P.
(1997). Processes of speech production: Application in a group of subjects chronically
exposed to organic solvents (II). G Ital Med Lav Ergon 19: 85-88.
Spirtas. R: Stewart. PA: Lee. JS: Marano. DE: Forbes. CD: Grauman. DJ: Pettigrew. HM: Blair.
A; Hoover, RN: Cohen, JL. (1991). Retrospective cohort mortality study of workers at an
aircraft maintenance facility: I. Epidemiological results. Br J Ind Med 48: 515-530.
http://dx.doi.0rg/10.1136/oem.48.8.515.
Stanley. LA: Blackburn. PR: Devereaux. S: Folev. J: Lord. PG: Maronpot RR: Orton. TC:
Anderson, MW. (1994). Ras mutations in methylclofenapate-induced B6C3F1 and
C57BL/10J mouse liver tumours. Carcinogenesis 15: 1125-1131.
Stauber, AJ: Bull, RJ. (1997). Differences in phenotype and cell replicative behavior of hepatic
tumors induced by dichloroacetate (DCA) and trichloroacetate (TCA). Toxicol Appl
Pharmacol 144: 235-246. http://dx.doi.org/10.1006/taap.1997.8159.
Steineck. G: Plato. N: Gerhardsson. M: Norell. SE: Hogstedt C. (1990). Increased risk of
urothelial cancer in Stockholm during 1985-87 after exposure to benzene and exhausts.
IntJ Cancer 45: 1012-1017. http://dx.doi.org/10.1002/iic.2910450605.
Stemhagen, A; Slade, J: Altaian, R: Bill, J. (1983). Occupational risk factors and liver cancer: A
retrospective case-control study of primary liver cancer in New Jersey. Am J Epidemiol
117:443-454.
Stephens. EA: Taylor. JA: Kaplan. N: Yang. CH: Hsieh. LL: Lucier. GW: Bell DA. (1994).
Ethnic variation in the CYP2E1 gene: Polymorphism analysis of 695 African-Americans,
European-Americans and Taiwanese. Pharmacogenetics 4: 185-192.
Stevens, J: Jakoby, WB. (1983). Cysteine conjugate beta-lyase. Mol Pharmacol 23: 761-765.
Stevens, JL. (1985). Isolation and characterization of a rat liver enzyme with both cysteine
conjugate beta-lyase and kynureninase activity. J Biol Chem 260: 7945-7950.
Stewart. RD: Erlev. PS: Schaffer. AW: Gay. HH. (1961). Accidental vapor exposure to
anesthetic concentrations of a solvent containing tetrachloroethylene. Ind Med Surg 30:
327-330.
Stewart, RD: Dodd, HC. (1964). Absorption of carbon tetrachloride, trichloroethylene,
tetrachloroethylene, methylene chloride, and 1, 1, 1- trichloroethane through the human
skin. Am Ind Hyg Assoc J 25: 439-446.
7-60
-------
Stewart, RD: Baretta, ED; Dodd, HC: Torkelson, TR. (1970). Experimental human exposure to
tetrachloroethylene. Arch Environ Health 20: 224-229.
Stewart. RD: Hake. CL: Forster. AJ: Lubrun. AJ: Peterson. JE: Wu. A. (1974).
Tetrachloroethylene: Development of a biologic standard for the industrial worker by
breath analysis. Atlanta, GA: National Institute of Occupational Safety and Health.
http://www.ntis. gov/search/product.aspx?ABBR=PB82152166.
Stewart. RD: Hake. CL: Wu. A: Kalbfleisch. J: Newton. PE: Marlow. SK: Vucicevic-Salama. M.
(1977). Effects of perchloroethylene/drug interaction on behavior and neurological
function. Milwaukee, WI: Allen-Bradley Medical Science Laboratory.
Stiteler. WM: Knauf LA: Hertzberg. RC: Schoeny. RS. (1993). A statistical test of compatibility
of data sets to a common dose-response model. Regul Toxicol Pharmacol 18: 392-402.
http://dx.doi.org/10.1006/rtph.1993.1065.
Storm, JE; Mazor, KA. (2004). Update of residential tetrachloroethylene exposure and decreases
in visual contrast sensitivity. Environ Health Perspect 112: A862-A864.
Storm. JE: Mazor. KA: Aldous. KM: Blount BC: Brodie. SE: Serle. JB. (2011). Visual contrast
sensitivity in children exposed to tetrachloroethylene. Arch Environ Occup Health 66:
166-177. http://dx.doi.org/10.1080/19338244.2010.539638.
Stott WT; Quasi JF; Watanabe, PG. (1982). The pharmacokinetics and macromolecular
interactions of trichloroethylene in mice and rats. Toxicol Appl Pharmacol 62: 137-151.
Stromberg, PC: Rojko, JL; Vogtsberger, LM; Cheney, C: Berman, R. (1983). Immunologic,
biochemical, and ultrastructural characterization of the leukemia cell in F344 rats. J Natl
Cancer Inst 71: 173-181.
Stromberg, PC. (1985). Large granular lymphocyte leukemia in F344 rats: Model for human T
gamma lymphoma, malignant histiocytosis, and T-cell chronic lymphocytic leukemia.
AmJPathol 119: 517-519.
Styles, JA; Wyatt I; Courts, C. (1991). Trichloroacetic acid: Studies on uptake and effects on
hepatic DNA and liver growth in mouse. Carcinogenesis 12: 1715-1719.
http://dx.doi.0rg/10.1093/carcin/12.9.1715.
Suarez, L; Weiss, NS: Martin, J. (1989). Primary liver cancer death and occupation in Texas. Am
JIndMed 15: 167-175.
Subramoniam. A: Goel SK: Pandva. KP: Seth. PK. (1989). Influence of trichloroethylene
treatment on phosphoinositides in rat brain. Toxicol Lett 49: 55-60.
http://dx.doi.org/10.1016/0378-4274(89)90101-X.
Sung. TI: Chen. PC: Lee. LJ. -H: Lin. YP: Hsieh. GY: Wang. JD. (2007). Increased standardized
incidence ratio of breast cancer in female electronics workers. BMC Public Health 7:
102. http://dx.doi.org/10.1186/1471-2458-7-102.
Sung, TI; Wang, JD: Chen, PC. (2008). Increased risk of cancer in the offspring of female
electronics workers. Reprod Toxicol 25: 115-119.
http://dx.doi.0rg/10.1016/j.reprotox.2007.08.004.
Sung, TI; Wang, JD: Chen, PC. (2009). Increased risks of infant mortality and of deaths due to
congenital malformation in the offspring of male electronics workers. Birth Defects Res
A ClinMol Teratol 85: 119-124. http://dx.doi.org/10.1002/bdra.20496.
Sverdrup, B; Kallberg, H; Bengtsson, C: Lundberg, I; Padyukov, L; Alfredsson, L; Klareskog, L;
Group, EIoRAS. (2005). Association between occupational exposure to mineral oil and
rheumatoid arthritis: results from the Swedish EIRA case-control study. Arthritis Res
Ther 7: R1296-R1303. http://dx.doi.org/10.1186/arl824.
7-61
-------
Swanson, GM; Burns, PB. (1995). Cancer incidence among women in the workplace: A study of
the association between occupation and industry and 11 cancer sites. J Occup Environ
Med 37: 282-287.
Swenberg, JA; Lehman-McKeeman, LD. (1999). alpha 2-Urinary globulin-associated
nephropathy as a mechanism of renal tubule cell carcinogenesis in male rats. In CC
Capen; E Dybing; JM Rice; JD Wilbourn (Eds.), Species differences in thyroid, kidney
and urinary bladder carcinogenesis (pp. 95-118). Lyon, France: International Agency for
Research on Cancer.
http://apps.who.int/bookorders/anglais/detartl.jsp?sesslan=l&codlan=l&codcol=73&cod
cch=147.
Swerdlow. SH: Campo. E: Harris. NL: Jaffe. ES: Piled. SA: Stein. H: Thiele. J: Vardiman. JW.
(2008). WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. In th
(Ed.). Geneva, Switzerland: WHO Press.
S winker, M; Burke, WA. (2002). Visual contrast sensitivity as a diagnostic tool. Environ Health
Perspect 110: A120-A123. http://dx.doi.org/10.1289/ehp.110-al20.
Szakmary, E: GUngvary, G: Tatrai, E. (1997). The offspring-damaging effect of
tetrachloroethylene in rats, mice, and rabbits. Central Eur J Occup Env Med 3: 31-39.
Tabacova, S. (1986). Maternal exposure to environmental chemicals. Neurotoxicology 7: 421-
440.
Tal ES: Collins. D: Robins. SJ: O'Connor. JJ: Bloomfield. HE: Ordovas. JM: Schaefer. EJ:
Brousseau, ME. (2006). The LI62V polymorphism at the peroxisome proliferator
activated receptor alpha locus modulates the risk of cardiovascular events associated with
insulin resistance and diabetes mellitus: The Veterans Affairs HDL Intervention Trial
(VA-HIT). Atherosclerosis 187: 153-160.
http://dx.doi.0rg/10.1016/j.atherosclerosis.2005.08.034.
Takashima, K; Ito, Y; Gonzalez, FJ: Nakajima, T. (2008). Different mechanisms of DEHP-
induced hepatocellular adenoma tumorigenesis in wild-type and Ppar alpha-null mice. J
Occup Health 50: 169-180.
Tanaka. K: Smith. PF: Stromberg. PC: Eydelloth. RS: Herold. EG: Grossman. SJ: Frank. JD:
Hertzog. PR: Soper. KA: Keenan. KP. (1992). Studies of early hepatocellular
proliferation and peroxisomal proliferation in Sprague-Dawley rats treated with
tumorigenic doses of clofibrate. Toxicol Appl Pharmacol 116: 11-11.
http://dx.doi.org/10.1016/0041-008X(92)90146-J.
Tanaka, S: Ikeda, M. (1968). A method for determination of trichloroethanol and trichloroacetic
acid in urine. Occup Environ Med 25: 214-219.
Tanaka, T; Ordovas, JM: Delgado-Lista, J: Perez-Jimenez, F; Marin, C; Perez-Martinez, P;
Gomez, P; Lopez-Miranda, J. (2007). Peroxisome proliferator-activated receptor alpha
polymorphisms and postprandial lipemia in healthy men. J Lipid Res 48: 1402-1408.
http://dx.doi.org/10.1194/ilr.M700066-JLR200.
Tang, ML; Kemp, AS: Thorburn, J: Hill, DJ. (1994). Reduced interferon-gamma secretion in
neonates and subsequent atopy. Lancet 344: 983-985. http://dx.doi.org/10.1016/S0140-
6736(94)91641-1.
Tanios, MA; El Gamal, H; Rosenberg, BJ: Hassoun, PM. (2004). Can we still miss
tetrachloroethylene-induced lung disease? The emperor returns in new clothes.
Respiration 71: 642-645. http://dx.doi.org/10.1159/000081768.
7-62
-------
Tao, L; Kramer, PM: Ge, R: Pereira, MA. (1998). Effect of dichloroacetic acid and
trichloroacetic acid on DNA methylation in liver and tumors of female B6C3F1 mice.
Toxicol Sci 43: 139-144. http://dx.doi.Org/10.1093/toxsci/43.2.139.
Tao, L: Wang, W; Li, L; Kramer, PM: Pereira, MA. (2004). Effect of dibromoacetic acid on
DNA methylation, glycogen accumulation, and peroxisome proliferation in mouse and rat
liver. Toxicol Sci 82: 62-69. http://dx.doi.org/10.1093/toxsci/kfh266.
Taskinen, H; Anttila, A; Lindbohm, ML; Sallmen, M; Hemminki, K. (1989). Spontaneous
abortions and congenital malformations among the wives of men occupationally exposed
to organic solvents. Scand J Work Environ Health 15: 345-352.
Tateishi, M; Suzuki, S: Shimizu, H. (1978). Cysteine conjugate beta-lyase in rat liver. A novel
enzyme catalyzing formation of thiol-containing metabolites of drugs. J Biol Chem 253:
8854-8859.
Tateishi, T; Nakura, H: Asoh, M: Watanabe, M: Tanaka, M: Kumai, T; Takashima, S: Imaoka,
S: Funae, Y; Yabusaki, Y; Kamataki, T; Kobayashi, S. (1997). A comparison of hepatic
cytochrome P450 protein expression between infancy and postinfancy. Life Sci 61: 2567-
2574. http://dx.doi.org/10.1016/50024-3205(97)01011-4.
Tenkanen, L: Manttari, M: Kovanen, PT; Virkkunen, H: Manninen, V. (2006). Gemfibrozil in
the treatment of dyslipidemia: an 18-year mortality follow-up of the Helsinki Heart
Study. Arch Intern Med 166: 743-748. http://dx.doi.Org/10.1001/archinte.166.7.743.
Terrier. P: Townsend, AJ: Coindre, JM: Triche, TJ: Cowan. KH. (1990). An
immunohistochemical study of pi class glutathione S-transferase expression in normal
human tissue. Am J Pathol 137: 845-853.
Teschke, K; Morgan, MS: Checkoway, H: Franklin, G: Spinelli, JJ: van Belle, G: Weiss, NS.
(1997). Surveillance of nasal and bladder cancer to locate sources of exposure to
occupational carcinogens. Occup Environ Med 54: 443-451.
Thai, SF; Allen, JW: DeAngelo, AB; George, MH; Fuscoe, JC. (2003). Altered gene expression
in mouse livers after dichloroacetic acid exposure. Mutat Res Rev Mutat Res 543: 167-
180. http://dx.doi.org/10.1016/S1383-5742(03)00014-0.
Thomas, J: Haseman, JK; Goodman, JI; Ward, JM: Loughran, TP, Jr; Spencer, PJ. (2007). A
review of large granular lymphocytic leukemia in Fischer 344 rats as an initial step
toward evaluating the Implication of the endpoint to human cancer risk assessment
[Review]. Toxicol Sci 99: 3-19. http://dx.doi.org/10.1093/toxsci/kfm098.
Thomas. KW: Pellizzarl ED: Perritt RL: Nelson. WC. (1991). Effect of dry-cleaned clothes on
tetrachloroethylene levels in indoor air, personal air, and breath for residents of several
New Jersey homes. J Expo Anal Environ Epidemiol 1: 475-490.
Tichenor. BA: Sparks. LE: Jackson. MD: Quo. Z: Mason. MA: Plunket CM: Rasor. SA. (1990).
Emissions of perchloroethylene from dry cleaned fabrics. Atmos Environ 24: 1219-1229.
http://dx.doi.org/10.1016/0960-1686(90)90087-4.
Till, C: Westall, CA; Rovet JF; Koren, G. (200la). Effects of maternal occupational exposure to
organic solvents on offspring visual functioning: A prospective controlled study.
Teratology 64: 134-141. http://dx.doi.org/10.1002/tera.1056.
Till, C: Koren, G: Rovet JF. (200Ib). Prenatal exposure to organic solvents and child
neurobehavioral performance. Neurotoxicol Teratol 23: 235-245.
http://dx.doi.org/10.1016/S0892-0362(01)00141-6.
7-63
-------
Till, C: Rovet JF: Koren, G: Westall, CA. (2003). Assessment of visual functions following
prenatal exposure to organic solvents. Neurotoxicology 24: 725-731.
http://dx.doi.org/10.1016/S0161-813X(02)00212-7.
Till, C: Westall, CA: Koren, G: Nulman, I; Rovet JF. (2005). Vision abnormalities in young
children exposed prenatally to organic solvents. Neurotoxicology 26: 599-613.
http://dx.doi.0rg/10.1016/i.neuro.2005.05.011.
Tinston, DJ. (1994). Perchloroethylene: A multigeneration inhalation study in the rat.
(CTL/P/4097, 86950000190). Cheshire, UK: Zeneca Central Toxicology Laboratory.
U.S. EPA TSCA Submission. http://www.ntis.gov/search/product.aspx?ABBR=OTS0557702
Tomisawa, H; Suzuki, S: Ichihara, S: Fukazawa, H; Tateishi, M. (1984). Purification and
characterization of C-S lyase from Fusobacterium varium: A C-S cleavage enzyme of
cysteine conjugates and some S-containing amino acids. J Biol Chem 259: 2588-2593.
Tomisawa, H: Ichihara, S: Fukazawa, H: Ichimoto, N; Tateishi, M; Yamamoto, I. (1986).
Purification and characterization of human hepatic cysteine-conjugate beta-lyase.
BiochemJ235: 569-575.
long, Z; Board, PG: Anders, MW. (1998a). Glutathione transferase zeta-catalyzed
biotransformation of dichloroacetic acid and other alpha-haloacids. Chem Res Toxicol
11: 1332-1338. http://dx.doi.org/10.1021/tx980144f.
Tong, Z; Board, PG: Anders, MW. (1998b). Glutathione transferase zeta catalyses the
oxygenation of the carcinogen dichloroacetic acid to glyoxylic acid. Biochem J 331: 371-
374.
Topping, MD; Forster, HW: Dolman, C: Luczynska, CM; Bernard, AM. (1986). Measurement of
urinary retinol-binding protein by enzyme-linked immunosorbent assay, and its
application to detection of tubular proteinuria. Clin Chem 32: 1863-1866.
Toraason, M: Clark. J: Dankovic, D: Mathias, P: Skaggs, S: Walker. C: Werren, D. (1999).
Oxidative stress and DNA damage in Fischer rats following acute exposure to
trichloroethylene or perchloroethylene. Toxicology 138: 43-53.
http://dx.doi.org/10.1016/S0300-483X(99)00083-9.
Toraason, M: Butler, MA; Ruder, A; Forrester, C: Taylor, L; Ashley, PL; Mathias, P: Marlow,
KL; Cheever, KL: Krieg, E: Wey, H. (2003). Effect of perchloroethylene, smoking, and
race on oxidative DNA damage in female dry cleaners. Mutat Res Genet Toxicol Environ
Mutagen 539: 9-18. http://dx.doi.org/10.1016/S1383-5718(03)00130-X.
Travier, N: Gridlev, G: De Roos, AJ: Plato. N: Moradi, T: Boffetta, P. (2002). Cancer incidence
of dry cleaning, laundry and ironing workers in Sweden. Scand J Work Environ Health
28: 341-348.
Trevisan, A; Cristofori, P: Fanelli, G. (1999). Glutamine synthetase activity in rat urine as
sensitive marker to detect S3 segment-specific injury of proximal tubule induced by
xenobiotics. Arch Toxicol 73: 255-262. http://dx.doi.org/10.1007/s002040050614.
Trevisan, A; Macca: I: Rui, F; Carried, M: Bartolucci, GB; Manno, M. (2000). Kidney and liver
biomakers in female dry-cleaning workers exposed to perchloroethylene. Biomarkers 5:
399-409.
Tschopp. C: Safran. AB: Vivianl P: Bullinger. A: Reicherts. M: Mermoud. C. (1998).
Automated visual field examination in children aged 5-8 years. Part I: Experimental
validation of a testing procedure. Vision Res 38: 2203-2210.
Tsuruta, H. (1989). Skin absorption of organic solvent vapors in nude mice in vivo. Ind Health
27: 37-47. http://dx.doi.org/10.2486/indhealth.27.37.
7-64
-------
Tu. AS: Murray. TA: Hatch. KM: Sivak. A: Miltnan. HA. (1985). In vitro transformation of
BALB/C-3T3 cells by chlorinated ethanes and ethylenes. Cancer Lett 28: 85-92.
http://dx.doi.org/10.1016/0304-3835(85)90096-5.
Tugwood, JD: Aldridge, TC: Lambe, KG: Macdonald, N; Woodyatt NJ. (1996). Peroxisome
proliferator-activated receptors: Structures and function. Ann N Y Acad Sci 804: 252-
265. http://dx.doi.0rg/10.llll/i.1749-6632.1996.tbl8620.x.
Tzeng, HF; Blackburn, AC: Board, PG: Anders, MW. (2000). Polymorphism- and species-
dependent inactivation of glutathione transferase zeta by dichloroacetate. Chem Res
Toxicol 13: 231-236. http://dx.doi.org/10.1021/tx990175q.
U.S. EPA (U.S. Environmental Protection Agency). (1986b). Guidelines for mutagenicity risk
assessment. (EPA/630/R-98/003). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum, http://www.epa.gov/iris/backgrd.html.
U.S. EPA (U.S. Environmental Protection Agency). (1986c). Guidelines for the health risk
assessment of chemical mixtures. (EPA/630/R-98/002). Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (1987). Health advisories for 25 organics.
Office of drinking water. U.S. Environmental Protection Agency, available through the
National Technical Information Service (NTIS), Springfield, VA, PB87-2355781AS.
U.S. EPA (U.S. Environmental Protection Agency). (1988). Recommendations for and
documentation of biological values for use in risk assessment. (EPA/600/6-87/008).
Cincinnati, OH: U.S. Environmental Protection Agency, Environmental Criteria and
Assessment Office. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.
U.S. EPA (U.S. Environmental Protection Agency). (1991a). Alpha-2u-globulin: Association
with chemically induced renal toxicity and neoplasia in the male rat. (EP A/625/3-
91/019F). Washington, DC: U.S. Environmental Protection Agency, Risk Assessment
Forum, http://www.ntis.gov/search/product.aspx?ABBR=PB92143668.
U.S. EPA (U.S. Environmental Protection Agency). (1991b). Guidelines for developmental
toxicity risk assessment. (EPA/600/FR-91/001). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum, http://www.epa.gov/iris/backgrd.html.
U.S. EPA (U.S. Environmental Protection Agency). (1992). Draft report: A cross-species scaling
factor for carcinogen risk assessment based on equivalence of mg/kg(3/4)/day. Fed Reg
57:24151-24173.
U.S. EPA (U.S. Environmental Protection Agency). (1994). Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry. (EPA/600/8-90/066F).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
and Development, Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid=71993.
U.S. EPA (U.S. Environmental Protection Agency). (1995). The use of the benchmark dose
approach in health risk assessment. (EPA/630/R-94/007). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/raf/publications/useof-bda-healthrisk.htm.
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Guidelines for reproductive toxicity
risk assessment. (EPA/630/R-96/009). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum.
http://www.epa.gov/raf/publications/pdfs/REPRO51 .PDF.
7-65
-------
U.S. EPA (U.S. Environmental Protection Agency). (1996b). Modeled ambient concentrations:
perchloroethylene (CASRN 127-18-4). National Air Toxics Assessment, Technology
Transfer Network, Office of Air and Radiation.
http://www.epa.gov/ttnatwO 1 /nata/pdf/perch cone.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (1998a). Dichloroacetic acid:
carcinogenicity identification characterization summary. (EPA 815-B-98-010).
Washington, DC: U.S. Environmental Protection Agency; National Center for
Environmental Assessment.
http://www.ntis. gov/search/product.aspx?ABBR=PB99111387.
U.S. EPA (U.S. Environmental Protection Agency). (1998b). Guidelines for neurotoxicity risk
assessment. (EPA/630/R-95/001F). Washington, DC: U.S. Environmental Protection
Agency, Risk Assessment Forum.
http://www.epa. gov/raf/publications/pdfs/NEUROTOX.PDF.
U.S. EPA (U.S. Environmental Protection Agency). (2000a). Benchmark dose technical
guidance document [external review draft]. (EPA/630/R-00/001). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/raf/publications/benchmark-dose-doc-draft.htm.
U.S. EPA (U.S. Environmental Protection Agency). (2000b). Science policy council handbook:
Risk characterization. (EPA 100-B-00-002). Washington, D.C.: U.S. Environmental
Protection Agency, Office of Research and Development, Office of Science Policy.
http://www.epa.gov/osa/spc/pdfs/rchandbk.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2000c). Supplementary guidance for
conducting health risk assessment of chemical mixtures. (EPA/630/R-00/002).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid=2053 3.
U.S. EPA (U.S. Environmental Protection Agency). (200la). Sources, emission and exposure for
trichloroethylene (TCE) and related chemicals. (EPA/600/R-00/099). Washington, DC:
U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment.
http ://www. epa. gov/nceawww 1 /pdfs/tce/TCEMarch2001. pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes. (EPA/630/P-02/002F). Washington, DC.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=51717.
U.S. EPA (U.S. Environmental Protection Agency). (2003). Neurotoxicity of tetrachoroethylene
(perchloroethylene): Discussion paper. (EPA/600/P-03/005A). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment.
http://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=75193.
U.S. EPA (U.S. Environmental Protection Agency). (2004). Summary report of the peer review
workshop on the neurotoxicity of tetrachloroethylene (perchloroethylene) discussion
paper. (EPA/600/R-04/041). National Center for Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=84351.
U.S. EPA (U.S. Environmental Protection Agency). (2005a). Guidelines for carcinogen risk
assessment. (EPA/630/P-03/001F). Washington, DC.
http://www.epa.gov/cancerguidelines/.
7-66
-------
U.S. EPA (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for
assessing susceptibility from early-life exposure to carcinogens. (EPA/630/R-03/003F).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/cancerguidelines/guidelines-carcinogen-supplement.htm.
U.S. EPA (U.S. Environmental Protection Agency). (2006a). Aging and toxic response: Issues
relevant to risk assessment. (EPA/600/P-03/004A). Washington, DC.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid= 156648.
U.S. EPA (U.S. Environmental Protection Agency). (2006b). A framework for assessing health
risk of environmental exposures to children. (EPA/600/R-05/093F). Washington, DC.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid= 158363.
U.S. EPA (U.S. Environmental Protection Agency). (2006c). Peer review handbook (3rd
edition). (EPA/ 100/B-06/002). Washington, DC: U.S. Environmental Protection Agency,
Science Policy Council, http://www.epa.gov/peerreview/.
U.S. EPA (U.S. Environmental Protection Agency). (2008). Child-specific exposure factors
handbook. (EPA/600/R-06/096F). Washington, DC: U.S. Environmental Protection
Agency, National Center for Environmental Assessment.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid= 199243.
U.S. EPA (U.S. Environmental Protection Agency). (2009). Benchmark dose software (BMDS).
Available online at http://www.epa.gov/NCEA/bmds (accessed December 19, 2009).
U.S. EPA (U.S. Environmental Protection Agency). (2010). Benchmark dose software (BMDS)
(Version 2.1.2 (Final Release)) [Computer Program]. RTF, NC: National Center for
Environmental Assessment. Retrieved from http://www.epa.gov/NCEA/bmds/new.html
U.S. EPA (U.S. Environmental Protection Agency). (201 la). Final PBPK model simulation files
[Database]. Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (201 Ib). Toxicological review of
Trichloroethylene (CASRN 79-01-6) in support of summary information on the
Integrated Risk Information System (IRIS). (EPA/635/R-09/011F). Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (201 Ic). Toxicological review of
Trichloroacetic acid (CAS No. 76-03-9) in support of summary information on the
Integrated Risk Information System (IRIS). (EPA/635/R-09/003F). Washington, DC.
Uhler, AD; Miller, LJ. (1988). Multiple headspace extraction gas chromatography for the
determination of volatile halocarbon compounds in butter. J Agric Food Chem 36: 772-
775. http://dx.doi.org/10.1021/if00082a025.
Umezu, T; Yonemoto, J: Soma, Y; Miura, T. (1997). Behavioral effects of trichloroethylene and
tetrachloroethylene in mice. Pharmacol Biochem Behav 58: 665-671.
http://dx.doi.org/10.1016/S0091-3057(97)00046-4.
Uttamsingh, V: Keller, DA; Anders, MW. (1998). Acylase I-catalyzed deacetylation of N-acetyl-
L-cysteine and S-alkyl-N-acetyl-L-cysteines. Chem Res Toxicol 11: 800-809.
http://dx.doi.org/10.1021/tx980018b.
Vagnarelli, P; De Sario, A; De Carli, L. (1990). Aneuploidy induced by chloral hydrate detected
in human lymphocytes with the Y97 probe. Mutagenesis 5: 591-592.
http://dx.doi.0rg/10.1093/mutage/5.6.591.
Vainio, H; Parkki, MG: Marniemi, J. (1976). Effects of aliphatic chlorohydrocarbons on drug-
metabolizing enzymes in rat liver in vivo. Xenobiotica 6: 599-604.
http://dx.doi.org/10.3109/00498257609151673.
7-67
-------
Valencia, R: Mason, JM: Woodruff, RC: Zimmering, S. (1985). Chemical mutagenesis testing in
Drosophila. III. Results of 48 coded compounds tested for the National Toxicology
Program. Environ Mutagen 7: 325-348. http://dx.doi.org/10.1002/em.2860070309.
Valic, E; Waldhor, T; Konnaris, C: Michitsch, A; Wolf C. (1997). Acquired dyschromatopsia in
combined exposure to solvents and alcohol. Int Arch Occup Environ Health 70: 403-406.
Vamvakas. S: Dekant W: Berthold. K: Schmidt S: Wild. D: Henschler. D. (1987). Enzymatic
transformation of mercapturic acids derived from halogenated alkenes to reactive and
mutagenic intermediates. Biochem Pharmacol 36: 2741-2748.
http://dx.doi.org/10.1016/0006-2952(87)90258-9.
Vamvakas, S: Dekant W: Henschler, D. (1989a). Assessment of unscheduled DNA synthesis in
a cultured line of renal epithelial cells exposed to cysteine S-conjugates of haloalkenes
and haloalkanes. Mutat Res 222: 329-335. http://dx.doi.org/10.1016/0165-
1218(89)90108-0.
Vamvakas. S: Kochling, A: Berthold, K: Dekant W. (1989b). Cytotoxicity of cysteine S-
conjugates: structure-activity relationships. Chem Biol Interact 71: 79-90.
Vamvakas, S: Dekant W: Henschler, D. (1989c). Genotoxicity of haloalkene and haloalkane
glutathione S-conjugates in porcine kidney cells. Toxicol In Vitro 3: 151-156.
http://dx.doi.org/10.1016/0887-2333(89)90058-1.
Vamvakas, S: Herkenhoff M; Dekant W: Henschler, D. (1989d). Mutagenicity of
tetrachloroethene in the ames test: Metabolic activation by conjugation with glutathione.
J Biochem Toxicol 4: 21-27. http://dx.doi.org/10.1002/jbt.2570040105.
Van Hummelen, P; Kirsch-Volders, M. (1992). Analysis of eight known or suspected aneugens
by the in vitro human lymphocyte micronucleus test. Mutagenesis 7: 447-455.
http://dx.doi.Org/10.1093/mutage/7.6.447.
van Lieshout EM; Peters, WH. (1998). Age and gender dependent levels of glutathione and
glutathione S-transferases in human lymphocytes. Carcinogenesis 19: 1873-1875.
van der Gulden, JWJ: Zielhuis, GA. (1989). Reproductive hazards related to perchloroethylene:
A review [Review]. Int Arch Occup Environ Health 61: 235-242.
http://dx.doi.org/10.1007/BF00381420.
van Wijngaarden, E; Hertz-Picciotto, I. (2004). A simple approach to performing quantitative
cancer risk assessment using published results from occupational epidemiology studies.
Sci Total Environ 332: 81-87. http://dx.doi.Org/10.1016/i.scitotenv.2004.04.005.
Vartiainen, T; Pukkala, E: Rienoja, T; Strandman, T; Kaksonen, K. (1993). Population exposure
to tri- and tetrachloroethene and cancer risk: Two cases of drinking water pollution.
Chemosphere 27: 1171-1181. http://dx.doi.org/10.1016/0045-6535(93)90165-2.
Vaughan, TL; Stewart, PA; Davis, S: Thomas, DB. (1997). Work in dry cleaning and the
incidence of cancer of the oral cavity, larynx, and oesophagus. Occup Environ Med 54:
692-695.
Veitch, JM: Murray. GI: Juronen, E: Kilty. CG: Doyle. S: Hawksworth, GM: Rodilla, V. (1997).
Theta-class glutathione S-transferases in human kidney and renal tumours. Biochem Soc
Trans 25: S605.
Verberk, MM; Scheffers, TML. (1980). Tetrachloroethylene in exhaled air of residents near dry-
cleaning shops. Environ Res 21: 432-437.
Verplanke, AJ: Leummens, MH; Herber, RF. (1999). Occupational exposure to tetrachloroethene
and its effects on the kidneys. J Occup Environ Med 41: 11-16.
7-68
-------
Vieira, I; Sonnier, M; Cresteil, T. (1996). Developmental expression of CYP2E1 in the human
liver: Hypermethylation control of gene expression during the neonatal period. Eur J
Biochem 238: 476-483. http://dx.doi.Org/10.llll/i.1432-1033.1996.0476z.x.
Vieira, V: Webster, T; Weinberg, J: Aschengrau, A; Ozonoff, D. (2005). Spatial analysis of lung,
colorectal, and breast cancer on Cape Cod: An application of generalized additive models
to case-control data. Environ Health Global Access Sci Source 4: 11.
http://dx.doi.org/10.1186/1476-069X-4-ll.
Vineis, P; Magnani, C. (1985). Occupation and bladder cancer in males: A case-control study.
Int J Cancer 35: 599-606. http://dx.doi.org/10.1002/iic.2910350506.
Volkel, W; Pahler, A; Dekant, W. (1999). Gas chromatography-negative ion chemical ionization
mass spectrometry as a powerful tool for the detection of mercapturic acids and DNA and
protein adducts as biomarkers of exposure to halogenated olefms. J Chromatogr A 847:
35-46. http://dx.doi.org/10.1016/S0021-9673(99)00022-9.
Volkel W: Friedewald, M: Lederer. E: Pahler. A: Parker. J: Dekant W. (1998).
Biotransformation of perchloroethene: Dose-dependent excretion of trichloroacetic acid,
dichloroacetic acid, and N-acetyl-S-(trichlorovinyl)-L-cysteine in rats and humans after
inhalation. Toxicol Appl Pharmacol 153: 20-27.
http://dx.doi.org/10.1006/taap.1998.8548.
von der Hude, W: Behm, C: Gurtler, R: Easier. A. (1988). Evaluation of the SOS chromotest.
MutatRes 203: 81-94. http://dx.doi.org/10.1016/0165-1161(88)90023-4.
Von Tungeln, LS: Yi. P: Bucci, TJ: Samokyszyn, VM: Chou, MW: Kadlubar. FF: Fu. PP.
(2002). Tumorigenicity of chloral hydrate, trichloroacetic acid, trichloroethanol,
malondialdehyde, 4-hydroxy-2-nonenal, crotonaldehyde, and acrolein in the B6C3F1
neonatal mouse. Cancer Lett 185: 13-19. http://dx.doi.org/10.1016/S0304-
3835(02)00231-8.
Vyskocil, A; Emminger, S; Tejral, J: Fiala, Z; Ettlerova, E; Cermanova, A. (1990). Study on
kidney function in female workers exposed to perchlorethylene. Hum Exp Toxicol 9:
377-380.
Wada, N; Marsman, DS: Popp, JA. (1992). Dose-related effects of the hepatocarcinogen, Wy-
14,643, on peroxisomes and cell replication. Fundam Appl Toxicol 18: 149-154.
Walgren, JE: Kurtz. DT: McMillan. JM. (2000). The effect of the trichloroethylene metabolites
trichloroacetate and dichloroacetate on peroxisome proliferation and DNA synthesis in
cultured human hepatocytes. Cell Biol Toxicol 16: 257-273.
http://dx.doi.Org/10.1023/A:1007638227821.
Walles, SAS. (1986). Induction of single-strand breaks in dna of mice by trichloroethylene and
tetrachloroethylene. Toxicol Lett 31: 31-35. http://dx.doi.org/10.1016/0378-
4274(86)90191-8.
Wang. JL: Chen. WL: Tsai, SY: Sung. PY: Huang. RN. (2001). An in vitro model for evaluation
of vaporous toxicity of trichloroethylene and tetrachloroethylene to CHO-K1 cells. Chem
Biol Interact 137: 139-154. http://dx.doi.org/10.1016/S0009-2797(01)00226-5.
Wang, S: Karlsson, JE: Kyrklund, T: Haglid, K. (1993). Perchloroethylene-induced reduction in
glial and neuronal cell marker proteins in rat brain. Basic Clin Pharmacol Toxicol 72:
273-278. http://dx.doi.0rg/10.llll/i.1600-0773.1993.tb01649.x.
Wanner, M: Lehmann, E: Morel, J: Christen, R. (1982). [The transfer of perchloroethylene from
animal feed to milk]. Mitteilungen aus dem Gebiete der Lebensmitteluntersuchung und
Hygiene 73: 82-87.
7-69
-------
Ward, JM; Reynolds, CW. (1983). Large granular lymphocyte leukemia: A heterogeneous
lymphocytic leukemia in F344 rats. Am J Pathol 111: 1-10.
Ward, JM; Rehm, S: Reynolds, CW. (1990). Pathology of tumours in laboratory animals.
Tumours of the rat. Tumours of the haematopoietic system. IARC Sci Publ625-657.
Ward, JM: Peters, JM: Perella, CM; Gonzalez, FJ. (1998). Receptor and nonreceptor-mediated
organ-specific toxicity of di(2-ethylhexyl)phthalate (DEFIP) in peroxisome proliferator-
activated receptor alpha-null mice. Toxicol Pathol 26: 240-246.
Ward. RC: Travis. CC: Hetrick. DM: Andersen. ME: Gargas. ML. (1988). Pharmacokinetics of
tetrachloroethylene. Toxicol Appl Pharmacol 93: 108-117.
http://dx.doi.org/10.1016/0041-008X(88)90030-0.
Warner, J: Hughes, T; Claxton, L. (1988). Mutagenicity of 16 volatile organic chemicals in a
vaporization technique with Salmonella typhimurium TA100. Environ Mol Mutagen 11:
111-112.
Warner. JA: Miles. EA: Jones. AC: Quint DJ: Colwell. BM: Warner. JO. (1994). Is deficiency
of interferon gamma production by allergen triggered cord blood cells a predictor of
atopic eczema? Clin Exp Allergy 24: 423-430. http://dx.doi.org/10.llll/U365-
2222.1994.tb00930.x.
Warr, T; Parry, E; Parry, J. (1993). A comparison of two in vitro mammalian cell cytogenetic
assays for the detection of mitotic aneuploidy using 10 known or suspected aneugens.
Mutat Res-Fundam Mol Mech Mutagen 287: 29-46. http://dx.doi.org/10.1016/0027-
5107(93)90143-4.
Warren. DA: Reigle. TG: Muralidhara. S: Dallas. CE. (1996). Schedule-controlled operant
behavior of rats following oral administration of perchloroethylene: Time course and
relationship to blood and brain solvent levels. J Toxicol Environ Health 47: 345-362.
http://dx.doi.org/10.1080/009841096161690.
Waskell, L. (1978). A study of the mutagenicity of anesthetics and their metabolites. Mutat Res-
Fundam Mol Mech Mutagen 57: 141-153. http://dx.doi.org/10.1016/0027-
5107(78)90261-0.
Watanabe, K; Satamoto, K; Sasaki, T. (1998). Comparisons on chemically-induced mutation
among four bacterial strains, Salmonella typhimurium TA102 and TA2638, and
Escherichia coli WP2/pKM101 and WP2 uvrA/pKMlOl: Collaborative study II. Mutat
Res 412: 17-31. http://dx.doi.org/10.1016/S1383-5718(97)00155-l.
Webler, T: Brown, HS. (1993). Exposure to tetrachloroethylene via contaminated drinking water
pipes in Massachusetts: A predictive model. Arch Environ Health 48: 293-297.
http://dx.doi.org/10.1080/00039896.1993.9936716.
Weichardt H; Lindner, J. (1975). [Health hazards caused by perchloroethylene in dry cleaning
plants, from the viewpoint of occupational medicine and toxicology]. Gefahrst Reinhalt
Luft 35: 416-420.
Werner, M; Birner, G: Dekant, W. (1996). Sulfoxidation of mercapturic acids derived from tri-
and tetrachloroethene by cytochromes P450 3 A: A bioactivation reaction in addition to
deacetylation and cysteine conjugate beta-lyase mediated cleavage. Chem Res Toxicol 9:
41-49. http://dx.doi.org/10.1021/tx950075u.
Westlind, A; Lofberg, L: Tindberg, N; Andersson, TB; Ingelman-Sundberg, M. (1999).
Interindividual differences in hepatic expression of CYP3A4: relationship to genetic
polymorphism in the 5'-upstream regulatory region. Biochem Biophys Res Commun 259:
201-205.
7-70
-------
Whalan, JE: Redden, JC. (1994). Interim policy for particle size and limit concentration issues in
inhalation toxicity studies. Washington, DC: U.S. Environmental Protection Agency,
Office of Pesticide Products, Health Effects Division.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid= 186068.
Wheeler, M; Kono, H; Yin, M; Nakagami, M; Uesugi, T; Arteel, G: Gabele, E; Rusyn, I;
Yamashina, S: Froh, M; Adachi, Y; limuro, Y; Bradford, B; Smutney, O: Connor, H;
Mason, R; Goyert, S; Peters, J: Gonzalez, F; Samulski, R; Thurman, R. (2001). The role
of Kupffer cell oxidant production in early ethanol-induced liver disease. Free Radic Biol
Med 31: 1544-1549.
White, IN: Razvi, N; Gibbs, AH; Davies, AM; Manno, M: Zaccaro, C; De Mattel s, F; Pahler, A;
Dekant W. (2001). Neoantigen formation and clastogenic action of HCFC-123 and
perchloroethylene in human MCL-5 cells. Toxicol Lett 124: 129-138.
http://dx.doi.org/10.1016/S0378-4274(00)00281-2.
WHO (World Health Organization). (1978). A co-operative trial in the primary prevention of
ischaemic heart disease using clofibrate. Report from the Committee of Principal
Investigators. British Heart Journal 40: 1069-1118.
WHO (World Health Organization). (1980). W.H.O. cooperative trial on primary prevention of
ischaemic heart disease using clofibrate to lower serum cholesterol: mortality follow-up.
Report of the Committee of Principal Investigators. Lancet 2: 379-385.
WHO (World Health Organization). (1984). WHO cooperative trial on primary prevention of
ischaemic heart disease with clofibrate to lower serum cholesterol: final mortality follow-
up. Report of the Committee of Principal Investigators. Lancet 2: 600-604.
WHO (World Health Organization). (2006). Concise international chemical assessment
document 68: Tetrachloroethene. Geneva, Switzerland: World Health Organization,
International Programme on Chemical Safety.
http://www.inchem.org/documents/cicads/cicads/cicad68.htm.
Wilcox, AJ: Horney, LF. (1984). Accuracy of spontaneous abortion recall. Am J Epidemiol 120:
727-733.
Williams. RL: Creasy. RK: Cunningham. GC: Hawes. WE: Norris. FD: Tashiro. M. (1982).
Fetal growth and perinatal viability in California. Obstet Gynecol 59: 624-632.
Wilson. R: Donahue. M: Gridlev. G: Adami. J: El Ghormli. L: Dosemeci. M. (2008). Shared
occupational risks for transitional cell cancer of the bladder and renal pelvis among men
and women in Sweden. Am J Ind Med 51: 83-99. http://dx.doi.org/10.1002/ajim.20522.
Windham, GC: Shusterman, D; Swan, SH; Fenster, L; Eskenazi, B. (1991). Exposure to organic
solvents and adverse pregnancy outcome. Am J Ind Med 20: 241-259.
http://dx.doi.org/10.1002/aiim.4700200210.
Windham, GC: Zhang, L: Gunier, R: Croen, LA; Grether, JK. (2006). Autism spectrum disorders
in relation to distribution of hazardous air pollutants in the San Francisco Bay area.
Environ Health Perspect 114: 1438-1444.
Wolf CJ: Takacs, ML; Schmid, JE: Lau, C: Abbott BD. (2008). Activation of mouse and human
peroxisome proliferator-activated receptor alpha by perfluoroalkyl acids of different
functional groups and chain lengths. Toxicol Sci 106: 162-171.
http://dx.doi.org/10.1093/toxsci/kfnl66.
Woods, CG: Burns, AM; Maki, A; Bradford, BU: Cunningham, ML; Connor, HP; Kadiiska,
MB; Mason, RP; Peters, JM; Rusyn, I. (2007). Sustained formation of alpha-(4-pyridyl-l-
oxide)-N-tert-butylnitrone radical adducts in mouse liver by peroxisome proliferators is
7-71
-------
dependent upon peroxisome proliferator-activated receptor-alpha, but not NADPH
oxidase. Free Radic Biol Med 42: 335-342.
http://dx.doi.0rg/10.1016/j.freeradbiomed.2006.10.053.
Xu, W; Adler, I. (1990). Clastogenic effects of known and suspect spindle poisons studied by
chromosome analysis in mouse bone marrow cells. Mutagenesis 5: 371-374.
Yang, 0: Ito, S: Gonzalez, FJ. (2007). Hepatocyte-restricted constitutive activation of PPAR
alpha induces hepatoproliferation but not hepatocarcinogenesis. Carcinogenesis 28: 1171-
1177. http://dx.doi.org/10.1093/carcin/bgm046.
Yang. Q: Nagano. T: Shah. Y: Cheung. C: Ito. S: Gonzalez. FJ. (2008). The PPAR alpha-
humanized mouse: a model to investigate species differences in liver toxicity mediated by
PPAR alpha. Toxicol Sci 101: 132-139. http://dx.doi.org/10.1093/toxsci/kfm206.
Yllner, S. (1961). Urinary metabolites of 14C-tetrachloroethylene in mice. Nature 191: 820.
Yokley, KA; Evans, MV. (2007). An example of model structure differences using sensitivity
analyses in physiologically based pharmacokinetic models of trichloroethylene in
humans. Bull Math Biol 69: 2591-2625. http://dx.doi.org/10.1007/sll538-007-9233-x.
Yoo, JS: Guengerich, FP: Yang, CS. (1988). Metabolism of N-nitrosodialkylamines by human
liver microsomes. Cancer Res 48: 1499-1504.
Yoon. JS: Mason. JM: Valencia. R: Woodruff. RC: Zimmering. S. (1985). Chemical
mutagenesis testing in Drosophila. IV. Results of 45 coded compounds tested for the
National Toxicology Program. Environ Mutagen 7: 349-367.
http://dx.doi.org/10.1002/em.2860070310.
Yoshioka, T; Krauser, JA: Guengerich, FP. (2002). Tetrachloroethylene oxide: hydrolytic
products and reactions with phosphate and lysine. Chem Res Toxicol 15: 1096-1105.
http://dx.doi.org/10.1021/tx020028i.
Youssef, J: Badr, M. (1998). Extraperoxisomal targets of peroxisome proliferators:
Mitochondrial, microsomal, and cytosolic effects: Implications for health and disease.
Crit Rev Toxicol 28: 1-33. http://dx.doi.org/10.1080/10408449891344182.
Yu. S: Cao. WQ: Kashireddv. P: Meyer. K: Jia. Y: Hughes. DE: Tan. Y: Feng. J: Yeldandi. AV:
Rao, MS: Costa, RH: Gonzalez, FJ: Reddy, JK. (2001). Human peroxisome proliferator-
activated receptor alpha (PPARA ) supports the induction of peroxisome proliferation in
PPARA -deficient mouse liver. J Biol Chem 276: 42485-42491.
Zeka, A; Zanobetti, A; Schwartz, J. (2006). Individual-level modifiers of the effects of
particulate matter on daily mortality. Am J Epidemiol 163: 849-859.
Zheng, T: Cantor, KP: Zhang, Y: Lynch, CF. (2002). Occupation and bladder cancer: A
population-based, case-control study in Iowa. J Occup Environ Med 44: 685-691.
Zhou, S: Wallace, K. (1999). The effect of peroxisome proliferators on mitochondrial
bioenergetics. Toxicol Sci 48: 82-89.
Zhou, YC: Waxman, DJ. (1998). Activation of peroxisome proliferator-activated receptors by
chlorinated hydrocarbons and endogenous steroids. Environ Health Perspect 106 Suppl 4:
983-988.
Zielhuis, GA: Gijsen, R: van der Gulden, JWJ. (1989). Menstrual disorders among dry-cleaning
workers [Letter]. Scand J Work Environ Health 15: 238.
Zingg, JM; Jones, PA. (1997). Genetic and epigenetic aspects of DNA methylation on genome
expression, evolution, mutation and carcinogenesis. Carcinogenesis 18: 869-882.
7-72
-------
Zordan, M; Osti, M; Pesce, M; Costa, R. (1994). Chloral hydrate is recombinogenic in the wing
spot test in Drosophila melanogaster. Mutat Res 322: 111-116.
http://dx.doi.org/10.1016/0165-1218(94)00017-4.
Zytner, R; Biswas, N; Bewtra, JK. (1989). Adsorption and desorption of perchloroethylene in
soils, peat moss, and granular activated carbon. Can J Civ Eng 16: 798-806.
http://dx.doi.org/10.1139/189-123.
7-73
-------
APPENDIX A. EPA RESPONSE TO MAJOR EXTERNAL PEER-REVIEW AND
PUBLIC COMMENTS
The 2008 external review draft (ERD) of EPA's Toxicological Review of
Tetrachloroethylene (Perchloroethylene) underwent a formal external peer review in accordance
with U.S. Environmental Protection Agency (EPA) guidance on peer review (U.S. EPA, 2006c).
The external peer review was performed by the National Research Council (NRC). NRC was
tasked with evaluating the adequacy of the EPA assessment, the data and methods used for
deriving the noncancer values for inhalation and oral exposures and the oral and inhalation
cancer unit risks posed by tetrachloroethylene; whether the key studies underlying the draft IRIS
assessment are of requisite quality, reliability, and relevance to support the derivation of the
reference values and cancer risks; and whether the uncertainties in EPA's risk assessment were
adequately described and, where possible, quantified. The major peer review comments below
(Sections A.I - A.3) include all comments from the Summary section (with introductory or
background text omitted), and several additional comments that potentially have a major impact
on the revision of the ERD, and are quoted verbatim from NRC (2010). Page numbers for each
quotation are also noted.
In addition, Sections A.4 and A.5 contain a summary of public comments and EPA's
responses. In revising the ERD Toxicological Review of Tetrachloroethylene, EPA utilized a
harmonized PBPK model that was developed in response to NRC (2010) recommendations.
EPA conducted a focused peer review on the application of the harmonized model to support its
use in the final document. The focused peer review report is publicly available
(www.epa.gov/iris). A summary of the focused peer review comments and EPA's responses can
be found in Section A.6.
A.I. Major NRC Introductory Comments and EPA Response
NRC Comment: [pp 3-4] The committee appreciates the extensive work that EPA has invested
in the development of its draft assessment of tetrachloroethylene. However, the committee has
identified concerns about some of the approaches that EPA used to evaluate the data on
tetrachloroethylene and subjects about which inadequate information or rationales are used to
support its risk assessment—factors that call into question the soundness and reliability of EPA's
proposed reference values and cancer risk estimates for tetrachloroethylene. One of the
overarching weaknesses of the draft assessment was a lack of critical analysis of the data on
which EPA relied in evaluating methodologic strengths and weaknesses. That lack was
particularly evident in the assessment of the epidemiologic data: study selection and conclusions
A-l
-------
appeared to be based heavily on results that showed positive associations, and other data and the
strengths and weaknesses of the selected studies were not adequately taken into consideration.
The committee observed similar problems in its review of EPA's evaluation of the genotoxicity
evidence, in which preference appeared to be given to studies that reported positive results.
Specifically, EPA did not analyze studies critically with respect to their methodologic strengths
and weaknesses, nor did it organize its discussion clearly to provide an integrated consideration
of the weight of evidence on the genotoxicity of tetrachloroethylene. Other mode of action
evaluations were also hampered in this way.
EPA Response: EPA agrees that a balanced critical analysis of the data is necessary, and has
significantly revised its assessment to make its evaluation of study methodological strengths
and weaknesses more organized and transparent, and to ensure impartial consideration of all
pertinent studies. Specific changes with respect to evaluation of epidemiologic data (A.3.1),
genotoxicity (A.3.6), and modes of action (A.3.4) are described in the more detailed
responses below.
A.2. Noncancer Assessment
A.2.1. Major NRC Comments on "Critical Noncancer End Point and Studies" and
EPA Responses
NRC Comment: [pp 4-5] The committee found that EPA adequately supported its selection of
neurotoxicity as the critical effect on which to base the RfC and RfD. The draft IRIS document
illustrates that neurotoxic effects are the most sensitive effects of tetrachloroethylene and that
reference values based on neurotoxic effects would be protective against other noncancer effects
that occur at higher concentrations.
EPA Response: EPA accepts these NRC recommendations, and continues to rely on
neurotoxicity as the critical effect (Section 5.1.1.1).
NRC Comment: [p 5] EPA provides descriptions of the relevant neurotoxicity studies, but its
evaluation of the epidemiologic literature could be improved by providing a critical evaluation of
the validity of study designs and evaluation of the methods used for data collection and analysis,
which the committee judges to be most important in selecting key studies.
EPA Response: EPA accepts these NRC recommendations. The rationale for selecting
principal studies of neurotoxicity has been more fully and transparently articulated (see
Section 5.1.1). Study strengths and weaknesses are judged according to the recommended
criteria (e.g., study populations, exposure durations, quality of neurotoxicological tests and
A-2
-------
exposure measurements) (see Table 5.1). EPA has also strengthened the presentation of
human and animal studies and reorganized them by the toxicological endpoint, particularly
(1) neurobehavior, (2) neurophysiology, (3) brain pathology, and (4) developmental
neurotoxicity. As also suggested by the NRC, the developmental neurotoxicity studies are
grouped together in one section, and a more robust discussion of these studies is provided.
EPA focuses on the neurotoxic effects (including developmental neurotoxicity) observed in
studies of tetrachloroethylene, and does not comprehensively review the neurotoxicity of
structurally related solvents. The MOA discussion for neurotoxic effects (see Section 4.1.3)
addresses mechanistic commonalities with other volatile organic solvents and alcohols but
likewise focuses on tetrachloroethylene. Hypothesized mechanisms for the different
neurological domains affected by tetrachloroethylene are addressed, as are potential
molecular targets. EPA addresses recent animal studies identified by NRC (Boyes et al.,
2009; Oshiro et al., 2008) and also includes three new epidemiological studies published or
available since the release of the 2008 ERD of EPA's Toxicological Review of
Tetrachloroethylene (Perchloroethylene), including the final peer-reviewed report of New
York State Department of Health study [NYSDOH (2010): published by Storm et al. (2011)1
that was presented to NRC during the committee deliberations.
NRC Comment: [p 5] EPA chose the study by Altmann et al. (1995) as the critical one for
determining the RfC and RfD because it involved an environmental exposure and used a
standardized computer-assisted testing battery. Those are reasonable bases for the choice, but
they do not outweigh methodologic deficiencies that seriously compromised the results of the
study. Most important, the referent group was not appropriate. The group had more education
than the exposed group and appeared to have pre-existing differences in cognitive abilities,
which could account for its better test results. Evidence of residual confounding by education
can be seen in the variability in reported results. For example, there was no association between
tetrachloroethylene and visual evoked potentials; this is important because changes in the visual
system and abnormalities in visual evoked potentials have been associated with
tetrachloroethylene and other related solvents, and they are essentially unrelated to education.
Other limitations of the study included the lack of a rationale for initial selection of study
subjects, inadequacy of exposure characterization, and lack of a dose-response relationship.
Finally, even though the test battery was performed properly, some of the tests have not been
well validated with regard to what they reveal about brain damage.
Thus, the committee disagrees with EPA's selection of the Altmann et al. (1995) study as
the basis of its risk calculations.
A-3
-------
EPA Response: EPA acccepts these NRC recommendations. In particular, based on
application of the criteria described above for conducting its critical review (see Table 5.2),
EPA acknowledges the limitations of Altmann et al. (1995) identified by NRC, and as
indicated in Section 5.1.1, relies on other studies as the basis for determining the RfC and
RfD (see next response).
NRC Comment: [p 5] In reviewing the database, the committee gave greater weight to studies
that had the strongest methods; it neither chose nor excluded studies on the basis of their results.
The set of studies that the committee judged to be more appropriate for supporting the RfC and
RfD include those of Altmann et al. (1990). Cavalleri et al. (1994), Gobba et al. (1998),
Echeverria et al. (1995), and Boyes et al. (2009).
EPA Response: EPA modified its approach based on these recommendations. EPA's
revised assessment relies on two of the chronic neurotoxicity studies recommended by the
NRC, Echeverria et al. (1995) and Cavalleri et al. (1994). Acute studies recommended for
consideration by NRC [the chamber study of Altmann et al. (1990) and the rodent study of
Boyes et al. (2009)1 are supportive, but were not selected for POD derivation because EPA
gives preference to quality studies of chronic human exposures for reference value
derivation. Gobba et al. (1998) evaluated the same cohort as Cavalleri et al. (1994), and the
earlier study was preferred due to its use of a control group and the clearer identification of a
POD (refer to section 5.1.1.3.2).
A.2.2. Major NRC Comments on "Derivation of Reference Values" and EPA
Responses
NRC Comment: [p 5] EPA derived sample inhalation reference values by using results from
several supporting neurotoxicity studies for comparison with its principal study by Altmann et al.
(1995). The committee found that some uncertainty factors (UFs) were applied inconsistently;
specifically, the application of the uncertainty factor to account for subchronic exposures in
epidemiologic studies should be justified better. In some cases, EPA did not use such a factor; in
other cases, it applied a value of 10 with weak justification.
EPA Response: EPA accepts these NRC recommendations, and has provided more thorough
justification for the selection of all UFs (Sections 5.1.3 and 5.2.3). With respect to the UF to
account for subchronic exposures, because each of the selected studies was of chronic
exposure, EPA did not apply this UF to any of the PODs. Comments regarding other UFs
are discussed in the response to comments that follow.
A-4
-------
NRC Comment: [Chapter 10, p 90] A factor of 10 was used consistently by EPA when a
lowest observed-adverse-effect level (LOAEL) from a study was used instead of a no-observed-
adverse-effect level (NOAEL). That is consistent with EPA policy. A benchmark dose (BMD)
can be treated as a NOAEL, but no studies of neurotoxicity that could support a BMD
calculation had been published when the draft was written. More recent studies of neurotoxicity
would support such a calculation (Benignus et al., 2009; Boyes et al., 2009; Oshiro et al., 2008).
EPA Response: EPA accepts these NRC comments, and has retained the value of 10 for the
UFL(LOAEL-to-NOAEL extrapolation) (Sections 5.1.3 and 5.2.3). Regarding studies of
neurotoxicity that would support dose-response modeling, NRC noted the lack of suitable
studies at the time of the 2008 external review draft, and recommended consideration of
more recent studies of neurotoxicity by Oshiro et al. (2008), Benignus et al. (2009) and
Boyes et al. (2009). The studies of Oshiro et al. (2008) and Boyes et al. (2009) are in rats
and covered shorter, acute exposure duration periods than the available human studies and
require extrapolation of animal observations to humans. The Benignus et al. (2009) analysis
of tetrachloroethylene included three acute exposure studies, the two rat studies above and
the acute-exposure study in humans of Altmann et al. (1992). While subjects in Altmann et
al. (1992) could serve as their own controls, there was not an unexposed group. Further,
these three studies were not considered principal studies given the availability of suitable
human data from chronic exposures.
NRC Comment: [Chapter 10, pp 90-91] The uncertainty factor for extrapolating animal data to
humans is considered to have toxicokinetic and toxicodynamic aspects. EPA judged that an
uncertainty factor of 3 was adequate to address these uncertainties. EPA applied that approach
consistently, but the rationale for doing so was not adequately described. Specifically, the draft
cites an EPA (1994) document, but it would have enhanced transparency if it summarized briefly
why an uncertainty factor of 3, rather than the default factor of 10, was used.
EPA Response: EPA accepts this NRC recommendation. With respect to the UF for
interspecies extrapolation from animals (UFA), because each of the selected studies was in
humans, EPA did not apply this UF to any of the PODs. For the "sample" RfCs and RfDs,
where some animal studies were used, it is explained that the PODs are expressed as human
equivalent concentrations, so the UF of 3 is applied to account for potential
pharmacodynamic differences (Sections 5.1.3 and 5.2.3).
NRC Comment: [Chapter 10, p 91] The application of a default factor of 10 to account for
interindividual variation is justified because of the paucity of data on sensitive populations,
A-5
-------
including developing and aging organisms. Its use is appropriate and in accordance with EPA
guidance.
EPA Response: EPA accepts these NRC comments, and has retained the value of 10 for the
uman variability) (Sections 5.1.3 and 5.2.3).
NRC Comment: [Chapter 10, p 92] In the derivation of RfCs on the basis of neurotoxicity,
EPA used a factor of 3 for database deficiencies because of the inadequacy of the experimental
literature designed to characterize hazard and dose-response. Key deficiencies identified were
inadequate data to address childhood or other life-stage susceptibility, a paucity of animal studies
(especially studies of developing animals and of chronic, low-level exposures) designed to
investigate neurotoxicity or to define and characterize dose-response relationships, and
inadequate database on cognitive testing. It was unclear whether a factor of 3 was adequate to
address these uncertainties because there was some overlap with the factor of 10 applied for
human variation, which also addressed developmental concerns.
The committee recommends that EPA revisit and defend more clearly its decision to
apply a factor of 3 for database deficiencies in light of new data and the committee's findings in
Chapter 3. New studies include, for example, recent papers from researchers in EPA's National
Health and Environmental Effects Research Laboratory provide excellent data from well-
designed studies using controlled, acute exposures that link deficits in visual function and signal
detection with atmospheric tetrachlorethylene concentrations and instantaneous concentrations in
the brain. This includes papers by Oshiro et al. (2008) and Boyes et al. (2009) investigating
function and by Shafer et al. (2005) on mechanisms, which is described in the IRIS document but
not fully integrated. These studies link neural or behavioral effects to actual brain concentrations
of tetrachloroethylene or to their estimated concentration using PBPK modeling. Thus, the
animal literature on controlled acute exposure is now stronger. Notable gaps in the animal
literature still include the paucity of studies of developmental or chronic exposures. Another
consideration is that the committee found the human study of exposed children (Schreiber et al.,
2002) to be methodologically flawed. The committee judges these to be serious gaps in the
database, which suggests that a factor of 3 may be inadequate to account for database
deficiencies.
EPA Response: EPA accepts these NRC recommendations. Based on concerns raised by
the NRC, EPA re-examined the adequacy of the database and increased the UFD from 3 to 10
(Sections 5.1.3 and 5.2.3). EPA's application of a UFo of 10 to address the lack of data to
adequately characterize the hazard and dose response in the human population is consistent
with EPA's A Review of the Reference Dose and Reference Concentration (U.S. EPA, 2002).
EPA provides scientific justification for choosing this UFo in Section 5 where the reference
A-6
-------
concentration (Section 5.1.3) and dose (Section 5.2.3) are derived. EPA's justification is
based on a number of data gaps identified from both the human and animal literature.
Regarding neurotoxicity, animal studies of chronic exposures (including in developing
animals) examining sensitive neurotoxic endpoints are lacking. Moreover, the most sensitive
neurotoxic endpoint associated with tetrachloroethylene exposure in humans—decrement in
visual contrast sensitivity—was identified in residential studies that were judged to be limited
for developing an RfC (Storm etal.. 2011 [previously reported in NYSDOH. 20101:
Schreiber et al., 2002; Altmann et al., 1995). This specific endpoint was not evaluated in any
of the occupational studies used for developing the RfC. Regarding sensitive endpoints other
than neurotoxicity, the available human and animal studies of immunologic and hematologic
toxicity [e.g., Emara et al. (2010): Marth (1987)1 are limited.
NRC Comment: [p 8] The committee derived candidate values by using the same studies as
EPA and additional studies. The committee found that the reference values from the strongest
studies were in the range of 6-50 ppb (or 0.04-0.34 mg/m3). That range is higher than the RfC of
0.016 mg/m3 derived by EPA and is further supported when considered in the context of the full
database (see further discussion below).
EPA Response: EPA revisited the above calculation based on NRC's annotation that their
exercise was illustrative, and that some candidate values were subject to change based on
implementation of their advice regarding the UFs. As discussed above, due to concerns
raised by the NRC, EPA increased the UFD from 3 to 10. With this change, the NRC-
suggested range would be lowered to 0.01-0.10 mg/m3, which fully encompasses EPA's
revised range of candidate RfCs (0.02 to 0.06 mg/m3) and EPA's selected RfC (0.04 mg/m3)
(see Section 5.1.3).
NRC Comment: [p 8] EPA extrapolated the results of inhalation studies to derive the oral RfD
for tetrachloroethylene. Physiologically based pharmacokinetic (PBPK) modeling was used to
support the route-to-route extrapolation. The rationale behind that approach is sound and
adequately explained by EPA, and the choice of dose metric (blood area-under-the-curve) was
appropriate and adequately supported by the available evidence. However, the three models
used by EPA were formulated and validated with data from inhalation exposures—none was
validated against blood concentrations that result from oral exposure. EPA empirically assumed
a value for the rate of oral absorption of tetrachloroethylene; this assumption is inferior to direct
estimation. Other PBPK models that use direct estimation are available, and their use may help
to reduce the uncertainty in the assumed values; or additional PBPK models could be developed
(see recommendation below for a harmonized PBPK model).
A-7
-------
EPA Response: EPA accepts these NRC recommendations. EPA followed the NRC
recommendations and developed a new harmonized PBPK model that incorporated available
oral data from which the oral absorption rate could be estimated (Section 3.5) (Chiu and
Ginsberg, 2011), which was used in the route-to-route extrapolations for the RfD calculations
(Section 5.2.2). The response to recommendations with respect to PBPK modeling is
discussed in more detail below (Section A.3.7).
A.2.3. Major NRC Comments on "Graphical Presentation" and EPA Responses
NRC Comment: [p 8] EPA provides graphical comparisons of reference values, values that
could be derived from supporting studies. Reference values derived from neurotoxicity data are
presented, as are values based on other noncancer effects to illustrate dose dependence of
multiple forms of observed toxicity. Overall, the committee supports the approach of presenting
the evidence in this visual format. However, the committee recommends some revisions to
improve illustration of the uncertainties being represented and to expand the presentation to
include the larger body of literature on a particular end point to show how the RfC compares
with sample reference values derived from studies that are methodologically sound but not
judged to be critical for the RfC. Consistency between the RfC and such studies would provide
additional support.
Figure S-l provides an example illustration developed by the committee. It shows that
the majority of sample values is centrally clustered, but there is a wide spread at the lower and
higher ends. The overall range of the 19 sample reference values is 0.03-333 ppb (0.0002-2.6
mg/m3), but the range is reduced to about 6-50 ppb (0.04-0.34 mg/m3) when consideration is
restricted to the five strongest studies. The RfC of 0.016 mg/m3 calculated by EPA on the basis
of the Altmann et al. (1995) study falls below the range. The figure shows that sample reference
values that could be derived from the full database of neurotoxicity studies provide some support
for the range.
EPA Response: EPA accepts these NRC recommendations. In particular, EPA agrees that
the graphical presentation of studies and resulting risk values is useful. EPA graphically
portrayed the PODs for all the tetrachloroethylene neurotoxicity studies considered for dose-
response analysis, with the studies used to derive candidate reference values highlighted (see
Section 5.1, Figure 5-1). Separately, for the studies used to derive candidate reference
values, EPA graphically presented the PODs and uncertainty factors used in the derivation of
the candidate noncancer RfCs (see Section 5.1, Figure 5-2). Additionally, in agreement with
NRC, EPA continues to provide "sample" reference values based on reproductive and
developmental, kidney, liver, immunological, and hematological noncancer endpoints.
A-8
-------
Sample PODs and composite UFs for noncancer effects other than the critical effect of
neurotoxicity are also graphically displayed (see Section 5.1, Figure 5-3), in accordance with
the NRC recommendations.
A.2.4. Major NRC Comments on "Reproductive and Developmental Effects" and EPA
Responses
NRC Comment: [Chapter 4, p 48] EPA's identification of the key animal and epidemiologic
reproductive and developmental studies of tetrachloroethylene appears to be complete, but the
committee recommends some reorganization and reconsideration of data to provide a more
transparent and balanced characterization of the data.
EPA Response: EPA accepts this NRC recommendation, and has made revisions throughout
Section 4.7. Consistent with NRC advice, the presentation of developmental and
reproductive toxicity studies was reordered, and developmental studies were separated from
reproduction studies, to emphasize the differences in exposure paradigm and types of
endpoints assessed. Study strengths and deficiencies are presented in the individual study
descriptions. Evidence from supportive in vitro and in vivo studies and the consistency of
outcomes across species and protocols are described. Findings of parental (including
maternal) toxicity and the treatment levels at which those effects were observed are also
described for each study.
NRC Comment: [Chapter 4, p 48] The committee agrees with the selection of the Tinston
(1994) two-generation reproductive-toxicity study and the Carney et al. (2006) developmental -
toxicity study as supportive of a point of departure and an RfV [reference value]. EPA's
derivation of a comparative RfV based on the developmental toxicity of tetrachloroethylene is an
important contribution to the tetrachloroethylene database.
EPA Response: EPA accepts this recommendation, and has included these studies in
developing comparative "sample" RfCs and RfDs (Sections 5.1.4 and 5.2.4).
NRC Comment: [Chapter 4, pp 48-49] However, the committee recommends that EPA revise
the chapter to address the specific deficiencies discussed above regarding information presented
on the animal reproductive and developmental studies. In particular, the revision should include:
(1) a critical analysis of the described studies, including an assessment of the relationship of
maternal toxicity to developmental toxicity and the strengths, limitations, and consistency of the
various study results; (2) characterization of maternal toxicity (e.g., mild or severe) associated
with the studies listed in Table 4-10 and use of consistent nomenclature (ppm or mg/m3) for
listing tetrachloroethylene concentrations; (3) the scientific basis for selecting the Tinston (1994)
A-9
-------
and Carney et al. (2006) studies as supportive of an RfV; (4) the scientific rationale for selecting
the Tinston (1994) study instead of the Carney et al. (2006) study for derivation of the
comparative RfV; (5) information on the mode of action for tetrachloroethylene-induced
developmental toxicity which addresses the apparent contradictions raised in the committee's
review that TCA may be the causative agent; and (6) characterization of the evidence for
tetrachloroethylene-induced reproductive and developmental toxicity in animals based on EPA
risk assessment guidelines. Stating explicitly whether the animal evidence is sufficient or
insufficient for these important end points will help risk managers and others to more readily
identify and protect against potential adverse health effects. It will also help to identify data gaps
in the tetrachloroethylene database.
EPA Response: EPA has made revisions based on the NRC recommendations throughout
Section 4.7. In particular, EPA accepts Recommendation (1), and has revised the individual
study descriptions to include the critical analysis elements noted. Regarding (2), EPA
clarified that there is no evidence in the tetrachloroethylene mammalian developmental or
reproductive toxicity study database of severe maternal toxicity that compromised or
confounded the evaluation of offspring toxicity, noting the difficulty in determining the
relationship between maternal and developmental toxicity in a developmental or reproductive
toxicity study. EPA accepts Recommendations (3) and (4) and has revised the basis and
rationale for selecting studies in Section 5.1. EPA accepts Recommendation (5) and has
expanded the discussion of the MOA hypotheses for developmental outcomes to address the
potential involvement of the metabolite TCA. EPA accepts Recommendation (6), and in
accordance with EPA risk assessment guidelines for reproductive and developmental
toxicity, explicitly states that the database of animal and human studies was sufficient for the
evaluation of developmental and reproductive toxicity.
NRC Comment: [Chapter 4, p 49] In addition to revising the chapter, the committee also
recommends that EPA consider conducting a bench-mark dose analysis and deriving an RfV
based on the Carney et al. (2006) study in addition to, or instead of, the Tinston (1994) study.
This will address the potential confounding effects of maternal toxicity at the 1,000 ppm
exposure level observed in the Tinston (1994) study.
EPA Response: Endpoints from both Tinston (1994) and Carney et al. (2006)—as well as
from Beliles et al. (1980) and Nelson et al. (1979)—were carried forward for potential RfC
development (see Table 4-49). Sample RfCs were derived for reproductive and
developmental effects (see Table 5-7) and were one order of magnitude greater than the
candidate RfCs derived for neurological effects (see Table 5-3). The possible—but
A-10
-------
uncharacterized—influence of maternal toxicity on offspring outcomes at the highest dose
tested (1,000 ppm) in the Tinston (1994) study would have no impact on the final noncancer
reference value derivation.
A.3. Cancer Assessment
A.3.1. Major NRC Comments on "Epidemiologic Evidence Pertaining to Cancer" and
EPA Responses
NRC Comment: [Chapter 9, p 85] One of the biggest difficulties in assessing the cogency of
the EPA's assessment related to cancer is how the data are organized in the tables and some parts
of the text. It would be much easier to evaluate the overall picture of results regarding
tetrachloroethylene and a particular cancer if the tables were organized by cancer type as
opposed to the current format, which organizes them by study design. The current format
requires the reader to jump between sections for cohort mortality, incidence, and case-control
studies. Studies are sometimes further categorized as to the type of worker included (for
example, dry-cleaner vs degreaser); this makes it extremely difficult to evaluate the overall
consistency or lack of consistency in results related to specific cancers.
EPA Response: EPA accepts these NRC recommendations. EPA has significantly
reorganized the data presentation by type of cancer as follows: kidney and bladder toxicity
and cancer (see Section 4.2); liver toxicity and cancer (see Section 4.3); esophageal cancer
(see Section 4.4); lung and respiratory cancer (see Section 4.5); immunotoxicity, hematologic
toxicity, and cancers of the immune system (see Section 4.6); developmental and
reproductive toxicity, and reproductive cancers (see Section 4.7). Epidemiologic
observations on tetrachloroethylene and breast cancer are included in Section 4.7.
NRC Comment: [Chapter 9, p 85] Errors in reporting results also occur occasionally. For
example, the draft reports (on page 4-150, lines 1-3), in relation to Hodgkin disease, "a
statistically significantly elevated risk for male [sic] with a job title of dry cleaner or laundry
worker (Costantini et al., 2001)." The result from Costantini et al. (2001) for that group in
relation to Hodgkin disease was an OR of 2.5 (95% CI, 0.3-24.6), which is not significant and
was based on a single case.
EPA Response: EPA accepts this recommendation and has corrected reporting errors.
NRC Comment: [Chapter 9, p 85] The overall impression is that data are presented to support
a positive association between tetrachloroethylene and cancer and that studies that found no such
A-ll
-------
association are criticized or minimized. EPA should provide a clearer discussion of criteria used
to identify studies of merit and a more balanced critique to strengthen the draft IRIS assessment.
EPA Response: In agreement with NRC, EPA has also now included an updated and more
balanced evaluation of the epidemiologic literature on tetrachloroethylene and cancer
(throughout Section 4, as described above). The revised discussion of epidemiologic
observations achieves a more balanced review by clarifying how EPA considered study
methodological strengths and weaknesses, including evaluation of the exposure-assessment
approach, study size, number of observed cancer events, choice of referent population,
presence or absence of an exposure-response relationship, and the potential for alternative
explanations such as chance, bias, or confounding. A synthesis of the epidemiologic cancer
data is provided considering evidence across cancer sites. Recent literature added to the
epidemiologic evaluation comprised 27 epidemiologic studies on occupational
tetrachloroethylene exposure and cancer, and one meta-analysis of bladder cancer and dry
cleaning. These studies were published since 2004, the date of the comprehensive literature
review in support of the 2008 ERD of EPA's lexicological Review of Tetrachloroethylene
(Perchloroethylene). As a supplement to the tabular summaries of epidemiologic
observations organized by cancer site in Section 4, Appendix B characterizes the design and
methods more fully.
A.3.2. Major NRC Comments on "Cancer Characterization" and EPA Responses
NRC Comment: [p 8] EPA classified tetrachloroethylene as "likely to be carcinogenic to
humans." The committee reviewed the classification guidance in EPA's 2005 Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005) and the bioassay data available on
tetrachloroethylene and concluded that EPA adequately documented that its classification has
been based on the results of bioassay s that found increased incidences of hepatocellular tumors,
mononuclear-cell leukemia (MCL), renal tumors, and hemangiosarcomas in laboratory animals
and to a lesser extent on epidemiologic evidence. EPA's decision to characterize
tetrachloroethylene as likely to be a human carcinogen as opposed to "carcinogenic to humans"
appropriately reflects the possibility that there are deficiencies or potential inaccuracies in
interpretation of the data. Some of the possible deficiencies and inaccuracies are discussed
below for each of the datasets.
EPA Response: EPA accepts these NRC recommendations, and continues to characterize
tetrachloroethylene as "likely to be carcinogenic to humans" (Section 4.10.3). EPA agrees
with NRC that the epidemiologic literature on cancers provides limited evidence that
tetrachloroethylene is carcinogenic in humans.
A-12
-------
NRC comments and responses regarding individual animal bioassay datasets are
discussed below.
NRC Comment: [pp 8-9] Mononuclear-Cell Leukemia
An increased incidence of MCL in F344 rats has been reported in two bioassays. The
biologic significance of the increases was debated by the committee because increases were
observed in only one strain of rat, which is known to have a high background incidence of MCL,
and because MCL's relevance to humans and the mode of action of tetrachloroethylene causing
it are not understood. In considering the high background of MCL, the committee found a
published assessment by Thomas et al. (2007) that applied statistical approaches (life-table
analyses) to bioassays of the National Toxicology Program (NTP) to interpret dose-response
relationships. Tetrachloroethylene was one of five chemicals of 500 tested by NTP that showed
statistically significant increases in MCL in both male and female rats despite the high
background rates. The publication advocated that such statistical evidence be supported with a
weight-of-evidence analysis of biologic data before conclusions were drawn.
The committee found some support from epidemiologic studies that suggested an
association between tetrachloroethylene and lymphoma, but the data were relatively weak and
inconsistent. A difficulty in interpreting the findings is a difference of opinion about the human
relevance of MCL. Some committee members judged that similarities between a form of human
leukemia (natural killer-cell large granular lymphocyte leukemia) and rat MCL and results of
mechanistic studies that the committee recommended be added to EPA's assessment were
adequate to establish human relevance; others believed that more research was needed to
establish the relevance. The committee agreed that there was little information on a mode of
action of tetrachloroethylene in increasing MCL and that it therefore was not possible to
determine whether exposure to tetrachloroethylene results in initiation of new tumors or
enhances the expansion or promotion of existing tumors.
EPA Response: EPA added the studies that NRC judged to provide indirect evidence that
tetrachloroethylene induces effects associated with MCL and with known leukemogens (see
Section 4.6). Particularly, EPA now includes studies of tetrachloroethylene exposure by
Marth et al. (1989; 1985) and Marth (1987) demonstrating hemolysis and a study by Seidel et
al. (1992) showing effects on bone marrow function. Additionally, EPA summarizes the
findings of Thomas et al. (2007) and presents and discusses the data relevant to interpreting
tetrachloroethylene effects according to the approach proposed by those authors, as
recommended by NRC (see Section 4.6). This includes a summary and statistical analyses of
the MCL findings for tetrachloroethylene in the NTP and JISA bioassays (including of the
A-13
-------
JIS A MCL data which appropriately considered time of death), as well as a presentation of
the historical control incidences of MCL for these laboratories.
NRC Comment: [p 9] Hepatic Cancer
Statistically significant increases in hepatic tumors were observed in male and female
mice after oral or inhalation exposure. As in the case of MCL, the biologic significance of the
increases was debated by the committee because B6C3F1 mice have a high background
incidence of hepatic cancer. However, the findings were reproduced in several studies
conducted in different laboratories and showed a dose-response relationship. There is also fairly
substantial information for characterizing potential modes of action of hepatic-tumor formation
relative to the data available on MCL and renal cancer. Although the committee recommended
that EPA revise its presentation of the mode-of-action evidence on tetrachloroethylene-related
hepatic cancer to clarify its position, most of the members agreed with EPA that the mode of
action is complex and remains to be established. The latter members also agreed that there was
insufficient evidence to rule out human relevance. One member objected to those conclusions
and to the committee's support of using hepatic cancer to quantify risk. He argued that in the
absence of evidence of other contributing modes of action, the evidence is sufficient to conclude
that the mode of action in mice is predominantly through activation of the peroxisome
proliferator-activated receptor-alpha, a mode of action that he considered to be of little relevance
to humans. His arguments are presented in a dissenting statement in Appendix B of the report.
EPA Response: EPA agrees with the majority of the NRC panel that the mode of action for
hepatic tumors observed in male and female mice is complex and remains to be established,
and that, therefore, there was insufficient evidence to rule out human relevance (Sections
4.3.5 and 4.10.5.3). EPA revised the presentation of hepatic mode of action evidence in
accord with the NRC panel suggestions to clarify its position.
NRC Comment: [p 9] Renal Cancer
Tetrachloroethylene caused a low rate of induction of renal tumors in rats. Although the
increases were not statistically significant when compared with concurrent controls, EPA has
used historical controls to calculate the chances of two of these rare carcinomas to occur by
chance to be less than 0.001. Furthermore, a dose-response trend was shown against the low
background and the tumors in the treated rats were malignant whereas the tumors in the controls
were not. EPA provided a strong evaluation of the potential modes of action for
tetrachloroethylene-induced kidney cancer. The committee agrees with EPA that the mode of
action of tetrachloroethylene tumorigenesis is not understood but that a mutagenic mode of
action cannot be ruled out. Thus, renal tumors observed in tetrachloroethylene-treated rats were
A-14
-------
considered relevant to humans although additional characterization of quantitative relevance is
desirable.
EPA Response: EPA agrees with the NRC panel that renal tumors observed in
tetrachloroethylene-treated rats are considered relevant to humans. EPA has performed
additional characterization of quantitative relevance through development of a harmonized
PBPK model for tetrachloroethylene that includes the glutathione conjugation pathway
(Section 3.5, and discussed below, Section A.3.7).
A.3.3. Major NRC Comments on "Selection of Tumor Type for Quantitative
Assessment" and EPA Responses
NRC Comment: [p 10] The committee was unable to reach consensus on the selection of the
critical cancer end point. The majority of the members judged that the uncertainties associated
with MCL (particularly the high background incidence, uncertainty about the dose-response
relationship, and poor understanding of mode of action) were too great to support using MCL
data rather than data on hepatic or renal cancer for determining quantitative estimates of risk.
Those members judged that the use of the MCL data could be justified only if it is EPA's policy
to choose the most conservative unit risk when considering options but that such justification
should be distinguished as a policy decision, not a scientific one. They believed that a more
scientifically defensible approach would be to use the data set that has the least uncertainty rather
than the dataset that yields the highest estimate of risk. In their judgment, the hepatic-cancer
data would have the least uncertainty, followed by the data on renal cancer and MCL.
Other members judged that the MCL data should be used for cancer-risk estimation.
Their opinions were based on the observation that reproducible, statistically significant increases
in MCL in male and female rats above the background incidence of MCL were found and that
MCL was the cancer end point with the highest magnitude of response. They believed that use
of the most sensitive response to quantify cancer risk decreases the uncertainty associated with
potential differences in metabolism and susceptibility to tetrachloroethylene among exposed
populations. They concluded that additional statistical analyses of the dose-response data and
the addition of supporting mechanistic information identified by the committee would strengthen
the existing support of the use of MCL in the draft assessment.
EPA Response: In accordance with the majority of the NRC peer review panel, the oral
slope factor and inhalation unit risk are now based on the male mouse hepatocellular tumor
data from the JISA (1993) bioassay as shown in Sections 5.3.4.2 and 5.3.4.3. EPA also
presents what the cancer risk estimates would be if they were based on the male and female
rat MCL data from the JISA (1993) bioassay.
A-15
-------
A.3.4. Major NRC Comments on "Mode-of-Action Considerations" and EPA
Responses
NRC Comment: [pp 10-11] The modes of action by which tetrachloroethylene produces
increases in MCL, hepatic cancer, and renal cancer were an important consideration in EPA's
and the committee's evaluations of the evidence. The analytic framework described in EPA's
cancer guidelines for considering hypothesized modes of action was best applied in the draft
IRIS assessment's consideration of renal cancer. The evaluation focused on synthesizing the
evidence to support the idea that multiple modes of action may play a role.
EPA Response: EPA accepts these NRC recommendations. In addition, EPA has included
text and tabular summaries of the relevant data for the MO A hypotheses to better and more
clearly support the conclusions.
NRC Comment: [p 11] However, for hepatic cancer, the committee found that the assessment
lacked the organization to present and provide appropriate context for the evidence clearly. It
therefore recommended that EPA revise its mode-of-action assessment for hepatic cancer to
support better the conclusions that were drawn. Specifically, the committee suggested that the
mode-of-action analyses would be improved by outlining the proposed sequence of hypothesized
tetrachloroethylene-associated key events (possibly with a diagram). Transparency would be
improved by presenting the details of experimental results in tabular form to allow the reader to
understand more easily the relative potency of tetrachloroethylene, or its metabolites, in inducing
both key events and tumors. In this context, species and strain differences could also be
considered more easily. The goals of the presentation should be to lay out the timeline of key
events explicitly in the context of dose, to evaluate concordance between early and late events,
and to consider the relative contribution of chemical-specific data compared with information on
categories of chemicals.
EPA Response: EPA accepts these NRC recommendations. EPA agrees with and has
followed NRC's recommendations in revising the discussion of supporting evidence for the
various hypothesized MO As. In particular, the presentation of the pertinent MO A data has
been reorganized, and additional analyses suggested by NRC have been included (Sections
4.3.5, 4.10.5.3, and Appendix C). For each MOA hypothesis, EPA identifies the proposed
sequence of hypothesized key events. The final assessment provides tabular summaries of
the relevant results for tetrachloroethylene to facilitate understanding of the relative potency
of tetrachlroethylene and its metabolites in inducing both hypothesized key events and
tumors. Information on species, strain, sex, dose and temporality of effect are presented so
that concordance between early and late events, and differences across experimental
A-16
-------
paradigms, can be recognized. In addition, data specific to tetrachloroethylene and on other
categories of chemicals are clearly delineated, to aid consideration of their relative
contribution in supporting the conclusions.
NRC Comment: [p 11] This approach should be applied to each hypothesized mode of action.
Even if the data are ultimately judged to be insufficient to support a hypothesis, the exercise can
be used to identify critical data gaps and to inform the direction of future research.
EPA Response: EPA accepts these NRC recommendations. In particular, EPA has
undertaken significant revisions to more clearly specify the hypothesized MO As for each
tumor endpoint, and to present and analyze the evidence available to support conclusions
about these hypothesized MOAs (Sections 4.2.4, 4.3.5, and 4.10.5.3). This includes
presentation of experimental details in tabular form.
A.3.5. Major NRC Comments on "Low-Dose Extrapolation" and EPA Responses
NRC Comment: [p 11] EPA's dose-response analyses of the various cancer datasets involved
using several models to extrapolate to doses below the experimental range. EPA considered six
data sets: hepatocellular adenoma or carcinoma in male and female mice, hemangiosarcoma in
male mice, MCL in male and female rats, and renal tumors in male rats. EPA used the
multistage model for each dataset because mode-of-action information was lacking or uncertain
and the model was able to fit a broad array of dose-response patterns. However, because the
studies used small numbers of dose groups and because the benchmark-dose software
automatically fixed some parameters to zero to obtain convergence in model-fitting, the fitted
models were nearly linear in the low-dose range. The imposed linearity explains the similarity
among the slopes of the models and among the unit risks derived from the models.
EPA Response: EPA would like to clarify that the earlier modeling did not impose linearity
on the subject data sets. EPA's software uses maximum likelihood estimation, a standard
method; the software merely allowed for the possibility of linearity in the chosen model and
did not select or fix parameters at zero. Although multistage model parameters are restricted
to be nonnegative, this only imposes monotonicity, not linearity. Also, the multistage model
can take on more curvilinear forms, even with first-order models. The methods used have
been clarified further in the assessment (Section 5.3.3.2.1).
NRC Comment: [p 11] In the case of hepatocellular adenoma and carcinoma in male mice and
MCL in female rats, EPA considered the fitted models acceptable solely on the grounds that
statistical tests for goodness of fit had nonsignificant results (p > 0.10). The committee considers
A-17
-------
this to be a weak rationale in that the statistical significance of goodness-of-fit tests may not
detect a poor fit when the number of animals per dose group is small.
EPA Response: EPA agrees that factors in addition to statistical significance should be taken
into account in assessing goodness-of-fit, and that adequate power of statistical testing is an
important consideration when interpreting lack or presence of statistical significance. For
instance, EPA prefers that the absolute values of the standardized residuals for the reported
fits be within the limit of ±2 units, though EPA acknowledges that this consideration was not
adequately documented in the external review draft. In addition, the visual fits for both data
sets were not satisfying, as pointed out by the committee. Moreover, in both cases, using the
multistage model, the mode led benchmark concentration for an extra risk of 10% (both the
maximum likelihood estimate and the 95% lower bound) was higher than the concentrations
at which 10% or more extra risk response was observed. EPA has revised its discussions of
the dose-response modeling to more completely document these considerations when judging
goodness of fit. Additional dose-response analyses to improve model fits to these data are
discussed in the response to the next comment.
NRC Comment: [pp 11-12] The questionable fitting of the multistage model to some candidate
datasets and insufficient consideration of alternative models contribute to underestimation of the
overall uncertainties.
EPA Response: EPA agrees that the dose-response relationships for the highlighted data sets
merited reanalysis, particularly in cases where the multistage model did not fit the data at
lower doses. EPA also agrees that in such cases, uncertainty may be underestimated if there
is insufficient consideration of alternative models (see response below under "Uncertainty
Analysis"). The additional analyses, discussed further below, resulted in better
characterization of these data sets.
NRC Comment: [p 12] EPA adopted linear low-dose extrapolation, the default option, with
several justifications. First, nonlinear, mechanistic models are unavailable for dose-response
modeling because mode-of-action information on tetrachloroethylene is insufficient and support
for dynamic models is unavailable. Second, because mathematical models are subject to
uncertainties for low-dose extrapolation beyond the experimental dose range, linear extrapolation
is more conservative than all sublinear (curvilinear) models. When individual thresholds in the
human population are plausible, wide variation in threshold values typically implies a curvilinear
shape of the dose-response relationship. Thus, linear extrapolation protects susceptible
subpopulations. Third, a few of the candidate data, especially the male-rat MCL data, exhibit a
linear dose-response relationship. Whereas those arguments are consistent with EPA's
A-18
-------
Guidelines for Carcinogen Risk Assessment, there is evidence in the candidate datasets that the
underlying dose-response relationship can be supralinear (for example, in MCL in female rats).
When that is the case, low-dose linear extrapolation is not conservative. EPA does not present
the full ranges of variation and uncertainty in relation to model choice, in large part because it
applied only linear or nearly linear dose-response models to all candidate datasets.
EPA Response: EPA agrees that there is evidence in the candidate datasets that the
underlying dose-response relationship can be supralinear, and that in this case low-dose
linear extrapolation is not conservative. Additional analyses were performed to address the
concerns with respect to dose-response fitting of supralinear data sets. Several options to
find better fitting models were considered, including: other model forms; substitution of
historical controls for concurrent controls; exclusion of exposure groups from the analysis,
starting with the highest exposure group; and consideration of dose-response analysis of
combined males and females. These analyses (provided in detail in Section 5.3.4.1) resulted
in use of a one-degree multistage model for the male hepatocellular tumors and in use of
combined male and female rat MCL data and an alternate model (Michaelis-Menten) for
adequate characterizations of the low-dose data. Responses to comments with respect to the
range of uncertainty and variability are addressed below under "Uncertainty Analysis."
A.3.6. Major NRC Comments on "Age-Adjustment Factor" and EPA Responses
NRC Comment: [p 12] EPA did not apply an age-adjustment factor to its cancer risk
assessment, because there is little evidence that tetrachloroethylene or its oxidative metabolites
directly damage DNA, because information about genotoxicity of glutathione (GSH) metabolites
in cell assays other than Salmonella or in vitro experiments is lacking, and because the mode of
action of tetrachloroethylene has not been established. In addition, there are no data on
differential sensitivity to tetrachloroethylene carcinogenicity among life stages. The committee
agrees that those are adequate reasons for not using an age-adjustment factor but suggests that
the rationale can be strengthened if EPA follows the committee's suggestions for improving its
analysis of the genotoxicity data and mode-of-action evidence.
EPA Response: EPA accepts these NRC recommendations. To better support its
conclusions, EPA has substantially revised the genotoxicity section (see Section 4.8) in
accord with NRC recommendations. Text and tabular study summaries of the available
genotoxicity studies of tetrachloroethylene and its metabolites are presented, organized by
test article (chemical entity) and further structured according to the assessed endpoint.
Missing and more recently peer-reviewed and published studies as identified by the NRC
A-19
-------
committee are included. Additionally, the data for each metabolite are summarized, and an
overall synthesis section is included.
A.3.7. Major NRC Comments on "Physiologically Based Pharmacokinetic Models"
and EPA Responses
NRC Comment: [pp 12-13] Tetrachloroethylene can be viewed as being metabolized by three
pathways. The predominant pathway is the cytochrome P-450 (CYP) pathway that produces
metabolites that have been associated with hepatic cancer. Two other pathways involve the GSH
conjugation pathway that produces metabolites that are further metabolized by the p-lyase
pathway or the p-lyase-independent pathway, each of which produce metabolites that have been
associated with renal cancer. To take those metabolic factors into account, EPA used three
PBPK models to estimate human equivalent doses from animal studies and to perform route-to-
route extrapolations. Each of the models used total metabolism of tetrachloroethylene as the dose
metric. In some instances, EPA used a single model; in others, it used all three. The justification
for using single or multiple models is not always clear. The committee observed that the models
could yield different results because they were calibrated with different datasets, so comparisons
among them were not straightforward. For consistency and to allow for better comparisons
among end points, the committee recommends that EPA use a single PBPK model for its
assessment. Ideally, the model would be a "harmonized" version of the three models used by
EPA or of other relevant models (that is, a single model that integrates multiple exposure routes
and tissue compartments).
The committee notes that the use of total metabolism as the dose metric for
carcinogenicity reflects primarily the CYP metabolic pathway because of large differences in the
flux of the metabolism between it and the GSH pathway. Using that dose metric does not reflect
the contribution of the GSH conjugation pathway, which has been implicated in the development
of renal cancer. EPA did not pursue the addition of the GSH pathway to any of the PBPK
models, arguing that data on GSH-dependent metabolism are from in vitro studies or constitute
measurements of urinary excretion products and do not represent toxic species in vivo. The
committee agrees that the available data on the GSH pathway are more limited than the available
data on the CYP pathway but notes that in vitro and urinary metabolite data were used in the
development of the CYP-based PBPK models chosen by EPA. Thus, better justification is
necessary to rule out modeling the GSH pathway.
The committee recommends that EPA explore the possibility of adding the GSH pathway
to a harmonized PBPK model. If such modeling is deter-mined to be infeasible, total
metabolism can be used as a reasonably conservative dose metric. The modeling exercise would
be useful in identifying data gaps that prevent successful modeling, which can be used to guide
A-20
-------
research that will allow more comprehensive PBPK models to be developed in support of the
next IRIS reassessment of tetrachloroethylene.
EPA Response: EPA accepts these NRC recommendations, agrees that a "harmonized"
PBPK model that includes data regarding the GSH pathway would be beneficial, and has
developed such a model that integrates multiple exposure routes and tissue compartments
(Section 3.5). Additionally, EPA followed the NRC advice of separating metabolism into
three pathways (oxidation, GSH-conjugation with further p-lyase metabolism, and GSH-
conjugation with further p-lyase-independent metabolism). The PBPK modeling analysis
showed that the GSH conjugation pathway in humans remains highly uncertain and/or
variable, and that additional data are needed to better quantify that pathway in humans (see
Section 3.5). Therefore, the assessment does not rely on quantitative estimates of GSH
pathway metabolism provided by the new PBPK model. Instead, the quantitative risk
estimates presented in the revised assessment rely on estimates of blood tetrachloroethylene,
oxidation of tetrachloroethylene, and route-to-route extrapolation from this new model.
These dose metric estimates from the new model are robust and consistent with prior models
and, thus, insensitive to model choice.
A.3.8. Major NRC Comments on "Uncertainty Analysis" and EPA Responses
NRC Comment: [p 13] EPA has clearly identified key sources of uncertainty as part of its
process of assessing the cancer risk posed by exposure to tetrachloroethylene, including human
population variation, low-dose extrapolation, dose metrics, extrapolation from animals to
humans, and the use of PBPK models for route-to-route extrapolation. The effect of
uncertainties on risk estimates is assessed qualitatively in most parts of the IRIS draft except in
dealing with such issues as the choice of dose-response models, the use of PBPK models, and, to
a small degree, variation between studies. That approach reflects the current state of practice of
uncertainty analysis.
EPA Response: EPA agrees with the NRC comments that its approach to uncertainty
analysis reflects the current state of practice, and that emerging new methods for
quantification of overarching uncertainty, of variability, and of their cumulative effects could
be considered when tetrachloroethylene is re-evaluated. In addition, as recommended by the
NRC committee, EPA has retained tabular presentation highlighting EPA's choices and their
effects on the determination of the upper bound of the risk estimate (section 5.3.5).
NRC Comment: [p 14] In a few respects, the committee disagrees with EPA's presentation on
uncertainties. For example, EPA notes narrow variation between cancer risks derived from four
A-21
-------
dose-response models. However, in its comparison, EPA used only data on male rats, and all
four models were linear or nearly linear at lower doses. Failure to consider a wider array of
feasible dose-response models, including multistage models of various orders, could lead to
inadequate quantification of uncertainty associated with the choice of dose-response model.
The committee supports EPA's quantitative assessments of uncertainty with regard to
choice of dose-response models, the use of PBPK models, and variation between studies. In
particular, the committee found EPA's consideration of uncertainty due to different forms of
dose-response models to be valuable, and it recommends that such quantitative evaluations be
extended to all candidate datasets so that a fuller array of uncertainties can be assessed.
EPA Response: EPA accepts these NRC recommendations and has extended the quantitative
evaluation of different models to all candidate data sets to a fuller array of uncertainties at the
point of departure (i.e., 10% extra risk level). EPA has conducted dose-response modeling
on the basis of administered concentration for each of the JISA candidate data sets using the
range of dichotomous dose-response models included in BMDS (Appendix D). The results
of the suite of models were evaluated for goodness-of-fit. For datasets exhibiting
supralinearity, models that led to both a better fit to the supralinear shape and a stable BMDL
were considered for further application using PBPK model-based dose metrics. The results
of this analysis showed that for datasets exhibiting supralinearity, the BMD estimated using
the multistage model may lead to an underestimation of risk, consistent with the NRC
comments. Moreover, in such cases, it can be challenging to obtain both a better fit than the
multistage model and a stable lower bound estimate for the BMD.
A.4. Response to Public Comments - Noncancer Assessment
A.4.1. Critical Noncancer End Point and Studies
Public Comments: Several public commenters recommended specifying the criteria used to
select studies of the best quality, to better support weight of evidence conclusions and principal
study selection. Several commenters critiqued Altmann et al. (1995) based on factors such as
small sample sizes, uncontrolled confounding, selection bias, the transient and subtle nature of
the effects, the relevance of exposure scenario and the statistical analysis. Another public
commenter submitted, and recommended consideration of (for RfC derivation and in choice of
UF), the final peer-reviewed report of New York State Department of Health study [NYSDOH
(2010): published by Storm et al. (2011)1. Other studies of neurological effects in residential
populations were identified for use either in supporting an RfC based on Altmann et al. (1995) or
in conduct of a meta-analysis together with Altmann et al. (1995). One commenter noted a lack
of concordance from high to low exposures in human studies, and from human to animal studies.
A-22
-------
EPA Response: As discussed above in Section A.2.1, EPA followed NRC
recommendations in more transparently presenting the rationale for evaluating and
selecting principal studies of neurotoxicity. The EPA re-evaluation included the final
peer-reviewed report of New York State Department of Health study [NYSDOH (2010):
published by Storm et al. (2011)1 that was provided to EPA in public comments and
presented to NRC during the committee deliberations.
Public Comments: A study by the Halogenated Solvents Industry Alliance that had been under
final review was cited as providing a lack of evidence of immune suppression with 28-day
inhalation of up to 1,000 ppm. One commenter noted that misclassification of effect and recall
bias limited conclusions from human studies of autoimmunity.
EPA Response: The study cited on immune suppression is still not available in final,
peer-reviewed form, and so was not addressed in the Toxicological Review. Limitations
regarding ascertainment of disease incidence and exposure assessment in population-
based studies are addressed in the discussion of autoimmune disease data in the
Toxicological Review (see Section 4.6.1.1.2).
A.4.2. Derivation of Reference Values
Public Comments: One commenter disagreed with selection of 3 for the UFD (database).
Another commenter noted that additional accounting of sensitivity and susceptibility of children
is needed in RfC derivation. One commenter remarked that median, rather than mean [adopted
by EPA], tetrachloroethylene concentration is more scientifically defensible as a POD. This
commenter also questioned the assumption of continuous exposure in the critical study
supporting the RfC and recommended EPA develop a time-weighted average exposure estimate
using an estimate of 75% time in residence from Schreiber et al. (2002), the
population/percentile estimates from EPA's Exposure Factor Handbook, or using a biologically-
motivated mathematical or PBPK-based approach.
EPA Response: Based on public comments and concerns raised by the NRC, EPA re-
examined the adequacy of the database and increased the UFD from 3 to 10 (See Sections
A.2.2, 5.1.2, and 5.1.3). EPA generally uses mean exposure, based on the argument of
Crump (1998) that arithmetic means are expected to represent total risk better than
A-23
-------
geometric means (see Section 5.1.1.3.2). The revised assessment uses mean exposures
from occupational studies, which EPA time-weighted to represent continuous exposures.
Public Comments: An objection was raised to citing Fredriksson et al. (1993) as a supporting
study, because no difference in responses was noted between doses which differed by 60-fold.
EPA Response: EPA agrees with concerns that the Fredriksson et al. (1993) is limited to
support dose-response analyses, and no longer includes it among the tetrachloroethylene
neurotoxicity studies considered for dose-response analysis.
A.4.3. Graphical Presentation
Public Comments: Several commenters endorsed graphical presentation of studies to illustrate
the support of reference values by multiple studies, and one commented that a distributional
quantitative uncertainty analysis should have been undertaken.
EPA Response: EPA has followed the NRC recommendations and agrees that the
graphical presentation of studies and resulting risk values is useful (see Section A.2.3).
As noted by the NRC, EPA's approach with respect to uncertainty analysis represents the
current state of practice. In agreement with the NRC advice, EPA will consider
expanding beyond the current state of practice for uncertainty analysis - such as use of
distributional quantitative uncertainty analysis for non-cancer effects - in future re-
evaluations of tetrachloroethylene.
A.4.4. Reproductive and Developmental Effects
Public Comments: Several commenters raised concern that the epidemiologic studies for
developmental and reproductive endpoints were not objectively reviewed. One commenter
raised concern about the studies by Szakmary et al. (1997) and Fredricksson et al. (1993), noting
that in the latter study effects were only reported at doses causing maternal toxicity. They also
raised concern about the presentation of a potential MOA for developmental toxicity, especially
regarding the potential role of the tetrachloroethylene metabolite TCA. Other concerns raised by
the commenter related to POD selection and lack of transparency in LOAEL and NOAEL
selection.
EPA Response: As presented in Section A.2.4, study strengths and deficiencies, and
evidence from supportive in vitro and in vivo studies and the consistency of outcomes
across species and protocols, are addressed. Findings of parental (including maternal)
A-24
-------
toxicity and the treatment levels at which those effects were observed are also described
for each study. The developmental neurotoxicity evaluation by Fredricksson et al. (1993)
is presented in the neurotoxicity section of the assessment, with a brief summary included
in the developmental toxicity section. EPA has also expanded the discussion of the MOA
hypotheses for developmental outcomes to address the potential involvement of the
metabolite TCA. Selection of PODs for endpoints selected for potential RfC
development (see Table 4-49) has also been more transparently described.
A.5. Response to Public Comments - Cancer Assessment
A.5.1. Epidemiologic Evidence Pertaining to Cancer
Public Comments: Several commenters were critical of the presentation and interpretation of
epidemiologic studies and the weight-of-evidence provided by these studies. A more clear,
comprehensive and balanced review was recommended. Particular comments concerned
community drinking-water studies [Aschengrau et al. (1993), as well as Lynge et al. (2006),
Ruder et al. (2001), and Ma et al. (2009)1.
EPA Response: As discussed in Section A.3.1, EPA has followed the NRC
recommendations and significantly reorganized the data presentation by type of cancer,
included updated and more comprehensive evaluation of the epidemiologic literature, and
clarified presentation of data from these studies, including those identified by public
commenters.
A.5.2. Cancer Characterization
Public Comments: Several commenters raised issues about the cancer characterization. Some
agreed with the characterization of tetrachloroethylene as "likely to be carcinogenic to humans";
others disagreed based on the lack of human relevance of animal tumors, and inconclusive
epidemiologic evidence.
EPA Response: As discussed in Section A.3.1, EPA has followed the NRC
recommendations and continues to characterize tetrachloroethylene as "likely to be
carcinogenic to humans." EPA's characterization was based on the results of bioassays
that found increased incidences of tumors and to a lesser extent on epidemiologic
evidence.
A-25
-------
A.5.3. Hepatic and Renal Toxicity and Cancer
Public Comments: Several commenters recommended improved transparency and clarity in the
presentation of hepatic and renal toxicity and carcinogenicity (including hepatocellular tumors
and hemangiosarcomas), and carcinogenic MOA information. With respect to hepatic endpoints,
one commenter recommended transparency with respect to presentation and analyses of the
hemangiosarcoma data of JISA (1993). Another was critical of the Kjellstrand et al. (1984)
study. Regarding kidney, commenters stated that the rodent and human data were not
comprehensively or critically evaluated. One commenter noted that hepatocellular and renal
tumors were specific to the rodent species and strains studied.
EPA Response: EPA has revised its presentation of the renal and hepatic toxicity,
carcinogenicity (including hepatocellular tumors and hemangiosarcomas), and
carcinogenic MOA information, following NRC recommendations. For liver toxicity,
lesser emphasis has been given to the hepatotoxicity findings in the shorter-term study of
Kjellstrand et al. (1984), which has also been more completely and accurately described.
A.5.4. Selection of Tumor Type for Quantitative Assessment
Public Comments: Several commenters were critical of calculating cancer potency based on
MCL, highlighting issues of susceptibility and that this endpoint is a poor model for human
responses. Others raised issues with the use of any of the rodent tissue endpoints based on their
apparent specificity with respect to species/sex combinations.
EPA Response: As discussed in Section A.3.3, the majority of the NRC peer review
panel recommended that the mouse hepatocellular tumors be used for cancer risk
estimation. Therefore, the oral slope factor and inhalation unit risk are now based on the
male mouse hepatocellular tumor data from the JISA (1993) bioassay.
A.5.5. Mode-of-Action Considerations
Public Comments: Several commenters raised issues about the MOA presentation and
conclusions, recommending improved transparency and clarity. With respect to genotoxicity,
several recommended a more comprehensive review of the available studies for
tetrachloroethylene and its metabolites, including tabular summaries of the available data, a
discussion of study strengths and weaknesses, and a summary discussion of the evidence.
Several criticized the clarity of the MOA presentation for hepatocellular tumors. Some
commenters agreed, while others disagreed, that PPARa is not the MOA. Some recommended
A-26
-------
more explicitly addressing tetrachloroethylene-specific studies, the role of metabolic activation,
and alternative MO As (cytotoxicity and hyperplasia). For renal cancers induced by
tetrachloroethylene, further consideration and discussion of the PPARa activation and sustained
cytotoxicity MO As was recommended.
EPA Response: As discussed in A.3.3, EPA has undertaken significant revisions to more
clearly specify the hypothesized MO As for each tumor endpoint, and to present and
analyze the evidence available to support conclusions about these hypothesized MO As.
In particular, the genotoxicity section (see Section 4.8) was substantially revised to
include text and tabular study summaries of the available genotoxicity studies of
tetrachloroethylene and its metabolites, and an overall synthesis section. For mouse liver
tumors, EPA has significantly revised the presentation of the PPARa activation MO A
and added discussion of epigenetic changes and cytotoxicity and secondary oxidative
stress. EPA presents quantitative analyses of TCA, DC A, other known peroxisome
proliferators, PPARa endpoints (including PPARa transactivation) and hepatic cancer.
Additional emphasis has been given to the deficiencies in the knowledge-base regarding
the MOA for tetrachloroethylene. Similarly, with respect to rat kidney tumors, EPA has
included text and tabular summaries of the relevant data for the MOA hypotheses
(including PPARa activation and sustained cytotoxicity) to better and more clearly
support the conclusions.
A.5.6. Low-Dose Extrapolation
Public Comments: Several commenters supported a (threshold) non-linear dose-response
analysis (or approach) based on hypothesized MO As (particularly, PPARa activation for liver
and kidney). Several were critical of the range of potencies; some noted that other factors in
addition to PBPK models should be considered, while others recommended selection of a point
estimate within the proposed range.
EPA Response: As described in Section A.3.5, EPA has retained linear low-dose
extrapolation from the POD below the observed range, based on NRC advice.
The NRC noted that this is in accordance with the U.S. EPA (2005) Guidelines for
Carcinogen Risk Assessment, and supported EPA's use of linear low-dose extrapolation.
In addition, as recommended by the NRC, EPA has evaluated multiple options for
improving the model fit in the observed range for data sets showing supra-linear dose-
response shapes. As discussed in Section A.3.5, the additional analyses resulted in better
characterization of the available data sets for cancer risk estimation.
A-27
-------
A.5.7. Physiologically Based Pharmacokinetic Models
Public Comments: Several commenters addressed the dose metrics and PBPK models used in
risk estimation. Regarding the dose metric, some commenters endorsed total metabolism;
another commenter remarked that dose metrics for cancer risk estimates do not address the
distribution or elimination of metabolites likely involved with carcinogenic process. One
commenter disagreed that BW3/4-scaling is appropriate for interspecies extrapolation because 1)
EPA has not established that a metabolite causes cancer; and 2) the unknown metabolite could be
a highly reactive intermediate. Several commenters were critical of the PBPK models used, with
some suggesting inclusion of the Clewell et al. (2005) model, one recommending inclusion of
Covington et al. (2007), and another recommended utilizing only Clewell et al. (2005) and
Gearhart et al. (1993). One commenter suggested using the upper 95th confidence limit of the
fraction metabolized in the Chiu and Bois (2006) analysis (i.e., 61% at a modeled exposure
concentration of 0.001 ppm). Limitations of the available PBPK models for predicting GSH
conjugation pathway metabolism were noted. Commenters recommended clarity in the
presentation of, and disagreed with some conclusions regarding, the oral and dermal metabolism
of tetrachloroethylene, the rates of metabolism through the GSH pathway conjugation, the
metabolism of TCA to DCA, the bioavailability of TCA, and the presentation of and selection
among the available PBPK models.
EPA Response: As discussed in Section A.3.7, EPA developed a "harmonized" PBPK
model (Chiu and Ginsberg, 2011), including implementation of the NRC advice to
separate metabolism into three pathways (oxidation, GSH-conjugation with further
P4yase metabolism, and GSH-conjugation with further p-lyase-independent metabolism).
The harmonized PBPK modeling analysis showed that the GSH conjugation pathway in
humans remains highly uncertain and/or variable (yielding an approximately 3000-fold
range in human estimates), and that additional data are needed to better quantify that
pathway in humans (see Section 3.5). Therefore, the assessment does not rely on
quantitative estimates of GSH pathway metabolism provided by the new PBPK model.
Instead, the quantitative risk estimates presented in the revised assessment rely on
estimates of blood tetrachloroethylene, oxidation of tetrachloroethylene, and route-to-
route extrapolation information derived from this harmonized model. These dose metric
estimates from the harmonized model are robust and consistent with prior models and,
thus, insensitive to model choice. EPA also revised its presentation of metabolism (see
Section 3) and dose metric selection (see Section 5), as well as presentation of an
A-28
-------
empirical analysis of the contribution of TCAto tetrachloroethylene-induced hepatic
tumorigenesis (see Appendix C).
Given the current understanding of tetrachloroethylene metabolism and cancer mode(s)
of action, EPA maintains that BW3/4-scaling of metabolites (oxidative metabolites for
hepatocellular tumors, GSH conjugates for kidney tumors) for extrapolation to human
cancer risk is supported by the principles previously outlined by U.S. EPA (1992). In this
revised assessment, risks extrapolated using tetrachloroethylene AUC in blood as the
dose metric were not scaled by BW3/4, also consistent with U.S. EPA (1992).
A.5.8. Uncertainty Analysis
Public Comments: Some public commenters recommended that a more clear and concise
summary of the limitations and uncertainties in the data and analyses would be more informative
than the uncertainty analysis presented by EPA. Others recommended expansion of EPA's
uncertainty analysis to include additional quantitative analyses.
EPA Response: EPA has followed the NRC recommendations and retained tabular
presentation highlighting EPA's choices and their effects on the determination of the
upper bound of the risk estimate. As discussed in Section A.3.8, EPA has followed the
NRC recommendations and extended the quantitative evaluation of different models to
all candidate data sets so as to more fully array the uncertainties at the point of departure.
A.6. Focused External Peer Review of the Application of Physiologically-Based
Pharmacokinetic (PBPK) modeling in the Toxicological Review for Tetrachloroethylene
The IRIS Toxicological Review of Tetrachloroethylene utilizes a harmonized PBPK model that
EPA developed in response to NRC (2010) recommendations. In particular, the NRC
recommended that EPA:
1. Pursue development of a single harmonized PBPK model for its assessment (as was done
for trichloroethylene), which synthesizes important aspects of the previously published
models. This includes the use of multiple exposure routes and inclusion of all relevant
tissue compartments, and utilization of data from all relevant species and routes of
exposures.
2. Explore the possibility of adding the GSH pathway (in addition to the P-450 metabolic
pathway) to a harmonized PBPK model. The initial goal should be to predict the fraction
A-29
-------
of an absorbed tetrachloroethylene dose that enters the GSH pathway and the fraction that
enters the P-450 pathway. That would permit the development of more discrete dose
metrics and should lead to a more rational and defensible selection of dose metrics for the
various cancer end points.
3. If such modeling is determined to be infeasible, then identify data gaps that prevent
successful modeling. These can be used to guide future research.
EPA accepted the NRC recommendations and developed a harmonized PBPK model for mice,
rats, and humans. In particular:
1. The model synthesizes the important aspects of the various previous models, and utilizes
a wider range of in vitro and in vivo data than any previous individual PBPK model of
tetrachl oroethy 1 ene.
2. The model separately predicts the fractions of absorbed tetrachloroethylene dose that
enter either the P-450 or the GSH pathways, in addition to blood levels of the parent
compound tetrachloroethylene.
3. The analysis showed that modeling the GSH pathway is feasible, but that predictions for
this pathway in humans remain highly uncertain and/or variable. Therefore, better
quantification of this pathway, whether at the individual or the population level, would
require more direct in vivo data.
This harmonized PBPK model was peer-reviewed and published by Toxicology and Applied
Pharmacology (Chiu and Ginsberg, 2011).
EPA conducted a focused peer review on the application of the published harmonized model to
support its use in the IRIS Toxicological Review of Tetrachloroethylene. Peer reviewers were
asked to consider whether the harmonized PBPK model 1) is clearly and transparently described,
and adequately responsive to the NRC recommendations; and 2) used appropriately in the dose-
response assessment.
The following peer reviewers were independently chosen by an EPA peer review contractor
according to scientific credentials and particular areas of expertise:
1. Janusz Z. Byczkowski, Ph.D., DABT - Independent Consultant, Fairborn, OH
2. Claude Emond, Ph.D. - University of Montreal, Quebec, Canada
A-30
-------
3. Moiz Mumtaz, Ph.D. - Computational Toxicology and Methods Development
Laboratory, Division of Toxicology and Environmental Medicine, Agency for Toxic
Substances and Disease Registry, Atlanta, GA
The charge questions, peer reviewer comments, and EPA responses are summarized below.
A.6.1. Charge Question #1, Peer Reviewer Comments, and EPA Responses
Charge Question:
1. NRC (2010) recommended that EPA develop a harmonized PBPK model for
tetrachloroethylene and metabolites for use in its risk assessment, and provided advice to
EPA on how such a model should be developed. Overall, is the harmonized PBPK
model developed by EPA clearly and transparently described and adequately
responsive to the NRC recommendations?
Peer Review Comments: All three peer reviewers commented that the PBPK model developed
by EPA appears to be adequately responsive to the NRC recommendations. Specifically, one
reviewer found the model to be "adequately responsive to the NRC recommendations from [a]
technical standpoint"; a second reviewer stated that EPA "appeared to respond clearly and
transparently to the National Research Council's (NRC's) recommendation"; and a third
reviewer stated that EPA "has used credible science and current information to adequately
respond to the recommendations of the NRC."
EPA Response: No response needed.
Peer Review Comments: Two reviewers recommended including more detail on the PBPK
model structure and changing the conceptual representation of the model. Specifically, one peer
reviewer commented that the model "is not clearly and transparently described," and
recommended additional documentation in the Toxicological Review of the model structure and
mechanics. A second reviewer found the model to be well presented, but having room for
improvement. In particular, this reviewer found the conceptual representation of the model
(Figure 3-5) did not provide a full understanding of the model's structure, and thought Tables 3-2
to 3-5 could benefit from presentation of more details. The third reviewer only provided
editorial comments on this point.
EPA Response: In response to the reviewers' recommendations, EPA has substantially
revised Figure 3-5 and its caption in order to more completely document the model
structure and associated parameters. EPA has also added text to the table notes of Tables
A-31
-------
3-2 to 3-5 to provide more details as to how the calculations were made. The
Toxicological Review contains approximately 17 pages of text and tables describing the
mechanics, parameterization, and predictions of the PBPK model. Further details of the
mechanics and structure of the model are contained in Chiu and Ginsberg (2011)
(particularly Appendix A of the article). To further address the peer reviewers'
comments, EPA has made the model code available for download via the internet. The
supporting publications and detailed code are both publicly available through EPA's
HERO database, and hyperlinks to these materials are contained in the Toxicological
Review.
A.6.2. Charge Question #2, Peer Reviewer Comments, and EPA Responses
Charge Question:
2. EPA's dose-response assessment includes development of a chronic inhalation reference
concentration (RfC) and oral reference dose (RfD) for non-cancer effects, and an
inhalation unit risk and oral slope factor for carcinogenic effects. The assessment uses
the following dose metric predictions from the harmonized PBPK model to conduct inter-
species and/or route-to-route extrapolation for use in the dose-response assessment:
a. For the critical non-cancer effect of neurotoxicity, in accordance with NRC
recommendations, the area-under-the-curve (AUC) of tetrachloroethylene in
blood is used as the preferred dose metric, and represents a surrogate for the AUC
of tetrachloroethylene in the brain.
b. For non-cancer hepatotoxicity and hepatocarcinogenesis, liver oxidative
metabolism was used as the preferred dose metric, due to weight of evidence that
oxidative metabolism plays a role in these endpoints for tetrachloroethylene.
Results for the AUC of trichloroacetic acid (TCA) in the liver were presented as
an alternative dose metric for comparison purposes.
c. For non-cancer nephrotoxicity and nephrocarcinogenesis, from a toxicological
perspective, glutathione (GSH) conjugation metabolism would have been the
preferred dose metric due to the weight of evidence that conjugative metabolites
play a role in these endpoints for tetrachloroethylene. However, due to the wide
range of PBPK model predictions for GSH conjugation in humans, the surrogate
dose metric of AUC of tetrachloroethylene in blood was preferred. Results for
A-32
-------
GSH conjugation were presented as an alternative dose metric for comparison
purposes.
d. For all other non-cancer endpoints (reproductive, developmental, immunological,
hematological toxicity) and cancer endpoints (hemangiosarcomas, mononuclear
cell leukemias, brain gliomas, and testicular interstitial cell tumors), the AUC of
tetrachloroethylene in blood was used as the preferred dose metric, due to the lack
of available data on what the active carcinogenic moiety(ies) may be for these
endpoints.
Is the harmonized PBPK model used appropriately for making predictions for these
dose metrics, and are the results appropriately characterized?
Peer Review Comments: The peer reviewers supported the technical soundness of the
application of the PBPK model and the numerical results. In particular, one reviewer found that
the technical soundness of the PBPK methodology and numerical results seems adequate. A
second reviewer found that PBPK modeling improves the quality of risk assessment predictions,
and that the current model will reduce the uncertainties as compared to using previous PBPK
models or the default approach. A third reviewer found that using the model will result in an
acceptable degree of the uncertainty in the assessment.
EPA Response: No response needed.
Peer Review Comments: One peer reviewer requested additional explanation of how the AUC
was calculated, and how and when the uncertainty factors are applied. This reviewer also
recommended following U.S. EPA (2006a) recommendations as to description and
documentation of the PBPK model and its use. Another peer reviewer requested information
about the dose metric selection for non-cancer effects.
EPA Response: EPA has added documentation as to how the AUC was calculated to
Tables 3-2 to 3-5. Documentation as to how and when the uncertainty factors are applied
in cases of PBPK modeling use is already contained in Sections 5.1.4, 5.2.2, and 5.2.4;
additional documentation has not been provided. With respect to the comment to follow
U.S. EPA (2006b) recommendations as to description and documentation of the PBPK
model and its use, EPA has, as stated above, added links to the full text of the Chiu and
Ginsberg (2011) article, the supplementary materials, and the model code and simulation
files, rather than reproducing the information in the Toxicological Review or its
appendices. Specifically, the documentation recommended by U.S. EPA (2006b) is as
A-33
-------
follows, with the location of such documentation for the tetrachloroethylene PBPK model
in square brackets:
• Graphical representation of the model [revised Figure 3-5]
• Well-annotated and complete documentation of the model code [HERO link
to model code and simulation files: (U.S. EPA, 2011)1
• All data (fully referenced) that were used to calibrate and/or test the model
[HERO link to model code and simulation files: (U.S. EPA. 2011)1
• A description of the calibration and testing procedures [HERO link to full text
and supplementary material: (Chiu and Ginsberg, 2011)1
• Full reference of sources for all parameter values (or the optimization
methods, results, and data used in optimizing parameters) [HERO link to full
text: (Chiu and Ginsberg. 2011)1
• Sensitivity analysis or other rationale that guided the choice of which
parameters were optimized [HERO link to full text and supplementary
material: (Chiu and Ginsberg, 2011)]
• Simulation run conditions [added as notes to Tables 3-2 to 3-5, as well as
contained in HERO link to model code and simulation files: (U.S. EPA,
2011)]
• Any additional analyses that help characterize or support the quality of the
model [HERO link to full text and supplementary material: (Chiu and
Ginsberg. 2011)1
• All supporting documentation that would be needed by an experienced
modeler to run the model and accurately reproduce any simulations used (or
submitted for use) in deriving reference values [HERO link to model code and
simulation files: (U.S. EPA. 2011)1
Finally, with respect to the requested information about the dose metric selection for
non-cancer effects, the following information has been added to Sections 5.1.4 and 5.2.4: "For
liver effects, the dose metric of liver oxidative metabolism was used, based on the view that
oxidative metabolites are involved in tetrachloroethylene-induced liver effects. For kidney
effects, while it is generally thought that GSH conjugation metabolites are involved, the large
uncertainty in estimates of human GSH conjugation preclude use of that dose metric. Instead,
the AUC of tetrachloroethylene in blood is used as a surrogate. For the other non-cancer effects,
A-34
-------
the AUC of tetrachloroethylene in blood was used as the preferred dose metric due to the lack of
data on what the active toxic moeity(ies) may be for those effects."
Peer Review Comments: One peer reviewer raised a question as to whether using
tetrachloroethylene in blood as a surrogate for tetrachloroethylene in brain implied that the
concentrations in brain and blood are similar, that tetrachloroethylene crosses the blood-brain
barrier, and that there is no metabolism in the brain. This reviewer questioned the
appropriateness of these assumptions, and that they could result in "an unrealistically
conservative RfC."
EPA Response: EPA notes that the inhalation RfC did not utilize PBPK modeling, so is
not impacted by the choice of dose metric. Tetrachloroethylene in blood is used as a
surrogate for tetrachloroethylene in brain only for route-to-route extrapolation from the
inhalation RfC to the oral RfD for neurotoxicity. For this application, the only
assumption is that the blood and brain levels are proportional, independent of route of
exposure, not that they are similar. Therefore, no assumptions are made as to the degree
to which tetrachloroethylene crosses the blood-brain barrier and the degree of metabolism
in the brain. Thus, any uncertainties regarding the blood-brain barrier and metabolism in
the brain do not impact the RfD derived from route-to-route extrapolation. As noted in
Section 3.2, tetrachloroethylene readily crosses the blood-brain barrier, with measured
brain concentrations 4-5 times higher than blood concentrations.
A.6.3. Editorial comments
Peer Review Comment: One reviewer provided editorial comments.
EPA Response: EPA has addressed the editorial comments identified by the reviewer.
A-35
-------
A.7. References
Acharya, S; Mehta, K; Rodrigues, S; Pereira, J: Krishnan, S; Rao, CV. (1995). Administration of
subtoxic doses of t-butyl alcohol and trichloroacetic acid to male Wistar rats to study the
interactive toxicity. Toxicol Lett 80: 97-104. http://dx.doi.org/10.1016/0378-
4274(95)03340-0.
Altmann, L; Bottger, A; Wiegand, H. (1990). Neurophysiological and psychophysical
measurements reveal effects of acute low-level organic solvent exposure in humans. Int
Arch Occup Environ Health 62: 493-499. http://dx.doi.org/10.1007/BF00381179.
Altmann, L; Wiegand, H; Bottger, A; Elstermeier, F; Winneke, G. (1992). Neurobehavioural and
neurophysiological outcomes of acute repeated perchloroethylene exposure. Appl
Psychol 41: 269-279. http://dx.doi.Org/10.llll/i.1464-0597.1992.tb00705.x.
Altmann, L: Neuhann, HF; Kramer, U; Witten, J: Jermann, E. (1995). Neurobehavioral and
neurophysiological outcome of chronic low-level tetrachloroethene exposure measured in
neighborhoods of dry cleaning shops. Environ Res 69: 83-89.
http://dx.doi.org/10.1006/enrs.1995.1028.
Aschengrau, A: Ozonoff. D: Paulu, C: Coogan, P: Vezina, R: Heeren, T: Zhang. Y. (1993).
Cancer risk and tetrachloroethylene-contaminated drinking water in Massachusetts. Arch
Environ Health 48: 284-292.
Beliles, RP: Brusick, DJ: Mecler, FJ. (1980). Teratogenic-mutagenic risk of workplace
contaminants: trichloroethylene, perchloroethylene, and carbon disulfide. (210-77-0047).
Cincinnati, OH: National Institute for Occupation Safety and Health.
Benignus, VA: Bushnell. PJ: Boves, WK: Eklund, C: Kenyon, EM. (2009). Neurobehavioral
effects of acute exposure to four solvents: Meta-analyses. Toxicol Sci 109: 296-305.
http://dx.doi.org/10.1093/toxsci/kfp063.
Boyes, WK: Bercegeay, M; Oshiro, WM; Krantz, QT; Kenyon, EM; Bushnell, PJ: Benignus,
VA. (2009). Acute perchloroethylene exposure alters rat visual-evoked potentials in
relation to brain concentrations. Toxicol Sci 108: 159-172.
http://dx.doi.org/10.1093/toxsci/kfn265.
Carney, EW: Thorsrud, BA; Dugard, PH; Zablotny, CL. (2006). Developmental toxicity studies
in Crl:CD (SD) rats following inhalation exposure to trichloroethylene and
perchloroethylene. Birth Defects Res B Dev Reprod Toxicol 77: 405-412.
http://dx.doi.org/10.1002/bdrb.20091.
Cavallerl A: Gobba. F: Paltrinierl M: Fantuzzl G: Righl E: Aggazzottl G. (1994).
Perchloroethylene exposure can induce colour vision loss. Neurosci Lett 179: 162-166.
http://dx.doi.org/10.1016/0304-3940(94)90959-8.
Chiu, WA; Bois, FY. (2006). Revisiting the population toxicokinetics of tetrachloroethylene.
Arch Toxicol 80: 382-385. http://dx.doi.org/10.1007/s00204-006-0061-9.
Chiu, WA; Ginsberg, GL. (2011). Development and evaluation of a harmonized physiologically
based pharmacokinetic (PBPK) model for perchloroethylene toxicokinetics in mice, rats,
and humans. Toxicol Appl Pharmacol 253: 203-234.
http://dx.doi.0rg/10.1016/i.taap.2011.03.020.
Clewell, HJ: Gentry, PR: Kester, JE; Andersen, ME. (2005). Evaluation of physiologically based
pharmacokinetic models in risk assessment: An example with perchloroethylene. Crit
Rev Toxicol 35: 413-433. http://dx.doi.org/10.1080/10408440590931994.
A-36
-------
Costantini, AS: Miligi, L; Kriebel D: Ramazzotti, V: Rodella, S; Scarpi, E; Stagnaro, E;
Tutnino, R; Fontana, A; Masala, G: Vigano, C: Vindigni, C: Crosignani, P; Benvenuti, A;
Vineis, P. (2001). A multicenter case-control study in Italy on hematolymphopoietic
neoplasms and occupation. Epidemiology 12: 78-87.
Covington, TR; Gentry, PR: Van Landingham, CB; Andersen, ME; Kester, JE; Clewell, HJ.
(2007). The use of Markov chain Monte Carlo uncertainty analysis to support a public
health goal for perchloroethylene. Regul Toxicol Pharmacol 47: 1-18.
http://dx.doi.0rg/10.1016/i.vrtph.2006.06.008.
Crump, KS. (1998). On summarizing group exposures in risk assessment: Is an arithmetic mean
or a geometric mean more appropriate? Risk Anal 18: 293-297.
Echeverria, D: White, RF; Sampaio, C. (1995). A behavioral evaluation of PCE exposure in
patients and dry cleaners: A possible relationship between clinical and preclinical effects.
J Occup Environ Med 37: 667-680.
Emara, AM; Abo El-Noor, MM; Hassan, NA; Wagih, AA. (2010). Immunotoxicity and
hematotoxicity induced by tetrachloroethylene in egyptian dry cleaning workers. Inhal
Toxicol 22: 117-124. http://dx.doi.org/10.3109/08958370902934894.
Fredriksson, A; Danielsson, BRG: Eriksson, P. (1993). Altered behaviour in adult mice orally
exposed to tri- and tetrachloroethylene as neonates. Toxicol Lett 66: 13-19.
http://dx.doi.org/10.1016/0378-4274(93)90074-8.
Gearhait JM: Mahle. DA: Greene. RJ: Seckel CS: Flemming. CD: Fisher. JW: IIL CH. (1993).
Variability of physiologically based pharmacokinetic (PBPK) model parameters and their
effects on PBPK model predictions in a risk assessment for perchloroethylene (PCE).
Toxicol Lett 68: 131-144. http://dx.doi.org/10.1016/0378-4274(93)90126-1.
Gobba, F; Righi, E: Fantuzzi, G: Predieri, G: Cavazzuti, L; Aggazzotti, G. (1998). Two-year
evolution of perchloroethylene-induced col or-vision loss. Arch Environ Health 53: 196-
198.
JISA (Japan Industrial Safety Association). (1993). Carcinogenicity study of tetrachloroethylene
by inhalation in rats and mice. Hadano, Japan.
Kjellstrand, P: Holmquist B; Kanje, M; Aim, P; Romare, S: Jonsson, I; Mansson, L: Bjerkemo,
M. (1984). Perchloroethylene: Effects on body and organ weights and plasma
butyrylcholinesterase activity in mice. Acta Pharmacol Toxicol 54: 414-424.
http://dx.doi.0rg/10.llll/i.1600-0773.1984.tb01951.x.
Lynge, E: Andersen, A; Rylander, L: Tinnerberg, H; Lindbohm, ML; Pukkala, E: Romundstad,
P: Jensen, P: Clausen, LB; Johansen, K. (2006). Cancer in persons working in dry
cleaning in the Nordic countries. Environ Health Perspect 114: 213-219.
http://dx.doi.org/10.1289/ehp.8425.
Ma, J: Lessner, L: Schreiber, J: Carpenter, DO. (2009). Association between residential
proximity to PERC dry cleaning establishments and kidney cancer in New York City. J
Environ Public Health 2009: 183920. http://dx.doi.org/10.1155/2009/183920.
Marth, E: Stunzner, D: Binder, H; Mose, JR. (1985). [Tetrachloroethylene: effect of low
concentrations of 1,1,2,2-tetrachloroethylene (perchloroethylene) on organisms in the
mouse. I. Laboratory chemical research]. Zentralbl Bakteriol Mikrobiol Hyg 181: 525-
540.
Marth, E. (1987). Metabolic changes following oral exposure to tetrachloroethylene in subtoxic
concentrations. Arch Toxicol 60: 293-299.
A-37
-------
Marth, E; Stiinzner, D; Kock, M; Mose, JR. (1989). Toxicokinetics of chlorinated hydrocarbons.
J Hyg Epidemiol Microbiol Immunol 33: 514-520.
Nelson. BK: Taylor. BJ: Setzer. JV: Hornung. RW. (1979). Behavioral teratology of
perchloroethylene in rats. J Environ Pathol Toxicol Oncol 3: 233-250.
NRC (National Research Council). (2010). Review of the Environmental Protection Agency's
draft IRIS assessment of tetrachloroethylene. Washington, DC: National Academies
Press.
NYSDOH (New York State Department of Health). (2010). Tetrachloroethylene (perc) exposure
and visual contrast sensitivity (VCS) test performance in adults and children residing in
buildings with or without a dry cleaner. Troy, NY.
Oshiro. WM: Krantz. QT: Bushnell. PJ. (2008). Characterization of the effects of inhaled
perchloroethylene on sustained attention in rats performing a visual signal detection task.
Neurotoxicol Teratol 30: 167-174. http://dx.doi.Org/10.1016/j.ntt.2008.01.002.
Ruder, AM; Ward, EM; Brown, DP. (2001). Mortality in dry-cleaning workers: An update. Am J
IndMed39: 121-132. http://dx.doi.org/10.1002/1097-0274(200102)39:2<121::AID-
AJIM1000>3.0.CO:2-H.
Schreiber. JS: Hudnell. HK: Geller. AM: House. DE: Aldous. KM: Force. MS: Langguth. K:
Prohonic, EJ: Parker, JC. (2002). Apartment residents' and day care workers' exposures to
tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect
110: 655-664.
Seidel HJ: Weber, L; Barthel E. (1992). Hematological toxicity of tetrachloroethylene in mice.
Arch Toxicol 66: 228-230. http://dx.doi.org/10.1007/BF01974021.
Shafer, TJ: Bushnell, PJ: Benignus, VA; Woodward, JJ. (2005). Perturbation of voltage-sensitive
Ca2+ channel function by volatile organic solvents. J Pharmacol Exp Ther 315: 1109-
1118. http://dx.doi.0rg/10.l 124/jpet. 105.090027.
Storm. JE: Mazor. KA: Aldous. KM: Blount BC: Brodie. SE: Serle. JB. (2011). Visual contrast
sensitivity in children exposed to tetrachloroethylene. Arch Environ Occup Health 66:
166-177. http://dx.doi.org/10.1080/19338244.2010.539638.
Szakmary, E; GUngvary, G: Tatrai, E. (1997). The offspring-damaging effect of
tetrachloroethylene in rats, mice, and rabbits. Central Eur J Occup Env Med 3: 31-39.
Thomas. J: Haseman. JK: Goodman. JI: Ward. JM: Loughran. TP. Jr: Spencer. PJ. (2007). A
review of large granular lymphocytic leukemia in Fischer 344 rats as an initial step
toward evaluating the Implication of the endpoint to human cancer risk assessment
[Review]. Toxicol Sci 99: 3-19. http://dx.doi.org/10.1093/toxsci/kfm098.
Tinston, DJ. (1994). Perchloroethylene: A multigeneration inhalation study in the rat.
(CTL/P/4097, 86950000190). Cheshire, UK: Zeneca Central Toxicology Laboratory.
U.S. EPA (U.S. Environmental Protection Agency). (1992). A cross-species scaling factor for
carcinogen risk assessment based on equivalence of mg/kg3/4/day. Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). (1994). Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry. (EPA/600/8-90/066F).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
and Development, Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid=71993.
A-38
-------
U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes. (EPA/630/P-02/002F). Washington, DC.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=51717.
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk
assessment. (EPA/630/P-03/001F). Washington, DC.
http://www.epa.gov/cancerguidelines/.
U.S. EPA (U.S. Environmental Protection Agency). (2006a). Approaches for the application of
physiologically based pharmacokinetic (PBPK) models and supporting data in risk
assessment (final report). (EPA/600/R-05/043F). Washington, DC: U.S. Environmental
Protection Agency, Office of Research and Development.
U.S. EPA (U.S. Environmental Protection Agency). (2006b). A framework for assessing health
risk of environmental exposures to children. (EPA/600/R-05/093F). Washington, DC.
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid= 158363.
U.S. EPA (U.S. Environmental Protection Agency). (2006c). Peer review handbook (3rd
edition). (EPA/ 100/B-06/002). Washington, DC: U.S. Environmental Protection Agency,
Science Policy Council, http://www.epa.gov/peerreview/.
U.S. EPA (U.S. Environmental Protection Agency). (2011). Final PBPK model simulation files
[Database]. Washington, DC.
A-39
-------
APPENDIX B. STUDY DESIGN CHARACTERISTICS OF
TETRACHLOROETHYLENE EXPOSURE AND CANCER
EPIDEMIOLOGICAL STUDIES
B.I. Cohort Studies
Tetrachloroethylene cohort studies have been organized by occupational sector with
summary of study design, exposure-assessment approach and statistical methodology. Table B-l
provides summaries of the study characteristics of each paper or group of papers.
B.I.I. Dry Cleaner and Laundry Worker Studies
B.l.1.1. Andersen et al. (1999)
Andersen, A.; Barlow, L.; Engeland, A.; Kjaerheim, K.; Lynge, E.; Pukkala, E.
(1999). Work-related cancer in the Nordic countries. Scand J Work Environ Health,
25,1-116. http://www.ncbi.nlm.nih.gov/pubmed/10507118
Summary: This cohort study examined work-related cancer in Denmark, Finland,
Norway, and Sweden. The Danish, Finnish, and Norwegian cohorts were identified through the
1970 Censuses, and the Swedish cohort was ascertained through the 1960 Census. Each of the
country-specific cohorts consisted of those individuals who were between 25 and 64 years of age
and still alive on January 1, 1971. Overall, the four cohorts included 10,101,711 people, of
which 2,346,134 were Danish, 2,115,691 were Finnish, 1,792,817 were Norwegian, and
3,847,069 were Swedish. Follow-up began in 1971 and ended with death, emigration, or end of
follow-up, whichever came first. The follow-up protocol for each country was as follows:
Denmark: 1971-1987 and linked for identifying deaths and/or emigration with the Central
Population Register; Finland: 1971-1990 and linked with Statistics Finland; Norway:
1971-1991 and linked with the Central Population Register; and Sweden: 1971-1989 and linked
with the cause-of-death register. Incident cancer cases were obtained through the national cancer
registries in each country. Of the more than one million cases, 228,456 were from Denmark,
197,305 were from Finland, 207,068 were from Norway, and 397,433 were from Sweden.
The censuses contained information on demographics as well as occupations and
industries that was obtained through descriptions provided by the heads of households for all
economically active members. These descriptions were then coded according to the Nordic
Occupational Classification in Finland, Norway, and Sweden. Denmark coded their inhabitants'
occupations according to their own standards. The researchers then receded all jobs based on a
set of 54 common occupational groups based on Nordic Occupational Classification standards
B-l
-------
and included 1 group for those who were economically inactive at the time of the census. This
scheme was used to evaluate occupation as a proxy for exposure. Group number 51, Code 95
consisted of launderers and dry cleaners and included 29,333 (0.3%) cohort members. There
were 9,873 (0.4%) within the Danish cohort, 4,949 (0.2%) within the Finnish cohort, 4,061
(0.2%) within the Norwegian cohort, and 10,450 (0.3%) within the Swedish cohort. This
occupational group contributed a total of 519,844 person-years, which were distributed as
follows: 159,156 in Denmark, 94,302 in Finland, 78,086 in Norway, and 187,580 in Sweden.
Overall, there were 3,254 incident cancer cases among the laundering and dry-cleaning worker
population. Of these, 964 occurred in Denmark, 429 in Finland, 545 in Norway, and 1,316 in
Sweden.
Age-standardized incidence ratios and their corresponding 95% confidence intervals
(CIs) were calculated for launderers and dry cleaners for all cancer sites, and for cancers of the
pancreas, lung and bronchus, cervix, kidney, nervous system, and lymphopoietic tissues
(non-Hodgkin lymphoma and multiple myeloma), stratified by country. Expected numbers of
cases were determined using the cancer incidence rates for each country's study population. A
Poisson distribution was assumed for all CIs whose standardized incidence ratios were calculated
with 100 cancer cases or fewer. Strengths of the study include the compulsory nature of the
1970 Census in all four countries, the 95-99% accuracy in cancer incidence data depending on
the country, and the linkage of census, mortality, and emigration, and cancer incidence data
based on personal identifiers. Limitations of the study include the lack of lifetime occupational
histories and the inability to differentiate between launderers and dry cleaners in the analyses.
B.l.1.2. Blair et al. (2003)
Blair, A.; Petralia, S. A.; Stewart, P. A. (2003). Extended mortality follow-up of a
cohort of dry cleaners. Ann Epidemiol, 13, 50-56. http://dx.doi.org/10.1016/S1047-
2797(02)00250-8
Summary: This study extended the follow-up of an earlier cohort (Blair et al., 1990) of
dry-cleaning workers for the purpose of providing more information on mortality and cancer risk
among those occupationally exposed to dry-cleaning solvents. The cohort was identified through
the dues records from the Local No. 161 (St. Louis) of the Laundry, Dry Cleaning, and Dye
House Workers' International Union. This particular union was composed entirely of dry
cleaners. The cohort consisted of male and female members who entered the union between
1945 and 1978, worked for at least 1 year, and had demographic information (race, sex, date of
birth, date of entry) available. Of the 11,062 union members identified, 5,369 met inclusion
criteria. Blair et al. (1990) followed-up subjects through January 1979, and Blair et al. (2003)
began in January 1979 and ended in December 1993, an addition of 14 years. Person-years were
B-2
-------
calculated starting at entry to the union or in 1948, whichever came later, and ended with death
or December 1993, whichever came first. Deaths were identified through the National Death
Index and all were coded according to the International Classification of Diseases (8th revision)
standards.
Dues records were used to obtain demographic and employment information. When
demographics were not accessible through this mechanism, attempts were made to identify this
information through driver's license records, social security files, health care finance
administration records, and credit bureaus (Blair et al., 1990). Tasks within the dry cleaning
occupation were used to assess exposure. Exposure indices were determined for four different
categories of jobs within the dry cleaning occupation: (1) cleaners who run the machines and
handle the clothes were deemed to have the highest exposure and assigned a time-weighted index
of 40 (based on an 8-hour day); (2) pressers, sewers, and counter workers who worked where the
dry cleaning occurred and were deemed to receive the bulk of their exposures through the air and
were assigned a time-weighted average exposure index of 7; (3) counter workers who were
employed at pick-up stations were determined to have minimal exposure and assigned an index
of 0; (4) maintenance workers who had high, short-term exposures were assessed to have a
time-weighted average exposure of 7. Although the authors did not report the numbers of
exposed within each category, there were 220 deaths from cancer among those with little/no
exposure (index of 0) and 316 deaths from cancer among those with medium/high exposure
(index of 7 or 40). Standardized mortality ratios (SMRs) and 95% CIs were estimated for all
causes of death, stratified by the initial follow-up period, the extended follow-up period, and the
full follow-up period. Expected SMRs were determined using U.S. population 5-year age and
mortality statistics. SMRs were also presented by exposure score—little or no exposure and
medium/high exposure—and by date of union membership, before 1960 or after 1960, a time
corresponding to widespread use of tetrachloroethylene for cleaning clothes. At the end of the
follow-up period in 1993, 43.8% of the cohort members were identified as deceased. The
authors did not report any strengths of their methodology; limitations include the lack of
information on potential confounders, the lack of detail on job history within the industry, the
study's inability to determine what proportion of their cohort were exposed to
tetrachloroethylene, the inability to attribute risk to occupational versus lifestyle factors, and
potential misclassification due to the use of death certificates in determining the cause of death,
and the notably limited ability to examine liver cancer due to disease misclassification biases.
B-3
-------
B.l.1.3. Cano and Pollan (2001)
Cano, M. I. and Pollan, M. (2001). Non-Hodgkin's lymphomas and occupation in
Sweden. Int Arch Occup Environ Health, 74, 443-449.
http://dx.doi.org/10.1007/s004200100248
Summary: This study used a historical cohort design to follow 2,881,315 Swedish men
and women from 1971 to 1989 to determine whether workers associated with certain occupations
had a higher risk of non-Hodgkin lymphoma. The researchers conducted the follow-up by
linking the Swedish cancer environment register, which provided information on cancer cases, as
well as demographic variables from the 1960 and 1970 Censuses, with a population register,
which housed information on occupation and residence in 1970 and occupation in 1960.
Incidence rate numerators were calculated using the Swedish cancer environment register, while
rate denominators were calculated using the population register. Person-years were counted
starting in 1971 until either that individual's date of death or 1989. A total of 278 occupations in
men and 263 occupations in women were counted.
A total of 7,610 non-Hodgkin lymphomas were reported in the study cohort, with
5,391 cases in men and 2,219 in women. Among male cases, 11 fell within the launderers and
dry cleaners occupational category (Code 943), and 22 were considered textile workers
(Code 701). There were no women classified as launderers/dry-cleaners or textile workers in
this study. The analysis consisted of the calculation of age-standardized incidence rates,
standardized incidence ratios, as well as relative risks, and their associated 95% CIs.
Age-standardized incidence rates were developed for each occupation for the entire time period
and used the standard European population as a reference. Standardized incidence ratios were
calculated for the 10 main occupational sectors, as well as each occupation, stratified by 5-year
age groups and 5-year calendar-year period. Log-linear Poisson models were used to compare
occupations against the overall cohort, adjusted for geographical area. Relative risks for each
occupation and of the 10 main occupational sectors were also calculated and adjusted for age,
period, and geographical category and using the other occupations in the general cohort as a
reference. A strength of this study is its inclusion of 1960 Census data leading to an improved
definition of exposure; a limitation was its lack of control for other potential confounders beyond
demographic information.
B-4
-------
B.l.1.4. Chow et al. (1995)
Chow, W. H.; McLaughlin, J. K.; Malker, H. S.; Linet, M. S.; Weiner, J. A.; Stone,
B. J. (1995). Esophageal cancer and occupation in a cohort of Swedish men. Am J
Ind Med, 27, 749-757. http://dx.doi.org/10.1002/aiim.4700270509
Summary: This study aimed to generate and refine hypotheses regarding occupational
risks for esophageal cancer by examining the esophageal cancer incidence by occupation and
industry in Sweden. The cohort was identified from the Swedish Cancer Environment Registry,
which linked employment and cancer information for all individuals registered in the 1960
Census and the National Swedish Cancer Registry. The linkage was performed using personal
identifiers. The authors do not report the final cohort size. The follow-up period was from 1961
to 1979. There were three cases of laundry workers, though exposure prevalence could not be
estimated because the authors did not report the total laundry worker population. Standardized
incidence ratios were calculated for the entire time frame, with expected numbers of cases based
on the 5-year birth cohort- and sex-specific rates for esophageal cancer in the general Swedish
population during that same time period. Only those occupations that had at least
500 individuals were examined. Statistical significance was evaluated assuming a Poisson
distribution. A strength of this study is its extended follow-up period, and a limitation was the
small number of exposed laundry worker cases.
B.I.1.5. Ji et al. (2005a, b), Ji and Hemminki (2005a, b, c), Ji and Hemminki (2006)
Ji, J.; Granstrom, C.; Hemminki, K. (2005a). Occupation and bladder cancer: A
cohort study in Sweden. Br J Cancer, 92,1276-1278.
http://dx.doi.org/10.1038/si.bic.6602473
Ji, J.; Granstrom, C.; Hemminki, K. (2005b). Occupational risk factors for kidney
cancer: A cohort study in Sweden. World Journal of Urology, 23, 271-278.
http://dx.doi.org/10.1007/s00345-005-0007-5
Ji, J. and Hemminki, K. (2005a). Occupation and upper aerodigestive tract cancers:
A follow-up study in Sweden. J Occup Environ Med, 47, 785-795.
http://dx.doi.org/10.1097/01.iom.0000165798.28569.b5
Ji, J. and Hemminki, K. (2005b). Occurrences of leukemia subtypes by
socioeconomic and occupational groups in Sweden. J Occup Environ Med, 47,1131-
1140. http://dx.doi.org/10.1097/01.iom.0000174302.63621.e8
Ji, J. and Hemminki, K. (2005c). Variation in the risk for liver and gallbladder
cancers in socioeconomic and occupational groups in Sweden with etiological
implications. Int Arch Occup Environ Health, 78, 641-649.
http://dx.doi.org/10.1007/s00420-005-0015-l
B-5
-------
Ji, J. and Hemminki, K. (2006). Socioeconomic/occupational risk factors for
lymphoproliferative diseases in Sweden. Ann Epidemiol, 16, 370-376.
http://dx.doi.0rg/10.1016/i.annepidem.2005.09.002
Summary: These six studies used a cohort obtained through the Swedish Family-Cancer
Database to examine the potential relationship between occupation and various cancers. The
database linked national censuses (1960, 1970, 1980, and 1990), mortality data, cancer incidence
data, and an administrative family register. The database was updated at two different time
periods (2002 and 2004) with information from the Swedish Cancer Registry. The update in
2002 covered the period 1961-2000, and the update in 2004 covered the period 1958-2002. The
cohort consisted of 1,644,958 men who were employed at the time of the 1960 Census and
1,154,091 women who were employed at the time of the 1970 Census. Follow-up began at
immigration or at one of the following dates: January 1961 for those in the 1960 Census, January
1970 for those in the 1970 Census or for those who reported the same occupation in both the
1960 and 1970 Censuses, or January 1980 for those who reported the same occupation in all
three censuses. Follow-up ended with cancer diagnosis, death, emigration, or on December
2000, whichever came first. Occupation was assessed as a proxy for exposure, with relevant
census information (employment status, job title, work industry) coded according to Nordic
Occupational Classifications. These codes corresponded to 53 occupational groups, which
included launderers and dry cleaners. Overall, there were 9,255 (0.6%) male and 14,974 (1.3%)
female launderers and dry cleaners. Standardized incidence ratios were calculated for each
occupation in each census subcohort (1960, 1960-1970, and 1960-1970-1980). Expected
numbers of site-specific cancer were estimated from 5-year-age, 10-year-period, and 6 group
socioeconomic status-specific standard incidence rates. Corresponding 95% CIs were estimated
assuming a Poisson distribution. Strengths of these studies include their population-based
design, extended follow-up period, and utilization of three different censuses. Limitations to the
studies include the inability to directly control for smoking as a potential confounder, the
exception was bladder cancer (Ji et al., 2005a), the low power resulting from a small proportion
of exposed cohort members that reported the same occupation in more than one census, and the
high proportion of women without occupational data in the 1960 Census, which made
comparisons with more than one census difficult. Also, Ji et al. (2005b) and Ji and Hemminki
(2006) limited their study to those over the age of 30 years, which may have biased the external
validity of the study because the findings could not be generalized to the <30 population.
Ji et al. (2005a) examined the relationship between occupation and first primary bladder
cancers using the information from the 2002 database update. Overall, 24,041 men and
3,405 women developed bladder cancer, which included 157 male launderers and dry cleaners
from the 1960 Census. There were 67 cases among male launderers and dry cleaners in both the
B-6
-------
1960 and 1970 Census, and 19 cases among male launderers and dry cleaners in the 1960, 1970,
and 1980 Censuses. The results for female launderers or dry cleaners were not reported.
Two different standardized incidence ratios were calculated for occupations stratified by gender.
The first estimate was adjusted for age and period, and its corresponding 95% CI was adjusted
for age, period, and socioeconomic status. In order to account for the effect of smoking on
bladder cancer, all standardized incidence ratios were divided by 35%, which was based on a
difference in bladder and lung cancer risks for smoking 20 cigarettes per day developed by the
International Agency for Research on Cancer (IARC, 2004b). This estimated a second,
smoking-corrected standardized incidence ratio as well as a smoking-corrected 95% CI. All
estimates were stratified by gender.
Ji et al. (2005b) examined the relationship between occupation and first primary kidney
cancer, including parenchymal cancer, pelvic cancer, and unspecified cancer. This study utilized
the data from the 2002 update. There were 61 cases (51 renal parenchyma, 7 renal pelvis, and
3 unspecified) among male launderers from the 1960 Census and 92 cases (79 renal parenchyma,
6 renal pelvis, and 7 unspecified) among female launderers from the 1970 census. There were
26 cases (21 parenchyma, 3 renal pelvis, 2 unspecified) that occurred among women who
reported being launderers in both the 1960 and 1970 Censuses and 3 cases (1 parenchyma,
1 renal pelvis, 1 unspecified) among those who reported being launderers in the 1960, 1970, and
1980 Censuses. All standardized incidence ratio estimates were adjusted for age, time period,
and socioeconomic status and stratified by gender.
Ji and Hemminki (2005a) assessed risk factors for first primary upper aerodigestive tract
(lip, tongue, month, pharynx, and larynx) cancers using the 2002 database update. There were
83 cases (9 lip, 9 tongue, 13 mouth, 24 pharynx, 28 larynx) that occurred among male launderers
in the 1960 Census, 32 cases (2 lip, 3 tongue, 6 mouth, 10 pharynx, 11 larynx) among male
launderers who were in both the 1960 and 1970 Censuses, and 13 cases (0 lip, 2 tongue, 2 mouth,
6 pharynx, 3 larynx) among male launderers who were in the 1960, 1970, and 1980 Censuses.
Among the female launderer population in the 1970 Census, there were 30 upper aerodigestive
cancers (10 lip, 2 tongue, 7 mouth, 6 pharynx, 5 larynx). The results for female launderers in
multiple censuses were not reported. All standardized incidence ratio estimates were adjusted
for age, period, and socioeconomic status and stratified by gender.
Ji and Hemminki (2005b) examined socioeconomic and occupational risks on leukemia
by histologic type after the database's 2004 update. This study was limited to cohort members
aged 31 or older and diagnosed with primary leukemia, including chronic lymphatic leukemia
(CLL), acute myeloid leukemia (AML), chronic myeloid leukemia (CML), and polychythemia
vera (PV). There were 47 cases of leukemia, of which 19 were CLL, 7 AML, and 5 CML,
among male launderers and dry cleaners from the 1960 cohort and 80 cases of leukemia, of
B-7
-------
which, 32 were CLL, 20 AML, and 2 CML, among female launders and dry cleaners from the
1970 cohort. The results for men and women employed as launderers and dry cleaners for more
than one census were not reported. Standardized incidence ratios were calculated for six socio-
economic groups reported in the 1960 Census, adjusted for age and period, as well as for each
occupational group, adjusted for age, period, and socioeconomic status. All estimates were
stratified by gender.
Ji and Hemminki (2005c) examined the relationship between occupation, socioeconomic
status, and liver and gallbladder cancers using information from the 2002 database update.
Overall, 7,620 men and 4,041 women developed liver and gall bladder cancer (4,211 men and
1,126 women with primary liver cancer), which included 53 male launderers and dry cleaners
from the 1960 Census and 86 female launderers and dry cleaners from the 1970 Census (25 men
and 25 women with primary liver cancer). All standardized incidence ratio estimates were
adjusted for age and socioeconomic status and stratified by gender.
Ji and Hemminki (2006) looked at the socioeconomic and occupational risks for
lymphoproliferative diseases, including non-Hodgkin lymphoma, chronic lymphatic leukemia,
and multiple myeloma, after the database's 2004 update. This study was limited to cohort
members aged 31 and older who were diagnosed with primary lymphoproliferative diseases,
including non-Hodgkin lymphoma (NHL), CLL, and multiple myeloma (MM). There were
59 cases of NHL, 19 cases of CLL, and 29 cases of MM among male launderers and dry cleaners
from the 1960 Census. Among female launderers and dry cleaners, there were 67 cases of NHL,
18 cases of CLL, and 36 cases of MM in the 1960 Census, 64 cases of NHL, 32 cases of CLL,
and 31 cases of MM among those in the 1970 Census, and 12 cases of NHL, 8 cases of CLL, and
9 cases of MM among those from both the 1960 and 1970 Censuses. Standardized incidence
ratios were calculated for six socio-economic groups reported in the 1960 Census, adjusted for
age and period, as well as for each occupational group, adjusted for age, period, and
socioeconomic status. All estimates were stratified by gender.
B.l.1.6. Lindbohm et al. (2009)
Lindbohm, M. L.; Sallmen, M.; Kyyronen, P.; Kauppinen, T.; Pukkala, E. (2009).
Risk of liver cancer and exposure to organic solvents and gasoline vapors among
Finnish workers. Int J Cancer, 124, 2954-2959. http://dx.doi.org/10.1002/iic.24309
Summary: This cohort study of economically active Finns born between 1906 and 1945
examined the relationship between job title reported on the 1970 Census and primary liver cancer
incidence between 1971 and 1995. The cohort consisted of 1.2 million economically active men
and women born between 1906 and 1945 who participated in the Finnish Population Census of
1970. There were 2,474 liver cancers diagnosed between 1971 and 1995 of which 9 occurred in
B-8
-------
launderers (2 male, 7 female). Exposure was defined as longest held occupation reported on the
1970 Census and assigned to subjects using industry code (850 for launderers) or as cumulative
exposure to "organic solvents" using the Finnish job exposure matrix (FINJEM) for every 5-year
birth cohort and 5-year calendar period. The exposure for each birth cohort was assumed to start
in the year when the average age of the birth cohort was 20 or at 65 years of age, whichever
came first, because occupational histories were not available. The annual average exposure was
the product of the proportion of exposed and the mean level of exposure in that occupation. A
lag period was incorporated in the cumulative estimate by omitting exposure from the 10 last
years. For launderers, averages of 10 ppm in 1945-1959 and 5.3 ppm in 1960-1984 were
assumed. In Finland, for the later time period, this would likely be for tetrachloroethylene
because this was the predominate solvent used in dry cleaning, accounting for roughly 85% of all
solvents at that time (Kauppinen et al., 2009; Lvnge et al., 2006; Johansen et al., 2005).
Standardized incidence ratios and 95% CIs were calculated by gender using a Poisson
regression, with expected number of cancer cases estimated using site-specific cancer incidence
rates of the larger Finnish population. Statistical analyses controlled for alcohol consumption,
smoking, and socioeconomic status. A strength of the study includes the use of census
information and the ability of statistical analysis to account for potential confounding from
smoking, alcohol consumption, and socioeconomic status. The few liver cancer deaths and
low-exposure prevalence and the classification of exposure based solely on census-reported
information rather than a full lifetime of employment are limitations.
B.I.1.7. Lynge and Thygesen (1990), Lynge et al. (1995)
Lynge, E; Thygesen, L. (1990) Primary liver cancer among women in laundry and
dry-cleaning work in Denmark. Scand J Work Environ Health 16(2):108-112.
Lynge, E; Carstensen, B; Anderson, O. (1995) Primary liver cancer and renal cell
carcinoma in laundry and dry-cleaning workers in Denmark. Scand J Work
Environ Health 21(4):293-295.
Summary: These studies used a retrospective cohort design to examine the relationship
between work in dry-cleaning shops where tetrachloroethylene was the main solvent used and
cancer in Denmark. The cohort consisted of 10,600 Danish men and women aged 20 to 64 years
who were registered in the 1970 Census as "laundries, cleaning and dyeing." This encompassed
industry Code 860 (laundries, cleaning, and dyeing) and occupational Codes 411 (laundry
worker, ironer) and 380 (factory hand), as well as those who reported themselves as self-
employed or family workers. There were 2,434 (23%) self-employed dry cleaners or launderers,
830 (7.8%) family workers, 6,837 (64.5%) laundry workers or ironers, and 499 (4.7%) factory
hands. Overall, there were 2,886 laundry and dry-cleaning shops in Denmark in 1970, of which
B-9
-------
695 were known dry-cleaning and dyeing shops where dry cleaning was the predominant
activity.
Lynge et al. (1990) studied the cancer incidence within the cohort during a 10-year
follow-up after the 1970 Census. The census data were linked to the Danish Cancer Registry for
the period 1970 to 1980, and 24 cancer sites were examined. There were a total of 510 observed
cancer cases. Standardized incidence ratios and their corresponding 95% CIs were calculated
assuming a Poisson distribution if the observed number of cases was <30 and a normal
distribution if the number was >30. Expected numbers were estimated by multiplying the
person-years at risk within each 5-year age group with the site-specific incidence rates that were
estimated for the full 1970 Census cohort. The authors do not report any strengths of their
methodology; limitations include the study's inability to separate laundries from dry-cleaning
shops and the lack of a sufficient period for cancer latency.
Lynge and Thygesen (1995) used a nested case-control study to differentiate the laundry
workers from the dry cleaners in their examination of liver and renal cell carcinoma within this
cohort. The cohort was followed from 1970 through 1987 for death, emigration, and incident
cancer. During this period, there were a total of 17 liver cancer cases and 16 renal cell
carcinoma cases. Controls were randomly selected from within the cohort and matched to cases
on gender, 5-year age group, and occupation. In order to be included in the study, controls were
required to be alive and living in Denmark at the time of the case's diagnosis; no other exclusion
criteria were adopted. There were five controls matched to each case, and the final sample
consisted of 33 cases (17 liver and 16 renal cell carcinoma) and 165 controls (85 liver and
80 renal cell carcinoma). Occupation was assessed as a proxy for exposure. The identification
numbers of each of the 198 participants were unencrypted to obtain personal addresses, which
were then used to retrieve the original census forms. These forms contained descriptions of the
occupations and workplaces and allowed the researchers to recede each individual as either a
launderer or a dry cleaner. The authors do not state if this assessment occurred blindly. Overall,
none of the liver cancer cases, 20 (24%) liver cancer controls, 3 (18%) renal cell carcinoma
cases, and 20 (29%) renal cell carcinoma controls worked as dry cleaners in 1970. Conditional
logistic regression was used to calculate relative risks and their corresponding 95% CIs. A
strength of this study is its ability to examine dry cleaners separately from laundry workers.
Limitations include the low-exposure prevalence, the lack of adjustment for alcohol and smoking
as possible confounders, the classification of exposure based solely on a census form rather than
a full lifetime of employment, and the use of controls with diseases potentially associated with
dry-cleaning exposure.
B-10
-------
B.l.1.8. Pukkala et al. (2009)
Pukkala, E.; Martinsen, J.; Lynge, E.; Gunnarsdottir, H.; Sparen, P.;
Tryggvadottir, L.,... Kjaerheim, K. (2009). Occupation and cancer - follow-up of
15 million people in five Nordic countries. Acta Oncol, 48, 646-790.
http://dx.doi.org/10.1080/02841860902913546
Summary: This cohort study, the Nordic Occupational Cancer Study, of 15 million
subjects, aged 30-64 years in the 1960, 1970, 1980/1981, and/or 1990 Censuses in Denmark
Finland, Iceland, Norway, and Sweden assessed cancer incidence through 2005 using national
cancer registries. Occupational title as recorded on census records was used as a surrogate for
exposure, and the investigators examined 54 broad occupational categories identified from
Nordisk Yrke Klassifisering (NYK) and the International Standard Classification of Occupations
(ISCO). In total, 43,496 dry-cleaners and laundry workers (n = 8,744 men, n = 34,752 women),
defined by NYK and ISCO Codes 95 (http://astra.cancer.fi/NOCCA/). Both tetrachloroethylene
and trichloroethylene were potential exposures to dry cleaners and launderers.
Tetrachloroethylene used in Finland was less than in the other Nordic countries in 1975-1994
(Kauppinen et al., 2009). A future effort of this project is an examination of cancer incidence
and 20 agents, including tetrachloroethylene and trichloroethylene (Kauppinen et al., 2009).
Follow-up began on January 1 of the year after the first available census, and
person-years were counted until the date of emigration, death, or to December 31 of the
following years: 2003 (subjects from Denmark and Norway), 2004 (subjects from Iceland), 2005
(subjects from Finland and Sweden). The study examined 49 cancer sites and 27 diagnostic
subgroups during the 13-45 year follow-up period. Standardized incidence ratios and their
corresponding 95% CIs were calculated for each site-specific cancer and occupational title with
expected number of site-specific cancers calculated from separate countrywide incidence rates.
Statistical analyses did not include examination of duration of employment, in this case,
appearing as a dry-cleaner or laundry worker on more than one census.
This is a large study with follow-up to account for a cancer latent period of >15 years,
and a strength is linkage with national population registries and cancer registries. The large
number of dry-cleaners and laundry workers is an advantage; however, occupational title as
dry-cleaner and laundry worker is broad, with subjects having differing potential to exposure
intensities and to multiple solvents. Despite the large number of subjects with occupational title
of dry-cleaner and laundry worker, statistical power may be compromised from the low-level
detail of the exposure-assessment approach for these reasons.
B-ll
-------
B.l.1.9. Ruder et al. (2001,1994), Calvert et al. (2011)
Ruder, A. M.; Ward, E. M.; Brown, D. P. (1994). Cancer mortality in female and
male dry-cleaning workers. J Occup Med, 36, 867-874.
http://www.ncbi.nlm.nih.gov/pubmed/7807267
Ruder, A. M.; Ward, E. M.; Brown, D. P. (2001). Mortality in dry-cleaning workers:
An update. Am J Ind Med, 39,121-132. http://dx.doi.org/10.1002/1097-
0274(200102)39:2<121::AID-AJIM1000>3.0.CO;2-H
Calvert, G. M.; Ruder, A. M.; Peter sen, M. R. (2011). Mortality and end-stage renal
disease incidence among dry cleaning workers. Occup Environ Med.
http://dx.doi.org/10.1136/oem.2010.060665
Summary: This retrospective cohort study examined the relationship between
occupational exposures and mortality in a cohort of dry cleaners and updates earlier studies
(Calvert et al., 2011; Ruder etal., 2001, 1994). An examination of end-stage renal disease
incidence (ESRD) was presented in (Calvert et al., 2011) using the Renal Management
Information System (REMIS) maintained by the U.S. Centers for Medicare and Medicaid
Services. The cohort was obtained from union dry-cleaning records in California, Illinois,
Michigan, and New York and included anyone employed for at least 1 year prior to 1960 in a
dry-cleaning shop that used tetrachloroethylene. Attempts were made to verify the solvent
exposure records with visits to the shops themselves. Follow-up of vital status was to 1990
(Ruder etal., 1994), 1996 (Ruder et al., 2001), and 2004 (Calvert et al., 2011). Of the cohort of
1,704 workers in the current follow-up analysis, 618 (36%) worked only in shops that used
tetrachloroethylene as the primary solvent cleaner, and 1,086 (64%) worked at shops where the
primary cleaner (tetrachloroethylene or Stoddard solvent) could not be verified or where other
solvents were known or suspected to be used instead of tetrachloroethylene. Calvert et al. (2011)
found four subjects in Ruder et al. (2001) had missing birthdates, and these subjects were not
included in their latest cohort follow-up. Calvert et al. (2011), additionally, was less successful
than Ruder et al. (2001) at obtaining causes of deaths; 8% of deaths were not obtained in the
latest follow-up compared to 3% in Ruder et al. (2001). As of 2004, 322 deaths had occurred.
Tetrachloroethylene exposure was estimated by duration of employment in the dry-
cleaning shops (1 to 5 years or more than 5 years) and by latency periods (time since
first employment was less than 20 years or 20 or more years). Person-years were calculated
from either January 1, 1940, or after 1 year of employment in a unionized tetrachloroethylene
shop, whichever came later, through their death, the date they were lost to follow-up, or the end
of 2004, whichever came earlier. SMRs and their corresponding 95% CIs were calculated for
each cause of death in the full cohort, for selected causes of death by duration of employment
and time since first employment, and for selected causes of death by the tetrachloroethylene-only
subcohort (618 workers) and the mixed cohort (1,704 workers) separately. The expected
B-12
-------
number of deaths was estimated using national rates. SMRs and their 95% CIs for each of the
four regions were also estimated using both county and national rates, though these data were not
shown. The National Death Index was used to obtain information on deaths that occurred in the
cohort.
Subjects employed since 1977, the date REMIS was first available, were followed for
ESRD incidence to 2004. A total of 1,296 subjects—494 in the tetrachloroethylene
cohort—were followed with 30 incident cases of ESRD identified. Standardized incidence ratios
and their corresponding 95% CIs were calculated for each ESRD type by the
tetrachloroethylene-only subcohort and the mixed cohort separately. The expected number of
deaths was estimated using all incident cases of ESRD available in REMIS as the numerators
and U.S. Census data as the denominators.
A strength of this study is its estimation of tetrachloroethylene exposure based on
duration and intensity. A limitation of the study is likely exposure-measurement error introduced
through inability to update work histories after 1982, potentially underestimating duration, and
lack of information on exposure intensity. Both aspects would tend to result in nondifferential
bias that would dampen risk estimates. Additionally, a full latent period has not passed for the
cohort, only 2% of the cohort had died at the end of follow-up in 2004, and only one-half of the
cohort had a latent period of >20 years. This is also valid for analysis of ESRD incidence as the
latent period is less than that for mortality. Another limitation is the lack of individual subject
information on smoking and alcohol consumption as potential confounders, although the authors
noted that the estimates for certain cancers were higher than what they would be if smoking was
the only significant factor, and potential for multiple solvents exposures with subjects whose first
employment date was before 1960.
B.l.l.lO.Selden and Ahlborg (2011)
Selden, A. I. and Ahlborg, G. (2011). Cancer morbidity in Swedish dry-cleaners and
laundry workers: Historically prospective cohort study. Int Arch Occup Environ
Health, 84, 435-443. http://dx.doi.org/10.1007/s00420-010-0582-7
Summary: This study examined cancer incidence in a cohort of 9,440 Swedish dry-
cleaning workers launderers, dry cleaners, and pressers identified by employers as working in
laundries or dry-cleaning shops during 1973 and 1983 for a study of pregnancy outcomes
(Ahlborg, 1990a). In mid-1980, a questionnaire was mailed to all washing establishments
recorded in the Swedish Postal Address Registry. Of the 1,254 employers that received the
questionnaire, 475 (37.9%) of the employers responded to the questionnaire and identified
10,389 employees. Data from 14 companies were lost from the original study, leaving workers
from 461 companies for the cancer incidence study. The size of companies participating in the
B-13
-------
study varied from small family businesses to large establishments. In addition to seeking
information on employee identities, the questionnaire sought details of production volumes,
washing techniques, and details of any chemicals used; no tetrachloroethylene exposure
information was provided on individual employees. Study authors verified subjects fulfilled
inclusion criteria of Ahlborg (1990a) and this study, excluding subjects who did not fulfill
criteria.
Exposure assignment to three categories was carried out using company-provided
information on type of business: tetrachloroethylene or PER subgroup, laundry subgroup, or
other. The PER subgroup was composed of employees of dry cleaner and laundries with a
proportion of dry-cleaning with tetrachloroethylene only; the laundry subgroup included
employees in laundry establishments, while the "other" subgroup contained employees of
businesses using a combination of chemicals for dry cleaning in addition to tetrachloroethylene
(chlorofluorocarbons, white spirit, naphtha, or trichloroethylene). Tetrachloroethylene had been
used in Sweden almost exclusively for dry-cleaning since the 1950s. Historical industrial
monitoring data indicated exposure levels on the order of 100-200 mg/m3 in the 1970s, with
tetrachloroethylene concentrations decreasing by 50% from 1980 to 1985, and an 8-hour TWA
rarely exceeding 50 ppm (Selden and Ahlborg, 2011; Johansen et al., 2005; Ahlborg, 1990b).
Of the 10,389 subjects reported by the companies and who were employed at least
1 month, 677 were excluded for either not fulfilling the original inclusion criteria or other
reasons and 272 were lost in the identification process. Overall, 9,440 subjects (2,810 men and
6,630 women) were followed for cancer incidence from January 1, 1985 to until 85 years of age,
death, emigration or to December 31, 2006, whichever can first. A total of 1,106 incident
cancers were identified from the Swedish Cancer Registry, 723 of which occurred in subjects
categorized in the PER exposure subgroup. Site-specific standardized incidence ratios and their
95% CIs were estimated using expected numbers of cancers estimated from cancer incidence
rates of the Swedish population. Additionally, SIR and 95% CIs are reported separately for each
exposure category, as well as by employment duration for subjects in the PER and laundry
categories.
This study differs from other included in this summary of Swedish or all-Nordic dry
cleaners and launderers in that it is based on employer-reported instead of census-reported
information. A strength of the study is the over 20-year follow-up. Also, the authors provide
some information to evaluate potential confounding from smoking and alcohol. Ahlborg (1990a,
b) collected smoking information on some of the women who also participated in the pregnancy
outcome study and reported a prevalence of daily smoking before conception of 66-70%, higher
than reported for women attending Swedish prenatal care centers in the early 1980s and for
national data (Selden and Ahlborg, 2011; Ahlborg and Bodin, 1991). The higher prevalence of
B-14
-------
smoking among women may potentially confound observations for smoking-related site-specific
cancers such as lung and bladder. With respect to alcohol consumption, a previous survey of
women in the pregnancy outcome study found a higher prevalence of "high" consumption
compared to women attending prenatal care centers (Selden and Ahlborg, 2011; Ahlborg and
Bodin, 1991; Ahlborg, 1990a). Selden and Ahlborg (2011) do not identify the average age of the
cohort. As some subjects in the cohort were included in a study of pregnancy outcomes, this is
not an "old" cohort, and expected cancer rates would be lower than for a cohort composed of
more aged subjects. Limitations identified by Selden and Ahlborg (2011) included lack of
quantitative exposure data, lack of a full occupational history, and low tetrachloroethylene
exposures. The expected lower background cancer rates and low tetrachloroethylene exposures
would lower the study's statistical power.
B.l.l.ll.Travier et al. (2002)
Travier, N.; Gridley, G.; De Roos, A. J.; Plato, N.; Moradi, T.; Boffetta, P. (2002).
Cancer incidence of dry cleaning, laundry and ironing workers in Sweden. Scand J
Work Environ Health, 28, 341-348. http://www.ncbi.nlm.nih.gov/pubmed/12432988
Summary: This cohort study examined cancer incidence in Swedish launderers, dry
cleaners, and pressers using a linked register that included the 1960 and 1970 Censuses, the
Swedish national cancer registry, and the national register of causes of death. Person years were
counted starting January 1971 until cancer diagnosis, death, or loss to follow-up December 1989,
whichever came first. All individuals with second primary neoplasms were excluded. The
authors did not report the total number of individuals included in the cohort. Launderers and dry
cleaners comprised Nordic Classification of Occupation Code 943, and pressers were Code 944;
laundry, ironing, and dyeing comprised Swedish Industrial Code 880 in 1960 and Code 9520 in
1970.
Exposure was classified into five categories: Group 1, launderer, dry cleaner, or presser
occupation employed in the laundry, ironing, or dyeing industries in either the 1960 or 1970
Censuses (543,036 person-years); Group 2, launderer, dry cleaner, or presser occupation
employed in the laundry, ironing, or dyeing industries in both the 1960 and 1970 Censuses
(46,934 person-years); Group 3, launderer, dry cleaner, or presser occupation employed in other
industries at the time of both censuses (18,960 person-years); and, Group 4, other occupational
titles employed in laundry, ironing, or dyeing industry (13,395 person-years); and, Group 5, not
employed in relevant industries or occupations during both censuses (69,540,184 person-years).
Multivariable Poisson regressions were used to calculate the relative risks and 95% CIs
of cancer for each category of exposure, adjusted for age, calendar period, geographic region,
urban setting, and gender. These analyses were also stratified by gender and adjusted for age,
B-15
-------
calendar period, geographic region, and urban setting. Travier et al. (2002) further assessed
temporal changes in solvent use, portraying relative risks by age in 1960 and noted subjects
under 40 years of age in 1960 presumably used mainly tetrachloroethylene and carbon
tetrachloride. A strength of this study is its detailed analysis of observed associations.
Limitations of this study include its low power and the use of self-reported occupational and
industrial codes to classify exposures.
B.l.1.12.Wilson et al. (2008)
Wilson, R.; Donahue, M.; Gridley, G.; Adami, J.; El Ghormli, L.; Dosemeci, M.
(2008). Shared occupational risks for transitional cell cancer of the bladder and
renal pelvis among men and women in Sweden. Am J Ind Med, 51, 83-99.
http://dx.doi.org/10.1002/aiim.20522
Summary: This study used a retrospective cohort design to examine if incident bladder
and renal pelvic cancers share similar occupational risk factors. It tested the hypothesis that
bladder and renal pelvis cancers were similarly elevated in each occupation and industry
category. The cohort consisted of 4,197,684 Swedish men and women employed during either
the 1960 or the 1970 Census and still living at the start of 1971. Person-years were counted
starting January 1, 1971, and ending with a cancer diagnosis, emigration, death, or December 31,
1989, whichever came first. Cancer information was obtained from the Swedish
Cancer-Environment Registry for the study period 1971 to 1989. Overall, there were a total of
70,083,912 person-years of follow up, with a mean time of 16.7 years. Within the cohort, there
were 1,374 incident renal pelvis cancers and 21,591 incident bladder cancers.
Occupation as noted on the 1960 and 1970 Censuses was assessed as a proxy for
chemical exposures, including tetrachloroethylene. Job titles reported in the censuses were
coded according to the National Swedish Classification of Occupations and Industries standards,
for which laundry and dry-cleaning workers were occupation Code 943, and the laundry, ironing,
and dyeing was industry Code 880 in 1960 and Code 9250 in 1970; 25,249 men and women
(0.6% of the cohort) were employed in this industry, for which there 110 observed bladder
cancer cases (55 female and 55 male) and 11 observed renal pelvic cancer cases (8 female and
3 male). A job exposure matrix was also used to assess exposure to indoor work and low
physical activity, among others.
Standardized incidence ratios and their associated 95% CIs were calculated for each
occupation and industry using expected site-specific cancer incidence of rates of the total
employed Swedish population. Strengths include the large sample size, use of well-validated
registries, adequate follow-up, and high case ascertainment. Limitations to the study include its
lack of adjustment for confounders including smoking, possible misclassification of exposure
B-16
-------
based on job title in 1 or 2 census years, and the lack of occupational history for each
participant.
B.1.2. Other Occupational Cohorts
B.l.2.1. Anttila et al. (1995)
Anttila, A.; Pukkala, E.; Sallmen, M.; Hernberg, S.; Hemminki, K. (1995). Cancer
incidence among Finnish workers exposed to halogenated hydrocarbons. J Occup
Environ Med, 37, 797-806. http://www.ncbi.nlm.nih.gov/pubmed/7552463
Summary: This cohort study assessed the incidence of cancer among employees who
were biologically monitored by the Finnish Institute of Occupational Health (FIOH) between
1965 and 1983 in comparison to the total Finnish population. The cohort consisted of workers
who had their blood and urine assessed for trichloroethylene, tetrachloroethylene, and
1,1,1-trichloroethane due to their employment in occupations that exposed them to hazardous
substances. Tetrachloroethylene was monitored in the blood of workers between 1974 and 1983,
and median blood tetrachloroethylene concentrations were 0.7 umol/L in males and 0.4 umol/L
in females. There were, on average, 3.2 blood tetrachloroethylene measurements per individual.
In addition to the measurement information, the FIOH database provided data on demographics,
date and time of sampling, workplace, solvent code, result, specific activity, and the laboratory in
which the sample was analyzed. Approximately 600 codes of workplaces or sampling
laboratories were included in the database.
Follow-up was conducted automatically with the Finnish Cancer Registry and began in
January 1967 or on the date of first measurement of the solvent, whichever was later, and ended
at emigration, death, or December 1992, whichever was first. Death and emigration were
ascertained through the Population Register Center; mortality was also followed-up using cause-
of-death data from the Central Statistical Office of Finland for the period 1956 to 1991. Of the
11,534 biological measurements taken between 1965 and 1983, 10,743 (93.1%) were linked to
personal identifiers, which corresponded to a total of 3,976 workers. After excluding those who
could not be completely identified or were not alive at the start of follow-up, the final sample
consisted of 3,974 individuals who contributed a total of 71,800 person-years. Follow-up time
averaged 18 years, with 27,547 person-years within the period 10-19 years after entry into the
cohort and 5,877 person-years within the period >20 years. There were 849 (21.4%) workers
monitored for exposure to tetrachloroethylene, and they contributed a total of 11,958
person-years.
The study examined 29 cancer sites during the 26-year follow-up period, which were
selected on the basis of their known or suspected association with the solvents. Of these, 8 sites
B-17
-------
(pancreas, lung/bronchus, cervix uteri, kidney, nervous system, non-Hodgkin lymphoma,
multiple myeloma, and all cancer sites) were specifically evaluated with respect to
tetrachloroethylene and contributed 31 observed cancer cases during the follow-up period of
1974 to 1992. Standardized incidence ratios and their corresponding 95% CIs were calculated
for all halogenated hydrocarbons, trichloroethylene, tetrachloroethylene, and
1,1,1-trichloroethane separately, with expected numbers of site-specific cancers calculated from
incidence rates of the Finnish population. Significance was evaluated with the Mantel-Haenszel
%2 test under the assumption that the observed cases followed a Poisson distribution. Exposure
duration was not examined for tetrachloroethylene subjects, given the few site-specific cancer
cases. Strengths of the study include its use of the same source for the calculation of observed
and expected cases, as well as its linkage with Finnish registries, which provided complete
ascertainment of death, emigration, and cancer incidence. Limitations include the study's low
power for its analysis of tetrachloroethylene, incomplete registration in earlier years, low
tetrachloroethylene concentrations [for comparison, Ferroni et al. (1992) reported median blood
tetrachloroethylene and atmospheric monitoring of dry cleaners of 874 umol/L and 15 ppm,
respectively], the study's inability to infer lifetime exposure based on blood tetrachloroethylene,
and the potential for multiple solvents exposures.
B.l.2.2. Boice et al. (1999)
Boice, J.; Marano, D.; Fryzek, J.; Sadler, C.; McLaughlin, J. (1999). Mortality
among aircraft manufacturing workers. Occup Environ Med, 56, 581-597.
http://dx.doi.0rg/10.1136/oem.56.9.581
Summary: This cohort mortality study conducted follow-up to evaluate cancer and other
diseases among aircraft workers. The cohort was identified through work history cards,
personnel files, and retirement records and consisted of individuals employed at Lockheed
Martin aircraft manufacturing factories for at least 1 year from January 1960 onwards. Those
with missing work history information or incorrect dates were excluded. The follow-up period
began January 1, 1960 or after 1 year of employment and ended with death, age 95 years, or
December 1996, whichever came first. The vital status of each cohort member at the end of the
follow-up period was obtained through a variety of methods, which included California death
tapes, the National Death Index, Pension Benefit Information Files, Social Security Death Index,
Health Care Financing Administration files, California Department of Motor Vehicles records,
employment work history cards, pension and retirement records, and obituaries from 1960 to
1996. Vital status could not be ascertained for 11,533 (15%) of cohort members and were
assumed to be alive. This assumption was examined using a random sample of 700 subjects and
demonstrated that approximately 95% of this sample was alive, and if representative of all
B-18
-------
subjects with missing vital status information, overall, lost to follow-up was estimated to be
0.7% of the cohort. The published paper lacks information to evaluate whether the random
sample was representative of all subjects lacking vital status. Of the 113,204 aircraft workers,
77,965 (68.9%) were included in the study and contributed a total of 1,889,795 person-years of
follow-up. The average follow-up per cohort member was more than 20 years.
Exposure was assessed through walk-through surveys of the closed factories or similar
factories, interviews with long-term employees, and industrial hygiene files or other historical
documents. From job code, job title, and job change information, the researchers were able to
group occupations with similar work activities, identifying job titles that may have indicated
chemical exposure. All administrative and technical jobs were classified as having "no
significant chemical exposure" and removed from the analysis. All factory-related jobs were
categorized based on a number of chemical exposures and a job exposure matrix, to assign
tetrachloroethylene exposure defined as routine (part of daily activities), intermittent (not routine
or on a daily basis), or minimal to no exposure. Limited data on tetrachloroethylene levels were
available, with few measurements before 1970 although tetrachloroethylene was used in vapor
degreasing starting in 1966 after trichloroethylene (TCE) was discontinued until the early 1990s
(Marano et al., 2000). Air sampling revealed that long-term air exposures to tetrachloroethylene
measures from 1987-1988 were 3 ppm (median) and 9.5 ppm (mean) [range: 0.06-27 ppm], and
short-term air exposures measured from 1978-1988 were 56 ppm (mean) and 17 ppm (median)
[range: 1.7-150 ppm]. Similarly, many factory workers were exposed to multiple substances.
For example, 4,421 (59%) subjects were also exposed to chromate, 2,262 (42%) were also
exposed to TCE, 5,830 (18%) were also exposed to mixed solvents, and 298 (24%) were also
exposed to asbestos. Among the factory worker subcohort, 2,631 (5.8%) employees were
assessed as having been exposed to routine levels, and another 3,199 (7.1%) subjects were
exposed to intermittent levels of tetrachloroethylene. The workers that were routinely exposed
contributed a total of 51,214 person-years at risk and had 476 observed deaths (all causes).
SMRs and their corresponding 95% CIs were calculated for routine exposed subjects
assuming the observed number of deaths followed a Poisson distribution. Expected numbers of
deaths among the Caucasian population were based on race, age, calendar year, and sex-specific
rates among the general population of California, while expected numbers of deaths among the
non-Caucasian population were based on the general population rates of the United States.
Poisson regression was used to estimate relative risks and their corresponding 95% CIs for a
combined grouping of routine and intermittent exposed subjects for duration of exposure,
adjusted for date of birth, date first employed, date of finishing employment, race, and sex. Tests
of linear trend were also performed to examine the potential effect of exposure duration for
subjects with routine or intermittent tetrachloroethylene exposure potential. In all analyses, the
B-19
-------
referent population consisted of all factory workers with incidental or no exposure. The
strengths of this study included its large size, although only 4% of the cohort was identified with
routine tetrachloroethylene exposure potential, extended (>37 years) follow-up period, exposure
assessment using a job exposure matrix, and use of an internal referent group in analyses
examining exposure duration. Limitations include the study's lack of adjustment for smoking,
lack of control for the healthy worker effect, and its finding that the assumption of living vital
status for approximately 7% of the cohort was incorrect. This bias would lead to an inflation of
the expected number of deaths due to the fact that they were assumed alive at the end of the
study period. Additionally, the inclusion of subjects with intermittent exposures who likely have
low-exposure potential may reduce the study's detection sensitivity. This may also introduce
differential bias in duration exposure-response analyses if intermittently exposed subjects had
longer employment duration than routinely exposed subjects. Marano et al. (2000) noted the
number of subjects with intermittent tetrachloroethylene exposure potential was 1.5 times larger
than the number of subjects assigned routine tetrachloroethylene exposure potential.
B.l.2.3. Bond et al. (1990; 1987)
Bond, G.; McLaren, E.; Cartmill, J.; Wymer, K.; Sobel, W.; Lipps, T.; Cook, R.
(1987). Cause-specific mortality among male chemical workers. Am J Ind Med, 12,
353-383. http://www.ncbi.nlm.nih.gov/pubmed/3674026
Bond, G.; McLaren, E.; Sabel, F.; Bodner, K.; Lipps, T.; Cook, R. (1990). Liver and
biliary tract cancer among chemical workers. Am J Ind Med, 18,19-24.
http://www.ncbi.nlm.nih.gov/pubmed/2378367
Summary: This nested case-control study, conducted as a follow-up to a cohort mortality
study (Bond et al., 1987), investigated liver and biliary cancer deaths of male employees working
at Dow Chemical's Midland/Bay City production, research, and headquarters units. The initial
cohort was identified through work history records and consisted of men and women employed
for 3 or more days between 1940 and 1982. Overall, the cohort consisted of 48,521 men and
women, of whom 96% were Caucasian, 77.7% were male, and 56.9% were paid by the hour
(Bond et al., 1987). Cases were identified through a review of death certificates and consisted of
all male, hourly employees who died between 1940 and 1982. Of the 6,259 cohort members
identified, 44 (0.7%) (11 primary liver cancer, 14 gallbladder/bile duct cancer, and
19 unspecified liver cancer) were considered eligible for and included in this study. The source
of death certificates was not identified by the authors, and it is not known whether they were
obtained through pension records or the National Death Index. Controls were randomly chosen
from among the cohort of male workers. Of the 21,437 hourly, male subjects, a random sample
B-20
-------
of 1,888 (8.8%) was selected as controls. Bond et al. (1990) do not identify if controls were
matched to cases on age, time period of first employment or end employment, or vital status.
Dow's work history records were used to determine the employee's work area
(administration, manufacturing, unknown manufacturing) as well as their possible exposure to
tetrachloroethylene and 10 other chemical agents. Overall, 6 (13.6%) cases and 213 (11.3%)
controls were potentially exposed to tetrachloroethylene during their time at Dow. The
Mantel-Haenszel method was used to estimate risk ratios for work areas and chemical exposures
separately, adjusted for birth year; Miettinen's method was used to calculate corresponding
95% CIs. Additional adjustment for period of hire produced similar results; as a result, only
those analyses controlling for birth year were presented. Individual analyses were conducted for
primary liver cancer and gall bladder/bile duct cancer separately. This study is of low prevalence
of tetrachloroethylene exposure, unable to determine whether the cancer was primary or
secondary in almost half of the cases, lacked information and statistical adjustment for
alcoholism as a potential confounder, is based on pension records or deaths known to the
employer, and deaths among nonpensioned employees are not used to identify deaths, used a
living control population, and may have misclassified exposure and disease given the use of
information on death certificates. Additionally, the study lacks description of source and process
for assigning tetrachloroethylene exposure potential. The authors do not report any strengths of
their methodology.
B.l.2.4. Chang et al. (2005; 2003), Sung et al. (2008; 2007)
Chang, Y.; Tai, C.; Yang, S.; Chen, C.; Shih, T.; Lin, R.; Liou, S. (2003). A cohort
mortality study of workers exposed to chlorinated organic solvents in Taiwan. Ann
Epidemiol, 13, 652-660. http://dx.doi.org/10.1016/S1047-2797(03)00038-3
Chang, Y.; Tai, C.; Yang, S.; Lin, R.; Sung, F.; Shih, T.; Liou, S. (2005). Cancer
incidence among workers potentially exposed to chlorinated solvents in an
electronics factory. J Occup Health, 47,171-180.
http://www.ncbi.nlm.nih.gov/pubmed/15824483
Sung, T.; Chen, P.; Jyuhn-Hsiarn Lee, L.; Lin, Y.; Hsieh, G.; Wang, J. (2007).
Increased standardized incidence ratio of breast cancer in female electronics
workers. BMC Public Health, 7,102. http://dx.doi.org/10.1186/1471-2458-7-102
Sung, T.; Wang, J.; Chen, P. (2008). Increased risk of cancer in the offspring of
female electronics workers. Reprod Toxicol, 25,115-119.
http://dx.doi.0rg/10.1016/i.reprotox.2007.08.004
Summary: After tetrachloroethylene and other substances were detected in the soil and
groundwater surrounding a closed Taiwanese electronics factory (Bechtel Environmental Inc.,
1990 and Target Environmental Services Inc., 1995), a series of retrospective cohort studies were
B-21
-------
conducted to examine the potential effect of employment in the factory. Strengths of all of these
studies are their large size and linkage with national data sets to assure that all cases had been
retrieved. Limitations include the use of employment in the factory as a proxy for exposure, and
subjects will have varying exposure potential to tetrachloroethylene.
Chang et al. (2005; 2003) identified the cohort through the Bureau of Labor and
Insurance's records for the years 1973 to 1997. To ensure completeness of the cohort, the
researchers also determined which employees had been hospitalized through labor-insurance
hospitalization data and obtained a list of those associated with the United Labor Association.
The cohort consisted of 86, 868 individuals (16,133 men and 70,735 women) who contributed a
total of 1,380,354 person-years. The average follow-up time was 14.3 years for men and
16.3 years for women; the average age of cohort members was 39.3 years. The cohort included
both white- and blue-collar workers.
Chang et al. (2003) linked the cohort with the National Mortality Database for their
13-year follow-up from 1985 to 1997. Person-years were counted starting when an individual
entered the cohort or on January 1, 1985, whichever came later, and ended with either that
person's death or December 31, 1997. The cohort experienced 1,357 deaths, 316 (24%) of
which were due to cancer. All cause mortality rate was 1.56%, and all cancer mortality rate was
0.36%. The analysis consisted of the calculation of SMRs. The number of deaths was stratified
by their underlying cause and compared with the expected numbers using the general Taiwanese
population as a reference. In order to better understand any dose-response relationships, the
cohort was stratified twice. First, it was stratified based on duration of employment: <1 year,
>1 year but <5 years, and >5 years. Then it was stratified based on the calendar year:
1985-1990 and 1991-1997. Duration of employment consisted of the period of employment
between the start and end of labor insurance coverage. Assumptions were made regarding the
duration of employment for those individuals with missing data. Limitations include reliance on
mortality rates from registration data sources, too brief of a follow-up time to allow for a
sufficient cancer latent period, data on employment were incomplete, and the cohort was very
young despite the mortality endpoint.
Chang et al. (2005) examined the cancer incidence from 1968 to 1992 by linking the
cohort with the National Cancer Registry, National Mortality Registry. Follow-up time was
calculated between the latter of employment start date or January 1, 1979, until the first of cancer
diagnosis, death, or December 31, 1997, with assumptions made regarding duration for those
with missing start or end dates. Overall, 998 individuals developed cancer. Standardized
incidence ratios were calculated comparing this exposed cohort to incidence rates in the general
population of Taiwan by age, calendar year, and sex. Latency periods of <3 months, 6 months,
and 1, 5, and 10 years were used. Trends were examined by duration of employment (<1 year, 1
B-22
-------
through <5 years, 5 through <10 years, and 10+ years) and period of employment (1979-1984,
1985-1990, and 1991-1997). Limitations include the study's reliance on registration data,
which meant that exposure could not be quantified, and the results were not adjusted for potential
confounders, such as smoking, alcohol consumption, reproductive history, or diet. The lack of
company personnel records prevented verification of the completeness of cohort identification,
as well as the need by the author's to make assumptions regarding length of employment. The
study was unable to assess potential individual exposures, and the cohort included white-collar
employees with limited potential exposure to organic solvents, which may reduce the study's
detection ability. Finally, the young average age of cohort subjects and short duration of follow-
up may reduce the study's sensitivity given low background cancer rates and inadequate latent
period.
Sung et al. (2007) tested the hypothesis of increased breast cancer among female workers
in the factory. Using a retrospective design, the cohort was identified through employment
records from the Bureau of Labor Insurance and consisted of women employed between 1973
and 1992 who worked for at least 1 day and whose cancer diagnosis occurred after employment
began. Of the 64,000 women employed during this time, 63,982 (99.97%) were eligible for and
included in the study, contributing a total of 1,403,824 person-years. Vital statistics information
was obtained from the Ministry of the Interior; cancer diagnoses were retrieved from the Taiwan
National Cancer Registry for the period 1979 to 2001 and linked with the cohort through
employee identification numbers. There were 29 cancer sites (oral, salivary, nasopharynx,
esophagus, stomach, small intestine, colon/rectum, liver/bile ducts, gall bladder, pancreas,
peritoneum, trachea/bronchus/lung, other respiratory, breast, cervix uteri, other uterus,
ovary/fallopian tube/broad ligament, other genital, kidney/urinary organs, bladder, skin, brain,
other nervous system, thyroid, bone, connective tissue, other/unspecified sites, leukemia, and all
sites) examined, and depending on the type of cancer, the latency periods were 5 years (thyroid
and leukemia), 15 years (breast and cervix uteri), or 10 years (all other cancer sites).
Employment in the factory was assessed as a proxy for exposure; duration of employment
(1 month; 1, 5, 10, 15, 20 years) was calculated based on the date that labor insurance started and
the date that employment ended. In the event of missing employment information,
two assumptions were made: (1) if the date of labor insurance was missing, this was assessed as
the earliest possible age (14 years); (2) if the date that employment ended was missing, this was
assessed as the date the factory closed in 1992. Periods of exposure were classified according to
government regulations that were issued in 1974, 1976, and 1978, as well as documents that
discussed factory violations with regard to proper ventilation. Pre-1974 was considered to be the
time of highest exposure, and there were 8,461 (13.2%) women who began working during this
time. Standardized incidence ratios and their corresponding 95% CIs were calculated assuming a
B-23
-------
Poisson distribution for each cancer site separately. Additional standardized incidence ratios and
95% CIs were calculated for breast, cervical, colorectal, and thyroid cancer stratified by calendar
year (pre- or post-1974) as well as duration of employment. Mests were used to compare
women with breast cancer who were employed either before or after 1974 on age at diagnosis,
age at first employment, and length of employment. Limitations included the lack of detailed
exposure information for both the factory and individuals within the cohort, the lack of control
for possible confounders, and the lack of detailed information related to the early 1970s.
Sung et al. (2008) investigated any possible link between maternal employment and
childhood cancer among first live born children. The factory employment records for all women
employed between 1973 and 1992 were obtained from the Bureau of Labor Insurance and linked
to the Taiwan Birth Registration database for the period 1978 to 2001. Children were required to
be first born singletons. Of the 103,506 children born to 47,348 women between 1978 and 2001,
40,647 children were eligible for and included in the study, contributing a total of
639,051 person-years. Demographics were obtained through the National Birth Registry and
included information on birth date, sex, single/multiple pregnancy, gestational age, and birth
weight, as well as parents' birth dates, education, marital status, and maternal parity. The
children's identification numbers were linked with the National Cancer Registry for the period
1979 to 2001 to ascertain how many were diagnosed with cancer. Employment at the factory
during the periconceptional time period was assessed as a proxy for exposure to
tetrachloroethylene and the other substances previously found in the soil and groundwater around
the factory (Bechtel Environmental Inc., 1990 and Target Environmental Services Inc., 1995).
Periconceptional exposure was defined as having been employed at the factory during 3 months
prepregnancy and 3 months after conception. Conception was calculated by subtracting the
length of gestation and an additional 14 days from the date of birth. Overall, there were
8,506 (20.9%) exposed children who contributed a total of 155,121 person-years. There were
11 cases of cancer (1 liver, 2 bone, 1 skin, 1 testis, 6 leukemia) in the exposed group and
36 cases (3 buccal cavity/pharynx, 1 liver, 1 bone, 3 connective/soft tissue, 1 skin, 2 breast,
2 ovary, 1 testis, 5 brain/other nervous system, 4 multiple myeloma, 9 leukemia, and 4 others) in
the nonexposed group. Poisson regression was used to calculate rate ratios and their
corresponding 95% CIs, adjusted for maternal age and education level, sex, and year of birth.
Strengths of this study include its use of an internal, nonexposed comparison group, which
reduced the potential for confounding and selection bias. Also, the researchers compared their
list of cancer cases to the death registry to verify that all cases had been retrieved. Limitations
include the study's inability to link exposures to individuals, and the study's inability to separate
the effects of different chemicals.
B-24
-------
B.l.2.5. Spirtas et al. (1991), Blair et al. (1998), Radican et al. (2008)
Spirtas, R.; Stewart, P. A.; Lee, J. S.; Marano, D. E.; Forbes, C. D.; Grauman, D. J.,
... Cohen, J. L. (1991). Retrospective cohort mortality study of workers at an
aircraft maintenance facility: I. Epidemiological results. Br J Ind Med, 48, 515-530.
http://dx.doi.0rg/10.1136/oem.48.8.515
Blair, A.; Hartge, P.; Stewart, P. A.; McAdams, M.; Lubin, J. (1998). Mortality and
cancer incidence of aircraft maintenance workers exposed to trichloroethylene and
other organic solvents and chemicals: Extended follow-up. Occup Environ Med, 55,
161-171. http://dx.doi.0rg/10.1136/oem.55.3.161
Radican, L.; Blair, A.; Stewart, P.; Wartenberg, D. (2008). Mortality of aircraft
maintenance workers exposed to trichloroethylene and other hydrocarbons and
chemicals: Extended follow-up. J Occup Environ Med, 50,1306-1319.
http://dx.doi.org/10.1097/JOM.Ob013e3181845f7f
Summary: A retrospective cohort mortality study of workers at Hill Air Force Base in
Utah was conducted with four aims: (1) to determine whether working at the aircraft
maintenance facility was associated with an increased risk of death; (2) to evaluate, in detail,
mortality risks associated with exposure to trichloroethylene; (3) to determine whether any raised
risks for specific causes of death were associated with specific chemical exposures; and (4) to
generate hypotheses for future research by evaluating the relation between various diseases and
specific chemicals. Individual earnings records from the National Personnel Records Center
were used to identify the cohort, which consisted of male and female civilian employees who had
worked at the base for at least 1 year between 1952 and 1956. Of the 14,457 eligible workers,
14,425 (99.8%) had official personnel folders that were able to be retrieved. These files
contained demographic information, as well as complete occupational histories, and were used to
create a "job dictionary" with 43,000 job titles. Industrial hygienists assessed exposure through
walkthrough surveys of the base, interviews with employees, industrial hygiene files, job
descriptions, and other historical documents including worker compensation files, telephone
books of the facility, organization charts, technical orders, and position descriptions (Spirtas et
al., 1991). Then, position descriptions with job titles and shops/departments were used as the
basis for evaluating chemical exposures. Any job that could not be linked to specific solvents
was coded as "mixed solvent" exposure, which consisted of 16 chemicals including
tetrachloroethylene (Stewart et al., 1991). Tetrachloroethylene consisted of a dichotomous
(yes/no) classification (Radican et al., 2008; Spirtas et al., 1991). Tetrachloroethylene was
primarily used to clean fabric in the parachute shop, replacing carbon tetrachloride in the late
1950s. In the accompanying paper in exposures at Hill Air Force Base, Stewart et al. (1991) do
not present industrial hygiene monitoring data on tetrachloroethylene concentrations; however,
Gold et al. (2008) noted, the arithmetic means for personnel measurements were 13 ppm for
l-hour samples for degreasing jobs in the aircraft and parts
B-25
-------
industry. Stewart et al. (1991) identified 851 (5.9%) subjects in the Hill Air Force Base cohort as
ever exposed to tetrachloroethylene (Stewart et al., 1991).
Observations for tetrachloroethylene and site-specific cancers are limitedly reported in
these studies given the study's primary focus on trichloroethylene exposure. Risk estimates for
tetrachloroethylene are presented for multiple myeloma and non-Hodgkin lymphoma in Spirtas
et al. (1991) and for breast cancer (women), multiple myeloma, and non-Hodgkin lymphoma in
Radican et al. (2008). Blair et al. (1998) did not present risk estimates for tetrachloroethylene.
Spirtas (1991) conducted a follow-up of this cohort through 1982. The data in the official
personnel folders were supplemented with vital status information, which was ascertained
through the Social Security Administration, the U.S. Office of Personnel Management, official
personnel folders, Veterans Administration records, motor vehicle bureau records, the National
Death Index, interviews with base personnel, and state vital statistics offices. Death certificates
were retrieved for all cohort members that died during the follow-up period, and the underlying
cause of death was assessed by a nosologist. Follow-up began in 1953 or 1 year after the start of
employment, whichever came later. Person-years at risk were stratified by race, sex, 5-year age
group, and calendar era. Person-years of exposure were calculated starting 1 year from the date
of first exposure or January 1953, whichever came later. Of those who were exposed to any
chemical or solvent and died from multiple myeloma, two (33.3%) women and no men were
exposed to tetrachloroethylene; of those who were exposed to any chemical or solvent and died
from non-Hodgkin lymphoma, two (20%) women and two (9.1%) men were exposed to
tetrachloroethylene. SMRs for the cohort of all white civilian employees at Hill Air Force Base
and for the tetrachloroethylene subcohort were estimated using Utah death rates as the basis for
determining the expected number of deaths. Corresponding 95% CIs were calculated assuming
the observed deaths followed a Poisson distribution. Estimates for the full cohort were adjusted
for age, sex, and calendar period. These estimates were then stratified by gender and adjusted
for age and calendar period only. All calculations were performed on the Caucasian population
only, which included those of unknown race. Strengths of this study include its size, analysis of
both genders, and use of a variety of mechanisms to assess exposure. Limitations include the
lack of adjustment for smoking and employees' exposure to multiple chemicals.
Blair et al. (1998) aimed to better understand the potential relationship between disease
risk and trichloroethylene and other organic solvents/chemicals. The follow-up period was
extended to December 1990 and conducted through linkage of the cohort with the National
Death Index and the Utah Tumor Registry. Person-years for the mortality analyses began
January 1, 1953 or 1 year after first employment and ended December 31, 1990, or date of death.
Person-years for incidence analyses began January 1, 1973, and ended December 31, 1990, or
the date of cancer diagnosis. Deaths were classified according to the International Classification
B-26
-------
of Diseases rules. Exposure was assessed using company personnel records from the first job to
the end of 1982. Individuals were evaluated as having ever (or never) been exposed to
chemicals. All cause mortality in the cohort was 40%, and all cancer mortality was 7%. Overall,
of those who died from non-Hodgkin lymphoma, 40 (81.6%) were exposed to any solvent; of
those who died from multiple myeloma, 24 (75.0%) were exposed to any solvent; and of those
who died from breast cancer, 28 (57.1%) were exposed to any solvent and died from breast
cancer. Relative risks and SMRs were estimated based on mortality in Utah. Rate ratios were
calculated for mortality and cancer incidence and compared between the exposed and the
unexposed using Poisson regression. Regression models adjusted for the following covariates:
date of birth, calendar year of death, and sex. The authors do not report strengths of this study.
Limitations include the lack of information on tetrachloroethylene, the lack of mutually
exclusive exposures, and lack of data on potential lifestyle confounders.
Radican (2008) extended the follow-up period to gain additional information about the
health risks associated with workplace exposures. The cohort was linked with the National
Death Index using personal identifiers and followed up for the period 1991 to 2000, an addition
of 10 years from Blair et al. (1998). Of those women who were exposed to any solvent and died
from breast cancer, 1 (2.6%) had been exposed to tetrachloroethylene. Of those who had been
exposed to any solvent and died from non-Hodgkin lymphoma, 5 (10%) men and 2 (16.7%)
women had been exposed to tetrachloroethylene. Of those who had been exposed to any solvent
and died from multiple myeloma, 3 (10%) men and 2 (25%) women had been exposed to
tetrachloroethylene. Of those who had been exposed to any solvent and died from nonmalignant
respiratory diseases, 46 (9%) men and 4 (51%) women had been exposed to tetrachloroethylene.
Cox proportional hazards regression was used to estimate hazard ratios and their corresponding
95% CIs using age as the time variable and race as the covariate. Analyses were stratified by
gender. The researchers also examined mortality using the Cox proportional hazards model for a
previously conducted study using a different follow-up period to compare the hazard ratios
between the two different statistical approaches. This was not performed for tetrachloroethylene,
though. Strengths of the study include its size, long follow-up, limited reporting bias due to
exposure assessment before the outcome was known, and its use of an internal comparison group
to minimize the healthy worker effect. Limitations include the small number of
tetrachloroethylene-exposed deaths and reduced statistical power, the inability to estimate risk of
one exposure while controlling for exposures to other chemicals, and the potential
misclassification of exposure based on job descriptions and other historical information.
B-27
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
LA. Dry-cleaner and laundry worker studies
Andersen et al.
(19991
Danish, Finnish, and Norwegian
cohorts from 1970 Censuses,
Swedish cohort from 1960
Census, men and women,
25-64 yr, alive on January 1,
1971; cancer cases from national
cancer registries in each country;
demographics, occupations,
industries from census
descriptions provided by the
heads of households for all
economically active members
Proxy — launderers and dry
plpjmprc
**tl^dlL^lJ
All cancers (incidence)
Full cohort:
10,101,711
Denmark:
2,346,134
Finland:
2,115,691
Norway:
1,792,817
Sweden:
3,847,069
Follow-up started
1971 and ended
with death,
emigration, or end
of follow-up;
Denmark,
1971-1987, linked
with Central
Population
Register; Finland,
1971-1990, linked
with Statistics
Finland; Norway,
1971-1991, linked
with Central
Population
Register; Sweden,
1971-1989, linked
with cause-of-death
register
Census descriptions coded according
to Nordic Occupational Classification
in Finland, Norway, Sweden;
Denmark coded according to own
standards; researchers then receded all
jobs based on a set of 54 occupational
groups based on Nordic Occupational
Classification standards; Group 51,
Code 95: launderers and dry cleaners
29,333 (0.3%) cohort members,
Denmark: 9,873 (0.4%), Finland:
4,949 (0.2%), Norway: 4,061 (0.2%),
Sweden: 10,450 (0.3%); Launderers
and dry cleaners: 519,844 person-
years; Denmark, 159,156; Finland,
94,302; Norway, 78,086; Sweden,
187,580
SIRs, 95% CIs, stratified by
cancer site, country,
adjusted for age; expected
numbers of cases from
cancer incidence rates for
each population; Poisson
distribution assumed for all
CIs whose SIRs had
<100 cases
cd
to
00
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Blair et al. (2003)
cd
to
Cohort: from dues records of
Local No. 161 (St. Louis) of
Laundry, Dry Cleaning, Dye
House Workers' International
Union; male and female dry
cleaners, entered union from
1945-1978, worked >1 yr; dues
records for demographic,
employment information, also
driver's license records, social
security files, health care finance
administration records, credit
bureaus; excluded if no
demographic information;
proxy—dry-cleaning tasks
All cancers (mortality)
11,062
members
identified, 5,369
met inclusion
criteria
Extended from
Blair et al. (1990).
which ended
January 1979;
started January
1979, ended
December 1993
(14 yr); person-
years start at entry
to union or 1948,
whichever later and
ended with death or
December 1993,
whichever came
first; deaths from
National Death
Index
Exposure indices for jobs within dry
cleaning: (1) run machines and handle
clothes (highest exposure),
TWA = 40; (2) pressers, sewers,
counter workers, TWA = 7;
(3) counter workers at pick-up stations
(minimal exposure), TWA = 0;
(4) maintenance workers (high, short-
term exposures), TWA = 7; entire
follow-up period (1948-1993):
220 deaths from cancer among those
with little/no exposure (index = 0),
316 deaths from cancer among those
with medium/high exposure (index = 7
or 40)
SMRs and 95% CIs to
examine relationship
between cancer and other
causes of death among dry
cleaners; Expected numbers
based on general U.S.
population 5-year age and
mortality statistics;
44% deceased at end of
entire follow-up period
Cano and Pollan
(2001)
Swedish men and women aged
25-64 yr in 1970 Census,
employed and counted in 1960,
followed 1971-1989. Over
200 occupational codes
examined including " launderers
and dry cleaners"
Non-Hodgkin lymphoma cancer
incidence
2,881,315
Followed
1971-1989 or date
of death; Swedish
Cancer
Environment
Register linked to
population register
Job title reported on 1960 and 1970
Censuses. Eleven of male cases were
launderers and dry cleaners
(occupational Code 943); no female
cases classified as launderers or dry
cleaners
Log-linear Poisson models
to compare occupations with
cohort, adjusted for
geographical area; PJls for
sectors, occupations,
adjusted for age, period, and
geographical category
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Chow et al. (1995)
cd
u>
o
Swedish Cancer Environment
Registry, which linked
employment and cancer
information for all individuals
registered in the 1960 Census
and National Swedish Cancer
Registry; linkage was performed
using personal identifiers
Esophageal cancer incidence
Not reported
Follow-up: 1961 to
1979
Job title reported on 1960 Census;
3 cases among laundry workers
SIRs, expected numbers
based on 5-year birth
cohort- and sex-specific
rates for esophageal cancer
in general Swedish
population during time
period; only occupations
with >500 individuals
examined; significance
evaluated assuming Poisson
distribution
Ji et al. (2005a. b);
Ji and Hemminki
(2006. 2005a. b, c)
Cohort: Swedish males and
females in Family-Cancer
Database linked national
censuses; cancer incidence data
from Swedish cancer registry
(1961-2000) Ji, et al. [(2005a):
bladder cancer] Ji and Hemminki
[(2005a): upper aerodigestive
tract cancer; (2005b): kidney
cancer; (2005c): liver and
gallbladder]
Additionally, subjects >31 yr age
and cancer incidence 1961-2002
[Ji and Hemminki (2006):
lymphoproliferative diseases;
(2005b): leukemia]
1,644,958
employed men
(9,255 dry
cleaners and
launderers) in
1960 Census
and 1,154,091
employed
women (14,974
dry cleaners and
launderers) in
1970 Census
Follow-up from
1961(1960
Census), 1970 (for
1970 Census or
those with same job
in 1960 +1970
Censuses), or 1980
(for those with
same job in
3 censuses) through
2000 or 2002
Relevant census information
(employment status, job title, work
industry) coded according to Nordic
Occupational Classifications; codes
merged into 53 occupational groups,
including launderers and dry cleaners;
9,255 (0.6%) male, 14,974 (1.3%)
female launderers and dry cleaners
SIRs for each occupation,
stratified by gender:
(1) adjusted for age period,
SES [aerodigestive tract
cancers, leukemia/lympho-
proliferative diseases,
kidney cancer, liver and gall
bladder cancer];
(2) smoking-corrected SIR
and smoking-corrected
95%CI[basedonIARC
(2004a): bladder cancer]
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Lindbohm et al.
(2009)
Cohort: Finnish males and
females participating in the 1970
National Population Census;
cancer incidence data from
Finnish Cancer Registry
Liver cancer incidence
1.2 million men
and women who
were born
between 1906
and 1945
Follow-up from
1971-1995
cd
Industry Code 850, and cumulative
exposure for organic solvent class.
Cumulative exposure based on
exposure for each birth cohort, starting
when average age of birth cohort was
20 to end of observation period or age
65 yr and included a 10-yr lag period.
If exposure took place before 1960,
FINJEM estimated use for 1945-1959
period, otherwise estimated for
1960-1984 period used
SIRs for each occupation,
stratified by gender from
Poisson regression models
adjusted for alcohol
consumption, smoking, and
socioeconomic status;
(smoking and alcohol
consumption by occupation
obtained for FINJEM from
the annual surveys of the
Finnish population in
1978-1991)
Lynge and
Thygesen (1990)
Cohort: Danish men and women,
20-64 yr, registered in 1970
Census as engaged in laundry
and dry-cleaning work; linked to
Danish Cancer Registry
Site-specific cancer incidence
Cohort: 10,600
Follow-up:
1970-1980 (Lynge
and Thygesen.
Industry Code 860 (laundries,
cleaning, and dyeing), occupational
Codes 411 (laundry worker, ironer)
and 380 (factory hand), and those who
reported as self-employed/family
workers; 2,434 (23%) serf-employed
dry cleaners/launderers, 830 (7.8%)
family workers, 6,837 (64.5%)
laundry workers/ironers, 499 (4.7%)
factory hands, 2,886 laundry/dry-
cleaning shops in 1970, 695 where dry
cleaning was the known predominant
activity
SIRs, 95% CIs, assuming
Poisson distribution if
observed cases <30 and
normal distribution if >30;
expected numbers from
multiplying person-years at
risk within each 5-year age
group with site-specific
incidence rates for full 1970
cohort (Lynge and
Thygesen. 1990)
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Pukkala et al.
(2009)
Ruder et al. (2001.
1994); Calvert et al.
(2011)
Study design
Cohort: Men and women in 1990
Census or prior census in
five Nordic countries (Denmark,
Finland, Iceland, Norway,
Sweden); dry cleaners and
launderers occupational title;
incident cases from national
cancer registries
Site-specific cancer incidence
Cohort: from union dry-cleaning
records in California, Illinois,
Michigan, New York; employed
>1 yr pre-1960 in dry-cleaning
shops using PCE; verified
records with visits to shops;
National Death Index for deaths
that occurred in cohort (site-
specific cancer mortality) or
REMIS for end-stage renal
disease incident cases.
Sample size
15 million
subjects total,
43,496 dry
cleaners and
launderers
1,704 workers
(mortality)
1 796 (end-
i^jLy\j ^tiiu
stage renal
. lse^se
mci ence)
Data collection
Person-years
started January 1
after the first
available census
and ended with
deaths, emigration,
or at end of
2003-2005,
whichever came
first, (depended on
country)
Person-years
started January 1,
1940, for mortality
or January 1, 1977,
for renal disease
incidence or after
1 year of
employment in
unionized shop,
whichever came
later, and ended
with death loss to
follow-up, or end
of 2004, whichever
came first.
Exposure assessment
Dry cleaner and launderer (Code 95)
according to Nordisk Yrke
Klassifisering and International
Standard Classification of Occupation
PCE exposure estimated by duration
of employment in dry-cleaning shops
(1-5 yr or >5 yr) and latency periods
(time since first employment <20 yr or
20+ yr); 618 (36%) worked only in
shops that only used PCE; 1,086
(64%) worked at shops where PCE
use is unable to be verified or where
other solvents are known/suspected to
be used instead
Statistical approach
SIRs and 95% CIs; expected
number of deaths using
national rates
SMRs (for deaths) and
95% CIs; expected number
of deaths estimated using
national rates; estimates for
each of 4 regions used
county and national rates for
expected numbers though
data not shown
SIRs (for end-stage renal
disease types) and 95% CIs;
expected number of deaths
estimated using REMIS
rates and nauonai
population estimates
cd
u>
to
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Selden and Ahlborg
(2011)
cd
Cohort: men and women
identified by employer as
working between 1973-1983 in
dry-cleaning and laundry
establishments for a previous
study of pregnancy outcome
(Ahlborg. 1990a). incident
cancers from Swedish National
Cancer Registry
Site-specific cancer incidence
10,389
employed
>1 mo
identified by
employers;
9,440 included
in follow-up
Person years
started January
1985 and ended
with cancer
diagnosis, death,
emigration, or end
of observation
period on
December 2006,
whichever came
first.
Jobs assigned to three exposure
categories: PCE (dry cleaners or
laundries with proportion of dry-
cleaning with PCE), laundries
(laundering only, no dry cleaning),
"other" (dry cleaning with PCE and
other solvents)
SIR and 95% CI using site-
specific cancer incidence
rate of Swedish population
Travier et al. (2002)
Cohort: Men and women
reporting work as launderers, dry
cleaners, and pressers in 1960 or
1970 Swedish Census, incident
cancers from Swedish national
cancer registry; all with second
primary neoplasms excluded
Site-specific cancer incidence
Authors did not
report total
number
included in
cohort;
543,036 person-
years from 1960
Census and
46,933 person-
years from 1970
Census
Person years
started January
1971 and ended
with cancer
diagnosis, death, or
loss to follow-up,
or December 1989,
whichever came
first.
Jobs coded by Nordic Classification of
Occupations and Swedish Industrial
codes; Group 1, launderer, dry cleaner,
or presser occupation employed in the
laundry, ironing, or dyeing industries
in either 1960 or 1970; Group 2,
launderer, dry cleaner, or presser
occupation employed in the laundry,
ironing, or dyeing industries in both
1960 and 1970; Group 3, launderer,
dry cleaner, or presser occupation
employed in other industries; Group 4,
other occupational titles employed in
laundry, ironing, or dyeing industries;
Group 5, not employed in relevant
industries or occupations
Multivariable Poisson
regressions, adjusted for
age, calendar period,
geographic region, urban
setting, gender. Analyses
also stratified by gender,
adjusted for age, calendar
period, geographic region,
urban setting, and by age in
1960 (<40yr, 40-59 yr,
>59 yr), adjusted for gender,
age, calendar period,
geographic regions, and
urban setting
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Wilson et al. (2008)
cd
Cohort: Swedish men and
women employed during 1960 or
1970 Census and still alive in
January 1971; cancer
information from Swedish
Cancer-Environment Registry
for 1971-1989
Proxy—laundry, ironing and
dyeing industries, laundry
workers and clothes pressing
occupations
Renal pelvis cancer incidence,
bladder cancer incidence
4,197, 684
cohort members
Person years began
January 1, 1971,
and ended with
cancer diagnosis,
emigration, death,
or December 31,
1989, whichever
came first.
70,083,912 person-
years of follow-up,
mean: 16.7 yr
Job titles in censuses coded according
to National Swedish Classification of
Occupations and Industries standards,
laundry workers: occupation Code 943
and clothes pressing: occupation Code
944. Laundry, ironing, and dyeing:
industry Code 880; 110 bladder cancer
cases and 11 renal pelvic cancer cases
with this industry code.
Job exposure matrix to assess
exposure to indoor work, low physical
activity, etc.
25,249 (0.6%) employed in industry
Code 880. 16,512 (0.4%) employed in
occupation 943, laundry worker
SIRs and 95% CIs for each
occupation and industry
using expected site-specific
cancer incidence rates of
total employed Swedish
population
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size Data collection
Exposure assessment
Statistical approach
I.B. Other occupational cohort studies
Anttila et al.
cd
Workers with blood:urine
biological monitoring; most of
TCE in urine; PCE in blood
1974-1983; demographics,
date/time sampling, result,
workplace, solvent code from
FIOH database; excluded if not
identified or deceased at start of
follow-up
Site-specific cancer incidence
(pancreas, lung/bronchus, cervix
uteri, kidney, nervous system,
non-Hodgkin lymphoma,
multiple myeloma, and all
cancer)
11,534
measurements
from
1965-1983,
10,743 (93.1%)
linked to
personal
identifiers,
which
corresponded to
3,976 workers.
Final sample:
3,974 subjects;
849 workers
with blood PCE
measurements
Follow-up started
January 1967 or
date of first
measurement;
ended with
emigration, death,
or December 1992,
whichever came
first; used Finnish
Cancer Registry,
Population Register
Center, Central
Statistical Office of
Finland; overall:
71,800 person-
years, averaged
18 yr
Blood measurements: median
0.7 umol/L in males, 0.4 umol/L in
females, average 3.2
measurements/individual; 849 (21.4%)
workers monitored for exposure to
PCE, contributed 11,958 person-years;
duration not examined for PCE, given
its few site-specific cancer cases
SIRs for 8 sites, 95% CIs,
expected numbers of
cancers from incidence rates
of Finnish population;
Mantel-Haenszel % test for
significance, assuming
observed cases followed
Poisson distribution
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Boice et al. (1999)
cd
Cohort: from work history cards,
personnel files, retirement
records, employed at aircraft
manufacturing factories >1 yr
from 1960 onwards; exclusions:
missing work history or incorrect
dates; vital status from California
death tapes, National Death
Index, Pension Benefit
Information Files, Social
Security Death Index, Health
Care Finance Administration
files, California Department of
Motor Vehicles records,
employment work history cards,
pension and retirement records,
obituaries from 1960-1996; if no
vital status information found,
assumed alive
JEM for PCE exposure
Site-specific cancer mortality
113,204
employees,
77,965 (68.9%)
included in
study
Follow-up: started
January 1, 1960 or
after 1 year
employment and
ended with death,
age 95 yr or
December 1996,
whichever came
first.
1,889,795 person-
years of follow-up;
average of >20 yr
per person
Exposure assessed via walk-through
surveys of factories, interviews with
employees, industrial hygiene
files/other historical documents; based
on job code, job title, job change
information; factory jobs only
assessed as routine (daily),
intermittent (not daily), minimal/no
exposure; duration (<1, 1-4, >5 yr)
based on dates of employment for
each job; overall: 5,830 (7.5%)
exposed to PCE; 2,631 (5.8%) routine
exposure, 3,199 (7.1%) intermittent
exposure, 51,214 person-years at risk
SMRs, 95% CIs, assuming
observed deaths followed
Poisson distribution,
expected number of deaths
among Caucasians based on
race, age, calendar year,
sex-specific California rates;
expected number of deaths
among non-Caucasian based
on general U.S. population
rates; Poisson regression for
duration, adjusted for date
of birth, date first employed,
date of end of employment,
race, sex; tests of trend to
examine duration of
exposure
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Bond et al. (1990)
Chang et al. (2005;
2003); Sung et al.
(2008: 2007)
Study design
Dow Chemical's Midland/Bay
City production, research, and
headquarters units
Nested case-control; cohort:
from work history records,
48,521 men and women
employed 3+d; Cases: from
death certificates, men who died
from 1940-1982. Controls:
randomly selected from among
the cohort of male employees
Proxy — factory employment
Primary liver cancer, cancer of
gallbladder/bile ducts, cancer of
liver not specified
Cohort: from Bureau of Labor
and Insurance's records
Proxy — employment in
electronics factory in Taiwan
Cancer, mortality
Sample size
44 (0.7%) liver
and biliary tract
deaths eligible
for and included
in study; 1,888
(8.8%) controls
selected
randomly from
cohort
(n = 21,437
males)
Final sample*
44 cases, 1,888
controls
Various (see
below)
Data collection
Follow-up period:
1940-1982
Various (see
below)
Exposure assessment
Work history records for exposure by
work area and exposure to 1 1
chemicals, including PCE; 6 (13.6%)
cases, 213 (11.3%) controls exposed
to PCE
Various (see below)
Statistical approach
Mantel-Haenszel for RRs,
adjusted for birth year;
Miettinen's method for
95% CI; primary liver
cancer and gall bladder/bile
duct assessed separately but
not presented; duration work
exposure failed to reveal any
significant trends
Various (see below)
cd
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Chang et al. (2003)
Cohort: identified from
1973-1997, men and women,
average age of cohort members:
39.3 yr; linked with National
Mortality Database; compared
with labor-insurance
hospitalization data
Site-specific cancer mortality
cd
u>
00
86, 868
individuals
(16,133 men
and 70,73 5
women) who
contributed
1,357 deaths,
316 (24%) due
to cancer. All-
cause mortality
rate was 1.56%,
and all cancer
mortality rate
was 0.36%
Person-years
started when
entered cohort or
January 1, 1985,
whichever later,
and ended with
death or December
31,1997,
whichever came
first; total of
1,380,354 person-
years; average
follow-up time:
14.3 yr for men and
16.3 yr for women
Duration of employment: period of
employment between the start and end
of labor insurance coverage, with
assumptions made for those with
missing data
SMRs, stratified by
underlying cause of death,
expected numbers based on
general Taiwanese
population as a reference.
For dose-response
assessment, cohort stratified
by duration employment:
<1 year, >1 year but <5 yr,
>5 yr and calendar year:
1985-1990, 1991-1997
Chang et al. (2005)
Cohort: identified from
1973-1997, men and women,
linked with National Cancer
Registry and National Mortality
Database; compared with labor-
insurance hospitalization data
Site-specific cancer incidence
86,868(16,133
men and 70,73 5
women) who
contributed a
total of
1,380,354
person-years;
998 incident
cancer cases
Follow-up: started
January 1, 1979, or
date of
employment,
whichever later,
and ended with
cancer diagnosis,
death, or December
31, 1997,
whichever came
first
Duration of employment: <1, 1-5,
5-10, and 10+ yr; assumptions made
for those with missing start or end
dates
Period of employment: 1979-1984,
1985-1990, 1991-1997
SIRs comparing exposed to
incidence rates in general
population of Taiwan by
age, calendar year, and sex;
trends examined by duration
and period of employment
Latency periods: <3 mo,
6 mo, and 1, 5, and 10 yr
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Sung et al.
cd
Cohort: identified from
1973-1992, women worked for
1+ d, cancer diagnosis after
employment began; Vital status
from Ministry of Interior; cancer
diagnoses from Taiwan National
Cancer Registry and linked with
cohort from 1979-2001
Site-specific cancer incidence:
29 cancer sites
64,000 women
employed,
63,982
(99.97%)
eligible for and
included in the
study,
contributing
1,403,824
person-years
Follow-up:
1979-2001
Duration of employment (1 mo, 1, 5,
10, 15, and 20 yr) based on date labor
insurance started and employment
ended; for missing employment
information: (1) if missing date labor
insurance, and assumed earliest
possible age (14 yr); (2) if date
employment ended missing assumed
factory closure in 1992. Exposure by
dates of government regulations:
pre-1974 had the highest exposure
8,461 (13.2%) women who started
working pre-1974
SIRs and 95% CIs,
assuming a Poisson
distribution for each cancer
site; SIRs and 95% CIs for
breast, cervical, colorectal,
thyroid cancers, stratified by
pre- or post-1974, duration
of employment; Mests for
breast cancer among those
employed pre- or post-1974
latency periods: 5 yr
(thyroid/leukemia), 15 yr
(breast/cervix uteri), 10 yr
(all others)
Sung et al.
Cohort: identified from
1973-1992, women who worked
in a factory, linked to the Taiwan
Birth Registration Database from
1978-2001; only firstborn
singletons, demographics from
National Birth Registry, children
linked with National Cancer
Registry from 1979-2001
Childhood cancers
103,506
children born to
47,348 women
from
1978-2001,
40,647 children
eligible for and
included in the
study,
contributing
63 9,051 person-
years; 11 cancer
cases among
exposed,
36 cancer cases
among
nonexposed
Follow-up:
1979-2001
Periconceptional exposure defined as
employed at factory during 3 mo
prepregnancy and 3 mo after
conception; conception calculated by
subtracting length of gestation +14 d
from the date of birth
8,506 (20.9%) exposed children who
contributed a total of 155,121 person-
years
Poisson regression for RRs
and 95% CIs, adjusted for
maternal age, maternal
education level, sex, year of
birth
-------
Table B-l. Summaries of characteristics of cohort studies (dry-cleaner and laundry workers, and other cohorts)
(continued)
Reference
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Spirtas et al. (1991):
Blair et al. (1998):
Radican et al.
(2008)
cd
-k
o
Cohort: male/female civilian
employees, worked 1+ yr at base
from 1952-1956, identified from
individual earnings records
Personnel folders, vital status
data from Social Security
Administration, U.S. Office of
Personnel Management,
veterans administration records,
motor vehicle records, National
Death Index, interviews, state
vital statistics; death certificates
for cohort members who died
during follow-up, underlying
cause of death assessed by
nosologist
Mortality: MM, NHL, breast
cancer, nonmalignant respiratory
diseases
14,425 with
personnel files
(99.8% of
14,455 eligible);
analysis
restricted to
10,461 men and
3,605 women
classified as
Caucasian.
Follow-up began
1953 or after 1 year
of employment
Follow-up study
through 1982
(Spirtas etal..
1991) or 2000
(Radican etal..
2008)
Job exposure matrix based on
industrial hygienists, walkthrough
surveys of base, interviews with
employees, industrial hygiene files,
job descriptions, historical documents.
Job titles/shops used as basis for
evaluating exposures, which for PCE
consisted of ever/never classification;
jobs unable to be linked to solvents
coded as "mixed solvents". 10,256
ever exposed to mixed solvents, 851
ever exposed to PCE. Of those
exposed to any chemical/solvent and
died from MM, 2 (33.3%) women,
0 men exposed to PCE; of those who
were exposed to any chemical/solvent
and died from NHL, 2 (20%) women,
2 (9.1%) men exposed to PCE
SMRs MM and NHL and
PCE. All calculations on
Caucasian population only,
including unknown race
(Spirtas etal.. 1991)
Cox proportional hazards
regression for hazard ratios
and 95% CIs using age as
time variable and race as
covariate, stratified by
gender (Radican et al..
2008); RR for breast cancer,
MM, NHL, nonmalignant
respiratory diseases, and
PCE
JEM = job-exposure matrices; RR = relative risk.
-------
B.2. Case-Control Studies
Tetrachloroethylene case-control studies have been organized by (1) multiple cancer-site
studies and (2) single cancer-site studies. Tables B-2 and B-3 provide summaries of the study
characteristics of each paper or group of papers.
B.2.1. Multiple Cancer-Site Studies
A number of case-control studies of multiple cancer-site studies have been conducted by
a single research group. These studies are discussed in this section given common
methodologies among the studies. The studies are organized by region (British Columbia and
Montreal in Canada, Massachusetts in the United States, New Zealand, Germany, and
four Nordic countries (Denmark, Finland, Norway, and Sweden).
B.2.1.1. British Columbia (Canada)
B.2.1.1.1. Band et al. (1999), MacArthur et al. (2009)
Band, P.; Le, N.; Fang, R.; Threlfall, W.; Gallagher, R. (1999). Identification of
occupational cancer risks in British Columbia. Part II: A population-based case-
control study of 1516 Prostatic cancer cases. J Occup Environ Med, 41, 233-247.
http://www.ncbi.nlm.nih.gov/pubmed/10224589
MacArthur, A.; Le, N.; Fang, R.; Band, P. (2009). Identification of occupational
cancer risk in British Columbia: A population-based case-control study of 2,998
lung cancers by histopathological subtype. Am J Ind Med, 52, 221-232.
http://dx.doi.org/10.1002/aiim.20663
Summary: A registry-based case-control study was undertaken to examine occupational
risk factors for cancer in British Columbia. Cases were identified through the British Columbia
Cancer Registry from 1983 to 1990 and consisted of men aged 20 or older with histologically-
confirmed cancer. All cases were mailed a self-administered questionnaire inquiring about
lifetime job descriptions, including duration and period of employment, as well as occupation
and industry titles. Participants were also asked about their ethnic origin, education, lifetime
smoking habits, and alcohol consumption. Data were collected for each cancer site until 1,000
completed questionnaires were returned for that site or until December 31, 1990. If the patient
was deceased, the spouse or closest living relative was asked to complete the questionnaire. A
total of 25,726 eligible cases were contacted, and 15,463 (60%) participated by returning the
questionnaire. Occupations and industries were used as a proxy for exposure and coded
according to the Canadian Standard Occupational Classification (SOC) and the Canadian
B-41
-------
Standard Industrial Classification (SIC). Laundries and dry cleaners comprised SIC Code 972.
The authors did not report the SOC code for dry cleaners. For each occupation and industry,
estimates for "ever" (whether or not a job within the given occupation or industry was ever held)
and "usual" (job with the longest held lifetime employment in a given occupation or industry)
occupations and industries were calculated. Strengths include complete ascertainment of cases
and occupational histories, adjustment for confounders, and examination of lung cancer
subtypes. Limitations include the lack of information on occupational exposures, small numbers
of exposed cases, self-reported lifestyle characteristics, and possible bias due to the use of other
cancer cases as controls. There may also be nonrepresentativeness of controls between early and
late responders due to the fact that the survey for each cancer site ended at 1,000 cases.
Band et al. (1999) used the data to conduct a matched case-control study examining the
occupational risks associated with prostate cancer in British Columbia. Cases consisted of
patients with histologically confirmed prostate cancer who returned the questionnaire. A total of
1,519 (9.8%) of the cases in the larger study were prostatic cancer cases. Controls were selected
from among the other cancer sites within the larger study, excluding lung cancers and cancers of
unknown primary sites, and were matched to cases based on age and year of diagnosis. The final
sample consisted of 1,516 cases matched to at least 1 of 4,994 controls. Proxy respondents
represented 19.9% of cases and 19.3% of controls. Overall, there were 7 (0.5%) cases who
reported "ever" employment and 2 (0.1%) cases who reported "usual" employment in the
laundries and cleaners industry. The authors do not report the number of controls that reported
"ever" or "usual" employment. Conditional logistic regression was used to estimate odds ratios
(ORs) and 90% CIs for each occupation and industry separately for each of two estimates of
exposure, adjusted for education, alcohol consumption, smoking duration, and respondent to
questionnaire.
MacArthur et al. (2009) evaluated the occupational risks for lung cancer. Of the
5,528 eligible, incident lung cancer cases, 2,998 (54.2%) returned the questionnaire. Controls
consisted of all other cancer cases, excluding those with unknown primary sites (708 other cases)
and were matched to cases based on age and year of diagnosis. Laundries and dry cleaners
comprised Code 972 and contained 10 (0.3%) cases of lung cancer (squamous cell carcinoma,
adenocarcinoma, and small cell lung cancer). Matched case-control analyses for industries and
occupations with at least three cases were performed to calculate maximum likelihood estimates
of ORs and their corresponding 90% CIs for "ever" and "usual" employment. Lung cancer
subtypes (squamous cell carcinoma, adenocarcinoma, small cell lung cancer, large cell lung
cancer) were also separately assessed. The estimates for all lung cancers combined were
adjusted for smoking, questionnaire respondent, alcohol, and education. Lung cancer subtype
estimates were separately adjusted for their own set of covariates. All subtypes were adjusted for
B-42
-------
questionnaire respondent; squamous cell carcinoma, adenocarcinoma, and large cell lung cancer
were adjusted for alcohol consumption status; adenocarcinoma, small cell lung cancer, and large
cell lung cancer were each adjusted for smoking duration (years); squamous cell carcinoma was
also adjusted for cigarette pack-years, marital status, pipe smoking status, and cigar smoking
status. Adenocarcinoma was also adjusted for ethnicity, and small cell lung cancer's additional
covariates included ethnicity and cumulative alcohol score. The final covariate for large cell
lung cancer was level of education.
B.2.1.1.1.1. Band et al. (2000)
Band, P. R.; Le, N. D.; Fang, R.; Deschamps, M.; Gallagher, R. P.; Yang, P. (2000).
Identification of occupational cancer risks in British Columbia: A population-based
case-control study of 995 incident breast cancer cases by menopausal status,
controlling for confounding factors. J Occup Environ Med, 42, 284-310.
http://www.ncbi.nlm.nih.gov/pubmed/10738708
Summary: This study is a population-based case-control study whose objective was to
examine the relationship between occupational risk and hormonal factors in breast cancer. Cases
were identified through the British Columbia Cancer Registry and consisted of all women under
the age of 75 years who were diagnosed with breast cancer between June 1988 and June 1989.
In order to be included in the study, cases needed to be Canadian citizens, residents of British
Columbia, English-speaking, and have no prior history of breast cancer. Controls were randomly
selected from the 1989 British Columbia Provincial Voters List, matched on age, and had no
history of breast cancer diagnosis before June 1989.
Participants were mailed a self-administered questionnaire inquiring about demographics,
lifetime smoking, lifetime alcohol consumption, current body weight, weight in late teens, age at
menarche, parity, age at first birth, history of breast biopsy before 1987, family history of breast
cancer, breastfeeding, birth control, estrogen replacement therapy, and lifetime occupational
history, including job descriptions, occupation and industry titles, duration, and period of
employment. Of the 1,489 eligible cases, 1,018 (68%) returned the questionnaire; of the
1,502 eligible controls, 1,025 (68%) returned the questionnaire. After matching and excluding
those with missing information on statistically significant confounders, a total of 995 cases and
1,020 controls were available for the analysis. Occupations and industries were coded according
to the Canadian SOC and the Canadian SIC; dry cleaning was included in SOC Code 6162 and
SIC Code 9721. Two surrogates of exposure were assessed: "usual" occupation/industry,
defined as the job with the longest held lifetime employment in a given occupation or industry,
and "ever" occupation/industry, defined as whether a job was ever held in the occupation or
industry in question. Overall, there were 12 (1.2%) cases "ever" exposed and 9 (0.9%) cases
with "usual" exposure to the laundry and dry cleaning occupation; there were also 23 (2.3%)
B-43
-------
cases with "ever" exposure and 10 (1.0%) cases with "usual" exposure to the power laundries
and/or dry-cleaners industry. The authors do not report the controls' exposures.
Conditional logistic regression was used to estimate ORs and 90% CIs for all occupations
and industries, stratified by menopausal status and "usual"/"ever" occupation. Covariates were
individually assessed using a forward methodology. The occupational analyses were adjusted
for the following three factors: (1) premenopausal women—cigarette pack-year groups, breast
biopsy, and family history of breast cancer in the mother and sisters; (2) postmenopausal
women—weights in 1986, family history of breast cancer in a first-degree relative, a history of
breast biopsy for benign breast diseases, and cumulative alcohol scores; and (3) all women
combined—both the pre- and postmenopausal confounders. Strengths of the study include its
population-based design, lifetime occupational history, and stratification by menopausal status.
Limitations include a lack of information on actual exposures, small number of cases in each
occupational category, chance occurrence, and lack of assessment of duration or intensity of
exposure.
B.2.1.1.1.2. Teschke et al. (1997)
Teschke, K.; Morgan, M. S.; Checkoway, H.; Franklin, G.; Spinelli, J. J.; van Belle,
G.; Weiss, N. S. (1997). Surveillance of nasal and bladder cancer to locate sources of
exposure to occupational carcinogens. Occup Environ Med, 54, 443-451.
http://www.ncbi.nlm.nih.gov/pubmed/9245952
Summary: This case-control study examined sources of occupational exposure to known
or probable carcinogens in British Columbia, Canada, with the aim of alerting regulatory
agencies and industrial health professionals about occupations that warranted occupational
hygiene exposure measurement and control. Cases were identified through the British Columbia
Cancer Agency and consisted of men and women aged 19 years or older with histologically
confirmed nasal cavity/sinus or urinary bladder cancers. Nasal cavity/sinus cancer cases were
obtained for the time period from 1990 to 1992, and bladder cancer cases were selected between
1990 and 1991. Bladder cancer cases born before 1916 were excluded from the study, as were
carcinomas in situ. Controls consisted of British Columbia residents aged 19 years or older.
They were randomly selected from the provincial voter list and matched to cases based on age
and sex. Any selected controls that were in prison or in a mental health institution by court order
were excluded from the study. Of the 54 eligible nasal cancer cases and 195 eligible nasal
cancer controls, 48 (88.9%) cases and 159 (81.5%) controls participated in the study. Of the
119 eligible bladder cancer cases and 173 eligible bladder cancer controls, 105 (88.2%) cases
and 139 (80.3%) controls participated in the study. The final sample consisted of 153 cases and
298 controls.
B-44
-------
Interviews were conducted with all cases and controls using a structured questionnaire
administered by a registered nurse who knew of their case or control status. In-person or
telephone interviews were conducted with all subjects who lived within a 6-hour (one-way) drive
of Vancouver. Telephone interviews were conducted with all participants residing more than
6 hours away (21% cases and 23% controls). Proxy interviews with relatives were conducted if
the individual was deceased, did not speak English well, or if he/she could not accurately
remember life events. This occurred with 26 (17%) cases and 41 (13.8%) controls. The
questionnaire inquired about occupational, residential, medical, smoking, and exposure histories;
a blinded industrial hygienist evaluated all completed interviews and asked the nurse to conduct
follow-up, asking clarification questions of the participant when necessary. Occupations and
industries were first coded according to standard occupational and industrial classifications and
then blindly grouped according to a previously established classification system. Assignment
into a group was based on whether the occupation or the industry was more likely to determine
the individual's exposure. In the event that both the occupation and the industry determined
exposure, the occupation was used. After that, all duties and exposures related to each
occupation were reviewed to verify the accuracy of all categorizations, and all groups with less
than 20 individuals were reviewed to determine if they could be combined with others. In total,
57 occupational groups were developed. Laundry personnel were part of the "other" category
for nasal cancer and contained no cases or controls; on the other hand, laundry personnel were
included in the "originally suspect" group for bladder cancer and contained five cases (3.3%) and
four (1.3%) controls who reported "ever" employment in the occupation.
Exact methods were used to estimate summary ORs and their corresponding 95% CIs
according to Breslow and Day (1980). In the event that nonoccupational risk factors were found
to be positively associated with any of the cancers, the odds ratios and their corresponding 95%
CIs were estimated using unconditional logistic regression, adjusted for these risk factors.
Latency times of 5, 10, and 15 years were also examined, though the results were not shown. All
odds ratios were adjusted for sex, age, and smoking. The influence of duration of employment
(6 months to 10 years, and 10 years or more) was also examined but only reported if the
estimates affected the results. Occupational groups were then assessed for their need for further
surveillance based on a set of criteria. Limitations of the study include its small sample size, the
grouping of jobs with different duties and exposures, and the exclusion of carcinomas in situ.
The authors do not report any strengths associated with the methodology of their study.
B-45
-------
B.2.1.2. Montreal (Canada)
B.2.1.2.1. Siemiatycki et al. (1991; 1987), Aronson et al. (1996), Parent et al. (2000)
Siemiatycki, J. (1991). Risk factors for cancer in the workplace. Boca Raton, FL:
CRC Press.
Siemiatycki, J.; Wacholder, S.; Richardson, L.; Dewar, R.; Gerin, M. (1987).
Discovering carcinogens in the occupational environment: Methods of data
collection and analysis of a large case-referent monitoring system. Scand J Work
Environ Health, 13, 486-492. http://www.ncbi.nlm.nih.gov/pubmed/3433050
Aronson, K.; Siemiatycki, J.; Dewar, R.; Gerin, M. (1996). Occupational risk factors
for prostate cancer: Results from a case-control study in Montreal, Quebec,
Canada. Am J Epidemiol, 143, 363-373.
http://www.ncbi.nlm.nih.gov/pubmed/8633620
Parent, M. E.; Hua, Y.; Siemiatycki, J. (2000). Occupational risk factors for renal
cell carcinoma in Montreal. Am J Ind Med, 38, 609-618.
http://www.ncbi.nlm.nih.gov/pubmed/11071683
Summary: Siemiatycki (1991) used a population-based case-control design to examine
the possible association between occupational exposures and cancer. Cases were identified from
hospitals in Montreal and consisted of male residents of Montreal aged 35 to 70 years who were
diagnosed or histologically confirmed with any of the following cancers between 1979 and 1985:
esophagus, stomach, small intestine, colon, rectum, gall bladder, pancreas, peritoneum, lung,
pleura, skin, prostate, penis, testes, bladder, kidney, eye, lymphoid tissue, and multiple myeloma.
Brain cancer, buccal cavity cancer, larynx cancer, and leukemia were excluded; due to limited
resources, lung cancer was excluded in Years 2, 3, and 6; rectal cancer was excluded in Years 1
and 2; prostate cancer was excluded in Years 4 and 5. All of the large hospitals in Montreal took
part, providing 97% population-based case ascertainment. Of the 4,576 cases identified,
3,730 (81.5%) participated in the interview. Response rates for individual cancers varied
between 78% and 85%. Two sets of controls were used. Population-based controls were
selected through electoral lists and random digit dialing. Of the 541 chosen from electoral lists,
375 (69.3%) were interviewed. Of 199 eligible participants identified through random digit
dialing, 158 (79.4%) participated in the interview. Overall, of 740 population controls selected,
533 (72%) were interviewed. The final sample consisted of 99 esophagus cases, 251 stomach
cases, 497 colon cases, 257 rectum cases, 116 pancreas cases, 857 lung cases, 449 prostate cases,
484 bladder cases, 177 kidney cases, 103 melanoma cases, and 215 lymphoma cases. In-person
interviews were conducted by trained interviewers with cases and controls through a two-part
questionnaire. The first section was structured and inquired about demographics; residential
history; lifetime consumption of cigarettes, alcohol, coffee, and tea; consumption of food
containing carotene; and height and weight. The second part was semi-structured, so as to
B-46
-------
acquire detailed information on each of the jobs held during the man's working lifetime.
Occupations and industries were coded according to the Canadian Classification and Dictionary
of Occupations 1971 and the SIC Manual, respectively. Exposure was classified by a team of
blinded chemists and hygienists, who used a checklist of 294 substances to determine the number
of potential exposures for each job. All classifications were based on a three-point scale: the
degree to which they believed the exposure had actually occurred (possible, probable, definite),
the frequency of exposure in a normal workweek (<5, 5-30, and >30%), and the level of the
concentration of the exposure (low, medium, high). Nonexposure was evaluated according to the
background levels of that particular substance. The 294 substances were combined with
98 occupations and 77 industries to make a total of 469 occupational circumstances. Exposure
was assessed as both direct exposure to tetrachloroethylene and proxy exposure through
employment as launderers and dry cleaners. There were 6 (1.2%) cases of colon cancer,
7 (0.8%) cases of lung cancer, and 9 (2.0%) cases of prostate cancer that were "ever" exposed to
tetrachloroethylene. Similarly, there were 4 (1.6%) cases of stomach cancer, 5 (1.0%) cases of
colon cancer, 5 (2.0%) cases of rectum cancer, 12 (1.4%) cases of lung cancer, 9 (2.0%) cases of
prostate cancer, 10 (5.6%) cases of kidney cancer, 3 (2.9%) cases of skin melanoma, and
3 (1.4%) cases of non-Hodgkin lymphoma among those who reported "ever" employment as
launderers or dry cleaners. The Mantel-Haenszel method was used to estimate odds ratios and
their corresponding 90% CIs for "ever" exposure and "substantial" exposure. All estimates were
adjusted for age, family income, and cigarette index. Additionally, stomach cancer was adjusted
for birthplace; colon and rectum cancers were adjusted for ethnic origin and beer index; lung
cancer was adjusted for ethnic origin, alcohol index, and respondent; prostate cancer was
adjusted for ethnic origin, Quetelet index, and respondent; and kidney cancer and skin melanoma
were adjusted for ethnic origin. Strengths of the study design include its detailed information on
potential confounders and occupational histories, its blind exposure assessment, use of
histologically confirmed cases and access to two different control groups. Limitations to the
study include the possible misclassification of exposure, the small numbers of exposed, the
examination of many chemicals and job categories, and the study's goal to identify risk factors
for further investigation.
Aronson et al. (1996) and Parent et al. (2000) used the data from Siemiatycki (1991) to
further examine associations with selected cancers. Aronson et al. (1996) examined the
association between occupations and prostate cancer. Of the 557 prostate cancer cases,
449 (81%) participated in the interview. The cancer controls included all other cancer cases
from Siemiatycki et al. (1991) except lung cancer. The final sample consisted of 449 cases,
1,550 nonprostate cancer controls, and 533 population controls. Overall, 55 (27 substances,
11 industries, and 17 occupations) of the 469 occupational circumstances initially reviewed in
B-47
-------
Siemiatycki et al. (1991) were examined in this study. Tetrachloroethylene exposure was
classified as "unexposed," "nonsubstantial," or "substantial." There were eight participants with
"substantial" exposure, but the authors failed to note whether these were cases or controls,
precluding a calculation of exposure prevalence. Unconditional logistic regression was used to
estimate odds ratios and their corresponding 95% CIs for exposures using partially adjusted and
fully adjusted models. The partially adjusted model controlled for age, ethnicity, socioeconomic
status, Quetelet index, and self-/proxy respondent status, while the fully adjusted model included
the covariates in the partially adjusted model, in addition to the all-core substances with 30 or
more exposed cases. Due to the fact that very few differences were found when analyzing the
control groups separately, the majority of the results were reported using the pooled group. In
the event that two substances were highly correlated, one was removed from the model.
Parent et al. (2000) examined occupation and renal cell cancer. Of the 227 eligible
kidney cases, 177 (78%) were interviewed and 142 of the 177 kidney cancers were renal cell
carcinoma. There were a total of 1,900 cancer controls, representing a participation rate of 78%.
Occupations and industries were assessed as a proxy for exposure and classified as any exposure
and duration of exposure >10 years. The laundry and cleaning industry had a total of four cases
(2.8%) that were "ever" exposed to the industry. Fewer than 4 cases were exposed for more than
10 years, and the results were not reported. The authors did not report exposure to
tetrachloroethylene in this study, although the predecessor study, Siemiatycki et al. (1991) did.
Unconditional logistic regression models were used to calculate odds ratios and their
corresponding 95% CIs for each occupation and industry, stratified by any exposure and duration
of exposure >10 years. Estimates for any exposure were adjusted for respondent status, age,
smoking, and BMI.
B.2.1.3. Massachusetts (United States)
B.2.1.3.1. Aschengrau et al. (1998; 1993), Paulu et al. (2002,1999)
Aschengrau, A.; Ozonoff, D.; Paulu, C.; Coogan, P.; Vezina, R.; Heeren, T.; Zhang,
Y. (1993). Cancer risk and tetrachloroethylene-contaminated drinking water in
Massachusetts. Arch Environ Health, 48, 284-292.
http://www.ncbi.nlm.nih.gov/pubmed/8215591
Aschengrau, A.; Paulu, C.; Ozonoff, D. (1998). Tetrachloroethylene-contaminated
drinking water and the risk of breast cancer. Environ Health Perspect, 106, 947-953.
http://www.ncbi.nlm.nih.gov/pubmed/9703477
Paulu, C.; Aschengrau, A.; Ozonoff, D. (1999). Tetrachloroethylene-contaminated
drinking water in Massachusetts and the risk of colon-rectum, lung, and other
cancers. Environ Health Perspect, 107, 265-271.
http://www.ncbi.nlm.nih.gov/pubmed/10090704
B-48
-------
Paulu, C.; Aschengrau, A.; Ozonoff, D. (2002). Exploring associations between
residential location and breast cancer incidence in a case-control study. Environ
Health Perspect, 110, 471-478. http://www.ncbi.nlm.nih.gov/pubmed/12003750
Summary: These population-based case-control studies of bladder cancer, kidney cancer,
and leukemia evaluated the relationship between various types of cancer and tetrachloroethylene
exposure through contaminated drinking water. From 1968 to 1980, tetrachloroethylene was
used in a vinyl liner for asbestos cement water distribution pipes throughout Massachusetts to
improve taste and odor. A substantial amount had been installed in five Upper Cape Cod towns,
including Barnstable, Bourne, Falmouth, Mashpee, and Sandwich. In 1980, it was discovered
that the tetrachloroethylene, which had been assumed to evaporate during the lining process, had
leached into drinking water supplies. Cases were identified from the Massachusetts Cancer
Registry and consisted of permanent residents of five Upper Cape Cod towns who were
diagnosed with cancer between 1983 and 1986. Controls were identified through three
mechanisms: living controls <65 years of age were obtained through random digit dialing, living
controls >65 years of age were randomly chosen from Health Care Finance Administration lists
using stratified sampling, and deceased controls were randomly selected from a Massachusetts
Department of Vital Statistics and Research file for the period from 1983-1989. Of the
2,236 controls <65 years, 249 (11.1%) were eligible and contacted; of these, 184 (73.9%) were
interviewed. Of the 611 controls >65 years, 537 (87.9%) were eligible and contacted; of these,
464 (86.4%) were interviewed. Of the 918 deceased controls, 794 (86.5%) were eligible and
ascertained; of these, 723 (91.1%) were interviewed with a proxy respondent. Control groups for
each of the cancer sites were selected through a two-step process. First, each cancer site was
stratified by age, vital status, year of death (if applicable), and gender. Then, all controls that fell
within a stratum with at least one case were chosen. Index years for each control group were
determined based on the median year of diagnosis for the case group. Controls that moved to the
Upper Cape Cod area after the index year, cases or controls with incomplete residential histories,
and controls for which no tetrachloroethylene data were available were subsequently excluded.
In-person (14%) and telephone (86%) interviews with participants, conducted by trained
interviewers, inquired about a 40-year residential history, demographics, smoking, medical and
occupational histories and exposures, bottled water consumption, and usual bathing habits. The
articles did not provide estimates of proxy interviews. Cases and controls were similar in race,
age, marital status, and religion.
Aschengrau et al. (1998; 1993) assessed exposure through relative delivered dose (RDD)
of tetrachloroethylene via contaminated water estimated using Webler and Brown's (1993)
algorithm, which was based on a tetrachloroethylene-leaching model by Demond (1982). The
algorithm accounted for information about the water pipe that supplied each person's home,
B-49
-------
including water flow and pipe characteristics. Inputs were determined using maps from local
water suppliers or the Massachusetts Department of Environmental Protection. The exposure for
cases and controls was assessed by one individual blinded to the individual's case/control status
with a high degree of intraobserver and interobserver agreement. An ordinal estimate of
exposure to tetrachloroethylene-contaminated water was defined as the estimated mass of
tetrachloroethylene that entered the home through the drinking water during a specified period.
The estimates were first categorized as "never exposed" (private wells) and "ever exposed," with
the latter further categorized as "low" (up to and including median) and "high" (above the 50th,
75th, and 99th percentiles). The estimates based on Webler and Brown (1993) were recently
found to correlate with historically measured tetrachloroethylene levels, demonstrating the
algorithm's value in epidemiological research (Spence et al., 2008). RDDs were calculated for
those that had more than one exposed residence and were categorized into low (<50th percentile
of cumulative exposure among the exposed women), >50th, >75th, and >90th percentiles.
Aschengrau et al. (1993) evaluated the relationship between tetrachloroethylene-
contaminated drinking water and bladder cancer, kidney cancer, and leukemia separately. Cases
consisted of men and women of all ages who were diagnosed with incident bladder cancer,
kidney cancer, or leukemia. Of the 79 bladder cancer cases, 42 kidney cancer cases, and
44 leukemia cases, 72 (91.1%) bladder cancer, 36 (85.7%) kidney cancer, and 38 (86.4%)
leukemia cases were eligible and contacted. Of these, 63 (87.5%) bladder cancer, 35 (97.2%)
kidney cancer, and 35 (92.1%) leukemia cases participated in the study. After employing the
two-step control selection process and the additional exclusion criteria, the final sample
consisted of 61 bladder cancer cases and 852 bladder cancer controls, 35 kidney cancer cases and
777 kidney cancer controls, and 34 leukemia cases and 737 leukemia controls. Industries and
job titles were coded according to standard industrial (1987) and occupational (1990)
classifications. Occupational exposure to tetrachloroethylene was based on industry and job
titles, as well as specific questions posed during the interview. Overall, 34.4% bladder cancer
cases, 26.2% bladder cancer controls, 25.7% kidney cancer cases, 25.2% kidney cancer controls,
35.3% leukemia cases, and 25.3% leukemia controls reported occupational exposure to solvents
including tetrachloroethylene. Overall, there were 13 (21.3%) bladder cancer cases, 127 (14.9%)
bladder cancer controls, 6 (17.1%) kidney cancer cases, 112 (14.4%) kidney cancer controls,
7 (20.6%) leukemia cases, and 94 (12.8%) leukemia controls with any exposure to
tetrachloroethylene through drinking water without considering a latency period. Unadjusted
odds ratios were estimated for all sites with at least two exposed cases, stratified by bottled water
consumption and bathing habits separately. The Fisher exact test was used to estimate
corresponding 95% CIs. These analyses were performed with and without the assumption of a
latency period of 15 years for bladder and kidney cancer and 5 years for leukemia. Multiple
B-50
-------
logistic regression was used to estimate odds ratios adjusted for sex, age at diagnosis for cases or
index year for controls, vital status at interview, education, and occupational exposures.
Additional potential confounders were included if present in at least three or more cases. This
consisted of prior medical treatment with irradiation in the leukemia analysis, usual number of
cigarettes smoked, and history of a urinary tract infection or stone in the kidney cancer analysis,
and usual number of cigarettes smoked, history of a urinary tract infection or stone, and history
of a cancer-associated job in the bladder analysis. Maximum likelihood estimates of the standard
errors were used to estimate corresponding 95% CIs. Strengths of this study include its ability to
control for a variety of potential confounders, including occupational exposures, and its
examination of the effect of latency periods on the different cancers. Limitations to the study
include a small bladder cancer sample size to examine the effect of a latency period, unknown
levels of exposures, and nonblinded interviews.
Both Aschengrau et al. (1998) and Paulu et al. (2002) examined the relationship between
tetrachloroethylene-contaminated drinking water and breast cancer in women diagnosed between
1983 and 1986. Of the 334 breast cancer cases, 295 (88.3%) were eligible, and 265 (89.8%)
were interviewed. There were 763 controls identified through the two-step control selection
process. After employing the additional exclusion criteria, the final sample consisted of
258 cases and 686 controls.
Aschengrau et al. (1998) reported 36 (14%) exposed cases and 81 (11.8%) exposed
controls without considering a latency period. Latency periods of 5, 7, 9, 11, 13, and 15 years
were also evaluated. Unadjusted odds ratios and their corresponding 95% CIs examined crude
associations and potential modifiers. Multiple logistic regression was used to calculate odds
ratios adjusted for age at diagnosis or index year, vital status at interview, family history of
breast cancer, age at first live birth or stillbirth, personal history of prior breast cancer and benign
breast disease, and occupational exposure to solvents. Maximum likelihood estimates of the
standard errors were used to estimate corresponding 95% CIs. A strength of the study is its
adjustment for a variety of potential confounders. Limitations include potential for measurement
error in exposure estimates, small numbers of women, and possible misclassification due to
inaccurate reporting on death certificate of control's address or cause of death.
Paulu et al. (1999) studied the relationship between tetrachloroethylene-contaminated
drinking water and colon-rectum, lung, brain, and pancreatic cancer cases between 1983 and
1986. Of the 420 colon-rectum, 326 lung, 42 brain, and 43 pancreatic cancer cases selected,
366 (87.1%) colon-rectum, 272 (83.4%) lung, 40 (95.2%) brain, and 39 (90.7%) pancreatic
cancer cases were contacted and eligible. Of these, 326 (89.1%) colon-rectum, 252 (92.6%)
lung, 37 (88.1%) brain, and 37 (86.1%) pancreatic cancer cases were interviewed for an overall
participation rate of 79%. The final sample consisted of 311 colon-rectum cancer cases and
B-51
-------
1,158 colon-rectum cancer controls, 243 lung cancer cases and 1,206 lung cancer controls,
36 brain cancer cases and 703 brain cancer controls, and 36 pancreatic cancer cases and
622 pancreatic cancer controls. Excluding any latent periods, exposure assessments were as
follows: colon-rectum cancer had 44 (14.1%) cases and 153 (13.2%) controls; lung cancer had
33 (13.6%) cases and 158 (13.1%) controls; brain cancer had 3 (8.3%) cases and 92 (13.1%)
controls; and pancreatic cancer had 3 (8.3%) cases and 81 (13.0%) controls. Due to their low
numbers of "ever exposed," unadjusted estimates of odds ratios and their corresponding 95% CIs
were calculated for brain and pancreatic cancer cases. Multiple logistic regressions was used to
estimate the odds ratios and 95% CIs for colon-rectum and lung-cancer cases, adjusted for age at
diagnosis or index year, vital status at interview, sex, and occupational exposure to
tetrachloroethylene and other solvents. Colon-rectum cancer was further adjusted for history of
polyps, inflammatory bowel disease, and occupational history associated with colon-rectum
cancer. Lung cancer was further adjusted for usual number of cigarettes smoked and history of
cigar/pipe use, living with a smoker, and occupational history associated with lung cancer.
Latency periods of 0, 5, 7, 9, 11, 13, and 15 years were considered in the analyses. Strengths of
this study are its adjustment for confounders and consideration of a latency period. Limitations
include a lack of measured tetrachloroethylene levels, lack of adjustment for smoking,
particularly for lung cancer, and low-exposure prevalence, particularly for brain and pancreatic
cancer cases.
Paulu et al. (2002) examined residential location using GIS-coded information. The
40-year residential history obtained during the interview included full addresses and calendar
years of residence. If the complete address was unknown, tax assessors' books were used to help
identify the geographical location. All participants were then blindly mapped onto an enlarged
version of a U.S. Geological Survey map, which was later converted into a digital format. The
Upper Cape Cod area was divided into subregions with two methodologies: the first employed
fixed, multiscale grids and coded each participant as ever exposed or unexposed for each grid
cell; the second used overlapping circles (adaptive k-smoothing) whose sizes were based on the
number of nearby cases and controls. Crude and adjusted odds ratios were estimated for both the
grid and k-smoothed methodologies, using map choropleths for visualization. These maps
facilitate visualization of "hot spots" for microscale residence. Multiple logistic regression was
used to estimate odds ratios for breast cancer, adjusted for age, parity, vital status, family history
of breast cancer in a first-degree female relative, age at first live birth or stillbirth, and prior
history of breast cancer or benign breast disease. No strengths were reported by the authors for
this study; a limitation was the study's lack of individual measurements of household
tetrachloroethylene exposures.
B-52
-------
B.2.1.3.2. Aschengrau et al. (2003), Vieira et al. (2005)
Aschengrau, A.; Rogers, S.; Ozonoff, D. (2003). Perchloroethylene-contaminated
drinking water and the risk of breast cancer: Additional results from Cape Cod,
Massachusetts, USA. Environ Health Perspect, 111, 167-173.
http://www.ncbi.nlm.nih.gov/pubmed/12573900
Vieira, V.; Aschengrau, A.; Ozonoff, D. (2005). Impact of tetrachloroethylene-
contaminated drinking water on the risk of breast cancer: Using a dose model to
assess exposure in a case-control study. Environ Health, 4, 3.
http://dx.doi.org/10.1186/1476-069X-4-3
Summary: These population-based case-control studies were conducted as a follow-up to
Aschengrau et al. (1998) and Paulu et al. (2002) to examine breast cancer and drinking water
exposure. Cases were identified through the Massachusetts Cancer Registry and consisted of
women diagnosed with breast cancer between 1987 and 1993, a period after that examined in the
earlier studies. In contrast to the two-stage control selection process in earlier studies, controls
were selected in three ways: (1) random-digit dialing (women <64 years), (2) random selection
from a Medicare beneficiary roster (>65 years), or (3) random selection from among death
certificates provided by the Massachusetts Bureau of Health Statistics, Research, and Evaluation.
Controls were matched to cases based on age and vital status at the time of identification. The
final sample consisted of 672 cases and 616 controls, of which 211 (31.4%) cases and
192 (31.2%) controls were nonproxy respondents. Structured interviews were conducted with
participants and next of kin to obtain information on demographics, confounders (age at
diagnosis, family history of breast cancer, personal history of prior breast cancer, age at first live
birth/stillbirth, occupational exposure to tetrachloroethylene, etc.), potential effect modifiers
(bathing habits, bottled water, and water filter use), as well as a 40-year residential history. The
authors do not state if these were in-person or blinded interviews.
Aschengrau et al. (2003) further examined the hypothesis that tetrachloroethylene
exposure via contaminated drinking water increases the risk of breast cancer. Overall, 672 cases
(81% selected and eligible cases) and 616 controls (157 [83%] random-digit dialed, 301 [76%]
of Medicare roster, and 158 [79%] deceased) were included in the analysis. ROD of
tetrachloroethylene via contaminated water was estimated using Webler and Brown's (1993)
algorithm, which was based on a tetrachloroethylene leaching model by Demond (1982). The
algorithm accounted for information about the water pipe that supplied each person's home,
including water flow and pipe characteristics. Inputs were determined using maps from local
water suppliers or the Massachusetts Department of Environmental Protection. The exposure for
cases and controls was assessed by one individual blinded to the individual's case-control status
with a high degree of intraobserver and interobserver agreement. An ordinal estimate of
exposure to tetrachloroethylene-contaminated water was defined as the estimated mass of
B-53
-------
tetrachloroethylene that entered the home through the drinking water during a specified period.
The estimates were first categorized as "never exposed" (private wells) and "ever exposed," with
the latter further categorized as "low" (up to and including median) and "high" (with
categorization as >50th, >75th, and >99th percentiles). The estimates based on Webler and Brown
(1993) were recently found to correlate with historically measured tetrachloroethylene levels,
demonstrating the algorithm's value in epidemiological research (Spence et al., 2008). Overall,
there were 155 (23.1%) cases and 136 (22.1%) controls exposed to tetrachloroethylene. Data
analysis included the following latent periods: 0, 5, 7, 9, 11, 13, 15, 17, and 19 years. Exposure
odds ratios and their corresponding 95% CIs estimated crude associations. Multiple logistic
regression was used to estimate odds ratios, adjusted for age at diagnosis or index year, vital
status at interview, family history of breast cancer, personal history of breast cancer, age at first
live birth or stillbirth, and occupational exposure to tetrachloroethylene. Maximum likelihood
estimates of the standard error were used to calculate corresponding 95% CIs.
Viera et al. (2005) further studied the associations between tetrachloroethylene exposure
and breast cancer. Due to the fact that the majority of the relevant interview information was
only collected from nonproxy cases and controls, proxy interviews were excluded from the
analyses, though included in comparisons with the total sample. Nonproxy information obtained
through the interviews included the daily consumption of tap water or drinks that used tap water
(number of drinks), bottled water consumption, and the temperature, frequency, and duration of
showers and baths. Data not collected in the interviews, such as inhalation rate, water flow rate,
and air exchange rate, were ascertained from the literature. The authors did not provide
references for these obtained rates. In contrast to Aschengrau et al. (2003), this study estimated
the personal delivered dose (FDD) for each participant by adding the amount inhaled, dermally
absorbed, and ingested together for each exposed residence. Inhalation was estimated from
reported temperature, frequency, and duration of baths and showers, as well as from the amount
of tetrachloroethylene in the bath/shower air. Dermal absorption was estimated according to
Pick's first law and used height and weight data to calculate each participant's surface area.
Ingestion was based on the volume of tap water the participant drank. RDDs were re-estimated
for the nonproxy participants only, and both the RDD and FDD were used to classify each
participant into nested exposure levels: <50th percentile, >50th, >75th, and >90th percentiles.
Latency periods of 0, 5, 7, 9, 11, 13, 15, 17, and 19 years were employed. Without considering a
latency period, the full sample contained 155 (23.1%) exposed cases and 136 (22.1%) exposed
controls, and the nonproxy sample contained 101 (21.9%) exposed cases and 88 (20.8%)
exposed controls. Crude and adjusted analyses were conducted for both the RDD and FDD
levels, though adjusted analyses were limited to those with at least three exposed cases and at
least three exposed controls. Multiple logistic regression was used to estimate adjusted odds
B-54
-------
ratios, controlling for the following confounders, which were identified a priori: age at diagnosis
or index year, family history of breast cancer, personal history of breast cancer, age at first live
birth or stillbirth, and occupational exposure to tetrachloroethylene. Maximum likelihood
estimates of the standard errors were employed to estimate corresponding 95% CIs. All
nonproxy estimates were subsequently compared to estimates for all subjects. The full sample's
adjusted odds ratios (ORs) further controlled for vital status at interview. A goodness-of-fit test
compared the RDD and the FDD to ascertain which was a better measure, and a nonparametric
rank test evaluated whether RDD and FDD exposures differed significantly from each other. A
strength of this study is its incorporation of personal behaviors in estimating exposure,
examination of nonproxy respondents, considered to provide more correct information than
proxy respondents and to reduce misclassification bias, and comparison of results from only
nonproxy respondents to results of all subjects. Limitations include its use of cumulative
exposures, which may mask the effect of intensity of exposure, recall bias for behavioral data,
and decreased sample size due to the use of nonproxy respondents only.
B.2.1.4. New Zealand
B.2.1.4.1. Corbin et al. (2011), Dryson et al. (2008), 't Mannetje et al. (2008), McLean et al.
(2009)
Corbin, M.; McLean, D.; Mannetje, A.; Dryson, E.; Walls, C.; McKenzie, F.,...
Pearce, N. (2011). Lung cancer and occupation: A New Zealand cancer registry-
based case-control study. Am J Ind Med, 54, 89-101.
http://dx.doi.org/10.1002/aiim.20906
Dryson, E.; 't Mannetje, A.; Walls, C.; McLean, D.; McKenzie, F.; Maule, M.,...
Pearce, N. (2008). Case-control study of high risk occupations for bladder cancer in
New Zealand. Int J Cancer, 122,1340-1346. http://dx.doi.org/10.1002/iic.23194
't Mannetje, A.; Dryson, E.; Walls, C.; McLean, D.; McKenzie, F.; Maule, M.,...
Pearce, N. (2008). High risk occupations for non-Hodgkin's lymphoma in New
Zealand: Case-control study. Occup Environ Med, 65, 354-363.
http://dx.doi.org/10.1136/oem.2007.035014
McLean, D.; Mannetje, A.; Dryson, E.; Walls, C.; McKenzie, F.; Maule, M.,...
Pearce, N. (2009). Leukaemia and occupation: A New Zealand Cancer Registry-
based case-control Study. Int J Epidemiol, 38, 594-606.
http://dx.doi.org/10.1093/iie/dvn220
Summary: The case-control studies of Dryson et al. (2008), Mannetje et al. (2008),
McLean et al. (2009), and Corbin et al. (2011) are part of an ongoing series of studies examining
the relationship between occupation and cancer in the New Zealand population. Cases were
identified through the New Zealand Cancer Registry from 2003 to 2004 in Dryson et al. (2008),
't Mannetje et al. (2008), and McLean et al. (2009) and from 2007 to 2008 in Corbin et al.
B-55
-------
(2011). Population-based controls were randomly chosen from the 2003 New Zealand Electoral
Roll and matched to cases based on age. Of 1,200 potential controls initially mailed letters of
invitation, 1,100 had valid addresses. Of these, 660 were able to be contacted and considered
eligible to participate. Overall, 473 controls were interviewed, with an overall response rate of
48%. After excluding controls with missing information for key variables, the final sample
consisted of 471 controls. Controls in Corbin et al. (2011) were identified from 2003 to 2008
with letters of invitation mailed to 2,000 individuals, 1,878 of whom had valid addresses. Of
these, 1,134 replied, and 796 were interviewed (48% response rate). In Dryson et al. (2008), 't
Mannetje et al. (2008), and McLean et al. (2009), cases and controls were similar in occupational
class, with the exception of the lowest class, which was more prevalent among cases than
controls. In Corbin et al. (2011), "ever" smoking was more frequent among cases than among
controls, as might be expected in a study of lung cancer, and the frequency of subjects 71 years
of age and older was higher among controls than among cases.
In-person interviews were conducted with a trained interviewer whose background was in
occupational health nursing. The questionnaire inquired about demographics, smoking, and
occupational history, and more detailed information was obtained on all jobs lasting longer than
1 year. Occupation was assessed as a proxy for exposure, and jobs were blindly coded according
to the 1999 New Zealand Standard Classification of Occupations and the Australian and New
Zealand SIC. The authors do not report who assigned the codes. Occupation Code 8264
consisted of textile bleaching, dyeing, and cleaning machine operators and was considered
a priori to be high risk. The authors did not refer specifically to dry-cleaners or laundry workers.
McLean et al. (2009) noted workers in this occupational group had similar exposures as laundry
and dry cleaning occupations. In addition, Corbin et al. (2011) presented analyses separately for
occupational titles of dry cleaner and launderer. Unconditional logistic regression was used to
calculate odds ratios and their corresponding 95% CIs for occupations and industries considered
a priori and a posteriori to be high risk. All estimates were adjusted for 5-year age group, sex,
smoking ("ever," "ex," "never"), Maori ethnicity, and occupational status. Semi-Bayes
adjustments were performed to minimize the risk of false positive results due to multiple
comparisons. These adjustments were performed using an estimate of the variation that was
determined a priori. Strengths of the study design include its population-based design, near
complete coverage of both incident cancers and the general population, and adjustment for
smoking. Additionally, interviews were conducted in person and obtained detailed occupational
histories. The limitations of the study include the lack of an exposure profile, information on
duration or length of employment for only certain occupations or chemicals, and possible
selection bias due to the low-response rates of cases and controls, though McLean et al. (2009)
noted the similarity between the distribution of occupations in the national census and the
B-56
-------
sample. Additional limitations include the study's low exposure prevalence and possible
exposure misclassification due to the use of broad occupational categories.
Dryson et al. (2008). 't Mannetje et al. (20081 and Corbin et al. (2011) included men and
women aged 25 to 70 years who were diagnosed with either bladder cancer or non-Hodgkin
lymphoma. The authors did not state if the cases were histologically confirmed. Dryson et al.
(2008) studied bladder cancer among selected occupations that may contribute to the risk of
bladder cancer. Of the 381 cases in Dryson et al. (2008) identified from the New Zealand
Cancer Registry, 232 (60.9%) were able to be contacted by mail and eligible to participate. In
total, 213 cases were interviewed for the study, with an overall response rate of 64%. The final
sample consisted of 213 cases and 471 controls. Approximately 77% of cases and 47% of
controls were male. "Current" and "ever" smoking were more prevalent among cases than
controls. There were 3 (1.4%) bladder cancer cases and 10 (2.1%) controls that reported
employment in the bleaching, dyeing, and cleaning machine occupations.
't Mannetje et al. (2008) aimed to assess whether previously reported associations (Reif
et al., 1989; Pearce etal., 1988; Pearce et al., 1987; Pearce et al., 1985) between occupations and
non-Hodgkin lymphoma persist, and to identify other occupations that may also contribute to the
risk of non-Hodgkin lymphoma in the New Zealand population. Of the 533 cases identified from
the cancer registry, 335 (62.9%) were able to be contacted and eligible to participate. In total,
291 cases were interviewed for the study, with a response rate of 69%. The final sample
consisted of 291 cases and 471 controls. Approximately 54% of cases and 47% of controls were
male, and current smoking was more common among cases than controls. There were 5 (1.7%)
cases and 10 (2.1%) controls that reported employment as a textile bleaching, dyeing, and
cleaning machine operators.
McLean et al. (2009) studied the relationship between occupation and leukemia (chronic
lymphocytic leukemia, acute myeloid leukemia, chronic myeloid leukemia, acute lymphoblastic
leukemia, and other forms of leukemia). Cases consisted of men and women aged 20 to 75 years
who were added to the registry between 2003 and 2004. The authors did not state if the cases
were histologically confirmed. Of the 391 eligible cases, 225 (57%) participated in the
interview; 11 (4.9%) of which were proxy interviews with next of kin. The final sample
consisted of 225 cases and 471 controls. Approximately 61% cases and 47% controls were male,
and a higher proportion of cases were current smokers than controls. Overall, 6 (2.7%) cases and
10 (2.1%) controls comprised the textile bleaching, dyeing, and cleaning machine occupation.
Corbin et al. (2011) examined lung cancer and occupation to support previously
identified risk factors and to identify new risk factors. Of 744 eligible lung cancer cases, aged
20-75 years in Corbin et al. (2011), 458 were interviewed (53% response rate). Among those
interviewed, 432 of the 796 cases were by phone, and all interviews were with living subjects.
B-57
-------
Face-to-face interviews were carried for the remaining 432 control subjects. Overall, 20 cases
and 13 controls were employed as textile bleaching, dyeing, and cleaning machine operators,
with 3 of these cases and 4 controls identified as dry cleaners. An additional 9 cases and
5 controls were identified as launderers.
B.2.1.5. Germany
B.2.1.5.1. Pesch et al. (2000a, b)
Pesch, B.; Haerting, J.; Ranft, U.; Klimpel, A.; Oelschlagel, B.; Schill, W. (2000a).
Occupational risk factors for renal cell carcinoma: Agent-specific results from a
case-control study in Germany. Int J Epidemiol, 29,1014-1024.
http://dx.doi.0rg/10.1093/iie/29.6.1014
Pesch, B.; Haerting, J.; Ranft, U.; Klimpel, A.; Oelschlagel, B.; Schill, W. (2000b).
Occupational risk factors for urothelial carcinoma: Agent-specific results from a
case-control study in Germany. Int J Epidemiol, 29, 238-274.
http://dx.doi.Org/10.1093/iie/29.2.238
Summary: Between 1991 and 1995, a population-based case-control study was conducted
in five regions of Germany to independently estimate the risk of urothelial cancer and renal cell
cancer as functions of exposure to aromatic amines, poly cyclic aromatic hydrocarbons (PAHs),
and other chlorinated hydrocarbons. Cases were identified through the large hospitals in each of
the regions and consisted of German men and women who were diagnosed with histologically
confirmed urothelial or renal cell cancer within the 6 months prior to the start of the study.
Controls were randomly selected from local residency registries and matched to cases on age,
sex, and region. In order to be included in the study, cases and controls were required to be
German nationals; there were no age limits during the recruitment process.
In-person interviews with trained interviewers occurred with cases in the hospital within
the first 6 months of diagnosis and with controls in their home. A structured questionnaire
inquired about demographics, lifestyle, and occupational exposures. The final sample consisted
of 1,970 cases (1,035 urothelial cancer and 935 renal cell cancer) and 4,298 controls, with
overall response rates of 84% for cases and 71% for controls. Exposure was assessed based on
the participant's reported occupational history, exposure to specific agents during tasks, and
average amount of time each day exposed. All jobs held for at least 1 year were coded according
to the International Standard Classification of Occupations. Lifetime exposure was calculated as
the total number of years spent at a specific job title; task- and agent-specific exposures were
estimated as weighted sums of years spent at that task or exposed to the agent in question. Job
exposure matrices (JEMs) and job-task exposure matrices (JTEMs) were also used for
calculating exposure to specific agents, including tetrachloroethylene. These matrices evaluated
the probability and intensity of exposure. The JEM, which assessed exposure based on job title,
B-58
-------
used both the British (Pannett et al., 1985) and the German (Robra and Seidler, 1994) versions.
The ITEM was developed by the researchers and adjusted for both region and time. Both studies
used conditional logistic regression to calculate odds ratios and their corresponding 95% CIs for
potential confounders, occupations and tasks, and substances separately, adjusted for age, study
center, and smoking. Potential confounders were also stratified by gender, and additional
analyses were adjusted for age and study center without smoking. Strengths of these studies
include population-based selection of controls and the use of a JEM and a ITEM to assess
substance exposure. Limitations include the lower response rate of controls compared to cases
and the reliance of self-reported information for exposure assessment.
Pesch et al. (2000a, b) estimated the urothelial cancer risk for occupational exposure to
aromatic amines, PAHs, and chlorinated hydrocarbons besides other suspected risk factors. This
study sample included 1,035 urothelial cancer cases and 4,298 controls. When
tetrachloroethylene was assessed using the German JEM, there were 183 (17.7%) cases with
medium exposure, 188 (18.2%) cases with high exposure, and 74 (7.1%) cases with substantial
exposure. The ITEM approach, however, only identified 37 (3.6%) cases with medium
exposure, 47 (4.5%) with high exposure, and 22 (2.1%) with substantial exposure.
Pesch et al. (2000a, b) examined the possible impact of occupation-related agents on
renal cell cancer development. The sample in this study consisted of 935 renal cell cancer cases
and 4,298 controls. When tetrachloroethylene was evaluated using the German JEM, there were
166 (17.8%) cases with medium exposure, 138 (14.8%) cases with high exposure, and 54 (5.8%)
cases with substantial exposure. The JTEM approach, however, identified only 52 (5.6%) cases
with medium exposure to tetrachloroethylene, 45 (4.8%) with high exposure, and 18 (1.9%) with
substantial exposure.
B.2.1.6. Nordic Countries
B.2.1.6.1. Lynge et al. (2006)
Lynge, E.; Andersen, A.; Rylander, L.; Tinnerberg, H.; Lindbohm, M. L.; Pukkala,
E.,... Johansen, K. (2006). Cancer in persons working in dry cleaning in the Nordic
countries. Environ Health Perspect, 114, 213-219. http://dx.doi.org/10.1289/ehp.8425
Summary: This study of a nested case-control design within a cohort examined eight site-
specific cancers (non-Hodgkin lymphoma, esophageal, gastric cardia, liver, pancreatic, cervix
uteri, kidney, and bladder), and job title, distinguishing between dry-cleaning workers, a proxy
for tetrachloroethylene, and other job titles such as laundry workers. The cohort from which cases
and controls arose consisted of 46,768 individuals identified as laundry and dry-cleaning workers
in the 1970 Censuses in Denmark, Finland, Norway, and Sweden. All were followed for death,
emigration, and incident cancer based on nationwide population, death, and cancer registries.
B-59
-------
Relevant cancer cases were identified as those that occurred during the period of November 1970
(Denmark) or January 1971 (Finland, Norway, Sweden) through 1997 to 2001. Controls were
randomly selected from the cohort and matched based on country, sex, 5-year age group, and
5-year calendar period at the time of diagnosis. All analyses were conducted at the level of the
record rather than person because a subject may have appeared as a case or as a cancer control in
the study more than once. Out of 4,014 records from 3,883 persons, 131 subjects were
considered both as a case and as a cancer control.
Lynge et al. (2006) used job title and occupational task identified in the 1970 Census to
identify tetrachloroethylene exposure potential. Differing recordkeeping systems and record
availability in each country necessitated a number of approaches for assigning exposure potential
to cases and controls; Johansen et al. (2005) provides an in-depth description of available records
for Danish subjects and Lynge et al. (2011) of available records in the four studied countries. In
Denmark and Norway, occupational task identified on the 1970 Census form was available and
used to identify subjects as (1) dry-cleaners or other workers in dry-cleaning shops with <10
workers, assumed to have high-exposure potential because of the shared work tasks and physical
proximity in small dry-cleaning shops; (2) other workers in dry-cleaning shops; (3) unexposed
laundry workers and other persons in dry cleaning, and (4) unclassifiable, a category for subjects
with missing employment information. Pension data from Denmark and Finland, as well as a
Danish biography of dry-cleaning shop owners, were used to identify length of employment, a
proxy for cumulative exposure, between 1964 and 1979, and size of workforce, for self-
employed subjects. For subjects from Norway and Sweden, a blinded telephone interview was
undertaken, given the lack of storage of the 1970 Census forms. The questionnaire asked about
occupational task for job title reported on the 1970 Census form, and if dry cleaning, questions
sought answers on employment length, number of employees, solvents used, and personal habits
of smoking and alcohol consumption. Interviews were obtained with 148 of 258 of cases (57%)
and 293 of 457 controls (64%) in Norway; for which 107 cases subjects (72%) and 123 control
subjects (42%) were with proxy respondents. For Swedish subjects, interviews were obtained
with 369 of 586 cases (63%) and 454 of 756 controls (60%) controls; for which 284 case
subjects (77%) and 177 control subjects (39%) were with proxy respondents.
In total, the study included 1,616 cases and 2,398 controls, with roughly two-thirds
(68%) of subjects from Denmark and Sweden. There were 695 (17.3%) cases and controls who
were exposed due to their work as dry cleaners, 183 (5%) exposed through other work in a
dry-cleaning shop, and 716 (18%) for whom information on employment and exposure potential
could not be obtained and were identified as "unclassifiable." The percentage of subjects
identified as "unclassifiable" varied by country, with no subjects from Denmark, 41% of all
subjects from Finland, 2% of all subjects from Norway, and 35% of all subjects from Sweden.
B-60
-------
Lynge et al. (2006) provided some exposure monitoring data, particularly for 1964-1979,
the period examined in analyses of exposure duration. Although a large variation in exposure
levels was observed in 168 samples from Nordic dry-cleaning shops, the median concentrations
over this period were relatively stable and appeared to range from 3-12 ppm. Lynge et al.
(2006) reported a mean of 24 ppm from 53 samples of >60 minutes in length. Lynge et al.
(2011) provided exposure monitoring from dry clean facilities for the period 1947-2000
(Denmark), 1956-1999 (Finland), 1976-2001 (Norway) and 1973-1995 (Sweden). Personal
monitoring was not available before 1978; roughly 90% of the stationary air measurements were
from the period after 1975The mean of stationary measurements over the monitoring period was
11.92 ppm (95% CI: 10.66, 13.51) and, for personal measurements, 7.27 ppm (95% CI; 6.78,
7.79). Based on the stationary monitoring measurements, the exposure of maintenance
workers>dry cleaners>shop assistants, with means of 35.94 ppm (95% CI: 20.92, 55.99), 13.20
ppm (95% CI:, 11.13, 15.24), and 7.50 (95% CI: 6.84, 8.02), respectively. Personal monitoring
measurements were lower than stationary monitoring measurements for dry cleaners and shop
assistants, no personal monitoring was available for maintenance workers, and suggested job title
was not a predictor of exposure intensity. Mean exposures were 7.50 ppm (95% CI: 6.73, 7.94)
for dry cleaners and 6.25 ppm (95% CI: 4.09, 8.93) for shop assistants. Exposure by job title of
dry cleaner and shop assistant varied in Denmark and Finland, whereas little or no difference was
indicated from monitoring data from Sweden and Norway.
Rate ratios (RRs) for dry cleaners versus unexposed controls were estimated using
logistic regression. RRs were also calculated for the other persons in dry cleaning and for the
unclassifiable persons, although the underlying hypothesis did not include these groups. RRs
were estimated for all countries together and for Denmark and Norway together given their lower
percentage of unclassifiable subjects compared to that for Finland (41%) or Sweden (35%). The
researchers adjusted for the matching criteria, as well as smoking and alcohol use (Norway and
Sweden only) in bladder cancer analyses that showed smoking as not greatly affecting observed
risk estimates. Strengths of the study include its coverage of the period where
tetrachloroethylene was used as the main solvent, its population-based design, its use of a series
of nested case-control studies within the cohorts to examine specific cancers, its control for
smoking in bladder cancer analysis, and its examination of dry cleaner versus other dry-cleaning
tasks. A limitation is a lack of exposure monitoring data on individual subjects as industrial
hygiene data from 1964-1979 showed a large variation in tetrachloroethylene concentrations
across shops. Additionally, the large number of next-of-kin interviews in cases from Sweden
and Norway; a control series which included cases with other cancers of a priori interest (8% of
case series); assessment of tetrachloroethylene exposure potential for one job, that was held in
1970, versus for the full employment history; and, censoring employment duration to 1979 rather
B-61
-------
than for the study's full period, to 1998 or 2001 (depending on country) likely introduces
misclassification bias. Finally, a large number of subjects from Sweden and Finland had missing
information. If differential reporting of job title and occupational tasks was associated with
exposure as a dry cleaner and status as a case or control, then a misclassification bias would be
introduced. Lynge et al. (2006) explored the magnitude of this potential bias on esophageal
cancer estimates, noting if all unclassified subjects were exposed as dry cleaners, observed odds
ratio (0.76; 95% CI: 0.34, 1.69) would increase (to 1.19; 95% CI: 0.67, 2.12), and, if unexposed,
would decrease (to 0.66; 95% CI: 0.30, 1.45).
B.2.2. Single Cancer-Site Studies
B.2.2.1. Bladder Cancer
B.2.2.1.1. Burns and Swanson (1991), Swanson and Burns (1995)
Burns, P. B. and Swanson, G. M. (1991). Risk of urinary bladder cancer among
blacks and whites: The role of cigarette use and occupation. Cancer Causes Control,
2, 371-379. http://dx.doi.org/10.1007/BF00054297
Swanson, G. M. and Burns, P. B. (1995). Cancer incidence among women in the
workplace: A study of the association between occupation and industry and 11
cancer sites. J Occup Environ Med, 37, 282-287.
http://www.ncbi.nlm.nih.gov/pubmed/7796194
Summary: This population case-control study is part of the Occupational Cancer
Incidence Surveillance Study examining occupation and 11 cancer sites. Burns and Swanson
(1991) examined cigarette smoking and occupational title and bladder cancer, with Swanson and
Burns (1995) focusing on occupation and cancer in women. The Metropolitan Detroit Cancer
Surveillance System (MDCSS) was used to identify cancer cases at 11 sites (lung, colon, rectum,
bladder, esophagus, liver, salivary gland, stomach, eye, melanoma, and mesothelioma) among
males and females aged 40 to 84 years, diagnosed between 1984-1991. In all, 2,160 bladder
cancer cases and 3,979 cancer controls were interviewed by telephone for response rates of 94%
and 95% for cases and controls, respectively. Colon and rectal cancer cases from the registry
were selected as controls and not matched to cases based on demographic variables. Of those
interviewed, 25% of case series and 27.6% of the controls series were proxy or next-of-kin
respondents. The high percentage of proxy interviews may introduce potential for recall bias of
detailed occupational history. The interview gathered information on complete lifetime
occupation history, including occupation and industry titles, lifetime smoking history, medical
history, residential history, and demographic information. Occupation and industry data were
coded according to the three-digit codes of the 1980 U.S. Census Bureau classification. The
paper does not identify if occupational coding was carried out blinded to case or control status.
B-62
-------
Exposure prevalence was low for holding an occupation as dry-cleaning worker, 0.4% for cases
and 0.4% for controls, or for working in the dry-cleaning or laundry industry, 0.6% for cases and
0.6% for controls. Association with bladder cancer and occupation was examined using
unconditional logistic regression adjusted for cigarette smoking habits, race, gender, and age at
diagnosis and usual industry or occupation, defined as the longest period of employment.
B.2.2.1.2. Colt et al. (2004)
Colt, J.; Baris, D.; Stewart, P.; Schned, A.; Heaney, J.; Mott, L.,... Karagas, M.
(2004). Occupation and bladder cancer risk in a population-based case-control
study in New Hampshire. Cancer Causes Control, 15, 759-769.
http://dx.doi.Org/10.1023/B:CACO.0000043426.28741.a2
Summary: This population case-control study examined a number of risk factors
including occupation exposures for primary bladder cancer among New Hampshire residents,
aged 25-74 years. To be eligible for the study, subjects were required to have a listed telephone
number and speak English. Six hundred eighteen (n = 618) cases diagnosed over a 4-year
period, between July 1, 1994 and June 30, 1998, were identified from the New Hampshire
Cancer Registry and histologically confirmed; 459 were subsequently interviewed
(74% participation rate). Controls, shared with a study of nonmelanoma skin cancer in the
period 1993-1995 and frequency matched based on age and sex, were selected from population
lists of the New Hampshire Department of Transportation, if <65 years old, and from New
Hampshire Centers for Medicare and Medicaid Services. The study augmented the control
group, adding controls for bladder cancer cases diagnosed between July 1, 1995 to June 30,
1997. Interviews were carried out with 665 of the 990 potential controls (67% participation
rate). Little age difference existed between cases and controls, although cases were more likely
than controls to have a history of cigarette smoking, with current smokers twice as prevalent
among the cases as controls.
Subjects who agreed to participate in the study underwent a detailed in-person interview,
usually at their home, with questions on sociodemographic information, tobacco use, medical
history prior to the diagnosis, and lifetime work history. Each job reported in the occupation
history was coded according to the Standard Occupation Classification Manual scheme, with
codes of 7658 and 7657 for occupations in dry-cleaning and laundry service. For each
occupation, bladder cancer risk was estimated separately for men and women for each job held
after age 15 using unconditional logistic regression models adjusted for age and smoking status.
Additionally, the authors conducted a separate analysis of a priori suspect high-risk occupations,
that included dry-cleaner and laundry workers. Only five male case and five male controls
reported a job title of dry-cleaner and laundry workers, and the study authors did not report the
B-63
-------
associated odds ratio because of the small numbers. The low-exposure prevalence for dry-
cleaning and laundry work, as is typical of population case-control studies, greatly reduces the
statistical power of this study to examine bladder cancer.
B.2.2.1.3. Colt et al. (2011)
Colt, J. S.; Karagas, M. R.; Schwenn, M.; Baris, D.; Johnson, A.; Stewart, P.,...
Silverman, D. T. (2011). Occupation and bladder cancer in a population-based case-
control study in Northern New England. Occup Environ Med, 68, 239-249.
http://dx.doi.org/10.1136/oem.2009.052571
Summary: This population case-control study examined occupation and industry as risk
factors for urothelial bladder cancer, among residents of Maine, New Hampshire, and Vermont,
aged 30-79 years. A focus of the study was exposure to metal working fluids. All residents
newly diagnosed with a histologically confirmed carcinoma of the urinary bladder (including
carcinoma in situ) between September 1, 2001 and October 31, 2004 (Maine and Vermont) or
between January 1, 2002 and July 31, 2004 (New Hampshire) were eligible for study. Cases
were identified through a rapid patient ascertainment in each state using data from hospital
pathology departments, hospital cancer registries, and state cancer registries. A total of
1,878 eligible cases were identified with in-person interviews obtained from 1,213 (65%).
Further pathologic review determined 43 subjects did not have bladder cancer or had
nonurothelial carcinoma, leaving 1,170 cases. Controls were randomly selected from state motor
vehicle records, if aged 30-64 years, or Medicare or Medicaid records, if aged 65+ years, and
frequency matched to cases by state, sex, and age at diagnosis or control selection. Interviews
were carried out with 1,418 controls, 594 identified from driver records (65% of eligible) and
824 identified through Medicare/Medicaid roles (65% of eligible).
Case and control subjects were first mailed a questionnaire with follow-up by a home
visit where a trained interviewer administered a computer-assisted questionnaire that sought
information on all jobs held for at least 6 months since age 16 years, demographic information,
tobacco use, and other exposures. For certain occupations held by subjects, a job-specific
questionnaire was administered, soliciting detailed information about exposures of interest. Each
job was coded blinded to case or control status to the 1980 SOC and the 1987 SIC scheme. Of
the 1,170 cases and 1,418 controls, 1,158 cases and 1,402 controls completed both
questionnaires.
For each occupation and industry, bladder cancer risk was estimated separately for men
and women for each job using unconditional logistic regression models adjusted for age, race,
Hispanic ethnicity, state, smoking status, and employment in a high-risk occupation. A high-risk
occupation was defined for men and women separately if odds ratios in the current study were
B-64
-------
1.5 or higher and 10 or more subjects were employed in the category examined in statistical
analyses. Additionally, the authors evaluated smoking effects, replacing smoking status with
smoking duration and found minimal changes in the estimated odds ratios; final statistical
models were thus adjusted for smoking status only. Interactions between smoking and
occupation were tested, adding cross-product terms to the logistic model. Additionally, the
authors examined employment duration for occupations and industries with a positive
association for "ever'V'never" employed. Tests of linear trend were performed by treating the
median duration of employment among controls for each duration category as a continuous
variable, with a value of zero assigned to subjects never holding a job in the subject category.
For occupations with observed risk estimates that increased with increasing employment
duration, an examination of initial year of employment and bladder cancer risk was examined.
Strengths of the study include the population-based design, ascertainment of complete
occupational histories from direct interview with study participants, blind assignment of
exposure, and ability to adjust for smoking, employment in high-risk occupations, and other risk
factors. Limitations include the low prevalence of exposure as laundering and dry-cleaning
machine operators (0.5% of all cases) and lower participation rate (65%) for cases and controls.
However, the authors concluded that because study participation likely did not differ between
cases and controls in an exposure-dependant manner, observations would not be biased. Some
exposure misclassification is likely, given limited information for some cases and controls, and
the lack of information on specific exposures by using job title. Biases are likely nondifferential
and lead to dampened risk estimates.
B.2.2.1.4. Gaertner et al. (2004)
Gaertner, R. R. W.; Trpeski, L.; Johnson, K. C. (2004; 1995). A case-control study
of occupational risk factors for bladder cancer in Canada. Cancer Causes Control,
15,1007-1019. http://dx.doi.org/10.1007/sl0552-004-1448-7
Summary: This population-based case control study of bladder cancer in seven Canadian
provinces (Newfoundland, Prince Edward Island, Nova Scotia, Manitoba, Alberta,
Saskatchewan, and British Columbia) made use of data collected in the Canadian National
Enhanced Cancer Surveillance System. The project collected data on cases of various cancers
and controls, with the intent of improving knowledge of environmental factors in cancer
development. Cases were identified through each province's cancer registry and consisted of
men and women aged 20 to 74 years who were diagnosed with histologically confirmed bladder
cancers between 1994 and 1997. Controls were randomly selected from the general population
of the seven provinces using simple random digit dialing (Newfoundland and Alberta) or
sampling from the provincial health insurance plan database (the remaining five provinces), and
B-65
-------
frequency matched to cases based on age and sex. Of the 1,499 eligible cases, 887 (59%)
completed the mailed questionnaire and participated in the study. The response rate among
controls was 62% (n = 2,847) of 4,604 eligible subjects.
The mailed questionnaire sought information on socio-demographics, lifetime smoking
history, dietary habits, and occupational history, including information on specific agents. Up to
12 occupations were categorized into SOC codes, and study investigators identified
9 occupations as suspect and ofa priori interest, including job title of dry cleaner. A total of
9 subjects, 4 cases, and 5 controls, reported "ever" holding an occupation as dry cleaner.
Employment duration was calculated from time period reported for each occupational activity
over a subject's lifetime. Unconditional logistic regression was used to calculate odds ratios and
their corresponding 95% CIs, for males and females separately, adjusted for age, province, race,
current smoking status, ex-smoking, consumption of fruit, fried food, and coffee, and "ever"
employed in suspect occupations. The authors did not present observations on employment
duration and "ever" holding a job as dry cleaner, likely due to the few exposed subjects. The
authors reported strengths of the methodology of their study as use of histologically confirmed
incident bladder cancer cases and the extensive information on nonoccupational factors.
Limitations include the study's small sample size, recall bias among cancer cases, low-response
rate among both cases and controls, and use of occupational title as surrogate for
tetrachloroethylene exposure potential.
B.2.2.1.5. Kogevinas et al. (2003)
Kogevinas, M.; 't Mannetje, A.; Cordier, S.; Ranft, U.; Gonzalez, C.; Vineis, P.,...
Boffetta, P. (2003). Occupation and bladder cancer among men in Western Europe.
Cancer Causes Control, 14, 907-914.
http://dx.doi.Org/10.1023/B:CACO.0000007962.19066.9c
Summary: This study used pooled data from 11 previously conducted European case-
control studies to examine the association between risk of bladder cancer and occupational
exposures in men. The case-control studies were Claude et al. (1988), Cordier et al. (1993),
Gonzalez et al. (1989), Hours et al. (1994), Jensen et al. (1987), Pesch et al. (2000a, b), Porru
et al. (1996), Rebelakos et al. (1985), Serra et al. (2000), and Vineis and Magnani (1985). These
case-control studies were published between 1976 and 1996 and included detailed information
on occupation as well as smoking. Cases and controls needed to fall within the 30-to-79-year
age range. Cases whose interview occurred more than 2 years after diagnosis were also
excluded. Of the 4,101 cases in the pooled dataset, 3,346 (81.6%) met these criteria and were
included in the analysis. Of the 7,365 controls in the pooled dataset, 6,840 (92.9%) were
included in the analysis. Three of the pooled studies used population controls; one used both
hospital and population controls; the remaining seven used hospital controls only. Cases and
B-66
-------
controls were matched on 5-year age group and geographic area. All occupational and industrial
information were coded according to ISCO-68 and International Standard Industrial
Classification of All Economic Activities (ISIC) rev2 standards, respectively. Launderers, dry
cleaners, and pressers fell within ISCO Code 56. A total of 19 (0.6%) cases and 30 (0.4%)
controls were launderers, dry cleaners, or pressers. The researchers did not consider this
occupation to be at high risk a priori, though it was identified in other studies to be at high risk.
Unconditional logistic regression was used to estimate odds ratios and their
corresponding 95% CIs, adjusted for 5-year age group, smoking, and study center. The
interaction between age and study center was found to be significant and was also included in all
of the models. Attributable risk for those occupations identified as high risk a priori was also
calculated, though this did not include launderers, dry cleaners, and pressers. A strength of the
study is its pooled nature, which allowed for a high power and the ability to determine whether
risks are similar in different populations. A limitation is the low-exposure prevalence among
both cases and controls.
B.2.2.1.6. Reulen et al. (2007)
Reulen, R.; Kellen, E.; Buntinx, F.; Zeegers, M. (2007). Bladder cancer and
occupation: A report from the Belgian case-control study on bladder cancer risk.
Am J Ind Med, 50, 449-454. http://dx.doi.org/10.1002/aiim.20469
Summary: This population-based case control study aimed to add to the data on
associations between occupation and bladder cancer, thereby strengthening the case for focused
research on specific occupational categories. Cases were identified through the Limburg Cancer
Registry and consisted of men and women aged 40 to 96 years who were diagnosed with
histologically confirmed transitional cell carcinoma of the bladder between 1996 and 2004.
Controls were randomly selected from the general population of Limburg through simple random
sampling and consisted of Caucasian men and women over the age of 50 years, with no previous
history of bladder cancer. The exclusion of individuals less than 50 years of age was due to the
researchers' finding that the majority of controls were over 50 years. Of the 2,230 eligible cases,
202 (9.1%) participated in the study. The response rate among controls was 26% and included
390 participants.
In-person interviews were conducted by three trained interviewers in the participants'
homes using a structured questionnaire. Information was obtained on socio-demographics,
lifetime smoking history, and lifetime occupational history of all jobs held for at least 6 months.
Lifetime occupational history was assessed as a proxy for exposure, and all occupations were
blindly coded according to the International Standard Classification of Occupations. Domestic
helpers, cleaners, and launderers comprised Code 913 and included a total of 14 (6.9%) cases
B-67
-------
and 20 (5.1%) controls. Unconditional logistic regression was used to calculate odds ratios and
their corresponding 95% CIs, adjusted for age, sex, current smoking status, years of cigarette
smoking, number of cigarettes smoked per day, and education. An interaction of sex and
occupation was also included in the model. Only those occupations with 15 or more participants
were reported. The authors do not report strengths of their study methodology. Limitations
include the study's small sample size, recall bias among cancer cases, and the low-response rate
among both cases and controls.
B.2.2.1.7. Schoenberg et al. (1984)
Schoenberg, J. B.; Stemhagen, A.; Mogielnicki, A. P.; Altman, R.; Abe, T.; Mason,
T. J. (1984). Case-control study of bladder cancer in New Jersey. I. Occupational
exposures in white males. J Natl Cancer Inst, 72, 973-981.
http://www.ncbi.nlm.nih.gov/pubmed/6585596
Summary: This population case-control study of New Jersey male residents, 21-84 years
of age, examined bladder cancer, including papilloma not specified as benign, and occupation.
Newly diagnosed incident bladder cancer cases were identified between 1978-1979 using a
mechanism whereby incident cases were reported within 72 hours of diagnosis or by searching
hospital pathology records (hospital number and local hospitals not identified in published
paper). No overlap occurs between this study and the large National Bladder Cancer Study,
which also included cases diagnosed between 1977 and 1978 (Silverman et al., 1990; Silverman
et al., 1989a: Silverman et al., 1989b: Smith etal., 1985). Age-stratified random samples of
male population controls were identified using random digit dialing, if 21-64 years old or,
records of the Health Care Financing Administration, if 65-84 years. Controls were not
frequency matched by county. To allow for potential county-specific comparison, additional
controls were identified, employed, and stratified by county so that the case-to-control ratio for
each age-county group would be at least 1:1. Of the 787 male cases and 1,608 controls meeting
the case or control definition, 706 cases (90%) and 1,392 controls (87%) were interviewed; all
cases and controls were alive at the time of the interview. Few subjects were non-Caucasian, and
analyses were restricted to Caucasian males, 658 cases, and 1,258 controls. Face-to-face
interviews were carried out using a structured questionnaire that sought information on
demographic, personal, and occupational risk factors. Information on all jobs held >6 months
was ascertained, and subjects were shown lists of industries, employers, and materials to elicit
information not initially recalled. All industry and job title information was coded to the 1970
Census Index System and based upon these codes; 19 employment categories were identified a
priori as known or suspected occupations or exposures; employment as dry-cleaning workers
was one of the 19 categories. Few cases and controls were identified with employment as a dry-
B-68
-------
clean worker: 7 cases (1.1%) and 10 controls (0.8%), which limited the statistical power of this
study to examine bladder cancer and dry-cleaning employment.
Odds ratios and 95% confidence limits were calculated using logistic regression with a
model including either 19 exposure terms or, as in the case of employment as a dry-cleaning
worker, a term for the specific exposure category. Statistical analyses were adjusted for age and
duration of cigarette smoking, placed into four categories. Other covariates, such as previous
bladder or kidney infection, family history of urinary tract cancer, coffee consumption,
education, and use of artificial sweeteners did not change the odds ratio estimate by more than
10% and, therefore, were not included in the final logistic regression model.
B.2.2.1.8. Smith et al. (1985), Silverman et al. (1990; 1989a; 1989b)
Smith, E. M.; Miller, E. R.; Woolson, R. F.; Brown, C. K. (1985). Bladder cancer
risk among laundry workers, dry cleaners, and others in chemically-related
occupations. J Occup Med, 27, 295-297.
http://www.ncbi.nlm.nih.gov/pubmed/3998883
Silverman, D. T.; Levin, L. I.; Hoover, R. N.; Hartge, P. (1989a). Occupational risks
of bladder cancer in the United States: I. White men. J Natl Cancer Inst, 81,1472-
1480. http://dx.doi.org/10.1093/inci/81.19.1472
Silverman, D. T.; Levin, L. L; Hoover, R. N. (1989b). Occupational risks of bladder
cancer in the United States: II. Nonwhite men. J Natl Cancer Inst, 81,1480-1483.
http://dx.doi.org/10.1093/inci/81.19.1480
Silverman, D. T.; Levin, L. L; Hoover, R. N. (1990). Occupational risks of bladder
cancer among white women in the United States. Am J Epidemiol, 132, 453-461.
http://www.ncbi.nlm.nih.gov/pubmed/2389750
Summary: These studies used data from the National Bladder Cancer Study (Hartge et al.,
1984), which was a large case-control study researching the relationship between occupation and
bladder cancer. Cases consisted of men aged 21 to 84 years who were diagnosed with
histologically confirmed urinary bladder cancer between 1977 and 1978 in 9 Surveillance,
Epidemiology, and End Results (SEER) reporting locations (Connecticut, Iowa, New Mexico,
Utah, Atlanta, Detroit, New Orleans, San Francisco, and Seattle) and one rapid reporting system
for bladder cancer, which was mandated by state law (New Jersey). Controls were randomly
selected from within each of the 10 geographical areas and matched to cases based on 5-year age
group and sex. Control selection occurred in two ways: men aged 21 to 64 years were randomly
digit dialed, and men aged 65 years and older were obtained from a stratified random sample of
Health Care Finance Administration lists. The random digit dialing telephone screening yielded
an 88% response rate, and the home interview response rates were 73% for cases and 83% for
controls (Hartge et al., 1984). In-person interviews were conducted by a trained interviewer
B-69
-------
within 3 months of diagnosis, reducing the need for proxy interviews. All interviews used a
structured questionnaire that inquired about artificial sweeteners, smoking, coffee consumption,
medical history, and occupational history for all jobs that lasted at least 6 months from 12 years
of age onwards. Job histories were coded according to the U.S. Bureau of the Census Index of
Industries and Occupations.
Smith et al. (1985) examined bladder cancer risk among individuals employed as laundry
workers and dry cleaners and in other occupations and industries with similar chemical
exposures and compared it with that of workers in occupations or industries that did not expose
them to these chemicals. The authors did not report the final number of cases and controls
included in the study. Participants were classified into one of three exposure categories:
(1) exposed through employment as laundry/dry-cleaning operatives for at least 6 months
(103 participants); (2) exposed through chemicals encountered in other occupations or industries
(5,776 participants); and (3) unexposed (1,869 participants). Duration of exposure among those
in the laundry/dry cleaning occupation was calculated as the total number of years employed in
that profession. Logistic regression was used to separately calculate the relative risks of
occupational exposure, adjusted for age and sex, and duration of exposure by age, sex, and
smoking status. One strength of this study is its large, population-based design with in-person
interviews and matched cases and controls. A limitation is its small exposed population.
Silverman et al. (1989a; 1989b) examined high risk occupations for bladder cancer
among and Caucasian and non-Caucasian men. The final sample consisted of 2,100 Caucasian
male cases, 126 non-Caucasian male cases, 3,874 male controls, and 383 non-Caucasian male
controls. Cases and controls were similar on occupational history variables, with the exception
of age at first employment where cases were younger than controls. Occupations were
subsequently grouped by their potential to have similar exposures, which aggregated 417 census
codes into 163 categories. Workers involved in "processing" within an industry were also
grouped together in one category within that industry. Dry cleaners, ironers, and pressers were
examined as miscellaneous a priori suspect occupations and contained 11 (8.7%) non-Caucasian
cases and 12 (3.1%) non-Caucasian controls. Exposure prevalence for occupation as dry
cleaners, ironers, and pressers is not presented in the published papers for Caucasian cases and
controls. The maximum likelihood method was used to estimate odds ratios for occupations.
The estimate for the dry cleaner, ironer, and presser occupation was adjusted for smoking and
employment in other high risk occupations. The estimates' corresponding 95% CIs were
calculated using Gait's interval estimation procedure. Maximum likelihood was also used to
estimate odds ratios for duration of exposure as a dry cleaner, ironer, or presser (<5 years, >5
years), adjusted for smoking and age, and a Mantel-Haenszel procedure was used to evaluate
one-tailed significance tests of trend. Finally, population attributable risks and their
B-70
-------
corresponding 95% CIs were calculated according to Whittemore (1983) and adjusted for age,
geographic area, and smoking. Due to the large number of analyses performed, only those
occupations for which there were a minimum of 15 exposed cases or controls and who met one
of three additional criteria (statistically significant risk, a priori category, or summary category)
were presented. The authors did not note any strengths of their methodology. A limitation of the
methodology is the potential for misclassification of exposure when grouping occupations for the
purposes of analysis.
B.2.2.1.9. Steineck et al. (1990)
Steineck, G.; Plato, N.; Gerhardsson, M.; Norell, S. E.; Hogstedt, C. (1990).
Increased risk of urothelial cancer in Stockholm during 1985-87 after exposure to
benzene and exhausts. Int J Cancer, 45,1012-1017.
http://dx.doi.org/10.1002/iic.2910450605
Summary: This population case-controls study of males residing in the county of
Stockholm 1985-1987 and born between 1911 and 1945 and population controls examined
occupational exposures and urothelial cancer. The source for identifying cases is not identified
in the published paper. Population controls were identified using random sampling of population
registers at four periods during case ascertainment. A total of 320 cases and 363 controls were
identified of which 256 cases and 287 controls were alive and completed the interview;
participation rates were 80% for cases and 79% for controls. Of the 256 cases, 243 were of the
urinary bladder, 5 of the renal pelvis, 5 of the ureter, and 3 of multiple sites. An additional
two cases had heavy exposure to aromatic amines and were excluded from the case series.
Occupational history was sought from case and control using a questionnaire with an
industrial hygienist blinded to case and control status classifying potential exposure to 38 agents
or groups of substances, including 17 categories of aromatic amines. Two cases and two
controls reported employment as a dry cleaner or in the dry-cleaning industry, with an exposure
prevalence of <1% for either cases or controls. The published paper does not discuss other
information obtained from the questionnaire, except smoking, for which a subject was
categorized as either a current smoker, former smoker, or never smoker. Some residual
confounding is likely given the use of these broad categories rather than pack years. Logistic
regression was used to estimate an odds ratio adjusted for birth year and smoking.
B-71
-------
B.2.2.1.10. Zheng et al. (2002)
Zheng, T.; Cantor, K. P.; Zhang, Y.; Lynch, C. F. (2002). Occupation and bladder
cancer: A population-based, case-control study in Iowa. J Occup Environ Med, 44,
685-691. http://www.ncbi.nlm.nih.gov/pubmed/12134533
Summary: This population case-control study used data from a larger study of drinking
water by-products to examine the relationship between occupation and bladder cancer. Cases
and controls were identified in two phases. In Phase 1, cases consisted of male Iowa residents
without a previous diagnosis of neoplasm, aged 40 to 85 years, and who were diagnosed with
one of six cancer sites (brain, kidney, pancreas, colon, rectum, and urinary bladder) between
1986 and 1987. Controls were randomly selected from the Iowa residents, and frequency was
matched to all cases based on 5-year age group and sex. Control selection occurred in two ways:
(1) men aged 64 aged years or younger were randomly digit dialed, and (2) men aged 65 years
and older were obtained from a stratified random sample of Health Care Finance Administration
lists. The control matching frequency to bladder cancer case series was -2.3:1. Phase 2 of case
and control ascertainment occurred between 1988 and 1989. Cases, aged 40 to 85 years, with in
situ and invasive bladder cancer (transitional cell carcinoma and papillary transitional cell
carcinoma) were identified among Iowa residents between 1998-1989 with controls frequency
matched to cases at a ratio of 1:1. The random digit dialing telephone screening yielded an
85% response rate for cases (1,452): 82% for controls younger than 65 years, and 80% for
controls >65 years. A total of 1,452 case (1,135 men, 317 women) and 2,434 control
(1,601 men, 833 women) participated in the study.
In-person interviews were conducted by a trained interviewer within 3 months of
diagnosis, reducing the need for proxy interviews. All interviews used a structured questionnaire
that inquired about artificial sweeteners, smoking, coffee consumption, medical history and
occupational history for all jobs held for 5 years or longer from 16 years of age onwards.
Proxies completed the questionnaires for 156 cases who had died or were not competent to
participate. All controls except two completed questionnaires in person. Job titles and industries
were reported by Standard Industry Classification and SOC Manual schemes using two-, three-,
and four-digit codes. The SOC code for occupation in laundering and dry cleaning was 7,658.
Zheng et al. (2002) reported three female cases, and one female control held an occupation in dry
cleaning and laundering; however, these authors did not report the number of male cases or
controls.
Odds ratios and 95% CIs were calculated using unconditional logistic regression adjusted
for age, lifetime pack-years of cigarette smoking, and having a first-degree relative with bladder
cancer. Other variables such as education, frequency of strenuous or moderate exercise, duration
of living in a residence served by chlorinated surface water, population size of places of
B-72
-------
residence, and other cancer in a first-degree relative were also examined in the statistical analysis
but did not result in material change to the observed association and were, therefore, not included
in the final logistic regression model. Duration of exposure was examined using a dichotomous
grouping of <10 years and >10 years employment duration.
Certain characteristics of this study strengthen the interpretation and included use of
histologically confirmed cases, use of lifetime job-exposure history, and relatively high-response
rates from both cases and controls. On the other hand, observations for dry cleaning and
laundering occupation are based on small numbers, limiting the study's sensitivity. Additionally,
exposure misclassification may have been introduced because the study did not specifically
identify tetrachloroethylene exposure intensity for individual subjects and is likely of a
nondifferential direction, which would be expected to attenuate the strength of estimated risks.
Additionally, use of employment duration is a crude surrogate for cumulative exposure,
particularly in light of any temporal changes in intensity.
B.2.2.2. Brain Cancer
B.2.2.2.1. Heineman et al. (1994)
Heineman, E. F.; Cocco, P.; Gomez, M. R.; Dosemeci, M.; Stewart, P. A.; Hayes, R.
B.,... Blair, A. (1994). Occupational exposure to chlorinated aliphatic
hydrocarbons and risk of astrocytic brain cancer. Am J Ind Med, 26,155-169.
http://dx.doi.org/10.1002/aiim.4700260203
Summary: This case-control study explored the potential association of brain cancer with
specific solvents, including tetrachloroethylene. Cases consisted of Caucasian men who had died
of brain or other central nervous system (CNS) tumors between 1978 and 1980 in Louisiana and
between 1979 and 1980 in New Jersey and Pennsylvania. Controls were Caucasian men who
had died of other causes, excluding cerebrovascular diseases, epilepsy, suicide, and homicide.
Controls were matched to cases based on age, year of death, and study area. Both cases and
controls were obtained from death certificates. Of the 741 cases and 741 controls selected, next
of kin were found for 654 (88%) cases and 612 (83%) controls. Of these, proxy interviews were
performed for 483 cases (74% of those contacted) and 386 (63% of those contacted) controls.
After excluding cases for which a hospital diagnosis was not reported and controls whose death
may have been associated with their occupation (e.g., lung cancer, liver cancer, leukemia, etc.),
the final sample included 300 cases and 320 controls.
Blinded, trained interviewers conducted interviews with next of kin regarding possible
risk factors for brain cancer as well as all occupations held by the case or control since the age of
15 years. Information collected included job title, tasks, name and location of the company, type
of industry, kinds of products, employment dates, and hours worked. A job exposure matrix
B-73
-------
(Gomez etal., 1994) was used to estimate exposures based on reported occupations and
industries. Occupations and industries were coded according to four digit U.S. SIC and SOC
codes, respectively. All of the four digit codes were assigned exposure estimates of probability
(i.e., low, medium, high) and intensity (i.e, 1, 2, 3) apriori. Intensity was defined as an average
of the concentration and frequency of exposure. Occupations were also assigned a category.
Jobs that fell within Category A, such as dry cleaner operators, had sufficient information to be
assessed for exposure, independent of their industry. For jobs that fell within Category B, the
probability of exposure depended entirely on the industry, and the intensity was weighted by
both the occupation and the industry. Those in Category C had their probability and intensity of
exposure fully determined by the industry within which the job fell. Time of employment was
accounted for in the matrix through a decade indicator. There were 111 (37%) cases and
106 (33.1%) controls "ever" exposed to tetrachloroethylene.
The analysis included maximum likelihood estimates of odds ratios and 95% CIs using
Gart (1970), adjusted for age and study area. Linear trends were examined using Mantel (1963),
and logistic regression was performed to estimate odds ratios and their corresponding 95% CIs,
controlling for age, study area, and employment in electronics-related occupations or industries.
A lag time of 10 or 20 years was included. A strength of the study is its blinded exposure
classification. Limitations include possible misclassification due to inaccurate reports from
proxy respondents, although cases and controls were dead, minimizing potential differential
reporting between cases and controls by proxy respondents, misclassification of exposure due to
the interchangeability of some solvents, and a high proportion of nonrespondents.
B.2.2.3. Breast Cancer
B.2.2.3.1. Peplonska et al. (2007)
Peplonska, B.; Stewart, P.; Szeszenia-Dabrowska, N.; Rusiecki, J.; Garcia-Closas,
M.; Lissowska, J.,... Blair, A. (2007). Occupation and breast cancer risk in Polish
women: A population-based case-control study. Am J Ind Med, 50, 97-111.
http://dx.doi.org/10.1002/ajim.20420
Summary: This study used data from a large case-control study in Poland to evaluate the
risk of breast cancer by occupation and industry. Cases were identified through a rapid case
ascertainment system organized by participating hospitals and were newly diagnosed
histologically confirmed in situ or invasive breast cancers in female residents of Warsaw and
Lodz, between 20-74 years of age, diagnosed 2000-2003. Population controls were identified
from the Polish Electronic System of Population Evidence and matched to cases by city of
residence and age within 5-year age groups.
B-74
-------
A structured questionnaire administered using in-person interviews collected data on
demographic, reproductive, and menstrual history; hormone use history; physical activity;
occupation history; smoking and alcohol use; diet; cancer history in female relatives; medical
and screening history; prenatal exposures; and history of weight and height development. With
respect to occupation history, all jobs held at least 6 months, including job title and possible
exposure to a list of chemicals potentially associated with breast cancer, were obtained, and
industry and occupation codes were assigned according to the SIC Manual and SOC Manual. Of
the 2,275 cases (79% response rate) and 2,424 controls (66% response rate) completing the
questionnaire, 28 cases and 32 controls were identified as working in the laundry, cleaning, and
garment services industry; exposure prevalence was 1% for cases and 1% for controls.
Peplonska et al. (2007) does not report the percentage of subjects with proxy interviews.
Unconditional logistic regression analyses were used to estimate odds ratio and 95% CIs
as the measure of association between occupation or industry and breast cancer risk.
Multivariate models included adjustment for age, age at menarche (<12, 13-14, >15, missing),
menopausal status (premenopausal; postmenopausal), age at menopause among postmenopausal
women (<45, 45-54, >55, missing), number of full-term births (<1, 2, >3), body mass index (25,
25-30, >30), breast cancer in first-degree relative (yes, no), education (less than high school,
high school, some college, professional training, college degree, missing), and city of residence.
The influence of oral contraceptive use, marital status, tobacco and alcohol use, age at first full-
term birth and breastfeeding, and recreational and occupational physical activity was also
evaluated; but, these factors had little impact on risk estimates and were not included in the final
models. Additionally, each specific white-collar job, using all other white-collar jobs as the
reference group, was analyzed to control for socioeconomic factors that could not be completely
captured by adjustment for education level.
Methodological strength of this study includes the size of the studied population and the
scope of information on lifetime occupational history that was collected, together with
comprehensive data on potential confounders, and effect modifiers including pre- and
postmenopausal status. Potential limitations are the small numbers for many occupational
groups, multiple jobs, and multiple comparisons.
B.2.2.4. Colon Cancer
B.2.2.4.1. Fredriksson et al. (1989)
Fredriksson, M.; Bengtsson, N. O.; Hardell, L.; Axelson, O. (1989). Colon cancer,
physical activity, and occupational exposures: A case-control study. Cancer, 63,
1838-1842. http://www.ncbi.nlm.nih.gov/pubmed/2702592
B-75
-------
Summary: This case-control study examined the relationship between occupational
exposures and colon cancer in Sweden. Cases consisted of men and women between 30 and
75 years of age who were diagnosed with large bowel cancer adenocarcinoma between 1980 and
1983. Cases were obtained from the Swedish Cancer Registry and needed to be living at the
time of the study (1984 to 1986), located within the admissions region of the Department of
Oncology in Umea, and medically able to complete a mailed questionnaire. Controls were
selected from the National Population Register, and two controls were matched to each case
based on county of residence, sex, and age. Controls were required to be living at the time of the
study and medically able to complete the questionnaire. Of the 402 cases and 717 controls
identified, 329 cases and 658 controls met inclusion criteria and were contacted with a mailed
questionnaire inquiring about occupational histories, occupational exposures, food and drinking
habits, previous diseases, and drug intake. Overall, 312 (94.8%) cases and 623 (94.6%) controls
participated in the study.
Occupations were assessed as a proxy for exposure by two physicians and one hygienist,
who independently classified exposure as either high or low grade. There were 5 (1.6%) female
cases and 5 (0.8%) female controls who reported employment in dry cleaning. The authors did
not report any dry cleaning information for men. Mantel-Haenszel methods were used to
calculate odds ratios, and Miettinen (1976) was used in the estimation of corresponding 95% CIs.
These analyses were performed for all occupations, including dry cleaning, stratified by age and
physical activity. The authors note that an advantage to limiting their study to living patients
only was the lack of information bias due to proxy responders. Although this is a
methodological strength, there is a potential bias created if occupational exposure is associated
with survival.
B.2.2.5. Liver Carcinoma
B.2.2.5.1. Austin et al. (1987)
Austin, H.; Delzell, E.; Grufferman, S.; Levine, R.; Morrison, A. S.; Stolley, P. D.;
Cole, P. (1987). Case-control study of hepatocellular carcinoma, occupation, and
chemical exposures. J Occup Med, 29, 665-669.
http://www.ncbi.nlm.nih.gov/pubmed/2821204
Summary: This case-control study studied the relationship between hepatocellular
carcinoma and occupational factors and chemical exposures encountered at work or in leisure
activities. Cases consisted of men and women aged 18 to 84 years, diagnosed with
hepatocellular carcinoma at 5 study centers, including the University of Alabama, Duke
University, University of Miami, University of Pennsylvania, and the Harvard School of Public
Health. The majority (93.0%) of cases were histologically confirmed, and the remainder were
B-76
-------
clinically confirmed. Controls consisted of patients admitted to the same hospitals for other
conditions that were diagnosed within 3 years of the interview, excluding bronchitis,
emphysema, primary liver disease, and the following cancers: lung, oral cavity, esophagus,
larynx, bladder, and pancreas. Controls were matched to cases based on gender, age, race, and
study center. The final sample consisted of 86 cases and 161 controls. Authors do not report
response rates.
Each participant's occupational history related to all jobs held 6 months or longer was
ascertained during the interview, and jobs were coded according to SIC and SOC manuals.
There were 0 cases and 4 (2.5%) controls who reported employment in the laundering and
cleaning industry. The authors did not report this industry's corresponding code. Conditional
likelihood methods were used in logistic regression models to estimate odds ratios and their
corresponding 95% CIs. Due to the small numbers of "exposed," the authors did not present the
results for the laundry and cleaning industry. The authors do not report strengths of their study.
A limitation is the small number of exposed, which precluded the analysis of participants
employed in the laundry and dry-cleaning industry.
B.2.2.5.2. Hernberg et al. (1988)
Hernberg, S.; Kauppinen, T.; Riala, R.; Korkala, M. L.; Asikainen, U. (1988).
Increased risk for primary liver cancer among women exposed to solvents. Scand J
Work Environ Health, 14, 356-365. http://www.ncbi.nlm.nih.gov/pubmed/3212412
Summary: This case-control study examined if previously reported findings of an
increased risk of primary liver cancer among women exposed to organic solvents (Hernberg et
al., 1984) were a true effect, due to chance, or reflective of an undetected systematic error. Cases
consisted of men and women diagnosed with primary liver cancer and reported to the Finnish
Cancer Register from 1976 to 1978, and also in 1981. The years 1979 and 1980 were excluded
from this study because they were previously examined (Hernberg et al., 1984). Two control
groups were used in this study: a control series of randomly selected stomach cancer patients
identified from the Finnish Cancer Register in 1977; the other included patients whose hospital
autopsy records noted that they had died of a coronary infarction in 1977. Coronary infarction
controls were matched to cases on sex, age, and hospital of diagnosis. The authors make no
mention of matching between cases and stomach cancer controls. All living patients were
excluded from the analyses, as were those with untraceable relatives. Of the 526 cases by proxy
who met inclusion criteria, 377 (71.7%) returned the questionnaire. After excluding those for
whom a diagnosis could not be confirmed, a total of 344 (65.4%) were included in the analysis.
Of the 654 stomach cancer controls and 558 coronary infarction controls who met the inclusion
criteria, 476 (72.8%) stomach cancer controls and 385 (69.0%) coronary infarction controls
B-77
-------
returned the questionnaire. The final sample consisted of 344 cases and 861 controls (476
stomach cancer and 385 coronary infarction).
A questionnaire mailed to proxy-respondents focused on obtaining information on work
history, including employers, work sites, jobs held, and calendar years of work. Information on
alcohol, tobacco, coffee, tea, medicines, leisure activities, and for women, history of oral
contraceptive use, was also obtained. Two occupational hygienists blindly assessed exposure,
based on the likelihood of the participants' industries, workplaces, and job titles, including
solvents or other agents. Exposures were classified as heavy, moderate, or light; dry-cleaning
exposures were based on 1950 records by the Finnish Institute of Occupational Health that noted
tetrachloroethylene exposure ranged from 34-600 ppm during that time. Any exposures that
could not be determined by the occupational hygienists were followed up with phone calls to the
workplace or the proxy respondent. Two cases (0.6%) were identified with possible chlorinated
hydrocarbon exposures: (1) a case assessed as having light, possible exposure to chlorinated
hydrocarbons in a laundry facility, and (2) another case estimated to have heavy exposure to
chlorinated hydrocarbons during 6 years employed as a dry cleaner. Two coronary infarction
controls (0.5%) were determined to have light exposure to tetrachloroethylene as a result of
employment in the dry-cleaning industry.
Likelihood-based odds ratios and 90% CIs were calculated according to Cornfield (1956)
for the association between primary liver cancer and solvent exposure and for the association
between primary liver cancer and heavy/moderate alcohol use. Both were stratified by sex using
methods by Gart (1970). A latency period of 10 years was included, and, thus, any exposures
that occurred before this time were excluded from the analysis. A strength of the study is its use
of a blinded exposure assessment. Limitations to the study include the potential for selection
bias due to the number of eligible cases and controls whose proxy respondents could not be
found or whose proxy respondents did not return the questionnaire. Moreover, misclassification
bias is likely, given the high percentage of proxy respondents. The authors also noted the need
for information on previous hepatitis B infection in order to control for it as a potential
confounder.
B.2.2.5.3. Houten and Sonnesso (1980)
Houten, L. and Sonnesso, G. (1980). Occupational exposure and cancer of the liver.
Arch Environ Health, 35, 51-53. http://www.ncbi.nlm.nih.gov/pubmed/7362270
Summary: This study used a hospital-based case-control design to study the occupational
associations of patients admitted to Roswell Park Memorial Institute. The 102 cases were men
and women with primary liver cancer between 1956 and 1965. Controls consisted of all other
cancer patients admitted to the Roswell Park Memorial Institute during the same time frame.
B-78
-------
The authors failed to mention how many controls were included in the study. Occupation was
assessed as a proxy for exposure, with a total of two cases (2%) employed in the laundry and
dry-cleaning industry. The analysis consisted of a $ goodness-of-fit test, where the distribution
of the cases was compared to controls by each industry. Limitations to the study include the size
of the sample; few exposed cases, which decreased the study's detection sensitivity; the use of
other cancer patients as controls; self-reported occupational information; and inadequate
reporting of study design and results.
B.2.2.5.4. Stemhagen et al. (1983)
Stemhagen, A.; Slade, J.; Altman, R.; Bill, J. (1983). Occupational risk factors and
liver cancer: A retrospective case-control study of primary liver cancer in New
Jersey. Am J Epidemiol, 117, 443-454.
http://www.ncbi.nlm.nih.gov/pubmed/6837558
Summary. This study used a case-control design to examine occupational associations
with liver cancer. Cases were identified through New Jersey hospital records, the New Jersey
State Cancer Registry, and death certificates and consisted of men and women living in New
Jersey who were diagnosed with histologically confirmed primary liver cancer between 1975 and
1980. The authors do not note any age restrictions in their methodology, though cases were aged
20 years and older. Controls were chosen from among men and women admitted to the same
hospitals as the cases, as well as from death certificates, and matched to cases on age, race, sex,
county of residence, and vital status. Potential controls were excluded from the study if they had
a history of liver cancer, hepatitis, cirrhosis, or other liver disease. Deceased controls whose
cause of death was homicide or suicide were also excluded from the study because of the
sensitivity of approaching next of kin. Of the 335 eligible cases, 296 were able to be contacted,
and of these, 265 (79.1%) were interviewed. Of the 825 eligible controls, 687 were able to be
contacted, and of these, 530 (64.2%) were interviewed. Demographics between cases and
controls were similar.
In-person interviews were conducted with all participants or their next of kin to obtain
information on lifetime residence, smoking habits, alcohol, medical history, and employment
since the age of 12 years. There were 254 (95.8%) proxy case interviews and 508 (95.8%) proxy
control interviews. Occupations held for at least 6 months were assessed as a proxy for
exposure. All industries and occupations were coded according to the Index of Industries and
Occupations standards developed by the Bureau of Census. The laundering, cleaning, and other
garment services industry included 10 male cases (3.8%) and 8 male controls (1.5%). The
authors further examined the laundry/dry-cleaning industry by occupations, though the results
B-79
-------
are not presented. No information was reported on female employment in the laundry, dry
cleaning, or garment service industry.
Mantel-Haenszel methods were used to estimate odds ratios and 95% CIs for males
employed at least 6 months in selected industries and occupations. The distribution of subjects
by calculated level of alcohol consumption were adjusted for age (women) and age and smoking
(men), but risk estimates for occupations and industries were not adjusted for potential
confounders. There were no differences in alcohol consumption between cases and controls.
The authors did not report any strengths in their study. Limitations include possible
misclassification of exposure due to proxy respondents, lack of adjustment for confounders such
as smoking and alcohol consumption, possible misclassification due to inaccurate information on
death certificates, and lack of assessment of intensity or duration of exposure.
B.2.2.5.5. Suarez et al. (1989)
Suarez, L.; Weiss, N. S.; Martin, J. (1989). Primary liver cancer death and
occupation in Texas. Am J Ind Med, 15,167-175.
http://www.ncbi.nlm.nih.gov/pubmed/2729281
Summary: This case-control study examined the risk of liver cancer among occupations
in the petroleum and chemical industry and other potentially high-risk occupations. Death
certificates, which listed usual occupation and business or industry, were obtained from the
Texas Bureau of Vital Statistics and used to identify cases and controls for the time period from
1969 to 1980. Cases consisted of men aged 20 years or older who were living in Texas and
whose underlying cause of death was liver cancer. Of the 1,771 potential cases, 1,742 were
eligible and included in the study. The same number of controls were randomly selected from
among the 537,000 death certificates, which represented all other causes of death, excluding
neoplasms, liver and gallbladder diseases, infectious hepatitis, and alcoholism. Controls were
matched to cases based on 5-year age group, race, ethnicity, and year of death.
Occupation was assessed as a proxy for exposure grouped according to the U.S. Census
Classified Index on industrial categories. Groupings were partially based on Hoar et al. (1980),
who categorized industries by product or exposure. In addition to the petrochemical industry,
22 other industries or product categories with at least 10 individuals were examined, including
dry-cleaning services. Occupations within these categories that had at least 10 individuals were
also analyzed and included dry-cleaning operators. There were a total of 11 cases and
12 controls employed in the dry-cleaning industry and 4 cases and 8 controls employed as dry-
cleaning operators. The published paper does not provide information regarding the total
number of controls included in the final sample (although they state that the number of controls
and cases are the same), precluding a calculation of exposure prevalence for this study.
B-80
-------
The Mantel-Haenszel method was used to calculate odds ratios, adjusted for race and
ethnicity. Corresponding 95% CIs were estimated using Miettinen's method. Limitations to the
study include the lack of control for potential confounders that were not included in the death
certificate information, such as alcohol consumption or hepatitis B infection, as well as the lack
of information on exposure and the possible misclassification of exposure based on occupation
and industry information provided on the death certificates. The authors do not report any
strengths of their study's methodology.
B.2.2.6. Lung and Upper Respiratory Tract Cancers
B.2.2.6.1. Brownson et al. (1993)
Brownson, R. C.; Alavanja, M. C.; Chang, J. C. (1993). Occupational risk factors
for lung cancer among nonsmoking women: A case-control study in Missouri
(United States). Cancer Causes Control, 4, 449-454.
http://www.ncbi.nlm.nih.gov/pubmed/8218877
Summary: This study used a population-based case-control design to evaluate the risk of
lung cancer in nonsmokers in relation to their specific occupations. Cases were Caucasian
females living in Missouri between 30 and 84 years and diagnosed with primary lung cancer
between 1986 and 1991. Cases needed to be either lifetime nonsmokers, ex-smokers who had
quit for at least 15 years prior to diagnosis, or ex-smokers that had smoked less than one pack per
year. The cases were selected from the Missouri Cancer Registry; hospitals participating in the
study were also visited to ensure all cases were documented. Of the 429 cases included in the
study, 333 (77%) were histologically confirmed. The 1,021 controls were chosen in two ways:
(1) through a sample of state driver's licenses of women under the age of 65 years, provided by
the Missouri Department of Revenue; and (2) through a roster of Medicare beneficiaries of
women aged 65 to 84 years, provided by the Health Care Finance Administration. Controls were
matched to the cases by age group at a 2.2:1 ratio. Of the 650 eligible cases, 618 (95%)
participated in the telephone interview, and 429 (69%) of these 618 also participated in the
second, in-person interview. Of the 429 cases included in the final analysis, 179 (42%) consisted
of interviews with the cases, and 250 (58%) involved interviews with the spouse or another
relative. Of the 1,527 eligible controls, 1,402 (92%) participated in the telephone interview, and
1,021 (73%) of these 1,402 also participated in the second, in-person interview. Overall, 30 of
the cases and 39 of the controls were employed in the dry-cleaning industry.
Both the telephone and in-person interviews were performed by trained interviewers.
The telephone interview inquired about residential history, passive smoke exposure, personal and
family health histories, and reproductive health history. The in-person interview consisted of
questions related to diet and occupation. Occupational risk factors were determined by
B-81
-------
28 questions, which were based on a review of the literature and focused on job title as well as
exposure. Subjects reported the years in which they worked at each job or with each exposure.
Analysis consisted of the calculation of odds ratios and 95% CIs with multiple logistic
regression, adjusted for age, active smoking (for ex-smokers), and history of previous lung
disease. In their examination of risk based on duration of employment, the researchers
ascertained their cutoff points by achieving an approximate equal distribution of controls in
"low" and "high" exposure categories.
Strengths of the study include its large sample size and the fact that it pathologically
reviewed all cases. On the other hand, the study's retrospective nature has limitations. For
example, there was a substantial difference in the proportion of proxy respondents between the
cases and controls. Here, 58% of the case interviews were conducted with surrogates, compared
with none of the controls. The authors noted the inclusion of proxy respondents would introduce
recall bias that would likely bias risk estimates towards the null. Additionally, the researchers
lacked information on the intensity and specific types of occupational exposures these women
experienced. Limitations include the study's low statistical power, small sample of
histologically confirmed cases, difficulty in assessing passive smoking retrospectively, and that
not all cases were eligible to be controls. In this study, 91% of cases under the age of 65 years
and 100% of controls had a current driver's license, suggesting that the case population may
have differed in some characteristics from the control population.
B.2.2.6.2. Consonni et al. (2010)
Consonni, D.; De Matteis, S.; Lubin, J. H.; Wacholder, S.; Tucker, M.; Pesatori, A.
C.,... Landi, M. T. (2010). Lung cancer and occupation in a population-based case-
control study. Am J Epidemiol, 171, 323-333. http://dx.doi.org/10.1093/aie/kwp391
Summary: This large population case-control study, part of the Environment And
Genetics in Lung cancer Etiology (EAGLE), was designed to explore various etiologic factors
for lung cancer risk factors using an integrative approach that combined epidemiologic, clinical,
and molecular data in a clearly defined population setting. Cases and controls were identified
from the Lombardy area in Italy and were from 5 cities and 216 municipalities. The study
included 1,943 incident lung cancer cases, 35-79 years of age, from 2002-2005, identified from
13 hospitals and 2,116 population controls through population databases (not identified in paper)
and frequency matched to case by residence, sex, and age. Cases could have any stage of
primary cancer of the trachea, bronchus, and lung as well as morphology that was verified with
tissue pathology (67%), cytology (28%), or review of clinical records (5%). Response rates were
92.5% for cases and 99.8% for controls. Controls had higher education and held more jobs
compared to cases. All subjects underwent a computer-assisted personal interview and blood
B-82
-------
sampling (or buccal rinse collection for a small percentage of study subjects), and they
completed self-administered questions available on the EAGLE Web site. Lung tissue sample
from cases were collected when available. The interview included lifetime history of jobs held
for >6 months. Industries and job titles were coded blindly by two of the study investigators,
following the International Standard Industrial Classification of All Economic Activities and the
ISCO. Codes were then translated into occupations as known (List A) or suspected (List B) lung
carcinogens. Two (0.2% exposure prevalence) male cases and 3 male controls and 12 (3%
exposure prevalence) female cases and 11 controls were identified as launderers, dry cleaners, or
pressers (ISCO Code 560). Odds ratios and 95% CIs were calculated using unconditional
logistic regression, separately by gender, with covariates for area, age, smoking pack-years, and
number of jobs held. Selected analyses were repeated, adding educational level as a surrogate of
socioeconomic status.
Major strengths of the study are the enrollment of incident cases, the large sample size,
high participation rates, and face-to-face interviews using a structured questionnaire. Jobs are
self-reported, potentially introducing bias; however, the authors noted reliability of self-reported
job history is usually considered good. The blind coding of job title could introduce
misclassification, a source of nondifferential bias. The low exposure prevalence of dry cleaner,
laundry worker, or presser, job titles, particularly among males, in this study reduces its
sensitivity.
B.2.2.6.3. Pohlabeln et al. (2000)
Pohlabeln, H.; Boffetta, P.; Ahrens, W.; Merletti, F.; Agudo, A.; Benhamou, E.,...
Jockel, K. H. (2000). Occupational risks for lung cancer among nonsmokers.
Epidemiology, 11, 532-538. http://www.ncbi.nlm.nih.gov/pubmed/10955405
Summary: This study used a case-control design to investigate the relationship between
occupational exposures and lung cancer in nonsmokers in Europe. A total of 12 study centers in
7 countries (France, Germany, Italy, Portugal, Spain, Sweden, UK) participated. Cases and
controls up to the age of 75 years were enrolled in the study between 1988 and 1994. Controls
were chosen from the community in six centers, from hospitals in five centers, and from both the
community and a hospital in one center. All hospital-based controls had diseases not related to
smoking. A nonsmoker was defined as an individual who has smoked less than 400 cigarettes
during his/her lifetime. The final sample consisted of 650 nonsmoking cases and
1,542 nonsmoking controls. Cases and controls were similar based on sex, age, and most
common histological subtype. With the exception of two centers in Germany and one center in
Portugal, whose response rates were below 50%, the response rates ranged between 55% and
95%.
B-83
-------
Demographics, diet, smoking exposure, smoking history, and occupational history were
collected for each participant through an in-person interview. Industry and occupation were
blindly assessed as a proxy for exposure and coded according to ISCO and ISIC standards. All
jobs that lasted at least 6 months were assessed according to Ahrens and Merletti (1998), who
categorized occupations based on either known (List A) or suspected (List B) associations with
lung cancer. Launderers and dry cleaners were classified into List B and included 20 (3.1%)
cases and 29 (1.9%) controls. Participants were then divided into one of three exposures: "ever"
List A, "ever" List B/"never" List A, and "never" List A or B (unexposed). Unconditional
logistic regression was used to calculate odds ratios and their corresponding 95% CIs of "ever"
working in a List A or List B occupation, adjusted for age and center and stratified by gender.
The inclusion of occasional smoking, residence, diet, and exposure to tobacco smoking as
confounders did not significantly affect the estimates and was not included in the final model.
No differences were found when considering the control groups separately, and the pooled
results were provided. A strength of this study is its size, and limitations include possible
misclassification of smoking status among very light smokers, different response rates among the
centers in Europe, and lack of assessment of duration or intensity of exposure.
B.2.2.6.4. Richiardi et al. (2004)
Richiardi, L.; Boffetta, P.; Simonato, L.; Forastiere, F.; Zambon, P.; Fortes, C.,...
Merletti, F. (2004). Occupational risk factors for lung cancer in men and women: A
population-based case-control study in Italy. Cancer Causes Control, 15, 285-294.
http://dx.doi.Org/10.1023/B:CACO.0000024223.91059.ed
Summary: This population case-control study was conducted in two regions in North
Italy, included subjects in a large international case-control study of nonsmoker lung cancer, and
was coordinated under the International Agency for Research on Cancer (Pohlabeln et al., 2000).
Richiardi et al. (2004) reported observations from the two Italian centers on lung cancers, adding
the smoking cases and occupational factors. Cases (n= 1,171) were incident primary
histologically or cytologically confirmed lung cancers among residents 75 years of age or
younger and identified from all hospitals in the study area. Controls (n = 1,569) were randomly
selected from the local population registries and were frequency matched (>1:1 ratio) with cases
by 5-year age groups and sex. The case series included a higher proportion of ever smokers,
heavy smokers, and lower education compared to the control series. The enrollment period was
1990-1991 (Eastern Venice) and 1991-1992 (Turin). Response rates for Turin and Venice
regions, respectively, were 86, 72, 85, and 74% among cases and controls, respectively.
In-person interviews with a standardized questionnaire gathered information on
demographic details, active and passive smoking, and lifetime occupational history for all jobs
lasting at least 6 months. No information is presented by the authors regarding the number of
B-84
-------
proxy interviews; however, the paper appears to suggest interviews were carried out directly with
subjects. Job title and industry were coded blindly to case-control status using the International
Standard Classification of Occupations and the International Standard Industrial Classification.
The occupational history of each subject was evaluated for employment in occupations and
industries a priori known (List A) or suspected (List B) to entail exposure to lung carcinogens;
List B included dry cleaner and laundry occupations. Three male cases (0.3% exposure
prevalence) and 9 female cases (5% exposure prevalence) were identified as holding dry cleaner
or laundry occupation.
Odds ratios (ORs) and 95% CIs were estimated using unconditional logistic regression
with analyses conducted separately for males and females for lung cancer histological types.
Covariates included in the models were age, study area, education, cigarette smoking,
consumption of other tobacco products, and total number of jobs.
Strengths of this study include the high-response rate and statistical control in analyses
examining occupational title for smoking with any residual confounding related to smoking
likely of a small magnitude. There was evidence of selection bias related to a higher
socioeconomic status among nonparticipant cases and lower among nonparticipant controls
compared to participant cases and controls, although the potential bias may be minimal, because
education was a covariate in statistical analyses and did not substantially change risk estimates.
The exposure-assessment approach based on job and industry titles is limited as a proxy for
cumulative exposure with potential for misclassification bias, usually, nondifferential and of a
downward direction.
B.2.2.6.5. Vaughan et al. (1997)
Vaughan, T. L.; Stewart, P. A.; Davis, S.; Thomas, D. B. (1997). Work in dry
cleaning and the incidence of cancer of the oral cavity, larynx, and oesophagus.
Occup Environ Med, 54, 692-695. http://www.ncbi.nlm.nih.gov/pubmed/9423585
Summary: This study used data collected from two population-based case-control studies
to examine whether employment in the dry-cleaning industry and its associated exposure to
tetrachloroethylene increased the risk of upper aerodigestive tract cancers. The authors do not
provide any references for the studies. Cases were identified through the Fred Hutchinson
Cancer Research Center, a population-based cancer registry encompassing 13 counties in
Washington state, and consisted of male and female residents within the three largest counties.
Cases were between 20 and 74 years of age and diagnosed with cancer of the oral cavity or
pharynx, larynx, esophagus, or gastric cardia between 1983 and 1987 or with adenocarcinoma of
the esophagus or gastric cardia between 1987 and 1990. The authors do not state if the cancer
cases were histologically confirmed. Response rates were 85.2% for the oral cavity, 80.8% for
B-85
-------
the larynx, and 82.9% for the esophagus and gastric cardia. Cases of nonepithelial and
nonspecified cancers were excluded, as were cases without telephones on the date of their
diagnosis. Controls were selected through random digit dialing and matched to cases based on
5-year age group and sex. Of those contacted, 95.4% were screened, and 80.3% of those eligible
were interviewed. The final sample included 1,130 cases (491 oral cavity, 235 larynx,
109 esophagus squamous cell, and 295 esophagus adenocarcinoma), of which, 10 had
two cancers, and 724 controls.
In-person interviews were conducted to gather detailed information on all occupations
that lasted at least 6 months, including employer, type of business, job title, typical activities
performed, and dates of employment. History of occupational exposure to solvents was also
obtained. Information on demographics, and tobacco and alcohol consumption was also
obtained. Proxy interviews with next of kin were conducted in 7.2% of the laryngeal cases,
18.7% of the oral and pharyngeal cases, and 33.2% of the esophageal and gastric cardia cases.
Exposure to tetrachloroethylene was assessed blindly by estimating the probability that the
solvent was used on the job and the 8-hour time weighted average exposure to
tetrachloroethylene on the job. The latter was based on findings in the literature but was not
validated within this population. Overall, 16 (1.4%) cases (7 oral cavity, 5 larynx, 2 esophagus
squamous cell, and 2 esophagus adenocarcinoma) and 8 (1.1%) controls reported "ever"
employment in the dry-cleaning industry. Exposure to tetrachloroethylene was determined to be
possible among 15 (1.3%) cases (7 oral cavity, 4 larynx, 2 esophagus squamous cell, and
2 esophagus adenocarcinoma) and 8 (1.1%) controls. Probable exposure to tetrachloroethylene
was determined for 8 cases (0.7%) and 3 controls (0.4%). Finally, duration of employment
(1-9 years and >10 years) in the dry-cleaning industry and cumulative exposure to
tetrachloroethylene (1-29 ppm/year and >30 ppm/year) were assessed, with the latter being the
product of the duration and the 8-hour time weighted average.
Conditional logistic regression was used to estimate odds ratios and their corresponding
95% CIs for those employed in the dry-cleaning industry and those exposed to
tetrachloroethylene. All estimates were adjusted for age, sex, education, study period, alcohol
consumption, and cigarette smoking. Including race among the potential confounders in the
analysis did not change the estimates and was not included in the final model. A strength of the
study is its detailed occupational history. Limitations included the low prevalence of exposed
cases and controls, the high proportion of proxy respondents, and the lack of information on
solvents used.
B-86
-------
B.2.2.7. Lymphopoietic cancers
B.2.2.7.1. Blair et al. (1993)
Blair, A.; Linos, A.; Stewart, P. A.; Burmeister, L. F.; Gibson, R.; Everett, G.,...
Cantor, K. P. (1993). Evaluation of risks for non-Hodgkin's lymphoma by
occupation and industry exposures from a case-control study. Am J Ind Med, 23,
301-312. http://dx.doi.org/10.1002/aiim.4700230207
Summary: This population-based case-control study examines occupational exposures—
particularly agricultural exposure—as risk factors for non-Hodgkin lymphoma. In Iowa, cases
were identified through the Iowa State Health Registry and consisted of Caucasian men who
were diagnosed with non-Hodgkin lymphoma between 1981 and 1983. In Minnesota, cases
consisted of Caucasian men diagnosed between 1980 and 1982 who were identified from a
surveillance of participating network hospitals that covered approximately 97% of the state. The
cities of St. Paul, Duluth, Minneapolis, and Rochester were excluded from the study. All
identified cases underwent pathology review. Controls included Caucasian men without
hematopoietic or lymphatic malignancies who were frequency matched by state, age, and year of
death for deceased cases. Controls for living cases who were under the age of 65 years at
diagnosis were obtained through random-digit dialing, and those for living cases who were
65 years or older at diagnosis were selected from computerized Medicare files from the Health
Care Finance Administration. Controls for deceased cases were chosen from state vital records
(death certificates). Of the 715 eligible cases, 622 (87.0%) participated in the interview. A total
of 1,245 controls (77% of random digit dialing, 79% of Medicare, and 77% of death certificate)
participated in the interview, though the authors do not provide the eligible population. Farmers
were excluded from the analysis, leaving a total of 546 cases and 1,087 controls.
In-person interviews were conducted by trained interviewers with a structured
questionnaire that inquired about sociodemographic characteristics; agricultural exposures;
exposures to chemicals through hobbies; residential, medical, and occupational histories; as well
as family history of cancer. Occupational histories were ascertained for all jobs held at least
1 year since the age of 18 years, as well as industry, name of employer, products produced, job
title, and duties. There were 184 (29.6%) proxy case interviews and 425 (34.1%) proxy control
interviews. Industries and occupations were coded according to SIC and the Dictionary of
Occupational Titles (DOT), respectively. Exposure was assessed blindly by an industrial
hygienist who used a job-exposure matrix to evaluate probability (4-point scale) and intensity
(3-point scale) of exposure. Laundry and garment workers comprised Code 721 and included
16 (2.9%) cases and 14 (1.3%) controls.
Polychomotous unconditional logistic regression was used to estimate odds ratios and
their corresponding 95% CIs, adjusted for age, state, direct or surrogate respondent, agricultural
B-87
-------
use of pesticides, postsecondary education, use of hair dyes, first-degree family member with
malignant lymphoproliferative diseases, and tobacco. Analyses were also conducted for the
three main types of non-Hodgkin lymphoma (follicular, diffuse, and other) for selected
exposures, occupations, and industries. Exposure-response relationships examined the risk of
non-Hodgkin lymphoma, or of the subtypes of non-Hodgkin lymphoma, by duration of
employment, intensity of employment, and by the probability of exposure. In this instance,
unexposed cases and controls consisted of those not employed in that particular occupation or
industry, or those who lacked the exposure of interest. A strength of the study is its use of a job-
exposure matrix. Limitations include its low-exposure prevalence, possible misclassification due
to limited exposure information, and high percentage of proxy respondents.
B.2.2.7.2. Clavel et al. (1998)
Clavel, J.; Mandereau, L.; Conso, F.; Limasset, J. C.; Pourmir, L; Flandrin, G.;
Hemon, D. (1998). Occupational exposure to solvents and hairy cell leukaemia.
Occup Environ Med, 55, 59-64. http://dx.doi.0rg/10.1136/oem.55.l.59
Summary: This study used a retrospective, hospital-based case-control design to examine
the relationship between occupational exposures and hairy cell leukemia in men in France.
Cases and controls were obtained from 18 hospitals throughout the country; cases included all
patients diagnosed between 1980 and 1990 who were still alive at the time of the study. Controls
consisted of patients admitted to the hospitals during this same time frame for other reasons.
Due to the researchers' need to find a restricted number of cases in the same age range in each
city, controls were predominantly chosen from the orthopedic and rheumatological departments.
Control exclusion criteria included patients admitted for malignant disease, diseases related to
occupations, and work-related accidents. Cases and controls were matched on birth date, sex,
admission date, and residence. Of the 378 cases identified, 278 were considered eligible (i.e.,
still alive at the time of the study). Of these, 226 (81.3%) participated by returning the
questionnaire. Of the 809 eligible controls, 465 (57.5%) participated by returning the
questionnaire. Of these, 40 were excluded because the case they were initially matched with
either died or did not respond, and they could not be matched to other cases. As a result,
425 (52.5%) of the eligible controls were included in the analysis. Efforts were made to match
2 controls with each case; 30% of cases were matched with 1 control, 56% were matched with
2 controls, and 14% were matched with 3 to 5 controls. The final sample consisted of 226 cases
and 425 controls.
Self-administered questionnaires were sent to all participants, inquiring about
sociodemographic characteristics, tobacco smoking, lifelong occupations, and leisure activities.
Additional questionnaires were sent to participants with suspected occupational exposures.
B-88
-------
Semi-structured interviews were also conducted to help experts assess exposures for those
involved in textile degreasing, among others.
Jobs were coded according to International Labor Organization (ILO) and ISIC standards.
Launderers and dry cleaners comprised ILO Code 5.6. There were 1 (0.4%) case and 2 (0.5%)
controls who reported employment as launderers or dry cleaners. Exposure was evaluated in
two ways. The first consisted of a blinded assessment by two of the study's researchers, which
was based on the consistency of the participant's statements, the type of industry in which they
worked, their job title, and the type of exposure. From this information, the researchers were
able to classify the type of solvent used as well as the intensity of exposure associated with each
job. The second method used a job-exposure matrix that was initially developed for a study by
the International Agency for Research on Cancer (Ferrario et al., 1988) to assess exposure to
solvents. The matrix used ILO and ISIC codes to classify each job into one of seven categories
based on the probability, intensity, and frequency of exposure. These categories included
unexposed, possibly exposed/unevaluable, probably exposed (2/3 exposed subjects), certainly exposed, and certainly highly exposed.
Conditional logistic regression was used to estimate odds ratios and their corresponding
95% CIs, adjusted for smoking and farming. These estimates were ascertained for job titles,
including launderers and dry cleaners, occupational tasks, including degreasing, and the main
chemical families of organic solvents. A strength of this study is its use of a job-exposure
matrix. Limitations include the study's retrospective recruitment, low response rate, lack of
verification of self-reported information, and lack of individual solvent assessment. Also, while
the study's inclusion of only living cases may have caused confounding by duration of survival,
exposure information was obtained directly by the case, precluding the use of proxy respondents,
who often do not provide as accurate information as that obtained directly from subjects. This
study was inadequately powered to evaluate dry-cleaning exposures, resulting from the
low-exposure prevalence among cases and one reported case as a launderer or dry cleaner.
Interviews were conducted by trained interviewers either in-person or by telephone. Of
the 430 cases interviewed, 76% were in-person with the cases themselves, 9% over the telephone
with the cases themselves, and 16% were proxy-interviews with next of kin when the case was
deceased or too ill to participate. Of the 1,683 controls interviewed, 81% were in-person with
the cases themselves, 18% were over the telephone with the cases themselves, and less than 1%
were proxy interviews with next of kin. The questionnaire inquired about multiple risk factors,
including chemical exposures, which were assessed blindly by the researchers and a toxicologist
into 20 categories. For statistical reasons, only those exposures with a minimum of 10 exposed
cases were analyzed, and this consisted of aliphatic hydrocarbons, aromatic hydrocarbons,
chlorinated hydrocarbons, and pesticides. Chlorinated hydrocarbons included dry-cleaning
B-89
-------
solvents, though only one case reported exposure to these particular solvents. The questionnaire
also inquired about employment in four industries: petroleum, rubber, dry cleaning, and meat
processing. Overall, 14 (3.3%) cases and 59 (3.5%) controls reported working for at least
6 months in the dry-cleaning industry.
Unconditional logistic regression was used to estimate odds ratios and their
corresponding 95% CIs. Both unadjusted and adjusted odds ratios were calculated for
all-respondents and self-respondents only (excluding proxy-respondents); adjusted odds ratios
controlled for race, 10-year age group, education, sex, and study site. The authors do not report
strengths of their methodology. Limitations include the study's lack of adjustment for smoking
and reliance on self-reported or proxy-reported occupational, which may have introduced recall
bias. Misclassification of exposure towards the null was possible, given the discrepancy between
those who reported working in the dry-cleaning industry and those who reported exposure to
dry-cleaning solvents. Misclassification may have also occurred when participants were required
to judge what, if anything, needed to be noted in the "other chemicals" portion of the chemical
exposures question. Additionally, frequency, intensity, and duration of exposure were missing,
which is mostly due to the fact that the questionnaire was developed to assess different risk
factors. The study group was enrolled for the purposes of measuring the effect of gene
influences on the immune system, rather than exposure to common chemicals. The study also
suffered from a small sample size, impacting the statistical power to examine dry-cleaning
exposures.
B.2.2.7.3. Fabbro-Peray et al. (2001)
Fabbro-Peray, P.; Daures, J. P.; Rossi, J. F. (2001). Environmental risk factors for
non-Hodgkin's lymphoma: A population-based case-control study in Languedoc-
Roussillon, France. Cancer Causes Control, 12, 201-212.
http://dx.doi.Org/10.1023/A:1011274922701
Summary: This population-based case-control study of NHL evaluated medical,
occupational, and environmental risk factors and the occurrence of malignant lymphomas. This
study was limited to French men and women aged 18 years or older who were living in
Languedoc-Roussillon, which is the French county with the highest incidence of non-Hodgkin
lymphoma. Cases were diagnosed with malignant lymphomas between 1992 and 1995 from
19 hospitals and a cancer research center. Controls were randomly chosen from electoral lists in
a two-phase approach. First, the municipalities were randomly selected based on their size and
the distribution of the population in the county. Second, individuals within each of the chosen
municipalities were randomly selected. There were two controls assigned to each case, though
the nonelectronic nature of the data prevented matching of cases and controls. Of the
B-90
-------
627 eligible cases and 1,962 eligible controls, a total of 517 (82.5%) cases and 1,025 (52.2%)
controls participated in an interview between 1992 and 1996. Of the 517 cases, 445 cases
(86.0%) presented with NHL and 72 cases (13.9%) with Hodgkin lymphoma. Overall, there
were more male cases (56.9%) than male controls (44.8%), and cases were older than controls.
Unblinded interviews were conducted by trained interviewers with cases and controls
either in-person or over the phone. The questionnaire inquired about general characteristics,
medical history, occupational history, environmental and occupational exposure to chemicals,
occupational exposure to electromagnetic radiation, and smoking. Age at first exposure,
duration of exposure, total number of days exposed, and time since first exposure were assessed
for each chemical, including dry-cleaning solvents. There were a total of 35 (6.8%) cases and
77 (7.5%) controls exposed to dry-cleaning solvents.
Mantel-Haenszel methods were used for estimating the odds ratios and 95% CIs
examining the effect of sociodemographic characteristics. Unconditional logistic regression
using a forward stepwise approach was used to estimate odds ratios and 95% CIs for the effect of
chemical exposures, occupational exposures to electromagnetic radiation, and cigarette smoking
individually on non-Hodgkin lymphoma, adjusted for age, gender, urban setting, and education
level. A lag time of 5 years prior to cancer diagnosis was included. Limitations to the study
include the recall bias, given all information was self-reported, a high rate of refusal to
participate in the control group, leading to potential selection bias, nondifferential
misclassification, and the use of a broad category of dry-cleaning solvents that included
tetrachloroethylene and other solvents. No strengths were reported by the authors.
B.2.2.7.4 Gold et al. (2010a; 2010b)
Gold, L. S.; Milliken, K.; Stewart, P.; Purdue, M.; Severson, R.; Seixas, N.,... De
Roos, A. J. (2010a). Occupation and multiple myeloma: An occupation and industry
analysis. American Journal of Industrial Medicine, 53(8), 768-779.
http://dx.doi.org/10.1002/aiim.20857
Gold, L. S.; Stewart, P. A.; Milliken, K.; Purdue, M.; Severson, R.; Seixas, N.,...
De Roos, A. J. (2010b). The relationship between multiple myeloma and
occupational exposure to six chlorinated solvents. Occupational and Environmental
Medicine, 68(6), 391-399. http://dx.doi.org/10.1136/oem.2009.054809
Summary: This population-based case-control study examined occupation exposures,
particularly solvents exposure, as risk factors for multiple myeloma, and was carried out in
two SEER sites, Seattle, WA, and Detroit, MI. Incident multiple myeloma cases (ICD-O-2/3,
9731 [plasmacytoma not otherwise specified] and 9732 [multiple myeloma]) eligible to
participate were 35-74 years old and were newly diagnosed between 2000 and 2002. Gold et al.
(2010a) reported on occupational and industry, with Gold et al. (201 Ob) reporting findings on
B-91
-------
6 chlorinated solvents: tetrachloroethylene, trichloroethylene, 1,1,1-trichloroacetic acid,
methylene chloride, chloroform, and carbon tetrachloride. Of the 365 cases eligible to
participate, 64 (18%) had died before they could be contacted, 28 (8%) were unable to be
located, and 18 (5%) were patients of physicans who refused to participate (71% participation
rate). Population controls were selected from a previous case-control study of NHL undertaken
at the same time in the same two SEER reporting sites (Charter]ee et al., 2004) and who (1) had
not been previously diagnosed with multiple myeloma, plasmacytoma, NHL, or HIV, (2) were
between 35-74 years of age, (3) were identified as residents of the Detroit or Seattle-Puget sound
areas between 1998-2002, and (4) spoke English. Controls under 65 years of age were
identified using random digit dialing; controls (65-74 years of age) were identified from
Medicare roles. Of the eligible 1,133 controls, 481 (52%) participated. Control participation
was not associated with study site or generation, but individuals in the 35-50- and 65-74-age
groups were less likely to have participated than subjects 51-64 years old.
In-person interviews were conducted using a computer-assisted personal interview
program. All interviews were carried out with the case; proxy interviewees could not complete
the interview but could aid in recalling details of occupational exposures. Information on all jobs
held since the age of 18 years for at least 1 year between 1941, for cases, and 1946, for controls,
and the study enrollment dates, was collected. Subjects were additionally administered
job-specific questionnaires for 20 occupations with potential solvent exposure. These modules
were administered only when participants held the relevant job for at least 2 years. All jobs were
coded blinded to case or control status according to the SOC system (Gold et al., 2010a) or
assessed for exposure to six chlorinated solvents using job-exposure matrices developed for each
decade for specific industries such as the chemical or rubber industries, occupations such as auto
mechanics or hair dressers, and tasks such as degreasing, gluing, and painting, through literature
reviews for trichloroethylene and tetrachloroethylene (Gold et al., 2008; Bakke et al., 2007).
Each job was assigned a score for probability (0-4) based on the percentage of subjects likely to
have had exposure, and for jobs with probability scores of 1 or higher, frequency (1~4), and
intensity (1-4) scores. All jobs were assigned a score for confidence levels (1-4). Probability
was scores as 0 = <1%; 1 = 1 through <10%; 2=10 through <50%; 3 = 50 through <90%;
4 + > 90%. Frequency was defined as the average hours per week of exposure: 0 = <15
minutes/week; 1 = 15 minutes through <1 hour/week; 2 = 1-10 hours/week; 3 = > 10-20
hours/week; 4 = >20 hours/week. The intensity score was the contraction of solvent estimated to
have been in the subject's breathing zone over the exposure period (not an 8-hour TWA):
1 = 1-10 ppm, 2 = >10-100 ppm, 3 = >100-200 ppm, 4 = >200 ppm. The confidence level was
assigned as 1 = literature contradictory or no information was available; 2 = one metric
(probability, frequency, or intensity) was based on the literature or self-report; 3 = two metrics
B-92
-------
were based on the literature or self-report; and 4 = all metrics based on the literature or directly
from self-report. Of the 180 cases and 481 controls interviewed, 9 (5%) cases and 4 (0.8%)
controls were identified as "ever" holding job as textile, apparel, and furnishing machine
operator or tender, of whom, 5 cases (3%) and 3 (0.7%) controls were dry cleaners. Regarding
specific exposures, 29 cases (19%) and 63 (13%) controls were assigned "ever" exposed to
tetrachloroethylene, of whom, 17 (3%) cases and 15 (3%) controls were assigned high
cumulative tetrachloroethylene exposure (>7,794 ppm-hours).
Statistical analyses consisted of unconditional logistic regression to estimate odds ratios
and their 95% CIs for associations between the risk of multiple myeloma and the exposure
surrogate ["ever" employed in occupation or industry (Gold et al., 2010a) or exposed to any of
the six chlorinated solvents or to each of the chlorinated solvents (Gold et al., 2010b)1. Other
surrogates examined were employment duration, cumulative exposure (for each exposed job, the
midpoint of intensity x the midpoint of frequency x total years worked, summed over all
exposed jobs), cumulative exposure for all jobs with a probability score of 2 or greater, and all
jobs with solvent exposures lagged 10 years. All models adjusted for sex, age, race, education,
and SEER site. As a sensitivity analysis, all analyses were repeated, assuming occupations with
confidence scores of 1 were considered as unexposed.
A strength of this study is its use of detailed occupational information to improve
assessment of solvent exposure compared to analyses based only on job title. Even so, exposure
misclassification was likely. Some limitations of this study were relatively low participation
rates among cases and controls, the inability to examine race or socioeconomic status, and if
associated with occupation and, potentially, solvents exposure, the potential for selection bias,
and small numbers of subjects with exposure to individual chlorinated solvents with limited
statistical power. Last, the study may reflect relationships between chlorinated solvents and less
severe forms of multiple myeloma due to the large proportion of cases who died before they
could be contacted or eligible subjects who refused to participate, particularly, if refusal was
related to being too ill.
B.2.2.7.4. Hardell et al. (1981)
Hardell, L.; Eriksson, M.; Lenner, P.; Lundgren, E. (1981). Malignant lymphoma
and exposure to chemicals, especially organic solvents, chlorophenols and phenoxy
acids: A case-control study. British Journal of Cancer, 43(2), 169-176.
http://www.ncbi.nlm.nih.gov/pubmed/7470379
Summary: This study used a case-control design to examine the possible relationship
between exposure to chemical classes (organic solvents, chlorophenols, and phenoxy acids) and
Hodgkin lymphoma and non-Hodgkin lymphoma. Cases consisted of men aged 25 to 85 years
B-93
-------
with histologically confirmed malignant lymphoma between 1974 and 1978. Living controls
were obtained from the National Population Registry and matched to cases on sex, age, and
municipality. Potential living controls were excluded if they did not live in the same
municipality at the time the case was diagnosed, if they were deceased, or if they had emigrated.
Deceased controls were obtained from the National Registry for Causes of Death and matched on
sex, age, municipality, and year of death. Potential deceased controls were excluded if their
death had occurred in 1978, was the result of suicide or malignant tumors, or if the date of last
employment did not occur within 5 years of the deceased case's last employment. There were
initially 8 living controls matched to each living case and 10 deceased controls matched to each
deceased case; in each instance, the two controls closest in age to cases were used in the analysis.
The final sample consisted of 169 cases (60 with Hodgkin lymphoma and 109 with non-Hodgkin
lymphoma) and 338 controls.
Self-administered questionnaires inquired about leisure-time activities, smoking/drug use,
exposure to chemicals, and occupational history (including time and place of employment). A
blinded individual evaluated each questionnaire and conducted telephone interviews with the
participants when information was unclear or incomplete. Exposure to organic solvents,
including tetrachloroethylene, was categorized into high grade and low grade (continuous
exposure of <1 week or repeated, brief exposure for <1 month). There were 10 (5.9%) cases and
31 (9.2%) controls who reported exposure to low-grade organic solvents, though the authors do
not report if this included tetrachloroethylene. Of the 40 (23.7%) cases and 47 (13.9%) controls
who reported high-grade exposure to organic solvents, only 1 (0.6%) case reported exposure to
tetrachloroethylene. Chi-square tests based on Miettinen (1970) were used to calculate $
estimates and odds ratios. Corresponding 95% CIs were determined according to Miettinen
(1976). Limitations to the study include its inability to independently evaluate the effect of
tetrachloroethylene within the chlorinated solvents category and possible misclassification due to
self-reported exposures. The authors do not report any strengths of their methodology.
B.2.2.7.5. Kato et al. (2005)
Kato, L; Koenig, K. L.; Watanabe-Meserve, H.; Baptiste, M. S.; Lillquist, P. P.;
Frizzera, G.,... Shore, R. E. (2005). Personal and occupational exposure to organic
solvents and risk of non-Hodgkin's lymphoma (NHL) in women (United States).
Cancer Causes Control, 16,1215-1224.
Summary. This study used a population-based case-control design to examine whether
exposures to solvents were associated with the risk of non-Hodgkin lymphoma in women. Cases
were identified through the New York State Cancer Registry and consisted of women aged 20 to
79 years living in New York State and diagnosed with non-Hodgkin lymphoma between 1995
B-94
-------
and 1998. Any potential cases with a previous history of hematologic cancers or without a valid
driver's license were excluded. Two sets of controls were used. Those under the age of 65 years
were obtained from an age-stratified random sample of driver's licenses from the New York
Department of Motor Vehicles (DMV), and those 65 years and older were identified from Health
Care Finance Administration (HCFA) beneficiary records. All eligible cases and DMV controls
were first sent a solicitation letter by the New York Cancer Registry. Only cases and DMV
controls that responded to the letter were contacted for an interview. Of the 722 eligible cases,
376 (56%) participated. The participation rates were 30% for DMV controls and 67% for HCFA
controls. The authors did not report response rates for cases or controls. The final sample
consisted of 376 cases and 463 (248 DMV and 215 HCFA) controls.
Blinded telephone interviews were conducted with both the cases and controls with a
structured questionnaire. Nearly 21% of the case interviews and just over 3% of the control
interviews were conducted with proxy respondents, in this case, next of kin. This occurred when
the participant was either deceased or medically incapable of answering the questions. The
median time between the cancer diagnosis and the interview was 1.2 years and ranged between
2 months and 3.3 years. A total of 50 (13.3%) cases and 48 (10.4%) controls reported
occupational exposure to degreasers and cleaning solvents, and 7 (1.9%) cases and 8 (1.7%)
controls reported occupational exposure to dry-cleaning fluids. To allow for a minimum lag
period of 1 year, an index date was determined for each case. Any exposures that occurred after
this date were excluded from the analysis.
Unconditional logistic regression estimated odds ratios and 95% CIs for occupational and
household exposures to solvents, adjusting for age at index date, family history of hematologic
cancer, college education, surrogate status, year of interview, BMI 10 years before interview,
average frequency of use of pain-relieving drugs, total number of episodes of systemic antibiotic
use, total number of uses of household pesticide products, and duration of work involving
pesticide exposures. This study's strength is its questionnaire that examined long-term
exposures by asking about the participant's whole personal history. Limitations include its self-
reported occupational history and the potential for recall bias in measuring exposures to
degreasers/cleaning solvents and dry-cleaning fluids. It also includes a limited number of
household products that contained organic solvents, which may have underestimated actual
exposure. Recall bias and the low response rate were additional methodological limitations to
this study.
B-95
-------
B.2.2.7.6. Malone et al. (1989)
Malone, K. E.; Koepsell, T. D.; Dating, J. R.; Weiss, N. S.; Morris, P. D.; Taylor, J.
W., Lyon, J. L. (1989). Chronic lymphocytic leukemia in relation to chemical
exposures. Am J Epidemiol, 130,1152-1158.
http://www.ncbi.nlm.nih.gov/pubmed/2589308
Summary: This study used previously collected data from a larger study (Koepsell et al.,
1987) to examine the relationship between selected occupations or chemical exposures and
leukemia. Cases were identified through SEER reporting sites in Washington state, Utah,
Michigan, and Georgia and consisted of men and women under the age of 80 years who were
diagnosed with chronic lymphocytic leukemia between 1977 and 1981. Of the eligible cases,
82.5% responded, and 430 were interviewed. The authors were unclear regarding the number of
eligible cases. Three of the cases were excluded because the interview failed to provide any
information on chemical exposures. Controls were randomly selected in one of two ways:
(1) random digit dialing in Utah, Michigan, and Georgia and (2) area sampling in Washington
state. Controls were matched based on sex, race, and/or age, depending on the location. Of the
2,028 eligible controls, 83% were interviewed. The final sample consisted of 427 cases and
1,683 controls.
B.2.2.7.7. Mester et al. (2006), Seidler et al. (2007)
Mester, B.; Nieters, A.; Deeg, E.; Eisner, G.; Becker, N.; Seidler, A. (2006).
Occupation and malignant lymphoma: A population based case control study in
Germany. Occup Environ Med, 63,17-26.
http://dx.doi.org/10.1136/oem.2005.020453
Seidler, A.; Mohner, M.; Berger, J.; Mester, B.; Deeg, E.; Eisner, G.,... Becker, N.
(2007). Solvent exposure and malignant lymphoma: A population-based case-
control study in Germany. J Occup Med Toxicol, 2, 2.
http://dx.doi.org/10.1186/1745-6673-2-2
Summary: A multicentre, population-based case control study was conducted in
six regions in Germany that is part of a larger multicountry lymphoma case-control study (the
EPILYMPH study). Cases were identified through physicians who played a role in the diagnosis
and treatment of malignant lymphoma in patients admitted to hospitals in each of the study areas.
Cases consisted of German residents (men and women) aged 18 to 80 years who were diagnosed
with either non-Hodgkin or Hodgkin lymphoma. Controls were identified from the population
registration office and matched to cases based on sex, region, and age. The participation rate
among controls was 44.3%; more than half (51%) of those who did not participate cited reasons
related to lack of interest. Additionally, in order to be included in the study, cases and controls
B-96
-------
needed to be familiar with the German language. The final sample consisted of 710 cases and
710 controls.
In-person interviews were conducted with trained interviewers and inquired about the
participant's medical history, lifestyle behaviors (smoking, alcohol, etc.) and activities, and
occupational history. The occupational history obtained information on dates of employment,
title, industry, and tasks associated with each job held for at least 1 year. Any participants who
reported potentially hazardous jobs (including dry cleaning) were asked additional questions
about their job tasks based on Bolm-Audorff et al. (1989).
Mester et al. (2006) aimed to identify occupations suspected to be associated with
lymphoma risk and to generate new hypotheses about occupational risks. Occupation was
assessed as a proxy for exposure, with job titles and industries blindly coded by two individuals
from the Frankfurt Institute for Occupational Medicine according to ISCO-68 and Statistical
Classification of Economic Activities in the European Community. Launderers, dry cleaners,
and pressers comprised ISCO-68 Code 56 and included 11 (1.5%) cases and 11 (1.5%) controls.
Conditional logistic regression was used to calculate odds ratios and their corresponding 95%
CIs, adjusted for smoking and alcohol consumption. Unconditional logistic regression was
employed to further examine lymphoma subentities (Hodgkin lymphoma, B-non-Hodgkin
lymphoma, T-non-Hodgkin lymphoma, B-non-Hodgkin lymphoma and Hodgkin lymphoma, and
other lymphomas), estimating odds ratios and their corresponding 95% CIs, adjusted for age,
sex, region, smoking, and alcohol consumption. All estimates were stratified by employment
duration (<10 years or >10 years). Additional analyses examined the effect of a latency period
of 10 years by only including exposures that occurred up until 10 years before diagnosis, though
the data were not reported. Limitations to the study include the low control response rate,
possible misclassification in the assessment of exposure through employment in specific
industries and occupations, small numbers of exposed, and lack of control for race/ethnicity or
immigration status. The authors do not report strengths associated with the methodology of their
study.
Seidler et al. (2007) examined the association between exposure to chlorinated
hydrocarbons and lymphoma on an in-depth expert assessment of solvent exposure. Intensity
and frequency of exposure was assessed by a blinded, trained industrial physician. Intensity was
evaluated as low (0.5-5 ppm), medium (>5-50 ppm), or high (>50 ppm). Frequency was
calculated as the percentage of weekly working time exposed and was categorized as low
(1-5%), medium (>5-30%), or high (>30%). Confidence in the exposure was classified as
possible, probable, or certain, and cumulative exposure (ppm-years) to each solvent for each
occupation was also calculated. Overall, there were 36 (5.1%) cases and 31 (4.4%) controls
exposed to tetrachloroethylene. Conditional logistic regression was used to calculate odds ratios
B-97
-------
and 95% CIs, adjusted for smoking and alcohol consumption. The authors reported only those
calculations with at least five participants reporting exposures. Tests for trend were analyzed by
including specific exposures as continuous variables in the logistic regression model.
Unconditional logistic regression was employed to estimate odds ratios and 95% CIs in an
unmatched analysis of the most frequent lymphoma subentities (Hodgkin lymphoma,
B-non-Hodgkin lymphoma, T-non-Hodgkin lymphoma, B-non-Hodgkin lymphoma and Hodgkin
lymphoma, and other lymphomas), adjusted for age, sex, region, smoking, and alcohol.
Strengths of the study include blinded exposure assessment, adjustment for potential
confounders, and an expert-based estimate of solvent exposure. A limitation of this study is the
low-exposure prevalence.
B.2.2.7.8. Miligi et al. (2006; 1999), Costantini et al. (2008; 2001)
Miligi, L.; Costantini, A. S.; Benvenuti, A.; Kriebel, D.; Bolejack, V.; Tumino, R.,
... Vineis, P. (2006). Occupational exposure to solvents and the risk of lymphomas.
Epidemiology, 17, 552-561. http://dx.doi.org/10.1097/01.ede.0000231279.30988.4d
Miligi, L.; Seniori, C. A.; Crosignani, P.; Fontana, A.; Masala, G.; Nanni, O.,...
Vineis, P. (1999). Occupational, environmental, and life-style factors associated with
the risk of hematolymphopoietic malignancies in women. Am J Ind Med, 36, 60-69.
http://dx.doi.org/10.1002/(SICI)1097-0274(199907)36:K60::AID-AJIM9>3.0.CO;2-
Z
Costantini, A. S.; Benvenuti, A.; Vineis, P.; Kriebel, D.; Tumino, R.; Ramazzotti, V.,
... Miligi, L. (2008). Risk of leukemia and multiple myeloma associated with
exposure to benzene and other organic solvents: Evidence from the Italian
Multicenter Case-control study. Am J Ind Med, 51, 803-811.
http://dx.doi.org/10.1002/aiim.20592
Costantini, A. S.; Miligi, L.; Kriebel, D.; Ramazzotti, V.; Rodella, S.; Scarpi, E.,...
Vineis, P. (2001). A multicenter case-control study in Italy on hematolymphopoietic
neoplasms and occupation. Epidemiology, 12, 78-87.
http://www.ncbi.nlm.nih.gov/pubmed/11138825
Summary: These four publications report on a large, population-based case-control study
examining pesticide or solvent exposures and hematolymphopoietic malignancies. The studies
were conducted in 12 different parts of Italy (Turin, Ragusa, Siena, Alessandria, Forli, Novara,
and Vercelli, as well as Florence, Verona, Imperia, Latina, and Varese provinces), but only 11 of
the locations had interviews available for analysis. The authors do not note which study site was
excluded. Cases consisted of men and women aged 20 to 74 years who were diagnosed with
hematolymphopoietic malignancies between 1991 and 1993. Cases were obtained through
surveys with public hospitals in the 12 study areas, as well as regional medical centers or
university-affiliated hospitals in Milan, Pavia, Rome, and Bologna to ensure complete collection
B-98
-------
of cases from all 12 locations. One location (Varese) found all of its cases through the local
cancer registry. Controls were randomly selected from the general population of residents in
each of the study locations, stratified by sex and 5-year age groups. Of the 3,357 eligible cases,
3,118 were able to be contacted, and 2,737 (88%) were interviewed, including
1,450 non-Hodgkin lymphoma, 365 Hodgkin lymphoma, 652 leukemia, and 270 multiple
myeloma cases. Of the 2,391 eligible controls, 2,196 were able to be contacted, and 1,779 (81%)
were interviewed.
In-person interviews were conducted to obtain information on education, lifestyle
behaviors, occupational history, extraoccupational exposure to solvents and pesticides, hair dye
use, lifelong residential history, medical history, and reproductive history. Proxy interviews
were conducted with spouses (45%), children (28%), parents (11%), or another relative (16%)
for 19% of the cases and 5% of the controls. The occupational history section of the
questionnaire was created by industrial hygienists and agronomists and inquired about the
participant's full working history as well as exposure to chemicals, solvents, and pesticides.
Industrial hygienists from each of the areas blindly assessed occupational exposures, evaluating
the probability and intensity of exposures to categories of solvents as well as individual
chemicals, including tetrachloroethylene. Probability was rated as low, medium, or high, and
intensity was classified as "very low," "low," "medium," and "high." To ensure consistency in
assessment, a job exposure matrix was created with the minimum overall consensus for those
jobs that were reported most frequently.
Costantini et al. (2001) investigated the associations between occupational exposures and
hematolymphopoietic neoplasms. This study used the full sample of 2,737 cases and
1,779 controls within the 11 available study areas. All jobs were coded according to
International Standard Classification of Occupations; launderers, dry cleaners, and pressers fell
within Code 56. There were 3 (0.2%) male non-Hodgkin lymphoma cases, 1 (0.3%) male
Hodgkin lymphoma case, and 2 (0.3%) male leukemia cases who were employed in the
dry-cleaning industry. Odds ratios and their corresponding 95% CIs for non-Hodgkin
lymphoma, Hodgkin lymphoma, leukemia, and multiple myeloma were calculated using the
Mantel-Haenszel approach adjusted for age. The results for men were presented and compared
with the results for men and women combined. Only those occupations that had a minimum of
five exposed cases in at least one gender were presented. The authors did not specifically
mention strengths associated with the methodology; a limitation is its use of occupation as a
proxy for examining risks associated with chemical exposures.
Miligi et al. (2006) evaluated the association between solvent exposure in the work
environment and non-Hodgkin lymphoma, including chronic lymphatic leukemia, and Hodgkin
lymphoma. Due to the fact that the industrial hygienists' exposure assessment was only
B-99
-------
completed in 8 areas, there were 1,719 eligible cases of non-Hodgkin lymphoma, 347 eligible
cases of Hodgkin lymphoma, and 2,086 eligible controls. The final sample consisted of
1,428 (83%) NHL cases, 304 Hodgkin lymphoma cases (88%), and 1,530 controls (73%). In-
person interviews were conducted with 85% of the non-Hodgkin lymphoma cases, 93% of the
Hodgkin lymphoma cases, and 97% of the controls. Intensity of exposure was classified as
"very low/low" or "medium/high." There were 18 (1.3%) non-Hodgkin lymphoma cases and
29 (1.9%) controls with "very low/low" exposure to tetrachloroethylene and 14 (1.0%)
non-Hodgkin lymphoma cases and 15 (1.0%) controls with "medium/high" exposure. Duration
of exposure was categorized as having lasted <15 years or >15 years. There were 10 (0.7%)
non-Hodgkin lymphoma cases and 10 (0.7%) controls who reported <15 years of exposure to
tetrachloroethylene and 3 (0.2%) non-Hodgkin lymphoma cases and 5 (0.3%) controls who
reported 15 years or more. Odds ratios and their corresponding 95% CIs were calculated for
non-Hodgkin lymphoma, non-Hodgkin lymphoma subtypes, and Hodgkin lymphoma,
individually. All were adjusted for sex, age, education, and area. Strengths of the study include
its large sample size and the exclusion of participants who were classified as having a low
probability of exposure. Limitations include the potential for misclassification of subjects by
individual chemical, the low-exposure prevalence to tetrachloroethylene, the high percentage of
proxy interviews among the case series, and the small sample of cases for each lymphoma
subtype, which for Hodgkin lymphoma, prevented the examination of potential associations by
individual chemical.
Costantini et al. (2008) examined the association between solvent exposure and
occurrence of leukemia subtypes and multiple myeloma. The final samples consisted of
586 cases of leukemia (acute myeloid leukemia and chronic lymphatic leukemia) and
1,278 controls collected from 7 of the locations, as well as 236 cases of multiple myeloma and
1,100 controls collected from 6 of the sites. Intensity of exposure was classified as "very
low/low" or "medium/high." There were 6 (1.0%) leukemia cases, 17 (1.3%) leukemia controls,
3 (1.3%) multiple myeloma cases, and 15 (1.4%) multiple myeloma controls with "very
low/low" exposure to tetrachloroethylene and 7 (1.2%) leukemia cases, 12 (0.9%) leukemia
controls, 2 (0.8%) multiple myeloma cases, and 12 (1.1%) multiple myeloma controls with
"medium/high" exposure. Duration of exposure was categorized as having lasted <15 years or
>15 years, though tetrachloroethylene was not specifically reported. Point odds ratios and their
corresponding 95% CIs were calculated for leukemia, leukemia subtypes, and multiple myeloma,
individually. All were adjusted for gender, age, education, and area. The authors did not report
the method used to derive estimates. A linear test for trend was also conducted using the
midpoints of all duration categories (0, 7.5, and 35 years). The authors did not note any
B-100
-------
strengths or limitations of their methodology, although those identified for Miligi et al. (2006)
are relevant for this study.
B.2.2.7.9. Schenk et al. (2009)
Schenk, M.; Purdue, M.; Colt, J.; Hartge, P.; Blair, A.; Stewart, P.,... Severson, R.
(2009). Occupation/industry and risk of non-Hodgkin's lymphoma in the United
States. Occup Environ Med, 66, 23-31. http://dx.doi.org/10.1136/oem.2007.036723
Summary: This study used a case-control design to examine the relationship between
occupation and development of non-Hodgkin lymphoma. Cases were identified through the
SEER registry and consisted of men and women aged 20 to 74 years, living in Iowa or selected
parts of California, Michigan, or Washington state and diagnosed with histologically confirmed
non-Hodgkin lymphoma between 1998 and 2000. Controls were selected in two ways. Those
under the age of 65 years were chosen from random digit dialing, and those 65-74 years were
chosen through Medicare files. Controls were matched to cases on 5-year age group, gender, and
race within each study center. All HIV-positive individuals were excluded from both the cases
and controls, as were controls with a previous diagnosis of non-Hodgkin lymphoma. Of the
2,248 eligible cases, 1,728 (77%) were contacted, and 1,321 participated in the interview,
yielding a response rate of 59% and a participation rate of 76%. Of the 2,409 eligible controls,
2,046 (85%) were contacted, and 1,057 participated, yielding a response rate of 44% and a
participation rate of 52%. After excluding those cases and controls who were never employed or
whose occupations were unknown, the final sample consisted of 1,189 cases (293 follicular,
366 diffuse large B-cell lymphoma, 487 other, 43 unknown) and 982 controls.
Initially, all participants were mailed a self-administered questionnaire inquiring about
either family and medical history or diet. Then participants were visited in their homes for a
computer-assisted interview. All participants were asked about demographics, hair coloring,
residential history since 1970, and occupational history (Chatterjee et al., 2004). The
occupational history asked about all jobs lasting 6 months or longer and obtained information on
location, dates of employment, job title, and number of hours worked (part-time or full-time).
Occupation was assessed as a proxy for exposure, and all jobs were blindly assigned occupation
and industry codes according to SOC and SIC conventions. Participants were considered
exposed if they had ever been employed in a particular occupation or industry and unexposed if
they had not. Launderers and ironers were assigned Code 503 and included a total of 12 (1.0%)
cases and 3 (0.3%) controls.
Unconditional logistic regression was used to estimate odds ratios and their
corresponding 95% CIs, adjusted for age, gender, ethnicity, and study center. Analyses were
also performed, stratified by gender and histological subtype separately. Strengths of the study
B-101
-------
include its population-based sample, the large number of cases, its detailed information on
multiple risk factors, as well as its ability to examine this disease by multiple histological
subtypes and by gender. Limitations to the study include its small number of exposed
participants and low power. The study also fails to examine intensity or duration of exposure
and may be subject to selection bias in light of its low participation rate.
B.2.2.7.10. Scherr et al. (1992)
Scherr, P. A.; Hutchison, G. B.; Neiman, R. S. (1992). Non-Hodgkin's lymphoma
and occupational exposure. Cancer Res, 52, 5503s-5509s.
http://www.ncbi.nlm.nih.gov/pubmed/1394164
Summary: This hospital-based case-control study of non-Hodgkin lymphoma examined
occupations and exposures associated with increased risk of all NHLs or of specific NHL
histological subtypes. The case series were patients diagnosed with NHL from January 1, 1980,
to May 31, 1982, treated at any of nine participating Boston hospitals, and residents of the
Boston Standard Metropolitan Statistical Area. A total of 379 NHL histologically confirmed
cases were identified, of which, 303 interviews with the living case, next of kin, or parent for
ages 17 years or younger (80% response rate). A pathology review of the cases confirmed the
NHL diagnosis, and this is one of the early studies to classify NHL subtypes according to the
Modified Rappaport or Working Formulation Classification, if tumors were nodular or diffuse,
or by cell type (B- or T-cell). The control series were randomly selected from residence lists for
all Massachusetts towns and, for controls 17 years of age or older, matched to cases based on sex
and age. For cases under 18 years of age, possible controls were identified from matching based
on the age and sex of a case's parent or guardian and interviewed to determine whether he or she
had a child of the same age and sex as the case. Of 423 potential controls, 303 were interviewed
(72% response rate). All interviews were carried out with the liver control. No statistically
significant differences between cases and controls were found for education, marital status,
current family income, and highest family income. Religion was found to differ between cases
and controls (p < 0.05).
Face-to-face interviews were carried out using a questionnaire that sought information on
current or most recent job, job held 15 years previously, major and second major occupation, and
exposure to a list of agents that included chlorinated solvents as a category. One-third of cases'
responses were from proxy or next-of-kin respondents. Each occupation was categorized by
occupation and industry and coded according to the Dictionary of Occupational Titles.
Nine cases (3% exposure prevalence) were identified as holding jobs in laundering, dry cleaning,
and leather products fabrication, and 73 cases (24% exposure prevalence) reported exposure to
chlorinated solvents.
B-102
-------
Statistical analyses were carried out using a hierarchal approach that aggregated
histological subtypes into groups with similar histological characteristics and exposure defined
as a function of calendar time (1901-1949, 1950-1959, 1960-1969, 1970 and later) or exposure
duration (10 years, 20 years). All exposure that showed consistent patterns within histological
categories over calendar time or over duration were considered as candidate variables for
conditional logistic models with covariates for age and sex.
A strength of the study is its examination of NHL subtypes, although different
classification schemes were used, and no attempt was made to harmonize across schemes. The
lack of a common scheme to classify NHLs might lead to potential for disease misclassification.
Another limitation of the study includes the high percentage of proxy respondents (33% cases,
no controls), which may have led to measurement error and misclassification bias in exposure
assignment. Last, the statistical analyses, using a hierarchal approach, could identify true
positive associations but was not specific, with a greater potential for some false negative
findings.
B.2.2.8. Childhood lymphopoietic cancers
B.2.2.8.1. Infante-Rivard et al. (2005)
Infante-Rivard, C.; Siemiatycki, J.; Lakhani, R.; Nadon, L. (2005). Maternal
exposure to occupational solvents and childhood leukemia. Environ Health
Perspect, 113, 787-792. http://dx.doi.org/10.1289/ehp.7707
Summary: This study used a population-based case-control design to examine the
possible relationship between childhood leukemia and maternal exposure to occupational
solvents. Cases consisted of children who were diagnosed with acute lymphoblastic leukemia
and were being treated in tertiary care centers in Quebec Province, Canada. Between 1980 and
1993, children aged 0 to 9 years were included in the study; between 1994 and 2000, cases were
children between 0 and 14 years of age. Controls were selected in two ways. The 1980-1993
controls were obtained from government records indicating all families who had received a
stipend for having children living legally in Canada. The 1994-2000 controls were chosen from
universal health insurance records for the Quebec Province. Both mechanisms provided the most
complete census of children during these time periods. Controls were matched to cases on sex
and age at diagnosis. Prospective participants were excluded if children were adopted or lived
with foster families, if neither French nor English was spoken, if they did not currently reside in
Canada, or if the parents were not available to be interviewed. Of the 848 eligible cases, 790
(93.1%) of parents were interviewed. Of the 916 eligible controls, 790 (86.2%) of parents were
interviewed. The final sample consisted of 790 cases and 790 controls.
B-103
-------
Telephone interviews with parents were conducted using a structured questionnaire,
which inquired about general risk factors, potential confounders, and maternal occupational risk
factors. The latter consisted of a complete occupational history (job title, industry, name,
address) provided by the mother for the time period, which began when she was 18 years of age
and ended with the birth of the child. All jobs held by the mother during the 2 years prior to the
birth of the child were further examined using a semi-structured questionnaire to obtain
information on the company's activities, raw materials and machines used, goods produced,
responsibilities, working conditions, activities of coworkers, and the presence of solvents and
other chemicals. Finally, for all jobs for which there was significant possible exposure,
including textile dry cleaners, more detailed questionnaires were used to inquire about the
specific tasks, the time spent at these tasks, the specific exposures associated with these tasks,
and the environment in which the tasks were carried out.
Exposures were classified by a team of blinded chemists and industrial hygienists. All jobs were
coded according to standard Canadian industrial titles (3-digit codes) and job titles (7-digit
codes). Then, the team determined whether or not participants were exposed to a list of over 300
chemicals, including tetrachloroethylene, based on the information provided by the respondent,
previous information on exposures in that geographical area, and the team's knowledge of
exposures in the industry in question. All jobs held in the 2-year time period before the
pregnancy were coded separately, based on the team's confidence that the exposure had occurred
(possible, probable, or definite), the frequency of the exposure during a normal workweek (<5%,
5-30%, or >30% time), and the level of the concentration (low/background, medium, high).
This methodology for exposure assessment has been validated and used in other research
publications. All chemicals were assigned 3 digit codes based on Siemiatycki (1991):
tetrachloroethylene was Code 243. This study did not provide sufficient detail to determine the
number of cases exposed to tetrachloroethylene to calculate a prevalence of exposure.
Conditional logistic regression was performed separately for each chemical using two time
periods, the 2 years before the child's birth and during pregnancy. Odds ratios and their
corresponding 95% CIs were estimated, adjusting for maternal age and level of schooling. A
strength of this study is the use of a detailed exposure assessment. Limitations include power
due to small sample size and nondifferential misclassification.
B-104
-------
B.2.2.8.2. Lagakos et al. (1986), Costas et al. (2002)
Lagakos, S. W.; Wessen, B. J.; Zelen, M. (1986). An analysis of contaminated well
water and health effects in Woburn, Massachusetts. J Am Stat Assoc, 81, 583-596.
Costas, K.; Knorr, R. S.; Condon, S. K. (2002). A case-control study of childhood
leukemia in Woburn, Massachusetts: The relationship between leukemia incidence
and exposure to public drinking water. Sci Total Environ, 300, 23-35.
http://dx.doi.org/10.1016/S0048-9697(02)00169-9
Summary: Lagakos et al. (1986) had two aims: (1) to assess the association between
access to contaminated water and the incidence rate of childhood leukemia, and (2) to determine
whether adverse pregnancy outcomes (fetal wastage, low birth weight, stillbirth, birth defects)
were correlated with exposure to water from the contaminated wells. Cases were identified
through the state cancer registry and the Dana-Farber Cancer Institute/Children's Hospital cancer
registry and consisted of male and female children aged 19 years and younger who were
diagnosed with leukemia in Woburn between 1964 and 1983. In total, 20 cases of childhood
leukemia were identified. This study was a precursor to a later case-control study of childhood
leukemia by Costas et al. (2002).
In 1982, telephone interviews were conducted by blinded, trained interviewers with
Woburn residents. Of the 8,109 telephone numbers, 7,134 (88%) were contacted. After
excluding for business, second phones, and disconnected numbers, the sample decreased to
6,219 households. Of these, 5,010 (80.6%) were interviewed. The questionnaire used during the
interviews inquired about all pregnancies 1960 and 1982, excluding elective and spontaneous
abortions, chronic and recurrent child health problems, and the residential history for each family
member (current and former) up through, but excluding the current address. Pregnancy
information obtained included the date the pregnancy ended, maternal age, and smoking status
during pregnancy, offspring vital status at delivery, offspring weight, gender, and congenital
anomalies. A 1983 study by Waldorf and Cleary estimated the monthly distribution of water
from the contaminated wells for the time the wells were in use (1964-1979). These estimates
were used to determine the proportion of annual water supplied by the contaminated wells for
each household between 1960 and 1982. Annual exposures for pregnancies were based on the
year the pregnancy ended; annual exposures for children were determined starting the first year
they lived in Woburn. Cumulative and binary metrics were used to characterize exposure to well
water in the study area. Of the 4,396 pregnancies that occurred during the study period and
4,978 children about whom information was obtained, approximately 16% of pregnancies, and
27% of children were estimated to have had some exposure to the contaminated wells.
A Cox hazards regression model was used to estimate whether the distribution of
childhood leukemia cases was associated with the contaminated wells. Individual risk sets
B-105
-------
consisted of children from the survey of adverse pregnancy outcomes and childhood disorders
who were matched to cases on year of birth and were residents of Woburn when the case was
diagnosed. These risk sets were used to estimate the expected cumulative exposure for each
case. Logistic regression using the maximum likelihood method was employed to estimate odds
ratios and their corresponding 95% CIs for adverse pregnancy outcomes including spontaneous
abortion, perinatal death, low birth weight, and musculoskeletal, cardiovascular, eye/ear, CNS,
chromosomal, and oral cleft anomalies separately. Each adverse outcome was adjusted for its
own set of risk factors. Overall, these were maternal age during pregnancy, smoking during
pregnancy, year pregnancy ended, and mother's pregnancy history, which included prior
spontaneous abortion, prior perinatal death, prior low birth weight, and prior musculoskeletal
anomaly. A survival time model with age of diagnosis as the time variable was used to estimate
relative risks of childhood disorders, including anemia/blood disorders, allergy/skin disorders,
kidney/urinary tract disorders, lung/respiratory tract disorders, neurologic/sensory disorders,
learning disabilities, and other disorders separately. These were each adjusted for their own set
of risk factors. Overall, these were year the pregnancy ended, age at pregnancy, SES, and sex.
A strength of this study is its large sample of pregnancy outcomes and child disorders. A
limitation is its use of annual exposure estimates, which may not have been accurate enough to
assess intensity of exposure, particularly for pregnancy outcomes. A second limitation included
potential bias of omission of families that moved from Woburn prior to 1982. Nonresponse of
eligible, omitted households that were not contacted may have introduced bias.
Costas et al. (2002) used a matched case-control design in their follow-up to a
Massachusetts Department of Health study (Cutler et al., 1986), which found a cluster of
leukemia cases in Woburn, Massachusetts. This study expanded on the initial research by using
water distribution models to assign exposure rather than location of residence (Costas et al.,
2002) and aimed to determine whether childhood leukemia was associated with exposure to
water from Wells G and H (MDPH, 1997). Cases consisted of children who were diagnosed
with leukemia before their 19th birthday between 1969 and 1989. Those diagnosed before 1982
were identified through pediatric health professionals and greater-Boston pediatric oncology
centers, and those diagnosed from 1982 onwards were obtained through the Massachusetts
Cancer Registry. Controls were randomly selected from Woburn Public School records and
matched to cases based on race, sex, and date of birth. Two controls were matched to each case.
Both cases and controls were required to be Woburn residents at the time of the case's diagnosis.
Of 21 eligible cases, 19 (90.5%) participated in the study. Of the 38 controls selected, one was
excluded from the study when it became known that they no longer fit the inclusion criteria. The
authors do not report response rates for controls. The final sample consisted of 19 cases and
B-106
-------
37 controls. Cases and controls were similar on family history of cancer, maternal smoking and
alcohol consumption, and potential exposure to 60 Hz electric and magnetic fields.
In-person interviews were conducted with both parents of cases and controls, except for
two instances where the father was interviewed via telephone. The maternal questionnaire
inquired about demographics, lifestyle characteristics, medical history, environmental and
occupational exposures, and use of public drinking water at home. The paternal questionnaire
inquired about occupational history and occupational exposures. A detailed residential history
was also ascertained for each mother and child for the period 2 years preconception through date
of case diagnosis, and all were evaluated for electromagnetic field exposure using a power
distribution wire code scheme (Kaune and Savitz, 1994). All estimates were based on well water
contaminant levels, which were measured just before the wells' closure in 1979. As a result,
exposure was determined by the potential for a residence to receive water from the contaminated
wells rather than the actual concentrations of contaminants. A water distribution model
developed by Murphy (1990) was used to estimate water distribution patterns through the
creation of exposure index values for each neighborhood in Woburn for each month that the
wells were in use between 1964 and 1979. Overall, seven cumulative exposure scores were
estimated for each participant for the entire etiologic period, preconception, each trimester,
overall pregnancy, and period between birth and diagnosis. Two exposures were assessed for
each participant: (1) cumulative exposure (summed for all months of residence in a particular
location), and (2) average exposure (consisted of water exposure data averaged over time).
During the full exposure time frame (2 years preconception through case diagnosis), 16 (84.2%)
cases and 24 (64.9%) controls had been exposed to water from the contaminated municipal
wells. Of these, 7 cases (36.8%) and 13 controls (35.1%) received the most exposure, and
9 cases (47.4%) and 11 controls (29.8%) received the least.
Conditional logistic regression with a proportional hazards model was used to calculate
odds ratios and their corresponding 95% CIs. Unadjusted odds ratios examined the relationship
between case-control status and effects of maternal alcohol consumption, breastfeeding, paternal
grandfather with cancer, paternal employment in a high-risk industry, and public water as
primary beverage. Odds ratios for four exposure time periods (2 years preconception through
case diagnosis, 2 years preconception, pregnancy, birth to diagnosis) examined "ever" exposure
and subcategories within "ever" exposure (i.e., "most" or "least") separately. Each was then
adjusted for a composite covariate according to Tukey (1991), who controlled for socioeconomic
status, maternal smoking during pregnancy, maternal age at birth of child, and breastfeeding.
Trends related to increasing exposure (i.e., "never," "least," "most") were evaluated for each
exposure time period separately using the %2 method. Strengths of this study include its
adjustment for potential confounders and its use of exposure estimates that were developed
B-107
-------
through an investigation of the distribution of municipal water throughout the city. A limitation
of the study is its lack of information on actual well water contamination levels during the time
the wells were in use and the study's small size, which leads to low statistical power and
imprecise estimates of risk.
B.2.2.8.3. Lowengart et al. (1987)
Lowengart, R. A.; Peters, J. M.; Cicioni, C.; Buckley, J.; Bernstein, L.; Preston-
Martin, S.; Rappaport, E. (1987). Childhood leukemia and parents' occupational
and home exposures. J Natl Cancer Inst, 79, 39-46.
http://www.ncbi.nlm.nih.gov/pubmed/3474448
Summary: This case-control study investigated possible etiologic factors for childhood
leukemia. Cases were identified through the Los Angeles County Cancer Surveillance Program
and consisted of children aged 10 years or younger at the time of their diagnosis between 1980
and 1984. In order to be included in the study, biological case mothers were required to be
available for an interview. Controls were selected in two ways: (1) friends of cases were
identified by case mothers and asked to participate, and (2) population-based controls were
chosen through random digit dialing when friends were not available. Of the 216 eligible cases,
202 (94%) were able to be contacted. Of them, 159 (79%) mothers were interviewed. There
were 154 case fathers also interviewed, of which, 30 cases where mother's provided proxy
information on paternal variables. There were five fathers who did not participate in the study.
There were 136 control mothers and 130 control fathers who participated in the interview. There
were 6 fathers for whom interviews could not be obtained, and 43 of the paternal interviews were
by proxy with mothers. The authors do not report control response rates. Controls were matched
to cases based on age, sex, race, and Hispanic origin (if the race was "white"), though 3
population-based controls were unable to be matched based on sex and 10 were unable to be
matched based on race. After further exclusions (4 cases and 5 controls) for incomplete
occupational histories, the final sample consisted of 123 case-control pairs for which complete
information was available about both parents.
Telephone interviews were conducted by two, nonblinded, trained interviewers using a
structured questionnaire that inquired about family and personal medical histories, alcohol and
tobacco use, household and personal products, X-ray exposure, and occupational history (job
title, industry, time period worked). The maternal questionnaire also asked about medical
complications, use of drugs, and diet during the index pregnancy, as well as the child's medical
history and exposure to ionizing radiation. The interviews occurred between 1983 and 1985.
Industries and occupations were coded according to 1970 U.S. Census classifications and
grouped based on potential hydrocarbon exposure. All occupations and exposures within 1 year
B-108
-------
of conception were excluded. There was 1 (0.8%) case father who reported exposure to
tetrachloroethylene in the year before pregnancy, 1 (0.8%) case father who reported exposure
during pregnancy, and 2 (1.6%) case fathers who reported exposures after delivery. No control
fathers were exposed to tetrachloroethylene. The authors did not report information on maternal
exposure to tetrachloroethylene. Conditional logistic regression was used to estimate odds ratios
and their corresponding 95% CIs. Limitations of the study include its lack of exposure
verification and inability to assess intensity of exposure.
B.2.2.8.4. Shu et al. (1999)
Shu, X. O.; Stewart, P.; Wen, W. Q.; Han, D.; Potter, J. D.; Buckley, J. D.,...
Robison, L. L. (1999). Parental occupational exposure to hydrocarbons and risk of
acute lymphocytic leukemia in offspring. Cancer Epidemiol Biomarkers Prev, 8,
783-791. http://www.ncbi.nlm.nih.gov/pubmed/10498397
Summary: This case-control study examined the association between parental
occupational exposure and the risk of childhood acute lymphocytic leukemia. Potential cases
and controls were required to meet the following criteria: have a telephone in their place of
residence and have their English-speaking, biological mother available for an interview. Cases
consisted of children aged 15 years and under who were diagnosed with acute lymphocytic
leukemia between 1989 and 1993 by 1 of 37 participating Children's Cancer Group members or
institutions from Australia, California, Canada, Colorado, Illinois, Indiana, Iowa, Michigan,
Minnesota, Missouri, Nebraska, New York, North Carolina, Ohio, Oregon, Pennsylvania,
Tennessee, Texas, Washington state, Washington, DC, Wisconsin, and Utah. Of the
2,081 eligible cases, 1,914 (92%) mothers were interviewed. Controls were chosen through
random digit dialing and matched to cases based on age, race, and telephone area code and
exchange. Of the 2,597 eligible controls, 1,987 (76.5%) mothers were interviewed. After
excluding the 72 cases for whom a matched control could not be found, the final sample
consisted of 1,842 cases and 1,986 controls.
Telephone interviews were conducted with case mothers and fathers using a structured
questionnaire. The maternal questionnaire inquired about demographics; personal habits;
household exposures before and during index pregnancy; exposure to environmental hazards;
and occupational, medical, reproductive, and family histories. The paternal questionnaire
inquired about personal habits; household exposures; and medical, occupational, and family
histories. Of the 2,081 eligible cases and 2,597 eligible controls, fathers were interviewed for
1,801 (86.5%) cases and 1,183 (69.8%) controls, yielding 1,618 matched sets. The majority
(83.4% cases and 67.7% controls) consisted of direct interviews with fathers; the remainder were
proxy interviews with the mothers.
B-109
-------
Maternal occupational histories were collected for all jobs that lasted at least 6 months
and occurred between the 2 years prior to the pregnancy and the case's diagnosis, while paternal
occupational histories were obtained for all jobs lasting at least 6 months from age 18 years
onwards. Both maternal and paternal occupational histories inquired about job titles, industries,
duties, dates of employment, and exposure to solvents/degreasers/cleaning agents, plastic
materials, paints, pigments/thinners, and oil/coal products. Any self-reported exposures that
were not included as part of the exposures listed in the questionnaire were blindly assessed by an
industrial hygienist and placed into the established exposure categories. Maternal and paternal
dates of employment were used to determine whether an exposure occurred during
preconception, pregnancy, or the postnatal period; duration of exposure within each of these time
frames was calculated and categorized using the control group's median time as the cutoff.
Maternal exposures to tetrachloroethylene occurred anytime in 4 (0.2%) cases and 9 (0.5%)
controls, during preconception in 3 (0.2%) cases and 2 (0.1%) controls, during pregnancy in
3 (0.2%) cases and 2 (0.1%) controls, and during the postnatal period in 4 (0.2%) cases and
8 (0.4%) controls. Paternal exposures to tetrachloroethylene occurred anytime in 25 (1.4%)
cases and 23 (1.9%) controls, during preconception in 21 (1.2%) cases and 22 (1.9%) controls,
during pregnancy in 8 (0.4%) cases and 14 (1.2%) controls, and during the postnatal period in
10 (0.6%) cases and 15 (1.3%) controls.
Conditional logistic regression was used to estimate odds ratios and their corresponding
95% CIs for maternal exposures, adjusted for maternal education, race, and family income.
Unconditional logistic regression was used to calculate odds ratios and their corresponding
95% CIs for paternal exposures, adjusted for paternal education, race, family income, age, and
sex of the case. Tests for trend were conducted by incorporating the categorical variables of
exposure as continuous variables in the models. No strengths were reported by the authors.
Limitations included self-reported information on exposure based on a list of specific exposures
provided to the participant, the lack of information on intensity or level of exposure, the lack of
specific information related to additional exposures that prevented their categorization, low
prevalence of maternal tetrachloroethylene exposure, and the high proportion of proxy paternal
interviews that likely results in an increased potential for misclassification bias.
B.2.2.9. Neuroblastoma
B.2.2.9.1 DeRoos et al. (2001)
De Roos, A.; Olshan, A.; Teschke, K.; Poole, C.; Savitz, D.; Blatt, J.,... Pollock, B.
(2001). Parental occupational exposures to chemicals and incidence of
neuroblastoma in offspring. Am J Epidemiol, 154,106-114.
http://dx.doi.0rg/10.1093/aie/154.2.106
B-110
-------
Summary: This case-control study evaluated the effects of parental occupational exposure
on neuroblastoma incidence in offspring. Cases consisted of male and female children aged
18 years or under who were diagnosed with neuroblastoma between 1992 and 1994 and
registered in one of 139 participating hospitals in either the United States or English-speaking
Canada. Of the 741 eligible cases, 538 (73%) were enrolled in the study. A control was selected
for each of 504 cases through random digit dialing and matched to cases based on birth date.
Approximately 71% of the eligible controls were recruited, and 74% of the households that were
screened participated. After excluding those with missing occupational exposures and all proxy
interviews, the final sample consisted of 537 case mothers, 405 case fathers, 503 control
mothers, and 302 control fathers.
Telephone interviews were conducted with both mothers and fathers and inquired about
demographics as well as lifetime occupational history, which included dates of employment,
names of employers, occupations, industries, job titles, specific duties, and hours per week.
Chemical exposure histories were requested for all jobs held within 2 years of the index child's
birth. Interviews were conducted with 537 case mothers, 472 case fathers, 503 control mothers,
and 445 control fathers, though 67 (14.2%) of the case father interviews and 141 (31.7%) of the
control father interviews were completed by mothers as the proxy respondent. All proxy
interviews were subsequently excluded from the analyses. Exposure was assessed in two ways.
First, the participant was asked to report his/her possible exposure to any of 65 substances, as
well as the form (liquid, gas, dust, smoke, solid) and route (inhalation, dermal, ingestion,
clothing) of the exposure, the activities being performed during the exposure, the number of
hours per week exposed, and the time frame during which the exposure occurred.
Second, these responses were then reviewed by a blinded industrial hygienist who
reclassified any improbable exposures as nonexposed. The hygienist did not review the
responses of participants who reported no exposure to any of the possible chemicals during their
jobs; as a result, jobs that may have had exposure potential were not reclassified. The substances
themselves were classified into five categories: halogenated hydrocarbons, a category which
included tetrachloroethylene, nonvolatile hydrocarbons, volatile hydrocarbons,
paints/inks/pigments, and metals/alloys/solders. Maternal exposure to halogenated hydrocarbons
was reported by 15 (2.8%) cases and 19 (3.8%) controls. After the industrial hygienist's review,
this decreased to 6 (1.1%) cases and 8 (1.6%) controls. Among the fathers, 8 (2.0%) cases and
11 (3.6%) controls reported exposure to tetrachloroethylene more specifically; the industrial
hygienist's review subsequently decreased this to 4 (1.0%) cases and 6 (2.0%) controls.
Unconditional logistic regression was used to calculate exposure odds ratios and their
corresponding 95% CIs for each of the five categories of substances as well as for each of the
individual chemicals, adjusted for the child's age, maternal race, maternal age, and maternal
B-lll
-------
education. Limitations to the study include possible misclassification of self-reported exposures,
lack of adjustment for smoking, and recall bias. Additionally, the researchers' focus on
correcting false positives means that the study may have included false negatives. No strengths
were reported by the authors.
B.2.2.10.Pancreatic Cancer
B.2.2.10.1. Kernan et al. (1999)
Kernan, G. J.; Ji, B. T.; Dosemeci, M.; Silverman, D. T.; Balbus, J.; Zahm, S. H.
(1999). Occupational risk factors for pancreatic cancer: A case-control study based
on death certificates from 24 U.S. States. Am J Ind Med, 36, 260-270.
http://dx.doi.org/10.1002/(SICI)1097-0274(199908)36:2<260::AID-
AJIM5>3.0.CO;2-P
Summary: This study used a case-control design to examine the risk of pancreatic cancer
by occupation, industry, and exposure to solvents, including tetrachloroethylene. Cases were
identified using International Classification of Disease Code 157 (pancreatic cancer) on death
certificates in 24 states (Maine, New Hampshire, New Jersey, Rhode Island, Vermont, Indiana,
Ohio, Wisconsin, Kansas, Oklahoma, Missouri, Nebraska, Kentucky, Georgia, North Carolina,
South Carolina, Tennessee, West Virginia, Colorado, Idaho, Nevada, New Mexico, Utah, and
Washington) that also included codes for occupation and industry, based on 1980 Census codes.
Controls were chosen from among those who died of nonpancreatic, noncancer causes within the
same time frame. Each case was matched to four controls based on state, race, gender, and
5-year age group. For the study period 1984-1993, 63,097 cases and 252,386 controls were
selected. JEMs were developed by industrial hygienists for the solvents, including
tetrachloroethylene. Indexes of probability and intensity of exposure to tetrachloroethylene were
estimated and scored as "low," "medium," and "high." Overall, there were 5,344 participants
exposed to "low" levels, 2,187 exposed to "medium" levels, and 903 exposed to "high" levels of
tetrachloroethylene. Although not cited in the paper, the author's affiliation with the National
Cancer Institute and the identified solvents make it likely that the JEM was that of Gomez et al.
(1994) and Dosemeci et al. (1994).
Race and gender-specific mortality odds ratios and their corresponding 95% CIs were
estimated for intensity and probability of exposure to solvents, including tetrachloroethylene,
adjusted for age, marital status, metropolitan status, and region of residence. A strength of the
study is its use of the JEM in exposure assessment. Limitations include the possibility of
missing information related to occupation and potential confounders on death certificates, as well
as the potential for misdiagnosis of pancreatic cancer.
B-112
-------
B.2.2.10.2. Lin and Kessler (1981)
Lin, R. S. and Kessler, 1.1. (1981). A multifactorial model for pancreatic cancer in
man: Epidemiologic evidence. JAMA, 245,147-152.
http://www.ncbi.nlm.nih.gov/pubmed/7452829
Summary: This case-control study aimed to collate information of malignant neoplasms
whose prevalence was so low as to render investigations in one institution—or even one city—
largely impractical. The study was conducted in over 115 hospitals in Buffalo, Detroit, Miami,
Minneapolis-St. Paul, and New York City and collected information on 13 (adrenal, gallbladder,
kidney, liver, nasopharynx, pancreas, ureter, urethra, breast, penis, testis/scrotum, vagina, and
vulva) cancers. Cases were identified through medical records and the pathology departments of
each hospital and consisted of men and women aged 15 and over. Controls were randomly
chosen from the admissions records of cancer-free patients of the same hospital as the case and
matched to cases based on age, sex, race, and marital status. The authors do not report the
response rates for cases and controls, though they note that 22% of those eligible were not
interviewed due to the extremity of their situation. Once these individuals were excluded, the
male and female response rates were 86.2% and 86.3%, respectively. The final sample consisted
of 109 case-control pairs (67 male pairs and 42 female pairs).
In-person interviews were conducted by blinded interviewers in the hospital or at the
participant's home. The majority took place in the hospital and inquired about demographics,
residential history, occupations, toxic exposures, animal contacts, smoking habits, diet, medical
history, medications, and family history. The occupational history encompassed all jobs that
were held full-time for at least 6 months or part-time for at least 1 year. Men were also asked
about their sexual practices and urogenital conditions, and women were questioned about their
marital, obstetric, and gynecologic histories. All medical conditions that were diagnosed within
1 year of the cancer diagnosis were excluded. Duration of exposure to dry cleaning and gasoline
derivatives was categorized into 0 years, <2 years, 3-5 years, 6-10 years, and >10 years.
Overall, there were 25 (37.3%) male cases and 23 (34.3%) male controls exposed to either
dry-cleaning or gasoline derivatives.
Chi-squares and t-tesis were used to examine the differences between cases and controls.
Odds ratios were calculated to estimate the relative risk for pancreatic cancer among men and
women who were exposed to a variety of risk factors, including occupational exposure to dry
cleaning. A strength of this study is its detailed questionnaire, inquiring about part-time and full-
time jobs, as well as a variety of possible confounders. A limitation is its failure to differentiate
between occupational exposures to dry-cleaning and gasoline derivatives, as well as control
group biases weighting for diabetics, which might have obscured the observed associations with
pancreatic cancer.
B-113
-------
B.2.2.11.Renal Cell Cancer
B.2.2.11.1. Asal et al. (1988)
Asal, N. R.; Geyer, J. R.; Risser, D. R.; Lee, E. T.; Kadamani, S.; Cherng, N. (1988).
Risk factors in renal cell carcinoma. II. Medical history, occupation, multivariate
analysis, and conclusions. Cancer Detect Prev, 13, 263-279.
http://www.ncbi.nlm.nih.gov/pubmed/3266567
Summary: This population-based case-control study examined risk factors of renal cell
carcinoma. Cases were identified from 29 hospitals in Oklahoma that agreed to participate in the
study. The authors do not report details regarding where the hospitals were located, but they do
note that they included Tulsa and Oklahoma City. Cases consisted of men and women with renal
cell cancer who had a tissue or radiological examination diagnosis between 1981 and 1984.
Two sets of controls were used in this study. The first comprised hospital-based controls
matched to cases based on age, sex, race, hospital, and date of admission. The authors do not
state how they identified the hospital-based controls, and it is not known whether controls were
drawn from the same hospitals as cases, though they excluded anyone diagnosed with a kidney
disease or a psychiatric illness. The second group consisted of population-based controls from
the general Oklahoma population and were chosen through random digit dialing according to
Waksberg (1978). These controls were matched to cases based on age and sex. Of 345
identified cases, 315 (91.3%) participated in the study. Those not included in the final sample
either refused or were unable to participate or did not notify the study on time. The authors did
not provide the response or participation rates for control groups. The final sample included
315 cases, 313 hospital-based controls, and 336 population-based controls.
Interviews were conducted in the hospital with cases and hospital controls and in the
home or business with population-based controls, inquiring about medical history, medications,
radiation exposure, occupational history for all jobs held at least 1 year, self-reported industrial
exposure, tobacco smoking, beverage use, artificial sweeteners, family history of disease, height
and weight at age 20, weight most recently, and highest weight. BMI was calculated from
reported height and weights, and the occupational history was used to identify the predominant
occupation or the job held the longest out of all reported occupations lasting 1 year or more.
Employment in occupations and industries were assessed as a proxy for exposure; dry cleaning
was examined as a high-risk industry and had 11 (3.5%) cases and 7 (1.1%) controls reporting at
least 1 year of employment. Cox linear logistic regression modeling was employed to estimate
odds ratios and 95% CIs for lifetime occupations and high-risk industries. All of the
predominant lifetime occupations in men were adjusted for age, smoking, and weight. The
authors do not report lifetime occupation calculations for women. The industry estimates varied
in their adjustment of confounders. Painting and welding only adjusted for age, while chemical
B-114
-------
manufacturing, machining, petroleum refining, dry cleaning, and metal degreasing adjusted for
age, smoking, and weight. All industry calculations were stratified by gender, though only
petroleum refining and dry cleaning reported estimates for both men and women. Strengths of
the study include its use of confirmed cases of renal cell carcinoma, its population-based design,
its use of two control groups, and its adjustment for smoking. Limitations include its low-
exposure prevalence and its inability to distinguish between jobs within the dry-cleaning
industry.
B.2.2.11.2. Auperin et al. (1994)
Auperin, A.; Benhamou, S.; Ory-Paoletti, C.; Flamant, R. (1994). Occupational risk
factors for renal cell carcinoma: A case-control study. Occup Environ Med, 51, 426-
428. http://dx.doi.0rg/10.1136/oem.51.6.426
Summary: This hospital-based case-control study examined the relationship between
occupation and renal cell carcinoma in France between 1987 and 1991. Cases consisted of
138 men and 58 women with histologically confirmed renal cell carcinoma in 1 of 10 hospitals.
Two controls, one with a malignant disease and one with a nonmalignant disease (excluding
tobacco related diseases), were matched for each case based on sex, age at interview, hospital,
and interviewer. Patients with alcohol-related cirrhosis or diabetes were excluded from the
study. Eligibility and matching criteria caused some recruitment difficulties, resulting in
151 cases being matched to 2 controls and 45 cases being matched to 1 control. In total, the
study consisted of 161 controls with cancer (107 men and 54 women) and 186 controls with
nonmalignant disease (128 men and 58 women). Only one of the eligible cases and two of the
eligible controls refused to participate in the interview.
Trained interviewers used a standardized questionnaire to obtain information on
education, height, weight, smoking habits, beverage consumption, and medication, as well as a
complete occupational history. In the occupational history, participants provided their duration
of employment for each job held (minimum 1 year). Although interviewers were not blinded to
the individual's case or control status, the job history data were coded blindly, according to the
International Standard Classification of Occupations. The authors did not report the code for
launderers or dry cleaners. The numbers of exposed were not reported for launderers or dry
cleaners, though the authors noted that the estimates for laundry workers could not be calculated
due to the small numbers of exposed.
Conditional logistic regression was used to estimate odds ratios and their 95% CIs for
occupations, including launderers and dry cleaners. Analyses looked at women and men
separately, and matched odds ratios were adjusted for the matching criteria (age, hospital,
interviewer). Covariates included educational level, cigarette smoking, and the Quetelet index.
After similar results were obtained for each of the control groups, the groups were pooled into
B-115
-------
one control group. The authors do not report strengths of their study. A limitation is the small
number of exposed laundry workers.
B.2.2.11.3. Delahunt et al. (1995)
Delahunt, B.; Bethwaite, P. B.; Nacey, J. N. (1995). Occupational risk for renal cell
carcinoma. A case-control study based on the New Zealand Cancer Registry. Br J
Urol, 75, 578-582. http://dx.doi.Org/10.llll/i.1464-410X.1995.tb07410.x
Summary: This registry-based case-control study investigated the risk for renal cell
carcinoma among various occupational groups. Cases consisted of men and women aged
20 years and older who were diagnosed with renal cell carcinoma and registered in the New
Zealand Cancer Registry between 1978 and 1986. Controls were randomly selected from among
all other cancer cases during this same time frame, excluding those with a primary tumor outside
of the urinary tract, and included men and women aged 20 years or older. Cases or controls
without active occupational codes in their New Zealand Cancer Registry files were excluded
from the analysis. Of the 1,060 identified cases, 914 (86.2%) were eligible for and included in
the study. The proportion of female participants with occupational information was low
(204 cases); as such, they were excluded from the analysis. The authors did not report any
information regarding how many controls were identified for inclusion. The final sample
consisted of 710 male cases and 12,756 male controls.
All information was obtained from the New Zealand Cancer Registry, which in 1978,
began recording patients' current or most recent occupations and smoking habits. The registry
coded all occupations according to the New Zealand Standard Classification of Occupations, and
the authors did not delineate the codes used for each of the occupations they examined. All of
the occupations included in this study were determined a priori due to their previously
established or potential association with renal cell carcinoma. This included dry cleaning, which
was classified within the occupational category of services. Overall, there were a total of
52 male cases (7.3%) and 737 male controls (5.8%) whose occupation was classified as a
service, including catering/lodging, hairdressers, firefighters, and policemen, in addition to dry
cleaners. The authors did not provide the numbers of case and control dry cleaners.
The Mantel-Haenszel method was used to estimate relative risks in stratified 10-year age
groups for each occupation, including dry cleaning, and Miettinen's approximation method was
used to calculate their associated 95% CIs. All were stratified by smoking history and 10-year
age groups. A strength of this study is its use of other cancer patients in the registry for the
selection of controls, which reduces information and selection bias. Limitations include
selection bias if other cancers are associated with the selected occupations and/or their
B-116
-------
exposures, the assumption that current or most recent occupation represented lifetime
occupation, and the lack of stratification of service jobs in terms of exposure prevalence.
B.2.2.11.4. Dosemeci et al. (1999)
Dosemeci, M.; Cocco, P.; Chow, W. H. (1999). Gender differences in risk of renal
cell carcinoma and occupational exposures to chlorinated aliphatic hydrocarbons.
Am J Ind Med, 36, 54-59. http://www.ncbi.nlm.nih.gov/pubmed/10361587
Summary: This case-control study evaluated the effects of organic solvents on renal cell
carcinoma risk in Minnesota. Cases were identified through the Minnesota Cancer Surveillance
System and consisted of Caucasian men and women aged 20 to 85 years who were diagnosed
with histologically confirmed renal cell carcinoma between 1988 and 1990 (Chow et al., 1994).
Of 796 eligible cases, 690 (87% response rate) were interviewed. Two groups of controls were
elicited: the first group included Caucasian men and women between 20 and 64 years of age who
were identified through age- and gender-stratified random digit dialing; the second group
consisted of Caucasian men and women aged 65 years and older who were identified through an
age- and gender-stratified systematic sample of Health Care Financing Administration lists.
Overall, 707 (86% response rate) controls were interviewed. The final sample for the
occupational analyses consisted of 438 cases and 687 controls.
In-person interviews were conducted with blinded, trained interviewers about
demographics, diet, smoking, and drug use, as well as medical, residential, and occupational
histories. Of the 690 case interviews completed, 241 (34.9%) were proxy with next of kin. The
occupational history inquired about recent and usual job and industry, activities performed, dates
of employment, and part-time or full-time status. Duration of employment was also obtained for
13 occupations and industries, as well as 7 occupations with specific exposures. A job exposure
matrix (Gomez etal., 1994) was used to estimate exposures based on reported occupations and
industries. Occupations and industries were coded according to four digit U.S. SIC and SOC
codes, respectively. All of the four digit codes were assigned exposure estimates of probability
(i.e., "low," "medium," "high") and intensity (1, 2, 3) a priori. Intensity was defined as an
average of the concentration and frequency of exposure. Occupations were also assigned a
category. Jobs that fell within Category A, such as dry cleaner operators, had sufficient
information to be assessed for exposure independent of their industry. For jobs that fell within
Category B, the probability of exposure depended entirely on the industry, and the intensity was
weighted by both the occupation and the industry. Those in Category C had their probability and
intensity of exposure fully determined by the industry within which the job fell. Time of
employment was accounted for in the matrix through a decade indicator. Overall, 48 (11%)
cases and 76 (11%) controls were identified as potentially "ever" exposed to tetrachloroethylene.
B-117
-------
Logistic regression using the Breslow and Day (1980) method was employed to estimate relative
risks and their corresponding 95% CIs, adjusted for age, smoking, BMI, and hypertension status
and/or use of diuretics and/or antihypertension drugs. All analyses were stratified by gender and
did not include subjects with proxy respondents. The authors did not examine duration of
employment. The authors did not report any strengths of their methodology; limitations include
the small number of exposed participants, potential survival bias, and the lack of a lifetime
occupational history.
B.2.2.11.5. Harrington et al. (1989)
Harrington, J. M.; Whitby, H.; Gray, C. N.; Reid, F. J.; Aw, T. C.; Waterhouse, J.
A. (1989). Renal disease and occupational exposure to organic solvents: A case
referent approach. Br J Ind Med, 46, 643-650.
http://dx.doi.0rg/10.1136/oem.46.9.643
Summary: This case-control study conducted a detailed "blind" exposure assessment to
identify occupational risk factors for renal cancer. Cases were identified through the West
Midlands Regional Cancer Registry and consisted of living men and women in West Midlands
with histologically confirmed renal adenocarcinoma that was diagnosed between May 1984 and
April 1985. Controls were randomly selected from among the patient loads of each of the case's
general practitioners and matched (one control per case) based on 5-year age group, sex,
ethnicity, geographical location, and socioeconomic status. Of the 101 eligible renal cancer
cases, 85 (84%) were allowed to be contacted, and of these, 59 (69%) cases agreed to be
interviewed. Due to the fact that 5 of the cases were unable to have matched controls, the final
sample decreased to 54 cases and 54 controls.
In-person interviews were conducted with each participant, inquiring about personal
habits, such as smoking, coffee, and alcohol consumption; medical history; and occupational
history. Exposure was assessed blindly by an experienced chemist/occupational hygienist using
an independent checklist of exposures to solvents; exposure indices were calculated by a
computer program that multiplied the exposure level by the duration of exposure. None of the
cases or controls reported exposure to dry-cleaning fluids, but it appears that 9 (16.7%) cases and
12 (22.2%) controls reported exposure to degreasing agents. Paired analyses were conducted to
calculate odds ratios and 95% CIs in two exposure categories using Schlesselman (1982) and
three exposure categories using Pike et al. (1975). There were no strengths reported by the
authors. Limitations to this study include its small sample size (low power), low prevalence of
exposure to dry-cleaning fluids, low response rate, unaddressed renal cancer latency, and
possible recall bias associated with self-reporting.
B-118
-------
B.2.2.11.6. Mandel et al. (1995)
Mandel, J. S.; McLaughlin, J. K.; Schlehofer, B.; Mellemgaard, A.; Helmert, U.;
Lindblad, P.,... Adami, H.-O. (1995). International renal-cell cancer study. IV.
Occupation. Int J Cancer, 61, 601-605. http://dx.doi.org/10.1002/iic.2910610503
Summary: This international, multicenter case-control study evaluated factors possibly
related etiologically to renal cell cancer (Mandel et al., 1995). Six centers in five countries
(Australia, Denmark, Germany, Sweden, and United States;one center in each country, with the
exception of Germany, which had two) participated in the study. Each center had different start
dates, which were not provided. Cases were identified through population-based cancer
registries in all locations except Germany, where they were obtained through a surveillance of all
departments where renal cell cancer was diagnosed or treated. Cases consisted of men and
women aged 20 to 79 years (20-75 in Heidelberg) who were diagnosed with histologically or
cytologically confirmed renal cell adenocarcinoma between 1989 and 1991. In all centers except
in Australia and the United States, participants were required to have been born in their
respective countries. Controls were ascertained through the following: population-based
registers in Denmark and Sweden, electoral rolls in Australia, residential lists in Germany, and
either random digit dialing for American controls <65 years of age or Health Care Finance
Administration lists for American controls >65 years. All controls were matched to cases based
on gender and 5-year age group. The final sample consisted of 1,732 cases (73.2% response
rate) and 2,309 controls (74.7% response rate); cases and controls were comparable in terms of
demographics: approximately 60% were men, and 62% were over the age of 60 years at the time
of their diagnosis or interview.
In-person interviews were conducted by trained interviewers either in the hospital
(German cases) or in the participant's home (German controls and all other countries) and
inquired about tobacco, diuretics, analgesics, antihypertensive drugs, diet pills, hormones and
alcohol, height and weight, physical activity, medical and reproductive histories, family history
of cancer, demographics, and occupational history. The two centers in Germany obtained
complete occupational histories, and the four other centers asked about industries, occupations,
and exposures of interest. Occupations and industries were coded according to various
standards, including the International Labour Office (1968, 1988), the UN Department of
Economic and Social Affairs (1968, 1971, 1990), the U.S. Department of Commerce (1980), and
the U.S. Office of Management and Budget (1987). Only those occupations, industries, or
exposures that were commonly reported by all study centers were included in the analysis. The
authors do not state if the codes were harmonized or if exposure was assessed blindly. Duration
of exposure was assessed as the total number of years worked or exposed and was subsequently
divided into tertiles based on the distribution among controls. Exposures to dry-cleaning
B-119
-------
solvents were stratified into duration categories of 1-7, 8-25, and 26-60 years. Participants
were determined to be "exposed" if they had been employed in the occupation or industry or had
been exposed to the chemical of interest for at least 1 year. There were 23 (1.3%) cases and
28 (1.2%) controls who reported "ever" working in the dry-cleaning industry and 302 (17.5%)
cases and 265 (11.5%) controls who reported "ever" exposure to dry-cleaning solvents.
Logistic regression was used to estimate odds ratios and their corresponding 95% CIs
stratified by gender and adjusted for age, smoking status, BMI, education, and study center.
These estimations were performed for industry, occupation, exposure, and duration of exposure
based on the categories stated above separately, though only the results for men were presented,
as there were fewer cases among women were exposed. Tests of heterogeneity were used to
assess differences between centers. Odds ratios and their corresponding 95% CIs were also
calculated to assess the effect of education, stratified by gender and adjusted for age, smoking,
BMI and hypertension, and study center. A strength of this study is its large sample size and
standardized methodology to collect information. Limitations are its failure to verify self-
reported data and its inability to examine specific chemical agents, which was carried out by
Dosemeci et al. (1999) for a subset of this study's cases from Minnesota.
B.2.2.11.7. McCredie and Stewart (1993)
McCredie, M. and Stewart, J. H. (1993). Risk factors for kidney cancer in New
South Wales. IV. Occupation. Br J Ind Med, 50, 349-354.
http://www.ncbi.nlm.nih.gov/pubmed/8494775
Summary: This case-control study sought to report the results of a New South Wales
study examining the relationship between occupational exposure and renal cell cancer, as well as
those pertaining to cancer of the renal pelvis. Cases were identified through urologists and the
New South Wales Central Cancer Registry and consisted of men and women aged 20 to 79 years
who were diagnosed with renal cell and renal pelvis cancer between 1989 and 1990. In order to
be included in the study, cases needed to be registered in the current electoral roll, have a
telephone number that could be found, and be able to speak English. Controls were selected
through a proportional random sample of electoral rolls. Of the 744 eligible renal cell cancer
cases and 200 eligible renal pelvis cancer cases, 503 (68%) renal cell and 149 (75%) renal pelvis
cancer cases were interviewed. Of the 725 eligible controls, 535 (74%) participated in the
interview. After excluding for those who completed self-administered questionnaires, the final
sample included 489 renal cell cancer cases, 147 renal pelvis cancer cases, and 523 controls.
Interviews were conducted by a trained interviewer, and all but 10 case interviews took
place within 1 year of diagnosis. Depending on the proximity of the participant to Sydney,
interviews consisted of one of three formats: in-person (256 renal cell cases, 71 renal pelvic
B-120
-------
cases, and 232 controls), telephone (233 renal cell cases, 76 renal pelvic cases, and 291 controls),
and self-administered (14 renal cell cases, 2 renal pelvic cases, and 12 controls). The
questionnaire inquired about demographics, chemical exposures, and employment in various
occupations and industries. Occupation was assessed as a proxy for exposure, and there were
16 (3.3%) renal cell cases, 8 (5.4%) renal pelvic cases, and 7 (1.3%) controls who reported
employment in the dry-cleaning industry.
Logistic regression was used to estimate relative risks and their corresponding 95% CIs,
adjusted for age, sex, method of interview, and smoking. Renal cell cancer estimates were also
adjusted for BMI, and renal pelvic cancer estimates were further adjusted for education and
phenacetin-containing analgesics. A strength of this study is its adjustment for smoking;
limitations included the small exposure prevalence, potential recall bias due to self-reported
exposures, and the study's lack of detailed occupational information, which prevented any
assessment of intensity of exposure.
B.2.2.11.8. Mellemgaard et al. (1994)
Mellemgaard, A.; Engholm, G.; McLaughlin, J. K.; Olsen, J. H. (1994).
Occupational risk factors for renal-cell carcinoma in Denmark. Scand J Work
Environ Health, 20,160-165. http://www.ncbi.nlm.nih.gov/pubmed/7973487
Summary: This study used a population-based case-control design to examine the
relationship between employment in specific occupations and risk of renal cell carcinoma. Cases
were selected from among the Danish Cancer Registry and consisted of men and women aged 20
to 79 years who were born and living in Denmark. Controls were chosen from the Central
Population Register and matched to cases based on gender and 5-year age group. After selecting
the controls, the researchers found that they failed to account for the structure of the Central
Population Register and had obtained an inaccurate representation of certain regions. To address
this problem, the researchers randomly removed controls from the regions that had been
overrepresented and randomly selected additional controls from the regions that had been
underrepresented. Of the 482 eligible cases, 368 (76%) were interviewed. Of the 500 eligible
controls, 396 (79%) were interviewed.
In-person interviews were conducted by trained interviewers who inquired about
occupation and occupational exposure histories, as well as demographics, smoking, medical
history, and diet. Jobs were coded according to the International Standard Classification of
Occupation, and industries were coded according to the International Standard Industrial
Classification. Although dry cleaning was among those identified a priori as a high risk
industry, the authors did not provide the specific code used. Exposures were assessed for jobs
held at least 1 year and occurred at least 10 years prior to the interview. A total of 4 (1.1%)
B-121
-------
cases and 2 (0.5%) controls were employed in the dry-cleaning industry. Unconditional logistic
regression was used to estimate odds ratios and 95% CIs for men and women separately,
adjusted for age, BMI, and smoking. Strengths of the study are its population-based design and
its response rate of nearly 80% for both cases and controls. A limitation to the study is its low
number of exposed participants.
B.2.2.11.9. Schlehofer et al. (1995)
Schlehofer, B.; Heuer, C.; Blettner, M.; Niehoff, D.; Wahrendorf, J. (1995).
Occupation, smoking and demographic factors, and renal cell carcinoma in
Germany. Int J Epidemiol, 24, 51-57. http://www.ncbi.nlm.nih.gov/pubmed/7797356
Summary: This population-based case-control study examined the demographic and
occupational risk factors, as well as the risk of smoking on the development of renal cell cancer
in the Rhein-Neckar-Odenwald area of Germany. Cases consisted of German men and women
with histologically confirmed renal cell cancer between 1898 and 1991. Of the 328 cases
identified, 277 (84.5%) participated in the study. Controls were randomly selected from the
population register of the Rhein-Neckar-Odenwald area and matched to cases based on age and
gender. Of the 381 controls identified, 286 (75%) participated in the study.
In-person interviews were conducted by trained interviewers with both cases and
controls. The majority (92%) of cases was interviewed in the hospital; all of the control
interviews took place at participants' homes. Efforts were made to interview matched cases and
controls within 6 months of the case's diagnosis. A standardized questionnaire was used to
collect information on demographics, smoking history, occupational history, medical history,
family history, physical activity, weight, and diet. Occupational information was obtained on
four levels: (1) all industries in which the subject was "ever" employed, (2) occupations in which
the subject was trained, (3) activities performed during employment, and (4) exposure to specific
substances. An individual was assessed as "exposed" to an industry, occupation, or substance if
it occurred for 5 years or more. Industries were coded, and industries, occupations, and activities
were grouped into different categories. Of the 51 substances examined for possible exposure,
22 were reported by at least 5% of male subjects and subsequently analyzed. This included
chlorinated solvents, which consisted of tetrachloroethylene and tetrachlorocarbonate, and
contained a total of 27 cases (14.6%) and 12 (13%) controls. Female exposures were not
prevalent and, therefore, not examined in this study.
Unconditional logistic regression was used to calculate odds ratios and their
corresponding 95% CIs for demographics, smoking, industry/occupation, and substance
exposure separately. Demographic calculations were adjusted for age and smoking; smoking
was adjusted for age; industry and occupational groups were adjusted for age, gender, and
B-122
-------
smoking; and substance groups were adjusted for age and smoking. Limitations to the study
include its inability to independently evaluate the impact of tetrachloroethylene versus
tetrachlorocarbonate within the chlorinated solvents category and possible misclassification due
to self-reported exposure. The authors do not report any strengths of their methodology.
B-123
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
British Columbia
Band et al. (1999);
MacArthur et al.
(2009)
Band et al. (1999)
Cases: from British Columbia
Cancer Registry from
1983-1990, men, >20 yr,
histologically confirmed
cancer; Controls: from all
other cancer sites examined,
matched to cases on age, year
of diagnosis
Proxy — laundry and dry
r-lporipr
^1^/d.llVl
Site-specific cancer incidence
(prostate, lung)
Cases: prostate cancer;
Controls: excluded lung
cancer, cancer of unknown
primary site
Prostate cancer incidence
25,726 eligible
cases contacted,
15,463 (60%)
returned
questionnaire
Final sample: 1,516
cases and
4,994 controls
Serf-administered
questionnaire:
demographics, lifetime
smoking habits, alcohol
consumption,
occupational history
(lifetime job
descriptions, duration
and period of
employment,
occupation/industry
titles)
See above
Proxy respondents for
19.9%, 19.3% controls
Occupations/industries coded
according to Canadian SOC and
the Canadian SIC; launderers
and dry cleaners SIC Code 972,
SOC code not reported; assessed
"ever" and "usual" (longest
employment)
occupations/industries
7 (0.5) cases ever, 2 (0.1%) cases
usual employment in dry-
cleaning industry; control
exposures not reported
Various (see below)
Conditional logistic
regression for ORs,
95% CIs for
occupations/industries,
adjusted for education,
alcohol consumption,
smoking duration
td
I
K>
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
MacArthur et al.
(2009)
Evaluate occupational
risks for lung cancer
Case-control
td
i
to
Cases: lung cancer; Controls:
excluded unknown primary
sites
Proxy—launderers and dry
cleaners
Lung cancer (squamous cell
carcinoma, adenocarcinoma,
small cell, large cell)
5,528 eligible lung
cancer cases,
2,998 (54.2%)
returned
questionnaire
See above
10 (0.3%) cases of lung cancer in
SIC Code 972 (laundries and dry
cleaners)
Matched case-control
analyses for maximum
likelihood estimates for
ORs, 90% CIs; lung cancer
subtypes separately
evaluated
Adjustments: all lung
cancers—smoking, alcohol,
education, questionnaire
respondent; subtypes—
varied
Band et al. (2000)
Cases: from British Columbia
Cancer Registry, women
<75 yr, diagnosed with breast
cancer from 1988-1989,
Canadian citizens, residents of
British Columbia, English
speaking, no prior history of
breast cancer; Controls :
randomly selected from 1989
from British Columbia
Provincial Voters List, no
history of breast cancer before
1989, matched on age
Proxy—dry cleaning
Breast cancer incidence
1,489 eligible
cases and
1,502 eligible
controls,
1,018 (68%) cases
and 1,025 controls
returned
questionnaire
Final sample:
995 cases,
1,020 controls
Serf-administered
questionnaire: lifetime
job descriptions,
duration/period
employment,
occupation/industry
titles, demographics,
smoking, alcohol
consumption, current
body weight, weight in
late teens, age at
menarche, parity, age at
first birth, history of
breast biopsy before
1987, family history of
breast cancer, breast
feeding, birth control,
estrogen replacement
therapy
Occupations/industries coded
according to Canadian SOC and
Canadian SIC; dry cleaning:
SOC Code 6162 and SIC Code
9721; assessed "usual" and
"ever" occupation; 12 (1.2%)
cases "ever" exposed, 9 (0.9%)
cases "usual" exposure to
laundry and dry cleaning
occupation; 23 (2.3%) cases
"ever" exposed, 10 (1.0%) cases
"usual" exposure to power
laundries and/or dry-cleaners
industry; no information on
control exposure
Conditional logistic
regression for ORs and
90% CIs for each
occupation for each
estimate of exposure,
stratified by menopausal
status and "ever"/"usual"
occupation; premenopausal
adjustment: cigarette pack
years groups, breast biopsy,
family history of breast
cancer in mother/sisters;
Postmenopausal adjustment:
weights in 1986, family
history of breast cancer in
first degree relative, history
of breast biopsy for benign
breast disease, cumulative
alcohol scores; all women
combined: both pre- and
postmenopausal covariates
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Teschke et al.
td
I
K>
Cases: from British Columbia
Cancer Agency, men and
women, 19+yr, histologically
confirmed nasal
cavity/sinus/urinary bladder
cancer from 1990-1992,
exclusions: bladder cancer
cases born before 1916 and
carcinoma in situ; Controls:
British Columbia residents,
19+yr, randomly selected
from provincial voter list,
matched to cases based on age,
sex; exclusions: in prison or
mental health institution
Proxy—laundry personnel
Nasal cavity or sinus cancer,
urinary bladder cancer
54 eligible nasal
cancer cases and
195 eligible nasal
cancer controls,
48 (88.9%) cases
and 159 (81.5%)
controls
interviewed
Ineligible
bladder cancer
cases and 173
eligible bladder
cancer controls,
105 (88.2%) cases
and 139 (80.3%)
controls
interviewed
Final sample:
153 cases and
298 controls
In-person or telephone
interview by nonblinded
RN; proxy interviews if
not an English speaker,
poor memory of life
events, or deceased
Structured
questionnaire:
occupational,
residential, medical,
smoking, exposure
histories
Blinded industrial
hygienist evaluated
interviews and asked
follow-up questions
when necessary
Occupations and industries
coded according to standard
classifications, blindly grouped;
assignment based on whether
occupation or industry more
likely to determine exposure; if
both, occupation used; all
reviewed to verify accuracy; all
groups with <20 reviewed for
combination with others; In total,
57 occupational groups created
No case/control laundry
personnel for nasal cancer;
5 cases (3.3%), 4 (1.3%) controls
of laundry personnel for bladder
cancer
Exact methods for summary
ORs and 95% CIs; if
nonoccupational risk factors
found positively associated,
unconditional logistic
regression for ORs and
95% CIs, adjusted for risk
factors; all adjusted for sex,
age, smoking
Latency times: 5, 10, 15,
20 yr; only 20 yr reported
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Montreal
Siemiatyki (1991):
Aronson et al. (1996):
Parent et al. (2000)
td
I
K>
Cases: from hospitals in
Montreal, male residents,
35-70 yr, diagnosed with
histologically confirmed
cancer from 1979-1985,
97% population based case
ascertainment; Controls:
(1) population controls from
electoral lists/random digit
dialing, (2) cancer controls
from all other cases
4,576 cases
identified, 3,370
(81.5%)
participated;
response rates
varied from
78-85% for cancer
sites;
740 population
controls,
533 (72%)
participated
In-person interviews by
trained interviewers
Structured
questionnaire:
demographics,
residential history,
lifetime consumption
cigarettes, alcohol,
coffee, tea, food with
carotene, height, weight
Semi-structured
questionnaire: detailed
occupational history
Occupations/industries coded
according to Canadian
Classification and Dictionary of
Occupations/SIC; blinded
chemists, evaluated for
confidence exposure occurred,
frequency, concentration, 294
substances + 98
occupations + 77 industries =
469 circumstances
See below
Siemiatyki (1991)
See above
PCE exposure and proxy—
launderers and dry cleaners
Site-specific cancer incidence
Cancer cases:
99 esophagus,
251 stomach,
497 colon,
257 rectum,
116 pancreas,
857 lung,
449 prostate,
484 bladder,
177 kidney, 103
melanoma, and
215 lymphoma;
533 controls
See above
6 (1.2%) colon, 7 (0.8%) lung,
9 (2.0%) prostate cases ever
exposed to PCE; 4 (1.6%)
stomach, 5 (1.0%) colon,
5 (2.0%) rectum, 12 (1.4%) lung,
9 (2.0%) prostate, 10 (5.6%)
kidney, 3 (2.9%) skin melanoma,
3 (1.4%) non-Hodgkin
lymphoma cases "ever"
employed as launderers or dry
cleaners
Mantel-Haenszel for ORs,
90% CIs for
"ever"/" substantial"
exposure; all adjusted for
age, family income,
cigarette index; stomach
cancer also adjusted for
birthplace; colon/rectum
cancers also adjusted for
ethnic origin, beer index;
lung cancer also adjusted
for ethnic origin, alcohol
index, respondent; prostate
cancer also adjusted for
ethnic origin, Quetelet
index, respondent;
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Siemiatyki (1991)
(continued)
Aronson et al. (1996)
Parent et al. (2000)
Study design
Cases: prostate cancer; Cancer
controls: lung cancer excluded
PCE exposure
Prostate cancer incidence
Cases: renal cell cancer;
Controls: (1) population
controls from electoral
lists/random digit dialing,
(2) cancer controls from all
other cases
Proxy — laundry and dry
cleaners
Renal cell cancer incidence
Sample size
557 prostate cancer
cases, 449 (81%)
participated
Final sampled:
J.J.Q fQcpc
^^y \~tci j\^j^
IS^O PHTifpr
2^f^f\J V^ClllV^tl
fnntrnl Q S^ ^
^UllllUlo^ —' J J
population controls
227 eligible cases,
177 (78%)
participated,
142 renal cell
carcinoma;
1,900 cancer
controls with
78% participation
rate
Final sample:
142 cases,
1,900 cancer
controls,
533 population
controls
Data collection
See above
See above
Exposure assessment
55 (27 substances, 11 industries
and 17 occupations) of
469 occupational circumstances;
PCE exposure classified as
unexposed, nonsubstantial, or
substantial (8 subjects with
substantial exposure)
4 cases (2.8%) ever employed in
laundry /dry -cleaning industry,
<4 cases exposed >10 yr and
data not reported
Statistical approach
kidney cancer and skin
melanoma also adjusted for
ethnic origin
Unconditional logistic
regress for ORs, 95% CIs
for exposures; partially
adjusted models controlled
for age, ethnicity,
socioeconomic status,
Quetelet index, respondent
status; fully adjusted
models also controlled for
core substances with >30
exposed cases; control
groups pooled
Unconditional logistic
regression for ORs,
95% CIs for each
occupation/industry,
stratified by exposure and
duration exposure >10 yr,
adjusted for respondent
status, age, smoking, BMI
td
to
oo
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Massachusetts (Cape Cod)
Aschengrau et al.
(1998: 1993): Paulu et
al. (2002. 1999)
td
to
VO
Cases: permanent residents of
5 Upper Cape Cod towns,
diagnosed with cancer from
1983-1986, from
Massachusetts Cancer
Registry; Controls: (1) living
<65 yr—random digit dialing,
(2) living >65 yr—randomly
from Health Care Finance
Administration lists,
(3) deceased—randomly from
Massachusetts Department of
Vital Statistics file; method:
(1) cancer site stratified by
age, vital status, year of death,
gender (Aschengrau etal..
1993). (2) all controls in
stratum with 1+ case chosen;
exclusions: moved after index
year, incomplete residential
histories, no PCE data
Proxy—residence near
contaminated water
Controls:
2,236 controls
<65 yr, 249
(11.1%)
eligible/contacted,
184 (73.9%)
interviewed;
611 controls
>65 yr, 537
(87.9%)
eligible/contacted,
464 (86.4%)
interviewed;
918 deceased
controls,
794 (86.5%)
eligible/
ascertained,
723(91.1%)
interviewed via
proxy respondent
In-person (14%) and
telephone (86%)
interviews by trained
interviewers
Questionnaire: 40-year
residential history,
demographics, smoking,
medical and
occupational histories
and exposures, bottled
water consumption,
usual bathing habits
Various (see below)
Various (see below)
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Aschengrau et al.
(1993)
Study design
Cases: men/women, all ages,
diagnosed with bladder cancer,
kidney cancer, or leukemia
PCE in drinking water:
Estimated quantity delivered
toT'PQiHpTl(~'P
ItolUtllVt
Site-specific cancer incidence:
Bladder cancer, kidney cancer,
leukemia
Sample size
79 bladder cases,
72 (9 1.1%) eligible
and contacted,
63 (87.5%)
interviewed;
42 kidney cases,
36 (85.7%) eligible
and contacted,
35 (97.2%)
interviewed;
44 leukemia cases,
38 (90.5%) eligible
and contacted,
35 (92.1%)
interviewed
T^itinl ^nmnlp*
A llldl OCllllJJl^ .
61 bluHHpr putippr
\J J. LfldUU^l **tCUL**t^l
cases 852 bladder
cancer controls,
35 kidney cancer
cases 777 kidney
cancer controls,
34 leukemia cases
737 leukemia
controls
Data collection
See above
Exposure assessment
Industries/job titles coded
according to standard industrial
(1987) occupational (1990)
classifications; exposure based
on industry, job titles, percentage
reporting occupational exposure
to solvents including PCE
# (%) with any exposure to PCE
in drinking water: 34.4% bladder
cases, 26.2% bladder controls,
25. 7% kidney cases,
25.2% kidney controls,
35.3% leukemia cases,
25.3% leukemia controls
Water exposure (Relative
Delivered Dose) via Webler and
Brown (1993) algorithm, based
on leaching model by Demond
(1982); blinded assessments;
13 (2 1.3%) bladder cases,
127 (4.9%) bladder controls,
6 (17.1%) kidney cases,
112 (14.4%) kidney controls,
7 (20.6%) leukemia cases,
94 (12.8%) leukemia controls
without latency period
Statistical approach
Unadjusted OR for sites
with 2+ exposed cases;
Fisher exact test for
95%CIs. Analyses with
and without latency periods
(15 yr: bladder/kidney
cancers, 5 yr: leukemia);
stratified by bottled water,
bathing habits; multiple
logistic regression for ORs
adjusted for sex, age at
diagnosis/index year, vital
status at interview,
education, job exposures;
other confounders if in
3+ cases; maximum
likelihood estimates of
standard errors for 95% CIs
td
i
OJ
o
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Aschengrau et al.
(1998)
Study design
Cases: women of all ages,
diagnosed with incident breast
cancer
PCE in drinking water:
Estimated quantity delivered
to residence
Breast cancer incidence
Sample size
334 cases,
295 (88.3%)
eligible and
contacted. Of
these, 265 (89.8%)
were interviewed.
2,236 population
controls identified
by random digit
dialing, vital
records for
deceased controls,
and HCFA records
if >65 yr. There
were 763 controls
identified through
the two-step
control selection
process
After employing
the additional
exclusion criteria,
the final sample
consisted of
258 cases and
686 controls
Data collection
See above
Exposure assessment
Water exposure (RDD) via
Webler and Brown (1993)
algorithm, based on a leaching
model by Demond (1982):
blinded assessments
RDD categorized into low
(^SO^percentile cumulative
exposure), >50th, >75th, and
>90thpercentiles
36 (14%) exposed cases,
81 (11.8%) exposed controls
without latency period
Statistical approach
Unadjusted ORs, 95%CIs
for crude
associations/modifiers;
multiple logistic regression
for ORs adjusted for age at
diagnosis or index year,
vital status at interview,
family history of breast
cancer, age at first live birth
or stillbirth, personal
history of prior breast
cancer and benign breast
disease, occupational
exposure to solvents;
maximum likelihood
estimates of the standard
errors for 95% CIs; latency
periods of 5, 7, 9, 11, 13,
and 15 yr
td
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Paulu et al. (1999)
Study design
Cases: diagnosed with colon-
rectum cancer, lung cancer,
brain cancer, pancreatic cancer
between 1983 and 1986
PCE in drinking water:
Estimated quantity delivered
toT'PQiHpTl(~'P
ItolUtllVt
Site-specific cancer incidence
Sample size
420 colon-rectum,
326 lung, 42 brain,
and 43 pancreatic
cancer cases
selected; 366
(87.1%) colon-
rectum, 272
(83.4%) lung,
40 (95. 2%) brain,
and 39 (90.7%)
pancreatic cancer
cases were
contacted and
eligible. Of these,
326 (89.1%) colon-
rectum,
252 (92.6%) lung,
37 (88.1%) brain,
and 37 (86.1%)
pancreatic cancer
cases were
interviewed for an
overall
participation rate of
79%
Final sample:
311 colon-rectum
cancer cases,
1,158 colon-rectum
cancer controls,
243 lung cancer
cases, 1,206 lung
cancer controls,
Data collection
See above
Exposure assessment
# (%) with any exposure to PCE
in drinking water: Excluding any
latent periods: 44 (14.1%) colon-
rectum cancer cases and
153 (13.2%) controls;
33 (13.6%) lung cancer cases
and 158 (13.1%) controls;
3 (8.3%) brain cancer cases and
92 (13.1%) controls; 3 (8.3%)
pancreatic cancer cases and
81 (13.0%) controls
Statistical approach
Unadjusted ORs, 95% CIs
for brain/pancreatic cancer.
Multiple logistic regression
for ORs, 95% CIs for
colon-rectum/lung-cancer,
adjusted for age at
diagnosis or index year,
vital status at interview,
sex, occupational exposure
to PCE and other solvents.
Colon-rectum cancer
further adjusted for history
of polyps, inflammatory
bowel disease, occupational
history associated with
colon-rectum cancer. Lung
cancer further adjusted for
usual number of cigarettes
smoked and history of
cigar/pipe use, living with a
smoker, occupational
history associated with lung
cancer. Latency periods: 0,
5,7,9, 11, 13, and 15 yr
td
i
OJ
to
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Paulu et al. (1999)
(continued)
36 brain cancer
cases, 703 brain
cancer controls,
36 pancreatic
cancer cases,
622 pancreatic
cancer controls;
overall
participation rate:
79%
td
Paulu et al.
Cases: women diagnosed with
breast cancer
PCE in drinking water: GIS
analysis
Breast cancer incidence
334 cases,
295 (88.3%)
eligible/contacted,
265 (89.8%)
interviewed
Final sample:
258 cases and
686 controls
40 year residential
history during interview
included Ml addresses
and calendar years of
residence; if complete
address unknown, tax
assessors' books used to
identify
All participants blindly mapped
onto U.S. Geological Survey
map, later converted into digital
format; Upper Cape Cod area
divided into subregions in 2
ways: (1) fixed, multiscale grids,
coding each participant as
exposed or unexposed for each
grid cell, (2) overlapping circles
(adaptive k-smoothing) with
sizes based on number of nearby
cases/controls
Crude and adjusted ORs for
both grid and k-smoothed
methods, using map
choropleths for
visualization
Multiple logistic regression
for OR, adjusted for age,
parity, vital status, family
history of breast cancer in a
first-degree female relative,
age at first live birth or
stillbirth, prior history of
breast cancer or benign
breast disease
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Aschengrau et al.
(2003): Viera et al.
(2005)
Study design
Follow-up to Aschengrau
et al., 1998; permanent
residents of 8 Cape Cod
Towns from 1987-1993;
Cases: women diagnosed with
breast cancer from
1 987- 1 993 ; identified via
Massachusetts Cancer
Registry; Controls:
(1) random-digit dialing
(<64 yr); (2) random selection
from a Medicare beneficiary
roster (>65 yr), (3) random
selection from roster of
deceased residents; Controls
matched to cases based on age,
vital status
PCE in drinking water:
Estimated quantity delivered
to residence
Breast cancer incidence
Sample size
Various (see
below)
Data collection
Structured interviews:
demographics, age at
diagnosis, family history
of breast cancer,
personal history of prior
breast cancer, age at first
live birth/stillbirth,
occupational exposure
to PCE, etc., bathing
habits, bottled water,
and water filter use,
40-year residential
history; proxy
interviews with
21 1(3 1.4%) cases,
192 (3 1.2%) controls
Exposure assessment
Various (see below)
Statistical approach
Various (see below)
td
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Aschengrau et al.
(2003)
See above
td
672 cases
(81% selected and
eligible cases) and
616 controls
(157 [83%]
random-digit
dialed, 301 [76%]
of Medicare roster,
and 158 [79%]
deceased) were
included in the
analysis
See above
ROD of PCE estimated using
Webler and Brown's (1993)
algorithm, which was based on
PCE leaching model by Demond
(1982): algorithm accounted for
water flow, pipe characteristics
for each home, inputs
determined using maps;
exposure assessed; estimates
categorized as "never exposed"
(private wells) and "ever
exposed," with "ever" as low
(<50thpercentile) or high (>50th,
>75th, and >99* percentiles)
EOR and 95% CIs for crude
associations. Multiple
logistic regression for ORs,
adjusted for age at
diagnosis or index year,
vital status at interview,
family history of breast
cancer, personal history of
breast cancer, age at first
live birth or stillbirth,
occupational exposure to
PCE; maximum likelihood
estimates of standard error
were for 95% CIs
Latent periods: 0, 5, 7, 9,
11, 13, 15, 17, 19 yr
Viera et al.
See above. Also, proxy
interviews excluded from
analyses, compared with
results from total sample; Data
not collected in interviews
(inhalation rate, water flow
rate, and air exchange rate)
from literature
Full sample: 672
cases, 616 controls
Nonproxy sample:
461 cases, 424
controls
Nonproxy information
obtained via interviews:
daily number of glasses
of tap water or drinks
with tap water, bottled
water consumption,
temperature, frequency,
duration of
showers^aths
Dose model estimated FDD
(inhalation + dermal + ingestion
for each exposed residence);
inhalation: reported temperature,
frequency, duration of
baths/showers, and amount of
PCE in bath/shower air; dermal:
estimated according to Pick's
first law, height and weight data
to calculate surface area;
ingestion: volume of tap water
participant drank
Adjusted analyses limited
to those with 3+ exposed
cases and 3+ exposed
controls
Multiple logistic regression
for ORs, adjusted for age at
diagnosis/index year,
family history of breast
cancer, personal history of
breast cancer, age at first
live birth/stillbirth,
occupational exposure to
PCE;
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Viera et al. (2005)
(continued)
td
RDD reestimated for nonproxy
participants only, and both RDD
and FDD were used to classify
into nested exposure levels:
iSO^percentile, >50th, >75th, and
>90thpercentiles. Without
latency, full sample:
155 (23.1%) exposed cases,
136 (22.1%) exposed controls,
nonproxy sample: 101 (21.9%)
exposed cases, 88 (20.8%)
exposed controls
Maximum likelihood
estimates standard errors
for 95% CIs; goodness-of-
fit test compared RDD and
FDD; nonparametric rank
test evaluated whether RDD
and FDD exposures
differed significantly
Latency periods: 0, 5, 7, 9,
11,13, 15, 17, and 19 yr
New Zealand
Corbin et al. (2011):
Dryson et al. (2008):
't Mannetje et al.
(2008): McLean et al.
(2009)
Cases: From New Zealand
Cancer Registry from
2003-2004 or 2007-2008
(Corbin et al., in press), men
and women, 25-70 yr,
diagnosed with bladder cancer
or non-Hodgkin lymphoma;
Controls: randomly chosen
from 2003 electoral roll,
matched to cases on age
Proxy—occupation in textile
bleaching, dyeing and cleaning
machine operators (all
four studies, and dry cleaners
and launderers (Corbin et al.,
in press)
1,200 potential
controls,
1,100 valid
addresses,
660 contacted and
eligible,
473 interviewed for
response rate of
48%
Final sample:
471 controls
In-person interviews
with trained interviewer
(occupational health
nursing background)
Questionnaire:
demographics, smoking,
occupational history,
detailed information on
all jobs lasting >1 year.
Jobs blindly coded according to
1999 New Zealand Standard
Classification of Occupations
and Australian/New Zealand
SIC; textile bleaching, dyeing,
cleaning machine operators:
occupation Code 8264, a priori
high risk
Unconditional logistic
regression for ORs and
95% CIs for
occupations/industries
considered a priori and a
posteriori to be high risk,
adjusted for 5-year age
group, sex, smoking, Maori
ethnicity, occupational
status; Semi-Bayes
adjustments to minimize
risk of false positives due to
multiple comparisons
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Corbin et al. (2011)
Identify occupations
that may contribute to
the risk of lung cancer
in the New Zealand
population
Case-control
Dryson et al. (2008)
Study design
See above
Lung cancer incidence
See above
Bladder cancer incidence
Sample size
Of 1,057 cases, 744
eligible,
458 interviewed.
Controls identified
from 2003-2008;
Of 2,000 potential
controls,
1,878 with valid
addresses,
1,134 replied,
796 interviewed.
Excluding
ineligible, case
response rate: 53%
control response
rate 48%
Final sample:
457 cases,
792 controls
Of 358 cases,
232 eligible,
213 interviewed.
Of 1,200 controls,
660 eligible,
473 interviewed.
Excluding
ineligible, case
response rate: 64%
Final sample:
213 cases,
471 controls
Data collection
See above
See above
Exposure assessment
20 (0.2%) cases, 13 (1.6%)
controls reported employment as
bleaching, dyeing, cleaning
machine operators; 3 cases
(0.7%), 4 controls (0.5%)
identified as dry cleaner; 9 cases
(2.0%), 5 controls (0.6%)
identified as launderer
3 (1.4%) cases, 10(2.1%)
controls reported employment as
bleaching, dyeing, cleaning
machine operators
Statistical approach
See above
See above
td
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
't Mannetje et al.
(2008)
See above
Non-Hodgkin lymphoma
incidence
533 cases, 335
contacted/eligible,
291 interviewed for
response rate of
69%
Final sample:
291 cases,
471 controls
See above
5 (1.7%) cases, 10(2.1%)
controls reported employment as
bleaching, dyeing, cleaning
machine operators
See above
McLean et al. (2009)
td
oo
See above
Leukemia incidence (chronic
lymphocytic leukemia, AML,
chronic myeloid leukemia,
acute lymphoblastic leukemia,
and other forms of leukemia)
391 eligible cases,
225 (57%)
participated;
11 (3.7%) proxy
with next of kin;
988 eligible
controls,
660 contacted,
473 (48%)
participated
In-person interviews by
trained interviewers
with background in
occupational health
nursing
Questionnaire:
demographics, smoking,
detailed occupational
history
All occupations coded according
to New Zealand Standard
Classification of Occupations;
textile, bleaching, dyeing,
cleaning machine: Code 8264
designated a priori as high risk
6 (2.7%) cases and 10 (1.0%)
controls were in textile,
bleaching, dyeing, cleaning
machine occupation
Unconditional logistic
regression for ORs and
95% CIs for "ever" vs.
"never" occupation,
adjusted for age, gender,
smoking; Semi-Bayes
adjustments to assess the
impact of multiple
comparisons
McCredie et al.
(1993)
Cases: New South Wales
residents, 20 to 79 yr,
identified from hospitals and
physicians, who were
diagnosed with renal cell or
renal pelvic cancer between
1989 and 1990, needed to be
registered in the current
electoral roll, have a telephone
number, speak English;
Controls: random sample of
electoral rolls
744 eligible renal
cell cancer cases
and 200 eligible
renal pelvic cancer
cases, 503 (68%)
and 149 (75%)
interviewed,
respectively;
725 eligible
controls,
74% participated in
interview
In-person (327 cases,
232 controls), telephone
(309 cases, 291
controls), serf-
administered (16 cases,
12 controls)
Questionnaire:
occupations, industries,
chemical exposures,
demographics
Exposures quantified in textiles
based on distribution in control
group, though not provided for
PCE or the dry-cleaning industry
16 (3.3%) renal cell cases,
8 (5.4%) renal pelvis cases,
7 (1.3%) controls reported
employment in dry-cleaning
industry
Logistic regression for RRs
and 95% CIs; renal cell
cancer adjusted for age, sex,
method of interview,
smoking, BMI; renal pelvis
cancer adjusted for age, sex,
method of interview,
smoking, education,
phenacetin-containing
analgesics
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
McCredie et al.
(1993) (continued)
Study design
Proxy — dry cleaning
Renal cell carcinoma and renal
pelvic carcinoma incidence
Sample size
Final sample: 489
renal cell
carcinoma 147
renal pelvic cancer,
523 controls
Data collection
Exposure assessment
Statistical approach
Germany
Pesch et al. (2000a, b)
Pesch et al. (2000b)
5 regions in Germany; Cases
via hospitals, 1991-1995,
German men/women with
histologically confirmed
urothelial or renal cell cancer
within the 6 mo of study start;
Controls: randomly selected
via local residency registries,
matched on age, sex, region;
cases and controls required to
be German nationals; no age
limits
JEM/JTEM for PCE
Site-specific cancer incidence
See above
JEM/JTEM for PCE
Urothelial cancer incidence
Response rates:
84% cases, 71%
controls
Final sample: 1,970
cases (1,035
urothelial cancer,
935 renal cell
cancer) and
4,298 controls
1,035 urothelial
cancer cases, 4,298
controls
In-person with trained
interviewers; cases:
hospital within 6 mo
diagnosis; controls:
home
Structured
questionnaire:
demographics, lifestyle,
occupational exposures
See above
Jobs held 1+ yr coded according
to ISCO; based on self-reported
occupational history, exposure to
specific agents during tasks,
average amount of time exposed
daily
(1) Lifetime exposure: total
number of years spent at job
title; weighted sum of years
spent at task or exposed to
specific agent; (2) JEM based on
job title; (3) JTEM adjusted for
region and time; JEM/JTEM
evaluated probability, intensity
of exposure
PCE by JEM: 183 (17.7%) cases
with "medium," 188 (18.2%)
cases with "high," 74 (7.1%)
cases with "substantial"
exposure
PCE by JTEM: 37 (3.6%) cases
with "medium," 47 (4.5%) cases
with "high," and 22 cases with
"substantial" exposure
Conditional logistic
regression for ORs and
95% CIs
Adjusted for age, study
center, smoking
See above
td
i
OJ
VO
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Pesch et al. (2000a, b)
See above
JEM/JTEMforPCE
Renal cell cancer incidence
935 renal cell
cancer cases,
4,298 controls
See above
td
I
o
PCEby JEM: 166 (17.8%) cases
with "medium," 138 (14.8%)
cases with "high," 54 (5.8%)
cases with "substantial"
exposure
PCE by ITEM: 52 (5.6%) cases
with "medium" 45 (4.8%) with
"high," 18 (1.9%) with
"substantial" exposure
See above
-------
Table B-2. Summaries of characteristics of case-control studies: multiple cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Nordic Countries
Lynge et al. (2006)
td
Nested case-control study
Non-Hodgkin lymphoma,
esophageal, gastric cardia,
liver, pancreatic, cervix uteri,
kidney, and bladder incident
cancer cases in cohort of
46,768 individuals with
occupational code "laundry
and dry-cleaning worker" or
industry code "laundry and dry
cleaning" in 1970 Censuses in
Denmark, Finland, Norway,
Sweden. Cases: from 1970
(Denmark) or 1971 (Finland,
Norway, Sweden) through
1997 to 2001; ascertained
from mortality and cancer
registries. Controls: randomly
selected from cohort, matched
based on country, sex, 5-year
age group, 5-year calendar
period at the time of diagnosis
(1:3 matching except 1:6 for
esophageal cancer cases)
Proxy—dry cleaner
4,014 records—
1,616 cases, 2,398
controls, 131
subjects were both
cases and controls.
Participation rates:
57% cases,
64% controls in
Norway,
63% cases,
60% controls in
Sweden
In Denmark and
Finland, occupational
task identified on the
1970 Census form. For
subjects from Norway
and Sweden, a blinded
telephone interview was
undertaken, as 1970
Census forms were
unavailable. The
questionnaire asked
about occupational task
for job title reported on
the 1970 Census form,
and if dry cleaning,
questions sought
answers on employment
length, number of
employees, solvents
used, and personal
habits of smoking and
alcohol consumption.
Proxy interviews: 76%
cases (Norway, 72%;
Sweden, 77%), 40%
controls (Norway, 42%;
Sweden, 39%)
Occupational classification:
(1) dry-cleaners or other workers
in dry-cleaning shops with
<10 workers, assumed to have
high-exposure potential as dry
cleaners because of the shared
work tasks and physical
proximity in small dry-cleaning
shops; (2) other workers in dry-
cleaning shops; (3) unexposed
laundry workers and other
persons in dry cleaning, and
(4) unclassifiable, a category for
subjects with missing
employment information
695 cases and controls were dry
cleaners, 183 were exposed
through other work in a dry-
cleaning shop, 716 were
unclassifiable
Logistic regression for RRs
and corresponding 95%
CIs, adjusted for matching
criteria plus smoking and
alcohol use for Swedish and
Norwegian cohorts only
All analyses were
conducted at the level of the
record rather than person
because a subject may have
appeared as a case or as a
control in the study more
than once
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Bladder cancer
Burns and Swanson
(1991): Swanson and
Burns (1995)
td
I
K>
Cases: Histologically confirmed
urinary bladder cancer cases in
men and women, aged 40-84 yr,
identified from MDCSS,
1984-1991 Controls:
Histologically confirmed colon
and rectal cancer cases in men
and women, 1984-1991, aged
40-84 yr, identified from
MDCSS
Proxy—dry cleaning
Bladder cancer incidence
2,160 bladder
cancer cases
interviewed
(94%);
3,979 cancer
controls (95%)
Telephone interview
Questionnaire for lifetime
occupational history,
lifetime smoking history,
medical history,
residential history,
demographic information
Proxy respondents:
25% of cases series;
27.6% of control series
1980 U.S. Census Bureau
classification to Code 3-digit level
job title and industry; dry-
cleaning worker occupation and
dry-cleaning and laundry industry
8 cases and 14 controls identified
as dry-cleaning workers
(0.4% prevalence cases,
0.4% prevalence controls) and
15 cases and 27 controls identified
as working in dry cleaners and
laundries (0.6% prevalence cases,
0.6% prevalence controls)
Unconditional logistic
regression for ORs and
95% CI, adjusted for
cigarette smoking, race,
gender, and age at
diagnosis
Colt et al.
Cases: Primary bladder cancer
cases, diagnosed 1994-1998, in
men and women, aged 25-74 yr,
identified from New Hampshire
Cancer Registry
Controls: population controls
identified with driver's licenses,
if <65 yr age, or from state
Medicare and Medicaid roles, if
>65 yr. Controls series from
previous melanoma study
(1993-1995) with additional
controls identified using same
process for period 1995-1997
459 bladder
cancer cases
interviewed of 618
eligible cases
(74%); 665
interviews among
990 eligible
controls (67%)
In-person interview
Questionnaire for
sociodemographic
information, tobacco use,
medical history, work
history since age 15
No proxy interviews
SOC Manual used to code to 2-,
3-, and 4-digit level job title; dry-
cleaner and laundry workers,
Codes 7657, 7658
5 cases and 5 controls identified
as dry-cleaner/laundry worker.
Exposure prevalence—cases
(1%), controls (0.08%)
Unconditional logistic
regression for ORs and
95% CIs, adjusted for
5-year age group and
smoking
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Colt et al. (20041
(continued)
Proxy—dry-cleaner/laundry
worker
Bladder cancer incidence
Colt et al.
td
Cases: Primary bladder cancer
cases, diagnosed 2001-2004
(Maine, Vermont) or 2002-2004
(New Hampshire), in men and
women, aged 30-79 yr,
identified from rapid patient
ascertainment systems; Controls:
population controls identified
with driver's licenses, if <65 yr
age, or from state Medicare and
Medicaid roles, if >65 yr.
Controls: series from previous
study, and frequency matched to
cases by state, sex, and diagnosis
age or control selection
1,170 cases
(65% participation
rate), 1,418
controls
(65% participation
rate). 1,158 cases
and 1,402 controls
completed
interview
Mailed questionnaire with
follow-up, in-person visit
to administer a computer-
assisted questionnaire for
information on all jobs
held since age 16 yr,
demographic information,
tobacco use, and other
exposures. For certain
occupations, ajob-
specific questionnaire
used for information on
exposures of interest
Proxy—ever employed as textile,
apparel and furnishings machine
operator, or tender (SOC Code,
765) for >6 mo (males, 46
exposed cases, 5%; females,
27 exposed cases, 10%), of which
6 cases were laundering and dry-
cleaning machine operators and
tenders (SOC Code, 7658) (0.5%)
or in laundry, cleaning, and
garment services (SIC Code, 721),
24 exposed cases (males,
14 cases, 13%; females, 10 cases,
3%)
Each job coded blinded to case or
control status to the 1980 SOC
and the 1987 SIC scheme
Unconditional logistic
regression for males and
females separately,
adjusted for age, race,
Hispanic ethnicity, state,
smoking status, and
employment in high-risk
occupation. Other analyses
examined duration of
employment, exposure-
response using test of
linear trend, year of first
employment, and potential
for interaction between
occupation and smoking
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Gaertner et al.
td
Cases: from? Canadian province
cancer registries, men and
women, 20-74 yr, histologically
confirmed, 1974-1997
Controls: population controls
recruited through random digit
dialing (2 provinces) or
identified from provincial health
insurance plan database
(5 provinces)
Proxy—dry cleaner
Bladder cancer incidence
1,499 cases, 887
completed
questionnaire
(59% response
rate); 4,604
controls, 2,847
completed
questionnaire
(62% response
rate)
Mailed questionnaire with
telephone follow-up when
necessary; questionnaire
sought socio-demographic
information, occupation
history (up to 12
occupations), smoking,
specific agent exposures
and dietary habits
Proxy—ever employed as dry
cleaner for >1 year
Up to 12 occupations categorized
into SOC codes with employment
duration calculated from time
period reported for each
occupational activity over
subject's lifetime; dry cleaner was
a suspect occupation
Questionnaire sought information
on individual agents, if >1 yr
exposure
4 (0.7%) male and no female
cases were reported with dry
cleaner job title
Unconditional logistic
regression for males and
females, separately,
adjusted for age, province,
race, smoking, ex-smoking,
consumption of fruit, fried
food, and coffee, and ever
employed in 8 other
suspect occupations
Kogevinas et al.
(2003)
Pooled data from 11 previous
European case-control studies
from 1976-1996; cases and
controls aged 30-79 yr, cases
excluded if interview occurred
>2 yr after diagnosis; 3 studies
used population controls; 1 used
hospital/population controls;
7 used hospital controls; controls
matched to cases on 5-year age
group, geographic area.
Proxy—launderers, dry cleaners,
pressers
Bladder cancer incidence
4,101 cases in
pooled dataset,
3,346 (81.6%) met
inclusion criteria;
7,365 controls in
pooled dataset,
6,840 (92.9%) met
inclusion criteria
None reported
All data coded according to
ISCO-68 standards; launderers,
dry cleaners, pressers: Code 56
19 (0.6%) cases and 30 (0.4%)
controls were launderers, dry
cleaners, or pressers
Unconditional logistic
regression for ORs and
95% CIs, adjusted for
5-year age group, smoking,
study center; interaction
between age and study
center found significant
and included in models;
attributable risk for
occupations identified
a priori as high risk
calculated, did not include
launderers, dry cleaners,
pressers
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Reulen et al.
td
Cases: fromLimburg Cancer
Registry, men and women,
40-96 yr, diagnosed with
histologically confirmed
transitional cell carcinoma of the
bladder from 1996-2004;
Controls: Caucasian men and
women, 50+ yr, no previous
history of bladder cancer,
randomly selected from general
population of Limburg through
simple random sampling
Proxy—domestic helpers,
cleaners, launderers
Bladder cancer incidence
2,230 eligible
cases, 202(9.1%)
participated in the
study; 390
controls (response
rate: 26%)
In-person interviews by
3 trained interviewers in
homes
Structured questionnaire:
sociodemographics,
lifetime smoking, lifetime
occupational history of all
jobs lasting 6+ mo
All occupations blindly coded
according to ISCO; domestic
helpers, cleaners, and launderers:
Code 913
14 (6.9%) cases, 20(5.1%)
controls were domestic helpers,
cleaners, launderers
Unconditional logistic
regression for ORs and
95% CIs, adjusted for age,
sex, current smoking
status, years of cigarette
smoking, number of
cigarettes smoked per day,
education; interaction term
of sex and occupation also
included in model; only
occupations with 15+
participants reported
Schoenberg et al.
(1984)
Cases: men, 21-84 yr, diagnosed
with histologically confirmed
urinary bladder cancer from
1978-1979 in New Jersey, a
rapid reporting location;
Controls: (1) 21-64 yr random
digit dialed, (2) >65 yr from
stratified random sample of
Health Care Finance
Administration lists, matched to
cases on age, sex
Proxy—dry-cleaner/laundry
worker
Bladder cancer incidence
787 eligible cases,
706 (90%)
participated in
study; 1,608
eligible controls,
1,392 (87%)
participated in
study
Analysis restricted
to 658 Caucasian
male cases and
1,258 Caucasian
male controls
In-person interviews using
structured questionnaire
that sought information on
demographic, personal
and occupational history
(all jobs held >6 mo and
self-reported list of
exposures)
All occupations coded to 1970
Census Index System with 19
a priori employment categories,
including dry-cleaning and
laundering (proxy exposure)
7 cases (1.1%) and 10 controls
(0.8%) identified employment in
dry cleaning or laundry
Logistic regression for ORs
and 95% CIs, adjusted for
age and cigarette smoking
duration
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Smith et al. (1985)
Silverman et al.
(1989a; 1989b)
td
Data from National Bladder
Cancer Study (Hartge etal.
1984). Cases: men, 21-84 yr,
diagnosed with histologically
confirmed urinary bladder cancer
from 1977-1978 in 9 SEER
reporting locations and 1 rapid
reporting location; Controls:
(1) 21-64 yr random digit
dialed, (2) >65 yr from stratified
random sample of Health Care
Finance Administration lists,
matched to cases on age, sex
Proxy—laundry, dry cleaning
occupation
Bladder cancer incidence
Overall study
response rates:
75% cases,
84% controls
<65 yr, 83%
controls 65+ yr
In-person interviews by
trained interviewer within
3 mo of diagnosis
Structured questionnaire:
artificial sweeteners,
smoking, coffee
consumption, medical
history, occupational
history for all jobs >6 mo
from age 12 yr onwards
Occupations/industries coded
according to U.S. Census Bureau
indices
Various (see below)
Smith et al.
Cases: transitional or squamous
cell carcinoma of urinary bladder
Total: 7,748
# cases and
controls not
reported
See above
Exposure categories:
(1) employed >6 mo as laundry or
dry-cleaning operative;
(2) chemicals in other occupations
or industries; (3) unexposed;
Duration of exposure: total
number of years in profession;
Exposed by category: 1:
103 subjects, 2: 5,776 subjects, 3:
1,869 subjects
Logistic regression for RRs
of occupational exposure,
adjusted for age, sex, and
duration of exposure,
adjusted for age, sex,
smoking
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Silverman et al.
O989a;1989b)
Cases and controls: non-
Caucasian men
Final sample:
126 cases, 383
controls
See above
Occupations grouped by potential
to have similar exposures; dry
cleaners, ironers, pressers:
11 (8.7%) cases, 12(3.1%)
controls
td
Maximum likelihood
method for OR; Cart's
interval estimation
procedure for 95% CIs;
Mantel-Haenszel for tests
of trend; Whittemore
(1983) for PARs, 95% CIs
All adjusted for smoking;
dry cleaners adjusted for
smoking, high risk
occupation; PAR adjusted
for age, geographic area,
smoking
Steineck et al.
Cases: men, born 1911-1945
and residing in a county of
Stockholm 1985-1987, source
not identified in published paper
Controls: population controls
randomly sampled at 4 periods
from population registers
between 1985 and 1987
Proxy—dry-cleaning worker
Bladder cancer incidence
254 cases, 287
controls
Response rates:
80% cases,
79% controls
Interview method not
identified in published
paper. Structured
questionnaire for
information on
occupational history and
smoking
Self-reported occupational title
2 (0.8%) cases, 2 (0.7%) controls
were dry cleaners or worked in
dry-cleaning industry
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Zheng et al. (2002)
td
oo
Cases: histologically confirmed
incident bladder cancer cases,
1986-1989, 40-85 yr, Iowa
State Health Registry
Controls: population controls
frequency-matched (1.7:1) by
sex and age and randomly
selected from state driver's
license records for subject <65 yr
age or from HFCA records if
>65 yr age
Proxy—dry-cleaner/laundry
worker
Bladder cancer incidence
1,452 cases, 2,434
controls
Response rates:
85% cases, 82%
controls <65 yr,
and 80% for
controls >65 yr
Mailed and telephone
interviews with structured
questionnaire for
information on each job
held >5 yr, demographic
factors, residence,
smoking, past medical
history, first-degree
family history of bladder
cancer, and other potential
risk factors
Proxy respondents for
156 cases (11%) and all
controls
Self-reported occupation title and
industry coded to SIC and SOC
Proxy—laundering and dry
cleaning occupation, SOC Code
7658
Employment duration: 10 yr,
>10yr
Unconditional logistic
regression for ORs,
95% CIs and adjusted for
age, lifetime pack-years of
cigarette smoking, and
having a first-degree
relative with bladder
cancer. Other variables
such as education,
frequency of strenuous or
moderate exercise, duration
of living in a residence
served by chlorinated
surface water, population
size of places of residence,
and other cancer in a first-
degree relative did not
result in material change in
association and was not
included in final statistical
model
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Brain Cancer
Heineman et al.
(1994)
td
From death certificates in New
Jersey, Pennsylvania; cases:
Caucasian men, died from
brain/other CNS tumors from
1978-1980, exclusions: no
hospital diagnosis; Controls:
Caucasian men, died from other
causes, exclusions: death
associated with occupation,
epilepsy, suicide, homicide,
cerebrovascular diseases,
matched to cases based on age,
year of death, location
JEMforPCE
Brain cancer mortality
741 cases, 654
(88%) contacted,
483 (74%)
interviewed; 741
controls, 612
(83%) contacted,
386 (63%)
interviewed
Final sample:
300 cases, 320
controls
Blinded interviews with
next of kin by trained
interviewers
Questionnaire: brain
cancer risk factors,
occupations held from age
15 yr onward (job title,
tasks, company name and
location, industry,
products, employment
dates, hours worked)
Occupations/industries coded
according to U.S. standards; all
codes assigned a priori estimates
of probability and intensity of
exposure; JEM by Gomez et al.,
1994 to estimate exposures to
PCE
111 (37%) cases, 106 (33.1%)
controls "ever" exposed to PCE
Maximum likelihood
estimates for ORs,
95% CIs using Gart (1970):
Linear trends using Mantel
(1963)
Lag time of 10, 20 yr
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Breast cancer
Peplonska et al.
(2007)
Cases: hospital cases newly
diagnosed histologically
confirmed in situ or invasive
breast cancers residents of
Warsaw and Lodz, between
20-74 yr of age, 2000-2003;
Controls: population controls
identified from the Polish
Electronic System of Population
Evidence and matched to cases
by city of residence and age
within 5-year age groups.
Proxy — laundry, cleaning, and
garment services industry
Breast cancer incidence
2,275 cases
(79% response
rate),
2,424 controls
(66% response
rate)
In-person interview with
cases and controls by
trained interviewer
Questionnaire: known and
suspected risk factors for
breast cancer,
reproductive history,
occupations held >6 mo
Occupation/industry coded to
SIC/SOC Manuals
28 (1%) cases and 32 (1%)
controls worked in laundry, dry
cleaning, and garment services
industry
Unconditional logistic
regression for ORs and
95% CIs adjusted for age,
age at menarche, age at
menopause, number of full-
term births, breast cancer
in first degree relative,
education, and city of
residence
Colon cancer
Fredriksson et al.
(19891
All: alive at time of study,
medically able to participate;
Cases: from Swedish Cancer
Registry, men and women,
30-75 yr, diagnosed with large
bowel adenocarcinoma from
1980-1983; Controls: from
National Population Register,
matched to cases based on
county of residence, sex, age
Proxy — dry cleaning
Colon cancer incidence
402 cases, 329
contacted/eligible,
312(94.8%)
participated;
717 controls,
658 contacted/elig
ible, 623 (94.6%)
participated
Mailed questionnaire:
occupational history,
occupational exposures,
food and drinking habits,
previous diseases, drug
intake
Occupational exposures assessed
by 2 physicians and 1 hygienist as
high grade, low grade
5 (1.6%) female cases, 5 (0.8%)
female controls reported
employment in dry cleaning
Mantel-Haenszel for ORs,
Miettinen (1976) for
95% CIs for all
occupations, stratified by
age, physical activity
td
I
o
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Liver cancer
Hernberg et al.
(1988)
Houten et al. (1980)
Cases: men and women,
diagnosed with primary liver
cancer reported to the Finnish
Cancer Register in 1976-1978
and 1981; Controls: (1)
randomly selected stomach
cancer patients reported to
Finnish Cancer Register in 1977;
(2) patients whose hospital
autopsy records noted death in
1977 due to coronary infarction;
Coronary infarction controls
matched to cases based on sex,
age, and hospital of diagnosis;
matching criteria for stomach
cancer controls not reported in
paper; cases excluded if no
confirmed diagnosis
Proxy — dry cleaning
Liver cancer incidence
Cases: men and women
diagnosed with primary liver
cancer from 1956-1965;
Controls: all other cancer
patients admitted to Roswell
Park Memorial Institute from
1956-1965
Proxy — laundry and dry cleaners
Liver cancer incidence
618 eligible cases,
526 contacted,
377 (71.7%)
responded;
772 eligible
stomach cancer
controls,
654 contacted,
476 (72.8%)
responded;
674 eligible
coronary
infarction
controls,
558 contacted,
385 (69.0%)
responded
Final sample
344 cases,
861 controls
102 cases, number
of controls not
reported
Mailed questionnaire:
occupational history
(employers, work sites,
jobs held, and calendar
years of work), alcohol,
tobacco, coffee, tea,
medicines, leisure
activities, and for women,
history of oral
contraceptive use
None reported
2 occupational hygienists blindly
assessed exposure, based on
industries, workplaces, job titles;
exposures not determined.
Followed up with phone calls to
workplace or proxy respondent;
exposure classified as "heavy,"
"moderate," "light"; Dry-cleaning
exposures based on 1950 records
by Finnish Institute of
Occupational Health, which noted
PCE exposure ranged from
34-600 ppm during that time
2 cases (0.6%): possible
chlorinated hydrocarbon
exposures (laundry facility, dry-
cleaning employment); 2 controls
(0.5%): light exposure to PCE
from dry-cleaning employment
Occupation assessed as a proxy
for exposure
2 cases (2%) employed in
laundry /dry -cleaning industry
Likelihood-based ORs and
90% CIs according to
Cornfield (1956) for
association between
primary liver cancer and
solvent exposure and for
association between
primary liver cancer and
heavy /moderate alcohol
use; both stratified by sex
using Gart (1970)
Latency period 10 yr
X2 goodness-of-fit test,
where distribution of cases
compared to controls by
each industry
td
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Stemhagen et al.
(19831
td
I
K>
Cases: from NJ hospital records,
NJ State Cancer Registry, death
certificates, men and women
living in NJ diagnosed with
histologically confirmed primary
liver cancer from 1975-1980;
Controls: men and women
admitted to same hospitals as
cases and from death certificates,
matched to cases based on age,
race, sex, county of residence,
vital status, excluded if history
of liver cancer, hepatitis, liver
disease; homicide, suicide
Proxy—laundering, cleaning,
and other garment services
Liver cancer incidence and
mortality
335 eligible cases,
296 contacted,
265 (79%)
interviewed;
96% cases
deceased, so
proxy interview
with next of kin;
825 eligible
controls,
687 contacted,
530 (64.2%)
interviewed
In-person interviews
Questionnaire: lifetime
residence, smoking,
alcohol, medical history,
and employment from age
12 yr onward
Occupations/industries coded
according to Index of Industries
and Occupations Standards
developed by Bureau of Census;
Occupations 6+ mo assessed as
proxy for exposure; laundering,
cleaning, other garment services
industry: 10 (3.8%) male cases
and 8 (1.5%) male controls
Authors examined laundry/dry-
cleaning industry by individual
occupations but method not
reported; no information for
females not reported
Mantel-Haenszel methods
for ORs and 95% CIs for
men employed 6+ mo in
selected industries and
occupations; also looked at
distribution of subjects by
level of alcohol
consumption, adjusted for
age, smoking
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Suarez et al. (1989)
Study design
Death certificates from
1969-1980; Cases: men, 20+ yr,
living in Texas, liver cancer
cause of death; Controls:
randomly selected from
population-based group of
537,000 who died of all other
causes, excluding neoplasms,
liver and gallbladder diseases,
infectious hepatitis, alcoholism;
matched to cases on 5-year age
group, race, ethnicity, year of
death
Proxy — dry cleaning
occupation/industry
Liver cancer mortality
Sample size
1,771 potential
cases, 1,742
(98.4%) eligible
and included in
study; did not
report total
number of
controls
Data collection
Not applicable
Exposure assessment
Occupations grouped according to
U.S. Census Classified Index,
groupings partially based on Hoar
et al. (1980)
11 cases, 12 controls employed in
dry-cleaning industry; 4 cases,
8 controls employed as dry-
cleaning operators; unable to
calculate exposure prevalence
Statistical approach
Mantel-Haenszel for ORs;
Miettinen's method for
95% CIs; adjusted for race,
ethnicity
Nonpetrochemical
categories, occupations
within categories with 10+
participants also analyzed
Lung and upper airway cancers
Brownson et al.
(1993)
Cases: from Missouri Cancer
Registry and participating
hospitals, Caucasian females,
30-84 yr, living in Missouri,
diagnosed with primary lung
cancer from 1986-1991,
nonsmokers/selected ex-
smokers; Controls: (1) <65 yr:
state driver's licenses, (2) 65-84
yr: Medicare roster, matched to
cases on age
Proxy — dry cleaning
Lung cancer incidence
650 eligible cases,
429 (66%)
participated,
1,527 eligible
controls, 1,021
(67%)
participated
T^ITlIll QHTTinlp*
1 iilCll OClllllJlt .
479 rases 1 071
T^ ^y V^Cloti::), \- j\J Ł* \.
controls
Telephone and in-person
interviews by trained
interviewers; Telephone:
residential history,
passive smoke, personal
and family health
histories, reproductive
health history; In-person:
diet, occupation
Occupational risk factors
determined by 28 questions,
based on review of literature,
focused on job title and exposure;
subjects reported years worked at
each job/with each exposure;
30 (7.0%) cases, 39 (3.8%)
controls employed in dry-
cleaning industry
Multiple logistic
regression for ORs and
95% CIs, adjusted for age,
active smoking (for ex-
smokers), history of
previous lung disease
td
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Consonni et al.
(2010)
td
Cases from 13 hospitals in
Lombardy region of Italy, part
of EAGLE study, 35-70 yr,
2002-2005; Controls:
population controls identified
through population databases,
frequency matched by residence,
sex, and age.
Proxy—dry cleaning and
laundry occupation
Lung cancer incidence
2,100 eligible
cases,
1,943 (92.5%)
participated,
2,120 eligible
controls,
2,116(99.8%)
participated
Computer-assisted
questionnaire with in-
person interview
including lifetime history
of jobs >6 mo
Occupations/industries blindly
coded according to ISCO and
ISIC; laundry and dry cleaners
included in list of suspected
occupations/industries
14 (0.7%) cases, 14 (0.7%)
controls employed in laundry and
dry cleaning occupation
Unconditional logistic
regression for ORs and
95% CIs for "ever"
worked in either known or
suspected occupations
associated with lung
cancer, stratified by
gender, adjusted for age,
area, education, smoking
pack-years, and number of
jobs held
Pohlabeln et al.
(2000)
Cases from 12 study centers in
7 countries, subjects <75 yr
enrolled from 1988-1994;
Controls: community and
hospital-based; hospital-based
controls had diseases not related
to smoking
Proxy—Launderers and dry
cleaners
Lung cancer incidence
650 nonsmoking
cases, 1,542
nonsmoking
controls; response
rates ranged
55-95%, except
2 German centers,
1 Portuguese
center (response
rates <50%)
In-person interview:
demographics, diet,
smoking exposure,
smoking history, and
occupational history
(6+ mo duration
minimum)
Occupations industries blindly
coded according to ISCO and
ISIC; laundry and dry cleaners
included in list of suspected
occupations/industries
20 (3.1%) cases, 29(1.9%)
controls employed in laundry and
dry cleaning occupation
Unconditional logistic
regression for ORs and
95% CIs for ever worked
in either known or
suspected occupations
associated with lung
cancer, stratified by
gender, adjusted for age,
center; no effect of
different sources of
controls, so pooled results
reported
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Richiardi et al.
(2004)
td
Cases: histologically (74%) or
cytologically (26%) confirmed
lung cancers from hospitals in
Turin and East Venice, Italy,
men and women, residents <75
yr, 1990-1991 (Venice) and
1991-1992 (Turin)
Controls: population control
from local registries frequency
matched on sex and age (>1:1
frequency)
Proxy—launderers and dry
cleaners
Lung cancer incidence
1,171 lung cancer
cases, 1,569
controls; response
rates for Turin
and Venice
regions,
respectively, 86%
and 72%, and
85% and 74%
among cases and
controls
Final analyzed
sample: 1,132
cases (956 men
and 176 women)
and 1,553 controls
(1,253 men and
300 women)
In-person interview:
demographics, diet,
smoking exposure,
smoking history, and
lifetime occupational
history for all jobs >6 mo
Occupations/industries blindly
coded according to ISCO and
ISIC; laundry and dry cleaners
included in list of suspected
occupations/industries
12 (1.1%) cases, 14(0.9%)
controls employed in laundry and
dry cleaning occupation
Unconditional logistic
regression for ORs and
95% CIs for "ever"
worked in either known or
suspected occupations
associated with lung
cancer, stratified by
gender, adjusted for age,
study center, cigarette
smoking, consumption of
other tobacco products,
education, and total
number of jobs
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Vaughan et al.
(1997)
td
Cases from Fred Hutchinson
Cancer Research Center
(population-based registry), men
and women, 20-74 yr,
diagnosed with cancer of oral
cavity/pharynx, larynx,
esophagus/gastric cardia from
1983-1987, with
adenocarcinoma of
esophagus/gastric cardia from
1987-1990, residents of
Washington state; exclusions:
nonepithelial and nonspecified
cancers, no telephones at
diagnosis; Controls: random
digit dialing, matched to cases
on 5-year age group, sex
PCE exposure for dry cleaning
Upper aerodigestive tract cancer
incidence
Case response
rates: 85.2% oral
cavity, 80.8%
larynx, 82.9%
esophagus/gastric
cardia. Controls:
95.4% contacted
were screened,
80.3% eligible
were interviewed
Final sample:
1,120 cases, 724
controls
In-person interviews
Questionnaire:
demographics, tobacco,
alcohol consumption,
occupational history
(6+ mo duration,
employer, business, job
title, typical activities,
dates, solvent exposures)
Proxy interviews: 7.2%
laryngeal cases, 18.7%
oral/pharyngeal cases,
33.2% esophageal and
gastric cardia cases
Blindly assessed by estimating
probability PCE was used and 8-h
time weighted average exposure
in the job; Duration of
employment and cumulative
exposure assessed
16 (1.4%) cases, 8(1.1%)
controls "ever" employed in the
dry-cleaning industry; 15 (1.3%)
cases, 8 (1.1%) controls
"possibly" exposed to PCE, 8
(0.7%) cases, 3 (0.4%) controls
"probably" exposed to PCE
Conditional logistic
regression for ORs and
95% CIs for those
employed in dry-cleaning
industry and those exposed
to PCE, adjusted for age,
sex, education, study
period, alcohol
consumption, cigarette
smoking
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Lymphopoietic cancer
Blair et al.
td
Cases: all pathology reviewed,
Iowa: from Iowa State Health
Registry, Caucasian men
diagnosed with non-Hodgkin
lymphomafrom 1981-1983,
Minnesota: from surveillance of
hospitals, Caucasian men
diagnosed from 1980-1982;
Controls: Caucasian men
without hematopoietic or
lymphatic malignancies,
matched on state, age, year of
death, (1) <65 yr from random
digit dialing, (2) 65+ yr from
Medicare files, (3) deceased
from state vital records; farmers
excluded
Proxy—laundry and garment
workers
Incidence of non-Hodgkin
lymphoma
715 eligible cases,
622 (87.0%)
participated;
1,245 controls
participated (77%
random digit
dialing, 79%
Medicare, 77%
deceased)
Final sample: 546
cases, 1,087
controls
In-person interviews with
trained interviewers
Structured questionnaire:
sociodemographic
characteristics,
agricultural exposures,
exposures to chemicals
through hobbies,
residential history,
medical history, family
history cancer,
occupational history (all
jobs held 1+ year from 18
yr onward, industry,
employer, products
produced, job titles,
duties)
Blinded exposure assessment by
an industrial hygienist;
occupations/industries coded
according to DOT and SIC
standards; Job-exposure matrix
used to evaluate probability
(4-point scale) and intensity
(3-point scale) exposure;
laundry/garment workers: Code
721, 16 (2.9%) cases, 14 (1.3%)
controls
unconditional logistic
regression for ORs and
95% CIs, adjusted for age,
state, direct or surrogate
respondent, pesticides,
tobacco, postsecondary
education, hair dye use,
first-degree family
member with malignant
lymphoproliferative
diseases; exposure-
response relationships for
risk of non-Hodgkin
lymphoma or subtypes by
duration, intensity,
probability exposure
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Clavel et al.
td
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Fabbro-Peray et al.
(2001)
td
All participants: French men and
women, 18+ yr, living in
Languedoc-Roussillon region of
France; Cases: from 19 hospitals
and 1 cancer research center,
diagnosed with malignant
lymphomas from 1992-1995,
HIV-negative; Controls:
randomly chosen from electoral
lists in randomly selected
municipalities based on size and
population distribution and
randomly selected individuals
within each municipality; not
matched to cases
Proxy—dry-cleaning solvents
Non-Hodgkin lymphoma
incidence
627 eligible cases,
517(82.5%)
interviewed;
1,962 eligible
controls, 1,025
(52.2%)
interviewed
Final sample: 445
NHL cases, 1,025
controls
Cases: in-person
interviews at hospital;
Controls: in-person at
home or telephone
Questionnaire: general
characteristics, medical
history, occupational
history, environmental
and occupational
exposures, smoking
Age at first exposure, duration of
exposure, and cumulative index
of exposure calculated for each
chemical for each participant;
classified as either not exposed,
lower than threshold, higher than
threshold
35 (6.8%) cases, 77 (7.5%)
controls exposed to dry-cleaning
solvents
Mantel-Haenszel methods
for ORs and 95% CIs for
effect of sociodemographic
characteristics, adjusted for
age, gender; unconditional
logistic regression using
forward stepwise approach
for ORs and 95% CIs for
effect of chemical and
other exposures on
non-Hodgkin lymphoma,
adjusted for age, gender,
urban setting, education
level
Lag time 5 yr
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Gold et al. (2010a:
2010b)
td
I
o
Cases: Males and females,
35-74 yr, reported to Seattle-
Puget Sound, 2000-2002, WA,
and Detroit, MI SEER registry,
alive at time of interview;
Controls: population control
identified previously for NHL
study, random-digit dialing (for
<65 yr) and Medicare roles (65+
yr) and frequency matched to
NHL cases, resident of two
areas, 1998-2004, 35-74 yr old,
spoke English
PCE exposure
Proxy—dry cleaner or launderer
Multiple myeloma incidence
(ICO-O-2/3, 9731,9732,
plasmacytoma not otherwise
specified or multiple myeloma)
255 eligible cases,
181 participated
(71%) and 180
interviewed;
1,133 eligible
controls, 481
participated and
interviewed
(52%)
Cases and controls: in-
person interviews using a
computer-assisted
personal interview
program
Questionnaire: all jobs
held for at least 1 year
between 1941 (cases) or
1946 (controls) and
enrollment date. Job-
specific module for
solvent exposures when
participant held relevant
job for >2 yr
Occupation, industry, ever
exposed to 6 chlorinated solvents
or to individual solvent (PCE,
TCE, methylene chloride, 1,1,1-
trichloroacetic acid, chloroform,
carbon tetrachloride), exposure
duration, cumulative exposure
9 (5%) cases, 4 (0.8%) controls
with occupation as textile, apparel
and furnishing machine operator
and tender, of whom, 5 cases
(3%) and 3 controls (0.7%) were
dry cleaners; 29 cases (19%) and
63 controls (13%) "ever" exposed
to PCE, of whom 17 (3%) cases
and 15 (3%) controls with high
cumulative PCE exposure
(>7,794 ppm-hours)
Unconditional logistic
regression for ORs and
95% CIs adjusted for sex,
age, race, education, and
SEER site. Sensitivity
analysis considered all
occupations with
confidence score >1 and
repeated all analyses (Gold
etal.. 2010b) Lag time 10
vr (Gold et al.. 2010b)
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Exposure to organic solvents,
including PCE, categorized into
"high-grade" and "low-grade"
10 (5.9%) cases, 31(9.2%)
controls reported "low-grade"
exposure to organic solvents,
PCE-specific exposure not
reported
40 (23.7%) cases, 47 (13.9%)
controls reported "high-grade"
exposure to organic solvents, only
1 case (0.6%) reported exposure
to PCE
Statistical approach
Hardell et al.
td
Cases: men, 25-85 yr,
histologically confirmed
malignant lymphoma from
1974-1978. Controls: (1)
living: from National Population
Registry, exclusions: not in
same municipality at case
diagnosis, deceased, emigrated,
(2) deceased: from National
Registry for Causes of Death,
exclusions: died in 1978,
suicide, malignant tumor, date of
last employment >5 yr from
case; living matched on sex, age,
municipality; deceased matched
on sex, age, municipality, year
of death
Exposure to organic solvents,
including PCE
Hodgkin lymphoma,
non-Hodgkin lymphoma
incidence
Final sample: 169
cases (60 with
Hodgkin
lymphoma and
109 with
non-Hodgkin
lymphoma), 338
controls
Serf-administered
Questionnaire: leisure
activities, smoking/drug
use, chemical exposures,
and occupational history
(including time/place of
employment)
Blinded reviewer
telephone interviews with
participants when
information
unclear/incomplete
X tests based on Miettinen
(1970) for x2 estimates and
ORs; Miettinen (1976) for
95% CIs
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Kato et al.
td
I
K>
Cases: women, 20-79 yr, no
prior history of hematologic
cancer, living in upstate New
York, diagnosed with
non-Hodgkin lymphoma from
1995-1998; Controls: <65 yr
from age-stratified random
sample of driver's licenses, >65
yr from Health Care Finance
Administration records
Proxy—worker exposure to
degreasers/cleaning solvents or
dry-cleaning fluids
Non-Hodgkin lymphoma
incidence
722 eligible cases,
376 (56%) cases,
248 (30%) DMV
controls, 215
(67%) HCFA
controls
Blinded telephone
interviews with cases and
controls with structured
questionnaire
21% case interviews and
>3% of control interviews
were conducted with
proxy respondents
Median time between the
cancer diagnosis and the
interview was 1.2 yr and
ranged between 2 mo and
3.3 yr
Occupational: hours exposed,
year of first and last exposure,
and total number of years/months
of exposure. Cumulative
exposure hours based on hours
per time unit and total exposure
duration
50 (13.3%) cases, 48 (10.4%)
controls reported exposure to
degreasers/cleaning solvents, 7
(1.9%) cases, 8 (1.7%) controls
reported exposure to dry-cleaning
fluids
Unconditional logistic
regression for ORs and
95% CIs, for occupational,
household exposures to
solvents, adjusted for age
at index date, family
history of hematologic
cancer, college education,
surrogate status, year of
interview, BMI 10-yr
preinterview, average
frequency of use of pain-
relieving drugs, total
number of episodes of
systemic antibiotic use,
total number of household
pesticides, duration of
work involving pesticide
exposures
Lag period 1 yr
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Malone et al.
td
Cases: from SEER reporting
sites in Washington state, Utah,
Michigan, Georgia, men and
women, <80 yr, diagnosed with
chronic lymphocytic leukemia
from 1977-1981; Controls:
random digit dialing in Utah,
Michigan, Georgia, random area
sampling in Washington;
controls matched to cases based
on sex, race, and/or age,
depending on location
Chlorinated hydrocarbons
(including, but not limited to
dry-cleaning solvents) and proxy
(dry-cleaning industry)
Leukemia incidence
83% eligible
cases responded,
430 interviewed
(total eligible not
stated and unclear
if responded
equals
interviewed); Of
2,028 eligible
controls, 83%
interviewed
Final sample: 427
cases, 1,683
controls
In-person or telephone
interviews by trained
interviewers
Questionnaire: chemical
exposures, other risk
factors, employment in
petroleum, dry cleaning,
rubber, meat processing
industries
Cases: 76% in-person,
9% telephone, 16% next
of kin; Controls: 81% in-
person, 18% telephone,
<1% next of kin
Chemical exposures assessed
blindly by researchers and
toxicologist into 20 categories;
exposures with 10+ cases
analyzed; chlorinated
hydrocarbons (dry-cleaning
solvents included) assessed; 1
case reported dry-cleaning
solvent exposure
14 (3.3%) cases, 59 (3.5%)
controls reported working for 6+
mo in dry-cleaning industry
Unconditional logistic
regression for ORs and
95% CIs for all
respondents and nonproxy
respondents only; adjusted
odds ratios controlled for
race, 10-year age group,
education, sex, study site
Mester et al. (2006)
Siedler et al. (2007)
Part of EPILYMPH study; 6
regions in Germany; Cases:
from physicians, diagnosed
1998-2003, German men and
women, 18-80 yr, diagnosed
with non-Hodgkin or Hodgkin
lymphoma; Controls: from
population registration office,
matched to cases on sex, region,
age; Exclusions: subjects who
did not speak German.
Lymphoma incidence
Participation rate
among controls:
44.3%;
participation rate
among cases not
reported
Final sample: 710
cases, 710
controls
In-person with trained
interviewers
Questionnaire: lifestyle,
medical history,
occupational history
(dates of employment,
title, industry, tasks) for
each job >1 year
Various (see below)
Conditional logistic
regression for ORs and
95% CIs, adjusted for
smoking, alcohol
consumption
Unconditional logistic
regression for ORs and
95% CIs in unmatched
analysis of most frequent
lymphoma subentities,
adjusted for age, sex,
region, smoking, alcohol
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Master et al. (2006)
Seidler et al. (2007)
Study design
Proxy — launderers, dry cleaners,
pressers
Chlorinated solvent, PCE
exposure
Sample size
See above
See above
Data collection
See above
See above
Exposure assessment
Job titles/industries blindly coded
according to ISCO-68 and
Statistical Classification of
Economic Activities in the
European Community;
launderers, dry cleaners, pressers:
ISCO-68 Code 56
11 (1.5%) cases, 11(1.5%)
controls were launderers, dry
cleaners, pressers
Blinded, trained industrial
physician
Intensity: "low" (0.5-5 ppm),
"medium" (>5-50 ppm), "high"
(>50 ppm); frequency: percentage
weekly working time exposed:
"low" (1-5%), "medium"
(>5-30%), or "high" (>30%);
confidence in exposure:
"possible," "probable," "certain";
cumulative exposure: ppm-years
36 (5.1%) cases, 31(4.4%)
controls exposed to PCE
Statistical approach
All estimates stratified by
duration of employment
(<10yr, >10yr)
Latency period of 10 yr,
though data not reported
Tests for trend used
exposures as continuous
variables in logistic
reOTPCC1OTl
gltSMUll
td
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Miligi et al. (2006:
1999): Costantini et
al. (2008: 20011
td
12 sites in Italy, 11 used for
analysis; Cases: from hospitals,
medical centers, local cancer
registry, men and women, 20-74
yr, diagnosed with
hematolymphopoietic
malignancies from 1991-1993,
Controls: randomly selected
from general population in each
site. Stratified by sex and age
(5-year groups)
3,357 eligible
cases, 3,118
contacted, 2,737
(88%) responded;
2,391 eligible
controls, 2,196
contacted, 1,779
(81%) responded
Final sample:
2,737 cases and
1,779 controls
In-person interviews and
proxy with next of kin
(19% cases, 5% controls)
Questionnaire:
residential, medical,
reproductive, and
occupational histories,
behaviors, education,
solvent exposure
Industrial hygienists blindly
assessed probability ("low,"
"medium," "high") and intensity
("very low," "low," "medium,"
"high") of occupational
exposures; job-exposure matrix
created with consensus for jobs
reported most frequently
Various (see below)
Costantini et al.
(2001)
Information from 11 sites
Proxy—launderers, dry cleaners,
and pressers
Incidence: NHL, HD, leukemia,
MM stratified by sex, age
Final sample:
2,737 cases
(1,450 NHL, 365
HD, 652
leukemia, 270
MM) and 1,779
controls
See above
Jobs coded according to
International Standard
Classification of Occupations;
launderers, dry cleaners, pressers:
Code 56: 3 (0.2%) NHL, 1 (0.3%)
HD, 2 (0.3%) leukemia cases
Mantel-Haenszel method
forORs, 95% CIs,
adjusted for age; reported
results for men and
compares with total
sample; reported jobs with
5+ exposed cases
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Miligi et al. (2006)
Information from 8 sites
PCE exposure
NHL. HD incidence
td
Oi
Oi
1,719 eligible
NHL cases, 1,428
(83%) responded;
347 eligible HD
cases, 304 (88%)
responded; 2,086
eligible controls,
1,530 (73%)
responded
Final sample:
1,732 cases (285
small
lymphocytic
NHL, 100
follicular NHL,
308 diffuse NHL,
315 other NHL,
304 HD), 1,530
controls
In-person interviews with
85% NHL cases, 93% HD
cases, 97% controls;
proxy interviews with
remaining cases and
controls
Intensity of exposure to PCE
(NHL): 18 (1.3%) cases, 29
(1.9%) controls with "very
low/low"; 14 (1.0%) cases, 15
(1.0%) controls with
"medium/high"; duration of
exposure to PCE (NHL): 10
(0.7%) cases, 10 (0.7%) controls
with<15 yr; 3 (0.2%) cases, 5
(0.3%) controls with 15+ yr
ORs and 95% CIs
calculated separately for
non-Hodgkin lymphoma,
non-Hodgkin lymphoma
subtypes, and Hodgkin
lymphoma; Adjusted for
sex, age, education, area
Costantini et al.
(2008)
Information from 6 sites
PCE exposure
Leukemia subtypes, MM
586 leukemia
cases and 1,278
controls; 236
multiple myeloma
cases and 1,100
controls
Final sample: 822
cases and 2,378
controls
See above
Intensity of exposure to PCE:
"Very low/low": leukemia—6
(1.0%) cases, 17 (1.3%) controls,
MM—3 (1.3%), cases, 15 (1.4%)
controls, "Medium/high":
leukemia—7 (1.2%) cases, 12
(0.9%) controls, MM—2 (0.8%)
cases, 12 (1.1%) controls
Individual point ORs, 95%
CIs for leukemia, leukemia
subtypes, and multiple
myeloma, adjusted for
gender, age, education,
area; Linear test for trend
using duration category
midpoints
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Schenk et al.
td
Cases: men and women 20 to 74
yr and diagnosed with
non-Hodgkin lymphoma
between 1998 and 2000, living
in Iowa, California, Michigan,
or Washington state and in
SEER registry; Controls:
random digit dialing for <65 yr,
Medicare files for >65 yr
Matched based on 5-year age
group, gender, and race within
each study center
Proxy—launderers and ironers
Non-Hodgkin lymphoma
incidence
2,248 eligible
cases, 1,728
(77%) contacted,
1,321 (59%)
interviewed;
2,409 eligible
controls, 2,046
(85%) contacted
and 1,057 (44%)
interviewed
Final sample:
1,189 cases (293
follicular, 366
diffuse large
B-cell lymphoma,
487 other, 43
unknown), 982
controls
Mailed, self-administered
questionnaire: family,
medical history, diet;
computer-assisted
questionnaire in home:
demographics, hair
coloring, residential
history since 1970,
occupational history
Jobs blindly assigned
occupation/industry codes
according to standard
conventions; launderers and
ironers: occupation Code 503, 12
(1.0%) cases, 3 (0.3%) controls
Unconditional logistic
regression for ORs, 95%
CIs, adjusted for age,
gender, ethnicity, study
center; stratified by gender
and histological subtype
separately
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Scherr et al.
td
oo
Cases: men and women,
including children diagnosed
with NHL between 1980 and
1982, histologically confirmed
residents of Boston Standard
Metropolitan Statistical Area,
treated in one of nine
participating hospitals; Controls:
randomly selected from town
and precinct population lists
and, if over 17 yr of age,
matched based on sex and age,
or, if case <17 yr, one parent or
guardian matched based on age
and sex to adult resident (with
interview to determine whether
child was living in household of
same age and sex as case)
Proxy—occupation in laundry,
dry cleaner, leather products
fabrication industries;
chlorinated solvents as a
category
NHL classified using Rapapport
or Working Formulation
classification system, diffuse or
nodular tumors, or B- or T-cell
202 cases, 303
controls
Response rates,
80% cases, 72%
controls
In-person interview
Questionnaire: current or
most recent job, job held
15 yr prior, major and
second major occupation,
exposure to 10 specific
agents or chemical classes
(including chlorinated
solvents)
Proxy respondents: 33%
cases, none for controls
Occupation and industries coded
according to standard
classification, or to any of 10
specific agents
3% of cases reported employment
in laundering, dry cleaning,
leather products fabrication
industries; 24% reported exposure
to chlorinated solvents
Hierarchal approach that
aggregated histological
subtypes into groups with
similar histological
characteristics and
exposure defined as a
function of calendar time
(1901-1949, 1950-1959,
1960-1969, 1970 and
later) or exposure duration
(10 yr, 20 yr). All
exposure that showed
consistent patterns within
histological categories
over calendar time or over
duration were considered
as candidate variables for
conditional logistic models
with covariates for age and
sex
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Childhood lymphopoietic cancer
Costas et al.
td
Follow-up to a study (Cutler et
al.. 1986) that found a cluster of
leukemia cases in Woburn, MA.
Cases: pre-1982 cases from
pediatric health
professionals/pediatric oncology
centers; post-1982 cases from
Massachusetts Cancer Registry,
children, diagnosed with
leukemia <19-yr-old from
1969-1989; Controls: randomly
selected from Woburn Public
School records, matched on
race, sex, and date of birth;
excluded if not Woburn resident
at case diagnosis
Proxy—drink contaminated
water
Leukemia incidence
21 eligible cases,
19 (90.5%)
participated; 38
controls selected,
1 excluded; no
control response
rates reported
Final sample: 19
cases, 37 controls
In-person interviews with
parents of cases and
controls, except 2 fathers
via telephone
Questionnaires:
Maternal—lifestyle,
demographics, medical
history, environmental/
occupational exposures,
public drinking water at
home; (2) Paternal—
occupational
history/exposures;
Residential history for
each mother/child for 2 yr
preconception to case
diagnosis
Based on well water contaminant
levels from before 1979 closure;
determined by potential for
residence to receive water from
contaminated wells using
distribution model by Murphy
(1990); 2 exposures assessed
(cumulative and average); 16
(84.2%) cases, 24 (64.9%)
controls "ever" exposed; 7 cases
(36.8%), 13 controls (35.1%)
"most" exposure, 9 cases
(47.4%), 11 controls (29.8%)
"least" exposure
Conditional logistic
regression with
proportional hazards
model for ORs, 95% CIs;
unadjusted ORs for effect
of possible confounders;
adjusted ORs for "ever"
and "most"/"least"/"never"
exposure, adjusted with
composite covariate
controlling for
socioeconomic status,
maternal smoking during
pregnancy, maternal age at
birth of child,
breastfeeding; trends
evaluated with %2 method
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Infante-Rivard et al.
(2005)
td
I
o
Cases: children 0-14 yr,
diagnosed with acute
lymphoblastic leukemia, treated
in tertiary care centers in
Quebec Province, 1980-2000;
Controls: (1) 1980-1993 from
family stipend records; (2)
1994-2000 from health
insurance records; Exclusions:
adopted/foster, French/English
not spoken at home, not in
Canada, parents unavailable;
matched to cases on sex, age
Proxy—occupational exposure
toPCE
Leukemia incidence
848 eligible cases,
790(93.1%)
parents
interviewed; 916
eligible controls,
790 (86.2%)
parents
interviewed
Final sample: 790
cases, 790
controls
Telephone interviews
Questionnaires: (1)
structured: risk factors
and confounders,
maternal job history from
1 Syr to child's birth; (2)
semi-structured: company
activities, raw materials
or machines, goods,
responsibilities, working
conditions, coworker
activities, solvents/
chemical presence, etc;
(3) detailed tasks: time,
exposures, environment
Blind classification by chemists
and industrial hygienists; Jobs
coded according to Canadian
industrial titles (3-digit) and job
titles (7-digit); assessed exposure
to chemicals through interview
responses, geographical
information, previous knowledge
industry exposures; then
chemicals assigned codes based
on Siemiatycki (1991). PCE
Code 243; jobs 2 yr before birth,
coded separately based on
confidence that exposure
occurred, frequency and
concentration of exposure
Not enough detail to calculate
exposure prevalence
Conditional logistic
regression for ORs and
95% CIs for each
chemical, stratified by time
period, adjusted for
maternal age, education
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Lowengart et al.
(1987)
td
Cases: from Los Angeles
County Cancer Surveillance
Program, <10 yr at diagnosis
from 1980-1984, biological
mothers required availability for
interview; controls: (1) friends
of cases identified by case
mothers, (2) population-based
controls via random digit
dialing, matched to cases on age,
sex, race, Hispanic origin (if
"white"); additional exclusions
if incomplete occupational
history
PCE exposure
Leukemia incidence
216 eligible cases,
202 (94%)
contacted, 159
(79%) case
mother
interviews; 154
case father
interviews; 136
control mother
interviews, 130
control father
interviews; 30
case and 43
control paternal
interviews proxy
with mothers;
control response
rates not reported
Final sample: 123
case-control pairs
Telephone interviews by
2 nonblinded, trained
interviewers
Structured questionnaire:
family and personal
medical histories, alcohol
and tobacco use,
household and personal
products, X-ray exposure,
occupational history (job
title, industry, time period
worked); maternal
questionnaire also asked
about use of drugs,
medical complications,
diet during index
pregnancy, child's
medical history, child's
exposure to ionizing
radiation
Industries/occupations coded
according to 1970 U.S. Census
classifications, grouped based on
hydrocarbon exposure;
occupations/exposures within 1
year conception excluded; 4 case
fathers reported exposure to PCE
1 year before pregnancy (1 case),
during pregnancy (1 case), or
after delivery (2 cases); no
control fathers reported exposure
to PCE; maternal exposure to
PCE not reported
Conditional logistic
regression for ORs and
95% CIs
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Shu et al. (1999)
Study design
Cases: children <15 yr,
diagnosed with acute
lymphocytic leukemia from
1989-1993 by Children's
Cancer Group member or
institution, exclusions: no
matched control; Controls:
random digit dialing, matched to
cases on age, race, telephone
area code and exchange;
Excluded if no telephone
residence; no biological,
English-speaking mother
available
PCE exposure
Leukemia incidence
Sample size
2,081 eligible
cases and 2,597
eligible controls;
Mothers: 1,914
(92%) case
mothers and 1,987
(76.5%) control
mothers
interviewed
Final sample:
1,842 cases and
1,986 controls
Fathers: 1,801
(86.5%) case
fathers and 1,183
(69.8%) control
fathers
interviewed;
16.6% cases and
32.3% control
interviews were
proxy with
mothers
Final sample:
1,842 cases, 1,986
controls
Data collection
Telephone interviews
with parents
Structured questionnaire:
("\} IVTutpmnl
^-L^ IVAdLtl lldl
demographics, personal
"hfihit^ Tinii^pTinlH
lldL'lLOj llVLlo^/lUJlU.
exposures before/during
pregnancy, environmental
hazards exposure,
medical/family /reproducti
ve/job histories; (2)
Paternal — personal
"hfihit^ Tinii^pTinlH
lldL'lLOj llVLlo^/lUJlU.
exposures,
medical/family /j ob
histories; occupational
history: job titles,
industries duties
employment dates,
exposures
Exposure assessment
Maternal — all jobs 6+ mo from 2
yr prepregnancy through cancer
diagnosis; Paternal — all jobs 6+
mo from age 18 yr onwards; Serf-
reported exposures not on
exposure list blindly assessed by
industrial hygienist; timing of
exposure (preconception,
pregnancy, postnatal): dates of
employment; duration of
exposure: control group's median
time as cut-off; Maternal
exposures to PCE: anytime: 4
(0.2%) cases, 9 (0.5%) controls;
preconception: 3 (0.2%) cases, 2
(0.1%) controls; pregnancy: 3
(0.2%) cases, 2 (0.1%) controls;
postnatal: 4 (0.2%) cases, 8
(0.4%) controls; Paternal
exposures to PCE: anytime: 25
(1.4%) cases, 23 (1.9%) controls;
preconception: 21 (1.2%) cases,
22 (1.9%) controls; pregnancy: 8
(0.4%) cases, 14 (1.2%) controls,
postnatal: 10 (0.6%) cases, 15
(1.3%) controls
Statistical approach
Maternal exposures:
Conditional logistic
regression for ORs, 95%
CIs, adjusted for maternal
education, race, family
income; Paternal
exposures: Unconditional
logistic regression for
ORs, 95% CIs, adjusted
for paternal education,
race, family income, age,
sex of case; Tests for
trend: add categorical
variables of exposure as
continuous variables in the
models
td
I
K>
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
DeRoos et al.
td
Cases: male and female
children,
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Pancreatic cancer
Lin and Kessler
(1981)
td
>115 hospitals in Buffalo,
Detroit, Miami, Minneapolis-St.
Paul, New York City; Cases:
from medical records and
hospital pathology departments,
men and women >15 yr;
Controls: randomly chosen from
admissions records of cancer-
free patients of same hospital as
case, matched to cases based on
age, sex, race, marital status
Proxy—unspecified
occupational exposure to PCE
and gasoline derivatives
combined
Pancreatic cancer incidence
Response rates
not reported but
22% eligible not
interviewed due
to extreme illness;
once excluded,
response rates:
86.2% men and
86.3% women
Final sample: 109
case-control pairs
(67 male pairs and
42 female pairs)
Blinded, in-person
interviews in hospital
(most) or at home
Questionnaire:
demographics, residential
history, occupations (all
jobs held full-time 6+ mo
or part-time for 1+ year),
toxic exposures, animal
contacts, smoking habits,
diet, medical history,
medicines, family history;
sexual practices (men),
urogenital conditions;
marital, obstetric,
gynecologic histories
(women)
Duration of exposure to dry-
cleaning and gasoline derivatives
categorized into 0, <2, 3-5, 6-10,
>10 yr; 25 (37.3%) male cases,
23 (34.3%) male controls exposed
to either dry-cleaning or gasoline
derivatives
^2s and Mests to examine
differences between cases
and controls; ORs for
relative risk for pancreatic
cancer among men and
women exposed to a
variety of risk factors,
including occupational
exposure to dry cleaning
Kernan et al.
1984-1993; Cases: from death
certificates, all International
Classification of Disease
Code 157 in 24 states, included
occupation/industry codes based
on 1980 Census. Controls: from
death certificates, nonpancreatic,
noncancer causes, matched to
cases based on state, race,
gender, 5-year age group
PCE
Pancreatic cancer mortality
63,097 cases and
252,386 controls
were selected
Not applicable
JEM developed by industrial
hygienists for solvents,
probability and intensity of
exposure estimated and scored as
"low," "medium," or "high";
5,344 exposed to "low" levels,
2,187 exposed to "medium"
levels, and 903 exposed to "high"
levels of PCE
Race and gender-specific
mortality odds ratios for
intensity and probability of
exposure to PCE, adjusted
for age, marital status,
metropolitan status, region
of residence
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Renal cancer
Asal et al.
td
Cases: from 29 hospitals in
Oklahoma; men and women
with renal cell cancer diagnosed
from 1981-1984; Controls: (1)
hospital-based: excluded kidney
disease/psychiatric illness,
matched to cases based on age,
sex, race, hospital, date of
admission, (2) population-based:
random digit dialing, matched to
cases based on sex, age
Proxy—dry cleaning
Renal cell cancer incidence
345 identified
cases, 315
(91.3%)
participated;
control
response/participa
tion rates not
reported
Final sample: 315
cases, 313
hospital-based
controls, 336
population-based
controls
In-person interviews in
hospital with cases and
hospital-based controls,
home with population-
based controls
Questionnaire: medical
history, medications,
radiation exposure,
occupational history for
all jobs held >1 year,
industrial exposures,
tobacco smoking,
beverage use, artificial
sweeteners, family history
disease, height, weight
Dry cleaning examined as high-
risk industry: 11 cases (3.5%), 7
controls (1.1%)
Cox linear logistic
regression for ORs, 95%
CIs for lifetime
occupations and high-risk
industries, adjusted for
age, smoking, weight
Auperin et al.
Cases: men and women with
histologically confirmed renal
cell carcinoma identified at 10
selected hospitals; Controls: (1)
with a malignant disease, (2)
with a nonmalignant disease,
excluded tobacco-related
diseases, matched for each case
based on sex, age at interview,
hospital, and interview. Patients
with alcohol-related cirrhosis or
diabetes excluded from study.
Proxy—laundry workers
Renal cell carcinoma incidence
151 cases
matched to two
controls, 45 cases
matched to 1
control; 161
controls with
cancer and 186
with
nonmalignant
disease
Unblinded, trained
interviewers
Questionnaire: education,
height, weight, smoking
habits, beverage
consumption, and
medication, complete
occupational history,
including duration of
employment for each job
held. Interviewers were
not blinded to the case or
control status
Blinded exposure assessment;
coded according to International
Standard Classification of
Occupations; minimum of 1 year
employment for each job held;
numbers of exposed laundry
workers not reported
Conditional logistic
regression for ORs, 95%
CIs, stratified by gender;
pooled control group;
adjusted for age, hospital,
interview, educational
level, cigarette smoking,
and the Quetelet index
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Delahunt et al.
(1995)
td
Cases and controls: men and
women in New Zealand Cancer
Registry from 1978-1986;
Cases: men and women, 20 yr
and older, diagnosed with
malignant neoplasm of kidney;
Controls: men and women, 20 yr
or older, diagnosed with primary
tumor outside of urinary tract;
women subsequently excluded
Proxy—dry cleaning
Renal cell carcinoma incidence
1,060 eligible
cases, 914
(86.2%)
information on
occupations
Final sample: 710
cases, 12,756
controls
New Zealand Cancer
Registry for current/most
recent occupations,
smoking habits
Occupations coded according to
New Zealand Standard
Classification of Occupations,
selected a priori
52 (7.3%) cases and 737 (5.8%)
controls in service occupation,
including but not limited to dry
cleaners
Mantel-Haenszel method
for RRs for each
occupation; Miettinen's
approximation method for
95% CIs; all stratified by
smoking history, 10-year
age groups
Dosemeci et al.
(1999)
Cases: from Minnesota Cancer
Surveillance System, Caucasian
men and women, 20-85 yr,
diagnosed with histologically
confirmed renal cell carcinoma
from 1988-1990; Controls:
Caucasian men and women, (1)
20-64 yr from random digit
dialing, (2) >65 yr from
systematic sample of Health
Care Financing Administration
lists
PCE
Renal cell carcinoma incidence
796 eligible cases,
690 (87%
response rate)
interviewed, 241
(34.9%) proxy
with next of kin;
707 (86%
response rate)
controls
interviewed
Final sample: 438
cases, 687
controls
In-person interviews with
blinded, trained
interviewers
Questionnaire:
demographics, diet,
smoking, drug use,
medical/residential
histories, occupational
histories recent and usual
job and industry,
activities, employment
dates, part-time/full-time
status. Duration of
employment for specific
industries, occupations,
exposures
Occupations/industries coded
according to SOC and SIC; linked
with JEM by Gomez et al., 1994;
11% cases, 11% controls exposed
to PCE
Logistic regression using
Breslow and Day (1980)
method for RRs and 95%
CIs
Adjusted for age, smoking,
BMI, and hypertension
status, and/or use of
diuretics and/or
antihypertension drugs
All stratified by gender
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Harrington et al.
(1989)
td
Cases: from West Midlands
Regional Cancer Registry, men
and women, living in West
Midlands, diagnosed with renal
adenocarcinoma from
1984-1985; Controls: randomly
selected from patients of general
practitioners, matched based on
5-year age group, sex, ethnicity,
geographical location,
socioeconomic status; excluded
if no matching controls
Proxy—dry-cleaning fluids,
degreasing agents
Renal cancer incidence
101 eligible renal
cancer cases, 85
(84%) contacted,
59 (69%)
interviewed
Final sample:54
cases, 54 controls
In-person interviews
Questionnaire: personal
habits (smoking, coffee,
and alcohol
consumption), medical
history, occupational
history
Exposure assessed blindly by
chemist/occupational hygienist
using checklist of exposures;
exposure indices calculated by
computer program (exposure
level x duration exposure)
No cases or controls reported
exposure to dry-cleaning fluids; 9
(16.7%) cases, 12 (22.2%)
controls reported exposure to
degreasing agents
Paired analyses for ORs
and 95% CIs for 2
exposure categories using
Schlesselman (1982) and 3
exposure categories using
Pike et al. (1975)
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Mandel et al. (1995)
td
oo
6 centers: Australia, Denmark,
Germany (2 centers), Sweden,
United States; Cases: from
population-based cancer
registries except Germany
(surveillance of
diagnosis/treatment
departments), men and women,
20-79 yr, diagnosed with
histologically or cytologically
confirmed renal cell
adenocarcinoma from
1989-1991; required in-country
birth (except Australia/United
States); Controls:
Denmark/Sweden: population-
based registers, Australia:
electoral rolls, Germany:
residential lists, United States:
Health Care Finance
Administration lists and random
digit dialing, matched on gender,
5-year age group
Proxy—dry-cleaning industry,
dry-cleaning solvents
Renal cell cancer incidence
Final sample:
1,732 cases
(73.2% response
rate) and 2,309
controls (74.7%
response rate)
In-person interviews by
trained interviewers in
hospital (German cases)
and homes (German
controls/all others)
Questionnaire: tobacco,
diuretics, analgesics,
antihypertensive drugs,
diet pills, hormones, and
alcohol, height and
weight, physical activity,
medical/reproductive
histories, family history
cancer, demographics,
occupational history
Industries/occupations coded
according to International Labour
Office (1968, 1988), UN
Department of Economic/Social
Affairs (1968, 1971, 1990), U.S.
Department of Commerce (1980),
U.S. Office of Management
Budget (1987); Duration: total
number years worked/exposed;
tertiles based on control
distribution; dry-cleaning solvents
duration: 1-7, 8-25, 26-60 yr; 23
(1.3%) cases, 28 (1.2%) controls
in dry-cleaning industry; 302
(17.5%) cases and 265 (11.5%)
controls exposed to dry-cleaning
solvents
Logistic regression for
ORs and 95% CIs for
industry, occupation,
exposure, stratified by
gender, adjusted for age,
smoking status, BMI,
education, study center;
only men reported
Only industries,
occupations, exposures
reported by all centers
analyzed; tests of
heterogeneity to assess
differences between
centers
-------
Table B-3. Summaries of characteristics of case-control studies: single cancer-site studies (continued)
Reference,
objective/hypothesis,
study type
Study design
Sample size
Data collection
Exposure assessment
Statistical approach
Mellemgaard et al.
(1994)
Cases: from Danish Cancer
Registry, women and men
20-79 yr, born and living in
Denmark; Controls: from
Central Population Register;
matched to cases on gender,
5-year age group
Proxy—dry cleaning
Renal cell carcinoma incidence
482 eligible cases,
368 (76%)
interviewed; 500
eligible controls,
396 (79%)
interviewed
In-person interviews with
trained interviewers
Questionnaire asked
about occupation,
occupational exposures,
medical history, diet,
smoking, demographics
Occupations and industries coded
according to ISCO and ISIC; dry
cleaning identified a priori as
high risk (code not provided);
exposures assessed for jobs held
1+ year and occurred 10+ yr prior
to interview
Odds ratios and 95% CIs
were calculated for men
and women separately
Adjusted for age, BMI,
and smoking
td
Schlehofer et al.
(1995)
Cases: German men and women,
histologically confirmed renal
cell cancer, from 1988-1991;
Controls: randomly selected
from population register of the
Rhein-Neckar-Odenwald area
Matched on age, gender
Chlorinated solvents
Renal cell cancer incidence
Of the 328 cases
identified, 277
(84.5%)
participated in the
study
Of the 381
controls
identified, 286
(75%)
participated in the
study
In-person interviews by
trained interviewers; 92%
case interviews in
hospital; 100% control
interviews at home
Questionnaire: medical,
smoking family, weight,
diet, demographics,
physical activity,
occupational history
(industry, occupation,
activities, chemical
exposures)
Industries coded; industries,
occupations, activities grouped
into different categories; 51
possible substances, 22 reported
by >5% male subjects and
analyzed, including chlorinated
solvents (PCE and
tetrachlorocarbonate); exposed if
5+ yr duration
27 (14.6%) male cases, 12 (13%)
male controls exposed to
chlorinated solvents; female
exposures not examined
Unconditional logistic
regression for ORs and
95% CIs for smoking, age,
and sex.
HD = Hodgkin lymphoma.
-------
B.3. Geographically Based And Other Studies
The following papers examined tetrachloroethylene using geographically based
(ecological) and other study designs. Summaries of the study characteristics of each paper are
provided in Table B-4.
B.3.1. Cohn et al. (1994)
Cohn, P.; Klotz, J.; Bove, F.; Berkowitz, M.; Fagliano, J. (1994). Drinking water
contamination and the incidence of leukemia and non-Hodgkin's lymphoma.
Environ Health Perspect, 102, 556-561.
http://www.ncbi.nlm.nih.gov/pubmed/9679115
Summary: This ecological incidence rate study examined the following four hypotheses:
(1) the incidence of leukemia is associated with exposure to trichloroethylene and/or
tetrachloroethylene; (2) childhood leukemia is associated with trichloroethylene and/or
tetrachloroethylene; (3) non-Hodgkin lymphoma is associated with trichloroethylene and/or
tetrachloroethylene; and (4) gender may be an effect modifier. Cases were identified through the
New Jersey State Cancer Registry and consisted of men and women residing in 1 of
75 municipalities diagnosed with primary leukemia (acute lymphocytic, chronic lymphocytic,
acute myelogenous, chronic myelogenous, other specified, and unspecified) or non-Hodgkin
lymphoma (low-grade, intermediate-grade, intermediate-grade/diffuse large cell/reticulosarcoma,
high-grade, and high-grade NHL/non-Burkitt's) between 1979 and 1987. Information was
supplemented with death certificates; any cases that were determined exclusively through death
certificates were excluded. Municipalities were chosen on the basis of their water supply. Only
those where at least 80% of the population received their water from a public water supply were
selected. In total, 1,190 cases of leukemia (118 acute lymphocytic, 354 chronic lymphocytic,
276 acute myelogenous, 146 chronic myelogenous, 61 other specified, 235 unspecified) and
1,658 cases of non-Hodgkin lymphoma (434 low-grade, 708 intermediate-grade,
402 intermediate-grade/diffuse large cell/reticulosarcoma, 69 high-grade, 51 high-
grade/non-Burkitt's) were included in the study.
Tetrachloroethylene exposure potential was based on water monitoring data—averages
from 1984-1985 by the New Jersey Department of Environmental Protection and Energy.
Although the authors do not explicitly state these were the same municipalities, due to the
mandatory nature of the monitoring data, this may be assumed. Samples were taken from water
treatment plants as well as tap sites within the distribution system. Tetrachloroethylene
exposures were categorized as <0.1, 0.1-5, and >5 ppb, based on EPA standards. Surveys
conducted by the New Jersey Department of Environmental Protection and Energy and the
B-180
-------
Department of Health between 1978 and 1983 were used to provide corroborating evidence for
use of the 1984-1985 estimates. The authors do not state if these were the same municipalities.
The latter data were used as the primary source of exposure estimation because the earlier
surveys were conducted primarily in response to known contamination, and the quality assurance
and quality control of the latter mandatory monitoring were reported to be better. The
1978-1983 data are further described elsewhere (Cohnetal.. 1994).
There were 440 cases of leukemia (37%) and 662 cases of non-Hodgkin lymphoma
(39.9%) that were assessed as exposed due to their residence in a municipality with
tetrachloroethylene levels between 0.1 and over 5 ppb. The remainder were determined to have
<0.1-ppb exposure. Log-linear regression models with the Poisson distribution were used to
estimate incidence rate ratios and their corresponding 95% CIs, adjusted for age and stratified by
sex with subjects, identified <0.1 ppb as referents. The rate ratios were also stratified by sex and
type of leukemia or non-Hodgkin lymphoma. The authors do not report strengths of their
methodology. Limitations included the lack of adjustment for possible confounders, potential
misclassification of exposure that biases the result away from the null, and the ecological study
design, which measured exposure and disease at the same time, assigned exposure and leukemia
incidence at the municipal level, and lacked information on individual exposure potential.
B.3.2. Lee et al. (2003)
Lee, L.; Chung, C.; Ma, Y.; Wang, G.; Chen, P.; Hwang, Y.; Wang, J. (2003).
Increased mortality odds ratio of male liver cancer in a community contaminated by
chlorinated hydrocarbons in groundwater. Occup Environ Med, 60, 364-369.
http://dx.doi.0rg/10.1136/oem.60.5.364
Summary: Exposure potential to chlorinated hydrocarbons was assigned in this
community case-control study of liver cancer in males >30 years of age using residency as coded
on death certificates obtained from local household registration offices. No information is
available to assess the completeness of death reporting to the local registration office. Of the
1,333 deaths between 1966 and 1997 in two villages surrounding a hazardous waste site, an
electronics factory operating between 1970 and 1992 in Taoyuan, Taiwan, 1,266 cancer deaths
were identified; 53 liver cancer deaths, 39 stomach cancer deaths, 26 colorectal deaths, and
41 lung cancer deaths. Controls were identified from 344 deaths due to cardiovascular and
cerebrovascular diseases, without arrhythmia; 286 were included in the statistical analysis.
Residents from a village north and northeast of the plant were considered exposed and residents
living south considered unexposed to chlorinated hydrocarbons. Additionally, death certificates
were obtained from the registration offices in two villages near the factory. These records
contained information on gender, age, date of birth and death, address, and cause of death. The
B-181
-------
underlying cause of death was then blindly assessed by a team of nosologists. Cases consisted of
any individual whose cancer was coded by the nosologists as either an underlying cause of death
or as a significant condition. All individuals were also linked to the Taiwan National Cancer
Registry to verify the accuracy of coded cancer cases.
Residence was assessed as a proxy for exposure, with exposed cases considered exposed
if living downstream of the factory and unexposed if living upstream of the factory.
Geographical exposure was confirmed through well water sampling. Between 1999 and 2000,
74 groundwater samples were collected from off-site residential wells near a factory whose soil
and groundwater had been previously found to be contaminated with chlorinated hydrocarbons,
including tetrachloroethylene. Overall, 20 (45.5%) of the wells downstream had concentrations
above the maximum contaminant level for tetrachloroethylene. No upstream wells were found to
be contaminated with tetrachloroethylene. Death certificates were also linked to the Labour
Insurance Bureau to ascertain those who had previously worked in the factory.
The Mantel-Haenszel method was used to estimate mortality odds ratios and their
corresponding 95% CIs, adjusted for age. Multiple logistic regression was also conducted,
adjusted for age and time period. The Cochran-Armitage test for trend calculated the effect of
time period for downstream and upstream villages. A latency period of 10 years was also
included.
One strength of this study is its linkage with the National Cancer Registry, which located
an additional 12 cancer cases that had not been recorded in the death certificates. Limitations to
the study include possible selection bias due to the use of only cardiovascular and
cerebrovascular deaths as controls, possible misclassification due to the use of residence as a
proxy for exposure status, and the lack of control for potential confounders such as Hepatitis C
virus, which was of high prevalence in this area.
B.3.3. Ma et al. (2009)
Ma, J.; Lessner, L.; Schreiber, J.; Carpenter, D. O. (2009). Association between
residential proximity to PERC dry cleaning establishments and kidney cancer in
New York City. J Environ Public Health, 2009,183920.
http://dx.doi.org/10.1155/2009/183920
Summary: The hypothesis tested in this study was living in an area with a high density of
tetrachloroethylene dry cleaners increases tetrachloroethylene exposure and the risk of kidney
cancer. Subjects were individuals 45 years of age or older with a principal or other diagnosis of
kidney cancer and identified from a New York State register of hospital discharges between 1993
and 2004. The database used to identify subjects did not include personal identifiers leading to
an inability to distinguish multiple hospital discharges by a single individual. A subject had the
B-182
-------
potential for multiple entries for this reason. The inclusion criteria were restricted to subjects
whose residence at the time of discharge was in a New York City zip code having a median
household income from $17,864 to $142,926. No information is provided in the paper on how
temporal a change as zip code's median income may have affected the inclusion criteria. Of the
total of 181 zip codes in New York City, 164 zip codes met the inclusion criteria; six zip codes
were not considered because population or income information was unavailable, and 10 zip
codes had median household incomes either below or above the inclusion criteria. A total of
674,519 discharges with a diagnosis of cancer, of which 10,916 were of kidney cancer, were
identified with a residence at the time of discharge within the 164 eligible zip codes. Population
estimates, year not identified by authorsm, by zip code were derived from U.S. Census data and
stratified by age, race, and sex.
Dry-cleaning establishments were identified from a listing maintained by the New York
State Department of Environmental Conservation. This listing included dry-cleaning
establishments who were required under state statutes to report their usage of
tetrachloroethylene. The authors do not provide information as to whether the statute identified a
minimal usage level that would lead to an underreporting of the number of dry-cleaning
establishments. The density of dry cleaners by zip code (number of dry cleaners per km2) was
estimated for each zip code from the number of dry-cleaning establishment and the population
density, based upon a zip code's population estimate and area. The authors do not provide
information in the paper for the source for estimating area of individual zip codes.
A negative binomial model was fit to the data to examine the rate of discharge rate for a
principal or other diagnosis for kidney cancer as functions of the densities of dry-cleaning
businesses. For each of the exposure strata, the authors examined different variables as possible
effect modifiers, and these included median household, age, and sex, with a finding that effect
modifiers differed for each exposure strata. The authors also used a Poisson regression model
but did not report findings because of an inadequate fit to the observed data.
This study is ecological in design, with associated limitations known as "ecological
fallacy" because variables of exposure and outcome measured on an aggregate level do not
represent association at the individual level. A significant shortcoming of this study is the
potential for a subject to have multiple discharges, inflating the numerator, but not the
denominator, for estimating the discharge rate, and its use of a crude exposure surrogate. The
authors did not validate how well the density of dry-cleaning businesses predicted atmospheric
concentrations of tetrachloroethylene for individual zip codes or potential exposure to individual
subjects. The authors noted New York City zip code densities varied by boroughs, particularly
Staten Island, which had large areas but lower population densities and lead to large variation in
the exposure surrogate. Furthermore, risk ratios from the negative binomial model are difficult
B-183
-------
to interpret because each exposure strata's rate ratio was based upon a different set of covariates.
On the other hand, the study was able to characterize disease distribution geographically, as well
as dry-cleaning business location.
B.3.4. Mallin (1990)
Mallin, K. (1990). Investigation of a bladder cancer cluster in northwestern Illinois.
Am J Epidemiol, 132, S96-106. http://www.ncbi.nlm.nih.gov/pubmed/2356842
Summary: This ecological study examined the incidence of bladder cancer to determine if
the high mortality rates also reflect high incidence rates, and if high incidence rates were found
in areas of known groundwater contamination. Incident cancer cases were identified through
medical records in 8 of the 9 counties within a region of northwestern Illinois. Resident cases
diagnosed or treated in the bordering states of Iowa or Wisconsin were also ascertained through
the Iowa State Health Registry and the Wisconsin Cancer Reporting System. Cases consisted of
men and women who were diagnosed with histologically confirmed bladder cancer between
1978 and 1985. Deaths due to bladder cancer during this time period were also ascertained.
Incidence data were stratified by county and zip code. After significantly higher risks of bladder
cancer incidence and mortality were found only in Winnebago County, the researchers searched
Illinois EPA and Department of Energy and Natural Resources documents and found well water
in this county was contaminated with tetrachloroethylene and other compounds.
Indirect standardization was used for the estimation of standardized incidence ratios and
SMRs, both adjusted for age. Expected numbers of cases were derived from age-specific rates
for 1978-1991 and 1982-1985. Their corresponding 95% CIs were calculated according to
Miettinen's exact limits; significance was evaluated using a ^ distribution. Where the expected
counts were less than 5, Fisher exact test limits were used for the estimation of 95% CIs, and a
Poisson distribution was assumed for the significance tests. The authors did not report any
strengths of their study. Limitations include the lack of control for potential confounders,
ecological design causing chance associations, lack of survival data, lack of medical treatment
data, and no data on water consumption among Winnebago inhabitants.
B.3.5. Morton and Marjanovic (1984)
Morton, W. and Marjanovic, D. (1984). Leukemia incidence by occupation in the
Portland-Vancouver metropolitan area. Am J Ind Med, 6,185-205.
http://www.ncbi.nlm.nih.gov/pubmed/6475965
Summary: An examination of occupational risks using incident leukemia cases identified
over a 15-year period, 1963-1977, was carried out in the Portland-Vancouver Metropolitan area,
Oregon. Cases [n = 1,622] were identified through a search of 24 hospitals in the four-county
B-184
-------
area of Portland-Vancouver and included a Veteran's Administration hospital and two closed
hospitals (whose records were accessible in storage). In addition, death certificates that
mentioned leukemia from the same period were searched, adding 244 cases, for a total of
1,866 leukemia cases. The finding of additional cases using death certificates suggests hospital
records may have been incomplete. Associations with job title as coded to usual occupation and
leukemia cases aged 16-74 years were carried out, examining age-standardized rates based on
the direct method of age standardization using the 1970 U.S. population census (midpoint of case
ascertainment period). Given census records group population estimates for individuals aged 65
or older, occupation-specific analyses truncated case inclusion at age 67. The determination of
significance of a deviation of an occupational leukemia rate from its respective area-wide rate for
all women or all men was based on the assumption that a rate is a mean for a distribution of
binomial events, and the distribution of such means was regarded as approximately normal.
Age-adjusted incidence rates for separate occupations were calculated for all leukemia, all
lymphatic leukemia, and all nonlymphatic leukemia. Lymphatic leukemia cases, including
chronic lymphatic leukemia, are now classified as subtypes of non-Hodgkin lymphoma (Morton
et al., 2005). One female dry cleaner and launderer was identified with both lymphatic and
nonlymphatic leukemia subtypes and was counted twice in the statistical analyses.
Occupations were broadly grouped into over 20 categories and included dry cleaners and
launderers, a grouping that contained 313 males and 1,298 females with associated exposure
prevalences (based on the 1970 population) of 0.1% and 0.3%, respectively. Morton and
Marjanovic (1984) do not identify the source for job title information and stated the trained coder
could identify "usual occupation," but little else. The lack of information on full job history, in
addition to possible misclassification of occupation on death certificates, suggests an incomplete
occupation history.
This study is less sensitive for identifying cancer hazard because case ascertainment may
be incomplete and because of possible selection bias associated with hospital records, inability to
identify specific exposures, and use of age-adjusted incidence rates, rather than a relative risk
estimate.
B.3.6. Vartiainen et al. (1993)
Vartiainen, T.; Pukkala, E.; Rienoja, T.; Strandman, T.; Kaksonen, K. (1993).
Population exposure to tri- and tetrachloroethene and cancer risk: Two cases of
drinking water pollution. Chemosphere, 27,1171-1181.
http://dx.doi.org/10.1016/0045-6535(93)90165-2
Summary: This parallel standardized incidence ratio study had three aims: (1) to find out
whether inhabitants in the villages of Oitti and Hattula had been exposed to tetrachloroethylene
B-185
-------
by analyzing urinary excretion; (2) to examine which compound(s) would provide the best index
of exposure to low levels of tetrachloroethylene; and (3) to determine whether the cancer
incidence was increased in the municipalities in question. The first part of this parallel study
consisted of residents of two villages (Oitti and Hattula) in Finland who consumed contaminated
drinking water. Of the 116 possible participants, 8 were excluded because they had not drunk
any contaminated water. The reference population was divided into two groups: (1) ground
water controls included volunteers residing in a nearby town whose drinking water came from
ground water, and (2) surface water controls included volunteers whose drinking water came
from bank-filtrated surface water. The final sample consisted of 108 exposed (87 from Oitti and
21 from Hattula) and 60 unexposed (45 ground water and 15 surface water). All participants
were interviewed about their source of drinking water and the approximate amount of water they
consumed each day. Gas chromatography was used to detect tetrachloroethylene in the urine
samples. These results were then compared among all four locations (2 exposed villages and
2 unexposed villages), though the authors do not report the methods they used to conduct the
comparison. The second part of the parallel study used the Finnish Cancer Registry to identify
the number of all cancers, liver cancer, non-Hodgkin lymphoma, multiple myeloma, and
leukemia in Hausjarvi (municipality where Oitti is located) and Hattula. Overall, there were
1,931 cancer cases identified (972 from Hausjarvi and 959 from Hattula). Standardized
incidence ratios, along with their corresponding 95% CIs, were calculated assuming a Poisson
distribution. Expected numbers of cancer cases were estimated based on annual age- and
sex-specific numbers for the whole of Finland for each year between 1953 and 1991. A strength
of the latter part of this parallel study was use of the Finnish Cancer Registry, which contained
all cancer cases since 1953 and increased the confidence in calculated estimates. Limitations
overall included the extended latency time between exposure to tetrachloroethylene and cancer
diagnosis, as well as the ambiguity related to the time period within which cases were exposed.
Exposure misclassification bias is likely, given the ecologic design of this study, and exposure
assignment to both cases and controls is not validated, given the lack of monitoring data for the
examined time period. Furthermore, the presentation of incidence rates for a presumed
unexposed population provide little insight on site-specific cancer incidence in the two presumed
exposed towns, given direct comparison of SIRs has methodological limitations due to
differences in population age structure.
B-186
-------
Table B-4. Summaries of characteristics of geographically based and other studies
Reference
Study design1"
Sample size
Data collection
Exposure assessment0
Statistical approach"1
Cohnetal. (1994)
Cases: identified from New Jersey
State Cancer Registry, men and
women, diagnosed with primary
leukemia or non-Hodgkin
lymphomafrom 1979-1987,
living in 1 of 75 municipalities,
information supplemented with
death certificates; only
municipalities where >80%
population received public water
supply selected
Proxy exposure surrogate—PCE
Leukemia incidence (acute
lymphocytic, chronic lymphocytic,
acute myelogenous, other
specified, unspecified), NHL
(low-grade, intermediate-grade,
high-grade)
1,190 cases
leukemia (118 acute
lymphocytic, 354
chronic
lymphocytic, 276
acute myelogenous,
146 chronic
myelogenous, 61
other specified, 235
unspecified), 1,658
cases of NHL (434
low-grade, 708
intermediate-grade,
402 intermediate-
grade/diffuse large
cell/
reticulosarcoma, 69
high-grade, 51 high-
grade/non-Burkitt' s)
Records based—
cancer registry data
and summary data
from state of
average drinking
water concentration
in each
municipality
PCE exposure based on water
monitoring data (average from
1984-1985 by NJ Department of
Environmental Protection and
Energy), calculated for each
municipality based on water
measurements and proportion of
water purchased elsewhere;
samples from treatment plants
and tap sites within distribution
system; PCE exposures
categorized as <0.1, 0.1-5, and
>5 ppb, based on EPA standards;
surveys from 1978-1983 used to
corroborate 1984-1985
estimates; 440 (37%) cases of
leukemia, 662 (39.9%) cases of
NHL exposed due to residences
in a municipality with PCE levels
from 0.1 through >5 ppb
Log-linear regression
models assuming
Poisson distribution for
incidence rate ratios,
95% CIs, adjusted for
age, stratified by sex
and type of leukemia or
non-Hodgkin
lymphoma
Lee et al. (2003)
Death certificates from 2 villages
near factory from 1966-1997;
nosologists blindly assessed cause
of death; Cases: cancer cause of
death, linked to Taiwan National
Cancer Registry; Controls:
cardiovascular or cerebrovascular
disease cause of death;
Exclusions: arrhythmia, all
noncancer diseases also used as
controls
Cancer mortality (liver, stomach,
colorectal, lung, all cancers)
1,333 decedents:
266 cancer cases;
344 cardiovascular-
cerebrovascular
controls
Not applicable
Proxy—residence near factory.
Exposed lived downstream of
factory; unexposed: lived
upstream
Death certificates linked to
Labour Insurance Bureau to find
previously employed in factory
Mantel-Haenszel for
mortality ORs, 95%
CIs, adjusted for age
Multiple logistic
regression for exposure
effect, adjusted for age,
time period
Cochran-Armitage test
for trend for effect of
time period
Latency period of 10 yr
B-187
-------
Table B-4. Summaries of characteristics of geographically based and other studies (continued)
Reference
Ma et al. (2009)
Mallin (19901
Study design1"
Cases discharged from hospital
with diagnosis of kidney cancer,
1993-2004, for New York City
zip codes with median household
income of $17,864-$ 142,926.
Population estimate by zip code
from U.S. Department of Census.
Proxy exposure surrogate: density
of dry-cleaning businesses
(number of dry cleaners/zip code
area in square kilometers)
Renal cell cancer prevalence
Cases: identified from medical
records in 8 counties of northwest
Illinois, men and women,
diagnosed with histologically
confirmed bladder cancer from
1978-1985, those diagnosed or
treated in Iowa or Wisconsin also
identified from Iowa State Health
Registry, Wisconsin Cancer
Reporting System; deaths from
bladder cancer obtained
Proxy — residence near
contaminated well water
Bladder cancer incidence and
mortality
Sample size
10,916 discharges,
1,458 discharges in
lowest exposure
category (referent
group)
Unit of analysis is
discharge, and a
subject could be
counted as many
times as discharged
from hospital
Cases and residence
from medical
records and cancer
registries
719 hlnHHpr punppr
/ AZ, UltfUUX/l V^ClllV^^l
cases among
CHIIPH^IUTI mpTi UTiH
V-^dLl^dolClll 111^-11 C111U
women
Data collection
Discharge
information
obtained from New
York Statewide
Planning and
Research
Cooperative
System
Not applicable
Exposure assessment0
Density of dry cleaners per zip
code (number of dry cleaners/zip
code area) proxy surrogate for
PCE exposure. Proxy exposure
surrogate not validated with
ambient monitoring data
Incidence data stratified by
county and zip code; Winnebago
county later found to have well
water contaminated with PCE
and other compounds
Statistical approach"1
Negative binomial
regression for
prevalence rate ratios,
95% CIs, adjusted for
population density, age,
race, and interactions
specific to individual
exposure level
Indirect standardization
for SIRs and SMRs,
each adjusted for age;
95% CIs using
Miettinen's exact
limits; %2 tests for
significance except
when expected counts
<5: Fisher exact test for
95% CIs, significance
assumed Poisson
distribution
B-188
-------
Table B-4. Summaries of characteristics of geographically based and other studies (continued)
Reference
Study design
Sample size
Data collection
Exposure assessment0
Statistical approach
Morton and
Marjanovic (1984)
Cases: Men and women leukemia
cases, diagnosed 1963-1977,
residing in Portland-Vancouver
Metropolitan Area, 16-74 yr old,
Referents: 1970 age-specific
leukemia rates for U.S.
population, direct method of age
standardization
Proxy—occupational title on
hospital record or on death
certificate
Leukemia incidence and mortality,
including subtype (lymphatic,
nonlymphatic)
975 leukemia cases
among males,
336,850 population
703 leukemia cases
among females,
102,310 population
Record-based
information—
Cases ascertained
from hospital
records and death
certificates.
Record source for
usual occupational
title not reported
Occupations grouped into 20
categories, 313 males and 1,298
females identified as dry cleaner
or launderer (0.1% and 0.3%,
respectively)
Comparison of directly
standardized age-
adjusted incidence rates
Vartiainen et al.
(1993)
1) Cases: 2 villages who
consumed contaminated water, 2
reference groups: drank ground
water, drank surface water
2) Identified from Finnish Cancer
Registry, all cancers, liver cancer,
non-Hodgkin lymphoma, multiple
myeloma, leukemia in 1 exposed
village and 1 exposed municipality
Proxy—residence in exposed
village
Cancer incidence (liver,
non-Hodgkin lymphoma, Hodgkin
lymphoma, multiple myeloma,
leukemia, all cancers combined)
1) 116 identified
cases, 108 exposed;
60 unexposed
references
2) 1,931 cases in
exposed villages
1) Interviewed
about drinking
water source and
daily water
consumption, urine
samples collected
2) Not applicable
1) Exposed: drank contaminated
water; gas chromatography to
detect PCE in urine samples
2) Residence in exposed village
or municipality
1) Compare urine levels
in villages;
methodology not
reported
2) SIR, 95% CIs,
assuming Poisson
distribution; expected
numbers based on age-
and sex-specific
number of cancer cases
for Finland each year
1953-1991
B-189
-------
Table B-4. Summaries of characteristics of geographically based and other studies (continued)
aA study's hypothesis is included here if explicitly stated; otherwise, only the objective is included.
bStudy design includes the overall approach, study population, relevant dates, type of exposure, and endpoint measured. For type of exposure, when a proxy is
not explicitly stated for PCE, an attempt was made to identify proxies based on relevant industries and occupations, including laundry/dry cleaners and textile
industry, and to some extent, metal industry, aerospace industry, appliance industry, automotive industry, and manufacturing of chloroflourocarbons (though
not used).
'Exposure assessment includes exposure assignment (e.g., was coding conducted and by whom), exposure approach (e.g., what kind of coding was used), and
exposure-assessment metrics (e.g., length of exposure).
dStatistical approach includes adjustment for covariates, latency, or lag period, and documentation of statistical analysis and observations.
B-190
-------
B.4. References
't Mannetje, A; Dryson, E; Walls, C: McLean, D; McKenzie, F; Maule, M; Cheng, S;
Cunningham. C: Kromhout H: BofFetta. P: Blair. A: Pearce. N. (2008). High risk
occupations for non-Hodgkin's lymphoma in New Zealand: Case-control study. Occup
Environ Med 65: 354-363. http://dx.doi.org/10.1136/oem.2007.035014.
Ahlborg, G: Bodin, L. (1991). Tobacco smoke exposure and pregnancy outcome among working
women. A prospective study at prenatal care centers in Orebro County, Sweden. Am J
Epidemiol 133:338-347.
Ahlborg, G, Jr. (1990a). Pregnancy outcome among women working in laundries and dry-
cleaning shops using tetrachloroethylene. Am J Ind Med 17: 567-575.
http://dx.doi.org/10.1002/aiim.4700170503.
Ahlborg, G, Jr. (1990b). Validity of exposure data obtained by questionnaire. Two examples
from occupational reproductive studies. Scand J Work Environ Health 16: 284-288.
Ahrens, W; Merletti, F. (1998). A standard tool for the analysis of occupational lung cancer in
epidemiologic studies. Int J Occup Environ Health 4: 236-240.
Andersen, A; Barlow, L; Engeland, A; Kjaerheim, K; Lynge, E; Pukkala, E. (1999). Work-
related cancer in the Nordic countries. Scand J Work Environ Health 25: 1-116.
Anttila, A; Pukkala, E: Sallmen, M; Hernberg, S: Hemminki, K. (1995). Cancer incidence among
Finnish workers exposed to halogenated hydrocarbons. J Occup Environ Med 37: 797-
806.
Aronson, K: Siemiatycki, J: Dewar, R; Gerin, M. (1996). Occupational risk factors for prostate
cancer: Results from a case-control study in Montreal, Quebec, Canada. Am J Epidemiol
143:363-373.
Asal NR: Gever, JR: Risser, PR: Lee. ET: Kadamani, S: Cherng, N. (1988). Risk factors in
renal cell carcinoma. II. Medical history, occupation, multivariate analysis, and
conclusions. Cancer Detect Prev 13: 263-279.
Aschengrau, A; Ozonoff D; Paulu, C: Coogan, P; Vezina, R: Heeren, T; Zhang, Y. (1993).
Cancer risk and tetrachloroethylene-contaminated drinking water in Massachusetts. Arch
Environ Health 48: 284-292.
Aschengrau, A; Paulu, C: Ozonoff D. (1998). Tetrachloroethylene-contaminated drinking water
and the risk of breast cancer. Environ Health Perspect 106: 947-953.
Aschengrau, A; Rogers, S: Ozonoff, D. (2003). Perchloroethylene-contaminated drinking water
and the risk of breast cancer: Additional results from Cape Cod, Massachusetts, USA.
Environ Health Perspect 111: 167-173.
Auperin, A; Benhamou, S: Ory-Paoletti, C: Flamant, R. (1994). Occupational risk factors for
renal cell carcinoma: A case-control study. Occup Environ Med 51: 426-428.
http://dx.doi.0rg/10.1136/oem.51.6.426.
Austin. H: Delzell. E: Grufferman. S: Levine, R: Morrison. AS: Stollev. PD: Cole. P. (1987).
Case-control study of hepatocellular carcinoma, occupation, and chemical exposures. J
Occup Med 29: 665-669.
Bakke, B; Stewart, P: Waters, M. (2007). Uses of and exposure to trichloroethylene in U.S.
industry: A systematic literature review. J Occup Environ Hyg 4: 375-390.
http://dx.doi.org/10.1080/15459620701301763.
B-191
-------
Band. P: Le. N: Fang. R: Threlfall. W: Gallagher. R. (1999). Identification of occupational
cancer risks in British Columbia. Part II: A population-based case-control study of 1516
Prostatic cancer cases. J Occup Environ Med 41: 233-247.
Band. PR: Le. ND: Fang. R: Deschamps. M: Gallagher. RP: Yang. P. (2000). Identification of
occupational cancer risks in British Columbia: A population-based case-control study of
995 incident breast cancer cases by menopausal status, controlling for confounding
factors. J Occup Environ Med 42: 284-310.
Blair. A: Stewart. PA: Tolbert PE: Grauman. D: Moran. FX: Vaught J: Ravner. J. (1990).
Cancer and other causes of death among a cohort of dry cleaners. Br J Ind Med 47: 162-
168. http://dx.doi.0rg/10.1136/oem.47.3.162.
Blair, A; Linos, A; Stewart, PA; Burmeister, LF; Gibson, R: Everett G: Schuman, L; Cantor,
KP. (1993). Evaluation of risks for non-Hodgkin's lymphoma by occupation and industry
exposures from a case-control study. Am J Ind Med 23: 301-312.
http://dx.doi.org/10.1002/aiim.4700230207.
Blair, A; Hartge, P: Stewart, PA; McAdams, M; Lubin, J. (1998). Mortality and cancer incidence
of aircraft maintenance workers exposed to trichloroethylene and other organic solvents
and chemicals: Extended follow-up. Occup Environ Med 55: 161-171.
http://dx.doi.0rg/10.1136/oem.55.3.161.
Blair, A; Petralia, SA; Stewart, PA. (2003). Extended mortality follow-up of a cohort of dry
cleaners. AnnEpidemiol 13: 50-56. http://dx.doi.org/10.1016/51047-2797(02)00250-8.
Boice. JD. Jr: Marano. D: Fryzek. J: Sadler. C: McLaughlin. JK. (1999). Mortality among
aircraft manufacturing workers. Occup Environ Med 56: 581-597.
http://dx.doi.0rg/10.1136/oem.56.9.581.
Bolm-Audorff U: Ahrens. W: Jockel KH: Greiser. E: Molik. B: Timm. J: Wichmann. HE.
(1989). Experience with supplementary questionnaires in a lung cancer case reference
study. In D Hemon; M Goldberg (Eds.), Proceedings of the Workshop on Methodology
of Assessment of Occupational Exposures in the Context of Epidemiology Detection of
Cancer Risks, Paris, 18-19 February 1988 (pp. 117-168). Luxembourg: Office for
Official Publications of the European Community.
Bond. G: McLaren. E: Cartmill. J: Wvmer. K: Sobel W: Lipps. T: Cook. R. (1987). Cause-
specific mortality among male chemical workers. Am J Ind Med 12: 353-383.
Bond, G: McLaren, E: Sabel, F; Bodner, K; Lipps, T; Cook, R. (1990). Liver and biliary tract
cancer among chemical workers. Am J Ind Med 18: 19-24.
Breslow, NE: Day, NE. (1980). Statistical methods in cancer research: Volume 1 - the analysis
of case-control studies. Lyon, France: International Agency for Research on Cancer.
http://www.iarc.fr/en/publications/pdfs-online/stat/sp32/index.php.
Brownson, RC: Alavanja, MC: Chang, JC. (1993). Occupational risk factors for lung cancer
among nonsmoking women: A case-control study in Missouri (United States). Cancer
Causes Control 4: 449-454.
Burns, PB; Swanson, GM. (1991). Risk of urinary bladder cancer among blacks and whites: The
role of cigarette use and occupation. Cancer Causes Control 2: 371-379.
http://dx.doi.org/10.1007/BF00054297.
Calvert, GM: Ruder, AM; Petersen, MR. (2011). Mortality and end-stage renal disease incidence
among dry cleaning workers. Occup Environ Med 68: 709-716.
http://dx.doi.org/10.1136/oem.2010.060665.
B-192
-------
Cano, MI; Pollan, M. (2001). Non-Hodgkin's lymphomas and occupation in Sweden. Int Arch
Occup Environ Health 74: 443-449. http://dx.doi.org/10.1007/s004200100248.
Chang. YM: Tai. CF: Yang. SC: Chen. CJ: Shih. TS: Lin. R: Liou. SH. (2003). A cohort
mortality study of workers exposed to chlorinated organic solvents in Taiwan. Ann
Epidemiol 13: 652-660. http://dx.doi.org/10.1016/S1047-2797(03)00038-3.
Chang. YM: Tai. CF: Yang. SC: Lin. RS: Sung. FC: Shih. TS: Liou. SH. (2005). Cancer
incidence among workers potentially exposed to chlorinated solvents in an electronics
factory. J Occup Health 47: 171-180. http://dx.doi.org/10.1539/ioh.47.171.
Chatterjee, N; Hartge, P; Cerhan, J: Cozen, W; Davis, S; Ishibe, N; Colt J; Goldin, L; Severson,
R, (2004). Risk of non-Hodgkin's lymphoma and family history of lymphatic,
hematologic, and other cancers. Cancer Epidemiol Biomarkers Prev 13: 1415-1421.
Chow. WH: McLaughlin. JK: Linet MS: Niwa. S: Mandel JS. (1994). Use of analgesics and
risk of renal cell cancer. Int J Cancer 59: 467-470.
http://dx.doi.org/10.1002/iic.2910590406.
Chow. WH: McLaughlin. JK: Malker. HS: Linet MS: Weiner. JA: Stone. BJ. (1995).
Esophageal cancer and occupation in a cohort of Swedish men. Am J Ind Med 27: 749-
757. http://dx.doi.org/10.1002/aiim.4700270509.
Claude, JC: Frentzel-Beyme, RR: Kunze, E. (1988). Occupation and risk of cancer of the lower
urinary tract among men: A case-control study. Int J Cancer 41: 371-379.
http://dx.doi.org/10.1002/iic.2910410309.
Clavel, J: Mandereau, L; Conso, F; Limasset JC: Pourmir, I; Flandrin, G: Hemon, D. (1998).
Occupational exposure to solvents and hairy cell leukaemia. Occup Environ Med 55: 59-
64. http://dx.doi.0rg/10.1136/oem.55.l.59.
Cohn, P: Klotz, J: Bove, F; Berkowitz, M; Fagliano, J. (1994). Drinking water contamination and
the incidence of leukemia and non-Hodgkin's lymphoma. Environ Health Perspect 102:
556-561.
Colt J: Baris, D: Stewart. P: Schned, A: Heanev, J: Mott L: Silverman, D: Karagas, M. (2004).
Occupation and bladder cancer risk in a population-based case-control study in New
Hampshire. Cancer Causes Control 15: 759-769.
http://dx.doi.Org/10.1023/B:CACO.0000043426.28741.a2.
Colt JS: Karagas, MR; Schwenn, M: Baris, D: Johnson, A; Stewart, P: Verrill, C; Moore, LE:
Lubin, J: Ward, MH; Samanic, C; Rothman, N: Cantor, KP; Beane Freeman, LE; Schned,
A; Cherala, S: Silverman, DT. (2011). Occupation and bladder cancer in a population-
based case-control study in Northern New England. Occup Environ Med 68: 239-249.
http://dx.doi.org/10.1136/oem.2009.052571.
Consonni, D: De Mattels, S: Lubin, JH; Wacholder, S: Tucker, M: Pesatori, AC: Caporaso, NE;
Bertazzi, PA; Landi, MT. (2010). Lung cancer and occupation in a population-based
case-control study. Am J Epidemiol 171: 323-333. http://dx.doi.org/10.1093/aje/kwp391.
Corbin, M: McLean, D: Mannetje, A; Dryson, E; Walls, C: McKenzie, F: Maule, M: Cheng, S:
Cunningham, C: Kromhout H; Blair, A; Pearce, N. (2011). Lung cancer and occupation:
A New Zealand cancer registry-based case-control study. Am J Ind Med 54: 89-101.
http://dx.doi.org/10.1002/aiim.20906.
Cordier, S: Clavel, J: Limasset JC: Boccon-Gibod, L: Le Moual, N: Mandereau, L: Hemon, D.
(1993). Occupational risks of bladder cancer in France: A multicentre case-control study.
Int J Epidemiol 22: 403-411. http://dx.doi.Org/10.1093/ije/22.3.403.
B-193
-------
Cornfield, J. (1956). A statistical problem arising from retrospective studies. In J Neyman (Ed.),
Proceedings of the third Berkeley Symposium on Mathematical Statistics and Probability
(Vol 4) (pp. 135-148). Berkeley, CA: University of California Press.
http://proiecteuclid.org/euclid.bsmsp/1200502552.
Costantini, AS: Miligi, L; Kriebel D: Ramazzotti, V: Rodella, S; Scarpi, E; Stagnaro, E;
Tumino, R; Fontana, A; Masala, G: Vigano, C: Vindigni, C: Crosignani, P; Benvenuti, A;
Vineis, P. (2001). A multicenter case-control study in Italy on hematolymphopoietic
neoplasms and occupation. Epidemiology 12: 78-87.
Costantini, AS: Benvenuti, A; Vineis, P: Kriebel, D: Tumino, R; Ramazzotti, V: Rodella, S:
Stagnaro, E: Crosignani, P; Amadori, D; Mirabelli, D: Sommani, L; Belletti, I; Troschel,
L: Romeo. L: Miceli, G: Tozzi, G: Mendico, I: Maltoni. S: Miligi. L. (2008). Risk of
leukemia and multiple myeloma associated with exposure to benzene and other organic
solvents: Evidence from the Italian Multicenter Case-control study. Am J Ind Med 51:
803-811. http://dx.doi.org/10.1002/aiim.20592.
Costas, K; Knorr, RS: Condon, SK. (2002). A case-control study of childhood leukemia in
Woburn, Massachusetts: The relationship between leukemia incidence and exposure to
public drinking water. Sci Total Environ 300: 23-35. http://dx.doi.org/10.1016/S0048-
9697(02)00169-9.
Cutler. J: Parker. G: Rosen. S: Prennev, B: Healev, R: Caldwell. G. (1986). Childhood leukemia
in Woburn, Massachusetts. Public Health Rep 101: 201-205.
De Roos, A: Olshan, A: Teschke, K: Poole, C: Savitz, D: Blatt J: Bondv, M: Pollock. B. (2001).
Parental occupational exposures to chemicals and incidence of neuroblastoma in
offspring. Am JEpidemiol 154: 106-114. http://dx.doi.Org/10.1093/aie/154.2.106.
Delahunt, B: Bethwaite, PB; Nacey, JN. (1995). Occupational risk for renal cell carcinoma. A
case-control study based on the New Zealand Cancer Registry. Br J Urol 75: 578-582.
http://dx.doi.Org/10.llll/i.1464-410X.1995.tb07410.x.
Demond, AH. (1982) A source of tetrachloroethylene in the drinking water of New England: An
evaluation of the toxicity of tetrachloroethylene and the prediction of its leaching rates
from vinyl-lined asbestos-cement pipe. (Master's Thesis). Massachusetts Institute of
Technology, Cambridge, MA. Retrieved from http://dspace.mit.edU/handle/1721.l/15478
Dosemeci, M; Cocco, P: Gomez, M; Stewart, PA; Heineman, EF. (1994). Effects of three
features of a job-exposure matrix on risk estimates. Epidemiology 5: 124-127.
Dosemeci, M: Cocco, P: Chow, WH. (1999). Gender differences in risk of renal cell carcinoma
and occupational exposures to chlorinated aliphatic hydrocarbons. Am J Ind Med 36: 54-
59. http://dx.doi.org/10.1002/(SICni097-0274(199907)36:K54::AID-AJIM8>3.0.CO:2-
0.
Dryson, E: 't Mannetje, A; Walls, C: McLean, D: McKenzie, F; Maule, M: Cheng, S:
Cunningham. C: Kromhout H: Boffetta. P: Blair. A: Pearce. N. (2008). Case-control
study of high risk occupations for bladder cancer in New Zealand. Int J Cancer 122:
1340-1346. http://dx.doi.org/10.1002/iic.23194.
Fabbro-Peray, P: Daures, JP: Rossi, JF. (2001). Environmental risk factors for non-Hodgkin's
lymphoma: A population-based case-control study in Languedoc-Roussillon, France.
Cancer Causes Control 12: 201-212. http://dx.doi.Org/10.1023/A:1011274922701.
Ferrario, F; Continenza, D: Pisani, P: Magnani, C: Merletti, F; Berrino, F. (1988). Description of
a job-exposure matrix for sixteen agents which are or may be related to respiratory
cancer. In C Hogstedt; C Reuterwall (Eds.), Progress in occupational epidemiology:
B-194
-------
Proceedings of the Sixth International Symposium on Epidemiology in Occupational
Health, Stockholm, Sweden, 16-19 August 1988 (pp. 379-382). Amsterdam, Holland:
Excerpta Medica.
Ferroni, C: Selis, L; Mutti, A; Folli, D; Bergamaschi, E; Franchini, I. (1992). Neurobehavioral
and neuroendocrine effects of occupational exposure to perchloroethylene.
Neurotoxicology 13: 243-247.
Fredriksson, M; Bengtsson, NO: Hardell, L; Axelson, O. (1989). Colon cancer, physical activity,
and occupational exposures: A case-control study. Cancer 63: 1838-1842.
Gaertner, RRW: Trpeski, L: Johnson, KC. (2004). A case-control study of occupational risk
factors for bladder cancer in Canada. Cancer Causes Control 15: 1007-1019.
http://dx.doi.org/10.1007/sl0552-004-1448-7.
Gaillard, Y; Billault F; Pepin, G. (1995). Tetrachloroethylene fatality: Case report and simple
gas chromatographic determination in blood and tissues. Forensic Sci Int 76: 161-168.
http://dx.doi.org/10.1016/0379-0738(95)01813-l.
Gart, JJ. (1970). A locally most powerful test for the symmetric folded binomial distribution.
Biometrics 26: 129-138.
Gold, LS: De Roos, AJ: Waters, M; Stewart, P. (2008). Systematic literature review of uses and
levels of occupational exposure to tetrachloroethylene [Review]. J Occup Environ Hyg 5:
807-839. http://dx.doi.org/10.1080/15459620802510866.
Gold. LS: Milliken, K: Stewart. P: Purdue. M: Severson, R: Seixas, N: Blair. A: Davis. S:
Hartge, P; De Roos, AJ. (2010a). Occupation and multiple myeloma: An occupation and
industry analysis. Am J Ind Med 53: 768-779. http://dx.doi.org/10.1002/aiim.20857.
Gold. LS: Stewart. PA: Milliken, K: Purdue. M: Severson, R: Seixas, N: Blair. A: Hartge. P:
Davis, S: De Roos, AJ. (201 Ob). The relationship between multiple myeloma and
occupational exposure to six chlorinated solvents. Occup Environ Med 68: 391-399.
http://dx.doi.org/10.1136/oem.2009.054809.
Gomez, MR; Cocco, P: Dosemeci, M: Stewart, PA. (1994). Occupational exposure to chlorinated
aliphatic hydrocarbons: Job exposure matrix. Am J Ind Med 26: 171-183.
http://dx.doi.org/10.1002/aiim.4700260204.
Gonzalez, CA; Lopez-Abente, G: Errezola, M: Escolar, A; Riboli, E; Izarzugaza, I; Nebot M.
(1989). Occupation and bladder cancer in Spain: A multi-centre case-control study. Int J
Epidemiol 18: 569-577. http://dx.doi.Org/10.1093/iie/18.3.569.
Hardell, L: Eriksson, M: Lenner, P: Lundgren, E. (1981). Malignant lymphoma and exposure to
chemicals, especially organic solvents, chlorophenols and phenoxy acids: A case-control
study. Br J Cancer 43: 169-176.
Harrington. JM: Whitbv, H: Gray. CN: Reid. FJ: Aw. TC: Waterhouse, JA. (1989). Renal disease
and occupational exposure to organic solvents: A case referent approach. Br J Ind Med
46: 643-650. http://dx.doi.Org/10.1136/oem.46.9.643.
Hartge. P: Cahill. JI: West D: Hauck. M: Austin. D: Silverman. D: Hoover. R. (1984). Design
and methods in a multi-center case-control interview study. Am J Public Health 74: 52-
56.
Heineman, EF; Cocco, P: Gomez, MR; Dosemeci, M: Stewart, PA; Haves, RB; Zahm, SH;
Thomas, TL; Blair, A. (1994). Occupational exposure to chlorinated aliphatic
hydrocarbons and risk of astrocytic brain cancer. Am J Ind Med 26: 155-169.
http://dx.doi.org/10.1002/aiim.4700260203.
B-195
-------
Hernberg, S; Korkala, ML; Asikainen, U; Riala, R. (1984). Primary liver cancer and exposure to
solvents. Int Arch Occup Environ Health 54: 147-153.
http://dx.doi.org/10.1007/BF00378517.
Hernberg, S: Kauppinen, T; Riala, R: Korkala, ML; Asikainen, U. (1988). Increased risk for
primary liver cancer among women exposed to solvents. Scand J Work Environ Health
14: 356-365.
Hoar, SK; Morrison, AS: Cole, P; Silverman, DT. (1980). An occupation and exposure linkage
system for the study of occupational carcinogenesis. J Occup Environ Med 22: 722-726.
Hours, M; Dananche, B; Fevotte, J: Bergeret A; Ayzac, L; Cardis, E; Etard, JF: Fallen, C; Roy,
P: Fabry, J. (1994). Bladder cancer and occupational exposures. Scand J Work Environ
Health 20: 322-330.
Houten, L; Sonnesso, G. (1980). Occupational exposure and cancer of the liver. Arch Environ
Health 35: 51-53.
IARC (International Agency for Research on Cancer). (2004a). Some drinking-water
disinfectants and contaminants, including arsenic. Lyon, France.
http://monographs.iarc.fr/ENG/Monographs/vol84/index.php.
IARC (International Agency for Research on Cancer). (2004b). Tobacco smoke and involuntary
smoking. Lyon, France.
Infante-Rivard, C: Siemiatycki, J: Lakhani, R: Nadon, L. (2005). Maternal exposure to
occupational solvents and childhood leukemia. Environ Health Perspect 113: 787-792.
http://dx.doi.org/10.1289/ehp.7707.
Jensen, OM; Wahrendorf, J: Knudsen, JB; Sorensen, BL. (1987). The Copenhagen case-referent
study on bladder cancer: Risks among drivers, painters, and certain other occupations.
Scand J Work Environ Health 13: 129-134.
Ji, J: Granstrom, C: Hemminki, K. (2005a). Occupation and bladder cancer: A cohort study in
Sweden. Br J Cancer 92: 1276-1278. http://dx.doi.org/10.1038/si.bic.6602473.
Ji, J: Hemminki, K. (2005a). Occupation and upper aerodigestive tract cancers: A follow-up
study in Sweden. J Occup Environ Med 47: 785-795.
http://dx.doi.org/10.1097/01.iom.0000165798.28569.b5.
Ji, J: Granstrom, C: Hemminki, K. (2005b). Occupational risk factors for kidney cancer: A
cohort study in Sweden. World Journal of Urology 23: 271-278.
http://dx.doi.org/10.1007/s00345-005-0007-5.
Ji, J: Hemminki, K. (2005b). Occurrences of leukemia subtypes by socioeconomic and
occupational groups in Sweden. J Occup Environ Med 47: 1131-1140.
http://dx.doi.org/10.1097/01.iom.0000174302.63621.e8.
Ji, J: Hemminki, K. (2005c). Variation in the risk for liver and gallbladder cancers in
socioeconomic and occupational groups in Sweden with etiological implications. Int
Arch Occup Environ Health 78: 641-649. http://dx.doi.org/10.1007/s00420-005-0015-l.
Ji, J: Hemminki, K. (2006). Socioeconomic/occupational risk factors for lymphoproliferative
diseases in Sweden. Ann Epidemiol 16: 370-376.
http://dx.doi.0rg/10.1016/j.annepidem.2005.09.002.
Johansen, K: Tinnerberg, H; Lynge, E. (2005). Use of history science methods in exposure
assessment for occupational health studies. Occup Environ Med 62: 434-441.
http://dx.doi.org/10.1136/oem.2004.016493.
Kato, I; Koenig, KL; Watanabe-Meserve, H; Baptiste, MS: Lillquist PP: Frizzera, G: Burke, JS:
Moseson, M; Shore, RE. (2005). Personal and occupational exposure to organic solvents
B-196
-------
and risk of non-Hodgkin's lymphoma (NHL) in women (United States). Cancer Causes
Control 16: 1215-1224.
Kaune, WT; Savitz, DA. (1994). Simplification of the Wertheimer-Leeper wire code.
Bioelectromagnetics 15: 275-282. http://dx.doi.org/10.1002/bem.2250150402.
Kauppinen, T; Heikkila, P; Plato, N; Woldbaek, T; Lenvik, K; Hansen, J; Kristjansson, V:
Pukkala, E. (2009). Construction of job-exposure matrices for the Nordic Occupational
Cancer Study (NOCCA). Acta Oncol 48: 791-800.
http://dx.doi.org/10.1080/02841860902718747.
Kernan, GJ: Ji, BT; Dosemeci, M; Silverman, DT; Balbus, J: Zahm, SH. (1999). Occupational
risk factors for pancreatic cancer: A case-control study based on death certificates from
24 U.S. States. Am J Ind Med 36: 260-270. http://dx.doi.org/10.1002/(SICI) 1097-
0274O99908)36:2<260::AID-AJIM5>3.0.CO:2-P.
Koepsell TD: Paling. JR: Weiss. NS: Taylor. JW: Olshan, AF: Lyon, JL: Swanson, GM: Child.
M. (1987). Antigenic stimulation and the occurrence of multiple myeloma. Am J
Epidemiol 126: 1051-1062.
Kogevinas, M: 'tMannetje, A; Cordier, S; Ranft, U; Gonzalez, C; Vineis, P: Chang-Claude, J:
Lynge, E; Wahrendorf J: Tzonou, A; Jockel, K; Serra, C; Porru, S; Hours, M: Greiser, E;
Boffetta, P. (2003). Occupation and bladder cancer among men in Western Europe.
Cancer Causes Control 14: 907-914.
http://dx.doi.Org/10.1023/B:CACO.0000007962.19066.9c.
Lagakos, SW: Wessen, BJ: Zelen, M. (1986). An analysis of contaminated well water and health
effects in Woburn, Massachusetts. J Am Stat Assoc 81: 583-596.
Lee. LJ. -H: Chung. CW: Ma. YC: Wang. GS: Chen. PC: Hwang. YH: Wang. JD. (2003).
Increased mortality odds ratio of male liver cancer in a community contaminated by
chlorinated hydrocarbons in groundwater. Occup Environ Med 60: 364-369.
http://dx.doi.0rg/10.1136/oem.60.5.364.
Lin, RS: Kessler, II. (1981). A multifactorial model for pancreatic cancer in man: Epidemiologic
evidence. JAMA 245: 147-152.
Lindbohm, ML; Sallmen, M: Kyyronen, P: Kauppinen, T: Pukkala, E. (2009). Risk of liver
cancer and exposure to organic solvents and gasoline vapors among Finnish workers. Int
J Cancer 124: 2954-2959. http://dx.doi.org/10.1002/iic.24309.
Lowengart, RA; Peters, JM; Cicioni, C: Buckley, J: Bernstein, L; Preston-Martin, S: Rappaport,
R (1987). Childhood leukemia and parents' occupational and home exposures. J Natl
Cancer Inst 79: 39-46.
Lynge, E: Thygesen, L. (1990). Primary liver cancer among women in laundry and dry-cleaning
work in Denmark. Scand J Work Environ Health 16: 108-112.
Lynge, E: Carstensen, B; Andersen, O. (1995). Primary liver cancer and renal cell carcinoma in
laundry and dry-cleaning workers in Denmark. Scand J Work Environ Health 21: 293-
295.
Lynge, E: Andersen, A; Rylander, L: Tinnerberg, H; Lindbohm, ML; Pukkala, E: Romundstad,
P: Jensen, P: Clausen, LB; Johansen, K. (2006). Cancer in persons working in dry
cleaning in the Nordic countries. Environ Health Perspect 114: 213-219.
http://dx.doi.org/10.1289/ehp.8425.
Lynge, E: Tinnerberg, H: Rylander, L: Romundstad, P: Johansen, K: Lindbohm, ML; Heikkila,
P: Westberg. H: Clausen. LB: Piombino. A: Thorsted. BL. (2011). Exposure to
B-197
-------
tetrachloroethylene in dry cleaning shops in the Nordic countries. Ann Occup Hyg 55:
387-396. http://dx.doi.org/10.1093/annhyg/meql01.
Ma, J: Lessner, L; Schreiber, J: Carpenter, DO. (2009). Association between residential
proximity to PERC dry cleaning establishments and kidney cancer in New York City. J
Environ Public Health 2009: 183920. http://dx.doi.org/10.1155/2009/183920.
MacArthur, A; Le, N; Fang, R; Band, P. (2009). Identification of occupational cancer risk in
British Columbia: A population-based case-control study of 2,998 lung cancers by
histopathological subtype. Am J Ind Med 52: 221-232.
http://dx.doi.org/10.1002/ajim.20663.
Mallin, K. (1990). Investigation of a bladder cancer cluster in northwestern Illinois. Am J
Epidemiol 132: S96-S106.
Malone. KE: Koepsell. TD: Paling. JR: Weiss. NS: Morris. PD: Taylor. JW: Swanson. GM:
Lyon, JL. (1989). Chronic lymphocytic leukemia in relation to chemical exposures. Am J
Epidemiol 130: 1152-1158.
Mandel JS: McLaughlin, JK: Schlehofer, B; Mellemgaard, A; Helmert U; Lindblad, P;
McCredie, M; Adami, HO. (1995). International renal-cell cancer study. IV. Occupation.
Int J Cancer 61:601-605. http://dx.doi.org/10.1002/iic.2910610503.
Mantel, N. (1963). Chi-square tests with one degree of freedom; extensions of the Mantel-
Haenszel procedure. J Am Stat Assoc 58: 690-700.
Marano, DE: Boice, JD, Jr: Fryzek, JP: Morrison. JA: Sadler. CJ: McLaughlin. JK. (2000).
Exposure assessment for a large epidemiological study of aircraft manufacturing workers.
Appl Occup Environ Hyg 15: 644-656. http://dx.doi.org/10.1080/10473220050075653.
McCredie, M; Stewart, JH. (1993). Risk factors for kidney cancer in New South Wales. IV.
Occupation. Br J Ind Med 50: 349-354.
McLean, D; Mannetje, A; Dryson, E; Walls, C; McKenzie, F; Maule, M: Cheng, S;
Cunningham, C; Kromhout H; Boffetta, P: Blair, A; Pearce, N. (2009). Leukaemia and
occupation: A New Zealand Cancer Registry-based case-control Study. Int J Epidemiol
38: 594-606. http://dx.doi.org/10.1093/iie/dvn220.
MDPH (Massachusetts Department of Public Health). (1997). Woburn childhood leukemia
follow-up study: Information booklet. Available online at
http://www.mass.gov/Eeohhs2/docs/dph/environmental/investigations/woburn_childhood
leukemia_follow.pdf (accessed
Mellemgaard, A; Engholm, G: McLaughlin, JK: Olsen, JH. (1994). Occupational risk factors for
renal-cell carcinoma in Denmark. Scand J Work Environ Health 20: 160-165.
Mester, B; Meters, A; Deeg, E; Eisner, G: Becker, N: Seidler, A. (2006). Occupation and
malignant lymphoma: A population based case control study in Germany. Occup Environ
Med 63: 17-26. http://dx.doi.org/10.1136/oem.2005.020453.
Miettinen, OS. (1970). Matching and design efficiency in retrospective studies. Am J Epidemiol
91: 111-118.
Miettinen, OS. (1976). Stratification by a multivariate confounder score. Am J Epidemiol 104:
609-620.
Miligi, L: Seniori, CA; Crosignani, P: Fontana, A; Masala, G: Nanni, O: Ramazzotti, V: Rodella,
S: Stagnaro, E: Tumino, R; Vigano, C: Vindigni, C: Vineis, P. (1999). Occupational,
environmental, and life-style factors associated with the risk of hematolymphopoietic
malignancies in women. Am J Ind Med 36: 60-69. http://dx.doi.org/10.1002/(SICI) 1097-
0274O99907)36:1<60::AID-AJIM9>3.0.CO:2-Z.
B-198
-------
Miligi, L; Costantini, AS: Benvenuti, A; Kriebel, D: Bolejack, V: Tutnino, R; Ramazzotti, V:
Rodella, S: Stagnaro, E; Crosignani, P; Amadori, D; Mirabelli, D: Sommani, L; Belletti,
I: Troschel L: Romeo. L: Micell G: Tozzl GA: Mendico. I: Vineis. P. (2006).
Occupational exposure to solvents and the risk of lymphomas. Epidemiology 17: 552-
561. http://dx.doi.org/10.1097/01.ede.0000231279.30988.4d.
Morton. LM: Hartge. P: Holford. TR: Holly. EA: Chiu. BC: Vineis. P: Stagnaro. E: Willett EV:
Franceschi, S: La Vecchia, C: Hughes, AM; Cozen, W: Davis, S: Severson, RK;
Bernstein, L: Mayne, ST; Dee, FR; Cerhan, JR; Zheng, T. (2005). Cigarette smoking and
risk of non-Hodgkin lymphoma: A pooled analysis from the International Lymphoma
Epidemiology Consortium (InterLymph). Cancer Epidemiol Biomarkers Prev 14: 925-
933. http://dx.doi.org/10.1158/1055-9965.EPI-04-0693.
Morton, W: Marjanovic, D. (1984). Leukemia incidence by occupation in the Portland-
Vancouver metropolitan area. Am J Ind Med 6: 185-205.
Murphy, JP. (1990). Exposure to Wells G and H in Woburn, Massachusetts. Boston, MA:
Massachusetts Health Research Institute, Inc. and Massachusetts Department of Public
Health.
http://www.mass.gov/Eeohhs2/docs/dph/environmental/investigations/woburn_exposure_
wells.pdf.
Pannett B; Coggon, D: Acheson, E. (1985). A job-exposure matrix for use in population based
studies in England and Wales. Br J Ind Med 42: 777-783.
Parent, ME; Hua, Y; Siemiatycki, J. (2000). Occupational risk factors for renal cell carcinoma in
Montreal. Am J Ind Med 38: 609-618.
Paulu, C: Aschengrau, A; Ozonoff D. (1999). Tetrachloroethylene-contaminated drinking water
in Massachusetts and the risk of colon-rectum, lung, and other cancers. Environ Health
Perspect 107: 265-271.
Paulu, C: Aschengrau, A; Ozonoff D. (2002). Exploring associations between residential
location and breast cancer incidence in a case-control study. Environ Health Perspect
110:471-478.
Pearce, NE; Smith, AH; Fisher, DO. (1985). Malignant lymphoma and multiple myeloma linked
with agricultural occupations in a New Zealand Cancer Registry-based study. Am J
Epidemiol 121:225-237.
Pearce, NE: Sheppard, RA; Smith, AH; league, CA. (1987). Non-Hodgkin's lymphoma and
farming: An expanded case-control study. Int J Cancer 39: 155-161.
http://dx.doi.org/10.1002/iic.2910390206.
Pearce, NE: Smith, AH; Reif, JS. (1988). Increased risks of soft tissue sarcoma, malignant
lymphoma, and acute myeloid leukemia in abattoir workers. Am J Ind Med 14: 63-72.
http://dx.doi.org/10.1002/ajim.4700140108.
Peplonska, B; Stewart, P: Szeszenia-Dabrowska, N; Rusiecki, J: Garcia-Closas, M; Lissowska, J:
Bardin-Mikolajczak, A; Zatonski, W: Gromiec, J: Brzeznicki, S: Brinton, L: Blair, A.
(2007). Occupation and breast cancer risk in Polish women: A population-based case-
control study. Am J Ind Med 50: 97-111. http://dx.doi.org/10.1002/ajim.20420.
Pesch. B: Haerting. J: Ranft U: Klimpel. A: Oelschlagel B: Schill. W. (2000a). Occupational
risk factors for renal cell carcinoma: Agent-specific results from a case-control study in
Germany. Int J Epidemiol 29: 1014-1024. http://dx.doi.Org/10.1093/iie/29.6.1014.
B-199
-------
Pesch. B: Haerting. J: Ranft U: Klimpel A: Oelschlagel B: Schill W. (2000b). Occupational
risk factors for urothelial carcinoma: Agent-specific results from a case-control study in
Germany. Int J Epidemiol 29: 238-247. http://dx.doi.Org/10.1093/ije/29.2.238.
Pike, MC: Casagrande, J: Smith, PG. (1975). Statistical analysis of individually matched case-
control studies in epidemiology: Factor under study a discrete variable taking multiple
values. Br JPrev Soc Med 29: 196-201.
Pohlabeln, H; Boffetta, P: Ahrens, W: Merletti, F: Agudo, A; Benhamou, E; Benhamou, S:
Bruske-Hohlfeld, I: Ferro, G: Fortes, C: Kreuzer, M; Mendes, A; Nyberg, F: Pershagen,
G: Saracci, R: Schmid, G: Siemiatycki, J: Simonato, L; Whitley, E; Wichmann, HE;
Winck, C: Zambon, P; Jockel, KH. (2000). Occupational risks for lung cancer among
nonsmokers. Epidemiology 11: 532-538.
Porru, S: Aulenti, V: Donato, F: Boffetta, P: Fazioli, R; Cosciani Cunico, S: Alessio, L. (1996).
Bladder cancer and occupation: A case-control study in northern Italy. Occup Environ
Med 53: 6-10.
Pukkala, E: Martinsen, J: Lynge, E: Gunnarsdottir, H; Sparen, P: Tryggvadottir, L: Weiderpass,
E: Kjaerheim, K. (2009). Occupation and cancer - follow-up of 15 million people in five
Nordic countries. Acta Oncol 48: 646-790.
http://dx.doi.org/10.1080/02841860902913546.
Radican, L: Blair, A; Stewart, P: Wartenberg, D. (2008). Mortality of aircraft maintenance
workers exposed to trichloroethylene and other hydrocarbons and chemicals: Extended
follow-up. J Occup Environ Med 50: 1306-1319.
http://dx.doi.org/10.1097/JOM.Ob013e3181845f7f
Rebelakos, A; Trichopoulos, D: Tzonou, A; Zavitsanos, X; Velonakis, E: Trichopoulos, A.
(1985). Tobacco smoking, coffee drinking, and occupation as risk factors for bladder
cancer in Greece. J Natl Cancer Inst 75: 455-461.
Reif, J: Pearce, N; Kawachi, I; Fraser, J. (1989). Soft-tissue sarcoma, non-Hodgkin's lymphoma
and other cancers in New Zealand forestry workers. Int J Cancer 43: 49-54.
Reulen, R: Kellen, E: Buntinx, F: Zeegers, M. (2007). Bladder cancer and occupation: A report
from the Belgian case-control study on bladder cancer risk. Am J Ind Med 50: 449-454.
http://dx.doi.org/10.1002/aiim.20469.
Richiardi, L: Boffetta, P: Simonato, L: Forastiere, F: Zambon, P: Fortes, C: Gaborieau, V:
Merletti, F. (2004). Occupational risk factors for lung cancer in men and women: A
population-based case-control study in Italy. Cancer Causes Control 15: 285-294.
http://dx.doi.Org/10.1023/B:CACO.0000024223.91059.ed.
Robra, BP; Seidler, A. (1994). A job-exposure matrix for exposure to solvents and other
industry-related agents. Available online at (accessed
Ruder, AM; Ward, EM; Brown, DP. (1994). Cancer mortality in female and male dry-cleaning
workers. J Occup Med 36: 867-874.
Ruder, AM; Ward, EM; Brown, DP. (2001). Mortality in dry-cleaning workers: An update. Am J
Ind Med 39: 121-132. http://dx.doi.org/10.1002/1097-0274(200102)39:2<121::AID-
AJIM1000>3.0.CO:2-H.
Schenk. M: Purdue. MP: Colt JS: Hartge. P: Blair. A: Stewart. P: Cerhan. JR: De Roos. AJ:
Cozen, W: Severson, RK. (2009). Occupation/industry and risk of non-Hodgkin's
lymphoma in the United States. Occup Environ Med 66: 23-31.
http://dx.doi.org/10.1136/oem.2007.036723.
B-200
-------
Scherr, PA; Hutchison, GB; Neitnan, RS. (1992). Non-Hodgkin's lymphoma and occupational
exposure. Cancer Res 52: 5503s-5509s.
Schlehofer, B; Heuer, C; Blettner, M; Niehoff D; Wahrendorf J. (1995). Occupation, smoking
and demographic factors, and renal cell carcinoma in Germany. Int J Epidemiol 24: 51-
57. http://dx.doi.0rg/10.1093/iie/24.l.51.
Schlesselman, JJ. (1982). Case control studies: Design, conduct, analysis. New York, NY:
Oxford University Press.
Schoenberg, JB; Stemhagen, A; Mogielnicki, AP; Altaian, R; Abe, T; Mason, TJ. (1984). Case-
control study of bladder cancer in New Jersey. I. Occupational exposures in white males.
J Natl Cancer Inst 72: 973-981.
Seidler. A: Mohner. M: Berger. J: Mester. B: Deeg. E: Eisner. G: Meters. A: Becker. N. (2007).
Solvent exposure and malignant lymphoma: A population-based case-control study in
Germany. J Occup Med Toxicol 2: 2. http://dx.doi.org/10.1186/1745-6673-2-2.
Selden, AI; Ahlborg, G. (2011). Cancer morbidity in Swedish dry-cleaners and laundry workers:
Historically prospective cohort study. Int Arch Occup Environ Health 84: 435-443.
http://dx.doi.org/10.1007/s00420-010-0582-7.
Serra, C: Bonfill, X; Sunyer, J: Urrutia, G: Turuguet D: Bastus, R: Rogue, M: 'tMannetje, A;
Kogevinas. M: Occidental WGotSoBCitCoV. (2000). Bladder cancer in the textile
industry. Scand J Work Environ Health 26: 476-481.
Shu. XO: Stewart. P: Wen. WO: Han. D: Potter. JD: Buckley. JD: Heineman. E: Robison. LL.
(1999). Parental occupational exposure to hydrocarbons and risk of acute lymphocytic
leukemia in offspring. Cancer Epidemiol Biomarkers Prev 8: 783-791.
Siemiatycki, J: Wacholder, S: Richardson, L; Dewar, R: Gerin, M. (1987). Discovering
carcinogens in the occupational environment: Methods of data collection and analysis of
a large case-referent monitoring system. Scand J Work Environ Health 13: 486-492.
Siemiatycki, J, . (Ed.). (1991). Risk factors for cancer in the workplace. Boca Raton, FL: CRC
Press.
Silverman, DT: Levin, LI; Hoover, RN: Hartge, P. (1989a). Occupational risks of bladder cancer
in the United States: I. White men. J Natl Cancer Inst 81: 1472-1480.
http://dx.doi.org/10.1093/inci/81.19.1472.
Silverman, DT: Levin, LI; Hoover, RN. (1989b). Occupational risks of bladder cancer in the
United States: II. Nonwhite men. JNatl Cancer Inst 81: 1480-1483.
http://dx.doi.org/10.1093/inci/81.19.1480.
Silverman, DT: Levin, LI; Hoover, RN. (1990). Occupational risks of bladder cancer among
white women in the United States. Am J Epidemiol 132: 453-461.
Smith, EM; Miller, ER; Woolson, RF; Brown, CK. (1985). Bladder cancer risk among laundry
workers, dry cleaners, and others in chemically-related occupations. J Occup Med 27:
295-297.
Spence. LA: Aschengrau. A: Gallagher. LE: Webster. TF: Heeren. TC: Ozonoff. DM. (2008).
Evaluation of the Webler-Brown model for estimating tetrachloroethylene exposure from
vinyl-lined asbestos-cement pipes. Environ Health Global Access Sci Source 7: 24.
http://dx.doi.org/10.1186/1476-069X-7-24.
Spirtas. R: Stewart. PA: Lee. JS: Marano. DE: Forbes. CD: Grauman. DJ: Pettigrew. HM: Blair.
A; Hoover, RN: Cohen, JL. (1991). Retrospective cohort mortality study of workers at an
aircraft maintenance facility: I. Epidemiological results. Br J Ind Med 48: 515-530.
http://dx.doi.0rg/10.1136/oem.48.8.515.
B-201
-------
Steineck. G: Plato. N: Gerhardsson. M: Norell. SE: Hogstedt C. (1990). Increased risk of
urothelial cancer in Stockholm during 1985-87 after exposure to benzene and exhausts.
IntJ Cancer 45: 1012-1017. http://dx.doi.org/10.1002/iic.2910450605.
Stemhagen, A; Slade, J: Altaian, R; Bill, J. (1983). Occupational risk factors and liver cancer: A
retrospective case-control study of primary liver cancer in New Jersey. Am J Epidemiol
117:443-454.
Stewart. PA: Lee. JS: Marano. DE: Spirtas. R: Forbes. CD: Blair. A. (1991). Retrospective
cohort mortality study of workers at an aircraft maintenance facility: II. Exposures and
their assessment. Br J Ind Med 48: 531-537.
Suarez, L; Weiss, NS: Martin, J. (1989). Primary liver cancer death and occupation in Texas. Am
J Ind Med 15: 167-175.
Sung. TI: Chen. PC: Lee. LJ. -H: Lin. YP: Hsieh. GY: Wang. JD. (2007). Increased standardized
incidence ratio of breast cancer in female electronics workers. BMC Public Health 7:
102. http://dx.doi.org/10.1186/1471-2458-7-102.
Sung, TI; Wang, JD: Chen, PC. (2008). Increased risk of cancer in the offspring of female
electronics workers. Reprod Toxicol 25: 115-119.
http://dx.doi.0rg/10.1016/j.reprotox.2007.08.004.
Swanson, GM; Burns, PB. (1995). Cancer incidence among women in the workplace: A study of
the association between occupation and industry and 11 cancer sites. J Occup Environ
Med 37: 282-287.
Teschke, K; Morgan, MS: Checkoway, H; Franklin, G: Spinelli, JJ: van Belle, G: Weiss, NS.
(1997). Surveillance of nasal and bladder cancer to locate sources of exposure to
occupational carcinogens. Occup Environ Med 54: 443-451.
Travier. N: Gridlev. G: De Roos. AJ: Plato. N: Moradi. T: Boffetta. P. (2002). Cancer incidence
of dry cleaning, laundry and ironing workers in Sweden. Scand J Work Environ Health
28: 341-348.
Tukey, JW. (1991). Use of many covariates in clinical trials. International Statistical Review 59:
123-137.
Vartiainen, T; Pukkala, E; Rienoja, T; Strandman, T; Kaksonen, K. (1993). Population exposure
to tri- and tetrachloroethene and cancer risk: Two cases of drinking water pollution.
Chemosphere 27: 1171-1181. http://dx.doi.org/10.1016/0045-6535(93)90165-2.
Vaughan, TL; Stewart, PA; Davis, S; Thomas, DB. (1997). Work in dry cleaning and the
incidence of cancer of the oral cavity, larynx, and oesophagus. Occup Environ Med 54:
692-695.
Vieira, V: Aschengrau, A; Ozonoff D. (2005). Impact of tetrachloroethylene-contaminated
drinking water on the risk of breast cancer: Using a dose model to assess exposure in a
case-control study. Environ Health 4: 3. http://dx.doi.org/10.1186/1476-069X-4-3.
Vineis, P; Magnani, C. (1985). Occupation and bladder cancer in males: A case-control study.
Int J Cancer 35: 599-606. http://dx.doi.org/10.1002/iic.2910350506.
Waksberg, J. (1978). Sampling methods for random digit dialing. J Am Stat Assoc 73: 40-46.
Webler, T: Brown, HS. (1993). Exposure to tetrachloroethylene via contaminated drinking water
pipes in Massachusetts: A predictive model. Arch Environ Health 48: 293-297.
http://dx.doi.org/10.1080/00039896.1993.9936716.
Whittemore, AS. (1983). Estimating attributable risk from case-control studies. Am J Epidemiol
117: 76-85.
B-202
-------
Wilson. R: Donahue. M: Gridlev. G: Adami. J: El Ghormli. L: Dosemeci. M. (2008). Shared
occupational risks for transitional cell cancer of the bladder and renal pelvis among men
and women in Sweden. Am J Ind Med 51: 83-99. http://dx.doi.org/10.1002/ajim.20522.
Zheng, T; Cantor, KP; Zhang, Y; Lynch, CF. (2002). Occupation and bladder cancer: A
population-based, case-control study in Iowa. J Occup Environ Med 44: 685-691.
B-203
-------
APPENDIX C. CONSISTENCY OF TETRACHLOROETHYLENE
AND TRICHLOROACETIC ACID HEPATOCARCINOGENICITY
Trichloroacetic acid (TCA), a metabolite of tetrachloroethylene, is associated with
hepatocarcinogenicity in male and female mice (DeAngelo et al., 2008; Bull et al., 2002; Pereira,
1996: Ferreira-Gonzalez et al.. 1995: Daniel etal.. 1993: JISA, 1993: Bull etal.. 1990: Herren-
Frevmdetal.. 1987: NTP, 1986: NCL 1977). There has been some suggestion that TCA does not
account for all of the toxicity observed with tetrachloroethylene exposure (Clewell et al., 2005:
Buben and O'Flaherty, 1985), while others have suggested that TCA can account for liver tumors
induced by tetrachloroethylene (Sweeney et al., 2009). The purpose of this investigation is to
examine quantitatively what fraction of tetrachloroethylene hepatocarcinogenicity may be
associated with TCA, using tetrachloroethylene and TCA bioassay data along with updated
physiologically based pharmacokinetic (PBPK) model-based predictions.
C.I. Methods
C.I.I. Response Data
Because of the more robust liver tumor response in males, and because the PBPK
modeling was calibrated exclusively to data in male mice, only response data in male mice were
considered. Table C-l provides the hepatocellular adenoma or carcinoma incidence data from
the two tetrachloroethylene inhalation bioassays considered in this assessment, National
Toxicology Program (NTP, 1986) and Japan Industrial Safety Association (JISA, 1993) (for
convenience, the studies will be referred to in the remainder of this appendix as the NTP and
JISA studies). These were previously described in Section 5, and so are not discussed further
here.
EPA generally emphasizes combining hepatocellular adenomas and carcinomas in
developing cancer risk values, for three reasons: (1) hepatocellular adenomas develop from the
same cell lines as carcinomas and can progress to carcinomas; (2) adenomas are often
distinguished from carcinomas only on the basis of size; and (3) histopathologic decision criteria
may vary between laboratories or over time.
Table C-2 summarizes data from the available TCA studies considered for carrying out
dose-response modeling. A number of these TCA studies lack information for a complete
comparison of hepatocarcinogenicity between tetrachloroethylene and TCA, either in terms of
exposure (i.e., drinking water intake not reported so total intake of TCA cannot be calculated) or
C-l
-------
in terms of responses. In particular, most of the TCA studies either did not consider adenomas or
did not report combined incidence of adenomas and carcinomas. Lacking data on adenomas, the
studies that only provided carcinoma incidence may under-represent hepatocellular tumor
incidence. For studies not reporting combined incidence of adenomas and carcinomas, there
could be some double-counting of animals when the separate totals of adenomas and carcinomas
are added together. Only the chronic (104-week) study of DeAngelo et al. (2008) reported data
on both total TCA intake as well as combined incidences of adenomas and carcinomas, and only
this study was considered further for comparing with the tetrachloroethylene bioassays in male
mice. However, one significant limitation of the DeAngelo et al. (2008) study is the high
incidence of liver tumors in control animals as compared to other TCA studies and as compared
to the historical background rates of these tumors in B6C3Fi mice. This raises concerns about
the representativeness of the DeAngelo et al. (2008) study for comparison to other studies.
Nonetheless, because this is the only available chronic-duration study that reports all the data
needed, it was used for this analysis.
Table C-l. Incidence of hepatocellular adenomas and carcinomas in male
B6C3Fi mice exposed to tetrachloroethylene in two inhalation bioassays
Sex
Male
Bioassay
NTP
(1986)
JISA
(1993)
Administered
exposures
(ppm)
0
100
200
0
10
50
250
Cumulative liver tumor incidence at Week 104
Adenomas
12
8
19
7
13
8
26
Carcinomas
7
25
26
7
8
12
25
Adenomas
or
carcinomas
17
31
41
13
21
19
40
Total at risk"
49
47
50
46
49
48
49
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.
C-2
-------
Table C-2. 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
et al. (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
0.06C
0.7C
Average TCA
intake (mg/kg-day)
330
0
170
330
0
NR
NR
0
NR
0
NR
0
6.7d
81.2d
N
11
35
11
24
20
20
20
22
22
16b
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
3b
8
26
14
32
Incidence of
adenomas or
carcinomas
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
o
oo
a Cumulative TCA exposures were provided in g/kg for the mice evaluated at 52 wk. Those exposures were converted to mg/kg-day by multiplying by
(1,000 mg/g)/(7 d/wk * 52 wk).
b Estimated from the reported proportion responding by selecting the smallest group size and incidence value consistent with the precision of the reported
proportion.
0 Measured concentrations—nominal concentrations were 0.05 and 0.5 mg/L.
d Calculated based on measured concentrations (different than those reported in the manuscript, which were based on nominal concentrations).
NR = not reported.
-------
C.1.2. Exposure-Level Conversions
TCA bioassay exposures were generally reported in terms of water concentration, in
mg/L or mmol/L. Table C-2 provides the exposure levels as reported by each set of authors.
Some reports provided mg/kg-day equivalents. To account for TCA bioavailability, the best
estimates of bioavailability from Chiu (2011) were used. These are modeled as an "effective"
concentration Ceff that changes as a function of actual concentration C: Ceff = Cmax x C/(d/2 + C),
where Cmax is a "maximal" effective concentration, and G/2 is the actual concentration where the
effective concentration is half the maximal value. Thus, the rate of TCA absorption will be
given by Ceff x Drinking water intake rate = (Ceg/C) x TCA intake. The best fit values from
Chiu (2011) of the parameters are Cmax = 1.34 mg/L and G/2 = 1.82 mg/L. The area under the
curve (AUC) of TCA in the liver is also calculated using the Chiu (2011) PBPK model for TCA.
This model is based on a model previously calibrated to TCA kinetic data from trichloroethylene
exposure and TCA oral gavage and i.v. exposures, and subsequently updated by Chiu (2011)
with TCA kinetic data from drinking water exposures [the same data used by Sweeney et al.
(2009)1. The Chiu and Ginsberg (2011) PBPK model was used to estimate the AUC of TCA in
the liver corresponding to the bioassay exposures in the NTP and JISA studies. As shown in
Table C-3, accounting for reduced TCA bioavailability, the exposures in the DeAngelo et al.
(2008) study lead to internal doses that are within the range of the internal doses predicted for the
tetrachloroethylene inhalation bioassays.
Table C-3. PBPK model-estimated TCA internal dose measures for
tetrachloroethylene and TCA bioassays used in analysis
Study
NTP
(1986)
JISA
(1993)
DeAngelo
etal.
(2008)
TV
49
47
50
46
49
48
49
56
48
51
Exposure
group
0 ppm PCE
lOOppmPCE
200 ppm PCE
0 ppm PCE
10 ppm PCE
50 ppm PCE
250 ppm PCE
0 g/L TCA
0.06 g/L TCA
0.7 g/L TCA
Proportion
with
adenomas or
carcinomas
0.35
0.66
0.82
0.28
0.43
0.40
0.82
0.55
0.44
0.71
TCA
absorbed"
(mg/kg-day)
-
-
-
—
-
-
-
0
5.9
53
TCA
produced
from PCEb
(mg/kg-day)
0
29
53
0
4.5
16
64
—
-
-
AUC of free
TCA in
plasmaa'b
(mg-hr/L-day)
0
454
834
0
73
260
1,043
0
74
666
AUC of TCA
in livera'b
(mg-hr/L-day)
0
487
895
0
78
280
1,121
0
58
526
a Calculated using PBPK model of Chiu (2011). using posterior mean parameter estimates and best-fit estimate of
the fractional absorption from drinking water (88% at 0.6 g/L and 66% at 0.7 g/L).
b Calculated using PBPK model of Chiu and Ginsberg (2011). using highest posterior mode parameter estimates.
C-4
-------
C.1.3. Dose-Response Modeling and Statistical Analysis
Due to significantly different tumor incidences in control groups across the TCA and
tetrachloroethylene data sets, combined analysis using standardized software such as BMDS is
not feasible. Therefore, to test the consistency of the dose-response relationships, a standard
statistical approach for binomial data is used. In particular, logistic regression was applied to the
various data sets, with potential independent variables of dose, chemical, study, and their
interactions (products). For a ^-length vector of independent variable jc, and a probability of
effectp(x), the logistic model is defined by
(*)]) (C-l)
z(x) = A, + /SlXl + #8*2 +...+$** (C-2)
with parameters fyj =1 ... k. This is equivalent to linear regression with binomial variance of
the variable z, which is the natural log of the odds of a effect \n\p/(l-p)].
For the purposes of this analysis, flo is called the "intercept," fi\ is called the "slope" with
respect to the dose metric x\, x% = 0 or 1 is either the study (NTP = 0, JISA = 1, for
tetrachloroethylene data) or the chemical (tetrachloroethylene = 0, TCA =1) with /?2 the
corresponding regression coefficient (i.e., "different intercepts"), and x3 = x\ x x2 is the
interaction term (i.e., "different slopes"). In this analysis, all ft coefficients are unconstrained.
This model was implemented using the "glm" function in the R statistical package
(version 12.2.1), which reports optimized estimates, standard errors, and ^-values for each
regression parameter. In addition, analysis of variance (ANOVA) using the "anova" function in
R (with a chi-squared test statistic) can also be used to determine whether additional regression
parameters produce significant improvement in model fit.
Of particular interest is the "interaction term" with regression coefficient/^. In the event
that this parameter is statistically significant (p-value < 0.05, confirmed by ANOVA), then this
would be evidence that tetrachloroethylene and TCA have different dose-responses as a function
of internal TCA dose (after accounting for possible differences in control incidences). In the
event that this parameter is not statistically significant, it is well known that failure to reject a
null hypothesis of no effect may simply be the result of low statistical power. While there is a
large literature on "postexperiment power calculations," Hoenig and Heisey (2001) show that
such an approach is fundamentally flawed and yield no further insights than confidence intervals
(CIs). Therefore, the CI of the /?3 parameter is used as an indication of the range of possible
contributions of TCA to tetrachloroethylene hepatocarcinogenesis.
C-5
-------
C.2. Results
C.2.1. Logistic Model Fits to Individual Data Sets
Each bioassay was fit individually (using an intercept/?o and a slope/?i) and goodness-of-
fit evaluated using the chi-squared test on the residuals (see Table C-4, Figures C-l A and C-2A).
All/>-values were >0.15, suggesting the logistic model is an adequate description of the data for
the purposes of this comparison.
C.2.2. Consistency of NTP and JISA Data
Analysis was conducted on the NTP and JISA studies to see if they are consistent, which
would increase statistical power in the subsequent analysis of the contribution of TCA.
Three logistic models were fit (see Table C-4, Figures C-1B, C-1C, and C-1D). Neither the
study intercept/^ nor the study-dose interaction term/?3 was statistically significant (CIs
overlapped with 0, /^-values from ANOVA were 0.17 and 0.14, respectively). Thus, the null
hypothesis that the two bioassays for tetrachloroethylene have a common slope and intercept
cannot be rejected.
As suggested by Hoenig and Heisey (2001), looking at the CIs gives insight into the
power to reject this null hypothesis. The difference in intercepts (fli) has a 95% confidence
region (based on ± 1.96 x standard error) of (-0.81, 0.14), implying that the odds of a tumor in
JISA control animals is between 0.44- and 1.16-fold that of NTP control animals. The difference
in slopes (fii) has a 95% confidence region of (-1.60, 0.23), implying that the odds ratio for
JISA-exposed animals is between 0.20- and 1.26-fold that of NTP animals with equivalent AUCs
of TCA in the liver. These ranges are quite large—up to twofold difference in background odds,
and up to a fivefold difference in exposed/unexposed odds ratios cannot be ruled out by the
available data. Nonetheless, for the purposes of further analysis, the NTP and JISA studies are
combined, as these provide greater statistical power for determining the extent to which the TCA
bioassay is consistent with the tetrachloroethylene bioassays on the basis of TCA internal dose.
C-6
-------
Table C-4. Logistic regression model fits—beta coefficients and standard
errors
Data analyzed
NTP (1986)
JISA (19931
DeAngelo et al.
(2008)
NTP (19861 +
JISA (1993)
NTP (1986) +
JISA (1993) +
DeAngelo et al.
(2008)
#>
(Intercept)
-0.599 ± 0.280
-0.763 ±0.199
-0.028 ± 0.207
-0.696 ±0.162
-0.482 ± 0.223
-0.709 ±0.162
-0.428 ±0.125
-0.665 ±0.155
-0.696 ±0.162
1,000 x/?t
(Dose
coefficient)
2.43 ±0.53
1.97 ±0.39
1.64 ±0.72
2.21 ±0.32
2.14 ±0.32
2.59 ±0.42
1.95 ±0.28
2.12 ±0.29
2.21 ±0.32
P2
(Study or
chemical
intercept)
Not included
Not included
Not included
Not included
-0.334 ± 0.244°
Not included
Not included
0.556 ±0.211
0.668 ± 0.262
1,000 x/?3
(Dose x [study
or chemical])
Not included
Not included
Not included
Not included
Not included
-0.683 ± 0.467°
Not included
Not included
-0.57 2 ±0.790°
Chi-
squared
goodness
-of-fit
/7-valueb
0.75
0.37
0.15
0.47
0.77
0.80
0.068
0.40
0.36
Figure
C-1A
C-1A
C-2A
C-1B
C-1C
C-1D
C-2B
C-2C
C-2D
aLogistic model: proportion responding =p (x) = 1/(1 + exp[-z(x)]), z(x) =/?0 + fi\x\ + $2X2+^3X1X2,.
where
xi = dose
x2 = study (NTP = 0, JISA = 1) for analysis of NTP (1986) + JISA (1993)
x2 = chemical (PCE [NTP or JISA] = 0, TCA [DeAngelo] = 1) for analysis of NTP (1986) + JISA
(1993) + DeAngelo et al. (2008).
b Chi-squared percentage point at the sum-squared residuals (weighted by inverse binomial variance), with degrees
of freedom equal to the number of data points minus the number of parameters.
0 Parameters in italics are not significant (p > 0.05) either by regression CI or by ANOVA.
C-7
-------
o
Q.
ion
ac
oo
o
CD
ci
C\l
ci
CI
ci
A
1
1
1
o
Q.
ion
ac
oo
ci
CD
ci
c\i
ci
p
ci
B
200 400 600
AUC TCA in liver
800
200 400 600 800 1000
AUC TCA in liver
o
oo
ci
D)
^
C CD
| o-
CD
c -sr
1 °-
ro
M—
CM
ci
O
ci
C
I s~'~'^L-
,-"'!..-•"' 1
T '''' ••'"""
.-'>••""'""
T ''' -''"''
rt--:"f
T
1
1 1 1 1 1 1
0 200 400 600 800 1000
o
oo
ci
D)
^
!= CD
| o -
CD
c -^r
1 °-
ro
M—
CM
ci
O
ci
D
-•!"''" I.-
''" i ,.-••"" i
1 ,'' _..-•-•"'
/- ..--"
}fs' m..--~~
r ^
'
x
i i i i i
0 200 400 600 800 1000
AUC TCA in liver
AUC TCA in liver
Figure C-l. Logistic regression dose-response fits to tetrachloroethylene data
[open circle: JISA (1993); filled circle: NTP (1986)1.
A: separate model fits to each dataset; B: single model fit to both data sets; C: model with separate
intercepts and common slope; D: model with common intercept and separate slopes.
See Table C-4 for parameter values, standard errors, and goodness-of-fit ^-values.
-------
C.2.3. Consistency of Tetrachloroethylene and TCA Data
Analysis was conducted on the combined NTP and JISA studies and the DeAngelo et al.
(2008) TCA study to see if they are consistent. Three logistic models were fit (see Table C-4,
Figures C-2B, C-2C, and C-2D). The chemical intercept term fa was statistically significant
(parameter CI did not overlap with 0,/>-value from ANOVA was 0.009). Thus, the null
hypothesis that the two bioassays for tetrachloroethylene have a common intercept is rejected.
The chemical slope term ^3 was not statistically significant (parameter CI overlapped with 0,
/7-value from ANOVA was 0.46). Thus, the null hypothesis that the TCA and
tetrachloroethylene bioassays have a common slope (after accounting for different background
rates) cannot be rejected.
As suggested by Hoenig and Heisey (2001), looking at the CIs gives insight into the
power to reject this null hypothesis. The difference in slopes (^3) has a 95% confidence region
of-2.12, 0.98, implying that the odds ratio for TCA exposed/unexposed animals is between
0.12- and 2.65-fold that of tetrachloroethylene-exposed/unexposed animals with equivalent
AUCs of TCA in the liver. These ranges are quite large—up to an eightfold difference in
exposed/unexposed odds ratios cannot be ruled out by the available data. Moreover, these CIs
assume that the tetrachloroethylene data reflect a common dose-response—relaxing this
assumption would lead to wider CIs for the relative odds ratios between TCA and
tetrachloroethylene bioassays.
C-9
-------
D)
C
T3
o
Q.
o
co
oq
o
CD
ci
CNI
ci
p
ci
A
I I I I I
200 400 600 800 1000
AUC TCA in liver
D)
C
T3
o
o.
oq
o
CD
ci
~ O
O
CO
it
CNI
ci
CD
ci
I I I I I
200 400 600 800 1000
AUC TCA in liver
8.
O
oo
ci
CD _
°
ci ~
CNI
ci
O
ci
C
.--•""I '"'J"
.--•"' ^f" 1
-•Tip i ' x''""
.-•'' ±T ^^""
-""" T *'
T .-" A x
IB .-•'' ^''
!--Tr x'''
pi
o
oo
ci
D)
^
!= CD
o _• —
Q. °
CD
Ł= ^]-
5 o -
O
M—
CNI
ci
O
ci
D
i ..•;•>••*"%'
'" 1
Tit--"'' !-'''
...-•T *"'
-'--" T **
T •-' /^x
m .--" x'
T--TV' x''
i1'"l
ci
1 1 1 1 1 1 1 1 1 1 1
0 200 400 600 800 1000 0 200 400 600 800 1000
AUC TCA in liver
AUC TCA in liver
Figure C-2. Logistic regression dose-response fits to TCA
[open square: DeAngelo et al. (2008)1 and combined TCA and tetrachloroethylene data [open
circle: JISA (1993): filled circle: NTP (1986)1. A: model fit to TCA data only; B: single model fit
to all data sets; C: model with chemical-specific intercepts and common slope; D: model with
chemical-specific intercepts and chemical-specific slopes. See Table C-4 for parameter values,
standard errors, and goodness-of-fit ^-values.
C-10
-------
C.3. CONCLUSIONS
This analysis suggests that TCA might explain the incidence of carcinomas observed in
the available tetrachloroethylene bioassays but that a wide range of possible contributions cannot
be ruled out by the available data. Specifically, a contribution of TCA from as little as 12 up to
100% cannot be ruled out, under the assumptions that the tetrachloroethylene NTP and JISA
bioassay data can be combined, and using the Chiu and Ginsberg (2011) PBPK model for
tetrachloroethylene and the Chiu (2011) PBPK model for TCA and TCA bioavailability. If
either of these assumptions is relaxed—i.e., given that residual uncertainties of about twofold
exist in the PBPK model predictions for TCA internal dose and that there may be some
underlying differences between the NTP and JISA bioassays—then the CIs will be greater.
Furthermore, the high control tumor incidence reported in the TCA bioassay of DeAngelo et al.
(2008) raise questions as to the representativeness of that bioassay for comparison to
tetrachloroethylene bioassays. Overall, as discussed in Chiu (2011) with regards to the
contribution of TCA to TCE-induced hepatomegaly, factors such as study-to-study experimental
variability in kinetics (e.g., metabolism, bioavailability) or in dynamics (e.g., background tumor
rates), different analytical methods used to quantify TCA in blood and tissues and uncertainty in
TCA dosing patterns in drinking water studies further limit the ability to discern the quantitative
contribution of TCA. A more precise quantitative measure of the relative contribution of TCA to
tetrachloroethylene-induced liver tumors requires an appropriately designed experiment to better
control for these factors.
C-ll
-------
C.4. References
Buben. JA; O'Flaherty. EJ. (1985). Delineation of the role of metabolism in the hepatotoxicity of
trichloroethylene and perchloroethylene: A dose-effect study. Toxicol Appl Pharmacol 78: 105-
122.
Bull. RJ; Sanchez. IM; Nelson. MA; Larson. JL; Lansing. AJ. (1990). Liver tumor induction in B6C3F1
mice by dichloroacetate andtrichloroacetate. Toxicology 63: 341-359.
http://dx.doi.org/10.1016/0300-483X(90)90195-M.
Bull. RJ: Orner. GA: Cheng. RS: Stillwell. L: Stauber. AJ: Sasser. LB: Lingohr. MK: Thrall. BD. (2002).
Contribution of dichloroacetate and trichloroacetate to liver tumor induction in mice by
trichloroethylene. Toxicol Appl Pharmacol 182: 55-65. http://dx.doi.org/10.1006/taap.2002.9427.
Chiu. WA; Ginsberg. GL. (2011). Development and evaluation of a harmonized physiologically based
pharmacokinetic (PBPK) model for perchloroethylene toxicokinetics in mice, rats, and humans.
Toxicol Appl Pharmacol 253: 203-234. http://dx.doi.Org/10.1016/i.taap.2011.03.020.
Chiu. WA. (2011). Trichloroacetic acid: Updated estimates of its bioavailability and its contribution to
trichloroethylene-induced mouse hepatomegaly. Toxicology 285: 114-125.
http://dx.doi.0rg/10.1016/i.tox.2011.04.009.
Clewell HJ; Gentry. PR; Kester. JE; Andersen. ME. (2005). Evaluation of physiologically based
pharmacokinetic models in risk assessment: An example with perchloroethylene. Crit Rev
Toxicol 35: 413-433. http://dx.doi.org/10.1080/10408440590931994.
Daniel. FB; Meier. JR; DeAngelo. AB. (1993). Advances in research on carcinogenic and genotoxic by-
products of chlorine disinfection: Chlorinated hydroxyfuranones and chlorinated acetic acids.
Ann 1st Super Sanita 29: 279-291.
DeAngelo. AB; Daniel. FB; Wong. DM; George. MH. (2008). The induction of hepatocellular neoplasia
by trichloroacetic acid administered in the drinking water of the male B6C3F1 mouse. J Toxicol
Environ Health A 71: 1056-1068. http://dx.doi.org/10.1080/15287390802111952.
Ferreira-Gonzalez. A; DeAngelo. AB; Nasim. S; Garrett CT. (1995). Ras oncogene activation during
hepatocarcinogenesis in B6C3F1 male mice by dichloroacetic and trichloroacetic acids.
Carcinogenesis 16: 495-500. http://dx.doi.Org/10.1093/carcin/16.3.495.
Herren-Freund. SL; Pereira. MA; Khoury. MD; Olson. G. (1987). The carcinogenicity of
trichloroethylene and its metabolites, trichloroacetic acid and dichloroacetic acid, in mouse liver.
Toxicol Appl Pharmacol 90: 183-189. http://dx.doi.org/10.1016/0041-008X(87)90325-5.
Hoenig. JM; Heisey. DM. (2001). The abuse of power: The pervasive fallacy of power calculations for
data analysis. Am Stat 55: 19-24.
JISA (Japan Industrial Safety Association). (1993). Carcinogenicity study of tetrachloroethylene by
inhalation in rats and mice. Hadano, Japan.
NCI (National Cancer Institute). (1977). Bioassay of tetrachloroethylene for possible carcinogenicity.
(NCI-CGTR-13; DREW Publication No. (NIH) 77-813). Bethesda, Md: National Institutes of
Health. http://ntp.niehs.nih.gov/ntp/htdocs/LT rpts/trO 13.pdf.
NTP (National Toxicology Program). (1986). Toxicology and carcinogenesis studies of
tetrachloroethylene (perchloroethylene) (CAS no. 127-18-4) in F344/N rats and B6C3F1 mice
(inhalation studies). (NTP TR 311). Research Triangle Park, NC: U.S. Department of Health and
Human Services, National Toxicology Program.
http://ntp.niehs.nih.gov/ntp/htdocs/LT rpts/tr311 .pdf
Pereira, MA. (1996). Carcinogenic activity of dichloroacetic acid and trichloroacetic acid in the liver of
female B6C3F1 mice. Fundam Appl Toxicol 31: 192-199.
http://dx.doi.org/10.1006/faat.1996.0091.
Sweeney. LM; Kirman. CR; Gargas. ML; Dugard. PH. (2009). Contribution of trichloroacetic acid to
liver tumors observed in perchloroethylene (perc)-exposed mice. Toxicology 260: 77-83.
http://dx.doi.0rg/10.1016/i.tox.2009.03.008.
C-12
-------
APPENDIX D. CANCER DOSE-RESPONSE MODELING
D.I. Model Selection Details for Tumor Sites from JISA (1993)
Table D-l. Model predictions for hepatocellular tumors in male mice (JISA,
1993)a, using several dose metrics and multistage cancer model
Model
stages
Goodness of fit
/7-valueb
Largest standardized
residual(s)
AIC
BMD10
BMDL10
Conclusion
Total liver oxidative metabolism (mg/kg° 75-day)
One
Two
Three
0.24
0.16
0.18
1.1, low-dose
-1.2, mid-dose
-0.7, control
1.1, low-dose
-0.7, control
1.0, low-dose
239.7
240.8
240.6
2.9
6.4
6.5
2.1
2.2
2.2
All three fits were adequate by
conventional criteria.13 There was no
statistical improvement in adding
higher-order coefficients (using
likelihood ratio test); one-stage fit
was selected.
TCA AUC in liver (mg-hr/L-day)
One
Two
Three
0.25
0.17
0.19
1.0, low-dose
-1.2, mid-dose
-0.7, control
1.1, low-dose
-0.7, control
1.0, low-dose
239.7
240.8
240.6
97.1
209.9
213.9
68.8
72.8
73.8
All three fits were adequate by
conventional criteria.13 There was no
statistical improvement in adding
higher-order coefficients (using
likelihood ratio test); one-stage fit
was selected.
Administered tetrachloroethylene concentration (ppm)
One
Two
Three
0.27
0.16
0.17
1.2, low-dose
-1.0, mid-dose
-0.8, control
1.1, low-dose
-0.8, control
1.1, low-dose
239.5
240.9
240.8
3.9
9.0
8.2
2.7
2.8
2.9
All three fits were adequate by
conventional criteria.13 There was no
statistical improvement in adding
higher-order coefficients (using
likelihood ratio test); one-stage fit
was selected.
a Incidence data and human equivalent continuous exposure estimates provided in Table 5-13.
b Goodness-of-fit^-values O.05 for a preferred model, or <0.10 when considering many models, fail to meet
conventional goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered. Best-fit
model is highlighted in bold; output for best-fit models provided in following pages.
AIC = Akaike's Information Criteria, BMD = benchmark dose, BMDL = lower bound benchmark dose.
D-l
-------
D.I.I. Modeling Output for Male Mice, Hepatocellular Tumors (JISA, 1993)
D.I.1.1. With total oxidative metabolism in liver as dose metric
Multistage Cancer Model with 0.95 Confidence Level
I
c
o
't5
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Linearextrapolation
40
45
Figure D-l One-degree multistage model fit to hepatocellular tumors in male
mice (JISA, 1993), with BMD and BMDL at 10% extra risk, using total
oxidative metabolism in liver (mg/kg°'75-day).
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*dose^l)]
The parameter betas are restricted to be positive
D-2
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.53
Beta(l) -0.53 1
Variable
Background
Beta(1)
3td. Err.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
Est. Prob.
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0. 95
2.91314
2.06187
4.49484
Taken together, (2.06187, 4.49484) is a 90
interval for the BMD
-------
D.l.1.2. With TCA AUC in liver as dose metric
Multistage Cancer Model with 0.95 Confidence Level
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Linear extrapolation
1000
Figure D-2. One-degree multistage model fit to hepatocellular tumors in
male mice (JISA, 1993), with BMD and BMDL at 10% extra risk, using TCA
AUC as dose metric (mg-hr/L-d).
Multistage Cancer Model. (Version: 1.7; Date: 05/16/2008)
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*doseAl)]
The parameter betas are restricted to be positive
Dependent variable = Response
Independent variable = Dose
Total number of observations = 4
Total number of records with missing values = 0
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
Default Initial Parameter Values
Background = 0.283935
Beta(l) = 0.00118591
Asymptotic Correlation Matrix of Parameter Estimates
D-4
-------
Background
Beta(l)
Background
1
-0.53
Beta(l)
-0.53
1
Parameter Estimates
Variable
Background
Beta(l)
Estimate
0.299803
0.0010848
Std. Err.
*
* - Indicates that this value is not calculated.
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
* *
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-116.442
-117.833
-132.99
239.666
# Param's
4
2
1
Deviance Test d.f.
2.78303
33.0977
P-value
0.2487
<.0001
Goodness of Fit
Dose
0.0000
78.4900
279.7000
1121.1000
Est . Prob .
0.2998
0.3570
0.4831
0.7925
Expected
13.791
17.491
23.186
38.832
Observed
13.000
21.000
19.000
40.000
Size
46
49
48
49
Scaled
Residual
-0.255
1.046
-1.209
0.411
ChiA2 =2.79
d.f. = 2
P-value = 0.2477
Benchmark Dose Computation
Specified effect =
Risk Type =
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0.95
97.1242
68.7915
149.76
Taken together, (68.7915, 149.76 ) is a 90
interval for the BMD
% two-sided confidence
Multistage Cancer Slope Factor =
0.00145367
D-5
-------
D.I.1.3. With administered tetrachloroethylene concentration (ppm) as dose metric
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linearextrapolation
35
40
45
Figure D-3. One-degree multistage model fit to hepatocellular tumors in
male mice (JISA, 1993), with BMD and BMDL at 10% extra risk, using
administered tetrachloroethylene concentration (ppm).
Multistage Cancer Model. (Version: 1.7; Date: 05/16/2008)
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*dose/xl)]
The parameter betas are restricted to be positive
Dependent variable = Response
Independent variable = Dose
Total number of observations = 4
Total number of records with missing values = 0
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
Default Initial Parameter Values
Background = 0.307193
Beta(l) = 0.0290723
D-6
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.48
Beta(l) -0.48 1
Interval
Variable
Limit
Background
Beta(l)
Estimate
0.316506
0.0273229
Parameter Estimates
Std. Err.
*
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
* - Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-116.442
-117.738
-132.99
239.476
# Param's
4
2
1
Deviance Test d.f.
2.59226
33.0977
P-value
0.2736
<.0001
Dose
0.0000
1.8000
9.0000
45.0000
Est. Prob.
0.3165
0.3493
0.4655
0.8001
Goodness of Fit
Expected Observed Size
14.559
17.116
22.344
39.206
13.000
21.000
19.000
40.000
46
49
48
49
Scaled
Residual
-0.494
1.164
-0.968
0.284
ChiA2 = 2.62
d.f.
P-value = 0.2704
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0.95
3.85613
2.70709
5.98909
Taken together, (2.70709, 5.98909) is a 90
interval for the BMD
two-sided confidence
Multistage Cancer Slope Factor =
0.03694
D-7
-------
Table D-2. Model predictions for hepatocellular tumors in female mice
(JISA, 1993), using several dose metrics3 and multistage cancer model
Model
stage
Goodness of fit
p-valueb
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Comments
Conclusions
Total liver oxidative metabolism (mg/kg° 75-day)
One-stage
Two-stage
Three-stage
0.14
0.82
0.82
-1.4, mid-dose
-0.18, low-dose
-0.18, low-dose
154.9
152.8
152.8
3.7
8.4
8.4
2.8
4.0
3.9
Adequate fit
Adequate fit
Adequate fit
Selected two-
degree multistage,
based on likelihood
ratio test.
TCA AUC in liver (mg-hr/L-day)
One-stage
Two-stage
Three-stage
0.13
0.82
0.82
-1.4, mid-dose
-0.18, low-dose
-0.18, low-dose
155.1
152.9
152.9
129
292
292
98
141
139
Adequate fit
Adequate fit
Adequate fit
Selected two-
degree multistage,
based on likelihood
ratio test.
Administered tetrachloroethylene concentration (ppm)
One-stage
Two-, three-
stage
0.36
0.83
-1.1, mid-dose
-0.1, low-dose
153.0
152.8
5.0
9.7
3.8
4.3
Adequate fit
Identical fits
resulted from both
models
Selected one-
degree multistage;
no statistical
improvement in
adding higher order
parameters.
"Incidence data provided in Table 5-13, and dose metrics provided in Table 5-17; both are included in following
output.
3 Values O.05 for a preferred model, or <0.10 when considering a suite of models, fail to meet conventional
goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered. Best-fit model is
highlighted in bold; output for best-fit models provided in following pages.
D-8
-------
D.1.2. Modeling Output for Female Mice, Hepatocellular Tumors (JISA, 1993)
D.l.2.1. With total oxidative metabolism in liver as dose metric
Multistage Cancer Model with 0.95 Confidence Level
o
I
c
o
'o
S5
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
BMDL
Multistage Cancer
Linear extrapolation
BMP
10
15
dose
20
25
30
Figure D-4. Two-degree multistage model fit to hepatocellular tumors in
female mice (JISA, 1993), with BMD and BMDL at 10% extra risk.
The form of the probability function is:
The parameter betas are restricted to be positive
Total number of observations = 4
Total number of records with missing value
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
D-9
-------
Default Initial Parameter Values
Background = 0.0554081
Beta(l) = 0.00569729
Beta(2) = 0.000883583
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l) Beta(2)
Background 1 -0.69 0.59
Beta(1) -0.69 1 -0.97
Beta(2) 0.59 -0.97 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(1)
Beta(2)
k - Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -73.398 4
Fitted model -73.4233 3 0.050713 1 0.821E
Reduced model -106.26 1 65.7232 3 <.0001
Chi^2 = 0.05 d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD =
BMDL =
BMDU =
Multistage Cancer Slope Factor = 0.0248549
D-10
-------
D.l.2.2. With TCA AUC in liver as dose metric
Multistage Cancer Model with 0.95 Confidence Level
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l) Beta(2)
Background 1 -0.69 0.6
Beta(l) -0.69 1 -0.97
Beta(2) 0.6 -0.97 1
Parameter Estimates
95.0% Wald Confidence Interval
Std. Err. Lower Conf. Limit Upper Conf. Limit
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -73.398 4
Fitted model -73.4249 3 0.0538645 1 0.8165
Reduced model -106.26 1 65.7232 3 <.0001
Scaled
Est. Prob. Expected Observed Size Residual
Chi"2 = 0.05 d.f. = 1 P-value = 0.8177
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 291.833
BMDL = 141.409
BMDU = 402.749
Taken together, (141.409, 402.749) is a 90 % two-sided confidence
interval for the BMD
D-12
-------
D.l.2.3. With administered tetrachloroethylene concentration (ppm) as dose metric
Multistage Cancer Model with 0.95 Confidence Level
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Linearextrapolation
gMDL
10 15
20 25
dose
30 35 40 45
Figure D-6. One-degree multistage model fit to hepatocellular tumors in
female mice (JISA, 1993), with BMD and BMDL at 10% extra risk.
Multistage Cancer Model. (Version: 1.7; Date: 05/16/2008;
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*dose^l)]
The parameter betas are restricted to be positive
Dependent variable = Response
Independent variable = Dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Default Initial Parameter Value
Background = 0.0124442
D-13
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.47
Beta(l) -0.47 1
Parameter Estimates
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -73.398 4
Fitted model -74.4575 2 2.11904 2 0.3466
Reduced model -106.26 1 65.7232 3 <.0001
Scaled
Residual
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD =
BMDL =
BMDU =
D-14
-------
Table D-3. Model predictions for hemangiomas or hemangiosarcomas in
male mice (JISA, 1993), using tetrachloroethylene AUC in blood and
administered tetrachloroethylene concentration as dose metrics3 and
multistage cancer model
Model
stage
Goodness of fit
p-valueb
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Conclusions
Tetrachloroethylene AUC in blood (mg-hr/L-day)
One-, two-,
three-stage
0.38
-1.0, low-dose,
0.9, mid-dose
142.0
63.0
34.3
Fits for all three models were the
same; only the first order term was >0.
Administered tetrachloroethylene concentration (ppm)
One-, two-,
three-stage
0.38
-1.0, low-dose,
0.9, mid-dose
142.0
24.4
13.3
Fits for all three models were the
same; only the first order term was >0.
a Incidence data and human equivalent continuous exposures provided in Table 5-13 and in the output below.
b Values <0.05 for a preferred model, or <0.10 when considering suite of models, fail to meet conventional
goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered. The best-fit model is
highlighted in bold; the output for best-fit models is provided in following pages.
D-15
-------
D.1.3. Modeling Output for Male Mice, Hemangiomas or Hemangiosarcomas (JISA, 1993)
D.l.3.1. With tetrachloroethylene AUC in blood as dose metric
T3
-2
O
g
"o
Multistage Cancer Model with 0.95 Confidence Level
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
Multistage Cancer
Linearextrapolation
BMDL
BMD
20
40
60
dose
80
100
120
Figure D-7. One-degree multistage model fit to hemangioma or
hemangiosarcoma incidence in male mice (JISA, 1993), with BMD and
BMDL at 10% extra risk.
The form of the probability function is:
P[response] = background + (1-background)* [1-EXP(
-betal^dose^l)]
The parameter betas are restricted to be positive
D-16
-------
Default Initial Parameter Values
Background = 0.0779832
Beta(l) = 0.00154747
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.6
Beta(l) -0.6 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(l)
Model
Full model -67.9801 4
Fitted model -69.0102 2
Reduced model -72.3399 1
Dose Est. Prob.
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level =
BMD =
BMDL =
BMDU =
Taken together, (34.3348, 191.7 ) is a 90 % two-sided confidence
interval for the BMD
D-17
-------
D.l.3.2. With administered tetrachloroethylene concentration (ppm) as dose metric
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
MultistageCancerModelwith 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
BMDL
BMD
10
15
20 25
dose
30
35
40
45
Figure D-8. One-degree multistage model fit to hemangioma or
hemangiosarcoma incidence in male mice (JISA, 1993), with BMD and
BMDL at 10% extra risk.
Multistage Cancer Model. (Version: 1.7; Date: 05/16/2008)
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*dose/xl)]
The parameter betas are restricted to be positive
Dependent variable = Response
Independent variable = Dose
Total number of observations = 4
Total number of records with missing values = 0
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
Default Initial Parameter Values
Background = 0.0770402
D-18
-------
Beta(l) = 0.00401128
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.6
Beta(l) -0.6 1
Parameter Estimates
Interval
Variable
Limit
Background
Beta(l)
Estimate
0.0723269
0.00432149
Std. Err.
*
*
Indicates that this value is not calculated.
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-67.9801
-69.001
-72.3399
142.002
# Param's
4
2
1
Deviance Test d.f.
2.04183
8.71962
P-value
0.3603
0.03326
Goodness of Fit
Dose
0.0000
1.8000
9.0000
45.0000
Est. Prob.
0.0723
0.0795
0.1077
0.2363
Expected
3.327
3.896
5.170
11.577
Observed
4.000
2.000
7.000
11.000
Size
46
49
48
49
Scaled
Residual
0.383
-1.001
0.852
-0.194
=1.91
d.f. = 2
P-value = 0.3843
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0.95
24.3806
13.3404
73.8608
Taken together, (13.3404, 73.8608) is a 90
interval for the BMD
two-sided confidence
Multistage Cancer Slope Factor =
0.00749601
D-19
-------
D.l.3.3. Modeling Output for Male Mice (JISA, 1993), Combined Risk of Hepatocellular
Tumors or Hemangiomas/Hemangiosarcomas, at 10% Extra Risk, using
Administered Concentration and Multistage Modeling (Discussed in Section
5.3.4.1)
MS_COMBO. (Version: 1.5 Beta; Date: 01/25/2011)
Input Data File: C:\Usepa\BMDS220\Data\SessionFiles\New.(d)
[For separate model fits of hepatocellular tumors and hemangiomas/hemangiosarcomas
using administered concentration, refer to Sections D.I.1.3 and D.I.3.2, respectively.
Duplicate output from MS_COMBO was omitted here.]
**** Start of combined BMD and BMDL Calculations.****
Combined Log-Likelihood -186.73874141530868
Combined Log-likelihood Constant 169.44438524661712
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 3.32952
BMDL = 2.4128
D-20
-------
Table D-4. Model predictions for male rat mononuclear cell leukemia
(MCL) (JISA, 1993), using tetrachloroethylene AUC in blood and
administered tetrachloroethylene concentration as dose metrics3 and
multistage model
Model
Goodness of fit
/7-valueb
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Conclusions
Tetrachloroethylene AUC in blood (mg-hr/L-day)
One-, two-,
three-stage
0.52
1.0, mid-dose
254.9
46.1
29.7
Fits for all three models were the
same; only the first order term was
>0.
Administered tetrachloroethylene concentration (ppm)
One-, two-,
three-stage
0.52
1.0, mid-dose
254.9
20.5
13.2
Fits for all three models were the
same; only the first order term was
>0.
a Incidence data and human equivalent continuous exposures provided in Table 5-15 and in the output below.
b Values <0.05 for a preferred model, or <0.10 when considering suite of models, fail to meet conventional
goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered. Best-fit model is
highlighted in bold; output for best-fit models provided in following pages.
D-21
-------
D.1.4. Modeling Output for Male Rats, MCL (JISA, 1993)
D.l.4.1. With tetrachloroethylene AUC in blood as dose metric
T3
s
I
c
o
Multistage Cancer Model with 0.95 Confidence Level
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Linear extrapolation
BMDL
BMD
50
100
150
200
250
dose
Figure D-9. One-stage model fit to MCL incidence in male rats (JISA, 1993),
with BMD and BMDL at 10% extra risk.
The form of the probability function is:
P[response] = background + (1-background)* [1-EXP(
-betal^dose^l)]
The parameter betas are restricted to be positive
Total number of observations = 4
Total number of records with missing value
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
D-22
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.63
Beta(l) -0.63 1
Parameter Estimates
95.0% Wald Confidence Interval
Std. Err. Lower Conf. Limit Upper Conf. Limit
Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
Dose
Est. Prob.
d.f. = 2
P-value = 0.51E
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
BMDU =
0.1
Extra risk
0. 95
46.0834
29.6814
90.5076
Taken together, (29.6814, 90.5076) is a 90
interval for the BMD
D-23
-------
D.l.4.2. With administered tetrachloroethylene concentration (ppm) as dose metric
Multistage Cancer Model with 0.95 Confidence Level
o
I
=5
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Linearextrapolation
BMDL
BMD
20
40
60
80
100
dose
Figure D-10. One-stage model fit to MCL incidence in male rats (JISA,
1993), with BMD and BMDL at 10% extra risk.
The form of the probability function is:
P[response] = background + (1-background ) * [ 1-EXP ( -betal*dose^l-beta2*dose^2-beta3*doŁ
The parameter betas are restricted to be positive
Dependent variable = mcl
Independent variable = hecdose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 4
Total number of specified parameters = 0
Degree of polynomial = 3
Default Initial Parameter Values
Background = 0.263191
Beta(l) = 0.00459397
Beta(2) = 0
Beta(3) = 0
D-24
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background
Background 1
Beta(l) -0.63
Variable
Background
Beta(l)
Beta(2)
Beta(3)
Std. Err.
Log(likelihood) # Param's Deviance Test d.f.
-124.787 4
-125.445 2 1.3173 2
-131.791 1 14.0088 3
Est. Prob.
11
14
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level =
BMD =
BMDL =
BMDU =
Multistage Cancer Slope Factor = 0.00756592
D-25
-------
Table D-5. Model predictions for female rat MCL (JISA, 1993),a using
administered tetrachloroethylene concentration (ppm)c and multistage model
Model
Goodness of fit
/7-valueb
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Tetrachloroethylene AUC in blood (mg-hr/L-day)
One-, two-,
three-stage
0.34
-1.0, control
1.0, low-dose
249.4
136
61
Administered tetrachloroethylene concentration (ppm)
One-, two-,
three-stage
0.34
-1.0, control
1.0, low-dose
249.4
60.4
26.8
Conclusion
All three models provided identical
fits, with only the first-order
parameter >0. However, model did
not adequately estimate responses at
control and low-dose.
Multistage model not selected
(output shown for administered
concentration only).
a Incidence data and human equivalent continuous exposure estimates provided in Table 5-15.
b When there is no preferred model, values <0.10 fail to meet conventional goodness-of-fit criteria. In addition,
visual fit and residuals (within +2 units) are considered.
0 Due to the proportionality of tetrachloroethylene AUC in blood, the preferred dose metric, to administered
concentration, only administered concentration was used as the dose metric until the final model selection was
made.
D-26
-------
Table D-6. Comparison of model predictions for female rat MCL (JISA,
1993),a using administered tetrachloroethylene concentration as dose metric*
Model
Goodness of fit
/rvalue0
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Comments
Conclusion
All dose groups
Michaelis-Menten
LogLogistic
Gamma, Weibull
Probit
Logistic
LogProbit
0.55
0.34
0.34
0.33
0.33
0.26
(0.0, control)
-0.5, mid-dose
-1.0, control
1.0, low-dose
-1.0, control
1.0, low-dose
-1.1, control
1.0, low-dose
-1.1, control
1.0, low-dose
-1.0, control
1.0, low-dose
249.6
249.4
249.4
249.5
249.5
250.0
5.3
56.2
60.4
67.6
68.4
88.0
NA
NA
NA
35.5
36.5
NA
Best visual fit (refer to
output)
Slope parameter
unrestricted0
Power parameters
unrestricted0
—
—
—
Poor fit to
control and
low-dose
responses.
Highest dose group dropped
Weibull
LogLogistic
Multistage (one-
degree)
Probit
Logistic
LogProbit
Gamma,
Michaelis-Menten
0.83
0.17
0.17
0.16
0.16
0.09
(0.0, control)
0.2, low-dose
-0.7, control
1.1, low-dose
-0.8, control
1.1, low-dose
-0.8, control
1.1, low-dose
-0.9, control
1.1, low-dose
181.0
182.7
182.7
182.8
182.8
NA
26.5
28.3
31.5
31.9
NA
NA
10.3
13.7
14.1
Power parameter
unrestricted"1; step-function
Slope parameter
unrestricted4
No statistical improvement
with higher order models
—
—
Implausible
fit
Poor fit to
control and
low-dose
responses
Inadequate
Insufficient degrees of freedom
Highest two dose groups dropped
Multistage
Other models
Fit statistics not relevant
4.9
2.3 | Adequate fit to control data
Insufficient degrees of freedom
aIncidence data and human equivalent continuous exposures provided in Table 5-15.
b Due to the proportionality of tetrachloroethylene AUC in blood, the preferred dose metric, to administered concentration, only
administered concentration was used as the dose metric until the final model selection was made.
0 When there is no preferred model, values <0.10 fail to meet conventional goodness-of-fit criteria. In addition, visual fit and
residuals (within +2 units) are considered. Best-fit model is highlighted in bold; output for best-fit models provided in
following pages.
d Slope or power parameters were initially limited to be >1 to avoid infinite slopes at zero dose. However, fits with restricted
parameters may not provide BMDLs with correct statistical coverage. Fits to these data with unrestricted power or slope
parameters did not provide BMDLs, effectively 0.
D-27
-------
D.1.5. Modeling Output for Female Rats, MCL (JISA, 1993), with administered
tetrachloroethylene concentration as dose metric
D.l.5.1. Multistage model fit
Multistage Cancer Model with 0.95 Confidence Level
T3
.ffi
o
I
C
O
••§
2
0.5
0.4
0.3
0.2
0.1
Multistage Cancer
Li near extrapolation
BMDLj
BMD
20
40
60
80
100
dose
Figure D-ll. One-stage multistage model fit to MCL incidence in female rats
(JISA, 1993), with BMD and BMDL at 10% extra risk.
The form of the
P[response] = background + ( 1-backgr
-betal*dose^l) ]
probability function is:
ckground)*[1-EXP(
L*dose^l)]
The parameter betas are res
Dependent variable = Response
Independent variable = Dose
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
D-28
-------
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.66
Beta(1) -0.66 1
Variable Estimate Std. Err.
Background
Beta(l)
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -121.619 4
Fitted model -122.71 2 2.18339 2
Reduced model -123.82 1 4.40312 3
Scaled
Dose Est. Prob. Expected Observed Size Residual
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 60.4331
BMDL = 26.8451
BMDU did not converge for BMR = 0.100000
BMDU calculation failed
BMDU = Inf
D-29
-------
D.l.5.2. Michaelis-Menten model fit
Dichotomous-Hill Model
Dichotomous-Hill
100 120
Figure D-12. Michaelis-Menten model (dichotomous Hill model with
exponent fixed at 1) fit to MCL incidence in female rats (JISA, 1993), with
BMD and BMDL at 10% extra risk.
The form of the probability function is:
P[response] = v*g +(v-v*g)/[1+EXP(-intercept-slope*Log(dose))]
where: 0 <= g < 1, 0
-------
intercept
g intercept
-0.53 -0.76
1 0.32
0.32 1
Parameter Estimates
intercept
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.231457 0.485107
0.193618 0.923338
-9.82655 8 . 76345
Model
Full model
Fitted model
Reduced model
Test d.f.
P-value
Dose
Est. Prob.
Scaled
Residual
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 1.74056
D-31
-------
Table D-7. Model predictions for combined male and female rat MCL (JISA,
1993),a using administered tetrachloroethylene concentrations as dose metric
Model
Goodness of fit
p-value
b
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Comments
Conclusions
Administered Tetrachloroethylene Concentration (ppm)
Michaelis-
Menten
LogLogistic
Multistage,
one-stage,
Gamma
Weibull
Probit
Logistic
LogProbit
0.68
0.35
0.28
0.96
0.98
0.20
0.19
0.07
-0.3, mid-
dose
-0.9, control
-1.1, control
-1.3, control
-1.3, control
-1.6, control
503.6
503.4
504.0
503.4
503.4
504.7
504.8
506.8
7.7
5.1
32.0
4.5
4.8
40.7
41.6
55.2
1.4
0.003
20.9
0.001
0.002
29.7
30.5
38.6
Best-fit to combined data.
Unrestricted slope parameter0
No statistical improvement
from higher order stages
Unrestricted power parameter
Unrestricted power parameter
—
—
Fit at control
response not
useful.
Inadequate overall fit
Tetrachloroethylene AUC in blood (mg-hr/L-day)d
Michaelis-
Menten6
0.68
-0.3, mid-
dose
503.6
17.4
3.0
Best-fit above repeated with preferred dose
metric.
a Incidence data and human equivalent continuous exposures provided in Table 5-15 and in output below.
b Values <0.10 fail to meet conventional goodness-of-fit criteria. In addition, visual fit and residuals (within +2
units) are considered. Best-fit model is highlighted in bold; output for best-fit models provided in following pages.
0 Lower limit for slope or power parameters is >1 to avoid infinite slopes at zero dose. However, fits with restricted
parameters may not provide BMDLs with adequate statistical coverage. Fits with unrestricted power or slope
parameters did not provide a usable BMDL.
Due to the proportionality of tetrachloroethylene AUC in blood, the preferred dose metric, to administered
concentration, administered concentration was used as the dose metric until the final model selection was made.
e Dichotomous-Hill model with slope fixed at 1.
D-32
-------
Analyses to evaluate combining the male and female rat MCL data:
1. Following the strategy of Stiteler et al. (1993) the data sets were evaluated for statistical
compatibility by applying the generalized likelihood ratio method to the results of fitting a
common dose-response model (Michaelis-Menten) to the separate and combined data sets
(using administered concentration as dose metric):
Maximum log-likelihoods (LLs), and sum
Overall LL from combined data set
Female Rat
MCLs
-121.795
(3df)
Male Rat
MCLs
-124.841
(3df)
X2 = 2 x absolute difference in LLs.
Females + Males
-246.636
(6df)
-248.79 (3 df)
X2 = 2.\5(p = 0.54)
2. Logistic regression was used to test whether the datasets differed significantly between males
and females. The advantage of this approach is that it does not require the assumption of a
specific functional form to represent the dose response relationship. Dose and sex were
treated as categorical variables using PROC LOGISTIC in SAS:
Effect
dose
sex
DF
3
1
Wald
Chi-Square
14.6302
1.6634
Pr > ChiSq
0.0022
0.1971
The/7-value of 0.197 for sex indicates no significant relationship of sex in the pattern of
responses.
D-33
-------
D.1.6. Modeling Output for Male and Female Rats, MCL (JISA, 1993)
D.l.6.1. With administered tetrachloroethylene concentration (ppm) as dose metric
Dichotomous-Hill Model with 0.95 Confidence Level
.Q
"o
2
0.6
0.5
0.4
0.3
0.2
0.1
100
120
dose
Figure D-13. Michaelis-Menten model (dichotomous-Hill model with
exponent fixed at 1) fit to MCL incidence in male and female rats (JISA,
1993), with BMD and BMDL at 10% extra risk.
Dichotomous Hill Model. (Version: 1.0; Date: 09/24/2006)
The form of the probability function is:
P[response] = v*g +(v-v*g)/[1+EXP(-intercept-slope*Log(dose))]
where: 0 <= g < 1, 0
-------
g =
intercept =
slope =
Asymptotic Correlation Matrix of Parameter Estimates
intercept
Parameter Estimates
Variable
g
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.310304 0.683283
0.230288 0.626157
-5.34015 -0.668408
Model
Full model
Fitted model
Reduced model
Log(likelihood) Deviance Test d.f.
-248.707
-248.79 0.167701 1
-256.414 15.4153 3
Est. Prob.
d.f. =1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 7.7341
BMDL = 1.35558
D-35
-------
D.l.6.2. With tetrachloroethylene AUC in blood as dose metric
0.6
0.5
0.4
0.3
0.2
0.1
Dichotomous-Hill Model with 0.95 Confidence Level
100 150
dose
200
250
Figure D-14. Michaelis-Menten model (dichotomous-Hill model with
exponent fixed at 1) fit to MCL incidence in male and female rats (JISA,
1993), with BMD and BMDL at 10% extra risk.
The form of the probability function is:
P[response] = v*g +(v-v*g)/[1+EXP(-intercept-slope^Log(dose))]
where: 0 <= g < 1, 0
-------
Asymptotic Correlation Matrix of Parameter Estimates
intercept
Parameter Estimates
intercept
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.31055 0.682655
0.230539 0.626245
-6.14833 -1.47795
Model
Full model
Fitted model
Reduced model
Log(likelihood) Deviance Test d.f.
-248.707
0.168004 1
15.4153 3
Dose
Prob.
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 17.382
BMDL = 3.04513
D-37
-------
D.2. Model Selection Details for Male Rat Tumors (NTP, 1986)
Table D-8. Model predictions for male rat tumors (NTP, 1986),a using
administered tetrachloroethylene concentration as dose metric and
multistage model
Model
Goodness of fit
p-valueb
Largest
standardized
residual(s)
AIC
BMD10C
(ppm)
BMDIV
(ppm)
Conclusions
Kidney tumors
One-, two-stage
0.75
0.3, low-dose
64.1
110
50
No statistical improvement
from adding higher-order
parameter; one-stage model
selected
Brain gliomas
One-stage
Two-stage
0.11
0.18
-1.3, low-dose
-1.3, low-dose
45.7
44.6
180
138
73
45
No statistical improvement from
adding higher-order parameter;
one-stage model selected.
Testicular interstitial cell tumors
One-, two-stage
0.40
0.7, low-dose
155.8
13.0
6.1
No statistical improvement from
adding higher-order parameter;
one-stage model selected
MCL
One-, two-stage
0.18
1.1, low-dose
184.8
12.1
6.5
Only a one-stage model resulted.
a Incidence data and human equivalent continuous exposures provided in Table 5-15 and in output below.
b Values O.05 fail to meet conventional goodness-of-fit criteria.
0 The highest response in these data sets was less than 10% extra risk; however, because the best-fit models were
linear, use of BMD10 and BMDL10 was equivalent to using a BMR within the data range.
D-38
-------
Multistage Cancer Model wfth 0.95 Confidence Level
Multistage Cancer Model wfth 0.95Conficence Level
Multistage Cancer
Linear extrapolation
a. One-stage model fit to kidney tumors.
Refer to section D.2. for model output.
b. One-stage model fit to brain gliomas.
See section D.2. for model output.
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer Model with 0.95 Confidence Level
MultistageCance
Linear extrapolatioi
30 40
dose
50 60 70
d. One-stage model fit to testicular
interstitial cell tumors. See section D.2.
for model output.
d. One-stage model fit to MCLs. See section D.2.
for model output.
Figure D-15. Multistage model fits to tumor incidences at multiple sites in
male rats—kidney tumors, brain gliomas, testicular interstitial cell tumors,
and MCL (NTP, 1986).
Graphs show BMD and BMDL at 10% extra risk.
D-39
-------
D.2.1. Modeling Output for Male Rats (NTP, 1986): MCLs, Brain Gliomas, Kidney
Tumors, Testicular Interstitial Cell Tumors and Combined Tumors at 10% Extra Risk,
Using Administered Concentration and Multistage Modeling (Discussed in Section 5.3.4.1)
D.2.2. Kidney Tumors
MS_COMBO. (Version: 1.5 Beta; Date: 01/25/2011)
Input Data File: C:\Usepa\BMDS21\New.(d)
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*dose^l)]
The parameter betas are restricted to be positive
Dependent variable = kidney
Independent variable = hec
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.78
Beta(l) -0.78 1
Parameter Estimates
Std. Err.
Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
D-40
-------
AIC: 64.05
Log-likelihood Constant
d.f. = 1
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 109.841
BMDL = 49.5786
D.2.3. Brain Gliomas
MS_COMBO. (Version: 1.5 Beta; Date: 01/25/2011)
Input Data File: C:\Usepa\BMDS21\New.(d)
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*dose^l)
The parameter betas are restricted to be positive
Dependent variable = gliomas
Independent variable = hec
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
D-41
-------
Background
Beta(1)
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(1)
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -18.8404 3
Fitted model -20.8724 2 4.06391 1 0.04381
Reduced model -21.8534 1 6.02604 2 0.04914
Log-likelihood Constant
d.f. = 1 P-value = 0.1119
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 179. 643
BMDL = 72.6378
D.2.4. Testicular Tumors
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(-betal*dose^l)
D-42
-------
Dependent variable = testtumor
Independent variable = hec
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(1)
Background 1 -0.71
Beta(1) -0.71 1
Variable Estimate Std. Err.
Background
Beta(1)
Indicates that this value is not calculated.
Log-likelihood Constant
Scaled
Est. Prob. Expected Observed Size Residual
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
D-43
-------
Confidence level =
BMD =
BMDL =
Taken together, (6.0591 , 8.43867e+014) is a 90 % two-sided confidence
interval for the BMD
D.2.5. MCLs
The form of the probability function is:
P[response] = background + (1-background ) * [ 1-EXP ( -betal*dose~l) ;
The parameter betas are restricted to be positive
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.72
Beta(l) -0.72 1
Variable Estimate Std. Err.
Background
Beta(l)
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
D-44
-------
Full model -88.7862 3
Fitted model -89.6897 2
Reduced model -91.7227 1
Scaled
Residual
Chi^2 = 1.76 d.f. = 1 P-value = 0.1843
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
Taken together, (6.54184, 1.67846e+007) is a 90
interval for the BMD
D.2.6. Combined BMD and BMDL for Male Rat Tumors
**** Start of combined BMD and BMDL Calculations.****
Combined Log-Likelihood -216.50768590007982
Combined Log-likelihood Constant 194.39505507650239
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 5.73369
BMDL = 3.48718
D.2.7. Modeling Output for Male Rats (NTP, 1986): MCLs, Brain Gliomas, Kidney
Tumors, Testicular Interstitial Cell Tumors and Combined Tumors at 10% Extra Risk,
Using Administered Concentration and Multistage Modeling (Discussed in Section 5.3.4.1)
D-45
-------
Note: For numerical stability, doses were modeled in units of g-hr/L-d rather than mg-hr/L-d.
Refer to Figure 5-15 for corresponding graphs.
D.2.8. Brain Gliomas
MS_COMBO. (Version: 1.5 Beta; Date: 01/25/2011)
The form of the probability function is:
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.7
Beta(l) -0.7 1
Std. Err.
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -18.8404 3
Fitted model -20.8459 2 4.01101 1 0.0452
Reduced model -21.8534 1 6.02604 2 0.04914
D-46
-------
Scaled
Residual
d.f. = 1
P-value = 0.1148
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.409021
BMDL = 0.166066
BMDU = Inf
D.2.9. Kidney Tumors
MS_COMBO. (Version: 1.5 Beta; Date: 01/25/2011)
The form of the probability function is:
Dependent variable = Effect
Independent variable = Dose
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.78
Beta(l) -0.78 1
D-47
-------
Parameter Estimates
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -29.9768 3
Fitted model -30.0318 2 0.110125 1 0.74
Reduced model -31.01 1 2.06651 2 0.3558
Chi~2 = 0.11 d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.253015
BMDL = 0.113761
BMDU did not converge for BMR = 0.100000
BMDU calculation failed
BMDU = 2.05051e+023
D.2.10. MCLs
The form of the probability function is:
Dependent variable = Effect
D-48
Scaled
Residual
-------
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(l)
Background 1 -0.72
Beta(l) -0.72 1
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background
Beta(1)
Indicates that this value is not calculated.
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -88.7862 3
Fitted model -89.7301 2 1.88778 1 0.1695
Reduced model -91.7227 1 5.87302 2 0.05305
Log-likelihood Constant
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
D-49
-------
BMD =
BMDL =
BMDU =
D.2.11. Testicular Tumors
The form of the probability function is:
Total number of observations = 3
Total number of records with missing values = 0
Total number of parameters in model = 2
Total number of specified parameters = 0
Degree of polynomial = 1
Default Initial Parameter Values
Background = 0.730195
Beta(1) = 3.0992
Asymptotic Correlation Matrix of Parameter Estimates
Background Beta(1)
Background 1 -0.71
Beta(1) -0.71 1
Parameter Estimates
Indicates that this value is not calculated.
D-50
-------
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -75.5554 3
Fitted model -75.9393 2 0.767742 1 0.3809
Reduced model -77.0228 1 2.93491 2 0.2305
Scaled
Residual
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0302396
BMDL = 0.0139853
BMDU did not converge for BMR = 0.100000
BMDU calculation failed
BMDU = Inf
D.2.12. Combined BMD and BMDL for Male Rat Tumors
**** Start of combined BMD and BMDL Calculations.****
Combined Log-Likelihood -216.54708443096888
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0133069
BMDL = 0.00805821
D-51
-------
D.3. Comparison of PODs Resulting from the Use of Models Alternative to the Multistage
Model, for Tumor Sites in the JISA (1993) Bioassay
Note: Refer to Section D. 1 for alternative modeling for female rat MCLs and combined
male and female rat MCLs.
Table D-9. Comparison of model predictions for hepatocellular tumors in
mice (JISA, 1993), using administered tetrachloroethylene concentration
(ppm) as the dose metric,3 across a range of dichotomous models
Model
Goodness of fitb
^j-value
Largest
standardized
residual(s)
AIC
BMD10
BMDL10
Comments0
Male mice (input data in Table 5-13)
Gamma
Weibull
Michaelis-Mentend
LogLogistic
LogProbit
Multistage
Logistic
Probit
0.14
0.15
0.13
0.14
0.14
0.27
0.35
0.36
-1.0, control
1.1, low-dose
-1.0, control
1.1, low-dose
-1.0, control
1.1, low-dose
-1.0, control
1.1, low-dose
-1.1, control
1.0, low-dose
1.2, low-dose
-1.0, mid-dose
-0.7, control
1.1, low-dose
-0.7, control
1.1, low-dose
241.1
241.0
241.0
241.1
241.1
239.5
238.9
238.9
10.2
9.5
2.5
10.6
11.0
3.9
6.0
6.1
0.4
0.7
1.3
1.5
2.1
2.7
4.7
4.8
Power parameter unrestricted
Power parameter unrestricted
—
Slope parameter unrestricted
—
Lowest residual at control
(-0.5)
—
—
Female mice (input data in Table 5-13)
Multistage
Gamma
Weibull
LogLogistic
Michaelis-Menten
LogProbit
Probit
Logistic
0.36
0.96
0.93
0.98
0.95
0.91
0.84
-1.1, mid-dose
-0.04, low-dose
-0.07, control
0.02, control
0.01, control
0.04, low-dose
0.4, mid-dose
0.5, mid-dose
152.9
152.8
152.8
152.8
154.8
152.8
151.0
151.2
5.0
9.5
9.5
9.5
9.5
9.5
11.8
13.1
3.8
4.3
4.3
4.7
4.7
5.0
9.7
10.6
—
—
—
Slope parameter unrestricted
—
—
—
—
a Only one dose metric used due to near proportionality of relevant dose metrics.
b Goodness-of-fit ^-values O.05 for a preferred model, or <0.10 when considering suite of models, fail to meet
conventional goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered.
0 Lower limit >lfor slope or power parameters avoids biologically implausible infinite slopes at zero dose. However,
fits with restricted parameters may not provide BMDLs with nominal (95%) statistical coverage.
d Dichotomous-Hill model with slope fixed at 1
D-52
-------
Table D-10. Comparison of model predictions for male mice, hemangiomas,
or hemangiosarcomas (JISA, 1993),a using administered tetrachloroethylene
concentration as dose metric,b across a range of dichotomous models
Model
Michaelis-
Mentene
LogLogistic
LogProbit
Weibull
Gamma
Multistage
Probit
Logistic
Goodness of fitc
/7-value
0.21
0.19
0.21
0.18
0.18
0.38
0.32
0.31
Largest
standardized
residual(s)
-1.0, low -dose
-1.0, low -dose
-1.0, low -dose
-1.0, low -dose
-1.0, low -dose
-1.0, low -dose
0.9, mid-dose
-1.1, low-dose
1.0, mid-dose
-1.1, control
1.0, low-dose
AIC
143.7
141.9
143.7
143.9
143.9
142.0
142.4
142.4
BMD10
16.1
20.7
19.2
21.3
21.7
24.4
30.6
31.6
BMDL10
4.1
5.4
5.4
5.5
5.6
13.3
20.9
22.1
Comments'1
Better fit than restricting or
unrestricting slope parameter
Slope parameter unrestricted
Power parameter unrestricted
Power parameter unrestricted
Lowest residual at control (0.4)
—
—
a Incidence data and human equivalent continuous exposure estimates provided in Table 5-13.
b Only one dose metric used, due to proportionality of relevant dose metrics.
0 Goodness-of-fit ^-values <0.05 for a preferred model, or <0.10 when considering suite of models, fail to meet
conventional goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered.
dLower limit >lfor slope or power parameters avoids biologically implausible infinite slopes at zero dose. However,
fits with restricted parameters may not provide BMDLs with nominal (95%) statistical coverage.
e Dichotomous-Hill model with slope fixed at 1.
D-53
-------
Table D-ll. Comparison of model predictions for MCL in male rats (JISA,
1993),a using administered tetrachloroethylene concentration as dose metric,b
across a range of dichotomous models
Model
Gamma
Weibull
LogLogistic
LogProbit
Michaelis-
Mentend
Multistage
Probit
Logistic
Goodness of fit
/7-valuec
0.51
0.54
0.59
0.63
0.74
0.51
0.34
0.32
Largest
standardized
residual(s)
-0.4, low-dose
0.4, mid-dose
-0.4, low-dose
0.4, mid-dose
0.4, mid-dose
0.4, mid-dose
(0.1, control)
-0.2, low-dose
1.0, mid-dose
-0.8, control
1.2, mid-dose
-0.8, control
1.2, mid-dose
AIC
256.0
256.0
255.9
255.8
255.7
254.9
255.7
255.8
BMD10
6.9
7.1
7.8
8.5
8.6
20.5
29.3
30.0
BMDL10
0.062
0.11
0.18
0.28
2.2
13.2
21.6
22.1
Comments
Power parameter unrestricted;
BMDL quite low
Power parameter unrestricted
Slope parameter unrestricted
—
—
—
—
—
a Incidence data and human equivalent continuous exposure estimates provided in Table 5-15.
b Only one dose metric used for comparing model predictions. Ouputs for the Michaelis-Menten model using either
administered concentration or tetrachloroethylene AUC in blood follow this table.
0 Goodness-of-fit ^-values <0.05 for a preferred model, or <0.10 when considering suite of models, fail to meet
conventional goodness-of-fit criteria. In addition, visual fit and residuals (within +2 units) are considered.
d Dichotomous-Hill model with slope fixed at 1.
D-54
-------
Dichotomous-Hill Model with 0.95 Confidence Level
0.7
0.6
0.5
< 0.4
0.3
0.2
0.1
Dichotomous-Hill
BMDL 3MD ,
20
40 60
dose
80
100
120
Dichotomous-Hill Model with 0.95 Confidence Level
8=
<
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Dichotomous-Hill
BMDL
BMD
50
100 150
dose
200
250
Figure D-16. Michaelis-Menten model (dichotomous-Hill model with
exponent fixed at 1) fit to MCL incidence in male rats (JISA, 1993), with
BMD and BMDL at 10% extra risk; with administered concentration as dose
metric (top) or tetrachloroethylene AUC in blood (bottom)
D-55
-------
Output for administered concentration:
Dichotomous Hill Model. (Version: 1.0; Date: 09/24/2006)
The form of the probability function is:
P[response] = v*g +(v-v*g)/[1+EXP(-intercept-slope*Log(dose))]
where: 0 <= g < 1, 0 < v <= 1
v is the maximum probability of response predicted by the model,
and v*g is the background estimate of that probability.
Total number of observations = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User Inputs Initial Parameter Values
v = 0.6
g = 0.38
intercept = -4
slope = -9999 Specified
intercept
-0. 9
0.47
1
Parameter Estimates
intercept
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.265395 1.07245
0.104107 0.536774
-6.10815 -1.3349
Model
Full model
Fitted model
Reduced model
Log(likelihood) Deviance Test d.f.
-124.787
-124.841 0.108028 1
-131.791 14.0088 3
D-56
-------
11
14
d.f. = 1
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 8. 63516
BMDL = 2.20116
Output for tetrachloroethylene AUC in blood:
Dichotomous Hill Model. (Version: 1.2; Date: 12/11/2009)
The form of the probability function is:
P[response] = v*g +(v-v*g)/[1+EXP(-intercept-slope*Log(dose))]
where: 0 <= g < 1, 0
-------
intercept
The model parameter(s) -slope
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
v g intercept
-0.7 -0.9
1 0.47
0.47 1
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.266453 1.07028
0.104578 0.536849
-6.91411 -2.14596
Model
Full model
Fitted model
Reduced model
Analysis of Deviance Table
Log(likelihood) # Param's Deviance Test d.f.
-124.787 4
-124.841 3 0.107785 1
-131.791 1 14.0088 3
P-value
Scaled
Residual
Chi~2 = 0.11
d.f. = 1
P-value = 0.7431
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 19.411
BMDL = 4.94596
D-58
-------
D.4. References
JISA (Japan Industrial Safety Association). (1993). Carcinogenicity study of tetrachloroethylene
by inhalation in rats and mice. Hadano, Japan.
NTP (National Toxicology Program). (1986). Toxicology and carcinogenesis studies of
tetrachloroethylene (perchloroethylene) (CAS no. 127-18-4) in F344/N rats and B6C3F1
mice (inhalation studies). (NTP TR 311). Research Triangle Park, NC: U.S. Department
of Health and Human Services, National Toxicology Program.
http://ntp.mehs.nih.gov/ntp/htdocs/LT_rpts/tr311 .pdf
Stiteler. WM: Knauf LA: Hertzberg. RC: Schoeny. RS. (1993). A statistical test of compatibility
of data sets to a common dose-response model. Regul Toxicol Pharmacol 18: 392-402.
http://dx.doi.org/10.1006/rtph.1993.1065.
D-59
------- |